Chromatographic Method

Development
Jenny Stanford Series on Pharmaceutical Analytics
Series Editors
Gregory K. Webster, J. Derek Jackson, and Robert G. Bell

Titles in the Series

Vol. 1 Vol. 3
Supercritical Fluid Chromatography: Chromatographic Method
Advances and Applications in Development
Pharmaceutical Analysis Gregory K. Webster and Laila Kott, eds.
Gregory K. Webster, ed. 2020
2014 978-981-4800-53-2 (Hardcover)
978-981-4463-00-3 (Hardcover) 978-0-429-20172-1 (eBook)
978-981-4463-01-0 (eBook)

Vol. 2
Poorly Soluble Drugs: Dissolution
and Drug Release
Gregory K. Webster, J. Derek Jackson,
and Robert G. Bell, eds.
2017
978-981-4745-45-1 (Hardcover)
978-981-4745-46-8 (eBook)
Jenny Stanford Series on Pharmaceutical Analytics — Volume 3

Chromatographic Method
Development

edited by
Gregory K. Webster
Laila Kott
Published by
Jenny Stanford Publishing Pte. Ltd.
Level 34, Centennial Tower
3 Temasek Avenue
Singapore 039190

Email: [email protected]
Web: www.jennystanford.com

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

Chromatographic Method Development

Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.

All rights reserved. This book, or parts thereof, may not be reproduced in any
form or by any means, electronic or mechanical, including photocopying,
recording or any information storage and retrieval system now known or
to be invented, without written permission from the publisher.

For photocopying of material in this volume, please pay a copying fee

through the Copyright Clearance Center, Inc., 222 Rosewood Drive,
Danvers, MA 01923, USA. In this case permission to photocopy is not
required from the publisher.

ISBN 978-981-4800-53-2 (Hardcover)

ISBN 978-0-429-20172-1 (eBook)
 Contents

Preface xxi

1. A Lesson Learned from UHPLC 1

Gregory K. Webster and Laila Kott

1.1 Introduction 1
1.2 Technology and Chromatography 2
1.2.1 New Goal of Method Development: Speed 3
1.3 The Chromatography Market 4
1.4 Chromatographic Method Development 4
1.4.1 Software for Chromatographic Method
Development 5
1.4.2 Flash Chromatography Method
Development 5
1.4.3 Orthogonality in Column Chromatography 5
1.4.4 Unified Approach to Reversed-Phase and
Normal-Phase Chromatography 6
1.4.5 Polysaccharide-Derived Chiral Stationary
Phases for the Separation of Enantiomers 6
1.4.6 Chiral Chromatography: Method
Development 6
1.4.7 Aqueous Normal-Phase Chromatography
Using Type-C Silica Columns 7
1.4.8 Ion Chromatography Columns 7
1.4.9 Ion Chromatography: Method Development 7
1.4.10 Fundamentals, Properties, and Applications
of Stationary Phases of Gas Chromatography
Method Development 8
1.4.11 Gas Chromatography Method Development 8
vi Contents

1.4.12 Method Transfer from HPLC to UHPLC 8

1.4.13 Chromatographic Method Validation 9
1.4.14 Method Development Strategies for Ion
Exchange Chromatography Using pH
Gradient Separation for mAbs, Antibody
Drug Conjugates, and Other Complex
Bio-Pharmaceuticals 9
1.4.15 Size Exclusion Chromatography Method
Development for Therapeutic Proteins 9
1.4.16 Ionic Liquids and Counter-Current
Chromatography: Is This the Future of
Process Purifications? 10

2. Software for Chromatographic Methods Development 13

Aoi Ariyasu and Eiki Maeda
2.1 Introduction 13
2.2 General Features of Software 15
2.2.1 Typical Procedure for the Systematic
Development of Analytical Methods and
the Effect of the Software Employed 15
2.2.2 General Software Features and
Applicability to the Various Stages of
Analytical Method Development 18
2.3 Application of Software to Analytical Method
Development 20
2.3.1 Structure-Based Retention Prediction 21
2.3.2 Chromatogram Simulation 22
2.3.3 Method Scouting/Screening 26
2.3.3.1 Online method scouting 26
2.3.3.2 Offline method scouting 29
2.3.4 Method Optimization 30
2.3.4.1 Method optimization using
chromatographic software 34
2.3.4.2 Method optimization using DOE
software 36
2.3.5 Robustness Evaluation 42
 Contents vii

2.4 Advantages and Disadvantages of Using

Available Method Development Software
Programs 44
2.5 Future Perspective 46

3. Flash Chromatography Method Development 53

J. Robert Bickler
3.1 Introduction 53
3.2 Goal Setting 54
3.3 Starting Out 54
3.3.1 Thin-Layer Chromatography 55
3.3.2 Using TLC to Scout Suitable Elution
Solvents 56
3.4 Determining Cartridge Load Capacity 58
3.5 Determining Elution Style 60
3.5.1 Isocratic Elution 60
3.5.2 Linear Gradient Elution 60
3.5.2.1 Linear gradients from 0% to 100% 62
3.5.3 Step Gradient Elution 62
3.6 Choosing between Normal- and Reversed Phase 63
3.7 Conclusion 64

4. Orthogonality in Chromatographic Methods 65

Zachary S. Breitbach and Gregory K. Webster

4.1 Introduction 65
4.2 Traditional View of Orthogonal LC Methods 66
4.3 Column Orthogonality in Reversed-Phase
Chromatography 66
4.3.1 Introduction 66
4.3.2 Hydrophobic Subtraction Theory 68
4.3.2.1 HST for column comparison 69
4.3.2.2 HST values for commercial
reversed-phase columns 69
4.3.2.3 HST for orthogonal reversed-phase
separations 70
viii Contents

4.4 Orthogonality in Traditional Normal-Phase LC 72

4.5 Orthogonality in Supercritical Fluid
Chromatography 74
4.6 Orthogonality in HILIC 77
4.7 Orthogonality in Chiral Chromatography 79
4.8 Orthogonality in Multi-Dimensional
Chromatography 79
4.9 Conclusion 82

5. Unified Approach to Reversed-Phase and Normal-Phase

Chromatographic Method Development 85
Gregory K. Webster
5.1 Introduction 85
5.2 Unified Approach to Method Development 86
5.2.1 “Like Dissolves Like” 86
5.2.2 Information and Samples Required 87
5.3 Reversed-Phase Chromatography 89
5.3.1 Control of Analyte Retention 89
5.3.2 Hydrophobic Subtraction Theory/Columns 90
5.3.3 Solvent Strength 90
5.3.4 Screening 91
5.3.4.1 System considerations: UHPLC
vs. Traditional 91
5.3.4.2 Traditional approach 93
5.3.4.3 Automated approach 95
5.3.5 Ion Pair Reagents 100
5.3.6 Reversed-Phase Screen Conditions 101
5.4 Normal-Phase Chromatography 101
5.4.1 Control of Analyte Retention 102
5.4.2 Hydrophobic Subtraction Theory 102
5.4.3 Solvent Strength 103
5.4.4 Ionic Samples 103
5.4.5 Screening 104
5.4.5.1 Mobile phases 104
5.4.5.2 Columns 104
 Contents ix

5.4.5.3 Automated Normal-Phase

Screening Systems 106
5.5 Hydrophilic Interaction Liquid Chromatography 107
5.5.1 Control of Analyte Retention 108
5.5.2 Hydrophobic Subtraction Theory 108
5.5.3 Solvent Strength 108
5.5.4 Ionic Samples 109
5.5.5 Screening 109
5.5.1.1 Mobile phases 109
5.5.1.2 Columns 110
5.5.1.3 Automated HILIC phase
screening systems 110
5.5.1.4 HILIC screen conditions 110
5.6 Conclusion 111

6. Polysaccharide-Derived Chiral Stationary Phases for

the Separation of Enantiomers 115
Tong Zhang and Pilar Franco

6.1 Introduction 115

6.2 Column and Mobile Phase Selection 118
6.2.1 Method Screening with Organic Mobile
Phases in HPLC 119
6.2.1.1 Immobilized CSPs under organic
mobile phase conditions 120
6.2.1.2 Coated CSPs under organic
mobile phase conditions 124
6.2.2 Method Screening with Aqueous or
Water Compatible Mobile Phases in HPLC 125
6.2.2.1 Immobilized CSPs in aqueous
or water compatible solvent
mixtures 125
6.2.2.2 Coated CSPs in aqueous or water
compatible solvent mixtures 128
6.2.3 Mobile Phases in SFC 128
6.2.3.1 Immobilized CSPs in SFC 128
6.2.3.2 Coated CSPs in SFC 130
 Contents

6.3 The Role of Additives 131

6.4 Column Temperature 132
6.5 Effects of Flow Rate and Particle Size 136
6.6 Influence of Analyte Structure and the
Substituent Effect 140
6.7 Column Maintenance, Cleaning, and
Regeneration 144
6.8 Conclusions 148

7. Chiral Chromatography: Method Development 157

Laila Kott and Gregory K. Webster

7.1 Introduction 157

7.2 Start at the End, with the Detectors 158
7.2.1 Polarimeters 159
7.2.2 Circular Dichroism Detector 161
7.2.3 When Chromatography Does Not Work:
An Alternative Technique 162
7.3 Next Step—the Type of Chromatography 163
7.3.1 Choice of Phase 163
7.3.2 Normal-Phase Screening 164
7.3.3 Supercritical Fluid Chromatography
Screening 165
7.3.4 Reverse-Phase Screening 167
7.3.5 Screening Success 170
7.4 The Columns 170
7.4.1 Polysaccharide-Type Columns 170
7.4.2 Pirkle-Type Columns 171
7.4.3 Antibiotic Type Columns 172
7.4.4 Columns for Ultrafast Chiral Separations 172
7.4.5 Other Phases for Chiral Separations 174
7.5 Conclusions 175

8. Aqueous Normal-Phase Chromatography Using

Type C Silica Columns 179
Joseph J. Pesek and Maria T. Matyska
8.1 Introduction 180
 Contents xi

8.2 Silica Hydride 181

8.3 Applications of Silica Hydride-Based Separation
Materials 186
8.3.1 Reversed Phase 186
8.3.2 Use of Aqueous Normal Phase for Polar
Compounds 189
8.3.3 Organic Normal Phase on Silica Hydride 197
8.4 Method Development Strategies on Silica
Hydride Phases 199
8.4.1 Mobile Phase Selection 200
8.4.2 Equilibration for Reversed-Phase 200
8.4.3 Reversed-Phase Gradient 201
8.4.4 Equilibrate for Aqueous Normal Phase 201
8.4.5 Aqueous Normal-Phase Gradient 201
8.4.6 Compare Reversed-phase and Aqueous
Normal-Phase Data 201
8.5 Conclusions 202

9. Ion Chromatography Columns 209

Christopher Pohl

9.1 Introduction 209

9.2 Cation-Exchange Phases 211
9.2.1 Surface-Sulfonated Cation-Exchange
Phases 211
9.2.2 Sulfonated Latex-Based Cation-Exchange
Phases 212
9.2.3 Polymer-Encapsulated Weak Acid
Cation-Exchange Phases 216
9.2.4 Polymer-Grafted Weak Acid Cation-Exchange
Phases 217
9.2.5 Polymer-Grafted Strong and Weak Acid
Cation-Exchange Phases 219
9.2.6 Cation-Exchange Phases for Transition
Metal Separations 220
9.3 Anion-Exchange Phases 224
xii Contents

9.3.1 Quaternary Latex-Based Anion-Exchange

Phases on Nonporous Substrates 224
9.3.2 Quaternary Latex-Based Anion-Exchange
Phases on Wide-Pore Substrates 227
9.3.3 Adsorbed Coating Anion-Exchange Phases 229
9.3.4 Polymer-Grafted Strong Base
Anion-Exchange Phases 230
9.3.5 Surface-Chloromethylated High
Cross-Link Macroporous Anion-Exchange
Phases 231
9.3.6 Polyvinyl Alcohol Backbone
Anion-Exchange Phases 232
9.3.7 Hyperbranched Anion-Exchange Phases
on Wide-Pore Substrates 233

10. Ion Chromatography: Method Development 237

Laila Kott

10.1 Introduction 237

10.2 Components of the System 238
10.2.1 Columns 239
10.2.1.1 Anionic columns 240
10.2.1.2 Cationic columns 241
10.2.2 Eluants 242
10.2.2.1 Anionic eluants 243
10.2.2.2 Cationic eluants 244
10.2.3 Suppressor Theory, Operation, and
Troubleshooting 244
10.2.4 Detectors 246
10.2.4.1 Conductivity detector 246
10.2.4.2 Other detectors and their
uses 251
10.3 Method Development 252
10.3.1 Sample Preparation and Diluents 252
10.3.2 Isocratic and Gradient Elution 253
10.3.3 Suppressor Current Settings 255
 Contents xiii

10.3.4 Temperature Effects 258

10.4 General System Troubleshooting 258
10.5 Applications 259

11. Fundamentals, Properties, and Applications of Stationary

Phases for Gas Chromatography Method Development 263
Omprakash Nacham and Jared L. Anderson

11.1 Introduction 264

11.2 Column Preparation and Evaluation 266
11.2.1 Materials and Methods Involved in
Preparation of WCOT Capillary Columns 266
11.2.2 Evaluation of Gas Chromatographic
Column Performance 268
11.3 Selectivity of Gas Chromatographic Stationary
Phases 270
11.3.1 Solvation Parameter Model 271
11.4 Stationary Phase Materials 272
11.4.1 Aliphatic and Aromatic Hydrocarbon-
Based Stationary Phases 273
11.4.2 Poly(Siloxane)-Based Stationary Phases 274
11.4.2.1 Preparation of poly(siloxane)-
based stationary phases for
open tubular columns 276
11.4.2.2 Modified poly(siloxane)
stationary phases for chiral
analysis 280
11.4.2.3 Chiral separations using
α-amino acid-based
poly(siloxanes) 280
11.4.2.4 Chiral separations using
cyclodextrin-based
poly(siloxanes) 281
11.4.3 Poly(Alkylene Oxide)-Based Stationary
Phases 282
11.5 Ionic Liquids and Polymeric Ionic Liquids as
Emerging Stationary Phase Materials 283
xiv Contents

11.5.1 Influence of Physicochemical Properties

of ILs on Column Performance 284
11.5.1.1 Melting point 284
11.5.1.2 Viscosity 285
11.5.1.3 Thermal stability 286
11.5.1.4 Surface tension 287
11.5.2 Imidazolium and Phosphonium-Based IL
Stationary Phases 287
11.5.3 Polymeric Ionic Liquid-Based Stationary
Phases 293
11.6 Conclusions and Future Directions 294

12. Gas Chromatography Method Development 309

Tien Ho and John C. Vinci

12.1 Gas Chromatography in the Pharmaceutical

Setting 309
12.2 Instrument and Column Optimization 310
12.2.1 Split/Splitless Inlet 311
12.2.1.1 Split injection 312
12.2.1.2 Splitless injection 313
12.2.2 Column Dimensions 314
12.2.2.1 Column inner diameter 314
12.2.2.2 Column length 315
12.2.2.3 Film thickness 316
12.2.3 Linear Velocity 320
12.3 Optimization of Separation Temperature 321
12.3.1 Programmed Temperature Separations 322
12.3.2 Isothermal Separations 324
12.4 Fast GC: A Culmination of Instrument, Column
and Temperature Optimization 325
12.5 Conventional GC Detectors Applied in
Pharmaceutical Analyses 328
12.5.1 Thermal Conductivity Detector 328
12.5.2 Flame-Ionization Detector 329
12.5.3 Electron Capture Detector 331
 Contents xv

12.6 GC/Mass Spectrometry 333

12.6.1 Ionization by GC/MS 335
12.6.1.1 Chemical ionization 335
12.6.1.2 Electron ionization 337
12.6.2 Systematic Approaches for Chemical
Structure Elucidation by EI-GC/MS 340
12.6.2.1 Interpretation of mass spectral
via database 340
12.6.2.2 Manual interpretation of mass
spectral data 342
12.7 Considerations for Using GC in the
Pharmaceutical Setting 346
12.7.1 Analysis of Residual Solvents 346
12.7.2 Analysis of Ordinary Impurities 351
12.7.3 Main Component (Assay) Methods 354
12.7.4 In-Process Testing 355
12.7.5 Mutagenic Impurities 356
12.8 Final Remarks 360

13. Method Transfer between HPLC and UHPLC 367

Susanne Fabel
13.1 Introduction 367
13.2 Method Transfer Strategy 368
13.2.1 Geometric Rules for Transfer from
HPLC to UHPLC 369
13.2.2 System Requirements 375
13.2.2.1 Column choice 375
13.2.2.2 Mobile phase 376
13.2.2.3 Flow rate 377
13.2.2.4 Injection volume 378
13.2.2.5 Extra-column volume 378
13.2.2.6 Data collection rate, injection
cycle 380
13.2.3 Method Transfer Applied to Isocratic
Methods 381
xvi Contents

13.2.4 Method Transfer Applied to Gradient

Methods 382
13.2.4.1 Gradient-delay volume 385
13.3 Method Transfer Tools 387
13.4 Limitations of Method Transfer 388
13.4.1 Frictional Heating 388
13.4.2 Stationary Phase Differences 390
13.4.3 Regulatory Limitations 392
13.5 Conclusion 394

14. Small-Molecule Pharmaceutical Impurities Test Method

Validation: Precision Acceptance Criterion 397
Russell L. Hertzler, Mark D. Johnson, Teodora Moldovan,
Michael D. Oberlander, James Reynolds, Holger van Lishaut,
and Yanbing Zheng

14.1 Introduction 397

14.2 Horwitz’s Empirical Observation of Test
Method Variability as a Function of Analyte
Concentration 400
14.2.1 Horwitz’s Acceptable Observed Test
Method Reproducibility: RSDR, Observed 407
14.2.2 Evaluation of AbbVie Test Method
Validation Repeatability, Intermediate
Precision Performance 408
14.2.3 Discussion and Conclusion 411

15. Method Development Strategies for Ion Exchange

Chromatography Using pH Gradient Separation for
mAbs, Antibody Drug Conjugates (ADCs) and Other
Complex Bio-Pharmaceuticals 415
Norman L. Fischer and Scott Allen

15.1 Introduction 415

15.1.1 Salt and pH Gradient IEC Separations 417
15.2 Approaches to pH Gradient Separations 418
15.2.1 Individual Acidic and Basic Buffers/No
pH Adjustment 418
 Contents xvii

15.2.2 Combined Buffers with pH Adjustment 419

15.2.3 Buffers Covering a Narrower pH Range 420
15.2.4 General Concept of pH Gradient
Separations 420
15.3 Impact of Ions on IEC Separations 421
15.3.1 Impact of Ions from pH Adjustment on
Salt Gradient Separations 422
15.3.2 Impact of Ions from pH Adjustment on
pH Gradient Separations 423
15.3.3 Impact of Ions from Buffer Concentration
on pH Gradient Separations 425
15.3.4 Further Evaluation of Buffer
Concentration 430
15.3.5 Buffer Composition 431
15.4 Molecule Information and Method Development 434
15.4.1 Mobile Phase Considerations for pH
Gradient Separations 434
15.4.2 mAb pI, Subclass, and Buffer
Concentration 435
15.4.3 mAbs with Higher pIs and Ionic
Strength Gradients 441
15.4.4 Ionic Strength Gradients and Fusion
Proteins 445
15.4.5 AEX pH Gradient Separations:
Vaccines 446
15.4.6 Parent mAb and ADC Separations 448
15.5 Benefits of Narrow vs. Wide pH Gradients 450
15.6 Platform Application of ph Gradients 453
15.7 Purification Using pH Gradient Separations 455
15.8 Mobile Phase Preparation 455
15.9 Sample Preparation 458
15.10 Columns and Column Temperature 460
15.11 Choosing the Separation Mode 462
15.12 System/Column Equilibration 465
xviii Contents

15.13 Summary 467

15.14 Conclusions 467

16. Size Exclusion Chromatography Method Development

for Therapeutic Proteins 471
Alexandre Goyon, Jean-Luc Veuthey, Davy Guillarme,
Szabolcs Fekete

16.1 Introduction to Size Exclusion Chromatography 471

16.2 Theoretical Aspects 472
16.3 Method Development of Proteins 477
16.3.1 Stationary-Phase Selection 477
16.3.1.1 Pore size, particle size, and
column dimensions 477
16.3.1.2 Stationary-phase chemistry 480
16.3.2 Mobile-Phase Selection
(Salts, Buffers, Organic Modifier) 483
16.3.3 Effect of Mobile-Phase Temperature 487
16.4 Applications for Protein Analysis 488
16.5 Perspectives 490

17. Countercurrent Chromatography Using Biphasic

Molecular Liquids and Ionic Liquids: Is This the
Future for Laboratory and Process Purifications? 497
Leslie Brown and Martyn J. Earle

17.1 Introduction 497

17.1.1 Countercurrent Chromatography/
Centrifugal Partition Chromatography
Instrumentation for Ionic Liquid
Applications 498
17.1.2 CCC/CPC Separations Using Ionic
Liquids 498
17.1.3 Aqueous Biphasic Solvent Systems
Using Ionic Liquids 499
17.2 Optimization of Ionic Liquid and Molecular
Biphasic Solvent Systems for CCC/CPC 501
17.2.1 Introduction 501
 Contents xix

17.2.2 Ionic Liquids in Biphasic Solvent

Selection for CCC 502
17.2.3 CCC/CPC Method Development 504
17.2.4 The “Attraction-Repulsion Principle” of
Biphasic Eluent Optimization 506
17.2.5 Ionic Liquid Solvent System Selection 513
17.2.6 Summary of the CCC Use of Ionic
Liquid–Liquid Chromatography 517
17.3 Conclusion 518

Index 531
 Preface

From Greg Webster

Years ago, when starting my industrial analytical chemist career,
the company I started working for directly out of graduate school
sent me to the HPLC Methods Development course by Lloyd Snyder
and Joseph Kirkland. Having spent most of my graduate years on
spectroscopy-related research, I was excited to “fill in the gaps”
from the several chromatography courses I had taken. I knew my
practical experience was lacking.
Snyder and Kirkland’s love of chromatography was evident in
their teaching. It was contagious and set the foundation for an
analytical career that has spanned several decades and many
chromatography applications. After moving back to Chicago, I was
asked to teach a graduate chromatographic methods development
class at Governors State University. The previous instructors
focused the class on regulatory applications and requirements of
chromatography. My view was that students need more method
development theory to be successful in industry—regulations
are not much help without suitable selectivity for the analytical
method. The standard text for liquid chromatographic development
is and remains Snyder and Kirkland’s Practical HPLC Method
Development and I used it for several classes.
I joked after finishing my second book with Pan Stanford
Publishing (now Jenny Stanford Publishing), Poorly Soluble Drugs:
Dissolution and Drug Release, that never again would I take on
such an undertaking. I also realized an update in text was needed
after driving home from a lecture at Governors State in which
I was getting tired of telling students that they needed to focus on
the principles of Snyder and Kirkland’s text, mostly because the
specific columns and instruments used in 1997 were no longer
Practical HPLC Method Development, 2nd edition, L. R. Snyder, J. L. Kirkland, J. L.
Glajich, ©1997, Wiley-Interscience, Print ISBN 978-0471007036.
Poorly Soluble Drugs: Dissolution and Drug Release, G. K. Webster, J. D. Jackson, R. G.

Bell (editors). ©2017, Pan Stanford Publishing, Print ISBN 978-981-4745-45-1.

xxii Preface

in use in current industrial laboratories. For example, few of us

will ever use a 250 × 4.6 mm column packed with 10 µm particles
again. I told them my true life experiences of the battles I personally
had with sister laboratories trying to get them to move from
5 to 3 µm particles and then ultimately having to do it all over
again with UHPLC packings. After further reflection, I called my
dear, very energetic and motivated friend and former co-worker
Laila to see if she was interested in undertaking what at this time
is truly to be my last chemistry book venture. I told Laila that no
one rewrites a “bible.” If we were going to do something, we had
to be different. We decided on a text that covers not just liquid
chromatographic methods development but all the methods that
an industrial chromatographer may see. Our idea was born. Laila
and I approached Jenny Stanford Publishing with our idea and
outline of the proposed chapters. With their interest assured, we
then contacted those we knew who were subject matter experts
in their assigned chapter.
Of course, life and work get in the way. A text we hoped to
complete in 2017 was finished in 2019. I thank Laila for the periods
where she drove the bus. We both thank our authors for their
expertise, patience, and efforts. My hope is that this text may help
an aspiring chromatographer, as Snyder and Kirkland’s text has
helped me.

From Laila Kott

My path to this point has some similarities to Greg’s and some
differences. As an undergrad at the University of Toronto, I took
my time to declare my major and after almost two years at the
university, chemistry it was and still is. The difference is that
there were always core functions of chemistry that I considered
just part of the chemists’ toolbox as I followed one research
project after another. I made the materials I needed to study; I
analyzed the materials I needed to study. I never considered myself
a certain type of chemist, just a chemist.
Several degrees later and having taken a sabbatical from
the scholarly life between my master’s and doctoral degree
to do chemistry in the former Soviet Union, I ended up in the
pharmaceutical industry, where I found my journey through many
different projects had afforded me a very strong background in
a variety of analysis techniques. Since, as Greg mentioned above,
 Preface xxiii

people often focus in one area or one technique, I found that

looking at a problem using several different techniques often gave
one a more complete picture of what you were looking at.
This was the exact time when Greg and I started collaborating
and we have been tackling interesting problems as well as
pharmaceutical industry standard problems ever since. We have
always pushed the newest columns, the newest technology. This
has its drawbacks as not all things new are readily accepted by
the regulatory bodies during new drug filings. Nevertheless,
we noted that there was a decided lack of method development
information for the modern chromatographer out there, and this
is our attempt to rectify the situation. We have had many
challenges during the compilation of this book, but our drive never
wavered as the need is out there. Happy separations!
Chapter 1

A Lesson Learned from UHPLC

Gregory K. Webstera and Laila Kottb

aAbbvie, 1 N. Waukegan Rd,

North Chicago, Illinois 60064, USA

bXenon, 200-3650 Gilmore Way, Burnaby, British Columbia V5G 4W8, Canada

[email protected], [email protected]

1.1 Introduction
In the late 1990s, it was not uncommon for laboratories to argue
over whether it was robust to use the new 3 µm packed particle
columns. Colleagues would argue over the merits of using this
“new” particle over the trusted industry standard 5 µm particle.
These memories bring a smile to those of us trying to bring Ultra
High-Performance Liquid Chromatography (UHPLC) and its 1.2 to
2 µm columns into routine use.
The differences between systems designed for the 3 µm and
UHPLC columns are significant. A change to 3 µm particle did not
require new instrumentation. Not only does a change to UHPLC

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
 A Lesson Learned from UHPLC

require a change in instrumentation, but it required analytical

chemists to go back and learn their chromatography theory again.
In order to use these smaller particle packings, mobile phase pumps
had to be designed that are jokingly able to “pump liquid through
a brick.” New autosamplers that could reproducibly inject nanoliter
volumes had to be devised and built. Detectors capable of tracking
much faster peak elution needed to be constructed. Probably
the biggest challenge for UHPLC was, and still is in many labs,
convincing management of the necessity of the required new capital
expenses needed to bring this technology into modern laboratories.
Fortunately, scheduled depreciation and instrument replacement
have caught up to manning most industry laboratories with the
capability of executing UHPLC methods.
Prior to the UHPLC “revolution,” liquid chromatography
had gotten stale. By this, we mean liquid chromatography had
become so routine and ubiquitous that research in this area and
in many universities has diminished. Methods used in industrial
laboratories became sloppy; void volumes were seldom considered,
plumbing was only worried about when a method transfer would not
work, and injection volumes and detection were solely considered
for linearity.
First and foremost, chemists had to decide if their system
was capable of running the smaller particle sizes that UHPLC called
for. In fact, the integration of this new technology forced chemists
to evaluate and decide if their system was capable of running
the smaller particle sizes that UHPLC called for. Manufacturers
of liquid chromatography instrumentation had to accelerate
their design plans to meet the new demands of UHPLC. Column
manufacturers had to defend their choice of stationary-phase
packings. Superficially porous particles became in vogue again
as they were capable of producing “UHPLC like” separations at
lower pressures and thus, reducing the need for a rush to new
capital expenses in instrumentation. UHPLC led chromatography
to become a focus of industrial and academic research again.

1.2 Technology and Chromatography

For many analytical chemists today, chirality was only a concept
introduced to us in our sophomore organic chemistry classes.
 Technology and Chromatography 

Now for those of us in pharmaceutical laboratories, chirality is

routinely included in determination of purity. The idea of using
chromatography to the ability to selectively resolve a compound’s
capability to rotate a plane of light is, at first, daunting. Yet, after
spending some time with the theory and techniques, one comes
to realize that chiral separation is just another application of the
separation theories we use every day. In other words, it is just an
evolving application of the technology and a notion of evolving
capabilities is something that chemists must sustain during their
career.
It is important to note that instrumental technologies will
continue to develop. Thin-layer chromatography (TLC) changed
to low-pressure column chromatography, which evolved to high-
performance liquid chromatography (HPLC), which has now
grown into UHPLC. However, in all these advancements, we still
have the same concept of stationary phases, mobile phases, and
plate height. Chromatographic theory is not abandoned; it has
just progressed with the changing technology. If one overlooks the
now outdated columns in stationary-phase packings in use at
the time, Snyder and Kirkland’s work on liquid chromatographic
development is just as relevant today as it was when it was
published [1]. In other words, regardless whether if it is TLC, flash
chromatography or supercritical fluid chromatography (SFC), the
principles of normal-phase chromatography apply.

1.2.1 New Goal of Method Development: Speed

The primary goal of chromatographic method development has
always been focused on resolution. One may argue that it is plate
height; however, plate theory revolves around achieving the
necessary resolution for the intended separation [2]. Sensitivity
requirements were added as a focus and industry standards for
toxicity of impurities came to light [3, 4].
Today, such method requirements are commonplace, and the
requirements for selectivity and sensitivity are expected. Method
development is about speed. Every industry is “on the clock” to
deliver results. This notion has trickled down to the analytical lab
in the urgency to get “results out the door” and “get the product
released” for its intended purpose. Thus, the pressure is on to
 A Lesson Learned from UHPLC

develop methods quickly and robustly for their intended use.

Trial-and-error approaches have given way to automated method
development schemes that typically yield a fit-for-purpose
method in much less time than previously required. However, the
fundamental theories behind chromatography have not changed.
Instead, the focus on speed in method development has come
from the application of automated technologies and approaches.
Method development will continue to evolve in this regard, and
more suitable technology developed to accompany that evolution.

1.3 The Chromatography Market

Chromatographic instrumentation and consumables is a
huge market. Separations technologies play a vital role in the
chemical, pharmaceutical, life science, food and beverage, and
environmental industries. The global chromatography instruments
market is projected to be valued at USD 7.86 Billion in 2017 and
is expected to grow at a compound annual growth rate (CAGR) of
6.9% to reach to USD 10.99 billion by 2022 [5]. Market growth is
fostered by ever increasing “food safety concerns, increasing use
of chromatography tests in the drug approval process, and growing
popularity of hyphenated chromatography techniques” [5].
Liquid chromatography sales is expected the reach USD 4.13 billion
by 2021 [6]. This forecast is followed by a gas chromatography
market prediction of USD 4.13 billion by 2021 [7] and a global
flash chromatography market to reach US$ 294.1 million in
2022 [8].

1.4 Chromatographic Method Development

The goal of this text was not to reintroduce theory. We leave
theory to instrumental textbooks. This book focuses on how
methods for each of the major separation techniques in use in
industry are currently being developed. Each chapter addresses
how a segment of chromatographic science, currently in routine
application in chemical, pharmaceutical, and environmental
industrial laboratories is being applied.
 Chromatographic Method Development 

1.4.1 Software for Chromatographic Method

Development
Modeling technologies are prevalent throughout the industry
today and the analytical lab is not an exception. Chromatography
theories along with design of experiments (DOE) approaches
are now available to automate, evaluate, and implement guided
method development approaches. Use of these software programs
has led to significant efficiency gains in reducing the time required
for developing chromatographic methods. Aoi Ariyasu and Eiki
Maeda of Takeda discuss the software approaches available to
industrial laboratories for chromatographic method development.

1.4.2 Flash Chromatography Method Development

In modern synthesis laboratories, flash chromatography is a
valuable tool reaction monitoring and characterization. Flash
chromatography has all but eliminated thin-layer chromatography
in this regard and is one of the core techniques used by chemists
to isolate synthesized or extracted compounds. Robert Bickler
of Biotage presents how flash chromatographic methods are
developed in this regard.

1.4.3 Orthogonality in Column Chromatography

Whenever a chromatographic method is being used for
separations, the same general questions always arise: (1) “How
do I know all my components are resolved?” and (2) “how do I
know all my components have eluted?” The traditional approach
for answering these questions is to use orthogonal separation
methods. While mass spectrometric detection has often been used
to address the first question, it cannot address the second. Thus,
the use of a differing separation theory/technology is applied to
track sample components. In addition, for reversed-phase
chromatography the notion of “orthogonality” has evolved into
the application of different stationary phases expected to yield
significant changes in separation. Jeff and Winton Caldwell
of Princeton Laboratories and Gregory Webster of AbbVie
 A Lesson Learned from UHPLC

discuss how the notion of orthogonality is being used in today’s

chromatographic method development.

1.4.4 Unified Approach to Reversed-Phase and

Normal-Phase Chromatography
As the most commonly used separation technique, liquid
chromatography is the furthest along in the theory of method
development and using technology to reduce development time.
For most laboratories today, reversed-phase chromatographic
method development is governed by an automated approach.
Normal-phase chromatography is typically performed by SFC.
When components of a sample are too polar for efficient reversed-
phase resolution, hydrophilic interaction liquid chromatography
(HILIC) is often applied. Gregory Webster of AbbVie discusses
the automated approach for method development using each of
these techniques.

1.4.5 Polysaccharide-Derived Chiral Stationary Phases

for the Separation of Enantiomers
Chiral chromatography has evolved rapidly from its introduction
as a routine separation technique in the late 1980s. When it
was found that differing enantiomers of a compound often had
significantly different biological activities, the pharmaceutical
industry felt it was best to address safety concerns by implementing
this technique as part of a compounds quality control. By far, the
most applied stationary phases for chiral separations have been
on polysaccharide platforms. Tong Zhang and Pilar Franco of
Chiral Technologies discuss stationary-phase technology for chiral
separations.

1.4.6 Chiral Chromatography: Method Development

As with routine reversed-phase chromatography, chiral
chromatography began with a trial-and-error approach based on
the structure of the chiral stationary phase. Method development
for chiral separations is based on first finding a stationary phase
 Chromatographic Method Development 

that exhibits chiral recognition to the sample(s) of interest. After

pinpointing an appropriate stationary phase, further method
development proceeds in a typical reversed-phase or normal-
phase path depending on which phase is being used. This was often
a time-consuming approach. Industrial laboratories have found
that screening approaches often lead to successful column hits.
Laila Kott of Xenon and Gregory Webster of AbbVie discuss
current automated approaches to chiral method development.

1.4.7 Aqueous Normal-Phase Chromatography Using

Type-C Silica Columns
Type-C silica stationary phases appear to be similar to HILIC
stationary phases and the two techniques are often confused.
Aqueous normal-phase executed on Type-C materials has a silicon
hydride (Si-C) support. HILIC is executed on Type-A or Type-B
material which has a silica (Si-O) support. Joseph Pesek and
Maria Matyska of San Jose State University distinguish the
mechanism of separation for these two technologies and discuss
how method development proceeds with aqueous normal-phase
chromatography.

1.4.8 Ion Chromatography Columns

While less commonly found in industrial laboratories than gas
or liquid chromatography, ion chromatography provides the
analytical chemists with a unique tool for detecting ions. Unique
stationary phases and capabilities for these columns are less
understood by the everyday chemist. Chris Pohl of Thermo Scientific
provides the background on the stationary phases used in ion
chromatography.

1.4.9 Ion Chromatography: Method Development

This chapter provides the process for developing an ion
chromatographic method. Method parameters for ion
chromatography are a little different than found in other liquid
chromatographic techniques and include selection of eluent
systems, detectors, sample preparation, gradient separations,
 A Lesson Learned from UHPLC

and choice of column chemistries. Laila Kott of Xenon discusses a

process for optimizing all the unique ion chromatography
parameters.

1.4.10 Fundamentals, Properties, and Applications

of Stationary Phases of Gas Chromatography
Method Development
Historically, gas chromatography is often the first instrumental
separation technique most chemists are trained on. Stationary
phases for gas chromatography have evolved from wide bore packed
columns made of glass and/or steel to the higher efficiency capillary
columns that are mostly used today. Gas chromatography is still
the primary way to separate and detect volatile components of a
sample. Residual solvent testing in the pharmaceutical industry
still primarily relies on this technique as well. Omprakash Nacham
and Jared Anderson of Iowa State University provide an updated
review of the stationary phases in use today for gas chromatography
and some new ionic liquid phases.

1.4.11 Gas Chromatography Method Development

Separation in gas chromatography is driven by volatility differences
in analytes as well as, like liquid chromatography, the interactions
of the analytes with the stationary phase. Despite the separation
occurring in the gas phase, gas chromatography can be used with a
variety of analytes, be that solid, liquid, or gas at room temperature.
In the pharmaceutical industry, GC is often used for residual
solvent analysis and the detection and quantification of genotoxic
impurities. Tien Ho and Cody Vinci of AbbVie discuss the
various detectors and parameters that can be optimized to make
GC a useful tool.

1.4.12 Method Transfer from HPLC to UHPLC

With the inception of UHPLC, many laboratories were tasked with
transferring methods from traditional HPLC to UHPLC capable
systems. As the UHPLC systems require different conditions,
software approaches were developed for this transfer based on
 Chromatographic Method Development 

traditional column chromatography theory. Susanne Fabel of

Thermo Fisher Scientific reviews the software algorithms applied
for identifying transferring HPLC method conditions to UHPLC.

1.4.13 Chromatographic Method Validation

For most industries, simply developing a separations method is
not good enough. Once developed, the method must show that it
can be executed in a repeatable and robust manner in the original
and other laboratories. Industry guidelines are commonly applied
to validate that the separations method is routinely capable of
meeting its intended purpose. Michael Oberlander and his team
from Abbvie discuss the industry-accepted guidelines to validate
suitability of the developed chromatographic method.

1.4.14 Method Development Strategies for Ion

Exchange Chromatography Using pH Gradient
Separation for mAbs, Antibody Drug Conjugates
(ADCs), and Other Complex Bio-Pharmaceuticals
Factors that affect ion exchange chromatography for both
monoclonal antibodies (mAbs) and antibody–drug conjugates
(ADCs) are salt concentration and gradient, pH gradients, the
pI of the protein, temperature, ion concentration, columns and
choice of separation mode. Norman L. Fischer, Scott Allen, and
Meg Ruesch from Pfizer show the importance of these parameters,
how they are optimized, and how they can affect ion exchange
chromatography of both mAbs and ADCs.

1.4.15 Size Exclusion Chromatography Method

Development for Therapeutic Proteins
Therapeutic proteins are not necessarily single components.
They can exist as aggregates and monomers. While size exclusion
chromatography (SEC) was initially developed to characterize
polymers, it has found major application in the biopharmaceutical
industry. SEC is now commonly used in research and quality
control to demonstrate the proteins in the intended product have
maintained their same character as the ones used in clinical trial
10 A Lesson Learned from UHPLC

studies. Alexandre Goyon, Jean-Luc Veuthey, Davy Guillarme, and

Szabolcs Fekete of the University of Geneva provide the platform
commonly used in industry to develop size exclusion methods for
proteins.

1.4.16 Ionic Liquids and Counter-Current

Chromatography: Is This the Future of Process
Purifications?
Current chromatography (CCC)/extraction (CCE) can be considered
to be an instrumentally created stationary liquid phase, rather
than a stationary phase brought about by chemical bonding of
a liquid phase to a solid support. By using high percentages of
ionic liquid in biphasic eluents in CCC/CCE (termed ionic liquid-
liquid chromatography or ILLC), several dimensions of choice of
mobile and stationary phases have been added to the technology.
The applications of molecular and ionic solvent systems in CCC/
CCE/ILLC include chiral and achiral organics, transition metals,
precious metals, actinides, lanthanides, bio-chemicals and natural
products, pharmaceuticals, and radionucleotides. In CCC/CCE,
compounds can also be partitioned between two mobile phases
(for example, in dual flow operation modes), and used to purify one
or more target compounds.
Method development in laboratory preparative and process
liquid-liquid, counter-current chromatography (CCC)/extraction
(CCE) follows the same criteria as for various other solid-liquid
forms of chromatography, such as solid-phase extraction (SPE), flash
chromatography (FC), and high-performance liquid chromatography
(HPLC). These techniques partition target compounds between
a stationary phase (SP) and a mobile phase (MP), or two mobile
phases in dual flow. SPE/FC/HPLC separations could, in the
reverse-phase mode, be considered a form of liquid-liquid
partitioning, since many bonded phases can be considered to
behave like liquid phases, which are chemically bonded to a solid
stationary phase. In normal-phase SPE/FC/HPLC separations,
liquid boundary layers also have a major effect on the separations
that can be achieved. Leslie Brown and Martyn J. Earle of AECS-
QuikPrep present the current technology for CCC/CCE using
ionic liquids.
 References 11

Disclosure
Laila Kott is an employee of Xenon. Gregory Webster is an
employee of AbbVie. The study contains no proprietary Xenon
or AbbVie data. Xenon and AbbVie jointly participated in the
writing, review, approval, and financial support of this publication.

References

1. L. R. Snyder, J. J. Kirkland, J. L. Glajch, Practical HPLC Method

Development, 2nd ed., Wiley-Interscience: Hoboken, NJ. 1997.
2. G. K. Webster, A. R. Diaz, D. S. Seibert, B. S. Weekley, J. D. Jackson,
Plate number requirements for establishing method suitability,
J. Chromatogr. Sci., 2005, 43, 67–72.
3. “Impurities in New Drug Products,” Q3B(R2), International
Conference on Harmonization of Technical Requirements for
Registration of Pharmaceuticals for Human Use (ICH), 2 June 2006.
4. “Assessment and Control of DNA Reactive (Mutagenic) Impurities
in Pharmaceuticals to Limit Potential Carcinogenic Risk,” M7(R1),
International Conference on Harmonization of Technical Requirements
for Registration of Pharmaceuticals for Human Use (ICH), 31 March
2017.
5. “Chromatography Instruments Market,” July 2017,
marketsetandmarkets.com. Report Code AST 3789.
6. “Global Flash Chromatography Market,” August 2017, 2007
Stratview Research, Telibandha, Raipur, 492001 India. Report
Code SRAM158.
7. “Gas Chromatography (GC) Market Analysis,” August 2017, Grand
View Research, San Francisco, CA 94105. Report ID 1-68038-970-8.
8. “High-performance Liquid Chromatography Market: Global Forecast
until 2021,” February 2017, ReportLinker, Lyon, 69002 France. ID
4753174.
Chapter 2

Software for Chromatographic Methods

Development

Aoi Ariyasu and Eiki Maeda

Pharmaceutical Sciences, Takeda Pharmaceutical Company Limited,
17-85, Jusohonmachi 2-chome, Yodogawa-ku, Osaka 532-8686, Japan
[email protected], [email protected]

2.1 Introduction
Analytical testing and its control strategy play critical roles during the
entire life cycle of the drug product in the pharmaceutical industry.
Chromatography, especially high performance liquid chromato-
graphy (HPLC), ultra-high performance liquid chromatography
(UHPLC), and gas chromatography (GC), represent the key techniques
employed throughout many aspects of pharmaceutical analyses,
including the determination of assays, dissolution, and impurity
profiles, process control testing, and drug metabolism studies
[1–3]. In general, the use of the trial-and-error approach has been
common practice in developing analytical methods for many
years. However, this method usually requires large numbers of

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
14 Software for Chromatographic Methods Development

experiments, and the final outcome is often not optimal, resulting

in the requirement for further development in later development
stages of the drug product. Thus, a more systematic approach to
analytical method development would enable identification of the
optimal method parameters, and contribute to faster and efficient
method development [4].
In this context, the application of quality by design (QbD)
principles, which plan or design quality into an analytical method,
can avoid such time-consuming processes and can enable the
implementation of a systematic approach to method development.
Beyond its traditional use in the pharmaceutical manufacturing
process, the concept of QbD has been extended to analytical
method development and its control strategy, and is known as
analytical QbD (AQbD). In many cases, design of experiments (DOE)
with DOE modeling software and/or chromatographic simulation
software is implemented and utilized from the aspect of QbD and/
or AQbD to achieve optimal method performance [5, 6]. A key
benefit of implementing the QbD approach is the development of an
understanding of multidimensional combinations, the interaction
of input variables, and their impact on quality. In the case of AQbD
for chromatographic method development and its control strategy,
all analytical method parameters (i.e., input variables) that have
a strong influence on chromatographic retention and selectivity
(i.e., the quality/performance of the method) should be studied
in combination, thus defining perfectly known multidimensional
correlations. In addition, an improved understanding of the
correlation between analytical method parameters and method
performance can aid in the recognition of optimal space of the
method parameters, i.e., Design Space. The Design Space, as defined
in the ICH Q8(R2) [7], is the “multidimensional combination and
interaction of input variables that have been demonstrated to
provide assurance of quality.” Determination of the optimal space
of the method parameters is therefore considered a key factor in
quality control (QC), method transfer, and future method changes.
Furthermore, the software, chromatographic simulation
software, and DOE modeling software are important for determining
the correlation between the analytical method parameters and
the method quality, and also in the identification of the optimal
space of the parameters by chromatogram simulation, experimental
 General Features of Software 15

design based on DOE, multivariate analysis calculation, and

other functions. Thus, the software enables the smooth and
efficient application of the AQbD approach in analytical method
development.
In this chapter, we summarize the application of general
commercially available software for HPLC/UHPLC or GC, i.e.,
chromatographic simulation software (e.g., DryLab®, Molnár
Institute, Berlin, Germany [8, 9], Fusion QbD® software, S-Matrix
Corporation, Eureka, California [10] ACD software, ACD/Labs,
Toronto, Canada [11], ChromSword®, ChromSword, Riga, Latvia
[12]) and DOE modeling software [13], which has the potential
for application in analytical method development. We expect that
this information will assist analysts who wish to employ AQbD,
and that it will aid in the effective development of optimized
and/or robust analytical methods.

2.2 General Features of Software

2.2.1 Typical Procedure for the Systematic Development

of Analytical Methods and the Effect of the
Software Employed
Software that is considered potentially useful in the development
of HPLC/UHPLC or GC analytical methods can be characterized
into two main types, namely chromatographic simulation software
(hereinafter referred to as chromatographic software) dedicated
to the modeling and simulation of retention times of target
peaks, and DOE modeling software (hereinafter referred to as
DOE software) for the modeling of multidimensional correlations
between analytical method parameters (input variables) and
method performance (output). Generally, chromatographic
software can simulate chromatograms through fewer experiments
when compared to DOE software, as its function is specialized for
the development of liquid chromatography (LC) or GC methods.
However, DOE software can be applied not only for the optimization
of various other analytical techniques, but also for demonstrating
the multidimensional correlation between sample preparation
procedures, equipment models and operators, and outputs.
16 Software for Chromatographic Methods Development

Thus, the typical strategy for analytical method development

is similar to that shown in Fig. 2.1. First, selection of input variables
to meet the desired goal of the analytical method takes place,
followed by experimental design and execution, response modeling,
and further optimization of the analytical method. Determination
of the goal is the key step during such method development,
as the desired analytical method performance must be defined
upon commencing method development. Goal examples include,
the complete separation of all peaks, a short analytical time,
and analytical robustness. Deep-dive steps employed during
chromatographic method development are listed in Table 2.1. Steps
1–3 are always conducted by analysts, based on their knowledge,
previous experience, and the results obtained during primary
method development. Subsequently, steps 4–7 can be carried
out using certain chromatographic software in combination with
online LC measurements, while step 5 often requires manual data
input into the chromatographic software (i.e., offline measurements).
Although DOE software can be used in steps 4, 6, and 7, step
5 must always be conducted by an analyst.

Figure 2.1 Typical steps of the chromatographic method development

procedure employed in the pharmaceutical industry. Reprinted from [14],
with permission from Elsevier (color altered).

Over the entire pharmaceutical development stages, analytical

methods are normally developed and optimized through repetition
of steps 2–7. In addition, a phased approach is often adopted
for systematic method development, and involves initiating the
investigation with a wide range of method parameters, and later
narrowing down the investigated range to finally yield the optimal
space or point in a phased manner, as shown in Fig. 2.2.
 General Features of Software 17

Table 2.1 Steps of chromatographic method development using software

Step Practical procedure Process

Step 1 Determination of the desired method Goal
performance that is affected by variations in determination
method parameters
Step 2 Determination of method parameters whose Experimental
variations are evaluated
Step 3 Determination of the investigated range of those design
parameters
Step 4 Generation of experimental design to examine the
impacts of parameter variation
Step 5 Execution of experiments based on the Execution of
determined experimental design, and collection experiments
of the method performance results for each
experiment
Step 6 Model development by correlation analysis Analysis
between the investigated method parameters and and method
the obtained method performance optimization
Step 7 Determination of the optimal space or points
of method parameters that result in acceptable
method performance

Figure 2.2 Phased approach for analytical method development.

18 Software for Chromatographic Methods Development

2.2.2 General Software Features and Applicability to the

Various Stages of Analytical Method Development
Analytical method development generally consists of two main
stages, namely method scouting and method optimization. In the
method scouting stage, the column/eluent combination considered
preferable for the analytical method is selected through rough
screening. This combination does not have to be the best strategy
but acts as a guide in the later optimization stages. During the
initial method scouting stage, software capable of roughly estimating
the impact of method parameters on method performance or
simulating analyte behavior is sufficient. Therefore, chromatographic
software that can run the experiments automatically under the
predesigned conditions can be employed. In contrast, in the method
optimization stage, software that assists the precise evaluation
of the influence of the method parameters on the target method
performance, while aiding in efficient optimization through
reducing the number of experiments based on DOE is preferable.
In this stage, critical method parameters and their combinations are
optimized by experimental design strategies using chromatographic
or DOE software, where the chromatographic software addresses
method development in a variety of manners using similar resolution
maps to that shown in Fig. 2.3 to clearly identify the optimal
conditions [15]. In this context, the simultaneous optimization
of many factors and seamless variation in the critical resolution
are of particular importance.
In addition to these two stages, structure-based prediction is
often applied prior to method scouting in early stage analytical
method development, while robustness evaluation is conducted
following method optimization in late stage analytical method
development. During all stages, steps 2–7 are repeated as
necessary.
The different functions of each type of software that are useful
in analytical method development are summarized in Table 2.2.
For example, chromatographic software can be employed in
structure-based retention prediction and chromatogram simulation,
in addition to method scouting, optimization, and robustness
evaluations. Alternatively, DOE software can be used in method
scouting, optimization, and robustness evaluations based on offline
analysis, which contrasts with chromatographic software that
 General Features of Software 19

can be used for these purposes based on online analysis. Further

details of these functions can be found in later sections and on
each vendor website.

Figure 2.3 Simulated resolution maps and chromatograms from

ChromSword® Auto. In the resolution map, the dotted line represents
the cursor, which can be moved horizontally across the resolution
map to obtain the predicted resolution ( y-axis) at the corresponding
mobile phase concentration (% B, x-axis). The chromatograms represent
simulated separation. Reprinted from [15], with permission from Elsevier.

Although a number of different chromatographic software

packages are available (e.g., DryLab®, Fusion QbD® software,
ACD software, and ChromSword®), the basic concept is common
for all products. However, each software program has a different
modeling algorithm and exhibits characteristic functions as
20 Software for Chromatographic Methods Development

developed by each individual company. Users should therefore

select the software that exhibits the desired functions for their
requirements. In addition, many other examples of DOE software
exist, and analysts can adopt their preferred software for their
desired analytical method development. For example, UnicornTM,
Design Expert®, MODDE, MATLAB, and JMP® are common
examples of DOE software. In general, each software is capable of
planning experiments and carrying out multivariate analysis to
determine the correlation between many input variables and a
single output. Indeed, the experimental planning function based
on DOE has been introduced in both chromatographic and DOE
software, while the modeling and simulation techniques differ
between various chromatographic and DOE software.

Table 2.2 General functions of the software employed for chromatographic

method development

(DOE-related)
Structure-
based Chromato Method
retention gram scouting/ Method Robustness
Function prediction simulation screening optimization evaluation
Chromato- Offline Online Online Online Online
graphic /Offline /Offline
simulation
software
DOE (n/a) (n/a) Offline Offline Offline
modeling
software
Online: Separation is conducted automatically for predesigned plans (combination
of parameters), thus no data input prior to simulation/calculation is required.
Offline: Separation is conducted independently of software, thus data input for
simulation/calculation is required.

2.3 Application of Software to Analytical

Method Development
In the following sections, the various features of the software
employed for analytical method development (see Table 2.2) are
introduced, and examples of their applications are provided.
 Application of Software to Analytical Method Development 21

2.3.1 Structure-Based Retention Prediction

To speed up analytical method development, knowledge and
information should be preserved and re-used. Thus, when
developing an analytical method from scratch, analysts must
consider the implementation of a structure-searchable database
that integrates existing screening results, method information,
and vendor applications. Such databases have been developed
and improved for both reverse-phase LC (RPLC) [16] and ion
chromatography (IC) [17]. In the case of RPLC methods, enormous
amounts of accumulated data and experience are available,
since LC methods have been key analytical methods in the
pharmaceutical industry for many years. By searching structure
similarity, one can quickly identify a good starting point (such
as column material, pH, and mobile phase) for target analytes
prior to analytical method development. In addition, the
physicochemical properties of a compound (e.g., pKa, log P, and
log D values) serve as a rule of thumb to aid in selecting
appropriate chromatographic techniques and method conditions,
such as the stationary phase, buffer pH, and mobile phase
additives for targeted and focused screening.
For example, Wang et al. adopted a structure-searchable
database tool, a physicochemical prediction tool, and an LC
simulator to create a holistic workflow for efficient HPLC
method development and lifecycle management for analytical
method development [18]. In particular, they established web-
accessible databases that are searchable by structure using the
commercially available ACD/Web Librarian software (ACD/
Labs, Toronto, Canada). The database includes achiral and chiral
separation methods obtained from application notes provided
by column manufacturers, literature resources, and analytical
methods generated within the company. Such analytical methods
for a structurally similar compound can provide a promising
starting point for method development when an analytical
method for the exact compound is not available. In addition, an
LC Simulator (ACD/Labs, Toronto, Canada) can provide calculated
pKa values based on the target structure, and as such, can
predict optimal separation conditions and retention times using
experimental data and physicochemical property predictions.
22 Software for Chromatographic Methods Development

ChromSword® is a further example of a software package

that can provide analysts with first guess analytical methods
predicted from structural formulae and column properties.
ChromSword® incorporates a database of commercially available
columns with retention and selectivity characteristics, which
have been experimentally calibrated in acetonitrile-water and
methanol-water mobile phases. As such, analysts can determine
alternative columns for their target compounds using the
ChromSword® column database. ACD/Column Selector (ACD/Labs,
Toronto, Canada) and ColumnMatch® (Molnár-Institute, Berlin,
Germany) [19] have similar functions. Following determination
of the tentative stationary and mobile phases using database
predictions, a few trial runs are performed to experimentally
verify the in silico selectivity predictions, and chromatographic
software is employed to simulate chromatographic separation
and aid in the rapid development of faster, greener, and more
robust methods.

2.3.2 Chromatogram Simulation

Chromatographic software dedicated to LC can be employed for

the simulation of chromatograms. Software tools allow the scientist
to model chromatographic separations based upon retention
data from a limited number of experiments (two or more runs).
In addition, preferred separation conditions can be predicted
and virtual chromatograms for user-selected conditions can be
simulated using such modeling software. This approach avoids
labor-intensive trial-and-error experiments, and potentially results
in significant improvements in the efficiency of analytical method
development and in the quality of the final analytical method.
Furthermore, to gain a better understanding of the method, detailed
investigations can be carried out using chromatographic software,
which is capable of simulating optimal conditions through a
combination of DOE and chromatogram simulations. The depth of
the evaluation depends on the goal or intended purpose of
the analytical method and the period provided for the method
development. Such method optimization using chromatographic
software is explained in greater detail in Section 2.3.4.
 Application of Software to Analytical Method Development 23

The simulation feature of chromatographic software such as

ChromSword®, DryLab®, ACD/LC and GC Simulators, and Osiris
are based on the linear solvent strength (LSS) theory [20, 21]. More
specifically, ChromSword® analyzes the LC retention data, builds
retention models, and simulates and predicts the separation. It
allows for the simultaneous optimization of two variables, such as
buffer pH and organic modifier concentration, or temperature and
organic modifier concentration. Zheng et al. reported its capability
in developing the ion-pair chromatographic method for both the
active pharmaceutical ingredients (API) and all known related
compounds in final pharmaceutical tablets [22]. Simultaneous
optimization of buffer pH and organic modifier (acetonitrile)
concentration was performed using ChromSword®. Based on
the obtained six chromatograms of the six isocratic runs for the
aqueous mobile phases at three different pH values and at two
different acetonitrile concentrations, a pair resolution map was
constructed with respect to changes in pH and organic modifier
concentration, and the simulation was performed based on the
map. Baseline separation for the API and all key related compounds
was achieved and the experimental results matched well with the
predicted results.
DryLab® has a similar function to ChromSword®. Kochling
et al. generated the resolution map shown in Fig. 2.4 by only
four experimental runs while varying both temperature and
gradient time [23]. At any point within this map, the simulated
chromatogram can be generated using this software. In this case,
they selected the final analytical method parameters based on
this resolution map, general knowledge, and past experiences.
In addition, Monks et al. generated 3D resolution models for a
constant sample using three different columns, with the aim
of developing a robust method with known tolerances (see
Fig. 2.5). For this purpose, four critical parameters were
examined, including gradient time, temperature, pH of eluent A,
and stationary phase [4]. Following construction of the resolution
cubes and experimental confirmation of the model precision,
working points were selected according to the triple criteria of
the critical resolution, robustness, and run time. As shown in
Fig. 2.6, good correlations between the predicted and experimental
results were observed.
24 Software for Chromatographic Methods Development

Figure 2.4 A heat map generated using DryLab® software which shows
the relationship between time and resolution (top). A simulated
chromatogram at the spot marked by the black dot in the upper image
(bottom). Reprinted from [23], with permission from Elsevier.

Figure 2.5 Experimental design for a 3D HPLC method optimiation.

Reprinted from [4], with permission from Elsevier.
 Application of Software to Analytical Method Development 25

Figure 2.6 Predicted (A) and experimental (B) chromatograms. Reprinted

from [4], with permission from Elsevier.

Alternative LC method development software packages (e.g.,

ACD/LC and GC Simulators, and Osiris) have similar functions,
and can be employed to enhance the efficiency of LC method
development in an analogous way [11, 24, 25]. However, simulation
using chromatographic software has some limitations, such as
its inapplicability to methods employing hydrophilic interaction
chromatography (HILIC) columns. As a number of software
packages apply the linear solvent strength theory or linear free
energy relationships, they are unable to simulate chromatograms
analyzed using an HILIC column, as in such cases, retention is
based on the mixed mode mechanism, i.e., reverse phase and normal
phase. However, Tyteca et al. attempted to develop the model to
predict the retention behavior in HILIC mode and reported its
applicability [26].
26 Software for Chromatographic Methods Development

2.3.3 Method Scouting/Screening

The first step of analytical method development, i.e., determination
of the critical parameters, is carried out either online (auto)
[12, 15, 27] or offline (manual) [28] to identify the factors
and interactions that have significant effects on the responses
among a large set of factors. This step is used to select critical
chromatographic parameters that affect retention and selectivity.
When scouting is performed online, the chromatographic software
or its combination with LC measurements is employed, which
allows control of the separation equipment (e.g., HPLC/UHPLC)
up to peak integration/identification and following simulation
and/or calculation without any manual data input by analysts.
In contrast, the offline method involves the use of software
alone to carry out simulations and/or calculations, i.e., analysts
must input the chromatographic parameters or results by hand.
In both cases, the assistance of computer simulation software
allows the interrelationships between critical parameters and
their impact on the analytical method performance to be studied
without the requirement for extensive laboratory experiments.
As such, preliminary optimized conditions are obtained through a
combination of analytical method parameters.

2.3.3.1 Online method scouting

ChromSwordAuto® (ChromSword, Riga, Latvia) is software
used in LC separation methods, and carries out the defined
tasks line by line, including automatically switching appropriate
columns, solvents, and buffers for scouting purposes. Hewitt
et al. investigated the utility of fully automated HPLC method
development by applying a variety of compounds, and also examined
the utility of “method scouting” using two different approaches
[12]. In their first approach (Approach One, Fig. 2.7), the starting
conditions were defined by analysts based on previous knowledge
and experience, and included selection of a column, buffer, and
organic solvent for the target sample (i.e., start without method
scouting). For their second approach (Approach Two, Fig. 2.7), 42
combinations of columns, aqueous buffers, and organic modifiers
were screened automatically by ChromSword®, the starting
condition was selected subjectively by analysts, and the selected
condition was applied within approximately 24 h without any
 Application of Software to Analytical Method Development 27

Figure 2.7 (a) Approach One gradient chromatogram (no pre-screening

of conditions); (b) Approach Two isocratic chromatogram (conditions
screened prior to optimization); (c) Approach Two gradient chromatogram
(conditions screened prior to optimization) Reprinted from [12], with
permission from Elsevier (color altered).

user assistance (i.e., start with method scouting). Subsequently,

automated method development using ChromSwordAuto® full
optimization was performed for both series. The first approach
identified a gradient method that sufficiently separated the
28 Software for Chromatographic Methods Development

analogs with adequate resolution (Fig. 2.7a) but in the absence

of an appropriate isocratic method. In the second approach, the
software was successful in identifying both an isocratic and
a gradient method of separation for all analogs (Fig. 2.7b,c).
The data suggest that the software provides a consistently
faster and more efficient method development option, while
screening various starting conditions (method scouting) prior to
the implementation of software optimization proved valuable in
obtaining a suitable method.

Figure 2.8 Chromatograms recorded during the screening design using

Fusion AETM for four columns (i.e., C18, C18 shield, fluorophenyl, and
phenyl), three mobile phase pH values (i.e., 3, 5, and 9) and two organic
modifiers (i.e., acetonitrile ACN, and methanol MeOH). Reprinted from [14],
with permission from Elsevier (color altered).

The Fusion® software package also incorporates the method

scouting function. More specifically, Fusion AETM software (S-
Matrix Corporation, Eureka, California), which is based on DOE and
multiple linear regression (MLR), is able to automatically build
programs in Empower and screen multiple columns and mobile
phases using templates. Indeed, DOE and MLR had already
been proven to be good alternatives to LSS for chromatographic
method development [29, 30]. Thus, Debrus et al. applied Fusion
AETM software to obtain 22 chromatograms in the separation of
16 compounds for their selection step to evaluate the stationary
 Application of Software to Analytical Method Development 29

phases of four columns, three mobile phase pH values, and two

organic modifiers [14]. All runs were performed using a generic
gradient from 5% to 95% organic modifier over 4 min, and
complete screening was performed overnight. The most promising
combination was obtained (see Fig. 2.8) where 15 peaks were
automatically integrated. Under this condition, the elution window
was narrow with a good distribution of compound, giving the
opportunity to easily improve the separation during the next
optimization step.

2.3.3.2 Offline method scouting

Boussès et al. reported the utilization of Nemrod-W (LPRAI,
Marseille, France) and JMP 11 (S.A.S. Institute Inc., Cary, NC, USA)
for offline method screening in UHPLC method development
to determine a single API [28]. The target analytical method
performance, e.g., method selectivity, impurity efficiencies, and
environmentally friendly properties, was initially determined, after
which, the analytical method parameters whose variability had an
impact on method performance and their investigated range were
chosen. In this case, a Plackett–Burman design composed of only
eight runs for four selected method parameters was chosen, and
the experiments were executed. The considered responses were
retention time, number of theoretical plates of impurities, and
total volume of mobile phase. Coefficient significance was evaluated
using a t-test comparing the coefficient estimate to its standard
deviation. Graphical analysis of the effects (see Fig. 2.9) indicated
that flow rate and column temperature had a significant negative
effect on retention time, whereas pH had a significant positive
effect on these responses. In the context of impurity efficiencies,
pH was the only significant factor that imparted a positive effect.
Finally, as anticipated, solvent consumption was significantly
affected by the flow rate (positive effect) and gradient slope
(negative effect). Hence, all four factors had an effect on at least
one response of interest. These four critical parameters were
therefore further studied by response surface methodology (RSM)
in subsequent method optimization studies.
A further example of the application of DOE software for
method scouting using Plackett–Burman DOE is introduced
30 Software for Chromatographic Methods Development

below. Wang et al. applied the DOE for analytical method

development for the solid phase extraction, high performance
liquid chromatography, and ultraviolet/evaporative light scattering
detection (SPE-HPLC-UV/ELSD) technique [31]. The combination
of sample preparation and chromatographic separation was then
examined, and the effects of 11 parameters on the resolutions
and signal-to-noise ratios of the chromatographic peaks were
investigated by 12 experimental runs. The resulting multiple
regression and Pareto ranking analysis (Fig. 2.10) indicated that
the six factors (sorbent mass, sample volume, flow rate, column
temperature, evaporator temperature, and gas flow rate) were
statistically significant ( p < 0.05) in this method. Subsequently,
the four parameters among the six factors were further applied
to DOE using the Box–Behnken design for method optimization.

Figure 2.9 Effects of the critical method parameters on the selected

method responses for experiments (a)–(f). Representation of the effects
(sorted by order of importance) of the 4 factors examined (i.e., flow
rate, temperature, pH, and gradient slope) and p-values (Pr(>t)) of
their significance tests for the responses. Reprinted from [28], with
permission from Elsevier (color altered).

2.3.4 Method Optimization

Method optimization is normally conducted following the
identification of critical analytical method parameters by method
 Application of Software to Analytical Method Development 31

scouting/screening as illustrated previously in Fig. 2.2. During

this process, the method parameters are optimized to maximize
the analytical method performance and to develop fit-for-purpose
and robust methods. In the method optimization stage, DOE is
often adopted to investigate the relationship between analytical
method parameters (input variables) and method performance
(output variables) in a systematic and efficient way. In addition,
chromatographic software utilizes the chromatogram simulation.
Although both chromatographic software and DOE software can
be applied for method optimization, the discussion is ongoing
regarding which calculates the optimized analytical method
parameters more accurately. A few reports [23] state that
chromatographic software, such as DryLab®, is significantly more
effective for method optimization than full or partial factorial DOEs;
however, further discussion is required based on accumulated
case studies.
Upon utilization of the software for method optimization,
analysts must consider the investigated range of each analytical
method parameter and the run time of each separation to
achieve reasonable analysis times, together with an acceptable
chromatographic performance. Analysts must therefore set the
method development target based on preferred retention times,
peak resolution, and an understanding of the physicochemical
properties of the analytes and the main degradants of the analytes.
As a result of method optimization, it is often difficult to obtain a
complete resolution for all peaks. In such cases, the resolution
between critical peaks, e.g., the main degradants and substances
which increase over time, takes precedence of that of minor peaks
such as small peaks and substances which was observed in the
early development stage but removed in the later development
stage. Thus, even when software is utilized for method
development, the planning and checks carried out by analysts prior
to and following method development are both necessary and
essential.
In this context, both types of method optimization, i.e.,
using chromatographic and DOE software, are introduced in the
following subsections, and a number of applications are
exemplified.
32 Software for Chromatographic Methods Development
 Application of Software to Analytical Method Development

Figure 2.10 Experimental design using the Plackett–Burman design and the resulting Pareto charts of factorial analysis showing
33

the effects of various factors on the modeled responses. Reprinted from [31], with permission from Springer (color altered).
34 Software for Chromatographic Methods Development

2.3.4.1 Method optimization using chromatographic software

A number of chromatographic software packages adopt the
combined DOE and chromatogram simulation function, where
various kinds of method optimization exist using the chromato-
graphic software, through on/offline optimization combined
with on/offline measurements. Although few chromatographic
software packages contain the online method optimization
function, ChromSword® is the one that does have this capability.
Following determination of the investigated method parameters
and their subsequent input into the software by analysts, the
software constructs the experimental design and automatically
runs LC measurements (online measurements). Based on the
obtained chromatograms, the software can generate the critical
resolution map and propose the optimal method parameters
(offline optimization), or it can search the optimal parameters
by the repetition of experiments in a stepwise fashion (online
optimization). Such online optimization is a trial-and-error approach
carried out with the aid of results from offline optimization.
Example results are shown in Fig. 2.7.
In contrast, DryLab® is a software package that combines
offline measurements with offline optimization. DOE studies are
conducted offline by analysts using full factorial design for the
predetermined range, and the obtained results (e.g., retention
times and peak areas) are input into the software, leading to the
simulation of chromatograms and the generation of a resolution
map for the points within the investigated range. Figure 2.11
shows the 3D resolution map representing the simultaneous
influence of three parameters on the selectivity and critical
resolution of the chromatogram. The minimum resolution value
(i.e., the value of critical resolution) at each point was calculated
by DryLab® and plotted. In this study, Molnár et al. observed
more than one point located in the middle of each robust
sphere (red region) representing regions where the critical
resolution was >1.5, indicating that all peaks were well separated
[32]. These points were therefore taken as the potential sets
of the optimal analytical method parameters.
In addition, both software packages incorporate the gradient
optimization function. This is especially important, as LC methods
for obtaining impurity profiles often require the application of a
gradient program to separate the numerous analyte peaks. For
 Application of Software to Analytical Method Development 35

example, based on the obtained results of two linear gradients

(i.e., increasing organic solvent concentration in the mobile
phase at varying times), the optimized gradient program can
be calculated by the software. The gradient program presented
as the dotted line in each panel of Fig. 2.12 indicates the
optimized gradient program calculated by ChromSword® in this
manner.

Figure 2.11 3D resolution spaces generated by DryLab®. (A) Gradient

time–temperature–ternary eluent composition and (B) gradient time–
temperature–pH models. Reprinted from [32], with permission from
Elsevier.
36 Software for Chromatographic Methods Development

Figure 2.12 Examples of the gradient optimization by ChromSword®,

indicating the optimized gradient separations of flavonoids using different
columns. Gradient profiles were provided by ChromSword® and are
indicated by dotted lines: (A) Phenyl, (B) Phenoxy, (C) Phenyl-amide, (D)
Phenyl-amine. Reprinted from [33], with permission from Elsevier.

2.3.4.2 Method optimization using DOE software

When general DOE software is applied to analytical method
development, analysts must select a suitable DOE type depending
on the stage of analytical method development. Common types
of DOE are summarized in Table 2.3 and illustrated in Fig. 2.13.
Box–Behnken and central composite designs are two of the most
widely used designs for method optimization, and enable the
generation of reliable contour plots or surface response plots as
outputs using fewer experiments than full factorial design. While
the contour plot presents the variation in analytical method
performance (output variables) with the contour line as a function
of two analytical method parameters (input variables), the surface
response plot is a 3D plot showing the relationship between
the method performance index and three method parameters.
 Application of Software to Analytical Method Development 37

Table 2.3 Commonly used types of DOE and their outcomes

Factor
Experimental General Response interaction
Types of DOE design application surface plot investigation

(1) Full factorial Nk design Robustness Possible Possible

confirmation (from >2
level design)

(2) Central 2k full factorial Robustness Possible Partly possible

composite design + center confirmation,
point + star method
points (including optimization
edge points)

(3) Box–Behnken Three-level Method Possible Partly possible

incomplete optimization
factorial design
(not including
the edge points)

(4) Plackett– Economical Method Not possible Not possible

Burman design screening

Figure 2.13 Schematic representations of the experimental design images.

38 Software for Chromatographic Methods Development

Furthermore, the Plackett–Burman design is also commonly

employed for method scouting/screening and for primary method
development, as this type of DOE can identify and rank the
critical analytical method parameters that have a large impact on
the target method performance. This again requires fewer
experiments than Box–Behnken or central composite designs (see
Section 2.3.3.2, Offline method scouting). The selected critical
method parameters based on the results of DOE using Plackett–
Burman design can then be investigated further in subsequent
method optimization studies.
Figure 2.14 shows one example of the contour plots obtained
as a result of DOE studies for method optimization.
DOE studies for method optimization. Boussès et al.
investigated the resolution between various sets of peaks, the
efficiency of critical peaks, the mobile phase volume, and the
volume of ethanol as the target analytical method performance
affected by the four critical analytical method parameters identified
in the previous method screening stage [28]. The relationship
between these method parameters and method performances
was examined using central composite design DOE with a mobile
phase pH range of 4.0–5.0, a gradient slope of 2.0–4.0%/min, a
column temperature of 30–50°C, and a flow rate of 0.2–0.4 mL/
min. The multivariate correlation analysis of the obtained method
performance results with method parameters using the selected
software produced the contour plots shown in Fig. 2.14. The
white regions in the plots indicate that the results of all target
method performances met an acceptable value. Furthermore, the
method operable design regions were identified as a combination
of parameters with ranges where any combination of parameters
could assure an acceptable method performance. The regions are
also known as the optimal space of method parameters.
The above example was obtained following the application
of DOE for an LC separation method. As previously mentioned,
DOE can also be applied for other chromatographic separation
methods, such as SFC methods. In this context, Landagaray
et al. developed the central composite design with three
factors (i.e., flow rate, pressure outlet, and percentage of
ethanol) to optimize the performance of the SFC method [34].
 Application of Software to Analytical Method Development 39

Figure 2.14 Contour plots of the seven method responses of interest as a

function of the 4 studied method parameters. The white zone corresponds
to the optimal space of method parameters where all responses
fulfilled the requirements, with (•) representing the optimum conditions.
Reprinted from [28], with permission from Elsevier (color altered).
40 Software for Chromatographic Methods Development

The results obtained through this experimental design were

input into the software, and the surface response plot outlining
the relationship between the three method parameters and
the method performance was calculated as shown in Fig. 2.15.
Based on these results, analysts can either select the appropriate
analytical method parameters or allow DOE software to select a
set of method parameters to maximize the output analytical
method performance. In this research, the optimization module
of the software was utilized and the optimal method parameters
were selected.

a b

c
Figure 2.15 Surface response modeling of the resolution for the SFC
method as a function of (a) the flow rate, (b) the percentage of ethanol
in the mobile phase, and (c) the outlet pressure. Reprinted from [34],
with permission from Elsevier.
 Application of Software to Analytical Method Development 41

Figure 2.16 Example 1 of the optimal space of method parameters

considering the probability concept. (Left) Sweet Spot plot obtained
by plotting acetonitrile content vs. aqueous phase pH as defined by the
requirements. Regions where only one criterion is met are colored in blue,
while regions where both criteria are met are colored in green. (Right)
Design space defined for the acetonitrile content and pH of the
aqueous phase in consideration of the probability. Reprinted from [35],
with permission from Elsevier.

Recently, the probability concept has also been considered

when determining the optimal space of analytical method
parameters, with one example being depicted in Fig. 2.16. The
left-hand panel shows the Sweet Spot plot, which was obtained
by multivariate analysis based on the obtained results, followed
by the construction of a prediction model by the DOE software.
42 Software for Chromatographic Methods Development

However, this model does not take into consideration the

variation in precision of the investigated analytical method
parameters and model errors. The border lines were therefore
reconsidered to assure a method performance with a probability
of 99%, and the resulting optimal/design space is presented
as the green region of the right-hand panel.
Hubert et al. also adopted the probability concept for
determination of the optimal space using Monte-Carlo simulations
[36]. The determined optimal space is presented as the area
within the blue rectangle in Fig. 2.17C, with the working point
indicated by the red circle. Based on this result, the robustness
was evaluated for the conditions within the optimal space in
terms of the preset target method performance.

Figure 2.17 Example 2 of the optimal space of method parameters

considering the probability concept: Two-dimensional qualitative
probability surfaces determined by Monte-Carlo simulations. (A) X.ACN
was fixed, pH and T were varied. (B) pH was fixed, X.ACN and T were
varied. (C) T was fixed, pH and X.ACN were varied. The area within the blue
rectangle represents the qualitative design space selected as the
operational space. The red circle corresponds to the reference point
selected for the formal validation. Reprinted from [36], with permission
from Elsevier.

2.3.5 Robustness Evaluation

The final stage of analytical method development involves
evaluation of the robustness of the optimized method. The
robustness of an analytical method is particularly important in
the case of methods employed over a number of years, such as
release testing methods for commercial products. As stated in ICH
Q2(R1) [37], robustness is a measure of reliability of an analysis
 Application of Software to Analytical Method Development 43

with respect to deliberate variations in method parameters and

should be evaluated during method development phase. Indeed,
robustness should be evaluated and confirmed at least in the
final stage of the method development phase. The conventional
approach for robustness evaluation is the one parameter approach,
i.e., the effect of changing a single analytical method parameter
on method performance is evaluated while the other parameters
remain unchanged. However, it is not possible to examine the
interaction effects of various factors using this approach. As such,
the application of multivariate analysis has recently become more
common, as its procedure for robustness evaluation is similar to
DOE studies for method optimization, with the only differences
being the investigated range of method parameters and DOE type.
A number of methods are available to carry out robustness
evaluations. For example, full factorial DOE is a basic technique
used to confirm that all edge points of an optimal space meet the
acceptance criteria. In the case of an analytical method developed
and optimized based on DOE studies that was consequently well
understood, and where the identification of the optimal space of
method parameters within the contour plot or surface response
plot had been carried out, the application of the partial factorial
DOE or Plackett–Burman design may be acceptable for the
subsequent robustness evaluation studies.
Furthermore, the functions of chromatographic software
packages have recently been expanded, including the inclusion of
automatic robustness evaluation. For example, some chromato-
graphic software packages can conduct DOE modeling and analysis
combined with the simultaneous operation of chromatographic
apparatus, i.e., steps 4–6 in Table 2.1, can be performed
automatically. For example, ChromSword® software coupled
with Agilent LC apparatus has the capability to carry out such an
evaluation. Following determination of the parameters and ranges
to be investigated for robustness evaluation, the various values
are input into the software, and the following procedures are
conducted automatically. After designing all experiments and
carrying out the corresponding LC measurements for all planned
experiments, the obtained results of the robustness evaluation are
plotted as a contour plot. Furthermore, robustness confirmation
can also be obtained based on simulated chromatograms, with
44 Software for Chromatographic Methods Development

one example including the Robustness Module of DryLab®.

Where a DOE study has previously been conducted to investigate
the impact of the target analytical method parameters on method
performance over a wide investigation range, the robustness
of the method performance within the space of a narrow range
of method parameters can be evaluated without additional
experiments. Finally, the chromatograms are simulated for all
points of the full factorial design experiments, and the robustness
is evaluated based on the simulated data.

2.4 Advantages and Disadvantages of Using

Available Method Development Software
Programs
As discussed above, the implementation of chromatographic
software and/or DOE software into analytical method development
is extremely beneficial. As software capabilities advance, such
application will be enhanced in future decades. However, at
present, the pros and cons of the software packages and their
implementation in analytical method development can be
summarized as follows:

Pros
 Possibility of significantly reducing experimental/analytical
time, resources, and number of experiments required
 Systematic approach enabling the identification and
statistical analysis of the optimal point or space of the
analytical method parameters
 Feasibility of chromatogram simulation without carrying
out any physical experiments (only for chromatographic
software)
 Easy to understand through a visualized outcome
 Easier transfer for knowledge and technical skills via
better understandings of analytical methods

Cons
 Possibility of inconsistency between the predicted (simulated)
and obtained chromatogram (Prediction accuracy)
 Disadvantages of Using Available Method Development Software Programs 45

 Necessity of training to use the software

 Inability to characterize the effect of injection solvent
mismatch and injection solvent volume overload [38].
Among the various points outlined above, a complementary
relationship can be established between the number of necessary
experiments and the prediction accuracy, with an increase in the
number of experiments increasing the prediction accuracy, and
vice versa. Therefore, each function of the software is employed
at a different development stage depending on the stage in question.
This relationship is summarized in Table 2.4. For example, in the
early development stage, a number of parameters and
multidimensional factors must be widely addressed. However,
the prediction accuracy is not critical at this stage, as no or
few experimental results are required. In contrast, in the late
development stage, only critical parameters over a narrower range
should be considered, and the prediction should be accurate and
based on a large amount of previous and/or ongoing development
outcomes.

Table 2.4 Comparison of software features for chromatographic method

development

Early stage                      Late stage

Development
stage

Type of Structure- Chromato- Method Method Robustness

function based retention gram scouting/ optimization evaluation
prediction simulation screening

Parameters/ Various Focused

interactions
to be
investigated

Prediction Not critical Critical

accuracy

Experiments No or a few experiments Many experiments

46 Software for Chromatographic Methods Development

2.5 Future Perspective

As described in this chapter, the implementation of software into
the analytical method development process provides a number of
advantages in terms of saving time and resources, and reducing
the number of experiments required. Over time, the functions
and features of software packages have been developed and
gradually improved according to scientific advances and recent
trends toward AQbD approaches. As the use of software has
become more common, a number of future perspectives for
the application of such software in chromatographic method
development have been outlined below.
In the middle of the analytical method life cycle, the parameters
of the developed method are often changed to improve separation,
reduce run times, achieve the desired robustness and accuracy at
later developmental stages, and due to the unavoidable changes
in column or critical reagents. Indeed, both major and minor
changes are often made even in post-approval stages. For such
cases, the AQbD approach established for chromatographic
method development by the appropriate software is the key for
such changes; an example would be changes in the analytical
method parameters within the developed optimal space, i.e.,
working point shifts within the established ranges. If the method
is developed without the AQbD approach, even such a minor
change requires additional experiments and method validation.
In contrast, when the method is developed using the AQbD
approach assisted by software allowing the determination of
an optimal space, further studies are generally not required
following such a change.
A further example is the unavoidable change in column vendor.
As the column vendor is not a continuous parameter such as pH
or ion strength, establishing an optimal space that incorporates
multiple column vendors may be difficult without the aid of
software. However, if the analytical method for the original column
is developed based on the AQbD approach with the support of
appropriate software, and that for the new column is developed in
a similar way, the two analytical methods can be easily compared,
and any differences between them identified. Such a comparison
 Future Perspective 47

can be performed either before or after adopting the method.

Figure 2.18 shows an example of the latter case, where the 2D
planes of the 3D resolution cubes are presented with a color gradient
from blue to red, which corresponds to low- to high-resolution
values of two peaks for each of the three columns. The similarities
of the optimal space between the first (original) and second
columns can be confirmed through a similar color distribution
pattern, while the differences between the third column and the
other two columns can be identified through different patterns.
This may suggest the possibility of easily shifting the optimal space
for the first column to that of the second column, and the difficulty
of shifting to the third column. Thus, such an AQbD approach
should ideally be adopted at the beginning of method development,
as it can permit simple and efficient changes in methods in the
later stages of development.

Figure 2.18 Comparison of three different 2-D planes of the 3-D resolution
spaces generated using DryLab® for each column. Reprinted from [4],
with permission from Elsevier.

Thus, in summary, in recent years, the chromatogram

simulation capability of chromatographic software has been
advanced, and the utilization of software for analytical method
48 Software for Chromatographic Methods Development

development has become more widespread. In addition, there

have been increasing reports of the QbD approach being adopted
for analytical method development (AQbD). This AQbD approach
is combined with analytical method development using software
or through DOE studies. It may therefore be possible that the
simulated or calculated results generated by the software can be
utilized for robustness evaluations or other areas of commercial
authorization application documents for analytical methods, thus
avoiding the necessity for obtaining actual experimental data
through a more traditional development strategy. It is therefore
expected that the utilization of chromatographic or DOE software
may be accepted by the regulatory authorities in the future
through the increased implementation of the QbD approach in
the pharmaceutical industry, i.e., representing a science- and
risk-based approach to pharmaceutical analytical method
development.

References

1. Lunn, G. (2005). HPLC Methods for Recently Approved Pharmaceuticals.

(Wiley, USA).
2. Snyder, L. R., Kirkland, J. J., and Glajch, J. L. (1997). Practical HPLC
Method Development. (Wiley, USA).
3. Swadesh, J. K. (1996). HPLC: Practical and Industrial Applications.
(CRC Press, USA).
4. Monks, K. E., Rieger, H.-J., and Molnár, I. (2011). Expanding the term
“Design Space” in high performance liquid chromatography (I).
Journal of Pharmaceutical and Biomedical Analysis, 56, pp. 874–879.
5. Rozet, E., Lebrun, P., Debrus, B., Boulanger, B., and Hubert, P.
(2013). Design spaces for analytical methods. Trends in Analytical
Chemistry, 42, pp. 157–167.
6. Snyder, L. R., Dolan, J. W., and Quarry, M. A. (1987). High-
performance liquid chromatographic method-development using
computer simulation. Trends in Analytical Chemistry, 6, pp. 106–111.
7. ICH Q8(R2). Pharmaceutical Development.
8. Molnar, I. (2002). Computerized design of separation strategies by
reversed-phase liquid chromatography: Development of DryLab
software. Journal of Chromatography A, 965, pp. 175–194.
 References 49

9. Snyder, L. R., Dolan, J. W., and Lommen, D. C. (1989). Drylab® computer

simulation for high-performance liquid chromatographic method
development: I. Isocratic elution. Journal of Chromatography A, 485,
pp. 65–89.
10. Karmarkar, S., Yang, X., Garber, R., Szajkovics, A., and Koberda, M.
(2014). Quality by design (QbD) based development and validation
of an HPLC method for amiodarone hydrochloride and its impurities
in the drug substance. Journal of Pharmaceutical and Biomedical
Analysis, 100, pp. 167–174.
11. Meng, M., Rohde, L., Čápka, V., Carter, S. J., and Bennett, P. K. (2010).
Fast chiral chromatographic method development and validation
for the quantitation of eszopiclone in human plasma using LC/MS/MS.
Journal of Pharmaceutical and Biomedical Analysis, 53, pp. 973–982.
12. Hewitt, E. F., Lukulay, P., and Galushko, S. (2006). Implementation of
a rapid and automated high performance liquid chromatography
method development strategy for pharmaceutical drug candidates.
Journal of Chromatography A, 1107, pp. 79–87.
13. Maeda, E., Urakami, K., Shimura, K., Kinoshita, M., and Kakehi, K.
(2010). Charge heterogeneity of a therapeutic monoclonal antibody
conjugated with a cytotoxic antitumor antibiotic, calicheamicin.
Journal of Chromatography A, 1217, pp. 7164–7171.
14. Debrus, B., Guillarme, D., and Rudaz, S. (2013). Improved quality-
by-design compliant methodology for method development in
reversed-phase liquid chromatography. Journal of Pharmaceutical
and Biomedical Analysis, 84, pp. 215–223.
15. Xiao, K. P., Xiong, Y., Liu, F. Z., and Rustum, A. M. (2007). Efficient method
development strategy for challenging separation of pharmaceutical
molecules using advanced chromatographic technologies. Journal of
Chromatography A, 1163, pp. 145–156.
16. Talebi, M., Schuster, G., Shellie, R. A., Szucs, R., and Haddad, P. R.
(2015). Performance comparison of partial least squares-related
variable selection methods for quantitative structure retention
relationships modelling of retention times in reversed-phase liquid
chromatography. Journal of Chromatography A, 1424, pp. 69–76.
17. Park, S. H., Shellie, R. A., Dicinoski, G. W., Schuster, G., Talebi, M.,
Haddad, P. R., Szucs, R., Dolan, J. W., and Pohl, C. A. (2016). Enhanced
methodology for porting ion chromatography retention data. Journal
of Chromatography A, 1436, pp. 59–63.
18. Wang, L., Zheng, J., Gong, X., Hartman, R., and Antonucci, V. (2015).
Efficient HPLC method development using structure-based
50 Software for Chromatographic Methods Development

database search, physico-chemical prediction and chromatographic

simulation. Journal of Pharmaceutical and Biomedical Analysis,
104, pp. 49–54.
19. Monks, K., Molnár, I., Rieger, H. J., Bogáti, B., and Szabó, E. (2012).
Quality by Design: Multidimensional exploration of the design space
in high performance liquid chromatography method development
for better robustness before validation. Journal of Chromatography A,
1232, pp. 218–230.
20. Horváth, C., Melander, W., and Molnár, I. (1976). Solvophobic
interactions in liquid chromatography with nonpolar stationary
phases. Journal of Chromatography A, 125, pp. 129–156.
21. Nikitas, P., and Pappa-Louisi, A. (2009). Retention models for isocratic
and gradient elution in reversed-phase liquid chromatography.
Journal of Chromatography A, 1216, pp. 1737–1755.
22. Zheng, J., and Rustum, A. M. (2010). Rapid separation of desloratadine
and related compounds in solid pharmaceutical formulation using
gradient ion-pair chromatography. Journal of Pharmaceutical and
Biomedical Analysis, 51, pp. 146–152.
23. Kochling, J., Wu, W., Hua, Y., Guan, Q., and Castaneda-Merced, J.
(2016). A platform analytical quality by design (AQbD) approach
for multiple UHPLC-UV and UHPLC-MS methods development for
protein analysis. Journal of Pharmaceutical and Biomedical Analysis,
125, pp. 130–139.
24. Euerby, M. R., Hulse, J., Petersson, P., Vazhentsev, A., and Kassam, K.
(2015). Retention modelling in hydrophilic interaction chromato-
graphy. Analytical and Bioanalytical Chemistry, 407, pp. 9135–9152.
25. Goga-Remont, S., Heinisch, S., and Rocca, J. L. (2000). Use of
optimization software to determine rugged analysis conditions in
high-performance liquid chromatography. Journal of Chromatography.
A, 868, pp. 13–29.
26. Tyteca, E., Guillarme, D., and Desmet, G. (2014). Use of individual
retention modeling for gradient optimization in hydrophilic
interaction chromatography: Separation of nucleobases and
nucleosides. Journal of Chromatography A, 1368, pp. 125–131.
27. Alden, P., Yurach, D., and Warren, P. (2010). An automated workflow
for accelerated development of a robust LC method within a QbD
framework. The Column, 6, pp. 15–19.
28. Boussès, C., Ferey, L., Vedrines, E., and Gaudin, K. (2015). Using an
innovative combination of quality-by-design and green analytical
chemistry approaches for the development of a stability indicating
 References 51

UHPLC method in pharmaceutical products. Journal of Pharmaceutical

and Biomedical Analysis, 115, pp. 114–122.
29. Debrus, B., Lebrun, P., Ceccato, A., Caliaro, G., Rozet, E., Nistor, I.,
Oprean, R., Rupérez, F. J., Barbas, C., Boulanger, B., and Hubert,
Ph. (2011). Application of new methodologies based on design of
experiments, independent component analysis and design space
for robust optimization in liquid chromatography. Analytica Chimica
Acta, 691, pp. 33–42.
30. Debrus, B., Lebrun, P., Kindenge, J. M., Lecomte, F., Ceccato, A., Caliaro, G.,
Mavar Tayey Mbay, J., Boulanger, B., Marini, R. D., Rozet, E., and Hubert,
Ph. (2011). Innovative high-performance liquid chromatography
method development for the screening of 19 antimalarial drugs based
on a generic approach, using design of experiments, independent
component analysis and design space. Journal of Chromatography A,
1218, pp. 5205–5215.
31. Wang, L., and Qu, H. (2016). Development and optimization of SPE-
HPLC-UV/ELSD for simultaneous determination of nine bioactive
components in Shenqi Fuzheng injection based on quality by
design principles. Analytical and Bioanalytical Chemistry, 408,
pp. 2133–2145.
32. Molnár, I., Rieger, H. J., and Monks, K. E. (2010). Aspects of the
“Design Space” in high pressure liquid chromatography method
development. Journal of Chromatography A, 1217, pp. 3193–3200.
33. Janas, P., Bocian, S., Jandera, P., Kowalkowski, T., and Buszewski,
B. (2016). Separation of flavonoids on different phenyl-bonded
stationary phases-the influence of polar groups in stationary phase
structure. Journal of Chromatography A, 1429, pp. 198–206.
34. Landagaray, E., Vaccher, C., Yous, S., and Lipka, E. (2016). Design
of experiments for enantiomeric separation in supercritical
fluid chromatography. Journal of Pharmaceutical and Biomedical
Analysis, 120, pp. 297–305.
35. Terzić, J., Popović, I., Stajić, A., Tumpa, A., and Jančić-Stojanovic,
B. (2016). Application of analytical quality by design concept
for bilastine and its degradation impurities determination by
hydrophilic interaction liquid chromatographic method. Journal of
Pharmaceutical and Biomedical Analysis, 125, pp. 385–393.
36. Hubert, C., Houari, S., Rozet, E., Lebrun, P., and Hubert, Ph. (2015).
Towards a full integration of optimization and validation phases:
An analytical-quality-by-design approach. Journal of Chromatography
A, 1395, pp. 88–98.
52 Software for Chromatographic Methods Development

37. ICH Q2(R1). Validation of Analytical Procedures: Text and

Methodology.
38. Jeong, L. N., Sajulga, R., Forte, S. G., Stoll, D. R., and Rutan, S. C.
(2016). Simulation of elution profiles in liquid chromatography-I:
Gradient elution conditions, and with mismatched injection
and mobile phase solvents. Journal of Chromatography A, 1457,
pp. 41–49.
Chapter 3

Flash Chromatography Method

Development

J. Robert Bickler
Biotage LLC,
10430 Harris Oaks Blvd., Suite C, Charlotte, North Carolina 28269, USA
[email protected]

3.1 Introduction
Flash chromatography (a.k.a. flash purification) is a rapid
separation and purification technique (hence the term flash)
used by organic, medicinal, and natural product chemists to
isolate one or more compounds in crude chemical mixtures. The
term Flash Chromatography was first coined by W. Clark Still et
al. in his seminal paper “Rapid Chromatographic Technique for
Preparative Separations with Moderate Resolution,” published
in 1978 [1].
Flash chromatography requires only a few items—a solid, inert
sorbent (e.g., silica), a sturdy, inert column (glass, polypropylene),

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
54 Flash Chromatography Method Development

solvents, and a pump or vacuum source. Today, though, most

chemists use automated flash systems with built-in pumps,
gradient making capability, in-line detectors, and fraction
collectors, all controlled by relatively simple-to-use software.

3.2 Goal Setting

As with any chromatographic technique, the results obtained
are dependent on several factors so in order to insure success
goals should be set. Chemists use flash chromatography for one
reason—to purify and isolate the compound synthesized.
Medicinal and organic chemists utilize flash chromatography
to purify synthesized intermediates while natural product
chemists use flash chromatography to separate extracts into as
many individual compounds as possible. In both cases, having
set goals and expectations will help in determining the best
route or routes to satisfactorily complete the task.
Typical goals include the following:
• Purity level
• Yield amount
• Throughput (mg/h or g/h)
• Scalability (from bench to pilot to production)
Note that even though a detector is frequently employed,
quantitation is not performed in this technique and typical goals
for analytical techniques such as precision and sensitivity are
not given any priority.

3.3 Starting Out

Flash chromatography itself is not an overly difficult technique;
however, it is the sample complexity that challenges most chemists.
Having the right tools and performing some preliminary tests
using thin-layer chromatography (TLC) with a variety of different
solvents will supply the information required for an efficient
flash purification. Most chromatography techniques rely on a series
of scouting runs to determine optimal conditions for a separation.
Because this application is to purify a newly synthesized compound
 Starting Out 55

only one time, TLC results are used as a substitute for scouting
runs.

3.3.1 Thin-Layer Chromatography

Thin-layer chromatography can be (and should be) a chemist’s
most important tool for purifying crude mixtures. With TLC,
chemists are able to scout various solvent mixtures that help
optimize both the separation and load amount.
TLC data is directly transferrable to flash chromatography.
A compound’s retardation factor (Rf) on TLC is inversely
proportional to the number of column volumes (CV) required
for its elution from a flash column using the identical solvent
blend (Table 3.1):

CV = 1/Rf (3.1)

Table 3.1 The relationship between TLC Rf values and the number of
required column volumes for elution from a flash column

Rf CV
0.1 10
0.2 5
0.3 3.33
0.4 2.5
0.5 2
0.6 1.67
0.7 1.43
0.8 1.25
0.9 1.11
1 1

In Fig. 3.1, we see how that relationship looks in terms of

both a TLC plate and a flash column.
For those familiar with other chromatography techniques
such as HPLC or GC, think of CV as k. However, the separation
performance with TLC and flash is based on the difference
between any two adjacent compounds’ CV. This value is known as
DCV:
56 Flash Chromatography Method Development

DCV = CV2 – CV1 (3.2)

A large DCV indicates a good separation and a high loading

capacity for any given cartridge or column.

Figure 3.1 The relationship between Rf from a TLC plate and the number
of CV required to elute the compound is illustrated here. The relationship
is reciprocal allowing the chemist to determine the required solvent
volume to elute each compound in the mixture being separated.

3.3.2 Using TLC to Scout Suitable Elution Solvents

When first developing a flash purification method an investigation
into suitable solvents is recommended. Solvents are classified
by their chemistry into selectivity classes and relative elution
strengths versus silica and/or alumina. These are separate
variables which need exploration separately, starting with
selectivity.
Table 3.2 provides a list of commonly used solvents and
their selectivity class. The key to using this data is to assist with
solvent evaluation. Create solvent blends from different selectivity
groups to evaluate their impact on the separation. As an example,
 Starting Out 57

choose hexane and ethyl acetate as well as a mix of hexane

and ether or hexane and DCM. More likely than not, a different
separation will be seen using TLC for each solvent blend.

Table 3.2 Common flash solvents and their selectivity group [1]

Solvent Selectivity group

Ether I
MeOH II
EtOH II
PrOH II
THF III
DCM V
Acetone VI a
EtOAc VI a
MeCN VI b
Toluene VII
Chloroform VIII
Hexane —
Heptane —
Isooctane —

Once you have found a solvent or solvent blend that best

separates your target compound from the other impurities and
by-products it is time to modify the blend’s strength. Strength
adjustment ensures your product elutes with maximum resolution
from the impurities in minimal time using minimal solvent.
Table 3.3 provides a listing of common flash solvents and
their strength relative to silica [2].
Elution strength is adjusted by manipulating the solvent
blend ratio so your target compound has a TLC Rf of between
0.10 and 0.4 (10 CV to 2.5 CV). This is true for both isocratic
and gradient elution. Usually it will not matter if the other
compounds in the crude sample are not separated from each
other as long as your targeted compound is separated from
everything else.
58 Flash Chromatography Method Development

To decrease Rf values, add a weaker solvent such as hexane

as it—and other saturated hydrocarbon solvents—has essentially
no elution strength. Once you have attained a suitable separation
from TLC, you can directly transfer the conditions to your flash
system.

Table 3.3 Common flash solvents and their strength relative to silica

Solvent Strength
MeOH 0.70
EtOH 0.68
PrOH 0.60
THF 0.53
MeCN 0.51
Acetone 0.50
EtOAc 0.43
Ether 0.40
DCM 0.32
Chloroform 0.26
Toluene 0.22
Hexane —
Heptane —
Isooctane —

3.4 Determining Cartridge Load Capacity

Before you can purify your crude material, you need to determine
the amount of media (silica, C18, etc.) needed in order to fully
separate your targeted compound from the others in the mix.
To do this, you can use your TLC’s Rf data for your target
compound and its two nearest neighbors—the one directly above
it and the one directly below it.
Convert each compound’s Rf value to a CV value using
Eq. 3.1 and then calculate the DCV for each pair using Eq. 3.2. Use
the lower DCV value for determining load capacity from Table 3.4.
 Determining Cartridge Load Capacity 59

For example, let us say you have TLC data showing three
spots with Rf values of 0.6, 0.3, and 0.1. Using Eq. 3.1, we can
convert each Rf into its CV, determine the DCV for each adjacent
pair.

CV1 = 1/0.6 = 1.67

CV2 = 1/0.3 = 3.33
CV3 = 1/0.1 = 10
CV2–CV1 = 1.66
CV3–CV2 = 6.67

Based on this data, the lower DCV is 1.66 and the suggested
load amount for any cartridge size is found in the column headed
by <2 in Table 3.4. For instance, if the sample size is 0.1 g, the
appropriate cartridge size is 10 g.

Table 3.4 Flash cartridge load estimate (in grams) based on TLC DCV data

DCV
Cartridge
size (g) <2 2–3.9 >4
5 <0.05 0.05–0.25 0.25–0.5
10 <0.1 0.1–0.5 0.5–1.0
25 <0.25 0.25–0.75 0.75–2.5
50 <0.5 0.5–2.5 2.5–5.0
100 <1.0 1.0–5.0 5.0–10
250 <2.5 2.0–7.5 7.5–25
500 <5 5–25 25–50
750 <7.5 6.0–22.5 22.5–75
1000 <10 10–50 50–100
1500 <15 12–45 45–150

Frequently chemists will take a shortcut and simply apply

the rule-of-thumb that for every 1 g of crude material to be
separated, 100 g of silica is needed; this is known as the 1% rule.
60 Flash Chromatography Method Development

While this approach can work, it ignores the fact that the DCV’s
usefulness is in determining the optimal amount of silica for a
particular separation.

3.5 Determining Elution Style

There are three basic elution styles used in flash chromatography:
• Isocratic
• Linear gradient
• Step gradient
Each of these typically provides a suitable separation, but
gradients elute the compounds in more highly concentrated
bands and often with better resolution between them.

3.5.1 Isocratic Elution

Isocratic elution just means that the solvent ratio remains
constant throughout the purification process (Fig. 3.2).
Your TLC provides you with the solvent ratio to use and the
Rf data. Converting the Rf values to CV and then DCV provides
separation quality information you can use to determine cartridge
size and loading. Although isocratic elution can be quick and
simple, it usually results in lower loading capacity and less than
ideal separations.

3.5.2 Linear Gradient Elution

With linear gradient elution, the mobile phase ratio changes
over time from a weak solvent blend to a stronger solvent blend.
The theory is that poorly retained compounds will retain and
separate better starting the elution with a weak solvent blend.
The better-retained compounds will then separate and elute by
increasing the solvent strength at a constant rate over a period
of time (Fig. 3.3).
A major benefit to linear gradient elution is its ability to
reduce the bandspreading seen with isocratic elution methods.
This concentration effect helps with both compound detection
and recovery. Load capacity is typically improved as well. Most
chemists today use linear gradients for flash purification.
 Determining Elution Style 61

Figure 3.2 An example of isocratic flash chromatography. The

solvent ratio remains constant throughout the chromatography
run and the compounds elute at volumes consistent with the predicted
values from TLC.

Figure 3.3 An example of linear gradient elution. The ratio of

strong solvent to weak solvent changes during the separation
enabling poorly retained compounds to separate better and well retained
compounds to elute in higher concentration.

The keys to a good linear gradient are the starting and

ending solvent ratios and the gradient slope. Many chemists use
a gradient that runs from 0% strong solvent to 100% strong
solvent over a specific time period or volume. While this
approach can work, it is not based on TLC data and often results
in poor separations using either too much or too little solvent.
A better approach is to base your linear gradient on your
TLC data using the algorithm below.
62 Flash Chromatography Method Development

¼ TLC%strong solvent for 1 CV

¼ TLC%strong solvent to 2× the TLC% strong solvent in 10 CV
2× the TLC%strong solvent for 2 CV
As an example, the linear gradient used in Fig. 3.2 was created
based on TLC using 20% ethyl acetate in hexanes. Following the
algorithm above, the gradient started with 5% ethyl acetate for
1 CV. This was followed by a linear gradient from 5% to 40%
ethyl acetate over 10 CV. The gradient ended with 40% ethyl
acetate being held for 2 CV. Gradients based on TLC data using
the algorithm above will elute all compounds with Rf values
from 0.1 to 1.

3.5.2.1 Linear gradients from 0% to 100%

It is common practice, but not supported by this author, among
many chemists to use linear gradients that start at 0% strong
solvent and run to 100% strong solvent over a span of 10–40 column
volumes. This approach skips any efforts to optimize the separation
with the assumption that by running over the entire range of
solvent blends, the compound of interest will come out at some
unpredictable point along the chromatogram.
The lack of predictability, waste of solvent, and necessity of
guessing the amount of silica that should be used for separation
make this shortcut approach quite undesirable.

3.5.3 Step Gradient Elution

The most powerful elution method is the step gradient. It can also
be the most difficult to create and optimize. Step gradients are
more powerful because each compound in the sample mix can
have its separation parameters optimized. This has been found
to be true in many cases, and Fig. 3.4 is a nice example using
the same five-component mix used in Figs. 3.2 and 3.3.
By using an optimized step gradient, the elution of each
compound is determined by step-wise solvent ratio changes.
Some commercially available automated flash chromatography
systems have capability to create an optimized step gradient
from a pair of TLC plate data. In Fig. 3.4, Rf data from 20% and
30% ethyl acetate in hexane TLC runs were used.
 Choosing between Normal- and Reversed-Phase 63

Figure 3.4 An example of step gradient elution. The ratio of strong

solvent to weak solvent changes step-wise during the separation.
If properly created, the amount of separation between each compound can
be optimized leading to higher fraction purity and increased loading.

3.6 Choosing between Normal- and Reversed-

Phase
Polar and ionizable compounds typically are not amenable to
normal-phase chromatographic methods using organic solvents
and polar sorbents such as silica. For these molecules, reversed-
phase chromatography is employed. The difference between
normal- and reversed-phase is that with reversed-phase the
solvents are polar and the sorbent is lipophilic, just the opposite
of normal phase.
In reversed-phase, compounds are retained and separated
primarily based on hydrophobicity. Reversed-phase method
development is also different. While TLC is an excellent method
development tool in normal phase, the results from reversed-
phase TLC plates do not reliably predict an optimal flash
chromatography method. This is due to stationary phase wettability
issues and mass-transfer kinetics which differ between reversed-
phase TLC plate development and reversed-phase column
chromatography.
Instead, methods and gradient shapes are developed by
reversed-phase HPLC using columns packed with the same sorbent
used in the flash cartridge or on a small flash cartridge since
they are re-usable.
64 Flash Chromatography Method Development

A good practice with reversed-phase method development

is to start with a solubility test. If the crude material is water
soluble, start your gradient in 10% organic solvent (typically
methanol or acetonitrile) and run a linear gradient to 100%
organic in 10 column volumes (CV). If the compounds elute too
fast, repeat the test but set the final organic concentration for
50%. This process can be continued until an acceptable separation
is achieved.
If the crude sample is most soluble in organic solvent
(methanol or acetonitrile), start your method development
at 50% organic and run a 10 CV linear gradient to 100% organic.
Make changes to either the starting solvent blend% or the final
blend% to obtain the desired separation.
Once a suitable gradient is determined, a loading study can
be conducted to determine the method’s limitations. This is best
accomplished by increasing the amount of sample loaded on the
flash cartridge. Maximizing sample concentration in the most
polar solvent available will help maximize load. Dissolution
solvents such as DMSO, DMF, and water perform best as they are
quite polar.

3.7 Conclusion
Flash chromatography is a useful technique for the separation
of a broad variety of compounds. It is one of the workhorse
techniques used by chemists who need to isolate compounds
that are recently synthesized, or extracted from a wide variety
of matrices.

References

1. Biotage Flash Consumables brochure (2015) p. 27.

2. http://www.sanderkok.com/techniques/hplc/eluotropic.html.
Chapter 4

Orthogonality in Chromatographic
Methods

Zachary S. Breitbach and Gregory K. Webster

AbbVie Inc., Research and Development, 1 N. Waukegan Rd.,
North Chicago, Illinois 60064, USA
[email protected]

4.1 Introduction
Throughout chromatographic method development, challenges
often arise when peaks elute with the void volume, co-elute with
each other, or do not elute at all. In these cases, it is beneficial
to evaluate and/or develop orthogonal separation methods.
Chromatographic orthogonality simply refers to a separation
technique, mode, or phase that offers unique selectivity. Early on,
this was most frequently seen as using alternate methods with
dissimilar mechanisms of separation to evaluate the selectivity
profile of the original method. This was often performed by using
a normal-phase method to evaluate a reversed-phase profile but
can also be done using TLC, GC, or capillary electrophoresis
techniques. Coupling mass spectrometric detection provides a

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
66 Orthogonality in Chromatographic Methods

great deal of selectivity information for the original method, but is

not considered an “orthogonal” evaluation.

4.2 Traditional View of Orthogonal LC Methods

When developing new LC methods, it is important to consider
what mode of chromatography is best employed. Table 4.1 lists
several unique modes of LC. Each column of the table (i.e., normal-
phase, reversed-phase, and others) can be considered orthogonal
to each other, while each mode within each column of the table
can also be considered orthogonal. For example, normal-phase
separations are always considered orthogonal to reversed-phase
separations. Yet, within the normal-phase mode, there is the
opportunity of unique selectivity when testing HILIC vs. SFC for
example. Further, within a given mode, there are selections of
columns which are intentionally developed to be orthogonal to
each other. As discussed later in the chapter, different approaches
of qualifying or quantifying orthogonality have been developed.
While no model is perfect and cannot overcome the need for column
screening, the systems described herein give the practitioner
direction in choosing columns or groups of columns to screen.

Table 4.1 Example of orthogonal modes of LC

Normal-phase Reversed-phase Others

Traditional normal- Traditional reversed- Ion-exchange
phase phase
Supercritical fluid Mixed mode Size exclusion
HILIC Ion-pairing Hydrophobic interaction
Polar organic

4.3 Column Orthogonality in Reversed-Phase

Chromatography

4.3.1 Introduction
For the last two decades, researchers have developed alternative
approaches to method development. The focus changed from
 Column Orthogonality in Reversed-Phase Chromatography 67

applying different chromatographic methods to evaluate impurity

methods. Instead, the primary method developed was challenged
by columns that offer differing reversed-phase performance.
The second “orthogonal” method is used to identify overlapping
peaks or missing peaks found in the initial method. Figure 4.1
illustrates the method development scheme proposed from
Pellet and co-workers [1].

Prelimimary
Method

ChangeBͲSolvent,
ColumnChange

AcidsorBases
Present?

Yes
No
AdjustTandeith%B
orgradienttimefor 2
Rsш 1.5

ChangepH3Ͳ5units ChangeColumn

Measure|ɷlogɲ|avg

2 2

|ɷlogɲ|avgш0.10 |ɷlogɲ|avg<0.10

1
“Orthogonal”Method

Figure 4.1 Orthogonal column scheme for methods development.

This scheme requires that the change in column selectivity

is large enough to resolve previously overlapping peaks. This
68 Orthogonality in Chromatographic Methods

requires a change in the absolute value of log a (|d log a|) of

≥ 0.05 units. Larger values of |d log a|avg increase the likelihood
of resolving overlapped peaks.

4.3.2 Hydrophobic Subtraction Theory

Several groups worked to develop a system of comparing
reversed-phase columns that ultimately came to be called the
“Hydrophobic Subtraction Theory” (HST) [1–7]. Other systems, such
as linear solvation energy relationships (LSER), have been developed
but, the HST system is currently most widely used. The HST
system for the determination of the separation factor is based on
Eq. 4.1:

log a = h¢H – s¢S* + b¢A + a¢B + k¢C, (4.1)

where the solute properties and the column properties are as

follows:
Solute Properties
η = solute hydrophobicity
σ’ = solute bulkiness
β’ = solute hydrogen-bond basicity
α = solute hydrogen-bond acidity
κ’ = charge on the solute molecule
Column Properties
H = hydrophobicity
S* = steric interaction
A = column hydrogen-bond acidity
B = column hydrogen-bond basicity
C = cation-exchange activity
The HST system applies five column-selectivity parameters
to correlate selectivity of commercial reversed-phase columns.
The parameters used are based on (a) steric interaction,
(b) column hydrogen-bond acidity, (c) column hydrogen-bond
basicity, (d) water sorption and (e) cation-exchange activity.
Hydrophobicity, H, is a measure of the hydrophobic interaction
between the analyte and the stationary phase, which is the
 Column Orthogonality in Reversed-Phase Chromatography 69

traditional model of the reversed-phase separation. Steric

interaction, S*, is defined by Snyder et al. [7] as a measure of the
resistance to insertion of bulky solute molecules into the stationary
phase. This parameter is similar to shape selectivity. Hydrogen-
bond acidity, A, is characteristic of the non-ionized silanols present
in the packing while hydrogen-bond basicity, B, comes from
water sorbed in the alkyl silica of Type B stationary phases.
Finally, column cation-exchange activity, C, comes from ionized
silanols. This last parameter can vary with mobile phase pH.
For an ideal C18 column, H would be considered equal to 1.00,
while S*, A, B, and C would be equal to 0.00. Solute parameters
for Eq. 4.1 have been derived from various authors and
manufacturers for many commercially available reversed-phase
columns and characteristic analytes of varying properties.

4.3.2.1 HST for column comparison

The wide availability of the column data has enabled an efficient
way to compare reversed-phase columns without making an
injection. The authors of the HST system further developed
a simple equation to compare columns called the “Column
Selectivity Function” (Eq. 4.2):

Fs =[( H2 – H1 )2 +( S 2* – S1* )2 +( A2 – A1 )2 +( B2 – B1 )2 +(C2 – C1 )2 ]1/2 (4.2)

Scientists do not have to derive these parameters for a majority

of the commercially available reversed-phase columns on the
market today. Databases are found in USP and many manufacturer
websites.

4.3.2.2 HST values for commercial reversed-phase columns

It is not unusual for laboratories to look for equivalent columns
to act as a backup for their methods. Sometimes, the original
method column is no longer in production and must be replaced.
Using a current website for HST comparisons [8], it is relatively
simple to find comparable columns. It is generally accepted that
comparable columns can be found when Fs for two columns is
less than or equal to three.
For example, three commercial core-shell columns have
column properties seen in Table 4.2.
70 Orthogonality in Chromatographic Methods

Table 4.2 Core-shell column HST column properties at pH = 7.0

Phenomenex Supelco Ascentis Macherey Nagel

Column Kinetex C18 Express C8 Nucleoshell RP 18
H 0.96 0.91 1.11
S* 0 0.0100 0.360
A –0.130 –0.110 0.0770
B –0.0100 0 –0.0400
C 0.120 0.170 0.266

Thus, when looking for an equivalent column to your method

using a Phenomenex Kinetex C18 at pH = 7.0, the Supelco
Ascentis Express C8 is likely to yield an equivalent method
profile. A Macherey Nagel Nucleoshell RP 18 likely will not.
Of course, once picked, this must be experimentally confirmed
and validated (Table 4.3).

Table 4.3 Core-shell column HST column selectivity function

Column Column selectivity function (Fs)

Phenomenex Kinetex C18, 100 A 0.00
Supelco Ascentis Express C8 1.95
Macherey Nagel Nucleoshell RP 18 10.94

4.3.2.3 HST for orthogonal reversed-phase separations

For impurity methods development and validation, the HST
system can be used to find orthogonal sets of columns using large
differences in the column selectivity function (Fs). Some researchers
consider this to be an Fs > 10. Using the same site as in the earlier
section, this time the goal is to find columns that have large
differences in the selective function. In Table 4.4, an alternative
column for a separation developed with a YMC Pack Pro column
is presented along with another commercial phase that has
significantly different column properties (Column X). From this
data, Fs is determined to be 59.93. As illustrated in Fig. 4.2, the
said orthogonal column resolves a previously unresolved peak
seen between peak “e” and “f” on Column X. In this example, the
HST system yields a true orthogonal separation.
 Column Orthogonality in Reversed-Phase Chromatography 71

Table 4.4 Core-shell column HST column properties at pH = 7.0

Column YMC Pack Pro C18 RS Column X

H 1.11 0.73
S* 0.0500 –0.190
A –0.0600 –0.340
B –0.0500 0.020
C –0.220 0.64

Figure 4.2 Orthogonal column profiles.

72 Orthogonality in Chromatographic Methods

The challenge of using large differences in column selectivity

function (Fs) is that column cation-exchange activity, C, can have
a disproportionate effect. This means ionized analytes in the
sample can have a significant effect on Fs. Thus, columns that
appear to have high orthogonality potential may not see a change
in neutral compounds. For separation of neutral compounds,
it is suggested to drop the cation-exchange term, C, from the
Fs equation.

4.4 Orthogonality in Traditional Normal-Phase

LC
The use of traditional normal-phase for analytical methods has
found diminished utility in recent years, in part because of the
versatility of reversed-phase HPLC (vide supra), but also because
supercritical fluid chromatography (SFC) has developed into a
greener, safer, and cheaper direct replacement. As such, there have
been nearly no advancements in normal-phase achiral columns
in recent years. Nonetheless, there are a handful of traditional
normal-phase columns (e.g., bare silica, diol, amino, and cyano) that
are available to offer orthogonality during method development.
The most common means to quantify orthogonality amongst
traditional normal-phase columns has been through the use
of LSERs. Equation 4.3 describes a typical LSER. Solute parameters
(known values for a set of carefully chosen and well defined
probe analytes) are given in capital letters. Once retention factors
for the set of probe analytes are determined, multiple linear
regression analysis can be employed to solve for the stationary
phase descriptors which can be used to compare properties
of stationary phases to one and other.

Log k = c + eE + sS + aA + bB + vV, (4.3)

where the solute properties and the column properties are as

follows:
Solute Properties
E = excess molar refraction
 Orthogonality in Traditional Normal-Phase LC 73

S = dipolarity/polarizability
A = solute hydrogen-bond acidity
B = solute hydrogen-bond basicity
V = molecular volume

Column Properties
e = interaction with n- and π-electrons
s = dipolar interactions
a = hydrogen-bond basicity
b = hydrogen-bond acidity
v = dispersive and cavity effects

One attempt to study silica using LSERs concluded that the LSER
methods typically used to classify differing column interactions
are not appropriate in predicting solute retention [9]. One reason
for this is the difficulty in defining the bulk stationary phase
in a bare silica column [9]. Other reports proved successful in
evaluating bonded normal-phase columns using the same LSER
approach [10] but did report negative values for dispersive and
cavity formation effects. Table 4.5 lists LSER coefficients which
have been determined for a series of traditional normal-phase
columns. This information can be useful in determining the
controlling factors that lead to differences between columns,
as well as, give an idea of which columns of similar phase type
are most unique. As seen in Table 4.5, interactions through
dipolarity and polarizability seem to be fairly consistent among
the tested columns (note, e values are not reported in the table
as they were noted to be insignificant on each test stationary
phase). Interactions through hydrogen bonding give the greatest
differences in the tested columns. For example, this could be
particularly useful in determining which cyanopropyl stationary
phases are most unique. In this case the Spherisorb CN appeared
to be orthogonal to the other cyanopropyl columns. Of course it
is important to note that mobile phase components can greatly
alter interaction types of normal-phase columns and the apparent
differences in selectivity from these evaluations may not always
remain.
74 Orthogonality in Chromatographic Methods

Table 4.5 LSER coefficients for bonded normal-phase columns (values

determined in hexane containing 1% (v/v) methanol)

Particle Surface
Stationary Size area
phase Ligand (µm) (m2/g) c s a b v
Hypersil Bare Silica 5 170 –1.21 1.06 2.23 1.56 –0.83
silica
Hypersil APS Aminopropyl 5 170 –1.14 0.94 2.94 1.20 –0.72
NH2
Lichrospher Dihydroxypropyl 5 250 0.94 1.07 2.37 1.47 –0.85
Diol
Lichrospher Cyanopropyl 5 350 –0.89 0.99 1.94 1.04 –0.64
CN
Ultrasphere Cyanopropyl 5 200 –1.10 1.03 1.97 1.08 –0.60
CN
Hypersil CN Cyanopropyl 5 170 –1.16 0.95 1.86 1.15 –0.61
Spherisorb Cyanopropyl 5 220 –0.99 1.00 2.13 1.69 –0.84
CN
Source: Adapted from reference [10].

4.5 Orthogonality in Supercritical Fluid

Chromatography
As noted earlier, SFC is an excellent replacement for normal-phase
chromatography and has seen continued growth in utility in
recent years as commercial instrumentation has become quite
robust and reliable. With the growing popularity of SFC, the
number of task-specific SFC stationary phases has increased.
Typically SFC columns are polar phases (e.g., bare silica, 2-
ethylpyridine, picolylamine) but some orthogonal separations
have been observed on C18 stationary phases in SFC.
As such, LSERs have been applied to dozens of stationary
phases in SFC to determine and quantify orthogonality [11–15].
The resulting system constants were used to plot different
phases on a five-sided diagram (see Chapter 5). An improved
LSER was developed (see Eq. 4.4) to also include electrostatic
interactions [16]. With the added anionic and cationic (d– and d+)
stationary phase descriptors, over 30 stationary phases have been
evaluated and their resulting polarity has been displayed on a
seven-sided diagram as shown in Fig. 4.3.
 Orthogonality in Supercritical Fluid Chromatography 75

Figure 4.3 Spider diagram based on LSER models calculated with seven
descriptors and Eq. 4.4, with the retention data measured for the 85
neutral species and 24 ionizable species, for 31 columns. Chromatographic
conditions: CO2–MeOH 90:10 (v/v), 25°C, 150 bar, 1 or 3 mL/min depending
on column dimensions. Reproduced with permission from [16].

Log k = c + eE + sS + aA + bB + vV + d–D– + d+D+, (4.4)

where the solute properties and the column properties are as

follows:
Solute Properties
E = excess molar refraction
S = dipolarity/polarizability
A = solute hydrogen-bond acidity
B = solute hydrogen-bond basicity
V = molecular volume
D– = negative charge
D+ = positive charge
Column Properties
e = interaction with n- and p-electrons
76 Orthogonality in Chromatographic Methods

s = dipolar interactions
a = hydrogen-bond basicity
b = hydrogen-bond acidity
v = dispersive and cavity effects
d– = electrostatic interaction with anions
d+ = electrostatic interaction with cations
Further, using over 100 probe analytes, a hierarchical cluster
map of 31 stationary phases was produced using normalized
retention data as shown in Fig. 4.4. This figure is a more
practical representation of column uniqueness and is useful in
understanding the orthogonality of such columns. In Fig. 4.4,
columns which are further away from each other should result in
the largest difference in selectivity.

Figure 4.4 Hierarchical cluster analysis (HCA) on the normalized

retention data (centered and reduced log k values) measured for 109
analytes. Retention dissimilarity in the abscissa reflects Euclidean
distance between the stationary phases. Chromatographic conditions:
CO2–MeOH 90:10 (v/v), 25°C, 150 bar, 1 or 3 mL/min depending on
column dimensions. Reproduced with permission from [16].
 Orthogonality in HILIC 77

4.6 Orthogonality in HILIC

HILIC has been used for the separation of polar analytes for
decades. Traditionally, bare silica was the main stationary phase
used for HILIC; however, in the past 10–15 years, a variety of
new HILIC columns have been introduced and have shown
orthogonal utility when compared with silica and other traditional
polar phases (Table 4.6).
Irgum and co-workers used a series of carefully chosen probe
analytes to evaluate the orthogonality of 22 HILIC stationary
phases by characterizing the phases’ ability to interact with
analytes via hydrophilic, hydrophobic, electrostatic, H-bonding,
dipole-dipole, and p-interactions as well as their interplay with
the eluent [17]. Principle component analysis was applied to
the column/analyte library and four main functional groups were
determined for HILIC column selectivity:

Table 4.6 Groups of HILIC stationary phases as determined in ref. [17].

Group Stationary phase

Cation-exchange Silica
Anion-exchange Amino
Dipole-dipole and multi-point Polymeric sulfobetaine and poly(2-
H-bonding sulfoethyl aspartamide)
Low specific interaction Hydroxyl, diol, amide, monomeric
zwitterionic

A similar but simplified version of evaluating HILIC stationary

phase orthogonality was developed by Lucy and co-workers
using just five specifically chosen analytes [18]. The ratio of
benzyltetramethylammonium chloride’s (BTMA) retention factor
to cytosine’s retention factor yields a measure of a HILIC columns
ion-exchange character. Similarly, the retention factor ratios of
cytosine/uracil and adenosine/adenine represent the hydrophilicity
and H-bonding abilities, respectively.
The following selectivity maps were assembled and are
excellent guides in understanding orthogonality between different
HILIC phases as well as orthogonality within a given group of
HILIC phases (Fig. 4.5).
78 Orthogonality in Chromatographic Methods

Figure 4.5 (A) Hydrophilicity vs. ion-exchange selectivity plot of HILIC

phases, (B) Ion-exchange vs. H-bond formation capability selectivity plot
of HILIC phases, and (C) The HILIC-Phase Selectivity Chart. Columns: 1
ZIC-HILIC, 2 ZIC-HILIC, 3 ZIC-HILIC, 4 ZIC-pHILIC, 5 Nucleodur HILIC, 6
PC HILIC, 7 TSKgel Amide 80, 8 TSKgel Amide 80, 9 PolyHYDROXYETHYL
A, 10 LiChrospher 100 Diol, 11 Luna HILIC, 12 PolySULFOETHYL A, 13
Chromolith Si, 14 Atlantis HILIC Si, 15 Purospher STAR Si, 16 LiChrospher
Si, 17 LiChrospher Si 60, 18 Cogent Type C Silica, 19 LiChrospher 100 NH2,
20 Purospher STAR NH2, 21 TSKgel NH2-100, 22 Atlantis HILIC, 23 Onyx
silica monolith, 24 Zorbax HILIC plus, 25 Silica monolith coated with AS9-
SC, 26 Zorbax RRHD HILIC plus, 27 Acclaim Trinity P1, 28 Cosmosil HILIC,
29 Acclaim HILIC-10, 30 Zorbax Eclipse XDB-C18, 31 XBridge C18, 32 YMC
Pro C18, and 33 Zorbax SB-aq. Adapted from [18].
 Orthogonality in Chiral Chromatography 79

4.7 Orthogonality in Chiral Chromatography

Chiral stationary phases (CSPs) for HPLC/SFC generally perform
best in either normal-phase/SFC mode or in reversed-phase and
polar organic mode. Polymeric and Pirkle-type columns tend
to be screened in normal-phase and SFC, whereas macrocyclic
glycopeptides, cyclofructans, and cyclodextrins are often best
screened in reversed-phase and polar organic modes. These groups
of columns are somewhat orthogonal in terms of the mechanism
that drives the enantiomeric recognition capabilities (e.g., p–p
interactions on Pirkle-type columns vs. H-bonding and inclusion
complexation on cyclodextrin based CSPs).
Because of the unique, multi-point, interaction mechanism
needed for chiral recognition, quantifying orthogonality of CSPs
is difficult and yet to be accomplished. Utilizing large libraries of
chiral compounds and columns tested under many conditions
has been the best way to qualitatively discover orthogonal
CSPs [19–20]. Ideally, one would have a system for reversed-phase
and polar organic mode screening and another for SFC screening,
each system equipped with unique CSPs.

4.8 Orthogonality in Multi-Dimensional

Chromatography
The heart of 2D-LC is having two orthogonal phases in the primary
and secondary dimensions. In this way, one can maximize the
2D separation space and take full advantage of the multiplicative
gains in peak capacity 2D separations can offer. See Fig. 4.6,
where panel A shows no orthogonality. Figure 4.6A describes the
relationship between the data (i.e., retention of analytes) in the x-
and y-axis (i.e., first and second dimension column) as a straight
line; when this is the case, the two columns are orthogonal to
each other. When two columns behave the same or similarly, this
line has a correlation coefficient close to unity. The further the
correlation coefficient deviates from one, the more orthogonal
the dimensions are. This is shown in panel B, which illustrates
theoretically perfect orthogonality, and panel C shows a realistic
representation of an ideal system.
80 Orthogonality in Chromatographic Methods

Figure 4.6 Geometric orthogonality concept. Hypothetical separation of

100 analytes In 10 × 10 normalized separation space. (A) Nonorthogonal
system, 10% area coverage represents 0% orthogonality. (B) Hypothetical
ordered system, full area coverage. (C) Random, ideally orthogonal,
system, area coverage is 63% representing the 100% orthogonality.
Reproduced with permission from [21].

A full mathematical understanding of orthogonal 2D-LC

phases is underdeveloped. However, coupling different columns
from the unique modes described throughout this chapter is the
best chance for success. That being said, using a similar phase
but different mobile phase conditions can also give rise to the
desired orthogonality. Giliar et al. used a geometric approach (as
described in Fig. 4.6) to examine orthogonal 2D configurations
for peptide analysis. In Fig. 4.7, SFC-RP, HILIC-RP, and SCX-RP all
yielded excellent utilization of the 2D separation space, as did a
C18-C18 configuration when operated at high pH in one
dimension at low pH in the other. Needless to say, finding optimal
2D configurations remains a matter of trial and error.
 Orthogonality in Multi-Dimensional Chromatography 81

Figure 4.7 Normalized retention time plots for selected 2D-LC systems.
Reproduced with permission from [21].
82 Orthogonality in Chromatographic Methods

4.9 Conclusion
Separation techniques, modes of chromatography, and
stationary phase material all can be orthogonal. Many models for
characterizing orthogonality can be studied and used to make
decisions on screening column sets for method development (as
discussed in Chapter 5). It should be noted that column vendors
specifically choose to commercialize sets of columns that are
orthogonal for their internal test analytes libraries. Thus, if a
column company has two HILIC columns available, they believe
them to be orthogonal. This, however, may not prove true in
all cases when applied to real world sample or using unique
mobile phases. At the end of the day, using the tools and maps
described herein can only guide educated guesses on column
orthogonality; it is still up to the practitioner to find the right
column for the job as is outlined in the proceeding chapter.

Disclosure
Zachary S. Breitbach and Gregory K. Webster are employees of
AbbVie. AbbVie provided no additional financial support for
this work. AbbVie participated in the writing, reviewing, and
approving the publication. The presentation contains no
proprietary AbbVie data.

References

1. Pellett, J., Lukulay, P., Mao, Y., Bowen, W., Reed, R., Mac, M., Mungerc, R.
C., Dolan, J. W., Wrisley, L., Medwide, K., Toltl, N. P., Chan, C., Skibic, M.,
Biswas, K., Wells, K. A., and Snyder, L. R., ’Orthogonal’ separations for
reversed-phase liquid chromatography, J. Chromatogr. A, 2006, 1101,
122–135.
2. Wilson, N. S., Nelson, M. D., Dolan, J. W., Snyder, L. R., Wolcott, R. G., and
Carr, P. W., Column selectivity in reversed-phase liquid chromatography
I. A general quantitative relationship, J. Chromatogr. A, 2002, 961,
171–193.
3. Wilson, N. S., Nelson, M. D., Dolan, J. W., Snyder, L. R., Wolcott,
R. G., and Carr, P. W., Column selectivity in reversed-phase liquid
chromatography II. Effect of a change in conditions, J. Chromatogr. A,
2002, 961, 195–215.
 References 83

4. Wilson, N. S., Nelson, M. D., Dolan, J. W., Snyder, L. R., Wolcott, R. G.,
and Carr, P. W., Column selectivity in reversed-phase liquid
chromatography III. The physico-chemical basis of selectivity, J.
Chromatogr. A, 2002, 961, 217–236.
5. Dolan, J. W., Maule, A., Bingley, D., Wrisley, L., Chand, C. C., Angodd,
M., Luntee, C., Kriskoe, R., Winstone, J. M., Homeiere, B. A., McCalley,
D. V., and Snyder, L. R., Choosing an equivalent replacement
column for a reversed-phase liquid chromatographic assay procedure,
J. Chromatogr. A, 2004, 1057, 59–74.
6. Marchand, D. H., Snyder, L. R., and Dolan, J. W., Characterization and
applications of reversed-phase column selectivity based on the
hydrophobic-subtraction model, J. Chromatogr. A, 2008, 1191, 2–20.
7. Snyder, L. R., Dolan, J. W., and Carr, P. W., A new look at the selectivity
of RPC columns, Anal. Chem., 2007, 79, 3254–3262.
8. http://www.hplccolumns.org/database/compare.php.
9. Cheong, W. J., and Choi, J. D., Linear solvation energy relationships
in normal-phase liquid chromatography based on retention data on
silica in 2-propanol/kexane eluents, Anal. Chim. Acta, 1997, 342,
51–57.
10. Park, J. H., Yoon, M. H., Ryu, Y. K., Kim, B. E., Ryu, J. W., and Jang, M. D.,
Characterization of some normal-phase liquid chromatographic
stationary phases based on linear solvation energy relationships,
J. Chromatogr. A, 1998, 796, 249–258.
11. West, C., and Lesellier, E., Characterization of stationary phases
in subcritical fluid chromatography by the solvation parameter
model I. Alkylsiloxane-bonded stationary phases, J. Chromatogr. A,
2006, 1110, 181–190.
12. West, C., and Lesellier, E., Characterisation of stationary phases
in subcritical fluid chromatography by the solvation parameter
model II. Comparison tools, J. Chromatogr. A, 2006, 1110, 191–199.
13. West, C., and Lesellier, E., Characterisation of stationary phases
in subcritical fluid chromatography with the solvation parameter
model III. Polar stationary phases, J. Chromatogr. A, 2006, 1110,
200–213.
14. West, C., and Lesellier, E., Characterisation of stationary phases
in subcritical fluid chromatography with the solvation parameter
model IV. Aromatic stationary phases, J. Chromatogr. A, 2006, 1115,
233–245.
15. West, C., and Lesellier, E., Characterisation of stationary phases
in subcritical fluid chromatography with the solvation parameter
84 Orthogonality in Chromatographic Methods

model V. Elaboration of a reduced set of test solutes for rapid

evaluation, J. Chromatogr. A, 2007, 1169, 205–219.
16. West, C., Lemasson, E., Bertin, S., Hennig, P., and Lesellier, E.,
An improved classification of stationary phases for ultra-high-
performance supercritical fluid chromatography, J. Chromatogr. A,
2016, 1440, 212–228.
17. Dinh, N. P., Jonsson, T., and Irgum, K., Probing the interaction mode
in hydrophilic interaction chromatography, J. Chromatogr. A, 2011,
1218, 5880–5891.
18. Ibrahim, M. E. A., Liu, Y., and Lucy, C. A., A simple graphical
representation of selectivity in hydrophilic interaction liquid
chromatography, J. Chromatogr. A, 2012, 1260, 126–131.
19. Akin, A., Antosz, F. J., Ausec, J. L., Greve, K. F., Johnson, R. L., Magnusson,
L. E., Ramstad, T., Secreast, S. L., Seibert, D. S., and Webster, G. K.,
An orthogonal approach to chiral method development screening,
Curr. Pharm. Anal., 2007, 3, 53–70.
20. Barhate, C. L., Joyce, L. A., Makarov, A. A., Zawatzky, K., Bernardoni, F.,
Schafer, W. A., Armstrong, D. W., Welch, C. J., and Regalado, E. L., Ultrafast
chiral separations for high throughput enantiopurity analysis, Chem.
Commun., 2017, 53, 509–512.
21. Gilar, M., Olivova, P., Daly, A. E., and Gebler, J. C., Orthogonality of
separation in two-dimensional liquid chromatography, Anal. Chem.,
2005, 77, 6426–6434.
Chapter 5

Unified Approach to Reversed-Phase and

Normal-Phase Chromatographic Method
Development

Gregory K. Webster
AbbVie Inc., Research and Development, 1 N. Waukegan Rd.,
North Chicago, Illinois 60064, USA
[email protected]

5.1 Introduction
Method development for liquid chromatography has not changed
substantially from the groundbreaking work of Snyder and
Kirkland [1] and updated texts [2, 3]. We still need selectivity
and sensitivity in validatable and transferable methods. The
requirements from the USP [4] and ICH [5] have not changed.
What has changed in this new millennium is the need for efficiency
(speed) over “academic” methods. The pharmaceutical industry
develops thousands of methods a year to support rapid drug
development. Methods need to meet their intended goal in their
role of moving the drug development pipeline forward. The

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
86 Unified Approach to Chromatographic Method Development

theory of chromatographic method development has not been

radically transformed; however, the complexity of the applications,
number of samples running at any given moment, and instrument
technology certainly have. Textbook analytical methods have given
way to fast turnaround.

5.2 Unified Approach to Method Development

There is no set standard approach to developing methods in
liquid chromatography. Older chemists tend to rely on experience
as a way to muscle through the process. However, with today’s
complex regulations regarding detection and selectivity of
impurities, many labs have adopted a more automated approach.
The principles set out by Lloyd Snyder and Joseph Kirkland [1]
are still currently relevant in method development. While column
and system technology has greatly advanced over the years, the
fundamentals have not changed and still hold true.

5.2.1 “Like Dissolves Like”

Fundamental to liquid chromatographic method development
is the understanding of how the process works. If one recalls the
differences between normal and reversed-phase chromatography
and the concept of “like dissolves like” taught in general
chemistry, you are 70% there. We have to recall that normal-
phase chromatography uses a polar stationary phase and a non-
polar mobile phase, and in this case the least polar compounds
elute first. As we increase the polarity of the mobile phase using
alcohol or another additive, there comes a point when the polar
compounds have equal to greater affinity for the mobile phase.
At this point, they traverse down the column. It is this changing
affinity from stationary to mobile phase that results in selectivity.
Nonpolar compounds are less strongly held by the polar stationary
phase and elute off the column first. Elution proceeds as the
compounds increase in polarity.
Conversely, reversed-phase chromatography uses a non-polar
stationary phase with a polar mobile phase. Non-polar compounds
tend to retain (“dissolve”) on the nonpolar stationary phase.
As we decrease the polarity of the mobile phase using acetonitrile
 Unified Approach to Method Development 87

or another additive, there again comes a point with non-polar

compounds having equal or greater affinity for the mobile phase,
and again the resulting traversal down the LC column results in
selectivity based on hydrophobicity. Polar compounds are less
strongly held by the non-polar stationary phase and elute off the
column first. Elution proceeds as the compounds decrease in
polarity (Table 5.1).

Table 5.1 LC elution

LC type Stationary phase Mobile phase

Normal phase Polar Non-polar  More polar
Reversed-phase Non-polar Polar  Less polar

The advantage of looking at method development in this

regard allows chromatographers to standardize a method
development approach regardless of developing a normal or
reversed-phase method. Method development simplifies down
to choosing the right stationary phase that will initially retain
the sample analytes and then manipulating the mobile phase so
that the analytes will traverse the column in a controlled and
reproducible manner.
The scenario in Table 5.1 is true when the analytes of
interest are not charged. To address this, chromatographers can
manipulate the mobile phases in the hopes of eliminating charge.
This is done by operating below the pKa, above the pKb, or
through the use of ion paring reagents. This is also why even
when analytes are suspected to be non-polar, a pH adjustment
to mobile phase is routinely applied in reversed-phase
chromatography.

5.2.2 Information and Samples Required

Ideally, chromatographic method development begins with a
targeted KPSS, or key predictive sample set. Chromatographers
need to know what they intend on separating. Not long ago,
most chromatographers could look at the molecule and know
the conditions that would likely get them 80% done with the
separation. Today, however, regulations often make this more complex.
The chromatographer has to know if the goal of the separation is
88 Unified Approach to Chromatographic Method Development

only to simply look at the profile of the compound, to determine

the potency, detect degradation products (“stability indicating”)
or to profile purity while being selective to synthetic impurities,
degradants, and/or toxic or medically significant impurities. Besides
the actual sample, the method development chemist would be
helped by knowing the number of compounds likely present
(Table 5.2). Additionally with each possible component in the
mixture, key information, such as molecular weight, structure,
pKa, solubility and concentration range, helps establish whether a
normal or reversed-phased method is appropriate and what
detection scheme is needed.
The choice of detection scheme defines what mobile phases
and additives can be used. Examples are given in Table 5.3. The
traditional phosphate buffers and perchloric acid ion pairing
reagents are used much less often in modern laboratories simply
because of their lack of compatibility with mass spectrometric
detection (MS). In regulated laboratories, a mass spectrometric
compatible LC method is desired even for UV quality control
methods. This is because, should an investigation be necessary,
no additional validation would be required for this method on the
MS detector. Knowing what detectors will be needed ultimately
limits what mobile phases can be used.

Table 5.2 What information do I ask for?

Structural information Structure/functional groups

pKa
Hydrophobicity
Chirality
Synthesis Information Synthesis route(s)
Chirality
Samples Drug
Precursors/Intermediates
Reaction-by products
Stressed samples
Initial methods Sample preparation information
LC conditions
Detectors
Goal for method Potency/purity
Stability indicating
 Reversed-Phase Chromatography 89

Table 5.3 Detection schemes

Detector technology Mobile phase restriction(s)

Absorption (UV-Vis, fluorescence) Minimize background chromophore
Refractive index (RI) No gradient
Evaporative light scattering (ELSD) Volatile mobile phases and additives
Mass spectrometry
Charged aerosol detection (CAD)
Electrochemical Charge carrying ions present
Chemical luminescence nitrogen No ACN or other nitrogen containing
detection (CLND) additives

5.3 Reversed-Phase Chromatography

5.3.1 Control of Analyte Retention

For reversed-phase liquid chromatography method development,
it has been shown that changes in organic modifier content yield
the most significant changes in separation [1–3, 13]. Choice and
concentration of the organic modifier is of primary importance.
Separations are affected by other conditions as well, but these
usually yield only small changes in resolution. Knowing this,
systematic method development can be used to minimize the
time it takes to develop methods suitable for their intended
purpose. When starting a method development project, many of
the common considerations are listed in Table 5.4.

Table 5.4 RPLC method considerations

Stationary phase Start with C18, unless early methods indicate

otherwise
Organic modifier Use ACN, MeOH and 1:1 ACN/MeOH; can look at
THF/EtOH/IPA as a second tier
Buffer pH control needed for ionizable compounds
Avoid using TFA/TEA Keep separation mechanism consistent. TFA/
in early screens TEA add ion pairing capabilities that may or
may not be needed
(Continued)
90 Unified Approach to Chromatographic Method Development

Table 5.4 (Continued)

Temperature Start with 30°C to maintain column above

ambient conditions
Gradient steepness Robustness of method
Delay volume on system Needs to be known for transfer

5.3.2 Hydrophobic Subtraction Theory/Columns

The combined use of solvent- and column-type selectivity may
be useful for the separation of extremely difficult samples.
Historically, solvent-type selectivity is first investigated for a C8
or C18 column. If the resulting profile is suitable, no additional
development is needed. If further selectivity is needed, a phenyl
or cyano (CN) columns can be applied. However, with a change
of column a new optimum in mobile phase may be needed.
RPLC retention for non-polar, non-ionic compounds generally
follows:

RtC18 ≈ RtC8 > RtPhenyl > RtC4 > RtC2 > RtCN >> Rtsilica

A chromatographer told me once when changing a column

for method development that “it is just dumb to pick up a random
C18 column and expect it to perform differently.” Choices of
screening columns should maximize their differences as
discussed in Chapter 4. Start with a column you have history with
and then select stationary phases known to behave differently
for the chemistry of the intended sample set.

5.3.3 Solvent Strength

Reversed-phase chromatography uses a non-polar stationary
phase with a polar mobile phase. Traditionally, this meant a C18
stationary phase bonded to a silica support. Today, this definition
has expanded to C1–C30, fluorinated, phenyl and porous carbon
stationary phases with varying endcapping and polar embedding.
Yet, these stationary phases remain relatively non-polar. Thus,
for reversed-phase chromatography, strong solvents are organics
such as acetonitrile (ACN), methanol (MeOH), or tetrahydrofuran
(THF) to “dissolve” the retained analyte from the stationary phase
 Reversed-Phase Chromatography 91

into the mobile phase. Weak solvents would be solvents which

do not induce the analyte into the mobile phase, such as water
(Table 5.5).

Table 5.5 Strong and weak solvents in RPLC

Reversed-phase strong solvents Reversed-phase weak solvents

Primary: Water
Acetonitrile (ACN) Aqueous buffers and ion pairing
Methanol (MeOH) solutions
Tetrahydrofuran (THF)
Secondary:
Ethanol (EtOH)
Isopropanol (IPA)

It should be noted that stationary phases that enhance

selectivity by p–p bonding interactions, such as phenyl phases,
need the presence of solvents such as MeOH to promote these
interactions.

5.3.4 Screening
If early methods are suitable to meet the separation goal, use
them. Otherwise, we need to do some injections! Since method
development is about critical pair resolution [6], not plate
number [7], we need to have the critical pair(s) in mind going
into development. All too often, advocates of computer-assisted
method development are too busy encouraging the use of their
software and they forget to promote that all development starts
with a few injections. We need initial information about the
elution characteristics of the sample and the critical pairs.
Depending on the chromatographer’s experience, there are several
starting points for this. However, they all must include looking
at extremes in pH and mobile phase. Most compounds tend to
have a region of pH where they have optimal resolution.

5.3.4.1 System considerations: UHPLC vs. Traditional

With chromatographic system technology changing at a fast pace,
industry has a mix of system capabilities at play. Older systems
92 Unified Approach to Chromatographic Method Development

still operate with a maximum pressure of ~400 bar while modern

ultrahigh performance liquid chromatography (UHPLC) systems
operate well over 1000 bar. While the maximum pressure of a
system delegates the size of columns, stationary phase particles
and system settings for the method conditions, the theory of
method development is independent of columns size, or system
pressures, with the notable exception that smaller stationary
phase packings operating at higher pressures generally exhibit
higher plate numbers and better resolution. Scaling methods
to older platforms generally work well (Chapter 13). The reverse,
however, generally does not [8, 9].

InitialScreen

ColumnDiameter

4.6mm 3.0mm 2.1mm

Flow=0.4 Flow=0.2
Flow=1.0

Dimensions:100x3.0mm Dimensions:50x2.1mm
Dimensions:150x4.6mm GradientTime:14min GradientTime:7min
GradientTime:20min K*2Ͳk*5:10.8Ͳ27.1 K*2Ͳk*5:4.7–11.7
K*2Ͳk*5:15.3Ͳ38.3

Scheme 5.1 System requirements.

No matter how much one believes in their method

development software, the process demands actual injections
 Reversed-Phase Chromatography 93

on real liquid chromatographs! The conditions in Table 5.6 are

generally where most method developers start.

Table 5.6 Representative starting conditions

Column – C18 C18

Detection1 UV Detection at 201 nm, 254 nm
Organic Component ACN
MeOH
1:1 ACN/MeOH or THF
Buffer pH = 3, pH = 6.8, pH = 9
Gradient2 Initial:
95% weak solvent/5% strong solvent
Final:
5% weak solvent/95% strong solvent
1Assuming existence of a chromophore, otherwise a CAD or ELSD detector with
volatile mobile phase adjustments can be used.
2Based on Scheme 5.1.

As a general rule led by the pharmaceutical industry,

reversed-phase liquid chromatography method development
is based on the notion of column orthogonality (Chapter 4).
Pellet et al. [10] came up with a process based on Scheme 5.2.
In this scenario, the developer optimized the preliminary method
through a series of column and solvent changes to optimize
resolution. The process is governed by the |d log a|avg term discussed
in Chapter 4. When this term reaches a value of 0.10 or greater,
an orthogonal method has been reached.
The Pellet et al. process can be tedious when done manually.
For this reason, the traditional trial and error approach has been
replaced through automation to achieve shorter development
times. Seen in many labs today is a (a) traditional approach, or
(b) varying automated approaches to reversed-phase method
development.

5.3.4.2 Traditional approach

The traditional approach to LC method development is to
systematically vary one factor, such as mobile phase composition
or selection of a column, with all other parameters being held
94 Unified Approach to Chromatographic Method Development

constant. This “trial and error” approach (seen in Figure 4.1 in

the last chapter) is rather slow, depending on the experience of
the method developer and knowledge of the KPSS. This approach
is repeated until an adequately performing instrument method
is obtained.

Prelimimary
Method

ChangeBͲSolvent,
ColumnChange

AcidsorBases
Present?

Yes
No
AdjustTandeith%B
orgradienttimefor 2
Rsш 1.5

ChangepH3Ͳ5units ChangeColumn

Measure|ɷlogɲ|avg

2 2

|ɷlogɲ|avgш0.10 |ɷlogɲ|avg<0.10

1
“Orthogonal”Method

Scheme 5.2 Preliminary orthogonal method development.

This approach generally starts with high levels of acetonitrile

in the mobile phase to ensure elution of all the analytes in the KPSS
 Reversed-Phase Chromatography 95

as well as a low wavelength in tandem with MS to monitor for

impurities. The level of the organic modifiers (or gradient slope) is
then backed off to look for improved critical pair resolution. This
would be repeated for other organic modifiers, pH and column
stationary phase. Temperature would be the last challenge to the
system, as it has been shown to have the least dramatic effect on
resolution for routine separations.
Oftentimes, ion pair reagents (TFA, TEA) are introduced too
early in the method development process. This will be discussed
in a later section of this chapter.
The application of this traditional approach has worked
historically. The issue is time. Modern laboratories want robust
methods with adequate selectivity for the intended purpose, but
the development of these methods is not commonly given time
on the product’s critical path to market. This leads not only to
potential delays but, often also comes with poor development
documentation and process understanding.

5.3.4.3 Automated approach

With the advent of computer controlled instrumentation, the
traditional approach is being replaced with programmed method
development approaches to rapidly develop and optimize separation
parameters. Use of these automated approaches can reduce
development time from weeks to days with full development report
documentation.

5.3.4.3.1 Model 1: Resolution optimization

Resolution optimization models use linear and nonlinear regression
to predict an analyte’s retention time (tr) as a function of multiple
method and instrument parameters. An example of this approach
is using Drylab®, or ChromSword® (see Chapter 2). The software
continuously solves the equations for the critical pairs to predict
their resolution [2]:

ln k = aV2/3 + b(DG) + c, (5.1)

where k is the retention factor of a solute, V the molecular

volume of a solute, DG the energy of interaction of the solute with
water, and a,b,c the characteristic parameters of the reversed-
phase column in the eluent being used.
 96
Unified Approach to Chromatographic Method Development

Figure 5.1 Model 1 approach with Chromsword®. Figure used with permission.
 Reversed-Phase Chromatography 97

The process has to start somewhere, so initial runs to adapt the

theoretical coefficients for prediction of proceeding experimental
conditions need to be run. The major limitation to this approach
is that the equation cannot be modified. It must be independently
applied to solve for other variables. Often it is simply assumed
that variables such as pH and gradient slope do not interact. This
application is detailed in Chapter 2.

5.3.4.3.2 Model 2: Design of experiments (DOE) optimization

The DOE approach screens columns based on the optimization
of factors such as pH, gradient slope, and organic solvent type
(Table 5.7). This practice has been well accepted in the
pharmaceutical industry [11, 12]. An example of this would be
using Fusion QBD®. Here a statistical experiment design is used
to systematically vary the factors together with experimental runs
to explore a multi-factor design space. The interaction effects of
these parameters are optimized with their linear and curvilinear
effects. Proprietary DOE designs allow a method development
approach to simultaneously look at multiple columns, pH, organic
mobile phase type and composition as well as temperature.
A second DOE tier can be added to investigate other parameters
to further optimize method performance.

Table 5.7 DOE optimizations

Tier 1 DOE Tier 2 DOE

Colum screening Temperature
pH Run time
Organic modifier Flow rate
Gradient Use of ion pair reagents

5.3.4.3.3 pH control
Automated systems for method development minimize the
number of mobile phases needed by taking advantage of the
Henderson-Hasselbalch equation to control buffer pH. Table 5.8
illustrates a six-level mass spectrometry detection-friendly pH
screen from pH 2.75 to 6.25 using formic acid and ammonium
formate. For sample sets requiring a low UV detection wavelength,
a similar series is easily generated with phosphate buffers.
Other pH series can be found in the literature and product websites.
98 Unified Approach to Chromatographic Method Development
Figure 5.2 Model 2 approach with Fusion®. Fusion QbD-LC Method Development White Paper, V3.2, S-Matrix Corporation. Used with
Normal-Phase Chromatography

permission.
99
100 Unified Approach to Chromatographic Method Development

Table 5.8 Automated pH control

pH 20.0 mM formic acid (%) 20.0 mM ammonium formate (%)

2.75 100 0

3.16 80 20

3.70 50 50

4.34 20 80

5.42 5 95

6.24 0 100

5.3.5 Ion Pair Reagents

A decade ago, if method development chemists saw aromatic
amine functionality in their KPSS, they would often simply default
to the use of an ion pairing reagent, usually perchloric acid or
trifluoroacetic acid, in the aqueous portion of their mobile phase.
Yet, many times simply buffering the mobile phase pH to a specific
range was all that was needed, as evidenced by the future use
of automated method development systems. Buffering mobile
phases are commonly easier to transfer, more robust and present
less fouling of mobile phase lines and columns than the use of ion
pairing reagents memory effects.
So, when should ion pairing reagents be introduced?
A traditional view of the mechanism for using ion paring reagents
was to cover the bonded phase surface to minimize the ion
exchange retention by absorbed analyte and the uncovered
stationary phase. If the reagent concentration of the ion pairing
reagent is sufficient to provide coverage of the bonded phase
surface, column selectivity becomes relatively unaffected and will
become similar for different columns [13]. However, the use of
polar embedded agents and endcapping in current stationary
phases have greatly reduced the need for this application.
In fact, much of today’s reversed-phase methods no longer require
ion paring reagents. The application can be minimized to only
applications to control acids or bases on a case-by-case basis.
 Normal-Phase Chromatography 101

5.3.6 Reversed-Phase Screen Conditions

Table 5.9 is an example of a preliminary set of conditions for an
orthogonal reversed-phase LC screen.

Table 5.9 Example preliminary reversed-phase screen conditions

Column phases 1. Symmetry C18

(Chapter 4) 2. Xterra C8 RP
3. Thermo CN
4. Discovery HS F5
5. ZirChrom-EZ
6. Bondclone C18
Organic phases 1. ACN
2. MeOH
3. ACN/MeOH
Buffer Formic acid/ammonium formate pH screen
Instrument Traditional 150 × 4.6 mm 5 to 95% organic
settings Columns 1 mL/min, 30°C
(3 µm fully porous particles) Gradient Time: 20 min
K*2-k*5 : 15.3–38.3.
Traditional 100 × 3.0 mm 5 to 95% organic
Columns 0.4 mL/min, 30°C
(2.6–3 µm particles) Gradient time: 14 min
K*2–k*5 : 10.8–27.1
UHPLC 50 × 2.1 mm Col- 5 to 95% organic
umns 0.2 mL/min, 30°C
(≤2.7 µm particles): Gradient time: 7 min
K*2–k*5 : 4.7–11.7.
Detection UV @ 254 nm, MS

5.4 Normal-Phase Chromatography

The traditional mode of normal-phase chromatography using
hexane, heptane, isooctane etc. is no longer a major player in
the world of chromatographic separations. It has been replaced
by modern supercritical fluid chromatography (SFC) [18].
102 Unified Approach to Chromatographic Method Development

Supercritical carbon dioxide behaves very similar to hexane in a

normal-phase mode. However, its use has many advantages such
as operation as a “green” technique, easier recovery of analytes
in preparative applications and greater chromatographic efficiency
in analytical applications. Because of this, our discussion of
normal-phase method development will be limited to achiral SFC
in this chapter and chiral applications in Chapter 7.

5.4.1 Control of Analyte Retention

For normal-phase liquid chromatography method development,
it has again been shown that changes in organic modifier content
yield the most significant changes in separation [1–3]. That is,
choice and concentration of the organic modifier remains of
primary importance. When starting a method development project,
many of the common considerations for normal-phase separations
are listed in Table 5.10.

Table 5.10 Achiral SFC method considerations

Stationary Start with silica unless early methods indicate otherwise

phase
Organic Use MeOH, EtOH or IPA and 1:1 MeOH/EtOH; can look at
modifier MeCl2 as a second tier
Buffer Buffers not used. Acetic acid, ammonium acetate, TFA and
alkyl amines for ionizable compounds
Avoid using Keep separation mechanism consistent. TFA/TEA add ion
TFA/TEA in pairing capabilities that may or may not be needed
early screens
Temperature Need to maintain column above supercritical or evaluate
sub-critical fluid conditions.
Gradient Robustness of method
steepness
Delay volume Needs to be known for transfer
on system

5.4.2 Hydrophobic Subtraction Theory

Because normal-phase columns are less in demand, formal
orthogonal column guides are generally not available online or
 Normal-Phase Chromatography 103

on vendor websites. Start with a column you have history with

and then select stationary phases known to behave differently
for the chemistry of the intended sample set.

5.4.3 Solvent Strength

Normal-phase chromatography uses a polar stationary phase
with a non-polar mobile phase. Commonly used strong and weak
solvents for normal phase chromatography are listed in Table 5.11
[14].

Table 5.11 Strong and weak solvents in NPLC

Normal-phase strong solvents Normal-phase weak solvents

Primary: Hexane/heptane
Isopropanol (IPA) Supercritical fluid carbon dioxide
Methanol (MeOH) (CO2)
Ethanol (EtOH) Supercritical fluid nitrous oxide
Secondary: (N2O)
Acetonitrile (ACN)
Methylene chloride (CH3Cl2)

Traditionally, this meant a silica stationary phase. Today, this

definition has expanded to cyano, diols and most importantly,
alky pyridines [15]. For normal-phase chromatography, strong
solvents are alcohols such as methanol (MeOH), ethanol (EtOH)
or isopropanol (IPA) to “dissolve” the retained analyte from the
stationary phase into the mobile phase. Weak solvents would be
solvents which do not induce the analyte into the mobile phase,
such as supercritical fluid carbon dioxide or hexane.

5.4.4 Ionic Samples

Since buffers are not effective in normal phases, the control of
ionization is typically controlled using acid and basic modifiers.
Acetic and trifluoroacetic acid are popular acidic modifiers with
ammonium acetate and alkyl amines for higher pHs. In addition,
low levels of water can be used as an acidic modifier in SFC. Unlike
with reversed-phase chromatography, trifluoroacetic acid and
alky amines are used in combinations for some normal-phase
applications.
104 Unified Approach to Chromatographic Method Development

5.4.5 Screening
5.4.5.1 Mobile phases
Elise et al. recently studied different mobile phase additives for
the achiral SFC separation of 160 proprietary drugs [16]. Using
methanol without additives as a reference, they studied the basic,
acidic and salt additives to methanol. The basic additives studied
were ammonium hydroxide, diethylamine, diethanolamine and
iso-propylamine. The acidic additives selected were trifluoroacetic
acid and water. The salt additives for this study were ammonium
acetate and ammonium formate. Each additive was evaluated for
their effect upon elution capability, peak shapes, UV baseline drift,
and UV and mass responses (signal-to-noise ratios) for the sample
set.
The study used Derringer desirability functions to rank the
performance of the selected additives. The authors concluded that
ammonium acetate at 20 mM was the most effective additive in
terms of chromatographic efficiency and detection. Diethylamine,
trifluoroacetic acid and water yielded surprisingly poor results.
The study also found that ammonium hydroxide provided useful
MS responses and low baseline drift with UV detection.
In a different screening study, Hicks et al. proposed that
diethylamine or formic acid were effective additives for their sample
sets [17].

5.4.5.2 Columns
In an adjacent study, Elise et al. also studied column orthogonality
with SFC for the achiral SFC separation of 160 proprietary
drugs [18]. They evaluated 1.7–2.5 µm fully porous or 2.6–2.7 µm
superficially porous particles, with a variety of stationary phase
chemistries. Because of the varying column dimensions used,
the linear speed and gradient conditions were adjusted to be nearly
equivalent for each of the columns.
In this study, a Derringer desirability functions was used to
rank the selected columns with regard to retention, retention
window and peak shapes. The results enabled the column set to
be segregated in four groupings: (i) 1-aminoanthracene and 2-
ethylpyridine phases, (ii) diol-type phases and the sole sulfobetaine
 Normal-Phase Chromatography 105

phase, (iii) diethylamine and 2-picolylamine phases and (iv) the

three bare silica phases and C18-bonded phase. For screening,
these groupings may be considered orthogonal to each other.

Figure 5.3 LSER classification of SFC columns. Figure used with permission
from ref. [15].

West and Lesellier [19–24] built a database for commercially

available SFC columns using a QSRR (quantitative structure-
retention relationship) approach. They used a five-dimensional
classification diagram (Fig. 5.3) that was based on a linear solvation
energy relationships (LSER) to represent column selectivity
similarity and differences. This work suggested that three major
groups of stationary phases can be defined: nonpolar, moderately
106 Unified Approach to Chromatographic Method Development

polar and very polar [19]. Alkyl-bonded stationary phases with no

polar functionalities represent the nonpolar group. SFC behaves
somewhat like the reversed-phase retention mode in this grouping.
The moderately polar group consists of stationary phases with
intermediate polarity, such as ODS phases containing a polar-
embedded group or a hydrophilic end-capping group, as well as
aromatic stationary phases. SFC behaves like a mixed mode that is
in between reversed-phase retention and normal-phase retention.
Finally, the polar group consists of more traditional normal phases
such as bare silica gel (Si), 3-aminopropyl bonded silica (NH2), 3-
cyanopropyl-bonded silica (CN) and propanediol-bonded silica
(DIOL) in addition to fluoroalkyl phase (FD), polymer-coated
phases (PEG and PVA) and alkylpyridines. SFC behaves in a more
traditional normal-phase retention mode with the polar phase.
It is recommended that the initial column screening is started
with picking one or two columns from each group. These stationary
phase combinations would provide a high probability to find
appropriate starting conditions for method optimization.

5.4.5.3 Automated Normal-Phase Screening Systems

Due to a smaller market demand, there are no commercially available
systems for automated normal-phase screening. However, the data
could be entered into Model 1 or Model 2 software systems and
evaluated from there.

5.4.5.3.1 Achiral SFC screen conditions

Table 5.12 is an example of a preliminary set of conditions for an
orthogonal achiral SFC screen.

Table 5.12 Example preliminary achiral SFC screening conditions

Column phases 1. Ethylpyridine

2. Diol
3. Sulfobetaine
4. 2-picolylamine
5. Silica
6. C18
 Hydrophilic Interaction Liquid Chromatography 107

Additives 1. Acetic Acid

2. Ammonium acetate
3. neat MeOH
Instrument Columns with 100 × CO2 with 5 to 50% MeOH in
settings 3.0 mm dimensions 10 min, flow rate at 1 mL/min,
(1.7–2.5 µm fully porous temperature at 25°C and out-
particles) let pressure at 150 bar
Columns with 150 × 4.6 CO2 with 5 to 50% MeOH in 15
mm dimensions (2.6 min, flow rate at 2.35 mL/min,
µm superficially porous temperature at 25°C and out-
particles) let pressure at 150 bar
Columns with 150 × 3.0 CO2 with 5 to 50% MeOH in
mm dimensions (2.7 µm 15 min, flow rate at 1 mL/min,
fully porous particles): temperature at 25°C and out-
let pressure at 150 bar
Detection UV @ 254 nm, MS

5.5 Hydrophilic Interaction Liquid

Chromatography
Hydrophilic interaction chromatography (HILIC) may be viewed
as a variation of normal-phase chromatography. Because the
mechanism for HILIC is not well understood, the technique is
being cataloged in its own section for this chapter. The technique
has become popular in its effectiveness to retain polar analytes
that often elute in the solvent front of reverse phase methods.
Oftentimes, polar impurities can be retained while the active
drug elutes early.
Many believe that the primary means of separation is through
the partitioning of analytes into a water layer held on the column
surface. This has been demonstrated by experimental and
simulation investigations [25]. In addition to simple partitioning,
other influences such as multipoint hydrogen bonding and
electrostatic interactions with bonded ionic groups have been
proposed.
108 Unified Approach to Chromatographic Method Development

5.5.1 Control of Analyte Retention

In normal-phase chromatography, the mobile phase is 100%
organic. In HILIC, the initial mobile phase is highly organic
and increasing levels of water are used to elute analyte from
water/polar packing surface. Water in the mobile phase attracts
the polar analytes bound to the stationary phase to be eluted.
Analytes are eluted in order of increasing hydrophilicity.
Considerations for starting a HILIC method development are
listed in Table 5.13.

Table 5.13 HILIC method considerations

Stationary phase Start with silica unless early methods indicate

otherwise
Organic modifier Use ACN/H2O
Buffer Buffers used to control ionization
Avoid using TFA/TEA in Keep separation mechanism consistent. TFA/
early screens TEA add ion pairing capabilities that may or
may not be needed
Temperature Typical held above ambient conditions.
Gradient steepness Robustness of method
Delay volume on system Needs to be known for transfer

5.5.2 Hydrophobic Subtraction Theory

As with SFC, formal orthogonal column guides are generally not
available online or on vendor websites. Start with a column you
have history with and then select stationary phases known to
behave differently for the chemistry of the intended sample set.

5.5.3 Solvent Strength

Hydrophilic interaction chromatography uses a polar stationary
phase with weak polar organic non-polar mobile phase.
Traditionally, this meant a silica stationary phase. Today, this
definition has expanded to bonded chemistries where the column
 Hydrophilic Interaction Liquid Chromatography 109

remains quite polar. Most HILIC methods have been developed

using acetonitrile and will not work if significant levels of alcohols
such as methanol (MeOH) or isopropanol (IPA) are used. The
strong solvent to “dissolve” the retained analyte from the stationary
phase into the mobile phase is water. Weak solvents would be
solvents which do not induce the analyte into the mobile phase,
such as acetonitrile or acetone (Table 5.14).

Table 5.14 Strong and weak solvents in HILIC

HILIC strong solvents HILIC weak solvents

Water Primary:
Aqueous buffers and ion pairing solutions Acetonitrile (ACN)
Methanol (MeOH) Secondary:
Acetone
Dioxane

5.5.4 Ionic Samples

Buffers or ion pairing additives are used to control ionization in
HILIC. Formic and trifluoroacetic acid are popular modifiers as
they maintain mass spectrometry compatibility of the mobile
phase. At pH 6, ionic effects from the stationary phase silanols
begin to influence selectivity [25].

5.5.5 Screening
5.5.1.1 Mobile phases
Traditionally, acetonitrile is used as the organic phase in
hydrophilic interaction chromatography. Use of small levels of
methanol and isopropanol has illustrated mixed results [25, 26].
Thus, the primary alternative to acetonitrile is the use of acetone.
Periat et al. found that for the 82 drugs they studied using
HILIC, pH 3 and pH 6 buffers at 10 and 50 mM were effective [26].
McCalley found additional selectivity with using trifluoroacetic or
heptafluorobutyric acid additives [27]. However, as these additives
can produce memory effects in columns and system tubing, they
may be best suited for a second tier screen.
110 Unified Approach to Chromatographic Method Development

5.5.1.2 Columns
While literature method development studies [25–28] have looked
to optimize columns, Periat et al. went a step further looking for
column orthogonality for HILIC separations [26]. Their approach
proposed four columns with sub-2 µm particles which seems
like an excellent starting point for early HILIC screening: Agilent
Zorbax RRHD (bare silica), Acquity BEH (hybrid silica), Thermo
Syncronis HILIC (zwitterionic phase) and Acquity BEH amide
(hybrid silica bonded with amide group).

5.5.1.3 Automated HILIC phase screening systems

Again, due to a smaller market demand, there are no commercially
available systems for automated HILIC screening. However, Tumpa
et al. [28] demonstrated a design of experiments (DOE) approach
using Mode1 (Umetrics, Umea, Sweden).

5.5.1.4 HILIC screen conditions

Table 5.15 is an example of a preliminary set of conditions for an
HILIC screen [26, 27].

Table 5.15 Example preliminary HILIC screen conditions

Column phases 1. Agilent Zorbax RRHD (bare silica)

2. Acquity BEH (hybrid silica)
3. Thermo syncronis HILIC (zwitterionic phase)
4. Acquity BEH amide (hybrid silica bonded with amide
group)
Mobile phases 1. ACN/ammonium formate, (pH 3.0, 10 mM)
2. ACN/ammonium formate, (pH 6.0, 10 mM)
3. ACN/trifluoroacetic Acid, (0.1%)
Instrument Columns with 50 × Initial isocratic hold at 5% ACN
settings 2.1 mm dimensions for 1 min, then to 65% ACN in
(1.7 µm fully porous 3 min and then 1 min at 65%
particles): ACN and re-equilibration during
3 min, flow rate at 0.5 mL/min,
temperature at 30°C
Detection UV @ 254 nm, MS
 References 111

5.6 Conclusion
Reversed-phase chromatography is once again rapidly evolving
with the advent of small particle columns with high pressure
systems optimized for minimal void volumes and fast detection.
Yet, the basics of method development for this mode is still
very much in accordance with the standards set by Snyder and
Kirkland years ago [1]. The principles set forth by this classic work
have simply been updated to reflect technology capabilities. In
addition, automated system opportunities have enabled statistical
and DOE algorithms to facilitate finding optimal separations
conditions.
Supercritical fluid chromatography has, to date, been
very successful in normal-phase applications, particularly in
pharmaceutical analysis. SFC provides several advantages over
traditional normal-phase liquid chromatography in terms of
resolution, throughput, and consumption of hazardous/expensive
solvents. However, these advantages have not been fully utilized
in achiral analysis. There is little doubt SFC will become more
and more popular in achiral analysis in the near future.
HILIC has found a very successful niche in chromatographing
polar impurities that are difficult to resolve by reversed-phase
separations. In particular, HILIC conditions are especially amenable
to mass spectrometric detection for low level sensitivity and high
selectivity.

Disclosure
AbbVie provided no financial support outside of Dr. Webster
being an employee of AbbVie. AbbVie participated in writing,
reviewing, and approving the publication. The chapter contains no
proprietary AbbVie data.

References

1. L. R. Snyder, J. J. Kirkland, J. L. Glajch: Practical HPLC Method

Development, 2nd ed., Wiley-Interscience: Hoboken, NJ. 1997.
2. HPLC Made to Measure: A Practical Handbook for Optimization, S.
Kromidas, ed., Wiley-VCH, Weinheim, 2006.
112 Unified Approach to Chromatographic Method Development

3. HPLC for Pharmaceutical Scientists, Y. Kazakevich and R. Lobrutto,

ed., John Wiley & Sons, Hoboken, NJ, 2007.
4. <621> Chromatography, United States Pharmacopeia and National
Formulary (USP 40-NF 35). Rockville, MD: United States Pharmacopeia
Convention, 2017.
5. The International Council for Harmonization of Technical
Requirements for Pharmaceuticals for Human Use (ICH), 1996,
Validation of Analytical Procedures: Text and Methodology.
6. G. K. Webster and C. L. Basel, Critical Pairs in Column Chromatography:
A Primer for Pharmaceutical Method Validations, LCGC, 2003, 21(3),
286–294.
7. G. K. Webster, A. R. Diaz, D. S. Seibert, B. S. Weekley, J. D. Jackson,
Plate Number Requirements for Establishing Method Suitability, J.
Chromatogr. Sci., 2005, 43, 67–72.
8. G. K. Webster and A. Elliott, Direct Method Scaling from UHPLC to
HPLC: Is This Feasible for Pharmaceutical Methods? Am. Pharm. Rev.,
2011, 14, 32–40.
9. G. K. Webster and M. A. Gragg, Scaling LC Methods Using Superficially
Porous Particle Stationary Phases, LCGC North Am., 2018, 36 (3),
184–193.
10. J. Pellett, P. Lukulay, Y. Mao, W. Bowen, R. Reed, M. Mac, R. C. Mungerc,
J. W. Dolan, L. Wrisley, K. Medwide, N. P. Toltl, C. Chan, M. Skibic, K.
Biswas, K. A. Wells and L. R. Snyder, ‘Orthogonal’ Separations for
Reversed-Phase Liquid Chromatography, J. Chromatogr. A, 2006, 1101,
122–135.
11. S. Karmarkar, R. Garber, Y. Genchanok, S. George, X. Yang, and R.
Hammond, Quality by Design (QbD) Based Development of a Stability
Indicating HPLC Method for Drug and Impurities, J. Chromatogr. Sci.,
2011, 49, 439–446.
12. B. Debrus, D. Guillarme, S. Rudaz, Improved Quality-by-Design
Compliant Methodology for Method Development in Reversed-Phase
Liquid Chromatography, J. Pharm. Biomed. Anal., 2013, 84, 215– 223.
13. N. S. Wilson, M. D. Nelson, J. W. Dolan, L. R. Snyder and P. W. Carr,
Column Selectivity in Reversed-Phase Liquid Chromatography II:
Effect of a Change in Conditions, J. Chromatogr. A, 2002, 961,
195–215.
14. Webster (ed.), Supercritical Fluid Chromatography: Advances and
Applications in Pharmaceutical Analysis, G. K. Pan Stanford: Singapore,
2014.
 References 113

15. J. W. Caldwell, W. B. Caldwell, G. K. Webster and Z. Wang Method

Development for Achiral SFC, Chapter 4, in Supercritical Fluid
Chromatography—Advances and Applications in Pharmaceutical
Analysis, Pan Stanford Publishing, 2014.
16. E. Lemassona, S. Bertin, P. Hennigb, H. Boiteuxc, E. Leselliera, C. West,
Development of an Achiral Supercritical Fluid Chromatography
Method with Ultraviolet Absorbance and Mass Spectrometric
Detection for Impurity Profiling of Drug Candidates. Part II. Selection
of an Orthogonal Set of Stationary Phases, J. Chromatogr. A, 2015,
1408, 217–226.
17. M. B. Hicks, Y. Zhang, D. Moore, A. Apedo, Advancing SFC Methods
Development with a Multi-Column Supercritical Fluid Chromatography
with Gradient Screening, Am. Pharm. Rev., 2011 14(6), 52, 54, 56,
58–60. http://www.americanpharmaceuticalreview.com/Featured-
Articles/37308-Advancing-SFC-Method-Development-with-a-
Multi-Column-Supercritical-Fluid-Chromatography-with-Gradient-
Screening/.
18. E. Lemassona, S. Bertin, P. Hennigb, H. Boiteuxc, E. Leselliera, C. West,
Development of An Achiral Supercritical Fluid Chromatography
Method with Ultraviolet Absorbance and Mass Spectrometric
Detection for Impurity Profiling of Drug Candidates. Part I: Optimization
of Mobile Phase Composition, J. Chromatogr. A, 2015, 1408, 227–235.
19. West, C., Lesellier, E. A Unified Classification of Stationary Phases for
Packed Column Supercritical Fluid Chromatography, J. Chrom. A, 2008,
1191, 21–39.
20. West, C., Lesellier, E. Characterization of Stationary Phases in
Supercritical Fluid Chromatography with the Solvation Parameter
Model, Adv. Chromatogr., 2010, 48, 195–253.
21. West, C., Lesellier, E. Characterization of Stationary Phases in
Subcritical Fluid Chromatography with the Solvation Parameter
Model IV. J. Chrom. A, 2006, 1115, 233–245.
22. West, C., Lesellier, E. Characterization of Stationary Phases in
Subcritical Fluid Chromatography by the Solvation Parameter Model,
J. Chrom. A, 2006, 1110, 181–190.
23. West, C.; Lesellier, E. Characterisation of Stationary Phases in
Subcritical Fluid Chromatography with the Solvation Parameter
Model. III. Polar Stationary Phases, J. Chrom. A, 2006, 1110, 200–213.
24. West, C., Lesellier, E. Orthogonal Screening System of Columns
for Supercritical Fluid Chromatography, J. Chrom. A, 2008, 1203,
105–113.
114 Unified Approach to Chromatographic Method Development

25. A. Kumar, J. C. Heaton, D. V. McCalley, Practical Investigation of

the Factors that Affect the Selectivity in Hydrophilic Interaction
Chromatography, J. Chromatogr. A, 2013, 1276, 33–46.
26. A. Periat, B. Debrus, S. Rudaz, D. Guillarme, Screening of the Most
Relevant Parameters for Method Development in Ultra-High
Performance Hydrophilic Interaction Chromatography, J. Chromatogr.
A, 2013, 1282, 72–83.
27. D. V. McCalley, Study of Retention and Peak Shape in Hydrophilic
Interaction Chromatography Over a Wide pH Range, J. Chromatogr. A,
2015, 1411, 41–49.
28. A. Tumpa, A. Stajić, B. Jančić-Stojanović, M. Medenicaba, Quality
by Design in the Development of Hydrophilic Interaction Liquid
Chromatography Method with Gradient Elution for the Analysis of
Olanzapine, J. Pharm. Biomed. Anal., 2017, 134, 18–26.
Chapter 6

Polysaccharide-Derived Chiral
Stationary Phases for the Separation of
Enantiomers

Tong Zhang and Pilar Franco

Chiral Technologies Europe, Parc Innovation,
160 Bd. Gonthier d’Andernach, F-67400 Illkirch, France
[email protected]

6.1 Introduction
Based on the research work from Prof. Okamoto’s group [1–3],
polysaccharide-derived and silica-supported chiral stationary
phases were introduced in the 1980s. They are considered as the
first choice of chiral stationary phases (CSPs) for enantiomer
separations by liquid chromatography (LC) and supercritical fluid
chromatography (SFC). A considerable number of publications
over three decades have reported and reviewed the applications of
these CSPs by scientists working in the field of enantioseparations
[4–24].

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
116 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

Currently many polysaccharide‑derived CSPs are commercially

available, either as coated or immobilized phases, with particle
sizes ranging from 20 µm for preparative separations over 5 µm
for conventional method development to 3 µm and sub‑2 µm for
ultrafast analysis of enantiomers [13, 25, 26]. The most advanced
series includes the CSPs by the optimized immobilization
technologies: CHIRALPAK® IA, CHIRALPAK® IB, CHIRALPAK® IC,
CHIRALPAK® ID, CHIRALPAK® IE, CHIRALPAK® IF, CHIRALPAK®
IG and CHIRALPAK IH (Fig. 6.1). These immobilized CSPs combine
the benefits of polysaccharide-based chiral selectors—that is their
broad application scope and their preparative potential—with the
advantages of the immobilization process, which leads to strong
CSP robustness and practically unrestricted solvent compatibility.
The second series contains the classical coated CSPs,
bearing many different substitution patterns: CHIRALPAK® AD,
CHIRALPAK® AS, CHIRALPAK® AY, CHIRALPAK® AZ, CHIRALCEL®
OD, CHIRALCEL® OJ, CHIRALCEL® OZ, and CHIRALCEL® OX
among others (Fig. 6.2).
OR OR
O O
O
RO RO
OR n Amylose-based OR Cellulose-based
O n

CSP R CSP R

CH3 CH3
H tris(3,5-dimethyl- H tris(3,5-dimethyl-
CHIRALPAK IA N N
phenylcarbamate) CHIRALPAK IB phenylcarbamate)
O CH3 O CH3
Cl
H
CHIRALPAK ID N tris(3-chloro- H tris(3,5-dichloro-
phenylcarbamate) CHIRALPAK IC N
phenylcarbamate)
O Cl
O Cl
Cl
H tris(3,5-dichloro-
CHIRALPAK IE N
phenylcarbamate)
O Cl

H tris(3-chloro-4-methyl-
CHIRALPAK IF N CH3
phenylcarbamate)
O Cl
CH3
H tris(3-chloro-5-methyl-
CHIRALPAK IG N phenylcarbamate)
O Cl

H3C tris((S)-α-methyl-
CHIRALPAK IH O (S) benzylcarbamate)
N
H

Figure 6.1 Structures of the immobilized polysaccharide–derived chiral

selectors.

One essential feature of these polysaccharide derivatives is

their broad application domain. Accordingly, a judicious selection
 Introduction 117

of CSPs allows the design of effective screening kits with a limited

number of columns and a high success rate. For many years this
has been the rational choice of four coated CSPs as primary
screening tools in the research laboratories and in pharmaceutical
industry. Concerning the immobilized CSP selection, primary
screening kit will contain the selectors with the broadest application
domains. The remaining selectors would be included in a secondary
screening.
OR OR
O O
O
RO RO
OR Amylose-based OR Cellulose-based
O n n

CSP R CSP R

CH3 O
H tris(3,5-dimethyl- CHIRALCEL OA triacetate
CHIRALPAK AD N
phenylcarbamate) C
H3
O CH3 O

H3C tris((S)-α-methyl- CHIRALCEL OB tribenzoate

CHIRALPAK AS O
N
(S) benzylcarbamate)
H
H
H 3C N
CHIRALCEL OC tris(phenylcarbamate)
H tris(5-chloro-2-methyl-
CHIRALPAK AY N O
phenylcarbamate)
O Cl CH3
H tris(3,5-dimethyl-
CHIRALCEL OD N
H tris(3-chloro-4-methyl- phenylcarbamate)
CHIRALPAK AZ N CH 3
phenylcarbamate) O CH3
O Cl
H tris(4-chloro-
N Cl
CHIRALCEL OF phenylcarbamate)
O
H
N CH 3 tris(4-methyl-
CHIRALCEL OG phenylcarbamate)
O

CHIRALCEL OJ CH3 tris(4-methyl-

O benzoate)
O

CHIRALCEL OK tricinnamate

H
N Cl tris(4-chloro-3-methyl-
CHIRALCEL OX
O CH3 phenylcarbamate)

H
CHIRALCEL OZ N CH3 tris(3-chloro-4-methyl-
phenylcarbamate)
O Cl

Figure 6.2 Structures of the coated polysaccharide–derived chiral selectors.

In this chapter, we overview and compile the screening

strategies or approaches for the separations of enantiomers on
these types of chiral supports. The main issues to be addressed
include the configuration of the column set for sample screening,
the separation technique (HPLC or SFC) and the choice of mobile
phase in different separation modes, i.e., normal phase (NP) mode,
118 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

polar organic (PO) mode and reversed phase (RP) mode to achieve
efficient, fast, and appropriate method which is matched to the
requirements of the project and the molecule.

6.2 Column and Mobile Phase Selection

The standard screening protocols adopted by laboratories or
research groups worldwide usually use three or four analytical
columns in the primary screening. Two main options for setting
up the primary screening, with either the four traditional coated
columns (CHIRALPAK AD and AS, CHIRALCEL OD and OJ) or
alternatively the immobilized phases for enhanced applicability
and robustness [27–34], including CHIRALPAK IA, IB, IC, and IG.
Such configurations with polysaccharide‑derived columns, tested
on a randomly composed sample pool, and with four solvent
mixtures, can afford complete enantiomer resolution for over 90%
of the racemates examined and enantiorecognition above 95%.
Although most separations can be achieved with a reduced
number of columns and mobile phases, there remains a need for
unique selectivities that allow the resolution of certain challenging
isomers or their impurities. This is where the “secondary” column
set can step in. It should be noted that continuous research effort
over the last decades has been devoted to the development new
and powerful selectors showing enantiomeric recognition ability
complementary to the primary column sets used for screening. A
common and typical characteristic of these new CSPs are cellulose
and amylose derivatives bearing methyl- and chloro-substituted
phenylcarbamate groups, instead of the dimethyl- or dichloro-
present in the standard systems either in coated or immobilized
fashion (Figs. 6.1 and 6.2 for structural features). Some of these
materials were initially developed in 20 µm particle size and
involved in our CSP library application solely for preparative
purposes but following their high success rates for nearly a
decade, they have been introduced in the product line in 20, 5, 3,
and now sub‑2 µm formats for LC and SFC separations.
Method screening with polysaccharide-derived CSPs should
consider several practical aspects and has to be designed based on
the needs of each particular application. Polysaccharide-derived
 Column and Mobile Phase Selection 119

CSPs exhibit chiral recognition abilities in organic and aqueous

mobile phases in LC, as well as under supercritical fluid (SFC)
conditions. Their enantioselectivity, retentivity and resolution
degree are differently modulated in those modes, as solvent
eluting strength, solvation effect, supra-molecular structure of the
polymer and interaction mechanism (such as hydrogen bonding,
dipole–dipole interaction, p‑p stacking, steric hindrance) will
not be equivalent in all of them. Therefore, the choice of the
chromatographic mode will be an essential element in the
experimental design. Moreover, when setting up a screening
strategy, it is important to consider the separation scale (analytical
or semi-preparative or preparative) that has to be achieved, the
time constraints, the compatibility of mobile phase with the CSP
and compound as well as environmental restrictions.
On this basis, the present section will aim to give a general
overview of the main principles to be applied, although it is
recommended to adapt these recommendations to the individual
needs. Three main areas will be covered:
1. Use of organic mobile phases in HPLC
2. Use of water compatible mobile phases in HPLC
3. SFC conditions

6.2.1 Method Screening with Organic Mobile Phases in

HPLC
Given the increasing number of chiral compounds in drug
development and other domains, efficient and automated HPLC
screening using columns packed with various CSPs has been a
challenging subject and discussed by several authors [27–33, 35].
The strategies to achieve enantiomer separations are quite
variable from one laboratory to another depending on the type
of substances in screening, the equipment available and the R&D
background. However, the motivation for the development of
efficient screening approaches remains common. We all search
for the easiest, the fastest, the most successful and reliable way
to analyze a series of molecules with diverse structures and
chemical properties in a reasonable time frame. This is the
standpoint for chiral compound screening.
120 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

6.2.1.1 Immobilized CSPs under organic mobile phase

conditions
Since their introduction in the market in 2004, a certain number
of research teams have published and described the use of
immobilized polysaccharide CSPs in HPLC for new and diverse
applications involving enantioseparations [16–18, 25, 26, 32–34,
36–81]. As already mentioned, immobilized polysaccharide‑
derived CSPs have several benefits that should be considered in
the screening process, but some guidelines may be needed to
direct the choice of solvents from their very broad range to compose
the mobile phase [16, 17, 32–38].
Based on exhaustive investigations on immobilized columns,
it was concluded that, when used as the mobile phase or mobile
phase components, a group of solvents and their mixtures can
lead to better selectivity values for most of the compounds
tested. These organic conditions could be considered as a
primary set of mobile phases for screening with these CSPs.
Primary solvent configurations for the analytical screening
system are shown in Table 6.1. In practice, these four mobile
phase systems can be set up using a quaternary HPLC screening
system.
It is well known that the mobile phase may have major
effects on the enantiomeric separations. The chromatograms in
Fig. 6.3a demonstrate how significantly the enantiomeric
separation can be dependent on the mobile phase composition.
On a CHIRALPAK IA‑U column (1.7 µm, 100 × 3.0 mm i.d.), the
resolution of N-CBZ-DL-Ala enantiomers is partial using the
mobile phase of hexane/EtOH/FA 80/20/0.1 (FA: formic acid). The
simple replacement of EtOH with isopropyl alcohol (2-PrOH) at
the same volume percentage leads to a much larger separation of
the same enantiomers. Figure 6.3b shows a case of reversal
of elution order of enantiomers on the same CHIRALPAK IA‑U
column, induced by the change from 2-PrOH to EtOH as the
polar mobile phase component.
Chlorinated solvents are usually submitted to a more rigorous
control due to environmental restrictions. Nevertheless, their use
at analytical level may be allowed in most laboratories, provided
that their waste residues are treated appropriately. If the use of
 Column and Mobile Phase Selection 121

dichloromethane (DCM) is indeed constrained, tetrahydrofuran

(THF)‑based mixtures could be implemented instead.

N-CBZ-DL-Alanine N-Fmoc-Phenylalanine (D/L 35:65)

O
H O HO
O N
OH O O
O N
H
L

(a-1) (b-1)
Hexane/2-PrOH/FA
D
Hexane/2-PrOH/FA
80/20/0.1 v/v/v 80/20/0.1 v/v/v
k1 = 1.33 k1 = 1.82
ɲ = 1.63 ɲ = 1.1
Rs = 6.9

(a-2) Hexane/EtOH/FA (b-2)

80/20/0.1 v/v/v Hexane/EtOH/FA
k1 = 1.29 80/20/0.1 v/v/v
D
ɲ = 1.08 k1= 1.89
Rs = 1.2 ɲ = 1.21
Rs = 2.5

0 1.0 2.0 3.0 4.0 0 1 2 3 4 5

Minutes Minutes

Figure 6.3 Effect of mobile phase on chiral separations. Column:

CHIRALPAK IA-U (1.7 µm, 100 × 3.0 mm i.d.); temperature: 25°C; flow
rate: 0.425 mL/min.

If additional screening of mobile phases was required, the

use of the following mixtures as a secondary set is recommended:
heptane or hexane/ethyl acetate EtOAc, heptane or hexane/THF,
acetonitrile (ACN) and methanol (MeOH) (Table 6.1).
Different alcohols, with shorter or longer aliphatic chains, e.g.,
EtOH, 1‑propanol (1‑PrOH) and 2‑PrOH, can play a “dual” role to
mix either with non-polar alkanes or with polar solvents (ACN,
MeOH), though alkane is immiscible to ACN and MeOH. However,
the solvent nature varies little; hence, the choice stays quite
limited if the viscosity factor is considered. In contrast, solvents
such as THF, EtOAc, MtBE, DCM, etc., belong to different solvent
families with distinct properties. Unlike heptane, methanol, and
acetonitrile, those solvents are all of mid-polarity and miscible
122 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

with both non-polar and polar organic solvents. This feature is

beneficial for fine-tuning composition of mobile phase, as it allows
easy adjustment in retention, selectivity and resolution.

Table 6.1 Mobile phases for sample screening with immobilized CSPs

Starting
condition Typical optimization
Solvent mixture (%) range (%)
Primary Alkane/EtOH 80/20 99/1 to 50/50
mobile phases Alkane/2-PrOH 80/20 99/1 to 50/50
Alkane/DCM/EtOH 50/50/2 85/15/0 to 0/80/20
Alkane/MtBE/EtOH 0/98/2 80/20/0 to 0/40/60
Secondary Alkane/EtOAc 50/50 80/20
mobile phases Alkane/THF 70/30 95/5 to 0/100
MeOH 100
ACN 100

The proposed mobile phase systems were selected to maximize

complementary behavior among the CSPs of the set and also
between solvents on the same column. It is important to note that
elution order of enantiomers may change in certain cases due
to the mobile phase (Fig. 6.3b) and due to the different selectors.
This phenomenon can also be observed for separation of certain
target sample and its impurities. Such systematic screening can be
a useful tool to choose the most suitable CSP-mobile phase
combination for the specific separation aims.
The first objective is to set the right conditions to start the
screening, leading to appropriate retention times and successful
chiral separations. As an example, and based on an important
number of experimental data, the proportions indicated in
Table 6.1 can be taken as starting point and possibly adjusted
according to the first results and the needs for optimization.
The retention factor of the compound will steer the adjustment of
the mobile phase for eluting strength. With these first data, one can
decide the best solvent(s) or solvent ratios to achieve the desired
separation.
 Column and Mobile Phase Selection 123

EtOAc, MtBE, THF, DCM, and chloroform (CHCl3) can

be used as mobile phases in their undiluted form or even
combined with alcohols. However, neat THF appears to be the
mobile phase with the highest eluting strength among them. Its
use in its pure form often leads to short or no retention of solutes,
although we have already seen certain exceptional resolutions in
THF/alcohol mixtures.
If an exhaustive investigation is to be carried out with a given
compound, or the results from the screening with the first set of
mobile phases are not satisfactory, it may be worth considering
expanding the solvent range. For example, acetone, 1,4‑dioxane,
toluene, methylal (dimethoxymethane) or ethylal (diethoxymethane)
have proved to be useful as mobile phase components for successful
chiral separations in certain cases.
It is important to point out that the conditions described in
Table 6.1 are directly applicable for neutral (bearing no charges)
compounds. If acidic or basic racemates have to be separated
into enantiomers, addition of an acidic or basic additive to these
mobile phases may be necessary, although the same principles
as described for the bulk solvent composition will remain valid.
The role and particular aspects of the additives will be discussed
later in this chapter.
No particular advice needs to be given for the transition
between different organic mobile phases, as long as miscibility
issues are taken into account. However, when switching from very
different mobile phase conditions, equilibration times have to be
adapted (e.g., at least 30 min at 1 mL/min for analytical columns
sized 150 × 4.6 mm i.d.) in order to ensure method reproducibility
as a result of complete equilibration of the column.
Taking into account the different nature and proportion of the
solvents described earlier, the use of a UV diode array detector may
not be highly recommended. Certain mobile phases containing
toluene, acetone and high proportions of DCM or EtOAc may
cause UV detection problems due to their high UV cut-off. In this
case, hyphenation of other type of detectors may be necessary,
such as CAD (charged aerosol detector), ELSD (Evaporative light
scattering detector), polarimeter, CD (circular dichroism), RI
(refractive index) or MS (mass spectrometry) detector.
124 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

The use of gradients could also be envisaged to select appropriate

solvent composition in terms of retention and selectivity [27]. It
would then be important to make sure that the HPLC system in use
is able to guarantee proper and accurate metering and mixing of
different mobile phase components, as this is not always the case
with DCM, EtOAc or THF mixtures.
Sample solubility in the mobile phase, enantioselectivity
and loading capacity are key factors in chiral preparative
chromatography to render a preparative separation practically
feasible and/or economically attractive. However, sample solubility
and stability in the injection solvent are also important at analytical
level. Such a limitation may be circumvented by using immobilized
CSPs. In contrast to the coated CSPs, the immobilized ones allow
method development directly guided by sample solubility with
the possibility to improve the enantioselectivity and separation.
From the method development viewpoint, a compromise of sample
solubility, enantioselectivity, resolution and retention time is
often needed.

6.2.1.2 Coated CSPs under organic mobile phase conditions

The approach of method development and optimization on the
coated CSPs can be based on a similar strategy to the one used
for the immobilized‑type phases. However, the choice of solvents
must be restricted to those being compatible with these supports.
This means that all solvents or solvent combinations that could
potentially dissolve, even partially, the polysaccharide derivative
must be avoided in the mobile phase. The appropriate solvent
set for screening will be then composed of alkane/EtOH, alkane/
2‑PrOH, alcohol‑based and ACN‑based mixtures (Table 6.2).
In principle, these mobile phase systems can be set up using the
same HPLC screening system. However, it is recommended that
ACN or its mixtures with alcohols will be carefully managed,
preferably separately, in order to avoid any potential accidental
wrong mixing between alkane and ACN. If such an event happens,
the coated columns could be irreversibly damaged. Obviously, a
good column flushing with a transition solvent having a broad
miscibility (e.g., EtOH [82]) must be inserted between the two
mobile phases which are not miscible with each other.
 Column and Mobile Phase Selection 125

Table 6.2 Mobile phases for sample screening with coated CSPs

Starting condition Typical optimization range

Solvent mixture (%) (%)
Alkane/EtOH 80/20 99/1 to 50/50
Alkane/2-PrOH 80/20 99/1 to 50/50
MeOH 100
ACN 100

6.2.2 Method Screening with Aqueous or Water

Compatible Mobile Phases in HPLC
The polysaccharide type CSPs seem to be most frequently used
with organic eluents (alkane-alcohol mixtures or the polar
solvents discussed above). However, resolution of enantiomers on
these columns by aqueous eluents has a history almost as long as
their applications with organic mobile phases [11]. Method
development in RP‑mode on polysaccharide‑derived CSPs has
been discussed in several articles [11, 83–85] and some new
applications have been recently developed [48, 54, 65].
The choice of RP‑mode is often dictated by the need for
direct injection of the samples of biological extracts. It is also an
alternative to certain compounds with low retention, or limited
solubility or lack of enantioselectivity with organic mobile phases.
An additional and increasing interest in RP methods is their
undeniable suitability for LC‑MS applications.
The chiral recognition mechanism of polysaccharides in the
presence of water differs from the one observed in organic mobile
phase. The solvation effect on the stationary phase induced by water
may be strong and will modify the retention and enantiorecognition
mechanisms. Therefore, reverse phase chromatography is often
complementary to normal phase separation mode and offers the
possibility of enhancing the resolution success.

6.2.2.1 Immobilized CSPs in aqueous or water compatible

solvent mixtures
As a logical sequence of the recent investigations with organic
mobile phases in HPLC and SFC, an intensive study on the
126 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

chromatographic behavior of the immobilized CSPs towards

various mobile phase systems was undertaken in our laboratories.
This has allowed the development of simple and straightforward
strategies for efficient method screening [18, 32, 33].
The nature of the molecules to be analyzed should be
considered for mobile phase selection. Under this primary
guideline, a “decision tree” can be defined to start the method
development (Fig. 6.4). The selected mobile phases do not really
differ from the proposals for achiral reverse phase screening.

Acidic compounds Neutral compounds Basic compounds

Aqueous solution pH 2 Water Aqueous solution pH 9

(FA) (20 mM NH4HCO3 - DEA)
+ + +
40% ACN 40% ACN 40% ACN
or or or
60% MeOH 60% MeOH 60% MeOH

Figure 6.4 HPLC screening strategy on polysaccharide-derived columns

using aqueous mobile phase.

The selection of the mobile phase for neutral compounds is

straightforward: plain water is suitable to be used as the aqueous
component. The most common and effective pH value of the
aqueous solution for the resolution of acidic molecules is in the
range 2.0–2.5 in which the ionization of most organic acids will
be suppressed. FA is a good choice where LC‑MS compatibility is
required while other acids (trifluoroacetic acid (TFA), acetic acid,
H3PO4) may be valuable for other applications [18, 33]. Based
on the same reasoning, basic aqueous solutions should normally
be in use to suppress the ionization of basic solutes. For this
purpose, certain buffers or aqueous media (phosphates, borates,
ammonium acetate or KPF6 aqueous solutions) at basic pH may
also be useful for enantiorecognition [18]. Where RP methods
suitable for MS detection for basic compounds are required,
ammonium bicarbonate solution adjusted to pH 9 is a good option.
The organic modifier plays an important role in regulating
retention and modulating enantioselectivity. The first choice of
 Column and Mobile Phase Selection 127

modifier to be combined with the aqueous solution is ACN and

MeOH. In this context, the RP‑mode could be sometimes considered
as an extension of the PO-mode when retention is not sufficient in
the absence of water.
MeOH can be a useful modifier in RP-mode but with weaker
eluting strength and leading to much more viscous aqueous mobile
phases. ACN has the advantages of low viscosity, low UV cut‑off,
adapted eluting strength on the polysaccharide‑type CSPs and the
outstanding ability to induce good enantioselectivity. The use of
other organic modifiers, such as EtOH, 1‑PrOH and 2‑PrOH can
also be considered. In our practice, however, these modifiers would
be tried only if no successful hit is observed with ACN and MeOH
in the mobile phase.
Additionally, and specifically to the immobilized supports,
THF can also be employed to compose the RP mobile phase. It may
sometimes be a good alternative to ACN and MeOH, but in a general
way it seems to be a less versatile organic modifier. The relative
eluting strength of these three organic modifiers can be ranked as
follows: THF > ACN > MeOH. Moreover, in case of need, sample can
be dissolved in DMSO, DMF or acetone for injection on the columns
packed with the immobilized CSPs.
As in NP-mode, the key point for RP method development
would be setting the right conditions to start the sample screening.
Based on a significant number of experimental data, the mobile
phase compositions indicated in Fig. 6.4 can be recommended as
the starting point. The typical starting conditions represent
actually the mobile phases of upper-middle eluting strength
(containing enough organic modifier and enabling elution of most
of the compounds in a reasonable analysis time frame). In many
cases, successful separations can be directly obtained issuing the
screening procedure. However, certain compounds or families of
compounds may be either extremely retained or too fast eluted
in these conditions. In this case, an optimization step should be
considered for the individual samples, by playing the nature and
the percentage of the organic modifiers. Gradient elution can
certainly be used for sample and mobile phase screening [86],
but isocratic elution mode seems to be more convenient as for
method reproducibility or robustness and for straightforward
system-to-system or lab-to-lab method transfer.
128 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

Another interesting feature of immobilized

polysaccharide‑derived supports is the possibility to use the
“strong” solvents (such as THF and DMF) for column flushing/
cleaning and for efficient transition from RP- to NP-mode and
vice versa. The flushing protocol is described in detail on the
instruction manual of the immobilized columns (discussed in
Section 6.7 in this chapter) but will not be compatible with the
previously existing coated CSPs.
It is worth mentioning the recent reports about chiral
separations carried out on the immobilized CSPs by incorporating
a low percentage of water (<40% in volume) in ACN or alcohol
(MeOH or EtOH) as mobile phase [87–90]. In these studies, HILIC
(hydrophilic interaction liquid chromatography) or HILIC-like mode
is suggested for effective retention and enantiomer separation of
certain polar and hydrophilic compounds.

6.2.2.2 Coated CSPs in aqueous or water compatible solvent

mixtures
Method development in RP-mode in the case of the coated supports
can follow exactly the same strategy as described in Fig. 6.4.
However, apart from alcohols and acetonitrile, no other alternative
organic solvents can be used, as they may irreversibly damage
the columns.

6.2.3 Mobile Phases in SFC

SFC applications in the separation of enantiomers are a growing
field [91–95] and many applications using polysaccharide-based
columns have been described [17, 33, 34, 40, 96–113].

6.2.3.1 Immobilized CSPs in SFC

Experiments to date with polysaccharide‑derived CSPs show
that success rates can be achieved by using alcohol co‑solvents
[17, 34, 96–104]. In particular, MeOH, EtOH, and 2‑PrOH are very
good initial co‑solvent choices (Fig. 6.5). ACN seems to be less
successful, but it may be an option for a secondary screening.
If the elution of the analytes is not fast enough when ACN is used,
an alcohol can be added to the CO2/ACN mixture. The screening
 Column and Mobile Phase Selection 129

can be performed in isocratic and gradient mode [33, 101,

103, 104].

Sample nature?

Choose mobile phase additive

CO2 / EtOH CO2 / MeOH CO2 / 2-PrOH

80/20 v/v 80/20 v/v 80/20 v/v

Figure 6.5 Primary screening strategy on polysaccharide-derived columns

in SFC mode.

To resolve basic compounds, add a small amount (normally

0.5–1% in the co‑solvent) of a basic additive (typically isopro-
pylamine (iPAm), diethylamine (DEA) or trimethylamine (TEA)).
For acidic compounds, an acidic additive (i.e., TFA, FA or acetic acid)
may be helpful for better resolution, although the acidity of CO2
could in certain cases be sufficient for good enantiomer resolution
acidic compounds [104].
Moreover, immobilized polysaccharide-derived CSPs are
compatible with a much wider range of solvents than are the
coated phases [114]. The use of such “extended range” solvents
was investigated for SFC screening in a similar manner as for LC
separations. Initially these solvents were used to enhance sample
solubility, to solve difficult separations or to process compounds
being unstable in alcohols. Subsequently we have found that
these solvents can also offer exceptional enantioselectivity profiles.
THF, MtBE, DCM, or EtOAc, for example, can be used as the CO2
modifier. In some cases, pure DCM, MtBE and EtOAc may not be
strong enough to elute certain compounds. In these cases, the
addition of small percentages of an alcohol mixed with the co-
solvent is advised (Fig. 6.6) [17, 33, 103].
As already mentioned, success rate is already very high with
the alcohol modifiers. Therefore, it has become established practice
to use either the immobilized columns or alternatively the coated
ones—in the column set for screening with the alcohol modifiers
130 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

as first strategy. Then the solvent range can be broadened on the

immobilized columns if no satisfactory separation is achieved
or sample solubility/stability becomes an issue. SFC and LC are
complementary techniques and comparable success rates are
achieved with the same set of columns. Nevertheless, recognition
patterns are sometimes different and leading to different retentive
and selectivity profiles. These two factors plus the likely sample
solubility issues will be determinant in the choice of one technique
or the other for separation of a given compound.

Samplenature?

Choosemobilephaseadditive

CO2 /ACN CO2 /THF CO2 /(DCM/MeOH90/10) CO2 /(MtBE/MeOH80/20) CO2 /(EtOAc/MeOH90/10)
70/30v/v 75/25v/v 80/20v/v 75/25v/v 80/20v/v

Figure 6.6 Secondary screening strategy on polysaccharide-derived

columns in SFC mode (only compatible with immobilized CSPs, except
ACN).

Analytes are preferably dissolved in the solvent mixture that

is used as the co‑solvent in SFC. If this is not possible, choose a
better sample diluent (e.g., ethanol) and monitor any potential
perturbance of the baseline due to the injection solvent. For sample
insoluble in alcohols or ACN, it is possible to dissolve these in
DCM or DCM mixtures for injection when using immobilized
phases. In such cases, caution should be taken to avoid on-line
precipitation of the compounds, but this should not be done with
coated phases without compromising life‑time of the column.

6.2.3.2 Coated CSPs in SFC

Coated CSPs are only compatible with alcohols and acetonitrile
mixtures in SFC. As discussed in the previous section, normally the
screening starts with alcohol screening and acetonitrile would be
only screened as a secondary option, either in pure form or mixed
with alcohols to enhance elution.
 The Role of Additives 131

6.3 The Role of Additives

For basic and acidic samples, it may be necessary to incorporate
an additive in the mobile phase in order to get reasonable elution
of the compounds with good peak shape and optimum resolution
degree. Thus, basic samples may require a basic additive (DEA,
butylamine (BA), 2-aminoethanol (AE), ethylenediamine (EDA))
and acidic compounds the addition of an acid (TFA, FA, or acetic
acid). The percentage needed is typically 0.1% (in volume) and
should normally not exceed 0.5% (see examples described in
the following sections) [16, 17, 33]. In SFC this amount is also
respected, although as added in the co‑solvent it may be 0.5–
1% to keep the overall percentage <0.5% in the final mobile
phase. In RP‑conditions, buffers will be used as discussed in the
corresponding section [18].
It has been found that certain amines, such as EDA and AE
induce much better behavior for certain compounds than the more
commonly used DEA. Resolution degree and peak symmetry can be
dramatically improved with this type of additives, as exemplified
in Fig. 6.7. For practical purposes, as EDA and AE may have
limited miscibility in certain organic solvents in the absence
of alcohol, it is better running screenings with DEA, but it will
be important in the optimization step to consider this fact.

OH
H
O N Propranolol

EDA 0.1% AE 0.1% DEA

0.3%
0.1%

0 2 4 6 8 10 12 min 0 2 4 6 8 10 12 min 0 2 4 6 8 10 12 min

Figure 6.7 Additive effect on enantiomer resolution. Column: CHIRALPAK

IB (5 µm, 250 × 4.6 mm i.d.); mobile phase: hexane/2‑PrOH/additive,
90/10/0.1 (or 0.3%) v/v/v; flow rate: 1.0 mL/min; temperature: 25°C.
132 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

6.4 Column Temperature

Column temperature is a critical factor impacting the thermodynamic
process and determining the kinetic aspects of separations by
chromatography. This is also true for enantiomer separation using
CSPs. The column temperature can induce multiple effects: it can
affect retention, enantioselectivity, column efficiency, resolution
degree, column backpressure, analysis speed, sample, and column
stabilities and even elution order of the enantiomers [115–124].
For CSPs based on polysaccharide derivatives, temperature could
also alter the conformation of the polymeric chiral selectors and
trigger different interaction mechanisms between the enantiomers
and the chiral selector.
The effect of column temperature on chromatographic
retention and selectivity can be expressed by van’t Hoff equations:

ln k = –DH°/RT + DS°/R + ln Ø (6.1)

ln a = –D(DH°)/RT + D(DS°)/R, (6.2)

where DH° and DS ° are, respectively, the standard enthalpy and

entropy changes for the analyte molecule transfer from mobile
phase to stationary phase; Ø is the phase ratio defined as the
ratio of the volume occupied by the stationary phase to the volume
of the mobile phase contained in a column; D(ΔH°) and D(DS°)
are, respectively the differences in the changes in standard
enthalpy and entropy of the enantiomers in separation; R is the gas
constant; T is the absolute temperature expressed in K (Kelvin).
The plots of ln k and ln a versus 1/T will reveal the mechanistic
information of the retention and enantiomer separation. The case
study below demonstrates the effects of temperature and the
approaches for method optimization.
Racemic Praziquantel was well resolved into enantiomers
on CHIRALPAK IA (5 µm, 250 × 4.6 mm i.d.) using a mobile phase
composed of n‑heptane/EtOH 45/55 v/v at 1.0 mL/min (Fig. 6.8).
The temperature was varied in the range of 25–50°C in 5°C
increments. As shown in Fig. 6.8c–d, van’t Hoff plots of the capacity
factors and selectivity were linear in the temperature range
of 50°C down to 30°C (1000/T = 3.09 to 3.30 K–1) but started to
 Column Temperature 133

bend at around 1000/T = 3.3 (30°C). As defined by van’t Hoff

equations, enantioselectivity is the balanced parameter between
differences in enantiomeric binding enthalpy and disruptive
entropic effects. A transition point breaking the linear relationship
between ln k or ln a and I/T in van’t Hoff plot indicates an
evolution of the retention and/or selective mechanisms governing
the separation. Very probably, a change in the conformation
of the chiral selector occurred under the given mobile phase
condition at approximately 30°C (or 3.30 1/°K), inducing the
mechanistic modification in the case of Praziquantel. Typically,
the conformational change of the chiral selector is thermo-
dynamically reversible if T is kept in the reasonable range,
ensuring the good reproducibility of the separation under
given chromatographic conditions. As an overall trend, the
enantioselectivity is slightly increasing with the decrease in
temperature. This is actually the scenario observed in enantiomer
resolution of Praziquantel on CHIRALPAK IA.
As far as the kinetic aspect is concerned, the increase in
temperature normally accelerates the mass transfer process,
leading to enhanced column efficiency (Fig. 6.8e) and sharp
shortening of the analysis time (Fig. 6.8f). As a balance of the
temperature effect on column efficiency, on enantioselectivity and
retention factor, the resolution degree of Praziquantel enantiomers
followed a bell-shaped curve against temperature and reached
the highest Rs value at 40°C (Fig. 6.8f).
The experimental evidence suggests that adjustment and
control of the temperature may be a very effective tool for optimizing
chiral separations, although the achievement of the maximized
resolution degree for a given pair of enantiomers would need
optimization efforts through a systematic study on temperature
effect. When combined with other parameters such as the reduced
particle size of the CSP, reduced column size and high flow rate,
the increase in temperature can accelerate the analysis to sub-
minute (fast or ultrafast analysis). This is the case of enantiomer
resolution of Indapamide on a short CHIRALCEL OD‑3 column
(50 × 4.6 mm i.d.) packed with 3 µm CSP particles. As shown in
Fig. 6.9a, the enantiomers were well resolved within 30 s of time
at 25°C with MeOH at 3 mL/min. By increasing the temperature
to 40°C, the analysis time could be shortened even further to about
20 s. The chromatograms in Fig. 6.9b illustrate the temperature
 134 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

effect on peak shape improvement. Raising the temperature

from 25 to 40°C led to drastic peak narrowing and sharpening of
Mephobarbital on CHIRALPAK AD‑3, most remarkably for the
second eluting enantiomer. In this case, the temperature increase
brought forth multiple benefits: ultra-fast analysis, enhanced
resolution degree (Rs from 4.7 up to 7.7) without mentioning the
potential gain in detection sensitivity (or signal-to-noise ratio) of
the compound.

(a) T = 25°C, Rs = 8.8

O (b) T = 50°C, Rs = 9.6
N (a) T = 25°C, (a)
Rs T==
8.8
25°C, Rs = 8.8
O O (b) T = 50°C, (b)
Rs T==
9.6
50°C, Rs = 9.6

N
N O N

N O N O

Praziquantel
Praziquantel Praziquantel

0
0 5 0
5 10 5
10
15 10 20 15
15 25 20
20 30 25
25
PLQ 30
30
PLQ
PLQ

2.5 2.5 0.80 0.80

2.5 ln k1 ln k1 0.80
0.78 0.78
ln k1
2.0 2.0ln k2 ln k2 0.76 0.76 0.78

2.0 ln k2 0.74 0.74 0.76

1.5 1.5 0.72 0.72
0.74
ln D

ln D
lnk

lnk

0.70 0.70
1.5 0.72
1.0 1.0 0.68 0.68
ln D
lnk

0.66 0.66 0.70

0.5 0.5 0.64 0.64

1.0 0.68
(c) (c) (d) (d)
0.62 0.62
0.66
0.0 0.0 0.60 0.60
0.5 3.05 3.10 3.053.15 3.103.20 3.153.25 3.203.30 3.25
3.35 3.30
3.40 3.35 3.40 3.05 3.10 3.053.15 3.10
3.20 3.15
3.25 3.203.30 3.25
3.35 3.30
3.40 3.35 3.40
0.64
1000/T (1/°K)1000/T (1/°K) (c) 1000/T (1/°K)1000/T (1/°K) (d)
0.62
0.0 9000 9000 11.0 11.0 0.60 35 35
3.05 3.10 N3.15 N1
3.20 3.25 3.30 3.35 3.40 3.05 3.10 3.15
8000 8000
1
Analysis 3.20 3.25time
time Analysis 3.30 3.35
30 30
N2 N2
1000/T (1/°K) 10.5 10.5
Rs Rs
1000/T (1/°K)
Anallysis time (min)

Anallysis time (min)

7000 7000
25 25
6000 6000 10.0 10.0
Plate count

Plate count

9000 5000 5000 11.0 20 20

N1
Rs

Rs

9.5 9.5
8000 4000 4000 15 Analysis
15 time
3000 N
3000
2 9.0 9.010.5 Rs
7000 10 10
2000 2000
8.5 8.5
6000 ((e)) ((e)) 10.0 ((f)) 5 ((f)) 5
1000 1000
Plate count

5000 0 0 8.0 8.0 0 0

20 25 20 30 25 35 30 40 35 45 40 50 45 55 50 55
Rs

20 25 9.530
20 25 35 30 40 35 45 40 50 45 55 50 55
4000 Temperature Temperature
(°C) (°C) Temperature Temperature
(°C) (°C)
3000 9.0

2000 Figure 6.8 Effect of temperature on enantiomer separation of

8.5
((e))
Praziquantel. Column: CHIRALPAK IA (5 µm, 250 × 4.6 mm i.d.); ((f))
1000

0 mobile phase: n‑heptane/EtOH 45/55 v/v; flow rate:

8.0 1.0 mL/min.
20 25 30 35 40 45 50 55
20 25 30 35 40 45 50
Temperature (°C) Temperature (°C)
 Column Temperature 135

15.0 s
8.1 s (a-2)
Indapamide ((a-1)
T = 25°C T = 40°C
O R = 3.2
Rs Rs = 2.7
O O
S
N
N NH2
H 20 4 s
20.4
Cl

26.4 s
CHIRALCEL OD-3
(3 —m, 50 x 4.6 mm i.d.)

MeOH 3.0 mL/min

0 30 60 0 30 60
Retention time (secon
nd) Retention time (second)

10.8 s 9.6 s

( -1))
(b ( )
(b-2)
O CH3 T = 25°C T = 40°C
Rss = 4.7 Rs = 7.7
NH

O N O
CH3

Mephobarbital

27.6 s
CHIRALPAK AD-3
(3 —m, 50 x 4.6 mm i.d.) 38.4 s

MeOH 5.0 mlLmin

0 30 60 0 30 60

Retention time (second) Retention time (second)

Figure 6.9 Method optimization by temperature.

In practice, the use of temperature higher than 50°C is not

advised in a general way with the polysaccharide-based columns
for chiral separations. From one side, raising the temperature up
to 40°C can usually fulfill the purpose of method optimization
without causing on-line sample degradation. On the other side,
the frequent use of some very volatile solvents (e.g., hexane) or
solvents of low boiling points (e.g., DCM and MtBE) in the mobile
phase will create inconvenient chromatographic conditions and
working environment if T > 40°C.
It should be noted that temperatures below ambient, typically
10–15°C, may also be envisaged in an attempt to reach good
enough separation of some difficult compounds by possibly
increase the enantioselectivity. This may be also a good option
when the thermal stability issue is raised for the sample to be
analyzed. As consequences of lowing temperature, one should
expect higher viscosity of the mobile phase, higher column
back pressure, longer retention times and reduced column
136 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

efficiency which may counter balance the gain in selectivity for

the separation.

6.5 Effects of Flow Rate and Particle Size

In rate theory of chromatography, flow rate is a variable affecting
the longitudinal diffusion and mass transfer process of the analyte
molecules [125], therefore the column efficiency (represented
by the plate count N) for a given separation on a column of given
dimension. Its role in column efficiency is well defined by Van
Deemter equation:

HETP = A + B/µ + Cµ (6.3)

HETP refers to the height equivalent to the theoretical plate

of the column. The smaller the HETP, the higher the column
efficiency, µ is the linear velocity of the mobile phase which can
be determined by the division of flow rate over the column cross-
sectional area. Coefficient A is the multi-path term for packed
columns in which the analyte molecules move through different
paths. The contribution of this term to chromatographic band
spreading can be reduced if the column packing is improved
and/or the column is packed with smaller particles. B is the
longitudinal diffusion term linked to the analyte diffusion from
the center to the edges of the chromatographic band. Increasing
the linear velocity will normally attenuate the longitudinal
diffusion of the analyte molecule as a consequence of its shorter
stay in the column. The C‑term in Van Deemter equation is related
to the mass transfer kinetics, more precisely to the speed of
adsorption-desorption process of the analyte towards the
stationary phase. Too high linear velocity will enforce the analyte
to elute too fast from the column without reaching or approaching
the mass transfer equilibrium. Thus, the higher the velocity, the
severer the band broadening becomes.
Figure 6.10 depicts the Van Deemter plots of trans-
Stilbene oxide on the pair of columns CHIRALPAK AD‑H (particle
size: 5 µm) and CHIRALPAK AD‑3 (particle size: 3 µm) in NP-
mode, as well as those of 5,5-Diphenyl-4-methyl-2-oxazolidinone on
CHIRALPAK IB (particle size: 5 µm) and CHIRALPAK IB‑3 (particle
 Effects of Flow Rate and Particle Size 137

size: 3 µm) in RP-conditions. They were traced with the data of

the first eluting peak, but similar scenarios were obtained for the
second eluting peak. These plots confirm the clear trend that
columns packed with smaller particles afford higher column
efficiency (smaller A‑term, represented by the lower position of
the curves) and have more favorable mass transfer kinetics
(the reduced C‑term, represented by the flatter slope of the curves
in the extended linear area).
25 21

(a) 19 (b)

20 17

15
HETP (—m)
HETP (—m)

15 13

11

10 9

5 5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 0.0 0.5 1.0 1.5 2.0
— (mm/s) — (mm/s)
H3C
H
N
CHIRALPAK AD-H (5 —m) O CHIRALPAK IB (5 —m)
O O
CHIRALPAK AD-3 (3 —m) CHIRALPAK IB-3 (3 —m)

trans-Stilbene oxide 5,5-Diphenyl-4-methyl-

2-oxazolidinone

Figure 6.10 Van Deemter plots: Column conditions: T = 25°C; column

size: 150 × 4.6 mm i.d; mobile phase: (a) hexane/2-PrOH 90/10 v/v,
(b) ACN/H2O 40/60 v/v.

These are characteristics of columns packed with particles of

small size and offer significant benefits of running the analysis at
high velocity without compromising significantly the column
efficiency [25]. Figure 6.11 shows the enantiomer separation of
trans-Stilbene oxide on 50 × 3 mm i.d. columns packed, respectively
with 5, 3, and 1.6 µm CSPs containing the same chiral selector:
immobilized amylose tris(3,5‑dimethylphenylcarbamate). It can
be observed that, in comparison with the 5 µm CHIRALPAK IA
column, (ultra) faster analysis could be obtained over the 3 µm
(CHIRALPAK IA‑3) to sub‑2 µm (CHIRALPAK IA‑U) by increasing
the flow rate (or µ) without compromising the resolution degree
of the enantiomers. Additional examples of sub‑1 min separations
138 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

CHIRALPAKIAͲU(subͲ2ђђm)
1.5mL/min(ђ=3.54mm
m/s)
Rs =10.0

30s
CHIRALPAKIAͲ3(3ђm)
0.6mL/min(ђ=1.42mm/s)
Rs =10.6

30s 60s 90s

CHIRALPAKIA(5ђm)
0.4mL/min(ђ=0.94mm/s)
Rs =8.8

30s 60s 90s 120s

ond)
Retention time (seco

Figure 6.11 Effect of particle size: Solute: trans-Stilbene oxide; column

size: 50 × 3 mm i.d; mobile phase: hexane/2-PrOH 90/10 v/v; temperature:
25°C; instrument: UHPLC (X‑LC, Jasco).

are shown in Fig. 6.12. It has to be noted that the instrument void
should be optimized when working with columns packed with
small CSP particles in order to maximize the gain in resolution
afforded by the high column efficiency. For this purpose, ultra-
high-performance of liquid chromatography (UHPLC) system is
strongly recommended to run separations with sub‑2 µm packed
columns.
For optimization of analytic method, the adjustment of
linear velocity (or the flow rate on a column with a given internal
diameter) may be helpful in search of the optimal column
efficiency, therefore the improved resolution degree. This is
particularly critical in the attempt to transfer a partial resolution
to a full one. Such an approach is normally more effective with
CSPs of larger particles than smaller particles due to the steeper
slope of the linear region in Van Deemter plot for the larger particle
size. As demonstrated in Fig. 6.13 for enantiomer resolution of
 Effects of Flow Rate and Particle Size 139

N OH
H O HO
O N
O 2N N CH3 O O
O
Cl O H O
H
OH O
Oxp
prenolol OH
Ornidazole
2,3-Dibenzoyl-
y
DL-tartaric acid

(a) (c)
(b)

0s 12s 24s 36s 48s 60s 0s 12s 24s 36s 48s 60s 0s 12s 24s 36s 48s 60s

Retention time (second) Retention time

e (second) Retention time (second)

Figure 6.12 Fast analysis at high flow rate on 3 µm‑packed columns.

Conditions: Flow rate = 5.0 mL/min; T = 25°C; column size: 4.6 × 50
mm; particle size: 3 µm; instrument: UPLC (I‑class, Waters); column
and mobile phase: (a) CHIRALPAK IA‑3, DCM/MeOH/2‑AE 98/2/0.1 v/v/v;
(b) CHIRALPAK IB‑3, hexane/THF/2‑AE 70/30/0.1 v/v/v; (c) CHIRALPAK
IC‑3, hexane/2‑PrOH/TFA 80/20/0.1 v/v/v.
14
1.4
N 2 / N 2’

1.3 N 1 / N 1’
Rs / Rs’
S
meter

1.2
Relative param

N
1.1
N

1.0

Mequitazine
0.9

0.8
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3

Flow rate
e (mL/min)

Figure 6.13 Effect of flow rate on relative column efficiency and

resolution degree of Mequitazine enantiomers, CHIRALPAK ID (5 µm,
250 × 4.6 mm i.d.). Mobile phase: hexane/2‑PrOH/DEA 80/20.0.1 v/v/v;
T = 25°C; detection UV 254 nm; N1, N2, and Rs: the values obtained
at 0.7–1.1 mL/min; N1, N2, and Rs: the values obtained at 1.2 mL/min.

Mequitazine on CHIRALPAK ID (5 µm), the reduction of the flow

rate from 1.2 to 0.7 mL/min allowed increasing the plate counts
at the level of 20%, resulting in enhancement of the resolution
degree at the level of 10%. It should be mentioned that the gain
140 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

of 10% in resolution is at the expense of much longer (about

80%) analysis time. In some cases, playing the mobile phase
composition to optimize the selectivity seems to be a more
attractive alternative for method optimization as the lack of
column efficiency can often be largely compensated by a high
selectivity [126].

6.6 Influence of Analyte Structure and the

Substituent Effect
It has been observed in experimental work that a given single
column can be peculiarly versatile for chiral separation of whole
series of structure-related compounds [51, 127–131]. Still, some
efforts would often be needed for optimizing the methods for
the given compound by varying the chromatographic conditions
(e.g., temperature, mobile phase composition) and even by changing
the chromatographic mode. This is undoubtedly related to the
complexity of the chiral recognition mechanism operating on the
polysaccharide-derived CSPs.
The interactions between analyte molecules and a chiral
stationary phase are multiple and of diverse natures, including
mainly hydrogen bonding, p–p associations, dipole–dipole stacking,
hydrophobic interactions and the steric fit of the asymmetric
portions of the analyte structure to the chiral environment, which
is defined by the molecular and supramolecular structure of
the polymeric chiral selector [132]. Depending on the analyte
molecular structure, the mobile phase in use and the solvation
state of the polymer, each of the interactions mentioned above
may play a role of higher or lower importance in retention and
in recognition of the enantiomers. Such a complexity makes the
prediction of chiral separation difficult and the trial-and-error
approach has been constantly adopted for method development
on polysaccharide-based CSPs. Even in the case where the column
and the chromatographic conditions are kept the same, the
slightest difference in the analyte molecular structure can induce
drastic change in chromatographic results. This is the case,
for instance, of the compound pair of Verapamil and
Methoxyverapamil.
 Influence of Analyte Structure and the Substituent Effect 141

Structurally, Verapamil and Methoxyverapamil differ from

each other merely from the additional methoxy group on the
benzeneacetonitril moiety in Methoxyverapamil. However,
such a small structural variation brought about quite different
chromatographic results on CHIRALPAK IE with Hexane/2-
PrOH/DEA 80/20/0.1 v/v/v as the mobile phase. As shown in Fig.
6.14a, the first eluting peak of both compounds had very similar
retention time (ca. 19 min). In contrast, the second peak of Methoxy-
verapamil was eluted much later than that of Verapamil, leading
to significant enhancement in enantioselectivity and in resolution
degree of the Methoxyverapamil enantiomers. It seems that the
presence of the third bulky methoxy group at 5‑position of the
phenyl group would intensify the steric selective effect but not
disturb interactions of all the other types between the analytes and
the CSPs. N‑CBZ-Valine and N‑CBZ-Norvaline are isobaric amino
acid derivatives. Their structural difference is in the aliphatic
side chain: isopropyl in the former; n-propyl in the later. It seems
that the isopropyl group has more favorable steric effect than
n-propyl group on IA‑U column, resulting to much larger
enantioselectivity for N‑CBZ-DL-Valine than for N‑CBZ-DL-Norvaline
enantiomers (Fig. 6.14b).

NC O
N O CH3
H
O N
OH
CH3O
O CH3
O
CH3O Verapamil
(aͲ1) (bͲ1)
NͲCBZͲDLͲVal
D =1.10 ɲ =1.62
Rs =1.9 Rs=6.8

NC O
N O CH3 H
O N
OH
CH3O
O CH3 O

CH3O O CH3

(bͲ2) NͲCBZͲDLͲNva
Methoxyverapamil
(aͲ2) ɲ =1.21
D =1.28 Rs=2.5
Rs =4.8

5 10 15 20 25 30
0 1 2 3 4
Minutes
Minutes

Figure 6.14 Substituent effect of the analytes on enantiomer separations.

(a) CHIRALPAK IE (5 µm, 250 × 4.6 mm i.d.); mobile phase:
hexane/2‑PrOH/DEA 80/20.0.1 v/v/v; flow rate: 1.0 mL/min; T = 25°C;
(b) CHIRALPAK IA‑U (1.7 µm, 100 x 3.0 mm i.d.); mobile phase:
hexane/EtOH/FA 80/20.0.1 v/v/v; flow rate: 1.425 mL/min; T = 25°C.
 142 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

If the substituent effects in the above examples could be quite

easily assessed in the enantiomer recognition, the consequence
of the structural change at the substituent level is not always
singular. In most cases, the substituent effect is more complex
and can hardly be predictable. In a systematic study on the
substituent effect, a series of ß‑blockers was injected on a
CHIRALPAK ID column using a single mobile phase (Hexane/
2‑PrOH/DEA 80/20/0.1 v/v/v). As summarized in Table 6.3, the
chromatographic OH data scattered very much for these compounds.
H
O N
Ar
Table 6.3 Enantiomer
OH separations of β-blockers on CHIRALPAK ID
H
O N
Ar Ar
OH t1 t2 k1 D
OH H
O OH HN
Bunolol Ar O N H 9.13
OH 9.78 1.97 1.11
Ar
Ar
O Ar NH
OH t1 H t2 k1 D
O N O N
Bunolol Ar Ar Ar t1 9.13 t2 9.78 k1 1.97 D 1.1
OAr t1 t2 k1 D
Bunolol Ar t1
9.13 t2
9.78 k1
1.97 D
1.11
Bunolol Ar 9.13
t1 9.78
t2 1.97
k1 1.11
D
Bunolol ArO 9.13
Ar9.13
t1 t2 k1 9.78 a t1 Rs 1.97 1.11
t2
Bunolol 9.78 1.97 1.11
PindololBunolol
BunololOO 9.13
6.56 9.78 1.97
6.94 1.119.131.44 1.13 9.78
1.11
O
HON
Pindolol 6.56 6.94 1.13 1.1
Pindolol HN
O 6.56 6.94 1.13 1.11
Pindolol 6.56 6.94 1.13 1.11
Pindolol
Pindolol HN 6.56
6.56 6.94 6.94 1.11
1.13 1.26 1.13 1.11
Pindolol HN 6.56 6.94 1.13 1.11
Propranolol HN 5.09 5.75 0.66 1.32
HN
Propranolol 5.09 5.75 0.66 1.3

Propranolol
Pindolol
Propranolol 5.09
5.09
5.75 6.56 0.66
5.75 1.32 3.500.66
6.94
1.32
1.32
Propranolol
Propranolol 5.09
5.09 5.75 0.66
5.75 0.66 1.32
Propranolol
Carazolol HN 7.51
5.09 5.75
7.93 0.66
1.44 1.32
1.09

Carazolol 7.51 7.93 1.44 1.0

Carazolol HN 7.51 7.93 1.44 1.09
CarazololCarazolol 7.51
7.51 7.93 1.44
7.93 1.09 1.121.44 1.09
Carazolol 7.51 7.93 1.44 1.09
Atenolol Propranolol 5.09 5.75
Carazolol HN 7.51 7.93 1.44 1.09
19.19 20.16 5.24 1.06
HN
HN
HN
Atenolol 19.19 20.16 5.24 1.0
HON
Atenolol 19.19 20.16 5.24 1.06
Atenolol Atenolol 19.19 20.16 1.06
19.19 20.16 5.24 0.865.24 1.06
Atenolol H2N
O 19.19 20.16 5.24 1.06
Atenolol 19.19 20.16 5.24 1.06
Alprenolol HO2O
N 4.82 4.82 0.57 1.00
CarazololH2N
O 7.51 7.93
Alprenolol H2N O
Alprenolol 4.82
4.82 4.82 0.574.821.00 0.00 0.57 1.0
H2N
H2N
Alprenolol 4.82 4.82 0.57 1.00
Alprenolol 4.82 4.82 0.57 1.00
Alprenolol 4.82 4.82 0.57 1.00
Alprenolol HN 4.82 4.82 0.57 1.00
Oxprenolol 5.87 13.34 0.91 3.67

Oxprenolol 5.87 13.34 0.91 3.6

Atenolol 19.19 20.1
Oxprenolol O 5.87 13.34 0.91 3.67
Oxprenolol 5.87 13.34 0.91 3.67
Oxprenolol 5.87 13.34 0.91 3.67
Oxprenolol O 5.87 13.34 0.91 3.67
Tiprenolol O 6.51 7.88 1.12 1.40
O
O
O
 H2N
Alprenolol 4.82 4.82 0.57 1.00
Oxprenolol 5.87 13.34 0.91 3
Alprenolol 4.82 4.82 0.57 1.00
Oxprenolol 5.87 13.34 0.91 3.67
Oxprenolol O
Influence of Analyte Substituent Effect 0.91
5.87Structure and the 13.34 143 3

Oxprenolol O 5.87 13.34 0.91 3.67

Tiprenolol O 6.51 7.88 1.12 1
Oxprenolol Ar t1
5.87 t2 k113.34 a Rs 0.91 3.67
Tiprenolol
Oxprenolol O
6.51
5.87 13.34 0.917.883.67 13.59 1.12 1.40
S
Tiprenolol 6.51 7.88 1.12 1
O
S
Tiprenolol S
6.51 7.88 1.12 1.40
Acebutolol
Tiprenolol 6.5115.707.88 1.12 17.14
1.40 4.91 4.11 1
O
Tiprenolol 6.51 7.88 1.12 1.40
Acebutolol S 15.70 17.14 4.11 1.11
O
Acebutolol
Acebutolol S
O 15.7017.14 4.11 17.14
15.70 1.11 1.44 4.11 1
O

Metoprolol
Acebutolol O 15.70 7.14 17.14 11.87 4.11 1.32 1.11 2
O O
Acebutolol
Metoprolol 15.70
7.14 17.14
11.87 4.11
1.32 1.11
2.16
Metoprolol O O 7.14 11.87 1.32 2.16 11.57
Metoprolol O 7.14 11.87 1.32 2
O
O
Metoprolol O 7.14 11.87 1.32 2.16
Bisoprolol
Bisoprolol 6.986.98 9.22 1.27 9.22 1.57 6.20 1.27 1
Metoprolol O 7.14 11.87 1.32 2.16
O O
Bisoprolol 6.98 9.22 1.27 1.57
Note: Column size: 250 × 4.6 mm i.d., Particle size: 5 µm. Mobile phase: 1.27
Bisoprolol O O 6.98 9.22 1
O
Hexane/2-PrOH/DEA 80/20/0.1
O v/v/v. Flow rate: 1.0 mL/min, T: 25°C.
O
Bisoprolol 6.98 9.22 1.27 1.57
Bisoprolol Although
O bearing
O the same general
6.98 or basic9.22
structure, these
1.27 1.57
ß‑blockers had veryO large variation in retention factor (k1 ranging
O
from 0.57 to 5.24) and in enantioselectivity (a spreading from
1.00 to 3.67) in function of the aromatic scaffold. The most
striking difference in enantiomer separation was observed
between Alprenolol and Oxprenolol. Both molecules have a single
substituent at position‑2 of the phenoxy group. The substituent is
an allyl group in Alprenolol but an allyloxy group in Oxprenolol.
The structural difference is very subtle but the chromatographic
results were extremely distinctive. Chromatographically, racemic
Alprenolol was eluted very fast in a single peak (Fig. 6.15a), while
racemic Oxprenolol was resolved into enantiomers to a very
large degree with slightly longer retention time for the first peak
(Fig. 6.15b). There is no doubt that the allyloxy group in Oxprenolol
played an essential role in the enantioselective mechanism under
the given chromatographic conditions. This phenomenon with
the same compounds may not be as significant as described here
if the chromatographic conditions are not the same (e.g., changes
in mobile phase and/or in column).
144 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

4.8 OH 5.9 OH
H H
O N O N

Oxprenolol
Alprenolol

13.3
(a) (b)
D = 1.00 D = 3.67
Rs = 0.0 Rs = 13.6

0 2 4 6 8 10 12 14 16 18 PLQ 0 2 4 6 8 10 12 14 16 18 PLQ

Figure 6.15 Substituent effect of the ß‑blockers on enantiomer

separations. Column: CHIRALPAK ID (5 µm, 250 × 4.6 mm i.d.); mobile
phase: hexane/IPA/DEA 80/20/0.1 v/v/v; flow rate: 1.0 mL/min; T: 25°C.

6.7 Column Maintenance, Cleaning, and

Regeneration
The column lifetime depends on daily care and the operator’s
awareness of the cautions to be taken for the columns in use. The
factors affecting the column performance in the time course are
multiple. Among them, solvent compatibility would be the most
important, especially when it comes to the coated-type columns
based on polysaccharide derivatives.
As previously described, in the coated CSPs the polysaccharide
derivative is not chemically bonded but physically deposited by
special technology onto the silica matrix. Therefore, all the mid‑
polar solvents, such as THF, DCM, chloroform, MtBE, acetone, and
EtOAc cannot be used as mobile phase or mobile phase components
as they are able to dissolve off the polymers and damage
irreversibly the columns. The solvents compatible to the coated
polysaccharide-CSPs are typically the mixtures of non-polar
solvents (e.g., hexane or heptane) with polar solvents (e.g., EtOH
and 2‑PrOH) for NP-mode. For PO-mode, alcohols (e.g., MeOH
and EtOH), ACN and their mixtures can be efficiently used.
When running the columns under RP conditions, all the above-
mentioned polar solvents can be employed as the organic modifier
for chiral separations. Any immiscible mixtures, for instance
hexane and ACN, must be avoided because they won’t allow getting
a homogeneous eluent medium and could also be destructive to
the coated-type columns.
 Column Maintenance, Cleaning, and Regeneration 145

In contrast, immobilized CSPs have no constrains in the choice

of the solvents, if all the solvents in use are miscible [16, 34–38].
Although the application of the non-miscible solvents does not
damage the immobilized columns, the solvent miscibility should
always be considered for the sake of reliable and reproducible
chromatographic conditions.
The columns can become “dirty” or the column inlet frit is
“clogged” if the non‑cleaned sample solutions are injected, or some
samples are too strongly retained on the column, or on-column
precipitations of the sample or the salt additives occurred. “Dirty”
column will lead to the baseline instability, need prolonged column
equilibration time and compromise the method performance.
“Clogged” column will induce higher and higher backpressure and
alter the column efficiency. In these cases, column cleaning may be
needed.
Ideally, the solvent affording the best solubility to the “dirty”
samples and the impurities should be the best for column cleaning.
However, the natures of the cumulated dirtying species are not
always clear and the free choice of the cleaning solvent is not
feasible with the columns of coated-type. According to our
experience, the most efficient cleaning method for columns of
coated-type is to flush (or carefully back-flush) the column with
pure EtOH at reduced flow rate for several hours (≥3 h). When
used in RP-mode, the column can be cleaned with salt-free and
water-enriched mobile phase, followed by ACN flushing at reduced
flow rate.
For columns of the immobilized-type, the column cleaning
may be more efficient by choosing the right solvent without
worrying about the CSP stability. A general protocol is proved to be
effective for column cleaning packed with the immobilized CSPs.
It consists of three steps: (1) EtOH for 30 min; (2) THF for 2 h;
(3) EtOH for 30 min. If the column cleaning is not successful
enough with this protocol, the column flushing can be repeated
but using DMF for 3 h instead of THF for 2 h. Again, the column
flushing is recommended to be carried out at reduced flow rate.
For columns of 4.6 mm internal diameter, for instance, the flow
rate can be set at 0.5 mL/min for EtOH and THF, at 0.3 mL/min
for DMF.
Another potential issue occasionally happening could be a
change in the polymer conformation in the column. When dealing
146 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

with immobilized-type phases, although the polysaccharide

derivatives cannot be stripped off by solvents, especially the
solvents of mid-polarity in use for the mobile phase and sample
solvent, the inter- and intra‑molecule interactions inside the
bulk structure of the polymer may be weakened or modified, the
polymer then swells to adopt a specific conformation by following
the lowest energy rule in the given solvent medium. As a result,
the new polymer conformation created may have an influence in
certain sensitive applications, by changing selectivity, resolution
and/or efficiency.
Figure 6.16 shows an example of conformation change on
the separation of a pair of enantiomers on CHIRALPAK IA (5 µm,
250 × 4.6 mm i.d.). In comparison with the separation initially
obtained (chromatogram (a)) using hexane/EtOH 90/10 v/v as
mobile phase, the resolution degree of racemic (4‑fluorophenyl)‑γ
‑butyrolactone was very much compromised after the application
of DCM-containing mobile phases (chromatogram (b)) and EtOAc-
containing mobile phases (chromatogram (c)). In both cases,
the separation deterioration was commonly characterized by
a significant reduction of enantioselectivity (the second peak
eluted earlier). In the case of (c), the reduction of selectivity
was accompanied by the broadened peak shape. As a part of
the typical characteristics of the immobilized CSPs, the column
performance deviation can be remedied by a specific regeneration
method. Once regenerated, the column can afford perfectly
reproducible separations regardless of its mobile phase
background (chromatogram (d)).
CSP regeneration could be particularly valuable while switching
from mobile phases containing the mid-polar solvents to the
ones composed of alkane-alcohol mixtures or of polar solvents.
Having the ability to dissolve the free polysaccharide derivatives,
all the solvents of mid polarity can condition the CSP by themselves,
erase the memory imprinted by other solvents and allow the
polymer to develop a specific solvated state or chain conformation
which, in its turn, determines the specific enantioselective properties
of the CSP.
The protocols to regenerate immobilized columns are quite
like the general column cleaning method described above. Thus,
for easy protocol handling, the column regeneration and column
cleaning can in principle be combined for these columns. If we
 Column Maintenance, Cleaning, and Regeneration 147

consider in detail the whole series, some differences can be

observed between cellulose and amylose derivatives. The amylose
derivatives (CHIRALPAK IA, CHIRALPAK ID, CHIRALPAK IE,
CHIRALPAK IF, and CHIRALPAK IH) have rather loose and flexible
structures due to the a‑1,4 linkage between the glucose units.
It is easier to get the amylose-based CSPs changed but also faster
to get the column performance regenerated. On the other side,
the columns based on cellulose derivatives (CHIRALPAK IB and
CHIRALPAK IC) usually undergo much less pronounced performance
change than the amylose-based ones, even with prolonged
exposure to the non-standard mobile phases. This may be attributed
to the intrinsic properties of the cellulose materials. In cellulose,
the glucose units are bonded together by β‑1,4 linkage, which
keeps the polymer chains from coiling. The straight and relatively
rigid chain conformation in cellulose or cellulose derivatives favors,
both in intensity and in strength, the inter- and intra‑chain
interactions by multiple attracting forces. It is this capacity for
self-assembling that firmly holds the polymer chains together
against the exertions for modification of the supra-molecular
structure. For this reason, the regeneration of CHIRALPAK IB
and CHIRALPAK IC is rather rarely needed.
13.2

12.6
15.8

13.5

O O
Non-standard
F
mobile phases
(4-Fluorophenyl)-
Ȗ-butyrolactone

(a) (b)

6 8 10 12 14 16 18 min
6 8 10 12 14 16 18 min
13.1
13.5

16.0

4.4
14

Regeneration
(d) (c)

6 8 10 12 14 16 18 min 6 8 10 12 14 16 18 min

Figure 6.16 Effect of the non-standard mobile phases and the regeneration:
CHIRALPAK IA (5 µm, 250 × 4.6 mm i.d.); flow rate: 1.0 mL/min;
mobile phase: hexane/EtOH 90/10 v/v. T: 25°C. (a) Initial separation.
(b) Separation obtained after application of DCM‑containing mobile
phase. (c) Separation obtained after application of EtOAc-containing
mobile phase. (d) Separation obtained after column regeneration.
148 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

6.8 Conclusions
The technology advancement in preparing polysaccharide-based
chiral stationary phases has provided an impetus to the field of
chiral separation by chromatography. It gives rise to CSPs with
higher performance, broader application scope, higher versatility
and enhanced robustness. It also makes the approaches of method
development evolve to reach higher success rate in analytical and
preparative separations of chiral compounds.
The establishment of the screening protocols by choosing
the appropriate chiral columns and the suitable mobile phase is
essential for efficient method development. The option to use 3 to
4 immobilized polysaccharide‑derived CSPs in combination with
the 4 most versatile mobile phase systems fits well the current
challenge of fast screening for an increased number of compounds
in a limited time frame. The choice in chromatographic mode
(SFC or HPLC, NP or RP, etc.) can be considered in connection with
the instrument availability, the application purpose and specific
aspects of the sample source and sample properties.
Based on the screening results, the optimization to obtain
the best suited method can be accomplished by fine‑tuning the
mobile phase composition, by using the most appropriate additives
in amount and in nature, by rational use of sample solvent, by
adjustment in temperature and in flow rate, by changing the
column dimension, by considering the particle size in the column
and, of course, without forgetting to try the alternative columns
which may exhibit complementary properties for chiral separation.
It is important to note that many methods issuing the first sample
screening can directly meet the separation objective with no need
of further optimization efforts. Fast or ultra‑fast chiral analysis
can be achieved by running short columns packed with 3 µm or
sub‑2 µm particles on optimized chromatographic systems such
as UHPLC or ultra-high-performance SFC.

References

1. Okamoto, Y., Kawashima, M., Yamamoto, K., and Hatada, K. (1984)

Chem. Lett., 13(5) p. 739.
 References 149

2. Okamoto, Y., Kawashima, M., and Hatada, K. (1984) J. Am. Chem. Soc.
106 p. 5357.
3. Okamoto, Y., Aburatani, R., and Hatada, K. (1987) J. Chromatogr. A 389
p. 95.
4. Dingenen, J. (1994) Polysaccharide phases in enantioseparations in
A Practical Approach to Chiral Separations by Liquid Chromatography,
ed., Subramanian, G., Chapter 6 (VCH, Weinheim).
5. Francotte, E. (1997) Chromatography as a separation tool for the
preparative resolution of racemic compounds in Chiral Separations,
Applications and Technology, ed., Ahuja, S., Chapter 10 (American
Chemical Society, Washington).
6. Okamoto, Y., and Kaida, Y (1994) J. Chromatogr. A 666 p. 403.
7. Francotte, E. (1994) J. Chromatogr. A 666 p. 565.
8. Okamoto, Y., and Yashima, E. (1998) Angew. Chem. Intl. Ed. Engl. 37
p. 1020.
9. Maier, N. M., Franco, P., and Lindner, W. (2001) J. Chromatogr. A 906
p. 3.
10. Yashima, E. (2001) J. Chromatogr. A 906 p. 105.
11. Tachibana, K., and Ohnishi, A. (2001) J. Chromatogr. A 906 p. 127.
12. Franco, P., Senso, A., Oliveros, L., and Minguillón, C. (2001) J. Chromatogr.
A 906 p. 155.
13. Francotte, E. (2001) J. Chromatogr. A 906 p. 379.
14. Aboul-Enein, H. Y. (2001) J. Chromatogr. A 906 p. 185.
15. Cox, G. (ed.) (2005) Preparative Enantioselective Chromatography
(Blackwell Publishing, Oxford).
16. Zhang, T., and Franco, P. (2007), Chiral Separation Techniques: A Practical
Approach (3rd revised and updated Edition) ed., Subramanian, G.,
Chapter 3 (VCH, Weinheim).
17. Franco, P., and Zhang, T. (2008) J. Chromatogr. B 875 p. 48.
18. Zhang, T., Nguyen, D., and Franco, P. (2010) J. Chromatogr. A 1217
p. 1048.
19. Lee, J., Lee, J. T., Watts, W. L., Barendt, J., Yan, T. Q., Huang, Y., Riley, F.,
Hardink, M., Bradow, J., and Franco, P. (2014) J. Chromatogr. A 1374
p. 238.
20. Khater, S., Lozac’h, M.-A., Adam, I., Francotte, E., and West, C. (2016)
J. Chromatogr. A 1467 p. 463.
21. Francotte, E. (2016) LCGC Eur. 29 p. 194.
150 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

22. Baudelet, D., Schifano-Faux, N., Ghinet, A., Dezitter, X., Barbotin, F.,
Gautret, P., Rigo, B., Chavatte, P., Millet, R., Furman, C., Vaccher, C., and
Lipka, E. (2014) J. Chromatogr. A 1363 p. 257.
23. Rizzo, S., Menta, S., Faggi, C., Pierini, M., and Cirilli, R. (2014) J.
Chromatogr. A 1363 p. 128.
24. Lämmerhofer, M. (2010) J. Chromatogr. A 1217 p. 814.
25. Zhang, T., and Franco, P. (2008) LC‑GC Europe 21(9) p. 430.
26. Cirilli, R., Ferreti, R., Gallinilla, B., La Torre, F., Mai, A., and Rotili, D.
(2006) J. Sep. Sci. 29 p. 1399.
27. De la Puente, M. L., White, C. T., Rivera-Sagredo, A., Reilly, J., Burton, K.,
and Harvey, G. (2003) J. Chromatogr. A 983 p. 101.
28. De la Puente, M. L. (2004) J. Chromatogr. A 1055 p. 55.
29. Andersson, M. E., Aslan, D., Clark, A., Roeraade, J., and Hagman, G.
(2003) J. Chromatogr. A 1005 p. 83.
30. Wetli, H. A., and Francotte, E. (2007) J. Sep. Sci. 30 p. 1255.
31. Matthijs, N., Maftouh, M., and Heyden, Y. V. (2006) J. Chromatogr. A
1111 p. 48.
32. Zhang, T., and Franco, P. (2010) LC‑GC Europe 23(6) p. 302.
33. Franco, P., and Zhang, T. (2013) Chiral Separations, Methods and
Protocols (Second edition), ed., Scriba, G., Chapter 6 (Humana Press),
pp. 113–126.
34. Zhang, T., Nguyen, D., and Franco, P. (2008) J. Chromatogr. A 1191
p. 214.
35. Younes, A. A., Ates, H., Mangelings, D., and Vander Heyden, Y. (2013)
J. Pharm. Biomed. Anal. 75 p. 74.
36. Zhang, T., Kientzy, C., Franco, P., Ohnishi, A., Kagamihara, Y., and
Kurosawa, H. (2005) J. Chromatogr. A 1075 p. 65.
37. Zhang, T., Nguyen, D., Franco, P., Murakami, T., Ohnishi, A., and
Kurosawa, H. (2006) Anal. Chim. Acta 557 p. 221.
38. Zhang, T., Nguyen, D., Franco, P., Isobe, Y., Michishita, T., Murakami,
T. (2008) J. Pharm. Biomed. Anal. 46 p. 882.
39. Zhang, T., and Franco, P. (2004) Int. Lab. 34‑5 p. 24.
40. Cox, G. B. (2008). A screening process for development of
enantioselective supercritical fluid chromatography separation
methods, LC.GC Application notes.
41. Amoss, C., Cox, G. B., Franco, P., and Zhang, T. (2008) CHIRALPAK®
IC™—An Immobilized Polysaccharide Chiral Stationary Phase with a
Unique Chiral Selector, LC.GC Application Notes.
 References 151

42. Zhang, T., Schaeffer, M., and Franco, P. (2005) J. Chromatogr. A 1083
p. 96.
43. Cox, G. B., Maier, N. M., Zhang, T., and Franco, P. (2009) LC GC Application
Notes.
44. Zhang, T., Nguyen, D., and Franco, P. (2006) J. Sep. Sci. 29 p. 1517.
45. Ghanem, A., and Aboul Enein, H. Y. (2005) Anal. Chim. Acta 548 p. 26.
46. Ghanem, A., and Naim, L. (2006) J. Chromatogr. A 1101 p. 171.
47. Ali, I., Naim, L., Ghanem, A., and Aboul Enein, H. Y. (2006) Talanta 69
p. 1013.
48. Cirilli, R., Simonelli, A., Ferretti, R., Bolasco, A., and Climenti, P. (2006)
J. Chromatogr. A 1101 p. 198.
49. Cirilli, R., Orlando, V., Ferretti, R., Luciana Turchetto, L., Silvestri, R.,
De Martino, G., and La Torre, F. (2006) Chirality 18 p. 621.
50. Andersson S. (2007) Chiral Separation Techniques: A Practical
Approach (3rd revised and updated Edition), ed., Subramanian, G.,
Chapter 17 (VCH, Weinheim).
51. Cirilli, R., Ferretti, R., Gallinella, B., De Santis, E., Zanitti, L., and La Torre,
F. (2008) J. Chromatogr. A 1177 p. 105.
52. Cirilli, R., Ferretti, R., Gallinella, B., Billa, A. R., Vincieri, F. F., and
La Torre, F. (2008) J. Sep. Sci. 31 p. 2206.
53. Cirilli, R., Ferretti, R., De Santis, E., Gallinella, B., Zanitti, L., and
La Torre, F. (2008) J. Chromatogr. A 1190 p. 95.
54. Ferretti, R., Gallinella, B., La Torre, F., Zanitti, L., Turchetto, L.,
Mosca, A., and Cirilli, R. (2009) J. Chromatogr. A 1216 p. 5385.
55. Pawar, R. S., Grundel, E., Mazzola, E., White, K. D., Krynitsky, A. J., and
J. I. Rader, J. I. (2009) J. Sep. Sci. 33 p. 200.
56. Wang, C., Armstrong, D. W., and Chang, C. D. (2008) J. Chromatogr. A
1194 p. 172.
57. García, P., Franco, P., Alvarez, R., and De Lera., U. R. (2011) J. Sep. Sci.
34 p. 999.
58. Zhang, T., Franco, P., Nguyen, D., Hamasaki, R., Miyamoto, S., Ohnishi, A.,
and Murakami, T. (2012) J. Chromatogr. A 1269 p. 478.
59. Ashok, S., Varma, M. S., and Swaminathan, S. (2012) J. Chromatogr.
Sci. 50 p. 799.
60. Zhou, J., Du, Q. Z., Zhao, S. Z., Sun, F., Li, X. Y., and Zhang, Z. Z. (2014)
J. Chromatogr. Sci. 53 p. 959.
61. Geryka, R., Kalíkováa, K., Vozkaa, J., Plecitáa, D., Schmidb, M. G., and
Tesărová, E. (2014) J. Chromatogr. A 1363 p. 155.
152 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

62. Peluso, P., Mamane, V., Aubert, E., and Cossu, S. (2014) J. Sep. Sci. 37
p. 2481.
63. Harrison, A., Melchionna, M., Franco, P., and Hlavac, J. (2014) J. Chem.
38 p. 5491.
64. Rizzo, S., Menta, S., Benincori, T., Ferretti, R., Pierini, M., Cirilli, R., and
Sannicolo, F. (2015) Chirality 27 p. 888.
65. Geryk, R., Kalíková, K., Vozka, J., and Tesarŏvá, E. (2015) Chromato-
graphia 78 p. 909.
66. Liu, N., Dong, F., Xu, J., Liu, X. G., Chen, Z. L., Tao, Y., Pan, X. L., Chen, X. X.,
and Zheng, Y. Q. (2015) J. Agric. Food Chem. 63 p. 6297.
67. Islam, M. F., and Lee, W. (2015) J. Chosun Natural Sci. 8 p. 111.
68. Rao, R. N., and Kumar, K. N. (2015) J. Chromatogr. Sci. 53 p. 295.
69. Menta, S., Carradori, S., Secci, D., Faggi, C., Mannina, L., Cirilli, R., and
Pierini, M. (2015) J. Org. Chem. 80 p. 11932.
70. López-Ram-de-Víu, P., Gálvez, I., and Díaz-de-Villegas, M. D. (2015)
J. Chromatogr. A 1390 p. 78.
71. Qian, Y., Zhao, X. Q., Zhao, L., Cui, L. L., Liu, L., Jiang, X. L., Liu, Y. J., Gao,
L. P., and Xia, T. (2015) J. Chromatogr. B 1006 p. 1.
72. Cirilli, R., Ferretti, R., Gallinella, B., Turchetto, L., Zanitti, L., and La
Torre, F. (2015) J. Pharm. Biomed. Anal. 50 p. 9.
73. Zanitti, L., Ferretti, R., Gallinella, B., La Torre, F., Sanna, M. L., Mosca, A.,
and Cirilli, R. (2015) J. Pharm. Biomed. Anal. 52 p. 665.
74. Sola, I., Viayna, E., Gómez, T., Galdeano, C., Cassina, M., Camps, P., Romeo,
M., Diomede, L., Salmona, M., Franco, P., Schaeffer, M., Colantuono,
D., Robin, D., Brunner, D., Taub, N., Hutter-Paier, B., and Muñoz-Torrero,
D. (2015) Molecules 20 p. 4492.
75. Moldovan, R. C., Dascal, G. S., Mirel, V., Bodoki, E., and R. Oprean, R.
(2015) Pharmacia 63 p. 909.
76. Adhikari, S., Kang, J. S., and Lee, W. (2016) Chirality 28 p. 789.
77. Srinivasu, G., Nagesh Kumar, K., Thirupathi, C., Lakshmi Narayana, C.,
and Parameswara Murthy, C. (2016) Chromatographia 79 p. 1457.
78. Saduttoa, D., Ferretti, R., Zanitti, L., Casulli, A., and Cirilli, R. (2016)
J. Chromatogr. A 1445 p. 166.
79. Gallinella, B., Ferretti, R., Zanitti, L., Sestili, I., Mosca, A., and Cirilli, R.
(2016) J. Pharm. Anal. 6 p. 132.
80. Kim, S. J., Nam, K. W., Park, B., Md. F. Islam, Md. F., and Lee, W. (2016)
Korean Soc. Biotechnol. Bioeng. J. 31 p. 186.
 References 153

81. Lotter, C., Poehler, E., Heiland, J., Mauritz, L., and Belder, D. (2016)
Lab Chip. 16 p. 4648.
82. Török, G., Goetelen, L., Luyckx, R., and Van Broeck, P. (2005) J. Pharm.
Biomed. Anal. 39 p. 425.
83. Huybrechts, T., Török, G., Vennekens, T., Sneyers, R., and Vrielynck, S.
(2007) LCGC Europe 20 p. 319.
84. Ishikawa, A., and Shibata, T. (1993) J. Liq. Chromatogr. 16 p. 859.
85. Zhou, L., Welch, C., Lee, C., Gong, X., Antonucci, V., and Ge, Z. (2009)
J. Pharm. Biomed. Anal. 49 p. 964.
86. Amoss, C., Coryell, B. R., Cox, G. B., and Maier, N. M. (2010) Poster at
the HPLC2010, Boston, USA.
87. Cirilli, R., Ferretti, R., Gallinella, B., and Zanitti, L. (2013) J. Chromatogr.
A 1304 p. 147.
88. Materazzo, S., Carradori, S., Ferretti, R., Gallinella, B., Secci, D., and
Cirilli, R. (2014) J. Chromatogr. A 1327 p. 73.
89. Gallinella, B., Bucciarelli, L., Zanitti, L., Ferretti, R., and Cirilli, R. (2014)
J. Chromatogr. A 1339 p. 210.
90. Ferretti, R., Zanitti, L., Casulli, A., and Cirilli, R. (2016) J. Sep. Sci. 39
p. 1418.
91. Petersson, P., Lundell, N., and Markldes, K. E. (1992) J. Chromatogr. 623
p. 129.
92. Wang, R. Q., Ong, T. T., Tang, W. H., and Ng, S. C. (1012) Trends Anal.
Chem. 37 p. 83.
93. Kalíková, K., Slechtová, T., Vozka, J., and Tesărová, E. (2014) Anal.
Chim. Acta 821 p. 1.
94. Regalado, E. L., and Welch, C. J. (2015) J. Sep. Sci. 38 p. 2826.
95. Lemasson, E., Bertin, S., and West, C. (2016) J. Sep. Sci. 39 p. 212.
96. Cox, G., Stringham, R., and Matabe A. (2002) LC GC North America,
The Application Notebook, August, p. 24.
97. Maftouh, M., Granier-Loyaux, C., Chavana, E., Marini, J., Pradines, A.,
Vander Heyden, Y., and Picard, C. (2005) J. Chromatogr. A 1088 p. 67.
98. Diehl, G., Meishammer, A., Huynh, D., and Francotte, E. (2006)
Presentation at SPICA 2006, Innsbruck, Austria.
99. Zhang, T., Nguyen, D., and Franco, P. (2007) Presentation at the
HPLC2007, Ghent, Belgium.
100. Cox, G. B. (2007) Presentation at the 1st International Conference
on SFC, Pittsburgh, Pennsylvania, USA (September 2007).
154 Polysaccharide-Derived Chiral Stationary Phases for the Separation of Enantiomers

101. Miller, L., and Potter, M. (2008) J. Chromatogr. B 875 p. 230.

102. Zhang, T., Nguyen, N., Franco, P., and Vollmer, M. (2011) Posters at
Chirality ISCD 23, July 2011 (Liverpool, UK) and at SFC 2011, July 2011
(New York, USA).
103. Zhang, T., Nguyen, N., Franco, P., and Vollmer, M. (2011) Agilent
application note (2011): http://www.chem.agilent.com/Library/
applications/5990-9315EN.pdf.
104. Zhang, T., Nguyen, N., Franco, P., and Vollmer, M. (2011) Agilent
application note (2011): http://www.chem.agilent.com/Library/
applications/5990-9459EN.pdf.
105. Ye, Y. K., Lynam, K. G., and Stringham, R. W. (2004) J. Chromatogr. A
1041 p. 211.
106. Hamman, C., Wong, M., Hayes, M., and Gibbons, P. (2011) J. Chromatogr.
A 1218 p. 3539.
107. Bhavyasri, K., Rambabu, D., Prasad, P. S. S., and Murali Balaram, V.
(2013) Am. J. Anal. Chem. 4 p. 51.
108. Kasten, S. A., Zulli, S., Jones, J. L., Dephillipo, T., and Cerasoli, D. M.
(2014) Chirality 26 p. 817.
109. Khater, S., Zhang, Y., and West, C. (2014) J. Chromatogr. A 1363 p. 278.
110. DaSilva, J. O., Coes, B., Frey, L., Mergelsberg, I., McClain, R., Nogle, L., and
Welch, C. J. (2014) J. Chromatogr. A 1328 p. 98.
111. Khater, S., and West, C. (2014) J. Chromatogr. A 1373 p. 197.
112. Khater, S., and West, C. (2015) J. Pharm. Biomed. Anal. 102 p. 321.
113. Jin, L. X., Gao, W. L., Yang, H. Y., Lin, C. M., and Liu, W. L. (2011) J.
Chromatogr. Sci. 49 p. 739.
114. Miller, L. (2012) J. Chromatogr. A 1256 pp. 261–266.
115. O’Brien, T., Crocker, L., Thompson, R., Toma, P. H., Conlon, D. A. Feibush,
B., Moeder, C., Blacker, G., and Grinberg, N. (1997) Anal. Chem. 69
p. 1999.
116. Wang, F., O’Brien, T., Dowing, T., Bicker, G., and Wyvratt, J. (2002)
J. Chromatogr. A 958 p. 69.
117. Sanaie, N., and Haynes, C. A. (2006) J. Chromatogr. A 1104 p. 164.
118. Lao, W., and Gan, J. (2006) J. Chromatogr. A 1131 p. 74.
119. Yao, B., Zhan, F., Yu, G., Chen, Z., Fan, W., Zeng, X., and Weng, W. (2009)
J. Chromatogr. A 1216 p. 5429.
120. Schlauch, M., and Frahm, A. W. (2001) Anal. Chem. 73 p. 262.
 References 155

121. Gasparrini, F., Marini, F. F., Pierini, M., and Villani, C. (1999) Enantiomer
4 p. 325.
122. Fulde, K., and Frahm, A. W. (1999) J. Chromatogr. A 858 p. 33.
123. Persson, B. A., and S., Andersson, S. (2001) J. Chromatogr. A 906
p. 195.
124. Balmer, K., Lagerström, P. O., and Persson, B. A. (1992) J. Chromatogr.
592 p. 331.
125. Majors, R. E. (2006) LC‑GC Europe 19 p. 352.
126. Butchart, K., Foster, G., and Temesi, D. (2007) Chromatogr. Today 1
p. 3.
127. Aboul‑Enein, H. Y., and Serignese, V. (1994) Chirality 6 p. 378.
128. Aboul‑Enein, H. Y., and Bakr, S. A. (1997) Chirality 9 p. 10.
129. Toribio, L., del Nozal, M. J., Bernal, J. L., Jimenez, J. J., and Alonso, C.
(2004) J. Chromatogr. A 1046 p. 249.
130. Toribio, L., del Nozal, M. J., Bernal, J. L., Alonso, C., and Jimenez, J. J.
(2005) J. Chromatogr. A 1091 p. 118.
131. Myrdal, P. B., Angersbach, B. S., Karlage, K., and Kuehl, P. J. (2006) J.
Chromatogr. A 1132 p. 315.
132. Gaffney, M. H., Stiffin, R. M., and Wainer, I. W. (1989) Chromatographia
27 p. 15.
Chapter 7

Chiral Chromatography: Method

Development

Laila Kotta and Gregory K. Websterb

aXenon, 200-3650 Gilmore Way, Burnaby, British Columbia V5G 4W8, Canada
bAbbvie, 1 N. Waukegan Rd, North Chicago, Illinois 60064, USA
[email protected], [email protected]

7.1 Introduction
In achiral chromatography, there are many understood interactions
and equilibria that take place between the analytes, the solid phase
of the column and the liquid (mobile) phase. These interactions
can be modeled (see Chapter 2) and often fall within well-
established theories and equations of chromatographic separations.
The same is not true for chiral separations. Chiral separations
depend initially on how well (points of contact) the chiral molecules
interact with the chiral selector on the solid phase of the column.
The packing of the column, the number of chiral selectors and

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
158 Chiral Chromatography

how they are attached (i.e., coated vs. immobilized) have effects on
the chiral separations and dictate how much method development
space there is within given chromatographic parameters. Many
chiral methods have very small design spaces as the chiral selectors
work under quite specific conditions. For example, in an achiral
assay, if one had two closely eluting peaks, one could adjust the
column (longer, smaller particle size, smaller internal diameter) to
increase the resolution. These parameters will all work for chiral
method development as well. In reverse-phase chromatography,
one can almost always slow down elution by slowing the change
from polar to non-polar liquids. This theoretically can be done ad
infinitium, as the peaks will continue to spread but nonetheless,
what happens to the peaks is highly predictable. Conversely,
in chiral chromatography one may only have a window of 5%
to 10% organic where the interactions with the chiral phase
favor chiral separation. Going below or above that range may
cause the two enantiomers to, once again, merge into one
peak. Overall, the goal of chiral chromatography is to achieve a
resolution (Rs) of at least 1.5 between enantiomer peaks. In some
cases when the 1.5 resolution is not achievable, the method can
still be used, if the small enantiomer peak elutes prior to the parent
peak. This order of elution is a second (and sometimes achievable)
goal when developing a chiral method. Once an enantioselective
stationary phase has been found, then much of the chiral method
development follows the same processes as developing an
achiral method.
These goals lead to the first and most useful part of chiral
chromatographic method development—column screening. This
chapter presents a roadmap for the development of a chiral purity
method.

7.2 Start at the End, with the Detectors

One tool available to aid in chiral analyses is detectors that
differentiate between enantiomeric pairs and provide unique,
distinguishable signals for each enantiomeric compound. In the
absence of a detector that can differentiate between enantiomers,
the analyst is left with only a diode array detector or a mass
detector to identify chiral pairs. The danger in relying solely on the
 Start at the End, with the Detectors 159

UV spectral data is that this type of data may not be selective to

one compound, whereas there is more confidence in mass
spectral data being unique to a compound, but not all chiral
separation buffers are mass spectrometry “friendly.” Many
compounds can give a similar UV spectrum, therefore trying to
discern enantiomeric pairs can be challenging in the presence
of other impurities of similar structure.
There are several detectors that can differentiate between
enantiomers and this is highlighted in the chromatographic traces
by the presence of positive and negative peaks. Currently there
are two types of detectors on the market that can be incorporated
into chromatographic systems: polarimeters and circular dichroism
(CD) detectors. Both detector types use Drude’s equation,
which describes the variation of optical rotation as a function of
wavelength.

7.2.1 Polarimeters
Normal light waves vibrate in many planes; however, plane
polarized light is generated when normal light is passed through
an optical polarizing filter. This effect results in a light beam
emerging that vibrates in a single plane (linearly polarized). Linearly
polarized light is rotated when passing through a compound which
is optically active. The degree of rotation depends on both the
concentration of a chiral compound and its molecular structure.
Every optically active substance has its own specific rotation
(degree of rotation in polarized light). The specific rotation of
the molecule, not the absorption characteristics, is what determines
the signal strength using the polarimeter.
Drude’s equation highlights the effect of wavelength on the
specific rotation [a]l, such that

al =
A i
, (7.1)
( l2 – li )

where Ai is a molecular constant and li is a constant wavelength

[1]. The equation also points out that the degree of rotation, as
a function of wavelength, is greatest at shorter wavelengths.
Therefore, to optimize the chiral response in a molecule, lower
source wavelengths should yield stronger responses.
160 Chiral Chromatography

Interestingly, at lower wavelengths right- and left-handed

circularly polarized light propagate at different velocities and are
absorbed by molecules to a different extent. (This phenomenon
is known as circular dichroism; see below.) When this happens,
it causes a deviation from Drude’s equation, known as the
“Cotton effect.”

Figure 7.1 Graphical representation of Drude’s equation showing the

Plain curve, the regions where there is the possibility of the Cotton effect,
and the theoretically optimal wavelength range. Reprinted from [2],
with permission from Elsevier.

The plain optical rotation dispersion (ORD) curve (or Plain

curve) shows the relationship between the magnitude of the
optical rotation versus wavelength (i.e., a representation of
Drude’s equation). The plain curve is affected by the Cotton effect
and Fig. 7.1 shows the optimum regions for analysis as well as
the flat region (a section of curve where there is little change in
[a]) but is a region with few optical interferences.

Circularly
polarized light: The two beams of linearly polarized light that are of equal
amplitude and are a quarter wave out of phase.
 Start at the End, with the Detectors 161

7.2.2 Circular Dichroism Detector

When an optically active compound preferentially absorbs right
or left circularly polarized light, the difference between the right
and left absorbances, which is often a very small value, is denoted
as the circular dichroism (CD) signal. CD wavelengths are
measured in the UV wavelength range. As with UV absorbance,
the CD signal is wavelength dependent. From the above discussion
of the Cotton effect, one could assume that a molecule should
have a chromophore with absorption in the range of 200 to
420 nm to have strong CD signal. This broad generalization was
somewhat disproven by Kott et al. [2], but considering that CD is an
absorbance based technique, chromophores will only add to the
signal. There are molecules that have CD spectra outside of this
wavelength range (i.e., porphyrins [3]), but most small active
pharmaceutical ingredients (APIs) fall within this wavelength
range.
Molecules have a distinctive CD spectrum, similar to having
unique UV spectra and for the best results, the CD detector should
be set to the enantiomeric pair’s maximum/minimum wavelength,
as one would choose a high absorbing wavelength for UV detection
in liquid chromatography.
If both the UV and CD data are collected, the anisotropy or
g-factor can be used in achiral chromatography to determine the
enantiomeric purity of a compound. Enantiomers do not have
different responses in the UV, therefore a 0.1 mg/mL sample of an
R-enantiomer, 0.1 mg/mL of an S-enantiomer of the same compound
and a 0.1 mg/mL solution of a racemic mix, will yield the same UV
signal intensity at a given wavelength. The CD signal, however, for
the R-enantiomeric solution will give the maximum positive signal
at the optimum CD wavelength, the S-enantiomeric solution will
give the maximum negative signal and racemic mix will give no
signal at all. The difference in these CD signals versus the UV signal
can also allow for the simultaneous UV detection and the
determination of the ratio of enantiomers.
Beer’s law states that the absorbance of a compound is
dependent on its molar absorptivity (e), solution concentration (c)
and path-length of the cell (l):

A=e×c×l (7.2)
162 Chiral Chromatography

and it has been discussed that the circular dichroism (CD) signal
is the difference between the absorbed right (A(r)) and left (A(l))
circularly polarized light:

CD = Ar – A1 (7.3)

Therefore, the anisotropy or g-factor is the ratio of the two

signals:

g-factor = CD/A, (7.4)

which can be reduced to a ratio of the difference in CD and

absorbance extinction coefficients:

( er – el ) × c × l e (7.5)
 =
e×c ×l e
which results in an expression independent of concentration and
path-length, provided they are constant for both measurements.
A g-factor of zero would then be indicative of a racemic mix.

7.2.3 When Chromatography Does Not Work: An

Alternative Technique
Vibrational circular dichroism (VCD) is similar to CD, but is a
spectroscopic technique which detects differences in absorption
of left and right circularly polarized infrared (IR) and near-IR
light, where vibrational transitions in molecules are observed.
VCD is an absorption measurement and its intensity is linearly
proportional to concentration and path-length in accordance to
Beer’s law. But unlike its parent IR, the VCD intensity is also linearly
proportional to enantiomeric purity of the sample, where the
signals for enantiomers are equal but opposite in sign. This feature
of VCD allows direct measurement of percent desired enantiomer
without a need for chromatographic separation [4].
The mid-IR region from 2000 to 800 cm–1 (5000 to 12500
nm) is typically were the small VCD signal is seen. Therefore, VCD
is fundamentally an IR measurement, which has an abundance of
spectroscopic bands, and works well with chemometric analyses.
 Next Step—the Type of Chromatography 163

For this type of analysis, the VCD spectra of mixtures of known

enantiomeric ratios are collected and analyzed. Chemometrics
allows for the use of many more data points than a standard
linear curve and, as such, allows one to use a defined wavelength
range instead of just a single wavelength to develop models.
With the extra data incorporated into the analysis, the error,
also known as the variability, is also calculated and included.
This error allows for a better understanding of the fit of the model
and how predictive it can be in the future, which is an indication
of how accurate and reliable these measurements are. Using
chemometrics for these percent enantiomer models leads to a
robust and sensitive approach to analysis of complex data sets
and is an alternative method of determining chiral purity in the
absence of chromatographic separation [4].

7.3 Next Step—the Type of Chromatography

7.3.1 Choice of Phase

A decision on the type of chromatography used is made based on
the solubility and stability of your molecule. The main choices are
the traditional normal phase, supercritical fluid chromatography
(SFC) or reverse phase. Polar organic chromatography has also
been used with varying degrees of success, as such it is not
highlighted in this chapter [5–7]. Other options include chiral GC
for small volatile compounds with no chromophore and chiral CE
for small charged molecules like peptides.
Choice of chromatographic separation often rests on the
solubility of the compound. More and more reverse-phase columns
are being developed (see Chapter 6), such that it is becoming the
first choice for chiral separations, as often the same or similar
preparation can be used for both the achiral and chiral assays.
If your compound is soluble in water, reverse phase is a
good place to start. If your compound is soluble in methanol,
supercritical fluid chromatography (SFC) may be useful staring
point, due to its quick throughput. If you are looking at a chiral
synthetic intermediate, which may be more soluble in strong
organic solvents or your compound is unstable under aqueous
164 Chiral Chromatography

conditions, then one tends to favor using normal-phase

chromatography. Normal-phase columns traditionally take a long
time to equilibrate, although gradients are also possible.
Recently there have been some major improvements in
analytical SFC. SFC, using mainly compressed liquid carbon dioxide
as one of the mobile phases, is a faster way to screen for chiral
separations over normal phase, providing one has the equipment.
Normal-phase chromatography can be run on almost any LC system,
whereas SFC requires a dedicated system. However, if available,
using SFC is much easier as the columns equilibrate much faster
and due to the low viscosity of liquid CO2 the flow rates are much
faster (typically between 1 and 2.5 mL/min for 4.6 mm id columns).
SFC has different selectivity, is a greener technology, is more
compatible with mass spectrometry and scales easily to preparation
scale.
Bottom line, the first step in developing a chiral method is to
screen different columns and modes of chromatography to find
the best enantiomeric separation(s) [8–10].

7.3.2 Normal-Phase Screening

Normal phase has traditionally been more selective for isomers
and has been the first choice for many as a chiral screen. Typically,
initial screens use hexane(s) or heptane as the weak eluting phase,
coupled with alcohols such as ethanol or isopropyl alcohol (IPA).
When using a combination of normal phase and SFC or converting
between the two, IPA is the preferred alcohol to use as elution
with IPA in normal phase correlates better to what is seen under
similar conditions in SFC than ethanol does. Unlike the reverse-
phase systems, gradients tend to be shallower or not run at all, as
the columns take a longer time to re-equilibrate under normal-
phase conditions. Steep gradients can be run as one needs to
elute the compounds, but care must be taken to ensure that
the systems properly re-equilibrate, for if the system has not
fully equilibrated, the chromatography achieved may not be
reproducible. Many normal-phase methods often are screened
using a gradient and when the optimum separation conditions are
found, they are converted to isocratic.
 Next Step—the Type of Chromatography 165

If both an analytical and preparatory chromato-graphic methods

are required, developing a normal-phase method yields fractions
with easily removable solvents for easier compound recovery.
For a non-exhaustive list of columns currently used in the
pharmaceutical industry, see the Introduction section of Chapter 6.

7.3.3 Supercritical Fluid Chromatography (SFC)

Screening
SFC is a normal-phase technique where supercritical fluid CO2
replaces the traditional alkane leading to much more efficient
chromatography [11]. Supercritical CO2 has a higher diffusivity and
a lower viscosity than traditional liquid chromatography
mobile phases [11], which allows for high flow rates and rapid
re-equilibration. These two factors lend well to SFC column
screening. Like normal-phase chromatography, SFC requires
screening several columns, although the mobile phase conditions
are usually restricted to the use of carbon dioxide and methanol.
Ethanol, isopropanol, and acetonitrile can also be used [12].
Table 7.1 compares example screening protocols based on the
column dimensions available. Note that other method parameters
(temp, back pressure regulator (BPR) settings, etc.) for method
screening should be based on standard settings for instrument
used.
Under SFC conditions, polysaccharide stationary phases have
the potential to provide separations for most compounds
representative of molecules in development by pharmaceutical
laboratories. Table 7.2 is a non-exhaustive list of currently used
chiral SFC columns used by the pharmaceutical industry.
Note that many of these may also be used in normal and
reverse-phase chromatography; however, it is strongly recommended
that once a column is used under a particular chromatographic
condition (reverse phase, normal phase, SFC) that column is
only to be used for that type of chromatography. Again, as with
the additives, the columns can become less reproducible when
forced to switch back and forth between conditions and different
solvents.
166 Chiral Chromatography

Table 7.1 Comparison of SFC screening methods based on column

dimensions

Column dimensions
Conditions 4.6 × 250 mm 4.6 × 250 mm 4.6 ×150 mm
Particle size (µm) 10 5 3
Flow rate (mL/min) 1.5 2 3
Starting % MeOH 4 4 1
Hold (min) 4 4 0
Ramp time to 40% 18 9 5
(min)
Hold (min) 3 2 1
Post equilibration time 5 1 0
(min)
Run cycle time (min) 30 16 6
Source: Reproduced with permission from [12].

Table 7.2 Chiral SFC columns used by the pharmaceutical industry

Typical
Particle dimensions
Manufacturer Type of phase Column name sizes (µm) (mm)
Waters polysaccharide Cell-1 Cell-2 2.5 3.0 × 100
(Trefoil series) Amy-1 (150)
Phemomenex polysaccharide Amylose-1, -2 3, 5 3.0 × 100
(Lux series) Cellulose-1, -2, (150)
-3, -4
iCellulose-5
iAmylose-1
Chiralcel polysaccharide See Figures Sub 2, 3, 5 2.1 (3.0) ×
6.1 and 6.2 in 150
Chapter 6
Regis Pirkle type Whelko (S,S) 1.8 4.6 (3.0) ×
and (R,R) (Whelko 250 (150)
Ulmo (S,S) and O1), 3.5,
(R,R) 5, 10
 Next Step—the Type of Chromatography 167

7.3.4 Reverse-Phase Screening

To make the most out of reverse-phase chiral screening it is
recommended that the screen spans the pH range and employs
multiple organic solvents for elution. A typical screen would include
a low pH (~2), a neutral pH (6.8 to 7.4) and a high pH (~9). For this
type of screen to be successful, verify that the columns used can
operate within this pH range. If not, then adjust the pH range of
columns.
As a first pass, it is recommended that methanol and
acetonitrile be the organic eluents. These solvents can be used
with either coated or immobilized chiral phases. If neither of
these phases proves to work, other organic solvents that are
miscible with water can be used (i.e., isopropyl alcohol, THF);
however, it should be verified whether such solvents are compatible
with the chosen chiral phases. Immobilized phases have a larger
solvent range than coated phases and certain solvents can strip
the (very expensive) chiral phase off coated columns.
For setting up a screen, it is best to use a system with a
column-switching valve (the more, the better; two will take a long
time to screen many columns) and, if possible, pumps with more
than the standard four lines. Ideally one would want a system
that can handle at least the three different pH aqueous phases
and two organic phases. Having lines to flush the columns between
pH changes helps prevent salting out and other damage that
sudden pH changes can cause. To expedite the screen process,
it is recommended to use a gradient, as one can essentially
perform several “isocratic” screens in one gradient run. Keep
in mind that when methanol and water mix, the mobile phase
heats up, but when acetonitrile and water mix the endothermic
enthalpy of mixing causes the mobile phase temperature to
decrease. If using high salt buffers, the aqueous solubility of the
buffer decreases with mobile phase temperature; care must be
taken when using acetonitrile to keep its maximum gradient
concentration low enough to not precipitate the salt in buffers.
Table 7.3 highlights typical gradient screening programs that
can be adjusted as needed.
168 Chiral Chromatography

Table 7.3 Typical reverse-phase gradient starting conditions for reverse-

phase chiral screening

Parameter Settings
pH’s screened 1 to 2 Neutral (6.8) 9.0
Organic phase, Methanol Acetonitrile
first pass
Gradient 90 (aqueous):10 90 (aqueous):10 (ACN) to
(methanol) to 20:80 30:70
Column 4.6 × 100 or 3.0 × 100 or 2.1 × 100 or 150
150 mm [3 µm 150 mm [3 µm mm [3 µm particle
particle size] particle size] size]
Flow rate 1.0 0.75 0.5
(ml/min)
Gradient time 20 to 40 min
Detection Collect all wavelengths, if possible, otherwise 220,
254, 280 nm
Detection Circular dichroism at estimated appropriate
(optional) wavelengths
Detection Mass spectrometry if aqueous phases used are volatile
(optional)

If sufficient column chemistries have been screened, often

there are ‘hits’ which show some degree of separation, which can
then be taken to a standard system and optimized. Unfortunately
there are times when no resolution is achieved. In this case, as
sometimes happens with strongly acidic or basic compounds, it
may be beneficial to use an additive in the mobile phase. In general,
additives are not the first choice to use on a screening system as
many of the additives used in chiral screen have memory effects.
These compounds may react/functionalize/or in some way alter
the stationary phase with their use and as such the stationary
phase does not return to its original condition. This then throws
into question all subsequent screens on a system with that column.
A separation may work on the screening system but not offline
with a fresh column not functionalized in the same way, the column
on the screen system is. Figure 7.2 shows a chart for choosing
an additive for the chiral mobile phases, if needed.
 Compund to be
analyzed

Acidic Neutral Basic

Check pKa Start out with Check pKa

or determine no additives or determine
theoretically theoretically

Choose additive based on pKa, choose a range Choose additive based on pKa, choose a range
2 pH units lower: 2 pH units higher:
• TFA: pH range <2.5 • Triethanolamine: pH range 6.8 to 8.8
• Formic acid: pH range 2.8 to 4.8 • Diethanolamine: pH range 7.9 to 9.9
• Acetic acid: pH range <2.5 • Ammonia: pH range 8.2 to 10.2
• Triethanolamine: pH range <2.5 • Ethanolamine: pH range 8.5 to 10.5
Next Step—the Type of Chromatography

Figure 7.2 Flow chart to aid in choosing an additive for chiral screening.
169
170 Chiral Chromatography

7.3.5 Screening Success

Screening is used extensively in the pharmaceutical industry,
especially once a standard in-house screen has been established.
Such screens are used as they can rapidly find enantiomeric
separations for 60% to 80% of the compounds in a time frame
much faster than trial and error alone. Unlike for the case of
achiral method development where some knowledge of the
structure can lead to a good suggestion for a chromatographic
starting point based on previous knowledge and experience
with different stationary phases, the separation of chiral molecules
is too complex to intuitively see the structure and choose the
best chiral phase. Polysaccharide columns have historically had
the best success rates and are the most prevalent columns in
most in-house screens. Following are summaries of some of
the different column types available.

7.4 The Columns

As discussed above and in Chapter 6, the chiral column is key to
the separation of the enantiomers. There are several types of
columns to consider; however, the polysaccharide columns
(cellulose and amylose) are often the first type of column tried
as their technology is continuously improving; these types of
columns have become more rugged and have been adapted for
use with UHPLC, for instance, the sub 2 µm from Chiralcel (see
Chapter 6). Other types of columns available include the Pirkle
(or brush type) and antibiotic-based columns.

7.4.1 Polysaccharide-Type Columns

Several manufacturers have polysaccharide-type columns.
Currently Chiral Technologies has the greatest selection of phases
(see Chapter 6), but other companies such as Phenomenex and
Waters are also developing polysaccharide chiral columns. These
columns are available as both coated, where the chiral selector
is coated on to the silica phase, and as an immobilized phase,
where the chiral selector is physically bonded to silica. There are
advantages and disadvantages to both types of column. Since the
 The Columns 171

silica is only coated with the chiral phase, choice of chromatographic

solvent is key as solvents other than alcohols and acetonitrile can
wash the chiral phase off of the silica, destroying the expensive chiral
column. Alternatively, immobilized phases are more rugged and
can be used with a wider variety of organic solvents but tend to
have a lower density of the chiral selector. Since polysaccharide
columns are the most prevalent chiral columns currently used by
the pharmaceutical industry, they are available in a large range
of particle sizes ranging from sub 2 µm for analytical use up to
20 µm for preparative separations, and the corresponding column
dimensions [13, 14].
There are traditionally two types of polysaccharide columns:
one based on amylose and one based on cellulose. It is of note
that it is these sugars that are the true chiral selector and that the
different column selectivity is based on the different derivatives
of these polysaccharides. For example, tris (3-chloro-4-methyl-
phenylcarbamate) and tris (3,5-dimethyl-phenylcarbamate)
are widely used derivatives of amylose and cellulose for chiral
separations, but these functional groups themselves are not
chiral, the selectivity comes from the combination of the added
derivative and the polysaccharide. For a list of currently used
derivatives of amylose and cellulose, see Chapter 6.

7.4.2 Pirkle-Type Columns

These phases were developed as p-electron acceptors, p -electron
donors and chiral selectors with both, p-donor and p-acceptor
attributes. Several studies have been conducted with individual
compound and column investigations using Pirkle-type stationary
phases [15–23]. Normally, the same mobile phase approaches
that work with polysaccharide stationary phases also work with
Pirkle-type chiral stationary phases. Because of the commercial
success of the polysaccharide columns, many pharmaceutical
laboratories simply choose not to screen Pirkle-type columns
when looking for enantioseparations. This is unfortunate as unique
separation opportunities may be missed. Past performance of
stationary phases may not be indicative of future enantioresolutions
with new pharmaceutical entities. Thus, maintaining screening
capabilities with alternative stationary phase choices is prudent.
172 Chiral Chromatography

Because Pirkle-type chiral phases are covalently bound to the

chromatographic support, they exhibit excellent stability. Many
polysaccharide phases in use today are non-covalent coatings that
can be washed off the support material if the chromatographic
conditions are not carefully chosen. Pirkle-type chiral stationary
phases are stable from pH 2.5 to 7.5 and can be used in both
the normal-phase mode and the reverse-phase modes. An
important characteristic of the Pirkle-type chiral stationary
phases is their availability in both enantiomeric and/or opposite
diasteriomeric forms. This choice of enantiomeric forms allows the
chromatographer to manipulate the separation for optimization,
as often, for ease of quantitation, the goal is to have the enanatiomer
elute in front of the parent peak. For example, in case of an
S,S-configured chiral stationary phase, the elution order of
separated enantiomers can be reversed simply by switching to the
R,R-configured chiral selector.

7.4.3 Antibiotic Type Columns

These phases are based on the chiral selectivity of several
macrocyclic antibiotic molecules. The most common antibiotics
used as chiral selectors are as follows [24, 25]:
• Vancomycin and its analogs
• Teicoplanin and its analogs, including Teicoplanin aglycone
(TAG)
• Ristocetin and its analogs
All these phases can be used in reverse-phase, normal-phase,
and polar organic modes. Table 7.4 highlights the selectivities of
these three types of phases.

7.4.4 Columns for Ultrafast Chiral Separations

Now that ultra-high-pressure liquid chromatography (UHPLC) is
found in most laboratories, chiral column manufacturers have been
bringing more UHPLC-type chiral columns on the market, allowing
for faster chiral separations [15]. Sub-minute separations are
possible using fully porous particles (FPP); however, the bonding
of chiral phases to superficially porous particles (SPP) has shown
 The Columns 173

to be even more effective for sub-minute and even sub-second

enantiomeric separations [26–28].

Table 7.4 Selectivities of several antibiotic based chiral columns

Chiral Column
selector Modification(s) name Selective for
Vancomycin None Chirobotic V Neutral molecules, amides,
acids, esters and cyclic
amines
Teicoplanin None Chirobiotic T Underivatized amino
acids, N-derivatized amino
acids, carboxy cyclic acids,
phenols, small peptides,
aromatic molecules, and
aliphatic and cyclic amines
Teicoplanin Sugars Chirobiotic Amino acids and other
aglycone removed TAG amphoteric molecules,
cyclic amino acids, and
neutral molecules such as
diazopenes, hydantoin and
oxazolidinones
Ristocetin A Chirobiotic R Anionic analytes (i.e., α-
hydroxy acids, substituted
aliphatic acids), chiral
alcohols, secondary and
tertiary amines

Several instrument settings and configurations need to be

considered for such quick separations. From a data collection
standpoint, the collection rate on the detector(s) should be at
80 Hz, as a minimum. Extra column volumes need to be reduced
which can be accomplished by replacing high volume mixtures
and using low volume fittings.
Some columns used for ultrafast separations are commercially
available, while others require special functionalization of packing
materials. Table 7.5 gives an overview of columns used for ultrafast
separations.
174 Chiral Chromatography

Table 7.5 Overview of chiral columns used for ultrafast separations

Phase/chiral Commercially
Column separator available Reference

LarihcShell-P Isopropyl bonded AZYP LLC 28

cyclofrutan CF6-P (Arlington, TX,
USA)
CDShell-RSP 2-hydroxypropyl-β- AZYP LLC 24, 26
cyclodextrin (Arlington, TX,
USA)
TeicoShell Teicoplanin AZYP LLC 28
(Arlington, TX,
USA)

VancoShell Vancomycin AZYP LLC 26

(Arlington, TX,
USA)

NicoShell Modified macrocyclic AZYP LLC

glycopeptide (Arlington, TX,
USA)

Chiralpak IA-U Polysaccharide Chiralcel See Fig. 6.1,

and IC-U Chapter 6

(S,S) and (R,R) Pirkle type Regis

Whelk-O1

7.4.5 Other Phases for Chiral Separations

This chapter and Chapter 6 have explored the more traditional
and available chiral columns, but this is not an exhaustive list.
There are many older chiral selectors that have fallen out of use
but, nonetheless, with ever-complicated separations could be worth
looking at if all other columns tried do not lead to a satisfactory
separation [29]. Additionally, new and interesting chiral phases
are emerging based on new and/or modified technologies [30].
This is especially true for ultrafast chiral chromatography, so
keeping on top of new chiral column technologies is very useful
for chiral method development.
 References 175

7.5 Conclusions
As discussed in this chapter and in Chapter 6, the key to quick
chiral method development is first and foremost to find the
conditions and column to achieve the enantiomeric separation.
This can be simplified by screening multiple columns under
multiple conditions, and even further simplified with the use of
chiral detectors. Once conditions that work have been found, then
the method development can be taken off-line from the screening
system for optimization, which can be similar to achiral optimization
at this point. Different approaches and techniques have been
discussed, and all work well for certain compounds. It may take
several attempts to find the right conditions for separation of
enantiomers, but options are many and usually some separation of
enantiomers is found.

Acknowledgments
The authors would like to thank Zachary Breitbach for review of
the chapter and providing useful insights.

Disclosure
Laila Kott is an employee of Xenon. Gregory Webster is an
employee of AbbVie. The study contains no proprietary Xenon or
AbbVie data. Xenon and AbbVie jointly participated in writing,
reviewing, and approving and providing financial support for
this publication.

References

1. Snatzke. G. (1968) Circular dichroism and optical rotatory

dispersion—Principles and application to the investigation of
the stereochemistry of natural products, Angewandte. Chemie
International Edition in English. 7(1), pp. 14–25.
2. Kott, L., Holzheuer, W. B., Wong, M., Webster, G. K. (2007) An evaluation
of four commercial HPLC chiral detectors: A comparison of three
polarimeters and a circular dichroism detector, J. Pharmaceutical
and Biomedical Analysis. 43(1), pp. 57–65.
176 Chiral Chromatography

3. Huang, X., Nakanishi, K., Berova, N. (2000) Porphyrins and

metalloporphyrins: versatile circular dichroic reporter groups for
structural studies, Chirality. 12, pp. 237–255.
4. Kott, L., Petrovic, J., Phelps, D., Roginski, R., Schubert, J. (2014)
Determination of a low-level percent enantiomer of a compound
with no ultraviolet chromophore using vibrational circular dichroism
(VCD): Enantiomeric purity by VCD of a compound with three
chiral centers, Applied Spectroscopy. 68(10), pp. 1108–1115.
5. Matthijs, N., Maftouh, M., Heyden, Y. V. (2006) Screening approach
for chiral separation of pharmaceuticals IV. Polar organic solvent
chromatography. Journal of Chromatography A, 1111, pp. 48–61.
6. Matthijs, N., Maftouh, M., Heyden, Y. V. (2006) Chiral separation
strategy in polar organic solve chromatography and performance
comparison with normal-phase liquid and supercritical-fluid
chromatography, Journal of Separation Science. 29, pp. 1353–1362.
7. Morante-Zarceno, S., Sierra, I. (2012) Comparative HPLC methods
for β-blockers separation using different types of chiral stationary
phases in normal phase and polar organic phase elution modes.
Analysis of propranolol enantiomers in natural waters, Journal of
Pharmaceutical and Biomedical Analysis. 62, pp. 33–41.
8. Akin, A., Antosz, F. J., Ausec, J. L., Greve, K. F., Johnson, R. L., Magnusson,
L. E., Ramstad, T., Secreast, S. L., Seibert, D. S., Webster, G. K. (2007)
An orthogonal approach to chiral method development screening,
Current Pharmaceutical Analysis. 3(1), pp. 53–70.
9. Wong, M. M., Holzheuer, W. B., Webster, G. K. (2008) A Comparison
of HPLC and SFC chiral method development screening approaches
for compounds of pharmaceutical interest, Current Pharmaceutical
Analysis. 4(2), pp. 101–105.
10. Holzheuer, W. B., Wong. M. M., G. K. Webster, (2009) Reverse
phase chiral method development screening for compounds
of pharmaceutical interest, Critical Pharmaceutical Analysis. 5,
pp. 345–357.
11. Webster, G. K. (2014) The SFC market: ’Yesterday, today, and
tomorrow, in Supercritical Fluid Chromatography: Advances and
Applications in Pharmaceutical Analysis, ed., Webster, G. K., Chapter 1
(Pan Stanford Publishing, USA), pp. 1–14.
12. Schafer, W., Chandrasekaran, T., Prizada, Z., Zhang, C., Gong, X.,
Biba, M., Regalado, E. L., Welch, C. J. (2013) Improved chiral SFC
screening for analytical method development, Chirality. 25(11),
pp. 799–804.
 References 177

13. Zhang, T., Franco, P. (2010) Finding the best separation for
enantiomeric mixtures, LCGC North America. 28(9), pp. 818–822.
14. Woiwode, U., Neubauer, S., Kaupert, M., Lindner, W., Lammerhofer,
M. (2017) Trends in enantioselective high performance liquid
chromatography, LCGC Europe. (Special issue) 30(6), pp. 34–42.
15. Kennedy, J. H. (1996) Comparison of chiral separations on
polysaccharide chiral stationary phases to an improved Pirkle-type
phase, Journal of Chromatography. 725, pp. 219–24.
16. Siluveru, M., Stewart, J. T. (1996) Stereoselective determination of
R(-)- and S(+)-prilocaine in human serum using a brush-type chiral
stationary phase, Journal of Pharmaceutical and Biomedical Analysis.
15, pp. 389–92.
17. Zhang, X., Ouyang, J., Yang, Y. (2001) A simple method for chiral
separation of ephedrines using (R)-1-naphthylglycine and 3,5,-
dinitrobenzoic acid as stationary phase, Analytical Letters. 34,
pp. 1851–1865.
18. Shao, B.-H., Xu, X.-Z., Wu, Q.-Z., Lu, J.-D., Fu, X.-Y. (2005) Comparative
enantioseparation of 2-arylpropionic acid esters on cellulose
derivative and (S,S)-Whelk-O 1 columns, Journal of Liquid
Chromatography & Related Technologies. 28, pp. 63–80.
19. Shao, B.-H., Xu, X.-Z., Wu, Q.-Z., Lu, J.-D., Fu, X.-Y. (2003) Comparison
of enantioseparation and chiral recognition mechanism of racemic
naproxen esters on (S,S)-Whelk-O 1 and CDMPC chiral columns,
Acta Chimica Sinica. 61, pp. 1635–1640.
20. Madhavan, P., Rao, B. M., Pravin, A., Kumar, P. R., Screenivasulu,
M., Chandrasekhar, K. B. (2007) A validated chiral HPLC method
for the determination of enantiomeric purity of R-β-amino-β-(4-
methoxyphenyl) propionic acid, Chromatographia. 65, pp. 81–84.
21. Szczerba, T. (2011) Summarization of screening hits on the
Whelk-O® 1, RegisPack and RegisCell chiral stationary phases (CSPs),
LCGC North America. (Suppl.), 24 (February) p. 24.
22. Holzheuer, W. B., Wong, M. M., Webster, G. K. (2009) Evaluation
of Pirkle-type stationary phases in chiral method development
screening for compounds of pharmaceutical interest, Critical
Pharmaceutical Analysis. 5, pp. 10–20.
23. McDevitt, T. F., Vicente, G., Webster, G. K., Szczerba, T. (2009) Column
coupling to solve a challenging API separation using the Whelk-O®
1 chiral stationary phase, LC/GC Application Notebook. September
2009 p. 23.
178 Chiral Chromatography

24. Lee, D., Web, M. L. (2009) Chiral analysis of pharmaceuticals, in

Pharmaceutical Analysis, eds., Lough, W. J., Chapter 3 (Wiley & Sons,
Oxford, UK), pp. 88–90.
25. Sigma-Aldrich (Millipore Sigma), https://www.sigmaaldrich.com/
analytical-chromatography/hplc/columns/chiral/chirobiotic.html
(accessed November 13, 2017).
26. Patel, D. C, Brietbach, Z. S., Wahab, M. F, Barhate, C. L., Armstrong,
D. W. (2015) Gone in seconds: Praxis, performance, and peculiarities
of ultrafast chiral liquid chromatography with superficially porous
particles, Analytical Chemistry. 87, pp. 9137–9148.
27. Wahab, M. F., Wimalasinghe, R. M., Wang, Y., Barhate, C. L., Patel, D. C.,
Armstrong, D. W. (2016) Salient sub-second separations, Analytical
Chemistry. 88, pp. 8821–8826.
28. Barhate, C. L., Joyce, L. A., Makarov, A. A., Zawatsky. K., Bernardoni,
F., Schafer, W. A., Armstrong, D. W., Welch, C. J., Regalado, E. L. (2017)
Ultrafast chiral separations for high throughput enantiopurity
analysis, Chemical Communications. 53, pp. 509–512.
29. Webster, G. K., Kott, L. (2011) Method development for pharmaceutical
chiral chromatography, in Handbook of Modern Pharmaceutical
Analysis, eds., Ahuja, S., Scypinski, S., Chapter 7 (Elsevier, San Diego),
pp. 251–282.
30. Brown, L., Webster, G. K., Kott, L., Rao, N. K. R., Luu, T. A., Shah, R.,
Henriques, L. (2011) Novel coated cellulose carbamate stationary
phase to enhance selectivity of compounds of pharmaceutical
interest, Pharmaceutica Analytica. 2, pp. 2–7.
Chapter 8

Aqueous Normal-Phase Chromatography

Using Type C Silica Columns

Joseph J. Pesek and Maria T. Matyska

Department of Chemistry, San Jose State University,
San Jose, California 95192, USA
[email protected]

This chapter presents the general principles of aqueous normal-

phase (ANP) chromatography, a technique that has features of both
reversed-phase (RP) and hydrophilic interaction chromatography
(HILIC). It is a unique chromatographic format and while some
types of stationary phases exhibit a limited ANP capability, it is
silica hydride-based materials that are most successful in this mode
and have the broadest range of applications. Thus, this review
will focus exclusively on the fundamental aspects of ANP, the
uniqueness of the silica hydride material and review applications
utilizing both the RP and ANP capabilities of these columns.

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
180 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

8.1 Introduction
A variety of approaches and stationary phases are available for the
separation of mixtures or analyses of specific compounds based
on the chemical and structural properties of the analyte. A large
fraction of current chromatographic methods utilize a format that
revolutionized separation science more than five decades ago;
chemically bonded stationary phases that are fabricated on silica
as a support material with hydrophobic moieties attached to the
surface. Columns made from these materials are referred to as
reversed-phase (RP) and the most common bonded group is the
straight chain C18 hydrocarbon. Other chain lengths between C4
and C30 have demonstrated the same type of retention based on
hydrophobic interactions with the degree of retention based on
the number of carbon atoms, the amount of the bonded phase and
the properties of the analytes. Other types of bonded moieties with
significant hydrophobic character can also provide RP retention and
many have been synthesized over the years but only a relatively
small number have merited commercial development. While other
contributions to retention can occur on these phases, very often
adsorption effects from residual silanols on the silica surface, the
hydrophobic effects are usually the dominant factor in these types
of chromatographic columns.
However, the analysis of polar (hydrophilic) compounds has
always presented a serious challenge due to the low retentivity
on typical RP stationary phases. In this case, the typical approach
for many analyses was to derivatize the analyte to make it less
polar so that it could be retained on a typical RP column. Only a
few stationary phases were polar enough to provide significant
retention for these types of compounds. The most common being
ion-exchange materials which could be used for analytes possessing
either a permanent positive or negative charge. This could often be
achieved by adjusting the pH of the mobile phase or dealing with
the salts of the parent compound. The relatively few good options
for polar compounds provided opportunities for new approaches
and materials to deal with these challenging types of analytes.
In recent years, there have been two types of chromatographic
materials developed more specifically designed for the analysis of
 Silica Hydride 181

hydrophilic compounds: mixed-mode and hydrophilic interaction

liquid chromatography (HILIC) stationary phases [1, 2]. In the
mixed-mode format, groups are bonded to the surface that have
some polar components as part of the bonded group as well as
a hydrophobic alkyl chain. Thus, polar compound retention is
determined by the nature of the functional group on the bonded
moiety. In HILIC, the most common phases are silica or silica with
a bonded group containing a polar functionality like cyano, amino
or diol. In this case, retention is due to a water layer that forms
on the surface of the stationary phase that results in a partitioning
effect between the adsorb water layer and the organo/aqueous
mobile phase. The mixed-mode phases often have limited
applicability and retention and the HILIC materials have other
issues that will be described in more detail later when specific
properties of the silica hydride phases are discussed.

8.2 Silica Hydride

There have been numerous reports in the literature describing
a broad array of stationary phases based on ordinary silica [1, 2].
The general approach to attaching an organic group to a silica
surface utilizes an organosilanization reaction. The organosilane
reagent can be either monofunctional or trifunctional which leads
to a single attachment to the silica surface or the formation of a
polymer layer on the support material. While other approaches
such as monoliths or different materials like zirconia or titania
have been developed over the years, they have not significantly
replaced ordinary silica as the predominant source for HPLC
stationary phases. Although some desirable properties of these
alternative materials have been found, the positive features are
often negated by other factors that have precluded them from
displacing silica except in a limited number of applications.
However, over the last two decades, one material has been
documented in the literature to have significantly different
properties than ordinary silica but still retains many of its good
features such as high surface area, excellent mechanical strength,
and the ability to fabricate the material in a broad range of particle
182 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

sizes. That material is silica hydride [3–7]. The essential difference

between ordinary silica and silica hydride is shown in Fig. 8.1.
The underlying structure in both materials is a polymeric series
of silicon-oxygen linkages. However, the surface of ordinary silica is
populated by polar silanols while silica hydride contains virtually
only mildly hydrophobic Si-H moieties. This significant difference
can be observed in the chromatographic properties of stationary
phases fabricated on these two support materials. However, a
simple example of the nature of the silica hydride surface is
shown in Fig. 8.2. In this case, simple nonpolar molecules are
retained on unmodified silica hydride using simple reversed-
phase conditions. These results illustrate the mildly hydrophobic
nature on the base material used for the fabrication of stationary
phases that are further modified.

Figure 8.1 Comparison of the surface structures of ordinary silica and

silica hydride.

The first important property of silica hydride that parallels a

desirable aspect of ordinary silica is its ability to be modified by a
simple and well-known chemical reaction. In the case of silica, it
is organosilanization. For silica hydride, the comparable reaction
is hydrosilation that involves the addition of Si-H to an olefin or
alkyne [4]. Thus, bonding to the surface is through a stable silicon-
carbon bond that provides durability and long lifetime to these
materials. Once the material was fabricated and through many
 Silica Hydride 183

prior investigations, a unique and distinctive property of all silica

hydride-based stationary phases is their ability to operate in both
the reversed-phase and normal-phase modes [8, 9]. However, in
this case normal-phase refers to the retention of polar/charged
compounds using a mobile phase containing at least a minimal
amount water. Hence, the term aqueous normal phase(ANP) is
used to describe this property. This mechanism is in contrast to
organic normal phase where the retention of moderately polar
compounds occurs using organic solvents in the mobile phase.
It should be noted that another interesting feature of silica hydride
stationary phases is that they can also function under organic
normal-phase conditions [10]. Their slight hydrophobic nature
creates a different type of material than the one most often used
in this separation mode: ordinary silica.

Figure 8.2 Chromatograms showing the retention of small hydrophobic

compounds on silica hydride using reversed-phase conditions.

An important clue to the nature of silica hydride has been

provided primarily by Jandera and coworkers, who measured
the amount of water on the surface of various chromatographic
materials [11]. A variety of silica hydride modified stationary phases
184 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

showed less than one-half of a monolayer of water on the surface.

This is contrast to ordinary silica and various modified silicas
where there was anywhere from 4 to 10 monolayers of water
on the surface. This fundamental difference between ordinary
silica and silica hydride materials leads to important variations in
chromatographic properties when comparing columns fabricated
on the two supports. There are at least two important
consequences of the difference in the amount of surface water. In
the instances where variations in the water layer occur, such as
gradients, this can often be a time-consuming process requiring
many column volumes before the stationary phase returns to
its original configuration. Since this is not an issue with silica
hydride, the time required to re-equilibrate the column is often
significantly shorter than with ordinary silica phases. The lack
of water on the surface also facilitates operation in the organic
normal-phase mode. On silica and other materials used for this
chromatographic mode, water can drastically affect retention of
analytes. Therefore, mobile phases must often be scrupulously
dried in order to achieve reproducible retention due adsorption
of water on the surface. Since silica hydride has a very low affinity
for water, its presence in the mobile phase has little effect on its
retention properties.
The absence of a water layer on silica hydride stationary
phases explains many of its unique properties. However, the one
feature of these materials that was difficult to explain involved
the retention of hydrophilic compounds. Based on a number of
different analytical methods, the surface was shown to contain no
more than 5% residual silanols [7]. In addition, the chromatographic
results described above demonstrate the general hydrophobic
nature of the silica hydride surface. The first step to formulating
an explanation for this feature of silica hydride materials was to
determine the physical state of the surface. This was accomplished
by measuring the zeta potential of the stationary phases [12]. The
zeta potential determines the net overall charge on the surface
and the region within a few nanometers. Surprisingly the
measurements showed a very high negative zeta potential,
comparable to ordinary silica, for the silica hydride materials. Since
the presence of silanols could not explain this property, another
theory was sought to account for the substantial negative charge
 Silica Hydride 185

and hence the retention of hydrophilic compounds. Because the

overall surface of a silica hydride particle is hydrophobic, a search
of the literature determined the best model for this material was an
oil droplet suspended in an organo/aqueous liquid environment.
A number of investigations have demonstrated that under these
conditions the droplet has a strong negative charge on the surface
[13–15]. The source of the negative charge was determined to be
hydroxide ions from the autodissociation of water [16]. Thus,
retention of polar compounds is based on either charge attraction
or displacement of hydroxide or other adsorbed mobile phase
components like buffers or additives.
The above description of the surface characteristics of silica
hydride and in particular the mechanism for the retention of polar
compounds, provides a sharp contrast to a currently common
method for their analysis, HILIC. In HILIC the compounds are
separated due to the water layer on the surface of the stationary
phase [17]. These water layers are comparable to those described
above for silica-based materials which are the base for many
HILIC stationary phases. Even HILIC columns in other formats
like polymeric materials still rely on the substantial water layer
at the surface. This layer acts as a pseudo stationary phase that is
more polar than the bulk mobile phase. Therefore, hydrophilic
compounds preferentially partition into this layer and are thus
retained on the column. The contrast between an adsorption type
mechanism on silica hydride and the partition mechanism in HILIC
have some very important consequences. For example, retention
generally increases on HILIC phases when the concentration of
additives in the mobile phase increases. However, just the opposite
effect is noticed for silica hydride, i.e., retention decreases as
additive concentration increases. The practical result is particularly
relevant when using mass spectrometry for detection. High
additive concentrations (some methods require from 20 to
100 mM) in the mobile often lead to deposits forming on the
ionization source that must be cleaned more frequently as well as
producing inconsistent ionization. Silica hydride almost exclusively
operates in the 5 to 15 mM additive concentration range. The
other consequence of the difference in mechanisms is the re-
equilibration time after gradient runs. Gradient runs are necessary
when complex samples with numerous analytes that are frequently
encountered in biological, clinical, pharmaceutical, forensic,
186 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

food and environmental determinations. At the end of the gradient

run, it often takes numerous column volumes (sometimes 20 or
more) to bring the stationary phase (water layer) back to the
original conditions at the start of the gradient. For silica hydride
materials, generally only two or three column volumes are
necessary to achieve the original conditions at the start of the
gradient. Practically, this means that many more analyses can
be done over a given time period and considerably less solvents
are used in the process.
The features described above de monstrate that silica hydride
and silica hydride-based stationary phases are unique separation
materials. They are clearly different from the most closely related
materials; ordinary silica and silica modified using traditional
organosilanization reactions. A number of advantages have
already been cited. These will be reiterated to some extent in the
examples given in the following sections as well as other features
that provide new approaches for the separation and analysis of
samples ranging from the determination of a single component
to multicomponent mixtures in both simple and complex matrices.

8.3 Applications of Silica Hydride-Based

Separation Materials
8.3.1 Reversed-Phase
As shown in the example above, bare silica hydride without any
further modification through hydrosilation possesses RP retention
properties for typical small hydrophobic molecules. As expected,
modification of silica hydride with nonpolar organic moieties
further enhances the reversed-phase capabilities of this material.
Thus, any silica hydride-based stationary phase can be used for RP
applications, but of course, the more hydrophobic attached organic
moieties provide the greatest RP retention.
Some typical examples of RP analyses done on a C18 modified
silica hydride material are shown in Fig. 8.3. In Fig. 8.3A, two
hydrophobic components in a sun screen formulation are analyzed
on this C18 column. The structures of these compounds are also
shown in the figure. The chromatogram demonstrates the high
efficiency and excellent peak shape that can be achieved in the
 Applications of Silica Hydride-Based Separation Materials 187

RP mode with silica hydride-based columns. Another feature of

these materials is also highlighted in this figure and that is the
high degree of reproducibility from batch to batch in the
fabrication of the stationary phase. A second example on this
same C18 column is shown in Fig. 8.3B. In this analysis of an
acetaminophen formulation, there is the primary pharmaceutical
component and two other minor constituents. Based on the
separation shown, both acetaminophen as the primary component
as well as caffeine and butalbital could be quantitated as needed
to assure the quality of the marketed product. In both examples
shown, peak shape is excellent that demonstrates the value of
the silica hydride surface being slightly hydrophobic and essentially
free of silanols for RP applications.

Figure 8.3 Examples of reversed-phase separations on a C18 modified

silica hydride stationary phase using an acetonitrile/water mobile phase.
(A) Sunscreen components oxybenzone (1) and octinoxate and (B)
pharmaceutical formulation containing acetaminophen (1), caffeine (2),
and butalbital (3).
188 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

A variety of applications have been reported on silica hydride

stationary phases that are primarily focused on the analysis of
hydrophobic compounds. One example relevant to forensic science
is the analysis of drugs of abuse [18]. Two types of silica hydride
columns were tested: C18 and phenyl. Both stationary phases
were successful in separating 17 common drugs of in less than
9 min. However, selectivity for the two columns was different, so
the proper choice for a particular analysis would depend on the
target compounds. The same combination of columns was used to
analyze the phenolic components in mesquite flour again showing
that both were adequate but had different selectivity [19]. The
hydride-based phenyl phase has also proved useful in the analysis
of mycotoxins [20]. Bean, rice, maize, and wheat samples were
stored under conditions conducive to fungal development and
the protocol developed could detect a variety of mycotoxins in the
matrices. The phenyl column has also been used to analyze for the
phenolic composition in pomegranate [21]. Another interesting
application is the determination of amoxicillin using a USP method
on a silica hydride C18 column [22]. The results showed both
improved reproducibility and peak shape using this column format.
Food analysis is becoming an increasingly important field for
HPLC that can be used to determine quality, nutrients and
authentication for certain high cost or rare products. One example
is the analysis of capsaicinoids in various hot sauces which are
responsible for the spicy flavor in this product as well as in
peppers [23]. The determination was made using both 2 µm and 4
µm C18 silica hydride columns. Another example of food analysis
using the same type of C18 column was the determination of
trans-resveratrol in wine [24]. The silica hydride column performed
better than a standard silica C18 column for the separation of the
target analyte from matrix components so that good quantitation
in real samples was obtained. The silica hydride-based C8 column
was used for the analysis of bisphenol A (BPA) in paper products
such as carbon paper and store receipts [25]. BPA can be transferred
by contact with the surface of these materials and has been
shown to be responsible for a number of physiological problems.
The protocol reported that included a sample preparation
procedure resulted in good retention and excellent reproducibility
so that both identification and quantification of BPA can be
done utilizing UV detection. An interesting application in the RP
 Applications of Silica Hydride-Based Separation Materials 189

mode has been reported using a capillary column format. The

analysis of nonsteroidal anti-inflammatory drugs and steroids was
reported using 50 and 100 µm columns packed with the C18 silica
hydride material [26]. A fast analysis (<3 min) having excellent
resolution for multicomponent mixtures was developed for a broad
range of these compounds.
One additional aspect of the RP uses for the silica hydride
materials is for preparative applications. One such example of a
viable format is one that was developed for the preparation
of carotenoids [27]. Carotenoids have been used as dietary
supplements and have been purported to have significant health
benefits. Another potential expansion of silica hydride phases
would be for the separation of chiral compounds. A preliminary
study utilizing a vancomycin stationary phase proved that
enantiomeric separation formats can be developed using silica
hydride as the support material [28]. Future investigations could
focus on different types of chiral selectors to be bonded to the
silica hydride surface.

8.3.2 Use of Aqueous Normal Phase for Polar

Compounds
Because there are often limited options for the analysis of polar
compounds or those that are available such as HILIC have
some substantial drawbacks, ANP utilizing silica hydride-based
stationary phases is becoming a more common technique [29].
HILIC, for example, is often limited by reproducibility of retention
times on both an intraday and interday basis, column lifetime
(robustness of the stationary phase), long equilibration times
for gradient methods, and the need for high concentrations of
additives to the mobile phase that make MS detection more difficult.
ANP on silica hydride materials is virtually free of all of these
chromatographic problems.
A dramatic example of the difference between ordinary
silica-based materials and silica hydride is shown in Fig. 8.4.
The compound analyzed in the chromatogram is methotrexate, an
extremely polar drug. The stationary phase used in this analysis is
the silica hydride C18. Thus, even though the amount of stationary
phase on the surface is similar to those phases produced through
organosilanization, it still retains the capability to operate in the
190 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

normal-phase mode. This is a unique feature of the silica hydride

columns. All phases can operate in either the RP or ANP modes,
and in this case even if the surface is modified with a very
hydrophobic moiety, it still retains normal-phase retention
capabilities. The retention for this compound is quite strong since a
very steep gradient that ends at 20% water (strong solvent) is need
to elute the compound in a reasonable period of time. Under these
experimental conditions, peak shape is excellent and quantitative
measurements, if needed, could easily be done. A similar result
on a stationary phase made on ordinary silica using organosilane
chemistry modification would result in little or no retention of
this compound. Another highly polar drug, tobramycin, has similar
behavior on hydrophobic silica hydride columns, even when using
methanol as the weak solvent.

Figure 8.4 Retention of the polar compound methotrexate in the aqueous

normal phase mode on a C18 modified silica hydride column.
 Applications of Silica Hydride-Based Separation Materials 191

While these are two interesting examples and illustrate the

unique behavior of the silica hydride phases, most analyses of
polar compounds utilize other phases with less hydrophobic
modifications, or in some cases very little or no modification at
all. The charged nature of the silica hydride surface as determined
by the zeta potential measurements cited above provide the
properties necessary for the retention of hydrophilic compounds.
The first significant report on the use of a silica hydride phase in
ANP focused on metabolites such as amino acids, organic acids,
and carbohydrates [30]. The stationary phase used is a material
referred to as the diamond hydride (DH); essentially a silica hydride
surface with only about 2% bonded carbon. Mass spectrometry
(MS) was used as the means of detection since most clinical
and pharmaceutical analyses for these compounds by HPLC are
increasingly done by MS. The advantage of MS detection is that
only isobaric compounds need to be separated in order to
determine the target analytes in the method. The requirements
for separation can be further reduced by using accurate mass
time-of-flight (TOF) or tandem mass (MS/MS) analyzers. Another
study using silica hydride columns reported on the analysis of
both hydrophobic and hydrophilic metabolites in serum [31]. An
important finding was that ion suppression occurred in a smaller
range of k values on the silica hydride columns than for traditional
columns whether the analysis was done in a single column or
tandem column format. Other metabolite studies have included
the monitoring of pathogens in cells [32] as well as the effect of
immunosuppressant drugs on cell pathogens [33]. It was also
possible to discern difference in metabolites found in the urine and
saliva of normal individuals and cancer patients [34]. Metabolic
profiling has also been used in the investigation of biomarkers
for bladder cancer [35]. Another interesting metabolic study
showed that elevated levels of hydrophilic sulfur amino acids
could be correlated with an increased risk cardiovascular disease
[36]. The ANP with MS/MS detection has excellent reproducibility
and quantitation in blood plasma. A sample calibration curve
(homocysteine) is shown in Fig. 8.5. An ANP method using the DH
column was used to identify aspartate as the source of nitrogen
for tuberculosis bacillus [37]. Metabolites associated with malaria
have also been studied using an ANP protocol [38]. The DH column
was also shown to be useful for global untargeted as well as
192 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

targeted metabolomic analysis [39]. Microbiota were fingerprinted

according to their small molecule constituents using an ANP
format. The role of succinate in the metabolism of Mycobacterium
tuberculosis was determined using an ANP method on the DH
column [40]. With this information, a subsequent investigation
determined that para-aminosalicylic acid could be used for an
effective treatment of tuberculosis [41]. An important class of
metabolites that often are difficult to analyze by HILIC and other
methods are nucleotides. Both the DH and a weak cation phase,
UDA (based on undecanoic acid) have shown to be effective for
the retention and separation of these compounds in the ANP
mode [42]. A comprehensive comparison of HILIC and ANP for
metabolomics was made and the advantages of the silica hydride
materials documented [43].

Figure 8.5 Calibration curve for homocysteine in blood plasma using the
diamond hydride stationary phase in the aqueous normal phase mode.
Reprinted with permission from ref. 36.

While ANP is now an established technology in metabolomics,

it has also proven to be effective in a variety of other analyses
involving hydrophilic compounds. For example, thiopurines that
are not readily analyzed by RP methods have been shown to be
potential biomarkers for some physiological conditions. These
compounds are easily retained and effective separations can be
obtained utilizing ANP formats [44]. Another potential clinical
application involves the analysis of the two isobaric compounds
 Applications of Silica Hydride-Based Separation Materials 193

symmetric and asymmetric dimethylarginine. These compounds

have been associated with a number of renal and cardiac
physiological conditions. Using exact mass TOF for detection a
separation protocol was developed for their analysis since both
compounds typically appear in any physiological sample [45].
Another approach is to use partial separation and a single quad
MS with in-source fragmentation to identify and quantitate the
two species. The actual separation of the two compounds (shown
in Fig. 8.6) done under isocratic conditions has not been reported
previously and was accomplished using an experimental silica
hydride material with 2-acrylamido-2-methylpropane sulfonic
acid (AMPS) bonded to the surface. Ethyl glucuronide (EtG) is used
as an indicator of long-term alcohol abuse and clinical testing for
it is common for public safety officers, airline pilots, and pregnant
women. A simple viable format for EtG was developed using
either the DH or UDA phases [46]. Another interesting application
is the use of ANP technology for the analysis of various acidic
lipids in brain and kidney extracts [47]. Another application in
the area of medical analysis is the determination of PEG 400 that
is used for intestinal permeability studies [48]. An evaporative
light scattering detection protocol was developed due to the high
salt content of the perfusate samples.

Figure 8.6 Separation of asymmetric and symmetric dimethylarginine on

an experimental AMPS column in the aqueous normal phase mode.
Reprinted with permission from ref. 45.
194 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

Figure 8.7 Analysis of fructose (1) and glucose (2) in various beverages
using the Diamond Hydride column in the aqueous normal phase mode.
(A) Cola, (B) red wine, and (C) grape juice. Reprinted with permission
from ref. 51.

A growing area that is becoming more reliant on HPLC and

LCMS methods is food and beverage analysis. Folic acid is found
in many food products, including juices and cereals. A protocol
 Applications of Silica Hydride-Based Separation Materials 195

including sample preparation and quantitation for these matrices

has been reported using the DH stationary phase [49]. Histamine,
a compound that can cause allergic reactions in many people, can
be found in wine, tuna, and cheese. A DH format at levels resulting
in physiological problems was designed for analyses in each of
these matrices [50]. Sugar levels in many food products are an
essential measurement for both quality and taste. A simple method
for fructose and glucose that is also useful for other sugars using
refractive index detection on a silica hydride amide column has
been reported [51]. Some beverage analyses are shown in Fig. 8.7.
Another type of sample where quality and regulation are
increasingly important is dietary supplements. A number of
investigations have been reported for these products which
are generally available over the counter. The compound 1,3-
dimethylamine, a component in dietary supplements, and a
number of similar hydrophilic species have been analyzed using
the DH column [52]. The substantial number of articles in the
food and beverage area utilizing ANP formats were described in a
review [53]. A silica hydride material was also used in a sample
preparation mode in the analysis of hydrophilic components in
dark chocolate [54].
Methods for the analysis of hydrophilic pharmaceutically
and related compounds have been reported based on the use of
silica hydride phases [55]. The powerful antibiotic, cycloserine,
was analyzed on a DH column [56]. The parent compound as well
as it degradation product and two impurities were successfully
determined with MS detection. The drug zanamivir, used for
treating influenza A and B, was analyzed in human serum using
several different columns [57]. The DH provided the best retention
and results under identical experimental conditions (Fig. 8.8). This
investigation also documented the reproducible retention times for
the drug on the DH vs. a HILIC column where tR varied from run to
run.
The versatility of the silica hydride phases is illustrated
(Fig. 8.9) by another application where even the hydrophobic Zoloft
with a log P value of 2.9 is retained under ANP conditions with a
gradient going from 95% to 50%, acetonitrile, over 6 min [58].
This same effect is seen in the analysis of peptides on the DH
column. A successful separation has been reported for a mixture
196 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

containing both hydrophobic and hydrophilic peptides [59].

Therefore, it should be possible to analyze an enzymatic digest in
a single run rather than the two-dimensional approaches currently
in use. A potential clinical analysis of the peptide bradykinin in
saliva has been demonstrated [60]. Bradykinin has been shown
to reduce blood pressure and a simple analysis in saliva, an easily
obtained physiological fluid, would make clinical determinations
more expedient. A series of collagen crosslinks having a complex
amino acid-like structure were analyzed on the DH column [61]
(Fig. 8.10). These collagen crosslinks were analyzed in skin, bone,
and tendons. The studies could be a precursor to understanding
normal and diseased tissue in a variety of abnormalities including
cancer as well as providing biomarkers for clinical analyses.


Figure 8.8 Retention of the drug zanamivir under optimal conditions
on several commercially available HPLC columns. (a) Scherzo SM C18,
(b) Agilent Zorbax SB-Aq, (c) Luna HILIC, (d) Cogent Bidentate C18 and
(e) Cogent Diamond Hydride. Reprinted with permission from ref. 57.
 Applications of Silica Hydride-Based Separation Materials 197

Figure 8.9 Simultaneous retention of a hydrophobic (sertraline) and

a hydrophilic (phenylglycine) compound under aqueous normal
chromatography conditions (linear gradient from 95% to 50% acetonitrile)
on the diamond hydride column.

8.3.3 Organic Normal Phase on Silica Hydride

This aspect of silica hydride technology has not been extensively
explored to date but does show promise to alleviate some of the
problems associated with conventional organic normal-phase
chromatography. The main problem in conventional organic
normal-phase chromatography, which typically uses ordinary silica
as the stationary phase, is associated with water. Because silica
198 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

Figure 8.10 Extracted ion chromatograms of crosslinks from deer skin

analyzed on the Diamond Hydride column under aqueous normal
phase conditions. Solutes: DHNLH = dihydroxylsinonorleucine;
HLNL = hydroxylsinonorleucine; LNL = lysinonorleucine; HHL =
histidinohydroxylsinonorleucine; 530.3276 = unidentified crosslink;
Des = desmosine; 427.2439 = unidentified crosslink; HHMD =
histidinohydroxymerodesmosine; 556.2031 = unidentified crosslink.
Reprinted with permission from ref. 61.
 Method Development Strategies on Silica Hydride Phases 199

is so hydroscopic, it readily absorbs water from any source. The

most common source is small traces of water in the mobile phase.
With the amount of water on the surface of silica being variable
due to the exact composition of the mobile phase, there often are
observed variable retention times in both inter- and intra-day
measurements. To somewhat overcome this problem, mobile phase
solvents are generally rigorously dried to try to maintain a constant
water content on the stationary phase. Because silica hydride is
mildly hydroscopic, it has low affinity for water as demonstrated in
the water layer determinations cited above. Therefore, the limited
studies available do not utilize any drying of the mobile phase
solvents and no significant variations in retention time on either a
short- or long-term basis were observed.
There are two studies on the use of silica hydride-based
stationary phases in organic normal phase. The significant feature
in both of these studies is that excellent results were obtained with
modified silica hydride surfaces. In one report, the pharmaceuticals
loratadine and carvone are separated on a C18 silica hydride
column using a gradient using hexane and dichloromethane as the
organic solvents [62]. The gradient goes from 95% to 50% hexane
in 6.5 min. A more extensive study was published that documented
the retention of various substituted phenols on several silica
hydride columns [10]. In this study, several different types of
organic solvents were tested as well as modified and unmodified
silica hydride materials. In all cases, it was possible to obtain
organic normal-phase retention and none of the solvents were
dried before use. Based on this data, it seems possible that this
mode could be used for both analytical and preparative
applications.

8.4 Method Development Strategies on Silica

Hydride Phases
The utility of these columns is found in the wide range of solvents
that can be used with them and the unique and strong retention
they can have for polar compounds while at the same time
retaining nonpolars. Selection of the column (C18, C8, cholesterol,
or diamond hydride) will be determined by the compounds to
200 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

be analyzed. Mixtures that will be highly polar and do not contain

nonpolar compounds might be better suited to be analyzed by
the diamond hydride, diol, or amide where mixtures with both
polar and nonpolar compounds might be better suited to the C18,
C8, cholesterol, or phenyl columns. The following sequence of
steps can be used as a generic method development strategy to
provide a viable analytical protocol for silica hydride stationary
phases. Further modification of results of these steps can be done
to formulate a validated method.

8.4.1 Mobile Phase Selection

The first step is to properly condition the column. This often
depends on the type of sample that will be analyzed. For urine,
plasma, or other biological samples, conditioning with 80%
acetonitrile/DI water/0.2% acetic acid (or 0.1% formic acid)
is recommended for 20 min. Cleaning with 50% MeOH/50%
water is recommended for every 10 injections for 5 min. If
standard samples are used, conditioning with 80% acetonitrile/DI
water/0.2% acetic acid (or 0.1% formic acid) is recommended
for 20 min. Subsequently cleaning with 50% MeOH/50% water
is recommended every 100 injections or sooner for 5 min if a
change in performance is noticed. After the column has been
installed and conditioned, it is a good idea to start with a typical
gradient run. This usually would involve an acidified mobile
phase of water and acetonitrile. Acidify the mobile phase with up
to 0.5% of an acid such as formic or acetic acid. If MS is not used
as a detector, trifluoroacetic acid (TFA) is another good candidate
as an additive.

8.4.2 Equilibration for Reversed-Phase

For columns that can operate in both modes or if only a reversed-
phase method is being considered, then equilibration with a
suitable mobile phase is necessary before beginning the method
development process. Typically this would be a mobile phase of
95:5 water/organic solvent containing any additives (formic acid,
acetic acid, TFA, etc.) to be used in the final method. The equilibration
should utilize approximately six column volumes of mobile phase
before the first sample injection.
 Method Development Strategies on Silica Hydride Phases 201

8.4.3 Reversed-Phase Gradient

In this step, a shallow gradient is run for 95% water to 40% water
over 20 min on a 75 mm or longer column. For longer columns, the
gradient should be increased proportionately. This shallow gradient
will provide an overall assessment of the type of compounds present
in a mixture and can be used to optimize gradient or isocratic
conditions for finalizing and analytical method or protocol. For
example, if less retention or sharper peaks are required, then a
steeper gradient from the same starting and final mobile phase
compositions can be created.

8.4.4 Equilibrate for Aqueous Normal Phase

Generally, at the end of a reversed-phase run the mobile phase
will already contain a high amount of acetonitrile. At this point,
the acetonitrile content in the mobile phase should be increased
to 100%. The pure acetonitrile should then be passed through
the column for at least 2 min with a flow rate of between 0.5 and
1.0 mL/min.

8.4.5 Aqueous Normal-Phase Gradient

Using the same mobile phase utilized for the reversed-phase
gradient, set up a gradient that runs from 90% to 40% acetonitrile
over 20 min for a 75 mm column. For a longer column, adjust the
gradient appropriately. As in the reversed-phase protocol, this type
of gradient allows for the determination of either an optimum
gradient or isocratic mobile phase composition. For sharper peaks
and less retention, a steeper gradient encompassing the same
starting and end points in mobile phase composition should be
used.

8.4.6 Compare Reversed-phase and Aqueous

Normal-Phase Data
The results from the two gradient runs should be compared to
determine the appropriate approach for analysis of the sample
tested. From these results it can concluded that either the reversed-
phase or aqueous normal-phase approach will provide the desired
202 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

separation required. In some cases, the comparison will demonstrate

that an isocratic mobile phase can retain and separate both polar
and nonpolar compounds.

8.5 Conclusions
Silica hydride-based separation materials are being recognized
as an alternative to traditional silica stationary phases.
Its unique properties provide chromatographic advantages that
can be exploited for improved and more efficient separations. The
low water content is one feature that has the broadest impact on
differentiating silica hydride from ordinary silica. The resulting
mildly hydrophobic surface surprisingly retains both polar and
nonpolar compounds. The ability to retain polar compounds has
been attributed to the strongly negative surface as determined by
zeta potential measurements. This negative charge is thought to
be similar to oil droplets in organo/aqueous solvents. In this case,
hydroxides from the auto dissociation of water are responsible for
the high negative zeta potential. The excellent analytical
reproducibility of the silica hydride materials is also another
distinguishing feature. The ANP capabilities of silica hydride phases
have produced the majority of applications to date since they
often have provided superior performance to HILIC. However,
rugged and reproducible reversed-phase and organic normal
phase analyses can also be done on silica hydride columns. It is
expected that use of these materials will continue to expand as
more scientists become aware of their unique properties.

References

1. MacNeill, R., and Taylor, P. (2015). The versatility in LC selectivity

attainable with the silica base and associated bonded phases,
Bioanalysis, 7, pp. 637–642.
2. Borges, E. M. (2015). Silica, hybrid silica, hydride silica and non-silica
stationary phases for liquid chromatography, J. Chromatogr. Sci.,
53, pp. 580–597.
3. Sandoval, J. E., and Pesek, J. J. (1989). Synthesis and characterization
of hydride modified porous silica material as an intermediate in
 References 203

the preparation of chemically bonded chromatographic stationary

phases, Anal. Chem., 61, pp. 2067–2975.
4. Sandoval, J. E., and Pesek, J. J. (1991). Hydrolytically stable bonded
chromatographic stationary phases prepared through hydrosilation
of olefins on a hydride-modified silica intermediate, Anal. Chem., 63,
pp. 2634–2641.
5. Montes, M. C., van Amen, C., Pesek, J. J., and Sandoval, J. E. (1994).
Chromatographic evaluation of alkyl-bonded phases through olefin
hydrosilation on a hydride-silica intermediate, J. Chromatogr., 688,
pp. 31–45.
6. Pesek, J. J., Matyska, M. T., Hemphala, H., Christensen, P. (1995).
High performance liquid chromatographic characterization of diol
bonded phases synthesized via a hydride intermediate, J. Liq.
Chromatogr., 18, pp. 2507–2525.
7. Pesek, J. J., Boysen, R. I., Hearn, M. T. W., Matyska, M. T. (2014).
Hydride-based stationary phases: A rapidly evolving technology
for the development of new bio-analytical methods, Anal. Methods,
6, pp. 4496–4503.
8. Brown, L., Ciccone, B., Pesek, J. J., Matyska, M. T. (2003). An evolution
in separation media for HPLC, Am. Lab., 35, pp. 23–29.
9. Pesek, J. J., and Matyska, M. T. (2006). How to retain polar and
nonpolar on the same column with an isocratic mobile phase, LCGC,
24, pp. 296–303.
10. Pesek, J. J., Matyska, M. T., and Sharma, A. (2008). Use of hydride-
based separation materials in organic normal phase chromatography,
J. Liq. Chromatogr. Rel. Technol., 31, pp. 134–147.
11. Soukup, J., and Jandera, P. (2014). Adsorption of water from aqueous
acetonitrile on silica-based stationary phases in aqueous normal
phase liquid chromatography, J. Chromatogr. A, 1374, pp. 102–111.
12. Kulsing, C., Yang, Y., Matyska, M. T., Pesek, J. J., Boysen, R. I., and Hearn,
M. T. (2015). Prediction of the zeta potentials and ionic descriptors
of a silica hydride stationary phase with mobile phases of different
pH and ionic strength, Anal. Chim. Acta, 859, pp. 79–86.
13. Beattie, J. K., and Djerdjev, A. M. (2004). The pristine oil/water interface:
Surfactant-free hydroxide-charged emulsions, Agnew. Chem., Int. Ed.,
43, pp. 3568–3571.
14. Zangi, R., and Engberts, J. B. (2005). Physisorption of hydroxide
ions from aqueous solution to a hydrophobic surface, J. Am. Chem.
Soc., 127, pp. 2272–2276.
204 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

15. Kudin, K. N., and Car, R. (2008). Why are water-hydrophobic

surfaces charged?, J. Am. Chem. Soc., 130, pp. 3915–3919.
16. Kulsing, C., Nolvachi, Y., Marriott, P., Boysen, R., Matyska, M., Pesek,
J. and Hearn, M. (2015). Insights into the origin of the separation
selectivity with silica hydride adsorbents, J. Phys. Chem., Part B, 119,
pp. 3063–3069.
17. McCalley, D. V. (2010). Study of the selectivity, retention mechanisms
and performance of alternative silica-based stationary phases
for separation of ionised solutes in hydrophilic interaction
chromatography, J. Chromatogr. A, 1217, pp. 3408–3417.
18. Pesek, J. J., Matyska, M. T., and Kim, A (2013). Evaluation of silica
hydride-based stationary phases for the analysis of drugs of abuse,
J. Sep. Sci., 36, pp. 2760–2766.
19. Nguyen, T., Felker, P., Hurst, J. W., Ly, C., Jarman, S., Diep, D., Pham,
C., Pesek, J. J., Matyska, M. T., Young, J, E., and Takeoka, G. (2016).
LC-MS characterization of phenolics in mesquite flour, LCGC North
Am., 34, pp. 28–31.
20. Pesek, J., Matyska, M., Hoffman, J. F., Madruga, N., Crizel, R. I., Elias,
M. C., Vanier, N. L., and Chaves, F. C. (2017). Analysis of mycotoxins
in grains using silica hydride-based stationary phases, J. Sep. Sci., 40,
pp. 1953–1959.
21. Young, J. E., Pan, Z., Teh, H. E., Menon, V., Modegerer, B., Pesek, J. J.,
Matyska, M. T., and Takeoka, G. (2017). Phenolic composition of
phenolic extracts using an LCMS approach with silica hydride
columns, J. Sep. Sci., 40, pp. 1449–1456.
22. http://kb.mtc-usa.com/article/AA-00789/0/.
23. Pesek, J. J., Matyska, M. T., Sieng, M., and Doan, L. (2016). Analysis of
capsaicinoids in hot sauces using a silica hydride-based stationary
phase, Curr. Chromatogr., 3, pp. 12–16.
24. Young, J. E., Lim, M. V., Topete, J., Hang, H., Gahol, M., Pesek, J. J.,
and Matyska, M. T. (2016). Improved sensitivity and specificity of
trans-resveratrol in red wine analysis with HPLC-UV and LC-MS,
LCGC North Am., 34, pp. 205–213.
25. Dang, A., Sieng, M., Pesek, J. J., and Matyska, M. T. (2015). Determination
of bisphenol A in receipts and carbon paper by HPLC-UV, J. Liq.
Chromatogr. Rel Technol., 38, pp. 438–442.
26. D’Orazio, G., Rocco, A., and Fanali, S. (2012). Fast liquid chromato-
graphy using columns of different internal diameters packed with
sub-2 µm silica particles, J. Chromatogr. A, 1228, pp. 213–220.
 References 205

27. Pesek, J. J., Matyska, M. T., and Lee, P. (2011). Synthesis of a preparative
C30 stationary phase on a silica hydride surface and its application
to carotenoid separation, J. Liq. Chromatogr. Rel. Technol., 34,
pp. 231–240.
28. Rocchi, S., Rocco, A., Pesek, J. J., Matyska, M. T., Capitani, M., and
Fanali, S. (2015). Use of a novel sub-2 µm vancomycin silica hydride
stationary phase, J. Chromatogr. A, 1381, pp. 149–159.
29. Pesek, J. J., Boysen, R. I., Hearn, M. T. W., and Matyksa, M. T. (2014).
Hydride- based stationary phases: A rapidly evolving technology
for the development of new bioanalytical methods, Anal. Methods, 6,
pp. 4496–4503.
30. Pesek, J. J., Matyska, M. T., Fischer, S. M., and Sana, T. R. (2008). Analysis of
hydrophilic metabolites by high-performance liquid chromatography-
mass spectrometry using a silica hydride-based stationary phase,
J. Chromatogr. A, 1204, pp. 48–55.
31. Chalkraft, K. R., and McCarry, B. E. (2013). Tandem LC columns for
the simultaneous retention of polar and nonpolar molecules in
comprehensive metabolomic analysis, J. Sep. Sci., 36, pp. 3478–3485.
32. Gouzy, A., and Larrouy-Maumus, G. (2014). Mycobacterium tuberculosis
exploits asparagine to assimilate nitrogen and resist acid stress
during infection, PLOS Pathogens, 10, e1003928, pp. 1–14.
33. Jenkins, S., Fischer, S. M., Chen, L., and Sana, T. R. (2013). Global LC/MS
metabolomics profiling of calcium stressed and immunosuppressant
treated Saccharomyces cerevisiae, Metabolites, 3, pp. 1102–1117.
34. Pesek, J. J., Matyska, M. T., Loo, J. A., Fischer, S. M., and Sana, T. R.
(2009). Analysis of hydrophilic metabolites in physiological fluids by
HPLC-MS using a silica hydride-based stationary phase, J. Sep. Sci., 32,
pp. 2200–2208.
35. Putluria, N., Shojaiegn, A., and Vasu, V. T. (2011). Metabolic profiling
reveals potential markers and bioprocesses altered in bladder
cancer progression, Cancer Res., 15, pp. 7376–7386.
36. Hellmuth, C., Koletzko, B., and Peissner, W. (2011). Aqueous
normal phase cysteine and methionine by liquid chromatography-
tandem mass spectrometry, J. Chromatogr. B, 879, pp. 83–89.
37. Gouzy, A., and Larrouy-Maumus, G. (2013). Mycobacterium tuberculosis
nitrogen assimilation and host colonization require aspartate,
Nat. Chem. Biol., 9, pp. 674–676.
38. Sana, T. R., Gordon, D. B., Fischer, S. M., Tichy, S. E., Kitagawa, N., Cai, C.,
Gosnell, W. L., and Chang, S. P. (2013). Global mass spectrometry based
206 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

metabolomics profiling of erythrocytes infected with Plasmodium

falciparum, PLUS ONE, 8, p. e60840.
39. Marcobal, A., Kashyap, P. C., Nelson, T. A., Aronov, P. A., Donia, M. S.,
Spormann, A., Fischbach, M. A., and Sonnenburg, J. L. (2013). A
metabolomic view of how the human gut microbiota impacts the
host metabolome using humanized and gnotobiotic mice, ISME J., 7,
pp. 1933–1943.
40. Eoh, H., and Rhee, K. Y. (2013). Multifunctional essentiality of
succinate metabolism in adaptation to hypoxia in mycobacterium
tuberculosis, Proc. Natl. Acad. Sci. U. S. A., 110, pp. 6454–6459.
41. Chakraborty, S., Gruber, T., Barry III, C. E., Boshoff, H. I., and Rhee, K. Y.
(2013). Para-aminosalicylic acid acts as an alternate substrate
of folate metabolism in mycobacterium tuberculosis, Science, 339,
pp. 88–91.
42. Pesek, J. J., Matyska, M. T., Duley, J., Zamzami, M., and Fischer, S. M.
(2010). Aqueous normal phase (ANP) retention of nucleotides on
silica hydride-based columns. Method development strategies for
analytes relevant to clinical analysis, J. Sep. Sci., 33, pp. 930–938.
43. Patti, G. (2011). Separation strategies for untargeted metabolomics,
J. Sep. Sci., 34, pp. 3460–3469.
44. Pesek, J. J., Matyska, M. T., and Young, J. E. (2014). Analysis of thiopurines
using aqueous normal phase chromatography, J. Pharm. Biomed.
Anal., 95, pp. 102–106.
45. Pesek, J. J., Matyska, M. T., Modereger, B., Hasbun, A., Phan, V. T.,
Mehr, Z., Guzman, M., and Watanable, S. (2016). The separation and
analysis of symmetric and asymmetric dimethylarginine and other
isobaric compounds using aqueous normal phase chromatography,
J. Chromatogr. A, 1441, pp. 52–59.
46. Pesek, J. J., Matyska, M. T., and Dang, A. (2015). Analysis of ethyl
glucuronide and ethyl sulfate using aqueous normal phase
chromatography/mass spectrometry, J. Sep. Sci., 38, pp. 1515–1520.
47. Cifkova, E., Hajek, R., Lisa, M., and Holcapek, M. (2016). Hydrophilic
liquid interaction chromatography-mass spectrometry of (lyso)-
phosphaditic acids, (lyso)phophaditylserines and other lipid classes,
J. Chromatogr. A, 1450, pp. 65–73.
48. Webster, G. K., Elliott, A., Dahan, A., and Miller, J. M. (2011).
Analysis of PEG 400 in perfusate samples aqueous normal phase
(ANP) chromatography with Evaporative light scattering detection,
Anal. Methods, 1218, pp. 742–744.
 References 207

49. Young, J. E., Matyska, M. T., and Pesek, J. J. (2011). Liquid

chromatography/ mass spectrometry compatible approaches for
the quantitation of folic acid in fortified juices and cereals using
aqueous normal phase conditions, J. Chromatogr. A, 1218, pp. 2121–
2126.
50. Dang, A., Matyska, M. T., and Pesek, J. J. (2013). The use of aqueous
normal phase chromatography as an analytical tool for food
analysis. Determination of histamine as a model system, Food Chem.,
141, pp. 4226–4230.
51. Young, J., Matyska, M. T., and Pesek, J. J. (2016). Robust HPLC-refractive
index analysis of simple sugars in beverages using silica hydride
columns, Curr. Nutr. Food Sci., 12, pp. 125–131.
52. Le, R., Young, J. E., Pesek, J. J., and Matyska, M. T. (2013). Separation of
1,3- Dimethylamine and other polar compounds in dietary supplement
formulations using aqueous normal phase chromatography with
mass spectrometry, J. Sep. Sci., 36, pp. 2578–2583.
53. Young, J. (2016). Advances in chromatographic analysis of foods and
beverages: Modern stationary phases for challenging compounds,
Agro FOOD Ind. Hi Tech., 27, pp. 14–17.
54. Nolvachai, Y., Kulsing, C., Boysen, R. I., Matyska, M. T., Pesek, J. J.,
Marriott, P. J., and Hearn, M. T. W. (2015). Comparison of the
performance of different silica hydride particles for the solid-
phase extraction of non-volatile analytes from dark chocolate with
analysis by gas chromatography-quadrupole mass spectrometry,
Food Chem., 174, pp. 434–439.
55. Pesek, J. J., Matyska, M. T., and Young, J. E. (2012). How silica hydride
HPLC columns can benefit the modern pharmaceutical laboratory,
Sep. Sci., 4, pp. 11–15.
56. Pesek, J. J., Matyska, M. T., and Dang, A. (2012). Analysis of
cycloserine and related compounds using aqueous normal phase
chromatography/mass spectrometry, J. Pharm. Biomed. Anal., 64–65,
pp. 72–76.
57. Gea, J., Liu, F., Holmes, E. H., Ostrander, G. K., and Li, Q. X. (2012).
Aqueous normal phase liquid chromatography coupled with tandem
time-of-flight quadrupole mass spectrometry for determination of
zanamivir in human serum, J. Chromatogr. B, 906, pp. 58–62.
58. http://kb.mtc-usa.com/article/AA-01460/90/.
59. Boysen, R. I., Yang, Y., Chowdhury, J., Matyska, M. T., Pesek, J. J., and
Hearn, M. T. W. (2011). Simultaneous separation of hydrophobic
208 Aqueous Normal-Phase Chromatography Using Type C Silica Columns

and hydrophilic peptides with a silica hydride stationary phase

using aqueous normal phase conditions, J. Chromatogr. A, 1218,
pp. 8021–8026.
60. http://kb.mtc-usa.com/article/AA-01232/0/.
61. Naffa, R., Holmes, G., Meekyung, A., Harding, D., and Norris, G. (2016).
Liquid chromatography electrospray ionization mass spectrometry
for the simultaneous quantitation of collagen and elastin crosslinks,
J. Chromatogr. A, 1478, pp. 60–67.
62. http://kb.mtc-usa.com/article/AA-00800/0/.
Chapter 9

Ion Chromatography Columns

Christopher Pohl
Chemistry R&D, Thermo Fisher Scientific,
1228 Titan Way, Sunnyvale, California 94085, USA
[email protected]

This chapter provides an overview of ion chromatography

columns including the early history of ion chromatography
columns and the state-of-the-art concerning the architecture of
ion chromatography columns.

9.1 Introduction
Ion chromatography is a highly specialized branch of liquid
chromatography, generally making use of columns specifically
designed for ion chromatography and intended to be used with ion
chromatography instrumentation. In general, ion chromatography
is performed with a specialized pumping system designed for the

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
210 Ion Chromatography Columns

delivery of highly acidic and highly basic mobile phases without risk
of corrosion as a consequence of the extreme pH associated with
such mobile phases. Also, although there are notable exceptions,
ion chromatography is commonly practiced while making use
of a conductivity detector. Because a conductivity detector is a
bulk property detector [1], the conductivity signal arising from
conductive mobile phases is highly susceptible to environmental
effects. For that reason, ion chromatography is usually practiced
in a temperature-controlled environment to minimize thermal
effects on the conductivity signal. In addition, the signal-to-noise
ratio associated with detection of ionic species against a background
conductivity of the mobile phase is relatively poor given the fact
that the concentration of the mobile phase is typically much
greater than the concentration of analytes being detected in the
presence of the mobile phase conductivity. For that reason, most
ion chromatography is practiced making use of a suppressor
device [2, 3]. The suppressor device is unique among analytical
techniques in that it reduces or eliminates the mobile phase
conductivity while simultaneously enhancing the conductivity of
the analyte ion. While the operation of the suppressor device is
beyond the scope of this chapter, it is relevant to the design and
architecture of ion chromatography columns. Because the best
conductivity detector performance is achieved when a strong
acid mobile phase is used for the analysis of cations and an alkali
metal hydroxide mobile phase is used for the analysis of anions,
this necessitates the use of stationary phase materials compatible
with long-term exposure to mobile phases at extreme pH. Thus,
while silica-based materials are the mainstay of HPLC, the poor
hydrolytic stability of silica at high pH and the poor ligand stability
at low pH render silica-based stationary phases a poor choice
for ion chromatography. In contrast, polymeric stationary phases,
especially polymers based on styrenic monomers are exceptionally
stable at high pH and low pH, making columns derived from such
monomers an excellent choice for use in ion chromatography.
For that reason, the bulk of the discussion in this chapter will
focus on polymeric stationary phases suitable for use in ion
chromatography.
 Cation-Exchange Phases 211

9.2 Cation-Exchange Phases

As previously mentioned, the earliest phases in ion chromato-
graphy were based on cross-linked polystyrene since they
provided the necessary chemical stability to withstand long-term
exposure to mineral acids. While there were examples of early
silica-based cation-exchange materials, these tended to be used
with non-suppressed ion chromatography given the relatively
poor chemical stability of silica-based materials when exposed to
extremes of pH. Although detection limits are 10–50 times worse
when operating in non-suppressed mode [3] than what can
be achieved using a suppressor, operation in non-suppressed
mode enables the use of mobile phases with more moderate pH
conditions that are not as corrosive to silica. Even so, silica-
based phases tend to be so unstable that even acidic samples can
damage the stationary phase. As a consequence, use of silica-
based stationary phases became increasingly uncommon as
newer high-performance polymeric phases came on the scene.

9.2.1 Surface-Sulfonated Cation-Exchange Phases

The earliest cation-exchange phases in ion chromatography [2]
were composed of 2% cross-link polystyrene particles that were
lightly surface sulfonated to introduce an extremely thin layer
of sulfonated sites on the exterior surface. Use of sulfuric acid as
the sulfonation reagent allows for a shell progressive reaction
as the reagent proceeds to sulfonate the particle in a highly
controllable manner. By adjusting the reaction time and reaction
temperature, a wide variety of strongly acidic cation-exchange
materials with a wide range of cation-exchange capacities suitable
for use in ion chromatography can be prepared.
The 2% cross-link used in early materials was optimized for
monovalent cation selectivity. Higher cross-link levels reduced
the resolution of ammonia from potassium. However, such low
cross-link levels rendered the substrate particle highly susceptible
to excessive swelling in the presence of organic solvent. Because
excessive swelling can cause column bed disturbances during
swelling and subsequent shrinking back to the original dimensions
212 Ion Chromatography Columns

when the solvent is removed from the mobile phase, such

materials were restricted to use with largely aqueous mobile
phases. At first glance, this might not seem like much of a problem,
but the addition of solvent is often required to remove highly
retained analytes from the stationary phase. Restrictions on the
amount of solvent made it difficult to remove such contaminants
from the stationary phase and reduced the long-term lifetime of
such columns.
One disadvantage of using sulfonated cation-exchange
materials for ion chromatography arises from the high-affinity
such materials have for divalent cations. The concentration of
hydronium necessary for elution of calcium or magnesium exceeded
the practical mobile phase concentration range suitable for use
with suppressor devices. As a result, the earliest suppressed ion
chromatography methods utilized different eluent systems for
monovalent cations than for divalent cations. Hydronium-based
eluent systems were used for the elution of monovalent cations
while acidic solutions of meta-phenylenediamine were used for
the elution of divalent cations.
While hydronium-based mobile phases were not problematic
for long-term use, meta-phenylenediamine-based eluent systems
were problematic both because they tended to degrade as used in
most common applications and because this eluent species was
difficult to remove from the cation-exchange phase after initial
use. Meta-phenylenediamine tends to undergo oxidation reactions,
producing polymeric byproducts difficult to remove from the
stationary phase.

9.2.2 Sulfonated Latex-Based Cation-Exchange Phases

The first commercial introduction of latex-based cation-exchange
phases took place in 1986 more than 10 years after the first
commercial introduction of ion chromatography. While anion-
exchange phases making use of an anion-exchange latex coating
on nonporous surface-sulfonated substrate particles had been
in use for more than a decade (see Section 9.3.1), preparation of
the cation-exchange variant arrived much later due to the
difficulty of synthesizing colloidally stable, fully sulfonated
cation-exchange latex particles. The first column of this type, the
 Cation-Exchange Phases 213

IonPac CS3 column was based on a three-layer electrostatically

bonded pellicular architecture (see Fig. 9.1 below).

Micro Anion Exchange Latex

Surface Sulfonated Core

2 XL Sulfonated Cation
Exchange
g Latex

Figure 9.1 Schematic representation of the CS3 stationary phase.

The construction of CS3 chromatographic material starts

with 2% cross-link polystyrene particles; the substrate is surface
sulfonated as described in Section 9.2.1. Sulfonation conditions
are adjusted to produce a very thin sulfonation layer to minimize
any potential cation-exchange retention after assembly of the
pellicular structure. Then, a layer of colloidal anion-exchange
particles was deposited on the surface-sulfonated substrate.
As a consequence of the minimal sulfonation of the substrate,
application of the anion-exchange particles eliminated all traces
of cation-exchange properties from the surface-sulfonated layer as
a consequence of the intercalation of the anion-exchange particles
into the surface-sulfonated cation-exchange layer at the particle
surface. Finally, a layer of fully sulfonated cation-exchange latex
was attached to the anion-exchange latex coating to produce the
final product. As with the previous coating step, only traces of
anion-exchange retention were detectable after application of the
cation-exchange latex coating.
Cation-exchange stationary phases based on sulfonated
latex provided significant improvements in chromatographic
performance compared to the surface-sulfonated materials.
Surface-sulfonated materials result in a cross-link gradient from
214 Ion Chromatography Columns

the outermost surface of the particle to the sulfonation boundary

adjacent to the edge of the sulfonated substrate particle which
comprises the majority the particle volume. Because swelling of
the sulfonated particle is constrained by the rigid core at the
boundary between the sulfonated zone in the un-sulfonated zone
of the particle, this effectively increases the cross-link at the
boundary between the two phases. In contrast, at the exterior
surface where swelling is relatively unconstrained by the influence
of the un-sulfonated substrate, the effective cross-link is close to
what one would expect for the homogeneously swollen sulfonated
polymer. Evidence of this cross-link gradient can be seen in
two different ways. First, when using latex particles which feature
cross-link homogeneously distributed throughout the particle,
5% cross-link is necessary to center ammonia between sodium
and potassium while in the case of surface-sulfonated substrate,
2% cross-link is optimal for providing the same selectivity. Second,
chromatographic performance for relatively poorly hydrated
cations is considerably better with the latex-based stationary
phases. Because the boundary between the sulfonated zone on
the un-sulfonated zone is relatively poorly hydrated, this tends
to be an area where weakly hydrated cations exhibit the highest
affinity. However, because this poorly hydrated boundary zone
is the farthest away from the particle surface, mass transport to
and from this region is relatively slow, resulting in relatively
poor peak shape for poorly hydrated cations.
A second innovation coincided with the commercial
introduction of cation-exchange materials utilizing cation-exchange
latex coatings: an improved eluent system, 2,3-diaminopropionic
acid. In contrast to the earlier HCl/HCl-meta-phenylenediamine
eluent systems which required separate dedicated columns, one
column for monovalent cations and a second column for divalent
cations, 2,3-diaminopropionic acid is relatively easy to elute
from a cation-exchange latex making it straightforward to switch
from the analysis of monovalent cations to the analysis of divalent
cations. The selectivity of this new eluent system was not optimal
for separation of common monovalent cations (lithium, sodium,
ammonium, and potassium) under conditions suitable for elution
of common divalent cations (magnesium and calcium), 2,3-
diaminopropionic acid. The best chromatography for monovalent
cations was still seen with simple mineral acid eluent system.
 Cation-Exchange Phases 215

The chief advantage of this new eluent system is that it made it

easy to switch back and forth from a mineral acid eluent optimal
for monovalent cations to the HCl and 2,3-diaminopropionic acid
mixture suitable for elution of divalent cations.

Figure 9.2 Separation of cations on an IonPac CS11 column. Flow rate

0.25 mL per min, 40 mM hydrochloric acid with 4 mM 2,3-diaminopropionic
acid, ambient temperature, suppressed conductivity detection, 2.5 µL
injection volume. Peaks: (1) 0.5 ppm lithium; (2) 2 ppm sodium; (3) 2.5 ppm
ammonium; (4) 5 ppm potassium; (5) 15 ppm magnesium; (6) 25 ppm
calcium.

The third innovation related to this class of stationary phases

was the introduction of latex particles composed the styrenic
monomers with dramatically different levels of reactivity
activation concerning sulfonation. By combining divinylbenzene
with a styrenic monomer with reactivity activated to sulfonation
216 Ion Chromatography Columns

and a second styrenic monomer with reactivity deactivated to

sulfonation, it was possible to achieve greater spacing between
the sulfonated groups within the sulfonated latex particle.
Columns based on this chemistry include the IonPac CS10 and
the IonPac CS11. Because retention of divalent ions is much
more sensitive to ion-exchange site concentration than the
retention of monovalent ions, this allows the manipulation of the
selectivity of divalent species relative to monovalent species. As
a consequence, for the first time it was possible to separate
the common cations (lithium, sodium, ammonium, potassium,
magnesium, and calcium) under isocratic conditions using an
HCl–2,3-diaminopropionic acid eluent system (see Fig. 9.2 above).
In addition to these phases, there are also several companies
(Wescan, Alltech, and Hamilton) making surface-sulfonated
macroporous cation-exchange resins. For the most part, these
columns were operated in nonsuppressed mode utilizing eluent
systems such as HCl-ethylenediamine. Such materials provided
relatively poor chromatographic performance due to the relatively
high cross-link of macroporous substrate particles.

9.2.3 Polymer-Encapsulated Weak Acid Cation-Exchange

Phases
In 1987, Prof. Schomburg and coworkers [4] developed an
entirely new method of stationary phase attachment. Their
work originally focused on developing new methods of attaching
stationary phases to alumina. Given the fact that alumina is
much more pH stable than silica, Prof. Schomburg and the team
was working on the development of methods for attaching a
stationary phase to alumina. Because alumina is not amenable
to direct silane bonding, which enables facile modification of the
surface chemistry of silica substrate, they devised a novel
encapsulation methodology involving deposition of a thin polymer
film which was subsequently cross-linked in situ. They chose
copolymers based on butadiene since such polymers contain
residual double bonds along the polymer backbone. Among the
polymers they investigated was butadiene-maleic anhydride
copolymer. They found that this polymer when fully hydrolyzed
and subsequently cross-linked in situ was useful for the separation
of both monovalent and divalent cations under isocratic
 Cation-Exchange Phases 217

conditions using a relatively simple mobile phase. By using a

weak acid cation-exchange site rather than a strong acid cation-
exchange site, the concentration of cation-exchange sites and the
corresponding monovalent-divalent selectivity can be manipulated
by adjusting the mobile phase pH. As they discovered, alumina
was not a very good substrate for the separation of inorganic cations
due to the underlying selectivity of alumina for inorganic cations.
For that reason, they ultimately ended up using silica as the
substrate for the butadiene-maleic anhydride copolymer. By
combining a silica substrate with suitable mobile phase pH, they
were able to achieve the best separations ever reported for the
common inorganic cations under isocratic conditions.
This development led to the wholesale shift of stationary
phase design to the use of weak acid cation-exchange phases for
ion chromatography. Several manufacturers, including Metrohm,
Shodex, Tosoh, and Dionex, developed silica-based analogs of the
phase described by Prof. Schomburg. However, there are significant
disadvantages of this synthetic approach when combined with a
suppressor. First, the underlying silica has a limited pH stability
which constrained the pH of samples that can be utilized with such
materials. Second, the in situ surface cross-linking methodology
developed by Prof. Schomburg exhibits a significant amount of
“bleed” arising from non-covalently attached polymer fragments
trapped within the cross-linked matrix. These fragments are
captured by the suppressor, slowly converting the suppressor to
a cation-exchange column, causing poor recovery and bad peak
shapes as polymer fragments accumulate on the suppressor. For
this reason, such materials are only recommended for use in
non-suppressed mode.

9.2.4 Polymer-Grafted Weak Acid Cation-Exchange

Phases
As a consequence of the polymer bleed problem associated
with encapsulated weak acid cation-exchange phases, Dionex
embarked upon a development process with the aim of replicating
the selectivity of Prof. Schomburg’s column while circumventing
the problems associated with the substrate and the attachment
method. Ultimately, they developed the first polymer-based weak
acid cation-exchange resin (the IonPac CS12 column) [5]. The CS12
218 Ion Chromatography Columns

column utilizes radical polymerization to covalently graft weak

acid cation-exchange polymers to the surface of a macroporous
substrate particle. Only a small fraction of the cation-exchange
sites in such a material is ionized at any one time, necessitating
a relatively high capacity to achieve sufficient retention under
suitable eluent conditions. A high-surface-area macroporous
polymer allows the preparation of material of high capacity
while still providing relatively good mass transport kinetics.
Later, improvements in the grafting chemistry and inclusion of a
vinyl phosphonate comonomer were used to prepare the IonPac
CS12A column exhibiting both improved peak shape for common
inorganic cations as well as better selectivity for magnesium,
manganese, and calcium (Fig. 9.3). The IonPac CS12A column is
still the world’s best selling cation-exchange ion chromatography
column, more than 20 years since its initial introduction in 1995.
It is cited in a wide variety of publications [6–8] and used in
many proprietary analytical methods throughout the world.

Figure 9.3 Separation of cations on a 4 mm ID IonPac CS12A column. Flow

rate 1 mL per min, 20 mM methanesulfonic acid, ambient temperature,
suppressed conductivity detection, 25 µL injection volume. Peaks:
(1) 0.5 ppm lithium; (2) 2 ppm sodium; (3) 2.5 ppm ammonium; (4)
5 ppm potassium; (5) 10 ppm diethylamine; (6) 2.5 ppm magnesium;
(7) 2.5 ppm manganese, (8) 10 ppm calcium.

With the initial success of such polymer-based weak acid

cation-exchange phases, leading manufacturers throughout the
world have introduced a variety analogous weak acid cation-
 Cation-Exchange Phases 219

exchange materials. A wide variety of such materials are

commercially available today [9]. Many of these materials were
specifically designed for use in nonsuppressed ion chromatography,
but most of these materials can be used in suppressed ion
chromatography with the exception those using the polymer
encapsulation methodology of Prof. Schomburg.

9.2.5 Polymer-Grafted Strong and Weak Acid

Cation-Exchange Phases
One disadvantage of using a weak acid cation-exchange phase
under acidic conditions is that mass transport is adversely affected
by the relatively poor hydration of carboxylic acids when largely
in a protonated state. Poor hydration is due in part to the fact that
the protonated carboxylic acid is significantly less hydrophilic
than the salt form of the corresponding acid. The second factor
that adversely affects the chromatographic performance of
carboxylic acid phases under conditions where the phase is largely
in a protonated state is the tendency of protonated carboxylic
acid groups to form stable hydrogen bonded dimers with adjacent
carboxylic acid groups. These hydrogen bonded dimers form
in situ cross-links which further dehydrate the polymer network.
The combination of these two factors results in relatively slow
mass transport within the stationary phase under acidic conditions.
One way to mitigate this problem is to copolymerize a carboxylic
acid monomer with more acidic monomers under acidic conditions.
An example of such a column is the IonPac CS20 column, shown in
Fig. 9.4. It uses a terpolymer composition consisting of styrene
sulfonate, vinyl phosphonate, and a carboxylic acid monomer. By
polymerizing the three monomers under acidic conditions such
that the carboxylic acid monomer is fully protonated, electrostatic
forces substantially reduce the probability of two of the more
acidic monomers ending up as adjacent monomers in the polymer
network. Such an approach greatly reduces the affinity of such
polymers for divalent species, eliminating the need for a divalent
eluent cation to elute alkaline earth cations. Such a stationary
phase essentially bridges the gap between older style sulfonated
cation-exchange materials which have excellent monovalent cation
selectivity but exhibit excessive retention for divalent cations and
220 Ion Chromatography Columns

weak acid cation-exchange phases which have relatively poor

monovalent cation selectivity but are easily eluted with hydronium.

Figure 9.4 Separation of cations on 4 mm ID IonPac CS20 and CG20

columns. Flow rate 1.2 mL per min, 2–4 mM methanesulfonic acid (MSA)
from 0 to 15 min, 4–15 mM methanesulfonic acid (MSA) from 15 to
20 min, 15–50 mM methanesulfonic acid (MSA) from 20 to 25 min,
50 mM methanesulfonic acid (MSA) from 25 to 37 min, 50°C, suppressed
conductivity detection, 10 µL injection volume. Peaks: (1) 0.25 ppm
lithium; (2) 1 ppm sodium; (3) 1.25 ppm ammonium; (4) 2.5 ppm
potassium; (5) 5 ppm dimethylamine; (6) 5 ppm trimethylamine;
(7) ethylmethylamine; (8) 5 ppm morpholine; (9) 5 ppm diethylamine;
(10) 5 ppm piperidine; (11) 1.25 ppm magnesium; (12) 2.5 ppm calcium.

9.2.6 Cation-Exchange Phases for Transition Metal

Separations
Although suppressed ion chromatography is commonly applied to
the analysis of alkali metals, alkaline earth metals, ammonia
and amines, suppressed ion chromatography is not compatible
with transition metal detection. Application of suppressed ion
chromatography to cations involves the use of a suppressor in
the hydroxide form. Hydroxide from the suppressor replaces the
strongly acidic anionic component of the mobile phase and
subsequently reacts with the hydronium counterion to produce
water. Also, the suppression reaction increases the conductivity
of the cation by converting it to its highly conductive hydroxide
form. Although this approach works for alkali metals and alkaline
earth metals, the same approach does not work well for transition
 Cation-Exchange Phases 221

metals because the hydroxides of transition metals are only sparingly

soluble in water. As a consequence, nearly all transition metal
detection is accomplished either by post-column reaction with
a chromophoric chelating agent or by use of a conventional
spectroscopic detector such as ICP or ICP-MS. Given the fact that
transition metal detection systems do not usually make use of a
suppressor, stationary phase design doesn’t focus on simple cation-
exchange retention mechanisms. Selectivity of both weak acid
cation-exchange materials and strong acid cation-exchange materials
is poor for transition metals due to the relatively similar hydration
properties of this class of metals.
Several approaches have been utilized to improve selectivity
transition metals. One approach is to add a chelating species to the
mobile phase that converts the metal to an anionic species. Cation-
exchange retention will be reduced in proportion to the extent
that the transition metal is in equilibrium with the anionic
form of the metal complex. By choosing an appropriate anionic
complexing agent, it is possible to manipulate the selectivity of
the separation if the retention time for a specific pair of transition
metals is identical when using a non-complexing mobile phase.
The best results are achieved when using a complexing agent with
widely differing formation constants with the transition metals of
interest as this will produce the highest selectivity. If conditions
are chosen such that all transition metals are 100% in the form
of the anionic complex, there will be no retention via cation-
exchange. Thus, such conditions should be avoided when operating
in cation-exchange mode. An alternative is to use anion-exchange
to separate the anionic complexes of transition metals. In this case,
the extent to which the transition metal forms an anionic complex
will enhance retention with maximum retention being achieved
under conditions where the complex is essentially 100% complex
with the anionic chelating agent. If multiple transition metals
have similar formation constants with the anionic complexing
agent, it is likely that the complexes will co-elute via anion-
exchange although some transition metals will be separated under
these conditions because different transition metals differ in
their coordination numbers. However, the best option is to avoid
conditions where the transition metals are 100% in the complexed
form when using anion-exchange.
222 Ion Chromatography Columns

Figure 9.5 Schematic of CS5A resin synthesis and structure.

A more powerful option is to use a material exhibiting both

anion-exchange and cation-exchange properties. Such a stationary
phase allows one to use either cation-exchange mode, anion-
exchange mode or a combination of the two modes. Operating with
a dual-mode stationary phase allows one to avoid poor retention
and poor selectivity associated with using anion-exchange or
cation-exchange alone. The Dionex IonPac CS5A column was
designed specifically to allow operation in dual ion-exchange mode.
The phase architecture is illustrated in Fig. 9.5. The substrate
particle is composed of 55% cross-link DVB-ethylvinylbenzene
nonporous particles. An initial, extremely thin anion-exchange
layer is produced by grafting the surface with a water-soluble
cationic monomer. A 140 nm 10% cross-link strong acid cation-
exchange latex coating is applied directly to the surface to
introduce cation-exchange sites. Due to the minimal nature of
the underlying anion-exchange layer, this retention layer is
 Cation-Exchange Phases 223

quantitatively negated by the cation-exchange nanoparticle

coating. A 76 nm 2% cross-link strong base anion-exchange latex
coating is applied to the previous layer to make a cation-exchange/
anion-exchange composite. The combination of the two outermost
ion-exchange latex layers results in useful retention zones for
transition metals in either cation-exchange or anion-exchange
mode or a combination of the two modes. An example separation
of the CS5A column operating in the anion-exchange mode is
shown in Fig. 9.6. In this case, the ligand pyridine-2,6-dicarboxylate
forms a relatively stable complex with transition metals and
the transition metals are separated as the anionic complex.

Figure 9.6 Separation of transition metal cations on 4 mm ID IonPac

CS5A and CG5A columns. Flow rate 1.2 mL per minute, MetPac PDCA
eluent, ambient temperature, absorbance detection at 530 nm with PAR
in MetPac post-column reagent diluent, 50 µL injection volume. Peaks:
(1) 1.3 ppm iron (III); (2) 1.3 ppm copper; (3) 2.6 ppm nickel; (4) 1.3 ppm
zinc; (5) 1.3 ppm cobalt; (6) 6 ppm cadmium; (7) 2.6 ppm manganese;
(8) 1.3 ppm iron (II).
224 Ion Chromatography Columns

9.3 Anion-Exchange Phases

Anion-exchange phases in ion chromatography are by far the
most diverse category of stationary phases. In the case of cations,
elemental spectroscopy instrumentation provides an alternative
to ion chromatography, diminishing the popularity of this mode
of ion chromatography. However, in the case of anions, elemental
spectroscopy is not capable of distinguishing between the
various anions commonly found in samples of analytical interest.
Considering the wide variety of anions found in nature and
industrial and pharmaceutical samples, a wide variety of different
phases have been developed to provide optimal selectivity for
critical separation needs. As in the case of cation-exchange phases,
the vast majority of all anion-exchange phases are based on
polymeric substrates. The extreme pH is associated with anion-
exchange separations when operated in conjunction with the
suppressor are incompatible with inorganic substrates. Although
there were a few silica-based phases developed in the early days
of ion chromatography, such phases have largely disappeared due
to the poor stability of such materials under typical high pH
operating conditions.

9.3.1 Quaternary Latex-Based Anion-Exchange Phases

on Nonporous Substrates
The earliest anion-exchange phases used in ion chromatography
[2] made use of surface-sulfonated 2% cross-link polystyrene
beads as a substrate. This was a natural starting point given the
fact that such materials had already been developed for separation
of cations (see Section 9.2.1). Preparation of the anion-exchange
phase was achieved by coating the surface-sulfonated substrate
with a layer of anion-exchange nanoparticles. Although Hamish
Small initially prepared the particle coating materials by grinding
macroporous anion-exchange materials, the chemists at Dow
later applied their considerable expertise in making emulsion
polymers to prepare anion-exchange materials via emulsion
polymerization. Preparation of the anion-exchange nanoparticles
involved the co-polymerization of vinylbenzyl chloride (VBC)
 Anion-Exchange Phases 225

with divinylbenzene. Subsequently, the active anion-exchange

nanoparticles were produced by reacting the cross-linked
vinylbenzyl chloride nanoparticles with a suitable tertiary amine,
forming particles with quaternary anion-exchange sites distributed
homogeneously throughout the nanoparticles. Finally, the anion-
exchange composite was prepared by mixing a suitable quantity of
anion-exchange nanoparticles with separately prepared surface-
sulfonated cross-linked polystyrene beads. Such composites form
essentially instantaneously upon mixing the oppositely charged
materials in an aqueous slurry medium. Although the attachment
mechanism is electrostatic, the large number of simultaneous
electrostatic interactions associated with the attachment of
the nanoparticles to the substrate surface preclude elution of
the electrostatically attached nanoparticles under any known
conditions producing an exceptionally stable composite.
The earliest phases using this synthesis approach used
relatively large particle size substrate particles (20–30 µm) and
relatively large diameter nanoparticles (200–300 nm) to prepare
the stationary phase coating. Such materials were of relatively low
capacity and low efficiency. The low capacity of these materials
was dictated by the relatively limited suppression capacity of packed
bed suppressors in use at the time. These low capacity stationary
phases allowed the use of relatively dilute mobile phases, thus
prolonging the life of the suppressor. However, the relatively low
efficiency of such materials necessitated relatively long analysis
times, generally greater than 30 min for a typical analysis. Later
developments in anion-exchange phases made use of much smaller
substrate particles and smaller anion-exchange nanoparticles.
By reducing the diameter of substrate particles and anion-
exchange nanoparticles while maintaining the same diameter
ratio for the particles, column capacities are held constant while
chromatographic efficiency is improved. Progressive reduction
in both particle diameters resulted in significant improvements
in chromatographic performance and speed of analysis.
Figure 9.7 shows an example chromatogram of one of the more
modern variants of this stationary phase architecture.
The final development for this class of ion-exchange materials
was the shift to the use of 55% divinylbenzene for the substrate
226 Ion Chromatography Columns

particles. As in the case of 2% cross-link polystyrene particles,

these particles are surface sulfonated before being coated with
anion-exchange nanoparticles. The chief advantage of such high
cross-link materials is that these materials are immune to swelling
in organic solvents due to the high cross-link. Use of 2% cross-
link polystyrene particles limited the use of organic solvent to no
more than 5% to avoid excessive swelling and column damage.
In contrast, these new highly cross-linked substrate particles
allowed the use of high concentrations of organic solvents to
clean hydrophobic contaminants from the active stationary phase
and the use of organic solvents to manipulate ion-exchange
selectivity. Figure 9.8 shows the influence of organic solvent on
selectivity with the AS4A-SC column which utilizes 55% cross-link
substrate particles.

Figure 9.7 Separation of anions on a 4 mm ID IonPac AS4A column.

Flow rate 2 mL per minute, 1.7 mM sodium bicarbonate and 1.8 mM
sodium carbonate, ambient temperature, suppressed conductivity
detection, 25 µL injection volume. Peaks: (1) 1 ppm fluoride; (2) 1.5 ppm
chloride; (3) 2.5 ppm nitrite; (4) 4 ppm bromide; (5) 4 ppm nitrate;
(6) 10 ppm phosphate; (7) 5 ppm sulfate.
 Anion-Exchange Phases 227

Figure 9.8 Separation of anions on a 4 mm ID IonPac AS4A-SC column.

Flow rate 2 mL per minute, 1.7 mM sodium bicarbonate and 1.8 mM
sodium carbonate, ambient temperature, suppressed conductivity
detection, 25 µL injection volume. Peaks: (1) 2 ppm fluoride; (2) 3 ppm
chloride; (3) 5 ppm nitrite; (4) 10 ppm bromide; (5) 10 ppm nitrate;
(6) 15 ppm phosphate; (7) 15 ppm sulfate.

9.3.2 Quaternary Latex-Based Anion-Exchange Phases

on Wide-Pore Substrates
From the time of initial commercialization of ion chromatography,
suppressor capacity had been somewhat limited. Initially, packed
bed suppressors needed to be periodically regenerated, so it was
desirable to reduce the frequency of regeneration by using relatively
low capacity columns. Subsequently, packed bed suppressors
were supplanted by fiber suppressors which allowed continuous
regeneration, but these too had limited dynamic capacity.
As a result, the maximum eluent concentration was constrained
to low millimolar concentrations. However, by the mid-1980s, the
first high-capacity suppressors were introduced. These suppressors
were capable of suppressing up to 100 mM eluents, greatly
expanding the range of eluent concentrations that could be utilized
228 Ion Chromatography Columns

Figure 9.9 Separation of anions on a 4 mm ID IonPac AS10 column.

Flow rate 1 mL per minute, 50–124 mM sodium hydroxide in 30 min,
ambient temperature, suppressed conductivity detection, 20 µL injection
volume. Peaks: (1) 1.2 ppm fluoride; (2) 30 ppm acetate; (3) 5 ppm
formate; (4) 10 ppm selenite; (5) 1.8 ppm chloride; (6) 9 ppm nitrite;
(7) 10 ppm sulfate; (8) 10 ppm oxalate; (9) 10 ppm selenate; (10) 9 ppm
phosphate; (11) 10 ppm bromide; (12) 6 ppm nitrate.

in ion chromatography. The availability of higher dynamic capacity

suppressors led to an interest in the development of new anion-
exchange phases with higher capacity. By using higher capacity
columns, a wider range of sample types could be directly injected
into the column without peak shape distortion associated with
high ionic strength samples. One of the earliest materials with high
capacity made use of quaternary anion-exchange latex particles
coating porous substrates rather than nonporous substrates.
Such an architecture enables the production of materials with
significantly higher capacities while using the same stationary
phase synthesis methodologies developed for use with nonporous
substrates. One challenge with such materials is that the pore
 Anion-Exchange Phases 229

size needs to be large enough to allow for anion-exchange

nanoparticles to coat both the interior and the exterior surface
of the particles. Preparation of a suitable material requires very
large pore size substrate particles with nanoparticles significantly
smaller than the nanoparticles used with nonporous substrates.
The first column made using this approach was the IonPac AS10
column (see Fig. 9.9). It utilized particles with an average pore
size of around 200 nm with 65 nm anion-exchange nanoparticles.
The wide-pore substrate particles used in the AS10 column were
prepared from 55% divinylbenzene which imparted resistance
to swelling under the influence of organic solvents. As mentioned
previously, using 2% cross-link polystyrene limited the use of
an organic solvent to less than 5%. Subsequently, this stationary
phase morphology has been used in a wide variety of different
columns including the IonPac AS9-HC, the IonPac AS11-HC,
the IonPac AS12A, and the IonPac AS16 targeting oxyhalides
in drinking water, organic acids in foods, analysis of common
anions in drinking water and analysis of perchlorate in drinking
water, respectively.

9.3.3 Adsorbed Coating Anion-Exchange Phases

The late 1980s saw the introduction of a new class of anion-exchange
phases, anion-exchange phases based on an adsorbed coating.
Examples of such materials include the Sarasep AN1 and AN300.
The main advantage of these materials is their relatively high
retention for fluoride under conditions that allow elution of more
strongly retained anions in under 12 min. Fluoride tends to be very
weakly retained by anion-exchange materials, eluting near the
void volume on most columns. Because of the proximity of
fluoride to the void volume, trace analysis of fluoride in real
samples is often difficult. The new columns from Sarasep somehow
circumvented this problem, greatly facilitating the quantitation
of fluoride in real drinking water samples. Although never
presented to the market as an adsorbed coating, treatment of these
columns with organic solvent removes the vast majority of the
stationary phase, demonstrating that the stationary phase is
not permanently attached to the substrate surface. Although the
exact nature of the stationary phase has not been revealed, it is
230 Ion Chromatography Columns

speculated that the stationary phase is prepared via deposition

of a low-molecular-weight cationic polymer onto the surface of a
high-surface-area hydrophobic substrate. Presumably, sufficient
retention in a purely aqueous environment provides stable
retention via a hydrophobic interaction with the substrate. While
such materials represented a major step forward anion-exchange
selectivity of columns designed for use with carbonate eluents,
the poor compatibility of such phases with organic solvent
containing mobile phases limited the utility of such materials.
As such, although these columns are still commercially available,
they have been for the most part supplanted by later materials
with similar fluoride selectivity by providing much better solvent
compatibility.

9.3.4 Polymer-Grafted Strong Base Anion-Exchange

Phases
Development of the Sarasep AN1 stationary phase demonstrated
the feasibility of providing improved resolution of fluoride from
the system void without sacrificing analysis time. However,
given the fact that solvent compatibility of the Sarasep AN1 was
poor, this drove research into the development of new classes
of anion-exchange materials with improved fluoride resolution
from the void while still maintaining solvent compatibility
common to other modern ion chromatography stationary phases.
Subsequent research showed that achieving this selectivity
required a stationary phase architecture enabling high levels
of hydration of the stationary phase. All of the previous anion-
exchange phases had been based on moderate levels of cross-
link which restricts hydration of the stationary phase. One way
around this problem is to graft enormous numbers of linear
polymer strands to the surface of a high-surface-area polymeric
substrate. Such a material provided the necessary retention even
in the absence of any cross-linking. The IonPac AS14 exemplifies
grafted materials of this architecture based on a pH stable
methacrylate polymer, while the IonPac AS14A exemplifies grafted
materials of this architecture based on an, even more, pH stable
styrenic polymer. This architecture provides greatly improved
fluoride resolution without sacrificing solvent compatibility
(see Fig. 9.10).
 Anion-Exchange Phases 231

Figure 9.10 Separation of anions on a 3 mm ID by 150 mm IonPac AS14A

column with 3 mm ID AG14A column. Flow rate 0.8 mL per min, 1 mM
sodium bicarbonate and 8 mM sodium carbonate, 30°C, suppressed
conductivity detection, 25 µL injection volume. Peaks: (1) 5 ppm fluoride;
(2) 20 ppm acetate; (3) 10 ppm chloride; (4) 15 ppm nitrite; (5) 25 ppm
bromide; (6) 25 ppm nitrate; (7) 40 ppm phosphate; (8) 30 ppm sulfate.

9.3.5 Surface-Chloromethylated High Cross-Link

Macroporous Anion-Exchange Phases
Early in the development of ion chromatography phases, Barron
and Fritz [10] developed an aqueous-based chloromethylation
methodology using paraformaldehyde in HCl. This approach offered
two potential benefits over conventional chloromethylation using
chloromethyl ether. First, it avoided the use of chloromethyl ether
which invariably contains low levels of bis-chloromethyl ether,
a potent carcinogen. Second, because the reaction takes place
in an aqueous environment, the reaction is largely restricted to
the surface of the substrate particle, improving chromatographic
232 Ion Chromatography Columns

performance. The first benefit (avoidance of bis-chloromethyl

ether) is highly suspect as it is likely that the paraformaldehyde-
HCl synthesis methodology does create bis-chloromethyl ether
in situ. The benefit of a surface constrained reaction because the
reaction was aqueous based also proved to be highly suspect as
the improved chromatographic performance was only observed
with aromatic eluents commonly employed in nonsuppressed ion
chromatography. Aromatic eluents tend to mask the high-energy
sites associated with poorly hydrated subsurface ion-exchange
sites, making it difficult to assess whether or not chloromethylation
sites are truly constrained to the surface.
While the details of the synthesis have not been published,
both the chromatographic performance and the resin
specifications suggest that Hamilton PRP-X100 is synthesized using
the above paraformaldehyde-HCl functionalization methodology.
The chromatographic performance with hydrophilic eluent species
such as carbonate-bicarbonate suggests that a significant fraction
of the ion-exchange sites are poorly hydrated subsurface ion-
exchange sites. In fact, Hamilton PRP-X110S is supplied in a
carbonate-bicarbonate storage solution pre-equilibrated with
thiocyanate, added to the solution to mask any poorly hydrated
ion-exchange sites. Using thiocyanate to mask poorly hydrated
sites presents several problems for the user. First, any thiocyanate
in the mobile phase adds to the background conductivity,
raising detection limits. Second, because the affinity of thiocyanate
for the stationary phase is much higher than the affinity of
either carbonate or bicarbonate for the stationary phase, long
equilibration times are associated with each new batch of
thiocyanate containing eluent which of necessity will be slightly
different regarding relative concentrations. Production of
stationary phase is free from such poorly hydrated ion-exchange
sites is far preferable to adding masking reagents to the mobile
phase.

9.3.6 Polyvinyl Alcohol Backbone Anion-Exchange

Phases
Anion-exchange materials based on polyvinyl alcohol substrates
represents another new class of anion-exchange materials widely
 Anion-Exchange Phases 233

used today in ion chromatography. Such materials were initially

used for the preparation of reversed-phase polymers after
extensively esterifying surface alcohols with activated long
chain carboxylic acids. Such materials offer many advantages
including good peak shapes for relatively poorly hydrated anions
or relatively hydrophobic anions. These materials also provide
excellent resolution of fluoride from the system void. Due to the
relatively unrestricted hydration of the ion-exchange sites on the
surface and the very hydrophilic polymer substrate, favorable
conditions are achieved for retention of fluoride. Shodex produces
columns using this architecture [9]. The most notable of these
is the SI-50-4E column which provides excellent selectivity and
peak shapes for the common anions. Metrohm also sells a number
of these phases as OEM products [9]. They are easily identified
in that their specifications exactly match those from the Shodex
website for the corresponding column.
One disadvantage of using polyvinyl alcohol as the basis for
constructing the substrate particle is the limited physical strength
of such materials. While highly cross-linked substrate particles
composed of styrenic monomers tend to be a very high strength,
with a typical operating flow rate in the 1–2 mL per minute range
when using 4 mm column hardware, analogous polyvinylalcohol
materials typically have a maximum flow rate of 0.8 mL per minute.
This lower flow rate limit is due to the lower physical strength of
polyvinyl alcohol-based materials. Such a flow rate constraint
limits the overall speed of analysis when using smaller particle size
media.

9.3.7 Hyperbranched Anion-Exchange Phases on

Wide-Pore Substrates
The most recently developed stationary phase synthesis architecture
utilizes synthesis of hyperbranched anion-exchange materials
that are electrostatically attached to wide pore substrates. Such
materials are synthesized using a novel synthesis methodology [11]
utilizing two main ingredients: a primary amine and a diepoxide
monomer. In the first part of the synthesis, the primary amine and
the diepoxide monomer are allowed to react in the presence of a
wide-pore surface-sulfonated substrate at elevated temperature.
234 Ion Chromatography Columns

A low-molecular-weight polymer with tertiary amine functionality

at regular intervals along the polymer backbone is formed under
these conditions. A fraction of this polymer is electrostatically
attached to the surface of the wide-pore substrate. The excess
free polymer is then removed. Subsequently, the polymer-coated
substrate is allowed to react with the diepoxide. The reaction of
the diepoxide with the polymer quaternizes the tertiary amine
groups distributed along the polymer backbone while generating
pendant epoxy groups at each reaction site. All unreacted
diepoxide is then removed and replaced with a solution containing
primary amine. The primary amine reacts with all of the pendant
epoxide moieties to produce pendant secondary amine groups.
The sequence of reacting the polymer first with the diepoxide
monomer and then with the primary amine is termed “a reaction
cycle.” Subsequently, the excess primary amine is removed and
replaced again with diepoxide monomer. This time the diepoxide
monomer reacts twice with the pendant secondary amine groups
to produce two pendant epoxide moieties for each secondary
amine group produced by the previous reaction. After removing
any unreacted diepoxide, the resulting polymer is again exposed to
the primary amine to produce secondary amine groups, completing
the second reaction cycle. Because each secondary amine group
can react with two epoxy monomers, this results in a doubling of
the number of ion-exchange sites at the end of the second
reaction cycle. Reaction cycles can be repeated multiple times
until the desired selectivity and capacity is achieved.
Of course, the above is only a general description of what is
possible using reactions between amines and epoxides. A variety
of different materials are readily produced using a variety of
reagents. In total, more than a dozen phases are commercially
available utilizing some version of this chemistry. An example
chromatogram from a column synthesized using this chemistry
platform is shown in Fig. 9.11. The combination of an extremely
hydrophilic stationary phase backbone with a relatively unrestricted
hydration of the stationary phase allows for the application of
such materials to an extremely wide variety of analytes ranging
from highly hydrated anions such as fluoride to poorly hydrated
polarizable anions such as hexafluorophosphate with little if any,
compromise in peak shape.
 References 235

Figure 9.11 Separation of anions on 4 mm ID IonPac AS19-4 µm and

AG19 columns. Flow rate 1 mL per minute, 10 mM potassium hydroxide
0–10 min, 10–45 mM potassium hydroxide 10–25 min, 30°C, suppressed
conductivity detection, 10 µL injection volume. Peaks: (1) 3 ppm fluoride;
(2) 10 ppm chlorite; (3) 20 ppm bromate; (4) 6 ppm chloride; (5) 15 ppm
nitrite; (6) 25 ppm bromide; (7) 25 ppm chlorate; (8) 25 ppm nitrate;
(9) carbonate; (10) 25 ppm sulfate; (11) 40 ppm phosphate.

References

1. Scott, R. P. W. (1986) Bulk Property Detectors, in Journal of

Chromatography Library 33, ed. Scott, R. P. W., Chapter 3 (Elsevier,
Amsterdam) pp. 49–87.
2. Small, H., Stevens, T. S., and Bauman, W. C. (1975) Novel ion exchange
chromatographic method using conductimetric detection, Anal.
Chem., 47, pp. 1801–1809.
236 Ion Chromatography Columns

3. Small, H. (1989) Ion Chromatography, Plenum, New York.

4. Kolla, P., Köhler, J., and Schomburg, G. (1987) Polymer-coated cation-
exchange stationary phases on the basis of silica, Chromatographia,
23, pp. 465–472.
5. Jensen, D., Weiss, J., Rey, M., and Pohl, C. (1993) Novel weak acid
cation-exchange column, J. Chromatogr. A, 640, pp. 65–71.
6. Rey, M., Pohl, C. (1996) Novel cation-exchange stationary phase
for the separation of amines and of six common inorganic cations,
J. Chromatogr. A, 739, pp. 87–97.
7. Richens, D., Simpson, D., Peterson, S., McGinn, A., Lamb, J. (2003) Use
of mobile phase 18-crown-6 to improve peak resolution between
mono- and divalent metal and amine cations in ion chromatography,
J. Chromatogr. A, 1016, pp. 155–164.
8. Kadnar, R. (1999) Determination of amines used in the oil and gas
industry (upstream section) by ion chromatography, J. Chromatogr. A,
850, pp. 289–295.
9. Pohl, C. (2005) Ion chromatography, in Separation Science and
Technology; Handbook of Pharmaceutical Analysis by HPLC 6, eds.
Ahuja, S., and Dong, M., Chapter 8 (Academic Press, Amsterdam)
pp. 219–254.
10. Barron, R., Fritz, J. (1984) Effect of functional group structure on
the selectivity of low-capacity anion exchangers for monovalent
anions, J. Chromatogr. A, 284, pp. 13–25.
11. Pohl, C., Saini, C., (2008) New developments in the preparation of
anion exchange media based on hyperbranched condensation
polymers, J. Chromatogr. A, 1213, pp. 37–44.
Chapter 10

Ion Chromatography: Method

Development

Laila Kott
Xenon, Burnaby, B.C., Canada
[email protected]

This chapter provides the methodology for ion chromatographic

method development, which includes selection of eluant systems
and detectors, as well as sample preparation. Discussion
also includes the choice of column chemistries and gradient
separations.

10.1 Introduction
The use of ion exchangers has been known throughout history.
Aristotle reported the desalination of seawater using sands and
soils, but the widespread use of synthetic ion exchangers did not
come about until scientists discovered the ion-exchange capabilities

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
238 Ion Chromatography

of crushed gramophone records [1]. Since then the uses of ion-

exchange resins have ranged from the desalination of water, to
recovery of toxic or low abundant materials from industrial waste
and mine tailings to the current use as columns used in ion
chromatography, as described in this chapter.
Ion chromatography as a commercially available system has
been around since the 1970s. In the pharmaceutical industry, it
is largely used for counterion analysis, but has also been widely
used with amino acids, peptides and other biological-based
components (i.e., proteins, carbohydrates). It has use outside of
the pharmaceutical industry for drinking and wastewater
analysis and sulfite analysis in the wine and the food industry
[2–5]. Many methods are developed to analyze or detect one or
two components, often making the overall method development
uncomplicated. However, in the presence of complicated matrices,
ion chromatographic method development can be challenging.
This chapter will cover the components of the ion chromatograph,
how to choose the components for a given analysis, method
development, and general system troubleshooting.

10.2 Components of the System

The setup of the ion chromatographic (IC) system is similar to the
liquid chromatographic (LC) system with some key differences.
First and foremost, an IC system is designed to detect ions, much
like mass spectrometry, such that the column, the suppressor, and
the eluant are all chosen depending on whether the analysis is
looking for anions or cations. As described in the previous chapter
(Chapter 9), there are many choices of columns, depending on
the target analyte, but unlike like LC systems the IC system needs
to be considered as a whole, not just what column to use.
An IC system consists of an injector, a pump, an eluant (which
can now be generated in situ by an eluant generator), a column and
usually a guard column, a suppressor (not 100% ubiquitous but
becoming more and more common) and a detector, all connected
to a chromatography data system (see Fig. 10.1). The components
that differ from liquid chromatography: column, suppressor,
eluant, and detector will be discussed. For the components that are
 Components of the System 239

similar to the LC, they are programmed and maintained, as they

would be on an LC system.

Column
Injector LCpump Detector(s)
andoven

Eluant Columnand
j
Injector p p
LCpump pp
Suppressor ()
Detector(s)
generator
t oven

Figure 10.1 A comparison of the main components in an ion

chromatography system, compared to those in a liquid chromatography
system. Note that the eluant can be routed back to the suppressor for
regeneration.

10.2.1 Columns
Unlike liquid chromatography columns, ion-exchange columns
are predominantly polymer based and due to the nature of the
analyte (ions) are designed to be stable under more extreme pH
conditions than typical (U)HPLC columns. This is in part due
to the elution of analytes in IC being accomplished by acids or
bases, typically methanesulfonic acid, potassium hydroxide or
potassium carbonate. Since the predominant mode of detection for
IC is conductivity, it is important to keep the overall ionic strength
of the eluant as low as possible, which in turn requires less
suppression of the signal. As such, columns with low ion-exchange
capacities are preferred for standard pharmaceutical analyses,
as they require less eluant and therefore, also less suppression
[6]. In IC the type and design of the stationary phase in particular
plays a significant role in controlling selectivity due to the range
of stationary phase chemistries available; this being in contrast
to reversed-phase liquid chromatography, which is dominated
by similar octadecylsilica-based stationary phases [7].
Column selection begins first and foremost with type of
analysis, anionic or cationic. Figure 10.2 shows a representation
of the functional groups found on both the cationic and anionic
column types.
240 Ion Chromatography

Figure 10.2 Schematic showing the monovalent and divalent ionic

interactions on anion and cation resins. Divalent ions require two hydroxyl
or hydronium ions to be eluted off the column, whereas monovalent
ions require only one.

10.2.1.1 Anionic columns

There are a wide variety of IC columns available for anion analysis.
A typical starting point for anion counterion analyses would be
 Components of the System 241

to consider the type of anion and the capacity of the column.

Lower capacity is better for suppressed systems as mentioned in
Section 10.2.1 earlier. The functional group for anion analysis is
usually quaternary amine based, with different groups attached to
the amine, giving different selectivities and capacities of exchanger
(weak, medium or strong).
Some suggested columns to try first are Dionex IonPac AS
analytical columns manufactured by Thermo Scientific, Metrosep A
Supp 17 by Metrohm, Agilent-AX and Hamilton PRP-X100 columns.
These columns and their equivalent guard columns come in both 2
and 4 mm diameters. Choice of column can also depend on detector
type (i.e., conductivity vs. UV) and if the system to be used is
equipped with suppressors or not. On systems using suppressors,
one must always remember that the column diameter must match
the suppressor diameter. The above-mentioned IC columns are
designed to separate and analyze most inorganic anions and organic
acids using hydroxide or carbonate as the eluant and conductivity
detection. It is of note that some columns are designed to work
with a specific eluant. Both weakly retained monovalent anions
such as fluoride, acetate, formate, chloride and strongly retained
multivalent ions, such as phosphate and citrate can be efficiently
eluted in the same run by using isocratic or gradient elution
typically within 10 to 20 min. If the ion chromatographic analysis
has a more complicated matrix, is to be used for analyzing waste
water, drinking water, specific species defined by the EPA or for
larger biologic molecules like polypeptides, oilgonucleotides
or proteins, then a little research and comparison of available
column technologies will help in choice of column.

10.2.1.2 Cationic columns

There are a wide variety of IC columns available for cation
analysis. A typical starting point for cation counterion analyses
would be to consider the capacity of the column. As mentioned
before, lower capacity is better for suppressed systems. The
functional group for cation analysis is usually carboxylic (weak
exchanger) or sulfonate (strong) based, giving different selectivities
and capacities of exchanger.
242 Ion Chromatography

Some suggested columns to try first are: Dionex IonPac CS

columns manufactured by Thermo Scientific Dionex, Metrosep C
Supp 1 by Metrohm Agilent-CX and Hamilton PRP-X200 columns.
As with the anionic columns, these columns and their equivalent
guard columns come in both 2 and 4 mm diameters, but one
must always remember that the column diameter must match the
suppressor diameter, if one is being used. All the considerations
above apply to choice of cationic column: use of suppressor, type
of detection, choice of eluant. These IC columns are designed for
the separation and analysis of alkali, alkaline earth metals and
ammonium with conductivity detection. Both monovalent and
mulitivalent cations can be efficiently eluted in the same run by
using isocratic or gradient elution typically within 10 to 20 min.
Similarly as for anions, if the cationic matrix is more complex,
is to be used for analyzing wastewater, drinking water, specific
species defined by the EPA, transition metals or alkylamines
and alkanolamines, then, again, a little research and comparison
of available column technologies will help in choice of column.

10.2.2 Eluants
In ion chromatography, once the ion is retained by the stationary
phase, it falls to the eluant to disrupt the interaction of the
sample ion and the stationary phase to elute the ion. Some ion
chromatographs come with the ability to generate the eluants
in situ/inline, whereas other systems require that eluants are
prepared as they would be for liquid chromatographic separations,
at fixed concentrations. Since pH and ion concentration play such
an important role in the IC elution process, special care must be
taken when preparing IC eluants.
For eluants prepared offline (not using the eluant generators),
they must be prepared with fresh de-ionized water, with
sufficiently low conductivity (18.2 MW/cm at 25°C is typical) and
should be free of organic compounds as not to compromise IC
separation sensitivity. If unsure of the water quality or conductivity,
boil the water to remove any dissolved CO2, which can convert
to carbonate and change the background conductivity of the
eluant. Other considerations when preparing eluants are as
follows:
 Components of the System 243

• Eluants should be prepared and stored in plastic bottles

to avoid contamination of leached ions from glass bottles.
• The eluant should be degassed online or helium purged
before use.
• To avoid clogging of the IC system and to extend the life of
the column, the eluant should be filtered using a chemically
inert filter with 0.45 μm pores or smaller. Cellulose
filters are not stable at high pHs, so filters such as the
“hydrophillic” versions of polyvinylidenedifluoride (PDVF)
or polytetrafluoroethylene (PTFE) are recommended. Note
if they are not the hydrophilic version of these filters, they
cannot be used for filtering aqueous solutions.
• The concentration of the eluants should be at a high enough
concentration to cover the elution of the most tightly bound
ion. Typical concentrations are
o carbonate/Bicarbonate up to 25 mM;
o hydroxide (sodium or potassium) up to 130 mM;
o methanesulfonic acid (MSA) up to 65 mM.
For in situ eluant generation, a generator cartridge needs to be
installed on the IC system and a source of fresh de-ionized water
is required, with sufficiently low conductivity (18.2 MW/cm at 25°C
is typical), free of organic compounds and stored in plastic mobile
phase bottles, as for offline eluant preparation above. For properly
calibrated systems, the in situ eluant generation uses the flow rate
(inverse relationship) and applied current to deliver the correct
concentration, following Faraday’s laws of electrolysis and using
the proportionality:
Applied Current(mA)
[Eluent charge concentration(mM)] (10.1)
Flow Rate(mL/min)

10.2.2.1 Anionic eluants

As introduced above, eluants elute solutes based on their ionic
strength. The common anion eluants in order of their ionic strength
are:

Borate < Hydroxide < Bicarbonate < Carbonate.

244 Ion Chromatography

In general, the predominant method of detection for inorganic

and organic anions is by conductivity and therefore, hydroxide
tends to be the preferred anion eluant due to its low background
conductivity.
Hydroxide is a monovalent ion therefore, higher concentrations
of the eluant would be required for it to successfully elute
multivalent ions. Most commercial suppressors can handle
concentrations above 130 mM, allowing for use of hydroxide as
an eluant for more highly charged anions. However, as a good
general rule, the valencies of the anions to be eluted and the eluant
itself should be comparable. For example, if a sample mixture
contains a many divalent anions, it would be wise to choose
an eluant that also contains bivalent anions, such as sodium
carbonate.

10.2.2.2 Cationic eluants

The typical cation eluants in order of ionic strength are

Hydrochloric acid < Nitric acid < Methanesulfonic

acid (MSA) < Sulfuric acid.

In general, MSA is the first choice due to its low background

of conductivity. Divalent cations show a high affinity for strong
ion-exchange columns and cannot be eluted with dilute mineral
acids, which would require higher acid concentrations, leading to
a higher background conductivity, thereby making MSA the best
starting point for method development. It is important to note
that organic additives in the eluant such as trifluroacetic acid (TFA)
that can act as ion pairing agents are not recommended, as they
will shield the charge required for separation.

10.2.3 Suppressor Theory, Operation, and

Troubleshooting
Not all systems running ion chromatography have suppressors.
For those that do, the suppressor plays a central role as it is used
to reduce the conductivity of the eluant to as near to zero as it can
eletrolytically, which enhances the conductivity of sample ions
and thereby increases the sensitivity of detector. Due to this signal
 Components of the System 245

enhancement by the suppressors, they are often considered to be

part of the detection system. Since the incoming sample, column,
and eluant are anion/cation specific, so must be the suppressors.
Suppressors work by converting the bulk of the effluent,
which is in the form of the eluant (i.e., MSA or hydroxide) and the
counterions of the analytes, to either hydroxide ions for cations or
hydronium ions for anions. The result of the conversions means
that only the analyte ions of interest are detected by the
conductivity detector. The conversions are effected by having
hydroxide ions forming at the cathode of a cation suppressor,
which then migrate through an anion-exchange membrane to
neutralize the MSA eluant. The MSA– ions migrate to the anode,
again through an anion-exchange membrane, where they combine
with locally generated hydronium ions, maintaining neutrality.
A bit of gas is generated at both electrodes, which goes to waste.
This is the cause of the apparent bubbles eluting out of the
suppressor from the regenerant out port and is an indication of
proper suppressor function. The same is true for the anion
suppressor with hydronium ions being produced at the anode
which migrate through a cation-exchange membrane to neutralize
the eluant (hydroxide or carbonate). The remaining counter ion
(typically sodium or potassium) migrates to the cathode where
it combines with some locally generated hydroxide and flows to
waste.
Currently the most common type of suppressors are the
membrane-based suppressors, as they have a high suppression
capacity and can be continuously regenerated. The electrolytic
self-regenerating suppressor uses the effluent from the conductivity
detector, which as described above contains only the analyte
of interest and hydronium or hydroxide ion, as the source of
regenerant. Some suppressors use an external chemical regenerant
or an external source of de-ionized water with a pump as a
regenerant. An external regenerant would be required if the
effluent is being diverted to a second detector, i.e., to a mass
spectrometer, if organic solvents are used in the eluant or if the
flow rate is so low (as can be the case with microbore columns)
that a higher flow is required for regeneration. When developing
a method, the column diameter should be matched with a
suppressor with a similar diameter.
246 Ion Chromatography

Figures 10.3 and 10.4 show a schematic of how the cation and
anion suppressors exchange out the highly charged eluant and
use the detector effluent to self-regenerate. The detector effluent,
which has only a few anions in it, is directed back through the
suppressor as a regenerant. The water in the regenerant aides in
the suppression and regeneration of the suppressor.

Figure 10.3 Schematic showing the ionic flows through the suppressor to
the detector and then from the detector back through the suppressor
for regeneration: Cation.

10.2.4 Detectors
10.2.4.1 Conductivity detector (CD)
The conductivity detector (CD) is the most common detector for
IC analysis, as all ions are electrically conducting, making this a
 Components of the System 247

universal detector for this type of chromatography. The CD is a

modular detector with an integrated cell and has a signal range up
to 15000 µS/cm. The detector is not ion specific per se, however,
since the columns and the suppressors are ion specific, the
counter ion to the analyte ion will be unretained on the column
and will only be part of the background conductance.

Figure 10.4 Schematic showing the ionic flows through the suppressor to
the detector and then from the detector back through the suppressor
for regeneration: Anion.

When a sample is eluted from the suppressor, the sample ions

enter the detector where, since they are already naturally charged,
they increase the conductivity in the detector. The response
above baseline will then be the result of the analyte(s) of interest
and their response will be proportional to their concentration.
248 Ion Chromatography

The signal from the detector will be positive or negative, but

the chromatography software will ensure that the data signals
are displayed as positive peaks. Table 10.1 shows some typical
specifications of a conductivity detector, based on sodium ion
measurements.

Table 10.1 Typical specifications of conductivity detectors using sodium

ion [8]

Sensitivity 5 × 10–9 g/mL

Linear dynamic range 5 × 10–9 to 1 × 10–6 g/mL
Response indexa 0.97 to 1.03
aWhich indicates that all monovalent ions have the same response, all divalent ions

have similar response, etc.

The conductivity of a solution is measured by applying an

alternating voltage between electrodes in a conductivity cell. The
more ions pass through the cell, the more current is conducted
through the solution, the greater the signal that is measured as
there is a higher electrical conductivity. The electrical conductivity
can be described using the following equations. Ohm’s Law
states that an applied potential (V in volts) is directly related to
the current (I in amps) and the resistance (R in ohms):

V=R (10.2)

The resistance of a solution depends on the type and

concentration of ionic species in the solution and temperature.
The conductance of a solution (G, units of Siemens, S) is the
reciprocal of resistance:
1 (10.3)
G=
R
The conductivity (k) which has units of S/cm is also related
to the reciprocal of resistance and is also dependent on the
cross-sectional area of the electrode (A) and the distance between
the electrodes (d):

d d
k= = G = KG , (10.4)
AR A
 Components of the System 249

where K is the cell constant for the detector, with units of cm–1.
Since conductance of electrolytes will vary with concentration
(C), the equivalent or molar conductance (L) is defined per
1000 cm3 as

1000k (10.5)
L=
C
Combining Eqs. 10.4 and 10.5, we get conductance defined as:

LC
G= (10.6)
1000K

Defining l+ and l– as the ionic conductance from both the

positive and negative ions, or in the case of an infinitely dilute
solution, the limiting equivalent ionic conductance for the
cationic and anionic species in solution, Eq. 10.6 becomes:

( l + l )C (10.7)
G=
1000K

Using the values for the limiting equivalent ionic conductance

for common eluant and solute species typically used in ion
chromatography (listed in Table 10.2), the background conductance
of eluants can be calculated. For example, for a 5 mM KOH eluant
in a cell with a cell constant K of 10 and using Eq. 10.7 and the
values in Table 10.2, the background conductance due to the
eluant would be:

(73.5 + 198.6) 0.005

G=
1000 × 10

(272.1) 0.005
G=
10000

G = 0.00013605 S = 136.05 mS

Therefore, without suppression, the background conductance

of an isocratic method using 5 mM KOH would be 136.05 μS. This
calculation can also give you an idea on the effectiveness of your
suppressor or if the magnitude of the (counter) current being
250 Ion Chromatography

applied to the suppressor is sufficient to negate the background

conductance from the eluant.

Table 10.2 Limiting equivalent ionic conductances of commonly used

ions in IC chromatography at 25°C [9, 10]

Ion ln (S . cm2/eq) [ per mol]

Hydroxide OH– 198.6
Fluoride F– 55.4
Chloride Cl– 76.3
Bromide Br– 78
Iodide I– 77
Nitrate – 71.4
NO3
Sulfate 2– 80 [160]
SO4
Phosphate 3– 69 [207]
PO4
Chlorite – 52
ClO2
Chlorate – 64.6
ClO3
Hydrogen carbonate – 44.5
HCO3
Carbonate 2– 72 [144]
CO3
Thiocyanate SCN– 66
Formate HCOO– 55
Acetate CH3COO– 41
Ethanolate CH3CH2COO– 36
Benzoate C6H5COO– 32
Oxalate 2– 74 [148]
C2O4
Hydronium H+ 349.8
Lithium Li+ 38.7
Sodium Na+ 50.1
Potassium K+ 73.5
Ammonium + 73.5
NH4
Magnesium Mg2+ 53.05 [106.1]
Tetraethylammonium N(C2H5)+ 33
Calcium Ca2+ 59.5 [119.0]
Zinc Zn2+ 52.8 [105.6]
 Components of the System 251

As long as the dissociation of the ions is complete (which is

usually true for low concentrations typically used in ion
chromatography), conductivity is linear with ion concentration.
Temperature also affects the conductivity of a solution.
Equation 10.8 best describes the effect, based on a comparison
of several equations to describe the temperature effects on
conductivity detection [11]:

 h K
G25 = GT  T  , (10.8)
 h25 

where G25 and GT are the conductances at 25°C and T°C and
h25 and hT are the viscosities of water at 25°C and T°C. K is a
constant.
To ensure reproducible results, it becomes important to
control the temperature of the detector.

10.2.4.2 Other detectors and their uses

UV detectors can be used in ion chromatography (more commonly
on non-suppressed systems), however, since most standard
counter ions do not have any UV activity, it is not one of the
more standard detectors. It can be useful when analyzing highly
aromatic ions, amino acids or ions that can become detectible
through derivatization. These detectors have a large analytical
range, provided that signal absorption remains below 1 AU, adhering
to Beer’s law.
Amperometric detection is a type of electrochemical detection
which uses a working electrode, a counter electrode and a
reference electrode to measure either the oxidation or reduction
of compounds. This detector is useful for compounds with high
pKs, yielding low dissociation which would make them difficult to
detect by conductivity. It is also useful for the detection of
nonconductive compounds and increases the detection limits
for ionic compounds. Amperometric detection has been applied
to analysis of cyanamide [12], sulfite ions [13] and carbohydrates
[2].
252 Ion Chromatography

10.3 Method Development

Method development can start in a similar way as with developing
a method for liquid chromatography. The first step in both
method development plans is to search the literature for a
good starting point and to determine the theoretical pKa of the
molecule(s) of interest.
From the literature, there are a few review articles for novel
ionic compounds, and data has been compiled using the “Virtual
Column” software to optimize and compile IC retention data for
different IC columns [14] however the current predominant use
of ion chromatography is mostly for methods that involve only a
few compounds to separate, so often a thorough literature
search is not required. Determining the theoretical pKa, however,
is very useful. In liquid chromatography, one is concerned about
keeping the analyte neutral, so that it interacts as predicted with
the stationary phase of the column. The opposite is true for ion
chromatography. The analyte needs to be charged so as to interact
with the stationary phase of the column, and only when the ion
concentration of the mobile phase reaches a point where the
analyte cannot compete and is less likely to interact with the
column than the eluant itself, will the analyte of interest elute.
As one might expect, monovalent compounds are more easily
displaced than multivalent ions, which would, in most cases,
require a higher concentration eluant.
Looking at the pKa curves also helps to highlight whether
the analyte should be analyzed in anion or cation mode and
whether a concentration or a pH gradient or both might be needed.
It is essential to understand this as anions and cations require
different columns, suppressors, and eluants as discussed above.
The first requirement of ion chromatography method development
is to determine what ion type (anion or cation) will be analyzed.

10.3.1 Sample Preparation and Diluents

In general, de-ionized water is the best solvent for IC samples;
organic solvent use in sample diluents is not preferred for IC
analysis and can cause issues with some of the columns and
suppressors. When preparing samples in de-ionized water, use of
sonication, stirring or shaking can improve the solubility of sample.
 Method Development 253

Plasticware, plastic pipettes, or pipette tips and plastic vials

should be used exclusively when making standards, samples, and
solutions in ion chromatography as ions such as sodium can be
leached from glass.
If the sample does not readily dissolve in water but the ion of
interest can be extracted into the aqueous phase, filtering off the
insoluble part of the sample and analyzing only the supernatant is
a sample preparation approach. This approach can also avoid any
interference from high concentrations of the parent and reduce
wear and tear on the guard and analytical columns. There are
a variety of plastic filter vials available, where filtration can be
performed directly in an IC vial which may then be placed directly
into the autosampler. Spiked recovery studies should be performed
into the appropriate matrix (e.g., drug substance for counterion
analysis) to verify recovery using this in-vial filtration sample
preparation approach.
If use of some organic diluent is unavoidable, this may require
some replumbing of the IC system to avoid contact of the organic
solvents with components susceptible to such solvents. Even if
only a small amount of organic solvent is required, the guard column
may need to be replaced more frequently. Organic solvents in the
sample preparation may also impact the suppressor, column, and
baseline. Additionally, sample precipitation from contact with
the 100% aqueous mobile phase could be an issue depending
on the solubility of sample. Consider a small amount of organic
solvents such as methanol, ethanol, isopropanol, acetonitrile, and
dimethylsulfoxide (DMSO) to improve the solubility of the sample
only if the above recovery experiments fail and it is deemed
necessary.
A final sample preparation possibility is to do an aqueous
extraction by an organic solvent that is not miscible with water such
as ethyl acetate, heptanes, and methylene chloride. Once dissolved,
the extraction can be performed to get the ion of interest into
an ion chromatographic compatible diluent, namely water.

10.3.2 Isocratic and Gradient Elution

The back pressures in ion-exchange columns are similar to
core shell columns, which, in general, means that flow rates for
254 Ion Chromatography

ion chromatography can be faster than those in reverse-phase

chromatography (with non-core shell columns). One must still
keep below the pressure limits of the column and suppressor being
used, but flow rates up to 2.5 mL/min can be employed.
For an isocratic method, the retention of monovalent and
divalent ions shift to earlier retention times with increasing
concentration of eluant. Consequently, if a shorter method is desired,
the eluant concentration can be increased, but one must keep in
mind the changes required for the suppressor (an increase in
suppression current is required) and depending on how high
the concentration of eluant used is, there may be more noise in
the baseline, resulting in higher limits of detection.
Gradients can also decrease the run time, and have the
advantage that they do not require the higher eluant concentrations
for the entire run. Gradient elutions can be adjusted for the
elution of monovalent species early on and then an increase in
eluant concentration can have the divalent ions elute quickly after.
However, for multivalent ions, elution depends more on pH as
the valency of these ions is pH dependent. Phosphate is a good
example of this.
Effect of pH on the speciation of phosphoric acid:

H3PO4  H+ + H2PO4– pK1 = 2.16

H2PO4–  H+ + HPO42– pK2 = 7.21

HPO42–  H+ + PO43– pK3 = 12.33

For these multivalent ions and also for biologic samples

such as amino acids, a pH gradient instead of a concentration
gradient (also known as a salt gradient) needs to be employed.
In these cases two (or more) mobile phases are prepared at
varying pHs and elution occurs as the pump gradient changes
from one to the other thereby increasing or decreasing the pH.
The gradients can be simple, changing gradually only one or two
pH units or can cover a larger range. Depending on the complexity
of the solution being analyzed, (i.e., two amino acids vs. a protein
digest) several pH buffers may need to be employed. One must
also consider ionic strength, as pH gradients can result in both
higher pH and ionic strength, essentially having two modes of
 Method Development 255

elution working together [15]. The choice of which gradient

type to use depends on how close the other components in the
sample elute. If there are only two or three components then a
combination gradient can be used for sharp peaks and a quick
method. If, however, the other components are all closely eluting
then a more controlled pH gradient, maintaining the ionic strength
of the eluant throughout the change in pH may be required.
Gradients can be further sped up by starting at higher initial
isocratic eluant concentrations. One, however, is not limited to
just one gradient. If the sample has a complex matrix and/or ions
of different valencies, a method may require two or more different
gradients combined with isocratic sections.
It is recommended to always have some length of isocratic
portion as the first step. This is because IC systems can vary in
dead volume (dwell time) and set up, which can make method
transfers more challenging as the gradients do not always match.
If there is an isocratic portion at the beginning of the method,
then the gradient start time can be used to compensate for the
difference in dead volume/dwell time. This is accomplished by
starting the gradient sooner or later in the run, depending on
whether the instrument transferring in has a larger or smaller
dead volume/dwell time.

10.3.3 Suppressor Current Settings

For electrolytic self-regenerating suppressors, it is very important
to select the proper operating current. A current setting set too high
may permanently damage the suppressor. The optimum current
setting produces just enough hydronium or hydroxide ions to
displace the eluant counterions, thereby neutralizing the eluant.
The optimum current for the suppressor can be calculated
by the following equation, which is based on Eq. 10.1:

Current(mA) = [Flow Rate(mL/min)] × [Eluent Concentration(mN)]

(10.9)

Some software packages will calculate it or give a range of

applicable currents for an isocratic separation based on the choice
of eluant, its concentration, and flow rate.
 256
Ion Chromatography

͊
Figure 10.5 Chromatogram showing the change in the baseline as the eluant gradient changes the effluent conductivity
above the suppressor setting for 1 mM hydroxide, which was set 7 mA.
 Method Development 257

Using the relationship in Eq. 10.9, the current setting on

the suppressor can be adjusted during method development to
compensate for the background conductance of the eluant in order
to maximize your analyte signal. During method development,
this parameter should be adjusted to give a zero baseline (or as
close to it as is reasonable).
When running a concentration gradient elution, one sees the
baseline rise as the eluant concentration increases. A gradient can
improve the separation and give sharper peaks, while running
in a shorter time. However, with a fixed suppressor value, as
the gradient background conductivity increases past the set
suppressor current one sees a rise in the baseline. This can result
in a lower detection limit as the peak of interest may now have to
be integrated over a sloped baseline. Figure 10.5 shows the effect
of a 1 mM to 15 mM gradient over 4 min and the resulting rise
and then subsequent fall of the baseline as the gradient is returned
to its starting concentration.
For gradient separations, a current “ramp” can be programmed
similar to the eluant concentration ramp by calculating the
appropriate current at each programmed gradient concentration
adding the current change at the same time or similar times as
the gradient is changed. There might be a need for a slight delay in
the time program needed to change the suppressor current as the
change in suppressor current will be immediate, but the gradient
concentration change will need time to migrate through the system.
The delay in the suppressor current change will be related to the
system volume and flow rate.
Suppressors, just like columns, will degrade over time and
are considered consumables. The question then arises: How do
we know when to change the suppressor? High backpressure,
decreased resolution, and non-Gaussian peaks are indicative of an
aging column, however, some of these symptoms may also be due
to the suppressor and it may be difficult to tell which part of the
system is at fault. Increased background, therefore having to
increase the current setting and loss of linearity or 100% noisy
baseline with no peaks are also indicative of a deteriorating
suppressor.
258 Ion Chromatography

10.3.4 Temperature Effects

Temperature can affect chromatographic separations on an ion
chromatography column in a similar manner as it does in liquid
chromatography. Elution will be quicker at higher temperatures
and compounds will elute later at lower temperatures.
Increasing the temperature decreases the viscosity of solutions
and increases ion mobility. This directly affects the conductivity
of a solution, and as such, temperature control of the CD is also
important. A general rule for setting the temperature on the CD is
that it should be at or above the column temperature, such that
there is not a cooling gradient across the solution when entering
the detector. Typical column temperatures for analysis in ion
chromatography are 25 to 35°C. A general starting detector
temperature for method development for that range of column
temperatures would be 35°C. The maximum temperature for
these detectors is generally around 60°C.

10.4 General System Troubleshooting

Troubleshooting in ion chromatography is similar to that in other
types of chromatography, and a lot of the first pass troubleshooting
one does applies to these systems as well. Some basic symptom/
errors and how to correct them are as follows:
• High back pressure
o Reduce flow rate
o Check for blockage in system and/or column
o Sample crashing out of solution along the flow path
• Baseline drift
o Wait for system to stabilize (this can take a while)
o Check for small leaks
o If running a gradient, try a suppressor gradient
• Large sodium peak
o Use plasticware and plastic pipettes
• Sensitivity has decreased
o Column, suppressor or both are spent
 Applications 259

• High background conductivity

o Eluant is contaminated
o Wrong suppressor setting; check to make sure that the
current setting is appropriate for the flow rate and eluant
concentration used
• High noise spikes with normal conductivity
o Air in the system; check valves and connections for leaks
• High background conductivity with high noise
o Replace suppressor

10.5 Applications
Most ion chromatographic applications in the pharmaceutical
industry are currently counterion analyses, analyses of ionic
impurities or detection of certain genotoxic impurities, which
often require the detection and separation of one or two ions
only, making method development fairly simple. However, as drug
development expands to new modalities, including compounds
such as nucleotides, amino acids, peptides, polypeptides, and
proteins, all of which are now becoming part of or in fact, the
active pharmaceutical ingredient, ion chromatography will be
further challenged with their analyses. Chapter 9 and Paull and
Nesterenko [7] have compiled lists of commercially available
columns for amino acid and analysis of other biologics, showing
that column developers are expanding on their stationary phase
chemistries, to keep up with the trend towards analyzing more
biologics.
Along with the newer horizons in drug development, there
is continued scrutiny on the environment. There are many EPA
and USP methods currently using ion chromatography and that
number is likely to rise. The monitoring of wastewaters from
sewage plants, mining operations, water runoff from farming,
air quality monitoring are all areas where ion chromatography
is being used. As acceptable limits for human exposure of metals
and other ionic toxins in foodstuffs and drinking water are
decreasing so must the limits of detection in the analytical
methodology. This may lead to more interest in IC-MS methodology,
260 Ion Chromatography

which is available commercially but has yet to be employed to

the same degree as other mass spectrometric combinations.

Acknowledgments
The author would like to thank Anthony Landry, Jelena Petrovic,
and Amy McCusker for their help.

References

1. Eccles, H., and Greenwood, H. (1992) Chelate ion-exchangers: The

past and future applications, A users view. Solvent Extraction and
Ion Exchange, 10, pp. 713–722.
2. Bhattacharyya, L., and Rohrer, J. S. (eds.) (2012) Applications of
Ion Chromatography for Pharmaceutical and Biological Products
(John Wiley & Sons, New Jersey, USA).
3. Herbranson, D. E., Eliason, M. E., and Karnatz, N. N. (1987) Development
of a high performance ion chromatographic (HPIC) method for the
determination of sodium metabisulfite in parenteral formulations.
Journal of Liquid Chromatography, 10(5), pp. 3441–3450.
4. Koch, M., Köppen, R., Siegel, D., Witt, A., and Nehls, I. (2010)
Determination of total sulfite in wine by ion chromatography after
in-sample oxidation. Journal of Agricultural and Food Chemistry, 58,
pp. 9463–9467.
5. Iammarino, M., and Di Taranto, A. (2013) Determination and
validation of an ion chromatography method for the simultaneous
determination of seven food additives in cheeses. Journal of Analytical
Sciences, Methods and Instrumentation, 3, pp. 30–37.
6. Weiss, J. (2004) Handbook of Ion Chromatography, Volume 1, 3rd ed.
(Wiley-VCH Verlag GmbH & Co., Germany).
7. Paull, B., and Nesterenko, P. N. (2005) Novel ion chromatographic
stationary phases for the analysis of complex matrices. The Analyst,
130, pp. 134–146.
8. Scott, P. W. R. (2013) Ion Chromatography: Theory and Practice in
Chromatographic Analysis of Cationic and Anionic Species, Chrom-Ed
Series Book 5 (Integritext, UK).
9. Shpigun, O. A., and Zoltov, Y. U. (1988) Ion Chromatography in Water
Analyses (Ellis Horwood, Chichester, UK) p. 85.
 References 261

10. A Practical Guide to Ion Chromatography. (2007) Sequant AB,

Sweden. ISBN 978-91-631-8056-9.
11. Haddad, P. R., and Jackson, P. E. (1990) Conductivity detection, in
Ion Chromatography (Journal of Chromatography Library Volume
46), Chapter 9 (Elsevier Science, Sydney, Australia).
12. Nair, J. B. (1994) Determination of trace levels of cyanamide in a novel
potassium channel activator bulk drug by pulsed electrochemical
detection. Journal of Chromatography A, 671, pp. 367–374.
13. Casella, I. G., and Marchese, R. (1995) Sulfite oxidation at a platinum
glassy carbon electrode. Determination of sulfite by ion exclusion
chromatography with amperometric detection. Analytica Chimica
Acta, 311(2), 199–210.
14. Park, S. H., Shellie, R. A., Dicinoski, G. W., Schuster, M. T., Haddadd,
P. R., Szucs, R., Dolan, J. W., and Pohl, C. A. (2016) Enhanced
methodology for porting ion chromatography retention data. Journal
of Chromatography A, 1436, pp. 59–63.
15. Wang, Y., Moreno, G. T., Zhang, B., Zhang, L., Farnan, D., Patapoff, T. Ionic
strength-mediated pH gradient ion exchange chromatography. WO
2014/078729 A l, 22 May (2014).
Chapter 11

Fundamentals, Properties, and

Applications of Stationary Phases
for Gas Chromatography Method
Development

Omprakash Nacham and Jared L. Anderson

Department of Chemistry, Iowa State University,
1605 Gilman Hall, Ames, Iowa, USA
[email protected], [email protected]

Gas chromatography (GC) is an important separation technique

that has experienced great popularity within numerous disciplines
of the chemical sciences. Technological advancements in column
fabrication as well as the development of new stationary phase
chemistries have played a significant role in increasing the
applicability of GC. To address the challenges associated with
complex samples, a wide range of stationary phases with different
chemical selectivities and thermal properties have been developed.

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
264 Stationary Phases for Gas Chromatography Method Development

The introduction of poly(siloxane) and polyethylene glycol

stationary phases has greatly improved the separation power of
GC. Importantly, the development of new stationary phase bonding
techniques to immobilize these stationary phase materials has
significantly increased the thermal stability, durability, and lifetime
of GC columns. Ionic liquids (ILs) and polymeric ionic liquids (PILs)
have emerged as promising stationary phase materials. They
demonstrate the potential to complement nonionic stationary
phases because of their interesting and useful physicochemical
properties. The structural tunability and high thermal stability
of ILs have provided potential alternatives to conventional
stationary phases, particularly in the analysis of polar compounds
(e.g., hydrogen bond donors) as well as for high-temperature GC
separations. Techniques that are employed during stationary
phase immobilization/deposition greatly influence the separation
efficiency of the GC column. As a result, a fundamental understanding
of stationary phase properties, column evaluation, and principles
involved in column fabrication methods is useful in GC analysis.
This chapter provides an overview on the properties, applications,
and procedures used in the fabrication of GC columns that are
essential in aiding GC method development.

11.1 Introduction
Gas chromatography (GC) is a separation technique that has been
widely used for the analysis of volatile and semi-volatile chemical
constituents. GC has become a workhorse separation method in
a variety of applications, including petroleum analysis [1], food
sample testing [2], environmental sample screening [3], and
drug analysis [4]. The remarkable work on GC in the 1940s by
A. J. P. Martin and R. L. M. Synge is considered a milestone in the
evolution of modern GC techniques.
Advances in column technology have played a significant role
in transforming the face of GC. Before optimizing any GC analysis,
the selection of an appropriate column type as well as stationary
phase is an important prerequisite. GC columns are broadly
divided into two classes, namely, packed columns and open tubular
columns. In packed columns, a solid material (possessing high
surface area) or a liquid coated on the solid support typically
 Introduction 265

functions as the stationary phase. High sample capacity, low phase

ratio (ratio of the volume of gas to liquid phase), and a plethora
of stationary phase materials that are compatible with column
fabrication procedures are characteristics of the classical packed
columns. However, lengthy analysis times, low column efficiency,
poor resolution, and band broadening are some drawbacks of
packed columns [5, 6]. Problems associated with classical packed
columns are partly alleviated by the introduction of micropacked
columns. They possess inner diameters lower than 1 mm with
particle sizes ranging from 40–100 µm. This type of column has been
successfully employed for trace level analysis and exhibit potential
for coupling with mass spectrometry (MS) [6]. Yet, relatively
low column lengths, demand for modified injection devices to
withstand high inlet pressures, and reduced throughput has
significantly limited practical applications of the micropacked
columns [7]. Despite their limitations, classical packed columns
and micro-packed are commonly utilized for preparative-scale
separations.
The growth and developments in open tubular column
technology have largely increased the applicability of gas-liquid
chromatography (GLC) to many areas of research within the
chemical sciences. The concept of open tubular columns and its
potential advantages over packed columns were first explained
by Golay in the late 1950s [8]. When a stationary phase (liquid or
adsorbent) is either coated on an inert support or directly applied
on the walls of packing material with capillary dimensions, the
resulting columns are called open tubular (capillary) GC columns.
High peak capacity, low peak dispersion, high plate numbers, good
reproducibility, and relatively low analysis times are examples of
attractive features for open tubular columns [9]. Among the different
open tubular column configurations, wall-coated open tubular
columns (WCOTs) have enjoyed much great success and exhibit
compatibility with a broad range of GC instrumentation. Owing to
their versatile nature and favorable stationary phase properties,
WCOTs are considered as a first choice of columns for most
analytical GC separations.
The selectivity of a stationary phase plays a crucial role in
separating multi-component samples. In parallel to advancements
in column technology, a significant amount of effort has been
invested towards the development of innovative new stationary
266 Stationary Phases for Gas Chromatography Method Development

phase materials that can impart the desired selectivity and high
thermal stability to satisfy the increasing demands of complex
samples. Today, a wide range of GC columns are available with
different chemical selectivities and thermal properties. As a result,
selection of an appropriate stationary phase that can provide high
selectivity as well as good separation efficiency is an important
task and holds significant impact in GC analysis.
This book chapter provides an overview on the various column
fabrication procedures and different types of stationary phases
utilized in GC analysis. The first section of the chapter focuses on
fundamentals involved in column preparation, evaluation, and their
influence on gas chromatographic separation. The second section
of this chapter encapsulates the properties, applications, and
limitations of current stationary phases such as poly(siloxanes) and
polyethylene glycols. Finally, emerging stationary phases including
ionic liquids (ILs) and PILs will be discussed along with their
advancements with respect to improved chemical selectivities and
thermal stabilities for providing solutions to separating complex
sample.

11.2 Column Preparation and Evaluation

11.2.1 Materials and Methods Involved in Preparation

of WCOT Capillary Columns
The tubing materials used in the fabrication of capillary GC columns
include stainless steel, nickel, glass, and fused silica [6]. Capillary
GC columns manufactured with fused silica and glass are more
common than other tubing materials. The introduction of the glass
capillary drawing machine in the early 1970s by Desty revolutionized
capillary column technology [10]. In comparison to other tubing
materials, glass is cheap, easily deformable, and has a homogenous
surface. For nearly a decade, glass capillary columns have remained
as a mainstay in GC separations. However, the poor wettability,
presence of high amounts of residual metallic impurities (sodium,
calcium, barium), and chemical reactivity with analytes present
serious challenges in the use of glass capillary columns [9].
Fused silica capillary columns were first introduced by
Dandeneau and Zerenner and co-workers in the 1970s [11]. Fused
 Column Preparation and Evaluation 267

silica capillary columns offer greater flexibility and chemical

inertness when compared to glass columns. They are mostly
composed of primary silanol groups and contain less than 1 part per
million (ppm) of metallic impurities and are supported externally
by a polymeric coating (e.g., polyimide). Today, most capillary GC
columns are prepared using fused silica tubing.
To coat a thin layer of liquid stationary phase on the surface of
tubing material, the surface tension of the stationary phase should
be lower than the critical surface tension of the coating surface.
Surface modification reactions are often employed to increase
the free surface energy (reduce the surface tension) as well as to
deactivate the free silanol groups. There are different types of
surface modification reactions available depending on the column
tubing material and stationary phase chemical composition. Acid-
leaching, etching with reactive gases [12], exhaustive silylation, and
particle-deposition [13] are some common methods to increase
the free surface energy of the glass and fused silica capillaries
[14, 15]. For example, acid-leaching using 20% (v/v) hydrochloric
acid at 100–170°C followed by silianization has greatly improved
the surface wettability of fused silica columns for coating non-polar
and moderately polar stationary phases. Most particle deposition
methods use a suspension of sodium chloride or barium carbonate
and are utilized for roughening glass surfaces. The surface activity
of fused silica or glass capillary columns can be reduced by
either high-temperature silylation reactions or condensation with
poly(alkylhydrosiloxanes) [16].
WCOT columns (glass or fused silica) are coated using either
the dynamic or static coating method. In the dynamic coating
method, a solution of the stationary phase (5–15% w/v) is typically
prepared in a suitable coating solvent and forced through the
column by an inert gas. Subsequent to coating, the temperature
of the gas is increased to evaporate the coating solvent. Various
modifications to this method were made to provide uniform and
reproducible coatings. Importantly, a short plug of mercury is
introduced between the coating solution and a driving gas. As
the mercury plug is moving through the column, a uniform and
thin layer of coating solution develops on the wall of the capillary
column [17]. This method has largely improved the reproducibility
of the coating. However, an important weakness associated
with the dynamic coating method is the difficulty in predicting
268 Stationary Phases for Gas Chromatography Method Development

the thickness of the stationary phase. Alternatively, static coating

methods enable the preparation of columns with a predictable
thickness. In static coating methods, the desired coating solution
(0.02–4.0% v/v) is prepared in a suitable volatile organic solvent
(e.g., dichloromethane, pentanes, etc.). After completely filling the
capillary, the coating solvent is evaporated by sealing one end of
the column and attaching another end to the vacuum source. As
the solvent is evaporated under controlled vacuum conditions, a
thin layer of stationary phase is deposited on the column surface.
The thickness of the stationary phase is related to concentration
of coating solution as well as inner diameter of the capillary
column [9, 18]. Although static coating methods are superior for
producing reproducible and thin stationary phases, these methods
can be time-consuming and laborious.
The demand for new column fabrication procedures requiring
less time and yielding more reproducible coatings is rapidly
increasing. Sol-gel coating techniques minimize the total number
of steps involved in the column fabrication procedure in which
surface roughening, deactivation, coating and/or immobilization
are performed in a single step. Wang et al. employed the sol-gel
coating technique for the fabrication of a PDMS column (efficiency
of 3400 plates/meter; 10 m × 250 µm i.d.) and demonstrated
its utility for separating a mixture of saturated and unsaturated
fatty acid methyl esters [19]. Using this technology, two columns,
SolGel-msTM and SolGel-WAXTM, have been recently commercialized.
The applicability of sol-gel techniques for different classes of
stationary phase materials is still under constant development.

11.2.2 Evaluation of Gas Chromatographic Column

Performance
The quality of open tubular GC columns is gauged based on three
important parameters, including efficiency, chemical inertness,
and thermal stability. Numerical values such as plate number (N),
effective plate number (Neff ), and plate height (H) are commonly
used to express column efficiency. These values are determined
by measuring the retention time of an inert chemical substance
(alkanes, aromatics, etc.) under isothermal conditions [20].
The selected probe should exhibit reasonable retention time and
enough solubility in the stationary phase. The stationary phase
 Column Preparation and Evaluation 269

film thickness, column dimensions, carrier gas type, and velocity

largely influence the column efficiency [21]. For example, as the
inner diameter of the column decreases, the efficiency of the
column increases. For a commercial WCOT column (30 m × 0.25 m
i.d.) with a film thickness of 0.25 µm, N values vary typically
from 5000–6000 plates/meter (non-polar stationary phase)
[9]. However, depending on the physicochemical properties (i.e.,
viscosity, polarity, surface tension) of the stationary phase and
applied coating method, efficiencies can be as low as 1000 plates/
meter. Often low column efficiency results from improper coating
techniques that can cause uneven film formation or a continuous
film with variable thickness resulting in low efficiency [21]. An
efficiency test is a quick and easy diagnostic tool to assess the
column quality, but it should be complemented with other tests
to obtain comprehensive information on column performance.
Subjecting the coated capillary columns to standardized
test mixtures enables access to information about the chemical
inertness and separation efficiency. The Grob test and Luong
test are commonly used to examine the secondary interactions
of the stationary phase [22, 23]. These interactions result in
peak broadening, peak tailing, irreversible adsorption, and
degradation. The Grob test mixture is a solution composed of
twelve different analytes with a wide range of polarities. On the
basis of peak shape and peak resolution of the test analytes,
the column performance is estimated. For example, based on the
peak shapes of alcohols (1-octanol and nonanol), column activity
due to residual silanol groups on non-polar stationary phase can
be determined.
Film instability and decomposition and/or volatilization of
the stationary phase at high column temperatures play a crucial
role in diminishing the column performance. Therefore, the
identification of maximum allowable operating temperature
(MAOT) of a stationary phase prior to GC separation is an
important prerequisite. To determine the MAOT, a thin layer of
the stationary phase is deposited on the capillary surface and
subjected to a programmed temperature run ranging from 40 to
400°C using a low-temperature ramp. During these measurements
and at a specific temperature or range of temperatures, chemical
constituents (intact stationary phase or decomposed products)
270 Stationary Phases for Gas Chromatography Method Development

that evolve from coating materials contribute to a significant

rise in the baseline which is considered as the MAOT of the
stationary phase [24–26]. Alternatively, the determination of a
column’s efficiency before and after holding the column at high
temperatures (approximately 12 h) provides valuable information
on the thermal stability and chemical inertness of the stationary
phase (isothermal GC method) [21]. Coupling any of these two
techniques to sensitive detectors such as mass spectrometry (MS)
enables the structural information of any decomposed products.
In research laboratories as well as in the production environment,
this data can be used to investigate decomposition mechanisms
and aids in tailoring the chemical structure to reduce overall
thermal decomposition. In some cases, data obtained from
thermogravimetric analysis (TGA) is used for establishing the
thermal stability of new stationary phase materials such as ionic
liquids. Due to lower sensitivity of the TGA method when compared
to GC-based methods, thermal decomposition temperatures
obtained from TGA are not considered a true estimation of MAOT
for a GC stationary phase.

11.3 Selectivity of Gas Chromatographic

Stationary Phases
Various models have emerged for the classification of stationary
phases and to facilitate the column selection process for GC method
development. Solvent strength (polarity) and solvent selectivity
are the primary basis for earlier models to classify stationary
phases. Since the solvation process is the result of multiple and
independent interactions between a solute and stationary phase,
the use of single value polarity scales does not always provide
reliable information on stationary phase properties. On the other
hand, solvent selectivity models are formulated by considering
various solute-solvent intermolecular interactions including
dispersion, induction, orientation, and hydrogen bonding. The
system of phase constants introduced by Rohrschneider (further
developed by McReynolds) and the Abraham solvation parameter
model are the two most common models used to classify stationary
phases based on selectivity [27–29].
 Selectivity of Gas Chromatographic Stationary Phases 271

The Rohrschneider–McReynolds system assumes that

intermolecular interactions are additive, and their individual
contribution to retention can be evaluated from the difference in
retention index values for a series of prototypical compounds on
the stationary phase to be classified. Within this model, squalane
is used a reference nonpolar phase and retention index values
are determined with respect to squalane at 120°C. Although
McReynolds phase constants are successfully used to characterize
various stationary phases, this method has some limitations. The
measurement of retention indices at 120°C results in minimum
interactions with some stationary phases and complicates
phase constant determination. In addition, a fewer number of
test solutes can be used to derive the phase constants and it is
assumed that a single dominant intermolecular interaction of
test solute is responsible for the solvation process. Selectivity
scales based on Gibbs free energy of solution have been employed
to understand the retention characteristics of the stationary
phase. However, these methods turned out to be unreliable
and provided inexplicable results for various stationary phases
chemistries [30].

11.3.1 Solvation Parameter Model

Today, a vast majority of new stationary phases are classified
using system constants generated from the Abraham solvation
parameter model. This model assumes that the solvation process
occurs in three steps. Initially, a cavity similar to a solute’s volume
is formed in the stationary phase. Subsequent to reorganization
of solvent molecules around the cavity, the solute is transferred
into the cavity and interacts with solvent molecules. It can be
assumed that the free energy associated with cavity reorganization
is negligible compared to other processes, and gas-solute and
solute-solute interactions in the gas phase are negligible. The
retention equation describing the solvent parameter model
describing all possible interactions between the solute and
stationary phase is given in Eq. 11.1.

log k = c + eE+ sS + aA + bB + lL (11.1)

272 Stationary Phases for Gas Chromatography Method Development

In this method, a set of probe molecules with various

functional groups is applied on stationary phase coatings, and the
dependent variable (log k) chromatographic retention factor for
each probe molecule is measured. The retention factor is a function
of interactions with the stationary phase and functional groups
on the probe molecule. As shown in Eq. 11.1, each probe molecule
has a unique set of solute descriptors defined as follows: E is
excess molar refraction, S is the solute dipolarity/polarizability
index, L is the solute gas-liquid distribution coefficient at 298 K,
A and B are solute hydrogen bond acidity and basicity, respectively.
The values for the solute descriptors can be obtained from
literature. The coefficients of solute descriptors are called the
system constants (e, s, a, l, b) which are temperature dependent
and provide a measure of interactions between the stationary
phase and solute molecules. The system constant e provides the
information of solute-stationary phase interaction via p–p and
n–p interactions, s is a measure of stationary phase dipolarity, l
defines the ability of the stationary phase to undergo dispersive
interactions with solute molecules, a and b are measure of
stationary phase hydrogen bond basicity and acidity, respectively.
These system constants are calculated by subjecting the retention
factors and solute descriptors to multiple linear regression
analysis [24, 31–34]. Tables 11.1 and 11.3 represent the system
constants of selected GC stationary phases. More information
regarding the system constants of various stationary phases
at different temperatures can be obtained from several sources
[35–38].

11.4 Stationary Phase Materials

Since the introduction of GC, numerous liquid materials have
been used as stationary phases. Unfortunately, most of the early
stationary phase materials were abandoned due to lack of
reproducibility in their production, undefined chemical composition,
poor retention reproducibility, and low thermal stability [9].
However, with greater advancements in chemical synthesis and
column fabrication techniques, liquid stationary phases that can
circumvent these challenges while providing improved selectivity
and chromatographic performance have been produced.
 Stationary Phase Materials 273

Table 11.1 System constants at 120°C for representative poly(siloxane)-

based stationary phases from different column manufacturers

Solvation parameter model

System constants
Stationary
phase e s a b l Ref.

DB-1 0 0.215 0.175 0 0.507 [35]

HP-5a 0.004 0.306 0.199 0 0.585 [35]

Rtx-20a 0.030 0.535 0.245 0 0.545 [35]

DB-35a 0.074 0.640 0.277 0 0.531 [35]

Rtx-65a 0.120 0.828 0.330 0 0.528 [35]

DB-225b –0.030 1.232 1.176 0 0.457 [35]

DB-200c –0.0335 1.041 0.147 0 0.465 [35]

DB-1: 100% PDMS.

aNumber indicates the percentage of diphenylsiloxane monomer.
b50%cyanopropylphenylsiloxane monomer.

cPoly(dimethyltrifluoropropylsiloxane).

11.4.1 Aliphatic and Aromatic Hydrocarbon-Based

Stationary Phases
Separations on hydrocarbon-based stationary phases are mostly
due to dispersive-type (cohesive) interactions. High-molecular-
weight hydrocarbons such as hexadecane, squalane, Apolane-87
and different grades of Apiezons are common hydrocarbon-
based stationary phases [39]. Attempts to prepare hydrocarbon
phases with some degree of branching as well as incorporation of
halogens as substituents has resulted in diminished selectivity [6].
This is likely due to orientation effects and reduced dispersion
interactions with the solute molecules. Squalane and Apiezon
require extensive chemical purification as their precursor materials
are often obtained from natural sources. There are several types
of Apiezon, including Apiezon A, B, C, L, M, MH, and W. Of all these
grades, Apiezon MH and L are commonly used because of their
better quality and reproducibility in production [9]. Apolane-87
is a well-characterized synthetic hydrocarbon phase possessing
274 Stationary Phases for Gas Chromatography Method Development

higher thermal stability than other hydrocarbon phases. Sojak

and co-workers developed WCOT columns with Apolane-87 and
subsequently used them for the separation of linear pentadecene
isomers. It has been claimed that Apolane-87 exhibited superior
selectivity and higher thermal stability than squalene [40].
Hydrocarbon phases are used for the analysis of petrochemical
samples and for the separation of perfluorocarbons from partial
or non-fluorinated hydrocarbon samples [6]. They are mostly
used as non-polar reference phases in various stationary phase
classification schemes. Important shortcomings of hydrocarbon
stationary phases include susceptibility to oxidation, mixed
retention mechanisms (providing active sites on the column walls),
poor wettability, volatility, and low MAOT. The chemical stability
of aliphatic phases was significantly improved by the introduction
of perfluorocarbon-based stationary phases [41]. These are
mostly used for the separation of reactive chemicals such as metal
fluorides, halogens, and halide compounds of hydrogen, sulfur,
and phosphorus where the conventional phases are prone to
degradation. In some cases, highly-fluorinated stationary phases
are used for the separation of isomeric perfluorocarbons and
Freons [42, 43].
To overcome problems related to volatility and thermal
stability of aliphatic hydrocarbons, polyaromatic compounds
were investigated as liquid stationary phases. Compounds such
as benzyldiphenyl, alkylnapthalenes, polyphenyl tar, and linear
polystyrene are among some earlier liquid stationary phases
used for the separation of aromatics and heterocyclic compounds.
However, the minimum column operating temperature required
for these stationary phases range from 60 to 230°C. This shortened
range has reduced their suitability as stationary phases towards
a wide variety of samples [6]. Most of the aliphatic and aromatic
hydrocarbon phases have been replaced by robust and thermally
stable poly(organosiloxane)-based liquid stationary phases.

11.4.2 Poly(Siloxane)-Based Stationary Phases

Poly(siloxanes) represent one of the most popular stationary
phase materials in capillary GC. The advances in silicone chemistry
have provided the opportunity to exploit poly(siloxanes) as
 Stationary Phase Materials 275

indispensable stationary phases for many complicated separations.

Relative to hydrocarbons, the high bond energy of Si-O and Si-C are
responsible for high thermal stability and low vapor pressure of
poly(siloxanes) [44, 45]. Depending on their chemical composition,
the viscosities of poly(siloxane) phases are slightly affected
by temperature. This represents an important advantage over
aliphatic and aromatic hydrocarbon stationary phases. In addition,
good wettability, low glass transition temperature, and stable
film formation capabilities have made these materials suitable
GC stationary phase candidates.
Figure 11.1 depicts representative chemical structures of
poly(siloxanes). Substituents groups such as methyl, vinyl, phenyl,
and 3-cyanopropyl are attached as R groups to tailor the chemical
selectivity of the resulting poly(siloxanes). Poly(dimethylsiloxane)
(PDMS; R = methyl) is the most commonly used non-polar
stationary phase. Because of its wide operating temperature range,
PDMS (DB-1; commercial identity) is capable of resolving samples
with a broad range of boiling points. In most cases, this is the first
liquid stationary phase of choice when there is little information

Figure 11.1 Common structural motifs of poly(siloxane)-based stationary

phases containing varying functional groups that result in different
chromatographic selectivity.
276 Stationary Phases for Gas Chromatography Method Development

on a sample’s composition. By increasing the percentage of

diphenyl or cyanopropyl monomer content in the PDMS phase,
poly(siloxanes) with a wide range of polarities are generated.
This effect can be clearly observed from the change in s (system
constant used to measure dipolarity of the stationary phase)
value of the stationary phases. For example, by introducing 65%
diphenylsiloxane monomer (Rtx-65) in PDMS, the s value increases
from 0.215 to 0.828, as shown in Table 11.1. The high dipolar-type
interactions in Rtx-65 are used for the separation of phenols and
fatty acid samples. Table 11.2 presents the commercial identities
and selected applications of various poly(siloxane) phases. For
Table 11.2, a representative column is chosen from each class of
poly(siloxanes) and their recent gas chromatographic separations
are reported.

11.4.2.1 Preparation of poly(siloxane)-based stationary

phases for open tubular columns
For the preparation of alkyl or phenyl substituted poly(siloxanes),
dichlorosilanes or dialkoxysilanes are usually subjected to acid/
base hydrolysis [44, 46]. In the presence of functional groups
such as esters and cyclodextrin (which are sensitive to hydrolysis
conditions), poly(siloxanes) phases are often prepared via
hydrosilylation techniques [47, 48]. Owing to the depolymerization
effects of acid/base catalysts on poly(siloxanes) via cyclosiloxane
formation at high temperatures, several purification methodologies
are employed following the synthesis of poly(siloxanes) [9].
The thermal stability and viscosity of poly(siloxane) phases
are highly dependent on the attached functional group and its
position relative to the siloxane bond. The presence of electron
withdrawing groups, such as cyano and fluoro, adjacent to the
siloxane bond greatly enhance the decomposition at lower
temperatures. To maintain the desired selectivity and thermal
stability, a minimum distance of three carbon atoms from the
siloxane bond is usually maintained. Furthermore, it was observed
that the incorporation of bulky groups such as phenyl (Fig. 11.1)
and corborane in the polymer backbone resulted in improved
thermal stability with minimal cyclosiloxane formation.
 Representative
Table 11.2 Applications of poly(siloxane)-based stationary phases

Stationary phase Commercial identity column Sample types Sample matrix Ref.
100% PDMS AT-1, BP-1, CB, CP-sil-5, DB-1, DB-1 Steroids, organochlorine Air; ginger extract; [81–87]
(non-polar) Equility-1, HP-1 ms, Optima-1 pesticides, mixture of essential crude oil; soil; shale
ms, oils, aromatic and saturated gas; tea.
DB-1 ms, ZB-, AT-1 ms, Sol-gel- hydrocarbons, pyrethroid
ms, HP-1, MDN-1, Optima-1, pesticides, chemical warfare
OV-1, agents
OV-101, SE-30, Rtx-1, SPB-1,
VF-1 ms
5% phenyl, 95% AT-5, AT-5 ms, BP-5, BPX-5 ms, DB-5 Alkylphenol, parabens, aroma Urine, blood; breast [76–80]
dimethysiloxane CP-Sil-8, CP-Sil-8, CB-ms, compounds, polychlorinated milk; particulate
(non-polar) DB-5, Equity-5, HP-5, HP-5 dibenzofurans, polychlorinated matter, pomegranate;
ms, MDN-5, MDN-5 s, MTX-5, dibenzo-p-dioxins, agricultural soils;
Optima-5, Optima-5 ms, OV-23, polybrominated dibenzo-p- fresh water.
Permabond SE-52, PB-5, RTX-5, dioxins, essential oils,
RTX-5 ms, SPB-5, Ultra-2, ZB-5, Polychlorinated biphenyls,
007-5 unsaturated isoprenoids
6% Cyanopropyl AT-624, BP-624, CP-624, CP-30, DB-624 Low-molecular-weight silanol Biogas; gas; [92–97]
phenyl, 94% DB-624, DB-30, DB-VRX, HP- and siloxanes; volatile organic municipal
dimethylsiloxane 624, compounds (VOCs); residual sludge sample;
(mid-polar) HP-30, Optima-624, Optima-30, solvents; odor-active sulfur and pharmaceutical drugs
PE-624, PE-30, RTX-624, RTX- nitrogen compounds; flavoring and intermediates;
Stationary Phase Materials

30, RTX-Volatiles, VF-624, ZB-5, excipients fermented sausage;

(Continued)
ZB-624,007-624, 007-502 pork liver powder.
277
 278

Table 11.2 (Continued)

Representative
Stationary phase Commercial identity column Sample types Sample matrix Ref.
35% phenyl 65% AT-35, BPX-35, DB-35, HP-35, DB-35 Pesticide residues; polycyclic Tobacco; fruit [88–91]
dimethylsiloxane MDN-35, RTX-35, SPB-35 aromatic compounds and samples; cigarette
(mid-polar) tobacco specific N-nitrosamines; smoke; urine; grape
methylmalonic acid (biomarker leaves.
for organic aciduria); plant
primary metabolite profiling

14% cyanopropyl BPX-10, CB-1701, CP-Sil-19CB, DB-1701 Organophosphorus pesticides; Plant; tobacco [101–
phenyl 86% DB-1701, DB-1701P, HP-1701, organochlorine pesticides; products; skimmed 103]
dimethylsiloxane Optima-1701, Rtx-1701, trihalomethanes; trimethyl silyl milk; drinking water;
(mid-polar) SPB-1701, ZB-1701, 007-1701 derived monosaccharides plant.

50% Cyanopropyl BP-225, BPx-50, HP-225, DB-225 Pectic polysaccharides; mono Fruit samples; [98–100]
Stationary Phases for Gas Chromatography Method Development

phenyl 50% DB-225, Optima-225, OV-225, unsaturated fatty acid profiling plasma; serum.
dimethylsiloxane RTX-50, RTX-225, SP-2250,
(polar) ZB-1701, 007-225
 Stationary Phase Materials 279

Earlier approaches used in the preparation of poly(siloxanes)

relied on the direct application of coating solutions on the wall of
capillary materials. These columns are unsuitable for long-term
use and their low thermal stabilities precluded their use in high-
temperature GC separations. The introduction of simultaneous
stationary phase immobilization and cross-linking techniques has
greatly improved film stability, durability, and thermal stability
of poly(siloxanes). Techniques such as thermal condensation and
radical initiated cross-linking methods are used to immobilize
poly(siloxanes) on the walls of open tubular columns. Most of the
modern GC capillary columns are produced using one of these two
techniques. In thermal immobilization techniques, non-polar or
moderately polar poly(siloxanes) with terminal silanol groups are
used as starting materials. A high column temperature is used to
cross-link the linear polymer chains and to react silanol-terminated
poly(siloxanes) with the walls of capillary materials (fused silica/
glass). Occasionally, a small of percentage of cross-linking agent
such as alkyltrimethoxysilanes is also included in the coating
mixture to enhance the reaction efficiency as well as improve
the thermal stability [49, 50].
In free radical immobilization methods, endcapped
poly(siloxanes) are cross-linked onto the deactivated column using
a variety of radical initiators. Dicumyl peroxide [51], azo-
compounds [52], g-radiations [53] are some examples of radical
generators. Unlike thermal techniques, organic substituents
attached to the poly(siloxanes) participate in cross-linking
reactions and no new siloxane bonds are formed with adjacent
polymer chains. However, the success of cross-linking largely
depends on the chemical nature of substituents. In general, neat
PDMS is a good substrate for cross-linking whereas the presence
of cyano, phenyl, and long alkyl chain lengths reduces its reaction
efficiency. In such cases, a small percentage of poly(siloxane)
monomers endcapped with functional groups such as tolyl and
vinyl are added to enhance the reaction efficiency. These
immobilization techniques not only help in obtaining columns
with high thermal stability, but also enable thicker coatings
(1.0–8.0 µm) and provide stationary phases that can tolerate
solvent rinsing procedures. The latter has a significant advantage
in extending the column lifetime and is especially useful for removing
280 Stationary Phases for Gas Chromatography Method Development

non-volatile components that can originate from complex sample

matrices [44, 54].
Immobilization approaches involving sol-gel technology
has reduced the total number of steps involved in conventional
immobilization methods. However, only a few column types
have been commercialized using this method. Nevertheless,
technological advancements made in stationary phase bonding
chemistry have enormously increased the types of applications
and samples that can be analyzed by GC.

11.4.2.2 Modified poly(siloxane) stationary phases for chiral

analysis
Since the separation of isomeric molecules with similar vapor
pressure is a formidable task using conventional poly(siloxanes),
chiral selectors are often either tethered or dissolved in
poly(siloxane) phases for chiral GC applications. Crown ethers
[55], chiral amino acids [56], calixarenes [57], resorcinarenes [58],
and cyclodextrins [59, 60] are some common chiral selectors
often used in chiral GC analysis. In chiral separations, hydrogen
bonding, coordination, steric interactions, and inclusion
formation are four important types of interactions responsible
for enantioselectivity [9]. The chemical cross-linking of chiral
selectors (amino acids, cyclodextrin, and metal bound ligands) to
poly(siloxanes) and subsequent immobilization and/or coating
on GC capillary has increased the applicability of chiral GC
analysis [61].

11.4.2.3 Chiral separations using α-amino acid-based

poly(siloxanes)
The inspiration for the design of amino acid-based chiral
stationary phases (CSPs) was conceived from the stereoselectivity
of enzymes which catalyze biochemical reactions to produce a
stereospecific product. The chiral poly(siloxane) stationary phase
containing valine diamide was termed as Chirasil-Val. Commercially,
this product is available in both enantiomeric L-Val and D-Val
forms. On D-Val column, L-amino acids elute earlier than D-amino
acids and vice versa. This is highly useful in performing method
validation and enantiomer labeling experiments [61]. These
columns possess thermal stability from 70–250°C and exhibit low
 Stationary Phase Materials 281

tendency for column bleeding. Chiral separations on diamide-

based stationary phases predominantly occur via hydrogen
bonding interactions. A variety of synthetic approaches have been
investigated in the preparation of different Chirasil-Val derivatives
to improve their enantioselectivty and thermal stability.
Recently, Levkin et al. prepared a new binary CSP comprising
valine diamide and β-cyclodextrin on poly(siloxane) backbone
using platinum-catalyzed hydrosilylation (Chirasil-DexVal-C11).
Interestingly, Chirasil-DexVal-C11 exhibited superior performance
in the enantiomeric separation of proline amino acid, which is
problematic using only neat valine-diamide-based CSPs [62].

11.4.2.4 Chiral separations using cyclodextrin-based

poly(siloxanes)
Enantiomeric separations using cyclodextrin phases is possible
due to inclusion complexation between the guest (analyte) and
the host (cyclodextrin) molecules. Cyclodextrins (CDs) are
oligosaccharides containing 6, 7, or 8 glucose units, designated
as α, β, γ cyclodextrins, respectively [9, 63]. The melting point
and solubility of CDs are typically varied by either O-alkylation or
O-acylation reactions. CDs are coated as a stationary phase in
either diluted or undiluted form. In the diluted approach, the
polarity of CD is adjusted according to dissolving solvent such as
poly(siloxanes) and subsequently the CD solution is coated using
static coating method. In the undiluted approach, low-melting
CD derivatives (per-O-pentylated and 3-O-acylated-2,6-di-O-
pentylated) are directly coated on the walls of open tubular
columns [64]. In the diluted methods, modified CDs such as
permethylated β-CD and 2,6-di-O-methyl-3-O-trifluoro-β-CD are
often dissolved in semipolar poly(siloxane) (OV-1701) for the
preparation of CD columns [65, 66].
Blum and co-workers developed second-generation CDs that
contain bulky tert-butyldimethylsilyl (TMDS) residues which are
dissolved in poly(siloxane) PS-086 and are used for chiral GC up to
250°C [67]. Compared to permethyl-β-CD, the TBDMS-derivatives
showed improved solubility in poly(siloxanes) and exhibited higher
enantioseparation factors. An important advancement in CD-based
CSP preparation is chemical linking of CDs to poly(dimethysiloxanes)
(Chirasil-Dex). This eliminates the requirement for a semi-polar
282 Stationary Phases for Gas Chromatography Method Development

solvent matrix. Using either thermal or radical initiated immobili-

zation techniques, CD linked poly(siloxanes) are immobilized on
fused silica and glass capillaries [59, 68]. Commercial products
such as Chirasil-Dex-CB involve a bonded chiral stationary phase
which has an operating temperature range from –25 to 200°C.
This type of CSP are successfully used in the separation of a
wide variety of racemic compounds. New developments in CD-
based CSPs include the emergence of linear dextrins based on
maltohepatose. These phases comprise linear oligosaccharides
(maltohepatose) with tert-butyl dimethylsilyl and acetyl groups
on primary and secondary hydroxyl sites. The chiral selectivity of
linear dextrin was demonstrated by separating α-amino acid
derivatives and halogenated compounds on fused capillary
columns [69]. Alternative solvents for dissolving the CD chiral
selector have also been explored. It was shown that ionic liquid-
based CDs exhibited broad enantioselectivities and greater
thermal stabilities [70]. Poly(siloxanes) modified with calixarenes,
benzo-crown ethers, and resorcinarene represent a new class of
CSPs. These are mainly limited to research laboratories and require
numerous developmental studies prior to commercialization.

11.4.3 Poly(Alkylene Oxide)-Based Stationary Phases

In addition to poly(siloxanes), poly(alkylene oxides) represent
the most popular contemporary GC stationary phases used for the
separation of polar compounds. These are polymers of ethylene/
propylene oxide units with terminal hydroxy groups. The hydrogen
bonding capabilities of these stationary phases are extensively
exploited for separating different classes of polar compounds.
By varying the molecular weight and chemical composition, the
selectivity and thermal stability of the resulting polymer can
usually be controlled [6]. Poly(ethylene oxide) (PEG) materials are
prepared by polymerizing ethylene oxide using metal hydroxide
as catalysts. The polydispersity (molecular weight distribution
index) of PEG has a significant impact on the viscosity, thermal
stability, and selectivity of the stationary phase. The low-molecular-
weight PEGs significantly interfere with analyte retention.
Previous reports suggest that conditioning the column at high
temperatures relieves the effects of low-molecular-weight PEGs.
 Ionic Liquids and Polymeric Ionic Liquids as Emerging Stationary Phase Materials 283

Often PEG phases are prone to auto oxidation at high temperature

due to the presence of ether linkages. This situation is exacerbated
in the presence of oxygen and leads to shorter column lifetimes
and low thermal stabilities (<200°C). The use of oxygen scavengers
in the flowing carrier gas has significantly improved the column
performance [71]. Since water molecules can hydrogen bond
with alkylene oxide chains, the use of moisture free carrier gas is
warranted to ensure reproducible analysis. Carbowax 20M is the
most commonly used PEG stationary phase in both packed and
open tubular columns. The column operating temperature of
Carbowax 20M (bonded to fused silica capillary) ranges from 60°C
to 250°C. For applications requiring high column temperatures,
more pure and high-molecular-weight Superox-20M can be used
[9, 72]. PEG phases are mostly used for separating mixture of
amines (primary, secondary, and tertiary) [73], alcohols from
hydrocarbons, and ketones, aldehydes from aromatics [6]. The
polarity of the PEG phase can be modified by chemical linking
to PDMS. DX-1 and DX-4 are PEG phases that contain 90% and
15% poly(dimethylsiloxne), respectively. The condensation of
terminal hydroxy groups of PEG with 2-nitroterephthalic yielded
a new stationary phase, FFAP (free fatty acid phase; commercial
identity: FFAP; SP-1000; Carbowax 20M-dinitroterephthalate)
mostly used for the separation of organic acids, alcohols, amides,
ketones, and lactones. Due to the presence of free carboxylic acid
groups within the substituent, the FFAP phase is not used for
separation of basic compounds, aldehydes, and other compounds
that can react with acidic groups.

11.5 Ionic Liquids and Polymeric Ionic Liquids as

Emerging Stationary Phase Materials
Although PEG and poly(siloxane)-based stationary phases have
been widely used in GC, the quest for new stationary phase materials
that can complement them as well as increase the chemical
separation space is highly desirable. ILs represent a new class of
non-molecular solvents with interesting physicochemical and
solvation properties. These are molten organic salts comprising
bulky organic cations and inorganic/organic anions that exhibit
284 Stationary Phases for Gas Chromatography Method Development

melting points at or below 100°C [74]. Since the introduction of

air-stable 1,3-dialkylimdazolium-based ILs, the applicability of
ILs has been greatly expanded to various areas in the chemical
sciences [75]. The structural tunability of ILs to tailor their
physicochemical properties is highly advantageous for designing
optimal GC stationary phases. In addition, the unique solvation
properties (e.g., hydrogen bond acidity) as well as high thermal
stability enable them to be used as potential alternatives to
conventional stationary phases for the analysis of samples that
require high temperatures.

11.5.1 Influence of Physicochemical Properties of ILs on

Column Performance
Important physicochemical properties of ILs including melting
point, viscosity, thermal stability, surface tension, and chemical
inertness play crucial roles in developing new IL-based stationary
phases. Consequently, these properties greatly influence the
separation performance of the column.

11.5.1.1 Melting point

The melting point of the IL determines the minimum operating
temperature of the column. When an IL is used as a GC stationary
phase at a temperature below its melting point, low column
efficiency usually results due to mixed retention mechanisms
(adsorption and partition) [32]. Therefore, room-temperature ILs
(RTILs; melting point/glass transition below room temperature)
are of significant interest since they extend the operating
temperature range of the stationary phase. Based on previous
studies, structural features such as low ion symmetry, charge
delocalization, and high conformational freedom yield ILs with
low melting points [104, 105]. Figure 11.2 represents typical
structures of cations and anions of low melting point ILs. For most
of the ILs containing the same cation composition, halide-
based anions usually exhibit higher melting points than the
corresponding bis[(trifluoromethyl)sulfonyl]imide [NTf–2],
–
trifluoromethanesulfonate [TfO–], tetrafluoroborate [BF4], and
hexafluorophosphate [PF–6] anions [32].
 Ionic Liquids and Polymeric Ionic Liquids as Emerging Stationary Phase Materials 285

Cations Anions

Figure 11.2 Representative structures of cations and anions often used

in preparation of room-temperature ionic liquids.

11.5.1.2 Viscosity
The viscosity of the IL is an important property that influences
the film stability of the stationary phase, particularly at high
temperatures. Owing to long-range columbic forces and
intermolecular interactions (e.g., hydrogen bonding), ILs exhibit
at least two to three orders of magnitude higher viscosities than
organic liquids at room temperature. Within the class of 1-alkyl-
3-methylimidazolium-based ILs, increasing the length of the alkyl
side chain causes a significant increase in viscosity [21, 106].
Effective charge delocalization tends to reduce the viscosity of
ILs, whereas substituents that reduce ion-symmetry often result
in increased viscosity. Halide-based anions enhance the viscosity
of the IL by participating in hydrogen bonding interactions
while [NTf–2] anions tend to reduce the viscosity due to charge
delocalization and poor hydrogen bonding capabilities [32, 107].
Other approaches to increase the viscosity involve incorporation
286 Stationary Phases for Gas Chromatography Method Development

of functional groups that can result in increased intermolecular

interactions (e.g., p–p and hydrogen bonding). However, these
approaches are only applicable if the chemical selectivity of
stationary phases is not compromised.

11.5.1.3 Thermal stability

In contrast to molecular liquids, the columbic interactions within
ILs prevent the escape of ions into the gas phase and subsequently
reduce their vapor pressure over a wide range of temperatures.
For this reason, the MAOT of an IL is mostly based on the film
instability and onset of thermal decomposition rather than its
vapor pressure. The examination of bleed profiles of IL-based
stationary phases at high temperatures using MS suggested that
thermal decomposition of the cation/anion is greatly responsible
for origin of the bleed components [108].
Apart from low vapor pressures, high thermal stability
is another important feature of ILs that has increased their
popularity in GC. The chemical properties of both the cation and
anion are critical in determining the thermal stability of the IL.
ILs comprising less coordinating or weak nucleophilic anions
(e.g., [NTf –2] and [TfO–]) exhibit higher thermal stability than halide-
based anions. In comparison to imidazolium cations, quaternary
ammonium cations possess low thermal stability due to their high
propensity to participate in Hofmann elimination and reverse
Menshutkin reactions at high temperatures [109, 110]. ILs derived
from phosphonium cations usually exhibit higher thermal stability
than their imidazolium and other cyclic quaternary ammonium
counterparts. It was shown that thermal decomposition of
imidazolium-based ILs progress via N-dealkylation by attack of
the anion to generate the amine [111, 112]. When compared to
monocationic systems, ILs with multiple cations (dication and
trication) demonstrated high thermal stabilities. For example,
the thermal stability of tricationic imidazolium-based ILs
containing [NTf –2] anions are approximately 100°C higher than
the corresponding monocationic analogs. Payagala and co-
workers prepared various tricationic ILs and investigated their
physicochemical properties and employed them as GC stationary
phases. It was observed that tripropylphosphonium-based ILs
containing [NTf –2] anions exhibited thermal stability as high as
 Ionic Liquids and Polymeric Ionic Liquids as Emerging Stationary Phase Materials 287

330°C (based on isothermal GC method, refer to Section 11.2.2)

[113].

11.5.1.4 Surface tension

Surface tension and wetting ability are important properties that
determine the film forming capacity of ILs. For 1-alkyl-3-methyl-
imidazolium-based ILs, an increase in the alkyl chain length
results in a gradual decrease in the surface tension [114]. Larger
anions such as [NTf –2] and [PF–6] usually yield ILs with low surface
tension [115]. In general, ILs with high surface tension often
pose challenges during the static coating process. In such cases,
the structure of cation and/or anion can be modified to produce
a lower surface tension. On the other hand, surfactant doping
can also be used to reduce the surface tension of the IL. However,
it should be emphasized that these approaches may alter the
chemical selectivity of the resulting IL stationary phase [116].
The purity of ILs largely influences its physicochemical
properties as well as separation efficiency when used as a
stationary phase. The presence of halide impurities in ILs often
results in a stationary phase with low thermal stability and tailing
peaks (especially for alcohols and acidic compounds). In addition,
water contamination often leads to lower viscosity and surface
tension, and presents severe challenges in the coating process.
To circumvent the coating challenges, ILs are often dried under
reduced pressure for longer times to remove volatile organic
impurities and water prior to coating.

11.5.2 Imidazolium and Phosphonium-Based IL

Stationary Phases
Owing to their relatively high thermal stability and stable film
forming capabilities, imidazolium and phosphonium cation motifs
are extensively exploited for generating IL-based stationary
phases. Indeed, most of the commercial IL columns are also
manufactured using these cations. For more details on other cation
systems (ammonium, pyridinium, piperidinium, etc.) and their
relevant GC applications, several publications have extensively
discussed this aspect [21, 32].
288 Stationary Phases for Gas Chromatography Method Development

The 1,3-dialkylimidazolium cations constitute the most

studied single cation IL-based GC stationary phases. The anionic
component of 1,3-dialkylimidazolium ILs largely influences the
thermal stability as well as melting point of the resulting ILs. In
most cases, ILs containing [NTf –2] anions yielded low melting
points and high thermal stability compared to structural
counterparts that contain halide-based anions. The applicability
of 1,3-dialkylimidazolium-based ILs as GC stationary phases was
first reported by Armstrong and co-workers in the late 1990s
[117]. The separation behavior of ILs for various organic molecules
revealed that ILs exhibit dual nature selectivity where they can
function as nonpolar stationary for separating nonpolar molecules
and as a polar stationary phase for separating polar molecules.
The chemical selectivity of IL stationary phases can be readily
varied by simple anion/cation exchange reactions. For example,
the 1-butyl-3-methylimidazolium (BMIM) IL containing chloride
anion showed a selectivity for polar organic molecules via
hydrogen bond donor/acceptor interactions whereas analogous
the [PF–6] anion-based IL demonstrated higher selectivity for
non-polar compounds [117]. Within the 1,3-dialkylimidazolium-
based ILs, the hydrogen bond basicity of the stationary phases is
largely determined by the anion. Based on the Abraham solvation
parameter model, the hydrogen basicity of various imidazolium-
based ILs comprising different anions follows: chloride >
octylsulfate > hexafluoroantimonate ≈ [TfO–] > [NTf–2] > [PF–6]. Polar
molecules such as alcohols and acids (proton donors) exhibited
severe peak tailing when separated using ILs containing [NTf–2]
anions. In these situations, ILs containing [TfO–] anions may
be potential alternatives to improve the chromatographic
performance [24].
In comparison to most of the nonionic stationary phases,
the cation functionalized imidazolium-based ILs yielded high
hydrogen bond acidic stationary phases. Anderson and co-workers
examined the solvation properties of various imidazolium and
pyrrolidinium ILs containing the tris(pentafluoroethyl)trifluo
rophosphate (FAP–) anion using solvation parameter model. It
was observed that in comparison to unfunctionalized cations, ILs
with 2-hydroxyethyl substituents possessed high hydrogen bond
 Ionic Liquids and Polymeric Ionic Liquids as Emerging Stationary Phase Materials 289

acidities and demonstrated superior selectivity for compounds

that contain proton acceptor groups. The tunability of imidazolium-
based ILs for structure modification in pursuit of desired chemical
selectivity was greatly exploited by incorporating different
functional groups [118].
Phosphonium-based ILs represent higher thermal stability
stationary phases relative to other cation systems. However,
structural tuning of phosphonium-based cations is not straight-
forward and often involves multiple synthesis procedures. As a
result, fewer of these stationary phases are reported compared to
imidazolium-based ILs. Breitbach and co-workers reported the
synthesis of eight monocationic phosphonium ILs with different
anions and investigated their solvation properties and thermal
stabilities. It was shown that with same anion composition, the
hydrogen bond basicity of phosphonium-based ILs is greater
than imidazolium ILs. Due to lack of acidic hydrogen atoms,
most of these ILs exhibited low or negative hydrogen bond
acidity system constants. In addition, phosphonium-based ILs
demonstrated superior capabilities in separating a mixture of
24 flavor and fragrance homologous ester compounds. For the
trihexyltetradecylphosphonium triflate [P66614+] [TfO–] IL a thermal
stability as high as 405°C was observed. Phosphonium-based ILs
tend to interact less with solutes through p–p or n–p interactions
[119]. Table 11.3 represents the properties, applications, and
system constants (determined at 100°C) of selected IL-based
columns.
Multi-cationic systems using either imidazolium or
phosphonium cations yielded ILs with high thermal stability and
stable films at high temperatures. Multi-cationic platforms provide
additional opportunities to impart desired chemical selectivity
and the ability to tune the physicochemical properties including
surface tension, viscosity, and thermal stability of the resulting
IL. Most of the commercial ILs columns are manufactured using
either dicationic imidazolium or phosphonium cation motifs
with [NTf –2] anions, as shown in Table 11.3. Many of the commercial
IL columns are non-bonded stationary phases. SLB-IL100 and
SLB-IL111 are extensively used for the separation of fatty acid
methyl esters (FAMES) from various sample matrices [35, 120, 121].
 290

Table 11.3 Chemical composition, MAOT, and system constants of selected ionic liquid-based stationary phases

Solvation parameter model

System constants
Ionic liquid and maximum operating
temperature (°C) e s a b l Application Ref.
1-benzyl-3-methylimidazolium [TfO]; 260 0 1.70 2.41 0 0.47 Selectivity and thermal stability [24]
1-(4-methoxyphenyl)-3-methylimidazolium 0.28 2.05 2.03 0.16 0.41 compared to different stationary
[TfO]; 260 phases

1,2-dimethyl-3-(3-cyanopropyl)imidazolium 0.25 1.81 1.25 0.30 0.35

[NTf2]; 250
Analysis of n-alkanes, esters,
1,2-dimethyl-3-(3-hydroxypropy)imidazolium 0.21 1.81 1.25 0.30 0.35 [127, 128]
ketones, and alcohols
[NTf2]; 250
1-butyl-1-methylpyrrolidinium [NTf2]; 300 0 1.44 1.25 0 0.48
Stationary Phases for Gas Chromatography Method Development

1-butyl-3-(2-hydroxycyclohexyl) imidazolium –0.29 0.67 3.11 0.31 0.49 Analysis of FAMES, alkanes, and [130]
[BF4]; 250 esters

1-(1,2-dimethyl)-3-propylimidazolium [NTf2]; — — — — — Analysis of n-alkanes, esters, [127]

320 and alcohols
2-methyl-1-propylpyridinium [NTf2–]; 250 0.06 1.62 1.19 0.17 0.31
Analysis of n-alkanes, alcohols,
3-methyl-1-propylpyridinium [NTf2]; 250 0.07 1.60 1.15 0.15 0.31 esters, phenols, and organic [128]
acids
4-methyl-1-propylpyridinium [NTf2–]; 250 0.10 1.57 1.13 0.10 0.32
 Solvation parameter model
System constants
Ionic liquid and maximum operating
temperature (°C) e s a b l Application Ref.
1,9-bis(2,3-dimethylimidazolium)nonane [NTf2]; 0.37 1.62 1.13 0.00 0.14 Analysis of alcohols, esters, [128]
250 phenols, and organic acids
1,14-di(3-methylimidazolium)-3,6,9,12- 0.18 1.77 1.79 0.14 0.41 Analysis of flavor and fragrance [129]
tetraoxopentadecane; 350 mixture
1,14-di(3-[2-hydroxyethyl]limidazolium)3,6,9,12 0.27 1.76 2.42 0.14 0.36 Analysis of flavor and fragrance [129]
-tetraoxopentadecane [TfO]; mixture
1,10-di(1-decylimidazolium)decane [NTf2]; –0.06 1.67 1.84 0.17 0.59 [132]
1,10-di(1-decylimidazolium)decane [FeCl4]; 0 1.78 2.32 0.50 0.72
Hydrocarbon analysis using
Trihexyl(tetradecyl)phosphonium [NTf2]; 300 –0.28 1.55 1.55 -0.15 0.75
two-dimensional GC [131]
Trihexyl(tetradecyl)phosphonium [FAP]; 290 –0.29 1.39 0.49 0.25 0.63
Trihexyl(tetradecyl)phosphonium [FeCl4]; 320 –0.26 1.43 1.23 0.02 0.69
Tetraoctylphosphonium [NTf2–]; 385 — — Flavor and fragrance analysis [119]
Trihexytetradecyllphosphonium bis(2,4,4-trimet — — — — — Analysis of n-alkanes, esters, [127]
hylpentyl)phosphinate; 300 and alcohols
Ionic Liquids and Polymeric Ionic Liquids as Emerging Stationary Phase Materials

(Continued)
291
 292

Table 11.3 (Continued)

Solvation parameter model

System constants
Ionic liquid and maximum operating
temperature (°C) e s a b l Application Ref.
Trigonal tripropylphosphonium [TfO] — — — — — Trace analysis of water [138]
1,12-di(tripropylphopshonium)dodecane [NTf2] — — — — — Analysis of haloanisoles, [135–137]
(SLB-IL59/60); 300 benzothizaoles, nitrosamines
1,12-di(tripropylphopshonium)dodecane[NTf2] — — — — — Analysis of octa-chlorinated [122]
[TfO] dibenzo-p-dioxins and
(SLB-IL61); 250 dibenzofurans
Tri(tripropylphosphoniumhexamido)triethylam — — — — —
ine (NTf2) Analysis of FAMEs [121]
(SLB-IL76); 270
Stationary Phases for Gas Chromatography Method Development

1,12-di(2,3-dimethylimidazolium) dodecane — — — — — Analysis of plasticizers and [134]

[NTf2] (SLB-IL82) synthetic musks
270
1,9-di(3-vinylimidazolium) nonane [NTf2] — — — — — Analysis of FAMEs [121, 133]
(SLB-IL100); 230
 Ionic Liquids and Polymeric Ionic Liquids as Emerging Stationary Phase Materials 293

It was reported that IL-based columns provided the best

separation performance for FAMES when compared to
bis(cyanopropyl)siloxane columns. Importantly, their ability to
separate cis-and trans-isomers is important for the identification
of trans fats in food products [21]. Recently, Do and co-workers
investigated the applicability of SLB-IL111, SLB-IL61 and
SLB-IL76 for the separation of 136 polychlorinated dibenzo-p-
dioxins and dibenzofurans (PCDD/Fs). Owing to their high polarity
and multiple solvation interactions, SLB-IL111 and SLB-IL61
separated about twice the number of PCDD/Fs congeners than
conventional nonionic stationary phases [122].

11.5.3 Polymeric Ionic Liquid-Based Stationary Phases

In addition to multi-cationic ILs, polymeric ionic liquid (PIL)-
based stationary phases have been investigated to improve
the film stability, viscosity, and thermal stability. Commonly,
imidazolium cations functionalized with polymerizable groups
(e.g., vinyl and vinylbenzene) are used to prepare the PIL phases.
These materials can be developed using either an in-column
approach [123] or polymerized outside of the column in a suitable
solvent system that can be directly coated onto the surface of the
capillary [124]. Using the static coating method, a mixture of
monomer, cross-linker, and a free radical initiator (e.g., AIBN) are
deposited onto the wall of capillary column. Subsequently, the
coated column is subjected to temperatures that can initiate the
polymerization reaction. The physicochemical properties (e.g.,
melting point, thermal stability) of PIL phases are regulated by
varying the percentage of cross-linker composition in the coating
solution [125]. For example, a partially cross-linked PIL phase
provided efficient chromatographic performance up to 280°C,
whereas a highly cross-linked analog was stable up to temperatures
as high as 360°C [123]. Recently, Anderson and co-workers developed
a series of functionalized PIL columns that contain different
amounts of cross-linkers. The developed PIL columns were
investigated in the analysis of kerosene and diesel samples using
comprehensive two-dimensional GC. It was observed that at the
optimized PIL composition, the thermal stability and resolution of
selected analytes are significantly higher for PIL-based stationary
phases compared to PEG stationary phase [125]. The solution-
294 Stationary Phases for Gas Chromatography Method Development

based polymerization approaches are used to generate PIL phases

(homopolymers) that do not contain cross-linkers. Compared
to the PIL columns that are prepared using the flexible cross-
linkers, homopolymers exhibited higher melting points and lower
column efficiency [21, 124].
Recently, two new polymeric phosphonium ILs were prepared
using an in-column polymerization approach. The thermal stability
of these PIL phases was evaluated based on a series of isothermal
measurements of retention time of naphthalene at different
temperatures (100 to 380°C). It was demonstrated that the
polymeric diphenylphosphonium cation that containing two allyl
groups with [NTf2–] anion, demonstrated high thermal stability
with no peak tailing at temperatures as high as 330°C [126].
Phosphonium-based PILs are relatively a new class of stationary
phases materials that have potential to yield high thermal
stability stationary phases compared to imidazolium counterparts.

11.6 Conclusions and Future Directions

Developments in column technology as well as progress in the
design/production of new stationary phase materials have
tremendously increased the applications of GC. The introduction
of new bonding and surface modification techniques has greatly
improved the thermal stability and enhanced the long-term use of
poly(siloxane) and PEG stationary phases. In addition, the ability
to incorporate a variety of functional groups within poly(siloxane)
motifs has yielded stationary phases with a wide range of
chemical selectivities. The chemical linking of chiral selectors to
poly(siloxanes) and subsequent immobilization onto the capillary
surface has significantly improved the separation efficiency of
chiral GC columns. Although poly(siloxane) and PEG-based GC
columns are useful in many cases, the stationary phases that can
demonstrate high thermal stability as well as unique chemical
selectivities are highly desirable.
Owing to their high thermal stability and multiple solvation
properties, ILs have been exploited as potential alternatives to
conventional GC stationary phases (poly(siloxanes) and PEGs).
Important physicochemical properties of ILs, including negligible
vapor pressure at room temperatures, tunable viscosity, and
 References 295

high thermal stability, are highly useful in extending the column

operating temperature range and for producing low-bleed
stationary phases. Functionalized IL-based stationary phases have
demonstrated superior capabilities for the analysis of complex
polar samples compared to nonionic stationary phases. ILs/PILs
have the potential to serve as orthogonal stationary phases in
multidimensional/comprehensive two-dimensional GC techniques.
The unique chemical selectivities of ILs (e.g., hydrogen bond
acidity and dispersive type interactions) can complement nonionic
stationary phases and can provide high separation space for
the analysis of complex samples. Most of the commercial GC
stationary phases are relatively thermally stable and are nonbonded
stationary phases. Therefore, an investigation on IL bonding
chemistry and methods that can immobilize them on capillary
surface can potentially extend the thermal stability as well as column
lifetimes.

References

1. Wang, Z., Fingas, M. (1997). Developments in the analysis of

petroleum hydrocarbons in oils, petroleum products and oil-spill-
related environmental samples by gas chromatography, J. Chromatogr.
A, 774, pp. 51–78.
2. Lehotay, S. J., Hajšlová, J. (2002). Application of gas chromatography
in food analysis, Trends Anal. Chem.; TrAC, 21, pp. 686–697.
3. Begerow, J., Jermann, E., Keles, T., Koch, T., Dunemann, L. (1996).
Screening method for the determination of 28 volatile organic
compounds in indoor and outdoor air at environmental concentrations
using dual-column capillary gas chromatography with tandem
electron-capture-flame ionization detection, J. Chromatogr. A, 749,
pp. 181–191.
4. Verenitch, S. S., Lowe, C. J., Mazumder, A. (2006). Determination of
acidic drugs and caffeine in municipal wastewaters and receiving
waters by gas chromatography–ion trap tandem mass spectrometry,
J. Chromatogr. A, 1116, pp. 193–203.
5. Grob, R. L., and Barry, E. F. (2004) Modern Practice of Gas
Chromatography, 4th ed. (John Wiley & Sons).
6. Rotzsche, H. (1991). Stationary Phases in Gas Chromatography, 1st ed.
(Elsevier).
296 Stationary Phases for Gas Chromatography Method Development

7. Herraiz, T., Reglero, G., Herraiz, M., Alonso, R., Cabezudo, M. D. (1987).
Micropacked columns: A suitable alternative to very thick capillary
columns, J. Chromatogr. A, 388, pp. 325–333.
8. Ettre, L. (1987). MJE Golay and the invention of open‐tubular (capillary)
columns, J. Sep. Sci., 10, pp. 221–230.
9. Poole, C. F. (2003). The Essence of Chromatography, 1st ed. (Elsevier).
10. Desty, D. (1987). The history and potentiality of capillary columns
in gas chromatography, J. Chromatogr. Sci., 25, pp. 552–563.
11. Dandeneau, R. D., and Zerenner, E. H. (1979). An investigation of
glasses for capillary chromatography, J. High Resolut. Chromatogr.,
2, pp. 351–356.
12. Peters, T. L., Nestrick, T. J., Lamparski, L. L. (1982). Etching borosilicate
glass capillary columns, Anal. Chem., 54, pp. 2397–2398.
13. Grob, K., and Grob, G. (1977). The barium carbonate procedure for
the preparation of glass capillary columns; further informations
and developments, Chromatographia, 10, pp. 181–187.
14. Grob, K. (1980). Persilylation of glass capillary columns. Part 4:
Discussion of parameters, J. High Resolut. Chromatogr., 3, pp. 493–
496.
15. Grob, K., Grob, G., Blum, W., Walther, W. (1982). Preparation on
inert glass capillary columns for gas chromatography: A revised,
comprehensive description, J. Chromatogr. A, 244, pp. 197–208.
16. Hetem, M., Rutten, G., Vermeer, B., Rijks, J., van de Ven, L., de Haan,
J., Cramers, C. (1989). Deactivation with polymethylhydrosiloxane:
A comparative study with capillary gas chromatography and solid-
state 29Si nuclear magnetic resonance spectroscopy, J. Chromatogr.
A, 477, pp. 3–24.
17. Novotný, M., and Bartle, K. (1974). Preparation of thick-film glass
capillary columns by the dynamic coating procedure, J. Chromatogr. A,
93, pp. 405–411.
18. Bouche, J., and Verzele, M. (1968). A static coating procedure for
glass capillary columns, J. Chromatogr. Sci., 6, pp. 501–505.
19. Wang, D., Chong, S. L., Malik, A. (1997). Sol−gel column technology for
single-step deactivation, coating, and stationary-phase immobilization
in high-resolution capillary gas chromatography, Anal. Chem., 69,
pp. 4566–4576.
20. Poole, C. F., and Poole, S. K. (2012). Chromatography Today, 5th ed.
(Elsevier).
 References 297

21. Poole, C. F., and Lenca, N. (2014). Gas chromatography on wall-

coated open-tubular columns with ionic liquid stationary phases,
J. Chromatogr. A, 1357, pp. 87–109.
22. Grob, K., and Grob, G. (1978). Comprehensive, standardized quality
test for glass capillary columns, J. Chromatogr. A, 156, pp. 1–20.
23. Luong, J., Gras, R., Jennings, W. (2007). An advanced solventless
column test for capillary GC columns, J. Sep. Sci., 30, pp. 2480–2492.
24. Anderson, J. L., and Armstrong, D. W. (2003). High-stability ionic
liquids. a new class of stationary phases for gas chromatography,
Anal. Chem., 75, pp. 4851–4858.
25. Qiao, L., Lu, K., Qi, M., Fu, R. (2013). Separation performance of
guanidinium-based ionic liquids as stationary phases for gas
chromatography, J. Chromatogr. A, 1276, pp. 112–119.
26. Sun, X., Zhu, Y., Wang, P., Li, J., Wu, C., Xing, J. (2011). High temperature
and highly selective stationary phases of ionic liquid bonded
polysiloxanes for gas chromatography, J. Chromatogr. A, 1218,
pp. 833–841.
27. Abraham, M. H., Poole, C. F., Poole, S. K. (1999). Classification of
stationary phases and other materials by gas chromatography, J.
Chromatogr. A, 842, pp. 79–114.
28. Poole, C. F., and Poole, S. K. (1989). Characterization of solvent
properties of gas chromatographic liquid phases, Chem. Rev., 89,
pp. 377–395.
29. Rohrschneider, L. (1998). Characterisation of GC stationary phases
in multilinear retention model, Chromatographia, 48, pp. 728–738.
30. Poole, C. F., Kollie, T. O., Poole, S. K. (1992). Recent advances in solvation
models for stationary phase characterization and the prediction
of retention in gas chromatography, Chromatographia, 34,
pp. 281–302.
31. Abraham, M. H. (1993). Scales of solute hydrogen-bonding: Their
construction and application to physicochemical and biochemical
processes, Chem. Soc. Rev., 22, pp. 73–83.
32. Anderson, J. L., Ding, J., Welton, T., Armstrong, D. W. (2002).
Characterizing ionic liquids on the basis of multiple solvation
interactions, J. Am. Chem. Soc., 124, pp. 14247–14254.
33. Baltazar, Q. Q., Leininger, S. K., Anderson, J. L. (2008). Binary ionic
liquid mixtures as gas chromatography stationary phases for
improving the separation selectivity of alcohols and aromatic
compounds, J. Chromatogr. A, 1182, pp. 119–127.
298 Stationary Phases for Gas Chromatography Method Development

34. Yao, C., and Anderson, J. L. (2009). Retention characteristics of

organic compounds on molten salt and ionic liquid-based gas
chromatography stationary phases, J. Chromatogr. A, 1216,
pp. 1658–1712.
35. Atapattu, S. N., Eggers, K., Poole, C. F., Kiridena, W., Koziol, W. W.
(2009). Extension of the system constants database for open-tubular
columns: System maps at low and intermediate temperatures for
four new columns, J. Chromatogr. A, 1216, pp. 1640–1649.
36. Atapattu, S. N., and Poole, C. F. (2008). Selectivity equivalence of
two poly(methylphenylsiloxane) open-tubular columns prepared
with different deactivation techniques for gas chromatography,
J. Chromatogr. A, 1185, pp. 305–309.
37. Atapattu, S. N., and Poole, C. F. (2008). Solute descriptors for
characterizing retention properties of open-tubular columns
of different selectivity in gas chromatography at intermediate
temperatures, J. Chromatogr. A, 1195, pp. 136–145.
38. Poole, C. F., and Poole, S. K. (2008). Separation characteristics of
wall-coated open-tubular columns for gas chromatography, J.
Chromatogr. A, 1184, pp. 254–280.
39. Vernon, F., and Ogundipe, C. O. E. (1977). Hydrocarbon stationary
phases for gas-liquid chromatography, J. Chromatogr. A, 132,
pp. 181–185.
40. Soják, L., Krupčík, J., Janák, J. (1980). Comparison of capillary columns
coated with C87 hydrocarbon and squalane in the analysis of
n-pentadecene isomers, J. Chromatogr. A, 191, pp. 199–206.
41. Pomaville, R. M., and Poole, C. F. (1987). Thermally stable, highly
fluorinated stationary phases for gas chromatography, Anal. Chim.
Acta., 200, pp. 151–169.
42. O’Mahony, T. K. P., Cox, A. P., Roberts, D. J. (1993). Gas chromatographic
separation of perfluorocarbons, J. Chromatogr. A, 637, pp. 1–11.
43. Varughese, P., Gangoda, M., Gilpin, R. (1988). Applications of
fluorinated compounds as phases and additives in chromatography
and their uses in pharmaceutical analysis, J. Chromatogr. Sci., 26,
pp. 401–405.
44. Ballschmiter, K., Mennel, A., Buyten, J. (1993). Long chain alkyl-
polysiloxanes as non-polar stationary phases in capillary gas
chromatography, Fresenius. J. Anal. Chem., 346, pp. 396–402.
45. Haken, J. K. (1984). Developments in polysiloxane stationary phases
in gas chromatography, J. Chromatogr. A, 300, pp. 1–77.
 References 299

46. Mayer, B. X., Zöllner, P., Lorbeer, E., Rauter, W. (2002). A new
75% diphenyl, 25% dimethyl polysiloxane coated on fused silica
capillary columns for high temperature gas chromatography, J. Sep.
Sci., 25, pp. 60–66.
47. Bradshaw, J. S., Schirmer, M. M., Juvancz, Z., Markides, K. E., Lee, M. L.
(1991). New polar polysiloxane stationary phases containing
cyano, nitrophenyl and 8-quinolinyl units attached to diethylene
oxide side groups, J. Chromatogr. A, 540, pp. 279–291.
48. Martin, S. D., Poole, C. F., Abraham, M. H. (1998). Synthesis and
gas chromatographic evaluation of a high-temperature hydrogen-
bond acid stationary phase, J. Chromatogr. A, 805, pp. 217–235.
49. Blomberg, L. G. (1990). Current aspects of stationary phase
immobilization in open tubular column chromatography, J. Microcolumn
Sep., 2, pp. 62–68.
50. Bottoli, C. B., Chaudhry, Z. F., Fonseca, D. A., Collins, K. E., Collins, C. H.
(2002). Poly (alkylmethylsiloxanes) thermally immobilized on silica
as stationary phases for high-performance liquid chromatography,
J. Chromatogr. A, 948, pp. 121–128.
51. Markides, K., Blomberg, L., Buijten, J., Wännman, T. (1983).
Cyanosilicones as stationary phases in gas chromatography: II. Gum
and rubber phases on fused silica, J. Chromatogr. A, 267, pp. 29–38.
52. Wright, B. W., Peaden, A., Lee, M. L., Stark, T. J. (1982). Free radical
cross-linking in the preparation of non-extractable stationay
phases for capillary gas chromatography, J. Chromatogr. A, 248,
pp. 17–34.
53. Schomburg, G., Husmann, H., Ruthe, S., Herraiz, M. (1982). Crosslinking
of alkylpolysiloxane filsm on various types of glass surfaces including
fused silica using g-radiation of a60cobalt-source. Comparison to
crosslinking by thermal peroxid treatment, Chromatographia, 15,
pp. 599–610.
54. Hubball, J., DiMauro, P., Barry, E., Lyons, E., George, W. (1984).
Developments in crosslinking of stationary phases for capillary gas
chromatography by cobalt-60 gamma radiation, J. Chromatogr. Sci., 22,
pp. 185–191.
55. Zhou, X.-C., Yan, H., Chen, Y.-Y., Wu, C.-Y., Lu, X.-R. (1996). Chiral crown
ether-anchored polysiloxanes as capillary gas chromatography
stationary phases, J. Chromatogr. A, 753, pp. 269–277.
56. Schurig, V. (1984). Gas chromatographic separation of enantiomers
on optically active metal-complex-free stationary phases. New
analytical methods (24), Angew. Chem. Int. Ed., 23, pp. 747–765.
300 Stationary Phases for Gas Chromatography Method Development

57. Ruderisch, A., Pfeiffer, J., Schurig, V. (2003). Mixed chiral stationary
phase containing modified resorcinarene and β-cyclodextrin
selectors bonded to a polysiloxane for enantioselective gas
chromatography, J. Chromatogr. A, 994, pp. 127–135.
58. Zhang, H., Dai, R., Ling, Y., Wen, Y., Zhang, S., Fu, R., Gu, J. (1997).
Resorcarene derivative used as a new stationary phase for capillary
gas chromatography, J. Chromatogr. A, 787, pp. 161–169.
59. Berthod, A., Li, W., Armstrong, D. W. (1992). Multiple enantioselective
retention mechanisms on derivatized cyclodextrin gas chromato-
graphic chiral stationary phases, Anal. Chem., 64, pp. 873–879.
60. Schurig, V., Schmalzing, D., Schleimer, M. (1991). Enantiomer
separation on immobilized chirasil-metal and chirasil-dex by gas
chromatography and supercritical fluid chromatography, Angew.
Chem. Int. Ed., 30, pp. 987–989.
61. Schurig, V. (2002). Chiral separations using gas chromatography,
Trends Anal. Chem.; TrAC, 21, pp. 647–661.
62. Levkin, P. A., Levkina, A., Schurig, V. (2006). Combining the
enantioselectivities of l-valine diamide and permethylated β-
cyclodextrin in one gas chromatographic chiral stationary phase,
Anal. Chem., 78, pp. 5143–5148.
63. Ding, J., Welton, T., Armstrong, D. W. (2004). Chiral ionic liquids
as stationary phases in gas chromatography, Anal. Chem., 76,
pp. 6819–6822.
64. König, W. A., Lutz, S., Mischnick-Lübbecke, P., Brassat, B., Wenz, G.
(1988). Cyclodextrins as chiral stationary phases in capillary gas
chromatography I. Pentylated α-cyclodextrin, J. Chromatogr. A, 47,
pp. 193–197.
65. Armstrong, D. W., Li, W., Chang, C. D., Pitha, J. (1990). Polar-liquid,
derivatized cyclodextrin stationary phases for the capillary gas
chromatography separation of enantiomers, Anal. Chem., 62,
pp. 914–923.
66. Nowotny, H. P., Schmalzing, D., Wistuba, D., Schurig, V. (1989).
Extending the scope of enantiomer separation on diluted methylated
β-cyclodextrin derivatives by high-resolution gas chromatography,
J. High Resolut. Chromatogr., 12, pp. 383–393.
67. Blum, W., and Aichholz, R. (1990). Gas chromatographic enantiomer
separation on tert-butyldimethylsilylated β-cyclodextrin diluted in
PS-086. A simple method to prepare enantioselective glass capillary
columns, J. High Resolut. Chromatogr., 13, pp. 515–518.
 References 301

68. Fischer, P., Aichholz, R., Bölz, U., Juza, M., Krimmer, S. (1990).
Permethyl-β-cyclodextrin, chemically bonded to polysiloxane: A
chiral stationary phase with wider application range for enantiomer
separation by capillary gas chromatography, Angew. Chem. Int. Ed.,
29, pp. 427–429.
69. Sicoli, G., Pertici, F., Jiang, Z., Jicsinszky, L., Schurig, V. (2007). Gas-
chromatographic approach to probe the absence of molecular
inclusion in enantioseparations by carbohydrates. Investigation of
linear dextrins (“acyclodextrins”) as novel chiral stationary phases,
Chirality, 19, pp. 391–400.
70. Huang, K., Zhang, X., Armstrong, D. W. (2010). Ionic cyclodextrins
in ionic liquid matrices as chiral stationary phases for gas
chromatography, J. Chromatogr. A, 1217, pp. 5261–5273.
71. Conder, J. R., Fruitwala, N. A., Shingari, M. K. (1983). Thermal
decomposition of polyethylene glycol 20 m and essential oils in
gas—liquid chromatography and the effect of traces of oxygen, J.
Chromatogr. A, 269, pp. 171–178.
72. Sandra, P., and Van Roelenbosch, M. (1981). Mixed phases of
Superox 20M and OV-1 in fused silica and glass capillary gas
chromatography, Chromatographia, 14, pp. 345–350.
73. Lindsay Smith, J. R., and Waddington, D. J. (1969). Gas chromatographic
analysis of aliphatic amines: The use of ethylene glycols as stationary
phases, J. Chromatogr. A, 42, pp. 183–194.
74. Azzouz, A., Rascón, A. J., Ballesteros, E. (2016). Simultaneous
determination of parabens, alkylphenols, phenylphenols, bisphenol
A and triclosan in human urine, blood and breast milk by continuous
solid-phase extraction and gas chromatography–mass spectrometry,
J. Pharm. Biomed. Anal., 119, pp. 16–26.
75. He, D., Simoneit, B. R. T., Xu, Y., Jaffé, R. (2016). Occurrence of
unsaturated C25 highly branched isoprenoids (HBIs) in a freshwater
wetland, Org. Geochem., 93, pp. 59–67.
76. Ren, Z., Xiao, X., Chen, D., Bi, X., Huang, B., Liu, M., Hu, J., Peng,
P. A., Sheng, G., Fu, J. (2014). Halogenated organic pollutants in
particulate matters emitted during recycling of waste printed
circuit boards in a typical e-waste workshop of Southern China,
Chemosphere, 94, pp. 143–150.
77. Tripathi, J., Chatterjee, S., Gamre, S., Chattopadhyay, S., Variyar, P. S.,
Sharma, A. (2014). Analysis of free and bound aroma compounds
of pomegranate (Punica granatum L.), LWT-Food. Sci. Technol., 59,
pp. 461–466.
302 Stationary Phases for Gas Chromatography Method Development

78. Zhu, Z.-C., Chen, S.-J., Zheng, J., Tian, M., Feng, A.-H., Luo, X.-J., Mai,
B.-X. (2014). Occurrence of brominated flame retardants (BFRs),
organochlorine pesticides (OCPs), and polychlorinated biphenyls
(PCBs) in agricultural soils in a BFR-manufacturing region of North
China, Sci. Total Environ., 481, pp. 47–54.
79. An, J., Zhu, B., Wang, H., Li, Y., Lin, X., Yang, H. (2014). Characteristics
and source apportionment of VOCs measured in an industrial
area of Nanjing, Yangtze River Delta, China, Atmos. Environ., 97,
pp. 206–214.
80. Hou, X., Zheng, X., Zhang, C., Ma, X., Ling, Q., Zhao, L. (2014).
Ultrasound-assisted dispersive liquid–liquid microextraction based
on the solidification of a floating organic droplet followed by gas
chromatography for the determination of eight pyrethroid pesticides
in tea samples, J. Chromatogr. B, 969, pp. 123–127.
81. López, L. (2014). Study of the biodegradation levels of oils from the
Orinoco Oil Belt (Junin area) using different biodegradation scales,
Org. Geochem., 66, pp. 60–69.
82. López, L., Lo Mónaco, S., Volkman, J. K. (2015). Evidence for mixed
and biodegraded crude oils in the Socororo field, Eastern Venezuela
Basin, Org. Geochem., 82, pp. 12–21.
83. Maguire-Boyle, S. J., and Barron, A. R. (2014). Organic compounds in
produced waters from shale gas wells, Environ. Sci. Process. Impact.,
16, pp. 2237–2248.
84. Nagashima, H., Kondo, T., Nagoya, T., Ikeda, T., Kurimata, N., Unoke,
S., Seto, Y. (2015). Identification of chemical warfare agents from
vapor samples using a field-portable capillary gas chromatography/
membrane-interfaced electron ionization quadrupole mass
spectrometry instrument with Tri-Bed concentrator, J. Chromatogr. A,
1406, pp. 279–290.
85. Yeh, H.-Y., Chuang, C.-H., Chen, H.-C., Wan, C.-J., Chen, T.-L., Lin, L.-Y.
(2014). Bioactive components analysis of two various gingers
(Zingiber officinale Roscoe) and antioxidant effect of ginger extracts,
LWT-Food. Sci. Technol., 55, pp. 329–334.
86. Aghamohammadi, M., Shahdousti, P., Harooni, B. (2016).
Ultrasound-assisted emulsification microextraction followed
by gas chromatography–flame ionization detection for urinary
methylmalonic acid determination, Microchem. J., 124, pp. 188–194.
87. Harb, J., Alseekh, S., Tohge, T., Fernie, A. R. (2015). Profiling of primary
metabolites and flavonols in leaves of two table grape varieties
 References 303

collected from semiarid and temperate regions, Phytochemistry, 117,

pp. 444–455.
88. Luo, Y.-B., Chen, X.-J., Zhang, H.-F., Jiang, X.-Y., Li, X., Li, X.-Y., Zhu, F.-P.,
Pang, Y.-Q., Hou, H.-W. (2016). Simultaneous determination of polycyclic
aromatic hydrocarbons and tobacco-specific N-nitrosamines in
mainstream cigarette smoke using in-pipette-tip solid-phase extraction
and on-line gel permeation chromatography-gas chromatography–
tandem mass spectrometry, J. Chromatogr. A, 1460, pp. 16–23.
89. Zhang, F., Wang, H., Zhang, L., Zhang, J., Fan, R., Yu, C., Wang, W.,
Guo, Y. (2014). Suspected-target pesticide screening using gas
chromatography–quadrupole time-of-flight mass spectrometry with
high resolution deconvolution and retention index/mass spectrum
library, Talanta, 128, pp. 156–163.
90. Corral, S., Leitner, E., Siegmund, B., Flores, M. (2016). Determination
of sulfur and nitrogen compounds during the processing of dry
fermented sausages and their relation to amino acid generation,
Food Chem., 190, pp. 657–664.
91. Kiene, F. E., Pein, M., Thommes, M. (2015). Orientation to determine
quality attributes of flavoring excipients containing volatile molecules,
J. Pharm. Biomed. Anal., 110, pp. 20–26.
92. Li, J., Peng, Y., Liu, Y., Li, W., Jin, Y., Tang, Z., Duan, Y. (2014). Investigation
of potential breath biomarkers for the early diagnosis of breast
cancer using gas chromatography–mass spectrometry, Clin. Chim. Acta,
436, pp. 59–67.
93. Moreno, A. I., Arnáiz, N., Font, R., Carratalá, A. (2014). Chemical
characterization of emissions from a municipal solid waste treatment
plant, Waste Manag., 34, pp. 2393–2399.
94. Nowak, T., Graffius, G. C., Liu, Y., Wu, N., Bu, X., Gong, X., Welch, C. J.,
Regalado, E. L. (2016). GC-FID method for high-throughput analysis
of residual solvents in pharmaceutical drugs and intermediates,
Green Chem., 18, pp. 3732–3739.
95. Raich-Montiu, J., Ribas-Font, C., de Arespacochaga, N., Roig-Torres,
E., Broto-Puig, F., Crest, M., Bouchy, L., Cortina, J. L. (2014). Analytical
methodology for sampling and analysing eight siloxanes and
trimethylsilanol in biogas from different wastewater treatment plants
in Europe, Anal. Chim. Acta., 812, pp. 83–91.
96. Cantu-Jungles, T. M., Maria-Ferreira, D., da Silva, L. M., Baggio, C. H.,
Werner, M. F. D. P., Iacomini, M., Cipriani, T. R., Cordeiro, L. M. C. (2014).
Polysaccharides from prunes: Gastroprotective activity and structural
elucidation of bioactive pectins, Food Chem., 146, pp. 492–499.
304 Stationary Phases for Gas Chromatography Method Development

97. Müller, D. C., Degen, C., Scherer, G., Jahreis, G., Niessner, R., Scherer,
M. (2014). Metabolomics using GC–TOF–MS followed by subsequent
GC–FID and HILIC–MS/MS analysis revealed significantly altered
fatty acid and phospholipid species profiles in plasma of smokers,
J. Chromatogr. A, 966, pp. 117–126.
98. Watkins, B. A., Kim, J., Tamez, H., Wenger, J., Thadhani, R., Friedman, A. N.
Serum phospholipid fraction of polyunsaturated fatty acids is the
preferred indicator for nutrition and health status in hemodialysis
patients, J. Nutr. Biochem., 38, pp. 18–24.
99. Qi, D., Fei, T., Sha, Y., Wang, L., Li, G., Wu, D., Liu, B. (2014). A novel fully
automated on-line coupled liquid chromatography–gas chromato-
graphy technique used for the determination of organochlorine
pesticide residues in tobacco and tobacco products, J. Chromatogr. A,
1374, pp. 273–277.
100. Siddique, A., Saied, S., Mumtaz, M., Hussain, M. M., Khwaja, H. A. (2015).
Multipathways human health risk assessment of trihalomethane
exposure through drinking water, Ecotoxicol. Environ. Saf., 116,
pp. 129–136.
101. Yang, Y., Kong, W., Zhao, L., Xiao, Q., Liu, H., Zhao, X., Yang, M. (2015).
A multiresidue method for simultaneous determination of 44
organophosphorous pesticides in Pogostemon cablin and related
products using modified QuEChERS sample preparation procedure
and GC–FPD, J. Chromatogr. B, 974, pp. 118–125.
102. Hallett, J. P., and Welton, T. (2011). Room-temperature ionic
liquids: Solvents for synthesis and catalysis. 2, Chem. Rev., 111,
pp. 3508–3576.
103. Ho, T. D., Zhang, C., Hantao, L. W., Anderson, J. L. (2014). Ionic liquids
in analytical chemistry: Fundamentals, advances, and perspectives,
Anal. Chem., 86, pp. 262–285.
104. Coker, T. G., Ambrose, J., Janz, G. J. (1970). Fusion properties of
some ionic quaternary ammonium compounds, J. Am. Chem. Soc., 92,
pp. 5293–5297.
105. Gordon, J. E. (1965). Fused organic salts. IV.1a characterization of
low-melting quaternary ammonium salts. Phase equilibria for
salt-salt and salt-nonelectrolyte systems. Properties of the liquid
salt medium, J. Am. Chem. Soc., 87, pp. 4347–4358.
106. Poole, C. F. (2004). Chromatographic and spectroscopic methods
for the determination of solvent properties of room temperature
ionic liquids, J. Chromatogr. A, 1037, pp. 49–82.
 References 305

107. Bonhôte, P., Dias, A.-P., Papageorgiou, N., Kalyanasundaram, K., Grätzel,
M. (1996). Hydrophobic, highly conductive ambient-temperature
molten salts, Inorg. Chem., 35, pp. 1168–1178.
108. Shashkov, M. V., and Sidelnikov, V. N. (2012). Mass spectral
evaluation of column bleeding for imidazolium-based ionic liquids
as GC liquid phases, Anal. Bioanal. Chem. 403, pp. 2673–2682.
109. Huddleston, J. G., Visser, A. E., Reichert, W. M., Willauer, H. D., Broker,
G. A., Rogers, R. D. (2001). Characterization and comparison of
hydrophilic and hydrophobic room temperature ionic liquids
incorporating the imidazolium cation, Green Chem., 3, pp. 156–164.
110. Kroon, M. C., Buijs, W., Peters, C. J., Witkamp, G.-J. (2007).
Quantum chemical aided prediction of the thermal decomposition
mechanisms and temperatures of ionic liquids, Thermochim.
Acta, 465, pp. 40–47.
111. Lebga-Nebane, J. L., Rock, S. E., Franclemont, J., Roy, D., Krishnan, S.
(2012). Thermophysical properties and proton transport mechanisms
of trialkylammonium and 1-alkyl-1H-imidazol-3-ium protic ionic
liquids, Ind. Eng. Chem. Res., 51, pp. 14084–14098.
112. Poole, C. F., and Poole, S. K. (2011). Ionic liquid stationary phases for
gas chromatography, J. Sep. Sci., 34, pp. 888–900.
113. Payagala, T., Zhang, Y., Wanigasekara, E., Huang, K., Breitbach, Z.
S., Sharma, P. S., Sidisky, L. M., Armstrong, D. W. (2009). Trigonal
tricationic ionic liquids: A generation of gas chromatographic
stationary phases, Anal. Chem., 81, pp. 160–173.
114. Law, G., and Watson, P. R. (2001). Surface tension measurements
of N-alkylimidazolium ionic liquids, Langmuir, 17, pp. 6138–6141.
115. Kilaru, P., Baker, G. A., Scovazzo, P. (2007). Density and surface
tension measurements of imidazolium-, quaternary phosphonium-,
and ammonium-based room-temperature ionic liquids: Data and
correlations, J. Chem. Eng. Data, 52, pp. 2306–2314.
116. Lantz, A. W., Pino, V., Anderson, J. L., Armstrong, D. W. (2006).
Determination of solute partition behavior with room-temperature
ionic liquid based micellar gas–liquid chromatography stationary
phases using the pseudophase model, J. Chromatogr. A, 1115,
pp. 217–224.
117. Armstrong, D. W., He, L., Liu, Y.-S. (1999). Examination of ionic liquids
and their interaction with molecules, when used as stationary
phases in gas chromatography, Anal. Chem., 71, pp. 3873–3876.
306 Stationary Phases for Gas Chromatography Method Development

118. Zhao, Q., Eichhorn, J., Pitner, W. R., Anderson, J. L. (2009). Using the
solvation parameter model to characterize functionalized ionic
liquids containing the tris (pentafluoroethyl) trifluorophosphate
(FAP) anion, Anal. Bioanal. Chem., 395, pp. 225–234.
119. Breitbach, Z. S., and Armstrong, D. W. (2008). Characterization
of phosphonium ionic liquids through a linear solvation energy
relationship and their use as GLC stationary phases, Anal. Bioanal.
Chem., 390, pp. 1605–1617.
120. Delmonte, P., Fardin-Kia, A. R., Kramer, J. K. G., Mossoba, M. M., Sidisky,
L., Tyburczy, C., Rader, J. I. (2012). Evaluation of highly polar ionic
liquid gas chromatographic column for the determination of the
fatty acids in milk fat, J. Chromatogr. A, 1233, pp. 137–146.
121. Zeng, A. X., Chin, S.-T., Nolvachai, Y., Kulsing, C., Sidisky, L. M., Marriott,
P. J. (2013). Characterisation of capillary ionic liquid columns for gas
chromatography–mass spectrometry analysis of fatty acid methyl
esters, Anal. Chim. Acta, 803, pp. 166–173.
122. Do, L., Liljelind, P., Zhang, J., Haglund, P. (2013). Comprehensive
profiling of 136 tetra- to octa-polychlorinated dibenzo-p-dioxins and
dibenzofurans using ionic liquid columns and column combinations,
J. Chromatogr. A, 1311, pp. 157–169.
123. Anderson, J. L., and Armstrong, D. W. (2005). Immobilized ionic
liquids as high-selectivity/high-temperature/high-stability gas
chromatography stationary phases, Anal. Chem., 77, pp. 6453–6462.
124. Hsieh, Y.-N., Ho, W.-Y., Horng, R. S., Huang, P.-C., Hsu, C.-Y., Huang, H.-H.,
Kuei, C.-H. (2007). Study of anion effects on separation phenomenon
for the vinyloctylimidazolium based ionic liquid polymer stationary
phases in GC, Chromatographia, 66, pp. 607–611.
125. Zhang, C., Park, R. A., Anderson, J. L. (2016). Crosslinked structurally-
tuned polymeric ionic liquids as stationary phases for the analysis
of hydrocarbons in kerosene and diesel fuels by comprehensive
two-dimensional gas chromatography, J. Chromatogr. A, 1440,
pp. 160–171.
126. Gonzalez-Alvarez, J., Arias-Abrodo, P., Puerto, M., Viguri, M. E.,
Perez, J., Gutierrez-Alvarez, M. D. (2013). Polymerized phosphonium-
based ionic liquids as gas chromatography stationary phases, RSC
Adv., 3, pp. 21377–21380.
127. Shashkov, M. V., and Sidel’nikov, V. N. (2012). Single cation ionic liquids
as high polarity thermostable stationary liquid phases for capillary
chromatography, Russ. J. Phys. Chem. A, 86, pp. 138–141.
 Preparation and Beginning of Performance 307

128. Shashkov, M. V., and Sidelnikov, V. N. (2013). Properties of columns

with several pyridinium and imidazolium ionic liquid stationary
phases, J. Chromatogr. A, 1309, pp. 56–63.
129. Huang, K., Han, X., Zhang, X., Armstrong, D. W. (2007). PEG-linked
geminal dicationic ionic liquids as selective, high-stability gas
chromatographic stationary phases, Anal. Bioanal. Chem., 389,
pp. 2265–2275.
130. Álvarez, J. G., Gomis, D. B., Abrodo, P. A., Llorente, D. D., Busto, E.,
Lombardía, N. R., Fernández, V. G., Álvarez, M. D. G. (2011). Evaluation
of new ionic liquids as high stability selective stationary phases in
gas chromatography, Anal. Bioanal. Chem., 400, pp. 1209–1216.
131. Hantao, L. W., Najafi, A., Zhang, C., Augusto, F., Anderson, J. L. (2014).
Tuning the selectivity of ionic liquid stationary phases for enhanced
separation of nonpolar analytes in kerosene using multidimensional
gas chromatography, Anal. Chem., 86, pp. 3717–3721.
132. Zhang, C., Ingram, I. C., Hantao, L. W., Anderson, J. L. (2015). Identifying
important structural features of ionic liquid stationary phases for
the selective separation of nonpolar analytes by comprehensive
two-dimensional gas chromatography, J. Chromatogr. A, 1386,
pp. 89–97.
133. Ragonese, C., Tranchida, P. Q., Dugo, P., Dugo, G., Sidisky, L. M., Robillard,
M. V., Mondello, L. (2009). Evaluation of use of a dicationic liquid
stationary phase in the fast and conventional gas chromatographic
analysis of health-hazardous C18 cis/trans fatty acids, Anal. Chem., 81,
pp. 5561–5568.
134. Sanchez-Prado, L., Lamas, J. P., Garcia-Jares, C., Llompart, M. (2012).
Expanding the applications of the ionic liquids as GC stationary
phases: plasticizers and synthetic musks fragrances, Chromatographia,
75, pp. 1039–1047.
135. Domínguez, C., Reyes-Contreras, C., Bayona, J. M. (2012). Determination
of benzothiazoles and benzotriazoles by using ionic liquid stationary
phases in gas chromatography mass spectrometry. Application to their
characterization in wastewaters, J. Chromatogr. A, 1230, pp. 117–122.
136. García Pinto, C., Pérez Antón, A., Pérez Pavón, J. L., Moreno Cordero,
B. (2012). Coupling of microextraction by packed sorbents with
gas chromatography with ionic liquid stationary phases for the
determination of haloanisoles in wines, J. Chromatogr. A, 1260,
pp. 200–205.
308 Stationary Phases for Gas Chromatography Method Development

137. Reyes-Contreras, C., Domínguez, C., Bayona, J. M. (2012). Determination

of nitrosamines and caffeine metabolites in wastewaters using gas
chromatography mass spectrometry and ionic liquid stationary
phases, J. Chromatogr. A, 1261, pp. 164–170.
138. Frink, L. A., and Armstrong, D. W. (2016). Determination of trace
water content in petroleum and petroleum products, Anal. Chem.,
88, pp. 8194–8201.
Chapter 12

Gas Chromatography Method

Development

Tien Ho and John C. Vinci

Abbvie, 1 N. Waukegan Rd,
North Chicago, Illinois 60064, USA
[email protected]

12.1 Gas Chromatography in the

Pharmaceutical Setting
Gas chromatography (GC) is a separation technique and analytical
tool that serves to provide quantitative and qualitative analysis.
The separation occurs by partitioning analytes through two
phases, a stationary phase and an inert gas moving phase. The
stationary phase may consist of solid or a liquid coating on a solid
substrate, and the column may be open tubular or packed with
particles. Separation by GC is achieved by volatility differences of
the analytes and by differing degrees of analyte-stationary phase
interaction. This analyte–stationary phase interaction is important
in distinguishing GC as a high-resolution separation technique.

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
310 Gas Chromatography Method Development

While classical distillation separates analytes based on vapor

pressure differences, GC offers the additional advantage of separation
through chemical interactions. By combing a high efficiency
separation with the potential for a high degree of selectivity, the
result can be a high-resolution separation that has the power
to operate as a high throughput technique.
Although the separation occurs in the gas phase, the analytes
amenable to GC may be solid, liquid or gas at room temperature
and pressure. The analytes of interest must be stable when
subjected to the GC inlet and column temperatures and must
be able to volatilize such that analysis in the gas phase can
occur. While these requirements create some limitations to the
number of analytes that may be analyzed by GC, this limitation is
overcome by the high separation efficiencies and ease of
operability for the end-user, along with innovative approaches to
deal with difficult to volatilize compounds (vide infra). When GC
is applicable to a given set of analytes, it is a very powerful
analytical tool worth having in the analyst’s toolbox.
The modern GC system in the pharmaceutical analysis lab is fully
automated with user-friendly software. A typical pharmaceutical
analysis lab will have GCs equipped with an autosampler and a
temperature programmable inlet, oven, and detector as the most
basic components of the system. The autosampler will allow for
either direct injection or headspace (HS) injection of the sample.
For direct injection, the sample is typically introduced onto
the GC column using a hot inlet, although cool on-column injection
is also possible. With the typical hot inlet sample introduction,
the sample is quickly vaporized allowing for fast transfer of sample
and for the separation to occur immediately.

12.2 Instrument and Column Optimization

The optimization of gas chromatographic separations can be a
cumbersome and complex task for the pharmaceutical analyst.
This is primarily due to the multitude of instrumental as well
as physical parameters that requires full attention to detail. The
interactions observed between each parameter also drastically
 Instrument and Column Optimization 311

complicate this process. Nevertheless, analysts can gain abundant

benefits when optimizing their chromatographic methods
with a particular goal(s) in mind. Perhaps, the method may be
optimized to achieve faster separation, resulting in higher analysis
throughput, and/or to gain sufficient resolution between critical
pairs. Ultimately, the optimization of chromatographic conditions
enables the analyst to gain better analytical performance in their
analyses (i.e., improving sensitivity, accuracy, and precision) and
utilize their laboratory equipment more effectively (i.e., reducing
instrument maintenance/down time and usage of consumables).
This section will provide the practicing pharmaceutical analyst
with a break-down of various physical and instrumental
parameters that can be optimized. Additionally, we will explore
the cause-and-effect(s) of varying each parameter and how they
can ultimately influence the outcome of an analysis.

12.2.1 Split/Splitless Inlet

The process of introducing samples into a GC can be highly
complicated, considering the variety of instrumental components
involved. Varying between sample injection modes, hardware, and
instrumental parameters can result in dramatic effects on both
the chromatography and analytical performance of a method.
In most pharmaceutical settings, the common sample introduction
method for GC is by liquid injection of the sample into a
split/splitless hot inlet. Other methods of liquid injection, such as
on-column and programmed-temperature vaporization (PTV)
are not as common and optimization for these modes will not be
discussed herein.
A schematic of a split/splitless inlet is shown in Fig. 12.1.
In split/splitless injections, liquid samples are withdrawn from
their source and injected (manually or automatically) into the
GC inlet via microsyringe. Once the liquid solution enters the hot
inlet, it readily vaporizes, mixes with the carrier gas, and is
transferred to a column. It should be noted that during an
injection, the split/splitless inlet can only be operated in either
split or splitless mode, not both.
312 Gas Chromatography Method Development

Figure 12.1 Schematic of a GC split/splitless inlet.

12.2.1.1 Split injection

Using split flow enables the analyst to address a major issue
associated with conventional wall-coated open tubular (WCOT)
capillary columns, namely, sample overloading. Due to their small
inner diameters, capillary columns may not possess the capacity
to withstand microliter(s) injections of a sample solution, leading
to overload. With sample overload, the chromatographer may
experience poor separation and/or asymmetrical peak shapes
(i.e., peak fronting). Injections of a concentrated sample without
splitting the gas flow may also result in analyte carry-over and/or
contaminate various components in the GC inlet.
Applying a split injection is synonymous to performing a
rapid dilution in the GC inlet. In a split injection, a fraction of the
injected sample is carried to the column, while the remaining
fraction is diverted to waste through a split vent. The ratio of the
volumetric flow rate of the gas fraction carried into the column
relative to that of the eluent exiting the split vent is deemed the
split ratio. The split ratio is controlled by varying the flow rate
of the carrier gas (calculated by the head pressure of the inlet)
and the split vent is operated via a needle valve. Increasing the
carrier gas flow rate, while opening the split vent, will result in a
higher split ratio and enable a narrower analyte band to enter
the column. As a result, optimizing the split ratio can also provide
 Instrument and Column Optimization 313

better peak shapes and improve signal sensitivity. It is necessary

to note that the split ratio is not entirely proportional to the split
flow (e.g., actual flow to column vs. split vent) due to a number
of factors, such as inlet temperature, solute physico-chemical
properties, injection amounts, etc. As such, the split ratio should
be understood as an approximation, rather than a quantitative
value [1].
Typical split ratios can range from 1:10 to 1:500. For optimization,
it is recommended that the analyst start with a split ratio of
1:5 to 1:20 for columns with larger inner diameters (e.g., 0.53 mm,
megabore) and higher stationary phase film thicknesses (e.g.,
>0.5 µm), as these columns possess better sample loading
capacities. For columns with smaller inner diameters (i.e., 0.25 mm,
0.18 mm, etc.) a higher split ratio, such as 1:50 and above,
is recommended as the starting point to mitigate overload.

12.2.1.2 Splitless injection

A common strategy to obtain good signal sensitivity for trace-
level analysis is to inject larger quantities of the sample into the
capillary column. One way that this can be achieved on a split/
splitless injector is by taking advantage of the splitless injection
mode. Contrary to a conventional split injection, a splitless
injection transfers most (~95%) of the sample to the capillary
column. In this injection mode, the split vent will remain closed
for the duration of an injection, as well as an additional period
immediately after the injection. When the split vent valve is
closed, the sample vapor will be directly transferred to the column
with minimal loss. Subsequently, the split vent is opened to flush
residual sample vapor from the inlet.
Unlike split injections, the transfer of vapor in a splitless
injection is significantly slower, resulting in band broadening of
the analyte peak(s). This is especially true for highly volatile
analytes with large thermal expansion coefficients. To address
this potential issue, the analyst can optimize the oven and inlet
temperatures, as well as the splitless injection time. In regards
to temperature, it is recommended that the analyst set the
inlet temperature to an appropriate temperature that enables
vaporization of the sample. The initial oven temperature should
be set at least 20°C below the solvent boiling point. By setting the
temperature as recommended, the solvent/analyte band can be
314 Gas Chromatography Method Development

condensed and concentrated at the head of the column, resulting

in a narrow band [2, 3]. The splitless injection time refers to the
total time during an injection where the purge valve is closed.
Setting the splitless injection time too short can reduce the signal
sensitivity of analytes (especially those with high-boiling points).
On the other hand, over-extending the splitless injection time
can result in peak degradation and a broad solvent peak tail [2].
Common splitless injection times are set from 20s to 90s [3].
In addition to optimizing inlet temperature and splitless
injection time, it is important to understand the thermal stability
of samples capable of undergoing thermal transformations/
degradations. Setting the inlet temperatures at or above the stable
temperatures of the sample may result in poor accuracy and
precision, in addition to inaccurate representations of its purity
profile. Over-extending the splitless injection time (at or above
stable sample temperatures) can result in similar issues.

12.2.2 Column Dimensions

When selecting a GC column, it is clear that there are multiple
types of stationary phases commercially available for the analyst.
It is even clearer that there are varieties of column dimensions
available for each stationary phase. Choosing the appropriate
column dimensions can be crucial for method development and
transfer. Herein, we provide a brief insight on the effects of varying
column dimensions and considerations for hand-picking the
appropriate column for the job. Since WCOT columns are currently
used for most GC analyses in the pharmaceutical industry, this
section will focus on WCOT columns.
It is worthy to note that selecting the appropriate stationary
phase is also vital for developing a fit-for-purpose separation. It
is necessary to fully understand the selectivity of the wide variety
of commercial columns currently available and how they can be
properly exploited during method development. As such, the
reader is referred to the in-depth discussion on column selection
written by Anderson and coworkers (Chapter 11).

12.2.2.1 Column inner diameter

A factor for selecting a column I.D. is the sample capacity needs.
As the sample capacity is proportional to the square of column
 Instrument and Column Optimization 315

radius, a larger column diameter exhibits higher peak capacity.

For samples with analytes possessing vastly different relative
concentrations, it may be beneficial to use larger diameter columns
to mitigate column overload. In regards to pressure, a higher
pressure drop is needed when a smaller column I.D., leading to
an increase in the inlet head pressure. Another factor that may
influence the selection of a suitable column I.D. is the separation
efficiency. Based on Eq. 12.1, as the column I.D. (dc) decreases,
the minimum plate height (hmin) will also be reduced. In other
words, decreasing the column I.D. can produce better separation
efficiency [4]. Therefore, if the relative concentration of all
analytes in the sample is similar, the analyst may choose to
use a smaller I.D. column to obtain better overall separation.
Equation 12.1 shows the relationship between column I.D. and
plate height:

1 + 6k +11k 2
hmin = dc , (12.1)
12(1 + k )2

where k is the retention factor, dc is the column diameter, and

hmin is the minimum plate height.
In most cases, an analyst can choose a 0.25 or 0.32 mm
I.D. column as a starting point for conventional split/splitless
injections. However, in some occasions, it may be necessary to
revert to 0.53 mm I.D. columns for volatile organic compounds
(VOCs) analyses (i.e., residual solvents), as these columns can
provide better robustness for method transfer and higher capacity
to withstand large variations in analyte concentrations in a
sample. For systems with vacuum outlets, such as those found in
GC-mass spectrometry, columns with I.D. at or below 0.25 mm
is recommended to maintain carrier gas flow rates in the vacuum.

12.2.2.2 Column length

Varying the column length can affect the run time, pressure drop,
peak capacity, and resolution. In terms of pressure, increasing
the length is similar to reducing the I.D. of a column, wherein a
higher pressure drop is needed when a longer column is used,
resulting in higher inlet head pressures. Increasing the length also
increases the peak capacity in a separation. More importantly,
316 Gas Chromatography Method Development

increasing the column length also increases peak-to-peak

resolution as a result of a higher number of theoretical plates
present. However, increasing the column length proportionally
affects the separation time (Eq. 12.2) [4]. Thus, it is important
that the analyst choose column lengths that are “fit-for-purpose.”
In other words, if peaks are sufficiently resolved and their
sensitivities (based on peak heights) are acceptable, shorter
columns are recommended to reduce the overall analysis time.
Equation 12.2 shows the relationship between analyte retention
time and column length:
L
t R = (k +1), (12.2)
u L
t R = (k +1),
where tR is the retention time, L is the column length, and u is
the average linear velocity.
On the other hand, if a complex sample contains a myriad of
components that require good separation resolution, it may be
more beneficial to resort to longer columns. Another factor to
consider is the cost of columns. Typically, the longer the column,
the more expensive it is for purchase. An analyst can choose to
start with column lengths at around 30 m as a starting point for
general separations. For screening samples and high-throughput
analyses, a shorter column may be used. For highly complex
samples, the column length may be increased to 60 m.

12.2.2.3 Film thickness

The thickness of the stationary phase affects the efficiency,
capacity, separation time, and in some cases, the background of a
separation.
The film thickness is inversely proportional to the number of
theoretical plates in a separation. As a result, when reducing the
film thickness, the efficiency is enhanced. Additionally, separation
time will be reduced, due to analytes retaining in the thicker
film for shorter periods of time. Column bleed (background
contribution) is also less pronounced in columns with smaller film
thicknesses. Also, thinner films are more applicable for separating
high-boiling analytes (i.e., petroleum samples, large-chain fatty
acids, polycyclic aromatic hydrocarbons, etc.) as their elution
temperature will be lower (faster elution) compared to larger
films. Thinner films (<1.0 um) should be used for GC/MS
 Instrument and Column Optimization 317

applications to minimize column bleed and contamination of the

ionization source.
On the other hand, since additional partitioning sites are
available during the separation when using a thicker film,
increasing the film thickness will result in an increase in the
peak capacity. This inherently means that column overloading
(observed as peak fronting) is less likely compared to thinner
films. Moreover, this also means that thicker films enable the
separation of samples containing large variations in analyte
concentrations. Active sites are less frequent in columns with
thicker films, which reduce the likelihood of polar compounds
adsorbing to the column wall. Analytes with relatively high
volatility are recommended to be separated using columns with
thicker films (>1.0 um), as they can more readily interact with the
stationary phase.
Taken into consideration the contributions from varying
the column dimensions, the analyst can refer to Table 12.1 as
guidance on the effects of altering each parameter.
As with all chromatography techniques, choosing an
appropriate mobile phase can be crucial for a successful separation.
Fortunately, the carrier gas choice for GC boils down to three
common inert gasses, namely, H2, He, and N2. All three gasses
possess various advantages and limitations, some of which can
be exemplified by plotting their respective van Deemter curves
(Fig. 12.2) [5].
As shown, the best separation efficiency (lowest height
equivalent to a theoretical plate, H) can be obtained by using N2
as a carrier gas. However, the speed of the separation (i.e., linear
velocity) must be relatively slower than He or H2 to achieve this
efficiency. Further, the efficiency deteriorates more rapidly as the
linear velocity increases for the N2 carrier gas compared to the
other carrier gases. The steeper slope of the curve (following
the curve minima) means that a small change in flow rate can
significantly impact the separation efficiency using N2 [6]. On the
other hand, the optimal separation efficiency using He and H2 is
slightly poorer than N2. However, the linear velocity (at optimal
efficiency) is faster using these carrier gases, meaning quicker
separation times. Overall, separations using lighter carrier gases
can be significantly faster, especially for H2, and can increase the
throughput of the analysis.
318 Gas Chromatography Method Development

Table 12.1 Benefits of modifying column I.D., length, and film thickness

Parameter Benefits of modification

Higher Lower
Column I.D. • Suitable for samples with large • Efficiency is
variations in concentration ranges enhanced
• Higher peak capacity • Suitable for GC/MS
• Lower head-pressure • Reduced column
• Reduced likelihood of column bleed
overloading
Column • Higher peak capacity • Lower head-
length • Resolution is enhanced pressure
• Shorter run time
Film • Higher peak capacity • Reduced column
thickness • Reduced likelihood of column bleed
overloading • Shorter run time
• Suitable for samples with large • Suitable for
variations in concentration ranges separating high-
• Less active sites – reduced boiling compounds
adsorption of polar compounds to • Suitable for GC/MS
wall
• Suitable for separating low-boiling
compounds

Figure 12.2 Van Deemter curves for N2, He, and H2. Column I.D. = 0.25 mm,
analyte retention factor (k) = 10. Reproduced from [5].
 Instrument and Column Optimization 319

The viscosity (i.e., resistance to flow) of the gases also influences

their respective linear velocities at a given temperature. In GC,
the viscosity of gases increases when the temperature is increased,
as shown in Fig. 12.3 [5]. As depicted, H2 possess the lowest
viscosity, compared to N2 and He, at every temperature studied.
As a result, this carrier gas can produce higher linear velocities.
The viscosity of the carrier gas also possesses a linear relationship
with the pressure drop (difference in carrier-gas pressure from
the column inlet to the outlet) of the system. A gas with a higher
viscosity will require higher pressure drops to sustain a specific
average linear velocity. Due to its low viscosity, H2 would be more
ideal for narrow-bore columns (<200 µm I.D.) where higher
pressure drops are required to maintain a suitable linear velocity.

Figure 12.3 Relationship between the carrier gas viscosity and temperature.
Reproduced from [5].

Generally, it is recommended that the analyst use either He

or H2 as a starting point for method development. N2 is least
preferred since its linear velocity operating range is highly
limited. However, if the cost of analysis is a major concern, then
using N2 may be beneficial as this gas can be produced by
320 Gas Chromatography Method Development

generators. Also, the efficiency loss observed when using N2 can be

compensated by modifying column dimensions [7] (as discussed
in Section 2.2). He is an ideal carrier gas choice for conventional
GC capillary columns and can provide exceptional separation
efficiency and peak resolution. H2 is more suited for high-speed
separations employing narrow-bore columns. It is important to
note that the analyst recognize all safety considerations in cases
where H2 is used as a carrier gas.

12.2.3 Linear Velocity

The linear velocity of a carrier gas can dramatically affect two
primary parameters in a separation, namely, efficiency and
retention time. Selecting a higher linear velocity will reduce the
interaction between the stationary phase and the analyte, reducing
the retention time. This can be advantageous when analysis
throughput is a critical factor during optimization. Conversely,
lowering the linear velocity will promote the stationary phase-
to-analyte interaction, resulting better separation efficiency to
a certain extent (optimum level). The trick is to select a linear
velocity that can strike a balance between throughput and
efficiency.
At a given column length, ID, and film thickness, the separation
efficiency can be simply understood as an inverse to the plate
height. As shown in Fig. 12.4, the lowest H values can be obtained
when the linear velocity is between 20–45 cm/s using He as the
carrier gas. As such, it is recommended that the analyst begin
their method development within this linear velocity window.
Further modifications to the linear velocity can be sought after
for improving the analysis throughput.
Table 12.2 displays the recommendations for selecting a flow
rate using He or H2 as a carrier gas, based on their corresponding
linear velocities [8]. At a constant length of 30 m, columns with
smaller IDs (i.e., narrow-bore) require lower He flow rates to
achieve a linear velocity of 20–45 cm/s. This dramatically differs
when increasing the ID to mega-bore type columns. It is worthy
to note that increasing the linear velocity correspondingly require
increasing the inlet pressure. It is recommended that the analyst
 Optimization of Separation Temperature 321

review the inlet pressure capabilities of their instrumentation

before modifying the linear velocity to extreme levels. In cases
where high velocities are desired (inlet pressures exceedingly
greater than 80 psi), the analyst can alternatively select hydrogen
as carrier gas.

Table 12.2 Recommendations for selecting initial flow rates, based on

the optimal linear velocity of hydrogen and helium

Helium Hydrogen
Column I.D. Linear velocity Flow rate Linear velocity Flow rate
(mm) (cm/s) (mL/min) (cm/s) (mL/min)
0.18 20–45 0.3–0.7 40–60 0.6–0.9
0.25 20–45 0.7–1.3 40–60 1.2–2.0
0.32 20–45 1.2–2.2 40–60 2.2–3.0
0.53 20–45 4.0–8.0 40–60 6.0–9.0

Source: Adapted from [8].

12.3 Optimization of Separation Temperature

One of the hallmarks of separation by GC is the ability to
manipulate analyte retention using temperature. Fundamentally
speaking, increasing the temperature during a separation will
reduce analyte partitioning to the stationary phase (i.e., decrease
retention), leading to shorter retention times. On the other hand,
reducing the temperature will result in the opposite effect (i.e.,
increase retention, longer retention times). Importantly, altering
the temperature will also alter analyte selectivity during a
separation. This means that peak elution order can potentially be
affected once temperature is manipulated during separation [4].
This section provides the analyst with practical suggestions on
developing initial temperature program/conditions to achieve
a “fit-for-purpose” separation. For fundamental and theoretical
discussions, the reader is advised explore the outstanding library
of literature available [4, 9–14], as well as the fundamental
theories established by John C. Giddings [15].
322 Gas Chromatography Method Development

12.3.1 Programmed Temperature Separations

Most GC methods today employ temperature programming over
isothermal separations for a number of reasons. Temperature
programming enables the separation of a mixture of analytes,
with broad boiling point ranges, in the shortest run times possible.
Sharper peaks can be obtained with the appropriate program as
a result of thermal focusing, especially for later-eluting analytes.
Also, applying a temperature gradient (and subsequent hold)
near the completion of a separation ensures that high-boiling
contaminants present in a sample are fully eluted and not carried-
over to subsequent injections. As separation depends not only
on temperature but also on a number of other parameters (i.e.,
stationary phase, column dimensions, carrier gas, etc.), the
optimization of separation temperature program may become more
or less an empirical task. However, with modern technology and
third-party software available via various GC consumable suppliers,
such as EZ-GC by Restek, choosing a separation temperature
can be less challenging.
A general and practical temperature scouting gradient was
recently described by Tony Taylor. The scouting gradient is shown
in Eq. 12.3 (adapted from [16]):

40°C hold for 1 min;

Ramp at 10°C/min to 300°C, hold for 10 min (12.3)

It is worthy to mention that the final temperature (300°C) should

be changed according to the maximum operating temperature
of the column.
From the scouting gradient, the analyst can tweak the program
according to their needs. The ramp rate should be altered within
a window of 4–25°C/min, due to the limitations of some GCs.
Higher ramp rates >25°C/min may sacrifice the retention time
reproducibility. If analytes are closely eluting, the ramp rate
should be reduced to better-resolve these peaks. On the other hand,
if the selectivity of analyte pairs is high, the ramp rate can be
increased to reduce separation time and peak broadening
(longitudinal band broadening). Note that altering the temperature
 Optimization of Separation Temperature 323

ramp rate can cause the elution order of analytes to change, as

shown in the separation of a mixture of chlorinated pesticides
(Fig. 12.4) [17]. As shown, drastic changes were observed
when increasing the temperature ramp from 1.5 to 20°C/min.
The retention times of all analytes were reduced from ~60 min
to ~12 min. The elution order of peak pairs J&K, M&N, P&Q
were reversed. Additionally, some peak pairs were better resolved
(i.e., J&K), others loss resolution (i.e., D&E).

Figure 12.4 Separation of a number of chlorinated pesticides using a

25 m × 0.53 mm methyl-35% phenylsilicone capillary column. The
oven temperature ramp rates were selected as (a) 1.5, (b) 4.0, (c) 8.0,
(d) 16.0°C/min. (A) α-BHC, (B) γ-BHC, (C) β-BHC, (D) heptachlor,
(E) δ-BHC, (F) aldrin, (G) heptachlor epoxide, (H) endosulfan I, (J) dieldrin,
(K) p,p-DDE, (L) eldrin, (M) endosulfan II, (N) p,p-DDD, (P) endrin
aldehyde, (Q) p,p-DDT, (R) endosulfan sulfate, (S) decomposition
product. Reproduced from [17].
324 Gas Chromatography Method Development

In addition to changing the temperature ramp, it is recommended

to add a mid-ramp isothermal hold for compounds closely eluting
with each other. The isothermal hold should be around a few min,
depending on empirical observations. Once all peaks have eluted,
the analyst can add an additional fast ramp to below the column
maximum operating temperature and hold for a few minutes to
ensure other matrix components are fully eluted from the column.

12.3.2 Isothermal Separations

Isothermal separations can be viable for samples with fewer
components and narrower volatility ranges. In this separation
mode, system cool-down and heat-up cycles are not needed, while
early peaks-of-interest can be eluted sooner. This means that
the flexibility for increasing analysis throughput is greater
for isothermal separation. Separation resolution can also be
meticulously optimized, and therefore enhanced, for critical pairs
in isothermal mode. This is especially true for chiral separations.
As with temperature-programmed separations, the optimal
temperature program may not be obtained without some empirical
observation and development.
A recommendation for choosing a starting point would be
using the scouting temperature program noted in Eq. 12.3 and
selecting an isothermal temperature that is around 45°C below
the elution temperature of the last analyte [16]. From the initial
starting point, modify the isothermal temperature based on the
method objectives. A lower temperature can be used to increase
analyte retention and may lead to better resolution, although
peak broadening will be more severe with longer retention
times. Vice versa, a higher temperature can be applied to reduce
analyte retention, run time, and peak broadening, but may
sacrifice separation resolution. When increasing the isothermal
temperature, it is worthy to note that the retention time of an
analyte will be reduced by approximately half if the temperature
is increased by 30°C [15]. Similar to temperature program, it
is recommended that the analyst add a fast ramp to below the
maximum operating temperature of the column with a subsequent
hold, once the analyte(s) of interest has eluted, to remove other
matrix components.
 A Culmination of Instrument, Column and Temperature Optimization 325

12.4 Fast GC: A Culmination of Instrument,

Column and Temperature Optimization
As stated previously, method optimization can revolve around
improving the separation efficiency and/or analysis speed. In
cases where separation speed is the primary focus, the analyst
can alter a variety of physical parameters to achieve “fast GC.”
Fast GC specifically denotes the separation of components in a
sample in a few minutes with peak widths ranging from 1–3 s [18].
In addition to significant gain in analysis speed, applying
fast GC for everyday analyses can provide additional benefits [19].
For example, since the capillary columns used for fast GC are
typically shorter compared to conventional GC columns, they are
relatively cheaper to purchase. Moreover, since the film thicknesses
of typical fast GC columns are thinner, these columns contribute
to lower column bleeding. Lastly, as peak elution is significantly
quicker compared to conventional GC, longitudinal band
broadening effects are less pronounced, resulting in better signal-
to-noise ratios (i.e., sensitivity). The primary limitations for
employing fast GC can be a loss in sample capacity and separation
efficiency. However, some of these issues can be addressed
by careful considerations in method development and/or by
employing selective detection techniques, such as MS, to deliver
peak resolution by mass or spectral differences. Figure 12.5
shows a chromatogram comparison of an analysis of nutmeg oil
using a conventional GC method and column (60 m × 0.25 mm;
1 um HP-1) versus a fast GC method and column (20 m × 0.1
mm; 0.4 um HP-1) [20]. The chromatographic profile is highly
similar; however, the retention time of the final peak was
drastically reduced from 74.9 to 13.3 min. As shown, the immense
gain in analysis throughput is the pinnacle of the potential
benefits gained from applying a fast GC method.
Fast GC can be accomplished by: reducing the column length,
I.D., and film thickness; increasing the linear velocity of the carrier
gas; using a reduced-pressure/vacuum outlet (i.e., GC/MS); and
employing a rapid temperature program [18, 21, 22]. As mentioned
previously, reducing the column length will decrease the retention
times of analytes, leading to a faster overall separation. Although
the resolution may also be reduced when using shorter columns,
326 Gas Chromatography Method Development

Figure 12.5 Comparison of an analysis of nutmeg oil using (A) a conventional

GC method and column (60 m × 0.25 mm; 1 um HP-1) versus (B) a fast GC
method and column (20 m × 0.1 mm; 0.4 um HP-1). Reproduced from [20].
 A Culmination of Instrument, Column and Temperature Optimization 327

careful method optimization can alleviate this potential issue

[23]. On the other hand, decreasing the ID and film thickness of
the column will also reduce the retention time, while providing
an enhancement in the separation efficiency.
Typical GC column dimensions used in fast GC applications
are as follows: column length < 20 m; I.D. < 0.25 mm; and film
thickness < 0.2. It should be noted that the dimensions stated
are not mutually exclusive to fast GC, as “fast GC” can theoretically
be accomplished using any column dimensions. However, the
recommendation provides a feasible starting point to lessen
the burden of choosing a suitable column for fast GC method
development.

Figure 12.6 Comparison of the chromatograms obtained in the analysis

of Japanese green tea extracts. (A) before and (B) after optimization of
the linear velocity and temperature program. Both methods employed
an RTX-5 column (10 m × 0.18 mm; 0.2 um). Reproduced from [24].

Achieving fast GC separation speeds can also be accomplished

by adjusting the linear velocity of the carrier gas and developing a
328 Gas Chromatography Method Development

temperature program which enables a “fit-for-purpose” separation.

This means using a rapid temperature ramp rate or a higher
starting oven temperature, and/or operating the carrier gas
linear velocity above the optimal/recommended speed (shown in
Table 12.2). Figure 12.6 shows a comparison of the chromatograms
obtained in the analysis of Japanese green tea extracts before
and after optimization of the linear velocity and temperature
program [24]. Using the same RTX-5 column (10 m × 0.18 mm;
0.2 um), the authors were able to reduce the retention time of
the last observed peak from 24 to 12 min. This was accomplished
just by modifying the linear velocity from 45 to 80 cm/s and
applying a relatively faster temperature program.
As mentioned previously, employing ­H2 as a carrier gas permits
the analyst to operate at higher linear velocities compared to He or
N2 due to its relatively low viscosity. It is therefore recommended,
but certainly not required, that the analyst exploit H2 as a carrier
gas for fast GC separations.

12.5 Conventional GC Detectors Applied in

Pharmaceutical Analyses
The GC technique has many detectors available at its disposal.
In fact, the wide array of detectors is a major advantage of GC over
other chromatographic techniques. In the pharmaceutical industry,
the flame-ionization detector (FID) and electron-capture detector
(ECD) are the most commonly used. Historically, the thermal
conductivity detector (TCD) is a valuable universal detector that
covers most permanent gases. While there are many more GC
detectors which have been reviewed [25, 26] the FID, ECD, and
TCD will be covered in this section as the most common and useful
that the practicing analytical chemist may routinely encounter.

12.5.1 Thermal Conductivity Detector

The TCD is a universal detector in that it responds to any analyte
with thermal conductivity that differs from the carrier gas. The
detector is most commonly applied to analysis of light, permanent
gasses. TCD has a wide linear range capable of extending over
5 orders of magnitude and is a concentration sensitive detector.
 Conventional GC Detectors Applied in Pharmaceutical Analyses 329

The TCD detector is non-destructive; analyte structure is maintained

after reaching the detector. In contrast to other detectors, the
TCD does not require a flame or flammable gasses, thereby making
it a simpler to use, safer detector. Also, the TCD power consumption
is low and the detector can be miniaturized to be compatible
with portable or chip-type analyses.
The TCD is based on the principle that resistance changes
can be measured on a heated filament when analytes pass by.
A second filament also measures this resistance for a reference
gas, which minimizes drift over the course of a run. In order to
maximize sensitivity of the TCD, the user should select a carrier
gas with a large difference in thermal conductivity between
carrier gas and analyte. Further, operating the TCD at lower
temperatures can help maximize sensitivity. As temperature goes
up, most analytes experience an increase in thermal conductivity.
Because the common carrier gasses Helium and Hydrogen have
some of the highest thermal conductivities, sensitivity decreases
as analytes’ thermal conductivity increases. However, operating
the TCD at lower temperatures should only be employed to the
point where peak shape and performance is maintained. If the TCD
is operated below the maximum GC oven temperature, then the
user runs the risk of peak tailing as the sample may interact with
the tungsten filament.

12.5.2 Flame-Ionization Detector

The FID is also a universal detector that responds to carbon in a
sample. The detector is mass sensitive. That is, the signal increases
proportionally to the carbon content in an analyte but would also
increase as the concentration of analyte increases. This is the
workhorse detector for many industries, and the most commonly
used GC detector for pharmaceutical analysis. In addition to being
a universal detector for almost all volatilizable pharmaceutical
compounds, the FID is relatively easy to use and maintain and
cost effective. The FID also boasts a wide linear range spanning up
to 7 orders of magnitude.
Further, the FID is inherently destructive; carbon in the analyte
is combusted in the FID to give a response. As seen in Fig. 12.7,
the schematic of the FID, analytes are flowed into the detector
where they are combusted in the presence of a hydrogen-oxygen
330 Gas Chromatography Method Development

flame. Charged particles are collected by applying a voltage between

the tip of the jet and a collector electrode. Signal on the order
of picoamps is amplified and collected as a voltage.

Figure 12.7 Schematic representation of the FID [27].

As carbon is combusted in the FID, the amount of carbon being

combusted per unit time in the detector is proportional to signal
response. As such, FID exhibits unit carbon response in the sense
that for each additional carbon present in an analyte molecule,
the signal response increases proportionally to the number of
carbon atoms in the molecule. One advantage of this property of
the FID is that response factors can be estimated, and standard
calibration curves may not be needed for every analyte. In
practice, however, analysts in the pharmaceutical industry would
typically still fully validate all known impurities or solvents for
a GC method, which would include calibration of all analytes.
While the unit carbon response property of FID is a useful
property to understand and apply, regulatory expectations and
 Conventional GC Detectors Applied in Pharmaceutical Analyses 331

risk management in the pharmaceutical industry dictate that the

amount of extra work involved in preparing additional calibration
curves during validation is minimal against the risk of an analyte
behaving in an unanticipated manner. However, for unspecified or
unknown impurities that cannot be validated, the more uniform
response is a major advantage to the FID. If using GC for an
impurities method in the pharmaceutical industry, the response
between the main component and related substance impurities
is similar for a GC-FID method. Because unknown impurities in a
product and method by nature cannot be validated, the similarity
in response factors between related substances when using an
FID helps minimize risk that an unknown impurity would be
quantitated vastly inaccurately.

12.5.3 Electron Capture Detector

The electron capture detector (ECD) is a non-destructive,
concentration-based detector that is suitable for analytes with
electron-capturing capability. The linear range of this detector
spans approximately 4–7 orders of magnitude. In principle, the
detector operates based on a reagent gas with low ionizability as
the background signal. A radioactive source ionizes this reagent
gas, and the resulting electrons that are emitted by the reagent
gas are collected as a signal. As analytes pass through the detector,
they are able to capture some of these electrons that exist as a
plasma from the reagent gas. The higher a given analytes’ ability to
capture electrons, the larger the background signal will be reduced
and thus, a higher overall absolute signal will be observed. The
ECD is traditionally highly sensitive toward halogenated analytes
due to high electronegativity properties that are conducive to
electron capture, and the signal response is greatest for iodine,
followed by bromine, chlorine, and fluorine. The more that a carbon
atom is substituted directly or adjacently with electronegative
heteroatoms, the higher the observed signal response.
As seen in the ECD configuration in Fig. 12.8, a nickel-63
radioisotope source is frequently seen as part of ECDs due to
high thermal stability as well as availability and practicality of
implementing this radioactive material. While designs may have
slight variations or be miniaturized, the principle of the ECD is
that GC column eluent is mixed with make-up gas and brought to
332 Gas Chromatography Method Development

the radioisotope source. In this chamber, secondary electrons

are generated when high-energy beta electrons collide with
carrier gas. These thermal electrons form a plasma, which also
includes radicals and positive ions, and a baseline current is
established for the detector. When electron-capturing analytes
pass through this plasma and capture the thermal electrons,
negative ions are generated. After neutralization by positive ions
present in the detector, the background plasma signal is altered
and signal response is generated proportional to the analytes’
ability to capture electrons.

Figure 12.8 Schematic representation of the radioisotope-ECD [27].

GC instruments with ECD setup have a wide range of

applications in the pharmaceutical industry due to the
 GC/Mass Spectrometry 333

frequency of good electron capturing groups on drug molecules.

For example, methods have been developed that can simultaneously
quantitate over 50 drugs in plasma or whole blood by utilizing
GC-ECD. The method was validated, and LOQs as low as 2 ng/mL
were achieved [28]. Mutagenic impurity analysis was also recently
accomplished by probing different ionic liquid-based diluent
and using GC-ECD in order to develop highly sensitive methods
down with LODs as low as 5 ppb, which is an improvement over
1,000 fold compared with GC-FID and comparable with LC/MS
and GC/MS [29].

12.6 GC/Mass Spectrometry

Hyphenated techniques, such as GC/MS and LC/MS, have become
essential in the pharmaceutical industry within the past decades.
Compared to other detection techniques, coupling GC to MS
provides the analyst with the ability to fully characterize a sample
by quantification and structural elucidation. Applications of GC/MS
in pharmaceutical analyses can vary from identification of
impurities, to peak purity analysis, to quantification of potent/
toxic compounds at trace levels in drug substances and products
[30–33]. Compared to other detectors, MS detectors can provide
better signal sensitivity and universal detection for most analytes.
Employing mass spectrometry also adds another dimension
of separation to sample analysis. Instead of a chromatographic
separation, the analyst can take advantage of separation of ions
based on their mass-to-charge ratio (m/z). By selecting the
proper ion(s) to analyze, it is possible to selectively detect
and quantify peaks which would otherwise co-elute by
chromatography [34]. For example, Xue et al. selected two distinct
ion fragments observed during the analysis of penehyclidine
(PH; 175 m/z) and its internal standard penehyclidine-d5
(PH-d5; 180 m/z) by GC-MS, as shown in Fig. 12.9 [35].
Although the chromatographic retention times of the drug and
its internal standards were the same (~8.2 min), specificity for
each analyte was obtained when only the ion fragment 175 m/z
for PH and 180 m/z for PH-d5 were extracted from the MS data
(Fig. 12.10), thereby enabling quantification [35].
334 Gas Chromatography Method Development

Figure 12.9 Mass spectra (electron ionization; EI) for: (A) penehyclidine
and (b) the internal standard, penehyclidine-d5. (A) EI-MS of m/z
315  175 for penehyclidine. (B) EI-MS of m/z 320  180 for penehyclidine-
d5. Reproduced from [35].
 GC/Mass Spectrometry 335

Figure 12.10 Extracted ion chromatograms of penehyclidine and

penehyclidine-d5. (A) Penehyclidine extracted at 175.00 m/z. (B)
Penehyclidine-d5 extracted at 180.00 m/z. Reproduced from [35].

12.6.1 Ionization by GC/MS

The mechanism of a GC/MS detector can be simplified to four
primary functions:
1. Ionization of volatile matrix components as they exit the
vacuum outlet of a GC column.
2. Acceleration and focusing of the ionized species and their
fragments to a mass analyzer.
3. Separation of the ionized species based on their relative
mass to charge ratios (m/z) in the mass analyzer.
4. Detection of the ionized species as they exit the mass
analyzer.
The ionization process is critical to generate detectable
components in a sample. As such, it is beneficial to briefly examine
the underlying principle of various ionization modes in GC/MS.

12.6.1.1 Chemical ionization

Commercially available GC/MS can be equipped with chemical
ionization (CI) sources capable of producing positively charged
or negatively charged ions. Ions generated via CI typically possess
the same composition as the parent compound, with one more
(or less) electron. These ion species are often referred to as
336 Gas Chromatography Method Development

molecular ions. As CI does not tend to generate additional ion

fragments from the parent molecular ion, it is considered a soft
ionization technique. Employing CI in GC/MS allows the analyst to
ascertain the molecular weight of an analyte. Mass spectral data
generated from CI experiments are relatively easier to interpret;
however, they do not provide structural information of the
compound.

12.6.1.1.1 Positive ion

Positive CI was introduced by Munson and Field in the
mid-1960s [36, 37]. Briefly, this technique involves the continuous
introduction of a reagent gas, such as methane, to the GC/MS
ionization source. During its residence in the source, the reagent
gas is ionized to a primary cation by an electron beam generated
from applying a potential to a tungsten filament. The primary
cations can also react with other the reagent gas molecules to
form secondary ions (Scheme 12.1). When a cloud of sample
molecules is introduced to the source, the reagent gas ions (primary
and secondary) undergo various chemical reactions with these
molecules to generate molecular ions (M+), see Scheme 12.2. It is
worthy to note that the molar ratio of the reagent gas is around
three orders of magnitude higher than the sample molecules. This
mitigates the direct ionization of the molecules by the electron
source.

Scheme 12.1 Ionization of methane reagent gas by positive CI. *The

primary cation (CH·4+) can react with bulk reagent gas to form secondary
cations, which can also further react.

12.6.1.1.2 Electron capture negative ionization

A complementary and less-understood CI technique is electron
capture negative ionization (ECNI). This technique was developed
by Stafford and Hunt in the mid-1960s [38, 39]. As the name
implies, this ionization technique produces negative molecular
 GC/Mass Spectrometry 337

ions of the sample gas. This ionization technique shares a number

of similarities to positive CI by requiring the use of a reagent
gas (i.e., methane, ammonia, isobutane, etc.) in a closed ionization
source to aid in the production of sample ions. The primary
difference is that ionization in ECNI is accomplished by the
capture/reaction of thermal electrons (low-energy; 0–5 eV) expelled
during the ionization of the reagent gas, rather than with the
reagent gas ions themselves (see Scheme 12.3 for ions generated
by ECNI). ECNI can be set up by changing the mode of detection
to anion mode and following the manufacturer’s recommended
settings for reagent gas ionization.

Scheme 12.2 Ionization pathways in positive CI using methane as a reagent

gas. (A) Charge transfer, (B) proton transfer, (C) hydride abstraction,
(D) adduct formation.

Scheme 12.3 Ionization pathways in ECNI. (A) Resonance electron

capture, (B) dissociative electron capture, (C) anion formation by
ion-pairing.

12.6.1.2 Electron ionization

Perhaps the most common ionization technique for GC/MS is
electron ionization (EI) [40]. This technique exhibits remarkable
utility for the universal detection and characterization of volatile
338 Gas Chromatography Method Development

compounds. EI is a hard ionization technique, which generates

the sample molecular ion and its corresponding fragments.
Based on the ion fragments formed during the ionization step,
the analyst can elucidate the chemical structure of the sample
molecule. An example mass spectra for a common synthetic
precursor, epichlorohydrin (MW: 92.52 g/mol), is shown in
Fig. 12.11 [41].

Figure 12.11 Mass spectra of epichlorohydrin generated by electron

ionization. Obtained from [41]. Reprinted (adapted) with permission
from Loda, C., Bernabe, E., Nicoletti, A., Bachi, S., and Dams, R. (2011)
Determination of epichlorohydrin in active pharmaceutical ingredients
by gas chromatography-mass spectrometry, Organic Process Research
& Development, 15, pp. 1388–1391. Copyright 2018 American Chemical
Society.

As shown, the mass spectral data obtain from EI can be

relatively complex and the molecular weight of the sample
may not be apparent using EI, due to fragmentation and/or
rearrangement [41, 42]. Therefore, EI, CI and ECNI should be
exploited as synergistic techniques to obtain the molecular weight
and chemical structure/composition of a sample molecule.
A schematic illustrating the EI source is shown in Fig. 12.12.
Similar to CI and ECNI, the EI source houses a heated metallic
 GC/Mass Spectrometry 339

filament which emits and accelerate electrons towards an anode

(via potential difference). The energy of the electrons generated
is typically 70 eV, a convention that ensures the ionization of
organic molecules. When sample molecules enter the source
from the GC column outlet, it is immediately exposed to the high-
energy electrons generated therein. The energy exerted by the
electrons consequently increases the energy level of a sample
molecule. To revert to a lower energy level, the molecule can
expel an electron and form a radical cation [M]+·, as shown in
Scheme 12.4. It is also possible for the molecule to abstract an
electron (Scheme 12.4); however, ions generated in this fashion
are much less abundant (~104 times) compared to radical cations.

Figure 12.12 Schematic of the EI source in GC/MS.

Scheme 12.4 Ionization pathways in EI. (A) Electron expulsion, and (B)
electron abstraction.

Even when sample molecule expels an electron to form

a radical cation and revert to a lower energy level, it may still
possess significant excess energy. Therefore, the molecular ion will
ultimately undergo unimolecular changes to reduce its energy level.
340 Gas Chromatography Method Development

These chemical changes include fragmentation or rearrangements

of the parent ion(s), and/or undergoing a unimolecular
decomposition to form a cation and a neutral-radical species
(Scheme 12.5).
  
0 )  1
Scheme 12.5 Formation of cation and a neutral-radical species by EI.

It is important to note that different molecules will fragment

in distinctive patterns, and some fragments are produced in
higher abundances than others. This essentially results in a unique
and highly reproducible mass spectral fingerprint for a sample
molecule and enables the use of a database/library to aid in their
identification [43].

12.6.2 Systematic Approaches for Chemical Structure

Elucidation by EI-GC/MS
In the modern era of pharmaceutical analyses, there has been
overwhelming interest on utilizing GC/MS for the rapid and
accurate characterization of unknown impurities generated in
a synthetic process. This is especially true for potentially toxic
impurities, such as carcinogens and mutagens (i.e., mutagenic
impurities) [31]. EI-GC/MS has thus become an essential tool
for this continuing endeavor. Mass spectral data interpretation
and fragmentation rules for EI techniques have previously been
well established in literature. It is advised that the analyst re-visit
these resources and consolidate information therein into their
own interpretation practices. Nevertheless, even with these
resources at one’s disposal, it may be time-consuming and difficult
to interpret unknown compounds without background knowledge
of the chemical process. In this section, we discuss simple and
effective strategies to tackle the challenging task of elucidating
and characterizing “unknown peaks” by EI-GC/MS.

12.6.2.1 Interpretation of mass spectral via database

The simplest approach to predicting the chemical makeup of an
unknown compound is to utilize the vast mass spectral libraries
 GC/Mass Spectrometry 341

available commercially. Since ionization by 70 eV energy produces

a highly reproducible fragmentation pattern, there has been
appreciable effort to harmonize the mass spectral data of a
numerous compounds into searchable databases [44, 45]. For
example, the National Institute of Standards and Technology
(NIST) and Wiley database store mass spectral data for hundreds
of thousands of chemical structures, and these libraries are rapidly
expanding. Additionally, many mass spectral data interpretation
software can also provide statistical analysis on the similarities of
an unknown compared to a known structure from the database
in mere seconds. As a result, commercially available mass
spectral databases are an outstanding starting point and a key
arsenal in the interpretation of unknown structures [46].
When using mass spectral libraries to identify an unknown
structure, care should be practiced in judging whether library
matches actually pertain to the unknown compound. The best
way to mitigate inaccurate structure identification is to obtain
a chemical standard of the predicted compound and perform
additional confirmation studies (i.e., comparison of retention times/
factors, spiked-recovery studies, etc.) for verification. However, in
situations where this is neither practical nor possible, the analyst
can consider the following recommendations to utilize the mass
spectral library more effectively.
(1) Perform background subtraction of the mass chromato-
graphic peak of interest at the apex position to account
for artifacts and interferences. Subtract mass spectral
data of components contributing to peak asymmetry (i.e.,
shouldering peaks, split peaks, peak fronting/tailing) from
the apex. This reduces the abundance of unrelated ion
fragments, resulting in a more accurate representation of
the sample mass spectrum.
(2) Consider the statistical results of the library matches and
verify whether multiple matches of the same compound
are reported. Multiple matches of the same compound may
indicate that the proposed structure is indeed the correct
structure. Additionally, utilize side-by-side comparisons
of the sample mass spectrum with respect to the library
match, such as difference and/or head-to-tail plots, to
observe variations in fragmentation patterns.
342 Gas Chromatography Method Development

12.6.2.2 Manual interpretation of mass spectral data

In cases where there are little-to-no logical library matches,
a systematic approach can be followed to interpret the mass
spectrum of the unknown peak [47, 48]:
(1) Determine the molecular weight from the spectrum by deducing
the molecular ion peak.
The molecular ion peak (M+·) is an essential puzzle piece
needed for accurate characterization of an unknown
structure, as it pertains to the molecular weight of the
structure. In general, potential M+· ions formed in EI-MS
should possess the highest m/z value, be an odd electron
(OE) fragment since neutral molecules lose a radical during
ionization, and rationally yield fragment masses by loss of
neutral molecules/species. Although EI can fragment the
molecular ion peak, residual abundances of the molecular ion
may be observed, depending on the structure of the sample
molecule. For example, the mass spectrum of analytes
possessing unsaturated and/or aromatic cores may show
a relatively high abundance of the M+· peaks, since its
molecular ion peaks can be stabilized through resonance.
Conversely, the molecular ion peak may be observed at
relatively low abundances (or not at all) in the mass spectrum
of analytes comprising aliphatic/saturated cores. If the
molecular ion peak is not observed in a mass spectrum,
it may be worthwhile to pursue alternative ionization
techniques (i.e., CI and ECNI) or reduce the EI electron
energy to obtain this information.
(2) Determine the elemental composition using the isotopic
abundances present in the molecule. Obtain further insight by
using the rings + double bonds rule and the nitrogen rule.
Isotopic Abundances: Once an M+· candidate has been
identified, the elemental composition of the unknown can
be deduced using the isotopic abundances of the molecule.
Peaks in a mass spectrum are represented in groups.
Within a group, the peak with the lowest m/z value is the
monoisotopic peak, while successive peaks with +1 or more
m/z are isotope peaks. Determining the relative abundance
 GC/Mass Spectrometry 343

of the isotope peak(s) with respect to the monoisotopic

peak give insight to the atoms comprised in the unknown
compound.
To determine the elemental composition, measure the relative
intensities of the isotope peaks and the monoisotopic peaks
and compare the results with the relative abundances
provided in Table 12.3 [47].

Table 12.3 Elements and their natural isotopic abundances

Abundance Abundance
Element (A) (%) A+1 (%) A+2 Abundance (%)
1H 99.99 2H 0.01 — —
12C 98.91 13C 1.1 — —
14N 99.6 15N 0.4 — —
16O 99.76 17O 0.04 18O 0.02
19F 100 — — — —
28Si 92.2 29Si 4.7 30Si 3.1
31P 100 — — — —
32S 95.02 33S 0.76 34S 4.22
35Cl 75.77 — — 37Cl 24.23
79Br 50.5 — — 81Br 49.5
127I 100 — —
Source: Adapted from [47].

Typical elements (A) may possess one or two isotopes (A+1

or A+2, respectively) in significant abundances. Elements
with no appreciable levels of A+1 or A+2 isotopes are H, F, P,
and I. Elements with A+1 isotopes are C and N. Elements
with both A+1 and A+2 isotopes are O, Si, and S. Elements
with only A+2 isotopes are Cl and Br. Interestingly, carbon
(12C) has a distinct isotopic distribution; its A+1 isotope
(13C) is 1.1% of compared to the abundance of the 12C
monoisotope. This means that it can be fairly easy to
deduce the number of carbon atoms in a fragment group.
For example, a fragment group with 1 carbon atom will
display an A+1 abundance of 1.1%, while a fragment group
344 Gas Chromatography Method Development

comprising 8 carbon atoms will display an A+1 abundance

of 8.8%, and so forth.
The Nitrogen Rule: The parity (even or odd numbers) of
nitrogen atoms comprised in the sample can be determined
by using the Nitrogen Rule. Specifically, if a compound is made
up of an even number of nitrogen atoms (or none at all),
its molecular weight will be an even number. Similarly, odd
molecular weights indicate an odd number of nitrogen
atoms. Utilizing the isotope abundance described in the mass
spectrum alongside the nitrogen rule provides sufficient
discernment (in many cases) to hypothesize the elemental
composition of the molecular ion.
The Rings + Double Bonds Rule: After the elemental
composition is deduced using the isotope abundance
and nitrogen rule, the degree of hydrogen deficiency
(unsaturation) of the sample molecule can be determined
using the Rings + Double Bonds (R + dB) Rule.
For a given formula with known valance for C, H, and N, the
total number rings and double bonds is calculated as follows:

  R + dB = C# – 12 H# + 12 N# + 1

Take pyridine for example: The molecular formula of

pyridine is C5H5N1. Thus, the R+dB for pyridine = 5 – 1/2(5)
+ 1/2(1) + 1 = 4, meaning that there can be up to three
double bonds and one ring.
It is important to note additional considerations for this
rule [47].
(3) Predict and interpret the connectivity of atoms in the molecule.
Once the chemical composition of the molecular ion peak
has been established, attention should be focused on
understanding the connectivity of the molecule. A strategy
to elucidate the connectivity is to determine the losses
of “common” fragments starting from the molecular ion
peak. Table 12.4 shows common ion fragments formed by
EI [49]. While identifying common ion fragment losses,
postulate the possible connectivity between them and
whether the connections form logical chemical substituents.
 GC/Mass Spectrometry 345

A number of references are available to provide in-depth

fragmentation and rearrangement mechanisms to aid in
interpretation [48, 50].

Table 12.4 Common fragment losses from the electron ionization of the
molecular ion

Parent ion-loss Fragment Parent ion-loss Fragment

M-1 H M-32 CH3OH, S
M-2 H2 M-34 H2S
M-15 CH3 M-35 Cl
M-16 O, NH2 M-43 CH3CO, C3H7
M-17 OH, NH3 M-44 CO2, CH3CHO
M-18 H2O M-45 CO2H, OC2H5
M-19 F M-46 C2H5OH, HCO2H
M-27 HCN M-60 CH3CO2H
M-28 CO, C2H4 M-73 (CH3)3Si
M-29 CHO, C2H5 M-79 Br
M-30 CH2O, C2H6 M-89 (CH)3SiOH
M-31 CH3O M-127 I
Source: Adapted from [49].

(4) Finalize and hypothesize possible structure(s) and compare

against known databases.
With data in hand, tally up all results gathered from
Steps 1–3 and hypothesize the structure(s) of the unknown
compound. Revisit literature to investigate whether
analytical/characterization data (i.e., NMR, IR, UV, XRD, etc.)
exists for the structure(s) proposed. Chromatographic data,
such as retention indices, may be available in the literature
and can also be valuable information [51]. Finally, obtain
and inject the standard compound(s) of interest to verify
the findings. It is worthy to note that modern mass
spectral libraries allow the user the option to input their
own structures and mass spectra to the existing database.
The analyst should take advantage of this option and
346 Gas Chromatography Method Development

populate their results into these libraries for future

reference.

12.7 Considerations for Using GC in the

Pharmaceutical Setting
GC analysis in the pharmaceutical environment is a powerful tool
to tackle complex analytical challenges. Within the realm of
industrial and pharmaceutical analysis, GC methods are used across
all aspects of chemical synthesis, processes, and manufacturing.
From starting materials, to intermediates, raw materials, and
final active pharmaceutical ingredients (APIs) and drug product
(DP), GC has found many applications in the pharmaceutical
industry. Perhaps the most routine application of GC is for residual
solvent analysis. However, GC can be useful for impurities (e.g.,
related substances) analysis, main component assay, and
chromatographic identification methods of chemicals that may
typically be analyzed by HPLC. Further, “in-process” testing, such
as reaction monitoring or solvent switching, can easily be handled
by GC in many cases. Lastly, as regulatory expectations tighten
around the understanding and low-level control of mutagenic
impurities (MIs) in products for human consumption, GC-based
methods have proven to be very useful for the analysis of these
highly scrutinized impurities that often require detection down
to ppm levels or lower.

12.7.1 Analysis of Residual Solvents

The development and implementation of residual solvent
methods for chemical analysis should be kept straightforward
and simple whenever possible. When faced with development of
a residual solvent method for GC, the analytical chemist should
not need to place much work around method development
because these types of methods are routine and typically already
exist, either at the company/laboratory or in the literature. For
methods that are going to be validated and implemented in a
quality control (QC) setting, the recommendation is to use a
GC-HS system instead of a direct injection setup. The reason for
this recommendation is that the main component or matrix (e.g.,
 Considerations for Using GC in the Pharmaceutical Setting 347

the drug substance or drug product) may degrade on the GC inlet

or already contain additional high-boiling components that can
show up on the GC chromatogram. For a QC lab, this can be
problematic depending on the quality assurance system that exists
because extraneous peaks can trigger investigations, even for
targeted impurities methods such as residual solvents analysis.
Direct injection is still routinely used for residual solvent analysis,
but care should be taken during method development to ensure
that extraneous peaks are either not present or documented and
explained.
A good rule of thumb is that any peak near or above the
practical quantitation limit of the method (default to the lowest
one in cases of multiple solvents) are addressed, documented, and
noted in the method before implementation in a QC setting.
However, the QC laboratory’s quality and compliance systems
(i.e., operating procedures) should be followed first and foremost.
Again, a good practice is to adopt GC-HS as the standard setup
for GC methods that will end up being validated and run in QC
labs to minimize the potential for extraneous peaks. Further, GC-
HS has an advantage of being more sensitive in practice. Because
of the nature of GC-HS, the main component is often not volatilized
and is therefore not brought to the inlet of the GC, which in the
case of direct inject GC, can limit sample concentration that
can be tolerated by a method. This benefit of GC-HS means that
much higher sample concentrations can be achieved relative to
direct injection, thus increasing the observed LOQ for the analyst
developing the method.
An example of a GC-HS method used for testing of residual
solvents in pharmaceutical products is shown below in Fig. 12.13.
In this example, the original proposed method (Panel A) is for
the quantitation of seven residual solvents using NMP as the diluent.
This GC-HS method achieves good resolution of the solvents of
interest. However, residual solvent analyses are often desired
in short time frames to determine if further drying is needed to
remove more a given solvent during in-process production, or to
quickly complete product release testing. As such, a 30 min method
is workable but far from ideal considering that one injection
sequence may include a few blanks, LOQ standard, six working
standards and multiple sample injections. By developing a fast GC-
HS method in 8 min (Panel B), the analysts were able to achieve
348 Gas Chromatography Method Development

a method with excellent resolution and similar accuracy and

LOQs. The chromatographic conditions of the fast GC-HS method
are shown in Table 12.5. As seen, instead of a method where the
chromatography occurs over 30 min plus re-equilibration time,
the fast GC-HS allowed for the chromatography to occur in under
8 min. In changing from the longer to the faster method, one big
change was the column length: 20 m instead of 60 m. Additionally,
a 0.18 mm ID column was used for the fast method versus a
0.25 mm ID column for the longer method. Compared to larger
inner diameter columns, smaller inner diameters will afford
higher plate counts. Therefore, one strategy for developing faster
GC methods is using smaller inner diameter columns, while also
using shorter length columns. While shorter length columns
will have less total theoretical plates, this is partially offset by
the plate count boost by using smaller inner diameter columns.

Figure 12.13 (A) Original GC-HS method for in-process testing of residual
solvents, and (B) Fast HS-GC method for in-process testing of residual
solvents. Elution orders are the same of solvents peaks in (B) as (A). Credit:
Harry Patel, Umesh Patel, Tien Ho (Process Research & Development,
AbbVie, Inc.).
 Considerations for Using GC in the Pharmaceutical Setting 349

Table 12.5 GC-HS conditions for the optimized (fast) GC-HS method

Carrier gas
(mL/min) 1.3 (constant) Temperature program
Split Ratio 30-1 35°C hold for 1.5 min, ramp at 25°C/
min to 90°C, hold for 0 min, ramp
Injector (°C) 210 35°C/min to 235°C, hold for 3.16 min
Detector (°C) 250
FID hydrogen flow 40
(mL/min) Vial incubation time: 10 min
FID air flow 40
(mL/min)
FID nitrogen 25 GC run time: 11 min/GC cycle time:
makeup (mL/min) 17 min
Column
GC-HS oven (°C) 100 624-type phase; 20 m × 0.18 mm ×
1.0 μm
GC-HS loop (°C) 110
GC-HS transfer 130
line (°C)

Additionally, fast direct inject GC methods are also useful

depending on equipment availability and ease of use and
preparations. As seen in Fig. 12.14, 37 solvents, including the
diluents, were separated in less than 7 min. Most solvents were
well resolved, which would allow for a single method to be used
to solve a variety of analytical challenges and with short runtime.
The chromatographic conditions for the fast GC direct inject
method are shown in Table 12.6.
For work outside of the QC environment, fast and user friendly
methods have been developed to aid in R&D activities in the
pharmaceutical development space. For example, Hamilton and co-
workers have implemented and automated GC-HS system for the
determination of residual solvents [52]. By applying this method
to both isolated solids and reaction slurries of pharmaceutically
relevant compounds, solvent analysis can be accomplished in a
fast and user-friendly manner. Recently, innovative approaches to
residual solvent analysis have allowed for much lower detection
limits to be obtained using a standard GC-FID configuration.
350 Gas Chromatography Method Development

Nacham and co-workers [53] showed that by using room-

temperature ionic liquids (RTILs), residual solvent analysis in drug
substances could be achieved with up to 25 times improvements
in LODs. By exploiting the high thermal stability and low vapor
pressure of RTILs as diluents, GC-HS methods could be pushed
to higher incubation temperatures with less interfering peaks
from the diluent over conventional organic solvent diluents.
LODs in the range of 6–20 ppm were achieved, an impressive
accomplishment that approaches LODs needed for low-level
detection of mutagenic impurities.

Figure 12.14 Fast GC direct inject method for residual solvent analysis.
Solvents: (1) Methanol, (2) Pentane, (3) Ethanol, (4) Diethylether, (5) Ethyl
Formate, (6) Acetonitrile, (7) Dichloromethane, (8) t-Butanol, (9) MTBE,
(10) n-Hexane, (11) 1-Propanol, (12) Nitromethane, (13) Methyl Ethyl
Ketone, (14) Ethyl Acetate, (15) Tetrahydrofuran, (16) Chloroform,
(17) Cyclohexane, (18) Carbon Tetrachloride, (19) 1,2-Dimethoxy ethane,
(20) Benzene, (21) Isopropyl Acetate, (22) 2-Methyl Tetrahydrofuran,
(23) n-Heptane, (24) Methyl Cyclohexane, (25) 1,4,-Dioxane, (26)
n-Propyl Acetate, (27) MIBK, (28) Pyridine, (29) Toluene, (30) 2-Hexanone,
(31) Dimethylformamide, (32) Chlorobenzene, (33) Ethyl benzene, (34)
Propylene Glycol/m-Xylene/p-Xylene, (35) o-Xylene, (36) DMSO, (37) NMP.
Credit: Harry Patel, Umesh Patel, Tien Ho (Process Research & Development,
AbbVie, Inc.).
 Considerations for Using GC in the Pharmaceutical Setting 351

Table 12.6 GC conditions for fast GC direct injection method for residual
solvent analysis

Carrier gas (mL/min) 1.3 Temperature program

Split ratio 25–1 35°C hold for 1.5 min, ramp at 25°C/min
Injector (°C) 210 to 90°C, hold for 0 min, ramp 35°C/min to
200°C, hold for 3.16 min
Detector (°C) 240
FID hydrogen flow 40 GC run time: 10 min
(mL/min)
FID air flow (mL/min) 400
FID nitrogen makeup 25
(mL/min)
Injection volume (μL) 1 Column
624-type phase; 20 m × 0.18 mm × 1.0 μm

12.7.2 Analysis of Ordinary Impurities

Impurity analysis by GC encompasses a wide range of possible
methods. In this section, ordinary impurities refer to non-MI, non-
solvent/reagent impurities whereas main component refers to
product (e.g., a starting material, intermediate or drug substance).
This includes related substances analysis, targeted chromatographic
impurities, and assay methods. In pharmaceutical analysis, one of
the most critical quality controls is measuring related substances
(impurities) of the main component, as well as assay of the main
component. Testing is performed against specifications in a QC
environment. These related substances measurements are critical
to patient safety and meeting regulatory requirements because
their allowable limit is directly related to process rejection
capability as well as toxicological data for safety.
Related substance measurements typically occur at many points
along the synthesis of a pharmaceutical product from starting
materials and raw materials, to intermediates, drug substance,
and drug product. Traditionally, HPLC-UV instrumentation and
detection is the standard scheme for a related substance
method. Understandably so, as the many column choices, good
352 Gas Chromatography Method Development

chromatographic resolution, robustness of the UV detector, and

good instrument precision allow for most related substance
analysis needs to be met. However, as pharmaceutical products
become more advanced and complex, there is a trend for regulatory
agencies to strengthen their oversight of the Good Manufacturing
Process (GMP) and pushback drug makers further upstream to
simpler, smaller molecule starting materials, often times commodity
chemicals. For example, a simple amino acid could be designated
a regulatory starting material (RSM) by pharmaceutical regulators,
even for pharmaceutical molecules that are much more complex
in structure. As such, validated, regulatory test methods are needed
for these small, simple starting materials and their respective
downstream intermediates. Traditional HPLC-UV methods are
not always the best choice, particularly when a good chromophore
may be lacking or there is low retention and resolution with
typical HPLC methods. In these cases, GC methods are more
sensitive, can provide more uniform response factors, and
can provide for better chromatographic resolution for related
substances analysis.
Consider a main component on an HPLC-UV method contains
an aromatic ring, and one unspecified impurity that was not
validated appears after a few years of commercial manufacture of
a pharmaceutical product. Because of response factor differences,
if this new unspecified impurity had a disruption or loss of
aromaticity, a 1% impurity could possibly be inaccurately
quantitated at 0.1% or 0.01%, for example. However, if the method
were GC-FID, then a 1% impurity would likely be quantitated
much closer to 1%, because the number of carbons between a
main component and a related substance impurity of the main
component are very similar, and FID offers more uniform response.
While electronegative functional groups, for example, can alter the
unit-carbon response, GC-based detection is still more uniform
than UV many cases where UV chromophore of a main component
or impurity are drastically different. While GC-FID for impurity
analysis of pharmaceuticals is rarely used for full profiling and
testing of related substance impurities, this is an area for expanded
research and implementation that should be further explored as
a means to mitigate risk of severely under- or over-quantitating
impurities by HPLC-UV methods.
 Considerations for Using GC in the Pharmaceutical Setting 353

A risk-based approach by analyzing the main component and

potential impurities should first be performed, and in cases where
GC-FID makes sense, this should be considered over HPLC-UV.
One example of where GC-FID would excel is for a starting material
or intermediate impurities method where the main component
has no significant chromophore. While an HPLC-UV detector at
200 nm might be a typical path forward, GC-FID offers significant
advantages in higher sensitivity and more uniform response.
If an aromatic impurity is present in the HPLC-UV method and
present at 0.01%, it is not unreasonable that due to response
factor differences this impurity would be quantitated at 1.0% and
cause unnecessary failures of product. With add-ons available to
GC-FID detectors to create more uniform response, such as the
Polyarc reactor that converts all species to into methane before
reaching the FID, there is much room for additional implementation
of GC-FID for standard impurities methods.
GC coupled with detectors other than the FID can also be useful
tools for impurity and related substance analysis. For example,
Belal and co-workers developed a GC-MS with direct injection
that was used for the analysis of Trimetazidine Dihydrochloride in
order to quantitate two of its related substances with a stability-
indicating method. In this method, an RTX-1 column was employed,
and an electron impact (EI) ionization source was used to
generate ions for MS analysis [54]. In other work from Balal and
co-workers [55], they showed the GC-MS chromatogram of the
antihistamine drug dimenhydrinate (DMH) and separation from
six of its related substances. Further, Balal and co-workers
[55] found that their GC-MS methodology was also potentially
suitable for main component assay analysis. By using GC-based
methodology, the analytical chemist has a powerful tool at his
disposal to tackle analytical challenges where HPLC-UV may have
significant deficiencies.
Reaction products and impurities are also well suited for GC
analysis. By using GC-MS technology, unknown products can be
fragmented with an electron impact (EI) ionization source and
fragments can be interpreted or compared against a library to
search for likely structures. Chemical ionization (CI) can also be
used to observe the molecular ion and pull a mass on intact analytes
to help identify a structure. Van den Berg and coworkers showed
354 Gas Chromatography Method Development

that acid-mediated transformation of cholesterol and phospholipids

can be studied by GC-MS. The authors notably found that no
chlorohydrin species were detected as would be expected with
unsaturated fatty acids. Instead, hydroxy- and keto-derivatives
were confirmed by GC-MS [56].

12.7.3 Main Component (Assay) Methods

In pharmaceutical analysis, an assay specification is critical to
ensure the quality of starting materials, intermediates, and drug
substance/product. Whereas related substance analysis will test
for the presence of low-level impurities, assay testing is a catch-all
approach to determine any impurity. For example, if a drug was
100% pure from all related substances, solvents, metals, and so
forth, but there was 15% of sodium chloride present, then the
assay testing would return a result of roughly 85%. If this drug
were being tested against an assay specification of not less than
98.0%, for example, then this 85% result would clearly fail. Again,
typically HPLC-UV would be used for main component assay
analysis, but the in certain cases, GC is more appropriate.
The most important consideration when developing a GC
method for assay is if the method relative standard deviation (RSD)
is suitable for the intended testing. As a good rule of thumb, the
statistical principle that three standard deviations on each side
will encompass 99.7% of normal variance should be applied
because this will ensure that specification failures do not routinely
occur due to normal method variance. For example, if a method
being developed is intended to be used for an assay specification
of not less than 95%, then the assay gap is 100% – 95% = 5%.
Then, calculate the standard deviation such that the assay gap
equals 3 standard deviations. In this example, 5%/3 = 1.7%.
This exercise tells the analyst that a method is needed that can
reliably achieve area %RSD of the standard of not more than 1.7%.
This is challenging to achieve for low concentration standards
with GC analysis, and can also be challenging to achieve due
to inherent variability in some detectors, such as MS where
ionization and signal capture of ions through a quadrupole is less
repeatable than GC-FID.
With GC-FID, and an appropriate concentration of standard,
area %RSD of around 1% can routinely be achieved to accomplish
 Considerations for Using GC in the Pharmaceutical Setting 355

a main component assay method. However, for tighter assay

specifications, the analyst will run into repeatability issues
when using GC. For example, when developing and assay by GC
method to test against a not less than 98% specification, the
analyst would want a method than can reliably hit area RSD of
not more than 0.67%. In this case, even with an optimized sample
concentration, inlet liner, column, peak shape, and so forth, this
criterion would still be challenging to regularly achieve by GC
analysis. In these cases where low area %RSD is needed, GC analysis
with an internal standard is recommended. Ideally, a structurally
similar compound that is resolved away from the main component
for assay would be used as an internal standard. By invoking an
internal standard, an optimized GC method that typically would
return a result of 1% RSD might now be expected to reliably
achieve results of <0.5% RSD, even as low as 0.1–0.2% as would
be possible on an HPLC-UV assay method. Because the internal
standard experiences similar variability as the main component
(e.g., injection precision, inlet liner interactions), the area ratio
of main component to internal standard provides a much better
degree of repeatability and thus low %RSD. Internal standards
can similarly be applied to impurities methods when traditional
GC impurities methods struggle to achieve repeatability, linearity,
accuracy, and so forth.

12.7.4 In-Process Testing

GC has been used in various industries as a means to monitor
various synthetic, chemical, or physical operations. These methods
are typically run “in-process” industrially, whereby the analytical
result is needed quickly to gain insight into the progress of an
operation in order to move on to the next step. For such methods,
time is of the essence, and GC is often a preferred choice.
For example, for a solvent switching operation, one solvent may
be distilled off while another solvent is added in to replace it. This
operation would be a natural fit for GC analysis because these
solvents might be transparent to a UV detector, and GC can provide
fast and accurate analysis.
In industrial processes, reaction products can also be
monitored by GC methods. For example, in the pharmaceutical
industry, most commercial drug manufacturing processes require
356 Gas Chromatography Method Development

methodology that monitors for reaction conversion at various

points along the chemical synthesis. GC has also been applied in
thermodynamic studies of complexation reactions. For example,
Skorka and co-workers found that chiral GC columns were
useful for studying complexes of monoterpenes by alpha- and
beta-cyclodextrins and experimentally determined distribution
constants, stability constants and thermodynamic parameters
[57]. In another study, Fiamegos and co-workers monitored the
reaction of amino acids with a derivatization-extraction approach.
The amino acids were converted to N(O,S)-ethoxycarbonyl amino
acid ethyl esters followed by a single-drop micro extraction
technique. By using GC-FID and GC-MS, detection limits as low
as 0.01 µg/mL and 0.26 ng/mL were achieved, respectively.
The kinetics of the reaction were studied by using GC data, which
allowed the authors to gain insight into the progress of the
reaction as shown in Fig. 12.15 [58].

Figure 12.15 Plot of the progress of an amino derivatization reaction using

GC data. The ratio of the amino acid derivatives peak height to internal
standard (n-Pentadecane) peak height was plotted over time. Adapted
from Fiamegos and co-workers [58].

12.7.5 Mutagenic Impurities (MIs)

MIs, which are synonymous with genotoxic impurities (GTIs),
are known to cause mutations of genes or rearrangements of
chromosomes. MIs are impurities that are also potentially
carcinogenic to humans [59], and adequate control of MIs
 Considerations for Using GC in the Pharmaceutical Setting 357

by pharmaceutical companies in their products has received

heightened scrutiny from regulators over recent years. This trend
has led to increased interest in analytical methods capable of
low detection limits for MIs. Further, both MIs and impurities
that are potentially MIs need to be addressed, which opens
up potentially hundreds of compounds to scrutiny during the
development of a drug candidate.
A typical workflow for addressing MIs and potential MIs
during the development of a drug would involve an in silico, expert
rule-based, and orthogonal statistical model in order to identify
impurities that are known MIs or potential MIs. After screening
up to hundreds of known and potential impurities, intermediates,
reagents, by-products, and starting materials, a list of known and
potential MIs would be compiled. Then, pharmaceutical companies
would address these alerting structures by a variety of mechanisms
such as Ames testing to prove or disprove mutagenicity, lab-based
process spike experiments to demonstrate adequate purge of
an impurity, or scientific justification based on process rejection
capability such as removal of impurities during crystallizations,
extractions, or reactivity arguments, for example.
For MIs and potential MIs that cannot be addressed adequately
by the previous mechanisms, then analytical testing is often a
last-line but highly important means to address these impurities.
Analytical testing is not preferred because of the high cost and
time commitment associated with specialty, targeted impurity
methods. However, when MIs and potential MIs need to be tested
by analytical methods, GC is considered the standard method of
choice for analysis of small molecule mutagenic impurities (MIs)
[60], and recent advances in GC for pharmaceutical applications
have seen breakthroughs in these types of analysis. LC-MS should
also be considered for MI analysis, such as when analytes are not
susceptible to GC analysis or due to instrument limitation in a
laboratory. A decision tree for when to use LC-MS and GC-MS based
methods for MI analysis can be helpful in this regard [61]. Lastly,
the importance of having appropriate analytical methods with
low enough detection limits for the testing of MIs and potential
MIs cannot be understated due to patient safety, and the case of the
withdrawal of Viracept (Nefinavir mesilate) in 2007 is an example
where elevated levels of a known MI, ethyl methanesulfonate,
were the cause of the drug withdrawal from the market [62].
358 Gas Chromatography Method Development

Based on current International Conference on Harmonisation

(ICH) guidelines for MIs, the maximum allowed MI in a drug
substance for a lifetime of dosing is 1.5 µg per day. For a hypothetical
2 g per day drug dose, this is an allowable amount of 0.75 ppm
in the drug substance. As such, it is easy to see why GC is a
preferred technique due to its wide array of column stationary
phases available, headspace format, and sensitive detectors that
are at the analyst’s disposal. A range of stationary phases allows
the user to select a column with good selectivity whereby the
MI is well resolved from main component and any potentially
interfering matrix components. Further, headspace format allows
for sensitive analysis with reduced potential for matrix interference.
In order to achieve good method sensitivity toward a MI, high
sample concentrations are often employed (e.g., 100 mg/mL). As
a result, direct injection can become problematic as sample builds
up on the inlet of the GC, which can lead to issues with accuracy,
tailing, extraneous peaks, and deteriorated detection limits.
Headspace-GC solves many of these challenges. And lastly, a
myriad of detectors in the GC realm allows the user flexibility to
select an appropriate detection scheme for a required challenge.
For GC analysis of MIs, a GC-FID setup is a good starting point due
to simplicity and ease of use in many labs. However, for low-level
detection needs, such as analysis in the range of lower ppm and
below, typically more sensitive detectors such as TCD or GC-MS
are more appropriate due to lower levels of quantitation (LOQ)
achievable compared to GC-FID. Each MI analysis need is different
and depends on the daily dose of the drug, the duration of dose, as
well as the analyte at hand and the matrix in which it is present.
For compounds that are not volatile enough to be amenable
to GC analysis, or in order to improve analysis, derivatization
approaches have been successful to afford GC analysis with useful
LOQs for MIs [63]. For example, pentafluorothiophenol derivatives
have been used by Alzaga and co-workers [64] in order to afford
some common alkylating agents analysis by headspace GC with
improved analysis time, LOQ, and analyte stability. Figure 12.16
shows various PFTP derivatives for the headspace GC-MS analysis
of various carbon chain length alkyl tosylates. Good LOQ is reached
at 1 ppm with impressive signal-to-noise. Another major advantage
of this approach is the headspace setup, whereby potentially
interfering sample matrix components are eliminated from GC
analysis through volatility discrimination.
 Considerations for Using GC in the Pharmaceutical Setting 359

Figure 12.16 HS-GC-MS selected ion chromatograms showing the

PFTP derivatives: (a) methyl (m/z+ = 214); (b) ethyl (m/z+ = 228);
(c) isopropyl (m/z+ = 242), and their corresponding deuterated IS’s (a)
(m/z+ = 217), (b) (m/z+ = 233) and (c) (m/z+ = 249). Selected analyte
was alkyl tosylate at 1.0 µg g−1 (1 ppm) relative to 50 mg of sample [64].
360 Gas Chromatography Method Development

Other strategies for MI analysis have employed GC-HS with an

ECD for detection of halogenated MIs. By using ionic liquids as
the diluent, Ho and coworkers were able to show LODs from
5–500 ppm for two MI classes: alkyl/aryl halides and nitro-
aromatics. These MIs are relatively high boilers, and by employing
ionic liquid-based diluent, the HS oven temperature could be
increased and tuned to allow for low-level detection compared
with a conventional DMSO diluent, for example [65]. In another
interesting study around GC for the quantitation of halogenated
MIs, Byun and coworkers used electron impact MS to study
halogenated cytosines and deoxycytidines. GC-EI-MS was compared
to GC with fast atom bombardment and electrospray (ESI)-MS
along with liquid chromatography-ESI. In this study, GC-EI-MS
was determined to be superior for low-level detection of these
halogenated MIs due to low sensitivity and wide dynamic range [66].
Outside of the pharmaceutical industry, Kataoka and coworkers
have used GC along with a nitrogen-phosphorous detector (NPD)
to detect mutagenic heterocyclic amines (HCAs). These MIs are
created during heating of protein-based foods through boiling,
frying, or grilling, and they can be formed in meats or fishes. The
authors used a derivatization approach to afford GC analysis, and
found that HCAs were present in higher amounts in certain fishes
compared with meats [67]. In another example of innovative
approaches to MI analysis by GC, David and coworkers employed
a two-dimensional GC system with MS detection. The heart-cut MI
fraction coming off of a 5% diphenyl-based column was transferred
to a more polar 50% diphenyl column, whereby the MI fraction
could further separate without interference from drug substance,
solvents or derivatizing agents that afforded the GC separation.
Michael acceptors and haloalcohols were quantitated with LODs
near the 0.01–0.1 ppm range [68].

12.8 Final Remarks

GC is a powerful analytical tool with a variety of uses across the
pharmaceutical industry and beyond. The range of columns,
detectors, modes of operation, and sample preparation techniques
all contribute to the overall versatility of this indispensable
equipment. Optimization of these parameters enables GC to be an
 References 361

essential means for rapid determination of a wide variety of small

molecules, with superb sensitivity and selectivity.
Within the pharmaceutical realm, GC has been well utilized
for residual solvent analysis, and exploited for determining
potent/toxic molecules, such as MIs, at trace and ultra-trace levels.
Recently, applications in the pharmaceutical industry have also
expanded towards assay determinations, impurities and related
substance characterizations, and in-process testing/monitoring.
As innovations in GC continue to cultivate and develop within the
pharmaceutical world, it is critical that we take full-advantage of
these progressions to guarantee quality in our products, minimize
supply chain risk, and ultimately ensure patient safety.

References

1. Hinshaw, J. V. (2015) Split-vent traps: GC’s dirty secret, LCGC N. Am.,

33, pp. 758–763.
2. Snow, N. H. (2004) Inlet systems for gas chromatography, in
Modern Practice of Gas Chromatography (Grob, R. L., and Barry, E. F.,
eds.), John Wiley & Sons, Inc., Hoboken, New Jersey, pp. 461–489.
3. Watson, D. W. (2016) How to optimize key variables in GC analysis:
Sample introduction, LCGC N. Am., 34, p. 446.
4. Hinshaw, J. V. (2004) Optimization of separations and computer
assistance, in Modern Practice of Gas Chromatography (Grob, R. L.,
and Barry, E. F., eds.), John Wiley & Sons, Inc., Hoboken, New Jersey,
pp. 193–228.
5. Hinshaw, J. V. (2001) Selecting carrier gases and conditions, LCGC N.
Am., 19, pp. 1056–1064.
6. Bartram, R. J. (2004) Gas management systems for gas chromato-
graphy, in Modern Practice of Gas Chromatography (Grob, R. L., and
Barry, E. F., eds.), John Wiley & Sons, Inc., Hoboken, New Jersey,
pp. 491–543.
7. Watson, D. W. (2017) Advantages of using nitrogen in capillary GC,
LCGC N. Am., 35, p. 142.
8. Barry, E. F. (2004) Columns: Packed and capillary; column selection
in gas chromatography, in Modern Practice of Gas Chromatography
(Grob, R. L., and Barry, E. F., eds.), John Wiley & Sons, Inc., Hoboken,
New Jersey, pp. 65–191.
362 Gas Chromatography Method Development

9. Sun, Y., Zhang, R., and Wang, Q. (1993) Programmed-temperature gas

chromatographic retention index, J. Chromatogr. A, 657, pp. 1–15.
10. Cavalli, E. J., and Guinchard, C. (1995) Emperical approach to the
dependence of plate height on temperature and gas velocity in gas
chromatography, J. Chromatogr. Sci., 33, pp. 573–577.
11. Blumberg, L. M. (2012) Theory of gas chromatography, in Gas
chromatography (Poole, C. F., ed.), Elsevier, Oxford, UK.
12. Blumberg, L. M. (2010) Temperature-Programmed Gas Chromato-
graphy, Wiley-VCH, Weinheim, Germany.
13. Castello, G., Moretti, P., and Vezzani, S. (2009) Retention models
for programmed gas chromatography, J. Chromatogr. A, 1216,
pp. 1607–1623.
14. Blumberg, L. M., and Klee, M. S. (2001) Quantitative comparison
of performance of isothermal and temperature-programmed gas
chromatography, J. Chromatogr. A, 933, pp. 13–26.
15. Giddings, J. C. (1962) Elementary theory of programmed temperature
gas chromatography, J. Chem. Ed., 39, pp. 569–573.
16. Taylor, T. (2015) GC temperature programming-10 things you
absolutely need to know, LCGC N. Am., 33, p. 438.
17. Hinshaw, J. V. (1991) Programmed-temperature optimization, LCGC, 9,
pp. 470–473.
18. Korytár, P., Janssen, H.-G., Matisová, E., and Brinkman, U. A. T. (2002)
Practical fast gas chromatography: Methods, instrumentation and
applications, TrAC, Trends Anal. Chem., 21, pp. 558–572.
19. Matisová, E., and Dömötörová, M. (2003) Fast gas chromatography
and its use in trace analysis, J. Chromatogr. A, 1000, pp. 199–221.
20. David, F., Gere, D. R., Scanlan, F., and Sandra, P. (1999) Instrumentation
and applications of fast high-resolution capillary gas chromatography,
J. Chromatogr. A, 842, pp. 309–319.
21. Klee, M. S., and Blumberg, L. M. (2002) Theoretical and practical
aspects of fast gas chromatography and method translation, J.
Chromatogr. Sci., 40, pp. 234–247.
22. Cramers, C. A., and Leclercq, P. A. (1999) Strategies for speed
optimisation in gas chromatography: An overview, J. Chromatogr. A,
842, pp. 3–13.
23. Snow, N. H. (2004) Fast gas chromatography with short columns:
Are speed and resolution mutually exclusive?, J. Liq. Chromatogr.
Related. Technol., 27, pp. 1317–1330.
 References 363

24. Jumtee, K., Bamba, T., and Fukusaki, E. (2009) Fast GC-FID based
metabolic fingerprinting of Japanese green tea leaf for its quality
ranking prediction, J. Sep. Sci., 32, pp. 2296–2304.
25. Yan, X. (2006) Unique selective detectors for gas chromatography:
Nitrogen and sulfur chemiluminescence detectors, J. Sep. Sci., 29,
pp. 1931–1945.
26. Ragunathan, N., Krock, K. A., Klawun, C., Sasaki, T. A., and Wilkins,
C. L. (1999) Gas chromatography with spectroscopic detectors, J.
Chromatogr. A, 856, pp. 349–397.
27. Poole, C. F. (2015) Ionization-based detectors for gas chromatography,
J. Chromatogr. A, 1421, pp. 137–153.
28. Gunnar, T., Mykkänen, S., Ariniemi, K., and Lillsunde, P. (2004)
Validated semiquantitative/quantitative screening of 51 drugs
in whole blood as silylated derivatives by gas chromatography-
selected ion monitoring mass spectrometry and gas chromatography
electron capture detection, J. Chromatogr. B, 806, pp. 205–219.
29. Ho, T. D., Yehl, P. M., Chetwyn, N. P., Wang, J., Anderson, J. L., and
Zhong, Q. (2014) Determination of trace level genotoxic impurities
in small molecule drug substances using conventional headspace
gas chromatography with contemporary ionic liquid diluents and
electron capture detection, J. Chromatogr. A, 1361, pp. 217–228.
30. Kosjek, T., Heath, E., Petrović, M., and Barceló, D. (2007) Mass
spectrometry for identifying pharmaceutical biotransformation
products in the environment, TrAC, Trends Anal. Chem., 26,
pp. 1076–1085.
31. David, F., Alegre, M. R., Vanhoenacker, G., and Sandra, P. (2013)
Determination of genotoxic impurities in pharmaceuticals, LCGC Eur.,
26, pp. 31–34.
32. Girijan, R. (2012) Gas chromatography-mass spectrometry
technique and its applications in pharmaceutical industry, Chem.
Ind. Dig., 25, pp. 59–63.
33. Gillespie, T. A., and Winger, B. E. (2011) Mass spectrometry for
small molecule pharmaceutical product development: A review,
Mass Spectrom. Rev., 30, pp. 479–490.
34. Ji, C., Dean, J. L., Flores, R. A., and Boisvert, S. M. (2009) Quantitative
determination of co-eluting compounds by using extract ion
chromatogram in GC-MS, Chem. Educator., 14, pp. 212–214.
35. Xue, M., Yuan, S., Ruan, J., Qiao, J., Xu, Y., Zhang, Z., and Liu, K. (2006)
Determination of penehyclidine by gas chromatographic-mass
364 Gas Chromatography Method Development

spectrometry and its application to pharmacokinetics in humans,

rabbits and mice, J. Chromatogr. B, 843, pp. 234–239.
36. Munson, B. (2000) Development of chemical ionization mass
spectrometry, Int. J. Mass Spectrom., 200, pp. 243–251.
37. Munson, M. S. B., and Field, F. H. (1966) Chemical ionization mass
spectrometry, J. Am. Chem. Soc., 88, pp. 2621–2630.
38. Hunt, D. F., and Crow, F. W. (1978) Electron capture negative
ion chemical ionization mass spectrometry, Anal. Chem., 50,
pp. 1781–1784.
39. Stafford, G. C. (1980) Instrumental aspects of positive and negative
ion chemical ionization mass spectrometry, Environ. Health Perspect.,
36, pp. 85–96.
40. Vaclavik, L., Krynitsky, A. J., and Rader, J. I. (2014) Mass spectrometric
analysis of pharmaceutical adulterants in products labeled as
botanical dietary supplements or herbal remedies: A review, Anal.
Bioanal. Chem., 406, pp. 6767–6790.
41. Loda, C., Bernabe, E., Nicoletti, A., Bachi, S., and Dams, R. (2011)
Determination of epichlorohydrin in active pharmaceutical
ingredients by gas chromatography-mass spectrometry, Org. Proc.
Res. Dev., 15, pp. 1388–1391.
42. Taylor, T. (2015) Understanding electron ionization processes for
GC-MS, LCGC N. Am., 33, p. 290.
43. Belas, F. J., and Blair, I. A. (2002) Mass spectrometry in pharmaceutical
analysis, in Handbook of Pharmaceutical Analysis (Ohannesian, L.,
and Streeter, A. J., eds.), Marcel Dekker, Inc., New York, New York,
pp. 151–185.
44. Mclafferty, F. W. (2009) Registry of Mass Spectral Data and Registry
Combined with NIST, 9th ed., Wiley-Blackwell, Hoboken, NJ.
45. Mclafferty, F. W. (2011) A century of progress in molecular mass
spectrometry, Ann. Rev. Anal. Chem., 4, pp. 1–22.
46. Little, J. L., Cleven, C. D., and Howard, A. S. (2013) Identifying “known
unknowns” in commercial products by mass spectrometry, LCGC N.
Am., 31, pp. 114–125.
47. Sparkman, D. O., Penton, Z. E., and Kitson, F. G. (2011) Mass spectral
data interpretation, in Gas Chromatography and Mass Spectrometry:
A Practical Guide, Elsevier, Oxford, UK, pp. 149–205.
48. Mclafferty, F. W., and Tureček, F. (1993) Interpretation of Mass
Spectra, University Science Books, Sausalito, CA.
49. Masucci, J. A., and Caldwell, G. W. (2004) Techniques for gas
chromatography/mass spectrometry, in Modern Practices of Gas
 References 365

Chromatography (Grob, R. L., and Barry, E. F., eds.), John Wiley & Sons,
Inc., Hoboken, New Jersey, pp. 339–401.
50. Smith, R. M. (2004) Understanding Mass Spectra: A Basic Approach,
2nd ed., John Wiley & Sons, Inc., Hoboken, New Jersey.
51. Zhang, L., Tang, C., Cao, D., Zeng, Y., Tan, B., Zeng, M., Fan, W.,
Xiao, H., and Liang, Y. (2013) Strategies for structure elucidation
of small molecules using gas chromatography-mass spectrometric
data, TrAC, Trends Anal. Chem., 47, pp. 37–45.
52. Hamilton, S. E., Rossington, M. D., and Bertrand, A. (2016) Development
of an automated headspace gas chromatography instrument for the
determination of residual solvents in pharmaceutical compounds
and reaction mixtures, Org. Proc. Res. Dev., 20, pp. 189–194.
53. Nacham, O., Ho, T. D., Anderson, J. L., and Webster, G. K. (2017)
Use of ionic liquids as headspace gas chromatography diluents for
the analysis of residual solvents in pharmaceuticals, J. Pharm. Biomed.
Anal., 145, pp. 879–886.
54. Belal, T. S., Awad, T., and Clark, C. R. (2014) Stability-indicating
determination of trimetazidine dihydrochloride in the presence of
two of its related substances using a direct GC/MS method, J. AOAC
Int., 97, pp. 1514–1518.
55. Belal, T. S., Abdel-Hay, K. M., and Clark, C. R. (2016) Selective
determination of dimenhydrinate in presence of six of its related
substances and potential impurities using a direct GC/MS method,
J. Adv. Res., 7, pp. 53–58.
56. van den Berg, J. J. M., Winterbourn, C. C., and Kuypers, F. A. (1993)
Hypochlorous acid-mediated modification of cholesterol and
phospholipid: Analysis of reaction products by gas chromatography-
mass spectrometry, J. Lipid Res., 34, pp. 2005–2012.
57. Skórka, M., Asztemborska, M., and Żukowski, J. (2005) Thermodynamic
studies of complexation and enantiorecognition processes of
monoterpenoids by α- and β-cyclodextrin in gas chromatography,
J. Chromatogr. A, 1078, pp. 136–143.
58. Fiamegos, Y. C., Nanos, C. G., and Stalikas, C. D. (2004) Ultrasonic-
assisted derivatization reaction of amino acids prior to their
determination in urine by using single-drop microextraction in
conjunction with gas chromatography, J. Chromatogr. B, 813,
pp. 89–94.
59. McGovern, T., and Jacobson-Kram, D. (2006) Regulation of genotoxic
and carcinogenic impurities in drug substances and products,
TrAC, Trends Anal. Chem., 25, pp. 790–795.
366 Gas Chromatography Method Development

60. Reddy, A. V. B., Jaafar, J., Umar, K., Majid, Z. A., Aris, A. B., Talib, J.,
and Madhavi, G. (2015) Identification, control strategies, and
analytical approaches for the determination of potential genotoxic
impurities in pharmaceuticals: A comprehensive review, J. Sep. Sci.,
38, pp. 764–779.
61. David, F., Alegre, M. R., Vanhoenacker, G., and Sandra, P. (2013)
Determination of genotoxic impurities in pharmaceuticals, LCGC Eur.,
pp. 31–34.
62. Elder, D. P., Delaney, E., Teasdale, A., Eyley, S., Reif, V. D., Jacq, K.,
Facchine, K. L., Oestrich, R. S., Sandra, P., and David, F. (2010) The
utility of sulfonate salts in drug development, J. Pharm. Sci., 99,
pp. 2948–2961.
63. Liu, D. Q., Sun, M., and Kord, A. S. (2010) Recent advances in trace
analysis of pharmaceutical genotoxic impurities, J. Pharm. Biomed.
Anal., 51, pp. 999–1014.
64. Alzaga, R., Ryan, R. W., Taylor-Worth, K., Lipczynski, A. M., Szucs, R.,
and Sandra, P. (2007) A generic approach for the determination
of residues of alkylating agents in active pharmaceutical ingredients
by in situ derivatization–headspace–gas chromatography–mass
spectrometry, J. Pharm. Biomed. Anal., 45, pp. 472–479.
65. Ho, T. D., Joshi, M. D., Silver, M. A., and Anderson, J. L. (2012)
Selective extraction of genotoxic impurities and structurally alerting
compounds using polymeric ionic liquid sorbent coatings in solid-
phase microextraction: Alkyl halides and aromatics, J. Chromatogr. A,
1240, pp. 29–44.
66. Byun, J., Henderson, J. P., and Heinecke, J. W. (2003) Identification
and quantification of mutagenic halogenated cytosines by gas
chromatography, fast atom bombardment, and electrospray ionization
tandem mass spectrometry, Anal. Biochem., 317, pp. 201–209.
67. Kataoka, H., Nishioka, S., Kobayashi, M., Hanaoka, T., and Tsugane,
S. (2002) Analysis of mutagenic heterocyclic amines in cooked food
samples by gas chromatography with nitrogen-phosphorus detector,
Bull. Environ. Contam. Toxicol., 69, pp. 682–689.
68. David, F., Jacq, K., Sandra, P., Baker, A., and Klee, M. S. (2010) Analysis
of potential genotoxic impurities in pharmaceuticals by two-
dimensional gas chromatography with Deans switching and
independent column temperature control using a low-thermal-mass
oven module, Anal. Bioanal. Chem., 396, pp. 1291–1300.
Chapter 13

Method Transfer between HPLC and

UHPLC

Susanne Fabel
Thermo Fisher Scientific, Dornierstr. 4, D-82110 Germering, Germany
[email protected]

13.1 Introduction
The method transfer between conventional HPLC and UHPLC is
done either to accelerate the method or to gain more resolution
compared to the HPLC method. The opposite transfer from
UHPLC to HPLC could also be a requirement, for example, if in
the R&D department only UHPLC equipment is available and the
methods have to be transferred to HPLC equipment in the QC
laboratory. Conveniently, method transfer for UHPLC to HPLC can
be achieved by applying the same rules and principles for HPLC
to UHPLC methods. In this chapter, the focus is on the method
transfer from conventional HPLC to UHPLC. The theory behind

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
368 Method Transfer between HPLC and UHPLC

the method transfer from HPLC to UHPLC is explained along with

instrument tips and tricks.

13.2 Method Transfer Strategy

The goals for HPLC method transfer to an UHPLC method can
focus on either maintaining or improving the HPLC separation
efficiency, or accelerating the HPLC method to increase sample
throughput (Fig. 13.1).

Figure 13.1 The focus of method transfer from HPLC to an UHPLC method
can be on high resolution for high peak capacity (left, up to 1000 peaks
separated in a single run) or on high speed for high sample throughput
(right, 10 peaks separated in 10 s). Reproduced with permission of
Thermo Fisher Scientific.

Theoretical HPLC to UHPLC calculations allow you to either

improve efficiency of the separation within the same analysis time
or shorten the analysis time while maintaining the same efficiency.
Compared to HPLC separations, UHPLC methods are more lenient
with method acceleration due to the ability to maintain resolution.
However, the trade-off for improved efficiency or throughput is
higher back pressure.
 Method Transfer Strategy 369

13.2.1 Geometric Rules for Transfer from HPLC to

UHPLC
Method transfer calculations differ between isocratic and gradient
separations. As a start, the basic geometric rules applying to
isocratic separations are discussed in this section. The arithmetical
continuation of the rules to gradient methods is discussed later
in Section 13.2.4.
One motivation for the transfer from an HPLC to UHPLC
method is to obtain better separation of the analytes. The quality
of a separation between two compounds is described with the
Resolution R of two neighboring eluting compounds, described
by the Purnell equations Eq. 13.1 [1], where N is the number
of theoretical plates, k the retention factor of the second peak,
and a the separation factor between two peaks.

1 k a –1
R =  N 2  (13.1)
4 1 + k2 a

The separation factor of two adjacent peaks is defined by

their k-values:

k2
a= (13.2)
k1

The separation factor a is influenced by column chemistry,

mobile phase chemistry and temperature. The retention factor k
is primarily influenced by mobile phase strength and temperature.
The number of theoretical plates N, which is also defined as the
efficiency factor, refers to the column conditions like length,
particle size, and flow rate. An increase of the efficiency by a factor
of 2 would generate a 1.4-fold increase in resolution due to the
square root dependency.
The number of theoretical plates N is correlated with the
length of the column L and the height equivalent to a theoretical
plate height H as described in Eq. 13.3.

L
N= (13.3)
H
370 Method Transfer between HPLC and UHPLC

Therefore, by keeping the column length constant, lower height

equivalents will generate a higher efficiency of the separation
and a better resolution. The factors influencing the theoretical
plate height are explained by considering the diffusion processes
during the separation in the column, which are expressed by the
Van Deemter equation (Eq. 13.4).

B
H = A+ + C u (13.4)
u

The A-term, eddy dispersion, describes the mobile phase

movement along random paths through the stationary phase
resulting in band broadening of the analyte. The longitudinal
diffusion of the analyte in and against the flow direction is expressed
by the B-term. The resistance of the analyte to the mass transfer
into the pores of the stationary phase is described by the C-term.
u is the linear velocity of the mobile phase in the column. The
connection between plate height H and particle size dp can be
simplified according to Halász [2]. This approximation is valid
for fully porous material.
2
6 dp u (13.5)
H = 2dp + +
u 20

The Van Deemter plots of plate height H against velocity u,

depending on the particle size dp, help visualize the role of small
particle sizes in UHPLC method transfer strategy (see Fig. 13.2).
The optimum flow rate and minimum achievable plate
height can be estimated by setting the first derivative of the Eq. 13.5
to zero. The optimal linear velocity uopt (in mm/s) can then be
estimated by Eq. 13.6.

10
uopt  (13.6)
dp

The minimum achievable plate height Hmin is subsequently

calculated by insertion of Eq. 13.6 in Eq. 13.5.

H min  3dp (13.7)

 Method Transfer Strategy 371

Figure 13.2 Smaller particles provide more theoretical plates as

demonstrated by the smaller minimum plate height H in the Van Deemter
plot and considering the relation of N and H in Eq. 13.3. Reproduced
with permission of Thermo Fisher Scientific.

The factor 3 in Eq. 13.7 is a conservative calculation. Excellent

packed modern UHPLC columns operated in a modern UHPLC
system could generate plate heights proportional to a factor 2 of
the particle size. With Eq. 13.7 it can be concluded that smaller
particles generate lower plate heights and higher efficiencies
according to Eq. 13.3 and better resolutions according to Eq. 13.1.
This is visualized in a simulated chromatogram in Fig. 13.3 for 5
and 2 µm particle size.
Following Eq. 13.3, shorter columns with smaller particles can
be used while maintaining efficiency. The use of shorter column
allows for analysis times to be shortened—achieving the second
motivating factor for UHPLC method, faster separations.
The analysis time is proportional to the column dead time,
which is proportional to the column length. Moving from an HPLC
method with column length L1, particle size dp1 and analysis time
372 Method Transfer between HPLC and UHPLC

t1, the calculated analysis time of the UHPLC method t2 with new
column parameters is described by Eq. 13.8.

Figure 13.3 The simulated chromatograms show that smaller particles,

under the same separation conditions, provide improved resolution.
Reproduced with permission of Thermo Fisher Scientific.

dp,2 L2
t 2 = t1   (13.8)
dp,1 L1

In addition to saving time, reducing solvent consumption is

another strong motivator for transferring conventional methods
to UHPLC. Solvent consumption is reduced by using shorter
columns with smaller void volumes and shorter analysis time.
Additional solvent can be conserved by scaling down the column
inner diameter. However, the linear velocity must be kept
constant to operate with the optimum Van Deemter conditions.
Subsequently, the flow rate has to be decreased with the new
geometric cross section of the column. By using smaller inner
diameter dc and shorter length L columns, the reduction in
solvent consumption can be calculated by Eq. 13.9.
2
dc,2 L
V2 = V1  2
 2 (13.9)
dc,1 L1
 Method Transfer Strategy 373

Another benefit of transferring a conventional method to

UHPLC is the ability to gain more mass sensitivity or consume
less sample by using more efficient column material and scaling
the injection volume accordingly, which will be discussed later in
Section 13.2.2.4.
While there are benefits to transferring an HPLC method to
UHPLC (i.e., improvements in efficiency or a decrease in analysis
time), there are additional factors that must be considered such
as the increase in back pressure in UHPLC. The pressure drop
across a column filled with porous medium is described by
Darcy’s law (Eq. 13.10) as a function of the linear velocity u,
viscosity η, length L, porosity Ф, and particles size dp.

uLÖ (13.10)
P =
dp2
The pressure drop generated by different particle sizes is
shown in Fig. 13.4. Operating a column filled with 5 µm particles
at 160 bar pressure drop would mean that the same column
filled with 2 µm particles, would in theory generate 1000 bar
pressure drop assuming that the other parameters like linear
velocity, column length, viscosity, and porosity are kept constant.
The optimum flow rate for HPLC to UHPLC method transfer
is influenced by the particle size (see Eq. 13.6). This dependency,
combined with Darcy’s law, allows the conclusion that the added
back pressure is inversely proportional by the third power of
particle size and directly related to the column length. The
added pressure drop for the method transfer from HPLC column
formats to UHPLC column formats can be calculated by Eq. 13.11.
3
dp,1 L
P2 = P1  3
 2 (13.11)
dp,2 L1

UHPLC instrumentation operating with smaller particles

has to be capable of operating at higher pressures.
Comparison of theoretical plates, savings in analysis time and
solvent consumption, and the expected back pressure increase by
changing the method from a conventional 4.6 × 150 mm column
filled with 5 µm particle to various UHPLC column formats is
shown in Table 13.1.
374 Method Transfer between HPLC and UHPLC

Figure 13.4 The simulated pressure drop on the column rises with inverse
square of the particle size according to Eq. 13.10 demonstrates the
trade-off for UHPLC separation targeting higher peak capacity or higher
analysis speed. Reproduced with permission of Thermo Fisher Scientific.

Table 13.1 Estimated efficiency, time saving, solvent reduction, and back
pressure increase for method transfer from HPLC method to UHPLC
method for different column geometries generating comparable efficiencies
and resolution

Column format Back

Particle Relative Reduction pressure
Length ID size Theoretical time in solvent increase
[mm] [mm] [µm] plates N saving consumption factor
150 4.6 5 10000 — — —
100 4.6 3 11111 60% 33% 3.1
75 4.6 2.2 11363 78% 50% 5.9
50 4.6 2.2 7575 85% 67% 3.9
50 4.6 1.5 11111 90% 67% 12.3
150 2.1 5 10000 0% 79% 1.0
100 2.1 3 11111 60% 86% 3.1
75 2.1 2.2 11363 78% 90% 5.9
50 2.1 1.5 11111 90% 93% 12.3
 Method Transfer Strategy 375

13.2.2 System Requirements

The instrument and column design as well as method details
like mobile phase, flow rate, and injection volume have to fulfill
certain requirements to accomplish an HPLC to UHPLC method
transfer by geometric rules.

13.2.2.1 Column choice

For successful HPLC to UHPLC method transfer, an adequate
column packed with smaller particles has to be carefully selected.
The column chemistry and selectivity of the stationary phase
should be maintained to get the closest prediction of the UHPLC
method.
Numerous tools can be found from column manufactures
and on the internet to get the most appropriate UHPLC column to
match the original HPLC column’s characteristics in hydrophobicity,
steric hindrance, hydrogen-bonding acidity, hydrogen-bonding
basicity, and cation-exchange capacity. An overview of 300 reversed-
phase columns by dedicated selectivity parameters has been
described by Snyder et al. [3]. If no selectivity equivalent UHPLC
column can be found, simple geometric method transfer rules
may not be sufficient to predict the method transfer.
The preferable dimension for the UHPLC column should be 3
or 2.1 mm. Inner diameters of 4.6 mm filled with small particles
are not recommended for UHPLC columns due to the high flow
rate needed to operate at the optimal linear velocity and the
resulting heating effects (see Section 1.4.1). Smaller inner diameters
than 2.1 mm need special care for minimizing the extra column
volume (see Section 1.2.2.4) and might disrupt the HPLC to
UHPLC method transfer by simply applying geometric transfer
rules.
The careful consideration of column length and suitable particle
size is required to meet the column efficiency needed for the
separation [4, 5].
If the analysis time is kept constant and the resolution of
a critical peak pair should be doubled, then the efficiency of
the column must be increased fourfold, due to the square root
dependency of efficiency and resolution (Eq. 13.1). If the objective
is to improve efficiency by reducing the particle size, then the
particle size must be reduced fourfold (Eq. 13.3). Keeping the flow
rate and column length constant results in a pressure drop that is
376 Method Transfer between HPLC and UHPLC

elevated 16-fold (Eq. 13.10), possibly succeeding the pressure of

the UHPLC instrumentation.
Another way to double the resolution is to increase the column
length fourfold; however, this would also increase the analysis
time four-fold, which is not wanted in most cases. Depending
on the analytes, separation efficiency could also be increased
by using high temperature chromatography or changing the
stationary phase morphology [6]. However, application of the
geometric method transfer rules to predict separations using high
temperature chromatography or adjustment of the stationary phase
morphology is difficult and beyond the scope of this chapter. The
combined information on separation quality and analysis time,
including information on permeability for different stationary
phase morphologies and geometries, can be combined to kinetic
plots to help find the most suitable column [7]. In this chapter,
the main focus is on choosing columns with comparable or slightly
better efficiency by using smaller particles to accelerate the
chromatographic method (see Table 13.1).

13.2.2.2 Mobile phase

For UHPLC methods, care has to be taken in the preparation
of the mobile phase to avoid clogging of the flow path. Smaller
particles in the UHPLC column require smaller frits than those
for the conventional HPLC columns, making UHPLC setups more
susceptible to clogging. HPLC columns generally have inlet frits
with a pore size of 2 µm, whereas the inlet frits of UHPLC columns
have a pore size of 0.5 to 0.2 µm for 3 µm and smaller particles. Next
to the column hardware, UHPLC instrument hardware is optimized
for low dispersion connections and therefore more receptive to
particle accumulation than conventional HPLC instrumentation.
Therefore, high-grade organic solvents and high quality water
should be used. Solvents should be filtered through a 0.2 µm
membrane, if it has not been done already by the supplier or the
lab water system. Buffer preparation and mobile phase additives
should also be added in high quality grade to achieve the best
UHPLC performance and robustness. Special care should be taken
to prevent organic growth by the usage of buffers, as it is already
the case for HPLC methods. This becomes even more important
 Method Transfer Strategy 377

for UHPLC methods to avoid problems. Inlet frits in the solvent

supply line of the UHPLC have to be used and changed from time to
time.

13.2.2.3 Flow rate

The mobile phase flow rate should be near the optimal linear
velocity value of the Van Deemter curve. The flow rate is directly
proportional to the square of the column diameter and depends
on the particle size of the stationary phase. A theoretical calculation
of the optimum flow rate is possible, but other parameters like
temperature and viscosity also influence the Van Deemter curve.
A rough estimation of the optimal linear velocity according to
Eq. 13.6 is possible. If that is not sufficient, recording a curve
with experimental data should be completed.
When transferring an HPLC method to an UHPLC method,
prediction of the new UHPLC flow rate F2 can be calculated
considering the original HPLC flow rate F1, the change in column
cross section dc and particle size dp according to Eq. 13.12.

 dc,2 2 dp,1
F2 = F1 
d   (13.12)
 c,1  dp,2

In addition, the flat C-term of the Van Deemter curve allows

for an acceleration of the flow rate without compromising
resolution with smaller particles. The new analysis time is then
calculated by Eq. 13.13.

2
F1 L2  dc,2 
t 2 = t1     (13.13)
F2 L1  
 dc,1 

The increasing back pressure generated by the higher flow

rates limit method acceleration. The back pressure is directly
proportional to the flow rate and the resulting analysis time is
indirectly proportional to the pressure drop. With a system capable
of 1500 bar, it is in theory possible to accelerate a 400 bar method
3.75-fold.
378 Method Transfer between HPLC and UHPLC

13.2.2.4 Injection volume

Compared to longer columns, shorter columns with smaller particles
cause a reduced zone dilution, and consequently sharper peaks.
However, an equivalent mass loading onto columns with smaller
inner diameters is recommended for method transfer. Scaling
down the injection plug adjusting to the change in the column
cross section avoids mass overloading and peak dispersion.
The new injection volume Vinj,2 can be calculated by taking into
account the changed cross section and reduced band broadening
by changing the particle diameter and column length according
to Eq. 13.14.

 dc,2 2 L2 dp,2
Vinj, 2 = Vinj,1 
d   (13.14)
 c,1  L1 dp,1

By adjusting the injection volume to load the smaller column

with a similar mass, sensitivity should be consistent in theory.
In praxis, the sensitivity could differ by changes in the HPLC
and UHPLC instrumentation like improved detector technology.
For example for UV detection, using the fiber-optics technology
compared to a normal flow-cell design, leads to sensitivity
increases.

13.2.2.5 Extra-column volume

The extra-column volume plays an important role for the HPLC to
UHPLC method transfer. The extra-column volume contributes
to band broadening of the analyte peak and is considered the
volume in the system through which the sample passes except
the column volume (Fig. 4.5). With the reduction of the column
internal volume by shorter lengths and smaller inner diameters,
the extra-column volume should be reduced proportionally to
limit band broadening effects and to maintain the efficiency of the
UHPLC stationary phase [8].
The total band broadening of the analyte peak is the sum
of the band broadening caused by the column and the extra-
column volume. Factors contributing to the extra-column volume
are the injection, the capillary connection and the detector flow
cell volume. The instrumentation has to be optimized to generate
 Method Transfer Strategy 379

the separation efficiency capable with small-particle stationary

phases [9].

Figure 13.5 The extra-column volume contributes to the band broadening

of the analyte peak. Extra-column volume should be minimized to exploit
the maximum possible efficiency of UHPLC columns. Reproduced with
permission of Thermo Fisher Scientific.

The injection volume should be scaled according to the

geometric rules as discussed above (see Eq. 13.14). As a rule
of thumb, the injection volume should not exceed 1–5% of the
column volume.
The capillary connections should be dead-volume free,
the length and the radius of the tubing should be kept as small
as possible to minimize the extra-column band broadening. A
compromise between generated back pressure and low volume
of the connecting capillaries has to be found (see Table 13.2).
For particles smaller than 2 µm, 0.1 mm inner diameter
capillaries should be used, and for particles smaller than 3 µm
capillaries with 0.13 mm inner diameter are appropriate for
UHPLC methods. The back pressure generated by the column
capillary inner diameter and other fluidic components, like the
flow cell, cannot be ignored during UHPLC method transfer. When
accelerating an UHPLC method, the backpressure generated from
all the fluidic components becomes even more relevant when
increasing the flow rate and operating in the flat C-term of the Van
Deemter curve.
In addition to adjusting the injection volume and the tubing
volume for UHPLC methods, the detector flow cell volume has to be
adapted to the peak volumes eluting from the UHPLC column. As
a rule of thumb, the detector flow cell volume should be no larger
than approximately 10% of the smallest eluting peak volume.
UHPLC columns with sub-2 µm particles, an inner diameter of 2.1
mm and operated around 600 µL/min produce very narrow peaks
with peak widths around 1 to 3 s at the base. The peak volumes
are therefore around 10 to 30 µL; accordingly, a typical flow cell
volume would need to be 2.5 µL or less in UHPLC.
380 Method Transfer between HPLC and UHPLC

Table 13.2 The pressure drop generated by flow through the capillary is
proportional to the fourth power of the inner diameter of the capillary
according to the Hagen–Poiseuille equation

ID Capillary Pressure drop [bar] (1 mL/

[µm] min, 0.894 cp, 1 m) Internal volume [µL]
75 192.0 4.4
100 60.7 7.9
130 21 13.3
180 6 25.4
250 2 49.1
Note: Small extra column volume generates a trade-off in pressure (prerequisites
viscosity of water at 25°C, flow rate 1 mL/min, capillary length 1 m).

The extra-column volume can be evaluated to verify

the compatibility of the UHPLC column with the instrumentation
[10, 11]. As a rule of thumb, the extra-column volume should
not be more than 10% of the column volume.
The extra-column volume should not be mixed up with the
gradient delay volume. The gradient delay volume is important
for the transfer of gradient methods and will be discussed in
Section 1.2.4.1.

13.2.2.6 Data collection rate, injection cycle

Small eluting peaks from the UHPLC column have an impact on
the detector settings. The data collection rate and time constant
have to be adapted to the narrow eluting bands. It is good
practice that every peak should be at least defined by 30 raw
data points. The detector settings have to be set to optimize the
data points per peak while reducing the short-term noise, but
still maintain the peak height, symmetry, and resolution.
Figure 13.6 shows an example how the detector settings influence
the quality of the chromatogram.
Assuming the optimal data rate setting for the HPLC method
is known, the data collection rate D of the UHPLC method can be
calculated taking into account the changes in column length,
particle diameter and peak dispersion [12].
 Method Transfer Strategy 381

Figure 13.6 Ideally 30 data points should be recorded for each peak
between the limits of integration. The time constant should be fit to
the set data collection rate. The best settings should be determined by
the narrowest peak. Poor peak symmetry can be improved by reducing
the time constant. High baseline noise can be improved by increasing
the time constant. Reproduced with permission of Thermo Fisher
Scientific.

D1
D2 = (13.15)
3
L2 dp,2
3
L1 dp,1

Chromatography data systems software wizards can help find

the optimal detector settings.

13.2.3 Method Transfer Applied to Isocratic Methods

The rules and equations discussed in Section 1.2.1 are all valid
for isocratic method transfer. Following the strategy to focus on
obtaining more efficiency or reducing analysis time, new column
dimensions for the UHPLC method must be chosen. Examples
describing the process of the isocratic method transfer in detail
can be found in literature [13].
382 Method Transfer between HPLC and UHPLC

An example of a method transfer from conventional HPLC to

UHPLC is described below and shown in Fig. 13.7. A separation
of three analytes according to United States Pharmacopeia (USP)
was accelerated using the HPLC to UHPLC guidelines described
above. Compared to the original HPLC method, the solvent
consumption was reduced 95% and the analysis time was
decreased 91% of the original method. The efficiency of the
UHPLC column was less than the HPLC column but still sufficient
to get the desired resolution between analytes.

Figure 13.7 Isocratic HPLC method (top) transferred to the UHPLC

method (bottom) applying geometric transfer rules. Separation of
mevastatin (1), lovastatin (2), and simvastatin (3), mobile phase 38
mM phosphate buffer, pH4.5/acetonitrile (35/65 v/v), T = 45°C; top:
Acclaim C18, 250 × 4.6 mm, 5 µm, 1.5 mL/min, Vinj = 10 µL; bottom:
Acclaim C18, 50 × 2.1 mm, 2.2 µm, 0.8 mL/min, Vinj = 0.6 µL. Reproduced
with permission of Thermo Fisher Scientific.

13.2.4 Method Transfer Applied to Gradient Methods

Beyond the simple geometric transfer rules for column dimension,
flow rate, and injection volume that apply to isocratic methods,
 Method Transfer Strategy 383

the scaling of the gradient profiles for new column formats and
flow rates are unavoidable to maintain separation performance.
The theoretical background is known as the gradient volume
principle, introduced by L. Snyder [14]. The gradient volume VG
is defined as the mobile phase volume flowing through a column
at a defined gradient time tG.

VG = F . tG (13.16)

To maintain the selectivity and relative position of the analyte

in the gradient, the ratio between gradient volume and column
volume is kept constant [15] when transferring the method to a
column of different dimensions. With this rule and assuming the
total porosity is constant for both columns, the gradient can be
transferred to different column formats. The chromatographic
separation only changes in terms of analysis time for different
gradient time intervals.

2
F1 L2  dc,2 
t g , 2 = t g ,1     (13.17)
F2 L1  
 dc,1 

The transfer of a conventional gradient separation of seven

soft drink additives to an UHPLC method is shown as example in
Fig. 13.8. As a first consideration, a column providing sufficient
efficiency has to be selected to resolve the most critical pair,
peak 5 and peak 6. Following the geometric transfer rules, the
11,000 theoretical plates of the conventional column should
be considered. With a current resolution of R5/6 being nearly
3.5 choosing a column with slightly fewer theoretical plates
while maintaining full resolution should be sufficient. Therefore,
a 2.1 × 50 mm, 2.2 µm column generating calculated 7600 plates
was selected. Using the relationship between R and the square
root of N, a resolution increase to 2.87 can be calculated.
Accelerating the flow rate to a higher linear velocity takes full
advantage of the optimal separation efficiency of the new UHPLC
column. By scaling down the inner diameter of the column, the
flow rate is decreased accordingly to keep the linear velocity
constant. The gradient steps are calculated by the gradient volume
384 Method Transfer between HPLC and UHPLC

principle (see Eq. 13.17). The resulting chromatogram shows a

successful separation of the critical peak pair with a resolution
of 2.91 (see Fig. 13.8B). A further acceleration of the flow rate
is possible, as the influence of the linear velocity on the plate
height is smaller than the influence of the linear velocity with
larger particles. A further acceleration of the flow rate by a factor
of 2.5 generates a satisfactory resolution of R5/6 = 2.56 (see the
zoom in Fig. 13.8C).

Figure 13.8 Method acceleration of seven soft drink analytes on to an

Acclaim C18 column applying gradient transfer rules from HPLC to
UHPLC and subsequent acceleration of flow rate [12]: Aceculfame K
(1), saccharin (2), caffeine (3), aspartame (4), benzoate (5), sorbate
(6), benzaldehyde (7). A: conventional column, 4.6 × 150 mm,
4.5 µm, 1.5 mL/min; B: UHPLC column, 2.1 × 50 mm, 2.2 µm,
0.639 mL/min; C: UHPLC column and accelerated flow rate, 2.1 × 50 mm,
2.2 µm, 1.599 mL/min. Reproduced with permission of Thermo Fisher
Scientific.
 Method Transfer Strategy 385

13.2.4.1 Gradient-delay volume

The gradient delay volume plays an important role for HPLC to
UHPLC method transfer. The gradient delay volume could result
in significant changes in selectivity if it is not correctly accounted
for. The gradient delay volume is also referred to as dwell volume
and corresponds to the volume between the mixing point of
solvents to the inlet of the column. As a result, the programmed
eluent compositions reach the column head after the dwell
volume has been flushed (see Fig. 13.9). Therefore, the sample
is confronted with an undesired isocratic step that could cause
retention-time changes, which may result in the loss of resolution
for early eluting peaks.

Figure 13.9 The dwell volume creates a delay before the programmed
solvent compositions reaches the column inlet. This adds an isocratic
hold at the beginning of the gradient. Reproduced with permission of
Thermo Fisher Scientific.

To overcome this, the gradient delay volume must also follow

the gradient volume principle of Snyder. Therefore, in a correct
gradient method transfer, the ratio between the gradient delay
volume and the column volume is kept constant, while changing
the column format and flow rate. The new gradient delay volume
can be calculated by

VGDV , 1 Vc ,2
VGDV , 2 = . (13.18)
Vc,1
386 Method Transfer between HPLC and UHPLC

Examples describing the process of the gradient method transfer

in detail, considering differences in column formats and gradient
delay volume, can be found in literature [16].
The volume going from a typical HPLC column, 4.6 mm ×
150 mm, to a 2.1 × 50 mm UHPLC column reduces the column
volume by a factor of 14 when considering the same porosity.
Therefore, the gradient delay volume should also be reduced by
a factor of 14.
The main contributors to the gradient delay volume are the
pump-mixing volume, the autosampler fluidics and all capillary
connections before the column inlet. The determination of the
gradient delay volume is described in the literature [17] and can
help optimize the UHPLC method.
For UHPLC gradient methods it is imperative to work with
UHPLC systems with low gradient delay volume. By its technical
design, high pressure mixing pumps are the preferred as they have
smaller dwell volume (see Fig. 13.10).

Figure 13.10 The gradient delay volume (dotted line) is defined as the
volume from the mixing of solvents to the column head. By its technical
operating principle the gradient delay volume is lower for high pressure
gradient pumps (top) than for low pressure gradient pumps (bottom).
Gradient delay volume must be considered for optimal gradient
method transfer from HPLC to UHPLC. Reproduced with permission
of Thermo Fisher Scientific.

The biggest contributor to the gradient delay volume is the

volume of the mixer. Therefore, the mixer must be considered
when optimizing the dwell volume. Instrument manufacturers
offer systems with adjustable gradient delay volumes to ease
method transfer.
 Method Transfer Tools 387

In practice, it is difficult to scale down the gradient delay

volume to exactly match the geometric calculation according to
Eq. 13.18. For high differences in the gradient delay volumes of
HPLC and UHPLC system the solution is either to add an isocratic
step or to delay the injection after the start of the gradient.

VGDV ,2 VGDV ,opt

tD = (13.19)
F2

The influence of different gradient delay volumes to a gradient

separation can be seen in Fig. 13.11. Transferring a method from
an instrument with a high gradient delay volume (bottom) to a
system with less gradient delay volume (top) shifts the peaks
toward earlier retention times. This shift in the early eluting polar
analytes is especially critical as they may be wrongly integrated
with peak picking automation.

Figure 13.11 The pseudo isocratic hold while transferring from an HPLC
system with high gradient delay volume (bottom, grey) to a system with
lower gradient volume (top, black) has to be considered for the smooth
transition from HPLC to UHPLC to avoid problems. Reproduced with
permission of Thermo Fisher Scientific.

13.3 Method Transfer Tools

The method transfer process from HPLC to UHPLC considering
all the geometric rules can be easily integrated into calculators.
388 Method Transfer between HPLC and UHPLC

These calculators are available from several sources. An overview

can be found in reference [18]. The original column dimensions,
flow rate, injection volume, and gradient profile can be programmed
in by the user to automatically determine the new UHPLC
parameters to be used. Additionally, parameters like resolution
prediction, adjusted data collection rate, back pressure prediction,
and consideration of gradient delay volumes are also integrated
to aid in successful method transfer.
These calculators are available as online tools [19],
downloadable offline tools [12, 20] or even embedded into the
chromatography data software, allowing for easy method transfer
by generating the transfer method parameters automatically [21].

13.4 Limitations of Method Transfer

Method transfer from HPLC to UHPLC has different limitations.
Working at very high pressures, the frictional heating phenomena
may affect method transfer. Furthermore, transferring the
stationary phase to a different morphology and the transfer
of instrumental parameters settings can be challenging. Other
theoretical method transfer options might be limited by regulatory
affairs and related restrictions.

13.4.1 Frictional Heating

Thermostatting the separation column and mobile phase is very
important for achieving a robust and reproducible separation
method. Small changes in temperature can have a huge effect on
the retention factor and in turn affect the selectivity of the method
as expressed by the logarithmic relationship of retention factor
k and temperature T in the Van’t Hoff equation (see Eq. 13.20).

H 0 1 S 0
ln k =  + + ln  (13.20)
R T R

In addition to external thermostatting of the UHPLC column,

internal heat generation must be considered. UHPLC separations
using columns packed with fine particles experience frictional
heating throughout the column. When mobile phase enters the
 Limitations of Method Transfer 389

column, frictional heat is generated inside the column by the

rubbing of viscous eluents over the packed particles at high
pressures and flow rates. The generated frictional heat within the
column can influence the retention time prediction of analytes
being transferred from an HPLC to UHPLC method [22, 23].
The dissipation of heat occurs along and across the column,
generating axial and radial temperature gradients, which can
influence not only the efficiency of the separation but also the
retention and selectivity [24]. The magnitude of the axial and
radial temperature distribution in the column depends on the
thermostatting conditions in the column compartment.
A radially heterogeneous distribution of temperature is mainly
observed with isothermal or forced air column thermostatting,
because the column wall temperature is controlled but the
temperature in the center of the column is increased. As a result, the
peak shape is distorted and a loss of efficiency may be observed.
An axially or longitudinally heterogeneous distribution of
temperature is mainly observed when the temperature at the
column wall is not kept constant because of adiabatic or still-air
thermostatting. As a result, the temperature at the column outlet is
higher than at the column inlet and changes in retention time and
selectivity may occur [25, 26].
An example of loss in efficiency, by using forced air thermostatting
compared to still air thermostatting while operating under
frictional heating conditions, is shown in Fig. 13.12.
Frictional heating also occurs in HPLC methods operating
at relatively low pressures. However, the influence of frictional
heat is more significant for methods using sub-2 µm particles and
column lengths of 50 to 100 mm operating at pressures above
400 bar. Frictional heat should not be neglected when aiming for
a smooth HPLC to UHPLC method transfer. This is especially true
if the thermostatting operating principles have to be adjusted
during method transfer due changing instrumentation. For UHPLC
methods working under frictional heat conditions it could be
suitable to set the column thermostatting at a lower temperature
to overcome the heat addition if allowed by the method [26].
The eluent pre-heating also influences the formation of the
temperature gradient. An example how the thermostatting of
column and pre-heating of mobile phase can influence the method
transfer result is shown in Fig. 13.13.
390 Method Transfer between HPLC and UHPLC

Figure 13.12 Frictional heating effects and generation of a longitudinal

temperature gradient in the column shows that still air operation mode
results in increased optimal flow rate and increased efficiency with a
slightly lower pressure drop of 860 bar versus 880 bar [25]. Column:
Thermo Scientific Hypersil Gold, 1.9 µm, 2.1 × 100 mm, acetonitrile/waters
(50:50 v/v). Thermostatting: 30°C in still air or forced air mode, active
eluent pre-heating: 30°C, sample: hexanophenone. At a linear velocity
of 8 mm/s the column efficiency by forced air mode is 40% lower than with
still air mode. The optimum linear velocity in still air mode 20% higher
and has 10% better efficiency than in forced air mode. Reproduced with
permission of Thermo Fisher Scientific.

13.4.2 Stationary Phase Differences

Transferring methods onto a column with similar stationary phase
chemistry, but different particle size can cause selectivity issues.
This is mostly related to a difference in silanol activity. Particle
size distribution of the stationary phase is another important
aspect to consider when applying geometric rules for transferring
methods. The effective and nominal particle size might differ. For
 Limitations of Method Transfer 391

example, a difference in 5 or 4.5 µm particle size can generate

different separation results.

Figure 13.13 The effect of thermostatting conditions on an UHPLC

separation experiencing frictional heating (>750 bar) Column: Acclaim
RSLC PA2, 2.1 × 150 mm, 2.2 µm, 0.55 mL/min, resulting in 760 bar
back pressure (Sample: Uracil (1), Dimethyl phthalate (2), Methyl
Paraben (3), methyl benzoate (4). Forced air thermostatting with passive
pre-heating resulted in a resolution of the critical peak pair above 1.5.
However, the plate number was below 13,000 plates. With still-air
thermostatting and active-preheater, the resolution of the critical peak
pair was insufficient, but the efficiency was improved. By setting the
active pre-heater at a lower temperature independent of the column
thermostatting resulted in an increase in both the resolution and
the plate number [26]. Reproduced with permission of Thermo Fisher
Scientific.
392 Method Transfer between HPLC and UHPLC

Another potential method transfer challenge may occur

when attempting to transfer a method from fully porous to core–
shell stationary phase. The change in the particle morphology
should be considered as well, especially when transferring
gradient methods. In the gradient volume equation (see Eq. 13.18),
the internal column volume filled with mobile phase should be
taken into consideration and adjusted by Eq. 13.21.

VM = r 2 Le (13.21)

The parameter e is the column interstitial porosity percentage.

As a rule of thumb e is 0.68 for fully porous material and 0.55
for core–shell particles. When transferring an UHPLC method
from a column filled with fully porous material to a core–shell
material change of the particle morphology effects, the volume
of mobile phase in the column and should be accounted for in
method transfer calculations [27].

13.4.3 Regulatory Limitations

The method transfer from HPLC to UHPLC or vice versa from
UHPLC to HPLC can be made by applying the rules discussed in
this chapter. However, validated methods and methods described
in monographs with the need to meet specific requirements are
somewhat restricted to HPLC to UHPLC method transfer. The
European Pharmacopeia (EP) and the United States Pharmacopeia
(USP) limit some modifications to the chromatography methods.
Chapter <621> of the USP defines an acceptable range where
method parameters can be adjusted without needing to revalidate
the method. The USP prerequisites include that the transferred
method still meets the system suitability requirements and
passes verification tests after the changes.
It is relatively straightforward to transfer isocratic methods
within certain range of variations. However, the transfer of
gradient methods allows very limited adjustments. Changes in
length, column inner diameter and particles size as well as changes
in flow rate are not allowed for gradient separations. Changes in
column packing while maintaining the chemistry of stationary
phase, changes in the duration of a prescribed isocratic hold
 Limitations of Method Transfer 393

and adjustments of the dwell volumes are in theory allowed,

but practically those changes alone do not support an HPLC
to UHPLC method transfer without the allowance of changing
column properties. Furthermore, changes in the gradient table are
perceived as uncertain and are more strongly limited for
modifications. The limitations described in USP chapter <621> [28]
are summarized in Table 13.3.

Table 13.3 Allowances and limitations for method transfer according to

USP [28]

Allowed adjustments by USP 38-NF 32

(December 1, 2015) Chapter <621>
Chromatographic
parameter Isocratic Gradient
Particle size L/dp ratio constant or No changes
N: –25% to +50% allowed
Column length L/dp ratio constant or No changes
N: –25% to +50% allowed
Flow rate F2 = F1 × [(dc22 × dp1)/ No changes
(dc21 × dp2)] allowed
and ± 50%
Column ID Any allowed if linear velocity is No changes
constant allowed
Injection volume Can be adjusted as consistent with precision,
linearity, and detection limits
Column temperature ±10°C
Mobile phase pH ±0.2 units
Concentration of
±10%
buffer salts
Composition of ±30% relative of the minor solvent (no other
mobile phase component is altered by more than 10% absolute)

Any changes of the method parameters beyond the ranges

described above require the full re-validation. A gradient method
transfer from HPLC to UHPLC with the motivation to speed-up
the method requires therefore a re-validation. However, this can
also be taken as an opportunity to fully optimize the method
with modern sub-2 µm particle columns by doing a fast UHPLC
394 Method Transfer between HPLC and UHPLC

method development on up-to-date UHPLC instrumentation and

validation supported by latest chromatography data systems.

13.5 Conclusion
The motivation for transferring a conventional HPLC method to
UHPLC can be better resolution or faster analysis time. Following
simple geometric rules and some practical guidelines allows for
easy transfer the conventional method. A systematic method
speed-up can be made by reducing the particle size, shortening
the column length and increasing the linear velocity of the mobile
phase. For high-resolution chromatography, all of the geometric
rules can greatly impact the separation and should be considered
accordingly. Method transfer tools embedded in chromatography
data software and available online or offline help to easily apply
the geometric rules.
Not only does the pressure capability of the UHPLC
instrument performance for method transfer matter, but the extra-
column volume, gradient delay volume, detection options, and
thermostatting characteristics have to be considered as well.
While there are many facets to consider for successful method
transfer, UHPLC separations give a greater freedom to exploit the
principles of chromatography and get the optimized method to
solve the analytical problem.

Acknowledgements
My thanks go to my colleagues for contribution to the chapter,
especially Holger Franz, Frank Steiner, Sabrina Patzelt, Michael
Heidorn, and Jenny Wong.

References

1. Purnell, J. H. (1959) Comparison of efficiency and separating power

of packed and capillary gas chromatographic columns, Nature, 184,
p 2009.
2. Halász, I., Endele, R., and Asshauer, J. (1975) Ultimate limits in high-
pressure liquid chromatography, J. Chromatogr. A, 112, pp. 37–60.
 References 395

3. Snyder, L. R., Dolan, J. W., and Carr (2004) The hydrophobic-subtraction

model of reversed-phase column selectivity, J. Chromatogr. A, 1060,
pp. 77–116.
4. Guillarme, D., Grata, E., Glauser, G., Wolender, J. L., Veuthey, J. L., and
Rudaz, S. (2009) Some solutions to obtain very efficient separations
in isocratic and gradient modes using small particles size and
ultra-high pressure, J. Chromatogr. A, 1216, pp. 3232–3243.
5. Dolan, J. W. (2017) Increasing resolution by increasing column
efficiency, LCGC Eur., 30(3), pp. 138–142.
6. Guillarme, D., Ruta, J., Rudaz, S., and Veuthey, J. L. (2010) New trends in
fast and high-resolution liquid chromatography: A critical comparison
of existing approaches, Anal. Bioanal. Chem., 397, pp. 1069–1082.
7. Desmet, G., Cabooter, D., and Broeckhoven, K. (2015) Graphical data
representation methods to assess the quality of LC column, Anal.
Chem., 87, pp. 8593–8602.
8. Fountain, K. J., Neue, U. D., Grumbach, E. S., Diehl, D. M. (2009) Effects
of extra-column band spreading, liquid chromatography system
operating pressure, and column temperature on the performance of
sub-2-μm porous particles, J. Chromatogr. A, 1216, pp. 5979–5988.
9. Gritti, F., Guichon, G. (2010) On the extra-column band-broadening
contributions of modern, very high pressure liquid chromatographs
using 2.1 mm I. D. columns packed with sub-2 μm particles, J.
Chromatogr. A, 1217, pp. 7677–7689.
10. Gritti, F., Guichon, G. (2014) Accurate measurements of the true column
efficiency and of the instrument band broadening contributions
in the presence of a chromatographic column, J. Chromatogr. A, 1327,
pp. 49–56.
11. Vanderlinden, K., Broeckhoven, K., Vanderheyden, Y., Desmet, G. (2016)
Effect of pre- and post-column band broadening on the performance
of high-speed chromatography columns under isocratic and
gradient conditions, J. Chromatogr. A, 1442, pp. 73–82.
12. Franz, H., Fabel, S. (2013) A universal tool for method transfer from
HPLC to UHPLC, Technical Note 75, Thermo Fisher Scientific, USA.
13. Guillarme, D., Nguyen, D. T. T., Rudaz, S., Veuthey, J. L. (2007) Method
transfer for fast liquid chromatography in pharmaceutical analysis:
Application to short columns packed with small particle. Part I:
Isocratic separation, Eur. J. Pharm. Biopharm., 66, pp. 475–482.
14. Snyder, L. R., Dolan, J. W., Gant, J. R. (1979) Gradient elution in
high-performance liquid chromatography: I. Theoretical basis for
reversed-phase systems, J. Chromatogr. A, 165, pp. 3–30.
396 Method Transfer between HPLC and UHPLC

15. Schellinger, A. P., Carr, P. W. (2005) A practical approach to transferring

linear gradient elution methods, J. Chromatogr. A, 1077, pp. 101–119.
16. Guillarme, D., Nguyen, D. T. T., Rudaz, S., Veuthey, J. L. (2007) Method
transfer for fast liquid chromatography in pharmaceutical analysis:
Application to short columns packed with small particle. Part II:
Gradient experiments, Eur. J. Pharm. Biopharm., 68, pp. 430–440.
17. Dolan, J. W. (2006) Dwell volume revisited, LCGC Eur., 24(5),
pp. 458–466.
18. Majors, R. E. (2011) Method translation in liquid chromatography,
29(6), pp. 476–484.
19. HPLC Method Transfer Calculator, Thermo Fisher Scientific, (https://
www.thermofisher.com/de/de/home/global/forms/industrial/lc-
method-transfer-calculator.html).
20. Easy HPLC to UHPLC Transfer (xlm–file), Thermo Fisher Scientific
(http://www.thermoscientific.com/content/dam/tfs/ATG/CMD/
cmd-documents/oper/tech/chrom/lc/sys/RSLC-Method-Transfer-
Tool-2-3.xlsm).
21. Hillbeck, D. (2016) Rapid screening method for statins using an
advanced UHPLC column and system, Application Note 21496, Thermo
Fisher Scientific, USA.
22. McCalley, D. V. (2014) The impact of pressure and frictional heating
on retention, selectivity and efficiency in ultra-high-pressure liquid
chromatography, Trends Anal. Chem., 63, pp. 31–43.
23. Grinias, J. P., Keil, D. S., Jorgenson, J. W. (2014) Observation of
enhanced heat dissipation in columns packed with superficially
porous particles, J. Chromatogr. A, 1371, pp. 261–264.
24. Heidorn, M. (2016) The role of temperature and column
thermostatting in liquid chromatography, White Paper 71499, Thermo
Fisher Scientific, USA.
25. Heidorn, M. (2016) A more flexible column thermostatting
technique in LC method transfer, White Paper 71500, Thermo Fisher
Scientific, USA.
26. Steiner, F., Patzelt, S., De Pra, M., Martin, M. M. (2016) Optimizing
selectivity and efficiency in critical UHPLC methods by separating
eluent pre-heating from column thermostatting, Poster Note 72103,
Thermo Fisher Scientific, USA.
27. Taylor, T. (2015) HPLC column transfer from fully porous to
core–shell in three simple equations, Column, 11 (9), pp. 2–4.
28. United States Pharmacopeia, USP 38-NF 33, chromatography <621>.
Chapter 14

Small-Molecule Pharmaceutical
Impurities Test Method Validation:
Precision Acceptance Criterion

Russell L. Hertzler,a Mark D. Johnson,a Teodora Moldovan,a

Michael D. Oberlander,a James Reynolds,a Holger van Lishaut,b
and Yanbing Zhenga
aAbbVie, 1 N. Waukegan Rd, North Chicago, Illinois 60064, USA
bAbbVie Deutschland GmbH & Co. KG
Knollstrasse, 67061 Ludwigshafen, Germany
[email protected]

14.1 Introduction
The International Conference on Harmonization (ICH) guideline
[1] for pharmaceutical small-molecule impurities test method
validation does not provide any guidance in setting numerical
validation performance criteria for linearity, accuracy, and
precision. It specifies a numerical minimum acceptance criterion

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
398 Small-Molecule Pharmaceutical Impurities Test Method Validation

for signal-to-noise performance respective of the test method’s

limit of detection and limit of quantitation. The lack of ICH
defined acceptance criteria for the other performance attributes
is most likely intentional, so the analytical procedures developed
can be validated for their intended purpose. However, for
regulatory submissions, the test method acceptance criteria used
to demonstrate that analytical procedures are suitable are required
to be reported. Regulatory Agency CMC reviewers may request
the sponsor to justify their test method validation acceptance
criteria.
For the main component (active ingredient assay) quantitation
test method, sponsors could select to set absolutely the precision
performance acceptance criterion as the tolerable variation
predefined as a statistically based fraction of the active ingredient
specification range [2]. A specification-based approach to setting
precision performance acceptance criterion for the active ingredient
assay test method is sensible and straight forward from the
scientific perspective and is less likely to be met with objection
during regulatory review.
Setting an objective, predefined, numerical acceptance
criterion for impurities test method precision performance is
more challenging in advance of the initial marketing application
in conjunction with Chemistry, Manufacturing, and Controls
(CMC) development of proprietary drugs. The challenge in using
a specifications-based approach to set precision acceptance
criterion for impurities and degradation products test methods
is due in part to uncertainties in negotiating the final impurities
and degradation products specifications with the regulatory
bodies. Additional limitations to using the specifications-based
approach could include a statistically limited knowledge of the
independent variation in the manufacturing process batch to
batch performance, as well as the product stability profile variation,
within their respective design space, at the time of test method
validation and method technical transfer which typically occur in
advance of the initial marketing application date.
A critical factor that should not be overlooked is that one of
the test method validation acceptance criteria should account
for the combined effect of precision and accuracy performance
[3], regardless of the nature of the quantitated analyte, active
 Introduction 399

ingredient, impurity, or inactive ingredient critical to the product

quality. For example, a test method systematic accuracy bias
would require corresponding improvement in the test method
precision performance in order to avoid excessive incidence of
reporting false passing or false failing results versus the material
specification limits.
The focus of this chapter does not include the accounting of
any systematic quantitation or manufacturing process bias, or
combinations of these, on test method precision requirements.
Instead this chapter, irrespective of quantitation and
manufacturing process bias, focuses on impurities and
degradation products test method validation precision in terms
of the comparative performance of a set of AbbVie test method
validations to an empirical observation that the literature credits
as being first investigated by Horwitz [4–6]. Without regard to
an approach of setting test method precision performance
requirements as a function of material specification limits, Horwitz
observed test method(s) inter-laboratory precision to generally
exhibit a systematic trend where the relative variability of the
analytical determination increases exponentially with decreasing
analyte concentration expressed as absolute mass fraction.
In terms of current trends applicable to small-molecule
pharmaceutical quality, Horwitz’s work has not been forgotten.
Without any direct explicit reference to Horwitz’s work, the
2015 Chinese Pharmacopoeia (Chapter 9101 Guidelines for
Validation of Analytical Method Adopted in Pharmaceutical Quality
Specification) can be reasonably inferred to have adopted the
Horwitz function as the tolerable level of test method inter-
laboratory variability and thus as the Chinese Pharmacopoeia
Acceptance Criterion for Reproducibility. A previously published
review of small-molecule pharmaceutical industry test method
precision performance has also recognized the work of Horwitz [7].
Hence, given our recent internal experiences in the context of
some external trends potentially signaling more comprehensive
expectations for objective predefined acceptance criteria, AbbVie
elected to make an assessment of our historical “minors” (“minors”
include manufacturing impurities, degradation products, and
functionally critical formulation excipients requiring testing
for control) quantitation test method performance against the
400 Small-Molecule Pharmaceutical Impurities Test Method Validation

published Horwitz Acceptance Criteria as the performance

parameter indicating the acceptability of methods of analysis
with respect to inter-laboratory precision (Reproducibility) [6].
This chapter does not present a comprehensive review of
the topic of defining suitable precision performance of impurities
test methods, which lies at the interface of statistical sciences
and the pragmatic implementation of robust quality control
analytical test method technology in the industrial pharmaceutical
CMC development and subsequent commercial manufacturing
setting.
This chapter presents a meaningful sample of AbbVie impurities
and degradation products test method validation precision
performance against the Horwitz Reproducibility acceptance
criterion.

14.2 Horwitz’s Empirical Observation of Test

Method Variability as a Function of
Analyte Concentration
William Horwitz published in 1980 an empirically observed
correlation between test method coefficient of variation and
analyte concentration from a collection of 7502 observations from
inter-laboratory collaborative studies (Reproducibility) spanning
multi-decades of analytical progress. The examined test method
sample matrices included pharmaceuticals, food, cosmetics, and
environmental samples of many kinds [4] (see Fig. 14.1).
Horwitz and coworkers found an empirical relation between
the absolute mass fraction concentration of the analyte quantitated,
c, and the observed value of the corresponding inter-laboratory
percentage relative standard deviation of the analyte determination.

RSDR, Horwitz = 21–0.5 log c = 2c–0.5 log 2 = 2c–0.1505

where log is base 10.

Equivalently in its linear transformation:
log (RSDR, Horwitz) = log(2) – 0.1505 log (c) = 0.3010 – 0.1505 log (c).

Note: The concentration, c, was expressed in terms of

absolute mass fraction to establish the relationship. For example,
 Test Method Variability as a Function of Analyte Concentration 401

a concentration of 1 g analyte per 100 g matrix corresponds to

c = 0.01, and using base 10, log c = –2. The above equations give
an RSDR, Horwitz of 2%, for c equal to 1; and an RSDR, Horwitz of 16%,
for a matrix concentration c of 0.000001, which corresponds to
an analyte concentration of one part per million.

Figure 14.1 The conceptual “Horwitz horn”. Inter-laboratory coefficient of

variation as a function of concentration. [4] Reprinted with permission
from AOAC INTERNATIONAL.

The formula became known as the “Horwitz horn” or “Horwitz

formula”. Qualitatively, the property reflected by the Horwitz
function [4–6] is familiar and intrinsic to analytical test method
precision performance as a function of a multi-fold reduction of
analyte concentration. The relative variation (Relative Standard
Deviation) of the analyte determination is generally found to
systematically increase with decreasing analyte concentration, c.
402 Small-Molecule Pharmaceutical Impurities Test Method Validation

See Fig. 14.2, noting that it plots the log of absolute standard
deviation versus log c.

Figure 14.2 Reproducibility standard deviation and concentration of

analyte in collaborative trials (7,502 observations). The line is the
Horwitz function Reprinted with permission from AOAC INTERNATIONAL
[5].

In our view, we note that the y-intercept (the %RSD criterion

corresponding to c = 1) of the log linear Horwitz function (or
equivalently the coefficient of c for the Horwitz function expressed
in its power form) could be justifiably re-scaled to value(s) less
than 2% RSD. Horwitz and Albert [6] do not cite any specific
experimental observations as forming the justification for their
assignment of this intercept. Hofer et al. [8] cite a pooled RSD
of 0.61% and an RSD range of 0.29% to 1.0% as representative
of intermediate precision performance for an HPLC-based drug
substance potency assay.
As well, in terms of the Horwitz function, a re-scaled intercept
value could be justifiably assigned based on the tolerable variability
predefined from a specification-based Capability Analysis, Cpk, to
setting the inter-laboratory precision acceptance criterion.
 Test Method Variability as a Function of Analyte Concentration 403

 USL – c( i )active; process c( i )active; process – LSL 

C pk,assay test = Minimize s assay test , 
 3s assay test 3s assay test 
 1.33

 USL – c( i )impurity;process c( i )impurity;process – LoQ 

C pk,impurity test = Minimize s impurity test , 
 3s impurity test 3s impurity test 
 1.33

The validated test method precision standard deviation,

si, according to the precision validation design and degrees of
freedom, is an estimate of the respective test method true
standard deviation 𝜎i, 𝜎impurity test method or 𝜎active ingredient test method.
USL, LSL and LoQ, respectively, are the Upper Specification Limit,
Lower Specification Limit and the impurities or degradation
products test method Limit of Quantitation applicable to the
determined analyte concentration level, ci, whose true value is,
in general, manufacturing process (i.e., variation batch to batch)
and stability storage time and temperature dependent in the case
of degradation products. Of course, the setting of the specification
limits, USL(s), for impurities and degradation products is strictly
according to the results of toxicological qualification studies
and, or permitted daily exposures specified by the respective
ICH Regulations [9–11], and in conjunction with the assessment
of manufacturing process capability. From the test method
performance perspective, Cpk is optimally ≥1.33.
For an unbiased drug substance manufacturing process
whose batch content is subject to an assay specification of 98.0
to 102.0%w/w, a suitable predefined acceptance criterion for
the validation inter-laboratory precision of an unbiased test
method would be s ≤ 0.50% RSD based on a Cpk of 1.33 and a
statistically robust precision validation design and degrees of
freedom.
Returning to the topic of impurities test method precision
performance, although analytical technology has very obviously
improved over time in terms of absolute concentration sensitivity,
Horwitz’s empirical observation indicates that the relative variability
of analytical determinations is meaningfully independent of these
404 Small-Molecule Pharmaceutical Impurities Test Method Validation

sensitivity improvements. Given that a first principles theoretical

model to support Horwitz’s empirical observation does not appear
to exist, we can only postulate, similarly to Ermer et. al. [7] who
stated, “the standard deviation decreases less rapidly than the
concentration, resulting in an increase of the relative standard
deviation for lower concentrations.”
At present, a broad multi-fold concentration range
collaborative inter-laboratory precision re-assessment might
be indicated, represented by the most contemporary practiced
analytical technology and classified according to the test method
intended purpose. For example, it can be reasoned that the set of
reproducibility performance observations evaluated by Horwitz
through 1980 would not have likely included a meaningful
sample of quantitative instrument-based chromatographic
pharmaceutical impurities and degradation product test methods.
The expectation for routine impurities quality control for
small-molecule pharmaceutical products was not firmly in place
until initiated by ICH in 1996 [12].
The essential elements of the Horwitz observed empirical
correlation between inter-laboratory relative variability of the
determination and the analyte weight fraction, c, are the following:
1. The relative variability (%CV or %RSD) is a power function
of c.
2. For a trained analyst, the relative variability is largely
independent of further experience.
3. There is no overriding influence of routinely practical
analytical technology progress on variability measured
according to the relative standard deviation of the analyte
determination.
4. There is no overriding reduction in the relative variability
performance over a broad multi-fold range of c as a function
of the selected analytical technique beyond the obvious course
of action that more sensitive detectors and sophisticated
techniques, for example sample pre-concentration extraction,
are selected for the determination of lower analyte absolute
mass concentrations.
The Horwitz function can be considered as an objective basis
for the acceptance criterion for variability in inter-laboratory
 Test Method Variability as a Function of Analyte Concentration 405

trials, also referred to as test method Reproducibility. A complete

summary can be found in Horwitz and Albert [6].

Figure 14.3 Example of a published inter-laboratory precision (Repro-

ducibility) acceptance criteria [13] for Drug Substance and that
corresponding to HorRat = 1.0 and HorRat = 2.0.

In addition and alternate to comparing the AbbVie test method

performance to the Horwitz acceptance criterion, an arbitrarily
selected example from a published [13] impurities test method
Reproducibility acceptance criterion is presented in Figs. 14.3
and Figs. 14.4. Although we present only two specific cases,
the Reproducibility acceptance criterion of Crowther et al. [13]
is general, based on the assignment of specified %RSD(s)
criterion respectively to analyte concentration ranges which are
distinguished as multiples of the test method’s intrinsic Limit
of Quantitation. This approach sets the Reproducibility acceptance
criterion as a function only of the test method intrinsic LoQ
and therefore, in our view, can be regarded as a Signal to
406 Small-Molecule Pharmaceutical Impurities Test Method Validation

Noise-based approach to setting impurities test method

precision performance. If our interpretation is appropriate, then
presumably the acceptable Signal to Noise at the intrinsic LoQ is
that defined by ICH [1], namely Signal to Noise equal to and not
meaningfully exceeding 10 to 1. Accordingly then, for any impurity
test method, Reproducibility performance for quantitation very
near the method intrinsic LoQ would be subject to an acceptance
criterion of ≤30%RSD [13], regardless of the impurity absolute
concentration corresponding to the intrinsic LoQ. As would
be expected, Crowther et al., [13] indicate a series of more
restrictive Reproducibility RSD(s) step wise as the concentration of
analyte increases above the test method’s intrinsic LoQ.

Figure 14.4 Example of a published inter-laboratory precision

(Reproducibility) acceptance criteria [13] for Drug Product and those
corresponding to Horwitz showing the dependence on formulation drug
load fraction, c2. Cases plotted for HorRat = 2.0 and c2 fractions of 0.5,
0.05 and 0.0005.
 Test Method Variability as a Function of Analyte Concentration 407

In Fig. 14.4, our depiction of Crowther’s et al. [13] drug product

Reproducibility acceptance criterion in direct comparison with
those of Horwitz must be viewed with caution. This is due to the
fact that for the Horwitz criterion, the concentration axis, c,
is to reflect the mass fraction of analyte in the whole matrix
mass, whereas we expect certainly that Crowther’s intended
concentration scale for drug product degradation products is the
mass fraction, or mass percent, of degradation product relative to
the mass of active ingredient formulated.
As reflected in the examples plotted in Fig. 14.4, in our
view a generalization of the Horwitz function can be considered
when analyte concentration is expressed relative to the active
ingredient mass, which is the expected basis of pharmaceutical
impurity and degradation products reporting for Regulatory
purposes.
The Horwitz mass fraction of analyte relative to the total
matrix mass, c, can be expressed as the product of the fraction of
impurity or degradation product mass to drug mass, c1, times
the fraction of drug mass to the total matrix mass, c2. The
Horowitz equation then becomes

RSDR, Horwitz = 2c–0.1505 = 2(c1· c2)–0.1505.

For drug substance, c2 equals one, and by definition for drug

product c2 is less than one.
In Section 14.2.1 we introduce Horwitz’s Ratio (HorRat) [6].

14.2.1 Horwitz’s Acceptable Observed Test Method

Reproducibility: RSDR, Observed
The “Horwitz-Ratio” or “HorRat” describes the quotient [6]:
Horwitz Ratio = HorRat = (RSDR, Observed)/(RSDR, Horwitz).
Horwitz allowed up to 2×RSDR, Horwitz or HorRat ≤ 2.0 as the
acceptable upper limit of variability, RSDR, Observed

RSDR, Observed
HorRat =  2.0
RSDR, Horwitz
408 Small-Molecule Pharmaceutical Impurities Test Method Validation

Therefore the upper limit of acceptable inter-laboratory

precision (Reproducibility) is:

RSDR, Observed ≤ 2xRSDR, Horwitz = 4c–0.1505 = 4(c1· c2)–0.1505.

Even though a rigorous justification of his cited 0.5 ≤ HorRat

≤ 2.0 acceptance range is not explicitly provided by Horwitz,
one can rationalize acceptable HorRat values falling within the
specified acceptance range 0.5 ≤ HorRat ≤ 2.0 as in likeness to
the statistical principle for calculation of confidence intervals
around the estimate, s, of the true standard deviation, s, of a
population. Generally the standard deviation estimate of the
population true value is associated with a confidence interval
calculated from the Chi-squared distribution according to a
chosen specified confidence and the finite degrees of freedom
of the precision experiment. See Table 4 in Horwitz and Albert [6].

14.2.2 Evaluation of AbbVie Test Method Validation

Repeatability, Intermediate Precision
Performance
The selected AbbVie test methods evaluated included
chromatographic ordinary and targeted impurities, and minor
formulation component (i.e., antioxidant, functional excipient) test
method validations. Here, a targeted impurity can be distinguished
as that, for example, requiring control to a very stringent
specification limit, such as a carcinogenic or mutagenic impurity
[11]. Ordinary impurities, for example, are those which have
undergone specific toxicological qualification, and, or impurities
which are to be quantified if present at or above the respective
ICH Reporting Threshold [14, 15], and therefore do not require as
highly sensitive analytical test methodology. New Chemical
Entity Drug Substance (DS), Drug Substance Manufacturing
Intermediates (DS-I), Drug Product Intermediates (DP-I) and Drug
Product (DP) test methods were included. The evaluated test
methods composition break down is summarized as follows:
 Test Method Variability as a Function of Analyte Concentration 409

Impurity-degradation product analytical Number of instances

method type included
HPLC-UV 53
HPLC-MS 6
GC-FID 11
GC-MS 2
Subtotal of (DS + DS-I) test methods 45
Subtotal of (DP + DP-I) test methods 27
Total of Number of test methods 72
Total number of analyte Concentration 813
levels in evaluation

RSDs corresponding to impurity accuracy study percentage

recoveries were also included, which are typically performed intra-
laboratory. Therefore the data is composed of inter-laboratory
precision, combined with data from accuracy studies, the latter
of which are properly described as “repeatability type” designs.
The test method variability assessment derived from the RSDs
from multi-level analyte recovery studies (Accuracy validation)
would be expected to have reduced variability as compared to
the variability that would be expected from an inter-laboratory
determination conducted at the same concentration level of
analyte. The data fractional composition of RSDs derived from
test method Accuracy validation is 644 of 813, or about 79%, and
the balance of 21% is data generated from inter-laboratory
precision.
For the formulated products, the range of drug load fractions
(w/w) included in the data is 3.38 × 10–6 to 0.938, and the median
drug load is 0.048.
The data fractional composition of impurity or minor
component concentration(s) that were adjusted for respective
drug product drug load in calculating the concentration, c, is 520
of 813, or about 64%.
In Fig. 14.5 there are no active ingredient potency test
methods, for drug substance or for drug product, plotted in our
410 Small-Molecule Pharmaceutical Impurities Test Method Validation

data evaluation versus Horwitz. If the active ingredient assay test

methods “class” were plotted, they would exhibit very distinct
superiority versus HorRat ≤ 1.0 acceptance criterion, even without
drug load adjustment for drug product active ingredient assay
test methods.
The AbbVie test method data is plotted in Fig. 14.5 along with
lines indicating a HorRat of 2.0 and the 95% Quantile Regression
line in conjunction with its evaluation according to a statistical
Quantile Regression (QR) analysis [16, 17]. The QR objective
function line and the HorRat = 2.0 line are nearly superimposed
for a high proportion of the data (i.e., for the quantile t = 0.95).
The vast majority of the AbbVie data is below both of the lines,
with only 2% of the data above the HorRat = 2.0 line. Some of
the points with HorRat > 2.0 are described in Section 14.2.3.
Adequate precision performance against the criterion of HorRat
≤ 2.0 is supported by the quantile regression of the AbbVie
data set.

Figure 14.5 HorRat = 2.0 line and the quantile regression of the AbbVie
data of seventy two impurity, degradation product or minor component
test method precision validations. The analyte concentration, c (%w/w),
is relative to the whole matrix mass.
 Test Method Variability as a Function of Analyte Concentration 411

14.2.3 Discussion and Conclusion

While a small fraction of the data RSDs (15 of 813) exceed the
HorRat ≤ 2.0 (the Horwitz Acceptance Criterion), these instances
generally exhibit an “attributable” cause. For example in one
instance, the subject impurity UV Response Factor was not found
repeatable inter-laboratory between different HPLC instrument
manufacturer(s) and in another, the cause is attributed as
intrinsic to the test method which employs a complex liquid-liquid
extraction sample preparation. In all cases, the respective test
method performance was suitable for its intended Regulatory use.
The following conclusions can be drawn from the data
evaluation versus Horwitz (HorRat ≤ 2.0 Acceptance Criterion).
The test method validation precision performance for both
drug substance and drug product targeted and ordinary impurities
and degradation products, and formulation critical excipient
component(s), is predominately in-line with, or superior to,
HorRat ≤ 2.0 acceptance criterion.
An advantage of HorRat ≤ 2.0 is principally that the slope of
the Horwitz function (or equivalently the exponent of c expressed
in power form of the Horwitz function) is an objective externally
recognized basis for acceptable test method Reproducibility, i.e.,
as inter-laboratory co-validated Intermediate Precision or
alternatively as test method transfer performance. The Horwitz
Acceptance Criterion also offers an advantage that the precision
acceptance criteria are defined as a continuous function of c.
Additionally the Horwitz Acceptance Criterion offer a means
to objectively define inter-laboratory Intermediate Precision
(Reproducibility) Acceptance Criterion in particular for cases of
“exceedingly low” drug load drug products, or for any drug product
drug load, c2, for that matter.
In general, however, the quality control and Regulatory
reported level, and the respective specification limits of impurity,
degradation product or formulated functional minor component
would remain calculated and specified relative to the weight of
active ingredient (or relative to the intended weight of formulated
functional minor component), c1, in the formulated drug product.
412 Small-Molecule Pharmaceutical Impurities Test Method Validation

Example of Formulation Drug Load Adjusted Degradation Products

Test Method Inter-laboratory Precision Acceptance Criteria
%RSD criteria for an example of c2 = 3.38 × 10–6 weight fraction
drug load formulation
(3.38 micrograms drug per gram formulation)
Impurity or Degradation Product HorRat ≤ 2.0 Acceptance Criterion for
Level Co-Validated Intermediate Precision
(%w/w, Relative to Active (or Reproducibility),
Ingredient Weight, c1) %RSD Limit
0.50% ≤59%
0.10% ≤75%
(the drug product applicable ICH
Reporting Threshold [15])

Acknowledgments
The authors thank the many Abbott and AbbVie analytical chemists
who endeavored to develop and validate the robust pharmaceutical
test methods reflected by the test method validation data set
characterized herein. The authors also thank AbbVie New Chemical
Entity Analytical Chemistry R&D and Non-Clinical Statistics
department senior management for their support of resources
for this standing functional interface within the AbbVie Global
Pharmaceutical R&D Development Sciences CMC function.

References

1. International Conference on Harmonisation of Technical

Requirements for Registration of Pharmaceuticals for Human Use.
Validation of Analytical Procedures: Text and Methodology Q2(R1),
Step 4, November 2005.
2. Debesis, E., Boehlert, J. P., Givand, T. E., Sheridan, J. C. (1982) Submitting
HPLC methods to the compendia and regulatory agencies. Pharm.
Tech., September 1982, pp. 120–137.
3. The United States Pharmacopeia, Chapter <1210>. http://app.uspnf.
com accessed 22 September 2018.
4. Horwitz W., Kamps L. R., Boyer K. W. (1980) Quality assurance in the
analysis of foods for trace constituents. J. Assoc. Off. Anal. Chem. 63(6),
pp. 1344–1354.
 References 413

5. Thompson, M., Lowthian, P. J. (1997) The Horwitz function revisited.

J. AOAC Int., 80(3), pp. 676–680.
6. Horwitz, W., Albert, R. (2006) The Horwitz ratio (HorRat): A useful
index of method performance with respect to precision. J. AOAC Int.
89(4), pp. 1095–1109.
7. Ermer, J., et al. (2005) Precision from drug stability studies.
Investigation of reliable repeatability and intermediate precision
of HPLC assay procedures. J. Pharm. Biomed. Anal., 38, pp. 653–663.
8. Hofer J. D., Olsen B. A., Rickard E. C. (2007) Is HPLC assay for drug
substance a useful quality control attribute? J. Pharm. Biomed. Anal.,
44, pp. 906–913.
9. International Conference on Harmonisation of Technical Requirements
for Registration of Pharmaceuticals for Human Use. ICH Harmonized
Guideline. Impurities: Guideline for Residual Solvents Q3C(R6),
Current Step 4 version, October 20, 2016.
10. International Conference on Harmonisation of Technical Requirements
for Registration of Pharmaceuticals for Human Use. ICH Harmonized
Guideline. Guideline for Elemental Impurities Q3D, Current Step 4
version, December 16, 2014.
11. International Conference on Harmonisation of Technical Requirements
for Registration of Pharmaceuticals for Human Use. ICH Harmonized
Guideline. Assessment and Control of DNA reactive (Mutagenic)
Impurities in Pharmaceuticals to Limit Potential Carcinogenic Risk.
M7(R1), Current Step 4 version, March 31, 2017.
12. ICH Topic Q3B, Impurities in New Medicinal Products, Step 4,
Consensus Guideline, November 6, 1996.
13. Crowther J. B., Jimidar M. I., Niemeijer N., Salomons P. (2000)
Qualification of laboratory instrumentation, validation, and transfer
of analytical methods. In Analytical Chemistry in a GMP Environment:
A Practical Guide. eds., Miller J. M., Crowther J. B., Wiley, New York,
pp. 423–458.
14. International Conference on Harmonisation of technical requirements
for registration of pharmaceuticals for human use. ICH harmonized
tripartite guideline. Impurities in New Drug Substances Q3A(R2),
Current Step 4 version, October 25, 2006.
15. International Conference on Harmonisation of technical requirements
for registration of pharmaceuticals for human use. ICH harmonized
tripartite guideline. Impurities in New Drug Products Q3B(R2),
Current Step 4 version, June 2, 2006.
414 Small-Molecule Pharmaceutical Impurities Test Method Validation

16. Gilchrist, W. G. (2000) Statistical Modelling with Quantile Functions,

Chapman & Hall/CRC Press.
17. Koenker, R., Hallock, K. F. (2001) J. Econ. Perspect. 15(4), pp. 143–156.
Chapter 15

Method Development Strategies for

Ion Exchange Chromatography Using pH
Gradient Separation for mAbs, Antibody
Drug Conjugates (ADCs) and Other
Complex Bio-Pharmaceuticals

Norman L. Fischer and Scott Allen

Biotherapeutics Pharm Sci, Pfizer Inc.,
700 Chesterfield Parkway West, Chesterfield, Missouri 63017, USA
[email protected]

15.1 Introduction
Recombinant proteins and monoclonal antibodies (mAbs),
although relatively complex, are becoming even more complex
with the advent of mAb conjugates (mAb + conjugated peptide or
protein) and antibody drug conjugates (ADCs) consisting of a
mAb + conjugated small molecules. These more complex molecules
are typically characterized by an array of orthogonal analytical

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
416 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

and biochemical methods which are repeatedly used as part of

the ongoing control system that ensures product quality and
manufacturing consistency throughout the commercial life of the
product [1, 2]. Ion exchange chromatography (IEC) has long been
one of the primary separation modes used for characterization
and stability monitoring of the charge heterogeneity within
these molecules. This approach has been used to help identify
and characterize product heterogeneity, including C-terminal
processing of lysine residues [3–5], N-terminal pyroglutamate
formation [6, 7], deamidation [8, 9] glycation (non-enzymatic
glucose addition) [10], amino acids sequence variation [4], and
non-covalent complexes [11]. With the increasing complexity of
these types of biological therapeutics, subsequent improvements
in separation technology are becoming necessary to meet
the additional challenges associated with the characterization
and stability monitoring of these molecules.
IEC chromatography represents a commonly used method for
profiling charge heterogeneity of mAbs and ADCs, and is the most
suitable approach for purification of charged variants to support
these characterization and validation activities. A 2008 review of
charge heterogeneity within mAbs, by Vlasak and Ionescu [12],
called cation exchange chromatography the gold standard for
charge-sensitive antibody analysis. Multiple approaches to this
type of separation exist including salt gradient, chromatofocusing,
and pH gradient. While salt gradient separations are most
prevalent, and may demonstrate sufficient resolving power and
robustness, they tend to be product specific and can be time
consuming to develop, requiring optimization of multiple factors
including mobile phase buffer and pH, column, mobile phase
additives, and salt concentration gradient. Additionally, the
sensitivity, resolution, and reproducibility of these salt gradient
separations may be unknowingly compromised due to several
factors such as buffer choice, buffer concentration, and subsequent
pH adjustment.
The approaches discussed in this chapter will focus on
strategies around mobile phase selection for the development
of pH gradients. This chapter will also cover the evaluation
and preparation of mobile phases, and incorporate molecule
information, such as molecule type and pI into method
 Introduction 417

development of pH gradients. The theoretical pI of a molecule

is determined from the number of acidic and basic residues in
the molecule and is representative of the pH at which the
molecule has a net charge of zero. In theory, during pH gradient
separations, the pH at which a molecule is eluted from the
stationery phase is similar to the pI of the molecule. Additionally,
in this chapter, mobile phase gradient strategies will also be
discussed which will provide a basis for developing both sensitive
and robust separation of the more complex molecules such as
mAbs, ADC’s, and fusion proteins.

15.1.1 Salt and pH Gradient IEC Separations

The basic principle of IEC separation, as it relates to this chapter,
uses electrostatic interactions between charged variants (of a
bio-molecule) and an oppositely charged stationery phase
(column). The charged variants of the bio-molecule are bound, to
the column, via electrostatic interactions which are facilitated by
the mobile phase pH and ionic strength. The charged variants are
eluted using mobile phases with increasing levels of counter ions
to either displace the molecule (salt gradient) or alter the charge
of the molecule (pH gradient) to enable separation of the charged
variants from the stationery phase.
Ion exchange chromatography can be accomplished with
mobile phases using salt gradients at a fixed pH, a pH gradient, a
single-buffer ionic strength gradient, or various combinations of
each. In general, the distinction between salt gradient at a fixed
pH and a pH gradient is as follows:
Salt Gradient: The salt gradient uses mobile phases at a
fixed pH with increasing levels of a counter ion, typically
sodium or chloride from NaCl, to displace the molecule
from the stationery phase.
pH Gradient: The pH gradient uses a change in mobile
phase pH to alter the charge state of the molecule in order
to facilitate its release from the stationery phase. However,
depending on how the mobile phases are prepared, the
elution of the charged variants, in pH gradient separations
may occur through a combination of both a change in
molecule charge state and displacement mechanisms similar
418 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

to salt gradients. In this chapter, we will focus on method

development of pH gradient and single-buffer ionic strength
separations due to their flexibility and ease of development.

15.2 Approaches to pH Gradient Separations

This chapter will focus on IEC method development using pH
and ionic strength gradients, since these strategies offer great
flexibility and represent the simplest approach with regard
to method development. The use of pH and ionic strength
gradients also provides a good foundation for developing more
appropriate mobile phases relative to the needs of the molecule.
This strategy starts with a dilute buffer solution suitable for
retaining the molecule, then gradually shifting the charge of the
mobile phase, in the form of pH and or ionic strength increases,
as necessary, to facilitate the separation. In this way, the addition
of unnecessary ions, which can compromise the resolution of the
separation, is limited. This simple approach to buffer selection
and mobile phase development will provide the basis for sensitive
yet robust IEC separations of these more complex molecules.
The development of pH gradients for use in IEC separations
can be achieved in several different ways.

15.2.1 Individual Acidic and Basic Buffers/No pH

Adjustment
Definition: The use of multiple buffers covering a wide pH
range (e.g., 4.5–10.5) consisting of a set of acidic buffers (e.g.,
acetic acid, 2-(N-morpholino)ethanesulfonic acid (MES), (4-(2-
hydroxyethyl)-1-piperazineethanesulfonic acid) (HEPES), ((1,1-
dimethyl-2-hydroxyethyl)amino)-2-hydroxypropanesulfonic acid
(AMPSO), and N-cyclohexyl-3-aminopropanesulfonic acid CAPS)
and a corresponding set of basic buffers (e.g., 1-methylpiperazine,
pyridine, Bis-Tris propane, AMP, piperidine) which cover a similar
pH range.
The advantage of this approach is that a relatively wide pH
range can be achieved which is useful as a platform method,
or for separation of mAb mixtures with a wide variance of pIs.
 Approaches to pH Gradient Separations 419

Additionally, for this approach using both acidic and basic buffers,
no pH adjustment is required. A disadvantage of this approach
is that care must be taken to ensure that ionic strength of the
individual buffers within the mobile phases are minimized to
avoid compromising the resolution of the separation due higher
ionic buffer strength from the combined buffers. Another
disadvantage of this approach is that it requires multiple
acidic buffers and multiple basic buffers, making mobile phase
preparation a more complex process requiring multiple acidic and
basic buffers. Additionally, more experiments are required during
development to insure that the buffering strength of the acidic
and basic buffers is appropriately balanced so that a reasonably
stable pH gradient is achieved. In general, the use of a wider pH
range may compromise the resolution of the separation, and
will generally require the use of shallower gradients to achieve
adequate separation. This is due to the increased rate of pH
change across the stationery phase associated with the wider pH
gradient. However, this drawback is easily overcome by reducing
the pH range of the gradient. The advantages of narrower pH
gradients will be discussed later in this chapter.

15.2.2 Combined Buffers with pH Adjustment

Definition: The use of multiple buffers covering a wide pH range
consisting of a set of either acidic or basic buffers combined in
a single bulk mobile phase which is then split and pH adjusted
to cover the buffering range of the buffers in the mobile phase.
This approach is presented as an alternative to the approach
listed in option one. In this case, only the acidic or basic buffers
would be used for both mobile phases (e.g., Acetic Acid, MES,
HEPES, AMPSO, CAPS at pH 4.5 and pH 10.5). This approach
would significantly reduce the number of buffers required for the
mobile phase preparation. The advantage of this approach over
the first is that the mobile phase development and preparation
are simpler, since the same buffers are used for both mobile
phases. As with the first option, the use of a wider pH range is
useful as a platform method, or for separation of mAb mixtures
with varying pIs. The disadvantages of this approach are similar
to those listed in option 15.2.1. Also, care must be taken to ensure
420 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

that the ionic strength of the individual buffers within the mobile
phases is minimized to avoid compromising the resolution
of the separation due higher ionic strength from the combined
buffers and subsequent pH adjustment. As with option 15.2.1
the drawbacks associated with a wider pH gradient range are
easily overcome by reducing the pH range of the gradient in the
mobile phases. This approach is outlined in the following section.

15.2.3 Buffers Covering a Narrower pH Range

Definition: The use of fewer buffers covering a narrower pH
range consisting of a single acidic and or basic buffer in each
mobile phase so that the pH and buffering range of the mobile
phases covers the pI of the molecule.
This option represents a scaled down version of the first
two options. The advantage of this approach over the first two is
that the mobile phase development and preparation are the
simplest, since fewer buffers are used for the mobile phases.
The mobile phases can use either a single buffer which is pH
adjusted to cover the desired pH range, (e.g., HEPES adjusted to
pH 6.8 and pH 8.2), or an acidic buffer in one mobile phase and a
basic buffer in the other mobile phase to generate the pH gradient.
Because of the narrower pH range, this approach will generally
yield better resolution. The primary disadvantage of a narrower
pH range is this separation may not be suitable for separating
multi-mAb mixtures if all of the mAb pIs do not fall within the
pH range of the buffers.
In this chapter, we will provide illustrations from these
options with the primary focus on the development of narrower
pH gradients (option 15.2.3), as well as ionic strength gradients
since these approaches are the simplest, from a development
standpoint, and may yield greater resolution and sensitivity than
the other approaches. We will also provide further discussion
on when the use of ionic strength gradients is most appropriate.

15.2.4 General Concept of pH Gradient Separations

Figure 15.1 provides a simple illustration of the concept of pH

gradient separations. Using a pH gradient approach for cation
exchange, the balance of positively charged H+ ions from the
 Impact of Ions on IEC Separations 421

acidic buffer in mobile phase A is shifted toward increasing levels

of electronegative ions from the basic buffers in mobile phase B.
This charge shift occurs through the delivery of the pH gradient
via the HPLC pump and serves to alter the charge of the bound
molecule on the stationery phase, from positive to negative for
cation exchange and negative to positive for anion exchange, thus
facilitating elution and separation of the charged variants.
For cation exchange separations, the starting pH of mobile phase A
will be below the pI of the molecule, and the pH of mobile phase B
will be above the pI of the molecule. For anion exchange, the
opposite is true, with the pH of mobile phase A being above the
pI of the molecule, and the pH of mobile phase B below the pI of
the molecule.
Chapter 15, Figure 15.1

Mobile Phases Consist of a Single Buěer or Group of Buěers Consisting of Weak A cids
or Bases Which BžěŽr Cover a Specięed pH Range
11
Buěer pH adjusted above molecule pI
10

pH gradient generated from

9
HPLC pump through mixing of
8 mobile phase buěers
Gradient pH

6
Buěer pH adjusted below molecule pI
5
5.5 6.5 7.5 8.5 9.5 10.5
Mobile Phase Buěering Range (e.g. pH 6 -10)

Figure 15.1 General concept of pH gradient separations.

As with salt gradients, the addition of ions through pH

adjustment can adversely impact resolution of the separation
with pH gradients as well.

15.3 Impact of Ions on IEC Separations

For IEC separations in general, an inverse correlation exists
between mobile phase ionic strength and molecule retention
time. For example, an increase in ionic strength of the mobile
phases will decrease molecule retention time while a decrease in
422 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

mobile phase ionic strength will increase molecule retention time.

The ion concentration in the mobile phase must be sufficient
to elute the molecule, but not so high as to cause poor retention
or resolution of the charged variants. Excessive ions in the
mobile phase will necessitate the use of shallower gradients to
compensate for the increased ion concentration in order to
achieve adequate separation. The need for use of shallow gradients
(< 0.5%/min), to achieve adequate separation, may be indicative
of excessive ions in the mobile phases that could result in
method robustness problems including significant retention time
shifts and changes in or loss of resolution. These problems with
method robustness are, in most cases related to a combination
of excessive ions in the mobile phase along with variance in
mobile phase preparation, and not the inability of the pump
to accurately deliver a mobile phase gradient of < 0.5 %/min.

15.3.1 Impact of Ions from pH Adjustment on Salt

Gradient Separations
The retention time shift associated with variations in mobile
phase preparation is illustrated in the experiment shown in
Fig. 15.2, for a cation exchange separation using a salt gradient.
The molecule in this experiment is a mAb with a theoretical pI
of ~8.1. In this experiment the mobile phases consisted of
10 mM HEPES, pH 8.0 ± 0.1 using 6N NaOH (to adjust pH) for
mobile phase A, and mobile phase B the same as A with
200 mM NaCl added. The gradient slope was 1%/min.
In this method, the retention time shift associated with the
allowable pH adjustment (pH 8.0 ± 0.1) in the method could
result in a retention time shift of approximately 6 min,
from 26–32 min. Additionally, the retention time shift between
pH 7.8 and 8.2 is also associated with a loss in resolution of the
acidic species. This experiment highlights the importance of
maintaining control of ion concentrations in the mobile phases
which are added during pH adjustment of the mobile
phases. This control is essential to develop robust methods which
deliver consistent results. The following experiment illustrates
the impact of additional of ions from pH adjustment for pH gradient
separations.
 Impact of Ions on IEC Separations 423
Chapter 15, Figure 15.2

0.10
µL NaOH RT µL NaOH RT µL NaOH RT µL NaOH RT
1400 22.9 1050 34.3 700 41.2 350 47.9
0.08 M.P. pH
pH = 8.2
8.2 M.P. pH
pH = 7.8
7.8 M.P. pH
pH = 7.4
7.4 M.P. pH
pH = 7.0
7.0

0.06
AU

Allowable
AllowablepH pH
AU

0.04 Range
RangeforforMethod
Method
pH
pH8.0
8.0±±0.1
0.1

0.02

0.00

10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 55.0 60.0
Minutes
Minutes

Figure 15.2 Impact of ion level in mobile phases associated with pH

adjustment.

15.3.2 Impact of Ions from pH Adjustment on pH

Gradient Separations
In this experiment, four molecules were separated with identical
IEC pH gradient methods using pH adjusted buffers and non
pH-adjusted buffers that covered a comparable pH range. The
purpose of this experiment was to evaluate the impact of ions
from pH adjustment on the resolution of the separation.
Three IgG2 mAbs and one IgG1 mAb-conjugate were evaluated
in this experiment. The theoretical pIs of these molecules
(labeled as mAb 1–4) range from 7.7–8.1.
Three buffers, listed in Table 15.1, were used for this
experiment. Two of the buffers were acidic and one was basic.
The buffering range for both pH adjusted and non-pH adjusted
mobile phases was comparable ranging approximately from pH
6.5–9.5. The mobile phase preparations were evaluated in the
cation exchange (CEX) separation mode. In this experiment,
2 combinations of mobile phases were evaluated as listed in
Table 15.2. Combination 1 uses pH adjustment and combination
2 does not use pH adjustment. Mobile Phase A for both combinations
is the same, consisting of 5 mM HEPES and 5 mM AMPSO, with
no pH adjustment. Mobile Phase B in combination 1 is 5 mM
HEPES and 5 mM AMPSO, pH adjusted to 9.3 with 2.5N NaOH.
424 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

Mobile Phase B in combination 2 is 5 mM Bis-Tris propane

without any pH adjustment. The sodium concentration associated
with the addition of NaOH to mobile phase B, in combination 1,
amounts to 9.4 mM. The mobile phase combinations are shown
in Table 15.2. For each mobile phase combination, a pH titration
curve was generated as shown in Fig. 15.3. The pH titration
curves for combinations 1 and 2 are nearly identical, and if
evaluated solely on pH gradient, one would expect that the
mobile phases with and without pH adjustment would yield
similar chromatographic profiles when used with the same method.

Table 15.1 Mobile phases for pH adjustment experiment

Mobile phase # Buffer description

Mobile phase 1 5 mM HEPES, 5 mM AMPSO, pH adjusted with 3.75 mL
of 2.5N NaOH
Mobile phase 2 5 mM Bis-Tris propane, with no pH adjustment
Mobile phase 3 5 mM HEPES, 5 mM AMPSO, with no pH adjustment

Table 15.2 Mobile Phase combinations to assess influence of pH

adjustment with NaOH on separation

Combination # pH adjustment Mobile phase A Mobile phase B

Combination 1 Yes Mobile phase 3 Mobile phase 1
Combination 2 No Mobile phase 3 Mobile phase 2

The column used in this experiment is a Dionex ProPac

WCX-10, 4 mm × 250 mm, 10 µm. The chromatographic profiles
for separation of mAbs 1–4, with all mobile phase combinations
are shown in Fig. 15.4. The separation with combination 2
(non-pH adjusted) is consistently better than that obtained with
combination 1 (pH adjusted).
This result suggests that the resolution difference between
pH adjusted and non-pH adjusted mobile phases may be related
to the addition of ions via pH adjustment of the mobile phases
and not a difference in the pH gradient. In this case, the pH
adjustment with NaOH may be facilitating molecule displacement
similar to a salt gradient, in addition to altering the molecule
charge on the stationery phase via the pH gradient. Also, this
result further suggests that minimizing the addition of ions from
 Impact of Ions on IEC Separations 425

pH adjustment in the mobile phases is a plausible strategy for

Chaptersignificantly improving the separation.
15, Figure 15.3

CEX pH Gradient (6.5 - 9.5): Comparison of Mobile

Phases With and Without pH Adjustment
10
9.5
9
8.5 No pH
8 adjustment
pH

7.5
pH
7 adjustment
6.5
6
5.5
0 20 40 60 80 100
% Mobile Phase B

Figure 15.3 The pH titration curves for combination 1 (pH adjustment)

and combination 2 (no pH adjustment) are shown. The pH gradients from
the two approaches are nearly identical.

15.3.3 Impact of Ions from Buffer Concentration on pH

Gradient Separations
The chromatographic profiles shown in Figures 15.2 and 15.4
A–D illustrate the potential impact of additional ions in the mobile
phases associated with pH adjustment for both salt and pH
gradients. The following experiment demonstrates the impact
of buffer concentration on a pH gradient separation using cation
exchange.
In this experiment, two sets of mobile phases were prepared
without pH adjustment using acidic buffers in mobile phase
A (HEPES and AMPSO) and a basic buffer in mobile phase B
(Bis-Tris Propane). The first buffer set comprised mobile phases
with buffer components at a concentration of 5 mM while the
second buffer set comprised the same acidic and basic mobile
phases with the buffer components at a concentration of 50 mM.
The mobile phase compositions are listed in Table 15.3.
426 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals
Chapter 15, Figure 15.4

0.20
mAb 1 A
CombinĂƟŽn 1
AU

0.10
(pH adjusted)
0.00

0.15
mAb 1
0.10
CombinĂƟŽn 2
AU

0.05
(non-pH adjusted)
0.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 40.00 42.00 44.00
Minutes
0.20
mAb 2 B
CombinĂƟŽn 1
AU

0.10

(pH adjusted)
0.00

0.15
mAb 2
0.10
CombinĂƟŽn 2
AU

0.05 (non-pH adjusted)

0.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 40.00 42.00 44.00
Minutes

0.20
mAb 3 C
CombinĂƟŽn 1
AU

0.10
(pH adjusted)
0.00

0.20
mAb 3
CombinĂƟŽn 2
AU

0.10
(non-pH adjusted)
0.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 40.00 42.00 44.00
Minutes

0.20
mAb 4 D
CombinĂƟŽn 1
AU

0.10
(pH adjusted)
0.00

0.15
mAb 4
0.10
CombinaƟon 2
AU

0.05 (non-pH adjusted)

0.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 38.00 40.00 42.00 44.00
Minutes

Figure 15.4 (A–D) Three mAbs (labeled 1–3) and one mAb conjugate
(labeled 4) were compared using combination 1 and combination 2.
The separation is improved for all molecules using combination 2, the
non-pH adjusted buffers.
 Impact of Ions on IEC Separations 427

Table 15.3 Mobile phase composition for ionic strength experiment

Acidic components (A) Basic components (B)

5 mM HEPES, 5 mM AMPSO 5 mM BIS-TRIS propane
Buffering range 6.8–9.7 Buffering range 6.3–9.5
50 mM HEPES, 50 mM AMPSO 50 mM BIS-TRIS propane
Buffering range 6.8–9.7 Buffering range 6.3–9.5

The column used in this experiment is a Dionex ProPac

WCX-10, 4 mm × 250 mm, 10 µm. A gradient slope, of 2.4%/min,
was used in this experiment. Mobile phase compositions were
adjusted to yield buffer concentrations of 5, 10, 25, 40, and
50 mM for each buffer component of the mobile phase.
Figure 15.5 shows the influence of mobile phase ionic strength
on the resolution of charged variants from mAb 2 in Fig. 15.4B.

Figure 15.5 Chromatographic overlay showing influence of mobile phase

ionic strength on a pH gradient separation of a monoclonal antibody.

For this experiment in general, the acidic components were

best resolved at lower concentrations of (5–10 mM) while the
basic species were resolved better at the higher concentrations
(25–50 mM). As a result, another condition was evaluated at a
buffer concentration of 16 mM, using the same gradient slope
to evaluate resolution. The chromatography is shown in Fig. 15.6.
428 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals



 ĞĐƌĞĂƐĞĚZĞƐŽůƵƚŝŽŶŽĨĐŝĚŝĐ^ƉĞĐŝĞƐ /ŵƉƌŽǀĞĚZĞƐŽůƵƚŝŽŶŽĨĂƐŝĐ^ƉĞĐŝĞƐ


$8






         
0LQXWHV 

Figure 15.6 mAb separation at 16 mM with gradient slope of 2.4%/min,

showing improved separation of the basic species relative to 10 mM
buffer concentration.

The chromatographic profile from the 16 mM buffer

concentration yielded the best overall separation and proved
to be a good compromise between the lower and higher buffer
concentrations, providing necessary balance between resolution
of acidic and basic species with a gradient slope of 2.4%/min.
Using the buffer concentration of 16 mM, the gradient slope was
reduced to 1.2%/min, and lower buffer concentrations down
to 5 mM were again evaluated. The profile from the optimized
conditions for this separation is shown in Fig. 15.7. Optimization
of buffer concentration and gradient slope resulted in acidic
and basic species which are more resolved from the previous
separation conditions.

Figure 15.7 Optimized separation at 7 mM buffer concentrations and

gradient slope of 1.2%/min.
 Impact of Ions on IEC Separations 429

This experiment demonstrates the relationship between

species charge and ionic buffer strength and shows how buffer
concentration and gradient slope can be used to optimize the
separation of acidic and basic species. It further illustrates a strategy
for identifying method conditions which yield suitable separation
of both the acidic and basic species.
While further optimization activities could be performed,
assessment of total area% for acidic, main, and basic peaks for
the various separation conditions evaluated, are shown in Table
15.4. The results for the area% for acidic, main, and basic species
show the robustness of this separation to be good at various buffer
concentrations ranging from 5–16 mM, and gradient slopes of
1.2%–2.4%/min with different run times.

Table 15.4 Area% data for acidic, basic, and main peak at various
separation conditions

Buffer
Run time conc. Grad. slope
Name (Min) (mM) (%/min) %Acidic % Main % Basic
mAb 2 45 16 2.4 35.64 37.63 26.73
mAb 2 75 16 1.2 34.40 37.46 28.14
mAb 2 75 10 1.2 36.80 36.68 26.52
mAb 2 75 7 1.2 36.12 37.33 26.56
mAb 2 75 5 1.2 37.06 36.89 26.05
mAb 2 45 7 1.2 36.72 37.36 25.92
Mean 36.12 37.23 26.65
Stdev 0.99 0.36 0.79
%RSD 2.74 0.97 2.98

This experiment illustrates the influence of the mobile phase

strength on the separation, and provides a viable strategy for
evaluating and establishing the optimal conditions for separating
the molecule. After identifying the appropriate mobile phase
conditions, the HPLC pump can be used to assess different slope
conditions to achieve optimal separation of both acidic and basic
variants.
430 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

15.3.4 Further Evaluation of Buffer Concentration

A titration curve of the pH gradient, for the 5 mM buffers and the

50 mM buffers used in Fig. 15.5, is shown in Fig. 15.8. The
titration curves show nearly identical pH values at various buffer
concentrations throughout the curve, and if evaluated solely on
pH gradient, one would expect that the mobile phases at both
concentrations would yield similar chromatographic profiles
when used with the same method.
Chapter 15, Figure 15.8

Comparison of pH Gradients from Mobile Phases Using IdenƟcal

Buīer Components at ConcentraƟons of 5mM and 50mM
11

10

5mM
8
50mM
pH

5
0 20 40 60 80 100
% Mobile Phase B

Figure 15.8 Titration curve showing pH gradients from mobile phases

using the same buffers at 5 and 50 mM concentration.

The consistency of the titration curves at 5 and 50 mM buffer

concentrations suggests the significant retention time shift
(16.2 min in Fig. 15.5) is not related to differences in mobile
phase pH but rather driven by the ionic strength of the mobile
phase being delivered by the HPLC pump. The results from this
experiment show that the ionic strength required to facilitate
the elution of the molecule from the stationary phase, occurs
in an earlier part of the pH gradient due to increased buffer
concentration in the in the mobile phases.
The overlay in Fig. 15.5 shows the impact of mobile phase
ionic strength on resolution of acidic and basic species. However,
 Impact of Ions on IEC Separations 431

there is also a clear impact of ionic strength on the elution

pH of the molecule. The information in Table 15.5 contains data
from the separation shown in Fig. 15.5.

Table 15.5 Impact of buffer concentration on retention time and

subsequent elution pH from Fig. 15.5

Mab # Buffer Main Peak %B (Based

(Table Conc. Rt, (Min) On Main Elution pH*
15.1) Molecule Pi (Mm) (Fig. 15.5) Peak Rt) (Fig. 15.5)
2 8.1 5 25.1 78 9.1
2 8.1 10 24.6 77 9.0
2 8.1 25 17.1 59 8.3
2 8.1 40 11.3 45 7.8
2 8.1 50 8.9 39 7.6

*Elution pH is calculated from main peak RT data in Fig. 15.5 with the gradient slope
of 2.4%/min to determine the mobile phase composition and subsequent elution
pH which is calculated from the average slope – intercept data from Fig. 15.8.

15.3.5 Buffer Composition

The buffers used in this experiment are zwitterionic, which are
neutral molecules containing both positive and negative charges.
The list and number of possible charged groups for each molecule
are listed in Table 15.6.

Table 15.6 Potential charged groups for zwitterionic buffers (HEPES,

AMPSO, and Bis-Tris propane

Sulfonic acid (RSO3H) Hydroxy groups Amine groups (N)/

Buffer groups/molecule (OH)/molecule molecule
HEPES 1 1 2
AMPSO 1 2 1
Bis-Tris 0 6 2
propane

Based on the number of potentially charged groups per

molecule in Table 15.6 along with the calculated mobile phase
composition at the elution pH in Table 15.5, one can calculate
432 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

the approximate concentration of charged groups at the point of

elution based on main peak retention time. Figure 15.9 shows
the sulfonic acid (RSO3H) groups from Table 15.5 as a percentage
of total charged groups in the mobile phases at the calculated
elution pH. This result shows a correlation between increased
sulfonic acid groups and lower calculated pH at the point of
elution. This result is to be expected considering the mobile phase
composition at the time of elution. However, this result is contrary
to the general expectation with pH gradient separations where
elution is expected at a pH near the pI of the molecule.

Figure 15.9 Impact of sulfonic acid concentration on elution pH at

5–50 mM buffer concentrations.

This data further confirms that pH is not the primary

factor driving molecule elution in this separation. A closer look
at the concentration of all possible charged groups is shown in
Fig. 15.10.
This figure shows increased sulfonic acid groups at the lower
retention times associated with increased levels of mobile phase A
at the higher buffer concentrations. However, under these
conditions, the concentration of the amine (N) and hydroxy (OH)
groups are more prevalent due to the higher buffer concentrations
along with the fact that the mobile phase comprises two buffers
 Impact of Ions on IEC Separations 433

which both contain these groups. It is plausible that the increased

levels of these groups are facilitating the decrease in retention
time in this experiment and may further explain the result of
decreasing elution pH as buffer concentration is increased. Given
the fact that there are a finite number of charged groups in the
sample matrix based on the primary structure and concentration
of the molecule, one could speculate that only a certain number
of oppositely charged groups in the mobile phase are necessary
to alter the charge on the molecule to facilitate its elution earlier
in the separation. This result further highlights the importance of
the HPLC pump in combination with minimizing ions in the mobile
phases, and thereby facilitating delivery of ions to the stationery
phase at the appropriate rate to yield optimal resolution during the
separation.

Figure 15.10 Impact of concentration of charged groups on molecule

retention time in CEX pH gradient.

This section illustrates the potential impact of excess ions

in the mobile phase associated with pH adjustment or excessive
buffer concentration. It also provides a framework for experimental
strategies to assess and optimize mobile phase concentrations
along with gradient slope to achieve a robust separation. The
434 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

following sections will begin to incorporate specific molecule

information into the method development strategy in order to
quickly develop a robust method with sufficient resolution of
acidic and basic species.

15.4 Molecule Information and Method

Development
Before starting the development of a pH gradient IEC separation
method, it is important to gather information about the molecule
in order to establish optimal separation conditions. The following
information will be helpful during the method development
process.
(1) pI of the molecule
(2) Molecule type: mAb (IgG1, IgG2) ADC (antibody drug
conjugate), fusion protein, vaccine, etc.
(3) Molecular weight
(4) Presence and abundance of glycosylated species
These molecule characteristics can be useful in guiding several
aspects of the method development including buffer selection
for mobile phases, establishing buffer strength of mobile phases,
selection of the appropriate separation mode (anion or cation), and
whether or not to consider incorporating alternative separation
strategies into the method. The information listed above will be
discussed in further detail later in this chapter.

15.4.1 Mobile Phase Considerations for pH Gradient

Separations
This section will look further at buffers and mobile phases with a
focus on mobile phase composition/concentration as it relates
to mAb subclass and pI. Mobile phase composition is most
important, since this will have the greatest impact on resolution
of the charged variants with pH gradient (IEC) separations. If care
is not taken to select the appropriate buffer and optimize the
mobile phases during method development, the separation may
be compromised to the point where altering any of the other
 Molecule Information and Method Development 435

possible variables (e.g., stationary phase, sample diluent, column

temperature, and gradient slope) will have minimal or no positive
impact. Additionally, poorly developed mobile phases will result
in methods which can be difficult to validate due the extreme
requirements placed on the instrument to compensate for
excessive ions in the mobile phases. A common example of this
are methods which use very shallow gradients (<0.5%/min) in
combination with mobile phases containing high levels of ions.
In most cases, a small portion of the mobile phases eluting
capacity is used for the separation, typically occurring within a
small percentage of the gradient, over a longer period of time. As a
result, a shallow gradient becomes necessary to compensate for the
excessive ions in the mobile phase in order to achieve adequate
resolution. Additionally, these methods may exhibit significant,
run to run, shifts in retention time along with inconsistent peak
resolution.
The first step in mobile phase development is choosing a
buffer with a buffering range that will bracket the pI of the
molecule being separated. The list of buffers for use in pH gradient
separations shown in Tables 15.1 and 15.7, is by no means
exhaustive, but provides some examples of suitable buffers for
pH gradient separation which will cover the pH range of
most mAbs.

Table 15.7 List of example buffers suitable for use in pH gradient

separations

Buffer pH range
MOPSO 6.2–7.6
HEPES 6.5–7.9
AMPSO 8.3–9.7
BIS-TRIS propane 6.3–9.5
Ammonium acetate 3.8–5.8
Ammonium formate 2.8–4.8

15.4.2 mAb pI, Subclass, and Buffer Concentration

The approach described in this chapter uses low molecular
weight acidic and or basic buffers to generate a pH gradient. The
436 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

buffers do not need to be pH adjusted if one of the buffers in the

separation is acidic and the other is basic (e.g., HEPES, BIS-TRIS
propane). However, if a single buffer is used for both mobile
phases, then at least one of the mobile phases will need to be
adjusted so that the mobile phase pH range covers the pI of
the molecule. It is acceptable to adjust the pH in both mobile
phases. However, the amount of acid or base used to adjust the pH
of mobile phase should be minimized, keeping in mind that the
mobile phase pH will need to bracket the pI of the molecule. Some
examples of this will be shown later in this chapter. In either case,
the pH is generated through the HPLC mixing of acidic and basic
mobile phases.
The primary goal, as it relates to mobile phase development
for IEC separations, is to minimize excessive ions in the
mobile phases to improve the sensitivity and resolution of the
separation. This is critical in developing a highly resolving,
sensitive and robust method for separation of these more complex
bio-pharmaceuticals. At this point, we will begin to discuss
buffer concentration with regard to mAb subclass and molecular
weight.
The isoelectric point (pI) of the molecule is a critical piece
of information which is important to the buffer selection and
should also be considered with regard to the separation mode
(cation or anion) that is used. In theory, for pH gradient separations,
the pI of a molecule should correlate somewhat with the elution
pH of the molecule. However, the data in Table 15.5 shows the
calculated elution pH at the various buffer concentrations to be
well above and below the pI of the molecule. This indicates that
buffer selection and concentration plays a greater role in molecule
elution than buffer pH.
Additionally, there are other factors which dictate the elution
pH of a molecule.
Figure 15.11 shows a graphical illustration from an actual
separation of an IgG1 and an IgG2 mAb with identical theoretical
pIs that were run with the same pH gradient method.
Table 15.8 shows the retention time and elution pH data
from the illustration in Fig. 15.11.
 Molecule Information and Method Development 437

Figure 15.11 Graphic Illustration from an actual separation of two different

mAbs with the same theoretical pI of 7.3 using the same pH gradient
method (actual profiles not disclosed).

Table 15.8 Retention time and elution pH information from 2 different

mAbs with the same theoretical pI of 7.3, which were run with
Bis-Tris propane mobile phases at 5 mM

mAb Buffer conc. Acidic RT Main RT Basic RT Elution

subclass pI (mM) (min) (min) (min) pH
IgG1 7.3 5 mM 30.0 31.0 32.5 8.3
IgG2 7.3 5 mM 12.5 16.5 25.0 7.4

In this experiment, both mAbs were run using the same pH

gradient (cation exchange) separation method. The mobile phases
used 5 mM Bis-Tris propane buffers, with mobile phase A pH
adjusted to 6.3 and mobile phase B at pH 9.5 (non-pH adjusted).
The pH gradient was run at 2%/min. The column used in this
experiment is a Thermo MAbPac SCX-10, 4 mm × 250 mm, 10 µm.
The significant retention time difference between these 2 mAbs
with identical pIs suggests that there are additional molecule
related factors which are also playing a role in the retention of the
438 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

charged variants during pH gradient separations. These factors

can include differences such as primary structure, amino acid
composition, and conformation of the molecule. In this case, the
early eluting mAb is an IgG2 and the late eluting mAb is an IgG1.
This data suggests that characteristics associated with mAb
subclass may play a role on retention time of the molecule.
Figure 15.12 shows the graphical illustration, and Table 15.9
shows retention time and elution pH data from the same molecules
in Fig. 15.11 and Table 15.8 separated using the same method, but
at an increased buffer concentration consisting of 10 mM Bis-Tris
propane.

Figure 15.12 Graphic Illustration from an actual separation of two different

mAbs with the same theoretical pI of 7.3, which were run using the same
pH gradient (cation exchange) method with Bis-Tris propane mobile
phases at 10 mM (actual profiles not disclosed).

Table 15.9 Retention time and elution pH information from two different
mAbs with the same theoretical pI of 7.3, which were run with Bis-Tris
propane mobile phases at 10 mM

mAb Buffer Conc. Acidic RT Main RT Basic RT

subclass pI (mM) (min) (min) (min) Elution pH
IgG1 7.3 10 mM 27.75 29.25 30.5 8.1
IgG2 7.3 10 mM 3.75 4.75 6.75 6.6
 Molecule Information and Method Development 439

In this experiment, the same two mAbs from Table 15.8

were run using the same pH gradient (cation exchange) method.
However the mobile phase concentration was increased from
5 mM to 10 mM. Mobile phase A was adjusted to pH 6.3 and Mobile
phase B was pH 9.5 (non-pH adjusted). The pH gradient was run
at 2%/min using the same column as shown in Table 15.8.
Results from Tables 15.8 and 15.9 show that a 5 mM increase
in buffer strength negatively impacted the retention of the IgG 2
mAb while having little impact on the retention of the IgG 1 mAb.
Also, the resolution of the IgG2 peaks was negatively impacted by
the 5 mM increase in buffer strength (actual profiles not shown).
These results from Tables 15.8 and 15.9 show the influence
of mAb subclass, on IEC separations, and how subsequent
molecule retention and resolution is impacted by mobile phase
concentration. Figure 15.13 shows the chromatographic profile
comparison of two mAbs consisting of an IgG1 (pI 8.5) and an IgG2
(pI 8.7) that were run.
In this experiment, 2 different mAbs with theoretical pIs of
8.5 (IgG1) and 8.7 (IgG2) were run using the same mobile
phase buffers, column, and separation method from Tables 15.8
and 15.9. Mobile phases used in the top overlay were at 5 mM
while mobile phases used in the bottom overlay were at 10 mM.
The buffer was Bis-Tris propane adjusted to pH 6.3 for mobile
phase A, and pH 9.5 (non-pH adjusted) for mobile phase B. The
pH gradient was run at 2%/min. Results from Fig. 15.11 show the
best separation for the IgG1 occurring at 10 mM while the best
separation for the igG2 occurred at a buffer concentration of 5 mM.
The comparison in Fig. 15.13 confirms the data shown in
Tables 15.8 and 15.9, showing that mAb characteristics related
to subclass have significant influence on molecule retention and
resolution. Tables 15.8 and 15.9, along with Fig. 15.13, show
that IgG 2’s are more sensitive to smaller changes in buffer
concentration, highlighting the need for consideration of mAb
subclass and careful preparation of mobile phases. These
results indicate that the mobile phase concentration for optimal
separation of IgG 2’s is around 5 mM and is lower than the
concentration required for optimal separation of IgG 1’s which
is around 10 mM or higher.
 440
Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

Figure 15.13 Chromatographic overlays of two different mAbs with theoretical pIs of 8.5 and 8.7. Both were run using the same mobile
phases and pH gradient (cation exchange) method used in Table 15.8 and 15.9.
 Molecule Information and Method Development 441

15.4.3 mAbs with Higher pIs and Ionic Strength

Gradients
However, in addition to mAb subclass, overall mAb charge as it
relates to the molecule pI can also influence the separation
necessitating the need for alternative strategies related to mobile
phase development and usage. Figure 15.14 shows another IgG
2 mAb with a pI of 9.1. In this experiment, a pH gradient with the
same buffer concentrations in both mobile phases was evaluated,
along with a combination approach using pH and ionic strength
gradients (e.g., 5 mM in MP A, and 10 or 20 mM in mobile
phase B). Mobile phases used 5, 10, or 20 mM Bis-Tris propane
buffers: Mobile phase A = pH 6.3 and Mobile phase B = pH 9.5
(non-pH adjusted). The pH gradient was run at 2%/min.
The column used in this experiment is a Dionex MAbPac SCX-10,
4 mm × 250 mm, 10 µm. In this case, we see that the mAb was
not eluted from the column at buffer concentrations up to 20 mM
in mobile phase B. This result is in direct contrast with the
results from Tables 15.8 and 15.9 and Fig. 15.13 for an IgG 2 mAb.
This result further illustrates the need for altering mobile
phase composition based on molecule characteristics.
Figure 15.15 shows a comparison of the different buffer
concentrations that were used in Fig. 15.14, for an IgG 1 molecule
with a pI of 8.9. In this case the molecule was not eluted at buffer
concentrations of 5 mM and 5–10 mM. At a buffer concentration
of 5–20 mM, the molecule was eluted at the end of the gradient,
indicating that conditions using higher buffer concentrations
should be evaluated.
Figure 15.16 shows optimized separation for both molecules
in Figs. 15.14 and 15.15. In this case, an ionic strength gradient
was used which consisted of low concentration, non pH adjusted
Bis-Tris propane in mobile phase A and higher concentration
non-pH adjusted Bis-Tris propane in mobile phase B. The column
used to generate the top profile in Fig. 15.16 is a Dionex MAbPac
SCX-10, 4 mm × 250 mm, 10 µm. The column used in the bottom
profile is a Waters Protein Pak, Hi Res, CM, 7 µm, 4.6 × 100 mm.
The use of an ionic strength gradient resulted in greatly improved
separations for both molecules. The advantage of the ionic
strength gradient applied in this manner is that it facilitates a very
efficient delivery of ions to the column due to the simplicity of
the mobile phases.
 442

Figure 15.14 Chromatographic overlay of a single mAb (IgG 2) with a theoretical pI of 9.1 run using a 5 mM pH gradient, or a 5–10
or 5–20 mM pH/ionic strength gradient (combination) in cation exchange separation mode.
Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals
Figure 15.15 Chromatographic overlay of a single mAb (IgG 1) with a theoretical pI of 8.9 run using a 5 mM pH gradient, a 5–10 or
5–20 mM pH/ionic strength gradient (combination) with cation exchange separation mode.
Molecule Information and Method Development
443
 444
Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

Figure 15.16 Chromatographic overlay of two mAbs from Figs. 15.14 and 15.15 (IgG 1, Top and IgG 2, Bottom) with theoretical pIs
of 8.9 and 9.1 respectively, run using a 3–35 or 3–50 mM ionic strength gradients in cation exchange separation mode.
 Molecule Information and Method Development 445

This experiment demonstrates the influence of molecule pI

on separation strategy. In this case we see that as pI increases to
an approximate pI of 9, the requirements for buffer strength also
increases significantly for both IgG 1’s and IgG 2’s. Additionally,
at the higher molecule pIs, an ionic strength gradient can be
used in place of a pH gradient. In fact, at pIs of ~9, even small
amounts of acidic buffers (HEPES and AMPSO) in the mobile
phases can significantly increase the molecule retention as well
as negatively impact the resolution. Evaluation using an ionic
strength gradient is recommended for molecules with a pI ~9, if
a suitable separation cannot be achieved using a pH gradient.

15.4.4 Ionic Strength Gradients and Fusion Proteins

The use of ionic strength gradients is appropriate for other mAb
like proteins as well. Figure 15.17 shows an overlay of the
separation of a 100 kDa fusion protein with a theoretical pI of 8.35.
In this separation a combination pH gradient/ionic strength
separation was developed using 3 mM Bis-Tris propane + 2 mL/L
of 1N HCl in mobile phase A and 60 mM Bis-Tris propane
(no pH adjustment) in mobile phase B. The gradient slope for this
separation was 4%/min. The column used in this experiment is a
Dionex MAbPac SCX-10, 4 mm × 250 mm, 10 µm. The top profile
in Fig. 15.17 was generated using the same separation, as the
bottom profile, with the addition of 5 mM NaCl in mobile phase B.
In this case the addition of 5 mM NaCl resulted in co-elution of
homo and hetero-dimer species. In the bottom chromatograph,
which did not use any NaCl in mobile phase B, we see well-resolved
separation of the hetero- and homo-dimer species.
This separation demonstrates that ionic strength and pH
gradients can be incorporated in the same method. In this case,
HCl was added to mobile phase A to facilitate retention of the
molecule at a low buffer concentration. It also shows that ionic
strength gradients can be successfully applied to molecules with
lower pIs. Another point about this separation is that the pI for
this molecule is lower (pI 8.35) but the concentration requirement
for mobile phase B was 60 mM. This is in direct contrast to
mAbs, which at a similar pI will require a much lower buffer
concentration to facilitate the separation. This requirement for
increased buffer strength may possibly be attributed to the lower
446 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

mass of the molecule (100 kDa) as compared to a mAb, or it may

be related to the overall number and availability of charged
groups, as it relates to conformation, on the molecule. Finally, it
clearly shows the negative impact on resolution that even a small
amount of additional ions can have on the separation. It further
emphasizes the need for minimizing ions in the mobile phases to
facilitate molecule retention, then adding ions as necessary, during
method development, to achieve an acceptable separation.

Figure 15.17 Separation of a fusion protein (100 kDa, pI 8.35) using CEX
ionic strength pH gradient separation. Top profile shows the impact of
5 mM NaCl in mobile phase B. Bottom profile shows separation (left to
right) of homo and heterodimer species.

15.4.5 AEX pH Gradient Separations: Vaccines

The next profile in Fig. 15.18 shows a profile from an AEX pH
gradient separation of a vaccine with a theoretical pI of ~4.0 and
molecular weight of ~55 kDa. The mobile phases used ammonium
formate with formic acid to adjust the pH. Mobile phase A
was 50 mM ammonium formate + 300 µL/L of formic acid
(approximate pH ~4.4). Mobile phase B was 50 mM ammonium
formate + 1.5 mL/L of formic acid (approximate pH ~3.4).
These mobile phases will buffer over an approximate pH range
of 2.5–5. The column used in this separation was a Dionex,
ProPac, SAX-10, 4 × 250 mm.
 Molecule Information and Method Development
447

Figure 15.18 AEX profile and expanded plot of vaccine with a theoretical pI of ~4.0, and molecular weight of ~55 kDa.
448 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

In this separation, the buffer concentration required is

somewhat higher (50 mM), than what is required for mAbs with
pIs of ~ 9. This information again suggests that the requirement
for higher buffer concentration may be attributed to the lower
molecular weight in combination with the number of charged
sites, along with molecule conformation. This separation was
further optimized and subsequently validated, confirming that
pH gradient separations if appropriately developed represent a
simple, scientifically sound strategy for developing robust IEC
methods.

15.4.6 Parent mAb and ADC Separations

The profiles in Figs. 15.19 and 15.20 show profile comparisons

from CEX pH gradient separations of parent mAb and the
subsequent ADC for mAbs consisting of an IgG 4 and an IgG 1
with theoretical pIs of 8.1 and 8.9, respectively. Additionally, the
payload used for each mAb is also different.

Figure 15.19 AEX IgG 4 parent mAb and ADC with pH gradient CEX
separation.
 Molecule Information and Method Development 449

Figure 15.20 IgG 1 parent mAb and ADC with pH gradient CEX separation.

Both of the profiles in Fig. 15.19 were run using CEX separation
mode. The mobile phases were the same for both separations.
Mobile phase A was 5 mM HEPES, 5 mM AMPSO no pH adjustment.
Mobile phase B was 5 mM Bis-Tris propane no pH adjustment.
The gradient slope for both was 1.5%/min. The column used for
the mAb was a Dionex ProPac WCX-10, 4 × 250 mm. The column
used for the ADC was a MabPac SCX-10, 4 × 250 mm.
The mAb/ADC comparison clearly shows the influence of the
conjugate on the separation. However, the separation conditions
in terms of mobile phases and gradient slopes are the same
for both showing in this case that the conjugation may significantly
impact the chromatographic profile, but separation conditions
are comparable.
Both of the profiles in Fig. 15.20 were also generated in the CEX
separation mode. Based on the higher pI of the molecule (pI 8.9),
an ionic strength gradient was used for these separations.
The mobile phases were the same for both the mAb and the ADC.
Mobile phase A was 3 mM Bis-Tris propane, no pH adjustment.
Mobile phase B was 50 mM Bis-Tris propane no pH adjustment.
The gradient slope for the mAb was 1.5%/min, while the gradient
slope used for the ADC separation was slightly greater at 1.7%/min.
The column used for the mAb and ADC was a MabPac SCX-10,
4 × 250 mm
In this comparison as with the comparison in Fig. 15.19 we see
the influence of conjugation on the separation. The separation
conditions in terms of mobile phases are the same for both
molecules and the gradient slopes are comparable. This further
450 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

demonstrates that while the conjugation may significantly impact

the chromatographic profile, the necessary method conditions
to facilitate a suitable separation, for the mAb and ADC, may be
comparable.

15.5 Benefits of Narrow vs. Wide pH Gradients

One of the benefits of using pH gradient separations is related to
the ease of optimizing the separation during development. The
important parameter to consider during optimization is the rate
of delivery of ions across the stationery phase, which has been
briefly discussed earlier in this chapter. Subsequently, there are
two variables which need to be considered during development
in order to optimize rate of ion delivery to achieve the best
resolution.
First and most, and importantly, is the amount of ions which
are included in the mobile phases, which has been discussed
previously in this chapter. Failure to consider this parameter can
compromise the resolution to the point that all other attempts
to optimize the separation will be minimized if not completely
ineffective. Further guidance on this parameter for the various
molecule types evaluated in this chapter is included in a summary
table (Table 15.12) later in this chapter. Following the guidelines
for buffer concentration, listed in the summary table in most cases
will provide a good starting point for optimizing the ions in the
mobile phases.
After the appropriate buffer concentration is established, the
gradient slope can be further decreased to improve the resolution
of the separation. However, another strategy should be considered
at this point in addition to decreasing the gradient slope.
This strategy is related decreasing the pH gradient range,
and can be easily accomplished in two ways. The first and most
effective is by decreasing the number of buffers used in the mobile
phase so that the pH range of your mobile phases is decreased,
for example, reducing the pH range of the buffers from 7 pH units
down to 3 pH units.
The following experiment demonstrates the impact of
decreasing the buffer pH range through the removal of buffers in
 Benefits of Narrow vs. Wide pH Gradients 451

the mobile phases. This experiment compared narrow and wide

pH gradients to assess their impact on chromatographic profile
in the cation separation mode. Figure 15.21 shows overlays of 2,
IgG 2 mAbs run with the same method and column (Dionex
MAbPac, SCX-10, 4 × 250 mm). The gradient slope for the separation
of both mAbs was 2%/min. The mobile phases for the wide pH
gradient covered a pH range of ~7 pH units. Mobile phase A
consisted of MES, HEPES, AMPSO, and CAPS all at a 6 mM
concentration. Mobile phase B consisted of Piperazine, Bis-Tris
propane, AMP, and Piperidine all at 3 mM concentration. The
mobile phases for the narrow pH gradient covered a pH range of
~3 pH units. Mobile phase A consisted of HEPES and AMPSO both
at 5 mM concentration. Mobile phase B consisted of Bis-Tris
propane at 5 mM concentration.
The above results show improved resolution associated
with narrower pH gradients and demonstrate that narrower pH
gradients, in general, would be preferable to wider pH gradients.
Additionally, the use of narrower pH gradients will facilitate use
of gradient slopes that are within a more preferable operating
range for the pump.
Another strategy for accomplishing this is through mobile
phase blending. Blending is accomplished by combining mobile
phase A and B to establish mobile phase conditions closer to
the start and end of molecule elution. For example, if the separation
is completed between 40% and 70% mobile phase B, then the
mobile phases can be blended so that mobile phase A is actually
a mix of A and B (e.g., 80% M.P. A, and 20% M.P. B).
Likewise, mobile phase B would also be a mix of mobile phase A
and B (e.g., 90 M.P. B & 10% M. P. A).
Mobile phase blending should be the last step in mobile phase
optimization, and may be not necessary in many cases. It should
be performed after the optimization of buffer concentration and
reducing the pH range of the mobile phases. The primary benefit
from mobile phase blending is that it will essentially further
stretch the gradient over the operating range of the pump, thereby
facilitating some further, but usually minor, improvement in
resolution or allowing the use of increased gradient slopes within
a more suitable range of the instrument.
 452
Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

Figure 15.21 Separate overlays of two IgG 2 mAbs comparing a wider pH gradient (~7 pH units) to a narrower pH gradient
(~3 pH units).
 Platform Application of pH Gradients 453

15.6 Platform Application of pH Gradients

This experiment was designed to demonstrate the platform nature
of pH gradient separations. In this experiment, 12 different mAbs
were separated using the same mobile phases and instrument
method. The mobile phases used mobile phases 2 and 3 from
Table 15.1. Mobile phase A: 5 mM HEPES, 5 mM AMPSO. Mobile
phase B: 5 mM Bis-Tris propane. The gradient slope for this
separation was 2.1% min. The column used in this experiment
was a Waters Protein-Pak, Hi Res CM, 7 μm, 4.6 × 100 mm column.
The chromatographic profiles along with a scatter plot with
linear regression, showing the correlation between retention
time and pI, associated with this pH gradient, are shown in
Fig. 15.22. The mAb subclass, retention time, and pI values for
this experiment are listed in Table 15.10.

Figure 15.22 Shows chromatographic profiles, along with RT/pI plot

(r2 = 0.8926), for 12 different mAbs separated using the same pH gradient
in CEX separation mode.

Results from this experiment clearly show the platform

nature of pH gradient IEC separations. It also, shows the general
454 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

correlation between main peak retention time and molecule pI,

further demonstrating the ability of this approach to separate
based on pI. This correlation between retention time and pI
for different molecules demonstrates the methods viability for
separation of charged variants within the molecule, based on
their pI. However, it is also clear, due to the lack of complete
agreement (r2 = 0.8926) for the linear regression, that there
are additional factors, which have been discussed earlier in
this chapter, that influence the retention time of the molecules
relative to their pI. These factors could include molecule
conformation, size, and the presence of conjugates or glycosylated
species on the molecule among other things. These observed
differences in retention time for molecules with the same
theoretical pI, point to the power of this approach to separate
intact molecules and charged variants, within a molecule, based
in part, on some characteristics which are not easily measured.
Finally, the ability of this approach to separate based on these
additional factors could prove useful for gaining further information
about the molecule. This experiment further demonstrates
the utility of this approach for separating complex mixtures
containing molecules with similar or even identical pIs.
Table 15.10 Shows the list of mAbs, used in this experiment, with
respective retention time and pI in order of elution

Type mAb subclass Retention time (min) pI

mAb IgG-2 6.4 7.2
mAb IgG-2 14.0 7.3
mAb IgG-2 20.9 7.8
ADC IgG-2 23.7 7.7
ADC IgG-1 23.4 7.7
mAb IgG-2 24.0 8.1
ADC IgG-2 25.1 8.4
mAb IgG-2 35.3 8.5
mAb IgG-1 37.1 8.5
mAb IgG-1 37.7 8.5
mAb IgG-1 39.5 8.9
ADC IgG-1 41.2 8.8
 Purification Using pH Gradient Separations 455

15.7 Purification Using pH Gradient Separations

The use of pH gradient separations is also suitable for purification
of charged variants to support validation and characterization
of charged variants. Figure 15.23 shows the separation from
2 different columns, with the top overlay showing the separation
from a column which could be used for purification of charged
variants (Agilent, PL-SCX, 4000Å, 150 × 4.6 mm, 8 µm) with a
5 mg load. The bottom profile shows the separation from an
analytical column (Dionex, MabPac, SCX-10, 4 × 250 mm) with
25 µg load run with the same method. The mobile phases used in
this experiment were HEPES, and Bis-Tris propane. Mobile phase
A was 5 mM HEPES, and mobile phase B was 1.5 mM HEPES and
3.5 mM Bis-Tris propane. The mAb is an IgG 2 with a theoretical
pI of 7.3. The method and gradient slopes used were the same
for both columns.
This method was originally developed for purification
purposes to support both characterization and possible validation
activities. The top profile from this method shows the separation
of a 5 mg load on an Agilent PL-SCX column. The bottom profile
is the same method run on a Dionex MAbPac column with a
25 µg load. The result from this experiment shows comparable
profiles from the different columns with significantly different
loads (5 mg vs. 25 µg). This experiment further demonstrates the
robustness of pH gradients for both analytical and purification
purposes.

15.8 Mobile Phase Preparation

One of the most critical parameters related to developing well-
resolved and robust IEC separations is the mobile phase. With
regard to mobile phase development, the choice of buffer and
buffer concentration must be made with consideration for the
following information which includes molecule pI, mAb subclass,
or molecule type, and molecular weight.
These factors have been discussed earlier in this chapter and
will not be discussed further.
 456
Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

Figure 15.23 Comparison of two profiles from the same method using columns from different manufacturers with loads of 5 mg (top
profile) and 25 µg (bottom profile).
 Mobile Phase Preparation 457

Another critical factor with regard to mobile phases is the

preparation. Failure to adequately control the level of ions in the
mobile phases, during preparation, can lead to excessive retention
time shifts and loss of resolution. As a result, all factors which
could contribute to variability of the ion concentration in the
mobile phases should be considered when writing mobile phase
preparation steps in the method.
The sources of variability during mobile phase preparation
include analyst, pH meter, balance, pipettes, and glassware. The
analyst represents the largest potential source of variability
in the mobile phase preparation. However, the majority of this
variability is related to either analyst training, or ambiguity in
the preparation instructions, leading to variability associated
with different mobile phase preparation techniques. Removing
most of the variability associated with differences in preparation
technique can be accomplished by adding an appropriate amount
of detail in mobile phase preparation to facilitate consistent
analyst-to-analyst mobile phase preparation. These variations can
be identified and minimized by having several analysts prepare
the mobile phases and running the method during method
development.
Another source of variability that can be minimized for IEC
separations, is the variability associated with the pH meter. As
seen in Fig. 15.2, a method with a pH criterion of ± 0.1 pH units can
result in significant retention time shifts and possible loss of
resolution. There are several factors which can influence the
accuracy of the pH meter including ambient temperature, calibration
standards, pH calibration range, variations in analyst calibration
procedure, improper maintenance of pH probe, in addition to
minor differences between manufacturers and models. All of
these factors are sources of variability which can influence
the accuracy of pH meter and subsequently the amount of pH
adjuster (ions) which are added to the mobile phases.
One of the primary advantages of pH gradients which use
acidic and basic buffers is that the mobile phase pH is self-adjusting
and occurs with the mixing of the mobile phases during the
separation. This approach to pH gradient separations will provide
the simplest mobile phase preparation along with the greatest
control over the amount of ions in the mobile phases. However, if
pH adjustment of the mobile phases is required, the analytical
458 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

method section for preparation of mobile phases should not

incorporate the use of a pH meter. The pH, and more importantly,
the ion concentration in the mobile phases can be precisely
controlled through the use of analytical balances, pipettes, and
calibrated glassware (e.g., graduated cylinders or volumetric
flasks). The pH meter should be limited to the development phase
of IEC methods and the pH should be linked to the weight or
volume of buffer in combination with the weight or volume of pH
adjuster added to the mobile phases. Once the appropriate weights
and volumes of mobile phase components are established, during
the development phase, there is no need to reference the mobile
phase pH in the analytical method. In other words, the mobile
phase composition should be always determined by the amount
of buffer and the amount of pH adjuster added as measured by
an analytical balance and a calibrated pipette, respectively.
For example, mobile phase preparation instruction might look
like this for mobile phase 1 in Table 15.1:

Mobile Phase B: 5 mM HEPES, 5 mM AMPSO + 3.75 mL of 2.5 N

NaOH

Add ~800 mL of HPLC grade water to a 1 L graduated cylinder.

Add 1.19 ± 0.01 g of HEPES, and 1.14 ± 0.01 g AMPSO. Using a
calibrated pipette, add 3.75 mL of 2.5 N NaOH. Q.S. to 1 L with
water and mix well. Filter through a 0.2 µm filter.
In this example the concentration of ions in the mobile phase,
will be more precisely controlled, from one preparation to the next,
than if one were to place a pH criteria of 9.5 ± 0.1 pH units on the
mobile phase preparation. In the experiment in Fig. 15.2, the criteria
of ± 0.1 pH units could result in an approximate range of ± 100 µL
of 6N NaOH adding unnecessary variability to the mobile phase
preparation.

15.9 Sample Preparation

Sample preparation is very important with regard to consistent
performance of the method. The goal for sample preparation is to
make sure that the sample is adequately diluted to reduce excessive
levels of ions in the sample matrix which can interfere with the
separation. Additionally, one must consider sample stability in the
 Sample Preparation 459

diluent as well as considering the final diluted sample concentration

as it relates to column loading. Failure to identify the appropriate
sample preparation instructions relative to the samples being
tested can result in, significant retention time shifts, poor profile
and loss of resolution.
The sample diluents used in the experiments in this chapter
consist of water or mobile phase A. Water is a good choice of
diluent since this will reduce the amount of ions in the sample
matrix while still maintaining some sample buffering from
the formulation buffer. However, if the pH of mobile phase A is
comparable to that of the sample matrix, then mobile phase A can
also be a good option. In either case, sample stability on the
instrument should be evaluated with the goal of maintaining
sample stability for an acceptable period of time.
For analytical samples, diluting to a concentration of 1 mg/
mL is adequate. However, if the starting sample concentration
is close to 1 mg/mL, it may be necessary to dilute the sample to
a lower concentration to eliminate matrix interference with the
separation. Typically a 2–3 fold dilution is adequate, but a lower
sample dilution should be more thoroughly evaluated to assess
the impact of increased levels of sample formulation matrix on
the separation. The dilution must be evaluated while keeping in
mind the necessity of loading an adequate amount of material on
the column. A typical range for loading is ~ 25–100 µg on the
column for analytical testing. If the initial sample concentration
is too low to achieve sufficient column loading with the necessary
dilution, lower column loads can be evaluated with more efficient
columns (e.g., narrow bore and or smaller particle size columns).
For purification purposes, it is important to maintain the
highest sample concentration while diluting just enough to
minimize the influence of ions from the sample matrix. The goal
in this case is to maximize the load on the column per injection
which may require one to minimize the dilution. A dilution as low
as 2–3-fold will usually be adequate to eliminate matrix
interference. However, a lower dilution, again, should be more
thoroughly evaluated for the negative profile influences which
can be associated with increased levels of sample matrix.
If the sample concentration is not high enough to achieve
the desired load on the column from a single injection, a short
loading method can be used load multiple injections. In this
460 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

case, starting conditions should be evaluated to insure that the

sample is retained at the column head throughout the loading
injections. If the sample is not adequately retained under starting
conditions, the profile will be negatively impacted, resulting
in repeating/ overlapping profiles from the multiple injections.
For in-process samples, it is most important to adequately
dilute the sample to eliminate or minimize the influence of the
sample matrix. For this reason, it is especially important to know
the sample concentration and the sample matrix of the samples
being tested. For in-process samples, if the sample concentration
is not high enough to achieve the necessary load on the column
from a single injection, a short loading method can be used to
load multiple injections. As mentioned above for multiple loading
injections, starting conditions should be evaluated to insure that
the sample is retained at the column head throughout the loading
injections.

15.10 Columns and Column Temperature

Table 15.11 is by no means exhaustive but is reflective of the
columns used for the experiments presented in this chapter as
well as other pH gradient–related work by the author.

Table 15.11 List of columns which have been used for IEC separations
in this chapter

Manufacturer Column desc.

Dionex ProPac WCX-10, 4 × 250 mm, 10 µm
Dionex MabPac SCX-10, 4 × 250 mm, 10 µm
Dionex ProPac SAX-10, 4 × 250 mm, 10 µm
Waters Protein-Pak, Hi Res CM, 4.6 × 100 mm, 7 μm
Agilent PL-SCX 4000 Å, 4.6 × 150 mm, 8 µm
Agilent PL-SAX 4000Å, 4,6 × 50 mm, 5 µm

Column temperature control should be specified in the

method since pH is influenced by temperature. Failing to control
temperature in the method will result in retention time shifts
as well as some minor changes in peak resolution which could
result in variations in quantitated peak area percent results.
 Choosing the Separation Mode 461

Overall, failing to control temperature will result in a less robust

method with greater variability with regard to precision. For all
of the separations listed in this chapter the column temperatures
were controlled at 35–40°C.
Figure 15.24 shows an example of the impact of column
temperature on retention time for an AEX pH gradient separation.
The profiles show that decreasing the column temperature was
associated with decreasing retention time.
Figure 15.25 shows the impact of column temperature on
resolution for an AEX pH gradient separation. The profiles show
that decreasing the column temperature was associated with
decreased resolution of the first acidic peak after the main peak.

Figure 15.24 Overlay showing impact of column temperature on general

retention time.

Figure 15.25 Overlay showing impact of column temperature on peak

resolution.
462 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

These results illustrate the impact of column temperature on

both retention time and resolution, and further demonstrate the
need for evaluating and controlling column temperature during
IEC separations.

15.11 Choosing the Separation Mode

The primary consideration for choosing the separation mode
during initial development is based on molecule pI. For example,
a molecule with a pI at < 7 would initially be evaluated using AEX
while a molecule with a pI of ~7 or above would be evaluated
using CEX. Molecules with pIs in the 6.5–7.5 range may separate
equally well in either mode and could be evaluated in both.
An interesting advantage of using mobile phases which contain
both an acidic buffer and a basic buffer with no pH adjustment
(e.g., HEPES, AMPSO, and Bis-Tris propane), is that these mobile
phases can be run in either the cation or anion exchange modes
by simply switching which mobile phase is used for eluting the
molecule.
This represents an advantage in method development
capabilities allowing for an efficient and more thorough evaluation
of both separation modes. In most cases, following the guidelines
above for choosing the separation mode will yield the best
separation. Additionally, due to the higher pI for most mAbs, the
preferred separation, based on profile, will be in the cation mode.
However, the presence of negatively charged structures on the
molecule, in the form of sialylated glycans or conjugates, can
facilitate adequate separation in either separation mode.
Figure 15.26 shows the IEC profile from a highly sialylated
mAb that was separated in both cation and anion exchange
modes using the same buffers. The mAb was an IgG2 with a pI of
7.3. The mobile phases used in this separation were the same as
what was used in Fig. 15.22 for the wide pH gradient. The column
used in the separation for CEX was a Varian PL-SCX 4000 Å,
4.6 × 150 mm, 5 µm. The column used in the separation for AEX
was a Varian PL-SAX 4000 Å, 4.6 × 150 mm, 5 µm.
Based on the profiles in Fig. 15.26, it appears that the
profiles from either separation mode are acceptable and could
be evaluated for further development.
 Choosing the Separation Mode
463

Figure 15.26 CEX (Top) and AEX (Bottom) profiles for a highly sialylated IgG2 with a theoretical pI of 7.3.
 464
Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

Figure 15.27 CEX (Bottom) and AEX (Top) profiles for an IgG2 mAb with a theoretical pI of 7.3.
 Summary 465

Figure 15.27 shows the IgG2 from Fig. 15.23, with a theoretical
pI of 7.3 which has been separated using the same mobile
phases used in Fig. 15.22 in both CEX and AEX separation modes.
In this case, the CEX separation is shown to be preferable over
the AEX separation mode. The gradient slope for the CEX mode
was 0.7%/min while the gradient slope for the AEX mode was
0.5%/min.

15.12 System/Column Equilibration

For IEC separations proper column equilibration is very important
in achieving consistent separation both within a sample sequence,
and from one instrument to another. There are several factors
which influence the time it takes to re-equilibrate a column, including
column size, flow rate, and system volume. All of these parameters
must be considered to establish an appropriate re-equilibration
time.
Symptoms of poor re-equilibration include retention time
shifts during the run or when switching from one instrument to
another, change in retention time, profile, and or loss of resolution.
However, proper re-equilibration is easily achieved by adding
a re-equilibration step as part of the method. The re-equilibration
step should occur after the separation gradient is complete, and
prior to the next injection. The re-equilibration time is established
based on system dead volume or volume of liquid in the system,
to include the column. In this case, the retention time of the void
peak from the chromatogram can be used to determine the
system void. The re-equilibration step in the method should be
at least 5 times the retention time of the void peak in the
chromatogram. The re-equilibration step can be easily evaluated
during a sequence by checking main peak retention times to
insure that the retention time has stabilized by the second or
third injection of the sequence, and that it is consistent with
the expected retention time in the method. If the retention time
continues to shift to the left after the second or third injection,
or if the profile is changing, additional time may need to be added
to the re-equilibration step in the method to stabilize retention
time.
 466

Table 15.12 Summary Table of separation conditions for mAbs, ADCs, and other bio-therapeutics

MW Buffer Buffer Buffer

Molecule pI KDa (low pH) (high pH) strength Mode Gradient type Figures in Chapter

mAb, IgG1 7 to < 9 150 HEPES/AMPSO Bis-Tris propane ~10–20 mM CEX pH 4D, 11, 12, 13, 23,

mAb IgG1 ≥9 150 N/A Bis-Tris propane 3–35 mM CEX Ionic Strength 16

ADC, IgG1 ~9 150 N/A Bis-Tris propane 3–35 mM CEX Ionic Strength 21

mAb IgG2 7 to < 9 150 HEPES/AMPSO Bis-Tris propane ~5–10 mM CEX pH 4A, 4B, 4C, 10, 11, 13,
23, 24,

mAb IgG2 ≥9 150 N/A Bis-Tris propane 3–35 mM CEX Ionic strength 16, 21

mAb, IgG4 8.1 150 HEPES/AMPSO Bis-Tris propane ~5 mM CEX pH 20

ADC, IgG4 8.1 150 HEPES/AMPSO Bis-Tris propane ~5 mM CEX pH 20

Vaccine 4 55 Ammonium Formate + Ammonium Formate + 50 mM AEX pH 19, 25

Formic Acid Formic acid

Fusion ~8.4 100 Bis-tris propane + HCL Bis-tris propane 3–60 mM CEX pH-Ionic strength 17
Protein combination

Fusion 6.2 40 Ammonium Ammonium 60 mM AEX pH 18

Protein acetate + Formic acid acetate + Formic acid
Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals
 Conclusions 467

15.13 Summary
Table 15.12 provides a summary of the information from the
separations in this chapter and should serve as a resource for
method development of other mAbs, ADCs, as well as other bio-
therapeutics. Matching the general information in the molecule
and pI columns to another mAb, ADC, or other molecule should
provide useful starting information for developing a method.
For example, if one wanted to develop a separation for a mAb,
IgG1 subclass, with a pI of 8.1, the conditions listed in the first row
of Table 15.12 will likely be suitable for generating an acceptable
separation to start with, and are summarized in Table 15.13.
Further information regarding the separations associated with
each row can be obtained through the figure references in the
far right hand column of Table 15.12.

Table 15.13 Shows initial method parameters using information from

Table 15.10 for an IgG 1 mAb with a pI of 7 to <9

Parameter Specification
Separation mode CEX
Columns Dionex, Mab Pac, SCX-10, 4 × 250,
Waters, Protein-Pak, Hi Res CM, 4.6 × 100 Agilent,
PL-SCX 4000 Å, 4.6 × 150 mm, 8 µm
Column temp. 35–40°C
Autosampler temp. 2–8°C
Flow rate ~1 mL/min
Detector wavelength 280 nm
Injection volume 50 µL
Sample dilution 1 mg/mL (water or mobile phase A)
Mobile phase A 15 mM HEPES/15 mM AMPSO
Mobile phase B 15 mM Bis-Tris propane

15.14 Conclusions
This chapter describes a simple approach to developing highly
resolving and robust pH gradient IEC methods for mAbs, ADC’s, and
468 Using pH Gradient Separation for mAbs, ADCs and Other Complex Bio-Pharmaceuticals

other bio-therapeutics such as fusion proteins and vaccines. The

flexibility, ease of development and use, along with the versatility
of pH gradients gives this type of IEC method an advantage over
traditional salt gradients for the separation of these more complex
molecules. The experiments in this chapter demonstrate the
negative impact from excessive ions in the mobile phases
providing rationale and strategies for easily controlling them to
optimize the separation. Additionally, the use of pH gradients for
IEC separations as described in this chapter provides additional
information regarding specific development strategies related to
mAb subclass, molecule pI, and molecular weight.
Because these strategies take into account additional
information with regard to molecule characteristics, they represent
an improved approach over current practices used for IEC pH
gradient separations. This approach will prove useful as a tool
for supporting analysis of bio-therapeutics, as well as supporting
characterization and validation activities. This approach also has
the potential for application within process development.

References

1. Flatman, S., Alan, I., Gerard, J., Mussa, N. J. Chromatogr. B, 2007, 848,
79–87.
2. Schenerman, M., Sunday, B. R., Kozlowski, S., Webber, K., Gazzano-
Sanotora, H., Mire-Sluis, A. BioProcess Int., 2004, 42–52.
3. Harris, R. J. J. Chromatogr., A, 1995, 705, 129–134.
4. Weitzhandler, M., Farnan, D., Horvath, J., Rohrer, J. S., Slingsby, R. W.,
Avdalovic, N., Pohl, C. J. Chromatogr. A, 1998, 828, 365–372.
5. Santora, L. C., Krull, I. S., Grant, K. Anal. Biochem., 1999, 275, 98–108.
6. Moorhouse, K. G., Nashabeh, W., Deveney, J., Bjork, N. S., Mulkerrin, M.
G., Ryskamp, T. J. J. Pharm. Biomed. Anal., 1997, 16, 593–603.
7. Chelius, D., Jung, K., Lueras, A., Rehder, D. S., Dillon, T. M., Vizel, A.,
Rajan, R. S., Li, T., Trauheit, M. J., Bondarenko, P. B. Anal. Chem., 2006
78, 2370–2376.
8. Donato, A., Ciardiello, M., Nigris, M., Piccoli, R., Mazzarella, L., D’Alessio,
G. J. Biol. Chem., 1996, 268, 4745–4751.
9. Hsu, Y.–R., Chang, W.-C., Mendiaz, E. A., Hara, S., Chow, D. T., Mann, M. B.,
Langley, K. E., Lu, H. S. Biochemistry, 1998, 37, 2251–2262.
 References 469

10. del la Guntińas, M., Wissiack, R., Bordin, G., Rodrigez, A. R. J. Chromatogr.
B, 2003, 791, 73–83.
11. Santora, L., Kaymakcalan, Z., Sakorafas, P., Krull, I., Grant, K. Anal.
Biochem., 2001, 299, 119–129.
12. Vlasak, J., Ionescu, R. Curr. Pharm. Biotechnol., 2008, 9, 468–481.
Chapter 16

Size Exclusion Chromatography Method

Development for Therapeutic Proteins

Alexandre Goyon, Jean-Luc Veuthey, Davy Guillarme,

Szabolcs Fekete
University of Geneva, University of Lausanne,
CMU—Rue Michel-Servet 1, Geneva, 1211, Switzerland
[email protected], [email protected]

16.1 Introduction to Size Exclusion

Chromatography
Size exclusion chromatography (SEC) is a mode of liquid
chromatography in which separation is based on the partial
exclusion of the solutes from the pores of the stationary phase [1].
The larger molecules elute earlier than the smaller ones, because
large solutes can penetrate into the pores only by a smaller fraction

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
472 Size Exclusion Chromatography Method Development for Therapeutic Proteins

(partially) [1]. Very large solutes (larger than the pores) can be
totally excluded from the pores and will elute within the interstitial
column volume. Based on its mechanism, SEC is applied to
separate solutes of different sizes (molecular weights). Hence, SEC
is commonly used to separate protein aggregates (size variants)
such as monomers, dimers, trimers, and other oligomers or
high-molecular-weight species (HMWS).
Among the different techniques applied for the characterization
of aggregates—such as sodium dodecylsulfate polyacrylamide gel
electrophoresis (SDS-PAGE), asymmetrical field flow fractionation
(AF4), fluorescence spectroscopy, circular dichroism (CD) or light-
scattering-based methods, like multi-angle laser light scattering
(MALLS)—SEC is considered today as the reference technique
for the quantification of protein aggregates. The main advantage
of this approach is the mild elution conditions (non-denaturing)
allowing the characterization of proteins with minimal impact
on their conformational structure and the local environment.
In SEC, we strive for separating the proteins under inert conditions.
Indeed, the separation should ideally depend only on the size of
the solutes. However, in practice, some secondary interactions
(hydrophobic and/or electrostatic) may occur between the proteins
and the stationary phase [2]. To limit such interactions and work
under inert conditions, the pH and ionic strength of the mobile
phase should be carefully controlled. Ideally, the mobile-phase pH
should be set close to the isoelectric point (pI) of the proteins but
it is not applied in practice for basic proteins and a physiological
pH is usually applied. Salts (sodium or potassium chloride) or
other additives (arginine) are also often added to the mobile
phase to avoid or at least limit electrostatic interactions, while a
small proportion of organic modifier (e.g., 10% 2-propanol) can
also be used to limit hydrophobic interactions as much as possible.
The aim of this chapter is to provide some guidelines to
develop the SEC method for proteins and to show the possibilities
of modern SEC (UHP-SEC) separations.

16.2 Theoretical Aspects

In SEC, biomolecules are separated based on their hydrodynamic
radius. The stationary phase consists of spherical porous particles
 Theoretical Aspects 473

(silica or polymer based) with a carefully controlled pore size and

pore size distribution, through which the solutes diffuse based on
their molecular size differences, using an aqueous buffer as the
mobile phase. Entropic processes without any physical or physico-
chemical adsorption ideally drive the partitioning between the
internal pores and interstitial volume.
To describe the elution of a peak in SEC, the so-called thermo-
dynamic retention factor (k)—or zone-retention factor—has to
be used, since there is no retention and therefore conventional
retention factor k comprised between –1 and 0 is not a good
measure of elution. Thus, k expresses the ratio of the
probabilities of the sample staying in the stagnant mobile phase
inside the pores and in the moving mobile phase between the
particles (interstitial volume) V[3]. eIt is easy to understand that
t0 – te
this k value is limited ( kmax
p
 ),= and p
= its =maximum , value is given by
Ve ee te
the ratio of the pore volume (Vp ) and the interstitial volume (Ve )
of a column, or similarly by the pore porosity (ep ) and the
interstitial porosity (ee ):

Velution – Ve t elution– te (16.1)

k = =
Ve te

Vp ep t0 – te
kmax
 = = = , (16.2)
Ve ee te

where Velution and telution are the elution volume and elution time,
respectively, while Ve and te are the total exclusion volume and
time (determined by the interstitial volume) and t0 is the column
dead time (determined Vp byep thet 0 –sum of interstitial and pore
te
volumes). The kmax  =value =
= then ,
defines the possible
Ve ee t Vp ep elution
t –t
window (tW) for the solutes. The elarger the kmax  , =the = 0the e ,
=better
Ve ee t
probability to separate the compounds. Figure 16.1 shows ea
typical SEC chromatogram and corresponding te, telution, t0 andV e
t –t
tW. Using currently available state-of-the-art columns, the kmax  =is p = p = 0 e ,
Ve ee te
around 1.4–1.8. This parameter—and the associated selectivity—
can be altered by packing a column more densely (or decreasing
particle size) or by using particles with larger pore volumes.
Another important characteristic of SEC columns is the
calibration curve, since it illustrates the size (weight) of analytes
that can be separated on a given column. The calibration curve
474 Size Exclusion Chromatography Method Development for Therapeutic Proteins

can also be used to estimate the molecular weight of an unknown

analyte. Typical calibration curves are based on proteins or
polymers of known molecular weights (M). By plotting log M vs.
the retention volume, one typically obtains a third order
polynomial with a linear region providing the highest resolution
and molecular weight accuracy [4]. The linear part of the curve
indicates the range of molecular weights that can be adequately
separated on the given stationary phase.

Figure 16.1 Typical SEC chromatogram of monoclonal antibody monomer

and aggregate, using a 150 × 4.6 mm column.

The slope of the line in the linear portion of the calibration

curve is a measure of the stationary-phase selectivity, which can
be defined by the following relationship:

logM = AVelution + B , (16.3)

where A and B are the slope and intercept of the line, respectively.
As the pore size distribution becomes narrower, the slope
becomes shallower which results in a greater selectivity to
discriminate analytes of similar size. The intercept of the linear
part depends on the average pore diameter. Figure 16.2 shows
 Theoretical Aspects 475

some model calibration curves related to different pore size

distributions and pore diameters.

Figure 16.2 Illustration of SEC calibration curves as log M vs. normalized

elution volume for columns possessing different average pore size and
pore size distribution.

There are some potential sources of error when measuring

proteins molecular weight based on a SEC calibration curve.
First, since proteins shapes could vary (e.g., globular, rod-like or
flexible chains), their Stokes radii do not correlate exactly with
molecular weight. Second, non-ideal adsorption may alter the
retention volume [5–7].
Ideally, SEC involves solely size-based separation. Thus,
there is no true retention compared to other sorptive techniques.
Nevertheless, apparent plate numbers (N) can be used as a
measure of the column performance, as introduced by Engelhardt
and used by others [8–10]. The efficiency (N) in SEC can also be
studied and compared by constructing plate height (H) vs. linear
velocity (u) plots.

B
H = Au01/2 + + C u0 , (16.4)
u0
476 Size Exclusion Chromatography Method Development for Therapeutic Proteins

where A corresponds to eddy-dispersion, B to longitudinal

diffusion and C to mass transfer resistance, which in this case
is related to the partitioning of the solute between the internal
pores and the interstitial volume. We thought that it is useful to
illustrate the expected shape of an H–u plot for large proteins in
SEC, since it could be quite different compared to the frequently
presented van Deemter curve shapes observed with small
molecules. Figure 16.3 shows the total plate height (Htotal) and
the plate height contributions of eddy-dispersion (HA), longitudinal
diffusion (HB), and mass transfer resistance (HC) as a function
of mobile-phase velocity.

Figure 16.3 Calculated H–u curve in SEC for a 150 kDa protein using a
column packed with 3.0 µm particles operating at 25°C. Reproduced
with permission from Elsevier.

As expected, the longitudinal diffusion (B term) is negligible

with large proteins. The plate height is determined by the eddy-
dispersion (A term), especially for more excluded species, and mass
transfer resistance (C term). To conclude, the lowest plate height
(highest efficiency, N) is reached at low mobile-phase velocity.
The higher the velocity (mobile-phase flow rate), the lower the
efficiency.
 Method Development of Proteins 477

By using current state-of-the art 150 × 4.6 mm SEC

columns (packed with sub-3 µm particles), N ~ 2000–10,000 can
be reached for a 150 kDa protein, while using conventional
300 mm × 6–8 mm column (packed with 5 µm particles),
N ~ 4000–15,000 can be performed in the practically useful
analysis time range. Such separations typically take 5–10 min on
150 × 4.6 mm and 10–40 min on 300 × 6–8 mm columns,
respectively. For smaller compounds (MW < 1 kDa, e.g., protein
formulation excipients), a higher efficiency is expected in most
cases, while very large proteins often elute in very broad and tailed
peaks. Small analytes eluting at t0 (totally penetrated through
the pores) possess similar plate numbers as can be observed in
reversed-phase liquid chromatography (RPLC). In average, an
efficiency of ~100,000 plates/meter can be attained with a column
packed with 5 µm particles, while a SEC material packed with
2 µm particles may perform up to 250000 plates/meter for common
small molecules such as uridine.

16.3 Method Development of Proteins

16.3.1 Stationary-Phase Selection
16.3.1.1 Pore size, particle size, and column dimensions
First of all, a stationary phase with appropriate pore size has to
be selected. The pore size should fit to the size (or molecular
weight) of the investigated protein and its aggregates. The log M
of the solutes should be included in the linear part of the
calibration curve. For most protein separations, typical pore sizes
between 150 and 500 Å are often employed. For the separation
of common proteins, a pore size of 150–200 Å is generally well
suited in the 0.5–450 kDa range, while a 200–300 Å pore size is
commonly applied for monoclonal antibodies or antibody drug
conjugates (MW ~ 150 kDa), to separate HMWS up to 1000 kDa.
For very large proteins (MW > 200 kDa, e.g., PEGylated proteins),
the 500–1000 Å phases offer probably the best selectivity up
to 7500 kDa. Obviously, not only the average pore size, but also
the pore size distribution has to be considered depending on
478 Size Exclusion Chromatography Method Development for Therapeutic Proteins

the range of molecular size/weight. If only the protein monomer

and dimer need to be separated, a stationary phase with narrow
pore size distribution is preferred. On the other hand, if a broad
range of compounds needs to be separated, a wide pore size
distribution material is more adequate.
Resolution between the aggregates and native protein peaks
can be improved by decreasing the column particle size. The
achievable plate number is indeed inversely proportional to the
particle size, while the resolution is inversely proportional to
the square root of the particle size. Historically, columns packed
with 5–10 µm particles were commonly used in routine SEC
applications, but several providers offer now 3.0, 2.7, 2.5 µm or
even sub-2 µm particles packings. By using these state-of-the-art
columns packed with such small particles, the resolution can be
significantly improved and/or the analysis time can be shortened
by applying shorter columns. Figure 16.4 shows experimentally
measured H–u curves for β-lactoglobulin (18.7 kDa) and a
therapeutic monoclonal antibody (150 kDa) when using SEC
columns packed with 5.0, 2.7, 2.0, and 1.8 µm particles. As
shown, the smaller the particle, the lower the plate height. For a
monoclonal antibody, about three times lower plate heights
can be obtained with sub-2 µm materials compared to conventional
5 µm packing.

Figure 16.4 Measured H–u curves for β-lactoglobulin (A) and a therapeutic
monoclonal antibody (B) by using columns packed with 5.0, 2.7, 2.0, and
1.8 µm particles, at ambient temperature. Reproduced with permission
from Elsevier.
 Particle Max. temperature Max. pressure
Table 16.1 Some popular state-of-the-art SEC columns and their properties

Name (provider) Material size (µm) Pore size (Å) (˚C) pH range (bar)
Acclaim SEC-300 hydrophilic 5, 7 300, 1000 60 2–12 80
(Thermo) polymethacrylate resin
Acquity UPLC BEH SEC diol modified hybrid-based 1.7, 2.5 125, 200, 450 60 2–8 400
(Waters)
AdvanceBio SEC (Agilent) surface-coated silica-based 2.7 130, 300 80 2–8.5 400
Bio SEC-3 (Agilent) surface-coated silica-based 3 100, 150, 300 80 2–8.5 240
Bio SEC-5 (Agilent) surface-coated silica-based 5 100, 150, 300, 80 2–8.5 240
500, 1000,
2000
SRT-SEC (Sepax) surface-coated silica-based 5 100, 150, 300, 80 2–8.5 240
500, 1000,
2000
TSKgel Ultra SW diol modified silica-based 3 300 30 2.5–7.5 120
aggregate (Tosoh)
TSKgel SW mAb (Tosoh) diol modified silica-based 4 250 30 2.5–7.5 120
TSKgel UP-SW3000 diol modified silica-based 2 250 30 2.5–7.5 250
(Tosoh)
Yarra (Phenomenex) silica-based 1.8, 3, 5 145, 150, 290, 60 2.5–7.5 (except 200 (except for
300, 500 for SEC-X150, 1.8 µm: 480 bar)
X300: 1.5–8.5)
Method Development of Proteins

YMC-Pack Diol-SEC (YMC) diol modified silica-based 3, 5 60, 120, 200, 40 5–7.5 200
300
479

Zenix-SEC (Sepax) surface-coated silica-based 3 80, 100, 150, 80 2–8.5 240

300
480 Size Exclusion Chromatography Method Development for Therapeutic Proteins

The very high pressure generated by small particles (P > 250

bar) in SEC may be critical for some pressure sensitive proteins,
which may suffer from on-column aggregation, and therefore
misleading results can be obtained [11]. However, this pressure
induced aggregation (or degradation) is protein dependent
and has to be experimentally evaluated. Columns packed with
2.0–3.0 µm particles are good compromise between efficiency
(or analysis time) and pressure. Some characteristics of the most
popular SEC columns on the market were summarized in
Table 16.1.
In conventional SEC, 300 mm long columns with internal
diameters (I.D.) of 6, 7.8, 8, or 10 mm are generally applied. These
SEC columns are referred to as standard bore columns. Now,
several vendors offer narrow-bore columns with 4.6-mm I.D.
and 150 mm length, packed with small particles. By using these
shorter columns, similar separation quality can be attained as
with 5 µm particles in 300 mm standard bore columns, but
the analysis time can be reduced by a factor of 3 to 4 [11].
When using the 4.6 mm I.D. columns, the extra-column (system)
band broadening could have a serious effect on the observed
efficiency in SEC. Those particular columns require the use of
optimized UHPLC systems with very low extra-column volume
(<10 µL). Otherwise, 30–60% loss of column efficiency may occur
on conventional HPLC systems, which possess large extra-
column band variance up to 100 µL2 [12]. Then, conventional HPLC
systems are only compatible with standard bore SEC columns,
to attain their full performance.

16.3.1.2 Stationary-phase chemistry

SEC columns packed with porous cross-linked dextran particles
were introduced in 1953 by Wheaton and Bauman [13].
Polyacrylamide-based gels were subsequently developed by
Hjertén and Mosbach in the 1960s [14]. The porosity of those
columns was controlled by adjusting the degree of cross-linking
degree. These polymeric stationary phases were effective to limit
non-specific interactions between the polymer (stationary phase)
and sample molecules (proteins). Several of these phases were
developed among the years, using various organic polymers such
as polymethacrylate, polystyrene, polydivinylbenzene, polyamide,
 Method Development of Proteins 481

or agarose. However, their performance was extremely limited

due to their low mechanical stability, preventing the use of small
particles and high mobile-phase flow rate. Therefore, these phases
were solely considered for low-resolution separations. In addition,
the polymeric resins used in SEC can also shrink and/or swell in
various solvents.
To overcome the mechanical stability of the polymeric-
based stationary phases, porous silica particles were introduced
later to perform SEC separations at higher pressures (few
hundred bars). However, the acidic residues of silica surfaces still
required modification to minimize the possibility of strong ionic
interactions between the proteins and the silanolates. The regular
improvement of silica materials (type B silica or hybrid silica)
enabled further developments of silica particles for routine SEC
analysis. Surface modification includes unbonded, diol (1,2-
propanediol) or tetramethylsilane (TMS) modifications to minimize
the amount of available residual silanolates. Residual acidic
silanols are detrimental in SEC applications since electrostatic
interactions may occur with the sample, influencing both elution
time and peak tailing. Bouvier et al. have proved that TMS was
more effective than diol to reduce the acidity of the silanols [15],
but shorter column lifetime may be experienced. Therefore,
diol bonding remains the most predominantly used silica-
surface chemistry modification in SEC when analyzing proteins.
In addition, diol modification increases the hydrophilic nature of
the stationary phase. This ligand blocks or reacts with many of
the acidic silanol groups and neutralizes the surface. Having a
hydrophilic surface in SEC is quite important, since many of the
proteins include some hydrophobic parts, which may interact
with the stationary phase through hydrophobic interactions, and
may affect peak tailing and elution time. It can be noted that some
differences can always be observed between two diol-bonded
phases in terms of possible electrostatic and hydrophobic
interactions, due to differences in the properties of the silica itself
and also different bonding density. A recent study compared the
inertness of current state-of-the-art silica based SEC columns, and
found some important differences between them, especially in
their capability to form electrostatic interactions with monoclonal
antibodies [16].
482 Size Exclusion Chromatography Method Development for Therapeutic Proteins

As mentioned, non-specific interactions may result in protein

adsorption, elution time shifting or asymmetrical band elution
(tailing). A shift in elution time may lead to false interpretation of
molecular weights, while peak tailing can decrease the resolution
between consecutively eluted peaks. Non-specific interactions
can include hydrophobic, electrostatic and hydrogen-bond
interactions. Hydrophobic interactions occur between the protein
and hydrophobic sites of the stationary phase. Electrostatic
interactions can be classified in two categories: (a) ion-exchange
when the protein and the stationary phase carry an opposite
charge (b) ion-exclusion when the protein and the stationary
phase carry the same charge. As an example, most monoclonal
antibodies are basic proteins possessing a pI around 7–9 and
ion-exchange often occurs between the positively charged
proteins and the negatively charged residuals silanolates of the
stationary phase at physiological mobile-phase pH conditions.
To select a suitable stationary phase, the change in peak
elution time, tailing factor and recovery (peak area) has to be
investigated using different amounts of mobile-phase salt additives
in order to evaluate non-specific interactions. When plotting the
elution time or peak tailing as a function of salt concentration
or buffer concentrations for several SEC columns, the optimal
stationary phase can be selected. Figure 16.5A shows SEC
chromatograms of an IgG1 monoclonal antibody, in presence of
various salt concentrations, while Fig. 16.5B shows the impact
of salt concentration for four different stationary phases by means
of relative change in k.
As shown in Fig. 16.5, the elution time of the IgG1 monomer
peak increased with the salt concentration. However, Fig. 16.5B
clearly illustrates some significant differences between SEC
stationary phases. As example, no significant change was observed
with increasing ionic strength in column B, while an important
shift in elution time was observed with columns C or D. Obviously,
this behavior is protein dependent, and has to be studied individually
during the method development. The same methodology can
be applied to compare possible hydrophobic interactions too,
by plotting k (or its relative change) as a function of mobile-
phase organic modifier content.
 Method Development of Proteins 483

Figure 16.5 Effect of salt (NaCl) concentration on the elution time of an IgG1
monoclonal antibody. Overlaid chromatograms correspond to different
salt concentrations on column C (A) while plots of relative change in k
vs. salt concentrations have been prepared for four different columns (B).
Reproduced with permission from Elsevier.

16.3.2 Mobile-Phase Selection (Salts, Buffers, Organic

Modifier)
To work under non-denaturing conditions in aqueous SEC, the
pH of the mobile phase is generally adjusted between 6.2 and
7.4 using phosphate buffer. Mobile-phase salt additives play a
major role to limit possible electrostatic interactions between the
proteins and the stationary phase, as well as to maintain proteins
under their native conformation.
To decrease protein–stationary phase interactions, the nature
and the concentration of the salt have to be carefully optimized.
A combination of phosphate buffer and sodium chloride (NaCl)
or potassium chloride (KCl) has been widely applied for protein
analysis in SEC [2, 11, 17]. The evaluation of three common
buffers with a wide range of proteins has proved that potassium-
and sodium-salts were more effective than Tris-HCl to improve
the recovery of proteins [18]. In addition, the authors evaluated
various ionic species ((NH4)2HPO4, Na2SO4, (NH4)2SO4, MgCl2,
NaCl) and found that better resolution factors were obtained for
six different protein samples with NaCl [18]. When dealing with
the analysis of monoclonal antibodies, the addition of NaCl in
the mobile phase can significantly increase the recovery of the
HMWS, as shown in Fig. 16.6 with rituximab (pI = 9.2).
484 Size Exclusion Chromatography Method Development for Therapeutic Proteins

Figure 16.6 Impact of salt concentration added to the mobile phase on

the recovery of rituximab HMWS. HMWS elute in the first peak on the
chromatogram and their amount varies with salt concentration, while
the amount of the monomer and fragment remains constant. Reproduced
with permission from Elsevier.

Recently, we demonstrated that potassium-based salts were

more effective than sodium-based salts to limit non-specific
interactions between the stationary phase and monoclonal
antibodies [19]. Indeed, larger cations such as potassium have a
higher ability to lose their hydration layer than sodium cations and
to interact strongly with the residual silanolates of the stationary
phase. The study highlighted an improvement in selectivity between
HMWS and the protein monomer when using a concentration of
potassium salt higher than 0.2 M, in comparison with sodium salts.
In addition, the use of potassium salt resulted in higher protein
recovery especially for the HMWS. These observations are in
agreement with the fact that electrostatic interactions are often
enhanced for HMWS vs. protein monomers [20].
However, high salt concentration can cause aggregates
formation that were not originally present in the sample [21].
In addition, high ionic strength increases the risk of salting-out
effects and can promote hydrophobic interactions between the
 Method Development of Proteins 485

proteins and the stationary phase. As an example, an increase

in elution time of very hydrophobic mAb (ipilimumab) has been
observed when increasing the amount of NaCl in the mobile
phase [16]. Finally, some buffers have also been reported to
promote non-specific interactions even on inert materials [6].
Except salt cations, the effect of salt anions (phosphate,
chloride) is also worth studying. A mobile phase containing
100 mM potassium phosphate buffer and 200 mM potassium
chloride is, in our opinion, an effective combination to minimize
both electrostatic and hydrophobic interactions. Increased salt
concentration can also improve the resolution between dimer and
monomer species, since dimers carry more charges and possess a
more hydrophobic nature than the monomer. Therefore, in most
cases, the elution time of dimer peaks decreases to a larger degree
than the one of monomers, when increasing salt concentration
and finally results in higher selectivity.
Arginine was also proposed as a suitable mobile-phase
additive to improve protein recovery during SEC analysis. A
concentration of 0.20–0.75 M arginine has been reported as
an effective range to increase proteins recovery and separate
soluble oligomers [22]. With a heat-stressed mAb sample, 22%
and 67% HMWS were observed in absence and in presence of
arginine, respectively. Arginine is able to limit both electrostatic
and hydrophobic interactions with most stationary phases [22].
Arginine stabilizes proteins by refolding their structure, it solubilizes
proteins from insoluble pellets, suppresses protein aggregation
and limit non-specific adsorption [22]. While in the past the use of
arginine raised concerns as protein-denaturant, lately it has been
proved that arginine does not cause denaturation of proteins and
was therefore suitable for SEC applications [23]. However, one
limitation of arginine is that it strongly absorbs at UV wavelengths
below 220 nm.
With the latest generation of SEC materials, hydrophobic
interactions seem to be more critical than electrostatic interactions
[16]. To limit possible hydrophobic interactions, small amounts
(up to 15% (v/v)) of organic modifiers (2-propanol, methanol,
ethanol, or acetonitrile) can be added to the mobile phase for
the analysis of the most hydrophobic proteins. Antibody–drug
conjugates (ADCs) are particularly challenging samples for SEC,
due to the hydrophobicity of their cytotoxic drugs and peptide
486 Size Exclusion Chromatography Method Development for Therapeutic Proteins

linkers. In the work of Wakankar et al., the addition of 15%

2-propanol significantly reduced hydrophobic interactions, when
analyzing trastuzumab emtansine and resulted in symmetrical
peak and slightly decreased elution time [24]. However, 2-propanol
drastically increases the mobile-phase viscosity and generates
high pressure which may increase the risk of on-column protein
aggregation. In addition, with very viscous mobile phase, the peak
width is often larger than with mobile phases of low viscosity
(water or acetonitrile). To overcome this limitation with 2-propanol,
less viscous organic modifiers may be preferred. In 1984, Malawer
et al. evaluated the effect of various polar solvents on the SEC of
PVP K-15, 30, 60, and 90 using diol-derivatized gel silica columns
and found that methanol was more efficient then acetonitrile
to limit adsorption [25]. Finally, the addition of organic solvent has
to be carefully evaluated as it may alter the native conformation
of proteins and interfere with SEC aggregate analysis.
Figure 16.7 shows an example on the effect of organic
modifier (acetonitrile) on the recovery and peak shape of an ADC
(trastuzumab emtansine).

Figure 16.7 Impact of mobile-phase organic modifier (acetonitrile) content

on the recovery of ADC (trastuzumab emtansine). By increasing the
acetonitrile content, the recoveries of both monomer and dimer species
increase and peak symmetry improves.
 Method Development of Proteins 487

16.3.3 Effect of Mobile-Phase Temperature

To increase the mass transfer rate, mobile-phase temperature
could be a valuable parameter. As the temperature increases, the
mobile-phase viscosity decreases, and the analytes diffusivity
improves. It is especially important for large molecules, which
possess very low diffusivity. Please note that in Eq. 16.3, the B
term is proportional to diffusion coefficient, while the C term is
inversely proportional to the diffusion coefficient. Furthermore,
elevated temperature allows the use of higher flow rates for a given
column dimension, and therefore faster analysis can be performed.
As expected, the symmetry of protein peak is often better at
higher temperature. It is not exclusively due to the improved
mass transfer properties, but also to the reduced secondary
interactions. Even if attempts are made to work under inert SEC
conditions, secondary interactions are often observed with large
therapeutic proteins. As the temperature increases, the strength
of secondary interactions becomes weaker. The decrease of
non-specific interactions manifests in less tailed peaks.

Figure 16.8 Recommended mobile-phase and stationary-phase conditions

for the SEC of proteins.

Another important effect of temperature in SEC is the

possible shift in elution time when changing the mobile-phase
temperature. In most cases, the higher the temperature, the lower
the elution time. However, in SEC, there is no adsorption
488 Size Exclusion Chromatography Method Development for Therapeutic Proteins

(and retention); therefore, the decrease of elution time with

temperature may be unexpected. The possible explanation is the
conformational change of proteins with temperature. At higher
temperature, proteins tend to unfold and their apparent size
becomes larger; then they will be more excluded from the pores.
On the other hand, working at elevated temperature in SEC
can be risky. Indeed, many proteins are thermo-labile and suffer
from aggregation and/or fragmentation at high temperature. It has
already been reported that the observed aggregate amount is often
larger at high temperature due to temperature induced on-column
aggregation [11].
Finally, Fig. 16.8 summarizes the conditions for a successful
SEC method development, when characterizing protein samples.

16.4 Applications for Protein Analysis

The main application for SEC is the relative quantitation of
biopharmaceuticals’ aggregates. In 2016, the United States
Pharmacopeia (USP) released the monograph <129> Analytical
Procedures for Recombinant Therapeutic Monoclonal Antibodies
[26]. This chapter contains biopharmaceuticals purity assessments
by SEC and provides validated procedures. SEC has been used in
numerous stability studies to understand the effect of temperature,
pH, formulation, light and glycosylation on the aggregation of
therapeutic proteins [27–31]. Except the formulation study from
Nicoud et al., where a polymeric stationary phase was selected
(dextran covalently bonded to agarose), others studies were
generally performed on silica-based SEC materials. The separation
of protein fragments (reduced protein, partially digested or
degraded) has also been successfully performed using sub-2 µm
UHP-SEC columns [17]. Columns packed with 1.7 μm BEH particles
were also employed to separate the aggregates of insulin within
7 min [32], and some sub–3 min separations (UHP-SEC) of test
proteins and mAb aggregates were reported with very small
particles based columns (Fig. 16.9) [11]. The very hydrophobic
ADC trastuzumab emtansine was analyzed using 15% of
2-propanol on a silica-based stationary phase [24]. More
recently, the HMWS of trastuzumab emtansine were successfully
 Applications for Protein Analysis 489

separated using the Tosoh TSKgel UP-SW3000 diol-bonded and

Agilent AdvanceBioSEC (proprietary bonding) columns using purely
aqueous mobile phase [16]. Finally, a combination of preparative
protein A and SEC analysis has been reported for the analysis of
mAb aggregates and host cell proteins (HCPs) [33].

Figure 16.9 UHP-SEC separation of Β-lactoglobulin (A) and chicken egg

ovalbumin (B) HMWS using 150 × 4.6 mm columns packed with 1.7 µm
particles operated at relatively high flow rate (0.85 mL/min). Reproduced
with permission from Elsevier.

The use of volatile salts has attracted attention in the last

few years, to make SEC compatible with mass spectrometry (MS).
Typically, 75 mM ammonium acetate is used as mobile-phase
additive for this hyphenation [34]. Combination of low amounts
of volatile salts (<200 mM, e.g., ammonium formate) and organic
modifier (<20%, e.g., acetonitrile) has also been investigated for
SEC-MS analysis of peptides and proteins under non-denaturing
conditions [35]. This allows direct coupling of SEC to electrospray
ionization time-of-flight mass spectrometry (ESI-TOF-MS) and
online characterization of the species. However, if quantitation is
required, the effect of the volatile salts on non-specific interactions
should be evaluated, as there might be a risk of underestimating
the HMWS due to the nature and the low concentrations of the
salt [20]. Liu et al. have successfully performed SEC-MS of a
reduced mAb using a mobile phase containing 20% acetonitrile
with TFA and formic acid as mobile-phase additives [36].
Lately, two-dimensional liquid chromatography (2D-LC)
emerged as an effective strategy for the analysis of complex
490 Size Exclusion Chromatography Method Development for Therapeutic Proteins

samples. 2D-LC methodologies have been applied using SEC in the

first dimension and RPLC in the second dimension, to quantify
the amount of free cytotoxic drug (small molecule) in antibody–
drug conjugate sample [37–38]. The first dimension also provided
information about the HMWS (monomer, dimer, oligomers).
In these setups, 20% (v/v) of acetonitrile was systematically
added to the mobile phase to improve the chromatographic
performance. Finally, on-line comprehensive 2D-LC using SEC in
the first dimension, mixed-mode chromatography in the second
dimension and evaporative light scattering detector (ELSD)
have been reported for the separation and detection of various
analytes (including protein, anionic, cationic, zwitterionic, and
nonionic excipients) in a biopharmaceutical drug product [39].

16.5 Perspectives
Since the introduction of UHPLC systems and columns, which
enabled high resolution, sensitivity, and peak capacity, there have
been a variety of new analytical advancements in LC and LC–MS
based on the use of low dispersion systems [40]. In SEC, the
trend is the same as in the other modes of liquid chromatography,
namely the decrease of particle sizes and column dimensions.
In conventional SEC, 300 mm × 6–8 mm columns are commonly
used, while the recent trend in UHP-SEC is to work with
150 mm × 4.6 mm columns. Many column providers offer such
UHP-SEC columns and efforts are also made to improve the
inertness of the stationary phases. However, further decrease in
SEC column dimension is not expected, due to the loss of intrinsic
column efficiency caused by commercially available LC systems.
This behavior is related to the nature of SEC separations. Indeed,
the band variance depends on the solute retention factor k¢, which
in SEC ranges between –1 and 0. It inherently results in very
low peak variances for columns of low volumes, which would be
too strongly affected by the system variance [12].
To improve the sensitivity of SEC separations or work with
very small amounts of available samples, the use of capillary
columns can be a promising approach. However, important
modifications to a commercially available LC systems are required
to drastically reduce the system volume. Until now, the number of
 References 491

applications in this field is rather limited, but 300 mm × 300 µm

I.D. SEC capillary columns were successfully applied for the
separation of mAb fragments [41]. Hyphenation of SEC with MS
is expected to attract still more attention in the coming years for
the characterization of biopharmaceutical proteins. With this
perspective, 2D-LC applications should rapidly grow to overcome
the current challenges, e.g., the analysis of HCPs in therapeutic
protein samples.
A possible way to improve throughput in SEC is to work with
two columns in parallel. A further reduction of analysis time
can be obtained by combining interlaced sample injections with
parallelization of two narrow bore columns packed with small
particles. With this strategy, both holdup times before and after
the peaks of interest can successfully be eliminated, resulting
in significant decrease of analysis time [42]. This methodology
is based on injecting a new sample before the ongoing analysis
of a previous sample has ended. To perform parallel interlaced
SEC—with two columns, two switching valves are required to
direct the flow alternately between the autosampler, the two
columns and the detector. The use of two columns and switching
valves require two distinct programs on which pumps, autosampler
and column compartment, including the switching valves, are
controlled.

References

1. Uliyanchenko, E., Schoenmakers, P. J., and Van der Wal, S. (2011).

Fast and efficient size-based separations of polymers using ultra-
high-pressure liquid chromatography. Journal of Chromatography A,
1218(11), 1509–1518.
2. Fekete, S., Beck, A., Veuthey, J. L., and Guillarme, D. (2014). Theory
and practice of size exclusion chromatography for the analysis of
protein aggregates. Journal of Pharmaceutical and Biomedical
Analysis, 101, 161–173.
3. Barth, H. G., Saunders, G. D., and Majors, R. E. (2012). The state
of the art and future trends of size-exclusion chromatography
packings and columns. LCGC North America, 30(7), 544-563.
4. Hong, P., Koza, S., and Bouvier, E. S. (2012). A review size-
exclusion chromatography for the analysis of protein biotherapeutics
492 Size Exclusion Chromatography Method Development for Therapeutic Proteins

and their aggregates. Journal of Liquid Chromatography & Related

Technologies, 35(20), 2923–2950.
5. Kopaciewicz, W., and Regnier, F. E. (1982). Nonideal size-exclusion
chromatography of proteins: Effects of pH at low ionic strength.
Analytical biochemistry, 126(1), 8–16.
6. Arakawa, T., Ejima, D., Li, T., and Philo, J. S. (2010). The critical role
of mobile phase composition in size exclusion chromatography
of protein pharmaceuticals. Journal of Pharmaceutical Sciences,
99(4), 1674–1692.
7. Golovchenko, N. P., Kataeva, I. A., and Akimenko, V. K. (1992).
Analysis of pH-dependent protein interactions with gel filtration
medium. Journal of Chromatography A, 591(1), 121–128.
8. Engelhardt, H., and Ahr, G. (1983). Optimization of efficiency in
size-exclusion chromatography. Journal of Chromatography A, 282,
385–397.
9. Mori, S., and Barth, H. G. (2013). Size Exclusion Chromatography.
Springer Science & Business Media.
10. Popovici, S. T., and Schoenmakers, P. J. (2005). Fast size-exclusion
chromatography—Theoretical and practical considerations. Journal
of Chromatography A, 1099(1), 92–102.
11. Fekete, S., Ganzler, K., and Guillarme, D. (2013). Critical evaluation
of fast size exclusion chromatographic separations of protein
aggregates, applying sub-2 μm particles. Journal of Pharmaceutical
and Biomedical Analysis, 78, 141–149.
12. Goyon, A., Guillarme, D., and Fekete, S. (2017). The importance of
system band broadening in modern size exclusion chromatography.
Journal of Pharmaceutical and Biomedical Analysis, 135, 50–60.
13. Wheaton, R. M., and Bauman, W. C. (1953). Non‐ionic separations
with ion exchange resins. Annals of the New York Academy of Sciences,
57(3), 159–176.
14. Hjertén, S., and Mosbach, R. (1962). “Molecular-sieve” chromatography
of proteins on columns of cross-linked polyacrylamide. Analytical
Biochemistry, 3(2), 109–118.
15. Bouvier, E. S., and Koza, S. M. (2014). Advances in size-exclusion
separations of proteins and polymers by UHPLC. TrAC Trends in
Analytical Chemistry, 63, 85–94.
16. Goyon, A., Beck, A., Colas, O., Sandra, K., Guillarme, D., and Fekete, S.
(2016). Evaluation of size exclusion chromatography columns
packed with sub–3 μm particles for the analysis of biopharmaceutical
 References 493

proteins. Journal of Chromatography A, DOI: 10.1016/

j.chroma.2016.11.056.
17. Yang, R., Tang, Y., Zhang, B., Lu, X., Liu, A., and Zhang, Y. T. (2015).
High resolution separation of recombinant monoclonal antibodies
by size-exclusion ultra-high performance liquid chromatography
(SE-UHPLC). Journal of Pharmaceutical and Biomedical Analysis, 109,
52–61.
18. Wu, C. S. (1999). Column Handbook for Size Exclusion Chromatography.
Academic Press.
19. Goyon, A., Beck, A., Veuthey, J. L., Guillarme, D., and Fekete, S. (2016).
Comprehensive study on the effects of sodium and potassium
additives in size exclusion chromatographic separations of protein
biopharmaceuticals. Journal of Pharmaceutical and Biomedical
Analysis. DOI: 10.1016/j.jpba.2016.09.031.
20. Carpenter, J. F., Bain, D. L., and Johnson, G. R. (2016). Use of analytical
ultracentrifugation as an orthogonal method for size exclusion
chromatography: Assuring quality for therapeutic protein products
and meeting regulatory expectations. Analytical Ultracentrifugation,
Springer Japan, pp. 389–395.
21. Philo, J. S. (2006). Is any measurement method optimal for all
aggregate sizes and types?. The AAPS journal, 8(3), E564–E571.
22. Ejima, D., Yumioka, R., Arakawa, T., and Tsumoto, K. (2005). Arginine
as an effective additive in gel permeation chromatography. Journal
of Chromatography A, 1094(1), 49–55.
23. Ishibashi, M., Tsumoto, K., Tokunaga, M., Ejima, D., Kita, Y., and Arakawa,
T. (2005). Is arginine a protein-denaturant?. Protein Expression and
Purification, 42(1), 1–6.
24. Wakankar, A., Chen, Y., Gokarn, Y., and Jacobson, F. S. (2011). Analytical
methods for physicochemical characterization of antibody drug
conjugates. mAbs, 3(2), 161–172.
25. Malawer, E. G., DeVasto, J. K., Frankoski, S. P., and Montana, A. J. (1984).
Size exclusion chromatography of poly (vinylpyrrolidone): I. The
chromatographic method. Journal of Liquid Chromatography, 7(3),
441–461.
26. Monograph <129> Analytical Procedures for Recombinant
Therapeutic Monoclonal antibodies, USP-NF PF39(3).
27. Hawe, A., Kasper, J. C., Friess, W., and Jiskoot, W. (2009). Structural
properties of monoclonal antibody aggregates induced by freeze–
thawing and thermal stress. European Journal of Pharmaceutical
Sciences, 38(2), 79–87.
494 Size Exclusion Chromatography Method Development for Therapeutic Proteins

28. Liu, B., Guo, H., Xu, J., Qin, T., Xu, L., Zhang, J., and Dai, J. (2016).
Acid-induced aggregation propensity of nivolumab is dependent on
the Fc. mAbs, 8(6), 1107–1117.
29. Nicoud, L., Cohrs, N., Arosio, P., Norrant, E., and Morbidelli, M.
(2015). Effect of polyol sugars on the stabilization of monoclonal
antibodies. Biophysical Chemistry, 197, 40–46.
30. Hernández-Jiménez, J., Salmerón-García, A., Cabeza, J., Vélez, C., Capitán-
Vallvey, L. F., and Navas, N. (2016). The effects of light-accelerated
degradation on the aggregation of marketed therapeutic monoclonal
antibodies evaluated by size-exclusion chromatography with
diode array detection. Journal of Pharmaceutical Sciences, 105(4),
1405–1418.
31. Zheng, K., Bantog, C., and Bayer, R. (2011). The impact of glycosylation
on monoclonal antibody conformation and stability. mAbs, 3(6),
568–576.
32. Koza, S., Hong, P., and Fountain, K. J. (2012). Size-exclusion ultra
performance liquid chromatography analysis of insulin (Application
note No. 720004271EN). Retrieved from Waters Corporation
website: http://www.waters.com/webassets/cms/library/docs/
720004271en.pdf.
33. Gjoka, X., Schofield, M., Cvetkovic, A., and Gantier, R. (2014).
Combined protein A and size exclusion high performance liquid
chromatography for the single-step measurement of mAb, aggregates
and host cell proteins. Journal of Chromatography B, 972, 48–52.
34. Haberger, M., Leiss, M., Heidenreich, A. K., Pester, O., Hafenmair,
G., Hook, M., and Bulau, P. (2016). Rapid characterization of
biotherapeutic proteins by size-exclusion chromatography coupled
to native mass spectrometry. mAbs, 8(2), 331–339.
35. Schmidt, A. C., Fahlbusch, B., and Otto, M. (2009). Size exclusion
chromatography coupled to electrospray ionization mass
spectrometry for analysis and quantitative characterization of
arsenic interactions with peptides and proteins. Journal of Mass
Spectrometry, 44(6), 898–910.
36. Liu, H., Gaza-Bulseco, G., and Chumsae, C. (2009). Analysis of
reduced monoclonal antibodies using size exclusion chromatography
coupled with mass spectrometry. Journal of the American Society
for Mass Spectrometry, 20(12), 2258–2264.
37. Li, Y., Gu, C., Gruenhagen, J., Zhang, K., Yehl, P., Chetwyn, N. P.,
and Medley, C. D. (2015). A size exclusion-reversed phase two
dimensional-liquid chromatography methodology for stability and
 References 495

small molecule related species in antibody drug conjugates. Journal

of Chromatography A, 1393, 81–88.
38. Li, Y., Stella, C., Zheng, L., Bechtel, C., Gruenhagen, J., Jacobson, F.,
and Medley, C. D. (2016). Investigation of low recovery in the free
drug assay for antibody drug conjugates by size exclusion—
reversed phase two dimensional-liquid chromatography. Journal
of Chromatography B, 1032, 112–118.
39. He, Y., Friese, O. V., Schlittler, M. R., Wang, Q., Yang, X., Bass, L. A.,
and Jones, M. T. (2012). On-line coupling of size exclusion
chromatography with mixed-mode liquid chromatography for
comprehensive profiling of biopharmaceutical drug product. Journal
of Chromatography A, 1262, 122–129.
40. Gazal, E. (2014). Can size exclusion chromatography (SEC) be done
on sub-3 um particles. In 17th annual meeting of the Israel Analytical
Chemistry Society, Tel Aviv, Israel.
41. Rea, J. C., Moreno, G. T., Vampola, L., Lou, Y., van Haan, B., Tremintin,
G., and Farnan, D. (2012). Capillary size exclusion chromatography
with picogram sensitivity for analysis of monoclonal antibodies
purified from harvested cell culture fluid. Journal of Chromatography
A, 1219, 140–146.
42. Farnan, D., Moreno, G. T., Stults, J., Becker, A., Tremintin, G., and van
Gils, M. (2009). Interlaced size exclusion liquid chromatography
of monoclonal antibodies. Journal of Chromatography A, 1216(51),
8904–8909.
Chapter 17

Countercurrent Chromatography Using

Biphasic Molecular Liquids and Ionic
Liquids: Is This the Future for Laboratory
and Process Purifications?

Leslie Brown and Martyn J. Earle

AECS-QuikPrep Ltd, Unit 5, Higher Trevonick Business Park,
Winnards Perch, St Columb Major, Cornwall, England TR9 6DH, United Kingdom
[email protected], [email protected]

17.1 Introduction
Countercurrent chromatography is a separation technique for
the isolation of a mixture using a liquid stationary phase held in
place by centrifugal force. Many texts are available that describe
countercurrent chromatography and related instrumentation
[1–6]. In addition, several papers by Berthod et al. discuss the use
of ionic liquids in centrifugal partition chromatography (CPC)

Chromatographic Method Development

Edited by Gregory K. Webster and Laila Kott
Copyright © 2020 Jenny Stanford Publishing Pte. Ltd.
ISBN 978-981-4800-53-2 (Hardcover), 978-0-429-20172-1 (eBook)
www.jennystanford.com
498 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

and CCC [7–12]. It has been suggested, that the high viscosity of
ionic liquids would be applicable in hydrodynamic (planetary
centrifuges) or hydrostatic (sun centrifuges), countercurrent
chromatography [8]. Berthod et al. reported that the use of ionic
liquids in CCC was found to have a number of drawbacks such
as “UV absorbance limiting the use of the convenient UV detector,”
“non-volatility precluding the use of the evaporative light-scattering
detector for continuous detection” and “high viscosity producing
pressure build-up” [8]. The authors conclude that CCC was
unsuitable at that time, as the instrumentation available “was not
able to cope with the high back-pressures generated by the large
viscosities of many ionic liquids” [8]. This view of the use of ionic
liquids in CCC was correct with the instrumentation available at
the time. However, with careful experimental design and the use
of wider coils and pipework, it has been possible to use ionic
liquids in CCC. Several papers have been published that show
the successful use of ionic liquids and related deep eutectic
solvents (DES) in CCC [13–19].

17.1.1 Countercurrent Chromatography/Centrifugal

Partition Chromatography Instrumentation for
Ionic Liquid Applications
Research into the fundamental understanding of differences and
similarities between 2D and 3D coil winding in CCC is ongoing.
This coil arrangement is unique in enabling multi-layer 2D coil
interactions between the phases of a biphasic solvent system to be
visualized and recorded. This 2D coil system can be interchanged
with 3D multi-layer coils of various beta vales (ratio of sun to
planet radii). It has been confirmed that ionic liquid solvent
systems behave differently than biphasic solvent systems created
from molecular solvents.

17.1.2 CCC/CPC Separations Using Ionic Liquids

Earlier publications stating that ionic liquids could not be easily
utilized with CCC/CPC [8] may have held back the development
of ionic liquid solvent systems in CCC/CPC. However, recently
biphasic IL’s were successfully employed in the direct separation
 Introduction 499

of a mixture of CoCl2, NiCl2, and CuCl2, the separation of two

monosaccharides from a disaccharide, and the extraction of
cumene from hexane [15]. By the use of hexane/ionic liquid
solvent system, the separation of hydrocarbons from oxygenated
hydrocarbons in vetiver oil was achieved [20]. In addition, ionic
liquid solvent systems have been studied in the countercurrent
chromatographic separation nonpolar lipids [16]. Cao et al. have
used dilute solutions of ionic liquids in CCC, in the determination
of Alternaria mycotoxins, and the determination of chlorophenols
in red wine [13, 14]. Also, four patents have been published
on the use of ionic liquid in CCC solvent systems in general [21],
in bio-organic [22], organic [23], and inorganic separations [24].
Other projects have carried out research into the use of
chiral ionic liquids [25, 26] as the stationary phase in chiral
separations. This research has been based on earlier chiral
separations using liquid–liquid chromatography in the literature.
Two examples of chiral separations by CCC are in the use of
solutions of chiral salts [27] (similar to an ionic liquid), derived
from cinchona alkaloid, in the separation of N-(3,5-dinitrobenzoyl)-
(+/-)-leucine. Similarly, cyclodextrin and a chiral copper complex
have been used in the enantioseparation of naringenin by
CCC [28].

17.1.3 Aqueous Biphasic Solvent Systems Using Ionic

Liquids
Aqueous biphasic solvent (ABS) systems [29] in CCC are used for
the separation of polar, or strongly hydrogen bonding solutes [30].
They are usually made from mixtures of concentrated inorganic
salts combined hydrophilic polymers, or alcohols. The two phases
generated in ABS systems both have water as a major component
in each phase. The first use of ionic liquid-based ABS systems has
been described by Berthod et al. [10] and more recently by Earle
et al. [15, 22]. The solvent system in these three publications used
was based on a mixture of 1-butyl-3-methylimidazolium chloride
([C4mim]Cl) and 1.5–3.0M K2[HPO4](aq) which forms two relatively
viscous liquid phases [10, 31]. The distribution ratios of solutes
can be adjusted by varying the concentration of K2[HPO4] present
in the mobile phase. This solvent system has been used in the
500 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

separation of glucose or fructose from sucrose [23] and in the

purification of the anti-cancer drug lentinan [22]. With a 2.1 mm
internal diameter (ID) coil, high backpressures can be encountered
at high flow rates and ambient temperatures [15]. In order for
the coils to withstand the 70 Bar backpressure generated at
>25 cm3 min–1 flow rate and temperatures at <40°C, the coils must
be constructed out of stainless steel, rather than the more
conventional PTFE tubing. The PTFE flying leads, PTFE tubing and
PEEK connectors must also be able to withstand this pressure,
and were tested by prolonged running (48 h), at maximum rotation
speed of the instrument while maintaining 70 Bar continuous
backpressure.
Lentinan, found in shiitake mushrooms (Lentinus edodes),
has been used in ancient Chinese medicine for the treatment of
cancer and is used as an adjunct to therapy in combination with
chemotherapeutic drugs such as fluorouracil in order to modulate
the body’s immune system activity [32]. Lentinan derived from
shiitake mushrooms is a β-1,3-1,6-glucan polymer with β-1,6
branching on a β-1,3-glucan backbone. Lentinan naturally exists
in a triple-helix form in water and salt solutions [33], but is easily
denatured. In one clinical trial, it was used to cure 30% of the
patients suffering from terminal gastric cancer, and approximately
doubled the life expectancy of 70% of the patients who did not
survive [32, 34]. The purification of lentinan is normally an
extremely difficult process involving up to 10 steps [35]. The
isolation not only requires resolution from crude mushroom extract
but also needs the maintenance of the polysaccharide’s three-
dimensional, triple-helix configuration to maintain its bioactivity.
Lentinan is denatured to a single helix form, which is not as
biologically or medicinally, active by heating at temperatures above
80°C. This denaturing of lentinan has been observed in DMSO-
based solvent systems at 25°C [36]. Lentinan has been purified
by HSCCC using an aqueous biphasic solvent system based on
polyethylene glycol (PEG) and aqueous mixtures of K[H2PO4] and
K2[HPO4] [37]. The main drawbacks with this process are that
the lentinan produced is contaminated with PEG and phosphate
salts. Also, the scale of this separation was less than 100 mg
per run, which is too low for commercial production. The
[C4mim]Cl/2.5M K2[HPO4](aq) (1:1) solvent system [10] has the
remarkable property that it does not damage or denature complex
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 501

biological molecules when they are dissolved, and this property

has been used in the purification of lentinan. The ionic liquid-
ABS process uses the K2[HPO4](aq) as the stationary phase and
[C4mim]Cl as the mobile phase. The lentinan was separated from
the [C4mim]Cl ionic liquid phase simply by the addition of ethanol
to the ionic liquid solution, which then precipitates the triple-
helix form of lentinan. In biological testing against cancer cells,
the lentinan produced by the ionic liquid process methodology
has the highest biological activity available [38].

17.2 Optimization of Ionic Liquid and Molecular

Biphasic Solvent Systems for CCC/CPC

17.2.1 Introduction
Ionic liquid research came about as a result of three discoveries:
(1) the adoption of water stable neutral ionic liquids such as
[C4mim][PF6] [39–41] and [C4mim][BF4] [42, 43], (2) the discovery
that a wide range of chemical reactions can be carried out in ionic
liquids [44–46], and (3) the realization that the number and range,
and diversity of potential ionic liquids that could be synthesized
is enormous [47].
Ionic liquids have unique properties that can help overcome
some of the problems associated with the use of conventional
organic solvents [44, 47]. Ionic liquids are materials composed
entirely of ions, with low melting points (usually <100°C) and
can dissolve a huge range of organic and inorganic compounds,
polymers, bio-molecules and other salts [48]. They have no
measurable vapor pressure at ambient temperatures and pressures
[49], resulting in their non-flammability and fire resistance [50].
These properties and their general stability make many ionic
liquids inherently safe and environmentally friendly solvents [47].
Their use in synthesis [44], catalysis [51] and separations [52]
has expanded significantly in recent years and they are now being
applied in a number of industrial processes [53]. Moreover, the
properties of ionic liquids can be tuned or adjusted by the choice
of the appropriate anion and cation allowing them to be optimized
for a particular application [47].
502 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

17.2.2 Ionic Liquids in Biphasic Solvent Selection for CCC

Ionic liquids phase behavior can be varied and used to produce
two, three [54] or four-phase solvent systems in combination
with conventional molecular solvents [55]. The most common
classification of ionic liquids is whether they are either hydrophobic
or hydrophilic. The term hydrophilic means that the ionic
liquid is miscible with water in all proportions and the term
hydrophobic corresponds to an ionic liquid that can form a
biphasic mixture with water. It should be noted that hydrophobic
ionic liquids, when mixed with water, often contain substantial
amounts of water in the ionic phase [56]. Another factor to note
is that many solvents which form biphasic mixtures with ionic
liquids, have very low solubility of the ionic liquid in the
molecular solvent phase. This is despite the ionic phase containing
appreciable amounts of the molecular solvent. The difference
in mutual solubility of ionic liquids and molecular solvents was
the miscibility of pentan-1-ol and [C6mim][BF4] at different
temperatures [57]. It has been observed that ionic liquids are not
necessarily miscible with each other, and biphasic ionic liquid
mixtures can be produced [55]. This generally occurs when
there is a considerable difference in the size and shape of both
the cation and anion. The mutual miscibility of two ionic liquids (1-
ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)amide)
[C2mim][NTf2] and (trihexyltetradecylphosphonium bis(trifluor
omethanesulfonyl)amide) are very different, with [C2mim][NTf2]
being 15 times more soluble in [P6 6 6 14][NTf2] than [P6 6 6 14][NTf2]
[C2mim][NTf2] [55].
The phenomena of the very low solubility of hydrophobic
ionic liquids in water and non-polar solvents like hexane [52] are
very important in ionic liquid–liquid chromatography. As a result,
the mobile water or organic solvent containing phase is not or
minimally contaminated with ionic liquid. This makes product
isolation very straightforward, and organic solvent recovery simply
involves distilling off the volatile solvent, which can then be reused.
The effect of dissolving a molecular solvent in an ionic liquid
dramatically reduces the ionic phase’s viscosity [58] to the point
where it can be effectively pumped by an HPLC. This has enabled
viscous hydrophobic ionic liquids such as [P6 6 6 14]Cl to be used in
CCC (viscosity = 1450 or 2729 cP at 25°C, depending on the source
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 503

of the measurement) [15, 24, 59]. An example of this (was measured

by Stark et al.) is seen in the change in viscosity and density of the
ionic liquid [C4mim][BF4] as water was added [58]. The result of
this was that the viscosity of the ionic was dramatically reduced
while the density showed a linear relationship with the change in
composition of the mixture. The addition of only 2 wt.% of water
to [C4mim][BF4] halved the viscosity of the ionic liquid phase. In
addition, the reduction in viscosity of the [C4mim][BF4] phase
is largely independent of the structure of the molecular solvent.
The viscosity of the ionic phase can therefore be predicted and
modeled by the equation h = hs exp (–xcs/a).
The biphasic solvent systems that employ ionic liquids for
use in CCC and CPC generally use mixtures with one or two
molecular solvents, rather than three to five components that
are commonly found in conventional CCC solvent systems.
The approach to solvent selection is to design or select an ionic
liquid with a chemical structure which is suited to a particular
separation. For polar, strongly hydrogen bonding solutes, ionic
liquids with small anions and cations, and for low polarity solutes
ionic liquids with long hydrocarbon chains are used. Examples of
ionic liquid solvent systems tested in separations are summarized
in Table 17.1 and the structure of the ionic liquids is shown
in Fig. 17.1.

Table 17.1 Phase combinations used for ILLC separations

Ionic liquid phase

Ionic liquid density and pump Mobile
phase direction phase Separation
[C12mim][NTf2] MDP Hexane Alkene/oxygenated
T to H terpenes in vetiver oil
[C10mim][OTf] MDP Hexane Alkane/aromatics
T to H
[C4mim][OTf] MDP Toluene/ Cyclohexanol from
T to H hexane cyclohexanone
[C2mim][OTf] MDP Toluene/ Polycyclic aromatic
T to H hexane hydrocarbons
[C4mim]Cl LDP 3.0 M Glucose/sucrose
T to H K2[HPO4]
(Continued)
504 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

Table 17.1 (Continued)

Ionic liquid phase

Ionic liquid density and pump Mobile
phase direction phase Separation
[C4mim]Cl LDP 2.5 M Lentinan
H to T K2[HPO4]
[P6 6 6 14]Cl/ LDP Water Aqueous Co(II), Ni(II),
ethyl ethanoate H to T Cu(II) chlorides
Note: Light gray = stationary phase; dark gray = mobile phase; MDP = most
dense phase; LDP = least dense phase; T to H = tail to head of coil pump
direction; H to T = head to tail of coil pump direction.

2 2

1  1 ) & 6 2 1  1
)& 6 2

2 2
>&PLP@>27I@ >&PLP@>27I@

2
>& PLP@>27I@
1  1 ) & 6 2

2 2

1  1 )& 6 1 6 &)
>&PLP@>17I @ 2 2


1  1 &O

>&PLP@&O

 
3 &O

>3  @&O

Figure 17.1 The structure of the ionic liquids used in Table 17.1.

17.2.3 CCC/CPC Method Development

CCC/CPC users should apply their fundamental chemical and
chromatographic knowledge to solvent systems and separations,
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 505

and not simply base their research on well-known biphasic mixtures.

The latter approach is still too common in CCC/CPC [60, 61].
There are very early research publications on the generic solvent
systems such as the n-hexane:ethyl acetate:methanol:distilled
water solvent systems, which, for a decade or more, was referred
to as the “Arizona system” [62] but was eventually renamed
to become the “HEMWat system” [63]. By varying the ratios of
n-hexane: ethyl acetate:methanol:water, these solvents can be used
to separate compounds with a large range of polarities. A ratio
of 1:0:1:0 (n-hexane: methanol), would be a suitable biphasic
solvent system for non-polar compounds. A ratio of 0:1:X:1 is a
good starting point for the separation of polar compounds, and a
ratio of 1:1:1:1 is a good starting position for a generic screen
of solute distribution ratios [64–66]. Changing methanol to
ethanenitrile is an option since this can improve settling times of
these biphasic solvent systems.
This use of complex mixtures approach to solvent selection
would be to assess first the 1:1:1:1 solvent mixture and then
use a sequence of pre-set, stated ratios of solvents to separate
compounds of different polarities, until a partition coefficient (K)
of between 0.5 and 3.0 is obtained [64–66]. A value of 1.0 for K is
often seen as optimal for CCC separations. K = 1 will in CCC/CPC
separations be equivalent to one coil or rotor volume of solvent.
Usually a K value of ≤ 0.5 is considered to give rise to excessively
early analyte elution and K ≥ 3 too excessively late values of
elution times for a solute or analyte. If no suitable K values
or solvencies are found, addition solvents, again with pre-set
ratios would be sequentially tried [64–66].
CCC/CPC separations need a more thought through and less
trial-and-error–based long-winded approach [65]. The principle
of LIKE/LIKES/LIKE [67] is fundamental to all chromatographic
method development. It is recommended that chromatographers
should always consider the chemistry and aim to choose
functionality, affinity, solvency (the extent to which a solute
dissolves in a given solvent (in units of g cm–3)) [68], and polarity
of solvent, to match the key functionality of target compound(s)
in the initial stages of the biphasic solvent choice. Tests should
be carried out (as would normally be carried out for recrystallization
506 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

of products), for not just solvency in single solvent, but rather

for solvency in a biphasic solvent system. These can be quite
different from single phase or single constituent solvents.
In preparative chromatography, solvency [68] is a key factor,
as higher solvency will result in a greater mass of crude product
per unit volume of the instrument coils or rotors that can be reliably
injected and separated. An example of this is in the purification
of lentinan [22], where the lentinan is much more soluble in the
ionic liquid-based solvent system 2.5 M K2[HPO4]/[C4mim]Cl,
than the more conventional aqueous K2[HPO4]/K[H2PO4]/PEG-
1000-based solvent system [37]. The result is that much higher
sample concentrations can be used, and the space time yield [69]
of the lentinan produced is increased by a factor estimated to be
5 to 10 times [22, 37].
Familiarity with the Snyder Polarity Triangle [70] and various
classifications of solvent polarity [71], and its implications can
be used in achieving orthogonal selectivity [72, 73]. The octanol:
water partition coefficients (OWPCs) [74] are available in many
publications, and websites, and an approximate value for these
OWPCs can be calculated from a compounds empirical formula
and other data [75–80]. If the functionality of a solute is known,
and a simple solvency value is obtained, plus, if the crude
determination of the OWPC’s is known, then the biphasic solvent
system should become obvious to chromatographers, without
the need for large-scale trial-and-error studies. Biphasic solvent
choice tests are very simple and described in Table 17.2.

17.2.4 The “Attraction-Repulsion Principle” of Biphasic

Eluent Optimization
The original attraction-repulsion principle (ARP) has been
recently updated to accommodate the different behavior of ionic
liquid containing biphasic eluents. ARP methodology is based on
fundamental chemical principles, rather than the utilization of
set ratios of pre-chosen solvents that make up biphasic solvent
systems of fixed composition, as is often utilized in CCC and
CPC [81].
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 507

Table 17.2 The procedure for developing CCC/CPC solvent systems

Step Procedure
1 A range of solvents of varying structure and polarity include n-
hexane, n-heptane, limonene, liquid carbon dioxide,* sub-critical
fluids such as R134a,* as relatively non-polar or low polarity choices;
ethyl ethanoate, MTBE, n-butanol, chloroform, dichloromethane
as mid polarity options; and ethanenitrile, ethanol, methanol and
distilled water as high polarity options. For unknown compounds
or analytes, screen the above solvents as a single liquid to
determine the solvency. Choose those solvents which form a rapid
settling biphasic mixture, and then test the solvency and partition
coefficient in this biphasic mixture.
2 In a 10 ml stoppered measuring cylinder, add a few milligrams of
the analyte to 2 ml of solvent to be used separately and note their
solvency.
3 Mix pairs of solvents known to form biphasic solvent systems when
combined (2 ml of each). For non-polar analytes, n-heptane or
n‑hexane, etc., mixed with ethanol/methanol/ethanenitrile/distilled
water gives some preferred solvent system options. For mid to high
polarity analytes, combinations such as n-butanol, ethyl acetate,
chloroform, dichloromethane, PEG, or dextran with distilled water
are options. Next, if appropriate, measure the partition coefficient
(K) by determining concentration of the target or analyte in both
the upper and lower phase, where K = [analyte]SP/[analyte]MP.
Also note the settling time of the biphasic solvent systems created.
This is accomplished by stoppering the measuring cylinder and
vigorously shaking for 60 s then recording the settling times and
volumes of upper and lower phase.
4 In over 80% of the cases, this simple procedure will accurately
predict the elution position of targets in CCC or CPC. Where
K = 1, the analyte will elute in one coil volume, where K = 0.5 is
half a coil volume, and for K = 2, 2 coil volumes, etc.
5 For cases where the above prediction fails, this is most often
caused by emulsion formation in the CCC or CPC instrument, and
experiments utilizing the instruments, in different modes and
using different injected masses may be needed to optimize the
method.
(Continued)
508 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

Table 17.2 (Continued)

Step Procedure
6 For biphasic solvent systems, where inappropriate or unacceptable
settling times and/or partition coefficients are found, the ARP is
used by the addition of a modifier solvent, which is miscible with the
least favored solvent phase, or a solvent which is miscible with the
most favored solvent can be added to adjust the partition coefficient
to preferred range.
7 For laboratory-only methods, simple optimization as in (1) to (6)
can be used with 1, 2, or even 3 additional solvents to rapidly achieve
the required partition coefficients, without the need to test a pre-set
number of set solvent ratios.
8 For process applications and recycling of solvents, the least number
of solvents necessary to obtain the desired distribution ratios is
preferred and often necessary. This can be achieved using ionic
liquid solvent systems that usually have only two or at most three
components. Here the ionic liquid is designed to suit the separation,
rather than using complex mixtures to give optimal separations.
This is where the use of ionic liquids can pay major dividends.
9 For CCC with molecular solvents, it is very important that the
biphasic solvent system chosen has a settling time of less than
60 s and ideally less than 30 s, to achieve good stationary phase
retention. For CPC, slower settling times for the solvent system
can be used successfully, but in our experience, this often achieves
a lower stationary phase retention than is normally found in CCC
for rapid settling biphasic eluents.
*Requires autoclave or high-pressure apparatus.
Abbreviations: MP, mobile phase; SP, stationary phase; MTBE, methyl t-butyl ether or
1-methoxy-1,1-dimethylethane; R134a, 1,1,1,2-Tetrafluoroethane.

When utilizing the ARP with molecular solvents and ionic

liquid-based biphasic eluents, consider using two immiscible
solvents as potential biphasic eluents. These are chosen based on
the known (or sometimes unknown) experimentally determined
solubility of the solute or analyte in a particular solvent (similar to
experimentation in recrystallization optimizations). With reference
to the settling time and partition coefficient of solutes (if the
target is known) (or use the polarities or bioactivity if the target
compound is not known), consider the addition of a partition
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 509

coefficient modifier (PCM) solvent or solvents. Based on the

procedure in Table 17.2, it will be found that certain solvents
will have a high affinity for compounds of interest, whilst other
solvents will display immiscibility or low solubility.
• For solvents with high solubility for the analytes, consider
these solvents as “Attractors.”
• For solvents with low solubility for the analytes, consider
these solvents as “Repulsors.”
For example, mixing n-hexane, n-heptane, cyclohexane,
pentane, toluene, chloroform, dichloromethane, ethyl acetate,
MTBE, n-butanol, etc., with distilled water will all create excellent
biphasic solvents. If methanol or ethanenitrile are added to these
biphasic mixtures, both will predominately distribute them
into the water phase (there are few exceptions, such as with
ethanenitrile preferentially dissolving in the organic phase of a
water-butanenitrile biphasic mixture). If both methanol and ethyl
acetate are added to a biphasic water-organic solvent mixture,
with which they are mutually soluble, then each will draw the
other into its preferred phase. Adding too much ethyl acetate and/
or methanol to a biphasic solvent system will result in the settling
time increasing, and eventually, at critical point, the solution
will become mono-phasic.
The solvency, K value, and settling times are all easily obtained
by adding crude mixture to be separated to the biphasic solvent
system (this only needs to be a few milligrams), and measuring
the concentration of each component of the mixture contained
in each phase. This can be achieved using the following
methodologies: GC (volatile compounds, but not ionic liquid
containing samples), HPLC, ion chromatography (for salts and
ionic liquids), 1H or 13C NMR (for organic compounds and ionic
liquids), UV-Vis spectrophotometry (for UV active or colored
compounds), polarimetry (for chiral compounds), and AA/ICP (for
metal compounds and salts). These measurements can also be
run on the individual components of the mixture providing that
they are available.
For the partitioning of the analyte into the starting biphasic
solvent system, which has been chosen to have high solubility of
the analyte, the ideal partition coefficient should be between
510 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

0.75 and 3.0. The partitioning of the analyte can be modified in

two ways. The addition of an “Attractor” solvent to the phase in
which the analyte is least soluble (in the biphasic mixture), and
which has high solvent to solvent affinity for the least favored
phase, to help draw the analyte into the least favored phase, to
better equate the partition coefficient of the analyte. For example,
if an analyte in a water/hexane solvent system is too soluble in
the water phase then the addition of ethyl ethanoate (which is
miscible with hexane and immiscible with water) will help draw
the analyte into the organic hexane/ethyl ethanoate phase.
Equally using the same logic, a “Repulsor” can be used to drive
the target compound from the more favored phase into the
least favored phase. An example with a moderately polar solute
dissolved in a water/ethyl ethanoate biphasic solvent system
that is preferentially soluble in the ethyl acetate phase. The solute
can be forced into the water phase by the addition of hexane to
the ethyl acetate phase. Both options can be utilized together as
required. The choice “attractor” or “repulsor” solvents or additives
will be driven by chemical structural knowledge of target (if
known), the solubility of the analyte in the solvents involved
(defined by very simple solubility tests), and by biphasic settling
tests (discussed below). The greater the stationary phase retention
value for biphasic solvent systems, the greater will be the resolution,
the amount of solute which can be separated, and the greater
flow rate/linear flow rate that can be achieved, which reduces
separation times.
To summarize, the appropriate biphasic solvent system can be
derived from solvency alone, since solvency is a key indicator of a
solutes polarity. As is the case with solid–liquid chromatography
used for ionic compounds, or with the use of a pH gradients in
separations, and in frontal chromatography, similar techniques
in CCC and CPC can be used, and have been found to be very
effective [82–84]. This was sometimes referred to as “pH Zone
Refining” for CCC or CPC, and this terminology has been generally
accepted by CCC/CPC chromatographers [85]. In both solid–liquid
and liquid–liquid chromatography, as the pH changes, compounds
will, at their pKas, change their mobility and move from the
stationary phase to the mobile phase. If a compound (ideally your
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 511

target) has a high enough relative loading, it will act as a buffer for
the eluting phase. Having buffered the eluent, initially all compounds
with that pKa will elute. If one solute (ideally your target) is in
relative excess, it will move only, at its buffered pH, giving a square
wave shape signal in the detector. The retention will be determined
by its solubility in the eluent, and will be the only component
to elute. Once this compound has been completely eluted, the
pH gradient will be reestablished and will not depend on other
compounds that can elute at later pHs. The use of this methodology
can often allow an increase in sample loading by 2 to 4 times,
over and above the already high loading of standard CCC/CPC
separations [86]. A list of simple solvent systems for use in CCC is
shown in Table 17.3. The solvent systems that are suitable depend
largely on the polarity of the compounds being separated.
One important factor, particularly with Flash and HPLC
chromatography, is how much mass of analyte can be retained
when complex mixtures are put onto silica or end-capped modern
reverse phase silica chromatography columns. This problem of
sample loss does not occur with CCC/CPC, when the elution/
extrusion separations are carried out. The full mass balance is
nearly always obtained with CCC/CPC type separations. Silica and
reversed-phase silica stationary phases, including fully end-capped
reverse phases, can change the nature of eluting compounds, with
respect to those injected. This can be due to on-column degradation,
and in particular, degradation of acid labile compounds [87].
Generally, CCC/CPC techniques do not decompose compounds,
assuming that the compounds are stable in the solvent system
used. For bio-active samples, if samples are run on a CCC/CPC
instrument, they generally remain bioactive. However, if samples
are run in solid-phase columns, such as Flash or HPLC columns,
the bioactive complex can be damaged, and the bioactivity lost.
CCC/CPC techniques and solvent systems can allow the retention
of the nature or structure of these molecules in the form that
they naturally occur, thereby maintaining the bioactivity. As many
natural bioactive compounds occur as complexes chromatographers
using solid–liquid chromatography denature the complexes found
in their natural state.
512 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

Table 17.3 The choice of standard biphasic solvent systems for molecular
solvents and ionic liquids (in gray background)

Polarity of analyte or solute Conventional solvent system

Ionic liquid solvent system
Very high polarity Aqueous biphasic solvents systems
(ABS), such as PEG or dextran/phosphate
salt/water
Very high polarity [C4mim]Cl/1.5 M–3.0 M K2[HPO4]
[choline][leucine]/25 wt.% K3[PO4] [88]
High polarity n-Butanol/methanol/water
(not readily recovered from n-Butanenitrile/ethanenitrile /water
above biphasic mixtures) Ethyl acetate/n-butanol/water
DMSO/ethanol: salt solution
Consider additives such as MTBE, DCM,
chloroform, etc., based on target function-
ality if known.
High polarity Water/[cholinium][NTf2], or
Water/methanol/[C2mim][NTf2]
[C4mim][OTf]/0.5 M–2.0 M K2[HPO4]
Medium Polarity Arizona (HEMWat) solvent system
[63, 66]. Consider additives such as
MTBE, DCM, Chloroform, etc., based on
target functionality if known.
Medium polarity [C2mim][OTf]/toluene
[C2mim][OMs]/ethyl acetate
Low polarity n-Hexane or heptane/methanol and/
ethanenitrile. Consider additives such as
MTBE, DCM, Chloroform, etc., based on
target functionality if known.
Low polarity [C10mim][OTf]/hexane
[C12mim][NTf2]/heptane
Very low polarity Not Available
Very low polarity [P6 6 6 14][NTf2]/hexane

As can be seen, CCC/CPC with standard solvents has

considerable potential in laboratory for both pilot and process
purifications. However, as the scale of separations increases,
the cost of recycling the solvents can become a key factor. Often,
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 513

in CCC/CPC, 3 or 4 (or more) solvents are used to form biphasic

solvent systems, which makes solvent recovery and recycling of
the solvents difficult and costly. An alternative to the use of
molecular solvents, supercritical fluids have been researched
for their use in CCC/CPC [89, 90]. Research with liquid carbon
dioxide and potentially other sub-critical fluids is ongoing, as a
means of separating mixtures without the use of volatile organic
solvents. Another approach to CCC separations is to simplify the
biphasic solvent system used to preferably two components [90].
This requires custom engineered, green, low risk, non-flammable,
low volatility, easily recyclable solvents, such as ionic liquids,
combined with a single molecular solvent or supercritical solvent
such as CO2 or subcritical solvents such as R134a (1,1,1,2-
tetrafluoroethane).

17.2.5 Ionic Liquid Solvent System Selection

For ionic liquid biphasic solvent systems, knowledge of analyte
chemistry and likely interactions with ionic liquids can be used
to choose readily available ionic liquids for solvent systems,
or this information can be used to custom synthesize the most
appropriate ionic liquid in terms of functionality, solvency and
the properties of the ionic liquid. For ionic liquids solvent system
development, a similar procedure to that in Table 17.2 is used.
For molecular solvents, it would be unusual to use only two
solvents, but this is quite common with ionic liquid solvent
systems. This is a large advantage particularly when considering
process applications since this and makes solvent recycling much
easier. Usually, the ionic liquid is selected or designed to have
the correct range of properties for a given separation. The term used
to describe the design of ionic liquids to suit a particular end use
is called solvent engineering.
For ionic liquid-based biphasic solvent systems in both CCC
and CPC, their successful use is very much less dependent on
rapid settling times. This is an advantageous situation for ionic
liquid-based methods, since many ionic liquids are surface active
and would be expected to form emulsions. Research is ongoing
as to why ionic liquid solvent systems are so much less sensitive
to rapid settling times that are required for molecular solvents
in CCC. This difference is thought to be due in part, to ionic liquid
514 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

solvent systems having a generally greater density contrast

between the two phases of the biphasic solvent system, which
helps the two phases separate under high “g” loadings in CCC/CPC
rotors.
Ionic liquids are said to be immiscible with other solvents
when there is a visible phase boundary between two phases
where at least one phase contains an ionic liquid. Immiscibility is
not however, the same as insolubility. CCC separations require
biphasic solvent systems with the phases chosen being either ionic
liquid-organic solvent, ionic liquid–water, or mixtures of two or
more immiscible ionic liquids. All the components of these biphasic
solvent systems will have a degree of solubility in each other,
but this mutual solubility can vary considerably. In most cases,
biphasic ionic liquid solvent systems have very low ionic liquid
solubility in the molecular solvent phase, and as a result,
molecular solvents are usually used as the mobile phase and
the ionic liquid containing phase as the stationary phase as in
Table 17.2. Solvent recovery often involves distillation of the
volatile molecular solvent from the mobile phase containing the
separated product, which allows the solvent to be reused [91].
Fewer volatile components are used in ionic liquid solvent
systems than the Arizona or HEMWat solvent systems [62, 92,
93], makes solvent recovery much more straightforward due to
the lower number of volatile components that require separation
[81]. Some solutes can remain dissolved in an ionic liquid phase,
after the volatile solvent recovery stage. These can be separated
from the ionic liquid by a number of different methods, such as
solvent extraction, precipitation with an antisolvent, or filtering
a solution of the ionic liquid through silica or charcoal, or a
combination of these methods [94].
The procedure for the determination of the suitability of a
biphasic ionic liquid solvent system for a given separation is as
follows: (1) add each solvent of the biphasic solvent system to the
sample to be separated, (2) measure and record the solvency in
each phase, (3) combine the two phases containing the sample and
vigorously shake for 60 s, then record settling time, (4) measure
the percentage of each compound dissolved in both the upper &
lower phase. With this information, the distribution ratio KD(X)
of each component (X) of the sample can be measured. KD(X) =
[Solute X]UP/[Solute X]LP where X = component of sample, UP = upper
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 515

phase and LP = lower phase. Conventionally, in CCC separations,

one of the phases (usually the ionic phase) is used as the
stationary phase. Therefore, KD(X) = [Solute X]SP/[Solute X]MP, where
KD(X) = distribution coefficient of Solute X, SP = stationary phase
and MP = mobile phase. In most cases, these simple tests can
be used to predict CCC/CPC elution times and the elution order
relative to the elution volume/coil volume. In some cases, the CCC/
CPC chromatogram will need to be run, before an optimization
of the solvent system composition can be carried out.
For a given biphasic solvent system, the value of KD can
be adjusted by employing the Attraction Repulsion Principle
described in Section 17.2.4 using an antisolvent to cause a given
solute to move into the opposite phase. As an example, the
distribution ratio of lentinan and saccharides in the aqueous
K2[HPO4]/[C4mim]Cl solvent system (Table 17.1) can be modified
by the addition of more K2[HPO4](s) to the water phase. The
K2[HPO4] salt is very water soluble and insoluble in the [C4mim]Cl
phase. This causes the lentinan or saccharides to become more
soluble in the ionic liquid phase; hence, K2[HPO4] can be said to
be a “repulsor” compound for the aqueous phase. These solute
attraction and repulsion effects can be utilized to fully and
predictably adjust K values. This approach, based on adjusting
the solvent system by adding to or modifying a solvent system
[95], contrasts with the systematic measurement of distribution
ratios from a large number of solvent systems described in
the literature [64, 81], or by using a computational approach
[96, 97].
The performance of CCC/CPC separations depends on differences
in the distribution ratios (KD) of compounds distributed between
two liquid phases. Also, mass transfer of solutes between these
two phases is a factor that affects separation quality. With ionic
liquid solvent systems, the viscosity of the two liquid phases is one
of the most important factors. Viscous phases in CCC/CPC can
result in excessive bandspreading [98] and high backpressures
which require low mobile phase flow rates. Since commonly used
ionic liquids have viscosities in the range of 35–2000 cP [99],
compared with molecular solvents with viscosities in the range of
0.3 to 10 cP, the use ionic liquids in CCC can be problematic. The
high viscosities encountered with ionic liquid solvent systems
can be minimized by running CCC/CPC separations at higher
516 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

temperatures than are conventionally used [100–102]. The

properties ionic liquids, including viscosity, density, solvent
behavior, and interfacial tensions are available in the literature
[103–106] and in the IST Standard Reference Database [107].
This data should be treated with caution, because when an ionic
liquid is mixed with a molecular solvent, the properties of the
subsequent mixture will be different and with the viscosity of
the ionic phase being greatly reduced [59, 108]. This can be
estimated using Eyring’s absolute rate theory [108].
For the separations in Table 17.1, the biphasic combinations
of ionic liquids and molecular solvents were chosen based on
previous research into ionic liquid phase behavior and solvent
extraction studies [21, 23, 24, 52, 53, 55, 109–114]. Factors
affecting the miscibility of solutes in an ionic liquid include the size
of the ions, ability of the ions to form hydrogen bonds, its
hydrophilicity/hydrophobicity, and the basicity of the anion. Large
mostly aliphatic cations (e.g., [P6 6 6 14]+) with very weakly basic
anions (e.g., [NTf2]– or [OTf]–) produce a phase which is highly
immiscible with polar organic compounds (e.g., water or water/
methanol mixtures), whereas small polar or strongly hydrogen
bonding cations (e.g., cholinium) with small or basic anions (e.g.,
chloride, ethanoate or hydroxide) are very miscible with polar
organic solvents solutes (e.g., water and methanol). The
hydrophobicity/hydrophilicity of ionic liquids can also be
calculated using software such as COSMO-RS [103].
For effective solvent selection, the distribution ratios of the
solutes between the two phases of the solvent system should be
in the range of 0.5 to 3.0. Distribution ratios (KD) can be measured
by several means, including electronic absorption, HPLC,
NMR spectroscopy of the solute dissolved in each phase. Gas
chromatography must not be used for ionic liquid containing
phases, because the ionic liquid will not be eluted from the column,
and the properties of the column will be altered. For simple
biphasic ionic liquid/molecular solvent mixtures, the addition of
a co-solvent (usually a second molecular solvent) will significantly
affect the distribution ratios of the solutes, as well as significantly
reducing the viscosity of the ionic phase [8]. This is a very effective
way of fine-tuning the solute distribution ratios to optimize
separations [8].
 Optimization of Ionic Liquid and Molecular Biphasic Solvent Systems for CCC/CPC 517

The combination of ionic liquid and molecular solvent(s)

chosen must have a large enough density difference between these
two phases (>0.05 g cm–3). The hydrophobic ionic liquid
[C10mim][OTf] (1.17 g cm–3 when pure) [115] cannot be used in
combination with water, because the density of the two phases
(water and water saturated ionic liquid) have similar densities
are slow to separate from each other. However, [C6mim][OTf]
(1.22 g cm–3 when pure), and [C8mim][OTf] (1.19 g cm–3 when
pure) are slightly more dense than water and readily separate [115].
Some combinations of ionic liquid and molecular solvent should
be avoided due to the formation of inseparable emulsions.
Although some ionic liquids can be used to break (demulsify)
emulsions [116], surface active ionic liquids like [C12mim]Cl [117],
when combined with halogenated solvents (CH2Cl2 or CHCl3), can
produce emulsions which will not separate during CCC operation.

17.2.6 Summary of the CCC Use of Ionic Liquid–Liquid

Chromatography
Ionic liquid solvent system advantages
• Can be designed for a specific separation
• Greater range of possible solvent systems
• Greater solute capacity
• Process intensification
• Greater versatility in separations and operating conditions
• Greater range of solutes that can be separated
• Greater range of biphasic solvent systems available
• Allows separations of previously insoluble materials
• Compatible with biological compounds
• Very polar and non-polar compound separations
• Direct metal separations
• Simple solvent systems
• Easy solvent recycling
• Ionic liquids can be recycled
• Best for process scale
Ionic liquid solvent system disadvantages
• Requires solvent engineering to design produce ionic liquids
518 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

• Lack of research and development of ionic liquid solvent

systems
• Higher viscosities and backpressures
• Bandspreading from higher viscosity of solvent system
Ionic liquid general properties
• Have no vapor pressure at atmospheric pressures [49]
• Millions of potential ionic liquids [47]
• Can form triphasic and tetraphasic solvent systems [40, 55]
• Can be designed for specific end use or phase behavior [105]
• High solute capacity allowing improved space time yields
[106]

17.3 Conclusion
The use of molecular solvents as biphasic eluents has been shown
to be excellent for a huge variety of applications in laboratory
scale preparations of up to kilo or multi-kilo of target. But the
complexity of molecular solvent biphasic eluents with three, four
or on occasions five solvents, becomes less favored in larger-
scale processes. For these larger-scale process applications, of tens
of kilos to many tons per annum, the unique properties of ionic
liquids can provide many advantages, which we have discussed
in the chapter. This can make the additional effort of developing
ionic liquid eluents well worthwhile. The use of ionic liquids in
countercurrent chromatography instruments was initially found
to be problematic due to high backpressures and consequent
low mobile phase flow rates. With careful design of the fluid flow
paths and by eliminating pressure bottlenecks, this problem has
now been solved. Currently there are very few papers in the
literature describing the use of solvent systems containing a high
percentage of ionic liquids in CCC or LLC. Control over the structure
and design of ionic liquids and ionic liquid containing solvent
systems (solvent engineering) has enabled us to select or alter
solute distribution ratios of solutes that are to be separated.
Peptides, proteins, monoclonal antibodies (mAbs), enzymes,
bioactive compounds, precious metals, actinides and lanthanides,
are all ideal candidates for ionic liquid CCC/CPC separation
processes, as all are high value, but can be potentially difficult
 References 519

to purify cost-effectively from their start crude matrices. Ionic

liquid-CCC/CPC/CCE/CPE science and research are likely going to
be driven by commercial applications rather than in the academic
arenas. Research groups are already in place in the USA and Europe
to develop the processes necessary for very large-scale process
applications.

Acknowledgements
We would like to express our gratitude to Invest NI for the
Collaborative Grants for R&D: RD0210506 and RD1111867, the
EPSRC, and the QUILL Research Centre for funding this work. We
would also like to thank Professor Paul Davey and Givaudan for
useful discussions.

References
1. Snyder, L. R., Kirkland, J. J. (1979). Introduction to Modern Liquid
Chromatography, 2nd ed., John Wiley & Sons Inc., New York.
2. Conway, W. D., Petroski, R. J. (1995). Modern Countercurrent
Chromatography, American Chemical Society, Washington, DC.
3. Degenhardt, A., Winterhalter, P. (2001). HSCCC–A powerful tool for
the preparative isolation of bioactive compounds, Pfannhauser W,
Fenwick GR, Khokhar S in Biologically-Active Phytochemicals in Food,
eds. Pfannhauser W, Fenwick GR, Khokhar S, pp. 143–146.
4. Berthod, A. (2009). Countercurrent Chromatography: From the
Milligram to the Kilogram, Grushka E, Grinberg N in Advances in
Chromatography, Vol 47, eds. Grushka E, Grinberg N, CRC Press-Taylor
& Francis Group, Boca Raton, pp. 323–352.
5. Cazes, J. (2009). Encyclopedia of Chromatography, 3rd ed., Taylor &
Francis, Boca Raton.
6. Sutherland, I. A. (2010). Industrial applications of CCC, Cazes J in
Encyclopedia of Chromatography, eds. Cazes J, 3 ed., Taylor & Francis,
Boca Raton, pp. 2116–2119.
7. Berthod, A., Carda-Broch, S. (2003). A new class of solvents for CCC:
Room temperature ionic liquids, J. Liq. Chromatogr. Relat. Technol., 26,
pp. 1493–1508.
8. Berthod, A., Carda-Broch, S. (2004). Use of the ionic liquid 1-butyl-
3-methylimidazolium hexafluorophosphate in countercurrent
chromatography, Anal. Bioanal. Chem., 380, pp. 168–177.
520 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

9. Berthod, A., Ruiz-Ángel, M. J., Huguet, S. (2005). Nonmolecular solvents

in separation methods: Dual nature of room temperature ionic
liquids, Anal. Chem., 77, pp. 4071–4080.
10. Ruiz-Ángel, M. J., Pino, V., Carda-Broch, S., Berthod, A. (2007). Solvent
systems for countercurrent chromatography: An aqueous two
phase liquid system based on a room temperature ionic liquid, J.
Chromatogr. A, 1151, pp. 65–73.
11. Berthod, A., Ruiz-Ángel, M., Carda-Broch, S. (2008). Ionic liquids in
separation techniques, J. Chromatogr. A, 1184, pp. 6–18.
12. Berthod, A., Ruiz-Ángel, M.-J., Carda-Broch, S. (2009). Ionic Liquids
as Stationary Phases in Countercurrent Chromatography, Koel M in
Ionic Liquids in Chemical Analysis, eds. Koel M, CRC Press, Boca Raton,
pp. 211–227.
13. Fan, C., Li, N., Cao, X. L. (2015). Determination of chlorophenols in
red wine using ionic liquid countercurrent chromatography as a
new pretreatment method followed by high-performance liquid
chromatography, J. Sep. Sci., 38, pp. 2109–2116.
14. Fan, C., Cao, X. L., Liu, M., Wang, W. (2016). Determination of
Alternaria mycotoxins in wine and juice using ionic liquid modified
countercurrent chromatography as a pretreatment method followed
by high-performance liquid chromatography, J. Chromatogr. A, 1436,
pp. 133–140.
15. Brown, L., Earle, M. J., Gîlea, M. A., Plechkova, N. V., Seddon,
K. R. (2017). Ionic liquid-liquid separations using countercurrent
chromatography: A new general purpose separation methodology,
Aust. J. Chem., 70, pp. 923–932.
16. Müller, M., Englert, M., Earle, M. J., Vetter, W. (2017). Development
of solvent systems with room temperature ionic liquids for the
countercurrent chromatographic separation of very nonpolar lipid
compounds, J. Chromatogr. A, 1488, pp. 68–76.
17. Roehrer, S., Bezold, F., Garcia, E. M., Minceva, M. (2016). Deep eutectic
solvents in countercurrent and centrifugal partition chromatography,
J. Chromatogr. A, 1434, pp. 102–110.
18. Ma, S. F., Hu, L. M., Ma, C. Y., Lv, W. P., Wang, H. X. (2014). Application
and recovery of ionic liquids in the preparative separation of four
flavonoids from Rhodiola rosea by on-line three-dimensional liquid
chromatography, J. Sep. Sci., 37, pp. 2314–2321.
19. Cai, M., Su, J., Jin, M., Song, Z. (2015). Extraction of triptolide and
tripdiolide from Tripterygium wilfordii using ionic liquid. Patent
CN104710502A.
 References 521

20. Brown, L., Earle, M. J., Gîlea, M. A., Plechkova, N. V., Seddon, K. R.
(2017). Ionic liquid-liquid chromatography: A new general purpose
separation methodology, Top. Curr. Chem., 74, pp. 1–41.
21. Earle, M. J., Seddon, K. R., Self, R., Brown, L. (2013). Ionic Liquid
Separations. Patent WO2013121218A1.
22. Earle, M. J., Gilea, M. A. (2013). Lentinan extraction process from
mushrooms using ionic liquid. Patent WO2013140185A1.
23. Earle, M., Seddon, K. (2013). Ionic liquid separations. Patent
WO2013121219A1.
24. Earle, M. J., Seddon, K. R. (2013). Ionic Liquid Separations. Patent
WO2013121220A1.
25. Foreiter, M. B., Gunaratne, H. Q. N., Nockemann, P., Seddon, K. R.,
Srinivasan, G. (2014). Novel chiral ionic liquids: physicochemical
properties and investigation of the internal rotameric behaviour in
the neat system, Phys. Chem. Chem. Phys., 16, pp. 1208–1226.
26. Payagala, T., Armstrong, D. W. (2012). Chiral ionic liquids:
A compendium of syntheses and applications (2005–2012), Chirality,
24, pp. 17–53.
27. Franco, P., Blanc, J., Oberleitner, W. R., Maier, N. M., Lindner, W.,
Minguillon, C. (2002). Enantiomer separation by countercurrent
chromatography using cinchona alkaloid derivatives as chiral
selectors, Anal. Chem., 74, pp. 4175–4183.
28. Wang, S. S., Han, C., Wang, S. S., Bai, L. J., Li, S. S., Luo, J. G., Kong,
L. Y. (2016). Development of a high speed counter-current
chromatography system with Cu(II)-chiral ionic liquid complexes
and hydroxypropyl-beta-cyclodextrin as dual chiral selectors for
enantioseparation of naringenin, J. Chromatogr. A, 1471, pp. 155–163.
29. Willauer, H. D., Huddleston, J. G., Rogers, R. D. (2002). Solvent
properties of aqueous biphasic systems composed of polyethylene
glycol and salt characterized by the free energy of transfer of a
methylene group between the phases and by a linear solvation
energy relationship, Ind. Eng. Chem. Res., 41, pp. 2591–2601.
30. Ishii, K., Tanaka, Y., Hata, K., Goto, M., Saitoh, K., Minamisawa, H.,
Shibukawa, M. (2004). Separation of inorganic ions by high-speed
countercurrent chromatography with an aqueous biphasic system,
Bunseki Kagaku, 53, pp. 911–917.
31. Freire, M. G., Claudio, A. F. M., Araujo, J. M. M., Coutinho, J. A. P.,
Marrucho, I. M., Lopes, J. N. C., Rebelo, L. P. N. (2012). Aqueous biphasic
systems: a boost brought about by using ionic liquids, Chem. Soc.
Rev., 41, pp. 4966–4995.
522 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

32. Ina, K., Furuta, R., Kataoka, T., Kayukawa, S., Yoshida, T., Miwa, T.,
Yamamura, Y., Takeuchi, Y. (2011). Lentinan prolonged survival in
patients with gastric cancer receiving S-1-based chemotherapy,
World Journal of Clinical Oncology, 2, pp. 339–343.
33. Xu, X. J., Wang, X. H., Cai, F., Zhang, L. N. (2010). Renaturation of triple
helical polysaccharide lentinan in water-diluted dimethylsulfoxide
solution, Carbohydr. Res., 345, pp. 419–424.
34. Ina, K., Kataoka, T., Ando, T. (2013). The Use of Lentinan for
Treating Gastric Cancer, Anti-Cancer Agents in Medicinal Chemistry,
13, pp. 681–688.
35. Chihara, G., Hamuro, J., Maeda, Y. Y., Arai, Y., Fukuoka, F. (1970).
Fractionation and purification of the polysaccharides with marked
antitumor activity, especially lentinan, from Lentinus edodes (Berk.)
Sing.(an edible mushroom), Cancer research, 30, pp. 2776–2781.
36. Xu, X. J., Zhang, X. F., Zhang, L. N., Wu, C. (2004). Collapse and
association of denatured lentinan in water/dimethlysulfoxide
solutions, Biomacromolecules, 5, pp. 1893–1898.
37. Jiang, Z. G., Du, Q. Z., Sheng, L. Y. (2009). Separation and purification of
lentinan by preparative high speed counter current chromatography,
Chin. J. Anal. Chem., 37, pp. 412–416.
38. Hughes, K., Paisley, N. (2015). Hughes Mushrooms Ltd.
39. Earle, M. J., McCormac, P. B., Seddon, K. R. (1998). Regioselective
alkylation in ionic liquids, Chem. Commun., pp. 2245–2246.
40. Carmichael, A. J., Earle, M. J., Holbrey, J. D., McCormac, P. B., Seddon,
K. R. (1999). The Heck reaction in ionic liquids: A multiphasic
catalyst system, Org. Lett., 1, pp. 997–1000.
41. Earle, M. J., McCormac, P. B., Seddon, K. R. (2000). The first high
yield green route to a pharmaceutical in a room temperature
ionic liquid, Green Chem., 2, pp. 261–262.
42. Holbrey, J. D., Seddon, K. R. (1999). The phase behaviour of 1-alkyl-
3-methylimidazolium tetrafluoroborates, ionic liquids and ionic
liquid crystals, J. Chem. Soc.-Dalton Trans., pp. 2133–2139.
43. Seddon, K. R., Stark, A., Torres, M. J. (2002). Viscosity and density of
1-alkyl-3-methylimidazolium ionic liquids, Abraham MA, Moens
L in Clean Solvents: Alternative Media for Chemical Reactions and
Processing, eds. Abraham MA, Moens L, Amer. Chem. Soc., Washington,
pp. 34–49.
44. Earle, M. J., Seddon, K. R. (2000). Ionic liquids. Green solvents for the
future, Pure Appl. Chem., 72, pp. 1391–1398.
 References 523

45. Earle, M. J. (2002). Ionic liquids: Solvents for the twenty-first century,
Rogers RD, Seddon KR in Ionic Liquids: Industrial Applications
for Green Chemistry, eds. Rogers RD, Seddon KR, Amer. Chem. Soc.,
Washington, pp. 90–105.
46. Earle, M. J., Seddon, K. R. (2002). Clean synthesis in ionic liquids,
Trulove PC, DeLong HC, Mantz RA, Stafford GR, Matsunaga M in
Molten Salts Xiii, eds. Trulove PC, DeLong HC, Mantz RA, Stafford GR,
Matsunaga M, Electrochemical Soc Inc, Pennington, pp. 177–189.
47. Freemantle, M. (2010). An Introduction to Ionic Liquids Royal Society
of Chemistry, Cambridge, UK.
48. Wasserscheid, P., Welton, T. (2008). Ionic Liquids in Synthesis, 2nd ed.,
Wiley-VCH, Weinheim, Germany.
49. Earle, M. J., Esperança, J. M. S. S., Gîlea, M. A., Lopes, J. N. C., Rebelo, L. P.
N., Magee, J. W., Seddon, K. R., Widegren, J. A. (2006). The distillation
and volatility of ionic liquids, Nature, 439, pp. 831–834.
50. Kim, H.-T., Kang, J., Mun, J., Oh, S. M., Yim, T., Kim, Y. G. (2016). Pyrrolinium-
based ionic liquid as a flame retardant for binary electrolytes of
lithium ion batteries, ACS Sustain. Chem. Eng., 4, pp. 497–505.
51. Earle, M. J., Wasserscheid, P., Schulz, P. C., Olivier-Bourbigou, H., Favre,
F., Vaultier, M., Kirschning, A., Singh, V., Riisager, A., Fehrmann, R.,
Kuhlmann, S. (2008). Organic Synthesis, Welton T, Wasserscheid P
in Ionic Liquids in Synthesis, eds. Welton T, Wasserscheid P, 2nd ed.,
Wiley-VCH, Weinheim, pp. 265–568.
52. Arce, A., Earle, M. J., Rodriguez, H., Seddon, K. R. (2007). Separation
of aromatic hydrocarbons from alkanes using the ionic liquid 1-ethyl-
3-methylimidazolium bis{(trifluoromethyl) sulfonyl} amide, Green
Chem., 9, pp. 70-74.
53. Plechkova, N. V., Seddon, K. R. (2008). Applications of ionic liquids in
the chemical industry, Chem. Soc. Rev., 37, pp. 123–150.
54. Earle, M. J., Katdare, S. P., Seddon, K. R. (2004). Paradigm confirmed:
The first use of ionic liquids to dramatically influence the outcome of
chemical reactions, Org. Lett., 6, pp. 707–710.
55. Arce, A., Earle, M. J., Katdare, S. P., Rodriguez, H., Seddon, K. R. (2006).
Mutually immiscible ionic liquids, Chem. Commun., pp. 2548–2550.
56. Martins, V. L., Nicolau, B. G., Urahata, S. M., Ribeiro, M. C. C., Torresi,
R. M. (2013). Influence of the Water Content on the Structure and
Physicochemical Properties of an Ionic Liquid and Its Li+ Mixture, J.
Phys. Chem. B, 117, pp. 8782–8792.
57. Wagner, M., Stanga, I., Schroer, W. (2004). Critical viscosity near the
liquid-liquid phase transition in the solution of the ionic liquid 1-
524 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

methyl-3-hexylimidazolium tetrafluoroborate in 1-pentanol, Phys.

Chem. Chem. Phys., 6, pp. 1750–1757.
58. Seddon, K. R., Stark, A., Torres, M. J. (2000). Influence of chloride,
water, and organic solvents on the physical properties of ionic
liquids, Pure Appl. Chem., 72, pp. 2275–2287.
59. Neves, C., Carvalho, P. J., Freire, M. G., Coutinho, J. A. P. (2011).
Thermophysical properties of pure and water-saturated
tetradecyltrihexylphosphonium-based ionic liquids, J. Chem.
Thermodyn., 43, pp. 948–957.
60. CamachoFrias, E., Foucault, A. (1996). Solvent systems in centrifugal
partition chromatography, Analusis, 24, pp. 159–167.
61. Sutherland, I. A. (2007). Review of centrifugal liquid-liquid
chromatography using aqueous two-phase solvent systems: Its
scale-up and prospects for the future production of high-value
biologicals, Curr. Opin. Drug Discov. Dev., 10, pp. 540–549.
62. Berthod, A., Hassoun, M., Ruiz-Angel, M. J. (2005). Alkane effect in
the Arizona liquid systems used in countercurrent chromatography,
Anal. Bioanal. Chem., 383, pp. 327–340.
63. Chadwick, L. R., Fong, H. H. S., Farnsworth, N. R., Pauli, G. F. (2005). CCC
sample cutting for isolation of prenylated phenolics from hops, J. Liq.
Chromatogr. Relat. Technol., 28, pp. 1959–1969.
64. Dubant, S., Mathews, B., Higginson, P., Crook, R., Snowden, M.,
Mitchell, J. (2008). Practical solvent system selection for counter-
current separation of pharmaceutical compounds, J. Chromatogr. A,
1207, pp. 190–192.
65. Brown, L., Luu, T. A. (2010). An introduction to discussions on
liquid-liquid/counter current/centrifugal partition chromatography
biphasic solvent selection methodologies and instrumentation for
target compound preparation from complex matrices, J. Sep. Sci., 33,
pp. 999–1003.
66. Liang, J. L., Meng, J., Wu, D. F., Guo, M. Z., Wu, S. H. (2015). A novel 9 x
9 map-based solvent selection strategy for targeted counter-current
chromatography isolation of natural products, J. Chromatogr. A, 1400,
pp. 27–39.
67. Carraher Jr., C. E. (2017). Introduction to Polymer Chemistry, 4th ed.,
CRC press, Boca Raton, FL.
68. Fleer, G. J., Skvortsov, A. M. (2005). Theory for concentration and
solvency effects in size-exclusion chromatography of polymers,
Macromolecules, 38, pp. 2492–2505.
 References 525

69. Kohlmann, C., Leuchs, S., Greiner, L., Leitner, W. (2011). Continuous
biocatalytic synthesis of (R)-2-octanol with integrated product
separation, Green Chem., 13, pp. 1430–1436.
70. Dolan, J. W. (2010). Selectivity in reversed-phase LC separations, Part
I: Solvent-type selectivity, LCGC North America, 28, pp. 1022–1027.
71. Lesellier, E. (2015). Sigma pider diagram: A universal and versatile
approach for system comparison and classification Application to
solvent properties, J. Chromatogr. A, 1389, pp. 49–64.
72. Tan, X., Tan, D. M., Brown, L., Yang, D. P., Zhu, L. P., Chen, Y. L., Chang, D.
H. T., Wang, D. M. (2013). Isolation and purification of four bioactive
constituents from alpinia officinarum hance utilizing high-speed
counter-current chromatography, J. Liq. Chromatogr. Relat. Technol.,
36, pp. 1355–1365.
73. Pellett, J., Lukulay, P., Mao, Y., Bowen, W., Reed, R., Ma, M., Munger, R.
C., Dolan, J. W., Wrisley, L., Medwid, K., Toltl, N. P., Chan, C. C., Skibic, M.,
Biswas, K., Wells, K. A., Snyder, L. R. (2006). “Orthogonal” separations
for reversed-phase liquid chromatography, J. Chromatogr. A, 1101,
pp. 122–135.
74. Kamath, G., Bhatnagar, N., Baker, G. A., Baker, S. N., Potoff, J. J. (2012).
Computational prediction of ionic liquid 1-octanol/water partition
coefficients, Phys. Chem. Chem. Phys., 14, pp. 4339–4342.
75. Soskic, M., Plavsic, D. (2005). Modeling the octanol-water partition
coefficients by an optimized molecular connectivity index, J. Chem Inf.
Model., 45, pp. 930–938.
76. Sedykh, A. Y., Klopman, G. (2006). A structural analogue approach to
the prediction of the octanol-water partition coefficient, J. Chem Inf.
Model., 46, pp. 1598–1603.
77. Xie, Y. J., Liu, H., Liu, H. X., Zhai, Z. C., Wang, Z. Y. (2008). Determination
of solubilities and n-Octanol/Water partition coefficients and QSPR
study for substituted phenols, Bull. Environ. Contam. Toxicol., 80,
pp. 319–323.
78. Hsieh, C. M., Lin, S. T. (2009). Prediction of 1-octanol-water partition
coefficient and infinite dilution activity coefficient in water from
the PR plus COSMOSAC model, Fluid Phase Equilib., 285, pp. 8–14.
79. Souza, E. S., Zaramello, L., Kuhnen, C. A., Junkes, B. D., Yunes, R. A.,
Heinzen, V. E. F. (2011). Estimating the octanol/water partition
coefficient for aliphatic organic compounds using semi-empirical
electrotopological index, Int. J. Mol. Sci., 12, pp. 7250–7264.
526 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

80. Bhatnagar, N., Kamath, G., Chelst, I., Potoff, J. J. (2012). Direct
calculation of 1-octanol-water partition coefficients from adaptive
biasing force molecular dynamics simulations, J. Chem. Phys., 137,
pp. 014502.
81. Garrard, I. J. (2005). Simple approach to the development of a CCC
solvent selection protocol suitable for automation, J. Liq. Chromatogr.
Relat. Technol., 28, pp. 1923–1935.
82. Kukula-Koch, W., Koch, W., Angelis, A., Halabalaki, M., Aligiannis, N.
(2016). Application of pH-zone refining hydrostatic countercurrent
chromatography (hCCC) for the recovery of antioxidant phenolics
and the isolation of alkaloids from Siberian barberry herb, Food
Chem., 203, pp. 394–401.
83. Kotland, A., Chollet, S., Autret, J. M., Diard, C., Marchal, L., Renault, J. H.
(2015). Modeling pH-zone refining countercurrent chromatography:
A dynamic approach, J. Chromatogr. A, 1391, pp. 80–87.
84. Vieira, M. N., Leitao, S. G., Porto, P. C. C., Oliveira, D. R., Pinto, S. C.,
Braz Filho, R., Leitao, G. G. (2013). Application of pH-zone-refining
countercurrent chromatography for the separation of indole
alkaloids from Aspidosperma rigidum Rusby, J. Chromatogr. A, 1319,
pp. 166–171.
85. Wang, T., Jiang, X., Yang, L., Wu, S. (2008). pH-gradient counter-current
chromatography isolation of natural antioxidant chlorogenic acid
from Lonicera japonica Thumb. using an upright coil planet centrifuge
with three multi-layer coils connected in series, J. Chromatogr. A, 1180,
pp. 53–58.
86. Ito, Y. (2013). pH-zone-refining counter-current chromatography:
Origin, mechanism, procedure and applications, J. Chromatogr. A,
1271, pp. 71–85.
87. Connolly, B. A. (1987). The synthesis of oligonucleotides containing
a primary amino group at the 5’-terminus, Nucleic acids research, 15,
pp. 3131–3139.
88. Wang, R., Chang, Y., Tan, Z., Li, F. (2016). Applications of choline
amino acid ionic liquid in extraction and separation of flavonoids
and pectin from ponkan peels, Sep. Sci. Technol., 51, pp. 1093–1102.
89. Yeh, C. S., Yu, T., Berthod, A. (1999). Separation of steroids by
countercurrent chromatography using supercritical fluid carbon
dioxide, J. Liq. Chromatogr. Relat. Technol., 22, pp. 345–356.
90. Yu, T., Chang, J. Y. (2001). Measurements of partition coefficients in
water-1,1,1,2-tetrafluoroethane by countercurrent chromatography,
J. Liq. Chromatogr. Relat. Technol., 24, pp. 1607–1615.
 References 527

91. Mai, N. L., Ahn, K., Koo, Y. M. (2014). Methods for recovery of ionic
liquids-A review, Process Biochem., 49, pp. 872–881.
92. Faure, K., Bouju, E., Doby, J., Berthod, A. (2014). Limonene in
Arizona liquid systems used in countercurrent chromatography. II
Polarity and stationary-phase retention, Anal. Bioanal. Chem., 406,
pp. 5919–5926.
93. Friesen, J. B., Pauli, G. F. (2005). GUESS - A generally useful estimate
of solvent systems in CCC, J. Liq. Chromatogr. Relat. Technol., 28,
pp. 2777–2806.
94. Earle, M. J., Gordon, C. M., Plechkova, N. V., Seddon, K. R., Welton, T.
(2007). Decolorization of ionic liquids for spectroscopy, Anal. Chem.,
79, pp. 758–764.
95. Costa, F. D., Leitao, G. G. (2010). Strategies of solvent system selection
for the isolation of flavonoids by countercurrent chromatography,
J. Sep. Sci., 33, pp. 336–347.
96. Hopmann, E., Arlt, W., Minceva, M. (2011). Solvent system selection
in counter-current chromatography using conductor-like screening
model for real solvents, J. Chromatogr. A, 1218, pp. 242–250.
97. Yin, L. H., Xu, L. N., Wang, X. N., Lu, B. N., Liu, Y. T., Peng, J. Y. (2010).
An economical method for isolation of dioscin from dioscorea
nipponica makino by HSCCC coupled with ELSD, and a computer-
aided UNIFAC mathematical model, Chromatographia, 71, pp. 15–23.
98. Berthod, A., Friesen, J. B., Inui, T., Pauli, G. F. (2007). Elution-
extrusion countercurrent chromatography: Theory and concepts
in metabolic analysis, Anal. Chem., 79, pp. 3371–3382.
99. Cocalia, V. A., Visser, A. E., Rogers, R. D., Holbrey, J. D. (2008).
Physicochemical Properties, Wasserscheid P, Welton T in Ionic Liquids
in Synthesis, eds. Wasserscheid P, Welton T, 2nd ed., Wiley-VCH,
Weinheim, pp. 89–102.
100. Domanska, U., Wlazlo, M. (2016). Thermodynamics and limiting
activity coefficients measurements for organic solutes and water in
the ionic liquid 1-dodecyl-3-methylimidzolium bis(trifluoromethylsu
lfonyl) imide, J. Chem. Thermodyn., 103, pp. 76–85.
101. Clough, M. T., Crick, C. R., Grasvik, J., Hunt, P. A., Niedermeyer, H.,
Welton, T., Whitaker, O. P. (2015). A physicochemical investigation
of ionic liquid mixtures, Chem. Sci., 6, pp. 1101–1114.
102. Tariq, M., Shimizu, K., Esperança, J. M. S. S., Lopes, J. N. C., Rebelo, L. P.
N. (2015). Viscosity minima in binary mixtures of ionic liquids plus
molecular solvents, Phys. Chem. Chem. Phys., 17, pp. 13480–13494.
528 Countercurrent Chromatography Using Biphasic Molecular Liquids and Ionic Liquids

103. Zhou, T., Chen, L., Ye, Y., Chen, L., Qi, Z., Freund, H., Sundmacher, K.
(2012). An overview of mutual solubility of ionic liquids and water
predicted by COSMO-RS, Ind. Eng. Chem. Res., 51, pp. 6256–6264.
104. Liu, W., Cheng, L., Zhang, Y., Wang, H., Yu, M. (2008). The physical
properties of aqueous solution of room-temperature ionic liquids
based on imidazolium: Database and evaluation, J. Mol. Liq., 140,
pp. 68–72.
105. Zhang, S. J., Sun, N., He, X. Z., Lu, X. M., Zhang, X. P. (2006). Physical
properties of ionic liquids: Database and evaluation, J. Phys. Chem.
Ref. Data, 35, pp. 1475–1517.
106. Pereiro, A. B., Araujo, J. M. M., Teixeira, F. S., Marrucho, I. M., Pineiro,
M. M., Rebelo, L. P. N. (2015). Aggregation behavior and total
miscibility of fluorinated ionic liquids in water, Langmuir, 31,
pp. 1283–1295.
107. NIST Standard Reference Database #147, (2016) http://ilthermo.
boulder.nist.gov/index.html (verified 19-12-2016).
108. Atashrouz, S., Zarghampour, M., Abdolrahimi, S., Pazuki, G.,
Nasernejad, B. (2014). Estimation of the viscosity of ionic liquids
containing binary mixtures based on the Eyring’s theory and a
modified Gibbs energy model, J. Chem. Eng. Data, 59, pp. 3691–3704.
109. Liu, W. W., Cheng, L. Y., Zhang, Y. M., Wang, H. P., Yu, M. F. (2008).
The physical properties of aqueous solution of room-temperature
ionic liquids based on imidazolium: Database and evaluation, J. Mol.
Liq., 140, pp. 68–72.
110. Arce, A., Earle, M. J., Rodriguez, H., Seddon, K. R., Soto, A. (2010).
Isomer effect in the separation of octane and xylenes using the ionic
liquid 1-ethyl-3-methylimidazolium bis{(trifluoromethyl)sulfonyl}
amide, Fluid Phase Equilib., 294, pp. 180–186.
111. Arce, A.; Earle, M. J., Rodriguez, H., Seddon, K. R., Soto, A. (2009).
Bis{(trifluoromethyl)sulfonyl}amide ionic liquids as solvents for
the extraction of aromatic hydrocarbons from their mixtures
with alkanes: effect of the nature of the cation, Green Chem., 11,
pp. 365–372.
112. Arce, A., Earle, M. J., Rodriguez, H., Seddon, K. R., Soto, A. (2008).
1-Ethyl-3-methylimidazolium bis{(trifluoromethyl)sulfonyl}amide as
solvent for the separation of aromatic and aliphatic hydrocarbons by
liquid extraction–extension to C-7- and C-8-fractions, Green Chem., 10,
pp. 1294–1300.
113. Arce, A., Earle, M. J., Rodriguez, H., Seddon, K. R. (2007). Separation
of benzene and hexane by solvent extraction with 1-alkyl-3-
 References 529

methylimidazolium bis{(trifluoromethyl)sulfonyl}amide ionic liquids:

Effect of the alkyl-substituent length, J. Phys. Chem. B, 111, pp. 4732–
4736.
114. Arce, A., Earle, M. J., Katdare, S. P., Rodriguez, H., Seddon, K. R. (2007).
Phase equilibria of mixtures of mutually immiscible ionic liquids,
Fluid Phase Equilib., 261, pp. 427–433.
115. Mantz, R. A., Trulove, P. C. (2008). Physicochemical Properties,
Wasserscheid P, Welton T in Ionic Liquids in Synthesis, eds.
Wasserscheid P, Welton T, 2nd ed., Wiley-VCH, Weinheim, pp. 72–88.
116. Abullah, M. M. S., Al-Lohedan, H. A., Attah, A. M. (2016). Synthesis
and application of amphiphilic ionic liquid based on acrylate
copolymers as demulsifier and oil spill dispersant, J. Mol. Liq., 219,
pp. 54–62.
117. Hezave, A. Z., Dorostkar, S., Ayatollahi, S., Nabipour, M., Hemmateenejad,
B. (2013). Investigating the effect of ionic liquid 1-dodecyl-3-
methylimidazolium chloride ([C12mim]Cl) on the water/oil interfacial
tension as a novel surfactant, Colloid Surf. A-Physicochem. Eng. Asp.,
421, pp. 63–71.
 Index

ABS, see aqueous biphasic solvent analyte determination 400–401,

acetic acid 102, 107, 126, 129, 131, 404
169, 200, 418–419 analyte ions 210, 245, 247
acetonitrile (ACN) 23, 64, 89–91, analytical methods 14–16, 18,
94, 101, 103, 109–110, 21–22, 42–44, 46–48, 72, 86,
121–122, 124–128, 130, 144, 184, 201, 218, 357, 458
165, 167–168, 171, 195, 197, development 14–18, 20–23,
200–201, 253, 350, 485–486, 25–27, 29–31, 33, 35–37, 39,
489–490 41–44, 48
ACN, see acetonitrile parameters 14–15, 26, 29, 31, 36,
active pharmaceutical ingredients 41, 44, 46
(APIs) 23, 161, 259, 338, 346 analytical procedures 398
ADCs, see antibody drug conjugates
analytical quality by design (AQbD)
AE, see 2-aminoethanol
14–15, 48
AEX pH gradient separation 446,
anion-exchange 77, 221–225, 227,
461
229–233
Alprenolol 142–144
materials 224, 229–230, 232
alumina 56, 216–217
nanoparticles 224–226, 229
amino acids 173, 191, 238, 251,
254, 259, 280, 356 phases 212, 224–225, 227, 229,
2-aminoethanol (AE) 131, 139 231, 233
ammonium acetate 102–104, 107, separations 224
126, 435 anions 76, 210, 224, 226–228, 231,
amylose 147, 170–171 235, 238, 240–242, 244–245,
analysis 247, 252, 284–289, 294, 434,
achiral 111 436, 502, 516
anion 210, 240–241 halide-based 285–286, 288
beverage 194–195 ANP, see aqueous normal-phase
cation 210, 241 antibody drug conjugates (ADCs)
chiral GC 280 9, 415–418, 420, 422, 424,
FAME 292 426, 428, 430, 432, 434, 436,
food 188 438, 440, 442, 444, 446,
ion chromatographic 241 448–450, 452, 454, 456, 458,
main component assay 353–354 460, 462, 464, 466–468, 477,
n-alkane 290–291 485–486
substance 351–354 antisolvent 514–515
532 Index

APIs, see active pharmaceutical cation-exchange phases 211–213,

ingredients 215, 217, 219, 221, 223–224
approach, in-column weak acid 217, 219–220
polymerization 294 cation exchange separation mode
approaches 442–444
derivatization 358, 360 CCC, see countercurrent
in-column 293 chromatography
phased 16–17 CCC/CPC separations 505, 515
AQbD, see analytical quality by CD, see circular dichroism
design centrifugal partition
aqueous biphasic solvent (ABS) chromatography (CPC) 497,
499–500, 512 503, 506–508, 510, 513
aqueous normal-phase (ANP) 7, chemical ionization (CI) 335–336,
179–180, 182–184, 186, 338, 342, 353
188–192, 194, 196, 198, chemical linking 281, 283, 294
chemical selectivities 263, 266,
200–202
275, 286–289, 294–295
arginine 472, 485
chemistry, manufacturing, and
controls (CMC) 398
chiral GC columns 294, 356
chiral selectors 132–133, 137, 157,
biomolecules 417, 501
170–172, 174, 189, 280, 294
biphasic mixtures 502, 505, 507,
chiral separations 3, 6, 121–123,
509–510, 512
128, 135, 140, 144, 148,
Bis-Tris propane 418, 425, 431,
157–158, 163–164, 171,
439, 451, 455, 462, 466 280–281, 324, 499
bisphenol A (BPA) 188 chiral stationary phases (CSPs) 79,
BPA, see bisphenol A 115–120, 122, 124–126, 128,
130, 132–134, 136, 138,
140–142, 144, 146, 148,
171–172, 280, 282
calcium 212, 214–216, 218, 220, CHIRALCEL 116–117, 170
250, 266 CHIRALPAK 116–117, 120–121,
carbon 329–330, 343, 352 131–134, 136–139, 141,
carbon atoms 180, 276, 330–331, 146–147, 174
343–344 chloroform 57–58, 123, 144, 350,
carbonate 232, 235, 241–243, 245, 507, 509, 512
250 chromatographic efficiency 102,
carrier gas 311–312, 317, 104, 225
319–322, 325, 327–329, 332, chromatographic methods,
351 normal-phase 63
cation-exchange 77, 211–212, chromatographic performance 31,
216–217, 219–220, 222–223, 213–214, 219, 225, 232, 272,
420–421, 423, 437–440 288, 490
 Index 533

chromatographic profiles 325, chromatography data systems 238,

424–425, 428, 430, 449–451, 394
453 chromophore 93, 161, 163
chromatographic resolution 352 CI, see chemical ionization
chromatographic retention 14, 132 circular dichroism (CD) 123,
chromatographic retention factor 159–162, 168, 246–247, 258,
272 281–282, 472
chromatographic separation, choice CMC, see chemistry, manufacturing,
of 163 and controls
chromatographic separation column chemistries 8, 168, 237,
methods 38 369, 375
chromatography 2–10, 13, 53, 55, choice of 8, 237
61, 66, 74, 79, 89, 101, 107, column chromatography theory,
132, 136, 148, 157, 162–165, traditional 9
167, 169, 179, 209, 237, 247, column flushing 128, 145
250, 258, 309, 311, 333, 348, column orthogonality 67, 69, 71,
394, 427, 471, 497–498, 517 82, 93, 110
achiral 157, 161 column regeneration 146–147
centrifugal partition 497 column screening 66, 90, 158
chiral 6, 79, 158, 160, 162, 164, column selectivity 67, 100, 171
166, 168, 170, 172, 174 column temperature control 460
counter-current 10 column thermostatting 389, 391
current 10 columns
gas-liquid 265 chiral 170, 174
high-resolution 394 chromatographic 180
high temperature 376 end-capped modern reverse
hydrophilic interaction 25, phase silica chromatography
107–109, 179 511
ion 7, 21, 209–212, 217, 224, fused silica capillary 266
227–228, 233, 238, 240, 242, HILIC 25, 82, 195
244, 246, 248–254, 256, HPLC 239, 376, 382, 511
258–260, 509 hydride-based 187
low-pressure column 3 immobilized 120, 128–130,
modes of 82, 164 145–146
preparative 124, 506 ion chromatography 209–210,
reverse phase 125 212, 214, 216, 218, 220, 222,
reverse-phase 158, 165, 254 224, 226, 228, 230, 232, 234,
reversed- phase 63 258
reversed-phase 5–6, 67, 69, 71, ion-exchange 239, 253
86–87, 89–91, 93, 95, 97, 103, liquid chromatography 239
111 micropacked 265
reversed-phase column 63 narrow-bore 319–320, 480
size exclusion 9, 471 normal-phase achiral 72
suppressed ion 220 phenyl 188, 200
534 Index

polysaccharide 170–171 diamond hydride (DH) 132,

reversed- phase 95 191–200
silica 179–180, 182, 184, 186, diethylamine 104–105, 129, 218,
188, 190, 192, 194, 196, 198, 220
200, 202 divinylbenzene 215, 225, 229
silica hydride 188, 190–191, 199,
202
solid-phase 511
UHPLC 1, 375–376, 379–380, ECD, see electron-capture detector
382, 384, 386, 388 ECNI, see electron capture negative
ultrafast separation 174 ionization
WCOT 267, 314 EI, see electron ionization
compounds electron-capture detector (ECD)
bioactive 518 328, 331, 360
hydrophilic 128, 181, 184–185, electron capture negative ionization
191–192
(ECNI) 336–338, 342
isobaric 191–192
electron ionization (EI) 334,
nonpolar 86–87, 200, 202, 288,
337–340, 342, 344–345, 353
505
ELSD, see evaporative light
conditions
scattering detector
inert 472
eluant 238–239, 241–245,
isocratic 193, 201, 216–217
249–250, 252, 254–255, 257,
thermostatting 389, 391
259
conductivity 210, 220, 239,
elution, isocratic 60
241–242, 244, 246–248, 251,
258, 328 emulsions 517
conductivity detector 210, enantiomeric purity 161–162
245–246, 248 enantiomeric separations 120,
countercurrent chromatography 170, 175, 281
(CCC) 10, 497–499, enantiomers 6, 115–118, 120,
502–503, 506–508, 510–511, 122–126, 128, 130, 132–134,
513, 515, 517–518 136–138, 140, 142–144, 146,
CPC, see centrifugal partition 148, 158–159, 161–162, 170,
chromatography 175
CSPs, see chiral stationary phases enantioselectivity 119, 124–125,
cyanopropyl 74, 277–278, 293 132–133, 135, 141, 143, 146,
cyclodextrins 79, 276, 280–281, 280, 282
499 enantioseparation of naringenin
499
enantioseparations 115, 120, 171
ethanenitrile 505, 507, 509, 512
de-ionized water 245, 252 ethanol 38, 40, 91, 103, 130,
detector flow cell volume 378–379 164–165, 253, 350, 485, 501,
DH, see diamond hydride 507
 Index 535

ethyl acetate 57, 62, 253, 350, 505, GC separations, high-temperature

507, 509, 512 264, 279
ethyl acetate phase 510 genotoxic impurities 8, 259, 356
evaporative light scattering GLC, see gas-liquid chromatography
detector (ELSD) 89, 123, 490 gradient
cross-link 213–214
reversed-phase 201
scouting 322
FAMEs, see fatty acid methyl esters temperature 322, 389
fatty acid methyl esters (FAMEs) gradient delay volume 380,
268, 289–290, 292–293 385–388, 394
FID, see flame-ionization detector gradient elution 57, 127, 241–242,
flame-ionization detector (FID) 254
328–331, 352–353 linear 60–61
flash chromatography method gradient method transfer 380, 386,
development 54, 56, 58, 60, 392–393
62, 64 gradient separations 7, 237, 257,
FPP, see fully porous particles 369, 387, 392
fully porous particles (FPP) 101,
107, 110, 172

HCAs, see hierarchical cluster

analysis
gas chromatographic column HCPs, see host cell proteins
performance 268 hierarchical cluster analysis (HCAs)
gas chromatography (GC) 8, 13, 76, 360
15, 55, 65, 263–264, 272, 280, high-molecular-weight species
283, 286, 294, 309–311, 319, (HMWS) 472, 483–485,
321, 325, 328, 346–347, 349, 489–490
351, 353–358, 360–361, 509, high-performance liquid
516 chromatography (HPLC) 1,
gas chromatography method 3, 8, 10, 13, 55, 117, 119–120,
development 264, 266, 268, 125, 148, 188, 191, 210, 346,
270, 272, 274, 276, 278, 280, 367–369, 392, 502, 509, 516
282, 284, 286, 288, 290, 292, HILIC, see hydrophobic interaction
294, 310, 312, 314, 316, 318, liquid chromatography
320, 322, 324, 326, 328, 330, HILIC column selectivity 77
332, 334, 336, 338, 340, 342, HILIC separations 110
344, 346, 348, 350, 352, 354, HMWS, see high-molecular-weight
356 species
gas-liquid chromatography (GLC) Hoff equations 132–133, 388
265 Horwitz acceptance criterion 405,
GC, see gas chromatography 411
536 Index

Horwitz function 399, 401–402, ion chromatography method

404, 407, 411 development 252
host cell proteins (HCPs) 489, 491 ion concentration 9, 242, 251–252,
HPLC, see high-performance liquid 422, 457–458
chromatography ion exchange chromatography (IEC)
HPLC pump 421, 429–430, 433 9, 416–417, 421, 434
HPLC to UHPLC method transfer ionic liquid phase 503–504,
373, 375, 378, 385, 392–393 514–515
HPLC-UV methods 352–353 ionic liquid phase density
hydrogen 219, 274, 321, 329 503–504
hydrophobic interaction liquid ionic liquids (ILs) 10, 264, 266,
chromatography (HILIC) 270, 283–289, 291, 293–295,
6–7, 25, 77, 107–111, 128, 350, 360, 497–504, 506,
179, 181, 185, 189, 192, 202 508–510, 512–518
hydrophobicity 63, 68, 87–88, 375, room-temperature 285
485 isocratic mobile phase composition
hydroxide 185, 202, 220–221, 201
243–245, 250, 516 isopropanol 91, 103, 109, 165, 253

IEC, see ion exchange Japanese green tea extracts,

chromatography analysis of 327–328
IEC pH gradient separations 468
IEC separations 417–418, 421,
436, 439, 455, 457, 460, 462,
465, 468
LC, see liquid chromatography
ILs, see ionic liquids
LC separation methods 26, 38
imidazolium 286–287, 289–290
lentinan 500–501, 504, 506, 515
inorganic cations 217
linear solvation energy
interactions
relationships (LSERs) 68, 72,
analyte-stationary phase 309
74, 105
electrostatic 74, 76, 107, 417,
linear solvent strength (LSS) 23,
481–482, 484–485
25, 28
hydrophobic 68, 140, 180, 230,
liquid chromatography (LC) 2, 6–8,
481–482, 484–485
15, 22, 66, 72, 85–86, 115,
intermolecular 271, 285
119, 125–126, 130, 138, 161,
protein-stationary phase 483
209, 238–239, 252, 258, 471,
solute-stationary phase 272
stationary phase-to-analyte 320 490
ion chromatography, hydrophilic interaction 107, 109,
nonsuppressed 211, 219, 128, 181
232 reversed-phase 239, 477
 Index 537

traditional normal-phase 111 reversed-phase 63–64, 93

two-dimensional 489 reversed-phase chromatographic
ultra-high performance 13 6
ultra-high-pressure 172 systematic 16, 89
LSERs, see linear solvation energy method optimization 17–18, 20,
relationships 22, 30–31, 34, 36–38, 43, 45,
LSS, see linear solvent strength 106, 132, 135, 140, 325, 327
method scouting 18, 20, 26–29, 45
method transfer 2, 14, 255, 315,
367–386, 388–394
limitations of 389, 391, 393
magnesium 212, 214–216, 218,
MIs, see mutagenic impurities
220, 250
MLR, see multiple linear regression
MAOT, see maximum allowable
mobile phase (MP) 10, 21–23,
operating temperature
28–29, 38, 86–91, 93–94, 100,
mass spectrometry (MS) 88–89,
103–104, 108–110, 117–148,
95, 97, 101, 107, 109–110,
164–165, 167–168, 180–181,
123, 159, 164, 168, 185, 191,
183–185, 199–202, 210–212,
200, 238, 265, 270, 315, 325, 217, 220–221, 232, 252–254,
333, 335, 337–339, 341, 343, 369–370, 375–377, 388–389,
345, 354, 360, 489, 491 392–394, 416–425, 427,
maximum allowable operating 429–439, 441, 445–446,
temperature (MAOT) 449–451, 453, 455, 457–459,
269–270, 286, 290 462, 467–468, 472–473,
metabolites 191–192 483–486, 489–490, 507–508,
methanol 64, 74, 90–91, 103–104, 514–515
109, 121, 163, 165, 167–168, mobile phase additives 21, 104,
253, 350, 485–486, 505, 507, 376, 416
509, 516 mobile phase conductivity 210
method development 3–4, 6–7, 10, mobile phase development
14, 16, 18, 21–22, 31, 47, 64, 418–420, 435–436, 441, 455
66, 72, 82, 85–87, 90–92, 97, mobile phase ionic strength
124, 126, 140, 148, 238, 244, 421–422, 427, 430
252–253, 255, 257–259, 314, mobile phase optimization 451
319–320, 325, 346–347, 418, mobile phase pH 69, 100, 217,
434–435, 437, 439, 441, 443, 393, 417, 421, 430, 436,
445–447, 457, 467, 482, 488 457–459
chiral 7, 158, 174–175 mobile-phase salt additives
chiral chromatographic 158 482–483
efficient 14, 148 mobile phase screening 127
ion chromatographic 237–238 mobile phase selection 119, 121,
liquid chromatographic 86 123, 125–127, 129, 416
normal-phase liquid mobile phase solvents 199
chromatography 102 mobile phase viscosity 486–487
538 Index

mobile phases conventional organic 197

acetonitrile/water 187 normal-phase columns 73, 102,
acidified 200 164
alkali metal hydroxide 210 bonded 73–74
aqueous 23, 119, 212, 253, 489 traditional 72–73
Bis-Tris propane 437–438 NP, see normal phase
bulk 185 NPD, see nitrogen-phosphorous
chiral 168 detector
conductive 210 nutmeg oil, analysis of 325–326
hydronium-based 212
isocratic 202
methanol-water 22
non-complexing 221 octanol: water partition coefficients
non-polar 86, 103 (OWPCs) 506
organic 119, 123, 125 organic mobile phase type 97
strong acid 210 organic modifier concentration 23
traditional liquid chromatography organic modifiers 23, 26, 28–29,
165 89, 95, 102, 126–127, 144,
viscous 486 472, 485–486, 489
viscous aqueous 127 organic solvents 128, 131, 167,
volatile 89 171, 199, 226, 229, 245, 253,
molecular ion peak 342, 344 513
molecular solvent phase 502, 514 orthogonality 5–6, 65–66, 72, 74,
76–77, 79–80, 82
monoclonal antibodies 9, 415, 427,
in chromatographic methods 66,
477–478, 481, 484, 518
68, 70, 72, 74, 76, 78, 80, 82
monovalent cations 212, 214–215
HILIC stationary phase 77
MS, see mass spectrometry
OWPCs, see octanol: water partition
multiple linear regression (MLR)
coefficients
28, 72, 272
Oxprenolol 142–144
mutagenic impurities (MIs) 340,
346, 350, 356–358, 360–361
halogenated 360
PEG, low-molecular-weight 282
penehyclidine 333–335
pH adjustment 87, 416, 419–425,
nitrogen-phosphorous detector
433, 445, 449, 457, 462
(NPD) 360 pH gradient CEX separation
normal phase (NP) 25, 63, 103, 448–449
117, 127–128, 136, 144, 148, pH gradient IEC separations 453
164–165, 183 pH gradient separation 416, 418,
normal-phase chromatography 3, 420, 422, 424, 426, 428, 430,
6, 74, 99, 101, 103, 105, 432, 434, 436, 438, 440, 442,
107–108, 164–165 444, 446, 448, 450, 452, 454,
 Index 539

456, 458, 460, 462, 464, 466, polymeric ionic liquids (PILs) 264,
468 266, 283, 285, 287, 289, 291,
pH gradient separations 417, 293
419–422, 425, 427, 432, polysaccharide 116, 118, 124–128,
435–436, 438, 448, 453, 455, 130, 140, 144, 171, 500
457 potassium 211, 214–216, 218,
pH gradients 220, 243, 245, 250, 484
narrower 419–420, 451–452 programmed-temperature
wider 419, 451–452 vaporization (PTV) 311
pharmaceutical analyses 13, 111, PTV, see programmed-temperature
vaporization
191, 329, 331, 333, 340, 346,
pyridine 344, 350, 418
351, 354
pharmaceuticals 10, 328, 346, 352,
397, 400
phases
QbD, see quality by design
624-type 349, 351 quality by design (QbD) 14
chiral 167, 171–172, 174
HILIC 77–78, 185
hydrocarbon 273–274
ion chromatography 231 refractive index (RI) 89, 123, 195
organic 109, 167–168, 509 relative standard deviation (RSD)
PEG 283 354–355, 400–407, 409, 429
silica hydride 181, 189, 191, 195, response surface methodology
199, 201–202 (RSM) 29, 352
solid 157 retention factor 72, 77, 95, 122,
phenyl 28, 36, 90, 188, 275–279 133, 143, 272, 315, 369, 388
phosphate 126, 226–228, 231, retention of hydrophilic compounds
235, 241, 250, 254, 485 184–185, 191
PILs, see polymeric ionic liquids retention time shifts 422, 430,
Plackett–Burman design 29, 33, 457, 459–460, 465
38, 43 retention volume 474–475
plasma 200, 278, 331–333 reversed phase (RP) 63, 118, 125,
polar analytes 77, 107–108 127–128, 131, 137, 145, 148,
179–180, 186, 188–190
polar compounds 86–87, 180, 183,
reversed-phase liquid
189, 191, 199, 202, 282, 318
chromatography (RPLC) 21,
analysis of 189, 264
239, 477, 490
retention of 185
reversed-phase liquid
separation of 282, 505
chromatography method
polar organic (PO) 66, 79, 108, development 89, 93
118, 122, 127, 144, 163, 172, RI, see refractive index
186, 288, 516 room-temperature ionic liquids
polar solvents 121, 125, 144, 146 (RTILs) 284–285, 350
540 Index

RP, see reversed phase hybrid 110, 481

RPLC, see reversed-phase liquid ordinary 181–184, 186, 190, 197,
chromatography 202
RSD, see relative standard deviation silica-based materials 185,
RSM, see response surface 210–211
methodology silica hydride 181–186, 189, 197,
RTILs, see room-temperature ionic 199, 202
liquids silica hydride-based separation
materials 186–187, 189,
191, 193, 195, 197
silica hydride materials 179, 184,
scouting 26, 29, 54–55 186, 189, 192, 195, 202
sensitivity 3, 54, 85, 244, 248, 258, silica hydride surface 182, 184,
316, 325, 329, 378, 416, 420, 189, 191
436, 490 siloxanes 264, 266, 273–277,
separation
279–283, 294
achiral SFC 104
software
cation exchange 421–422
chromatographic simulation
gas chromatographic 266, 310
14–15
high-resolution 310
chromatography data 388, 394
isothermal 322, 324
software for chromatographic
low-resolution 481
methods development 14,
normal-phase 66, 102
16, 18, 20, 22, 24, 26, 28, 30,
orthogonal 70, 74
32, 34, 36, 38, 40, 42, 44, 46,
reversed-phase 66, 69, 111, 187
single-buffer ionic strength 418 48
temperature-programmed 324 solid-phase extraction (SPE) 10
ultrafast 173 solvation parameter model 271,
separation efficiency 264, 269, 273, 290–292
287, 294, 315, 317, 320, 325, solvent recovery 513–514
327, 376, 379, 383 SP, see stationary phase
separation methods SPE, see solid-phase extraction
orthogonal 5, 65 squalane 271, 273
pH gradient IEC 434 static coating methods 267–268,
SFC, see supercritical fluid 293
chromatography stationary phase (SP) 5–8, 10,
silica 7, 53, 56–60, 62–63, 77–78, 68–69, 72–74, 76–77,
102, 106, 108, 170–171, 86–87, 89–92, 100, 102–106,
179–184, 186–189, 197, 199, 108–109, 115–116, 124–126,
202, 210–211, 216–217, 473, 132, 136, 140, 148, 168,
481, 511, 514 170–172, 179–192, 199–200,
bare 72, 74, 77, 110 210–217, 219, 221–222,
fused 266–267, 279, 282 224–226, 228–230, 232–234,
 Index 541

239, 263–295, 309, 313–314, polysaccharide 165, 171

316–317, 320–322, 375–379, polysaccharide-based chiral 148
390, 392, 471–472, 480–485, polysaccharide-derived chiral
508, 510–511, 514–515 116, 118, 120, 122, 124, 126,
stationary phase bonding chemistry 128, 130, 132, 134, 136, 138,
280 140, 142, 144, 146, 148
stationary phase chemistry 104, reversed-phase silica 511
239, 259, 390 silica 7, 103, 108
stationary phase coatings 225, 272 silica-based 210, 488
stationary phase descriptors 72, silica hydride 183–184, 188, 200
74 silica hydride-based 183, 186,
stationary phase dipolarity 272 189, 199
stationary phase materials 82, silica-supported chiral 115
210, 264–265, 268, 272–273, step gradients 60, 62
275, 277, 279, 281 stilbene oxide 136–138
stationary phase morphology 229, supercritical fluid chromatography
376 (SFC) 3, 6, 66, 72, 74–75,
stationary phase packings 2–3 79, 101, 103–104, 106, 108,
stationary phases 111, 115, 117, 119, 125, 128,
aromatic 106 130–131, 148, 163–165
cation-exchange 213
chiral 6, 79, 115, 140
column 95, 358
complement nonionic 264, 295 TCD, see thermal conductivity
conventional 264, 284
detector
conventional nonionic 293
temperature program 324,
gas chromatographic 270–271,
327–328, 349, 351
284, 286, 288
temperature ramp 323–324
high thermal stability 294
test method variability 401, 403,
HILIC 7, 77, 185
405, 407, 409, 411
hydrocarbon 274
thermal conductivity detector
liquid 267, 272, 274, 497
(TCD) 328–329, 358
modern ion chromatography 230
modified silica hydride 187 thermal stability 264, 266, 268,
nonionic 288, 295 270, 272, 274, 276, 279–284,
nonpolar 86–87, 90, 269 286–290, 293–295, 314
octadecylsilica-based 239 thin-layer chromatography (TLC)
PEG 293–294 3, 54–56, 58, 60–63
Pirkle-type chiral 171–172 TLC, see thin-layer chromatography
polar 86, 103, 108, 267, 288 toluene 57–58, 123, 350, 509, 512
polyethylene glycol 264 trifluoroacetic acid 100, 103–104,
polymeric 210, 480–481, 488 109, 126, 200
polymeric ionic liquid-based 293 tubing materials 266–267
542 Index

UHPLC, see ultra high-performance Van Deemter curves/plots

liquid chromatography 136–138, 318, 370–371, 377,
UHPLC methods 368, 371–372, 379
375–377, 379–383, 386, 389, vapor pressure 286, 518
392 VOCs, see volatile organic
UHPLC separations 374, 388, 391, compounds
394 volatile mobile phase adjustment
ultra high-performance liquid 93
chromatography (UHPLC) volatile organic compounds (VOCs)
1–4, 6, 8–10, 13, 91–92, 101, 277, 315
138, 148, 170, 172, 367, 369,
372–373, 377, 379, 382, 392,
394
United States Pharmacopeia (USP) wastewaters 242, 259
69, 85, 382, 392–393, 488 water, drinking 229, 241–242, 259,
USP, see United States 278
Pharmacopeia

zwitterionic phase 110

vaccines 434, 446–447, 468