Technische Universität München

TUM School of Life Sciences

Understanding deep eutectic systems from solid–liquid equilibria

perspective: Experimental study and thermodynamic modeling

Ahmad Alhadid

Vollständiger Abdruck der von der TUM School of Life Sciences der Technischen Universität

München zur Erlangung eines

Doktors der Ingenieurwissenschaften (Dr.-Ing)

genehmigten Dissertation.

Vorsitz:

Prof. Dr. Heiko Briesen

Prüfende:

1. Prof. Dr. Mirjana Minceva

2. Prof. Dr. Jakob Burger

3. Prof. Dr. Piotr Cysewski

Die Dissertation wurde am 02.06.2022 bei der Technischen Universität München eingereicht

und durch die TUM School of Life Sciences am 26.09.2022 angenommen.

Table of Contents
Acknowledgments ..................................................................................................................... 2
Abstract ..................................................................................................................................... 3
Kurzzusammenfassung .............................................................................................................. 4
1. Introduction ........................................................................................................................ 6
2. Theory................................................................................................................................10
2.1 Eutectic mixtures .............................................................................................................10
2.2 Types of phase diagram ..................................................................................................11
2.3 Experimental determination of SLE .................................................................................13
2.3.1 Visual and centrifugation methods ............................................................................13
2.3.2 DSC ..........................................................................................................................14
2.4 Thermodynamics of SLE .................................................................................................21
2.4.1 Melting properties of pure components ....................................................................23
2.4.2 Activity coefficients ...................................................................................................25
2.4.3 Modeling SLE phase diagram ...................................................................................30
3. Results ...............................................................................................................................33
3.1 Paper I .............................................................................................................................34
3.2 Paper II ............................................................................................................................64
3.3 Paper III ...........................................................................................................................80
3.4 Paper IV...........................................................................................................................99
3.5 Paper V..........................................................................................................................118
3.6 Paper VI.........................................................................................................................143
4. Discussion .......................................................................................................................164
4.1 Measuring and modeling SLE ........................................................................................164
4.2 DES formation ...............................................................................................................167
5. Conclusion .......................................................................................................................171
6. Outlook ............................................................................................................................172
Abbreviations ..........................................................................................................................174
Symbols ..................................................................................................................................175
List of figures ..........................................................................................................................177
References ..............................................................................................................................178

1
Acknowledgments
I am sincerely appreciative to Prof. Minceva for her supervision, guidance, time, and
consideration during the work on this thesis. I thank Prof. Burger and Prof. Cysewski for their
time and consideration in evaluating the thesis and Prof. Briesen for accepting the role of the
chair of the examination committee.

Words cannot express my gratitude to Dr. Mokrushina, the mastermind behind the proposed
question of the thesis. A sincere thank goes to Dr. Jandl for the great and fruitful collaboration
over the last two years of the thesis.

I came to know great people during my stay at the Biothermodynamics group, Anja, Heinrich,
Yasemin, Caren, Franzi, Simon, Martin, Raena, Andy, Rana, Lukas, Ana Maria, Kamila, Marko,
Vlad, Melanie, Vanessa, Jan, Benedikt, Daili, Pei, and Sahar. I thank you all for the great time we
had, and I’m looking forward to continuing my next years in the group. I’m grateful for the
opportunity to work with and supervise many awesome and hardworking students, Lea, Irsa,
Deniz, Joel, Maeliss, and Christoforos. Many thanks to Thomas, Sharad, and Ali for sharing their
wisdom and expertise in academic life.

I dedicate this thesis to my parents, my dearests and support, to my only love and best friend,
Manal, and to my joy givers, Nawal and Yusuf. I would like to thank my brother, sisters, and
family for their love and support. Lastly, I would be remiss in not mentioning our dear friends
Mohammad and Meli for the great moments even during the difficult CORONA times.

2
Abstract
Deep eutectic solvents (DES) are a class of designer solvents that can be prepared by mixing
two or more compounds to form a solution with a freezing temperature significantly lower than
that of pure constituents. Because the DES physicochemical properties correlate with those of
pure constituents, the solvent properties can be tuned by selecting the constituents and their
molar ratio. Having a liquid solution at the desired operating temperature is of utmost importance
to employing DESs as solvents in various applications. The melting temperature of the mixture
at any molar ratio can be obtained from the solid–liquid equilibria (SLE) phase diagram. However,
as DESs can be prepared from a large pool of substances, obtaining the SLE phase diagram
experimentally becomes a demanding task that often accompanies many difficulties. Therefore,
understanding DES formation is a vital goal to preselect the constituents and define the suitable
molar ratio range to tune the solvent properties accordingly.

This thesis aims to understand the reason for the formation of eutectic mixtures with a significant
depression of the melting temperature at the eutectic point from a SLE perspective. To achieve
this, SLE in binary eutectic mixtures was studied based on experimental and theoretical
investigations. The SLE phase diagram of 11 eutectic systems of various complexity was
measured and modeled. Moreover, seven cocrystals not reported before were identified in four
different eutectic systems. The SLE phase diagrams reported in this thesis were obtained by
combining rigorous thermal analysis using differential scanning calorimetry and solid phase
characterization using X-ray diffraction, where the latter is often lacking in literature studies
leading to incorrect interpretation of SLE phase diagrams. Furthermore, the capability of
correlative and predictive activity coefficient models, namely, the non-random two-liquid (NRTL)
model and conductor-like screening model for realistic solvation (COSMO-RS), in describing the
nonideality in strongly nonideal eutectic systems with complex phase diagrams was evaluated.

Based on the thesis results, DESs were found to form as a result of significant negative deviation
from ideal behavior and very low melting enthalpy and entropy of constituents. The very low
melting entropy is a unique character of substances with highly disordered and symmetrical
crystal structures, i.e., plastic crystalline materials. Therefore, the selection of DES constituents
is narrowed down from arbitrary salts and organic substances to plastic crystalline materials.
The thesis findings will aid in selecting DES constituents to improve the design approaches and
the utilization of DES in various applications.

3
Kurzzusammenfassung
Tief eutektische Lösungsmittel (DES) stellen eine Klasse von Designerlösungsmitteln dar, die
durch Mischen von zwei oder mehr Verbindungen hergestellt werden können. Dadurch entsteht
eine Lösung, die eine Gefriertemperatur aufweist, welche deutlich unter der der reinen
Bestandteile liegt. Da die physikochemischen Eigenschaften von DES mit denen der reinen
Bestandteile zusammenhängen, können die Lösungsmitteleigenschaften durch Auswahl der
entsprechenden Bestandteile und Molverhältnisse eingestellt werden. Um bei gewünschter
Betriebstemperatur eine flüssige Lösung vorliegen zu haben, ist der Einsatz von DES als
Lösungsmittel in verschiedenen Anwendungen von größter Bedeutung. Die Schmelztemperatur
des Gemischs bei einem beliebigen Molverhältnis lässt sich anhand des Phasendiagramms des
Fest-Flüssig-Gleichgewichts (SLE) ablesen. Da DESs jedoch aus einer Vielzahl an Substanzen
hergestellt werden können, ist die experimentelle Vorgehensweise zur Ermittlung des SLE-
Phasendiagramms eine anspruchsvolle Aufgabe, die oft mit vielen Schwierigkeiten verbunden
ist. Daher ist ein Verständnis der Bildung von DES für die Vorauswahl der Bestandteile und die
Festlegung des geeigneten Molverhältnisses von entscheidender Bedeutung, um die
Lösungsmitteleigenschaften entsprechend einstellen zu können.

Ziel dieser Arbeit ist es, die Ursache für die Bildung eutektischer Gemische mit signifikant
niedriger eutektischer Temperatur aus der Perspektive der SLE zu verstehen. Zu diesem Zweck
wurden die SLE in binären eutektischen Mischungen auf der Grundlage von experimentellen und
theoretischen Untersuchungen betrachtet. Das SLE-Phasendiagramm von 11 eutektischen
Systemen mit unterschiedlicher Komplexität wurde erfasst und modelliert. Darüber hinaus
wurden in vier verschiedenen eutektischen Systemen sieben bisher nicht angegebene Kokristalle
identifiziert. Die in dieser Arbeit dargestellten SLE-Phasendiagramme konnten anhand einer
Kombination einer strengen thermischer Analyse und einer Festphasencharakterisierung
erhalten werden, wobei letztere in Literaturstudien oft fehlt, was zu einer falschen Interpretation
der SLE-Phasendiagramme führt. Darüber hinaus wurde die Fähigkeit von korrelativen und
prädiktiven thermodynamischen Modellen zur Beschreibung der Nichtidealität in stark nicht-
idealen eutektischen Systemen mit komplexen Phasendiagrammen bewertet.

Ausgehend von den Ergebnissen der Dissertation konnte festgestellt werden, dass die Bildung
von DES das Ergebnis einer signifikanten negativen Abweichung vom idealen Verhalten und einer
sehr niedrigen Schmelzenthalpie und -entropie der Bestandteile ist. Die sehr niedrige

4
Schmelzentropie ist ein Alleiniges Merkmal von Stoffen mit einer hochgradig ungeordneten und
symmetrischen Kristallstruktur, d. h. von plastischen kristallinen Materialien. Daher wird die
Auswahl der DES-Bestandteile von beliebigen Salzen und organischen Substanzen auf
plastisch-kristalline Materialien begegrenzt. Die Ergebnisse dieser Arbeit werden bei der Auswahl
von DES-Bestandteilen helfen, um die Designansätze und die Nutzung von DES in
verschiedenen Anwendungen zu verbessern.

5
Introduction

1. Introduction
The establishment of green and sustainable chemistry concepts in process applications calls for
the utilization of green solvents.1 Green solvents should not only comply with the principles of
green chemistry but also be potential alternatives to conventional solvents while being
economically feasible.2

3-6
One of the first classes of solvents that has been proposed to be green is Ionic liquids (ILs).
ILs are salts with relatively low melting temperature–compared to conventional salts–that contain
7
organic cations and organic or inorganic anions. Being salts, ILs possess negligible vapor
8
pressure and are nonflammable. The variety of combinations of cations and anions that can
9
form ILs suggests that the properties of ILs can be tailored for a specific task. Although ILs
fulfill some of the requirements for green solvents, they bring some concerns. First, each IL
should be chemically synthesized, necessitating the use of conventional organic solvents and
producing wastes that may be potentially hazardous. 10 Second, from an economic point of view,
ILs are still significantly more expensive than conventional solvents, which is undoubtedly the
main drawback of using ILs in industrial-scale applications. 11

12
In 2001, Abbott et al. performed the first trials to liquefy quaternary ammonium halide salts
with the formula [Me3NC2H4Y]Cl (Y = OH, Cl, OC(O)Me, OC(O)Ph) by mixing them with zinc or tin
chlorides. Out of several quaternary ammonium halide salts investigated in the study, mixtures
containing choline chloride (ChCl) had the lowest melting temperature. These mixtures were
called deep eutectics due to the large depression in the melting temperature of the mixture at
the eutectic point relative to pure constituents’ melting temperature. The significant depression
in the melting temperature was attributed to the formation of complex zinc or tin chloride ions,
e.g., [ZnCl3]–, [Zn2Cl5]–, and [Zn3Cl7]–. The deep eutectic mixtures resemble ILs in terms of
properties however overcome some of their drawbacks, e.g., preparation by chemical synthesis
and feasibility.

The search for other so-called “complexing agents” continued by the same group. Abbott et al.13
showed in 2003 that urea could replace zinc and tin chlorides to form deep eutectics. Further
studies showed that various carboxylic acids could also be used, suggesting that any hydrogen
14
bond donor (HBD) can form deep eutectics when mixed with ChCl. Later, the term deep
eutectic solvents (DES) was introduced and defined as a mixture of a hydrogen bond acceptor

6
Introduction

(HBA) and HBD with a eutectic temperature significantly lower than the melting temperature of
pure constituents. 15

Soon after, the number of studies related to understanding the formation of DES and the search
16-28
for new and improved DESs increased sharply. The main hypothesis of these studies was
that hydrogen bonding and complex formation are the reason for DES formation. Moreover,
DESs have been postulated as a special type of liquids that are microscopically heterogeneous
29
and complex. Therefore, investigations relied on quantum mechanical, molecular simulation,
and experimental methods to unravel the microscopic nature of DES and understand their
16, 17, 20, 23-25, 30-46
formation. However, it was impossible to quantify the eutectic ratio and
temperature from theoretical and experimental microscopic studies. 18

Later, after a series of papers investigating the solid–liquid equilibria (SLE) in DES, Coutinho and
47
coworkers proposed another definition for DES. First, they defined DES as physical mixtures
from a thermodynamic point of view. Second, they noticed that the formation of DES could be
due to strong hydrogen bonding and negative deviation of the system from ideal solution
behavior. Third, DES is a term that describes the mixture at any composition and is not specific
to the mixture at the eutectic composition, which probably might not be of a specific ratio, such
as 1:1 or 1:2 molar ratios. Accordingly, Martins et al.47 suggested that only eutectic mixtures with
negative deviation from ideal behavior should be called DES.

The main feature of the proposed definition of Coutinho and coworkers, which was also adopted
48, 49
by other groups, is the rationalization of the understanding and designing of DES, i.e., the
selection of the constituents and their ratio. SLE phase diagram provides information on the
melting temperature of the mixture at any ratio between constituents–including the eutectic
point–as well as on the intermolecular interactions between unlike molecules. However, due to
the large number of substances that can form eutectic mixtures, it is impossible to measure the
phase diagram of all possible DES. Therefore, the number of DES with available SLE data is
significantly lower than the number of DES studied in the literature.

Although the study of DES from the SLE perspective can provide direct interpretation for the
melting temperature of the mixture, SLE studies are difficult to be comprehended and executed
by nonexperts in thermodynamics. Consequently, the term DES is usually used arbitrarily for any
eutectic mixtures forming a liquid at room temperature, though in some, one or both constituents
are liquid at or near room temperature. Out of many DES constituents proposed in the literature,

7
Introduction

ChCl and a few other quaternary ammonium halide salts are the only constituents that can form
eutectic mixtures with a eutectic temperature significantly below the melting temperature of pure
constituents. This might hint that the observed significant depression in the mixture melting
temperature at the eutectic point in ChCl-based DES is due to the unique character of ChCl.
Hence, it might be possible to obtain a DES by an appropriate selection of constituents without
the need for SLE or microscopic structure studies.

Regardless of their definition or nature, DES has been proved to outperform conventional
solvents in various applications. DESs have been shown to efficiently extract bioactive
50-56
compounds from various natural sources. From a process application perspective, DESs
57-61 62-67
have showen great potential as solvents for gas capture, liquid chromatography, and
extractive distillation. 68-71 Besides using DES in their liquid states, the depression in the melting
temperature of the DES has been employed to improve the solubility of active pharmaceutical
compounds and modify their crystal structure. 72-79 Because a large pool of substances can form
DESs, they can be utilized in a vast number of applications, of which only a part has been
investigated so far.

Although the physicochemical properties of eutectic mixtures may correlate with intermolecular
interactions between molecules in the liquid solution, the reason why the eutectic mixtures could
replace conventional solvents in various applications does not necessarily originate from a
specific microscopic structure or strong intermolecular interactions in the liquid solution. Instead,
understanding the physicochemical properties of eutectic mixtures and correlating them with the
pure constituents’ physicochemical properties is required to form eutectic mixtures with tunable
properties as alternative solvents for a specific application. However, the superiority of DESs
over common eutectic mixtures is the large depression in the mixture melting temperature, which
provides a wider selection of systems that can be liquid at or near room temperature to be used
as solvents.

This thesis aims to investigate the reason for the formation of eutectic mixtures with significantly
low eutectic temperatures from a SLE perspective. A comprehensive and systematic
experimental and theoretical study of SLE in various eutectic mixtures was performed to
investigate why a DES can be formed. First, the proper experimental determination and modeling
of SLE were carried out to obtain SLE data of eutectic mixtures of various complexity. Second,
the parameters influencing the phase diagram and the position of the eutectic point of eutectic

8
Introduction

mixtures, namely, melting properties, nonideality of the liquid phase, and solid complex
formation, were examined.

The thesis is organized into six sections. In Section 2, the pertinent theory is introduced, covering
the basic definitions for eutectic mixtures, thermal characterization, and thermodynamic
modeling. The results are presented in the form of six published papers in Section 3 along with
a summary of each paper. A thorough discussion of the papers findings and comprehensive
analysis of the thesis topic is given in Section 4. Section 5 contains the general conclusions. In
Section 6, the outlook for the future work is presented.

9
Theory

2. Theory
The current section covers the theory and the methods used to obtain the findings of the thesis.
First, an overview of eutectic mixtures is presented in Section 2.1. Second, a brief description of
the SLE phase diagram of different complexity is covered in Section 2.2. Third, the thermal
characterization techniques used to measure SLE are presented in Section 2.3, including a
throughout discussion about differential scanning calorimetry (DSC) analysis. Fourth, models
and approaches used to model SLE in eutectic mixtures of different complexity are presented in
Section 2.4.

2.1 Eutectic mixtures

Eutectic mixtures are mixtures of two or more components that crystallize as immiscible or
80
partially immiscible solids. Figure 1 shows a schematic representation of the SLE phase
diagram of a binary mixture in which the components crystallize as a miscible solid phase (Figure
1A) or immiscible solid phases (Figure 1B). The liquidus lines (blue lines) represent the
temperature at which the last solid disappears upon heating, i.e., the melting temperature of the
mixture. The solidus lines (orange lines) depict the temperature at which the first drop of the
liquid appears upon heating. The liquidus and solidus lines define the two-phase region in which
a liquid phase and a solid phase are in equilibrium. As seen in Figure 1A, the liquidus temperature
of the mixture at any composition lies between the melting temperature of the pure components.
In contrast, as seen in Figure 1B, the melting temperature of the eutectic mixture can be lower
than the melting temperature of the pure components. The lowest possible melting temperature
of the eutectic mixture is observed at the eutectic point (marked with a red circle in Figure 1B).
At the eutectic point, three phases are in equilibrium, solid A, solid B, and a liquid solution,
according to the following reaction

𝐴 𝐵 ↔𝐿 (1)

The reaction in Equation (1) is called the eutectic reaction, from which the name eutectic system
was derived (means easily melted in ancient Greek). 80

10
Theory

(A) (B)
Tm,A
Tm,A

L
T/K

L+S

T/K
Tm,B
L
S SA + L
L + SB
Tm,B SA + SB
A B A B
Figure 1. Schematic representation of solid–liquid phase diagram of (A) binary mixture with miscibility
in the solid state and (B) binary eutectic mixture.

2.2 Types of phase diagram

As described in Section 2.1, the melting temperature of the eutectic mixture at any composition,
i.e., liquidus temperature, can be obtained from its SLE phase diagram. In the following, the
possible shapes of the binary eutectic mixture phase diagram are presented. The complexity of
measuring and modeling SLE is commonly attributed to the complexity of the solid rather than
81
the liquid phase. Figure 2 shows the probable types of the SLE phase diagram of eutectic
mixtures. Figure 2A shows the simplest case possible for any binary eutectic mixture. Systems
having this type of SLE phase diagram are called simple eutectic systems. In the simple eutectic
type, the constituents crystallize as pure solids. Moreover, the system has only one eutectic
point. The eutectic point position depends on the course of the two liquidus lines of pure
constituents.

The phase diagram shown in Figure 2B is also of the simple eutectic type. However, pure
component A undergoes a solid–solid transition at 𝑇 →
. The melting properties of solid SA and

solid S'A may be different, which leads to a different course of the liquidus line of component A
at temperatures higher and lower than the solid–solid transition temperature.

11
Theory

(A) (B)
Tm,A Tm,A

→
S'A + L L
T/K

Tm,B Tm,B

T/K
SA + L
L
SA + L
SA + SB
L + SB

SA + SB
A B A B

(C) (D)
Tm,A Tm,A

L Tm,B
L Tm,
T/K
T/K

B SA + L
L + SB
α α+ L L+β
β
SA + Co Co + SB
α+β

A xα xβ B A Co B

(E)

Tm,A

SA + L
T/K

L Tm,B
Co + L
SA + Co
Co + SB

A Co B

Figure 2. Schematic representations of solid–liquid phase diagrams of different types, defining the
crystallized solid phases in each binary eutectic system.

Figure 2C shows the schematic representation for a binary eutectic mixture with partial miscibility
in the solid phase. In the example shown in Figure 2C, the eutectic system has two solid solution
regions. The α solid phase has the same crystal structure as pure component A. However, some
of the molecules of component A are replaced by component B molecules in the α solid phase
while preserving the same crystal structure of component A. The solubility of component B in

12
Theory

solid component A is temperature-dependent and has a maximum limit of xα. On the other hand,
component A is also soluble in the solid component B forming the β solid phase. Similarly, the
solubility limit of component A in solid component B is xβ. The formation of solid solution regions
is excepted when the constituents have very similar molecular and crystal structures. Thus, solid
solution regions in organic binary eutectic mixtures are quite uncommon. 82

The phase diagrams presented in Figures 2D and 2E depict the formation of solid compounds,
i.e., a cocrystal between components A and B. The cocrystal formed between components A
and B in the phase diagram shown in Figure 2D is a congruently melting cocrystal, i.e., a
cocrystal that melts to a binary liquid solution with the exact stoichiometry as that of the
cocrystal. In contrast, the cocrystal formed between components A and B in the phase diagram
shown in Figure 2E is an incongruently melting cocrystal, i.e., a cocrystal that melts to a solid A
and a liquid solution with a different stoichiometry than that of the cocrystal. As seen in Figure
2D, the eutectic system with a congruently melting cocrystal has two eutectic points. Principally,
the number of eutectic points for a system with congruently melting cocrystals equals the
number of cocrystals plus one. On the other hand, the number of incongruently melting cocrystal
does not increase the number of eutectic points for the system.

It is worth mentioning that combining any of the previously discussed cases is possible. For
example, a eutectic system containing a component that undergoes a solid-solid transition might
show the formation of several congruently and incongruently melting cocrystals in addition to
solid solution regions. The complexity of measuring and modeling SLE in such eutectic systems
is discussed in the following sections.

2.3 Experimental determination of SLE

2.3.1 Visual and centrifugation methods
Thermal analysis methods can be used to obtain the SLE data of binary eutectic mixtures by
determining the melting temperature of the mixture at various compositions. Several thermal
analysis methods have been proposed in the literature to determine the SLE data of binary
83, 84
eutectic mixtures. The melting temperature of the mixture can be determined by visual
methods, such as a melting point apparatus 85-90 or microscopy.91 The advantage of using visual
methods is that the interpretation of the melting temperature of the mixture is straightforward.
However, several important phase transitions, such as solid–solid transition or solidus

13
Theory

temperature, cannot be observed visually. Hence, visual methods might be only suitable for
eutectic mixtures of the simple eutectic type.

84
Van der Bruinhorst et al. proposed a more sophisticated method to obtain the solid–liquid
phase diagram of eutectic mixtures. The proposed method depends on equilibrating the eutectic
mixture of a certain composition at a temperature in the two-phase region. The liquid and solid
phases in equilibrium at a specific composition are separated by centrifugation, and the liquid
phase composition is analyzed to determine the liquidus line. The method is advantageous for
eutectic mixtures prone to supercooling and glass formation. However, homogenization of the
liquid phase is critical before its analysis, which can complicate the sampling procedure.
Moreover, the method requires an additional analytical method to determine the composition of
the liquid phase. Thus, in contrast to visual methods, the determination of the liquidus line is not
straightforward. On the other hand, analogous to visual methods, no information about the
solidus temperature of the eutectic mixture can be obtained. Thus, the centrifugation method
could be only used to determine the SLE in eutectic mixtures of the simple eutectic type.
Nevertheless, the SLE phase diagram of the eutectic system might not be of the simple eutectic
type, and hence, a more comprehensive method should be employed. In the following section,
the DSC analysis technique used to determine the experimental SLE data for eutectic systems
of various complexity is discussed in detail.

2.3.2 DSC
DSC is a robust thermal analysis method that can be used to obtain the phase diagram of
83
eutectic mixtures. Figure 3 shows a schematic representation of the DSC analysis principle.
Two crucibles, one is empty, and the other contains the sample, are placed inside a furnace and
subjected to cooling or heating. The temperature of the two crucibles is measured. When the
sample undergoes a phase transition, such as melting, crystallization, or solid–solid transition,
heat is absorbed by the sample as latent heat, so the temperature of the sample crucible remains
constant. The DSC signal represents a measure of the difference between the temperature of
the two crucibles, which is proportional to the heat flow. DSC provides information about any
phase transition occurring in the mixture, such as melting and solid–solid transition
temperatures. In addition to the transition temperature, DSC analysis can be used to determine
the phase transition enthalpy.

14
Theory

Furnace/cooling device

Reference/empty crucible Sample crucible

Temperature sensor

T T

t t

Figure 3. Schematic structure of differential scanning calorimetry chamber.

2.3.2.1 Interpretation of DSC curves

In the following, the development of DSC curves of the samples covering the entire composition
range of a hypothetical binary eutectic mixture with complex character is discussed. Figure 4
shows a schematic representation of the phase diagram of a hypothetical binary eutectic mixture
with the formation of various cocrystals and the observed DSC curve at each labeled point. The
DSC curve of point a represents the melting temperature of pure component A. The peak
corresponding to the melting of a pure substance is usually sharp and symmetrical. By increasing
the amount of component B in the mixture, two peaks are observed in the DSC curve at point b.
The first peak is sharp and symmetrical and corresponds to the decomposition temperature of
the incongruently melting cocrystal, while the second peak is asymmetrical and corresponds to
the liquidus temperature of the mixture at the composition of point b. At point c, the observed
DSC curve shows three distinct peaks related to the eutectic melting, incongruent melting, and
liquidus temperature. At the eutectic point (point d), only one sharp and symmetrical peak
appears in the DSC curve. The DSC curve of point e corresponding to congruently melting
cocrystal resembles the DSC curves observed for component A (point a) and eutectic melting
(point e). As seen in Figure 4, component B undergoes a solid–solid transition. Accordingly, the
solid–solid transition of component B is observed in the DSC curve at point f as a small peak
followed by a liquidus peak. Eventually, the DSC curve of component B shows two peaks
respective to a solid–solid transition and melting.

15
Theory

endo
a endo
b endo
c

Tm,A a
b L gT
m,B
f
SA + L c e
endo
d endo
e endo
f
S'B + L
T/K

d
SA + Co14

Co11 + Co14 SB + Co11

endo
g

A Co14 Co11 B

Figure 4. Schematic representation of a phase diagram of a hypothetical binary eutectic mixture and
the observed differential scanning calorimetry curve at each labeled point.

Despite the ability of DSC analysis to determine any phase transition temperature and enthalpy,
the interpretation of the DSC signal is rather not straightforward. As seen in Figure 4, eutectic,
incongruently, and congruently melting peaks are analogous in shape, which can be confusing
to assign the peak to the corresponding phase transition. Thus, additional analysis methods,
such as X-ray diffraction (XRD), are required to comprehend the crystallized solid phases in each
composition region and, accordingly, assign the peaks observed in the DSC curve of the
samples to the proper transition.

Figure 5 shows DSC curves of a pure component and a sample in a eutectic system. The phase
transition enthalpy hm is determined by the area of the corresponding peak. By definition, the
solidus temperature is the temperature at which the first drop of liquid appears (See Section 2.1).
Thus, the onset temperature of the corresponding peak well represents the eutectic, incongruent
and congruent melting (Tm and Tsolidus in Figure 5). On the other hand, the liquidus temperature is
the temperature at which the last solid disappears. Thus, the peak maximum temperature is
commonly used for determining the liquidus temperature (see Figure 5B). 83

16
Theory

(A) (B)
endo endo

2 3

+
+
1 4
liquid
𝑇
+
+
+ 𝑇 𝑇

Figure 5. Determining the melting, solidus, and liquidus temperatures and phase transition enthalpy from
a differential scanning calorimetry curve.

Tammann’s plot can be used to determine the eutectic composition, solubility limit in the solid
phase, and the cocrystal stoichiometry experimentally from the DSC analysis. Tammann’s
diagram is constructed by plotting the eutectic phase transition enthalpy (the area of the solidus
peak) as a function of the mole fraction of one of the components. 83 Figure 6 shows a schematic
representation of the SLE phase diagram of the simple eutectic type (Figure 6A), a eutectic
system with partial miscibility in the solid phase (Figure 6B), and a eutectic system with a
cocrystal formation (Figure 6C) along with the corresponding Tammann’s plot obtained from the
area of the solidus peak. As shown in Figure 6, the eutectic transition enthalpy depends linearly
on the mole fraction of the component. For eutectic systems of the simple eutectic type (Figure
6A), the eutectic point can be obtained from the intercept of the two lines regressed to eutectic
transition enthalpy in the hypoeutectic (composition range on the left side of the eutectic point)
and hypereutectic (composition range on the right side of the eutectic point) regions. For a
eutectic system with partial miscibility in the solid phase (Figure 6B), the lines regressed to the
eutectic transition enthalpy in the hypoeutectic and hypereutectic regions intercept the x-axis at
the solubility limits in the solid phase of component A in B and vice versa. Thus, Tammann’s plot
does not only provide the solubility limit but can also provide information about the phase
diagram type. In the case of cocrystal formation (Figure 6C), the stoichiometry of the cocrystals
and the eutectic points of the system can be determined from Tammann’s plot. Despite the

17
Theory

valuable information that can be obtained from Tammann’s plot, the reliability of the obtained
information depends largely on the quality of the determined phase transition enthalpy. In many
cases, the peaks corresponding to different phase transitions might overlap, limiting the
possibility of constructing Tammann’s plot for the eutectic system. 83, 92

(A) (B) (C)

L
L L
T/K

T/K
T/K
SA + L SB + L +L +L SA + L SB + L
SA + SB +
SA + Co SB + Co

B xα xet xβ xet,1 xco xet,2

A xet A B A B
h / J g−1

h / J g−1
h / J g−1

xet x xet x xet,1 xco xet,2

0 1 0 1 0 1

Figure 6. Schematic representation of solid–liquid phase diagrams of different types and the obtained
Tammann’s plot for each type.

2.3.2.2 DSC analysis protocol

The DSC analysis parameters are the heating and cooling rates, sample mass, and measurement
temperature range. The heating and cooling rates used in DSC analysis can affect the shape of
the solidus and liquidus peaks. Slower heating rates eliminate the effect of heat conductivity of
the sample and provide more time for the sample to reach equilibrium. However, slower heating
rates affect the sensitivity of the DSC analysis, which might leave out the detection of some
transitions, especially those with low transition enthalpy, such as liquidus temperatures near the
eutectic point. 93 The commonly used heating rates are between 1–10 K min–1. Cooling rates are
usually selected based on the capacity of the cooling device as the transition temperature and

18
Theory

enthalpy are determined from the heating runs only. However, cooling rates can be critical for
94
substances existing in different polymorphs. Fast cooling might lead to supercooling, glass
formation, and crystallization of metastable solid phases. 73, 95

The DSC instrument is calibrated using standard calibration substances with known phase
transition temperature and enthalpy before measurements. The transition enthalpy is used to
calibrate the instrument sensitivity, i.e., the area of the observed peak in the DSC curve of the
standard material. The onset temperature of the transition peak of the standard materials is used
for temperature calibration. The calibration procedure is performed at the same heating rate
planned for the measurement. Consequently, pure components, eutectic, incongruently, and
congruently melting as well as transition enthalpies are less sensitive to the heating rate, as the
heating rate is considered in the calibration procedure. In contrast, the heating rate can affect
the measured liquidus temperature. Van den Bruinhorst et al. 96 analyzed the effect of the heating
rate on the determined liquidus temperatures and concluded that the inflection point is more
suitable for determining the liquidus temperature than the peak maximum temperature.
Alternatively, the effect of the heating rate on the determined liquidus temperature can be ruled
out by performing the DSC analysis at different heating rates and determining the liquidus
97, 98
temperature as the peak maximum temperature extrapolated to zero heating rate.
Nevertheless, the influence of the heating rate on the determined liquidus temperature might be
within the uncertainty of the DSC data. Hence, selecting a convenient heating rate, such as 2 or
5 K min–1, can be sufficient to determine the SLE phase diagram of most eutectic systems
properly.

93
Saeed et al. showed that the influence of the sample mass is more pronounced on the
transition enthalpy than on onset or peak maximum temperatures. Therefore, for SLE
measurements, the effects of the sample mass might not be critical. The temperature range
should be defined to detect all possible phase transitions occurring in the sample. Hence, the
temperature range can be specifically tuned for each sample to ensure time efficiency and
convenience only during the experimental analysis.

Figure 7 shows a schematic representation of DSC curves of a sample analyzed by DSC using
the typical DSC protocol. The typical DSC analysis protocol involves heating the sample to 10 K
above melting temperature, an isothermal run at the final temperature for sufficient time to ensure
sample homogeneity, and then a cooling run around 50 K below the crystallization temperature

19
Theory

of the sample. 99 The sample is then subjected to second heating run to 10 K above the melting
temperature of the sample. The transition temperature and enthalpy are determined from the
second run or further heating/cooling cycles. 96-98 This protocol is used to clear the sample history
and ensure the homogeneity of the sample. However, in situ crystallization in the DSC (Figure
73, 95, 100
7A) could lead to metastable polymorphs, metastable solid phases, or glass formation.
94
Corvis et al. overcame the formation of metastable polymorphs in the lidocaine/L-menthol
eutectic system by annealing the DSC samples for several months prior to measurements.
Moreover, supercooling could occur during the cooling run, and the sample is crystallized during
99, 101
the heating run instead, i.e., cold crystallization. Cold crystallization in the vicinity of the
solidus temperature (Figure 7B) can limit the proper determination of the onset temperature of
91
the corresponding peak. In conclusion, the typical DSC protocol can be unsuitable for
determining the SLE in eutectic systems showing kinetic limitations in crystallization or the
formation of different polymorphs. Therefore, in such a case, methods to aid crystallization and
overcome glass formation in eutectic systems should be sought to determine the SLE phase
diagram.

(A) (B)
endo endo

second heating run second heating run

cold crystallization
cooling run
cooling run

in situ crystallization

first heating run first heating run

T/K T/K
Figure 7. Schematic representation of a differential scanning calorimetry (DSC) curve of a sample
analyzed using the typical DSC protocol: first heating, cooling, and second heating runs. In (A), the
sample in situ crystallizes in the DSC during the cooling run, while in (B), the sample undergoes cold
crystallization, i.e., crystallization during the heating run.

20
Theory

2.4 Thermodynamics of SLE

As seen in the previous section, experimental determination of SLE is a tedious task, especially
for eutectic systems with complex phase diagrams. Thus, the standard approach to obtaining
the SLE phase diagram is to measure the SLE data of the mixture at selected compositions,
usually with a mole fraction step of 0.1. The experimental SLE data are then correlated using a
thermodynamic model to obtain the phase diagram over the whole composition range.

The equilibrium condition for component 𝑖 between the solid (S superscript) and liquid (L
superscript) phases is as follows 102

𝑓 𝑓 (2)

𝑓 𝑓 (3)

where 𝑓 and is the fugacity, 𝑓 is the standard state fugacity, is the mole fraction, and is
the activity coefficient of component 𝑖. The standard states are defined as the pure solid and
pure liquid at the liquidus temperature. The ratio between the standard state fugacity of
component 𝑖 in the liquid and solid phases can be calculated from the thermodynamic cycle
shown in Figure 8. First, the pure solid phase is heated from the liquidus temperature T to its
melting temperature Tm. Second, the pure solid is melted at Tm. Third, the pure liquid is cooled
to the liquidus temperature T. The Gibbs energy change from point 1 to 4 ( 𝑔 → ) is defined as
follows

𝑓
𝑔 → 𝑅𝑇 ln → 𝑇 →
(4)
𝑓

where 𝑅 is the gas constant, 𝑇 is the liquidus temperature, → is the enthalpy change, and

→ is the entropy change.

21
Theory

Tm 2 3

T 1 4
solid liquid
Figure 8. Thermodynamic cycle used to derive the solid–liquid equilibrium equation.

The enthalpy and entropy changes from point 1 to 4 is defined as the sum of the enthalpy and
entropy changes along the path as follows

1→4 1→2 2→3 3→4

(5)

𝑠1→4 𝑠1→2 𝑠2→3 𝑠3→4 (6)

The enthalpy change from point 1 to 4 is the summation of the enthalpy change upon heating
the pure solid from T to Tm, the melting enthalpy , of pure component 𝑖, and the enthalpy
change upon cooling pure liquid from Tm to T. In equation form

𝑇 ,𝑖 𝑇
𝑆 𝐿 (7)
1→4 𝑐𝑃, 𝑑𝑇 , 𝑐𝑃, 𝑑𝑇
𝑇 𝑇 ,𝑖

where 𝑐𝑃𝑆 and 𝑐𝑃𝐿 are the constant pressure heat capacity of pure component 𝑖 in the solid and
liquid states. Similarly, the change in entropy from point 1 to 4 is

𝑇 ,𝑖 𝑐 𝑆 𝑇 𝐿
𝑐𝑃,
𝑃, (8)
𝑠1→4 𝑑𝑇 𝑠 , 𝑑𝑇
𝑇 𝑇 𝑇 ,𝑖
𝑇

where 𝑠 , is the melting entropy of component 𝑖. Substituting equations (7) and (8) in (3) and
(4) yields

𝑥𝐿𝛾𝐿 𝑓0𝐿 , 𝑇 1 𝑇
1 𝑇
𝑐𝑃 , 𝑖
ln ln 1 𝑐𝑃, 𝑑𝑇 𝑑𝑇 (9)
𝑥 𝑆𝛾 𝑆 𝑓0𝑆 𝑅𝑇 𝑇 , 𝑅𝑇 𝑇 ,𝑖 𝑅 𝑇 ,𝑖 𝑇

22
Theory

where 𝑐 𝑐 𝑐 . As seen, the ratio between standard state fugacities is determined by the
melting properties of pure component 𝑖. Therefore, the activity coefficients of components in the
liquid and solid phases are interrelated with the pure components melting properties. As seen in
Equation (9), activity coefficients of the component in the liquid and solid phases and its pure
melting properties–melting enthalpy, temperature, and the heat capacity of pure solid and liquid
states–are the prerequisites for modeling SLE in eutectic systems. The following sections
discuss the melting properties and the activity coefficients as well as the modeling procedure for
SLE phase diagrams of different patterns (Figure 2).

2.4.1 Melting properties of pure components

The right-hand side of Equation (9) can be divided into terms, the first melting enthalpy term and
the last two heat capacity terms. The melting enthalpy, entropy, and temperature of a pure
component are interrelated by the following simple equation

𝑇 (10)

Based on Equation (10), a high melting temperature of a substance results from high melting
enthalpy or low melting entropy. The melting enthalpy of a pure substance is a function of the
103
intermolecular interactions between molecules in the crystal lattice. The intermolecular
interactions depend on the type of the interacting groups, i.e., ionic or hydrogen bonding, and
the distance between interacting groups. The latter is influenced by the lattice parameters and
104
crystal packing, i.e., crystal structure. High melting enthalpy values indicate strong
intermolecular interactions and efficient crystal packing. In the case of unavailable melting
enthalpy, the melting enthalpy of pure substances can be predicted using group contribution
103-107
(GC) methods. However, GC methods predictions for melting enthalpy are sometimes
unreliable. Moreover, GC methods fail to differentiate between isomers, which can possess
108
significantly different melting enthalpies. The unsuitability of GC methods in predicting the
melting enthalpy of pure substances is a direct consequence of the influence of the crystal
structure on the melting enthalpy.

In contrast to melting enthalpy, melting entropy is better correlated to the molecular structure.
109
Yalkowsky, 110 Dannenfelser and Yalkowsky,111 and Jain et al. 112 proposed non-GC methods
to estimate the melting entropy of pure substances. The methods are based on calculating the

23
Theory

melting entropy from the rotational , conformational , and transitional

entropies as follows 104

W (11)

where W is Walden's rule constant which is equal to around 57 J mol–1 K–1. 113
The rotational
entropy is estimated from the rotational symmetry number as follows

R ln (12)

The rotational symmetry number is the number of similar structures that the molecule can
produce when rotated. The conformational entropy is calculated using the flexibility number Φ
as follows 103

R ln Φ (13)

Φ 2.435 (14)

𝑆𝑃 0.5𝑆𝑃 0.5𝑅𝑖𝑛𝑔 1 (15)

where 𝑆𝑃 , 𝑆𝑃 , and 𝑅𝑖𝑛𝑔 are the numbers of non-ring SP3 atoms, non-ring SP2 atoms, and
single, fused, or conjugated ring systems, respectively. The transitional entropy is calculated
104, 105, 114
from the eccentricity as follows

R ln (16)

where is the number of atoms in aromatic and aliphatic rings. Despite the simplicity of these
methods, good predictions have been obtained in many cases. 104

Next, the two heat capacity difference terms are discussed. Several approaches exist to account
96, 115-117
for the heat capacity difference. The simplest case is to neglect the two heat capacity
terms on the right-hand side of Equation (9). 115, 118 This approach is usually misunderstood to be
based on neglecting the value of 𝑐 . However, the approach is based on the fact that the two
heat capacity terms in the right-hand side of Equation (9) have opposite signs, which tend to
102, 118
cancel each other in case T is not far from Tm. The heat capacity terms are commonly
neglected in the absence of experimental data on the heat capacity of pure components.

24
Theory

The second approach is to assume a constant heat capacity difference estimated at the melting
temperature 𝑐 , , which reduces Equation (9) to the following Equation

, 𝑇 𝑐 , 𝑇 , 𝑇 ,
ln 1 1 ln (17)
𝑅𝑇 𝑇 , 𝑅 𝑇 𝑇

The value of 𝑐 , can be determined from experimental heat capacity data. The pure liquid and
solid heat capacities as a function of temperature are linearly extrapolated to the melting
temperature of the pure component. However, experimental data on the heat capacity of pure
components are scarce. In this case, several methods or approximations have been proposed
118-125
to estimate the difference in the heat capacity at the melting temperature. Hildebrand
approximation states that the difference in the heat capacity value can be assumed equal to the
120 119 124
melting entropy. Pappa et al. , Růžička and Domalski , and Kolská et al.125 used GC
methods to estimate the heat capacity difference. Wu and Yalkowsky 121 used a simple approach
to calculate the heat capacity difference based on the molecular flexibility, molecular symmetry,
and hydrogen bond numbers calculated from the molecular structure using simple empirical
115
correlations. However, Mishra and Yalkowsky found that neglecting or considering the heat
capacity terms would rather lead to an error within the uncertainty of the solubility
measurements. Moreover, unrealistic heat capacity difference values would lead to an
47
unphysical course for the calculated liquidus line. Therefore, in case of unavailable heat
capacity data for pure solid and liquid states, it is preferable to assume the simplest approach,
i.e., neglecting the two terms on the right-hand side of Equation (9).

The third and most sophisticated approach considers the temperature dependence of the heat
capacity of the pure solid and liquid. 96, 117 A linear temperature dependence is usually sufficient.
However, this approach requires data on pure solid and liquid heat capacities over a wide range
of temperatures, which are challenging to obtain experimentally. This approach could be justified
in the case of a significant difference between the T and Tm and when experimental data for the
heat capacity of pure liquid and solid are available.

2.4.2 Activity coefficients

The activity coefficients of components represent the intermolecular interactions between unlike
molecules in the liquid phase. 102 Activity coefficient values of components equal to one indicate
that the system behaves ideally, i.e., similar intermolecular interactions between like and unlike

25
Theory

molecules. On the other hand, activity coefficients smaller or larger than one indicate favored or
unfavored intermolecular interactions between unlike molecules, respectively. The activity
coefficients can be calculated using empirical, semi-empirical, or predictive activity coefficient
models.

2.4.2.1 Empirical models

The activity coefficients of components in liquid mixtures can be described by a polynomial with
empirical parameters that are characteristic of the studied system. Redlich-Kister (RK)
polynomial is used to describe the excess Gibbs energy (gE) in any binary mixture as follows 102

𝑔 𝐴 𝐵 𝐶 𝐷 ⋯. (18)

Accordingly, the activity coefficient can be calculated as follows

𝑅𝑇 ln 𝑎 𝑏 𝑐 𝑑 ⋯ (19)

𝑅𝑇 ln 𝑎 𝑏 𝑐 𝑑 ⋯ (20)

where the parameters in equations (19) and (20) are interrelated by the four empirical parameters
in Equation (18) as follows

𝑎 𝐴 3𝐵 5𝐶 7𝐷 𝑎 𝐴 3𝐵 5𝐶 7𝐷

𝑏 4 𝐵 4𝐶 9𝐷 𝑏 4 𝐵 4𝐶 9𝐷

𝑐 12 𝐶 5𝐷 𝑐 12 𝐶 5𝐷

𝑑 32𝐷 𝑑 32𝐷

The simplest empirical activity coefficient model is the two-suffix Margules equation 102

A (21)
ln 1
𝑅𝑇

Because Equation (21) is symmetrical, the two-suffix Margules equation is only suitable for
modeling the nonideality in binary mixtures of components with similar size, shape, and chemical
nature. Empirical models can describe any liquid mixture regardless of its complexity. However,
the availability and quality of fitted experimental data determine the number of empirical

26
Theory

parameters. For nonelectrolyte organic binary mixtures, two parameters are usually sufficient.
102, 126
Empirical activity coefficient models have been successfully used to model SLE data in
several nonideal eutectic mixtures. 96, 97

2.4.2.2 NRTL

In empirical models, the activity coefficients are expressed with a polynomial whose parameters
127
can be obtained by fitting the experimental data. Wilson derived an algebraic function to
express the activity coefficients using the local composition concept. A few years later, Renon
128
and Prausnitz used the same concept of local composition to derive the nonrandom two-
liquid (NRTL) equations for calculating the activity coefficients of components as follows

𝐺 𝐺
(22)
𝐺 𝐺

𝐺 exp – 𝐺 exp – (23)

𝑔 𝑔 𝑔 𝑔
(24)
𝑅𝑇 𝑅𝑇

Unlike the Wilson model, the NRTL model can describe completely miscible and partially
miscible liquid systems. The advantage of semi-empirical over empirical models is that the
models can be extended to describe the nonideality in multicomponent systems. The binary
interaction parameters can then be used to derive the interaction parameters of multicomponent
systems. 102 The NRTL model has been shown to describe the SLE in strongly nonideal eutectic
mixtures successfully. 87-89, 96, 126

2.4.2.2 COSMO-RS
Despite the ability of the RK-polynomial and NRTL model to describe the nonideality of the
eutectic mixture constituents in the liquid phase, the models are correlative, in which
experimental SLE data are needed to fit the model parameters. Therefore, it is desirable to
employ predictive thermodynamic models to obtain the SLE phase diagram without the need for
experimental SLE data.

129-132
The conductor-like screening model for realistic solvation (COSMO-RS) has been
developed as an efficient tool to predict various thermodynamic properties. The model is based

27
Theory

on quantum mechanics and statistical thermodynamic calculations. Density functional theory

(DFT) calculations are used to obtain the screening charge surface in an ideal conductor with an
infinite dielectric constant. The screening charge surface is discretized into segments with known
surface charge density (SCD) . The pairwise interactions between segments with SCD are used
to derive the chemical potential of component 𝑖 in the liquid phase by statistical thermodynamics
calculations. The calculated chemical potential of component 𝑖 is used to calculate various
thermodynamic properties.

In the COSMO-RS model, the liquid phase solution is considered an ensemble of packed
screened molecules, where the two molecules interact in the contact region. The two types of
interactions between segments in the contact regions are: (1) misfit interactions (𝑒 , )
from the electrostatic interactions (2) hydrogen bonding interactions 𝑒 , . The total
interaction energy function 𝑒 , in COSMO-RS is

𝑒 , 𝑒 , 𝑒 , (25)

The chemical potential of segment in the liquid solution is calculated as follows

𝑅𝑇 𝑎
ln 𝑒 𝑒 , 𝑑 (26)
𝑎 𝑅𝑇

where 𝑎 is the effective contact area and is the probability distribution. The probability
distribution is the -profile obtained from the quantum mechanical calculations. is called
the -potential, which is a measure of the affinity of the liquid solution to a surface segment of
polarity . Figure 9 shows the -profile and -potential of oxalic acid calculated using Turbomole
version 6.6 and COSMOtherm X19, respectively.

28
Theory

(A) (B)

Figure 9. (A) -profile and (B) -potential of oxalic acid calculated using density functional theory and
COSMO-RS, respectively.

The pseudo chemical potential of component 𝑖 in the liquid solution is calculated from the
of its all segments

, (27)
𝑑

,
where is the combinatorial contribution due to the differences in the molecular size and
shape between component 𝑖 and the liquid solution molecules. The combinatorial contribution
is calculated as follows

, (28)
𝑅𝑇 𝑙𝑛 𝐴 𝐿 𝐿

𝐴 𝐴
𝐿 1 𝑙𝑛 (29)
𝐴 𝐴

𝑉 𝐴 𝑉 𝐴
𝐿 1 ln (30)
𝑉 𝐴 𝑉 𝐴

where , and are adjustable internal model parameters; 𝐴 and 𝑉 are the COSMO
molecular surface area and volume, and subscript 𝑖 and 𝑆 refer to component 𝑖 and the solvent,
respectively. The liquid solution COSMO molecular surface area and volume are calculated from
those of all components in the liquid solution as follows

29
Theory

𝐴𝑆𝐶 𝑆
𝑥 𝐴𝐶 𝑆
(31)

𝑉𝑆𝐶 𝑆
𝑥 𝑉𝐶 𝑆
(32)

The pseudo-chemical potential in Equation (26) (𝜇𝑃𝑆 ) is used to calculate the chemical potential
(partial molar Gibbs energy) of component 𝑖 as follows

𝜇 𝜇𝑃𝑆 𝑅𝑇 𝑙𝑛 𝑥 (33)

As seen, the only input required for COSMO-RS calculations, which allows for predicting many
thermodynamic properties, is the molecular geometry obtained from DFT calculations. COSMO-
90, 133-136
RS has been applied in several studies concerning DESs to predict SLE, liquid–liquid
137, 138 66, 139, 140
equilibria, partitioning and solubility of bioactive compounds, and in gas capture
applications. 141, 142

2.4.3 Modeling SLE phase diagram

2.4.3.1 Simple eutectic system

In simple eutectic systems, the pure components crystallize in pure form, i.e., 𝑥 𝑆 𝛾 𝑆 1 (Figure
2A). In the simple eutectic type, Equation (9) reduces to the following equation

𝑇 𝑇
, 𝑇 1 1 𝑐𝑃,
ln 𝑥 𝐿 𝛾 𝐿 1 𝑐𝑃, 𝑑𝑇 𝑑𝑇 (34)
𝑅𝑇 𝑇 , 𝑅𝑇 𝑇 𝑅 𝑇 𝑇

In case one of the constituents undergoes a solid–solid transition (Figure 2B), the liquidus line of
the constituent below its solid–solid transition temperature, neglecting the difference between
the heat capacities of the two solids, is calculated as follows

𝑇 𝑇
, 𝑇 , 𝑇 1 1 𝑐𝑃,
ln 𝑥 𝐿 𝛾 𝐿 1 1 𝑐𝑃, 𝑑𝑇 𝑑𝑇 (35)
𝑅𝑇 𝑇 , 𝑅𝑇 𝑇 , 𝑅𝑇 𝑇 𝑅 𝑇 𝑇

where and 𝑇 are the transition enthalpy and temperature, respectively.

2.4.3.2 solid solution regions

Miscibility in the solid phase is expected when two components of the same chemical nature
possess similar molecular and crystal structures. 143 The SLE phase diagram of eutectic systems

30
Theory

with partial miscibility in the solid phase is modeled using Equation (9). In this case, the activity
coefficients of components in the solid phase are needed. Empirical models such as the two-
suffix Margules or predictive models have been used in the literature to calculate the activity
coefficients of alkanes and paraffines in the solid solution. 144-147 However, models for describing
the nonideality in the solid solution of salts and substances forming hydrogen bonding networks
are yet unavailable.

2.4.3.3 cocrystal formation

The formation of congruently (Figure 2D) or incongruently (Figure 2E) cocrystals with the
stoichiometric coefficients and for components A and B, respectively, can be described
via the following chemical reaction

𝐴 𝐵 ↔ 𝐴 𝐵 (36)

The equilibrium constant 𝐾 of the chemical reaction in Equation (35) is described as follows

𝐾 𝑎 (37)

As the cocrystal AB crystallizes in pure form, i.e., 1, Equation (36) is simplified to

𝐾 (38)

To determine the equilibrium constant at different temperatures, the following Gibbs-Helmholtz

equation is used

1 1 (39)
ln 𝐾 ln 𝐾
𝑅 𝑇 𝑇

where and 𝑇 are the melting enthalpy and temperature of the cocrystal. 𝐾 is the
equilibrium reaction constant determined at 𝑇 and cocrystal stoichiometry as follows

𝐾 (40)

where and are calculated at the cocrystal stoichiometry. As seen, to model SLE in eutectic
systems with cocrystal formation, the melting properties and stoichiometry of the cocrystal as

31
Theory

well as the activity coefficients of the components in the liquid phase are required. The melting
properties of the cocrystals can be measured by DSC, similar to pure components. In contrast,
determining the stoichiometry of the cocrystal is challenging. Powder XRD analysis over the
whole composition range can hint at the stoichiometry of the cocrystal. However, the cocrystal
stoichiometry can only be verified by single-crystal XRD (SC-XRD) or structure solution from
powder XRD data. The latter requires high-quality XRD data and expertise. Moreover, the crystal
structure obtained from powder XRD data is usually of poor quality. Thus, in many cases, the
SC-XRD technique is inevitable for the proper determination of cocrystal stoichiometry.

32
Results

3. Results
The results of the thesis are presented in six published papers. In Paper I, a parameter study
was performed to understand the influence of the melting properties of pure components and
their activity coefficients in the liquid solution on the position of the eutectic point. Paper II reports
SLE data for six different eutectic systems formed by mixing L-menthol with linear and cyclic
monocarboxylic acids to study the influence of the molecular structure of constituents on the
eutectic temperature. Paper III investigates glass formation and polymorphism in eutectic
systems, which can lead to misinterpretations of the phase diagram and the eutectic
temperatures of the eutectic systems. In Paper IV, an efficient sample preparation method was
proposed, allowing for measuring the phase diagram of the L-menthol/thymol eutectic system.
Paper V highlights cocrystal formation in ChCl-based DES and evaluates the melting properties
of ChCl estimated in the literature. Finally, In Paper VI, cocrystal formation in the L-
menthol/phenol eutectic system was invistigated experimentally and modeld using correlative
and predictive thermodynamic models.

33
3.1 Paper I

3.1 Paper I
Modeling of Solid–Liquid Equilibria in Deep Eutectic Solvents: A Parameter Study

A. Alhadid, L. Mokrushina and M. Minceva, Molecules, 2019, 24, 2334.

Author contribution: The thesis author conceptualized the paper’s idea, performed the
calculations, and interpreted the results. He wrote the manuscript draft and completed the
editing of the manuscript.

Summary: Paper I aims to investigate the parameters that affect the SLE in binary eutectic
mixtures of the simple eutectic type, namely, the melting properties of pure components and the
nonideality in the liquid solution, and their interrelation. The parameter study was performed on
hypothetical and selected real binary eutectic systems. Although previous works have performed
a similar parameter study, 47, 49, 74 Paper I was the first to quantify the influence and provide direct
interpretation for each parameter value. The depression at the eutectic point was described as
the normalized difference between the eutectic temperature and the melting temperature of the
low melting component.

The influence of the melting properties of pure components on the depression at the eutectic
point was studied by screening different melting temperatures of components at constant
melting entropy and assuming ideal solution behavior. It was found that the depression at the
eutectic point decreases as the difference between the melting temperature of both components
increases. Therefore, the largest depression at the eutectic point at constant melting entropy is
observed when both components possess the same melting temperature. However, the
influence of the difference between the melting temperature of both components is relatively
insignificant compared to the absolute values of the melting enthalpy and entropy of pure
components. Accordingly, ideal eutectic mixtures with large depression at the eutectic point can
be formed by mixing constituents possessing low melting enthalpy and entropy values.

Next, the influence of the activity coefficients of components in the liquid phase on the
depression at the eutectic point was studied. The activity coefficients in the liquid phase were
calculated using the two-suffix Margules equation by screening different binary interaction
parameter values. It was found that a large depression at the eutectic temperature cannot result
from the substantial negative deviation from ideal behavior only but rather from a combination
of low melting enthalpy of components and negative deviation from ideality. The influence of the

34
3.1 Paper I

melting enthalpy and nonideality of the high melting component on the depression at the eutectic
point is more prominent than those of the low melting component.

Further, the interrelation between pure components melting properties and their activity
coefficients in the liquid phase was studied to evaluate the melting properties of thermally
unstable substances estimated in the literature. It was found that any combination of melting
enthalpy and the activity coefficient model parameters can describe the SLE data of the thermally
unstable salt. However, the calculated activity coefficients of the thermally unstable salt in the
liquid phase were different in each case, which lead to a different interpretation of the nonideality
of the component in the same liquid solution. Therefore, it was concluded that due to the
interrelation between the melting properties of pure components and their activity coefficients,
the method for obtaining the melting properties of thermally unstable salts, such as ChCl, using
the SLE data could be unreasonable.

35
 molecules
Article
Modeling of Solid–Liquid Equilibria in Deep Eutectic
Solvents: A Parameter Study
Ahmad Alhadid 1 , Liudmila Mokrushina 2 and Mirjana Minceva 1, *
1 TUM School of Life and Food Sciences Weihenstephan, Technical University of Munich, Biothermodynamics,
Maximus-von-Imhof-Forum 2, 85354 Freising, Germany; [email protected]
2 Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Separation Science & Technology, Egerlandstr. 3,
91058 Erlangen, Germany; [email protected]
* Correspondence: [email protected]

Academic Editors: Mert Atilhan and John M. Herbert !"#!$%&'(!

!"#$%&'
Received: 5 April 2019; Accepted: 24 June 2019; Published: 25 June 2019

Abstract: Deep eutectic solvents (DESs) are potential alternatives to many conventional solvents in
process applications. Knowledge and understanding of solid–liquid equilibria (SLE) are essential to
characterize, design, and select a DES for a specific application. The present study highlights the main
aspects that should be taken into account to yield better modeling, prediction, and understanding of
SLE in DESs. The work is a comprehensive study of the parameters required for thermodynamic
modeling of SLE—i.e., the melting properties of pure DES constituents and their activity coefficients
in the liquid phase. The study is carried out for a hypothetical binary mixture as well as for selected
real DESs. It was found that the deepest eutectic temperature is possible for components with low
melting enthalpies and strong negative deviations from ideality in the liquid phase. In fact, changing
the melting enthalpy value of a component means a change in the di↵erence between solid and liquid
reference state chemical potentials which results in di↵erent values of activity coefficients, leading to
di↵erent interpretations and even misinterpretations of interactions in the liquid phase. Therefore,
along with reliable modeling of liquid phase non-ideality in DESs, accurate estimation of the melting
properties of their pure constituents is of clear significance in understanding their SLE behavior and
for designing new DES systems.

Keywords: deep eutectic solvents; solid–liquid equilibria; modeling phase equilibria; melting
properties; activity coefficient models

1. Introduction
Deep eutectic solvents (DESs) are a class of eutectic mixtures of two or more compounds
that have a eutectic point far below the melting temperatures of the individual components [1].
The large depression in the melting temperature of the mixture is commonly attributed to
strong hydrogen bonding interactions between DES constituents. The recent trend toward green
solvents has increased the potential of DESs in many applications [2–5]—e.g., as solvents for
active pharmaceutical ingredients [6–8], fuel production [9,10], solid–liquid and liquid–liquid
extraction [11–16], and solid–support free liquid–liquid chromatography [17–20].
Similar to ionic liquids (ILs), DESs are considered designer solvents because their properties
can be tailored by choosing di↵erent combinations of its constituents [2]. However, the ratio of
constituents—i.e., the eutectic composition—is not as easy to determine as it is for ILs, in which the
ratio of cations and anions is determined by mixture electroneutrality. Many molecular simulation
studies can be found in the literature; these aimed to understand DES systems at the molecular level
and determine their eutectic composition [21–35]. It should be noted, however, that these studies were
mainly performed using DESs of a fixed composition corresponding to a supposed hydrogen bond

Molecules 2019, 24, 2334; doi:10.3390/molecules24122334 www.mdpi.com/journal/molecules

Molecules 2019, 24, 2334 2 of 19

complex. Therefore, the studied compositions are not representative of the entire range of possible
compositions. Even with fixed ratios between DES constituents, no specific microstructure was found,
but rather a network of hydrogen bonding [35]. It has been postulated that the hydrogen bonding
network structure could be correlated to eutectic temperature [34]. Molecular simulation studies are
a helpful tool in modeling and understanding interactions between DES constituents at a molecular
level, but they are so far unable to provide sufficient macroscopic information required for process
applications. Although knowledge and understanding of solid–liquid equilibria (SLE) in DES systems
are essential to designing and selecting the appropriate system for a particular application, only a
small fraction of DES related literature contains information about their SLE [1,36–47].
Eutectic systems are mixtures of two or more components in which the components show complete
or partial immiscibility in the solid state at the mixture melting temperature [48]. Although eutectic
mixtures have been long studied, no agreement for strict criteria to distinguish between eutectic
mixtures and “deep” eutectic mixtures can be found in literature [49]. In recent publications, the term
DES or natural DES (NADES) is typically used for any mixture that forms a eutectic mixture with a low
freezing temperature. Smith et al. [4] defined DESs as systems that have a large di↵erence between
the melting temperature of the mixture at eutectic composition and the weighted sum of the melting
temperatures of the pure components. They assumed that the deep eutectic point is associated with the
formation of a new complex via hydrogen bonding between DES constituents; therefore, the eutectic
composition represents the stoichiometry of this complex. Based on this definition, the search for
eutectic compositions has been commonly performed by trial and error using a specific ratio between
constituents such as 1:1, 1:2, etc. Despite the fact that in some DES systems the eutectic composition
represents a specific molar ratio between the constituents—such as 1 choline chloride ([Ch]Cl):2 urea
or 1 [Ch]Cl:1 oxalic acid—the eutectic point of any phase diagram is determined by the intercept of the
liquidus lines of the components, even if the solution behaves ideally. Therefore, complex formation
might not be the only reason for deep eutectic formation and the eutectic composition may not accord
with a specific molar ratio between the constituents.
Martins et al. [49] proposed a more comprehensive definition for DES systems. According to
these authors, DES systems are mixtures of two or more components—i.e., not a new compound. At
the same time, DES systems should be di↵erentiated from other simple eutectic systems based on
their strong negative deviations from ideality—i.e., the eutectic temperature is lower than that of the
ideal eutectic. In addition, the term DES should not refer exclusively to eutectic composition but to
any composition at which the mixture is liquid at operating temperature. Based on this definition,
studying the non-ideality of the liquid phase is essential to determining whether or not a system is
a DES. Following this, most recent literature studies on modeling SLE in DESs have focused mainly
on modeling non-ideality in the liquid phase as a reason behind the formation of DESs [42–46,49,50].
From a thermodynamic perspective, SLE do not depend only on the intermolecular interactions in the
liquid phase but also on a correct estimate of the reference state—i.e., the ratio of the fugacities of the
pure components in the solid and subcooled liquid phases at system temperature. Although this was
clearly and explicitly stated many years ago in Prausnitz et al. [51], it continues to be overlooked in
the literature. Only consideration of the interplay of both pure component melting properties and
solution non-ideality can result in reliable modeling of SLE in DESs, which will help in understanding
their nature. Moreover, activity coefficients and melting properties are not independent since the
value of the reference state chemical potential the activity coefficients refer to is directly related to the
melting properties.
Although melting enthalpy and melting temperature can be measured using di↵erential scanning
calorimetry, it is still a difficult issue because of polymorphism, possible metastable solid phases, kinetic
limitations due to high liquid viscosities close to the melting temperature, etc. For some substances, it
is impossible to measure them owing to their thermal instability. These issues result in limited and
inaccurate data, especially on the enthalpy of fusion. Despite the fact that the literature reports several
Molecules 2019, 24, 2334 3 of 19

theoretical methods for the estimation of melting properties, these methods usually su↵er from many
limitations and are hardly applicable to ionic compounds [52].
Recently, Kollau et al. [47] and Martins et al. [49] studied the e↵ect of melting properties and
non-ideality on SLE phase diagrams. Defining DES as mixtures with high negative deviation from
ideality, they demonstrated that the SLE phase diagram should be known to distinguish DES from
other eutectic mixtures and that the non-ideality in DESs can be quantitatively described by the
regular solution theory or the PC-SAFT equation of state. However, the focus of these studies was
to convince the community that only eutectic mixtures with high negative deviations from ideality
should be considered as DES. A number of issues covered in the present study were not discussed or,
if considered, then indirectly and not in detail.
The present work presents a systematic study of both melting properties and solution non-ideality,
with the goal of understanding DES formation. Consideration is first made based on a hypothetical
model eutectic system to find the link between eutectic point properties—eutectic temperature and
composition—with both the melting properties and activity coefficients of the components. Then,
simplified modeling of [Ch]Cl/ILs eutectic systems is carried out to evaluate the e↵ect of small melting
enthalpies on the interpretation of the modeling results. Further, DES systems composed of quaternary
ammonium chlorides and fatty acids are modeled to evaluate the e↵ect of overestimation of melting
enthalpies of components and to support the initial findings. The only criteria for selecting the
activity coefficient model used in this work—namely, Redlich–Kister polynomial—is its simplicity
and flexibility in describing curves of di↵erent shape and complexity. It should be pointed out that
this work does not propose or recommend this model to be used for SLE calculations in DES. The
calculations done in this work do not aim at improved modeling of published SLE data; instead, the
aim is to encourage discussion and further interpretation of these data.
The general objective of this work is to provide the deeper understanding essential to establishing
the basic steps required in defining and designing DES systems based on their SLE behavior. Similar to
molecular simulation approaches, thermodynamic modeling of SLE can be a valuable tool to assess the
behavior of components in the liquid phase—i.e., intermolecular interactions. Designing DES systems
based on their SLE estimated from activity coefficient models or equation of state might have an
advantage over molecular simulation approaches owing to the direct estimation of eutectic temperature,
eutectic composition, and the composition range at which the DES is a liquid at operating temperature.

2. Theory
Most studied DESs have been considered as simple eutectic systems in which pure constituents
crystallize in the form of pure components at the mixture liquidus temperature. Figure 1 shows a
schematic of a solid–liquid phase diagram of a simple eutectic system. Curve (a’) and (a) are the liquidus
lines of Component 1 and 2, respectively. They intercept at the eutectic point, which corresponds to the
lowest melting temperature—the eutectic temperature (Te )—of a mixture of eutectic composition (xe ).
Molecules 2019, 24, 2334 4 of 19
Molecules 2019, 24, x FOR PEER REVIEW 4 of 20

Figure
Figure 1. Schematic representation
1. Schematic representation of
of aa solid–liquid
solid–liquid phase
phase diagram
diagram of
of aa simple
simple eutectic
eutectic system.
system.

The
The liquidus
liquidus line
line of
of component
component i isiscalculated
calculatedasas[51]
[51]
! !
µS0i µL0i f0iS Dhm,i T DcP Tm,i Tm,i (1)
ln xLi L
i = = ln L
= 1 1 + ln (1)
RT f0i RT Tm,i R T T
where and are the mole fraction and activity coefficient of component in the liquid solution,
L
respectively; L
where xi and i are the moleand fraction
are the referenceand chemical
activity coefficient
potentials of pure component
componenti in thein liquid solution,
the subcooled
respectively;
liquid and solid L
µ0i states S
and µ0iatare the reference
liquidus temperature,chemical potentials of pure
respectively; and component i in the subcooled
are the reference fugacities
liquid
of pureand solid states in
component at the
liquidus temperature,
subcooled liquid and solid statesf0iLatand
respectively; f0iS aretemperature,
liquidus the reference respectively;
fugacities of
pure component
and in the
iare thesubcooled liquid andand
melting enthalpy solidthe
states at liquidus
melting temperature,
temperature of purerespectively;
component Dhm,i,
and Tm,i are the
respectively; T ismelting enthalpy
the liquidus and the meltingistemperature
temperature; the difference of pure
in heat component
capacity of i, respectively;
pure component T is
ithe
in liquidus
the solidtemperature;
and liquid states DcP, i is the di↵erence in heat capacity of pure component i in the solid and
at constant pressure; R is the universal gas constant.
liquidThe states atterm
constant pressure;
on the right-hand R isside
the universal
of Equation gas(1)constant.
is typically of low value in comparison with
The Dc term.
P term
In on the
addition, right-hand
the side
values of Equation
are usually (1) is typically
unavailable from of previous
low value in comparison
experiments and
with the Dhm,i
are difficult to term.
estimate In addition, the DcPaccuracy
with reasonable values are usually
because nounavailable
liquid existsfrom previous experiments
at temperatures below the
and
meltingare difficult
temperature.to estimate with reasonable
As a result, the second accuracy
term onbecause no liquidside
the right-hand exists
of at temperatures
Equation (1) hasbelow
been
the meltingneglected
commonly temperature. As a result,
in literature andthe thesecond
followingtermsimplified
on the right-hand
equation side of Equation
is used to calculate(1) has
the
been
liquiduscommonly
lines neglected in literature and the following simplified equation is used to calculate the
liquidus lines ! (2)
L L Dhm,i T
ln xi i = 1 (2)
RT Tm,i
According to Equation (2), the position of the eutectic point, as well as the whole SLE phase diagram,
According to Equation (2), the position of the eutectic point, as well as the whole SLE phase diagram,
depend on both the melting properties—melting enthalpy and melting temperature—of the pure
depend on both the melting properties—melting enthalpy and melting temperature—of the pure
components and their activity coefficients in solution. When the solution is assumed as ideal (
components and their activity coefficients in solution. When the solution is assumed as ideal ( Li ),
), the SLE phase diagram can be modeled based on the melting properties of the pure components
the SLE phase diagram can be modeled based on the melting properties of the pure components
only. However, if the system deviates from ideal behavior, the activity coefficients of components in
only. However, if the system deviates from ideal behavior, the activity coefficients of components in
the liquid phase can be calculated using well-developed gE models or equations of state, which are
the liquid phase can be calculated using well-developed gE models or equations of state, which are
available in the literature.
available in the literature.
3. Results
3. Results and
and Discussion
Discussion

3.1. Parameter Study on a Hypothetical Binary System

In the
the present
present section,
section, aa hypothetical
hypothetical binary
binary DES
DES system
system isis considered.
considered. The high melting
melting
component is labeled
labeled as
as 11 and
and the
the low
low melting
melting component
component is
is labeled
labeled as
as 2.
2.

3.1.1. Melting Properties of Pure Components

Molecules 2019, 24, 2334 5 of 19

Molecules 2019, 24, x FOR PEER REVIEW 5 of 20

3.1.1. Melting Properties of Pure Components
The
Themelting
meltingproperties
propertiesofofthe thepure
pure components—melting
components—melting enthalpy
enthalpy andandmelting
meltingtemperature—are
temperature—
prerequisites for SLE
are prerequisites diagram
for SLE diagram calculations
calculations evenevenif the solution
if the solution is is
assumed
assumedideal.ideal.Melting
Meltingenthalpy
enthalpy
and
andentropy
entropydepend
dependon onthethemolecular
molecular arrangement
arrangement in in the
the crystal
crystal lattice—i.e.,
lattice—i.e., molecular
molecular symmetry,
symmetry,
conformational
conformationaldiversity,
diversity, andand intermolecular
intermolecular forces [52]. The
forces [52].melting
The meltingtemperature is the ratio
temperature between
is the ratio
melting enthalpy and entropy of the component; therefore, a high melting
between melting enthalpy and entropy of the component; therefore, a high melting temperature of a temperature of a compound
corresponds to a high melting
compound corresponds to a highenthalpy,
meltinga low meltingaentropy,
enthalpy, low melting or both [53]. Highly
entropy, or bothordered solid
[53]. Highly
structures usually
ordered solid have high
structures melting
usually temperatures
have high melting as atemperatures
result of a high asmelting
a resultenthalpy.
of a high However,
melting
disordered solid structures
enthalpy. However, disordered maysolid alsostructures
have highmay melting temperatures
also have high melting as atemperatures
result of lowasmeltinga result
entropy. In thisentropy.
of low melting case, theIn lowthismelting
case, the entropy corresponds
low melting entropytocorresponds
association to in association
the liquid phase—as
in the liquid a
result of strong
phase—as hydrogen
a result of strongbonding
hydrogen [54]—or due to
bonding a symmetrical
[54]—or due to aand rigid molecular
symmetrical structure
and rigid [53].
molecular
Therefore, it is Therefore,
structure [53]. very difficultit is to
very correlate
difficultthe melting enthalpy
to correlate the melting values of compounds—or
enthalpy values of compounds— even to
obtain
or evenantoestimate
obtain anbased
estimate on based
melting on temperature
melting temperatureonly—withoutonly—withoutinformation relatingrelating
information to (i) their
to (i)
crystal and molecular structures and (ii) interactions between like-molecules
their crystal and molecular structures and (ii) interactions between like-molecules in the liquid phase. in the liquid phase.
However,
However,according
accordingto toBondi’s
Bondi’sobservation,
observation,melting meltingentropy
entropyof ofcomponents
componentscan canbebebetter
betterrelated
relatedtoto
the
the molecular
molecular structure
structure [55],
[55], although
although itit cannot
cannot be bemeasured
measured and and isisestimated
estimatedfrom fromexperimental
experimental
melting
meltingenthalpy
enthalpyand andtemperature
temperatureof ofcompounds.
compounds.Therefore,
Therefore,two twocases
casesareareconsidered
consideredin inthis
thissection
section
to
torepresent
representcomponents
components of ofdi↵erent
differentcrystal
crystalstructures.
structures. In In the
the first
first case,
case, aa high
high melting
melting entropy
entropy isis
assumed
assumedof 54.4J mol 1 K 1 (Ds
of54.4 ( m /R ⇡ 7)—also )—also known
known as as the
the Walden
Walden rule rule (Equation
(Equation (3))—to
(3))—to represent
represent
compounds
compoundswith withordered
orderedcrystals
crystalsofofrigid
rigid molecules
molecules [53].
[53].InIn
thethesecond
second case, a low
case, melting
a low melting entropy of
entropy
of J mol 1 K 1 (Dsm( /R ⇡ 2.4) is assumed
20.0 ) is assumed to represent
to representcompounds
compounds with disordered
with disordered crystals of rigid
crystals of
molecules [53]. In this section, the e↵ect of melting properties on the eutectic
rigid molecules [53]. In this section, the effect of melting properties on the eutectic point properties— point properties—eutectic
temperature (Te ) and(T
eutectic temperature composition
e) and composition (xe )—of(xae)—of
hypothetical ideal binary
a hypothetical system
ideal binary is studied,
system is studied,withwith
the
results shown in Figures 2–4. 2–4.
In allIncases, SLE SLE
are modeled usingusing
Equation (2) assuming L = 1.
the results shown in Figures all cases, are modeled Equation (2) assuming i = 1.

Figure2.2.E↵ect
Figure Effectof
ofmelting
meltingtemperature
temperatureof
ofpure
purecomponents
componentswithwithconstant
constantmelting
meltingentropy
entropyonon(a)
(a)the
the
eutectic
eutectictemperature
temperature(b)
(b)the
thenormalized
normalizeddepression
depressionat
ateutectic
eutectictemperature
temperaturerelative
relativeto
tothat
thatof
ofthe
thelow
low
melting
meltingComponent
Component2.2.
Molecules 2019, 24, x FOR PEER REVIEW 6 of 20
Molecules 24,2334
2019,24,
Molecules2019, x FOR PEER REVIEW 66 of
of19
20

Figure 3. Effect of the melting enthalpies of pure components with constant melting entropy on the
Figure3.3.E↵ect
Figure Effectofofthe
themelting
meltingenthalpies
enthalpiesofofpure
purecomponents
componentswith withconstant
constantmelting
meltingentropy
entropyon
onthe
the
normalized depression at eutectic temperature relative to that of Component 2.
normalized depression at eutectic temperature relative to that of Component
normalized depression at eutectic temperature relative to that of Component 2. 2.

Figure 4.4. E↵ect

Figure Effect of
of (a)
(a)melting
meltingtemperature
temperature (b)
(b) melting
melting enthalpy
enthalpy of
of pure
pure components
components on
on the
the eutectic
eutectic
Figure 4. EffectComponent
composition of (a) melting temperature (b) melting enthalpy of pure components on the eutectic
composition of of Component1. 1.
composition of Component 1.
Figure
Figure 22 represents
represents the the e↵ect
effect of
of melting
melting temperatures
temperatures of of pure
pure components
components on on the the eutectic
eutectic
temperature
Figure of the system.
2 represents As
the seen
effect from Figure
of melting 2a, the absolute
temperatures of value
pure of
temperature of the system. As seen from Figure 2a, the absolute value of the eutectic temperatures the eutectic
components ontemperatures
the eutectic
changes
changesas
temperature
as the
the melting
melting temperature
temperatureof
of the system.
ofthe
As seen from
thelowlowand andthe
Figure 2a,
the high
high melting
melting components
the absolute
componentschanges.changes. The
value of the eutectic
The higher
temperatures
higher
the values
changes of
as the
the melting
melting temperatures
temperature of of both
the low components,
and the high the higher
melting
the values of the melting temperatures of both components, the higher the eutectic point of the the eutectic
components point
changes. of the
Thesystem.
higher
Asthethe di↵erence
values
system.
of thebetween
meltingthe
As the difference components
temperatures
between
ofincreases, theincreases,
both components,
the components absolutethe eutectic
higher
the absolutetemperature
the eutecticincreases
point ofand
eutectic temperature
the
reaches
system.
increases a constant
As the
and reaches value
difference that is close
between
a constant valuetothat
thethe melting
components
is close to temperature
the meltingoftemperature
increases, Componentof2.Component
the absolute The low value
eutectic temperature of
2. The
melting
increases entropy
and reachesof the acomponents
constant value at constant melting
that is close to thetemperatures further
melting temperature
low value of melting entropy of the components at constant melting temperatures further decreases decreases
of Component the eutectic
2. The
temperature
low value ofofmeltingthe system—the
the eutectic temperature
entropy ofblue curves blue
the components
of the system—the in Figure 2a in
at constant
curves areFigure
always
meltinglower than the
temperatures
2a are always lowerblack
further
than curves.
decreases
the black As
seen
the
curves.in Figure
As seen2b,
eutectic
in if
temperaturethe normalized
Figure
of the di↵erence difference
system—the
2b, if the normalized
blue between the
curves in eutectic
Figure
between
2a
thetemperature
are always and the melting
lower
eutectic temperature
than the black
and the
temperature
curves. As seen of theinlow melting
Figure 2b, ifcomponent
the normalizedis selected as
difference a
melting temperature of the low melting component is selected as a measure for the measure
between for the
the depression
eutectic at eutectic
temperature
depression point,
and the
at
the absolute
melting values
temperature of the melting
of the temperatures
low melting componentof both iscomponents
selected as have no
a measure
eutectic point, the absolute values of the melting temperatures of both components have no effect on e↵ect
foron thethe depression
depression at
at
theeutectic
eutectic point.
point,
depression at This
the e↵ect
absolute
eutectic is a result
values
point.
of of fixing
the
This effect
melting the melting
temperatures
is a result entropy
of fixing
of bothof components—increasing
the melting
components have no
entropy of components—
effect the
on
melting temperatures
the depression of
at eutectic components
point. This at constant
effect melting
is a result entropy
of fixing the means
increasing the melting temperatures of components at constant melting entropy means an increase
meltinganentropy
increase of in the melting
components—
enthalpy
increasing of components.
in the melting Therefore,
the melting temperatures
enthalpy of components. theof e↵ect of melting
components
Therefore, the temperatures
at constant melting
effect of melting of components is negligible
entropy means
temperatures if the
an increase
of components
melting
in the entropy
melting and similarly
enthalpy of the melting
components. enthalpy
Therefore, of components
the effect of is high.
melting
is negligible if the melting entropy and similarly the melting enthalpy of components is high. Therefore,
temperatures at constant
of value
components
ofismelting
negligible
Therefore, entropy,
if thethe
at constant largest
melting
value of depression
melting is
entropy and observed
entropy, theif largest
similarly the themelting
melting temperatures
enthalpy
depression of bothif components
of components
is observed
is high.
the melting
are equal.
Therefore,
temperatures
at constant value of melting
of both components are equal.
entropy, the largest depression is observed if the melting
temperatures of both components are equal.
Molecules 2019, 24, 2334 7 of 19

In Figure 3, the e↵ect of the melting enthalpies of pure components on the normalized depression
at eutectic point is studied. The di↵erent curves correspond to di↵erent melting entropy values. In
contrast to melting temperatures of pure components, the absolute values of the melting enthalpies
of components a↵ect the normalized depression at eutectic point. The lower the values of melting
enthalpies, the higher the depression at eutectic point. As can be noticed from Figure 3, if the value of
the melting enthalpy of Component 1 is high no depression is observed. In addition, only the low value
of melting enthalpy (5 kJ mol 1 ) of both components can a↵ect the depression at eutectic point—this
can be seen by the shift between solid and dashed lines, while the dotted and the dashed lines are
the same. In conclusion, for ideal eutectic systems in which the components possess di↵erent melting
temperatures, only low value of melting enthalpies of components can lead to deep eutectics.
Figure 4 shows how the eutectic composition depends on melting properties for the systems shown
in Figures 2 and 3. As can be seen in Figure 4, an equimolar eutectic composition is possible if both the
melting enthalpies and melting temperatures of the components are equal. The eutectic composition
is located very close to the pure Component 2 side (x1,e = 0) if the di↵erence between components
melting temperatures or the melting enthalpy of Component 1 is of a high value. The low melting
enthalpy value of Component 1 shifts the eutectic composition to a higher content of Component 1
(toward higher x1,e values). The change in the absolute values of the melting enthalpies of components
shifts the eutectic composition to di↵erent values. However, the e↵ect of melting properties on the
eutectic composition is lower compared to that on the depression at eutectic point—compare the shift
between black and blues curves in Figures 2 and 3 to that in Figure 4.

3.1.2. Non-Ideality of Components in Liquid Phase

The e↵ect of both parameters—namely the activity coefficients and melting properties—is
demonstrated here for a hypothetical binary system. As shown in the previous section, the absolute
values of the melting temperatures of components have no e↵ect on the normalized depression at
eutectic point at fixed melting entropies. Therefore, arbitrary melting temperatures of the high melting
Component 1 and the low melting Component 2 are set to 600 K and 300 K, respectively. The melting
enthalpy and entropy is used to investigate the e↵ect of melting properties. Four cases are present in
Figure 5. In the first case, the Walden rule (Equation (3)) is used to calculate the melting enthalpies.
In the second case, the melting enthalpy of Component 1 (12.0 kJ mol 1 ) is estimated assuming
a low value of melting entropy of 20 J mol 1 K 1 , whereas the melting enthalpy of Component
2 (16.32 kJ mol 1 ) is calculated using the Walden rule. In the third case, the melting enthalpy of
Component 2 (6.0 kJ mol 1 ) is estimated using a low value of melting entropy of 20 J mol 1 K 1 ,
whereas the melting enthalpy of Component 1 (32.64 kJ mol 1 ) is calculated using the Walden rule.
In the fourth case, the melting enthalpies of both components are estimated using a low value of
melting entropy of 20 J mol 1 K 1 . SLE are calculated using Equation (2) with activity coefficients
screened using RK-polynomials (Equation (4)). The values of the parameters a(i) in Equation (4) for
each
⇣ component
⌘ are varied in the range 30–0 to account for the negative deviation from ideality
L
(ln i < 0). It should be mentioned that these parameters have no significant physical meaning, and
they are serving just as a measure for the value of the activity coefficients—i.e., the lower the value of
parameter a(i) , the lower the value of Li at the eutectic point.
Molecules 2019, 24,
Molecules
Molecules
Molecules 233424,
2019,
2019,
2019, 24,
24,xxxFOR
FOR
FORPEER
PEER
PEERREVIEW
REVIEW
REVIEW 8 of 19888 of
of
of 20
20
20
Molecules 2019, 24, x FOR PEER REVIEW 8 of 20

E↵ect
FigureFigure
5.
Figure
Figure 5.of
5.
5. non-ideality
Effect
Effect
Effect of
of of components
ofnon-ideality
non-ideality
non-ideality of
ofcomponents
of components
componentson the onnormalized
on
onthe
the
thenormalized
normalized
normalized depression at eutectic
depression
depression
depression at temperature
ateutectic
at eutectic
eutectic temperature
temperature
temperature
Figure 5. Effect of non-ideality of components on the normalized depression at eutectic temperature
relative to the to
relative
relative
relative melting
to
tothe
the temperature
themelting
melting
melting temperature
temperatureof Component
temperature of
ofofComponent
Component2 assuming
Component 222assuming
assuming
assuming di↵erent melting
different
different
different melting
melting
melting enthalpy values
enthalpy
enthalpy
enthalpy values
valuesof
values of
of
relative to the melting temperature of Component 1 2 assuming different 1melting enthalpy values of1
of components (a) Dhm,1 = 32.64 kJ mol and Dhm,2 = 16.32 kJ mol (b) Dhm,1 = 12.00 kJ mol
components
components
components (a)
(a)
(a) and
and
and (b)
(b)
(b) and
and
and
components (a) and (b) and
and Dhm,2 = 16.32 kJ mol 1(c) (c)(c) Dhm,1 = 32.64 kJ mol and
(c) 1
and
and and Dhm,2 = 6.00 kJ(d) (d) 1
mol
(d)
(c) and (d)
(d) Dhm,1 = 12.00 kJ mol and1 and Dhm,2 = 6.00 kJ...mol
and
and The
The 1melting
. The melting
Themelting
melting temperatures
temperatures
temperatures temperatures
of
ofComponents
of Components of Components
Components 111and
and
and222are are
areset
set
set
and . The melting temperatures of Components 1 and 2 are set
1 and 2to
are
to
to setK
600
600
600 K
Ktoand
600 300
and
and K
300and
300 K,300
K,
K, K, respectively.
respectively.
respectively.
respectively. The The parameter
Theparameter
The parameter
parameter value
value
value valueis
isisonlyis only
only
only a measure
aaameasure
measure
measure forfor
for
for non-ideality
non-ideality
non-ideality
non-ideality with
withno
with no
no
to 600 K and 300 K, respectively. The parameter (1) value is only(1) a measure for (1) non-ideality with no
with nophysical
physicalsignificance.
physical
physical significance. Legend:
significance.
significance. Legend:
Legend:
Legend:(1) aaa(1)
(1)
(1)= 30; (1)
== 30;
30;
30; aaa(1)
(1)
(1)==== 18;
18;
18; (1) aaaa(1)
(1) === 12;
(1)= 12;
12; (1) aaa(1)
a(1)
(1) ===0.
(1)= 0.
0.
0.
physical significance. Legend: a = 30; a = 18; a = 12; a = 0.
FigureFigure
5 demonstrates
Figure
Figure 555demonstrates
demonstrates
demonstrates how solution
how
howsolution
how non-ideality
solution
solution non-ideality
non-ideality
non-ideality a↵ectsaffects depression
affects
affects depression
depression
depression at the at eutectic
at
atthe
the temperature
theeutectic
eutectic
eutectic temperature
temperature
temperature
Figure 5 demonstrates how solution non-ideality affects depression at the eutectic temperature
for di↵erent
for
for melting
fordifferent
different
different enthalpies
melting
melting
melting of the of
enthalpies
enthalpies
enthalpies components.
of
ofthe the
thecomponents.
components.
components. As in As the
As
Asin case
in theofcase
inthe
the theof
case
case ideal
of themixture,
ofthe
the ideal the lower
idealmixture,
ideal mixture,
mixture, the
thelower
the the
lower
lowerthe the
the
for different melting enthalpies of the components. As in the case of the ideal mixture, the lower the
melting enthalpy
melting
melting
melting of Component
enthalpy
enthalpy
enthalpy of
of 1, the larger
of Component
Component
Component 1,
1,1, the
thethelarger
the normalized
larger
larger the depressiondepression
the normalized
the normalized
normalized at eutectic temperature.
depression
depression at
at eutectic
at eutectic When
eutectic temperature.
temperature.
temperature.
melting enthalpy of Component 1, the larger the normalized depression at eutectic temperature.
the melting
Whenenthalpy
When
When the
themelting
the melting
melting of Component
enthalpy
enthalpyof
enthalpy of1Component
of is of a low value—see
Component
Component 111isis
isofof
ofaaalowlow
low Figure
value—see
value—see
value—see 5b,d—depression
Figure of the meltingof
Figure5b,d—depression
Figure 5b,d—depression
5b,d—depression of
ofthe
the
the
When the melting enthalpy of Component 1 is of a low (1) value—see Figure 5b,d—depression of the
point melting
occurs,
meltingeven
melting point
point
point if occurs,
Component
occurs,
occurs,even even1ifififbehaves
even Component
Component
Component ideally (a = 0;ideally
111 behaves
behaves
behaves orange(a
ideally
ideally curves
(a
(a =
(1) =
(1)
(1) = 0;
0; in
0; Figurecurves
orange
orange
orange 5b,d). in
curves
curves However,
in
in Figure
Figure
Figure 5b,d).
5b,d).
5b,d).
melting point occurs, even if Component 1 behaves ideally (a(1) = 0; orange curves in Figure 5b,d).
when However,
the melting
However,
However, when
when
when enthalpythe
the of Component
themelting
melting
melting enthalpy
enthalpyof
enthalpy 1ofis
of of a high value—see
Component
Component
Component 111isis
isof
of
ofaaahigh
high
high Figure
value—see
value—see
value—see 5a,c—depression
Figure
Figure5a,c—depression
Figure of the
5a,c—depression
5a,c—depression
However, when the melting enthalpy of Component 1 is of a high value—see Figure 5a,c—depression
meltingof
of temperature
of the
the
the melting
melting is possible
melting temperature
temperature
temperature is only is if the
is possible
possible activity
possible only only coefficients
only ififif thethe
the activity of
activity both components
activity coefficients
coefficients
coefficients of of deviate
of both
both considerably
both components
components
components deviate deviate
deviate
of the melting temperature is possible only if the activity coefficients of both components deviate
from unity—i.e.,
considerably
considerably
considerably a(1)from < 0. Despite
from
from unity—i.e.,
unity—i.e.,
unity—i.e., the fact aaa (1)that
(1)
(1) <<
< 0.a Despite
0.
0. low
Despite
Despite melting the
the
the enthalpy
fact
fact
fact that
that
that value
aa
a low
low
low of melting
Component
melting
melting enthalpy2 leadsvalue
enthalpy
enthalpy to
value
value of
of
of
considerably from unity—i.e., a(1) < 0. Despite the fact that a low melting enthalpy value of
an even more
Component
Component pronounced
Component222leads leads
leadsto decrease
to
toanan
aneven
even in
evenmore the
more eutectic
morepronounced
pronounced temperature—as
pronounceddecrease decrease
decreasein in shown
inthethe
theeutectic in
eutectic Figure 5c,d—this
eutectictemperature—as
temperature—as
temperature—asshown occurs shown
shownin in
in
Component 2 leads to an even more pronounced decrease in the eutectic temperature—as shown in
only ifFigure
the activity
Figure
Figure 5c,d—this
5c,d—this
5c,d—this coefficients
occursof
occurs
occurs onlyComponent
only
only ifififthe
the
theactivity 1 arecoefficients
activity
activity significantly
coefficients
coefficients oflower
of
of Component
Component
Component than unity. 111are
are
aresignificantly
significantly
significantlylower lower
lowerthan than
than
Figure 5c,d—this occurs only if the activity coefficients of Component 1 are significantly lower than
Generally,
unity.
unity.
unity. low component activity coefficients lead to a significant depression of the eutectic
unity.
temperature of the
Generally,
Generally,
Generally, mixture,
low
low independent
low component
component
component of the coefficients
activity
activity
activity melting
coefficients
coefficients enthalpylead
lead to
lead values.
to
to However,
aaa significant
significant
significant when theof
depression
depression
depression melting
of
of the
the eutectic
the eutectic
eutectic
Generally, low component activity coefficients lead to a significant depression of the eutectic
enthalpy of Component
temperature
temperature
temperature of
ofthe
of the 1mixture,
is of a independent
themixture,
mixture, high
independent
independent value, only of
ofofthe
thea small
the melting
melting
melting depression
enthalpy
enthalpyvalues.
enthalpy is observed
values.
values. However,whenwhen
However,
However, its activity
when
when the
themelting
the melting
melting
temperature of the mixture, independent of the melting enthalpy values. However, when the melting
coefficients
enthalpy
enthalpy
enthalpy are ofequal
of
ofComponent (or close)111to
Component
Component isisunity
is of
of aaa high
of (a(1)value,
high
high = 0, orange
value,
value, onlyaaacurves
only
only small in
small depression
small Figure 5a,c).
depression
depression is
is In contrast,
is observed
observed
observed whenwhen
when
when its
its activity
its activity
activity
enthalpy of Component 1 is of a high value, only(1) a small depression is observed when its activity
Component 1 significantly
coefficients
coefficients
coefficients are
are equal
are equaldeviates
equal (or
(or close)
(or close)
close) fromto
to ideality,
to unity
unity
unity (a a===larger
(a
(a (1)
(1) 0,
0, orange
0, orange
orange e↵ectcurvescan bein
curves
curves inseen
in Figure
Figure
Figure even if Component
5a,c).
5a,c).
5a,c). In
InIn contrast,
contrast,
contrast, 2 when
when
when
coefficients are equal (or close) to unity (a(1) = 0, orange curves in Figure 5a,c). In contrast, when
behaves ideally (compare
Component
Component
Component an orange
111 significantly
significantly
significantly deviates
deviates
deviates curve from
from
from with the lowest
ideality,
ideality,
ideality, aaa larger
largerpoint
larger of acan
effect
effect
effect blue
can
can be
bebecurve
seen at
seen even
seen evena(2)ififif=Component
even 0). In
Component
Component 222
Component 1 significantly deviates from ideality, a larger effect can be seen(1) even if Component 2
addition, comparing
behaves
behavesideally
behaves ideally the
ideally(compare change
(compare
(comparean an in
anorangeeutectic
orange
orangecurve curvetemperature
curvewith with
withthe the depression
thelowest
lowest
lowestpoint point
pointof when
of
ofaaablue
bluea
bluecurve equals
curve
curveat at 30
ataaa ===0).
(2)
(2)
(2) and
0).
0).InIn 0 at
Inaddition,
addition,
addition,
behaves ideally (compare an orange curve with the lowest point of a blue (1) curve at a(2) = 0). In addition,
a(2) = comparing
30(the y-axis
comparing
comparing the
the intercepts
thechange
change
changein in of the blue
ineutectic
eutectic
eutectic and orange
temperature
temperature
temperature curves) when
depression
depression
depression and
when
when when
aaa a(2) equals
equals
equals
(1)equals
(1) 30
30
30 and
and
and 30000 and
at
at
at aaa (2)0=
(2)
(2) ==at30(the
30(the
30(the
comparing the change in eutectic temperature depression when a(1)(2) equals 30 and 0 at a(2) = 30(the
y-axis
y-axisintercepts
y-axis intercepts
interceptsof of
ofthe
the
theblue
blue
blueandand
andorange orange
orangecurves) curves)
curves)and and
andwhen when
when a
aa equals
equals
(2)equals
(2) 30
30
30 and
and
and 0
00 at
at
at aaa (1)=
(1)
(1) == 30
3030 (the
(the
(the sides
sides
sides
y-axis intercepts of the blue and orange curves) and when a(2) equals 30 and 0 at a(1) = 30 (the sides
 Molecules 2019, 24, x FOR PEER REVIEW 9 of 20
Molecules 2019, 24, 2334 9 of 19

of the blue curves) indicates a weaker contribution of the non-ideality of Component 2 to decreasing
the
a(1)eutectic temperature
= 30 (the sides of the(as
bluecompared with that
curves) indicates of Component
a weaker contribution 1). This
of theshows that compared
non-ideality with
of Component
Component 2, the non-ideality of Component 1 is more significant in the formation
2 to decreasing the eutectic temperature (as compared with that of Component 1). This shows that of deeper eutectic
mixtures.
compared with Component 2, the non-ideality of Component 1 is more significant in the formation of
Figure
deeper 6 shows
eutectic the effect of the non-ideality of the liquid phase on eutectic composition at
mixtures.
different melting
Figure 6 shows enthalpies. The
the e↵ect of thelow activity coefficients
non-ideality of the liquid of Component
phase on eutectic 1 composition
significantlyatshift the
di↵erent
eutectic point to a higher content of Component 1 (toward higher x1,e values), whereas an opposite
melting enthalpies. The low activity coefficients of Component 1 significantly shift the eutectic point to
effect is observed
a higher content offor the low 1activity
Component (towardcoefficients
higher x1,e of Component
values), whereas 2 an
(toward lower x1,e values).
Molecules 2019, 24, xopposite
FOR PEERe↵ect
REVIEW is observed
However,
for the lowthe effect coefficients
activity of the activity coefficients2 of
of Component Component
(toward lower 8x8181,e
is significantly largerthethan that of
,,xxxFOR
4, FOR
FORPEER
PEER
PEERREVIEW
REVIEW
REVIEW ofvalues).
of
of 20
20
20 However, e↵ect of the
Component 2 (see Figure 5c).
activity coefficients of Component 1 is significantly larger than that of Component 2 (see Figure 5c).

Figure 5. Effect of non-ideality of components on the normal

Effect
Effect
Effectof of
ofnon-ideality
non-ideality
non-idealityof
of Figure
ofcomponents
components
components 6. E↵ect
on
on
on theof
the
the activity
normalized
normalized
normalized coefficients and
depression
depression
depression melting
at
at
at enthalpies
eutectic
eutectic
Figure 6. Effect of activity coefficients and melting enthalpies of theeutectic temperature
temperature
temperature of the components
relative onthe
theeutectic
to the melting
components on eutectic mole of Component 2 assumin
temperature
mole
othe
the
themelting
melting
meltingtemperature
temperature
temperature fractionofofComponent
of
of
of Component
fractionComponentComponent11 assuming
Component 22
2 assuming
assuming
assuming assuming different
different
different
different di↵erent melting
melting
melting
melting melting enthalpy
enthalpy
enthalpy
enthalpy enthalpy values
values
values
values of
of
ofvalues of
components of components
components (a)
(a) (a) Dhm,1 = and
1 1 1 1
nts
ts (a)
nts (a)
(a) 32.64 and kJ
andmol and
and and Dhm,2 = 16.32(b) kJ mol (b) Dhm,1
(b)
(b) (b) = 12.00and kJ mol and Dhm,2 =
and
and and16.32
(c)kJ mol and
1 1 1
(c)
(c)
(c) (c) Dhm,1 = 32.64 (c) kJ mol
and
and
and and Dh m,2 =and 6.00 kJ mol
(d)
(d)
(d) (d) Dh m,1 =
(d) 12.00 kJ mol
and and Dh m,2
and= . The melting temp
1melting
and
and
and 6.00 kJ...mol
The
The . The melting
Themelting
melting temperatures
temperatures
.temperatures
temperatures
The melting of
of of Components
ofComponents
Components
Components
temperatures 122and
of11Components
1and
and
and are2set
2are
are are
set
set
to set
600to2K600
1 and K
and
are setand
to 300
300 K, respectively.
K, respectively.
600 K and 300 K,The parameter value is on
and
nd 300
and 300
300K,K,
K,respectively.
respectively. The
The parameter
Theparameter
respectively.respectively.
parameter value
Thevalue
value
value is
parameter isonly
is only
only
only aaaameasure
valuemeasure
measure
measure
is only for
foranon-ideality
for
for non-idealityforwith
non-ideality
non-ideality
measure with
with
withnononophysical
no
non-ideality withsignificance.
physical no physicalLegend:
significance. significance. a = 30;
(1) a(1) = 1
ignificance.
gnificance.
ignificance.Legend:
Legend:
Legend: (1)
aaaa ==== 30;
(1)
(1)
(1) 30; (1)
30; a(1) = aaaa30;==== 18;
(1)
(1)
(1) 18; (1)
18; a(1) =aaa18;
(1)
(1)
(1)==
== 12;
12;
12; (1)
12; a(1) =aaa 12;
(1)
(1)
(1)===
=0.0.
0.
0.
Legend: a(1) = 0.
Figure 5 demonstrates how solution non-ideality affect
demonstrates
demonstrates
demonstrateshow how
howsolution To
solution
solution summarize,
non-ideality
non-ideality
non-ideality the melting
affects
affects
affects properties
depression
depression
depression
To summarize, the melting properties as well as the activity coefficients at
at
at as
the
thewell
the as
eutectic
eutecticthe
eutectic activity
temperature
temperature
temperature
for coefficients
different melting ofenthalpies
of thecomponents
the components
of the can can
components. As in th
melting
melting
meltingenthalpies
enthalpies a↵ect
of
enthalpiesaffectof
ofthethe
the the position
components.
components.
components. of As
Asthe
As in
ineutectic
in the
the
the case
case
case point.
of
of
of the
the
theEutectic
ideal
ideal
ideal composition
mixture,
mixture,
mixture, the
the
the position of the eutectic point. Eutectic composition shifts to the side of the component thatthe lowershifts
lower
lower
melting the
the
theto the
enthalpy sideofof the component
Component 1, that
the larger the normaliz
lpy
lpy of
alpy of
of Component
Component
Component has
has 1,
1,
1,lower
the
the
lower activity
the larger
larger
larger
activity the
the
the coefficients
normalized
normalized
normalized
coefficients asasdepression
wellas
depression
depression
well asaalower
lower
at
at melting
at eutectic
eutectic
eutectic
melting enthalpy.
temperature.
temperature.
temperature.
When the
enthalpy. IfIfmelting
bothcomponents
both components
enthalpy of have similar 1 is of a low va
Component
have similar
lting
ting
ltingenthalpy
enthalpy
enthalpyof of activity
Component
Component
ofactivity
Component coefficients
111is
coefficients is
isof of aand
ofaand
alow
low
low melting
value—see
value—see
value—see
melting enthalpies,
Figure
Figure
enthalpies,Figure the eutectic
eutecticcomposition
5b,d—depression
5b,d—depression
5b,d—depression
the melting
composition of
of
ofthe
the
themight
pointmight occur
occurs,
occur ateven
a molar
at ratio
if Component
a molar of 1:1.
ratio of 1 behaves ideally
occurs,
occurs,
occurs,even even Low
evenifififComponent
Component
Component melting 1
11 enthalpies
behaves
behaves
behaves and
ideally
ideally
ideally activity
(a
(a
(a (1)=
(1)
(1) =
= coefficients
0;
0;
0; orange
orange
orange
1:1. Low melting enthalpies and activity coefficients of both components lead to the greatest of
curves
curves
curvesboth
in
in
incomponents
Figure
Figure
Figure
However,5b,d).
5b,d).
5b,d). lead
when to the
the greatest
melting depression
enthalpy of of
Component 1 is of a
en
en
enthe the
themelting
melting
meltingenthalpy the
enthalpy
enthalpy
depressioneutectic
ofof temperature
ofComponent
Component
Component
of the eutectic 111isis
isof of
of
of the
aaahigh
high mixture.
highvalue—see
temperature value—see
value—see The melting
Figure
Figure properties
Figure5a,c—depression
of the mixture. 5a,c—depression
5a,c—depression and
of the melting
The melting non-ideality
properties temperatureof the high melting
is possible
and non-ideality only if the activity
of the
ggtemperature
temperature
temperatureis is component
ispossible
possible
possible only
only
only haveif
if
if thea
the
the more
activity
activity
activity pronounced
coefficients
coefficients
coefficients e↵ect
of
of
of on
both
both
both eutectic temperature
components
components
components
high melting component have a more pronounced effect on eutectic temperature depression. Despite the fact deviate
deviate
deviate
considerably depression.
from Therefore,
unity—i.e., the
a (1) search
< 0.
from
from
from unity—i.e.,
unity—i.e., aaa(1)
unity—i.e.,Therefore, (1) <
(1) << 0.0.
0. Despite
Despite
Despite
the search the
the
theforfact
fact
newthat
fact that
that
deep aaa eutectic
low
low
low melting
melting
melting
mixtures enthalpy
enthalpy
enthalpy
should value
value
likely of
value
Component of
of based
be 2 leadsontothean identification
even more pronouncedof decrease i
leads
leads
leadsto to
toan
an
aneven
even
evenmoremore
more pronounced
pronounced
pronounced decrease
decrease
decrease in
in
in the
the
the eutectic
eutectic
eutectic temperature—as
temperature—as
temperature—as
compounds with low melting enthalpy, especially if they also possess a high melting temperature. shown
shown
shown
Figure in
in
in
5c,d—this occurs only if the activity coefficients of C
this
his
thisoccurs
occurs
occursonly
only
onlyifififthe
the
theactivity
activity
activitycoefficients
coefficients
coefficientsof of
ofComponent
Component
Component111are are
aresignificantly
significantly
significantlylower lower
lowerthan
unity. than
than
Generally, low component activity coefficients lead to
y,
y,
y,low low
lowcomponent
component
componentactivity activity
activitycoefficients
coefficients
coefficientslead lead
leadto to
toaaasignificant
significant
significantdepressiondepression
depressionof of
ofthe the
the eutectic
eutectic
eutectic of the mixture, independent of the melting enth
temperature
ff the
of the
themixture,
mixture,
mixture,independent
independent
independentof of
ofthe
the
themelting
melting
meltingenthalpy enthalpy
enthalpyvalues. values.
values.However,
However,
However,when when
whenthe the
the melting
melting
melting
enthalpy of Component 1 is of a high value, only a small d
Molecules 2019, 24, 2334 10 of 19

for new deep

Molecules eutectic
2019, 24, x FOR PEERmixtures
REVIEW should likely be based on the identification of compounds with 10 oflow
20
melting enthalpy, especially if they also possess a high melting temperature.
3.2. Parameter Study on Real Binary Systems
3.2. Parameter Study on Real Binary Systems
3.2.1. [Ch]Cl-Based Binary Systems
3.2.1. [Ch]Cl-Based Binary Systems
Choline chloride ([Ch]Cl) is a common constituent (HBA) of DESs. The melting properties of
Choline
[Ch]Cl are notchloride ([Ch]Cl)
available is a commonowing
experimentally constituent
to its(HBA)thermal of DESs. The melting
instability. properties of
The decomposition
[Ch]Cl
temperature of [Ch]Cl is around 575 K. Recently, Fernandez et al. [56] made an indirect temperature
are not available experimentally owing to its thermal instability. The decomposition assessment
ofof[Ch]Cl is around
the melting 575 K.ofRecently,
properties Fernandez etSLE
[Ch]Cl. Experimental al. [56]
datamade an indirect
in binary mixtures assessment
of [Ch]Cl with of thedifferent
melting
properties
ILs were of [Ch]Cl.
fitted to theExperimental
ideal solution SLE data in binary
model, and the mixtures
[Ch]Clofmelting[Ch]Cl with di↵erent
properties ILs calculated
were were fitted
tocorrespondingly.
the ideal solution Themodel,
valuesand the [Ch]Cl
obtained were melting properties
melting enthalpy were calculated
= 4.3 andcorrespondingly.
melting temperature The
1
values
= 597 obtained
. were melting enthalpy = 4.3 kJ mol and melting temperature = 597 K.
According
According to to Equation
Equation (2) (2) applied
appliedtotoananideal
idealsolution
solution ( ( Li = 1), the
), the values
values of ln(xof 1ln(x 1 ) plotted
) plotted as a
asfunction
a function of the reciprocal
of the reciprocal temperature ( temperature (1/T) should fit a straight line with the
) should fit a straight line with the y-axis intercept equal y-axis intercept
to to (Dhm,1 /(RTm,1
equal andslope
and)) the the slope
equal to to ( Dhm,1)./R).
equal TheThe corresponding
corresponding plot plot
forfor1010[Ch]Cl/IL
[Ch]Cl/IL
systems
systemsused usedfor forfitting
fittingthe
themelting
melting properties
properties ofof [Ch]Cl
[Ch]Cl is is shown
shown in in Figure
Figure 7a. 7a. TheThepoints
pointsrepresent
represent
the
theexperimental
experimentaldata dataforforall
all 10
10 [Ch]Cl/IL
[Ch]Cl/IL systems,
systems, and the line is is the
the fitted
fitted straight
straightline lineused
usedfor forthe
the
calculation
calculationofofthe the[Ch]Cl
[Ch]Clmelting
melting properties
properties given
given above. As can can bebe seen
seen from
fromFigure
Figure7a, 7a,the
thevalue
valueofof
determination
determinationcoefficient
coefficient (R (R22) indicates
indicates a rather good linear linear regression.
regression. However,
However,the thepoints
pointsfor forthe
the
individual binaries are systematically distributed
individual binaries are systematically distributed to the sides of the
sides of the fit. For example, all points for
For example, all points for
[C[C2 OHmim]Cl
2OHmim]Cl lie lie above
above thethe fit,
fit, whereas
whereas those for [BzCh]Cl
[BzCh]Cl lie lie below
below the the fit.
fit. ItItisisworth
worthnoting
notingthatthat
the y-axisshows
they-axis showslogarithmic
logarithmic valuesvalues andand the x-axis
x-axis reciprocal values; values; therefore,
therefore, aasmall smallshift
shiftfrom
fromthe the
fitted line corresponds to a significant di↵erence in the calculated T and x. As shown in Figure 7b, the
fitted line corresponds to a significant difference in the calculated T and x. As shown in Figure 7b,
the lines
lines fittedfitted
to five to single
five single binaries
binaries [Ch]Cl/IL
[Ch]Cl/IL systems
systems havehave different
di↵erent slopes slopes
and and intercepts;
intercepts; therefore,
therefore, they
they lead to different [Ch]Cl melting properties. The melting temperature
lead to di↵erent [Ch]Cl melting properties. The melting temperature is in the range 588.9–630.8 K and is in the range 588.9–630.8
K and melting
melting enthalpy enthalpy
is in theisrange
in the3.85–5.63
range 3.85–5.63
kJ mol kJ 1 , mol
being, being
1
withinwithin
around around
7% and 7%40%andfrom40% from the
the mean
meanrespectively.
value, value, respectively.See Table SeeS1 Table S1 in supplementary
in supplementary material material for detailed
for detailed results.results.

Figure (a) Solubility

Figure7.7. (a) Solubilityofofcholine
cholinechloride
chloride([Ch]Cl)
([Ch]Cl)
in in several
several ILsILs regressed
regressed to ato a linear
linear function.
function. (b)
(b) Solubility
Solubility of of [Ch]Cl
[Ch]Cl in selected
in selected ILs,ILs,
eacheach regressed
regressed to atolinear
a linear function.
function. DataData are taken
are taken fromfrom
[56][56].
.

The
Theideal
idealsolution
solutionmodel
modelaccounts
accountsforfor
the melting
the properties
melting propertiesof of
one component
one component only, ignoring
only, any
ignoring
dissimilarity of the second components in the mixture, which can be described by
any dissimilarity of the second components in the mixture, which can be described by activity activity coefficients.
Incoefficients.
order to analyze
In orderthe
to possible
analyze thenon-ideality of [Ch]Cl inofthe
possible non-ideality liquid
[Ch]Cl phase,
in the SLE
liquid of binary
phase, SLE ofmixtures
binary
ofmixtures
[Ch]Cl and ILs have
of [Ch]Cl and been modeled
ILs have assuming
been modeled di↵erent
assuming values values
different of [Ch]Cl melting
of [Ch]Cl enthalpy
melting using
enthalpy
the RK-polynomial with two parameters to estimate the activity coefficients. The
using the RK-polynomial with two parameters to estimate the activity coefficients. The values of values of [Ch]Cl
melting
[Ch]Cl temperature estimated
melting temperature from single
estimated from[Ch]Cl/IL mixturesmixtures
single [Ch]Cl/IL scatter within
scatter7% only, 7%
within andonly,
as shown
and asin
Section
shown 3.1.1 for the
in Section hypothetical
3.1.1 ideal system,
for the hypothetical idealthe absolute
system, values ofvalues
the absolute melting temperatures
of melting have on
temperatures
have on effect on the eutectic point properties. Hence, the value of 597 K estimated by Fernandes et
al. [56] has been used in the present work. Figure 8a shows the obtained results for the [Ch]Cl liquidus
Molecules 2019, 24, 2334 11 of 19

e↵ect on the eutectic point properties. Hence, the value of 597 K estimated by Fernandes et al. [56] has
Molecules 2019, 24, x FOR PEER REVIEW 11 of 20
been used in the present work. Figure 8a shows the obtained results for the [Ch]Cl liquidus line in the
[Ch]Cl/choline acetate [Ch][Ac] binary system as an example. Figure 8b represents the corresponding
line in the [Ch]Cl/choline acetate [Ch][Ac] binary system as an example. Figure 8b represents the
data for activity coefficients
corresponding of [Ch]Cl,
data for activity wherein
coefficients the points
of [Ch]Cl, correspond
wherein the pointsto the values calculated
correspond to the values from
thecalculated
experimentalfrom the experimental data using Equation (2) and the lines give the best fit using the RK-As
data using Equation (2) and the lines give the best fit using the RK-polynomial.
shown in Figure
polynomial. As8a, it is possible
shown in Figureto8a,
model the SLE to
it is possible data withthe
model rather
SLE good agreement
data with rather assuming di↵erent
good agreement
values of [Ch]Cl melting enthalpy, except in the cases of very low (1 kJ mol 1 ) values. However, as
assuming different values of [Ch]Cl melting enthalpy, except in the cases of very low (1 kJ mol 1)
seen fromHowever,
values. Figure 8b,asthe seenexperimental
from Figure activity coefficients of
8b, the experimental [Ch]Cl
activity in the liquid
coefficients phaseinchange
of [Ch]Cl as the
the liquid
melting
phase enthalpy
change as of the[Ch]Cl
meltingchanges.
enthalpyThis is because
of [Ch]Cl a variation
changes. in the melting
This is because enthalpy
a variation in theindicates
melting a
change in the
enthalpy value aofchange
indicates the reference state
in the value of chemical potential
the reference to which
state chemical the activity
potential coefficients
to which refer.
the activity
1
[Ch]Cl shows ideal
coefficients refer. behavior when the
[Ch]Cl shows melting
ideal behaviorenthalpy
when hasthe amelting
low value of 5 kJ mol
enthalpy has a ,low
a value
valueclose
of to
that
5 found by , a Fernandez
value close et to al.
that[56].
found by Fernandez et al. [56].

Figure
Figure8. 8.(a)
(a)Liquidus lineof
Liquidus line of[Ch]Cl
[Ch]Clinin a [Ch]Cl/choline
a [Ch]Cl/choline acetate
acetate binary
binary mixture
mixture modeledmodeled using
using the RK-the
RK-polynomial assuming di↵erent melting enthalpy values. (b) Activity coefficients of [Ch]Cl the
polynomial assuming different melting enthalpy values. (b) Activity coefficients of [Ch]Cl in in the
liquid phase
liquid phase calculated using
calculated the RK-polynomial
using the RK-polynomial assuming di↵erent
assuming melting
different enthalpy
melting values.values.
enthalpy Melting
temperature of [Ch]Cl of
Melting temperature is assumed to be 597to
[Ch]Cl is assumed K.be
Experimental data aredata
597 K. Experimental taken
arefrom from [56].
taken[56].

Results
Results for
forother
otherstudied
studied[Ch]Cl/IL
[Ch]Cl/ILsystems
systems can
can be found in in Figures
FiguresS1–3
S1–S3 inin
thethe supplementary
supplementary
material. Figure
material. Figure9 9shows
showsthetheabsolute
absolute average deviation(AAD)
average deviation (AAD)obtained
obtainedwhen
when screening
screening thethe melting
melting
enthalpy of [Ch]Cl in all studied [Ch]Cl/IL binary mixtures. The high error in modeling the
enthalpy of [Ch]Cl in all studied [Ch]Cl/IL binary mixtures. The high error in modeling the solubility of solubility
of [Ch]Cl
[Ch]Cl using using
a veryalow very low melting
melting (1 kJ mol(11 )kJindicates
enthalpyenthalpy mol 1) indicates
that suchthat such
a value a value
might might be
be unreasonable.
At the same time, experimental data can be modeled with low AAD when the value of Dhthe
unreasonable. At the same time, experimental data can be modeled with low AAD when value
m,1 is in the
of ∆hm,1 is in the range
range 5–40 kJ mol 1 . 5–40 kJ mol 1.

To further explore [Ch]Cl behavior in the liquid phase, the value of the activity coefficients of
[Ch]Cl at the experimental eutectic point can be analyzed (see Figure 10). As seen from the figure, at
low melting enthalpy values of [Ch]Cl (<5 kJ mol 1 ), the activity coefficients are close to one in all
previously studied binary systems (as found by Fernandez et al. [56]). However, at higher [Ch]Cl
melting enthalpies (>20 kJ mol 1 ), the values of the activity coefficients change (see Figure 10). As seen
in Figure 11, the molecular structures of ILs significantly di↵er from each other and from [Ch]Cl, that
would result in di↵erent hydrophobicities. Owing to the long alkyl chains in [P4444]+ and [N4444]+
cations, these cations are more hydrophobic than the [Ch]+ cation; hence, the interactions in solution
would di↵er from those of the pure components. At low values of [Ch]Cl melting enthalpy, the
di↵erence in the interactions is concealed, leading to quasi-ideal behavior. However, at higher values,
the di↵erence in molecular size and hydrophobicity appears obvious.
Molecules 24,2334
2019,24,
Molecules2019, x FOR PEER REVIEW 12
12ofof1920

Figure 9. AAD between calculated and experimental liquidus temperatures for binary mixtures of
[Ch]Cl and ILs assuming different melting enthalpies for [Ch]Cl.

To further explore [Ch]Cl behavior in the liquid phase, the value of the activity coefficients of
[Ch]Cl at the experimental eutectic point can be analyzed (see Figure 10). As seen from the figure, at
low melting enthalpy values of [Ch]Cl (<5 kJ mol 1), the activity coefficients are close to one in all
previously studied binary systems (as found by Fernandez et al. [56]). However, at higher [Ch]Cl
melting enthalpies (>20 kJ mol 1), the values of the activity coefficients change (see Figure 10). As seen
in Figure 11, the molecular structures of ILs significantly differ from each other and from [Ch]Cl, that
would result in different hydrophobicities. Owing to the long alkyl chains in [P4444]+ and [N4444]+
cations, these cations are more hydrophobic than the [Ch]+ cation; hence, the interactions in solution
would differ from those of the pure components. At low values of [Ch]Cl melting enthalpy, the
difference in the interactions is concealed, experimental
leading to quasi-idealtemperatures
behavior. However, at higher values,
Figure9.9.AAD
Figure AADbetween
betweencalculated
calculatedand
and experimentalliquidus
liquidus temperaturesfor
forbinary
binarymixtures
mixturesofof
the difference in molecular size and hydrophobicity appears obvious.
[Ch]Cl and ILs assuming di↵erent melting enthalpies for [Ch]Cl.
[Ch]Cl and ILs assuming different melting enthalpies for [Ch]Cl.

To further explore [Ch]Cl behavior in the liquid phase, the value of the activity coefficients of
[Ch]Cl at the experimental eutectic point can be analyzed (see Figure 10). As seen from the figure, at
low melting enthalpy values of [Ch]Cl (<5 kJ mol 1), the activity coefficients are close to one in all
previously studied binary systems (as found by Fernandez et al. [56]). However, at higher [Ch]Cl
melting enthalpies (>20 kJ mol 1), the values of the activity coefficients change (see Figure 10). As seen
in Figure 11, the molecular structures of ILs significantly differ from each other and from [Ch]Cl, that
would result in different hydrophobicities. Owing to the long alkyl chains in [P4444]+ and [N4444]+
cations, these cations are more hydrophobic than the [Ch]+ cation; hence, the interactions in solution
would differ from those of the pure components. At low values of [Ch]Cl melting enthalpy, the
difference in the interactions is concealed, leading to quasi-ideal behavior. However, at higher values,
the difference in molecular size and hydrophobicity appears obvious.

Figure 10. The

Figure 10. The activity
activity coefficients
coefficients of
of [Ch]Cl
[Ch]Cl at
at the
the experimental
experimental eutectic
eutectic point
point estimated
estimated assuming
assuming
di↵erent
Molecules 2019,melting
different melting enthalpies
enthalpies
24, x FOR of
of [Ch]Cl.
PEER REVIEW [Ch]Cl. 13 of 20

Figure 10. The activity coefficients of [Ch]Cl at the experimental eutectic point estimated assuming
different melting enthalpies of [Ch]Cl.

Figure 11.Chemical
Figure11. Chemical names and structures
names and structuresof
ofthe
the[Ch]Cl
[Ch]Cland
andILs
ILsstudied.
studied.

Despite the fact that predicting [Ch]Cl solubility in ILs is possible by assuming an ideal solubility
model with a low melting enthalpy of [Ch]Cl, system-specific information such as the interactions of
[Ch]Cl with different ILs can be hidden by the system ideality assumption. Martins et al. [49] also
suggested that the deep depression of the eutectic temperature in [Ch]Cl-based mixtures is due to the
low value of its melting enthalpy. Assessing these important features of such systems is an advantage
Molecules 2019, 24, 2334 13 of 19

Despite the fact that predicting [Ch]Cl solubility in ILs is possible by assuming an ideal solubility
model with a low melting enthalpy of [Ch]Cl, system-specific information such as the interactions of
[Ch]Cl with di↵erent ILs can be hidden by the system ideality assumption. Martins et al. [49] also
suggested that the deep depression of the eutectic temperature in [Ch]Cl-based mixtures is due to the
low value of its melting enthalpy. Assessing these important features of such systems is an advantage
of modeling the non-ideality in the liquid phase. An apparent changing of the melting enthalpies of
the components leads to di↵erent activity coefficients in the liquid phase. This leads to a di↵erent
conclusion about the intermolecular interactions and the behavior of the components in the system,
which might be beneficial in decreasing the complexity of SLE modeling. However, a discussion of
SLE results might be impractical based on this empirical approach.

3.2.2. Quaternary Ammonium Chloride Salts/Fatty Acids DES

The SLE of binary mixtures of di↵erent quaternary ammonium chlorides and fatty acids have
been studied in detail [46]. These systems are considered DESs owing to strong negative deviation
from ideality in quaternary ammonium chloride liquidus lines. Pontes et al. [46] used PC-SAFT to
model liquid phase non-ideality. The modeling results are in good agreement with experimental data.
The melting properties of these salts reported in Pontes et al. [46] are summarized in supplementary
materials. Pontes et al. [46] confirmed the decomposition of [N1111 ]Cl, and its melting properties were
estimated using SLE data and activity coefficients calculated using COSMO-RS. For the other two salts
namely, [N2222 ]Cl and [N3333 ]Cl, the melting properties were measured in Büchi apparatus and DSC.
However, in earlier literature, these salts were reported to decompose at temperatures below their
melting temperatures [57,58]. As a result of decomposition, the melting enthalpy measured by DSC for
the salts might represent the heat of decomposition and not the melting enthalpy of these salts.
In this work, the di↵erent melting enthalpy values of quaternary ammonium chloride are screened
and the activity coefficients are modeled using the Redlich–Kister polynomial with three parameters.
The [N3333 ]Cl/palmitic acid system shows the strongest negative deviation from ideality compared
with the other systems studied in Pontes et al. [46]; therefore, its SLE data are analyzed here in detail
(see Figure
Molecules 12).
2019, 24, Results forREVIEW
x FOR PEER the other systems can be found in Figures S5–S7 in supplementary material.
14 of 20

Figure 12. (a) [N3333 ]Cl liquidus line for a [N3333 ]Cl/palmitic acid binary mixture modeled using
Figure 12. (a) [N3333]Cl liquidus line for a [N3333]Cl/palmitic acid binary mixture modeled using the
the Redlich–Kister polynomial with three parameters assuming di↵erent melting enthalpy values.
Redlich–Kister polynomial with three parameters assuming different melting enthalpy values. (b)
(b) Activity coefficients of [N3333 ]Cl in the liquid phase calculated using the Redlich–Kister polynomial
Activity coefficients of [N3333]Cl in the liquid phase calculated using the Redlich–Kister polynomial
assuming di↵erent melting enthalpy values. Experimental data taken from [46].
assuming different melting enthalpy values. Experimental data taken from [46].
As can be seen in Figure 12a, the liquidus lines of [N3333 ]Cl estimated using the RK-polynomials
As can be seen in Figure 12a, the liquidus lines of [N3333]Cl estimated using the RK-polynomials
are in good agreement with experimental data for any melting enthalpy assumed, except for the very
are in good agreement with experimental data for any melting enthalpy assumed, except for the very
low melting enthalpy (1 kJ mol 1 ). As for [Ch]Cl/IL systems, the melting enthalpy of 5 kJ mol 1 leads
low melting enthalpy (1 kJ mol ). As for [Ch]Cl/IL systems, the melting enthalpy of 5 kJ mol 1 leads
1
to ideal solubility for [N3333 ]Cl (see the yellow curve in Figure 12b). However, the larger the value
to ideal solubility for [N3333]Cl (see the yellow curve in Figure 12b). However, the larger the value of
of melting enthalpy, the stronger the negative deviations are from ideality required to reproduce
melting enthalpy, the stronger the negative deviations are from ideality required to reproduce the
experimental data. Therefore, assuming a very high melting enthalpy would overestimate the non-
ideality, making it rather difficult to account for using well-established and relatively simple-to-use
activity coefficient models.
Table 1 shows the eutectic point properties for all quaternary ammonium chloride/fatty acid
binary mixtures studied in [46]. Along with the experimental eutectic composition and temperature
Molecules 2019, 24, 2334 14 of 19

the experimental data. Therefore, assuming a very high melting enthalpy would overestimate the
non-ideality, making it rather difficult to account for using well-established and relatively simple-to-use
activity coefficient models.
Table 1 shows the eutectic point properties for all quaternary ammonium chloride/fatty acid
binary mixtures studied in [46]. Along with the experimental eutectic composition and temperature
(xe,acid and Te ), the table shows the normalized di↵erence between the eutectic temperature predicted
by PC-SAFT in [46] and the melting temperature of the acid (DTe ). As can be seen from Table 1,
the largest depression in melting temperature of the mixture at the eutectic point relative to the
melting temperature of the acid is for [N3333 ]Cl/palmitic acid system (DTe = 0.1096). In addition, the
[N3333 ]Cl/palmitic acid system has the eutectic composition with the highest salt mole fraction, as
compared with any other quaternary ammonium chloride/fatty acid systems (xe,salt = 0.517). This
seems to be due to the strong interactions between [N3333 ]Cl and palmitic acid, as was shown in [46]
(see also the parameter study section for the link between activity coefficients and eutectic composition).
As seen from Table 1, DTe values show that quaternary ammonium chlorides/ fatty acids DESs do not
show deep eutectic behavior based on the normalized depression at eutectic point. However, based
on the definition of Kollau et al. [47] and Martins et al. [49] for deep eutectic solvents, these systems
would be considered as deep eutectic systems. The small depression at eutectic temperature relative
to the acid melting temperature might be a result of the high melting enthalpies of the fatty acids.
As shown in the parameter study, melting properties of components a↵ect the depression at eutectic
point in a larger extent compared to the eutectic composition. In contrast, activity coefficient values
significantly shift the eutectic composition to the side of the component with low activity coefficients.
Therefore, the eutectic composition gives an indication for the component with the larger negative
deviation from ideality.

Table 1. Comparison of eutectic composition, eutectic temperature, and the normalized di↵erence
between the eutectic temperature estimated using PC-SAFT and the melting temperature of the acids.
Data are taken from [46].

[N1111 ]Cl [N2222 ]Cl [N3333 ]Cl

xe,acid Te DTe xe,acid Te DTe xe,acid Te DTe
Capric acid 0.808 299.91 0.0161 0.614 297.40 0.0247 0.710 288.46 0.0565
Lauric acid 0.630 303.01 0.0478 0.640 301.44 0.0532 0.644 297.05 0.0688
Myristic acid 0.712 321.65 0.0167 0.658 317.88 0.0288 0.571 304.54 0.0738
Palmitic acid 0.615 328.51 0.0254 0.607 318.78 0.0567 0.483 303.56 0.1096
Stearic acid 0.634 337.89 0.0171 0.556 317.02 0.0841 0.562 322.82 0.0646

To summarize, changing the melting enthalpy value of quaternary ammonium chloride changes
the activity coefficient values and thus hides or increases the actual complexity of the systems.
Eutectic systems with a high deviation from ideality do not always have a large depression at eutectic
temperature; rather, they would have a large shift in the eutectic composition away from pure
components sides.

4. Method
To study the e↵ect of melting properties and intermolecular interactions on eutectic point
properties, a hypothetical binary system is assumed. The system is considered to be either ideal
( Li = 1) in order to only study the e↵ect of melting properties, or non-ideal (with negative deviations
from ideality; Li < 1) to consider the e↵ect of intermolecular interactions. Di↵erent values of melting
enthalpy and melting temperature are screened for each of the components. In some cases, the melting
enthalpy values have been estimated using the Walden rule. The Walden rule states that the melting
Molecules 2019, 24, 2334 15 of 19

entropy of rigid molecules is constant, with a value of 54.4 J mol 1 K 1

[52]. The melting enthalpy can
thus be calculated as
Dhm = 54.4 Tm (3)

To model the non-ideality of components, the activity coefficients of components in the liquid phase
are calculated using Redlich–Kister (RK) polynomial [51]

RT ln i = a(i) x2j + b(i) x3j + c(i) x4j . . . . (4)

When the activity coefficients are screened (Section 3.1.2), only the first term with one parameter is
used. In this case, when parameter a(i) is equal to zero, the ideal case is described; whereas a(i) < 0
describes favored interactions between unlike molecules in the liquid phase—i.e., negative deviation
from ideality. Positive deviation from ideality leads to a higher liquidus temperature compared with
the ideal case; therefore, values of parameter a(i) > 0 are not considered in this work. The parameters
are empirical and do not provide any physical significance. They are just used as a measure of di↵erent
values of activity coefficients of components.
For modeling the non-ideality in [Ch]Cl/ILs binary systems (Section 3.2.1), two terms with two
parameters are used. Binary mixtures of quaternary ammonium chlorides with fatty acids (Section 3.2.2)
are more complex; hence, three terms with three parameters are used. In the latter two cases, for a given
melting enthalpy value, the RK-parameters are fitted to the SLE experimental data by minimizing the
objective function
X⇣ ⌘
cal exp 2
OF = 1 1
(5)

where cal
1
is the calculated activity coefficient of Component 1 in the liquid phase using the
exp
RK-polynomial and 1 is the activity coefficient of Component 1 in the liquid phase estimated
using Equation (2) and measured SLE taken from the literature.
To evaluate the quality of the calculated SLE data, the AAD between the calculated (Tical ) and
exp
experimental (Ti ) liquidus temperatures is calculated as

N
1 X cal exp
AAD = Ti Ti (6)
N
i

5. Conclusions
The parameter study carried out in the present work shows that (i) obtaining the accurate melting
properties of pure constituents and (ii) adequate modeling of non-ideality in the liquid phase are
essential for prediction of the SLE in DESs. Deep eutectics can be a result of low melting enthalpies,
low activity coefficients, or both. When analyzing the modeled SLE diagrams, the estimate of eutectic
composition can provide a better indication of system non-ideality. Components with lower activity
coefficient values have a higher mole fraction at the eutectic point. In systems of two components
with similar melting enthalpies and similar activity coefficients, the eutectic composition is observed
at around a 1:1 molar ratio of the two components. Therefore, designing DES systems should
simultaneously take into account melting properties as well as interactions between unlike molecules.
However, strong negative deviation from ideality is not the only reason for DES formation; substances
with low melting enthalpies could also form rather deep eutectics.
Recent work on the SLE of [Ch]Cl-based DES systems has reported that under an assumption of a
low melting enthalpy value (4.3 kJ mol 1 ), [Ch]Cl behaves quasi-ideally in most such systems. However,
molecular experiments and simulations in the recent literature have revealed that [Ch]Cl-based systems
behave in a complex manner. As demonstrated in the present study, the solubility of [Ch]Cl in di↵erent
ILs can be adequately modeled assuming di↵erent melting enthalpy values. Small values lead to a
nearly ideal behavior in the liquid phase, whereas it is necessary to account for activity coefficients
at higher melting enthalpy values. Owing to the absence of any experimental data on the melting
Molecules 2019, 24, 2334 16 of 19

enthalpy of [Ch]Cl for comparison, it is not possible to determine which of the values is true. However,
owing to the di↵erences in size and hydrophobicity of the studied ILs, di↵erences in their interactions
with [Ch]Cl should be expected. Thus, the low melting enthalpy value may conceal the complexity
of interactions in the liquid phase and should be used with caution. Therefore, modeling the SLE of
[Ch]Cl-based DESs remains unclear; more research is necessary to fully understand the behavior of
these systems.
For DES systems with strong intermolecular interactions between unlike molecules, the depression
in melting temperature at the eutectic point is not related only to activity coefficient values. The
studied quaternary ammonium chloride/fatty acid systems show strong negative deviation from
ideality with small depression at eutectic point relative to the acid melting temperature. At the same
time, the systems show a shift in the eutectic composition to the side of the salt. Therefore, the
eutectic composition can provide an indication of strong interactions—e.g., hydrogen bonds—between
DES components.
From the application viewpoint when designing a DES, no strict definition of a DES system is
necessary; rather, a better understanding of the system behavior and intermolecular interactions,
reliable data for the SLE of mixtures and pure components, and an assessment of system suitability
for an actual application are required. In fact, any simple eutectic mixture with a large depression in
eutectic temperature relative to the melting temperatures of the pure constituents sufficient to form
a liquid at operating temperature could be referred to as a DES (or NADES). In searching for new
DESs, high melting compounds with low melting enthalpies—which have strong interactions with
the second component in the liquid mixture but not in its pure solid state—should be sought. The
non-ideality of the liquid phase can thereby be directly estimated using well-developed and easy-to-use
thermodynamic tools such as activity coefficient models and equations of state.

Supplementary Materials: The following are available online at http://www.mdpi.com/1420-3049/24/12/2334/s1.

Figure S1: Liquidus line and activity coefficients in the liquid phase of [Ch]Cl in a binary mixture of [Ch]Cl and
(a) [Ch][Ac], (b)[Ch][Prop], and (c)[Ch][Buta], modeled using Redlich–Kister polynomial. Figure S2: Liquidus
line and activity coefficients in the liquid phase of [Ch]Cl in a binary mixture of [Ch]Cl and (a) [Ch][NTf2 ], (b)
[BzCh]Cl, (c) [C2 mim]Cl, and (d) [C2 OHmim]Cl, modeled using Redlich–Kister polynomial. Figure S3: Liquidus
line and activity coefficients in the liquid phase of [Ch]Cl in a binary mixture of [Ch]Cl and (a) [C4 mpyr]Cl, (b)
[N4444 ]Cl, and (c) [P4444 ]Cl, modeled using Redlich–Kister polynomial. Figure S4: Chemical names and structures
of [Ch]Cl and ILs studied. Figure S5: Liquidus line and activity coefficients of [N1111 ]Cl in a binary mixture of
[N1111 ]Cl and (a) capric acid, (b) lauric acid, (c) myristic acid, (d) palmitic acid, and (e) stearic acid, modeled using
the Redlich–Kister polynomial with three parameters assuming di↵erent melting enthalpy values. Figure S6:
Liquidus line and activity coefficients of [N2222 ]Cl in a binary mixture of [N2222 ]Cl and (a) capric acid, (b) lauric
acid, (c) myristic acid, (d) palmitic acid, and (e) stearic acid, modeled using the Redlich–Kister polynomial with
three parameters assuming di↵erent melting enthalpy values. Figure S7: Liquidus line and activity coefficients
of [N3333 ]Cl in a binary mixture of [N3333 ]Cl and (a) capric acid, (b) lauric acid, (c) myristic acid, (d) palmitic
acid, and (e) stearic acid, modeled using the Redlich–Kister polynomial with three parameters assuming di↵erent
melting enthalpy values. Table S1: Melting properties of [Ch]Cl obtained from linear regression of the solubility of
[Ch]Cl in di↵erent ILs. Table S2: Empirical parameters of RK-polynomial for calculating activity coefficients of
salts in binary mixtures of quaternary ammonium chloride salts and capric acid. Table S3: Empirical parameters of
RK-polynomial for calculating activity coefficients of salts in binary mixtures of quaternary ammonium chloride
salts and lauric acid. Table S4: Empirical parameters of RK-polynomial for calculating activity coefficients of salts
in binary mixtures of quaternary ammonium chloride salts and myristic acid. Table S5: Empirical parameters of
RK-polynomial for calculating activity coefficients of salts in binary mixtures of quaternary ammonium chloride
salts and palmitic acid. Table S6: Empirical parameters of RK-polynomial for calculating activity coefficients of
salts in binary mixtures of quaternary ammonium chloride salts and stearic acid. Table S7: Melting properties of
quaternary ammonium chloride and fatty acids.
Author Contributions: Conceptualization, generating data, analyzing the results and preparing the manuscript,
A.A.; revising and discussing the manuscript, L.M.; supervising, revising and discussing the manuscript, M.M.
Funding: This work was supported by the German Research Foundation (DFG) and the Technical University of
Munich (TUM) in the framework of the Open Access Publishing Program.
Conflicts of Interest: The authors declare no conflict of interest.
Molecules 2019, 24, 2334 17 of 19

References
1. Abbott, A.P.; Capper, G.; Davies, D.L.; Rasheed, R.K.; Tambyrajah, V. Novel solvent properties of choline
chloride/urea mixtures. Chem. Commun. 2003, 70–71. [CrossRef]
2. Tang, B.; Row, K. Recent developments in deep eutectic solvents in chemical sciences. Mon. Chem. 2013, 144,
1427–1454. [CrossRef]
3. Zhang, Q.; De Oliveira Vigier, K.; Royer, S.; Jerome, F. Deep eutectic solvents: Syntheses, properties and
applications. Chem. Soc. Rev. 2012, 41, 7108–7146. [CrossRef] [PubMed]
4. Smith, E.L.; Abbott, A.P.; Ryder, K.S. Deep Eutectic Solvents (DESs) and Their Applications. Chem. Rev. 2014,
114, 11060–11082. [CrossRef] [PubMed]
5. Liu, P.; Hao, J.-W.; Mo, L.-P.; Zhang, Z.-H. Recent advances in the application of deep eutectic solvents as
sustainable media as well as catalysts in organic reactions. RSC Adv. 2015, 5, 48675–48704. [CrossRef]
6. Abbott, A.P.; Ahmed, E.I.; Prasad, K.; Qader, I.B.; Ryder, K.S. Liquid pharmaceuticals formulation by eutectic
formation. Fluid Phase Equilib. 2017, 448, 2–8. [CrossRef]
7. Aroso, I.M.; Craveiro, R.; Rocha, Â.; Dionísio, M.; Barreiros, S.; Reis, R.L.; Paiva, A.; Duarte, A.R.C. Design
of controlled release systems for THEDES—Therapeutic deep eutectic solvents, using supercritical fluid
technology. Int. J. Pharm. 2015, 492, 73–79. [CrossRef] [PubMed]
8. Aroso, I.M.; Silva, J.C.; Mano, F.; Ferreira, A.S.; Dionísio, M.; Sá-Nogueira, I.; Barreiros, S.; Reis, R.L.; Paiva, A.;
Duarte, A.R.C. Dissolution enhancement of active pharmaceutical ingredients by therapeutic deep eutectic
systems. Eur. J. Pharm. Sci. 2016, 98, 57–66. [CrossRef]
9. Hayyan, A.; Ali Hashim, M.; Mjalli, F.S.; Hayyan, M.; AlNashef, I.M. A novel phosphonium-based deep
eutectic catalyst for biodiesel production from industrial low grade crude palm oil. Chem. Eng. Sci. 2013, 92,
81–88. [CrossRef]
10. Hayyan, A.; Hashim, M.A.; Hayyan, M.; Mjalli, F.S.; AlNashef, I.M. A novel ammonium based eutectic
solvent for the treatment of free fatty acid and synthesis of biodiesel fuel. Ind. Crop. Prod. 2013, 46, 392–398.
[CrossRef]
11. Hadj-Kali, M.K.; Mulyono, S.; Hizaddin, H.F.; Wazeer, I.; El-Blidi, L.; Ali, E.; Hashim, M.A.; AlNashef, I.M.
Removal of Thiophene from Mixtures with n-Heptane by Selective Extraction Using Deep Eutectic Solvents.
Ind. Eng. Chem. Res. 2016, 55, 8415–8423. [CrossRef]
12. García, A.; Rodríguez-Juan, E.; Rodríguez-Gutiérrez, G.; Rios, J.J.; Fernández-Bolaños, J. Extraction of
phenolic compounds from virgin olive oil by deep eutectic solvents (DESs). Food Chem. 2016, 197, 554–561.
[CrossRef] [PubMed]
13. Peng, X.; Duan, M.-H.; Yao, X.-H.; Zhang, Y.-H.; Zhao, C.-J.; Zu, Y.-G.; Fu, Y.-J. Green extraction of five target
phenolic acids from Lonicerae japonicae Flos with deep eutectic solvent. Sep. Purif. Technol. 2016, 157,
249–257. [CrossRef]
14. Rodriguez, N.R.; Ferre Guell, J.; Kroon, M.C. Glycerol-Based Deep Eutectic Solvents as Extractants for
the Separation of MEK and Ethanol via Liquid–Liquid Extraction. J. Chem. Eng. Data 2016, 61, 865–872.
[CrossRef]
15. Zhang, H.; Wang, Y.; Xu, K.; Li, N.; Wen, Q.; Yang, Q.; Zhou, Y. Ternary and binary deep eutectic solvents as
a novel extraction medium for protein partitioning. Anal. Methods 2016, 8, 8196–8207. [CrossRef]
16. Pena-Pereira, F.; Namieśnik, J. Ionic Liquids and Deep Eutectic Mixtures: Sustainable Solvents for Extraction
Processes. Chem. Sus. Chem. 2014, 7, 1784–1800. [CrossRef] [PubMed]
17. Roehrer, S.; Bezold, F.; Garcia, E.M.; Minceva, M. Deep eutectic solvents in countercurrent and centrifugal
partition chromatography. J. Chromatogr. A 2016, 1434, 102–110. [CrossRef]
18. Bezold, F.; Weinberger, M.E.; Minceva, M. Assessing solute partitioning in deep eutectic solvent-based
biphasic systems using the predictive thermodynamic model COSMO-RS. Fluid Phase Equilib. 2017, 437,
23–33. [CrossRef]
19. Bezold, F.; Weinberger, M.E.; Minceva, M. Computational solvent system screening for the separation of
tocopherols with centrifugal partition chromatography using deep eutectic solvent-based biphasic systems.
J. Chromatogr. A 2017, 1491, 153–158. [CrossRef]
20. Bezold, F.; Minceva, M. A water-free solvent system containing an L-menthol-based deep eutectic solvent for
centrifugal partition chromatography applications. J. Chromatogr. A 2019, 1587, 166–171. [CrossRef]
Molecules 2019, 24, 2334 18 of 19

21. Atilhan, M.; Aparicio, S. Behavior of Deep Eutectic Solvents under External Electric Fields: A Molecular
Dynamics Approach. J. Phys. Chem. B 2017, 121, 221–232. [CrossRef] [PubMed]
22. Das, A.; Das, S.; Biswas, R. Fast fluctuations in deep eutectic melts: Multi-probe fluorescence measurements
and all-atom molecular dynamics simulation study. Chem. Phys. Lett. 2013, 581, 47–51. [CrossRef]
23. Das, A.; Das, S.; Biswas, R. Density relaxation and particle motion characteristics in a non-ionic deep eutectic
solvent (acetamide + urea): Time-resolved fluorescence measurements and all-atom molecular dynamics
simulations. J. Chem. Phys. 2015, 142, 034505. [CrossRef] [PubMed]
24. Ferreira, E.S.C.; Voroshylova, I.V.; Pereira, C.M.; Cordeiro, M.N.D.S. Improved Force Field Model for the
Deep Eutectic Solvent Ethaline: Reliable Physicochemical Properties. J. Phys. Chem. B 2016, 120, 10124–10137.
[CrossRef] [PubMed]
25. García, G.; Atilhan, M.; Aparicio, S. The impact of charges in force field parameterization for molecular
dynamics simulations of deep eutectic solvents. J. Mol. Liq. 2015, 211, 506–514. [CrossRef]
26. Hammond, O.S.; Bowron, D.T.; Edler, K.J. Liquid structure of the choline chloride-urea deep eutectic solvent
(reline) from neutron di↵raction and atomistic modelling. Green Chem. 2016, 18, 2736–2744. [CrossRef]
27. Morrow, T.I.; Maginn, E.J. Density, local composition and di↵usivity of aqueous choline chloride solutions:
A molecular dynamics study. Fluid Phase Equilib. 2004, 217, 97–104. [CrossRef]
28. Perkins, S.L.; Painter, P.; Colina, C.M. Experimental and computational studies of choline chloride-based
deep eutectic solvents. J. Chem. Eng. Data 2014, 59, 3652–3662. [CrossRef]
29. Sun, H.; Li, Y.; Wu, X.; Li, G. Theoretical study on the structures and properties of mixtures of urea and
choline chloride. J. Mol. Model. 2013, 19, 2433–2441. [CrossRef] [PubMed]
30. Zahn, S. Deep eutectic solvents: Similia similibus solvuntur? Phys. Chem. Chem. Phys. 2017, 19, 4041–4047.
[CrossRef] [PubMed]
31. García, G.; Atilhan, M.; Aparicio, S. An approach for the rationalization of melting temperature for deep
eutectic solvents from DFT. Chem. Phys. Lett. 2015, 634, 151–155. [CrossRef]
32. Gilmore, M.; Moura, L.M.; Turner, A.H.; Swadźba-Kwaśny, M.; Callear, S.K.; McCune, J.A.; Scherman, O.A.;
Holbrey, J.D. A comparison of choline:urea and choline:oxalic acid deep eutectic solvents at 338 K. J. Chem.
Phys. 2018, 148, 193823. [CrossRef] [PubMed]
33. Ashworth, C.R.; Matthews, R.P.; Welton, T.; Hunt, P.A. Doubly ionic hydrogen bond interactions within the
choline chloride-urea deep eutectic solvent. Phys. Chem. Chem. Phys. 2016, 18, 18145–18160. [CrossRef]
[PubMed]
34. Stefanovic, R.; Ludwig, M.; Webber, G.B.; Atkin, R.; Page, A.J. Nanostructure, hydrogen bonding and
rheology in choline chloride deep eutectic solvents as a function of the hydrogen bond donor. Phys. Chem.
Chem. Phys. 2017, 19, 3297–3306. [CrossRef] [PubMed]
35. Mainberger, S.; Kindlein, M.; Bezold, F.; Elts, E.; Minceva, M.; Briesen, H. Deep eutectic solvent formation:
A structural view using molecular dynamics simulations with classical force fields. Mol. Phys. 2017, 115,
1309–1321. [CrossRef]
36. Fischer, V. Properties and Applications of Deep Eutectic Solvents and Low-Melting Mixtures Dissertation.
Ph.D. Thesis, Universität Regensburg, Regensburg, Germany, 2015.
37. Abbott, A.P.; Boothby, D.; Capper, G.; Davies, D.L.; Rasheed, R.K. Deep Eutectic Solvents Formed between
Choline Chloride and Carboxylic Acids: Versatile Alternatives to Ionic Liquids. J. Am. Chem. Soc. 2004, 126,
9142–9147. [CrossRef] [PubMed]
38. Morrison, H.G.; Sun, C.C.; Neervannan, S. Characterization of thermal behavior of deep eutectic solvents
and their potential as drug solubilization vehicles. Int. J. Pharm. 2009, 378, 136–139. [CrossRef]
39. Kareem, M.A.; Mjalli, F.S.; Hashim, M.A.; AlNashef, I.M. Phosphonium-Based Ionic Liquids Analogues and
Their Physical Properties. J. Chem. Eng. Data 2010, 55, 4632–4637. [CrossRef]
40. Guo, W.; Hou, Y.; Ren, S.; Tian, S.; Wu, W. Formation of Deep Eutectic Solvents by Phenols and Choline
Chloride and Their Physical Properties. J. Chem. Eng. Data 2013, 58, 866–872. [CrossRef]
41. Meng, X.; Ballerat-Busserolles, K.; Husson, P.; Andanson, J.-M. Impact of water on the melting temperature
of urea + choline chloride deep eutectic solvent. New J. Chem. 2016, 40, 4492–4499. [CrossRef]
42. Silva, L.P.; Fernandez, L.; Conceição, J.H.F.; Martins, M.A.R.; Sosa, A.; Ortega, J.; Pinho, S.P.; Coutinho, J.A.P.
Design and Characterization of Sugar-Based Deep Eutectic Solvents Using Conductor-like Screening Model
for Real Solvents. ACS Sustain. Chem. Eng. 2018, 6, 10724–10734. [CrossRef]
Molecules 2019, 24, 2334 19 of 19

43. Martins, M.A.R.; Crespo, E.A.; Pontes, P.V.A.; Silva, L.P.; Bülow, M.; Maximo, G.J.; Batista, E.A.C.; Held, C.;
Pinho, S.P.; Coutinho, J.A.P. Tunable Hydrophobic Eutectic Solvents Based on Terpenes and Monocarboxylic
Acids. ACS Sustain. Chem. Eng. 2018, 6, 8836–8846. [CrossRef]
44. Crespo, E.A.; Silva, L.P.; Martins, M.A.R.; Bülow, M.; Ferreira, O.; Sadowski, G.; Held, C.; Pinho, S.P.;
Coutinho, J.A.P. The Role of Polyfunctionality in the Formation of [Ch]Cl-Carboxylic Acid-Based Deep
Eutectic Solvents. Ind. Eng. Chem. Res. 2018, 57, 11195–11209. [CrossRef]
45. Crespo, E.A.; Silva, L.P.; Martins, M.A.R.; Fernandez, L.; Ortega, J.; Ferreira, O.; Sadowski, G.; Held, C.;
Pinho, S.P.; Coutinho, J.A.P. Characterization and Modeling of the Liquid Phase of Deep Eutectic Solvents
Based on Fatty Acids/Alcohols and Choline Chloride. Ind. Eng. Chem. Res. 2017, 56, 12192–12202. [CrossRef]
46. Pontes, P.V.A.; Crespo, E.A.; Martins, M.A.R.; Silva, L.P.; Neves, C.M.S.S.; Maximo, G.J.; Hubinger, M.D.;
Batista, E.A.C.; Pinho, S.P.; Coutinho, J.A.P.; et al. Measurement and PC-SAFT modeling of solid–liquid
equilibrium of deep eutectic solvents of quaternary ammonium chlorides and carboxylic acids. Fluid Phase
Equilib. 2017, 448, 69–80. [CrossRef]
47. Kollau, L.J.B.M.; Vis, M.; van den Bruinhorst, A.; Esteves, A.C.C.; Tuinier, R. Quantification of the liquid
window of deep eutectic solvents. Chem. Commun. 2018, 54, 13351–13354. [CrossRef] [PubMed]
48. Gamsjäger, H.; Lorimer John, W.; Scharlin, P.; Shaw David, G. Glossary of terms related to solubility (IUPAC
Recommendations 2008). Pure Appl. Chem. 2008, 80, 233–276. [CrossRef]
49. Martins, M.A.R.; Pinho, S.P.; Coutinho, J.A.P. Insights into the Nature of Eutectic and Deep Eutectic Mixtures.
J. Solut. Chem. 2018, 1–21. [CrossRef]
50. Pyykkö, P. Simple Estimates for Eutectic Behavior. Chem. Phys. Chem. 2019, 20, 123–127. [CrossRef]
51. Prausnitz, J.M.; Lichtenthaler, R.N.; Azevedo, E.G.d. Molecular Thermodynamics of Fluid-Phase Equilibria;
Prentice Hall PTR: Upper Saddle River, NJ, USA, 1999.
52. Katritzky, A.R.; Jain, R.; Lomaka, A.; Petrukhin, R.; Maran, U.; Karelson, M. Perspective on the Relationship
between Melting Points and Chemical Structure. Cryst. Growth Des. 2001, 1, 261–265. [CrossRef]
53. Brown, R.J.C.; Brown, R.F.C. Melting Point and Molecular Symmetry. J. Chem. Educ. 2000, 77, 724. [CrossRef]
54. Gavezzotti, A. Are Crystal Structures Predictable? Acc. Chem. Res. 1994, 27, 309–314. [CrossRef]
55. Jain, A.; Yalkowsky, S.H. Estimation of Melting Points of Organic Compounds-II. J. Pharm. Sci. 2006, 95,
2562–2618. [CrossRef]
56. Fernandez, L.; Silva, L.P.; Martins, M.A.R.; Ferreira, O.; Ortega, J.; Pinho, S.P.; Coutinho, J.A.P. Indirect
assessment of the fusion properties of choline chloride from solid–liquid equilibria data. Fluid Phase Equilib.
2017, 448, 9–14. [CrossRef]
57. Sawicka, M.; Storoniak, P.; Błażejowski, J.; Rak, J. TG-FTIR, DSC, and Quantum-Chemical Studies on the
Thermal Decomposition of Quaternary Ethylammonium Halides. J. Phys. Chem. A 2006, 110, 5066–5074.
[CrossRef] [PubMed]
58. Sawicka, M.; Storoniak, P.; Skurski, P.; Błażejowski, J.; Rak, J. TG-FTIR, DSC and quantum chemical studies
of the thermal decomposition of quaternary methylammonium halides. Chem. Phys. 2006, 324, 425–437.
[CrossRef]

Sample Availability: Samples not available.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
 Supplementary Material
Modeling of Solid-liquid Equilibria in Deep Eutectic
Solvents: A Parameter Study
Ahmad Alhadid 1, Liudmila Mokrushina 2 and Mirjana Minceva 1,*
1 Technical University of Munich, Biothermodynamics, TUM School of Life and Food Sciences Weihenstephan,
Maximus-von-Imhof-Forum 2, 85354 Freising, Germany; [email protected]; [email protected]
2 Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Separation Science & Technology, Egerlandstr.

3, 91058 Erlangen, Germany; [email protected].

* Correspondence: [email protected]

Figure S1. Liquidus line (left) and activity coefficients in liquid phase (right) of [Ch]Cl in binary mixture
of [Ch]Cl and (a) [Ch][Ac] (b)[Ch][Prop] (c)[Ch][Buta] modeled using Redlich–Kister polynomial.
Experimental data are taken from Fernandez et al. [1].

1
Figure S2. Liquidus line (left) and activity coefficients in liquid phase (right) of [Ch]Cl in binary mixture
of [Ch]Cl and (a) [Ch][NTf2] (b) [BzCh]Cl (c) [C2mim]Cl (d) [C2OHmim]Cl modeled using Redlich–
Kister polynomial. Experimental data are taken from Fernandez et al. [1].

2
Figure S3. Liquidus line (left) and activity coefficients in liquid phase (right) of [Ch]Cl in binary mixture
of [Ch]Cl and (a) [C4mpyr]Cl (b) [N4444]Cl (c) [P4444]Cl modeled using the Wilson equation. Experimental
data are taken from Fernandez et al. [1].

3
 Figure S4. Chemical names and structures of [Ch]Cl and ILs studied.

Table S1. Melting properties of [Ch]Cl obtained from linear regression of solubility of [Ch]Cl in different ILs.

IL Tm / K hm / kJ mol-1
[Ch][Ac] 588.92 4.86
[Ch][Prop] 594.39 3.90
[Ch][Buta] 607.94 5.06
[Ch][NTf2] 630.83 4.34
[BzCh]Cl 628.26 4.41
[C2mim]Cl 591.61 4.52
[C2OHmim]Cl 589.51 3.95
[C4mpyr]Cl 593.32 5.63
[N4444]Cl 594.22 4.11
[P4444]Cl 607.16 3.85
Average 602.61 4.46

4
Figure S5. Liquidus line (left) and activity coefficients (right) of [N1111]Cl in binary mixture of [N1111]Cl
and (a) capric acid (b) lauric acid (c) myristic acid (d) palmitic acid (e) stearic acid modeled using
Redlich-Kister polynomial with three parameters assuming different melting enthalpy values.
Experimental data are taken from [2].

5
Figure S6. Liquidus line (left) and activity coefficients (right) of [N2222]Cl in binary mixture of [N2222]Cl and (a)
capric acid (b) lauric acid (c) myristic acid (d) palmitic acid (e) stearic acid modeled using Redlich-Kister
polynomial with three parameters assuming different melting enthalpy values. Experimental data are taken from
[2].

6
Figure S7. Liquidus line (left) and activity coefficients (right) of [N3333]Cl in binary mixture of [N3333]Cl and (a)
capric acid (b) lauric acid (c) myristic acid (d) palmitic acid (e) stearic acid modeled using Redlich-Kister
polynomial with three parameters assuming different melting enthalpy values. Experimental data are taken from
[2].

7
 Table S2. Empirical parameters of RK-polynomial for calculating activity coefficients of salts in binary
mixtures of quaternary ammonium chloride salts and capric acid.

hm / kJ mol−1 [N1111]Cl [N2222]Cl [N3333]Cl

(1) (1) (1) (1)
a (1)
b c a(1)
b c a (1)
b(1) c(1)
1 12.23 -5.31 -3.97 10.97 -6.66 -2.63 14.91 -24.67 15.95
5 7.93 -10.69 4.71 3.20 -7.56 7.66 3.40 -9.77 11.33
10 -2.70 -0.75 2.88 -13.09 15.14 -0.36 -7.49 -2.29 14.09
20 - - - -49.75 75.31 -29.38 -37.25 38.00 0.19
30 -49.22 51.72 -14.25 -44.63 -16.63 75.13 -66.63 77.13 -12.84
40 -46.00 -5.13 40.13 -98.31 105.69 -8.50 -47.75 -36.63 91.13
Experimental -32.75 44.56 -19.59 -96.88 18.38 95.00 -170.38 208.75 -51.88

Table S3. Empirical parameters of RK-polynomial for calculating activity coefficients of salts in binary
mixtures of quaternary ammonium chloride salts and lauric acid.

hm / kJ mol−1 [N1111]Cl [N2222]Cl [N3333]Cl

a(1) b(1) c(1) a(1) b(1) c(1) a(1) b(1) c(1)
1 26.33 -48.76 20.49 10 -18 19.5 12.13 -5.17 -9.59
5 8.63 -10.90 -9.88 0 0 4.5 2.68 1.91 -7.20
10 -10.63 23.44 -33.81 0 0 -13.5 -10.48 16.09 -9.34
20 - - - -35.5 62 -48.25 -42.06 65.38 -33.56
30 -124.94 325.88 -305.63 -68.5 116 -76.75 -72.50 110.19 -53.63
40 -108.25 153.25 -98.25 -110.5 205 -138.75 -104.75 161.63 -79.38
Experimental -54.13 109.53 -99.13 -148.5 269 -173.75 -170.50 219.75 -73.50

Table S4. Empirical parameters of RK-polynomial for calculating activity coefficients of salts in binary
mixtures of quaternary ammonium chloride salts and myristic acid

hm / kJ mol−1 [N1111]Cl [N2222]Cl [N3333]Cl

a(1) b(1) c(1) a(1) b(1) c(1) a(1) b(1) c(1)
1 12.83 -3.81 -9.66 17.84 -28.73 14.94 19.55 -39.47 24.72
5 8.74 -15.07 7.05 6.02 -14.03 12.09 1.45 -6.17 6.78
10 -0.93 -12.67 13.61 -6.00 10.00 -10.00 -17.99 20.95 0.38
20 - - - -41.50 72.00 -43.00 -66.80 120.27 -61.87
30 -31.69 -32.06 65.25 -79.00 139.00 -78.50 -68.44 5.09 111.50
40 -77.50 67.94 -4.25 -108.00 176.00 -88.00 -161.44 305.31 -171.47
Experimental -30.09 24.00 0.08 -141.50 218.00 -96.50 -230.75 295.00 -35.88

8
 Table S5. Empirical parameters of RK-polynomial for calculating activity coefficients of salts in binary
mixtures of quaternary ammonium chloride salts and palmitic acid.

hm / kJ mol−1 [N1111]Cl [N2222]Cl [N3333]Cl

(1) (1) (1) (1)
a (1)
b c a(1)
b c a (1)
b(1) c(1)
1 20.54 -29.42 7.66 17.55 -33.27 22.19 14.81 -24.81 11.25
5 6.47 -9.34 -0.71 0.29 -5.85 12.72 -11.00 23.20 -6.38
10 -11.72 18.42 -13.98 -21.98 31.63 -2.62 -45.94 97.00 -45.47
20 - - - -46.50 23.03 51.75 -116.00 245.31 -124.25
30 -89.75 151.44 -89.13 -109.28 172.75 -53.94 -185.75 392.38 -202.00
40 -126.50 208.69 -117.63 -158.56 267.22 -104.25 -257.00 545.88 -286.00
Experimental -26.56 -20.25 55.63 -223.00 431.00 -235.25 -441.50 932.50 -488.00

Table S6. Empirical parameters of RK-polynomial for calculating activity coefficients of salts in binary
mixtures of quaternary ammonium chloride salts and stearic acid.

hm / kJ mol−1 [N1111]Cl [N2222]Cl [N3333]Cl

a(1) b(1) c(1) a(1) b(1) c(1) a(1) b(1) c(1)
1 18.97 -29.02 12.78 17.55 -33.27 22.19 19.63 -35.17 18.81
5 1.98 -0.83 -0.48 0.29 -5.85 12.72 5.37 -12.75 10.11
10 -11.47 3.36 12.94 -21.98 31.63 -2.62 -12.38 15.33 -1.19
20 - - - -46.50 23.03 51.75 -32.39 0.12 55.52
30 -101.56 165.25 -73.88 -109.28 172.75 -53.94 -90.22 157.97 -78.78
40 -145.50 259.00 -141.75 -158.56 267.22 -104.25 -130.98 237.61 -126.69
Experimental -61.25 98.13 -41.81 -196.75 301.19 -87.13 -239.94 452.41 -257.81

Table S7. Melting properties of quaternary ammonium chloride and fatty acids. Data are taken from [2]

Compound Tm / K hm / kJ mol−1 Sm / J mol−1 K−1

[N1111]Cl 612.87 20.49 33.43
[N2222]Cl 526.78 51.24 97.27
[N3333]Cl 503.07 66.58 132.35
Capric acid 304.75 27.5 90.24
Lauric acid 317.48 37.83 119.16
Myristic acid 327.03 41.29 126.26
Palmitic acid 336.84 51.02 151.47
Stearic acid 343.67 61.36 178.54

References

[1] L. Fernandez, L.P. Silva, M.A.R. Martins, O. Ferreira, J. Ortega, S.P. Pinho, J.A.P. Coutinho, Indirect
assessment of the fusion properties of choline chloride from solid-liquid equilibria data, Fluid Phase
Equilib., 448 (2017) 9-14.
[2] P.V.A. Pontes, E.A. Crespo, M.A.R. Martins, L.P. Silva, C.M.S.S. Neves, G.J. Maximo, M.D.
Hubinger, E.A.C. Batista, S.P. Pinho, J.A.P. Coutinho, G. Sadowski, C. Held, Measurement and PC-SAFT
modeling of solid-liquid equilibrium of deep eutectic solvents of quaternary ammonium chlorides
and carboxylic acids, Fluid Phase Equilib., 448 (2017) 69-80.

9
3.2 Paper II

3.2 Paper II
Design of Deep Eutectic Systems: A Simple Approach for Preselecting Eutectic Mixture
Constituents

A. Alhadid, L. Mokrushina, and M. Minceva, Molecules, 2020, 25, 1077.

Author contribution: The thesis author conceptualized the paper’s idea, performed the
investigations and formal analysis, interpreted the results, and wrote the manuscript.

Summary: Paper II proposed a simple approach for preselecting eutectic mixture constituents
from a pool of substances sharing the same chemical nature using their molecular structure. The
approach is based on the conclusion of Paper I by selecting constituents with small melting
entropy values. The melting entropy model proposed by Jain et al.112 was used to calculate the
melting entropy of various compounds from their molecular structure.

It was found that constituents with symmetrical and rigid molecular structures possess low
melting entropy values. Accordingly, constituents with symmetrical and rigid molecular
structures are expected to form eutectic mixtures with lower eutectic temperatures compared to
constituents with flexible and asymmetric molecular structures.

To confirm the proposed approach, the SLE phase diagrams of six eutectic systems containing
L-menthol and monocarboxylic acids were measured. Pairwise comparison between the eutectic

temperature of the studied systems was performed, wherein in each pair of systems, the two
acids have the same melting temperature, but the structure of the hydrocarbon chain is either
cyclic or linear. The comparison showed that the eutectic temperature of systems containing L-
menthol and cyclic acids is lower than that of systems with linear acids. Thus, it was concluded
that when substances have the same chemical nature and melting temperature, those with more
rigid and symmetrical molecular structures should be selected as constituents to form eutectic
mixtures with lower eutectic temperatures.

64
 molecules
Article
Design of Deep Eutectic Systems: A Simple Approach
for Preselecting Eutectic Mixture Constituents
Ahmad Alhadid 1 , Liudmila Mokrushina 2 and Mirjana Minceva 1, *
1 Biothermodynamics, TUM School of Life Sciences Weihenstephan, Technical University of Munich,
Maximus-von-Imhof-Forum 2, 85354 Freising, Germany; [email protected]
2 Separation Science & Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Egerlandstr. 3,
91058 Erlangen, Germany; [email protected]
* Correspondence: [email protected]
!"#!$%&'(!
Received: 5 February 2020; Accepted: 25 February 2020; Published: 28 February 2020 !"#$%&'

Abstract: Eutectic systems o↵er a wide range of new (green) designer solvents for diverse applications.
However, due to the large pool of possible compounds, selecting compounds that form eutectic
systems is not straightforward. In this study, a simple approach for preselecting possible candidates
from a pool of substances sharing the same chemical functionality was presented. First, the melting
entropy of single compounds was correlated with their molecular structure to calculate their melting
enthalpy. Subsequently, the eutectic temperature of the screened binary systems was qualitatively
predicted, and the systems were ordered according to the depth of the eutectic temperature. The
approach was demonstrated for six hydrophobic eutectic systems composed of L-menthol and
monocarboxylic acids with linear and cyclic structures. It was found that the melting entropy of
compounds sharing the same functionality could be well correlated with their molecular structures.
As a result, when the two acids had a similar melting temperature, the melting enthalpy of a rigid
acid was found to be lower than that of a flexible acid. It was demonstrated that compounds with
more rigid molecular structures could form deeper eutectics. The proposed approach could decrease
the experimental e↵orts required to design deep eutectic solvents, particularly when the melting
enthalpy of pure components is not available.

Keywords: eutectic mixtures; deep eutectic solvents; solid–liquid equilibria; hydrophobic DESs;
melting properties

1. Introduction
Eutectic systems are mixtures of two or more compounds that exhibit partial immiscibility or
negligible mutual solubility in the solid phase [1]. Deep eutectic solvents (DESs) are eutectic mixtures
characterized by a large depression of the melting temperature of the mixture at the eutectic point
relative to the melting temperature of the pure components [2,3]. DESs are analogous to ionic liquids
(ILs) in terms of being designer solvents and possessing low vapor pressure. However, DESs are usually
less toxic, easier to prepare, and less expensive than ILs. These advantages have led to the recent
increase in applications of DESs, for example, as solvents in diverse separation methods [4–8], media for
chemical [9–14], electrochemical [15–19], and biological reactions [20,21], in polymer chemistry [22–24],
and for increasing the solubility of active pharmaceutical ingredients [25–28].
Although their preparation may be easier than that of ILs, DESs are more difficult to design. The
ratio of the ions in ILs is defined by the electroneutrality of the solution. In contrast, the ratio of DES
components is not fixed and can be of any value. One of the first pieces of information required when
designing DESs for a specific application is the eutectic temperature and eutectic composition of the
system. Thus far, the design of DESs has been performed primarily using a trial and error approach. In

Molecules 2020, 25, 1077; doi:10.3390/molecules25051077 www.mdpi.com/journal/molecules

Molecules 2020, 25, 1077 2 of 11

most published works, preselected components are mixed at several fixed molar ratios, such as 1:1
or 1:2, and mixtures that remain liquid at room temperature are selected for further testing [29,30].
To determine the system composition and melting temperature at the eutectic point, the solid–liquid
equilibria (SLE) of the eutectic systems should be known. The SLE also provides information about the
melting temperature of the system at any specific composition.
The experimental determination of SLE phase diagrams of eutectic systems is non-trivial and is
often accompanied by difficulties and limitations. For example, the hygroscopic nature of some DES
components [31], the high viscosity and paste-like consistency of some DESs close to their melting
temperature [32], the decomposition of DES constituents before melting, and the chemical reaction
between DES constituents after storage [33]. Owing to the previously mentioned difficulties, predictive
methods are required. Abranches et al. [34] proposed Conductor like Screening Model for Real Solvents
(COSMO-RS) to predict the SLE of eutectic mixtures. Wolbert et al. [35] used UNIFAC (Do) to model
the activity coefficient of constituents in binary eutectic mixtures.
The SLE of simple eutectic systems, whose components show negligible mutual solubility in the
solid phase, is commonly calculated in the literature using the following simplified equation:
!
Dhm,i 1 1
ln xLi Li = (1)
R Tm,i T

where xLi and Li are the mole fraction and activity coefficient of the component i in the liquid solution,
respectively; Dhm,i and Tm,i are the melting enthalpy and melting temperature of the pure component i,
respectively; T is the liquidus temperature (i.e., melting temperature of the mixture at the mole fraction
xLi ); and R is the universal gas constant. As seen in Equation (1), SLE calculations require information
about the pure components melting properties, namely, the melting enthalpy Dhm,i and temperature
Tm,i , as well as information about the behavior of the components in the liquid phase (i.e., activity
coefficients Li ).
Strong intermolecular interactions between unlike molecules in the liquid phase—low activity
coefficient values of components—and/or low melting enthalpy values of pure components lead to a
deep depression of the melting temperature of the mixture at the eutectic point [35–38]. The melting
enthalpy of components is not always easily measured because of polymorphism, kinetic limitations,
and/or thermal instability. As a result, the melting enthalpy of many components is unavailable. The
aim of this work was to demonstrate the correlation between the molecular structure and melting
enthalpy of a component to simplify the selection of components, especially when no experimental
data on melting enthalpy is available.
At the melting point of a pure component, the solid and liquid phases are in equilibrium. The
melting temperature Tm of a pure component is the ratio between the melting enthalpy Dhm and the
melting entropy Dsm :
Dhm
Tm = (2)
Dsm
Melting enthalpy is the energy required to melt solid crystals [39] and depends on the type
of interactions between molecules in the lattice structure [39,40]. Melting entropy is the increase
in the disorder and randomness upon melting [41] and depends on the molecular symmetry and
conformational degrees of freedom of the molecule [41–44]. Despite several attempts to correlate the
melting properties of a component with its molecular structure, there is no generic model that can
predict the melting properties of pure components [45].
The melting temperature of pure components is difficult to predict [43]; however, unlike the
melting enthalpy, experimental data on the melting temperature are, in many cases, available. However,
depending on the method of determination as well as the purity of the components, the reported
values may deviate by several degrees Celsius from the actual melting temperature. However, as
seen from Equation (1), an uncertainty of several degrees Celsius in the melting temperature of pure
components would have a small e↵ect on the SLE of the mixture. According to Bondi [46], the melting
Molecules 2020, 25, 1077 3 of 11

entropy of compounds can be better related to the molecular structure than the melting enthalpy. The
melting entropy calculated from the molecular structure with the available melting temperature can be
used to estimate the melting enthalpy of the components with Equation (2). Using this information,
the eutectic temperature can be approximately calculated using Equation (1) under the assumption of
ideal behavior ( Li = 1).
The objective of this study was to test a simple approach that could be used to select potential
eutectic system constituents based on their melting enthalpies. It is assumed that mixtures of
components with a lower melting enthalpy would result in eutectic mixtures with a larger melting
temperature depression, as previously demonstrated [35–38]. This approach aims to reduce the
experimental e↵orts required to measure pure components’ melting enthalpy, as well as the SLE
of eutectic mixtures. To evaluate the proposed approach, binary eutectic mixtures of L-menthol
with six di↵erent monocarboxylic acids were considered in this work. The goal was to predict the
eutectic temperature of each system relative to other systems. Although the eutectic temperatures
were calculated with the unitary activity coefficient, it was not claimed that the components should
behave ideally. The proposed approach was based on the assumption that any component in its
binary solutions with other components sharing the same type and number of functional groups
behaves similarly (i.e., in any binary mixture with monocarboxylic acids, L-menthol behaves in a
similar manner). The latter assumption has been validated for many eutectic systems, for example, in
[Ch]Cl/sugar [32], [Ch]Cl/dicarboxylic acids [47], [Ch]Cl/fatty acids or alcohols [48], and thymol/fatty
acids [49].

2. Results and Discussion

2.1. Melting Properties of Pure Components

The melting entropy of acids was calculated using the model proposed by Jain et al. [44] (Equations
(4)–(6)). The model parameters and calculated melting entropies are presented in Table 1. Linear
acids possessed higher melting entropy than cyclic acids. This was a result of the higher flexibility
of a linear chain compared to a ring structure. The lowest predicted melting entropy was for
cyclohexanecarboxylic acid with a flexibility number F equal to zero. The model predicted the same
melting entropy for 3-phenylpropionic acid and 3-cyclohexylpropionic acid. This was because the
model did not di↵erentiate between phenyl and aliphatic ring and predicted the same flexibility
number F for both components. Due to the asymmetry of the carboxylic acid group, the symmetry
number of all acids tested in this work was one.

Table 1. Symmetry number , flexibility number F, and calculated melting entropy Dsm of components
using the model proposed by Jain et al. [44].

Compound SP3 SP2 Ring ⌧ F Dsm (J mol–1 K–1 )

3-cyclohexylpropionic acid 1 2 1 1 2 5.93 64.80
caprylic acid 1 6 1 0 5.5 133.58 90.69
cyclohexanecarboxylic acid 1 0 1 1 0 1 50
capric acid 1 8 1 0 7.5 792.03 105.49
3-phenylpropionic acid 1 2 1 1 2 5.93 64.80
lauric acid 1 10 1 0 9.5 4696.13 120.29

Table 2 presents the melting properties that were experimentally determined in this study using
DSC. The obtained values were in good agreement with those reported in the literature. To the best of
our knowledge, the melting properties of cyclohexylpropionic acid have not been measured before. As
seen in Table 2, the melting enthalpy of l-menthol had a low value, thus making l-menthol a good
candidate for designing deep eutectic systems. In general, linear acids have higher melting enthalpies
than cyclic acids. The melting enthalpy of linear acids increases by increasing the chain length; for
Molecules 2020, 25, 1077 4 of 11

example, lauric acid > capric acid > caprylic acid. The lowest melting enthalpy was observed for
cyclohexanecarboxylic acid, which is the component with the most rigid molecular structure.

Table 2. Comparison of melting enthalpies Dhm and temperatures Tm measured in this study and
reported in the literature.

Tm (K) Dhm (kJ mol–1 )

Compound
This Work * Lit. This Work * Lit.
l-menthol 314.6 ± 0.1 315.68 [49] 13.74 ± 0.5 12.89 [49]
3-cyclohexylpropionic acid 291.3 ± 0.1 – 16.96 ± 0.5 –
caprylic acid 288.0 ± 0.7 288.20 [49] 21.43 ± 0.3 19.80 [49]
cyclohexanecarboxylic acid 299.4 ± 1.1 301.9 [50] 10.69 ± 0.2 9.20 [50]
capric acid 303.9 ± 0.1 304.75 [51] 28.39 ± 0.7 27.50 [51]
3-phenylpropionic acid 321.6 ± 0.1 321.2 [52] 15.11 ± 0.1 15.68 [52]
lauric acid 316.6 ± 0.1 317.48 [51] 35.81 ± 0.4 37.83 [51]
* Uncertainties are considered as the standard deviation of three measurements.

Table 3 presents a comparison between the melting entropies predicted by the model by
Jain et al. [44] (Equations (4)–(6)) and the experimental melting entropies calculated with Equation (2)
using the experimentally determined melting enthalpy and melting temperatures of acids. As seen in
Table 3, the predicted melting entropy of all acids was overestimated.

Table 3. Comparison of predicted Dsm predicted and experimental Dsm experimental melting entropy of
monocarboxylic acids.

Compound Dsm predicted Dsm experimental

3-cyclohexylpropionic acid 64.80 58.22

caprylic acid 90.69 74.41
cyclohexanecarboxylic acid 50 35.72
capric acid 105.49 93.44
3-phenylpropionic acid 64.80 46.97
lauric acid 120.29 113.12

Figure 1 depicts the predicted melting entropy values in comparison to the experimental values.
The linear correlation between the predicted melting entropies indicated that the model proposed
by Jain et al. [44] could provide a reasonably good estimation of the melting entropy of compounds
sharing the same chemical functionality.

Figure 1. Experimental melting entropies measured in this study in comparison with predicted
melting entropies.

2.2. Solid–Liquid Equilibria

To validate the proposed approach, the SLE of the six eutectic systems was measured using DSC.
Figure 2 presents the measured SLE data for l-menthol with six di↵erent monocarboxylic acids. To
Molecules 2020, 25, 1077 5 of 11

simplify the comparison between systems, the systems containing acids of similar melting temperatures
are presented next to each other (Figure 2A–F). The approximated melting temperature of acids in each
pair increased from Figure 2A,B (⇡ 15 C) to Figure 2E,F (⇡ 45 C). Martins et al. [49] measured the SLE
of binary eutectic mixtures of L-menthol with caprylic acid, capric acid, and lauric acid. As seen from
Figure 2A,C,E, the determined eutectic temperatures were in good agreement with the results reported
by Martins et al. [49]. The slightly higher liquidus temperatures measured in this study might be due to
the greater heating rate; in this study, a heating rate of 5 K min–1 was used, while in Martins et al. [49],
1 K min–1 was used.

Figure 2. Solid–liquid phase diagrams of binary eutectic mixtures consisting of l-menthol and (A)
caprylic acid, (B) cyclohexylpropionic acid, (C) capric acid, (D) cyclohexanecarboxylic acid, (E) lauric
acid, and (F) phenylpropionic acid. The melting properties presented are experimentally determined
values. Legend: liquidus temperature measured in this study, • the experimental eutectic temperature
measured in this study, ⇥ Martins et al. [49].

As demonstrated in previous studies [35–38], the lower the melting enthalpy of the pure
components, the higher the depression at the eutectic point. This could be confirmed by comparing
the eutectic temperatures of eutectic systems formed between l-menthol and acids presented in
Figure 2. The molecules with cyclic structures (Figure 2B,D,F) had lower flexibility than molecules
with linear structures (Figure 2A,C,E). Therefore, the cyclic compounds possessed lower melting
entropies. Because the melting temperatures of each pair of acids were similar (caprylic acid and
cyclohexylpropionic acid ⇡ 15 C, capric acid and cyclohexanecarboxylic acid ⇡ 30 C, and lauric acid
and phenylpropionic acid ⇡ 45 C), the melting enthalpy of cyclic compounds was lower than that of
Molecules 2020, 25, 1077 6 of 11

linear ones. As a result, the eutectic temperature of a system formed by a cyclic acid was lower than
that of a system formed by a linear acid when both acids had the same melting temperature.
As seen in Figure 2, as the di↵erence in the melting entropy of acids increased, the di↵erence
in the eutectic temperature increased. For example, the melting entropy of capric acid was almost
three times higher than that of cyclohexanecarboxylic acid (see Figure 2C,D). This resulted in a eutectic
temperature for l-menthol/cyclohexanecarboxylic acid (Figure 2C), which was approximately 15 K
lower than that of l-menthol/capric acid (Figure 2D). For l-menthol/caprylic acid (Figure 2A) and
l-menthol/3-cyclohexylpropionic acid (Figure 2B), the di↵erence between their eutectic temperatures
was only 4.5 K. This might be the result of a small di↵erence between the melting entropies of caprylic
acid and 3-cyclohexylpropionic acid (see Figure 2A,B). In a previous study [38], it was demonstrated
that the eutectic composition was shifted toward the component that had a lower melting enthalpy.
Comparing the eutectic composition between the systems revealed that the lower the melting enthalpy
compared to that of l-menthol (13.74 kJ mol 1 ), the higher the mole fraction of the acid at the
eutectic point.
It could be concluded that a lower melting enthalpy of pure components led to a larger melting
temperature depression at the eutectic point of all l-menthol/monocarboxylic acid systems studied
in this work. The melting enthalpy of components could be correlated with the molecular structure
of acids possessing the same melting temperature. Therefore, the eutectic systems formed between
l-menthol and cyclic acids exhibited a deeper eutectic point than that of systems formed with linear
acids. Because the depression at the eutectic point was related to the di↵erence in the melting entropy
of the pure components, the relative depression at the eutectic point between the systems could also be
predicted. Thus, the approach of selecting components by assessing the flexibility of their molecular
structures could be used to design deeper eutectic systems.

3. Materials and Methods

3.1. Prediction of Melting Entropy

The melting entropy of the pure components was calculated as the sum of the rotational Dsrot
m ,
con f expan
conformational Dsm , and expansional entropies Dsm as follows [41]:

con f expan
Dsm = W + Dsrot
m + Dsm + Dsm (3)

where W is a constant. In this study, the model proposed by Jain et al. [44] was used to predict
the melting entropy of the pure components. The melting entropy in J mol–1 K–1 was calculated as
follows [44]:
con f
Dsm = 50 + Dsrot
m + Dsm (4)

Dsrot
m = R ln (5)
con f
Dsm = R ln F (6)

where is the symmetry number, F is the flexibility number, and R is the universal gas constant.
Readers are directed to the original paper [44] for more information about the determination of the
symmetry number of the components. The flexibility number F was calculated as follows [44]:

F = 2.435⌧ (7)

⌧ = SP3 + 0.5SP2 + 0.5Ring 1 (8)

where SP3 is the number of non-ring SP3 atoms (CH2 , CH, C, NH, N, O, S), SP2 is the number of SP2
atoms (=CH,=C,=N, C=O), and Ring is the number of independent single, fused, or conjugated ring
systems. If ⌧ is less than zero, the flexibility number F is set to 1.
Molecules 2020, 25, 1077 7 of 11

In the search for components with low melting enthalpy values, components with low melting
entropy were sought. According to Equations (4)–(6), components with symmetrical (large symmetry
number ) and/or more rigid molecular structures (small flexibility number F) should possess lower
melting entropy. According to Equations (7) and (8), components with ring systems and double bonds
should have a lower flexibility number F than single bond chains.

3.2. Eutectic Mixture Constituents

Binary eutectic mixtures of l-menthol with six di↵erent monocarboxylic acids were considered.
The acids were sorted into three pairs according to the melting temperatures reported by the suppliers.
The acids of each pair had close melting temperatures. The acids pairs were as follows:

(i) Caprylic acid and 3-cyclohexylpropionic acid with a melting temperature of approximately 15 C,
(ii) Capric acid and cyclohexanecarboxylic acid with a melting temperature of approximately 30 C,
(iii) Lauric acid and 3-phenylpropionic acid with a melting temperature of approximately 45 C.

In each pair, the acid with a more rigid molecular structure was expected to have lower melting
entropy, and according to Equation (2), lower melting enthalpy. Therefore, the eutectic systems formed
between l-menthol and rigid acids were expected to have a deeper eutectic temperature as a result
of their lower melting enthalpy. Figure 3 illustrates the di↵erences in the molecular structures of the
acids, paired based on the melting temperatures obtained from the suppliers.

Figure 3. Molecular structures of monocarboxylic acids considered in this study. The melting
temperatures are approximate values reported by the suppliers used to select each pair of acids.

3.3. Eutectic Mixture Preparation

Table 4 lists the chemicals used in this study, along with their purity, as declared by the suppliers.
All chemicals were used as received without further purification. The binary eutectic mixtures were
prepared by weighing the pure components in a glass vessel. The mixture was heated to 45 C while
stirring with a magnetic stirrer until a homogenous liquid was obtained.
Molecules 2020, 25, 1077 8 of 11

Table 4. Chemicals used in this study.

Name CAS Number Supplier Purity *

l-menthol 2216–51–5 Sigma Aldrich Chemie GmbH 99 %
3-cyclohexylpropionic acid 701–97–3 ThermoFisher (Kandel) GmbH > 98 %
caprylic acid 124–07–2 Merck KGaA 99 %
cyclohexanecarboxylic acid 98–89–5 ThermoFisher (Kandel) GmbH 98 %
capric acid 334–48–5 Alfa Aesar GmbH 99 %
3-phenylpropionic acid 501–52–0 Alfa Aesar GmbH 99 %
lauric acid 143–07–7 Merck KGaA 99 %
* As declared by the supplier.

The liquid samples were weighed in aluminum DSC crucibles using a syringe and then sealed
by cold welding. Depending on the density of the eutectic mixture, the mass of the samples in the
crucibles ranged between 4 and 6 mg.

3.4. Di↵erential Scanning Calorimetry

SLE data and the pure components’ melting properties were measured using DSC (NETZSCH
DSC 200 F3). Temperature and sensitivity calibrations were performed prior to the measurements
using five calibration standards with a purity of over 99.999%. The calibration standards included
adamantane, bismuth, indium, zinc, and tin. The uncertainties after the temperature and sensitivity
calibration were < 0.1 K and < 3%, respectively.
The DSC measurements were performed in an inert nitrogen environment. First, a cooling cycle
with a rate of 10 K min 1 to a final temperature of 80 C was performed. Then, the sample was heated
at a heating rate of 5 K min–1 up to approximately 10 K above the sample liquidus temperature. The
eutectic temperature, as well as the pure components’ melting temperature, were determined as the
onset temperature of the respective thermal event. The liquidus temperatures were determined as the
peak temperature, and the melting enthalpy of the pure components was determined as the peak area
of the respective thermal event. The DSC curves of the samples obtained during the heating cycle are
shown in Figures S2–S4 in Supplementary Materials.

4. Conclusions
In this study, a simple approach was proposed to select constituents for eutectic systems. This
approach could be used to select constituents from a pool of substances sharing the same functionality
and melting temperature based on their melting enthalpy. If the melting enthalpy is not available, it
can be estimated from the melting entropy, which is correlated with the molecular structure, using the
simple non-group contribution model proposed by Jain et al. [44].
The proposed approach was used to predict the relative depression of the melting temperature at
the eutectic point of l-menthol/monocarboxylic acid systems. It was demonstrated that components
with more rigid molecular structures possessed lower melting entropy. For linear and cyclic acids with
similar melting temperatures, the cyclic acids possessed lower melting enthalpy due to their lower
melting entropy. As a result, deeper eutectic systems could be formed by cyclic acids than by linear
acids sharing the same melting temperature. Furthermore, the larger the di↵erence in the melting
entropy between the acids in each pair, the higher the relative depression at the eutectic point between
the two eutectic systems.
From this study, it could be concluded that in the search for new DES systems, constituents with
more rigid and symmetrical structures should be pursued. This could narrow the pool of possible
components to be screened for a specific application based on eutectic temperature (i.e., when eutectic
mixtures that are liquid at room temperature are investigated). It should be mentioned, however,
that for quantitative predictions, experimental melting properties should be used along with activity
coefficient modeling, using, for example, excess Gibbs energy models (gE ) or equation of states.
Molecules 2020, 25, 1077 9 of 11

Supplementary Materials: The following are available online. Figure S1. Solid-liquid phase diagrams
of binary eutectic systems consisting of l-menthol and (A) cyclohexylpropionic acid (B) caprylic acid (C)
cyclohexanecarboxylic acid (D) capric acid (E) phenylpropionic acid (F) lauric acid. Dashed lines are ideal liquidus
lines of components modeled using the Schöder-van-Laar equation and using experimental melting properties.
Figure S2. DSC curves of l-menthol/3-cyclohexylpropionic acid and l-menthol/caprylic acid systems. Figure S3.
DSC curves of l-menthol/cyclohexanecarboxylic acid and l-menthol/capric acid systems. Figure S4. DSC curves
of l-menthol/3-phenylpropionic acid and l-menthol/lauric acid systems.
Author Contributions: Designing and performing experiments, analyzing the results, and preparing the
manuscript A.A.; revising and discussing the manuscript, L.M.; supervising, revising, and discussing the
manuscript, M.M. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported by the German Research Foundation (DFG) and the Technical University of
Munich (TUM) in the framework of the Open Access Publishing Program.
Acknowledgments: The authors would like to thank Lea Kefalianakis for helping in performing a part of the
DSC experiments as a part of her B. Sc. thesis.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Gamsjäger, H.; Lorimer J., W.; Scharlin, P.; Shaw David, G. Glossary of terms related to solubility (IUPAC
Recommendations 2008). Pure Appl. Chem. 2008, 80, 233.
2. Abbott, A.P.; Boothby, D.; Capper, G.; Davies, D.L.; Rasheed, R.K. Deep Eutectic Solvents Formed between
Choline Chloride and Carboxylic Acids: Versatile Alternatives to Ionic Liquids. J. Am. Chem. Soc. 2004, 126,
9142–9147. [CrossRef] [PubMed]
3. Abbott, A.P.; Capper, G.; Davies, D.L.; Rasheed, R.K.; Tambyrajah, V. Novel solvent properties of choline
chloride/urea mixtures. Chem. Commun. 2003, 70–71. [CrossRef] [PubMed]
4. Abbott, A.P.; Cullis, P.M.; Gibson, M.J.; Harris, R.C.; Raven, E. Extraction of glycerol from biodiesel into a
eutectic based ionic liquid. Green Chem. 2007, 9, 868–872. [CrossRef]
5. Abbott, A.P.; Harris, R.C.; Ryder, K.S.; D’Agostino, C.; Gladden, L.F.; Mantle, M.D. Glycerol eutectics as
sustainable solvent systems. Green Chem. 2011, 13, 82–90. [CrossRef]
6. Bezold, F.; Minceva, M. A water-free solvent system containing an L-menthol-based deep eutectic solvent for
centrifugal partition chromatography applications. J. Chromatogr. A 2019, 1587, 166–171. [CrossRef]
7. Gouveia, A.S.L.; Oliveira, F.S.; Kurnia, K.A.; Marrucho, I.M. Deep Eutectic Solvents as Azeotrope Breakers:
Liquid–Liquid Extraction and COSMO-RS Prediction. ACS Sustain. Chem. Eng. 2016. [CrossRef]
8. Roehrer, S.; Bezold, F.; Garcia, E.M.; Minceva, M. Deep eutectic solvents in countercurrent and centrifugal
partition chromatography. J. Chromatogr. A 2016, 1434, 102–110. [CrossRef]
9. Alonso, D.A.; Baeza, A.; Chinchilla, R.; Guillena, G.; Pastor, I.M.; Ramón, D.J. Deep Eutectic Solvents: The
Organic Reaction Medium of the Century. Eur. J. Org. Chem. 2016, 2016, 612–632. [CrossRef]
10. Hayyan, A.; Ali Hashim, M.; Mjalli, F.S.; Hayyan, M.; AlNashef, I.M. A novel phosphonium-based deep
eutectic catalyst for biodiesel production from industrial low grade crude palm oil. Chem. Eng. Sci. 2013, 92.
[CrossRef]
11. Hayyan, A.; Hashim, M.A.; Hayyan, M.; Mjalli, F.S.; AlNashef, I.M. A new processing route for cleaner
production of biodiesel fuel using a choline chloride based deep eutectic solvent. J. Clean. Prod. 2014, 65.
[CrossRef]
12. Khandelwal, S.; Tailor, Y.K.; Kumar, M. Deep eutectic solvents (DESs) as eco-friendly and sustainable
solvent/catalyst systems in organic transformations. J. Mol. Liq. 2016, 215, 345–386. [CrossRef]
13. Wagle, D.V.; Zhao, H.; Baker, G.A. Deep eutectic solvents: Sustainable media for nanoscale and functional
materials. Acc. Chem. Res. 2014, 47. [CrossRef] [PubMed]
14. Liu, P.; Hao, J.-W.; Mo, L.-P.; Zhang, Z.-H. Recent advances in the application of deep eutectic solvents as
sustainable media as well as catalysts in organic reactions. RSC Advances 2015, 5, 48675–48704. [CrossRef]
15. Abbott, A.P.; Griffith, J.; Nandhra, S.; O’Connor, C.; Postlethwaite, S.; Ryder, K.S.; Smith, E.L. Sustained
electroless deposition of metallic silver from a choline chloride-based ionic liquid. Surf. Coat. Technol. 2008,
202, 2033–2039. [CrossRef]
Molecules 2020, 25, 1077 10 of 11

16. Boisset, A.; Menne, S.; Jacquemin, J.; Balducci, A.; Anouti, M. Deep eutectic solvents based on
N-methylacetamide and a lithium salt as suitable electrolytes for lithium-ion batteries. PCCP 2013, 15,
20054–20063. [CrossRef]
17. Chen, Z.; McLean, B.; Ludwig, M.; Stefanovic, R.; Warr, G.G.; Webber, G.B.; Page, A.J.; Atkin, R. Nanostructure
of Deep Eutectic Solvents at Graphite Electrode Interfaces as a Function of Potential. J. Phys. Chem. C 2016,
120, 2225–2233. [CrossRef]
18. Gómez, E.; Cojocaru, P.; Magagnin, L.; Valles, E. Electrodeposition of Co, Sm and SmCo from a Deep Eutectic
Solvent. J. Electroanal. Chem. 2011, 658, 18–24. [CrossRef]
19. Zaidi, W.; Boisset, A.; Jacquemin, J.; Timperman, L.; Anouti, M. Deep Eutectic Solvents Based on
N-Methylacetamide and a Lithium Salt as Electrolytes at Elevated Temperature for Activated Carbon-Based
Supercapacitors. J. Phys. Chem. C 2014, 118, 4033–4042. [CrossRef]
20. Pätzold, M.; Siebenhaller, S.; Kara, S.; Liese, A.; Syldatk, C.; Holtmann, D. Deep Eutectic Solvents as Efficient
Solvents in Biocatalysis. Trends Biotechnol. 2019, 37, 943–959. [CrossRef]
21. Oh, Y.; Park, S.; Yoo, E.; Jo, S.; Hong, J.; Kim, H.J.; Kim, K.J.; Oh, K.K.; Lee, S.H. Dihydrogen-bonding deep
eutectic solvents as reaction media for lipase-catalyzed transesterification. Biochem. Eng. J. 2019, 142, 34–40.
[CrossRef]
22. Jablonsk˛, M.; Škulcová, A.; Šima, J. Use of Deep Eutectic Solvents in Polymer Chemistry–A Review.
Molecules 2019, 24, 3978. [CrossRef] [PubMed]
23. Roda, A.; Matias, A.A.; Paiva, A.; Duarte, A.R.C. Polymer Science and Engineering Using Deep Eutectic
Solvents. Polymers 2019, 11, 912. [CrossRef] [PubMed]
24. Gómez, A.V.; Biswas, A.; Tadini, C.C.; Furtado, R.F.; Alves, C.R.; Cheng, H.N. Use of natural deep eutectic
solvents for polymerization and polymer reactions. J. Braz. Chem. Soc. 2019, 30, 717–726. [CrossRef]
25. Abbott, A.P.; Ahmed, E.I.; Prasad, K.; Qader, I.B.; Ryder, K.S. Liquid pharmaceuticals formulation by eutectic
formation. Fluid Phase Equilib. 2017, 448, 2–8. [CrossRef]
26. Aroso, I.M.; Craveiro, R.; Rocha, Â.; Dionísio, M.; Barreiros, S.; Reis, R.L.; Paiva, A.; Duarte, A.R.C. Design
of controlled release systems for THEDES—Therapeutic deep eutectic solvents, using supercritical fluid
technology. Int. J. Pharm. 2015, 492, 73–79. [CrossRef]
27. Aroso, I.M.; Silva, J.C.; Mano, F.; Ferreira, A.S.; Dionísio, M.; Sá-Nogueira, I.; Barreiros, S.; Reis, R.L.; Paiva, A.;
Duarte, A.R.C. Dissolution enhancement of active pharmaceutical ingredients by therapeutic deep eutectic
systems. Eur. J. Pharm. Sci. 2016, 98, 57–66. [CrossRef]
28. Jeong, K.M.; Ko, J.; Zhao, J.; Jin, Y.; Yoo, D.E.; Han, S.Y.; Lee, J. Multi-functioning deep eutectic solvents as
extraction and storage media for bioactive natural products that are readily applicable to cosmetic products.
J. Clean. Prod. 2017, 151, 87–95. [CrossRef]
29. Van Osch, D.J.G.P.; Dietz, C.H.J.T.; van Spronsen, J.; Kroon, M.C.; Gallucci, F.; van Sint Annaland, M.;
Tuinier, R. A Search for Natural Hydrophobic Deep Eutectic Solvents Based on Natural Components. ACS
Sustain. Chem. Eng. 2019, 7, 2933–2942. [CrossRef]
30. Francisco, M.; van den Bruinhorst, A.; Kroon, M.C. New natural and renewable low transition temperature
mixtures (LTTMs): Screening as solvents for lignocellulosic biomass processing. Green Chem. 2012, 14,
2153–2157. [CrossRef]
31. Van den Bruinhorst, A.; Kollau, L.J.B.M.; Kroon, M.C.; Meuldijk, J.; Tuinier, R.; Esteves, A.C.C. A centrifuge
method to determine the solid–liquid phase behavior of eutectic mixtures. J. Chem. Phys. 2018, 149, 224505.
[CrossRef]
32. Silva, L.P.; Fernandez, L.; Conceição, J.H.F.; Martins, M.A.R.; Sosa, A.; Ortega, J.; Pinho, S.P.; Coutinho, J.A.P.
Design and Characterization of Sugar-Based Deep Eutectic Solvents Using Conductor-like Screening Model
for Real Solvents. ACS Sustain. Chem. Eng. 2018. [CrossRef]
33. Rodriguez Rodriguez, N.; van den Bruinhorst, A.; Kollau, L.J.B.M.; Kroon, M.C.; Binnemans, K. Degradation
of Deep-Eutectic Solvents Based on Choline Chloride and Carboxylic Acids. ACS Sustain. Chem. Eng. 2019,
7, 11521–11528. [CrossRef]
34. Abranches, D.O.; Larriba, M.; Silva, L.P.; Melle-Franco, M.; Palomar, J.F.; Pinho, S.P.; Coutinho, J.A.P. Using
COSMO-RS to design choline chloride pharmaceutical eutectic solvents. Fluid Phase Equilib. 2019, 497, 71–78.
[CrossRef]
35. Wolbert, F.; Brandenbusch, C.; Sadowski, G. Selecting Excipients Forming Therapeutic Deep Eutectic
Systems-A Mechanistic Approach. Mol. Pharm. 2019, 16, 3091–3099. [CrossRef] [PubMed]
Molecules 2020, 25, 1077 11 of 11

36. Kollau, L.J.B.M.; Vis, M.; van den Bruinhorst, A.; Esteves, A.C.C.; Tuinier, R. Quantification of the liquid
window of deep eutectic solvents. Chem. Commun. 2018, 54, 13351–13354. [CrossRef] [PubMed]
37. Martins, M.A.R.; Pinho, S.P.; Coutinho, J.A.P. Insights into the Nature of Eutectic and Deep Eutectic Mixtures.
J. Solut. Chem. 2018. [CrossRef]
38. Alhadid, A.; Mokrushina, L.; Minceva, M. Modeling of Solid-Liquid Equilibria in Deep Eutectic Solvents: A
Parameter Study. Molecules 2019, 24. [CrossRef]
39. Williams, D.H.; O’Brien, D.P.; Bardsley, B. Enthalpy/Entropy Compensation as a Competition between
Dynamics and Bonding: The Relevance to Melting of Crystals and Biological Aggregates. J. Am. Chem. Soc.
2001, 123, 737–738. [CrossRef]
40. Jain, A.; Yalkowsky, S.H. Estimation of Melting Points of Organic Compounds-II. J. Pharm. Sci. 2006, 95,
2562–2618. [CrossRef]
41. Yalkowsky, S.H. Carnelley’s Rule and the Prediction of Melting Point. J. Pharm. Sci. 2014, 103, 2629–2634.
[CrossRef] [PubMed]
42. Dearden, J.C. The QSAR prediction of melting point, a property of environmental relevance. Sci. Total
Environ. 1991, 109, 59–68. [CrossRef]
43. Brown, R.J.C.; Brown, R.F.C. Melting Point and Molecular Symmetry. J. Chem. Educ. 2000, 77, 724. [CrossRef]
44. Jain, A.; Yang, G.; Yalkowsky, S.H. Estimation of Total Entropy of Melting of Organic Compounds. Ind. Eng.
Chem. Res. 2004, 43, 4376–4379. [CrossRef]
45. Preiss, U.P.; Beichel, W.; Erle, A.M.; Paulechka, Y.U.; Krossing, I. Is universal, simple melting point prediction
possible? Chemphyschem 2011, 12, 2959–2972. [CrossRef] [PubMed]
46. Bondi, A. Physical Properties of Molecular Crystals, Liquids, and Glasses; John Wiley and Sons: New York, NY,
USA, 1968.
47. Crespo, E.A.; Silva, L.P.; Martins, M.A.R.; Bülow, M.; Ferreira, O.; Sadowski, G.; Held, C.; Pinho, S.P.;
Coutinho, J.A.P. The Role of Polyfunctionality in the Formation of [Ch]Cl-Carboxylic Acid-Based Deep
Eutectic Solvents. Ind. Eng. Chem. Res. 2018, 57, 11195–11209. [CrossRef]
48. Crespo, E.A.; Silva, L.P.; Martins, M.A.R.; Fernandez, L.; Ortega, J.; Ferreira, O.; Sadowski, G.; Held, C.;
Pinho, S.P.; Coutinho, J.A.P. Characterization and Modeling of the Liquid Phase of Deep Eutectic Solvents
Based on Fatty Acids/Alcohols and Choline Chloride. Ind. Eng. Chem. Res. 2017, 56, 12192–12202. [CrossRef]
49. Martins, M.A.R.; Crespo, E.A.; Pontes, P.V.A.; Silva, L.P.; Bülow, M.; Maximo, G.J.; Batista, E.A.C.; Held, C.;
Pinho, S.P.; Coutinho, J.A.P. Tunable Hydrophobic Eutectic Solvents Based on Terpenes and Monocarboxylic
Acids. ACS Sustain. Chem. Eng. 2018, 6, 8836–8846. [CrossRef]
50. Domańska, U.; Morawski, P.; Piekarska, M. Solubility of perfumery and fragrance raw materials based on
cyclohexane in 1-octanol under ambient and high pressures up to 900MPa. J. Chem. Thermodyn. 2008, 40,
710–717. [CrossRef]
51. Pontes, P.V.A.; Crespo, E.A.; Martins, M.A.R.; Silva, L.P.; Neves, C.M.S.S.; Maximo, G.J.; Hubinger, M.D.;
Batista, E.A.C.; Pinho, S.P.; Coutinho, J.A.P.; et al. Measurement and PC-SAFT modeling of solid-liquid
equilibrium of deep eutectic solvents of quaternary ammonium chlorides and carboxylic acids. Fluid Phase
Equilib. 2017, 448, 69–80. [CrossRef]
52. Monte, M.J.S.; Hillesheim, D.M. Thermodynamic study on the sublimation of 3-phenylpropionic acid and of
three methoxy-substituted 3-phenylpropionic acids. J. Chem. Thermodyn. 2001, 33, 837–847. [CrossRef]

Sample Availability: Samples of the compounds are no longer available from the authors.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
 Supplementary Material
Design of Deep Eutectic Systems: A Simple Approach
for Preselecting Eutectic Mixture Constituents
Ahmad Alhadid 1, Liudmila Mokrushina 2 and Mirjana Minceva 1,*
1 Technical University of Munich, Biothermodynamics, TUM School of Life and Food Sciences Weihenstephan,
Maximus-von-Imhof-Forum 2, 85354 Freising, Germany; [email protected]; [email protected]
2 Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Separation Science & Technology, Egerlandstr.

3, 91058 Erlangen, Germany; [email protected].

* Correspondence: [email protected]

Figure S1. Solid–liquid phase diagrams of binary eutectic systems consist of L-menthol and (A)
cyclohexylpropionic acid (B) caprylic acid (C) cyclohexanecarboxylic acid (D) capric acid (E) phenylpropionic acid
(F) lauric acid. Dashed lines are ideal liquidus lines of components modeled using the Schöder–van–Laar equation
and using experimental melting properties.
 L-menthol/3-cyclohexylpropionic acid L-menthol/caprylic acid

xacid= 0.1

xacid= 0.2

xacid= 0.3

xacid= 0.4

xacid= 0.5

xacid= 0.6

xacid= 0.7

xacid= 0.8

xacid= 0.9

Figure S2. DSC curves of L-menthol/3-cyclohexylpropionic acid and L-menthol/caprylic acid systems
 L-menthol/cyclohexanecarboxylic acid L-menthol/capric acid

xacid= 0.1

xacid= 0.2

xacid= 0.3

xacid= 0.4

xacid= 0.5

xacid= 0.6

xacid= 0.7

xacid= 0.8

xacid= 0.9

Figure S3. DSC curves of L-menthol/cyclohexanecarboxylic acid and L-menthol/capric acid systems
 L-menthol/3-phenylpropionic acid L-menthol/lauric acid

xacid= 0.1

xacid= 0.2

xacid= 0.3

xacid= 0.4

xacid= 0.5

xacid= 0.6

xacid= 0.7

xacid= 0.8

xacid= 0.9

Figure S4. DSC curves of L-menthol/3-phenylpropionic acid and L-menthol/lauric acid systems
3.3 Paper III

3.3 Paper III

Formation of glassy phases and polymorphism in deep eutectic solvents

A. Alhadid, L. Mokrushina and M. Minceva, J. Mol. Liq., 2020, 314, 113667.

Author contribution: The thesis author conceptualized the paper’s idea, performed the
investigations and formal analysis, interpreted the results, and wrote the manuscript.

Summary: The goal of Paper III is to study and highlight the influence of glass formation and
polymorphism on measuring and modeling SLE in binary eutectic systems. L-menthol-based
eutectic systems was selected because L-menthol is known to have two main polymorphs, the
stable α-polymorph and the metastable β-polymorph. Moreover, several studies on SLE in L-
menthol-based eutectic systems have not reported the eutectic temperature but rather only the
glass transition temperature of the mixture.

In Paper III, the SLE data for three eutectic systems containing L-menthol with linear
monocarboxylic acids were measured using two different sample preparation methods: in-situ
crystallization during the DSC run and annealing the DSC crucibles for one month prior to DSC
analysis. A comparison between the experimental data reported in Paper III and the literature
showed that in the letter, solidus peaks corresponding to the metastable polymorph were
misinterpreted as solidus peaks for the stable polymorph. In Paper III, the observed solidus
peaks were assigned to the corresponding polymorph by thermodynamic modeling of the SLE
data using the two polymorphs’ melting properties. It was concluded that the set of activity
coefficients model parameters fitted to the stable polymorph could be used to predict the phase
diagram of the metastable polymorph. Next, the glass formation in L-menthol-based eutectic
mixtures was studied. Annealing the samples for one month before DSC analysis aided the
crystallization, allowing for detecting the solidus peak for the L-menthol/thymol and L-
menthol/carvacrol systems. It was found that the observed glass transition temperature of the
mixture does not depict the eutectic temperature observed for the system. Moreover, glass
formation is not always attributed to strong intermolecular interactions, as glass formation was
observed in ideal eutectic mixtures. Further experimental investigations showed that glass
transition is attributed to the influence of the cyclohexyl ring in L-menthol and cyclohexyl
carboxylic acid as well as the nonideality in the liquid phase, which both contribute to the high
viscosity of the mixture. 148

80
3.3 Paper III

Reprinted with permission from Alhadid et al., J. Mol. Liq., 2020, 314, 113667. DOI:
10.1016/j.molliq.2020.113667. Copyright 2020 Elsevier.

81
 Journal of Molecular Liquids 314 (2020) 113667

Contents lists available at ScienceDirect

Journal of Molecular Liquids

journal homepage: www.elsevier.com/locate/molliq

Formation of glassy phases and polymorphism in deep eutectic solvents

Ahmad Alhadid a, Liudmila Mokrushina b, Mirjana Minceva a,⁎
a
Biothermodynamics, TUM School of Life Sciences Weihenstephan, Technical University of Munich, Germany
b
Separation Science & Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany

a r t i c l e i n f o a b s t r a c t

Article history: Deep eutectic solvents (DESs) are a class of eutectic mixtures that have very low melting temperature at the eu-
Received 23 April 2020 tectic point, which limits crystallization or lead to formation of metastable polymorphs. This can produce a false
Received in revised form 15 June 2020 estimate of the actual melting temperature of the mixture. The present work focuses on the formation of meta-
Accepted 22 June 2020
stable phases in eutectic systems containing L-menthol. Differential scanning calorimetry (DSC) was applied to
Available online 25 June 2020
measure the solid–liquid equilibria (SLE) of the eutectic mixtures. Two sample-preparation methods were
Keywords:
employed, namely annealing and in situ crystallization during the DSC run. We found that the eutectic tempera-
Deep eutectic solvents ture of the stable mixture is much higher than the observed glass-transition temperature of the mixture. More-
Solid–liquid equilibria over, the eutectic temperature of a mixture with a stable polymorph is different from that of a metastable
Glass-transition temperature polymorph. The measured SLE data were correlated through non-ideality modeling using the non-random
Polymorphism two-liquid (NRTL) equation. Our results show that SLE modeling is a useful tool for predicting the melting tem-
peratures of stable mixtures without the need for time-consuming annealing methods. This work serves as a
guide for reporting stable mixture properties when dealing with eutectic mixtures that form metastable phases.
© 2020 Elsevier B.V. All rights reserved.

1. Introduction SLE measurements in eutectic systems are not always straightfor-

ward. In many cases, no crystallization but a glass transition is observed
Green solvents are sought to replace conventional solvents in chem- on the DSC curve of a eutectic mixture near the eutectic point [22]. Such
ical and process applications in the modern environmental-protection eutectic mixtures are sometimes called Low-Transition Temperature
framework. Eutectic mixtures can be used as solvent systems that com- Mixtures (LTTM) [23,24]. However, the glass-transition temperature is
ply with the twelve principles of green chemistry [1]. Deep eutectic sol- not a first-order transition that occurs at a specific temperature [25].
vents (DESs) are mixtures of two or more compounds that are capable The glass transition is the transition of the amorphous solid to a
to form a liquid mixture with a melting temperature considerably subcooled liquid with increasing temperature. Amorphous solids and
lower than that of the pure compounds [2,3]. The DES compounds can subcooled liquids are both metastable phases; upon further cooling or
be biodegradable, nontoxic, inexpensive, and possess a very low vapor heating, transformation to a stable crystalline solid or a stable liquid, re-
pressure. Eutectic solvents have been used as solvents in separation spectively, occurs. When a glass phase is formed near the eutectic point,
methods [4–8] and in chemical and biochemical reactions [9–16]. measuring the eutectic temperature of the mixture is not a
Knowledge of the solid–liquid equilibria (SLE) of eutectic mixtures is straightforward task.
essential in order to design DESs for specific applications [17–20]. SLE Polymorphism refers to the existence of a pure component in differ-
data of eutectic mixtures are usually obtained through visual methods ent solid forms under different conditions [26]. Different polymorphs
such as a melting point device, or by thermal analysis methods such as have different melting properties. Transformation from one polymorph
differential scanning calorimetry (DSC). Other techniques have been to another can occur under certain conditions. L-menthol, which has
proposed in the literature, such as the centrifuge method for determin- been used to form eutectic mixtures, offers an interesting example in
ing the solid–liquid behavior in eutectic mixtures [21]. However, all this respect. L-menthol has four identified polymorphs: α, β, γ, and δ
these methods are time consuming and/or require special equipment. [27,28]. Corvis et al. [29] showed that the metastable β-polymorph of
Therefore, the amount of available SLE data for eutectic systems is low L-menthol can exist for sufficiently long time, so that its melting proper-
compared to the number of eutectic mixtures studied, proposed, or po- ties can be measured.
tentially useful. In this work, we investigate eutectic mixtures that show ki-
netic limitations in crystallization, polymorphism, and the forma-
⁎ Corresponding author. tion of metastable phases. The goal is to show that proper sample
E-mail address: [email protected] (M. Minceva). preparation and careful analysis of SLE data are essential in order

https://doi.org/10.1016/j.molliq.2020.113667
0167-7322/© 2020 Elsevier B.V. All rights reserved.
2 A. Alhadid et al. / Journal of Molecular Liquids 314 (2020) 113667

to avoid reporting metastable mixture properties. We use DSC to cycle. The DSC curves of the annealed samples were obtained in the
measure the SLE of selected eutectic mixtures using appropriate first heating cycle. The DSC curves of the second heating cycle repre-
methods of sample preparation. Furthermore, the non-ideality sented the in situ crystalized samples (common practice in DSC
of the mixture components in the liquid phase is modeled measurements).
using the non-random two-liquid (NRTL) equation on the mea- The DSC curves were analyzed using the NETZSCH Proteus soft-
sured SLE data. ware, version 6.1. The pure components' melting temperatures and
the eutectic temperatures were determined as the onset tempera-
2. Materials and method ture of the associated peak. The liquidus temperatures were taken
as the maximum temperature of the peak associated with the re-
2.1. Sample preparation spective liquidus thermal event. The reported values for each eutec-
tic mixture with a specific composition were the average values of
The chemicals used to prepare the eutectic mixtures in this work are the three DSC samples. The standard uncertainties of the DSC were
listed in Table 1. L-menthol (purity ≥ 99%) and carvacrol (purity 99%) 0.1 K and 3% for the measured temperatures and the transition en-
were purchased from Sigma Aldrich. Caprylic acid (purity 99%) and thalpies, respectively.
lauric acid (purity 99%) were purchased from Merck. Thymol (pu-
rity N 99%) was purchased from VWR International. Capric acid (purity 2.3. Thermodynamic modeling
99%) was purchased from Alfa Aesar. All chemicals were used as re-
ceived without further purification. The water content of pure compo- The SLE phase diagram of a simple eutectic system, i.e., whose com-
nents was measured in triplicates using Karl-Fischer coulometric ponents show negligible mutual solubility in the solid phase, can be cal-
titrator (Hanna instrument, USA). The results are listed in Table 1. In culated as follows [32].
the same table, the measured melting properties are compared with ! " ! "
data from the literature. ∆hm,i T ∆cP T T
ln xLi γLi ¼ − 1− − 1− m,i þ ln m,i ð1Þ
Each eutectic mixture was prepared by weighing the pure compo- RT T m,i R T T
nents in a glass vessel using a balance with a precision of 1 × 10−4 g
(Sartorius, Germany). The vessel was closed directly after introducing where xLi and γLi are the mole fraction and the activity coefficient of com-
the components to avoid possible sublimation/evaporation and mois- ponent i in the liquid solution, respectively; ∆hm, i and Tm, i are the melt-
ture absorption. The sample was heated to 318.15 K under continuous ing enthalpy and the melting temperature of pure component i,
stirring until a homogenous clear liquid was obtained. At least nine sam- respectively; R is the universal gas constant; T is the liquidus tempera-
ples were prepared for each eutectic mixture, covering the mole fraction ture; and ∆cP, i is the difference in heat capacities of the pure component
range from 0.1 to 0.9. The liquid mixtures were introduced into DSC alu- i between the solid state and the liquid state at constant pressure. The
minum crucibles using syringes. The DSC samples were prepared in second term on the right-hand side of this equation is usually of a low
triplicates for each mixture composition. The DSC samples of all eutectic value compared to the first term, and it therefore is usually neglected
systems were then transferred directly to 193.15 K and kept at this tem- [33]. Hence, the following simplified equation is typically used to calcu-
perature overnight. The DSC samples were stored for at least one month late the SLE
at constant temperature, i.e., annealing. The annealing temperature for ! "
the DSC samples of the acid-based eutectic mixtures, namely L- ∆hm,i T
ln xLi γLi ¼ − 1− ð2Þ
menthol/caprylic acid, L-menthol/capric acid, and L-menthol/lauric RT T m,i
acid, was 253.15 K, whereas for the L-menthol/thymol and L-menthol/
carvacrol systems, it was 193.15 K. The activity coefficients can be calculated using excess Gibbs energy
(gE) models or the thermal equation of state. In this work, the NRTL
2.2. Differential scanning calorimetry equation was used to calculate the activity coefficients of component i
in the liquid phase as follows [32].
Thermal characterization was carried out using DSC (NETZSCH
" ! "2 #
DSC 200 F3, Germany). The device was calibrated with five calibra- Gji τij Gij
tion standards: adamantane, bismuth, indium, zinc, and tin. The ln γi ¼ x2j τji þ# $2 ð3Þ
xi þ xj Gji xj þ xi Gij
measurements were done under a continuous nitrogen flow of
150 ml min−1. # $ # $
The DSC chamber was precooled to around 50 K below the eutec- Gij ¼ exp –ατ ij Gji ¼ exp –ατ ji ð4Þ
tic temperature of the mixture. Then, the annealed samples were in-
troduced into the DSC chamber. A heating cycle with a rate of g ij −g jj gji −g ii
τij ¼ τji ¼ ð5Þ
5 K min−1 was conducted until we achieved a temperature of ap- RT RT
proximately 10 K higher than the liquidus temperature of the sam-
ple. Next, a cooling cycle with a cooling rate of 5 K min−1 was run The value of the non-randomness parameter α was set to 0.3. The bi-
until 193.15 K on the same sample, followed by a second heating nary interaction parameters (gij − gjj) and (gji − gii) were fitted to

Table 1
Chemicals used to prepare the eutectic mixtures, including their water content, and their melting temperatures and enthalpies measured in this work and available from the literature.

Name Water content/mg g−1 Tm/K Δhm/kJ mol−1

This work Lit. This work Lit.

L-Menthol 0.328 ± 0.035 – 314.6 [30] – 13.74 [30]

Thymol 0.0749 ± 0.0093 322.70 ± 0.10 323.9 [31] 20.64 ± 0.41 19.65 [31]
Carvacrol 4.158 ± 0.011 274.20 ± 0.20 – 11.49 ± 0.29 –
Caprylic acid 1.253 ± 0.019 – 288.0 [30] – 21.43 [30]
Capric acid 0.167 ± 0.028 – 303.9 [30] – 28.39 [30]
Lauric acid 0.084 ± 0.010 – 316.6 [30] – 35.81 [30]
 A. Alhadid et al. / Journal of Molecular Liquids 314 (2020) 113667 3

experimental liquidus temperatures by minimizing the following objec- menthol/thymol system, the eutectic peak can only be observed if the
tive function samples have been annealed before the DSC measurements. As seen in
Fig. 1, the difference between the glass-transition peak and the onset
0% &2 112 of the eutectic peak is around 30 K. Moreover, the difference between
exp cal
B T i −T i C the onset temperature and the peak maximum temperature is around
F ðT Þ ¼ ∑@ A ð6Þ
n 22 K. Fig. 2 shows the SLE data of the annealed samples measured in
this work compared to the data reported in Abranches et al. [22]. The
liquidus temperatures measured in this work are in good agreement
where Texp
i and Tcal
i are the experimental and calculated liquidus tem- with those reported by Abranches et al. [22].
peratures, respectively; and n is the number of experimental points. The experimental eutectic composition can be determined from the
The model was used to calculate the eutectic composition as the inter- Tammann plot. However, it is impossible to make the Tammann plot if
cept of the two liquidus lines, since it is impossible to construct metastable phases occur in the eutectic thermal event [34]. Alterna-
Tammann plots for systems with metastable phases. tively, the eutectic composition can be determined by modeling the ac-
tivity coefficients of the components in the liquid phase. In this work,
3. Results and discussion we used the NRTL model (Eqs. (3)–(5)) to model the activity coeffi-
cients of the components in the liquid phase, and the SLE was calculated
We investigate the formation of metastable phases, either glassy according to Eq. (2). Fig. 3 shows the experimental and calculated SLE
phases or metastable polymorphs, in several eutectic mixtures. Two data of the L-menthol/thymol eutectic system. The eutectic mixture
sample-preparation methods were applied: annealing for a long time shows a strong negative deviation from the ideal behavior. The NRTL
at a temperature below the eutectic temperature of the mixture, and model with two fitted parameters provides a good representation of
in situ crystallization in the DSC run. The measured DSC curves and eu- the measured SLE data of the L-menthol/thymol system. Moreover, the
tectic temperatures of annealed and in situ crystalized DSC samples eutectic temperature predicted by the NRTL model (241.46 K) is in
were compared in each case. Next, the non-ideality of the components good agreement with the eutectic temperature determined from the
in the liquid phase was modeled using the NRTL equation, and the onset of the eutectic peak of the annealed samples (241.45 K). The eu-
liquidus curves of the components were calculated using Eq. (2). The tectic composition predicted by the NRTL model is xNRTL
e, thymol = 0.4326,
eutectic point was estimated from the intersection of the calculated slightly shifted toward L-menthol that has the lower melting enthalpy
liquidus lines. compared to thymol [18]. Abranches et al. [22] determined the eutectic

3.1. Glass formation

Several eutectic mixtures with very low glass-transition tempera-

tures have been reported in literature. Abranches et al. [22] reported
the glass-transition temperature of L-menthol/thymol near the eutectic
composition, which is between xthymol = 0.4 and 0.5, since no melting
transition could be observed as a result of kinetic limitations. In this
work, the L-menthol/thymol system was studied in detail. To measure
the melting temperature of the mixture near the eutectic point, the
samples were annealed before the DSC experiment. Fig. 1 shows the
DSC curves of the annealed (blue line) and the in situ crystalized
(dashed orange line) samples for the L-menthol/thymol system at
xthymol = 0.5000 (Fig. 1A) and xthymol = 0.4008 (Fig. 1B). The in situ
crystalized samples show only a glass transition at around 217 K. On
the other hand, an endothermic broad peak is observed in the annealed
samples, which indicates the transition of several metastable phases
[34]. The kinetic hindrance during the crystallization of L-menthol and
Fig. 2. Measured solid–liquid equilibria (SLE) data the of L-menthol/thymol system in
thymol near the eutectic point is the reason for the glassy phase forma- comparison to SLE data reported in Abranches et al. [22]. The eutectic temperature
tion. There have been reports in the literature on the kinetic hindrance (Te = 241.45 K) presented is an average value of temperatures determined at different
of other eutectic mixtures containing thymol [31,35]. In the L- composition.

Fig. 1. Differential scanning calorimetry (DSC) curves of the L-menthol/thymol system at (A) xthymol = 0.5000 and (B) xthymol = 0.4008. The DSC curves are shifted for clarity. The DSC
curves of in situ samples are scaled by a factor of five for clarity. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
4 A. Alhadid et al. / Journal of Molecular Liquids 314 (2020) 113667

Fig. 3. Measured solid–liquid equilibria (SLE) data of the L-menthol/thymol eutectic

system modeled assuming the ideal solution model and using the NRTL equation.

temperature by modeling the SLE data using the COSMO-RS model,

which predicted a very low eutectic temperature (226.90 K) with an eu-
COSMO
tectic composition of xe, thymol = 0.4654. Although the liquidus temper-
atures predicted by COSMO-RS were in good agreement with the
experimental data, the model underestimated the eutectic temperature
of the system.
To investigate whether or not the kinetic limitation of the L-
menthol/thymol system depends only on its strong negative deviation
from the ideal behavior, we studied the SLE of the L-menthol/carvacrol
eutectic mixture. Fig. 4 shows the chemical structures and melting
properties of carvacrol and thymol, which were measured in this
work. Although carvacrol and thymol are isomers, their melting proper-
ties are very different. Carvacrol is liquid at room temperature with a
melting temperature of 274.20 K. Moreover, the melting enthalpy of
carvacrol is almost half that of thymol. The lower melting enthalpy
and melting entropy might indicate a disordered crystal structure of
carvacrol. To the best of our knowledge, no crystallographic information
is available on carvacrol in the literature to confirm this.
Fig. 5 shows the DSC curves of the annealed samples of L-menthol/
carvacrol systems with different molar compositions. The top five DSC
curves in the orange box correspond to the mixtures that are rich in car-
vacrol, while the bottom two DSC curves correspond to mixtures that
are rich in L-menthol. As seen in Fig. 5, the eutectic peak is broad, similar
to that observed in the L-menthol/thymol system. Moreover, a glass
transition is observed at around Tg = 217.85 K. However, in mixtures
with a mole fraction of carvacrol lower than 0.8, the liquidus peak of car-
vacrol cannot be properly distinguished because of its overlapping with
the eutectic peak. In contrast, a melting peak corresponding to a crystal-
line phase of pure L-menthol is observed in the DSC curves of the mix-
tures rich in L-menthol (the DSC curves inside the green box in Fig. 5).
Fig. 6 shows the measured SLE data for the annealed samples as well
as the calculated liquidus lines using the ideal solution model and the Fig. 5. Differential scanning calorimetry (DSC) curves of the annealed L-menthol/carvacrol
mixtures with different mole fractions of carvacrol. The DSC curves in the orange and the
green boxes correspond to mixtures rich in carvacrol and L-menthol, respectively. The DSC
curves are shifted for clarity. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)

NRTL model. The NRTL model provides a good representation of the

measured SLE data of the L-menthol/carvacrol eutectic mixture. The eu-
tectic temperature predicted by the NRTL model (240.15 K) is in good
agreement with the onset of the eutectic peak in the annealed mixtures
(243.35 K). The eutectic composition predicted by the NRTL model is xe,
NRTL
carvacol = 0.5804. This shows that the eutectic composition is shifted
toward the component that possesses the lowest melting enthalpy, i.e.,
carvacrol (∆hm = 11.49 and 13.74 kJ mol−1 for carvacrol and L-menthol,
respectively).
Fig. 4. Chemical structures and melting properties of thymol and carvacrol measured in The L-menthol/thymol and L-menthol/carvacrol systems have simi-
this work. lar eutectic temperatures (Figs. 3 and 6). The deviation from the ideal
 A. Alhadid et al. / Journal of Molecular Liquids 314 (2020) 113667 5

deviation from ideality justifies the observed low glass-transition tem-

perature. On the other hand, crystallization at low temperatures can
lead to inefficient crystal packing in the crystalline phase of pure con-
stituents. In order to build the crystal lattice, molecules should adapt a
proper conformation that can be hindered at low temperatures. Sub-
stances with low melting enthalpies can also form deep eutectics.
Thus, the combined effect of the attraction between unlike molecules
and the inefficient crystal packing in the crystal structure of a pure com-
ponent can lead to the formation of glassy states. The latter case is ob-
served in the L-menthol/carvacrol, L-menthol/3–cyclohexylpropionic,
and L-menthol/cyclohexanecarboxylic acid systems. It can be concluded
that the formation of glassy phases at very low temperatures occurs not
only due to the strong negative deviations from the ideal behavior, i.e.,
strong attraction between unlike molecules, but also due to the crystal
Fig. 6. Measured solid–liquid equilibria (SLE) data of the L-menthol/carvacrol eutectic structure and melting properties of the pure components. Besides, the
system modeled assuming an ideal solution and using the NRTL equation. value of the glass transition temperature does not give a factual estimate
of the value of the eutectic temperature. However, in case of kinetic lim-
liquidus line of L-menthol in L-menthol/thymol (Fig. 3) is larger com- itation during crystallization near the eutectic point, the position of the
pared to that in L-menthol/carvacrol (Fig. 6). The thymol liquidus line latter can be estimated by SLE modeling using well-established thermo-
in the L-menthol/thymol system shows a negative deviation from the dynamic models.
ideal behavior (Fig. 3). In contrast, the carvacrol liquidus line shows a
quasi-ideal behavior (Fig. 6). This emphasizes that eutectic mixtures
with very low eutectic temperatures can be formed when their compo- 3.2. Polymorphism
nents possess low melting enthalpies, even if these components behave
ideally. As noticed from the DSC curves inside the orange box in Fig. 5, a As previously mentioned, L-menthol has four polymorphs. The melt-
glass transition (Tg ≈ 215.50 K) is observed in the mixtures rich in car- ing properties of L-menthol, which are reported in the literature and
vacrol, even if the liquidus line deviates from the ideal liquidus line only measured using common DSC practice, are for stable α-polymorph
slightly. [28,36]. However, Corvis et al. [29] showed that crystallization of L-
Kinetic hindrance during crystallization has also been noticed in menthol in the metastable β-polymorph is also possible, if the molten
other eutectic systems containing L-menthol [30]. Fig. 7 shows the DSC L-menthol is quenched at temperatures below 288.15 K. The β-
curves of annealed samples (blue curves) and in situ crystalized samples polymorph was stable for more than 8 h at 193.15 K, and all the β-
(orange dashed curves) of L-menthol/3–cyclohexylpropionic acid and L- polymorph transformed into the stable α-polymorph after 80 min at
menthol/cyclohexanecarboxylic acid near the eutectic composition of 288.15 K [29]. The melting properties measured by Corvis [29] were
each eutectic system. Crystallization at the eutectic point is hindered ∆hm = 14.1 and 11.0 kJ mol−1, and Tm = 316.05 and 308.45 K for the
in both systems. Although the glass transition is observed at lower tem- α-polymorph and the β-polymorph, respectively. In eutectic mixtures,
peratures (around 205 K) compared to those required in the L-menthol/ the mixture has a lower melting temperature than its pure components.
thymol and L-menthol/carvacrol systems (around 217 K), the eutectic Therefore, when in situ crystallization of the samples is performed in the
temperatures of L-menthol/3–cyclohexylpropionic acid (Te = DSC run, and if the eutectic temperature of the mixture is lower than
262.10 K) and L-menthol/cyclohexanecarboxylic acid (Te = 265.50 K) 288.15 K, L-menthol crystalizes at temperatures below its pure melting
are higher than those of L-menthol/thymol (Te = 241.45 K) and L- temperature. As a result, part of the solid L-menthol may crystalize in
menthol/carvacrol (Te = 243.35 K) systems. The SLE phase diagrams the form of the metastable β-polymorph.
for both eutectic systems are shown in Fig. S1 of the Supplementary ma- The SLE of several binary eutectic mixtures based on L-menthol can
terial. In L-menthol/3–cyclohexylpropionic acid, the L-menthol and 3– be found in the literature [36–38]. Corvis et al. [38] reported the phase
cyclohexylpropionic acid liquidus lines deviate negatively from the diagram of L-menthol/lidocaine, where L-menthol crystallized as the
ideal liquidus line. In contrast, the cyclohexanecarboxylic acid liquidus β-polymorph. The eutectic temperatures of mixtures of L-menthol as
line deviate positively from the ideal behavior. α-polymorph and β-polymorph differ by 12 K [38] due to different
Strong negative deviations from the ideal behavior can surely lead to melting properties of the two polymorphs.
a large temperature drop at the eutectic point and thus to kinetic hin- In the L-menthol/thymol and L-menthol/carvacrol systems, the
drance at low temperatures. In L-menthol/thymol, the strong negative metastable polymorphs cannot be observed explicitly in the in situ or

Fig. 7. Differential scanning calorimetry (DSC) curves of (A) the L-menthol/3–cyclohexylpropionic acid and (B) the L-menthol/cyclohexanecarboxylic acid at xacid = 0.5000. The DSC curves
are shifted for clarity. The DSC curves of the in situ samples are scaled by a factor of five for clarity.
6 A. Alhadid et al. / Journal of Molecular Liquids 314 (2020) 113667

the annealed samples. In the in situ measurements, only the glass tran- menthol/lauric acid (Te = 291.12 K) [30,37]. Therefore, we conclude
sition can be found near the eutectic point. Annealing leads to formation that, in eutectic systems containing L-menthol, L-menthol can crystalize
of the stable polymorph mostly. Thus, to study polymorphism, eutectic in the metastable β-polymorph if the samples are in situ crystalized in a
systems without kinetic limitations should be considered. In the present conventional DSC run.
study, we targeted the L-menthol eutectic systems with caprylic acid, Fig. 9 shows the SLE data of the annealed and in situ crystallized sam-
capric acid, and lauric acid. The DSC experiments were carried out for ples measured in this work in comparison to the SLE data reported by
the in situ crystallized samples as well as for the samples annealed in ad- Martins et al. [37]. We observe that the eutectic temperature of the in
vance for at least one month at 253.15 K. situ crystalized samples is lower than that of the annealed samples of
Fig. 8 shows the DSC curves of the annealed (blue curve) and the in the same mixture. The liquidus temperatures on the acid side are not
situ crystalized (orange dashed curve) samples of L-menthol/caprylic expected to differ between the in situ crystalized samples and the
acid (Fig. 8A), L-menthol/capric acid (Fig. 8B), and L-menthol/lauric annealed samples, because the liquidus line of the acid depends on
acid (Fig. 8C) systems near their respective eutectic points (xacid, the melting properties of the acid and the activity coefficient of the
e ≈ 0.5, 0.4, and 0.3 for L-menthol/caprylic acid, L-menthol/capric acid, acid in the liquid phase (see Eq. (2)). However, this should not be the
and L-menthol/lauric acid, respectively) [30]. The annealed and in situ case for the liquidus temperatures on the L-menthol side, as the melting
crystalized samples of the same mixture produce different DSC curves. properties of α- and β-polymorphs are different. The liquidus tempera-
Annealed samples (blue curves) show one large peak (peak maximum tures of the in situ crystalized samples are different from that of the
at T = 270.80 K, 283.15 K, and 292.84 K for L-menthol/caprylic acid, L- annealed samples only for L-menthol/caprylic acid (Fig. 9A) and L-
menthol/capric acid and L-menthol/lauric acid, respectively), whereas menthol/capric acid (Fig. 9B) at around xacid = 0.1000. The metastable
in situ crystalized samples (orange dashed curve) show two peaks polymorph can only be detected for the in situ crystalized samples at
(peak maximum at T = 266.45 K and 269.80 K, 278.08 K and the eutectic point or for the mixtures rich in L-menthol (Fig. 9). In Mar-
282.30 K, 287.45 K and 292.22 K, for L-menthol/caprylic acid, L- tins et al. [37], the samples were in situ crystalized, and the eutectic
menthol/capric acid and L-menthol/lauric acid, respectively). Annealing points reported in Martins et al. [37] were closer to those of the in situ
leads to almost complete disappearance of the first peak (Fig. 8), indi- crystalized samples measured in the present study (Fig. 9). The liquidus
cating that the first peak is related to a metastable polymorph of L-men- temperatures measured in this work are in good agreement with those
thol, supposedly the β-polymorph. Furthermore, the two peaks are reported in Martins et al. [37]. As can be seen in Fig. 9, the eutectic tem-
much closer in L-menthol/caprylic acid (Fig. 8A) compared to L- peratures of stable polymorph (annealed samples) are higher than
menthol/capric acid (Fig. 8B) and L-menthol/lauric acid (Fig. 8C). In ad- those of metastable polymorph (in situ crystalized samples). Thus, dif-
dition, the area of the first peak in Fig. 8A in L-menthol/caprylic acid is ferent sample preparation leads to different values of eutectic tempera-
larger compared to the second peak, whereas, the area of the two tures related either to the metastable polymorph, in case of the in-situ
peaks are almost the same in the L-menthol/capric acid (Fig. 8B) and measurements, or to the stable polymorph, if the samples are annealed
L-menthol/lauric acid (Fig. 8C) systems. This indicates that the amount prior to the DSC run.
of the metastable polymorph related to the first peak is greater in L- To confirm that the first peaks in Fig. 8 are the transition peaks at the
menthol/caprylic acid. Corvis et al. [29] showed that the ratio between eutectic point corresponding to metastable β-polymorph, x-ray diffrac-
the stable and metastable polymorphs in L-menthol depends on the tion measurements are needed. Alternatively, thermodynamic model-
quenching temperature. This justifies the relative position of the peaks ing can be used. The values of the activity coefficients depend on the
and the ratio between the peak areas observed in the present study, as composition and temperature of the solution and the latter is different
the eutectic temperature of L-menthol/caprylic acid (Te = 265.84 K) is for the stable and metastable polymorphs at liquidus conditions. The bi-
lower compared to L-menthol/capric acid (Te = 278.64 K) and L- nary interaction parameters of activity coefficient models, such as NRTL,

Fig. 8. Differential scanning calorimetry (DSC) curves of the (A) L-menthol/caprylic acid at xacid = 0.5058 (B) L-menthol/capric acid at xacid = 0.3939 (C) L-menthol/lauric acid xacid =
0.3002. The DSC curves are shifted for clarity.
 A. Alhadid et al. / Journal of Molecular Liquids 314 (2020) 113667 7

Fig. 9. Solid–liquid equilibria (SLE) in of (A) L-menthol/caprylic acid, (B) L-menthol/capric acid, and (C) L-menthol/lauric acid measured in this work for annealed and in situ crystalized
samples in comparison with SLE data reported by Martins et al. [37]. The eutectic temperature of annealed samples (Tannealed
e ) and in situ crystalized samples (Tin
e
situ
) presented are the
average values of temperatures determined at different compositions.

provide the values of activity coefficients of its components in the liquid polymorph, the metastable SLE phase diagram corresponding to the β-
phase in the whole composition range and in the appropriate range of polymorph can be calculated. Fig. 10 shows the calculated SLE of L-
temperatures (used for fitting the binary interaction parameters) [32], menthol/caprylic acid (Fig. 10A), L-menthol/capric acid (Fig. 10B), and
independent of whether such a solution is or is not in equilibrium L-menthol/lauric acid (Fig. 10C) eutectic systems modeled using the
with one or another phase. Therefore, the NRTL binary interactions NRTL equation and the melting properties of L-menthol in α- and β-
parameters (Eqs. (3)–(5)) fitted to the liquidus temperatures of the polymorphs, in comparison to the measured SLE data for the annealed
α-polymorph, i.e., the liquidus temperatures of the annealed samples, and in situ crystalized samples. We observe that the eutectic tempera-
can be used to calculate the activity coefficients of L-menthol in the ture measured for the in situ crystalized samples can be predicted
liquid phase. Then, using Eq. (2) and the melting properties of the β- from the melting properties of the β-polymorph and the activity

Fig. 10. Measured solid–liquid equilibria (SLE) in eutectic mixtures containing L-menthol with (A) caprylic acid, (B) capric acid, and (C) lauric acid compared to those modeled using the
NRTL equation.
8 A. Alhadid et al. / Journal of Molecular Liquids 314 (2020) 113667

coefficients calculated using the binary interaction parameters fitted to References

the α-polymorph liquidus temperatures. This confirms that the first
[1] P. Anastas, N. Eghbali, Green chemistry: principles and practice, Chem. Soc. Rev. 39
transition peak measured for the in situ crystalized samples corresponds (2010) 301–312, https://doi.org/10.1039/B918763B.
to the eutectic phase of the mixture with the metastable β-polymorph. [2] Q. Zhang, K. De Oliveira Vigier, S. Royer, F. Jerome, Deep eutectic solvents: syntheses,
Moreover, the liquidus temperatures determined at around xacid = 0.1 properties and applications, Chem. Soc. Rev. 41 (2012) 7108–7146, https://doi.org/
in L-menthol/caprylic acid (Fig. 10A) and L-menthol/capric acid 10.1039/C2CS35178A.
[3] E.L. Smith, A.P. Abbott, K.S. Ryder, Deep eutectic solvents (DESs) and their applica-
(Fig. 10B) are predicted well. It can be concluded that SLE modeling is tions, Chem. Rev. 114 (2014) 11060–11082, https://doi.org/10.1021/cr300162p.
a useful tool for predicting the SLE of components having different [4] T. Cai, H. Qiu, Application of deep eutectic solvents in chromatography: a review,
polymorphs. TrAC Trends Anal. Chem. 120 (2019) 115623, https://doi.org/10.1016/j.trac.2019.
115623.
[5] J. Huang, X. Guo, T. Xu, L. Fan, X. Zhou, S. Wu, Ionic deep eutectic solvents for the ex-
traction and separation of natural products, J. Chromatogr. A 1598 (2019) 1–19,
4. Conclusions https://doi.org/10.1016/j.chroma.2019.03.046.
[6] X. Li, K.H. Row, Development of deep eutectic solvents applied in extraction and
In this work, the formation of metastable phases in several eutectic separation, J. Sep. Sci. 39 (2016) 3505–3520, https://doi.org/10.1002/jssc.
201600633.
systems was studied in detail. SLE data were obtained using DSC with [7] I. Wazeer, M. Hayyan, M.K. Hadj-Kali, Deep eutectic solvents: designer fluids for
two different samples preparation methods, annealing and in situ crys- chemical processes, J. Chem. Technol. Biotechnol. 93 (2018) 945–958, https://doi.
tallization in the DSC run. For the in situ crystallized samples, glass tran- org/10.1002/jctb.5491.
[8] H.F. Hizaddin, M. Sarwono, M.A. Hashim, I.M. Alnashef, M.K. Hadj-Kali, Coupling the
sition or formation of metastable polymorphs was observed. The SLE
capabilities of different complexing agents into deep eutectic solvents to enhance
data were correlated by modeling the non-ideality of the components the separation of aromatics from aliphatics, J. Chem. Thermodyn. 84 (2015)
in the liquid phase using the NRTL equation. 67–75, https://doi.org/10.1016/j.jct.2014.12.024.
The eutectic temperature peak was not detected in the L-menthol/ [9] P. Xu, G.W. Zheng, M.H. Zong, N. Li, W.Y. Lou, Recent progress on deep eutectic sol-
vents in biocatalysis, Bioresources Bioprocess. 4 (2017) 34, https://doi.org/10.1186/
thymol and L-menthol/carvacrol systems by applying DSC on the in s40643-017-0165-5.
situ crystallized samples. However, the samples annealed in advance [10] B.L. Gadilohar, G.S. Shankarling, Choline based ionic liquids and their applications in
for a sufficient amount of time showed a eutectic peak. We propose organic transformation, J. Mol. Liq. 227 (2017) 234–261, https://doi.org/10.1016/j.
molliq.2016.11.136.
that a possible reason for the low glass-transition temperature of the
[11] S. Khandelwal, Y.K. Tailor, M. Kumar, Deep eutectic solvents (DESs) as eco-friendly
in situ crystalized samples may be the very low eutectic temperature and sustainable solvent/catalyst systems in organic transformations, J. Mol. Liq.
of the system or the inefficient crystal packing in the crystalline phase 215 (2016) 345–386, https://doi.org/10.1016/j.molliq.2015.12.015 (Review).
of the pure constituents. The low eutectic temperature can be a result [12] D.A. Alonso, A. Baeza, R. Chinchilla, G. Guillena, I.M. Pastor, D.J. Ramón, Deep eutectic
solvents: the organic reaction medium of the century, Eur. J. Org. Chem. 2016
of strong negative deviations from the ideal behavior or low melting en- (2016) 612–632, https://doi.org/10.1002/ejoc.201501197.
thalpy values. The latter can also indicate inefficient crystal packing of [13] J. García-Álvarez, Deep eutectic mixtures: promising sustainable solvents for metal-
the pure constituents. catalysed and metal-mediated organic reactions, Eur. J. Inorg. Chem. 2015 (2015)
5147–5157, https://doi.org/10.1002/ejic.201500892.
Modeling of activity coefficients of components in the liquid phase
[14] P. Liu, J.W. Hao, L.P. Mo, Z.H. Zhang, Recent advances in the application of deep eu-
using NRTL equations provides good estimates of the eutectic tempera- tectic solvents as sustainable media as well as catalysts in organic reactions, RSC
tures and compositions of eutectic systems that show glass-transition Adv. 5 (2015) 48675–48704, https://doi.org/10.1039/c5ra05746a.
temperatures near their eutectic points. Therefore, in case no crystalized [15] M. Pätzold, S. Siebenhaller, S. Kara, A. Liese, C. Syldatk, D. Holtmann, Deep eutectic
solvents as efficient solvents in biocatalysis, Trends Biotechnol. 37 (2019)
eutectic phase can be detected, thermodynamic modeling offers an eas-
943–959, https://doi.org/10.1016/j.tibtech.2019.03.007.
ier and more time-efficient option for estimating the position of the eu- [16] A. Hayyan, M.A. Hashim, M. Hayyan, F.S. Mjalli, I.M. AlNashef, A new processing
tectic point compared to the time-consuming annealing method. route for cleaner production of biodiesel fuel using a choline chloride based deep
In case one of the eutectic system constituents possesses several eutectic solvent, J. Clean. Prod. 65 (2014) 246–251, https://doi.org/10.1016/j.
jclepro.2013.08.031.
polymorphic phases, care should be taken when analyzing the DSC [17] M.A.R. Martins, S.P. Pinho, J.A.P. Coutinho, Insights into the nature of eutectic and
curves. Eutectic peaks can be misinterpreted as liquidus peaks. Metasta- deep eutectic mixtures, J. Solut. Chem. 48 (2019) 962–982, https://doi.org/10.
ble and stable polymorphs have different melting properties that result 1007/s10953-018-0793-1.
[18] A. Alhadid, L. Mokrushina, M. Minceva, Modeling of solid–liquid equilibria in deep
in different phase diagrams. However, we demonstrated that the binary
eutectic solvents: a parameter study, Molecules 24 (2019) 2334, https://doi.org/
interaction parameters of activity coefficient models fitted to the SLE 10.3390/molecules24122334.
data measured for the stable polymorph can be used to predict the [19] I.I.I. Alkhatib, D. Bahamon, F. Llovell, M.R.M. Abu-Zahra, L.F. Vega, Perspectives and
SLE in the corresponding system with the metastable polymorph. guidelines on thermodynamic modelling of deep eutectic solvents, J. Mol. Liq. 298
(2020)https://doi.org/10.1016/j.molliq.2019.112183 112183.
We showed here that prior to designing process applications using
[20] L.J.B.M. Kollau, M. Vis, A. van den Bruinhorst, G. de With, R. Tuinier, Activity model-
DES, it is necessary to establish a better understanding of polymorphism ling of the solid–liquid equilibrium of deep eutectic solvents, Pure Appl. Chem. 91
and the formation of metastable phases in DES systems. This is essential (2019) 1341–1349, https://doi.org/10.1515/pac-2018-1014.
to ensure the recyclability of the solvent as well as the reproducibility of [21] A. van den Bruinhorst, L.J.B.M. Kollau, M.C. Kroon, J. Meuldijk, R. Tuinier, A.C.C.
Esteves, A centrifuge method to determine the solid–liquid phase behavior of eutec-
the reported properties of DES. tic mixtures, J. Chem. Phys. 149 (2018) 224505, https://doi.org/10.1063/1.5051515.
[22] D.O. Abranches, M.A.R. Martins, L.P. Silva, N. Schaeffer, S.P. Pinho, J.A.P. Coutinho,
Phenolic hydrogen bond donors in the formation of non-ionic deep eutectic sol-
CRediT authorship contribution statement vents: the quest for type V DES, Chem. Commun. (Camb.) 55 (2019)
10253–10256, https://doi.org/10.1039/c9cc04846d.
[23] E. Durand, J. Lecomte, P. Villeneuve, From green chemistry to nature: the versatile
Ahmad Alhadid: Investigation, Formal analysis, Writing - original role of low transition temperature mixtures, Biochimie 120 (2016) 119–123,
draft. Liudmila Mokrushina: Writing - review & editing. Mirjana https://doi.org/10.1016/j.biochi.2015.09.019.
[24] M. Francisco, A. van den Bruinhorst, M.C. Kroon, Low-transition-temperature mix-
Minceva: Writing - review & editing, Supervision.
tures (LTTMs): a new generation of designer solvents, Angew. Chem. Int. Ed. 52
Declaration of competing interest (2013) 3074–3085, https://doi.org/10.1002/anie.201207548.
[25] T. Koop, J. Bookhold, M. Shiraiwa, U. Poschl, Glass transition and phase state of or-
ganic compounds: dependency on molecular properties and implications for sec-
The authors declare no conflict of interest. ondary organic aerosols in the atmosphere, Phys. Chem. Chem. Phys. 13 (2011)
19238–19255, https://doi.org/10.1039/c1cp22617g.
Appendix A. Supplementary data [26] U.P. Preiss, W. Beichel, A.M. Erle, Y.U. Paulechka, I. Krossing, Is universal, simple
melting point prediction possible? Chemphyschem 12 (2011) 2959–2972, https://
doi.org/10.1002/cphc.201100522.
Supplementary data to this article can be found online at https://doi. [27] F.E. Wright, The crystallization of menthol, J. Am. Chem. Soc. 39 (1917) 1515–1524,
org/10.1016/j.molliq.2020.113667. https://doi.org/10.1021/ja02253a001.
 A. Alhadid et al. / Journal of Molecular Liquids 314 (2020) 113667 9

[28] Y. Corvis, A. Wurm, C. Schick, P. Espeau, New menthol polymorphs identified by [34] L. Rycerz, Practical remarks concerning phase diagrams determination on the basis
flash scanning calorimetry, CrystEngComm 17 (2015) 5357–5359, https://doi.org/ of differential scanning calorimetry measurements, J. Therm. Anal. Calorim. 113
10.1039/C5CE00697J. (2013) 231–238, https://doi.org/10.1007/s10973-013-3097-0.
[29] Y. Corvis, P. Négrier, S. Massip, J.-M. Leger, P. Espeau, Insights into the crystal struc- [35] F. Bezold, M. Minceva, Liquid-liquid equilibria of n-heptane, methanol and deep eu-
ture, polymorphism and thermal behavior of menthol optical isomers and race- tectic solvents composed of carboxylic acid and monocyclic terpenes, Fluid Phase
mates, CrystEngComm 14 (2012) 7055–7064, https://doi.org/10.1039/C2CE26025E. Equilib. 477 (2018) 98–106, https://doi.org/10.1016/j.fluid.2018.08.020.
[30] A. Alhadid, L. Mokrushina, M. Minceva, Design of deep eutectic systems: a simple [36] M.A.R. Martins, L.P. Silva, N. Schaeffer, D.O. Abranches, G.J. Maximo, S.P. Pinho, J.A.P.
approach for preselecting eutectic mixture constituents, Molecules 25 (2020) Coutinho, Greener terpene–terpene eutectic mixtures as hydrophobic solvents, ACS
1077, https://doi.org/10.3390/molecules25051077. Sustain. Chem. Eng. 7 (2019) 17414–17423, https://doi.org/10.1021/
[31] F. Wolbert, C. Brandenbusch, G. Sadowski, Selecting excipients forming therapeutic acssuschemeng.9b04614.
deep eutectic systems-a mechanistic approach, Mol. Pharm. 16 (2019) 3091–3099, [37] M.A.R. Martins, E.A. Crespo, P.V.A. Pontes, L.P. Silva, M. Bülow, G.J. Maximo, E.A.C.
https://doi.org/10.1021/acs.molpharmaceut.9b00336. Batista, C. Held, S.P. Pinho, J.A.P. Coutinho, Tunable hydrophobic eutectic solvents
based on terpenes and monocarboxylic acids, ACS Sustain. Chem. Eng. 6 (2018)
[32] J.M. Prausnitz, R.N. Lichtenthaler, E.G.d. Azevedo, Molecular Thermodynamics of
8836–8846, https://doi.org/10.1021/acssuschemeng.8b01203.
Fluid-phase Equilibria, Prentice Hall PTR, Upper Saddle River, NJ, 1999.
[38] Y. Corvis, P. Négrier, M. Lazerges, S. Massip, J.-M. Léger, P. Espeau, Lidocaine/l-
[33] S.H. Yalkowsky, M. Wu, Estimation of the ideal solubility (crystal–liquid fugacity menthol binary system: cocrystallization versus solid-state immiscibility, J. Phys.
ratio) of organic compounds, J. Pharm. Sci. 99 (2010) 1100–1106, https://doi.org/ Chem. B 114 (2010) 5420–5426, https://doi.org/10.1021/jp101303j.
10.1002/jps.21897.
 Supplementary Materials
Formation of Glassy Phases and Polymorphism in Deep Eutectic
Systems
Ahmad Alhadid 1, Liudmila Mokrushina 2, and Mirjana Minceva 1*
1
Biothermodynamics, TUM School of Life Sciences Weihenstephan, Technical University of Munich, Germany
2
Separation Science & Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany
*
Corresponding author e-mail: [email protected]

(A) (B)
330 330
3-cyclohexylpropionic acid cyclohexanecarboxylic acid

310 310

T/K
T/K

290 290

270 270

250 250
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x acid x acid

Figure S1. Solid–liquid equilibria data for eutectic mixture containing L-menthol with (A) 3–cyclohexylpropionic acid and
(B) cyclohexanecarboxylic acid. Data are taken from [1]. Legend: liquidus temperature; solidus temperature;
ideal liquidus line
Table S1. Experimental liquidus temperatures T , onset temperature of eutectic peak T , and calculated liquidus
temperatures T using NRTL model for L-menthol/thymol eutectic system.

xthymol T /K T /K T /K

0.8978 319.55 237.80 316.15

0.8003 308.15 239.90 306.19

0.7032 285.65 244.15 292.51

0.6015 278.55 239.80 276.52

0.5000 – 247.95 –

0.4008 – 242.55 –

0.3010 269.90 238.00 272.43

0.2019 293.30 241.45 291.35

0.1000 305.70 – 305.68

0.9035 317.95 – 316.61

0.8009 309.00 239.25 306.29

0.7011 286.30 237.50 292.26

0.6012 277.80 239.80 276.40

0.3011 271.80 241.85 272.58

0.0987 307.15 238.00 305.81

Table S2. Experimental liquidus temperatures T , onset temperature of eutectic peak T , and calculated liquidus
temperatures T using NRTL model for L-menthol/carvacrol eutectic system.

xcarvacrol T /K T /K T /K

0.8997 270.05 – 267.66

0.7996 263.05 245.25 260.51

0.6943 – 243.55 –

0.6019 – 244.25 –

0.5038 – 244.05 –

0.4013 266.25 244.95 267.88

0.2979 285.58 241.08 283.67

0.2030 296.35 242.70 295.80

0.0998 305.85 240.95 306.82

Table S3. Experimental liquidus temperatures T , onset temperature of eutectic peak T , and calculated liquidus
temperatures T using NRTL model for L-menthol/caprylic acid eutectic system.

xacid T /K T /K T /K

0.9000 287.70 263.35 286.64

0.8021 283.55 264.15 284.90

0.6974 279.55 266.65 280.55

0.6013 275.00 265.55 274.29

0.5058 – 266.75 –

0.4019 276.40 266.65 278.55

0.3025 286.90 265.60 290.07

0.1996 298.75 267.65 300.02

0.0997 308.50 266.20 307.99

0.9002 286.65 266.65 286.65

0.8001 284.35 265.60 284.84

0.7012 279.65 267.65 280.75

0.5999 275.60 266.20 274.22

0.4989 – 266.45 –

0.4035 277.15 263.90 278.39

0.3030 287.75 264.85 290.04

0.2001 299.40 263.95 299.98

0.0998 308.20 260.65 307.99

Table S4. Experimental liquidus temperatures T , onset temperature of eutectic peak T , and calculated liquidus
temperatures T using NRTL model for L-menthol/capric acid eutectic system.

xacid T /K T /K T /K

0.8990 302.85 276.95 302.14

0.8007 299.55 277.25 300.35

0.7039 296.10 278.55 297.35

0.6066 292.65 279.20 292.86

0.4937 287.20 279.55 285.76

0.3939 – 279.75 –

0.3003 289.55 279.25 290.88

0.1997 298.60 278.65 300.21

0.1015 307.10 – 307.91

0.8995 303.95 276.35 302.14

0.8025 300.15 276.50 300.40

0.7014 296.60 277.00 297.26

0.5995 292.30 277.50 292.47

0.4998 287.85 277.70 286.22

0.4037 – 278.10 –

0.3023 288.40 275.90 290.88

0.2016 298.85 275.85 300.06

0.1082 307.65 – 307.42

Table S5. Experimental liquidus temperatures T , onset temperature of eutectic peak T , and calculated liquidus
temperatures T using NRTL model for L-menthol/lauric acid eutectic system.

xacid T /K T /K T /K

0.8997 315.68 290.58 314.80

0.8010 312.45 291.35 313.14

0.6996 310.38 290.68 310.65

0.6007 306.58 289.95 307.15

0.4998 302.65 290.58 302.39

0.4007 296.88 291.18 296.33

0.3002 – 292.52 –

0.1999 297.88 291.35 300.49

0.1014 307.45 291.85 307.98

Table S6. NRTL binary interaction parameters and root-mean-square deviation (RMSD) between experimental liquidus
temperatures and calculated liquidus temperatures.

NRTL binary interaction parameters

Eutectic system g g g g RMSD* / K

L-menthol/thymol 4.3333 4.0648 3.5

L-menthol/carvacrol 3.1532 4.5610 1.9

L-menthol/caprylic acid 4.3718 11.1196 1.5

L-menthol/capric acid 4.2923 9.7161 1.2

L-menthol/lauric acid 4.0097 8.5889 1.2

∑ −
* 𝑅𝑀𝑆𝐷/ 𝐾
−
Table S7. Glass transition temperature Tg , average onset temperature of eutectic peak T , as well as eutectic temperature
T T and eutectic composition x , T calculated by NRTL of studied systems.

Eutectic system Tg / K T /K T T
/K x ,
T

L-menthol/thymol 217.55 241.45 241.46 0.4326

L-menthol/carvacrol 217.85 243.35 240.15 0.5804

L-menthol/3–cyclohexylpropionic acid 205.35 262.10 [1] – –

L-menthol/cyclohexanecarboxylic acid 204.65 265.00 [1] – –

L-menthol/caprylic acid α-polymorph 265.84 265.08 0.4992

–
-polymorph 258.53 259.05 0.4431

L-menthol/capric acid α-polymorph 278.64 278.82 0.4091

–
-polymorph 272.67 273.53 0.3531

L-menthol/lauric acid α-polymorph 291.12 289.91 0.3173

–
-polymorph 285.93 285.02 0.2637

References

1. Alhadid, A.; Mokrushina, L.; Minceva, M. Design of Deep Eutectic Systems: A Simple Approach for
Preselecting Eutectic Mixture Constituents. Molecules 2020, 25, 1077, doi:10.3390/molecules25051077.
3.4 Paper IV

3.4 Paper IV
Experimental Investigation and Modeling of Cocrystal Formation in L‑Menthol/Thymol
Eutectic System

A. Alhadid, C. Jandl, L. Mokrushina and M. Minceva, Cryst. Growth Des., 2021, 21, 6083-
6091.

Author contribution: The thesis author conceptualized the paper’s idea, performed the
investigations and formal analysis (excluding powder and SC-XRD), interpreted the results, and
wrote the manuscript.

Summary: Paper IV objective is to investigate cocrystal formation in the L-menthol/thymol

eutectic systems. The system was selected for two reasons. First, it is one of the most studied
nonionic DES in the literature. Second, the observed solidus peaks for the system in Paper III
were broad and asymmetric, hinting at the formation of several metastable phases. To overcome
glass and metastable phases formation and allow for combined DSC and powder XRD analyses,
an efficient sample preparation method was proposed. The liquid solutions obtained by mixing
and heating pure constituents in different ratios were quenched at 193 K for several hours, and
later stored at 253 K for several days until complete crystallization was observed. The obtained
solid was ground to a fine powder within a cold room at 253 K using mortar and pestle. DSC and
powder XRD analyses were performed on samples covering the entire composition range of the
mixture to obtain the SLE phase diagram and investigate cocrystal formation.

Paper IV showed that the phase diagram of the L-menthol/thymol eutectic system is more
100
complicated than previously anticipated, whether in Paper III or the literature. The formation
of two cocrystals and two solid solution regions was observed.

The heat capacity terms are usually neglected in the literature when modeling SLE in DES, mainly
due to unavailable data for the heat capacity of pure liquid and solid. To test this assumption in
the studied system, the three common approaches to account for the heat capacity terms were
evaluated. It was found that assuming a constant difference between pure liquid and solid heat
capacities calculated at the melting temperature provided the lowest deviation between
calculated and experimental liquidus data of pure constituents. Nevertheless, the influence of
the approach used to account for the heat capacity difference on the obtained NRTL interaction
parameters was insignificant. Therefore, if no experimental data are available on the heat

99
3.4 Paper IV

capacity of pure components in the solid and liquid states, neglecting the heat capacity terms
could be sufficient for modeling of SLE.

The SLE data were correlated using the NRTL model and considering the two cocrystals and the
partial miscibility in the solid phase. The two calculated eutectic temperatures using the NRTL
model were in good agreement with the experimentally observed ones, which demonstrates the
postulated stoichiometries and the melting properties of the two cocrystals as well as the ability
of the NRTL model to describe the nonideality of the components in the liquid solution.

100
3.4 Paper IV

Reprinted with permission from Alhadid et al., Cryst. Growth Des., 2021, 21, 6083-6091.
DOI:10.1021/acs.cgd.1c00306. Copyright 2021 American Chemical Society.

Article full text: http://pubs.acs.org/articlesonrequest/AOR-3FTJ2S7WIHUD5KXIKNZ2

101
 pubs.acs.org/crystal Article

Experimental Investigation and Modeling of Cocrystal Formation in

L‑Menthol/Thymol Eutectic System
Ahmad Alhadid,* Christian Jandl, Liudmila Mokrushina, and Mirjana Minceva
Cite This: Cryst. Growth Des. 2021, 21, 6083−6091 Read Online

ACCESS Metrics & More Article Recommendations *

sı Supporting Information

ABSTRACT: Solid−liquid equilibria (SLE) of the L-menthol/

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

thymol eutectic system were studied in detail using a combination

of differential scanning calorimetry (DSC) and powder X-ray
diffraction (XRD). The existence of two cocrystals with
stoichiometric ratios 1:3 and 3:2 for L-menthol:thymol was
Downloaded via TU MUENCHEN on May 24, 2022 at 07:40:20 (UTC).

monitored by performing XRD on samples of different

compositions. Moreover, the existence of two solid solution
regions of L-menthol in thymol and L-menthol in the 1:3 cocrystal
was observed. The nonrandom two-liquid (NRTL) and two-suffix
Margules models were applied to model the measured SLE data.
The two eutectic points of the system were determined at Tcal e1 =
271.7 K, xcal cal cal
thymol,e1 = 0.48 and Te2 = 273.1 K, xthymol,e2 = 0.33. The
complex character of the obtained phase diagram of the system
shows that not all deep eutectic systems can be assumed to be of a
simple eutectic type with immiscible solid phases. For the accurate
determination of the SLE of eutectic systems, a combination of several experimental techniques and thermodynamic modeling is
needed.

1. INTRODUCTION Δhm, i
∫T
T 1 T
ln xiLγi L = − 1− − Δc p(T ) dT
Deep eutectic solvents (DES) have emerged as a new class of RT Tm, i RT m
designer solvents. Prepared from natural, nontoxic, and
Δc p(T )
∫T
1 T
sustainable components, they have been considered a greener + dT
alternative to ionic liquids.1,2 Several studies have shown that R m T (1)
DES can be used in several applications such as drug
delivery,3,4 liquid chromatography,5,6 and reaction media.7 A where xLi and γLi are the mole fraction and activity coefficient of
considerable amount of research has been dedicated to component i in the liquid solution, respectively, Δhm,i and Tm,i
studying intermolecular interactions in the liquid phase using are the melting enthalpy and melting temperature of pure
molecular simulation or thermodynamic models.8,9 The component i, respectively, R is the universal gas constant; T is
motivation for exclusively studying the liquid phase in DES the liquidus temperature, and Δcp,i is the difference between
is the assumption that strong hydrogen bonding in the liquid the liquid and solid state heat capacities of pure component i at
phase is the reason for the depression of the mixture melting constant pressure. Several approaches can be found in the
temperature.10,11 literature considering the treatment of the heat capacity term.18
A knowledge of solid−liquid equilibria (SLE) in DES is However, due to the opposite signs of the last two terms in eq
essential to the design of eutectic mixtures.12,13 Studying SLE 1, their contribution is insignificant if Tm/T < 1.4. Further, the
in DES requires measuring the melting temperature of the difference between the liquid and solid states heat capacities is
mixture at different compositions. The mixture melting difficult to obtain experimentally and is only available for a
temperature can be measured by visual methods or by
differential scanning calorimetry (DSC). The main advantage Received: March 19, 2021
of DSC is the possibility to identify different phase transitions, Revised: October 11, 2021
such as the solid−solid transition and the glass transition, Published: October 25, 2021
which is not possible with visual methods.14−16 To generate
the phase diagram over the full composition range, the
obtained SLE data are modeled using the equation17
© 2021 The Authors. Published by
American Chemical Society https://doi.org/10.1021/acs.cgd.1c00306
6083 Cryst. Growth Des. 2021, 21, 6083−6091
Crystal Growth & Design pubs.acs.org/crystal Article

limited number of substances. As a result, the heat capacity 150 mL min−1. The standard uncertainties of the temperature and
term is usually neglected in modeling SLE. sensitivity measurements were 0.1 K and 0.3%, respectively.
In eq 1, the primary assumption is that the components The DSC chamber was precooled to 243 K before introducing the
crystallize as pure solids in their most stable polymorph: i.e., a sample. Then, a cooling run with a rate of 5 K min−1 down to 193 K
was performed. Finally, a heating run up to 323 K was conducted. The
simple eutectic mixture. Although this assumption might be solidus phase transition temperatures were determined as the onset
applicable to many DES, experimental investigations to prove temperatures at a heating rate of 5 K min−1. To determine the
its validity are lacking. It has been shown that polymorphism liquidus temperatures, the DSC measurements were performed at
can be observed in many eutectic systems.19−22 Corvis et al.20 three different heating rates: namely, 1, 2, and 5 K min−1. The
reported the formation of a 1:1 molar ratio cocrystal in the liquidus temperatures at zero heating rate were determined by
lidocaine/L-menthol binary eutectic system. Recently, Hall et extrapolation of peak maximum temperature to the zero heating rate.
al.23 used synchrotron X-ray powder diffraction (XRD) and Figure S1 in the Supporting Information shows the extrapolation
DSC to prove the formation of several metastable solid phases procedure for selected samples. DSC experiments were performed in
in different eutectic systems. Polymorphism and the formation duplicate for the same mixture. The difference in the determined
temperatures for different samples of the same mixture was less than 1
of congruently or incongruently melting compounds will affect K (the average absolute deviation is 0.5 K).
the measured melting temperature of the mixture and thus the 2.2. X-ray Diffraction. Powder XRD experiments were performed
phase diagram. using a Stadi P diffractometer (Stoe & Cie, Germany) with Debye−
The L-menthol/thymol eutectic system has attracted much Scherrer geometry equipped with a Mo fine-focus sealed tube, a
attention in the literature because of its very low eutectic curved Ge monochromator selecting Kα1 radiation (λ = 0.70930 Å),
temperature, low viscosity at room temperature, and poor and a Mythen2 R 1K detector (Dectris, Switzerland). The data were
miscibility with water.24−28 The SLE data for this system are collected with a step size of 0.015° 2θ per data point. For
available in the literature.19,24 In a previous work,19 it was measurements performed at 253 K, a cryostream (Oxford 800 series,
shown that the peaks observed in the DSC curves of the L- UK) was used. The samples were prepared by grinding using a mortar
and pestle in a cold room (at a temperature of 253 K) and then filled
menthol/thymol eutectic system, which were interpreted as into a glass capillary (Hilgenberg, Germany) with a 1.0 mm diameter.
eutectic peaks, were broad, indicating a possibility of several For single-crystal XRD (SC-XRD) measurements, a single crystal
thermal events. The present study investigates the SLE of the with the approximate dimensions 0.442 mm × 0.528 mm × 0.547 mm
L-menthol/thymol eutectic system in detail to comprehend was used for analysis. The X-ray intensity data were measured on a D8
these thermal events and characterize the formed solid phases. Venture Duo IMS system (Bruker, USA) equipped with a Helios
For this purpose, DSC measurements were coupled with low- optic monochromator and a Mo IMS microsource (λ = 0.71073 Å).
temperature XRD experiments. The nonrandom two-liquid The measurements were performed at 100 K. Details of the structure
(NRTL) and the two-suffix Margules equations were applied and refinement can be found in the Supporting Information.
to model the nonideality in the liquid and solid phases, 2.3. Thermodynamic Modeling. In this work, SLE data were
modeled by considering the miscibility in the solid phase. Thus, the
respectively, to generate the phase diagram in the whole activities of components in the solid phase should be considered. The
composition range and determine the eutectic points. following equation was used to calculate the liquidus temperature at
different mole fractions of components17
2. METHODS
xiLγi L Δhm, i
∫T
T 1 T
2.1. Solid−Liquid Equilibria. L-menthol (purity ≥99%, Sigma- ln =− 1− − Δc p(T ) dT
Aldrich) and thymol (purity ≥99%, Sigma-Aldrich) were weighed xiSγiS RT Tm, i RT m
(precision 1 × 10−4 g, Sartorius, Germany) in various ratios in glass
Δc p(T )
∫T
1 T
vials. Then, the vials were tightly closed, and the mixtures were gently
+ dT
heated with continuous stirring until a homogeneous clear liquid was R m T (2)
obtained.
Two sample preparation methods were used. The first sample where xSi
and γSi
are the composition and the activity coefficient of
preparation methodlater referred to as slow crystallization because component i in the solid phase, respectively. The activity coefficients
the samples crystallized after long storagethe samples were of components in the liquid phase γLi were calculated using the NRTL
introduced in DSC crucible pans as a liquid. The DSC pans were model as follows:17
hermetically sealed and stored at a temperature below their freezing
2
temperature observed previously in the literature.19 For samples with Gji τijGij
xthymol ≥ 0.70, crystallization was observed at 277 K after a few days or ln γi L = xj2 τji +
weeks, depending on the sample composition. However, due to xi + xjGji (xj + xiGij)2
(3)
kinetic limitations, no crystallization was observed in samples with
xthymol < 0.70 even at 253 K. The second preparation methodlater
referred to as rapid crystallization since the crystallization was Gij = exp(− ατij), Gji = exp(− ατji) (4)
observed shortly after annealing at 253 Kincluded quenching the
liquid samples inside plastic tubes at 193 K and then keeping them at gij − gjj gji − gii
this temperature for several hours. Then, the tubes were transferred to τij = , τji =
RT RT (5)
a freezer at 253 K. Crystallization was observed after a few hours or
days, depending on the sample composition. Later, the solid was The value of the nonrandomness parameter (α) was set to 0.3. The
ground inside a cold room at 253 K using a mortar and pestle. The activity coefficients of components in the solid phase γSi were
samples were introduced into DSC aluminum crucible pans as a fine calculated using the two-suffix Margules equation as follows:17
powder, and the DSC crucibles were hermetically sealed.
SLE data were measured using DSC (NETZSCH DSC 200 F3, Aij
Germany). The instrument was calibrated before measurements using ln γiS = (xjS)2
RT (6)
the standard procedure based on the onset temperature of the
transition of six calibration standards at a heating rate of 5 K min−1: The binary interaction parameters Aij, gij − gjj, and gji − gii were fitted
adamantane, bismuth, cesium chloride, indium, tin, and zinc. to experimental liquidus data (Texp
i ) by minimizing the objective
Measurements were performed under nitrogen with a flow rate of function

6084 https://doi.org/10.1021/acs.cgd.1c00306
Cryst. Growth Des. 2021, 21, 6083−6091
Crystal Growth & Design pubs.acs.org/crystal Article

n 1/2
(Tiexp − Tical)2
F(T ) = ∑
i=1
n (7)
where Tcal
i is the calculated liquidus temperature and n is the number
of data points. The liquidus and solidus curves can be calculated
similarly to dew and bubble curve calculations.29
In a binary system with a cocrystal, the formation of a cocrystal
between component A and component B can be described using the
chemical reaction30,31
ϑAA (L) + ϑBB(L) ↔ A ϑA B ϑB
(S) (8)
where ϑA and ϑB are the stoichiometric coefficients of components A
and B in the cocrystal, respectively. The equilibrium constant (Ka) of
the chemical reaction in eq 2 is defined via activities (ai) as
Figure 1. Solid−liquid equilibria of the L-menthol/thymol system
(xALγAL)ϑA (x BLγBL)ϑB
Ka = ∏ aiϑi = S S
measured in this work (filled symbols) for samples prepared using two
i (xABγAB) (9) different methods compared to literature data (cross symbols), with
the corresponding differential scanning calorimetry curves for slow
In the case where the melting temperature of the cocrystal is close to crystallization (yellow curves) and rapid crystallization (blue curves)
the eutectic temperature, the effect of the temperature on the melting samples. Literature data were taken from Abranches et al.24 and
enthalpy of the cocrystal can be neglected, and the equilibrium Alhadid et al.19 Legend: slow crystallization liquidus (yellow circles);
constant of the chemical reaction at different temperatures can be rapid crystallization liquidus (blue circles); slow crystallization solidus
calculated using the Gibbs−Helmholtz equation (yellow diamonds); rapid crystallization solidus (blue diamond);
Alhadid et al. (black star); Abranches et al. (black cross).
Δhref 1 1
ln K a = ln K aref + −
R T ref T (10)
slow crystallization. It is worth mentioning that a glass
ref
where T is a reference temperature, Kref
is the equilibrium constant
a transition at 217 K was observed for samples with xthymol =
at Tref, and Δhref is the melting enthalpy of the cocrystal at Tref. In this 0.75 and 0.70 prepared by slow crystallization (the DSC curves
work, Tref was considered to be the maximum temperature at which
are shown in Figure S2 in the Supporting Information). This
the cocrystal is stable. Tref and Δhref were measured using DSC. The
equilibrium constant at Tref was calculated as can be attributed to incomplete crystallization in the samples
with a lower thymol content prepared by the slow
(xAref γAref )ϑA (x Bref γBref )ϑB crystallization method due to their high viscosity.26
K aref = ref ref To better understand the DSC curves and the phase
(xAB γAB ) (11)
diagram, powder XRD measurements are needed. For samples
ref ref
where x and γ are the mole fraction and the activity coefficient of with xthymol = 0.70 prepared by slow crystallization, it was
the component at which Tref and Δhref were measured, respectively. impossible to obtain a completely crystallized sample to
perform powder XRD. Therefore, single-crystal XRD (SC-
3. RESULTS AND DISCUSSION XRD) analysis was performed on crystals isolated from the
3.1. Experimental Solid−Liquid Equilibria. For the sake liquid phase during storage at 277 K. The obtained crystal
of simplicity, the discussion of the measured phase diagram is structure, shown in Figure S5 and Table S5 in the Supporting
divided below into two parts: first, the thymol-rich region Information, corresponds to the reported crystal structure for
(xthymol from 0.70 to 1.0) is considered, and then, the L- thymol at room temperature, thus confirming that the solid
menthol-rich region (xthymol ≤ 0.67). obtained by slow crystallization consists of pure thymol
Figure 1 shows the measured SLE data compared to those crystals.32 Therefore, the solidus and liquidus temperatures
found in the literature along with the DSC curves for samples of slow crystallization samples correspond to thymol liquidus
in the composition range xthymol from 0.70 to 1.0. In this and solidus lines.
composition range, samples were prepared using two different Powder XRD was performed on several mixtures of different
methods: (i) slow crystallization or (ii) rapid crystallization. As compositions in the range xthymol > 0.67. The samples were
shown in Figure 1, the liquidus temperatures of samples prepared by rapid crystallization. Figure 2 presents the SLE
prepared by slow crystallization (yellow circles) or rapid data and the powder XRD pattern at 253 K of the samples
crystallization (blue circles) are similar and are in good marked with circles in Figure 2A. As shown in Figure 2B, the
agreement with data found in the literature (black star and powder XRD pattern of pure thymol (point a) matches that of
cross symbols). For samples prepared by slow or rapid the sample with xthymol = 0.83 (point b). The absence of the
crystallization with xthymol < 0.80, the solidus temperatures do pure L-menthol solid phase indicates that it is dissolved in the
not change with composition. In contrast, in samples with solid phase of thymol, forming a solid solution (α solid phase).
xthymol > 0.80 prepared by rapid crystallization, the solidus Therefore, the solidus temperatures in Figure 2A for samples
temperatures depend strongly on the composition. As can be with xthymol > 0.83 correspond to the α solid phase. When the
seen in Figure 1, the DSC curves of the sample with xthymol = thymol mole fraction is decreased further, a different solid
0.88 prepared by slow crystallization (yellow curve) and rapid phase appears at xthymol = 0.80 (point c). This solid phase does
crystallization (blue curve) are similar. However, the DSC not correspond to pure L-menthol, as seen in Figure 2B
curves of the sample with xthymol < 0.80 prepared by rapid (bottom black powder XRD pattern). On comparison of the
crystallization and slow crystallization are different. The solidus powder XRD patterns at xthymol = 0.75 (point d) and 0.80
temperature of the samples with xthymol < 0.80 prepared by (point c), it becomes clear that the solid phase at xthymol = 0.80
rapid crystallization is lower than that of samples prepared by is a mixture of the α solid phase and the solid phase of the
6085 https://doi.org/10.1021/acs.cgd.1c00306
Cryst. Growth Des. 2021, 21, 6083−6091
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 2. (A) Solid−liquid equilibria of the L-menthol/thymol system and (B) powder X-ray diffraction pattern obtained at 253 K for the solid
formed by rapid crystallization. (C) Powder X-ray diffraction patterns of the sample with xthymol = 0.75 obtained at 253 and 298 K in comparison to
that of pure thymol. The intensities are shifted for a better comparison.

sample xthymol = 0.75. As can be seen in Figure 2B, the same

powder XRD patterns are observed for samples with xthymol =
0.75, 0.70, and 0.67. The absence of pure L-menthol indicates
the formation of a second solid solution (β solid phase). The
solidus temperatures in the range 0.67 < xthymol < 0.80
correspond to the β solid phase.
Although the solid phases of samples with xthymol = 0.75,
0.70, and 0.67 are similar, a liquidus peak is only observed at
xthymol = 0.75, as seen in Figure 1. To further investigate which
solid phase the liquidus peak corresponds to, powder XRD was
performed on samples with xthymol = 0.75 at 253 K (below the
solidus temperature) and 298 K (above the solidus temper- Figure 3. Solid−liquid equilibria data of L-menthol/thymol for
ature). As seen in Figure 2C, the powder XRD pattern of the samples prepared by rapid crystallization after quenching the liquid
sample with xthymol = 0.75 at 298 K corresponds to pure mixture at 193 K (filled symbols) compared to literature data (cross
thymol: i.e., α solid phase. This indicates that the solid formed symbols). Literature data were taken from Abranches et al.24 and
in the composition range of 0.67 < xthymol < 0.80 is unstable, Alhadid et al.19
and above the solidus temperature, it decomposes to an α solid
phase and a liquid phase. Indexing of powder data at xthymol = measured in this work compared to published literature data.
0.75, 0.70, and 0.67 showed that the number of molecules per As can be noticed, a significant negative deviation from the
unit cell is approximately 16 (results for xthymol = 0.67 are ideal behavior (dashed line) is observed in the part of the
shown in Table S4 in the Supporting Information). The phase diagram rich in L-menthol. The liquidus temperatures of
number matches a cocrystal with a stoichiometric ratio of 1:3 the L-menthol liquidus line measured in this work are in good
for L-menthol:thymol (1:3 cocrystal). As seen in Figure 2A, the agreement with the data found in the literature.
cocrystal melts incongruently at around 285 K. Moreover, the As seen in Figure 3, in the region 0.33 < xthymol ≤ 0.50, only
β solid phase is a solid solution of L-menthol and the 1:3 solidus temperatures are observed. The DSC curves in this
cocrystal. region are discussed in the following to understand the phase
In the following, the second part of the phase diagram in the diagram better. Figure 4A shows the DSC curves of samples
L-menthol-rich region, i.e., xthymol ≤ 0.67, is discussed. As with different thymol mole fractions. The DSC curve of the
already mentioned, it was impossible to crystallize the L- sample with xthymol = 0.67 shows the liquidus peak of the 1:3
menthol-rich samples by slow crystallization due to the high cocrystal and the solidus peak of the β solid phase as discussed
viscosity of the mixture close to its melting temperature.26 above. As the mole fraction of thymol decreases, the solidus
Therefore, only the second preparation method, i.e., rapid and liquidus peaks shift slightly to lower temperatures. At
crystallization, was considered. Figure 3 shows SLE data xthymol = 0.50, only a solidus peak is observed. Although at
6086 https://doi.org/10.1021/acs.cgd.1c00306
Cryst. Growth Des. 2021, 21, 6083−6091
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 4. (A) DSC curves of samples with different thymol mole fractions. Powder X-ray diffraction was performed on samples shown by solid
curves. (B) Powder X-ray patterns of samples with different mole fractions and of pure L-menthol.

Table 1. Melting Properties Used to Model the Solid−Liquid Equilibria of the L-Menthol:Thymol System
Δcp = a + bT (J mol−1 K−1)
a b
substance Δhm (kJ mol−1) Tm (K) liquid solid liquid solid
L-menthol 13.7436 314.636 −195.2037 −68.037 1.79537 1.09237
thymol 20.6419 322.719 147.5038 12.73938 0.57138 0.76738
3:2 cocrystal 44.90 ± 1.50a,b 274.7 ± 0.10a

Δhref (kJ mol−1) Tref (K)

a
1:3 cocrystal 45.0 ± 0.20 285.7 ± 0.10a
a b
Measured in this work using DSC. Uncertainties are the standard deviation of three samples.

xthymol = 0.40 only a solidus peak is observed, it is different additional peaks (marked with pentagons in Figure 4B) are
from that observed at xthymol = 0.50. At a lower thymol mole observed in the sample with xthymol = 0.33. As shown in Figure
fraction, a solidus peak with a similar onset temperature is 4B, these additional peaks are related to pure L-menthol (blue
observed with no obvious liquidus peaks. At xthymol = 0.30, a curve). This implies that the sample with xthymol = 0.33 is a
small liquidus peak appears. mixture of the solid phase observed at xthymol = 0.40 and pure L-
To help understand the different solidus peaks observed in menthol.
different samples in Figure 4A, powder XRD was performed on In conclusion, the formation of a second solid phase was
samples showing only one solidus peak. Figure 4B shows the observed in samples rich in L-menthol prepared by rapid
powder XRD patterns of mixtures with xthymol = 0.67, 0.50, crystallization after quenching the liquid mixture at 193 K. The
0.40, and 0.33 and pure L-menthol. The powder XRD patterns XRD pattern of the sample with xthymol = 0.40 depicted the
of samples with xthymol = 0.67 (black curve) and xthymol = 0.50 absence of a β solid phase and of pure L-menthol. According to
(red curve) have several similar peaks. However, additional the DSC curves shown in Figure 4A, it is highly probable that
peaks (marked with crosses in Figure 4B) are observed for the the second solid is a cocrystal with a stoichiometric ratio of 3:2
sample with xthymol = 0.50, indicating the presence of a different for L-menthol:thymol (3:2 cocrystal). An analysis of the
solid phase. By a comparison of the powder XRD pattern of powder XRD patterns and the DSC curves in Figure 4 shows
the sample with xthymol = 0.50 (red curve) with those of xthymol that the solidus peaks at xthymol = 0.50 and 0.33 with onset
= 0.67 (black curve) and xthymol = 0.40 (yellow curve), it can be temperatures of 271.6 and 272.7 K, respectively, represent the
seen that the solid phase of the sample with xthymol = 0.50 is a two eutectic points of the system.33−35
mixture of the β solid phase and the solid phase of the sample The melting properties of the 3:2 cocrystal and the reference
with xthymol = 0.40. At xthymol = 0.40, the β solid phase and pure temperature and enthalpy for the 1:3 cocrystal were measured
L-menthol are not observed. by DSC. Liquid mixtures with molar ratios of 3:7 and 3:2 L-
By a comparison of the powder XRD pattern of the samples menthol:thymol were quenched at 193 K for 1 day and then
with xthymol = 0.40 (yellow curve) and 0.33 (green curve), annealed at 253 K for 1 week. Then, the samples were ground
6087 https://doi.org/10.1021/acs.cgd.1c00306
Cryst. Growth Des. 2021, 21, 6083−6091
Crystal Growth & Design pubs.acs.org/crystal Article

using a mortar and pestle in a cold room at 253 K. The samples the binary interaction parameters in each case can be attributed
were then filled in DSC crucibles in triplicate. The melting to the small value of Tm/T for the system. The obtained
temperatures and enthalpies of the 3:2 cocrystal were RMSD was smaller when a constant heat capacity term was
determined as the onset temperatures and the areas of the considered. In contrast, the highest RMSD was obtained
corresponding peaks, respectively. The reference temperature assuming a temperature-dependent Δcp. Therefore, the
and enthalpy of the 1:3 cocrystal were determined as the peak constant heat capacity term was considered for further
maximum temperature and the area of the corresponding peak, calculations.
respectively. The melting properties of the 3:2 cocrystal, pure The liquidus lines of thymol (xthymol > 0.67) and L-menthol
thymol, and L-menthol, as well as the reference temperature (xthymol < 0.33) were calculated using eq 2 and their melting
and enthalpy of the 1:3 cocrystal, are presented in Table 1. properties from Table 1. Because no solid solution region is
3.2. Thermodynamic Modeling. The measured SLE data formed in the L-menthol-rich region, the liquidus line can be
were correlated to obtain the phase diagram of the mixture calculated as a simple eutectic system: i.e., xSi γSi = 1. In
over the whole composition range. eq 2 was used to calculate contrast, calculating the thymol liquidus line should consider
the liquidus temperatures at different thymol mole fractions. the composition and activity coefficients of components in the
The activity coefficients of components in the liquid solution solid phase.
were calculated using the NRTL model. The solid solution was The liquidus and solidus lines in the middle range of
considered either ideal or nonideal, and in the latter case, the composition (0.33 < xthymol < 0.67), corresponding to the
two-suffix Margules equation was used to calculate the activity crystallization of the two cocrystals, were calculated using eqs
coefficients of components in the solid phase. Thymol liquidus 8−11 and the measured melting properties from Table 1, and
temperatures (measured in the range xthymol = 0.75−1.0) and L- the activity coefficients in the liquid phase were calculated
menthol liquidus temperatures (measured in the range xthymol = using eqs 3−5. The xref thymol values for the 1:3 cocrystal and 3:2
0.0−0.30) were used to obtain the binary interaction cocrystal are 0.70 and 0.40, respectively. The Krefa values for the
parameters of the NRTL and two-suffix Margules equations. two cocrystals calculated using eq 11 are reported in Table S3
Three approaches were applied on consideration of the heat in the Supporting Information. In the case of 3:2 cocrystal
capacity term, namely, Δcp = 0, constant Δcp calculated at the liquidus lines, xAB S S
γAB = 1 because the 3:2 cocrystal is
melting temperature of pure components, and linear temper- immiscible with L-menthol and the 1:3 cocrystal in the solid
ature dependence of Δcp using the coefficients in Table 1. The state. In contrast, the composition and the activity coefficients
binary interaction parameters and RMSDs obtained using the of the 1:3 cocrystal in the solid phase should be considered in
three different approaches are shown in Table 2. Considering calculating the liquidus and solidus lines. The β solid phase was
assumed to be ideal to avoid fitting additional experimental
Table 2. Binary Interaction Parameters and Root-Mean- data.
Square Deviation (RMSD) Obtained Using Different The results of the SLE modeling are discussed in the
Approaches Considering the Heat Capacity Term following to depict the SLE behavior of the system and attain
g12 − g22 g21 − g11 Aαij RMSD
the position of the eutectic points. Figure 5 shows the
(kJ mol−1) (kJ mol−1) (kJ mol−1) (K) complete phase diagram of the L-menthol/thymol eutectic
Δcp = 0 −4.7068 −2.4656 −3.9724 0.6 system modeled considering an ideal (Figure 5A) and nonideal
constant Δcp −4.6067 −1.5759 −2.3691 0.4 (Figure 5B) α solid solution. The points correspond to
temperature- −4.6548 −1.6443 −2.1716 2.7 experimental data from the present study obtained by rapid
dependent Δcp crystallization, and the solid blue lines represent the results of
the SLE modeling. As seen, the calculated liquidus lines of pure
or neglecting the heat capacity term leads to slightly different thymol and L -menthol are in good agreement with
binary interaction parameters. The small difference between experimental data. Moreover, the SLE data in the middle

Figure 5. Solid−liquid equilibria data of the L-menthol/thymol mixture modeled using the NRTL equation considering cocrystal formation and
(A) ideal solid solution and (B) nonideal solution. Legend: (yellow ●) liquidus temperatures; (red ◆) solidus temperatures; (yellow ■) 3:2
cocrystal melting temperature; (blue line) modeled lines; (blue dashed line) predicted eutectic temperature; (dotted blue line) extension of
modeled thymol liquidus and solidus lines; (gray line) expected liquidus and solidus lines.

6088 https://doi.org/10.1021/acs.cgd.1c00306
Cryst. Growth Des. 2021, 21, 6083−6091
Crystal Growth & Design pubs.acs.org/crystal Article

composition range corresponding to the liquidus lines of the regions and indicates the reliability of the NRTL equation to
two cocrystals as well as the two eutectic points of the system capture the nonideality in the liquid phase. The two eutectic
are well predicted. This further supports the stoichiometric points of the system obtained by SLE modeling are Tcal e1 =
ratio of the cocrystals and the presence of a solid solution 271.7 K, xcal cal cal
thymol,e1 = 0.48 and Te2 = 273.1 K, xthymol,e2 = 0.33.
region between L-menthol and the 1:3 cocrystal (β solid In general, this study shows that immiscibility in the solid
phase).31,39 Additionally, this indicates the reliability of the phase cannot always be assumed in measuring and modeling
NRTL equation to capture the nonideality in the liquid phase. SLE in DES. The actual behavior of the system can be
The two eutectic temperatures estimated by SLE modeling are obtained from detailed studies of samples prepared by different
Tcal cal cal cal
e1 = 271.7 K, xthymol,e1 = 0.48 and Te2 = 273.1 K, xthymol,e2 = methods. DSC measurements should be coupled with XRD
0.33. when possible to acquire information regarding the solid
The dotted blue lines are extrapolations of the thymol phases formed. Thermodynamic modeling is a useful tool to
liquidus and solidus lines wherein the formation of the 1:3 generate and understand the solid−liquid phase diagram in the
cocrystal is not considered. These liquidus and solidus lines do whole range of compositions and can provide the position of
not describe the system behavior in the middle range of eutectic points that can often be difficult to obtain
composition, and their use would lead to an incorrect experimentally.
estimation of the eutectic point. The gray lines represent a To determine the exact type of crystalline phases and their
possible course of the liquidus and solidus lines if cocrystal stoichiometry, more specific techniques such as 3D electron
formation is considered. Obviously, because of the extreme diffraction and SC-XRD are recommended. Nevertheless, the
change in the liquidus temperature within a very narrow range sample preparation methods required to form the cocrystals of
of system compositions corresponding to the gray lines, it is the L-menthol/thymol system, i.e., quenching and annealing,
hardly possible to verify the course of the gray lines by are not suitable for obtaining samples for SC-XRD.
measurements.
Due to the small solid solubility limit and similarity between
thymol and L-menthol, it is reasonable to assume the solid
■
*
ASSOCIATED CONTENT
sı Supporting Information
solution (α solid phase) as an ideal solution, i.e., γSthymol = 1. The Supporting Information is available free of charge at
Nevertheless, the phase diagram was modeled by assuming https://pubs.acs.org/doi/10.1021/acs.cgd.1c00306.
ideal or nonideal solid phases. The obtained binary interaction
Solid−liquid equilibria data, models of binary interaction
parameters and RMSDs are shown in Table S3 in the
parameters, calculated eutectic point properties, and
Supporting Information. Although a significant difference was
RMSD between experimental and calculated liquidus
observed in the binary interaction parameters when the
temperatures, DSC curves of samples with xthymol = 0.7
nonideality in the solid phase was considered, the RMSD
and 0.75 prepared by different methods, DSC curves of
was similar in each case. As seen in Figure 5, the predictions for
samples with a 1:1 molar ratio mixture prepared by
the solidus lines of α and β solid phases could not be improved
different methods, DSC curves of samples with different
by considering the nonideality in the solid phase.
thymol mole fractions, indexing of powder data of a
sample with xthymol = 0.67, and single-crystal X-ray
4. CONCLUSION
diffraction data (PDF)
In this work, detailed SLE data of the L-menthol/thymol
eutectic system were obtained using DSC. Different sample Accession Codes
preparation methods were applied to comprehend the SLE CCDC 2069355 contains the supplementary crystallographic
behavior of the studied system. Due to kinetic limitations, the data for this paper. These data can be obtained free of charge
complete phase diagram of the system could only be obtained via www.ccdc.cam.ac.uk/data_request/cif, or by emailing
if samples were prepared by quenching the liquid mixture at [email protected], or by contacting The Cam-
193 K, followed by annealing at 253 K. DSC measurements bridge Crystallographic Data Centre, 12 Union Road,
were coupled with XRD measurements to characterize the Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
formed solid phases. The SLE diagram was found to have a
character more complex than that previously reported.
The formations of an incongruently melting cocrystal of 1:3
■ AUTHOR INFORMATION
Corresponding Author
menthol:thymol and a congruently melting cocrystal of 3:2 L- Ahmad Alhadid − Biothermodynamics, TUM School of Life
menthol:thymol were observed. The melting properties of the Sciences, Technical University of Munich (TUM), 85354
two cocrystsals were measured in this work using DSC. Freising, Germany; orcid.org/0000-0003-1443-1517;
Additionally, the presence of two solid solution regions was Email: [email protected]
confirmed by DSC analysis and powder XRD measurements.
The obtained SLE data were modeled by considering the Authors
formation of the two cocrystals and the solid solution regions. Christian Jandl − Catalysis Research Center, Department
The activity coefficients of the components in the solid and Chemie, Technical University of Munich (TUM), 85748
liquid phases were calculated by the two-suffix Margules and Garching bei München, Germany
NRTL models, respectively. The data obtained on the liquidus Liudmila Mokrushina − Separation Science & Technology,
and solidus lines of the pure components and the two Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU),
cocrystals are in good agreement with the measured data. 91058 Erlangen, Germany
Further, the two eutectic points of the system are in good Mirjana Minceva − Biothermodynamics, TUM School of Life
agreement with the solidus temperatures measured in the Sciences, Technical University of Munich (TUM), 85354
middle composition range. This supports the stoichiometric Freising, Germany
ratio of the cocrystals and the formation of the solid solution Complete contact information is available at:
6089 https://doi.org/10.1021/acs.cgd.1c00306
Cryst. Growth Des. 2021, 21, 6083−6091
Crystal Growth & Design pubs.acs.org/crystal Article

https://pubs.acs.org/10.1021/acs.cgd.1c00306 (17) Prausnitz, J. M.; Lichtenthaler, R. N.; Azevedo, E. G. d.

Molecular Thermodynamics of Fluid-Phase Equilibria; Prentice-Hall:
Author Contributions 1999.
(18) Do, H. T.; Chua, Y. Z.; Habicht, J.; Klinksiek, M.; Hallermann,
A.A.: conceptualization, methodology, formal analysis, and M.; Zaitsau, D.; Schick, C.; Held, C. Melting properties of peptides
writingoriginal draft. C.J.: single-crystal and powder XRD and their solubility in water. Part 1: dipeptides based on glycine or
experiments and writingreview and editing. L.M.: writing alanine. RSC Adv. 2019, 9 (56), 32722−32734.
review and editing. M.M.: writingreview, editing, and (19) Alhadid, A.; Mokrushina, L.; Minceva, M. Formation of glassy
supervision. phases and polymorphism in deep eutectic solvents. J. Mol. Liq. 2020,
314, 113667.
Notes
(20) Corvis, Y.; Négrier, P.; Lazerges, M.; Massip, S.; Léger, J.-M.;
The authors declare no competing financial interest. Espeau, P. Lidocaine/l-menthol binary system: Cocrystallization
versus solid-state immiscibility. J. Phys. Chem. B 2010, 114 (16),
■ REFERENCES
(1) van Osch, D. J. G. P.; Dietz, C. H. J. T.; van Spronsen, J.; Kroon,
5420−5426.
(21) Hamilton, V.; Andrusenko, I.; Potticary, J.; Hall, C.; Stenner,
R.; Mugnaioli, E.; Lanza, A. E.; Gemmi, M.; Hall, S. R. Racemic
M. C.; Gallucci, F.; van Sint Annaland, M.; Tuinier, R. A search for conglomerate formation via crystallization of metaxalone from volatile
natural hydrophobic deep eutectic solvents based on natural deep eutectic solvents. Cryst. Growth Des. 2020, 20 (7), 4731−4739.
components. ACS Sustainable Chem. Eng. 2019, 7 (3), 2933−2942. (22) Potticary, J.; Hall, C.; Hamilton, V.; McCabe, J. F.; Hall, S. R.
(2) Perna, F. M.; Vitale, P.; Capriati, V. Deep eutectic solvents and Crystallization from volatile deep eutectic solvents. Cryst. Growth Des.
their applications as green solvents. Curr. Opin. Green Sustain. Chem. 2020, 20 (5), 2877−2884.
2020, 21, 27−33. (23) Hall, C. L.; Potticary, J.; Hamilton, V.; Gaisford, S.; Buanz, A.;
(3) Emami, S.; Shayanfar, A. Deep eutectic solvents for Hall, S. R. Metastable crystalline phase formation in deep eutectic
pharmaceutical formulation and drug delivery applications. Pharm. systems revealed by simultaneous synchrotron XRD and DSC. Chem.
Dev. Technol. 2020, 25 (7), 779−796. Commun. 2020, 56, 10726.
(4) Á lvarez, M. S.; Zhang, Y. Sketching neoteric solvents for (24) Abranches, D. O.; Martins, M. A. R.; Silva, L. P.; Schaeffer, N.;
boosting drugs bioavailability. J. Controlled Release 2019, 311−312, Pinho, S. P.; Coutinho, J. A. P. Phenolic hydrogen bond donors in the
225−232. formation of non-ionic deep eutectic solvents: the quest for type V
(5) Olives, A. I.; González-Ruiz, V.; Martín, M. A. Sustainable and DES. Chem. Commun. 2019, 55 (69), 10253−10256.
eco-friendly alternatives for liquid Chromatographic Analysis. ACS (25) Martins, M. A. R.; Silva, L. P.; Schaeffer, N.; Abranches, D. O.;
Sustainable Chem. Eng. 2017, 5 (7), 5618−5634. Maximo, G. J.; Pinho, S. P.; Coutinho, J. A. P. Greener terpene-
(6) Cai, T.; Qiu, H. Application of deep eutectic solvents in terpene eutectic mixtures as hydrophobic solvents. ACS Sustainable
chromatography: A review. TrAC, Trends Anal. Chem. 2019, 120, Chem. Eng. 2019, 7 (20), 17414−17423.
115623. (26) Alhadid, A.; Mokrushina, L.; Minceva, M. Influence of the
(7) Hooshmand, S. E.; Afshari, R.; Ramón, D. J.; Varma, R. S. Deep molecular structure of constituents and liquid phase non-ideality on
eutectic solvents: cutting-edge applications in cross-coupling reac- the viscosity of deep eutectic solvents. Molecules 2021, 26 (14), 4208.
tions. Green Chem. 2020, 22 (12), 3668−3692. (27) Rodríguez-Llorente, D.; Cañada-Barcala, A.; Á lvarez-Torrellas,
(8) González de Castilla, A.; Bittner, J. P.; Mü l ler, S.; S.; Á gueda, V. I.; García, J.; Larriba, M. A review of the use of eutectic
Jakobtorweihen, S.; Smirnova, I. Thermodynamic and transport solvents, terpenes and terpenoids in liquid-liquid extraction processes.
properties modeling of deep eutectic solvents: A review on gE-models, Processes 2020, 8 (10), 1220.
equations of state, and molecular dynamics. J. Chem. Eng. Data 2020, (28) Lemaoui, T.; Darwish, A. S.; Attoui, A.; Abu Hatab, F.;
65 (3), 943−967. Hammoudi, N. E. H.; Benguerba, Y.; Vega, L. F.; Alnashef, I. M.
(9) Alkhatib, I. I. I.; Bahamon, D.; Llovell, F.; Abu-Zahra, M. R. M.; Predicting the density and viscosity of hydrophobic eutectic solvents:
Vega, L. F. Perspectives and guidelines on thermodynamic modelling towards the development of sustainable solvents. Green Chem. 2020,
of deep eutectic solvents. J. Mol. Liq. 2020, 298, 112183. 22, 8511.
(10) Smith, E. L.; Abbott, A. P.; Ryder, K. S. Deep eutectic solvents (29) Domanska, U.; Groves, F. R.; McLaughlin, E. Solid-liquid phase
(DESs) and their applications. Chem. Rev. 2014, 114 (21), 11060− equilibria of binary and ternary mixtures of benzene and polynuclear
11082. aromatic compounds. J. Chem. Eng. Data 1993, 38 (1), 88−94.
(11) Martins, M. A. R.; Pinho, S. P.; Coutinho, J. A. P. Insights into (30) Lange, L.; Sadowski, G. Polymorphs, hydrates, cocrystals, and
the nature of eutectic and deep eutectic mixtures. J. Solution Chem. cocrystal hydrates: Thermodynamic modeling of theophylline
2019, 48 (7), 962−982. systems. Cryst. Growth Des. 2016, 16 (8), 4439−4449.
(12) Alhadid, A.; Mokrushina, L.; Minceva, M. Modeling of solid- (31) Lange, L.; Sadowski, G. Thermodynamic modeling for efficient
liquid equilibria in deep eutectic solvents: A parameter study. cocrystal formation. Cryst. Growth Des. 2015, 15 (9), 4406−4416.
Molecules 2019, 24 (12), 2334. (32) Thozet, A.; Perrin, M. Structure of 2-isopropyl-5-methylphenol
(13) Wolbert, F.; Brandenbusch, C.; Sadowski, G. Selecting (thymol). Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem.
excipients forming therapeutic deep eutectic systems-A mechanistic 1980, 36 (6), 1444−1447.
approach. Mol. Pharmaceutics 2019, 16 (7), 3091−3099. (33) Chelouche, S.; Trache, D.; Pinho, S. P.; Khimeche, K.
(14) Abdallah, M. M.; Müller, S.; González de Castilla, A.; Gurikov, Experimental and modeling studies of binary organic eutectic systems
P.; Matias, A. A.; Bronze, M. D.; Fernández, N. Physicochemical to be used as stabilizers for nitrate esters-based energetic materials.
characterization and simulation of the solid-liquid equilibrium phase Fluid Phase Equilib. 2019, 498, 104−115.
diagram of terpene-based eutectic solvent systems. Molecules 2021, 26 (34) Chelouche, S.; Trache, D.; Neves, C. M. S. S.; Pinho, S. P.;
(6), 1801. Khimeche, K.; Benziane, M. Solid + liquid equilibria and molecular
(15) Trache, D.; Khimeche, K.; Benelmir, R.; Dahmani, A. DSC structure studies of binary mixtures for nitrate ester’s stabilizers:
measurement and prediction of phase diagrams for binary mixtures of Measurement and modeling. Thermochim. Acta 2018, 666, 197−207.
energetic materials’ stabilizers. Thermochim. Acta 2013, 565, 8−16. (35) Chelouche, S.; Trache, D.; Pinho, S. P.; Khimeche, K.;
(16) Benziane, M.; Khimeche, K.; Dahmani, A.; Nezar, S.; Trache, Mezroua, A.; Benziane, M. Solid-liquid phase equilibria, molecular
D. Experimental determination and prediction of (solid + liquid) interaction and microstructural studies on (N-(2-ethanol)-p-nitroani-
phase equilibria for binary mixtures of heavy alkanes and fatty acids line + N-(2-acetoxyethyl)-p-nitroaniline) binary mixtures. Int. J.
methyl esters. J. Therm. Anal. Calorim. 2013, 112 (1), 229−235. Thermophys. 2018, 39 (11), 129.

6090 https://doi.org/10.1021/acs.cgd.1c00306
Cryst. Growth Des. 2021, 21, 6083−6091
Crystal Growth & Design pubs.acs.org/crystal Article

(36) Alhadid, A.; Mokrushina, L.; Minceva, M. Design of deep

eutectic systems: A simple approach for preselecting eutectic mixture
constituents. Molecules 2020, 25 (5), 1077.
(37) Corvis, Y.; Espeau, P. Interpretation of the global heat of
melting in eutectic binary systems. Thermochim. Acta 2018, 664, 91−
99.
(38) Zhu, P.; Chen, Y.; Fang, J.; Wang, Z.; Xie, C.; Hou, B.; Chen,
W.; Xu, F. Solubility and solution thermodynamics of thymol in six
pure organic solvents. J. Chem. Thermodyn. 2016, 92, 198−206.
(39) Tumakaka, F.; Prikhodko, I. V.; Sadowski, G. Modeling of
solid-liquid equilibria for systems with solid-complex phase formation.
Fluid Phase Equilib. 2007, 260 (1), 98−104.

6091 https://doi.org/10.1021/acs.cgd.1c00306
Cryst. Growth Des. 2021, 21, 6083−6091
 Supporting Information

Experimental Investigation and Modeling of Cocrystal Formation in

L-Menthol/Thymol Eutectic System
Ahmad Alhadid* 1, Christian Jandl 2, Liudmila Mokrushina 3, and Mirjana Minceva 1
1
Biothermodynamics, TUM School of Life Sciences, Technical University of Munich (TUM), Germany
2
Catalysis Research Center, Department Chemie, Technical University of Munich (TUM), Germany
3
Separation Science & Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU),
Germany
*
Corresponding author e-mail: [email protected]

Table S1. Solid liquid equilibria data of L-menthol/thymol mixture for samples prepared by rapid crystallization.
xthymol Te / K T/K
0.94 319.6
0.91 306.1 317.7
0.88 299.0 315.8
0.85 292.1 312.9
0.83 291.2 312.2
0.80 282.9 310.2
0.75 284.4 306.1
0.70 285.7
0.67 272.7 285.2
0.65 271.7 284.7
0.60 272.3 282.9
0.59 272.3 282.9
0.55 271.7 280.1
0.51 272.1
0.50 271.6
0.46 271.2
0.44 272.7
0.40 274.8
0.35 274.6

1
 xthymol Te / K T/K
0.33 273.7
0.32 273.3
0.30 273.5 280.0
0.25 273.1 288.9
0.20 271.7 296.2
0.15 271.5 303.1
0.10 271.1 307.0

Table S2. Solid liquid equilibria data of L-menthol/thymol mixture for samples prepared by slow crystallization.
xthymol Te / K T/K
0.88 297.2 317.4
0.80 295.7 309.9
0.75 293.2 307.5
0.70 295.4

2
Table S3. Binary interaction parameters, calculated eutectic point properties, and root-mean-square deviation
(RMSD) between experimental and calculated liquidus temperatures obtained in this work compared to data reported
in the literature.

Non-ideal solid Ideal solid Alhadid et al. 1

1 4.6067 3.8712 4.3333

𝑔12 𝑔22 / kJ mol

1 1.5759 1.7768 4.0648

𝑔21 𝑔11 / kJ mol

1
𝐴 / kJ mol 2.3691 0 –

3 2
𝐾 (1:3 cocrystal) 8.60 10 1.27 10 –

3
𝐾 (3:2 cocrystal) 7.06 10 1.10 10 –

0.50 0.48
xe 0.43
0.34 0.33
270.4 271.7
Te / K 241.5
272.9 273.0

RMSD / K 0.4 0.5 3.5

xthy=0.90 xthy = 0.80 xthy = 0.65

xthy = 0.30 xthy = 0.20
330
y = 0.3154x + 317.74
320
y = 0.1808x + 310.23
310
T/K

300 y = 0.1192x + 296.17

290 y = 0.3885x + 284.68

280 y = 0.8346x + 279.96

270
0 1 2 3 4 5 6
heating rate / K min 1

Figure S1. Determination of the liquidus temperature by extrapolating the peak maximum to zero heating rate.

3
 (A) (B)
xthymol = 0.75 xthymol = 0.70

294.9 K

293.7 K

284.2 K
284.4 K

200 220 240 260 280 300 320 200 220 240 260 280 300 320
T/ K T/ K

Figure S2. Differential scanning calorimetry curves of samples with (a) xthymol = 0.75 and (b) xthymol = 0.70 prepared
by slow crystallization (yellow curve) and rapid crystallization (blue curve). The circles mark the glass transition
temperature. Curves are shifted and scaled for clarity.

247.7 K

271.7 K

200 220 240 260 280 300

T/ K
Figure S3. Differential scanning calorimetry curve of samples with 1:1 L-menthol:thymol molar ratio obtained by
different conditions. Curves are shifted and scaled for clarity. Legend: (orange dashed line) in-situ crystalized at
the DSC; (orange line) samples stored at 193 K for more than one month; (blue line) samples quenched at 193 K
and annealed at 253 K.

4
 endo

xthymol = 0.88

xthymol = 0.85

xthymol = 0.83

xthymol = 0.75

xthymol = 0.67

xthymol = 0.60

xthymol = 0.50

xthymol = 0.40

xthymol = 0.33
xthymol = 0.30
xthymol = 0.20

xthymol = 0.10

250 260 270 280 290 300 310 320 330

T/K
Figure S4. Differential scanning calorimetry curves of samples with different thymol mole fractions.

Table S4. Cell parameters from indexing and Pawley refinement of a powder X-ray diffraction pattern of a sample
with xthymol = 0.67
Monoclinic P
a 11.9451 Å
b 19.1600 Å
c 16.7199 Å
99.428°
V 3774.96 Å3
Z ca. 16

X-ray Crystallographic Details

Data were collected on a Bruker D8 Venture single crystal x-ray diffractometer equipped with a CPAD
detector (Bruker Photon II), an IMS micro source with MoK radiation (λ = 0.71073 Å) and a Helios optic
using the APEX3 software package.2 Measurements were performed on single crystals coated with
perfluorinated ether. The crystals were fixed on top of a Kapton micro sampler and frozen under a stream
of cold nitrogen. A matrix scan was used to determine the initial lattice parameters. Reflections were
corrected for Lorentz and polarisation effects, scan speed, and background using SAINT.3 Absorption
correction, including odd and even ordered spherical harmonics, was performed using SADABS.3 Space

5
group assignments were based upon systematic absences, E statistics, and successful refinement of the
structures. The structures were solved using SHELXT with the aid of successive difference Fourier maps
and were refined against all data using SHELXL in conjunction with SHELXLE.4,5,6 Hydrogen atoms
except on heteroatoms were calculated in ideal positions as follows: Methyl H atoms were refined as part
of rigid rotating groups, with a C H distance of 0.98 Å and Uiso(H) = 1.5·Ueq(C). Non-methyl H atoms were
placed in calculated positions and refined using a riding model, with methylene, aromatic, and other C H
distances of 0.99 Å, 0.95 Å and 1.00 Å, respectively, and Uiso(H) = 1.2·Ueq(C). Non-hydrogen atoms were
refined with anisotropic displacement parameters. Full-matrix least-squares refinements were carried out
by minimizing w(Fo2 - Fc2)2 with the SHELXL weighting scheme.4 Neutral atom scattering factors for all
atoms and anomalous dispersion corrections for the non-hydrogen atoms were taken from International
Tables for Crystallography.7 Images of the crystal structure were generated with PLATON.8
CCDC 2069355 contains the supplementary crystallographic data for this paper. These data are provided
free of charge by The Cambridge Crystallographic Data Centre.

Figure S5. Molecular structure of thymol in the solid state at 100 K as obtained from single-crystal x-ray diffraction
data with ellipsoids at the 50% probability level (disorder of the C10-methyl group is omitted). The structure is
analogous to a previously published structure at room temperature. 9

Table S5. Structure and refinement data

Deposition number CCDC 2069355
Chemical formula C10H14O
Crystal description colourless fragment
Formula weight 150.21
Temperature 100 ± 2 K
Wavelength 0.71073 Å
Crystal size 0.442 × 0.528 × 0.547 mm
Crystal system trigonal
Space group R 3
Unit cell dimensions a = 14.4552 ± 4 Å a = 90°
b = 14.4552 ± 4 Å b = 90°
c = 23.0330 ± 10 Å c = 120°
Volume 4168.0 ± 3 Å3
Z 18
Density (calculated) 1.077 g cm 3

6
 Absorption coefficient 0.067 mm 1
F(000) 1476
a ge 2.4-26.7°
Absorption correction multi-scan (SADABS)
Tmin, Tmax 0.712, 0.745
Reflections measured 41260
Independent reflections 1971
Ref ec I > 2 (I) 1909
Rint 0.021
Refinement method full matrix least squares on F2
Data, restraints, parameters 1971, 0, 108
R1 [I > 2 (I)] 0.045
wR2 (all data) 0.112
Goodness of fit 1.06
Weighting scheme W = 1/[ 2(FO2) + (0.0449P)2 + 5.2473P] where
P = (FO2 + 2FC2)/3
Largest difference peak and hole 0.27 and -0.21 eÅ-3

References

1. Alhadid, A.; Mokrushina, L.; Minceva, M., J. Mol. Liq. 2020, 314, 113667.
2. APEX suite of crystallographic software, APEX 3, Version 2019-1.0, Bruker AXS Inc., Madison,
Wisconsin, USA, 2016.
3. SAINT, Version 8.40A and SADABS, Version 2016/2, Bruker AXS Inc., Madison, Wisconsin, USA,
2016/2019.
4. Sheldrick, G. M. Acta Crystallogr. Sect. A 2015, 71, 3 8.
5. Sheldrick, G. M. Acta Crystallogr. Sect. C 2015, 71, 3 8.
6. Hübschle, C. B.; Sheldrick, G. M.; Dittrich, B. J. Appl. Cryst. 2011, 44, 1281 1284.
7. International Tables for Crystallography, Vol. C (Ed.: A. J. Wilson), Kluwer Academic Publishers,
Dordrecht, The Netherlands, 1992, Tables 6.1.1.4 (pp. 500 502), 4.2.6.8 (pp. 219 222), and 4.2.4.2 (pp.
193 199).
8. Spek, A. L. Acta Crystallogr. Sect. D 2009, 65, 148 155.
9. Thozet, A.; Perrin, M. Acta Crystallogr. Sect. B 1980, 36, 1444-1447. Deposition number: CCDC
1180630.

7
3.5 Paper V

3.5 Paper V
Cocrystal Formation in Choline Chloride Deep Eutectic Solvents

A. Alhadid, C. Jandl, L. Mokrushina, and M. Minceva, Cryst. Growth Des., 2022, 22, 3, 1933–
1942.

Author contribution: The thesis author conceptualized the paper’s idea, performed the
investigations and formal analysis (excluding powder and SC-XRD), interpreted the results, and
wrote the manuscript.

Summary: ChCl is one of the most studied HBA in the DES literature. Because ChCl is thermally
unstable, its melting properties are unavailable. As shown in Paper I, modeling of the nonideality
of thermally unstable salts is implausible. The objectives of Paper V are to highlight cocrystal
formation in ChCl-based DES and examine the melting properties and nonideality of ChCl. ChCl
was mixed with two coformers (dihydroxybenzenes) to form ChCl-based cocrystals. DSC and
powder XRD analyses were employed to obtain the SLE data of the two systems. The
investigations revealed the formation of two cocrystals in ChCl/catechol and one cocrystal in
ChCl/hydroquinone. The cocrystals' stoichiometry and structure were obtained by the SC-XRD
technique. It was concluded that not all ChCl-based eutectic systems could be assumed to be
of the simple eutectic type, and extensive experimental investigations are needed to detect
cocrystal formation.

Next, the SLE data of the coformers and cocrystals liquidus lines were used to obtain the NRTL
binary interaction parameters. The activity coefficients of ChCl in the liquid phase and the SLE
data of its liquidus line were used to assess the melting properties of ChCl. Surprisingly, no
liquidus temperature could be observed in the DSC curves of the samples in the ChCl-rich region
above the solid–solid transition temperature up to the decomposition temperature of the mixture.
The presence of the ChCl solid phase in the samples above the solid–solid transition temperature
was confirmed by variable temperature XRD (VT-XRD). Without liquidus data for pure ChCl above
the solid–solid transition temperature, no distinctive melting properties could be determined for
ChCl. However, the estimated ChCl melting properties estimated in the literature are not within
the range of possible values estimated in this work, indicating that those found in the literature
could be unreasonable.

118
3.5 Paper V

Paper V confirmed that the unique character of ChCl to form nonideal eutectic mixtures with a
large depression at the eutectic point is its small melting entropy. The small melting entropy of
ChCl is attributed to the symmetrical and disordered crystal structure of the high-temperature
polymorph of ChCl. Therefore, Paper V provided evidence for the conclusions drawn in Paper I
that DES are nonideal eutectic mixtures formed by mixing constituents with low melting enthalpy
and entropy values.

119
3.5 Paper V

Reprinted with permission from Alhadid et al., Cryst. Growth Des., 2022, 21, 6083-6091. DOI:
10.1021/acs.cgd.1c01477. Copyright 2022 American Chemical Society.

Article full text: http://pubs.acs.org/articlesonrequest/AOR-UVXRX3M2KU4DDTIVN2MA

120
 pubs.acs.org/crystal Article

Cocrystal Formation in Choline Chloride Deep Eutectic Solvents

Ahmad Alhadid, Christian Jandl, Liudmila Mokrushina, and Mirjana Minceva*
Cite This: Cryst. Growth Des. 2022, 22, 1933−1942 Read Online

ACCESS Metrics & More Article Recommendations *

sı Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via TU MUENCHEN on May 24, 2022 at 07:45:07 (UTC).

ABSTRACT: Deep eutectic solvents (DESs) are eutectic mixtures representing a new generation of tunable solvents. The majority
of the DESs studied in the literature contains choline chloride (ChCl). The knowledge of solid−liquid equilibria (SLE) in the DESs
is crucial for the identification of the composition range in which the mixture is liquid below process-relevant temperatures. ChCl-
based DES phase diagrams are usually assumed to be simple eutectics with no cocrystal formation. However, the simple eutectic
assumption is questionable without a detailed characterization of the crystallized solid phases. This study investigates the formation
of cocrystals in ChCl-based DESs. The SLE in ChCl/catechol and ChCl/hydroquinone eutectic mixtures were studied using powder
X-ray diffraction (XRD) and differential scanning calorimetry (DSC). The research demonstrated the formation of two cocrystals in
the ChCl/catechol system and one cocrystal in the ChCl/hydroquinone system; the crystal structures were obtained using single-
crystal XRD. The solid−solid transition of ChCl in the mixture was observed by DSC and variable-temperature XRD. The SLE data
were correlated using the nonrandom two-liquid (NRTL) model and the melting properties of the pure components and cocrystals.
Modeling the SLE data enabled one to evaluate the melting properties of ChCl estimated in the literature.

1. INTRODUCTION the melting temperature for a few mixture compositions and

Recently, eutectic mixtures have been used to form tunable and then correlate the SLE data assuming that the system is of the
green solvents, often called deep eutectic solvents (DESs).1,2 simple eutectic type, i.e., the components crystallize as pure
solids.13,14 However, the simple eutectic assumption should be
DESs are prepared by mixing hydrogen bond acceptors (HBAs)
verified. In various eutectic systems, polymorphism, the
with hydrogen bond donors (HBDs).3 The eutectic temperature
formation of metastable solid phases, and cocrystal formation
of the mixture is considerably lower than the melting
have been identified by the characterization of the crystallized
temperature of the pure constituents. The reasons for the low
solid phases.15−19 Therefore, the crystallized solid phases should
eutectic temperature of the DESs are the negative deviation from
be analyzed to determine and model the SLE phase diagram
the ideality in the liquid phase and the low melting enthalpy
properly.
value of the constituents.4
Choline chloride (ChCl) was the first HBA used in DESs and
A prerequisite for using a DES as a solvent is that the melting
the one present in over half of the DESs reported in the
temperature of the mixture is below the process temperature.
literature.1,20−23 The distinctive character of ChCl in forming
The melting temperature of the mixture at any composition is
DESs with a considerably low eutectic temperature has not been
obtained from the solid−liquid equilibria (SLE) phase diagram.4 well-explored so far. ChCl has been shown to undergo a solid−
Hence, the knowledge of SLE is essential for the selection of the
DES constituents and their molar ratios. Recently, the SLE
phase diagram of numerous binary eutectic mixtures has become Received: December 17, 2021
available in the literature to aid in the design of the DES for a Revised: January 24, 2022
certain application.5−12 However, experimental determination Published: February 8, 2022
of SLE data over the entire composition range is not a
straightforward task. The most commonly used approach for the
determination of the complete SLE phase diagram is to measure
© 2022 The Authors. Published by
American Chemical Society https://doi.org/10.1021/acs.cgd.1c01477
1933 Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

solid transition at around 352 K.24 However, the solid−solid of 0.015° 2θ per data point. Due to ChCl radiation sensitivity, the
transition of ChCl in its eutectic mixtures was indirectly measurements were performed at 253 K to avoid the decomposition of
observed only in a few cases as the change in the slope of the ChCl. A cryostream (Oxford 800 series, UK) was used to perform
ChCl liquidus line.10,25 Accordingly, it was argued that the ChCl measurements at various temperatures.
The cocrystal’s stoichiometry and structure were determined by
low-temperature polymorph is stabilized, and thus, the solid− single-crystal XRD (SC-XRD) analysis. Single crystals with sufficient
solid transition cannot be observed in most ChCl-based DESs, size and quality for SC-XRD were obtained by dissolving the cocrystal
relying only on the measurements of the liquidus temperature of powder in acetonitrile at 318 K in a 1:1 mass ratio, and the solution was
ChCl.5 However, it is hardly probable that the liquid phase could stored at 253 K for around 1 week until suitable crystals appeared.
stabilize ChCl low- or high-temperature polymorphs because Details regarding the SC-XRD method can be found in the Supporting
the liquid phase is only in contact with the surface but not with Information.
the bulk of the solid ChCl. Therefore, it is required to directly 2.3. DSC. DSC analysis was used to get the SLE data for the two
examine the solid−solid transition of ChCl in ChCl-based DESs eutectic systems. Before measurement, the DSC instrument
by analysis methods such as X-ray diffraction (XRD). (NETZSCH DSC 200 F3, Germany) was calibrated using six
calibration standards, namely, adamantane, bismuth, cesium chloride,
The formation of cocrystals in ChCl-based DESs has not been indium, tin, and zinc, at a heating rate of 5 K min−1. Measurements were
thoroughly disclosed.26−30 Various HBDs used in ChCl-based performed in an inert environment using nitrogen with a flow rate of
DESs, such as urea, thiourea, catechol, and hydroquinone, are 150 mL min−1. The standard uncertainties of the temperature and
known to be cocrystal formers.31 However, without systematical sensitivity measurements were 0.3 K and 0.4%, respectively.
characterization of the crystallized solid phases in ChCl-based After filling the DSC crucible pans with the ground powder in
DES, the formation of cocrystals can be ruled out. This study triplicates, they were hermetically sealed. Before adding the sample, the
presents a detailed investigation of SLE in two ChCl-based DSC chamber was precooled to 253 K. Then, a heating run with a
DESs, namely, ChCl/catechol and ChCl/hydroquinone, with an heating rate of 5 K min−1 was conducted until 373 K for ChCl/catechol
extensive analysis of the crystallized solid phases. The crystal- and 443 K for ChCl/hydroquinone systems. The solidus phase
transition and the liquidus temperatures were determined as the onset
lized solids were characterized by performing powder XRD on and peak maximum temperatures. DSC experiments were performed in
samples of different compositions. The SLE data of the two triplicate for each mixture composition being studied. The average
eutectic systems were measured using differential scanning standard deviation of the obtained liquidus and solidus temperatures
calorimetry (DSC) over the entire composition range. The was found to be 0.5 K.
obtained SLE data were modeled using the nonrandom two- 2.4. Thermodynamic Modeling. The SLE data obtained from the
liquid (NRTL) equation. DSC analysis were modeled to obtain the complete solid−liquid phase
diagram and the eutectic points. The pure constituent liquidus line was
2. METHODS modeled using the following equation32
2.1. Sample Preparation. Figure 1 shows the molecular structure Δhm, i T
of the pure substances. Before the eutectic mixtures were prepared, ln xiLγi L = − 1−
RT Tm, i (1)
xLi γLi
where and are the mole fraction and the activity coefficient of the
component i in the liquid phase, Δhm,i and Tm,i are the melting enthalpy
and temperature of pure component i, R is the gas constant, and T is the
liquidus temperature. Since ChCl undergoes a solid−solid transition,24
the ChCl liquidus line below the solid−solid transition temperature was
calculated as follows33
Δhm, i T Δhtr, i T
Figure 1. Molecular structure of the studied pure substances. ln xiLγi L = − 1− − 1−
RT Tm, i RT Ttr, i (2)
ChCl (Alfa Aesar, <98%) was dried overnight under vacuum at 343 K, where Δhtr,i and Ttr,i are the solid−solid transition enthalpy and
and its water content after drying was measured using a Karl Fischer temperature, respectively. The formation of a cocrystal containing
Coulometer (Hanna Instrument, USA) and found to be <1000 ppm. component A with a stoichiometric coefficient of ϑA and component B
The eutectic mixtures were prepared by mixing ChCl with catechol with a stoichiometric coefficient of ϑB in a binary eutectic mixture was
(Acros Organics, <99%, water content ∼ 500 ppm) or hydroquinone described as a chemical reaction.34
(Merck, <98%, water content ∼ 700 ppm) in different ratios. The pure ϑAA (L) + ϑBB(L) ↔ A ϑA B ϑB(S)
components were weighed (Sartorius, Germany, precision 1 × 10−4 g) (3)
and introduced into glass vials. The vials were directly sealed, and the The equilibrium constant of the chemical reaction (Ka) in eq 3 was
mixture was gently heated under continuous stirring until a clear liquid defined as
was formed. To aid sample crystallization, the liquid solutions were
quenched at 193 K and annealed for 1 day at 253 K. Using a mortar and (xALγAL)ϑA (x BLγBL)ϑB
pestle, the solid was ground to a fine powder inside a cold room at 253 Ka = ∏ aiϑi = S S
K. The water content of the powder was measured after grinding and i (xABγAB) (4)
found to be less than 500 ppm. The solid phase is a pure cocrystal; hence, = 1. The xSABγSAB
2.2. XRD. Powder XRD was used to characterize the crystalline solid temperature effect on the melting enthalpy of the cocrystal was
phases across the entire composition range. The powder XRD samples neglected in a narrow temperature range, and Ka at different
were prepared by grinding the powder at 253 K and directly adding it temperatures was calculated using the Gibbs−Helmholtz equation
into a glass capillary (Hilgenberg, Germany) with a 1.0 mm diameter.
The powder XRD experiments were performed using a Stadi P Δhref 1 1
diffractometer (Stoe & Cie, Germany) with Debye−Scherrer geometry ln K a = ln K aref + −
R T ref T (5)
equipped with a Mo fine-focus sealed tube, curved Ge monochromator
ref
selecting for Kα1 radiation (λ = 0.70930 Å), and a Mythen2 R 1K where T is a reference temperature, Kref
is the equilibrium constant at
a
detector (Dectris, Switzerland). The data were collected with a step size Tref, and Δhref is the melting enthalpy of the cocrystal at Tref. Tref and

1934 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 2. X-ray diffraction patterns of choline chloride (ChCl) and catechol (cat) binary eutectic mixtures for samples of different ratios. The
measurements were performed at 253 K. Characteristic peaks are marked as follows: catechol (blue dotted lines), 1:2 cocrystal (orange dotted lines),
1:1 cocrystal (green dotted lines), and ChCl (red dotted lines).

Δhref are the melting temperature and enthalpy of the cocrystals 3. RESULTS
measured in this work using DSC. The equilibrium constant at Tref was
calculated as follows 3.1. XRD. The formation of cocrystals was monitored by
powder XRD analysis. Figure 2 shows the powder XRD pattern
Karef = (xAref γAref )ϑA (x Bref γBref )ϑB (6) obtained at 253 K for pure catechol, ChCl, and a ChCl/catechol
ref ref
where x and γ are the mole fraction and the activity coefficient of the mixture of different ratios. Starting with the bottom curves,
component at which Tref and Δhref were measured, i.e., at the cocrystal which depict catechol-rich samples, the XRD pattern of the 1:3
stoichiometry. Equations 4−6 were solved simultaneously to obtain the ratio sample shows the main peaks of the catechol XRD pattern
liquidus lines of the cocrystal. (blue dotted lines) with additional peaks that do not correspond
The activity coefficients of the components in the liquid phase (γLi ) to the peaks of the pure ChCl XRD pattern. When the XRD
are prerequisites in calculating the pure components (eqs 1 and 2) and patterns of the 1:3 and 1:2 ratio samples were compared, the
cocrystal (eqs 4−6) liquidus lines. The activity coefficients of the additional peaks observed in the 1:3 ratio sample match with
components in the liquid phase were calculated using the NRTL model those in the 1:2 ratio sample XRD pattern (orange dotted lines),
as follows32 indicating that the crystallized solid of the 1:3 ratio sample
Gji
2
τijGij
consists of pure catechol and the solid phase of the 1:2 ratio
L 2 sample. The absence of pure catechol peaks in the XRD pattern
ln γi = xj τji +
xi + xjGji (xj + xiGij)2 of the 1:2 ratio sample indicates the formation of a cocrystal with
(7)
a 1:2 ratio. SC-XRD analysis was performed on a single crystal
Gij = exp(− ατij) Gji = exp(− ατji) (8) obtained from the 1:2 ratio sample to confirm the formation of a
1:2 cocrystal according to the procedure described in Section
gij − gjj gji − gii
τij = τji =
2.2. The cocrystal formation with a stoichiometric ratio of 1:2 for
RT RT (9) ChCl/catechol was confirmed by SC-XRD analysis; the crystal
The nonrandom parameter α was set to 0.3, and the binary structure details are shown in Figures S1 and S2 and Table S3.
interaction parameters (gij − gjj) and (gji − gii) were fitted to the When one moves toward the middle composition range, a
experimental liquidus data (Texp
i ) by minimizing the following objective
comparison between the XRD patterns of the 1:2 and 2:3 ratio
function samples reveals additional peaks in the 2:3 ratio sample XRD
n 1/2 pattern (green dotted lines), which again do not match with the
(Tiexp − Tical)2 peaks of pure catechol or ChCl. These additional peaks
F(T ) = ∑ correspond to the 1:1 ratio sample XRD pattern (green dotted
i=1
n (10)
lines). Thus, the crystallized solid of the 2:3 ratio sample consists
where Tcal
is the calculated liquidus temperature by the NRTL model
i of the 1:2 cocrystal and the solid phase of the 1:1 ratio sample.
and n is the number of experimental data points. The powder XRD pattern of the 1:1 ratio does not match or
The SLE data on the mixtures rich in ChCl were used to determine
the melting properties of thermally unstable ChCl. For this, eq 2 was
contain any peaks corresponding to pure catechol, ChCl, or the
rearranged as follows33 1:2 cocrystal, which may indicate the formation of a second
cocrystalline phase at the 1:1 ratio. Accordingly, SC-XRD
Δhm, i Δhtr, i 1 Δhm, i Δhtr, i analysis was performed on a single crystal obtained from the 1:1
ln xiLγi L = − − × + +
R R T RTm, i RTtr, i (11)
ratio sample. The results confirmed the formation of a cocrystal
with a stoichiometric ratio of 1:1 for ChCl/catechol; the crystal
Given that the NRTL model is available for the calculation of the structure details are shown in Figures S3 and S4 and Table S4.
activity coefficients of ChCl in the liquid phase, experimental ln xLi γLi Finally, the peaks of the pure ChCl XRD pattern (red dotted
values were plotted as a function of 1/T. Accordingly, the melting lines) can be seen in the 2:1 ratio sample XRD pattern, showing
enthalpy and temperature of ChCl were determined from the slope
that the crystallized solid of the 2:1 ratio sample consists of pure
(− Δhm, i
R
−
Δh tr, i
R ) and intercept ( Δhm, i
RTm, i
+
Δh tr, i
RTtr, i ). ChCl and the 1:1 cocrystal. We could perform quantitative
1935 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 3. X-ray diffraction pattern of choline chloride (ChCl) and hydroquinone (hyd) binary eutectic mixture for samples of different ratios. The
measurements were performed at 253 K. Characteristic peaks are marked as follows: hydroquinone (blue dotted lines), 1:1 cocrystal (green dotted
lines), and ChCl (red dotted lines).

phase analysis via Rietveld refinement of the powder XRD data systems. The melting properties of catechol, hydroquinone, and
for the samples shown in Figure 2 within the ChCl/catechol the three cocrystals as well as the solid−solid transition enthalpy
system (see Tables S6 and S7) because we had determined the and temperature of ChCl were measured using DSC. The
crystal structures of the 1:2 and 1:1 cocrystals, which confirmed corresponding DSC curves can be found in Figure S7. Table 1
the above assignments.
Next, powder XRD investigations were performed on the Table 1. Pure Constituents and Cocrystal Melting Properties
ChCl/hydroquinone eutectic system samples covering the Measured in This Work Compared to the Data Found in the
entire composition range. Figure 3 shows the powder XRD Literature
pattern obtained at 253 K for pure hydroquinone, ChCl, and
Δhm/kJ mol−1 Tm/K
ChCl/hydroquinone samples of varying ratios showing distinct
powder XRD patterns. Starting with hydroquinone-rich solid phase this work literature this work literature
samples, the XRD pattern of the 1:2 ratio sample shows the catechol 21.99 ± 0.61 22.8736 377.1 ± 0.1 377.636
peaks corresponding to the pure hydroquinone XRD pattern 1:2 ChCl:cat 39.54 ± 0.61 325.7 ± 0.1
(blue dotted lines). Conversely, the additional peaks in the 1:2 1:1 ChCl:cat 34.15 ± 0.11 327.4 ± 0.4 325.235
ratio sample XRD pattern do not correspond to the peaks seen in hydroquinone 28.73 ± 0.15 27.2336 445.7 ± 0.1 445.1
the pure ChCl XRD pattern. These peaks correspond to the 1:1 ChCl:hyd 31.97 ± 0.38 332.7 ± 0.1 332.235
XRD pattern of the 1:1 ratio sample (green dotted lines). Thus, solid−solid transition
the crystallized solid phase of the 1:2 ratio sample consists of Δhtr/kJ mol−1 Ttr/K
pure hydroquinone and the solid phase of the 1:1 ratio sample. this work literature this work literature
As seen, no hydroquinone or ChCl peaks were observed in the
choline chloride 16.35 ± 0.67 16.5337 352.3 ± 0.1 35137
1:1 ratio sample XRD pattern, hinting at the possibility of a
cocrystal formation at the 1:1 ratio. Thus, SC-XRD analysis was
performed on a single crystal obtained from the 1:1 ratio sample. shows the measured melting properties compared to those
The formation of a cocrystal with a stoichiometric ratio of 1:1 for reported in the literature. As seen, the melting temperatures of
ChCl/hydroquinone (Figure S5 and Table S5) was confirmed the pure components measured in the present work agree with
by the SC-XRD results. When one refers to Figure 3, the 2:1 the literature values. Abbott et al.35 reported the melting
sample ratio XRD pattern contains the peaks of pure ChCl and temperatures of the 1:1 molar ratio mixture in the two eutectic
the 1:1 cocrystal, which indicates that the crystallized solid phase systems as eutectic temperatures. The reported values match
of the 2:1 ratio sample consists of pure ChCl and the 1:1 well with the melting temperatures of the 1:1 cocrystals obtained
cocrystal. Consequently, one cocrystal can be identified in the in the present work but were misinterpreted as eutectic points.
ChCl/hydroquinone eutectic system. Finally, quantitative phase Figure 4 shows SLE data of the two eutectic systems over the
analysis by Rietveld refinement of the powder XRD data of the whole composition range measured using DSC in this work
samples shown in Figure 3 within the ChCl/hydroquinone compared to the data found in the literature (indicated by
system confirmed the above assignments (see Tables S6 and crosses). As seen in Figure 4A, the ChCl/catechol eutectic
S8). system shows three solidus temperatures (∼314, 319, and 323
3.2. DSC. The powder XRD analysis revealed the formation K) attributed to the formation of two cocrystals. As shown in
of two cocrystals in the ChCl/catechol (1:2 and 1:1 cocrystals) Figure 4B, the ChCl/hydroquinone eutectic system shows the
and one cocrystal in the ChCl/hydroquinone (1:1 cocrystal) formation of one cocrystal and two different solidus temper-
eutectic systems. Thus, the phase diagram is not of a simple atures (∼318 and 326 K).
eutectic type, and the melting properties of the cocrystals are For mixtures rich in ChCl, no liquidus temperature could be
needed for adequate modeling of the SLE in the two eutectic observed above a certain temperature value. This temperature
1936 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 4. Measured solid−liquid equilibria data for (A) choline chloride (ChCl)/catechol (cat) and (B) ChCl/hydroquinone (hyd). Literature data
were taken from Abbott et al.35 (C) Differential scanning calorimetry curves of samples from the ChCl/catechol eutectic system with xcat = 0.38 and
0.34. The curves were shifted for clarity.

Figure 5. Variable temperature powder X-ray diffraction for the sample with xcat = 0.34 from choline chloride (ChCl)/catechol. The solid−solid (S−S)
transition and eutectic temperature (Te) in the ChCl-rich region are shown.

(T = 352.3 K) corresponds to the solid−solid transition of ChCl. the mixture (T > 550 K). The same behavior was observed for all
The DSC curves of the samples with xcat = 0.38 and 0.34 are samples with xChCl > 0.66 in the ChCl/catechol and ChCl/
shown in Figure 4C. Although the liquidus temperature of the hydroquinone eutectic systems. The data found in the literature
sample with xcat = 0.38 is 342.3 K, no liquidus temperature at xcat for the ChCl/hydroquinone eutectic system (cross symbols in
= 0.34 could be measured above the solid−solid transition Figure 4B) hint at the observed behavior, but the solid−solid
temperature of ChCl up to the decomposition temperature of transition temperature of ChCl has been misinterpreted as
1937 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

liquidus temperatures. The variable temperature XRD (VT- compositions and temperatures for the two eutectic systems.
XRD) experiment carried out on pure ChCl (Figure S8) The binary interaction parameters with negative values indicate
confirmed the fact of the solid−solid transition. a significant negative deviation from the ideal behavior, i.e.,
Further investigations were performed to validate the DSC strong intermolecular interactions between unlike molecules in
observations. VT-XRD was performed on a sample from the the liquid phase (γi < 1). Figure 6 shows the calculated liquidus
ChCl/catechol eutectic system with xcat = 0.34 in the (solid blue lines) and solidus lines (dashed blue lines) compared
temperature range from 263 to 403 K; the results are shown to the measured data (points). The calculated liquidus and
in Figure 5. As seen in Figure 5, the XRD pattern remains the solidus lines are in good agreement with the experimental data,
same from 263 to around 323 K. At 323 K, several peaks in the as seen in Figure 6. The model can adequately describe the
XRD pattern disappear while the remaining peaks fade slightly. formation of the cocrystals and determine the position of the
The vanishing and fading of the XRD pattern peaks indicate eutectic points.
partial melting of the solid phase of the sample. Thus, this As shown in the previous section, the liquidus temperatures of
temperature corresponds to the eutectic temperature observed ChCl above its solid−solid transition temperature could not be
in the ChCl-rich region, i.e., the melting of the 1:1 cocrystal and measured. Nevertheless, the activity coefficients of ChCl in the
part of pure ChCl. At temperatures higher than 323 K, only the ChCl-rich liquid phase can be calculated by the NRTL model
XRD peaks corresponding to pure ChCl are observed. The using the binary interaction parameters from Table 2. Thus, the
solid−solid transition of ChCl can be observed as a change in the ChCl liquidus line above the solid−solid transition could be, in
XRD pattern at around 353 K. Above the solid−solid transition principle, modeled if the melting properties of ChCl were
temperature, no vanishing of the peaks of the high-temperature known. However, the melting properties of ChCl are not
polymorph of ChCl is observed, indicating that pure ChCl available experimentally due to its thermal instability. Fernandez
remains as a solid in the sample. et al.38 estimated the melting properties of ChCl as Tm = 597 K
Nevertheless, liquidus temperatures for ChCl above its solid− and Δhm = 4.3 kJ mol−1. The estimation was made based on the
solid transition temperature were reported for other ChCl-based SLE data of ten binary eutectic mixtures containing ChCl with
DESs.5,7−10,38 However, the reported data were obtained by a various ionic salts and assuming an ideal solution behavior of
melting-point device and not by DSC. This method relies on ChCl in the liquid solutions, i.e., γChCl = 1. Vilas-Boas et al.39
visual determination of the mixture melting temperature, and used ChCl water solubility data and the PC-SAFT or COSMO-
thus, some of the phase transitions, such as solid−solid solidus RS models to estimate the melting enthalpy of ChCl as Δhm =
(eutectic) transitions, cannot be captured or can be misinter- 7.67 kJ mol−1, assuming the melting temperature estimated in
preted. We recommend the detailed reinvestigation of the Fernandez et al.38 In both studies, the systems were assumed to
crystallized solid phases in ChCl-based DESs to confirm that be of the simple eutectic type without accounting for the solid−
ChCl undergoes the solid−solid transition and does not melt solid transition of ChCl.
above its solid−solid transition temperature regardless of the Figure 7 shows the ChCl-rich region (xChCl > 0.5) in the phase
diagram of ChCl/catechol (Figure 7A) and ChCl/hydro-
HBD.
quinone (Figure 7B) eutectic mixtures. The blue lines in the
3.3. Thermodynamic Modeling. The correlation of
figure represent the liquidus lines of the 1:1 cocrystals as
experimental SLE data is a valuable tool to obtain the complete
correlated with the NRTL model. The red and orange lines
solid−liquid phase diagram and determine the position of the
indicate the liquidus lines of ChCl calculated using the literature
eutectic points. As the melting properties of ChCl are unknown,
values of the melting properties of ChCl found in Fernandez et
the correlation of the liquidus line SLE data (xChCl > 0.6) to
al.38 (Δhm = 4.30 kJ mol−1, Tm = 597 K) and Vilas-Boas et al.
obtain the NRTL binary interaction parameters is not possible.
(Δhm = 7.67 kJ mol−1, Tm = 597 K),39 respectively. As seen, none
Thus, for each system being studied, the experimental data of the published melting properties of ChCl can adequately
obtained for the liquidus lines of catechol, hydroquinone, and describe the measured data in the composition range of xChCl >
the cocrystals were correlated using the NRTL model and based 0.60. Therefore, the reported ChCl melting properties seem
on the melting properties of the pure components and the inadequate for modeling its liquidus line accurately in the ChCl/
cocrystals from Table1. Table 2 displays the obtained binary catechol and ChCl/hydroquinone eutectic systems.
interaction parameters, the RMSD between experimental and Although the melting properties of ChCl are unknown, its
estimated liquidus temperatures, and the calculated eutectic crystal structure may provide a premise about its melting
properties. ChCl has a highly disordered crystal structure above
Table 2. NRTL Binary Interaction Parameters, RMSD, the solid−solid transition.24 The solid−solid transition entropy
Estimated Eutectic Point from Solid−Liquid Equilibria
Correlation, and the Absolute Difference between Measured
and Calculated Eutectic Temperatures
of ChCl (Δs tr =
Δh tr
Ttr
= 46.41 J mol−1 K−1 ) resembles the
melting entropy of many other solids, indicating the high
entropy of the high-temperature ChCl solid phase. Thus, the
ChCl/catechol ChCl/hydroquinone
entropy values of the solid and the liquid phases at the melting
(gij − gjj)/kJ mol−1 −9.8501 −11.4683 temperature should be similar, and therefore, the melting
(gji − gii)/kJ mol−1 −9.3442 −10.8958 entropy should be of a low value. The melting entropy of
RMSD/K 2.0 9.6 substances with highly disordered crystal structure ranges
xe 0.74
0.59 between 0.5 and 2 R (4.16 to 16.63 J mol−1 K−1).40 The
0.60 melting temperature is the ratio between the melting enthalpy
Te/K 312.1 and entropy of the component, and thus, the melting properties
315.5
316.0 are interrelated.
|Texp
e − e |/K
Tcal 3.0 To estimate the melting properties of ChCl, the experimental
1.7
2.9 SLE data on the ChCl liquidus line below the solid−solid
1938 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 6. Solid−liquid phase diagram of (A) choline chloride (ChCl)/catechol (cat) and (B) ChCl/hydroquinone (hyd) eutectic systems in the
hydrogen bond donor-rich region. The liquidus and solidus lines were calculated using the NRTL model.

Figure 7. Solid−liquid phase diagram of (A) choline chloride (ChCl)/catechol and (B) ChCl/hydroquinone. The ChCl liquidus line was calculated
using the nonrandom two-liquid (NRTL) model and the melting properties determined in Fernandez et al.38 (Δhm = 4.30 kJ mol−1, Tm = 597 K) or
Vilas-Boas et al.39 (Δhm = 7.67 kJ mol−1, Tm = 597 K).

transition in each eutectic system were fitted by eq 11 at various Simultaneously, these values would correspond to high values
melting entropy values in the range from 0.5 to 2 R; the results of melting temperatures. Furthermore, the melting temperature
are shown in Figure 8. Each value of the predicted melting of ChCl at the highest possible value for the melting entropy of
entropy yields a pair of melting temperature and enthalpy values substances with a disordered crystal structure like ChCl (2 R) is
due to a lack of experimental evidence regarding the liquidus 675.1 and 717.9 K, which are higher than the melting
temperatures of the high-temperature polymorph of ChCl. As temperature reported by Fernandez et al.38 (Tm = 597 K) and
shown in Figure 8A,B, any of these pairs adequately describe the correspond to the melting enthalpy of 11.22 and 11.94 kJ mol−1.
liquidus lines of the low-temperature polymorph of ChCl, but as Thus, the reported values of the melting properties of ChCl
expected, they lead to distinct courses of a high-temperature seem unreasonable. Hence, it is necessary to find another
polymorph’s liquidus line. Thus, the estimation of any definite approach to estimate the melting properties of ChCl.
values of the ChCl melting properties without the experimental
SLE data for the high-temperature polymorph seems unattain-
4. CONCLUSION
able. However, the obtained data allow one to analyze the
melting properties of the high-temperature polymorph. This work presents a detailed study of the SLE in ChCl/catechol
As seen in Figure 8C,D, the obtained melting enthalpy and ChCl/hydroquinone eutectic systems. The SLE data of the
increases with an increase in the melting entropy, while the two binary eutectic systems were measured over the whole
melting temperature, on the contrary, decreases. As shown in composition range by DSC analysis. Two cocrystals with
Figure 8C, the minimum feasible value of melting enthalpy of stoichiometric ratios of 1:2 and 1:1 in the ChCl/catechol
ChCl, i.e., at Δsm → 0, is about 6.26 or 5.58 kJ mol−1, which is eutectic system and one cocrystal with a stoichiometric ratio of
still higher than that estimated by Fernandez et al.38 1:1 in the ChCl/hydroquinone eutectic system were observed.
1939 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 8. Solid−liquid phase diagram of (A) choline chloride (ChCl)/catechol and (B) ChCl/hydroquinone. Obtained (C) melting enthalpy and (D)
temperature of choline chloride from its solid−liquid equilibria data assuming different melting entropy values.

Solid ChCl was shown to undergo the solid−solid transition at

352 K in the two studied systems. The formation and structure
■
*
ASSOCIATED CONTENT
sı Supporting Information
of the cocrystals and the ChCl solid−solid transition were
confirmed by SC-XRD and VT-XRD analysis, respectively. The Supporting Information is available free of charge at
The experimental data were correlated with the NRTL model https://pubs.acs.org/doi/10.1021/acs.cgd.1c01477.
for both studied systems. Substantial negative deviation from the Single crystal X-ray diffraction (SC-XRD) details, solid−
ideal behavior was observed for both constituents in both liquid equilibria data of choline chloride/catechol, details
systems. Due to the lack of experimental data on the liquidus on SC-XRD, solid−liquid equilibrium data of the ChCl/
lines of the ChCl high-temperature polymorph, the melting catechol and ChCl/hydroquinone eutectic system, SC-
properties of ChCl could not be estimated on the basis of the XRD data for the ChCl/catechol cocrystal with ratios of
measured SLE data. However, the close analysis of the ChCl 1:1 and 1:2, SC-XRD data for the ChCl/hydroquinone
crystal structure could show that the values of its melting cocrystal with a ratio of 1:1, DSC curves of cocrystals and
properties estimated in the literature could not be considered pure components of ChCl/catechol and ChCl/hydro-
reasonable. quinone eutectic systems, Rietveld refinement results for
It was found in this study that the prevalent assumption that powder X-ray data, and VT-XRD study of ChCl (PDF)
all ChCl-based DESs are of the simple eutectic type should be
examined. The quest to model the phase diagram of DESs Accession Codes
should be based on reliable melting properties of pure CCDC 2125060−2125062 contain the supplementary crystallo-
constituents and on verified SLE data of the system. The latter graphic data for this paper. These data can be obtained free of
could be obtained by a combination of DSC analysis with a charge via www.ccdc.cam.ac.uk/data_request/cif, or by email-
rigorous XRD characterization of the formed solid phases. To
conclude, we encourage the reinvestigation of the SLE data of ing [email protected], or by contacting The Cam-
ChCl-based DESs available in the literature guided by the bridge Crystallographic Data Centre, 12 Union Road, Cam-
procedure of the present study. bridge CB2 1EZ, UK; fax: +44 1223 336033.
1940 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

■ AUTHOR INFORMATION
Corresponding Author
(11) Kollau, L. J. B. M.; Tuinier, R.; Verhaak, J.; den Doelder, J.; Filot,
I. A. W.; Vis, M. Design of Nonideal Eutectic Mixtures Based on
Correlations with Molecular Properties. J. Phys. Chem. B 2020, 124
Mirjana Minceva − Biothermodynamics, TUM School of Life (25), 5209−5219.
Sciences, Technical University of Munich (TUM), Freising (12) Kollau, L. J. B. M.; Vis, M.; van den Bruinhorst, A.; Tuinier, R.; de
85354, Germany; Email: [email protected] With, G. Entropy models for the description of the solid−liquid regime
of deep eutectic solutions. J. Mol. Liq. 2020, 302, 112155.
Authors
(13) González de Castilla, A.; Bittner, J. P.; Mü l ler, S.;
Ahmad Alhadid − Biothermodynamics, TUM School of Life Jakobtorweihen, S.; Smirnova, I. Thermodynamic and Transport
Sciences, Technical University of Munich (TUM), Freising Properties Modeling of Deep Eutectic Solvents: A Review on gE-
85354, Germany; orcid.org/0000-0003-1443-1517 Models, Equations of State, and Molecular Dynamics. J. Chem. Eng.
Christian Jandl − Catalysis Research Center, Department Data 2020, 65 (3), 943−967.
Chemie, Technical University of Munich (TUM), Garching (14) Alkhatib, I. I. I.; Bahamon, D.; Llovell, F.; Abu-Zahra, M. R. M.;
85748, Germany Vega, L. F. Perspectives and guidelines on thermodynamic modelling of
Liudmila Mokrushina − Separation Science & Technology, deep eutectic solvents. J. Mol. Liq. 2020, 298, 112183.
Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), (15) Alhadid, A.; Mokrushina, L.; Minceva, M. Formation of glassy
Erlangen 91058, Germany phases and polymorphism in deep eutectic solvents. J. Mol. Liq. 2020,
314, 113667.
Complete contact information is available at: (16) Alhadid, A.; Jandl, C.; Mokrushina, L.; Minceva, M. Experimental
https://pubs.acs.org/10.1021/acs.cgd.1c01477 Investigation and Modeling of Cocrystal Formation in L-Menthol/
Thymol Eutectic System. Cryst. Growth Des. 2021, 21 (11), 6083−
Author Contributions 6091.
Conceptualization: A.A. Formal analysis: A.A. and C.J. Writing, (17) Hamilton, V.; Andrusenko, I.; Potticary, J.; Hall, C.; Stenner, R.;
original draft: A.A. and C.J. Writing, review and editing: L.M. Mugnaioli, E.; Lanza, A. E.; Gemmi, M.; Hall, S. R. Racemic
and M.M. Supervision: M.M. Conglomerate Formation via Crystallization of Metaxalone from
Notes Volatile Deep Eutectic Solvents. Cryst. Growth Des. 2020, 20 (7),
4731−4739.
The authors declare no competing financial interest. (18) Potticary, J.; Hall, C.; Hamilton, V.; McCabe, J. F.; Hall, S. R.

■ REFERENCES
(1) Abbott, A. P.; Capper, G.; Davies, D. L.; Rasheed, R. K.;
Crystallization from Volatile Deep Eutectic Solvents. Cryst. Growth Des.
2020, 20 (5), 2877−2884.
(19) Hall, C. L.; Potticary, J.; Hamilton, V.; Gaisford, S.; Buanz, A.;
Tambyrajah, V. Novel solvent properties of choline chloride/urea Hall, S. R. Metastable crystalline phase formation in deep eutectic
mixtures. Chem. Commun. (Camb) 2003, 70−71. systems revealed by simultaneous synchrotron XRD and DSC. Chem.
(2) Smith, E. L.; Abbott, A. P.; Ryder, K. S. Deep Eutectic Solvents Commun. (Camb) 2020, 56 (73), 10726−10729.
(DESs) and Their Applications. Chem. Rev. 2014, 114 (21), 11060− (20) Martins, M. A. R.; Pinho, S. P.; Coutinho, J. A. P. Insights into the
11082. Nature of Eutectic and Deep Eutectic Mixtures. J. Solution Chem. 2019,
(3) Zhang, Q.; De Oliveira Vigier, K.; Royer, S.; Jerome, F. Deep 48 (7), 962−982.
eutectic solvents: syntheses, properties and applications. Chem. Soc. Rev. (21) Che Zain, M. S.; Yeoh, J. X.; Lee, S. Y.; Shaari, K.
2012, 41 (21), 7108−7146. Physicochemical Properties of Choline Chloride-Based Natural Deep
(4) Alhadid, A.; Mokrushina, L.; Minceva, M. Modeling of Solid− Eutectic Solvents (NaDES) and Their Applicability for Extracting Oil
Liquid Equilibria in Deep Eutectic Solvents: A Parameter Study. Palm Flavonoids. Sustainability 2021, 13 (23), 12981.
Molecules 2019, 24 (12), 2334. (22) Abbasi, N. M.; Farooq, M. Q.; Anderson, J. L. Investigating the
(5) Silva, L. P.; Martins, M. A. R.; Conceiçaõ , J. H. F.; Pinho, S. P.; Variation in Solvation Interactions of Choline Chloride-Based Deep
Coutinho, J. A. P. Eutectic Mixtures Based on Polyalcohols as Eutectic Solvents Formed Using Different Hydrogen Bond Donors.
Sustainable Solvents: Screening and Characterization. ACS Sustain.
ACS Sustain. Chem. Eng. 2021, 9 (35), 11970−11980.
Chem. Eng. 2020, 8 (40), 15317−15326.
(23) Golgoun, S.; Mokhtarpour, M.; Shekaari, H. Solubility
(6) Martins, M. A. R.; Silva, L. P.; Schaeffer, N.; Abranches, D. O.;
Enhancement of Betamethasone, Meloxicam and Piroxicam by Use
Maximo, G. J.; Pinho, S. P.; Coutinho, J. A. P. Greener Terpene−
of Choline-Based Deep Eutectic Solvents. Pharm. Sci. 2021, 27 (1),
Terpene Eutectic Mixtures as Hydrophobic Solvents. ACS Sustain.
Chem. Eng. 2019, 7 (20), 17414−17423. 86−101.
(7) Silva, L. P.; Araújo, C. F.; Abranches, D. O.; Melle-Franco, M.; (24) Shanley, P.; Collin, R. L. The crystal structure of the high
Martins, M. A. R.; Nolasco, M. M.; Ribeiro-Claro, P. J. A.; Pinho, S. P.; temperature form of choline chloride. Acta Crystallogr. 1961, 14 (1),
Coutinho, J. A. P. What a difference a methyl group makes − probing 79−80.
choline−urea molecular interactions through urea structure modifica- (25) Crespo, E. A.; Silva, L. P.; Lloret, J. O.; Carvalho, P. J.; Vega, L. F.;
tion. Phys. Chem. Chem. Phys. 2019, 21 (33), 18278−18289. Llovell, F.; Coutinho, J. A. P. A methodology to parameterize SAFT-
(8) Silva, L. P.; Fernandez, L.; Conceiçaõ , J. H. F.; Martins, M. A. R.; type equations of state for solid precursors of deep eutectic solvents: the
Sosa, A.; Ortega, J.; Pinho, S. P.; Coutinho, J. A. P. Design and example of cholinium chloride. Phys. Chem. Chem. Phys. 2019, 21 (27),
Characterization of Sugar-Based Deep Eutectic Solvents Using 15046−15061.
Conductor-like Screening Model for Real Solvents. ACS Sustain. (26) Zahn, S. Deep eutectic solvents: similia similibus solvuntur? Phys.
Chem. Eng. 2018, 6 (8), 10724−10734. Chem. Chem. Phys. 2017, 19 (5), 4041−4047.
(9) Crespo, E. A.; Silva, L. P.; Martins, M. A. R.; Bülow, M.; Ferreira, (27) García, G.; Atilhan, M.; Aparicio, S. An approach for the
O.; Sadowski, G.; Held, C.; Pinho, S. P.; Coutinho, J. A. P. The Role of rationalization of melting temperature for deep eutectic solvents from
Polyfunctionality in the Formation of [Ch]Cl-Carboxylic Acid-Based DFT. Chem. Phys. Lett. 2015, 634, 151−155.
Deep Eutectic Solvents. Ind. Eng. Chem. Res. 2018, 57 (32), 11195− (28) Abranches, D. O.; Silva, L. P.; Martins, M. A. R.; Coutinho, J. A.
11209. P. Differences on the impact of water on the deep eutectic solvents
(10) Abranches, D. O.; Silva, L. P.; Martins, M. A. R.; Pinho, S. P.; betaine/urea and choline/urea. J. Chem. Phys. 2021, 155 (3), 034501.
Coutinho, J. A. P. Understanding the Formation of Deep Eutectic (29) Morrison, H. G.; Sun, C. C.; Neervannan, S. Characterization of
Solvents: Betaine as a Universal Hydrogen Bond Acceptor. thermal behavior of deep eutectic solvents and their potential as drug
ChemSusChem 2020, 13 (18), 4916−4921. solubilization vehicles. Int. J. Pharm. 2009, 378 (1−2), 136−139.

1941 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
Crystal Growth & Design pubs.acs.org/crystal Article

(30) van den Bruinhorst, A.; Kollau, L. J. B. M.; Kroon, M. C.;

Meuldijk, J.; Tuinier, R.; Esteves, A. C. C. A centrifuge method to
determine the solid−liquid phase behavior of eutectic mixtures. J.
Chem. Phys. 2018, 149 (22), 224505.
(31) Kitaigorodskii, A. I. Mixed crystals; Springer-Verlag: Berlin, 1984.
(32) Prausnitz, J. M.; Lichtenthaler, R. N.; Azevedo, E. G. d. Molecular
Thermodynamics of Fluid-Phase Equilibria; Prentice Hall PTR: Upper
Saddle River, NJ, 1999.
(33) Lohmann, J.; Joh, R.; Gmehling, J. Estimation of Enthalpies of
Fusion, Melting Temperatures, Enthalpies of Transition, and
Transition Temperatures of Pure Compounds from Experimental
Binary Solid−Liquid Equilibrium Data of Eutectic Systems. J. Chem.
Eng. Data 1997, 42 (6), 1176−1180.
(34) Lange, L.; Sadowski, G. Polymorphs, Hydrates, Cocrystals, and
Cocrystal Hydrates: Thermodynamic Modeling of Theophylline
Systems. Cryst. Growth Des. 2016, 16 (8), 4439−4449.
(35) Abbott, A. P.; Ahmed, E. I.; Prasad, K.; Qader, I. B.; Ryder, K. S.
Liquid pharmaceuticals formulation by eutectic formation. Fluid Phase
Equilib. 2017, 448, 2−8.
(36) Verevkin, S. P.; Kozlova, S. A. Di-hydroxybenzenes: Catechol,
resorcinol, and hydroquinone: Enthalpies of phase transitions revisited.
Thermochim. Acta 2008, 471 (1), 33−42.
(37) Petrouleas, V.; Lemmon, R. M. Calorimetric studies of choline
chloride, bromide, and iodide. J. Chem. Phys. 1978, 69 (3), 1315.
(38) Fernandez, L.; Silva, L. P.; Martins, M. A. R.; Ferreira, O.; Ortega,
J.; Pinho, S. P.; Coutinho, J. A. P. Indirect assessment of the fusion
properties of choline chloride from solid-liquid equilibria data. Fluid
Phase Equilib. 2017, 448, 9−14.
(39) Vilas-Boas, S. M.; Abranches, D. O.; Crespo, E. A.; Ferreira, O.;
Coutinho, J. A. P.; Pinho, S. P. Experimental solubility and density
studies on aqueous solutions of quaternary ammonium halides, and
thermodynamic modelling for melting enthalpy estimations. J. Mol. Liq.
2020, 300, 112281.
(40) Brown, R. J. C.; Brown, R. F. C. Melting Point and Molecular
Symmetry. J. Chem. Educ. 2000, 77 (6), 724.

1942 https://doi.org/10.1021/acs.cgd.1c01477
Cryst. Growth Des. 2022, 22, 1933−1942
 Supporting Information
Cocrystal Formation in Choline Chloride Deep Eutectic Solvents
Ahmad Alhadid, a Christian Jandl, b Liudmila Mokrushina c and Mirjana Minceva* a
a
Biothermodynamics, TUM School of Life Sciences, Technical University of Munich (TUM), Germany
b
Catalysis Research Center, Department Chemie, Technical University of Munich (TUM), Germany
c
Separation Science & Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU),
Germany
*
Corresponding author e-mail: [email protected]

Single-crystal X-ray diffraction

Crystals suitable for SC-XRD were prepared by dissolving the cocrystal powder in hot acetonitrile
in a 1:1 mass ratio. The solution was stored at 253 K until suitable crystals appeared. The average
time to obtain suitable crystals was around one week.
Data were collected on a Bruker D8 Venture single crystal x-ray diffractometer equipped with a CMOS
detector (Bruker Photon-100), a TXS rotating anode with MoK radiation (λ = 0.71073 Å) and a Helios
optic using the APEX3 software package.1 Measurements were performed on single crystals coated with
perfluorinated ether. The crystals were fixed on top of a Kapton micro sampler and frozen under a stream
of cold nitrogen. A matrix scan was used to determine the initial lattice parameters. Reflections were
corrected for Lorentz and polarisation effects, scan speed, and background using SAINT. 2 Absorption
correction, including odd and even ordered spherical harmonics, was performed using SADABS. 2 Space
group assignments were based upon systematic absences, E statistics, and successful refinement of the
structures. The structures were solved using SHELXT with the aid of successive difference Fourier maps
and were refined against all data using SHELXL in conjunction with SHELXLE. 3,4,5 Hydrogen atoms
except on heteroatoms were calculated in ideal positions as follows: Methyl H atoms were refined as part
of rigid rotating groups, with a C H distance of 0.98 Å and Uiso(H) = 1.5·Ueq(C). Non-methyl H atoms were
placed in calculated positions and refined using a riding model, with methylene, aromatic, and other C H
distances of 0.99 Å, 0.95 Å and 1.00 Å, respectively, and Uiso(H) = 1.2·Ueq(C). Non-hydrogen atoms were
refined with anisotropic displacement parameters. Full-matrix least-squares refinements were carried out
by minimizing w(Fo2 - Fc2)2 with the SHELXL weighting scheme.4 Neutral atom scattering factors for all
atoms and anomalous dispersion corrections for the non-hydrogen atoms were taken from International
Tables for Crystallography.6 Images of the crystal structures were generated with PLATON and Mercury.7,8
CCDC 2125060-2125062 contains the supplementary crystallographic data for this paper. These data are
provided free of charge by The Cambridge Crystallographic Data Centre.

1
Table S1 Solid liquid equilibria data of choline chloride/catechol eutectic system.
catechol mole fraction Tliq / K Te / K

0.89 362.7 ± 0.1 314.6

0.85 353.6 ± 0.9 315.6

0.83 345.0 ± 1.1 315.5

0.81 331.8 ± 0.1 315.3

0.79 328.4 ± 0.1 315.7

0.75 315.1

0.71 322.5 ± 0.1 314.7

0.69 323.5 ± 0.1 314.3

0.68 324.3 ± 0.3 314.7

0.65 323.8 ± 0.3 317.9

0.60 319.5

0.55 325.5 ± 0.3 319.1

0.47 352.2 ± 0.3

0.44 322.3

0.40 336.5 ± 0.3 323.8

0.38 342.3 ± 0.9 323.0

0.34 350.5 ± 0.1* 320.5

0.30 351.3 ± 0.1* 324.9

*solid-solid transition

2
Table S2 Solid liquid equilibria data of choline chloride/hydroquinone eutectic system.
Hydroquinone mole fraction Tliq / K Te / K

0.89 434.2 ± 0.9 317.8

0.81 418.9 ± 0.2 318.5

0.75 386.8 ± 1.6 318.6

0.71 358.3 ± 0.1 318.9

0.70 315.3

0.67 341.2 ± 0.9 318.7

0.63 317.4

0.60 315.3

0.57 324.3 ± 0.1 313.3

0.46 324.5

0.45 325.7

0.41 326.2

0.40 327.1

0.37 340.9 ± 0.2 326.5

0.32 351.3 ± 0.1* 326.7

0.30 352.1 ± 0.2* 326.4

*solid-solid transition

3
Figure S1: Asymmetric unit of the crystal structure of the choline chloride/catechol cocrystal with a ratio of 1:2. Ellipsoids are
displayed at the 50% probability level.

Figure S2: Packing diagram of the choline chloride/catechol cocrystal with a ratio of 1:2 viewed along the b-axis.

4
Table S3: Structure and refinement details for the choline chloride/catechol cocrystal with a ratio of 1:2.
Deposition number CCDC 2125062
Chemical formula C17H26ClNO5
Crystal description colourless fragment
Formula weight 359.84
Temperature 100 K
Wavelength 0.71073 Å
Crystal size 0.39 × 0.34 × 0.17 mm
Crystal system monoclinic
Space group P21/c
Unit cell dimensions a = 10.7901(14) Å a = 90°
b = 9.9623(13) Å b = 99.194(4)°
c = 17.303(3) Å c = 90°
Volume 1836.1(5) Å3
Z 4
Density (calculated) 1.302 g cm 3
Absorption coefficient 0.23 mm 1
F(000) 768
a ge 2.4-27.5°
Absorption correction multi-scan (SADABS)
Tmin, Tmax 0.681, 0.746
Reflections measured 58129
Independent reflections 4196
Ref ec i I > 2 (I) 3949
Rint 0.030
Refinement method full matrix least squares on F2
Data, restraints, parameters 4196, 0, 240
R1 [I > 2 (I)] 0.029
wR2 (all data) 0.073
Goodness of fit 1.05
Weighting scheme W = 1/[ 2(FO2) + (0.0312P)2 + 0.8905P] where
P = (FO2 + 2FC2)/3
Largest difference peak and hole 0.31 and -0.25 eÅ-3

5
Figure S3: Asymmetric unit of the crystal structure of the choline chloride/catechol cocrystal with a ratio of 1:1. Ellipsoids are
displayed at the 50% probability level.

Figure S4: Packing diagram of the choline chloride/catechol cocrystal with a ratio of 1:1 viewed along the a-axis.

6
Table S4: Structure and refinement details for the choline chloride/catechol cocrystal with a ratio of 1:1.
Deposition number CCDC 2125061
Chemical formula C11H20ClNO3
Crystal description colourless fragment
Formula weight 249.73
Temperature 123 K
Wavelength 0.71073 Å
Crystal size 0.40 × 0.19 × 0.04 mm
Crystal system orthorhombic
Space group P212121
Unit cell dimensions a = 6.9117(4) Å a = 90°
b = 10.7352(7)Å b = 90°
c = 17.5585(10) Å c = 90°
Volume 1302.81(14) Å3
Z 4
Density (calculated) 1.273 g cm 3
Absorption coefficient 0.29 mm 1
F(000) 536
a ge 3.2-26.0°
Absorption correction multi-scan (SADABS)
Tmin, Tmax 0.678, 0.745
Reflections measured 22895
Independent reflections 2569
Ref ec i I > 2 (I) 2387
Rint 0.044
Refinement method full matrix least squares on F2
Data, restraints, parameters 2569, 0, 160
Absolute structure parameter (Flack, Parsons)9 0.01(2)
R1 [I > 2 (I)] 0.025
wR2 (all data) 0.061
Goodness of fit 1.07
Weighting scheme W = 1/[ 2(FO2) + (0.0301P)2 + 0.3073P] where
P = (FO2 + 2FC2)/3
Largest difference peak and hole 0.16 and -0.21 eÅ-3

7
Figure S5: Asymmetric unit (with completed fragments) of the crystal structure of the choline chloride/hydroquinone cocrystal
with a ratio of 1:1. Ellipsoids are displayed at the 50% probability level. Symmetry code to create equivalent position: a) 1 - x, -y,
1 z; b) 2 - x, 2 - y, 1 - z.

Figure S6: Packing diagram of the choline chloride/hydroquinone cocrystal with a ratio of 1:1 viewed along the b-axis.

8
Table S5: Structure and refinement details for the choline chloride/hydroquinone cocrystal with a ratio of 1:1.
Deposition number CCDC 2125060
Chemical formula C11H20ClNO3
Crystal description colourless fragment
Formula weight 249.73
Temperature 123 K
Wavelength 0.71073 Å
Crystal size 0.56 × 0.42 × 0.24 mm
Crystal system monoclinic
Space group P21/c
Unit cell dimensions a = 10.7084(10) Å a = 90°
b = 6.9493(6) Å b = 94.833(3)°
c = 17.8691(16) Å c = 90°
3
Volume 1325.0(2) Å
Z 4
Density (calculated) 1.252 g cm 3
Absorption coefficient 0.28 mm 1
F(000) 536
a ge 2.3-26.4°
Absorption correction multi-scan (SADABS)
Tmin, Tmax 0.701, 0.745
Reflections measured 21818
Independent reflections 2678
Ref ec i I > 2 (I) 2534
Rint 0.025
Refinement method full matrix least squares on F2
Data, restraints, parameters 2534, 0, 160
R1 [I > 2 (I)] 0.027
wR2 (all data) 0.072
Goodness of fit 1.07
Weighting scheme W = 1/[ 2(FO2) + (0.0335P)2 + 0.5679P] where
P = (FO2 + 2FC2)/3
Largest difference peak and hole 0.26 and -0.19 eÅ-3

9
Rietveld refinements

Rietveld refinements were performed using TOPAS.10 Refinements of the single phases with fixed atomic
sites were used to determine the cell parameters at the measurement temperature of 253 K. Then,
quantitative phase analysis was performed with fixed unit cell parameters of these phases.

Table S6: Resulting unit cell parameters at 253 K of the relevant single phases from Rietveld refinement.
a [Å] b [Å] c [Å] [°] [°] [°] Rwp
choline chloride (ChCl) 5.8835(4) 11.1249(8) 11.5895(8) 90 90 90 10.919
catechol (cat) 10.0107(9) 5.5331(5) 10.9268(8) 90 118.840(6) 90 11.358
hydroquinone (hyd) 38.462(2) 38.462(2) 5.6424(5) 90 90 120 12.117
ChCl/cat 1:1 cocrystal 6.9641(5) 10.8219(8) 17.669(1) 90 90 90 9.299
ChCl/cat 1:2 cocrystal 10.8770(8) 10.0382(7) 17.394(1) 90 99.202(7) 90 8.123
ChCl/hyd 1:1 cocrystal 10.7953(8) 6.9764(6) 17.989(2) 90 94.991(5) 90 11.260

Table S7: Results of quantitative phase analysis via Rietveld refinement in the cholin chloride/catechol system.
% weight
Rwp
ChCl ChCl/cat 1:1 cocr. ChCl/cat 1:2 cocr. cat
choline chloride (ChCl) 99.5(6) 0.0(2) 0.5(4) 0.0(3) 11.472
catechol (cat) 0.0(2) 0.4(4) 0.3(4) 99.3(6) 11.698
ChCl/cat 1:1 mixture 0.5(2) 94.6(4) 4.3(3) 0.6(2) 8.103
ChCl/cat 1:2 mixture 0.4(2) 0.1(2) 99.0(4) 0.5(2) 9.126
ChCl/cat 2:3 mixture 0.2(1) 42.2(3) 57.3(3) 0.2(1) 6.849
ChCl/cat 1:3 mixture 0.5(2) 0.2(2) 69.8(4) 29.6(4) 8.747
ChCl/cat 2:1 mixture 45.7(3) 51.3(4) 2.6(2) 0.4(1) 7.018

Table S8: Results of quantitative phase analysis via Rietveld refinement in the cholin chloride/hydroquinone system.
% weight
Rwp
ChCl ChCl/hyd 1:1 cocr. hyd
choline chloride 97.5(7) 0.8(4) 1.7(6) 11.408
(ChCl)
hydroquinone (hyd) 0.5(2) 0.0(3) 99.5(4) 12.057
ChCl/hyd 1:1 0.8(3) 97.1(5) 2.1(5) 11.400
ChCl/hyd 1:2 0.0(1) 77.8(4) 22.1(3) 8.369
ChCl/hyd 1:3 0.0(1) 53.0(3) 47.0(3) 6.675
ChCl/hyd 2:1 30.0(3) 69.9(4) 0.1(3) 8.027
xhyd = 0.85 0.2(1) 67.4(3) 32.5(3) 6.372

10
 (A) (B)

endo
1:1 cocrystal endo

1:1 cocrystal

1:2 cocrystal

catechol hydroquinone

300 340 380 420 300 360 420 480

T/K T/K

Figure S7. Differential scanning calorimetry curve of cocrystals and pure components of (A) choline chloride (ChCl)/catechol
and (B) ChCl/hydroquinone eutectic systems.

choline chloride solid-solid transition

T ≈ 78 °C

Figure S8: Variable-temperature X-ray powder diffraction study of choline chloride.

11
References
1. APEX suite of crystallographic software, APEX 3, Version 2019-1.0, Bruker AXS Inc., Madison,
Wisconsin, USA, 2019.
2. SAINT, Version 8.40A and SADABS, Version 2016/2, Bruker AXS Inc., Madison, Wisconsin, USA,
2016/2019.
3. Sheldrick, G. M. Acta Crystallogr. Sect. A 2015, 71, 3 8.
4. Sheldrick, G. M. Acta Crystallogr. Sect. C 2015, 71, 3 8.
5. Hübschle, C. B.; Sheldrick, G. M.; Dittrich, B. J. Appl. Cryst. 2011, 44, 1281 1284.
6. International Tables for Crystallography, Vol. C (Ed.: A. J. Wilson), Kluwer Academic Publishers,
Dordrecht, The Netherlands, 1992.
7. Spek, A. L. Acta Crystallogr. Sect. D 2009, 65, 148 155.
8. C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington, P. McCabe, E. Pidcock, L. Rodriguez-Monge,
R. Taylor, J. van de Streek, P. A. Wood, J. Appl. Cryst. 2008, 41, 466 470.
9. Flack, H. D. Acta Crystallogr. Sect A 1983, 39, 876 881; Parsons, S.; Flack, H. D.; Wagner, T. Acta
Crystallogr. Sect B 2013, 69, 249 259.
10. TOPAS, Version 6, Bruker AXS Inc., Madison, Wisconsin, USA, 2016.

12
3.6 Paper VI

3.6 Paper VI
Cocrystal Formation in L-Menthol/Phenol Eutectic System: Experimental Study and
Thermodynamic Modeling

A. Alhadid, C. Jandl, L. Mokrushina, and M. Minceva, Cryst. Growth Des. 2022, 22, 6, 3973–
3980.

Author contribution: The thesis author conceptualized the paper’s idea, performed the
investigations and formal analysis (excluding powder and SC-XRD), interpreted the results, and
wrote the manuscript.

Summary: Paper VI demonstrates the potential of predictive and correlative thermodynamic

models to describe SLE in eutectic systems with cocrystal formation. The SLE data of the L-
menthol/phenol eutectic system were obtained by a combination of DSC and XRD
measurements. Two cocrystals were identified in the L-menthol/phenol eutectic system. The
crystal structure of the two cocrystals was solved by the SC-XRD technique and from powder
XRD data. The obtained crystal structures showed that hydrogen bonds are the main feature in
the cocrystal forming helical chains between the OH groups of the molecules.

This study is the first to use COSMO-RS to model SLE in eutectic systems with cocrystal
formation. COSMO-RS calculations were performed using TZVP and TZVPD_FINE
parameterization. It was found that COSMO-RS calculations at the TZVPD_FINE level provides
better predictions for the liquidus lines of pure constituents and cocrystals compared to the
TZVP level. The two eutectic points predicted by the NRTL and COSMO-RS models were in
good agreement with the experimentally determined eutectic temperatures.

Paper VI emphasizes the importance of experimental investigation of cocrystal formation for SLE
modeling, whether using correlative or predictive thermodynamics models. The proper
evaluation of the performance of thermodynamic models in predicting the eutectic temperature
of eutectic mixtures with cocrystal formation should rely on SLE data obtained by a combination
of DSC and XRD analyses.

Despite the significant negative deviation from ideal behavior, the measured eutectic
temperature of the system was not far from the eutectic temperature calculated using the ideal
solution model. This was attributed to the formation of the two cocrystals. If no cocrystals are
formed, the eutectic temperature determined by NRTL and COSMO-RS was significantly lower

143
3.6 Paper VI

than the experimentally observed one. Therefore, cocrystal formation could revoke the influence
of nonideality on the depression of the melting temperature at the eutectic point.

144
3.6 Paper VI

Reprinted with permission from Alhadid et al., Cryst. Growth Des., 2022, 22 (6), 3973-3980. DOI:
10.1021/acs.cgd.2c00362. Copyright 2022 American Chemical Society.

Article full text: http://pubs.acs.org/articlesonrequest/AOR-3FTJ2S7WIHUD5KXIKNZ2

145
 pubs.acs.org/crystal Article

Cocrystal Formation in L‑Menthol/Phenol Eutectic System:

Experimental Study and Thermodynamic Modeling
Ahmad Alhadid, Christian Jandl, Liudmila Mokrushina, and Mirjana Minceva*
Cite This: https://doi.org/10.1021/acs.cgd.2c00362 Read Online

ACCESS Metrics & More Article Recommendations *

sı Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via TU MUENCHEN on May 24, 2022 at 07:40:37 (UTC).

ABSTRACT: Deep eutectic solvents (DESs) are eutectic mixtures containing a hydrogen-bond acceptor and donor, forming a
mixture with a significantly lower melting temperature than those of its pure constituents. DESs containing cyclohexyl and phenolic
alcohols draw particular attention due to the observed large depression in the melting temperature of the mixture. The present study
investigates in detail the solid−liquid equilibria (SLE) in the L-menthol/phenol eutectic system. Differential scanning calorimetry
and powder X-ray diffraction (XRD) analysis were employed to obtain the phase diagram. Two cocrystals were identified with 1:2
and 2:1 ratios. The crystal structures were determined by single-crystal and powder XRD techniques. The SLE data were correlated
using the nonrandom two-liquid (NRTL) model and the conductor-like screening model for realistic solvation (COSMO-RS). The
two eutectic points determined by the NRTL and COSMO-RS models are xNRTL e,1 = 0.69, TNRTL
e,1 = 273.1 K, xNRTL
e,2 = 0.47 TNRTL
e,2 =
COSMO COSMO COSMO COSMO
261.3 K, xe,1 = 0.70, Te,1 = 272.8 K, and xe,2 = 0.48 Te,2 = 261.5 K, which are in good agreement with the
experimental eutectic temperatures.

1. INTRODUCTION essential to select the constituents, determine the DES eutectic

Deep eutectic solvents (DESs) are eutectic mixtures formed by point, or select the mixture composition for a certain
mixing two or more inexpensive and widely available application.17 So far, the SLE data available for nonionic
substances to produce a novel solvent with unique properties. DESs show that they follow ideal solution behavior,18−21 with
an exception for DESs formed by mixing cyclohexyl and
Soon after the term DES was introduced, the interest and
phenolic alcohols, such as L-menthol/thymol and L-menthol/
research on using them as a new class of green and tunable
carvacrol.22,23 Due to substantial negative deviation from
solvents in various applications increased significantly.1
ideality, DESs formed by mixing cyclohexyl and phenolic
Majority of studied DESs contain ionic constituents, which
alcohols are suspected of glass formation.14 Moreover, some
possess some advantages of ionic liquids (ILs), such as
phenolic alcohols are known coformers that have been shown
negligible volatility and nonflammability.1−6 However, ionic
to form cocrystals with several substances.22,24−27 Glass and
DESs possess a significant drawback of being hygroscopic.7
cocrystal formations complicate measuring the SLE phase
The hygroscopic nature of ionic DESs affects their
diagram of eutectic mixtures containing cyclohexyl and
physicochemical properties and limits their applicability. On
phenolic alcohols, which limited the number of SLE studies
the other hand, nonionic DESsusually referred to as type V
DEScould be formed by mixing simple organic substan-
ces.8−13 Nonionic DESs have several advantages over ionic Received: March 26, 2022
DESs, such as lower cost, melting temperature, and Revised: April 27, 2022
viscosity.14−16
The knowledge of solid−liquid equilibria (SLE) in DESs
provides information about the melting temperature of the
mixture at any molar ratio between constituents, which is
© XXXX The Authors. Published by
American Chemical Society https://doi.org/10.1021/acs.cgd.2c00362
A Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design pubs.acs.org/crystal Article

Table 1. Melting Properties of Pure Components Used to Model the Solid−Liquid Equilibria of the L-Menthol/Phenol
Eutectic System
Δcp = a + bT/J mol−1 K−1
a b
substance Δhm/kJ mol−1 Tm/K liquid solid liquid solid
L-menthol 13.7423 314.623 −195.2050 −68.050 1.79550 1.09250
phenol 12.36 ± 0.07a,b 313.9 ± 0.10a 100.86c 10.038c 0.3203c 0.392c
a
Measured in this work using DSC. bUncertainties are the standard deviation of three samples. cCalculated using the experimental data from
Andon et al.49 (see Figure S1 in the SI).

available in the literature for this distinctive type of eutectic mixtures were heated gently to 315 K under continuous stirring inside
mixtures.22,23,28 closed glass vials to obtain clear homogeneous liquids.
Modeling SLE allows for the determination of the eutectic Because the eutectic system is suspected of glass formation,14,23 the
point of the system, which is difficult to determine samples were prepared by quenching and annealing as previously
described.22 In short, liquid mixtures were quenched at 193 K for
experimentally. Prerequisites for modeling SLE are activity several hours and then annealed at 253 K for several days to aid the
coefficients, which are a measure of intermolecular interactions crystallization. Generally, the annealing time to observe the complete
between molecules in the liquid phase. The activity coefficients solidification of the sample is between 1 and 5 days. The obtained
can be calculated using thermodynamic models. Correlative solid was ground to a fine powder using a mortar and pestle within a
models, such as regular solution theory or the nonrandom two- cold room at 253 K and then directly introduced into DSC crucibles
liquid (NRTL) model, have been successfully used in the in triplicate. The DSC crucibles were immediately hermetically sealed.
literature to model SLE in nonideal eutectic systems.29−33 The DSC (NETZSCH DSC 200 F3, Germany) was calibrated
However, predictive models are preferable to prevent the need prior to measurements using adamantane, bismuth, cesium chloride,
for experimental data to model SLE. The conductor-like indium, tin, and zinc with a mass fraction purity of 99.999%. The DSC
chamber was precooled to 253 K before introducing the samples. A
screening model for realistic solvation (COSMO-RS) was cooling step to 193 K with a cooling rate of 10 K min−1 was
proposed to calculate the activity coefficients and thus predict performed first. Then, a heating run to 320 K was performed with a
the eutectic point and the phase diagram of ionic and nonionic heating rate of 5 K min−1. Measurements were performed under a
DESs.20,34,35 SLE calculations using COSMO-RS have been nitrogen atmosphere. The melting temperatures of pure components
exclusively performed assuming that the constituents crystallize and the solidus temperatures were determined as the onset
in pure form, i.e., the simple eutectic behavior.20,34,35 However, temperatures and the liquidus temperatures as the peak maximum
cocrystal formation has been identified in some DESs.22,25 temperatures. The average standard uncertainty of the measured
Cocrystal, hydrate, and solvate formations in binary eutectic liquidus temperatures was 0.4 K.
mixtures have been successfully modeled using the perturbed- 2.2. XRD. Powder XRD analysis was performed to identify the
crystallized solid phases. The crystallized samples with different
chain statistical associating fluid theory (PC-SAFT).36−39 To compositions were finely ground within a cold room at 253 K and
the best of our knowledge, no study has evaluated the directly filled into 0.70 or 1.5 mm diameter glass capillaries
performance of COSMO-RS in predicting the phase diagram (Hilgenberg, Germany). Powder XRD patterns were recorded in
of binary eutectic systems with cocrystal formation, as Debye−Scherrer geometry on a Stadi P diffractometer (Stoe & Cie,
COSMO-RS has been only used in several studies to screen Germany) equipped with a Mo fine-focus sealed tube, a curved Ge
and select suitable coformers apriori.40−46 Loschen and monochromator selecting Kα1 radiation (λ = 0.7093 Å), and a
Klamt47 used COSMO-RS to predict the SLE phase diagram Mythen2 R 1K detector (Dectris, Switzerland). All data were
of several ternary systems containing an active ingredient, collected with a step size of 2θ = 0.015° per data point. The
coformer, and solvent. However, due to the strong association measurements were performed at 253 K using an 800 series
cryostream (Oxford, UK) unless stated otherwise.
in the liquid phase, an additional binary interaction parameter The crystal structures of the observed cocrystals were solved by
was introduced and fitted to experimental data, which single-crystal (SC)-XRD analysis and from powder XRD data. An SC
withdrew COSMO-RS predictive ability. with sufficient size and quality for SC-XRD analysis was obtained by
The present study investigates the SLE in the binary eutectic mixing 1 g of the liquid 2:1 mixture with 1 mL of acetonitrile. The
system L-menthol/phenol. The SLE data were measured using liquid mixture was stored at 253 K until suitable crystals appeared.
differential scanning calorimetry (DSC). Powder X-ray Details on sample preparation and structure solution are provided in
diffraction (XRD) analysis was performed on samples covering the Supporting Information (SI).
the whole composition range of the mixture to evaluate the 2.3. Thermodynamic Modeling. The phase diagram and the
formation of cocrystals. The melting properties of the eutectic points of the eutectic mixture can be obtained by modeling
experimental SLE data. For the regions where pure L-menthol and
cocrystals were measured by DSC. The SLE data were phenol crystallize, the liquidus lines were calculated as follows48
modeled considering the cocrystal formation using the
correlative (NRTL) and predictive (COSMO-RS) thermody- Δhm , i
∫T
T 1 T
namic models to obtain the solid−liquid phase diagram of the ln xiLγi L = − 1− − ΔcP(T )dT
RT Tm , i RT m
eutectic system.
ΔcP(T )
∫T
1 T
+ dT
2. MATERIALS AND METHODS R m T (1)
2.1. Experimental SLE. L-Menthol (water content 107.5 ± 9.0 where xLi
and γLi
are the mole fraction and activity coefficient of
ppm, purity ≥99%, Sigma-Aldrich) was mixed with phenol (water component i in the liquid solution, respectively; Δhm,i and Tm,i are the
content 82.4 ± 5.0 ppm, purity ≥99%, Alfa Aeser) in different ratios. melting enthalpy and melting temperature of pure component i,
The pure constituents’ water content was analyzed using Karl− respectively; R and T are the universal gas constant and the
Fischer Coulometer (Hanna Instrument) in triplicate. The eutectic temperature, respectively; and ΔcP,i is the difference between the

B https://doi.org/10.1021/acs.cgd.2c00362
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 1. (A) Solid−liquid equilibria data for L-menthol/phenol eutectic system. (B) Powder X-ray diffraction (XRD) patterns of pure L-menthol
(men), phenol (ph), and mixtures of different ratios marked with circles in (A). XRD measurements were performed at 253 K. The characteristic
peaks are marked as follows: phenol (red dotted line), 1:2 cocrystal (green dotted lines), 2:1 cocrystal (blue dotted lines).

constant pressure heat capacities of component i in the liquid and n 1/2

(Tiexp − Tical)2
solid states. Three approaches were used for considering the last two F(T ) = ∑
terms on the right-hand side of eq 1, namely, neglecting ΔcP, i=1
n (9)
assuming constant ΔcP value calculated at the melting temperature, or
temperature-dependent ΔcP. Experimental data on the heat capacity where Tcal
i is the liquidus temperature of pure constituents calculated
of pure L-menthol and phenol in the solid and liquid states were taken using the NRTL model and n is the number of experimental data
from the literature.49,50 The pure components’ melting properties points. The value of the nonrandomness parameter (α) in eq 7 was
used to model the SLE data using eq 1 are shown in Table 1. set to 0.3.
To model the cocrystal liquidus lines, the formation of a cocrystal To predict the phase diagram of the eutectic system using
was considered via the following chemical reaction22,38 COSMO-RS, the activity coefficients of components were calculated
ϑAA(L) + ϑBB(L) ↔ A ϑA B ϑB(S) from the molecular structure.51−53 The molecular geometry and the
(2) screening surface charge density were obtained by DFT calculations
where ϑA and ϑB are the stoichiometric coefficients of components A using the BP86 functional and the def-TZVP and def-TZVPD basis
and B in the cocrystal, respectively. The equilibrium constant (Ka) of sets for TZVP and TZVPD_FINE levels, respectively, employing
the chemical reaction was defined as follows Turbomole54 (TURBOMOLE GmbH, version 6.6). The activity
coefficients of components were calculated by COSMOtherm55
Ka = ∏ aiϑi = (xALγAL)ϑA (xBLγBL)ϑB (Dassault Systèmes, version 19) with the BP_TZVP_19.ctd and
i (3) BP_TZVPD_FINE_19.ctd parameter sets for TZVP and
TZVPD_FINE levels. The obtained activity coefficients were used
The equilibrium constant of the chemical reaction at different
to calculate the liquidus lines of pure constituents (eq 1) and
temperatures was calculated using the Gibbs−Helmholtz equation
cocrystals (eqs 3−5).
Δhref 1 1
ln Ka = ln Karef + −
R T ref T (4) 3. RESULTS
ref
where T is a reference temperature, Kref
is the equilibrium constant
a 3.1. Experimental SLE. DSC analysis was used to measure
at Tref, and Δhref is the melting enthalpy of the cocrystal at Tref. The the solidus and liquidus temperatures of the L-menthol/phenol
equilibrium constant at Tref was calculated as follows binary eutectic system. The obtained experimental data and
the DSC curves are found in Table S1 and Figure S2 in the SI,
Karef = (xAref γAref )ϑA (xBref γBref )ϑB (5) respectively. Figure 1A shows the experimental liquidus and
where xref and γref are the mole fraction and the activity coefficient of solidus temperatures (points). As seen in Figure 1A, the system
the component at the cocrystal stoichiometry, respectively. shows three different solidus temperatures. The solidus
The activity coefficients of components γLi for modeling the temperatures in the ranges xphenol = 0.67−1.0, 0.33−0.67,
liquidus lines of pure constituents (eq 1) and cocrystals (eqs 3−5) and 0−0.33 are ∼273, 262, and 268 K, respectively.
were calculated using the NRTL and COSMO-RS models. The The three different solidus temperatures observed for the
NRTL equation was used to calculate the activity coefficients as
follows48 system hint at cocrystal formation. Figure 1B shows the
powder XRD patterns of pure components (points a and f)
2
Gji τijGij and samples with 1:3 (point b), 1:2 (point c), 1:1 (point d),
ln γi L = xj2 τji + and 2:1 (point e) for L-menthol:phenol ratios (marked with
xi + xjGji (xj + xiGij)2
(6) circles in Figure 1A). Dotted lines connect the matching peaks
observed in the XRD patterns of samples. Starting from
Gij = exp(− ατij) Gji = exp(− ατji) (7) samples in the phenol-rich composition range, it is clear from
Figure 1B that the 1:3 ratio sample XRD pattern is the sum of
gij − gjj gji − gii
τij = τji = the XRD patterns of pure phenol (red dotted lines) and the 1:2
RT RT (8) ratio sample (green dotted lines), indicating that the solid
The NRTL binary interaction parameters (gij−gjj) and (gji−gii) were phase of the 1:3 ratio sample consists of both pure phenol and
fitted to the experimental data on pure components’ liquidus lines the solid phase of the 1:2 ratio sample. The 1:2 ratio sample
(Texp
i ) by minimizing the following objective function XRD pattern does not contain the XRD pattern peaks of pure
C https://doi.org/10.1021/acs.cgd.2c00362
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design pubs.acs.org/crystal Article

phenol or L-menthol. Thus, it is postulated that a cocrystal with

a ratio of 1:2 is formed.
For the 1:2 cocrystal, it was not possible to obtain SCs
suitable for SC-XRD analysis. As an alternative, powder XRD
data were used for the structure determination. During a long-
term measurement of the sample divided into eight identical
runs of approximately 90 min duration, distinct changes over
the first three runs in the powder XRD pattern were observed,
indicating a phase transition (see Figures S3 and S4 in the SI).
Thus, it can be concluded that the primary cocrystal formed at
1:2 and 1:3 ratios (Figure 1B) is metastable and transforms
into the more stable structure that is observed in further runs.
Figure 2. Overlay of the asymmetric units of the structures of the 2:1
This could also justify the difficulty in obtaining an SC suitable
L-menthol:phenol cocrystal from single-crystal (red) and powder
for SC-XRD analysis. However, powder XRD data of the stable (blue) X-ray diffraction data. Unit cells are shifted for better
polymorph were sufficient for structure determination and comparison.
confirmed the 1:2 ratio of L-menthol and phenol (see Table 2).

Table 2. Unit Cell Data of the Crystal Structures of the 2:1

and 1:2 L-Menthol:Phenol Cocrystals Solved from Powder
Diffraction Dataa
L-menthol/phenol 2:1 L-menthol/phenol 1:2
formula 2(C10H20O), C6H6O C10H20O, 2(C6H6O)
molecular weight 406.63 344.48
(g mol−1)
Bravais lattice monoclinic P orthorhombic P
space group P21 P212121
a (Å) 6.099 20.3936
b (Å) 21.7903 16.8032
c (Å) 10.4267 6.2151
α (°) 90 90
β (°) 100.9126 90
γ (°) 90 90
Figure 3. Packing diagram of the structure of the 1:2 L-
volume (Å3) 1360.649 2129.77
menthol:phenol cocrystal viewed along the c axis. The structure was
Z 2 4 obtained from powder X-ray diffraction data.
a
See Table S3 in the SI for more details.

Based on powder XRD analysis, the solidus temperatures

observed at 1:3 (point b, ∼272 K) and 1:2 (point c, ∼273 K)
ratio samples represent the eutectic temperature in the phenol-
rich region (xphenol > 0.67) and the 1:2 cocrystal melting
temperature, respectively.
Examining the 1:1 ratio sample XRD pattern, it can be seen
that the solid phase of the 1:1 ratio sample consists of the 1:2
cocrystal and the solid phase of the 2:1 ratio sample. The 2:1
ratio sample XRD pattern does not contain the peaks of pure
phenol, L-menthol, or the 1:2 cocrystal. This hints at the
formation of a second cocrystal at a 2:1 ratio. The structure of
the 2:1 L-menthol:phenol cocrystal was obtained from powder
XRD data and was confirmed by single-crystal XRD (see Table
2 and Table S2 in the SI). Figure 2 shows the overlay of the
asymmetric units of the structures of the 2:1 cocrystal obtained
by SC-XRD and powder XRD data. As seen, a good agreement
between the two obtained crystal structures is observed, which
proves the reliability of our structure solutions from powder
XRD data. The slight difference in unit cell dimensions is due
to the temperature at which the measurements were
performed, i.e., SC-XRD at 100 K and powder XRD at 253 Figure 4. Packing diagram of the structure of the 2:1 L-
K. Moreover, the small variation in the molecular geometries of menthol:phenol cocrystal viewed along the a-axis. The structure was
constituents observed in Figure 2 is attributed to the input obtained from single-crystal X-ray diffraction data.
geometries of the molecular fragments used for the powder
structure solution and rigid-body refinement. Figures 3 and 4 cocrystals. As seen, hydrogen bonds are the dominant feature
show the packing diagrams of the structures of the 1:2 and 2:1 forming helical chains between the OH groups of the
D https://doi.org/10.1021/acs.cgd.2c00362
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design pubs.acs.org/crystal Article

molecules along the c- and a-axes in the 1:2 and 2:1 cocrystals, component, and temperature-dependent ΔcP. Table 3 shows
respectively. The hydrophobic parts of the molecules are then the obtained NRTL binary interaction parameters and RMSD
facing outwards, and these strains of molecules are arranged in
parallel along the respective axis. Table 3. Obtained Binary Interaction Parameters and
As seen in Figure 1A, at the 2:1 ratio (point e), along with RMSD from Modeling the Pure Components Liquidus
the observed solidus temperature, a liquidus temperature is Lines Using Different Approaches Considering the Heat
observed, which indicates that the 2:1 cocrystal melts Capacity Term
incongruently (peritectic transition). To confirm the peritectic
(g12−g22)/kJ mol−1 (g21−g11)/kJ mol−1 RMSD/K
transition, powder XRD analysis was performed on the 2:1
ratio sample at two temperatures, lower and higher than the Δcp = 0 −3.1628 −2.3630 2.3
solidus temperature observed in the L-menthol-rich region constant Δcp −2.9026 −1.4834 1.9
(268 K). Figure 5 shows the powder XRD pattern of pure L- temperature- −2.8372 −1.5449 2.0
dependent Δcp

between calculated and experimental liquidus temperatures of

pure components. As seen in Table 3, the smallest RMSD is
observed when considering constant or temperature-depend-
ent ΔcP. Moreover, the obtained binary interaction parameters
are quite similar in both cases. On the other hand, the largest
RMSD is observed when neglecting the ΔcP term. Therefore,
the binary interaction parameters obtained assuming constant
Δcp valuesbest fitwere used further to obtain the phase
diagram in the middle composition range.
The phase diagram in the middle composition range
Figure 5. Powder X-ray diffraction pattern for pure L-menthol at 253 representing the crystallization of the two cocrystalswas
K and 2:1 cocrystal at 253 and 273 K.
modeled using eqs 3−5. The melting properties of the
cocrystals were measured using DSC, and the results are
menthol at 253 K, 2:1 cocrystal at 253 K, and 2:1 cocrystal at shown in Table 4. The activity coefficients were calculated
273 K; blue dotted lines extended the characteristic peaks of
the L-menthol XRD pattern. As seen in Figure 5, it is clear that Table 4. Melting Properties of the 1:2 and 2:1 for L-
no peaks corresponding to pure L-menthol are observed in the Menthol:Phenol Cocrystals Used in Modeling their
2:1 cocrystal XRD pattern at 253 K. In contrast, the Liquidus Lines
characteristic peaks of pure L-menthol can be observed in
the 2:1 ratio sample XRD pattern at 273 K. Thus, the 2:1 cocrystal Δhm/kJ mol−1 Tm/K
cocrystal decomposes at the solidus temperature observed in 1:2 cocrystal 27.48 ± 0.24 273.3 ± 0.1
the L-menthol-rich region (∼268 K) to a liquid phase and solid 2:1 cocrystal 20.35 ± 0.25 269.4 ± 0.1
L-menthol. Accordingly, the solidus temperatures observed in
Figure 1A at 1:1 (point d, ∼262 K) and 2:1 (point e, ∼268 K)
ratio samples correspond to the eutectic temperature in the using the NRTL model with binary interaction parameters
middle composition range and the decomposition temperature obtained when assuming constant Δcp (Table 3). Figure 6
of the 2:1 cocrystal. shows the phase diagram of the L-menthol/phenol eutectic
In conclusion, the DSC and powder XRD analyses system modeled considering cocrystal formation. As seen in
performed on samples of the L-menthol/phenol eutectic Figure 6A, the NRTL model provides an adequate description
system covering the complete composition range showed of the liquidus temperatures (circles) of pure components and
that the system is not of the simple eutectic type. Two cocrystals. Moreover, the eutectic points determined by the
cocrystals with melting temperatures of 273.3 and 269.4 K NRTL model (xNRTL e,1 = 0.69 TNRTL
e,1 = 273.1 K and xNRTL
e,2 = 0.47
NRTL
were observed at 1:2 and 2:1 ratios. The system has two Te,2 = 261.3 K) are in good agreement with the measured
eutectic temperatures: the first is around 272 K in the phenol- solidus temperatures (triangles) in the phenol-rich and middle
rich composition range, and the second is around 263 K in the composition regions (Texp exp
e,1 = 272.3 K and Te,2 = 262.5 K).
middle composition range. Figure 6B shows the COSMO-RS predictions at the TZVP
3.2. Thermodynamic Modeling. The experimental (blue lines) and TZVPD_FINE (red lines) levels for the SLE
liquidus data and the information about the formed cocrystals phase diagram of the L-menthol/phenol eutectic system
acquired by DSC and XRD analyses were used to model the considering the formation of the two cocrystals. COSMO-RS
phase diagram of the L-menthol/phenol system. The satisfactorily predicts the course of the phenol liquidus line at
correlative thermodynamic model NRTL was shown to be the TZVPD_FINE (red lines) level, while at the TZVP level
successful in describing the SLE in eutectic mixtures with (blue lines), COSMO-RS underestimates the nonideality of
cocrystal formation. 22,25 In this work, the predictive phenol in the liquid phase. In contrast, the COSMO-RS model
thermodynamic model COSMO-RS performance was eval- at the TZVP level well describes the nonideality of the L-
uated in describing cocrystal formation in eutectic mixtures. menthol in the liquid phase, while at the TZVPD_FINE level,
The NRTL parameters were obtained by fitting the COSMO-RS underestimates the nonideality of L-menthol.
experimental liquidus data of pure components and assuming Accordingly, COSMO-RS better describes the peritectic point
three approaches regarding the heat capacity terms in eq 1, of the 2:1 cocrystal at the TZVP level compared to the
namely, neglecting the heat capacity terms, assuming constant TZVPD_FINE level. However, the COSMO-RS predictions
ΔcP calculated at the melting temperature of the pure for SLE in the middle composition range corresponding to the
E https://doi.org/10.1021/acs.cgd.2c00362
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design pubs.acs.org/crystal Article

Figure 6. Solid−liquid phase diagram of L-menthol/phenol eutectic system modeled using the (A) NRTL and (B) COSMO-RS models.

crystallization of the cocrystals are significantly improved at the a reasonable temperature range. Therefore, it is evident that for
TZVPD_FINE level compared to the TZVP level. The two a proper determination of the eutectic point of eutectic
eutectic points predicted by COSMO-RS at the systems, cocrystal formation should be investigated before-
TZVPD_FINE level are xCOSMO e,1 = 0.70, TCOSMO
e,1 = 272.8 K, hand. Moreover, the unreliable COSMO-RS predictions for
COSMO
and xe,2 COSMO
= 0.48 Te,2 = 261.5 K. Bearing in mind that the the eutectic point of some DESs observed in previous studies
model is purely predictive, COSMO-RS predictions at the found in the literature might result from ruling out cocrystal
TZVPD_FINE level for the two eutectic points are quite formation.34,35 Furthermore, as seen in Figure 7, the observed
reasonable, comparable with the correlative thermodynamic eutectic temperature of the system in the middle composition
model predictions, and agree well with those observed region (Te,2exp = 262.5 K) is not far from the eutectic
experimentally. temperature calculated by the ideal solution model (Teideal =
Nevertheless, a good estimation for the eutectic points of the 270.8 K). The small difference between the ideal and observed
system using the NRTL and COSMO-RS models was obtained eutectic temperatures of the system despite the observed
when considering the formation of the two cocrystals. Figure 7 significant negative deviation from ideal behavior is attributed
to the formation of the two cocrystals.
shows COSMO-RS predictions neglecting cocrystal formation.
As seen in Figure 7, the estimated eutectic temperature by the
4. CONCLUSIONS
COSMO-RS model at the TZVPD_FINE level is xCOSMO e =
0.55 and TCOSMO
e = 228.0 K, while COSMO-RS calculations at This study reports the phase diagram of the L-menthol/phenol
the TZVP level fail to estimate the eutectic temperature within eutectic system. The SLE data were measured using DSC, and
the solid phases were characterized by powder XRD. The
formation of two cocrystals with the ratios 1:2 and 2:1 for L-
menthol:phenol was observed. The crystal structure of the two
cocrystals was solved by SC-XRD analysis and the structure
solution from powder XRD data. It was found that the 1:2
cocrystal melts congruently while the 2:1 cocrystal melts
incongruently.
The phase diagram was obtained by modeling the measured
SLE data using the NRTL and COSMO-RS models. It was
found that assuming simple eutectic behavior, the estimated
eutectic temperatures by both models are significantly lower
than the experimentally observed eutectic temperatures.
However, both models can satisfactorily describe the SLE in
the L-menthol/phenol eutectic system when considering the
formation of the two cocrystals.
This work shows that a reasonable estimation for the
eutectic point of DES systems could be obtained by predictive
thermodynamic models. However, possible cocrystal formation
should always be experimentally investigated beforehand and
accounted for while modeling the SLE.

Figure 7. Solid−liquid phase diagram of L-menthol/phenol eutectic

■
*
ASSOCIATED CONTENT
sı Supporting Information

system modeled using the COSMO-RS models assuming the simple The Supporting Information is available free of charge at
eutectic type. https://pubs.acs.org/doi/10.1021/acs.cgd.2c00362.
F https://doi.org/10.1021/acs.cgd.2c00362
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design pubs.acs.org/crystal Article

Solid−liquid equilibria data for the L-menthol/phenol (5) Abbott, A. P.; Capper, G.; Davies, D. L.; Munro, H. L.; Rasheed,
eutectic system; heat capacity of pure phenol in the solid R. K.; Tambyrajah, V. Preparation of novel, moisture-stable, Lewis-
and liquid states; DSC curves for the L-menthol/phenol acidic ionic liquids containing quaternary ammonium salts with
eutectic system at different phenol mole fractions; functional side chains. Chem. Commun. 2001, 2010−2011.
(6) Abbott, A. P.; Barron, J. C.; Ryder, K. S.; Wilson, D. Eutectic-
variation in the powder XRD of the 1:2 L-menthol:phe- Based Ionic Liquids with Metal-Containing Anions and Cations.
nol cocrystal with measuring time; powder XRD pattern Chem. - Eur. J. 2007, 13, 6495−6501.
of the 1:2 L-menthol:phenol cocrystal at different runs; (7) Abranches, D. O.; Silva, L. P.; Martins, M. A. R.; Coutinho, J. A.
SC-XRD details; SC-XRD data for the 2:1 L - P. Differences on the impact of water on the deep eutectic solvents
menthol:phenol cocrystal; procedure for the structure betaine/urea and choline/urea. J. Chem. Phys. 2021, 155, No. 034501.
solution from powder data; structure and refinement (8) van Osch, D. J. G. P.; Dietz, C. H. J. T.; Warrag, S. E. E.; Kroon,
details of the crystal structures of the 2:1 and 1:2 L- M. C. The Curious Case of Hydrophobic Deep Eutectic Solvents: A
menthol:phenol cocrystals solved from powder diffrac- Story on the Discovery, Design, and Applications. ACS Sustainable
tion data (PDF) Chem. Eng. 2020, 8, 10591−10612.
(9) van Osch, D. J. G. P.; Dietz, C. H. J. T.; van Spronsen, J.; Kroon,
Accession Codes M. C.; Gallucci, F.; van Sint Annaland, M.; Tuinier, R. A Search for
CCDC 2153851−2153853 contain the supplementary crys- Natural Hydrophobic Deep Eutectic Solvents Based on Natural
tallographic data for this paper. These data can be obtained Components. ACS Sustainable Chem. Eng. 2019, 7, 2933−2942.
free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by (10) Lee, J.; Jung, D.; Park, K. Hydrophobic deep eutectic solvents
emailing [email protected], or by contacting The for the extraction of organic and inorganic analytes from aqueous
environments. TrAC, Trends Anal. Chem. 2019, 118, 853−868.
Cambridge Crystallographic Data Centre, 12 Union Road,
(11) Dwamena, A. K. Recent advances in hydrophobic deep eutectic
Cambridge CB2 1EZ, UK; fax: +44 1223 336033. solvents for extraction. Separations 2019, 6, No. 9.

■ AUTHOR INFORMATION
Corresponding Author
(12) Florindo, C.; Branco, L. C.; Marrucho, I. M. Quest for Green-
Solvent Design: From Hydrophilic to Hydrophobic (Deep) Eutectic
Solvents. ChemSusChem 2019, 12, 1549−1559.
Mirjana Minceva − Biothermodynamics, TUM School of Life (13) Abranches, D. O.; Coutinho, J. A. P. Type V deep eutectic
Sciences, Technical University of Munich (TUM), Freising solvents: Design and applications. Curr. Opin. Green Sustainable Chem.
85354, Germany; Email: [email protected] 2022, 35, No. 100612.
(14) Alhadid, A.; Mokrushina, L.; Minceva, M. Influence of the
Authors Molecular Structure of Constituents and Liquid Phase Non-Ideality
Ahmad Alhadid − Biothermodynamics, TUM School of Life on the Viscosity of Deep Eutectic Solvents. Molecules 2021, 26,
Sciences, Technical University of Munich (TUM), Freising No. 4208.
(15) Gajardo-Parra, N. F.; Cotroneo-Figueroa, V. P.; Aravena, P.;
85354, Germany; orcid.org/0000-0003-1443-1517
Vesovic, V.; Canales, R. I. Viscosity of Choline Chloride-Based Deep
Christian Jandl − Catalysis Research Center, Department Eutectic Solvents: Experiments and Modeling. J. Chem. Eng. Data
Chemie, Technical University of Munich (TUM), Garching 2020, 65, 5581−5592.
85748, Germany (16) Lemaoui, T.; Darwish, A. S.; Attoui, A.; Abu Hatab, F.;
Liudmila Mokrushina − Separation Science & Technology, Hammoudi, N. E. H.; Benguerba, Y.; Vega, L. F.; Alnashef, I. M.
Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Predicting the density and viscosity of hydrophobic eutectic solvents:
Erlangen 91058, Germany towards the development of sustainable solvents. Green Chem. 2020,
22, 8511−8530.
Complete contact information is available at: (17) Alhadid, A.; Mokrushina, L.; Minceva, M. Modeling of Solid−
https://pubs.acs.org/10.1021/acs.cgd.2c00362 Liquid Equilibria in Deep Eutectic Solvents: A Parameter Study.
Molecules 2019, 24, 2334.
Author Contributions (18) Alhadid, A.; Mokrushina, L.; Minceva, M. Design of Deep
A.A. contributed to conceptualization. Investigation and formal Eutectic Systems: A Simple Approach for Preselecting Eutectic
analysis were conducted by A.A. and C.J. A.A. performed Mixture Constituents. Molecules 2020, 25, 1077.
writingoriginal draft. Writingreview & editing was (19) Martins, M. A. R.; Crespo, E. A.; Pontes, P. V. A.; Silva, L. P.;
performed by C.J., L.M., and M.M. Supervision was conducted Bülow, M.; Maximo, G. J.; Batista, E. A. C.; Held, C.; Pinho, S. P.;
by M.M. Coutinho, J. A. P. Tunable Hydrophobic Eutectic Solvents Based on
Terpenes and Monocarboxylic Acids. ACS Sustainable Chem. Eng.
Notes 2018, 6, 8836−8846.
The authors declare no competing financial interest. (20) Abdallah, M. M.; Müller, S.; González de Castilla, A.; Gurikov,

■ REFERENCES
(1) Smith, E. L.; Abbott, A. P.; Ryder, K. S. Deep Eutectic Solvents
P.; Matias, A. A.; Bronze, M. D.; Fernández, N. Physicochemical
Characterization and Simulation of the Solid−Liquid Equilibrium
Phase Diagram of Terpene-Based Eutectic Solvent Systems. Molecules
(DESs) and Their Applications. Chem. Rev. 2014, 114, 11060−11082. 2021, 26, No. 1801.
(2) Huang, J.; Guo, X.; Xu, T.; Fan, L.; Zhou, X.; Wu, S. Ionic deep (21) Corvis, Y.; Négrier, P.; Lazerges, M.; Massip, S.; Léger, J.-M.;
eutectic solvents for the extraction and separation of natural products. Espeau, P. Lidocaine/l-Menthol Binary System: Cocrystallization
J. Chromatogr. A 2019, 1598, 1−19. versus Solid-State Immiscibility. J. Phys. Chem. B 2010, 114, 5420−
(3) González de Castilla, A.; Bittner, J. P.; Mü l ler, S.; 5426.
Jakobtorweihen, S.; Smirnova, I. Thermodynamic and Transport (22) Alhadid, A.; Jandl, C.; Mokrushina, L.; Minceva, M.
Properties Modeling of Deep Eutectic Solvents: A Review on gE- Experimental Investigation and Modeling of Cocrystal Formation in
Models, Equations of State, and Molecular Dynamics. J. Chem. Eng. L-Menthol/Thymol Eutectic System. Cryst. Growth Des. 2021, 21,
Data 2020, 65, 943−967. 6083−6091.
(4) Martins, M. A. R.; Pinho, S. P.; Coutinho, J. A. P. Insights into (23) Alhadid, A.; Mokrushina, L.; Minceva, M. Formation of glassy
the Nature of Eutectic and Deep Eutectic Mixtures. J. Solution Chem. phases and polymorphism in deep eutectic solvents. J. Mol. Liq. 2020,
2019, 48, 962−982. 314, No. 113667.

G https://doi.org/10.1021/acs.cgd.2c00362
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design pubs.acs.org/crystal Article

(24) Jamróz, M. E.; Palczewska-Tulińska, M.; Wyrzykowska- (44) Cysewski, P. Efficacy of bi-component cocrystals and simple
Stankiewicz, D.; Szafrański, A. M.; Polaczek, J.; Dobrowolski, J. C.; binary eutectics screening using heat of mixing estimated under super
Jamróz, M. H.; Mazurek, A. P. The urea−phenol(s) systems1. Fluid cooled conditions. J. Mol. Graphics Modell. 2016, 68, 23−28.
Phase Equilib. 1998, 152, 307−326. (45) Przybyłek, M.; Ziółkowska, D.; Mroczyńska, K.; Cysewski, P.
(25) Alhadid, A.; Jandl, C.; Mokrushina, L.; Minceva, M. Cocrystal Applicability of Phenolic Acids as Effective Enhancers of Cocrystal
Formation in Choline Chloride Deep Eutectic Solvents. Cryst. Growth Solubility of Methylxanthines. Cryst. Growth Des. 2017, 17, 2186−
Des. 2022, 22, 1933−1942. 2193.
(26) A. I., Kitaigorodskii, Mixed Crystals, Springer-Verlag: Berlin, (46) Cysewski, P. In silico screening of dicarboxylic acids for
1984. cocrystallization with phenylpiperazine derivatives based on both
(27) Hall, C. L.; Potticary, J.; Hamilton, V.; Gaisford, S.; Buanz, A.; cocrystallization propensity and solubility advantage. J. Mol. Model.
Hall, S. R. Metastable crystalline phase formation in deep eutectic 2017, 23, No. 136.
systems revealed by simultaneous synchrotron XRD and DSC. Chem. (47) Loschen, C.; Klamt, A. Cocrystal Ternary Phase Diagrams from
Commun. 2020, 56, 10726−10729. Density Functional Theory and Solvation Thermodynamics. Cryst.
(28) Abranches, D. O.; Martins, M. A. R.; Silva, L. P.; Schaeffer, N.; Growth Des. 2018, 18, 5600−5608.
Pinho, S. P.; Coutinho, J. A. P. Phenolic hydrogen bond donors in the (48) Prausnitz, J. M.; Lichtenthaler, R. N.; Azevedo, E. G. d.
formation of non-ionic deep eutectic solvents: the quest for type V Molecular Thermodynamics of Fluid-Phase Equilibria; Prentice Hall
DES. Chem. Commun. 2019, 55, 10253−10256. PTR: Upper Saddle River, NJ, 1999.
(29) van den Bruinhorst, A.; Kollau, L. J. B. M.; Vis, M.; Hendrix, M. (49) Andon, R. J. L.; Counsell, J. F.; Herington, E. F. G.; Martin, J. F.
M. R. M.; Meuldijk, J.; Tuinier, R.; Esteves, A. C. C. From a eutectic Thermodynamic properties of organic oxygen compounds. Part 7.
Calorimetric study of phenol from 12 to 330 °K. Trans. Faraday Soc.
mixture to a deep eutectic system via anion selection: Glutaric acid +
1963, 59, 830−835.
tetraethylammonium halides. J. Chem. Phys. 2021, 155, No. 014502.
(50) Corvis, Y.; Espeau, P. Interpretation of the global heat of
(30) Kollau, L. J. B. M.; Tuinier, R.; Verhaak, J.; den Doelder, J.;
melting in eutectic binary systems. Thermochim. Acta 2018, 664, 91−
Filot, I. A. W.; Vis, M. Design of Nonideal Eutectic Mixtures Based on
99.
Correlations with Molecular Properties. J. Phys. Chem. B 2020, 124, (51) Eckert, F.; Klamt, A. Fast solvent screening via quantum
5209−5219. chemistry: COSMO-RS approach. AlChE J. 2002, 48, 369−385.
(31) Kollau, L. J. B. M.; Vis, M.; van den Bruinhorst, A.; de With, G.; (52) Klamt, A.; Jonas, V.; Bürger, T.; Lohrenz, J. C. W. Refinement
Tuinier, R. Activity modelling of the solid−liquid equilibrium of deep and Parametrization of COSMO-RS. J. Phys. Chem. A 1998, 102,
eutectic solvents. Pure Appl. Chem. 2019, 91, 1341−1349. 5074−5085.
(32) Kollau, L. J. B. M.; Vis, M.; van den Bruinhorst, A.; Esteves, A. (53) Klamt, A. Conductor-like Screening Model for Real Solvents: A
C. C.; Tuinier, R. Quantification of the liquid window of deep eutectic New Approach to the Quantitative Calculation of Solvation
solvents. Chem. Commun. 2018, 54, 13351−13354. Phenomena. J. Phys. Chem. A 1995, 99, 2224−2235.
(33) Alkhatib, I. I. I.; Bahamon, D.; Llovell, F.; Abu-Zahra, M. R. M.; (54) TURBOMOLE V6.6 2014, a development of University of
Vega, L. F. Perspectives and guidelines on thermodynamic modelling Karlsruhe and Forschungszentrum Karlsruhe GmbH, 1989-2007;
of deep eutectic solvents. J. Mol. Liq. 2020, 298, No. 112183. TURBOMOLE GmbH, https://www.turbomole.org.
(34) Abranches, D. O.; Larriba, M.; Silva, L. P.; Melle-Franco, M.; (55) BIOVIA COSMOtherm, Release 2019; Dassault Systèmes,
Palomar, J. F.; Pinho, S. P.; Coutinho, J. A. P. Using COSMO-RS to http://www.3ds.com.
design choline chloride pharmaceutical eutectic solvents. Fluid Phase
Equilib. 2019, 497, 71−78.
(35) Song, Z.; Wang, J.; Sundmacher, K. Evaluation of COSMO-RS
for solid−liquid equilibria prediction of binary eutectic solvent
systems. Green Energy Environ. 2021, 6, 371−379.
(36) Lange, L.; Sadowski, G. Polymorphs, Hydrates, Cocrystals, and
Cocrystal Hydrates: Thermodynamic Modeling of Theophylline
Systems. Cryst. Growth Des. 2016, 16, 4439−4449.
(37) Lange, L.; Sadowski, G. Thermodynamic Modeling for Efficient
Cocrystal Formation. Cryst. Growth Des. 2015, 15, 4406−4416.
(38) Tumakaka, F.; Prikhodko, I. V.; Sadowski, G. Modeling of
solid−liquid equilibria for systems with solid-complex phase
formation. Fluid Phase Equilib. 2007, 260, 98−104.
(39) Lange, L.; Heisel, S.; Sadowski, G. Predicting the Solubility of
Pharmaceutical Cocrystals in Solvent/Anti-Solvent Mixtures. Mole-
cules 2016, 21, No. 593.
(40) Wu, D.; Zhang, B.; Yao, Q.; Hou, B.; Zhou, L.; Xie, C.; Gong,
J.; Hao, H.; Chen, W. Evaluation on Cocrystal Screening Methods
and Synthesis of Multicomponent Crystals: A Case Study. Cryst.
Growth Des. 2021, 21, 4531−4546.
(41) Dash, S. G.; Thakur, T. S. Computational Screening of
Multicomponent Solid Forms of 2-Aryl-Propionate Class of NSAID,
Zaltoprofen, and Their Experimental Validation. Cryst. Growth Des.
2021, 21, 449−461.
(42) Wu, D.; Li, J.; Xiao, Y.; Ji, X.; Li, C.; Zhang, B.; Hou, B.; Zhou,
L.; Xie, C.; Gong, J.; Chen, W. New Salts and Cocrystals of
Pymetrozine with Improvements on Solubility and Humidity
Stability: Experimental and Theoretical Study. Cryst. Growth Des.
2021, 21, 2371−2388.
(43) Loschen, C.; Klamt, A. Solubility prediction, solvate and
cocrystal screening as tools for rational crystal engineering. J. Pharm.
Pharmacol. 2015, 67, 803−811.

H https://doi.org/10.1021/acs.cgd.2c00362
Cryst. Growth Des. XXXX, XXX, XXX−XXX
 Supplementary Information

Cocrystal Formation in L-Menthol/Phenol Eutectic System:

Experimental Study and Thermodynamic Modeling
Ahmad Alhadid 1, Christian Jandl 2, Liudmila Mokrushina 3, and Mirjana Minceva* 1
1
Biothermodynamics, TUM School of Life Sciences, Technical University of Munich (TUM), Germany
2
Catalysis Research Center, Department Chemie, Technical University of Munich (TUM), Germany
3
Separation Science & Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU),
Germany
*
Corresponding author e-mail: [email protected]

Table S1. Solid liquid equilibria data for L-menthol/phenol eutectic system.
Phenol mole fraction Tliq / K Te / K Solid phase *

0.90 306.1 ± 0.4 272.1 Phenol

0.85 300.3 ± 0.5 271.5 Phenol

0.80 291.7 ± 1.1 272.2 Phenol

0.75 284.9 ± 1.0 272.3 Phenol

0.70 272.3 Phenol + 1:2 cocrystal

0.67 273.3 1:2 cocrystal

0.65 273.7 ± 0.2 262.6 1:2 cocrystal

0.60 272.8 ± 0.9 262.8 1:2 cocrystal

0.55 270.7 ± 0.3 262.9 1:2 cocrystal

0.51 263.0 1:2 cocrystal + 2:1 cocrystal

0.46 262.6 1:2 cocrystal + 2:1 cocrystal

0.39 266.7 ± 0.9 261.7 2:1 cocrystal

0.35 269.7 ± 0.1 262.3 2:1 cocrystal

0.33 280.0 ± 0.2 269.4 L-menthol

0.30 284.0 ± 0.3 269.4 L-menthol

1
 0.23 294.1 ± 0.1 268.8 L-menthol

0.20 299.2 ± 0.4 268.4 L-menthol

0.16 302.9 ± 0.3 267.9 L-menthol

0.12 307.2 ± 0.4 267.3 L-menthol

*
the crystallized solid phase at the liquidus temperature. Standard uncertainties are u (T) = 0.4 K and ur(x) = 0.005

(A) (B)
160 210

y = 0.3203x + 100.86
y = 0.392x + 10.038 208

cp / J mol−1 K−1
cp / J mol−1 K−1

R² = 0.9571
120 R² = 0.99

206
80
204

40
202

0 200
0 100 200 300 400 315 320 325 330 335 340
T/K T/K
Figure S1. Heat capacity of pure phenol in the (A) solid and (B) liquid states. Data were taken from Andon et al. 1

2
 endo

xphenol = 0.90

xphe nol = 0.80

xphe nol = 0.70

xphe nol = 0.67

xphenol = 0.60

xphe nol = 0.51

xphe nol = 0.39

xphe nol = 0.33

xphe nol = 0.30

xphenol = 0.20

xphe nol = 0.12

250 260 270 280 290 300 310 320

T/K
Figure S2. Differential scanning calorimetry curves for L-menthol/phenol eutectic system at different phenol mole
fractions.

3
Figure S3. The variation in the powder X-ray diffraction pattern of the 1:2 L-menthol:phenol cocrystal with
measuring time.

4
 Figure S4. Powder X-ray diffractions pattern of the 1:2 L-menthol:phenol cocrystal at different runs (range).
The changing peaks were magnified for clarity.

SC-XRD
Data were collected on a single crystal X-ray diffractometer equipped with a CMOS detector
(Bruker Photon-100), a TXS rotating anode with MoK radiation (λ = 0.71073 Å) and a Helios
optic using the APEX3 software package.2 The measurements were performed on single crystals
coated with perfluorinated ether. The crystals were fixed on top of a kapton micro sampler and
frozen under a stream of cold nitrogen. A matrix scan was used to determine the initial lattice
parameters. Reflections were corrected for Lorentz and polarisation effects, scan speed, and
background using SAINT.3 Absorption correction, including odd and even ordered spherical
harmonics was performed using SADABS.3 Space group assignment was based upon systematic
absences, E statistics, and successful refinement of the structure. The structures were solved using
SHELXT with the aid of successive difference Fourier maps, and were refined against all data
using SHELXL in conjunction with SHELXLE.4 6 Hydrogen atoms (except on heteroatoms) were
calculated in ideal positions as follows: Methyl hydrogen atoms were refined as part of rigid
rotating groups, with a C H distance of 0.98 Å and Uiso(H) = 1.5·Ueq(C). Non-methyl H atoms were
placed in calculated positions and refined using a riding model with methylene, aromatic, and other
C H distances of 0.99 Å, 0.95 Å, and 1.00 Å, respectively, and Uiso(H) = 1.2·Ueq(C). Non-

5
hydrogen atoms were refined with anisotropic displacement parameters. Full-matrix least-squares
efi eme e e ca ied b mi imi i g (Fo2 Fc2)2 with the SHELXL weighting scheme.4
Neutral atom scattering factors for all atoms and anomalous dispersion corrections for the non-
hydrogen atoms were taken from International Tables for Crystallography.7 Images of the crystal
structure were generated with Mercury and PLATON.8 10
Deposition Number 2153853 contains
the supplementary crystallographic data for this paper. These data are provided free of charge by
the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access
Structures service www.ccdc.cam.ac.uk/structures.

2:1 Cocrystal L-Menthol/Phenol (CCDC 2153853)

Figure S5: Asymmetric unit of the structure of the 2:1 L-menthol:phenol cocrystal from single-crystal XRD data
with ellipsoids at the 50% probability level.

Table S2. Structure and refinement details for the 2:1 L-menthol:phenol cocrystal.
Deposition number CCDC 2153853
Chemical formula C26H46O3
Crystal description colourless block
Formula weight 406.63
Temperature 100 K
Wavelength 0.71073 Å
Crystal size 0.24 × 0.27 × 0.45 mm
Crystal system monoclinic
Space group P 21
Unit cell dimensions a = 6.0315(4) Å α = 90°
b = 21.5189(14) Å β = 100.277(2)°

6
 c = 10.3083(7) Å = 90°
Volume 1316.46(15) Å3
Z 2
Density (calculated) 1.026 g cm 3
Absorption coefficient 0.065 mm 1
F(000) 452
a ge 2.76-26.73°
Absorption correction multi-scan (SADABS)
Tmin, Tmax 0.985, 0.972
Reflections measured 38249
Independent reflections 5579
Reflec i I > 2 (I) 5440
Rint 0.02
Refinement method full matrix least squares on F2
Data, restraints, parameters 5579, 1, 280
R1 [I > 2 (I)] 0.029
wR2 (all data) 0.076
Goodness of fit 1.02
Weighting scheme W = 1/[σ2(FO2) + (0.0475P)2 + 0.1413P] where
P = (FO2 + 2FC2)/3
Largest difference peak and hole 0.17 and -0.14 eÅ-3

P-XRD
Samples were finely ground within a cold room at 253 K and directly filled into 0.70 or 1.5 mm
diameter glass capillaries (Hilgenberg, Germany). Powder XRD patterns were recorded in
Debye-Scherrer geometry on a Stadi P diffractometer (Stoe & Cie, Germany) equipped with a Mo
fine-f c ealed be, a c ed Ge m ch ma elec i g K 1 adia i ( = 0.7093 Å), and a
Mythen2 R 1K detector (Dectris, Switzerland) using the WinXPOW software package.12 All data
were collected as overlapping step scans so that every data point is measured 4 times during one
run with a data point resolution of 0.015° 2θ. Multiple consecutive runs were performed and the
step duration as well as the number of runs (3-8) were adjusted to the diffraction intensity of the
sample. The measurements were performed at 253 K using an 800 series cryostream (Oxford, UK),
unless stated otherwise. Pattern fitting and indexing for cell determination was performed using
WinXPOW. Pawley fitting, space group determination and cell refinement were performed using
DASH.13 Structure solution via simulated annealing was performed with DASH using molecular
geometries from the Cambridege Structural Database (L-menthol from ZEGDIA [CCDC 832350],
phenol from PHENOL03 [CCDC 1232404]) and a fixed torsion angle for the isopropyl-group on
menthol.12 16
Initial rigid body Rietveld refinement was performed with DASH and the hydroxy-
groups were oriented in the direction of likely hydrogen bonds.13 Final rigid body Rietveld

7
refinement was performed using Jana2006 with global isotropic displacement parameters for C
and O atoms and riding isotropic displacement parameters for H atoms Uiso(H) = 1.2·Uiso(C/O).17
Deposition Numbers 2153851-2153852 contain the supplementary crystallographic data for this
paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre
and Fachinformationszentrum Karlsruhe Access Structures service
www.ccdc.cam.ac.uk/structures.

8
Table S3. Structure and refinement details of the crystal structures of the 2:1 and 1:2 L-menthol:phenol cocrystals
solved from powder diffraction data.
L-Menthol/Phenol 2:1 L-Menthol/Phenol 1:2
Deposition number 2153851 2153852
Formula 2(C10H20O), C6H6O C10H20O, 2(C6H6O)
1
Molecular weight /g mol 406.63 344.48
Bravais lattice monoclinic P orthorhombic P
Spacegroup P21 P212121
a /Å 6.099 20.3936
b /Å 21.7903 16.8032
c /Å 10.4267 6.2151
/ 90 90
/° 100.9126 90
/° 90 90
3
Volume /Å 1360.649 2129.77
Z 2 4
3
Density (calc.) /g cm 0.992 1.074
F(000) 452 752
2 a ge / 1.5 26.0 1.5 27.0
Background 30 Legendre polynomials 30 Legendre polynomials
Profile function Pseudo-Voigt Pseudo-Voigt
Rp 0.0436 0.0252
Rwp 0.0580 0.0336
Rexp 0.0148 0.0125
R(all) 0.0796 0.0342
wR(all) 0.0734 0.0373
R 0.0700 0.0322
wR 0.0727 0.0372
GOF 3.93 2.69

9
References
1. R. J. L. Andon, J. F. Counsell, E. F. G. Herington and J. F. Martin, Transactions of the Faraday
Society, 1963, 59, 830-835.
2. APEX suite of crystallographic software, APEX 3, Version 2019-1.0, Bruker AXS Inc., Madison,
Wisconsin, USA, 2019.

3. SAINT, Version 8.40A and SADABS, Version 2016/2, Bruker AXS Inc., Madison, Wisconsin, USA,
2016/2019.

4. G. M. Sheldrick, Acta Crystallogr. Sect. A 2015, 71, 3 8.

5. G. M. Sheldrick, Acta Crystallogr. Sect. C 2015, 71, 3 8.

6. C. B. Hübschle, G. M. Sheldrick, B. Dittrich, J. Appl. Cryst. 2011, 44, 1281 1284

7. International Tables for Crystallography, Vol. C (Ed.: A. J. Wilson), Kluwer Academic Publishers,
Dordrecht, The Netherlands, 1992, Tables 6.1.1.4 (pp. 500 502), 4.2.6.8 (pp. 219 222), and 4.2.4.2 (pp.
193 199).

8. C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington, P. McCabe, E. Pidcock, L. Rodriguez-Monge,

R. Taylor, J. van de Streek, P. A. Wood, J. Appl. Cryst. 2008, 41, 466 470.

9. A. L. Spek, Acta Crystallogr. Sect. D 2009, 65, 148 155.

10. H. D. Flack, Acta Crystallogr. Sect A 1983, 39, 876 881.

11. S. Parsons, H. D. Flack, T. Wagner, Acta Crystallogr. Sect B 2013, 69, 249 259.

12. WinXPOW, Version 3.6.0.1, Stoe & Cie GmbH, Darmstadt, Germany, 2018.

13. W. I. F. David, K. Shankland, J. van de Streek, E. Pidcock, W. D. S. Motherwell, J. C. Cole, J. Appl.

Cryst. 2006, 39, 910-915.

14. C. R. Groom, I. J. Bruno, M. P. Lightfoot, S. C. Ward, Acta Crystallogr. Sect B 2016, 72, 171 179.

15. Y. Corvis, P. Negrier S. Massip, J.-M. Leger, P. Espeau, Cryst. Eng. Commun. 2012, 14, 7055 7064.

16. V. E. Zavodnik, V. K. Bel'skii, P. M. Zorkii, Zh. Strukt. Khim. 1987, 28, 793 795.

17. V. Petricek, M. Dusek, L. Palatinus, Z. Kristallogr. 2014, 229, 345 352.

10
Discussion

4. Discussion
The results of Papers I–VI are discussed in two sections, one with the focus on measuring and
modeling methods to obtain SLE data for eutectic mixtures of various complexity (Section 4.1),
and the other on the characteristics of the pure constituents and the mixture leading to the
formation of eutectic mixtures with deep depression at the eutectic point (Section 4.2).

4.1 Measuring and modeling SLE

Since the term DES was introduced, considerable efforts have been made to understand DES
formation. Despite the numerous approaches employed to unravel the unique DES characters,
SLE studies have provided the most valuable insights. The SLE phase diagram is used to
determine the melting temperature of the DES at any ratio between constituents. Moreover,
modeling SLE data provides information about intermolecular interactions between unlike
molecules. Nevertheless, as depicted in Papers III–VI, measuring and modeling SLE is not a
straightforward task and often accompanies many difficulties. These difficulties are usually
attributed to the solid rather than the liquid phase. For eutectic systems with a large depression
in the melting temperature, the high viscosity of the liquid solution in the vicinity of its melting
temperature might hinder the crystallization, preventing the determination of the solidus or the
liquidus temperatures of the eutectic system. Furthermore, crystallization of constituents at low
temperatures might lead to the formation of metastable polymorphs. For modeling SLE data, the
melting properties of pure constituents are required, which can be unavailable due to their
thermal instability or polymorphism–the pure constituent can exist in different crystal structures.
Moreover, information on the solid phases crystallized from the liquid solution is a prerequisite
for adequate modeling of SLE. Nevertheless, obtaining such information becomes very
complicated in the case of glass or cocrystal formations or if the melting temperature of the
mixture is significantly lower than room temperature, which consequently requires special
handling and preparation for the samples. Therefore, it is evident that measuring and modeling
SLE in many DES are rather complicated and nontrivial tasks.

This thesis provides several advancements in measuring and modeling SLE in binary eutectic
mixtures. In the literature, SLE data of DES were commonly measured by visual methods or DSC.
Nevertheless, these measurements were obtained assuming that the system is of the simple
eutectic type without carefully analyzing the formed solid phases. The conventional DSC
protocol–in situ crystallization of the liquid solution at the DSC (see Section 2.3.2.2)–was found

164
Discussion

unsuitable for determining the SLE phase diagram of eutectic systems prone to glass formation.
As shown in Paper III, annealing the DSC samples prior to measurements allows for determining
the solidus temperature of the mixture. Nevertheless, the samples only partially crystallized even
after annealing for one month. To overcome long annealing times, an efficient sample preparation
method was proposed in Paper IV and applied in Papers V and VI, which allowed for fully
crystallizing the samples within days instead of months. Compared to the conventional DSC
protocol, the method used in Papers IV–VI allows for overcoming glass and metastable phases
formation, enabling the proper determination of SLE data for the systems. By obtaining the
solidus temperature and SLE data over the entire composition range of the system, cocrystal
formation can be adequately investigated. Moreover, the method allows for combined DSC and
powder XRD analyses, which aid in accurately interpreting DSC curves by assigning the
observed peaks to the corresponding solid observed in the XRD pattern.

The eutectic composition can be obtained experimentally from the DSC analysis by constructing
Tammann’s plot. Tammann’s plot can be constructed by plotting the area of the solidus peak as
a function of the mole fraction of one constituent. The determined solidus peak area at each
composition is fitted linearly in the hypereutectic and hypoeutectic regions, and the eutectic
composition corresponds to the intercept of the two lines, i.e., supposedly, the composition at
which the sample contains only pure eutectic phase. Nevertheless, determining the eutectic
point composition experimentally using Tammann’s plot is quite tedious, especially in the case
of glass formation or polymorphism. The components might crystallize partially or in two different
polymorphs, and the determined eutectic transition area might not correspond to the actual mole
fraction of the eutectic phase in the sample. Moreover, for eutectic systems with cocrystal
formation in which the cocrystal melting temperature is not far from the eutectic temperature,
the solidus and liquidus peaks might overlap. Hence, the area of the solidus peak cannot be
determined correctly. As an alternative to Tammann’s plot, Papers III–VI show that correlative
thermodynamic models (NRTL) can reasonably predict the eutectic composition.

Thermodynamic models are used for describing the nonideality of the liquid phase to obtain the
SLE phase diagram and the eutectic point of the eutectic system. However, as shown in Papers
IV–VI, not all eutectic systems are of the simple eutectic type. Therefore, experimental
investigation for cocrystal formation is needed to model the SLE phase diagram. For modeling
SLE in eutectic systems with cocrystal formation, the melting properties and stoichiometry of the
cocrystal are required. The melting properties can be obtained by DSC, and a combination of

165
Discussion

powder XRD and DSC analyses can hint at the cocrystal stoichiometry. Nevertheless, confirming
the postulated stoichiometry of the cocrystal from powder XRD and DSC is not straightforward.
SC-XRD technique provides definite proof for the cocrystal stoichiometry. However, obtaining a
single-crystal suitable for SC-XRD analysis can be impossible in many cases, mainly due to
kinetic limitations in crystallization. In the L-menthol/thymol eutectic system, no single crystals
could be obtained for the observed cocrystals, and the crystal structure could not be solved
from the powder XRD data. In contrast, the crystal structure of four never reported before
cocrystals was solved by SC-XRD, namely, ChCl:catechol 1:2 and 1:1 ratios, ChCl:hydroquinone
1:1 ratio, and L-menthol:phenol 2:1 ratio. The crystal structure of the L-menthol:phenol 1:2 and
2:1 cocrystals was obtained from powder XRD data.

COSMO-RS has been used in the literature to predict the SLE phase diagram of eutectic
systems. However, as COSMOtherm software was used to perform the SLE calculations, the
calculations were limited to eutectic systems of the simple eutectic type. Moreover, the solid–
solid transition of pure constituents was neglected in many cases due to the limitations of the
implemented SLE calculations algorithm in COSMOtherm software for considering the solid–
solid transition. Paper VI was the first to employ COSMO-RS to calculate SLE in eutectic systems
with cocrystal formation, and satisfactory predictions were obtained using the TZVPD_FINE
parameterization. Besides showing that COSMO-RS can be a valuable tool to predict the SLE
phase diagram of nonideal eutectic systems with cocrystal formation, the study highlights two
important outcomes. First, COSMOtherm SLE calculations limitations and assumptions should
not restrict using the model for predicting the SLE phase diagram of eutectic systems with a
complex character or to justify the simplification of SLE calculations by assuming simple eutectic
behavior and neglecting the solid–solid transition. Second, both TZVP and TZVPD_FINE
parameterizations should be used, as COSMO-RS predictions might differ significantly at the
two levels.

Eventually, SLE studies are challenging, particularly for eutectic systems with a significant
depression in the melting temperature besides solid complex formation. To obtain the SLE phase
diagram of eutectic systems with a complex character, rigorous thermal and solid
characterization analyses and advanced thermodynamic modeling should be combined.
However, the sample preparation method is the crucial factor that would allow or hinder the
experimental determination of SLE data. Without experimental SLE data over the entire
composition range and information about the formed solid phases, the efforts for modeling the

166
Discussion

liquid phase nonideality could be unavailing. Therefore, in SLE studies, experimental

investigations and thermodynamic modeling are complementary and should be carried out
jointly.

4.2 DES formation

Two different definitions for DES have been proposed in the literature. The first definition states
that DESs are a unique type of solvents formed by mixing HBA and HBD at a specific ratio. In
contrast, the second definition states that DES are common nonideal eutectic mixtures with a
significant difference between the eutectic temperature and the ideal eutectic temperature.

In Paper I, the influence of the melting enthalpy and entropy as well as the nonideality of the
liquid phase on the eutectic temperature of the eutectic mixture was studied. It was
demonstrated that both the activity coefficients and the melting entropy and enthalpy of
components influence the depression at the eutectic point similarly. As an example, the
depression at the eutectic point in the ideal eutectic system L-menthol/cyclohexanecarboxylic
acid ( , 49.6 K, Paper II) is similar to that observed in L-menthol/thymol ( ,

41.5 K, Paper IV) and L-menthol/phenol ( , 41.8 K, Paper VI), though the latter two
systems show significant negative deviation from ideality. This is attributed to the lower melting
enthalpy of cyclohexanecarboxylic acid ( h 10.69 kJ mol ) than thymol ( h
20.64 kJ mol ) and phenol ( h 12.36 kJ mol ) and cocrystal formation in L-menthol/thymol
and L-menthol/phenol. However, in none of these systems, the depression at the eutectic point
was significant, i.e., < 100K, as that observed in ChCl-based DES. Based on Paper I, significant
depression at the eutectic point, as depicted in ChCl-based DES, can only be realized in nonideal
eutectic mixtures formed by constituents with low melting enthalpy and entropy. Therefore, it is
evident that neither the nonideality of the mixture nor the melting properties of pure constituents
alone aid in the formation of DES but rather a combination of both.

A simple approach was proposed in Paper II based on Paper I conclusions to select the eutectic
system constituents with low melting entropy from a pool of chemical substances sharing the
same chemical nature and melting temperature. It was shown in Paper II that constituents with
a more rigid and symmetrical molecular structure could form eutectic mixtures with larger
depression at the eutectic point compared to constituents with flexible and asymmetrical
molecular structures. However, as shown in Section 2.4.1, substances with low melting entropy
would have higher melting temperatures. Moreover, molecules that are rotationally symmetrical

167
Discussion

are packed more efficiently in the crystal lattice, leading to a more ordered crystal structure and
higher melting enthalpy. The interrelation between rotational symmetry and melting temperature
149
and enthalpy can justify the odd-even effect in alkanes and dicarboxylic acids and the
difference in the melting properties of sugar alcohol stereoisomers. 108 For example, sorbitol and
mannitol are stereoisomers, but their melting properties are quite different. The melting
temperatures, enthalpies, and entropies are 372.0 and 441.0 K, h 35.9 and 59.5 kJ
mol–1, and s 96.5 and 134.9 J mol–1 K–1 for sorbitol and mannitol, respectively. Despite the
mannitol being rotationally symmetrical and sorbitol being asymmetrical, mannitol has a higher
melting entropy than sorbitol, which could be attributed to the more ordered crystal structure of
mannitol, depicted by its higher melting enthalpy. The entropy gain due to the transition from the
more ordered crystal of mannitol to a liquid is expected to be higher than the transition from the
less ordered crystal structure of sorbitol to a liquid. The approach proposed in Paper II does not
take into account the influence of the crystal structure of constituents on their melting enthalpy,
which can largely affect the eutectic temperature of their mixtures. Hence, the low melting
entropy might not result from the rotational symmetry and rigidity of the molecular structure but
from the disorder of the crystal structure.

As mentioned previously, the majority of DES reported in the literature are ChCl-based. To
understand the unique character of ChCl as a HBA forming DES, it is necessary to assess its
melting properties and activity coefficients in the liquid solution. However, ChCl decomposes
before melting; hence its melting properties are unavailable. Without the melting properties of
ChCl, the activity coefficients cannot be calculated from experimental SLE data. In Paper V, an
alternative method to calculate the activity coefficients of ChCl was proposed by mixing ChCl
with known coformers to form eutectic mixtures with cocrystal formation. The liquidus data of
the formed cocrystals were used to fit the binary interaction parameters of the correlative
thermodynamic model (NRTL) to describe the nonideality of ChCl in the liquid phase.
Accordingly, ChCl nonideality was evaluated in two eutectic systems, namely, ChCl/catechol
and ChCl/hydroquinone. The calculated activity coefficients of ChCl showed that ChCl shows a
negative deviation from ideal behavior in the liquid solution, contrary to the ideal or near-ideal
solution behavior observed in the literature studies. This contradictory observation is attributed
to the approach used in the literature to estimate the melting properties of ChCl. In the literature
approach, ChCl melting properties were estimated assuming the ideal solution model, which, as
shown in Paper I, can reasonably predict the SLE data of the ChCl liquidus line. However, the

168
Discussion

estimated melting enthalpy assuming the ideal solution model may conceal the nonideality of
ChCl in the liquid solution.

As widely known and explicitly shown in Paper V, ChCl undergoes solid–solid transition at around
352 K, which was disregarded in most SLE studies of ChCl-based DES in the literature, mainly
due to the lack of proper solid phase characterization and measuring SLE visually. To estimate
ChCl melting properties, SLE data above the solid–solid transition are needed. However, no
experimental liquidus temperatures for ChCl above its solid–solid transition temperature could
be measured. Accordingly, no distinct melting properties of ChCl could be estimated in Paper V.
Instead, the melting entropy range of substances with a highly disordered crystal structure like
ChCl high-temperature form was screened to obtain a pair of melting enthalpy and temperature
values at each melting entropy. The SLE of the ChCl liquidus line below the solid–solid transition
temperature can be well described using any combination of melting enthalpy, entropy, and
temperature. It was shown that the lowest possible melting temperature of ChCl, estimated at
the highest possible melting entropy of a substance with a disordered crystal structure, is still
higher than the estimated melting temperature of ChCl found in the literature. In addition, the
lowest possible melting enthalpy of ChCl, estimated at the lowest possible melting entropy of a
substance with a disordered crystal structure, is higher than the estimated melting enthalpy of
ChCl found in the literature. Thus, it was concluded that the melting properties of ChCl found in
the literature could be unreasonable.

Although the same argument about the interrelation between melting properties and activity
coefficients can be used to question the range of melting properties of ChCl estimated in Paper
V, various evidence can support the outcomes of Paper V regarding the nonideality and melting
properties of ChCl over the ones in the literature. First, ChCl solid–solid transition was not
considered in the literature for modeling SLE data of the ChCl liquidus line. Second, liquidus
temperatures above ChCl solid–solid transition were used to estimate the melting properties of
ChCl in the literature, which, as shown in Paper V, could not be measured. Therefore, the
estimated melting properties of ChCl could be more reasonable than those of the literature
studies. Accordingly, the evaluation of the liquid phase nonideality of ChCl-based DES in Paper
V could be more realistic than that found in the literature. Given the melting properties and activity
coefficients of ChCl found in Paper V, it is clear that the reason why ChCl can form DES is its
estimated low melting entropy and enthalpy and the negative deviation from ideality in the liquid
phase. These findings well confirm those of Paper I.

169
Discussion

Substances with symmetrical and disordered crystal structures are called plastic crystalline
materials. Plastic crystalline materials have very low melting entropy and enthalpy but high
150, 151
melting temperature compared to their chemical isomers with ordered crystal structures.
152
The melting entropy of plastic crystalline materials is between 0.5–2 R. Various plastic
crystalline materials have been used in the literature as DES constituents besides ChCl.
However, the influence of the solid phase plasticity on the observed depression at the eutectic
143
point has not been addressed. Martins et al. studied the eutectic mixtures containing L-
menthol or thymol with camphor, borneol, and sobrerol. Camphor and borneol are known plastic
crystalline materials. Although all the eutectic systems show ideal solution behavior and the
melting temperature of camphor and borneol is higher than sobrerol, the eutectic temperatures
observed for eutectic systems containing either camphor or borneol are significantly lower than
those containing sobrerol. The eutectic temperatures are: L-menthol/camphor (Te ~ 283 K), L-
menthol/borneol (Te ~ 283 K), and L-menthol/sorbrerol (Te ~ 331 K). Therefore, it is evident that
plastic crystalline materials can form eutectic mixtures with a large depression at the eutectic
point.

Eventually, for the purpose of forming eutectic mixtures with a large depression in the melting
temperature of the mixture, constituents with low melting entropy should be selected. The low
melting entropy can result from the rigidity and rotational symmetry of molecular structures
(Paper II) or the disorder and symmetry of the crystal structure (Paper V). The large depression
in the melting temperature of the mixture might aid in extending the number of constituents that
can be used or the ratio at which the mixture is liquid below the desired application temperature.

170
Conclusion

5. Conclusion
The work done in this thesis aims to understand the formation of eutectic mixtures with
significant depression in the melting temperature of the mixture at the eutectic point. To achieve
this goal, SLE in eutectic systems of different complexity was studied based on various
theoretical and experimental methods. The SLE phase diagram of 11 binary eutectic systems
was reported in the six published papers. The SLE phase diagrams were acquired by rigorous
thermal analysis by DSC, solid characterization by powder XRD and SC-XRD, and
thermodynamic modeling using correlative and predictive thermodynamic models. In addition to
the reported SLE phase diagrams, seven cocrystals in four eutectic systems were identified, five
of which were deposited at the Cambridge Crystallographic Data Centre (CCDC).

This thesis shows that the typical DSC protocol used to estimate the SLE data in binary eutectic
systems, i.e., in situ crystallization within the DSC run, is unsuitable for measuring the SLE data
of eutectic systems prone to glass or cocrystal formation or in the case of polymorphism. In
contrast, the proposed sample preparation method in this thesis allowed for measuring the SLE
in eutectic systems with large depression at the eutectic point and with significantly low liquidus
and solidus temperatures. Moreover, the proposed method allows for combined DSC and XRD
analyses, unraveling polymorphism, solid–solid transition, solid solution region, and cocrystal
formation.

Correlative and thermodynamic models were found to be capable of well predicting the eutectic
point and the SLE phase diagram of eutectic systems of various complexity. However, for
modeling SLE in eutectic systems with cocrystal formation, the melting properties and the
stoichiometry of the cocrystals are needed, which can be acquired by powder and SC-XRD.
Therefore, solid characterization techniques complement SLE modeling to obtain good
predictions for the SLE phase diagram and eutectic point of eutectic systems.

By combining the obtained experimental results with the performed theoretical studies, the
thesis shows that the reason for DES formation is a combination of strong negative deviation
from ideality as well as small melting entropy of pure constituents. The eutectic mixture
constituents selection can be based on their molecular structure (rigid and rotationally
symmetrical) or crystal structure (plastic crystalline materials).

171
Outlook

6. Outlook
The majority of DES studied in the literature are ChCl-based eutectic systems. In this thesis, the
SLE phase diagram of two ChCl-based DES was reported. Two cocrystals were identified in
ChCl/catechol, and one cocrystal was identified in ChCl/hydroquinone. The SLE data of the
coformer and cocrystals liquidus lines were used to evaluate the liquid phase nonideality in the
two eutectic systems. It was shown that ChCl-based eutectic mixtures show a deep depression
at the eutectic point due to strong negative deviation from ideality and low melting enthalpy and
entropy of ChCl. However, ChCl is hygroscopic and thermally unstable. Moreover, ChCl-based
eutectic mixtures are highly viscous. Therefore, other plastic crystalline materials that can be
used as DES constituents while overcoming the drawbacks of ChCl should be sought.

Based on the findings of this thesis, exploring eutectic mixtures containing plastic crystalline
materials is encouraged. The complex phase behavior of plastic crystalline materials and their
eutectic mixtures, i.e., multiple solid–solid transitions, require expertise in thermal and solid
characterization techniques. The measuring and modeling methods presented in this thesis
facilitate further research on the phase behavior of eutectic mixture containing plastic crystalline
materials. Plastic crystalline materials have already been used in several applications, for
153, 154
example, as phase-change and conducting materials. Therefore, as the melting
temperature and physicochemical properties of eutectic mixtures or cocrystals containing plastic
crystalline materials could be tuned, further potential applications for eutectic mixtures
containing plastic crystalline materials could be found.

So far, there is no robust method to predict the melting properties of thermally unstable
substances. In the literature, the melting properties of a thermally unstable substance can be
estimated from binary SLE or solubility data. However, estimating the melting properties of
thermally unstable substances from solubility or SLE data requires a robust thermodynamic
model to predict the nonideality of the liquid solution. The proposed method in Paper V applied
to ChCl can be used to indirectly assess the melting properties and activity coefficients of
thermally unstable substances. A cocrystal of the thermally unstable substance with a coformer
can be prepared, and the liquidus or solubility data of the cocrystal can be used to obtain the
correlative thermodynamic model parameters. Then, the SLE or solubility data and the calculated
activity coefficients of the thermally unstable substance can be used to estimate the melting

172
Outlook

properties of the thermally unstable substance indirectly. The method can be further applied to
bioactive thermally unstable substances, such as vitamins and amino acids.

Cocrystal stoichiometry and melting properties are needed to use correlative or predictive
thermodynamic models to model SLE in eutectic systems with cocrystal formation. In the case
of unavailable melting enthalpy of cocrystals, two approaches have been proposed in the
155
literature for their estimation. In the first approach, the melting enthalpy of the cocrystal is
calculated using the melting enthalpy and temperature of pure components and the
stoichiometry of the cocrystal. However, it was shown that this approach provides unreliable
estimates in many cases. In the second approach, the cocrystal melting properties are estimated
by regressing the experimental liquidus data. Obviously, the second approach restricts the
advantage of using predictive models for SLE calculations. Therefore, methods for predicting
the melting enthalpy of cocrystals are needed to allow for fully predictive modeling of SLE in
eutectic systems with cocrystal formation.

In the literature, PC-SAFT was used to successfully model cocrystal solubility in various solvents
155-158
by calculating the ternary phase diagram of a compound, a coformer, and a solvent
COSMO-RS was also used in the literature to predict the ternary SLE phase diagram with
159 159 160
cocrystal formation. Loschen and Klamt employed COSMO-DARE method, which
includes two additional model parameters. The COSMO-DARE parameters were obtained by
fitting experimental cocrystal solubility data, which ultimately eliminated the predictive ability of
159
COSMO-RS. Nevertheless, as the work of Loschen and Klamt is the only study employing
COSMO-RS for predicting ternary SLE phase diagrams with cocrystal formation, further studies
are encouraged to assess COSMO-RS ability to predict binary and ternary SLE phase diagrams
with cocrystal formation.

173
Abbreviations

Abbreviations
CCDC Cambridge Crystallographic Data Centre

ChCl choline chloride

COSMO-RS conductor-like screening model for realistic solvation

DES deep eutectic solvent

DSC differential scanning calorimetry

GC group contribution

HBA hydrogen bond acceptor

HBD hydrogen bond donor

IL ionic liquid

NRTL non-random two-liquid

RK Redlich-Kister

SLE solid–liquid equilibria

SC-XRD single-crystal X-ray diffraction

XRD X-ray diffraction

174
Symbols

Symbols
𝐴 COSMO molecular surface area

𝑎 Effective contact area

Solid

Solid

𝐶𝑜 Cocrystal

𝑐 Constant pressure heat capacity

Activity coefficient

𝑐 , Heat capacity difference at the melting temperature

𝑔 Gibbs energy

enthalpy

Melting enthalpy

Cocrystal melting enthalpy

Entropy

Melting entropy

Rotational entropy

Conformational entropy

Transitional entropy

𝑒 , Total interaction energy in COSMO-RS

𝑒 , Hydrogen bonding interaction energy

𝑒 , Misfit interaction energy

Eccentricity

175
Symbols

𝑓 Fugacity

𝑓 Standard state fugacity

𝑔 Excess Gibbs energy

𝐾 Equilibrium constant

𝐿 Liquid

Pseudo-chemical potential

Chemical potential

Stoichiometric coefficient

σ Probability distribution

Rotational symmetry number

𝑅 Universal gas constant

𝑆 solid

𝑇 Temperature

𝑇 Melting temperature

𝑇 Cocrytal melting temperature

𝑉 COSMO molecular volume

Φ Flexibility number

Mole fraction

176
List of figures

List of figures
Figure 1. Schematic representation of solid–liquid phase diagram of (A) binary mixture with
miscibility in the solid state and (B) binary eutectic mixture. .....................................................11
Figure 2. Schematic representations of solid–liquid phase diagrams of different types, defining
the crystallized solid phases in each binary eutectic system. ...................................................12
Figure 3. Schematic structure of differential scanning calorimetry chamber. ............................15
Figure 4. Schematic representation of a phase diagram of a hypothetical binary eutectic mixture
and the observed differential scanning calorimetry curve at each labeled point. ......................16
Figure 5. Determining the melting, solidus, and liquidus temperatures and phase transition
enthalpy from a differential scanning calorimetry curve. ...........................................................17
Figure 6. Schematic representation of solid–liquid phase diagrams of different types and the
obtained Tammann’s plot for each type. ..................................................................................18
Figure 7. Schematic representation of a differential scanning calorimetry (DSC) curve of a sample
analyzed using the typical DSC protocol: first heating, cooling, and second heating runs. In (A),
the sample in situ crystallizes in the DSC during the cooling run, while in (B), the sample
undergoes cold crystallization, i.e., crystallization during the heating run. ................................20
Figure 8. Thermodynamic cycle used to derive the solid–liquid equilibrium equation. ..............22
Figure 9. (A) -profile and (B) -potential of oxalic acid calculated using density functional theory
and COSMO-RS, respectively. .................................................................................................29

177
References

References
1. P. Anastas and N. Eghbali. Green Chemistry: Principles and Practice. Chem. Soc. Rev.
2010, 39 (1), 301-312. DOI: 10.1039/B918763B.
2. N. Winterton. The green solvent: a critical perspective. Clean Technol. Environ. Policy
2021, 23 (9), 2499-2522. DOI: 10.1007/s10098-021-02188-8.
3. J. F. Brennecke and E. J. Maginn. Ionic liquids: Innovative fluids for chemical processing.
AlChE J. 2001, 47 (11), 2384-2389. DOI: 10.1002/aic.690471102.
4. K. N. Marsh, A. Deev, A. C. T. Wu, E. Tran and A. Klamt. Room temperature ionic liquids
as replacements for conventional solvents – A review. Korean J. Chem. Eng. 2002, 19 (3),
357-362. DOI: 10.1007/BF02697140.
5. R. D. Rogers and K. R. Seddon, Ionic Liquids as Green Solvents, American Chemical
Society, 2003.
6. K. N. Marsh, J. A. Boxall and R. Lichtenthaler. Room temperature ionic liquids and their
mixtures—a review. Fluid Phase Equilib. 2004, 219 (1), 93-98.
DOI:10.1016/j.fluid.2004.02.003.
7. S. Werner, M. Haumann and P. Wasserscheid. Ionic Liquids in Chemical Engineering.
Annu. Rev. Chem. Biomol. Eng. 2010, 1 (1), 203-230. DOI: 10.1146/annurev-
chembioeng-073009-100915.
8. P. Wasserscheid. Volatile times for ionic liquids. Nature 2006, 439 (7078), 797-797. DOI:
10.1038/439797a.
9. C. Jork, C. Kristen, D. Pieraccini, A. Stark, C. Chiappe, Y. A. Beste and W. Arlt. Tailor-
made ionic liquids. J. Chem. Thermodyn. 2005, 37 (6), 537-558.
DOI:10.1016/j.jct.2005.04.013.
10. B. Tang and K. Row. Recent developments in deep eutectic solvents in chemical
sciences. Monatsh. Chem. 2013, 144 (10), 1427-1454. DOI: 10.1007/s00706-013-1050-
3.
11. P. E. Rakita, in Ionic Liquids as Green Solvents, American Chemical Society, 2003, 856
(3), 32-40.
12. A. P. Abbott, G. Capper, D. L. Davies, H. L. Munro, R. K. Rasheed and V. Tambyrajah.
Preparation of novel, moisture-stable, Lewis-acidic ionic liquids containing quaternary

178
References

ammonium salts with functional side chains. Chem. Commun. 2001, (19), 2010-2011.
DOI: 10.1039/B106357J.
13. A. P. Abbott, G. Capper, D. L. Davies, R. K. Rasheed and V. Tambyrajah. Novel solvent
properties of choline chloride/urea mixtures. Chem. Commun. 2003, (1), 70-71. DOI:
10.1039/b210714g.
14. A. P. Abbott, D. Boothby, G. Capper, D. L. Davies and R. K. Rasheed. Deep Eutectic
Solvents Formed between Choline Chloride and Carboxylic Acids: Versatile Alternatives
to Ionic Liquids. J. Am. Chem. Soc. 2004, 126 (29), 9142-9147. DOI: 10.1021/ja048266j.
15. E. L. Smith, A. P. Abbott and K. S. Ryder. Deep Eutectic Solvents (DESs) and Their
Applications. Chem. Rev. 2014, 114 (21), 11060-11082. DOI: 10.1021/cr300162p.
16. S. L. Perkins, P. Painter and C. M. Colina. Experimental and computational studies of
choline chloride-based deep eutectic solvents. J. Chem. Eng. Data 2014, 59 (11), 3652-
3662. DOI: 10.1021/je500520h.
17. G. García, M. Atilhan and S. Aparicio. The impact of charges in force field
parameterization for molecular dynamics simulations of deep eutectic solvents. J. Mol.
Liq. 2015, 211, 506-514. DOI:10.1016/j.molliq.2015.07.070.
18. G. García, M. Atilhan and S. Aparicio. An approach for the rationalization of melting
temperature for deep eutectic solvents from DFT. Chem. Phys. Lett. 2015, 634, 151-155.
DOI:10.1016/j.cplett.2015.06.017.
19. D. V. Wagle, C. A. Deakyne and G. A. Baker. Quantum Chemical Insight into the
Interactions and Thermodynamics Present in Choline Chloride Based Deep Eutectic
Solvents. J. Phys. Chem. B 2016, 120 (27), 6739-6746. DOI: 10.1021/acs.jpcb.6b04750.
20. S. Zahn, B. Kirchner and D. Mollenhauer. Charge Spreading in Deep Eutectic Solvents.
Chem. Phys. Chem. 2016, 17 (21), 3354-3358. DOI: 10.1002/cphc.201600348.
21. S. Zhu, H. Li, W. Zhu, W. Jiang, C. Wang, P. Wu, Q. Zhang and H. Li. Vibrational analysis
and formation mechanism of typical deep eutectic solvents: An experimental and
theoretical study. J. Mol. Graphics Modell. 2016, 68, 158-175. DOI:
10.1016/j.jmgm.2016.05.003.
22. T. Aissaoui, Y. Benguerba and I. M. AlNashef. Theoretical investigation on the
microstructure of triethylene glycol based deep eutectic solvents: COSMO-RS and
TURBOMOLE prediction. J. Mol. Struct. 2017, 1141, 451-456. DOI:
10.1016/j.molstruc.2017.04.009

179
References

23. M. Atilhan and S. Aparicio. Behavior of Deep Eutectic Solvents under External Electric
Fields: A Molecular Dynamics Approach. J. Phys. Chem. B 2017, 121 (1), 221-232. DOI:
10.1021/acs.jpcb.6b09714.
24. R. Stefanovic, M. Ludwig, G. B. Webber, R. Atkin and A. J. Page. Nanostructure,
hydrogen bonding and rheology in choline chloride deep eutectic solvents as a function
of the hydrogen bond donor. PCCP 2017, 19 (4), 3297-3306. DOI: 10.1039/C6CP07932F.
25. S. Zahn. Deep eutectic solvents: similia similibus solvuntur? PCCP 2017, 19 (5), 4041-
4047. DOI: 10.1039/C6CP08017K.
26. V. Alizadeh, F. Malberg, A. A. H. Pádua and B. Kirchner. Are There Magic Compositions
in Deep Eutectic Solvents? Effects of Composition and Water Content in Choline
Chloride/Ethylene Glycol from Ab Initio Molecular Dynamics. J. Phys. Chem. B 2020, 124
(34), 7433-7443. DOI: 10.1021/acs.jpcb.0c04844.
27. S. Kaur, M. Kumari and H. K. Kashyap. Microstructure of Deep Eutectic Solvents: Current
Understanding and Challenges. J. Phys. Chem. B 2020, 124 (47), 10601-10616. DOI:
10.1021/acs.jpcb.0c07934.
28. L. Percevault, A. Jani, T. Sohier, L. Noirez, L. Paquin, F. Gauffre and D. Morineau. Do
Deep Eutectic Solvents Form Uniform Mixtures Beyond Molecular Microheterogeneities?
J. Phys. Chem. B 2020, 124 (41), 9126-9135. DOI: 10.1021/acs.jpcb.0c06317.
29. A. P. Abbott, K. J. Edler and A. J. Page. Deep eutectic solvents—The vital link between
ionic liquids and ionic solutions. J. Chem. Phys. 2021, 155 (15), 150401. DOI:
10.1063/5.0072268.
30. S. Kaur, A. Gupta and H. K. Kashyap. Nanoscale Spatial Heterogeneity in Deep Eutectic
Solvents. J. Phys. Chem. B 2016, 120 (27), 6712-6720. DOI: 10.1021/acs.jpcb.6b04187.
31. O. S. Hammond, D. T. Bowron and K. J. Edler. Liquid structure of the choline chloride-
urea deep eutectic solvent (reline) from neutron diffraction and atomistic modelling.
Green Chem. 2016, 18 (9), 2736-2744. DOI: 10.1039/C5GC02914G.
32. I. I. I. Alkhatib, D. Bahamon, F. Llovell, M. R. M. Abu-Zahra and L. F. Vega. Perspectives
and guidelines on thermodynamic modelling of deep eutectic solvents. J. Mol. Liq. 2020,
298, 112183. DOI:10.1016/j.molliq.2019.112183.
33. S. Mainberger, M. Kindlein, F. Bezold, E. Elts, M. Minceva and H. Briesen. Deep eutectic
solvent formation: a structural view using molecular dynamics simulations with classical
force fields. Mol. Phys. 2017, 115 (9-12), 1309-1321. DOI:
10.1080/00268976.2017.1288936.

180
References

34. T. Zhekenov, N. Toksanbayev, Z. Kazakbayeva, D. Shah and F. S. Mjalli. Formation of

type III Deep Eutectic Solvents and effect of water on their intermolecular interactions.
Fluid Phase Equilib. 2017, 441, 43-48. DOI:10.1016/j.fluid.2017.01.022.
35. L. Yang, C.-W. Chang and S.-T. Lin. A novel multiscale approach for rapid prediction of
phase behaviors with consideration of molecular conformations. AlChE J. 2016, 62 (11),
4047-4054. DOI: 10.1002/aic.15290.
36. Y. Cui, K. D. Fulfer, J. Ma, T. K. Weldeghiorghis and D. G. Kuroda. Solvation dynamics of
an ionic probe in choline chloride-based deep eutectic solvents. PCCP 2016, 18 (46),
31471-31479. DOI: 10.1039/C6CP06318G.
37. A. D. Pauric, I. C. Halalay and G. R. Goward. Combined NMR and molecular dynamics
modeling study of transport properties in sulfonamide based deep eutectic lithium
electrolytes: LiTFSI based binary systems. PCCP 2016, 18 (9), 6657-6667. DOI:
10.1039/C5CP05811B.
38. M. Jasik and B. Szefczyk. Parameterization and optimization of the menthol force field
for molecular dynamics simulations. J. Mol. Model. 2016, 22 (10), 234. DOI:
10.1007/s00894-016-3082-1.
39. A. R. Harifi-Mood, R. Ghobadi, S. Matić, B. Minofar and D. Řeha. Solvation analysis of
some Solvatochromic probes in binary mixtures of reline, ethaline, and glyceline with
DMSO. J. Mol. Liq. 2016, 222, 845-853. DOI:10.1016/j.molliq.2016.07.036.
40. V. Lesch, A. Heuer, B. R. Rad, M. Winter and J. Smiatek. Atomistic insights into deep
eutectic electrolytes: the influence of urea on the electrolyte salt LiTFSI in view of
electrochemical applications. PCCP 2016, 18 (41), 28403-28408. DOI:
10.1039/C6CP04217A.
41. R. Ullah, M. Atilhan, B. Anaya, M. Khraisheh, G. Garcia, A. ElKhattat, M. Tariq and S.
Aparicio. A detailed study of cholinium chloride and levulinic acid deep eutectic solvent
system for CO2 capture via experimental and molecular simulation approaches. PCCP
2015, 17 (32), 20941-20960. DOI: 10.1039/C5CP03364K.
42. A. Das, S. Das and R. Biswas. Density relaxation and particle motion characteristics in a
non-ionic deep eutectic solvent (acetamide + urea): Time-resolved fluorescence
measurements and all-atom molecular dynamics simulations. J. Chem. Phys. 2015, 142
(3), 034505. DOI:10.1063/1.4906119.

181
References

43. D. Shah and F. S. Mjalli. Effect of water on the thermo-physical properties of Reline: An
experimental and molecular simulation based approach. PCCP 2014, 16 (43), 23900-
23907. DOI: 10.1039/C4CP02600D.
44. H. Sun, Y. Li, X. Wu and G. Li. Theoretical study on the structures and properties of
mixtures of urea and choline chloride. J. Mol. Model. 2013, 19 (6), 2433-2441. DOI:
10.1007/s00894-013-1791-2.
45. F. Dommert, K. Wendler, R. Berger, L. Delle Site and C. Holm. Force Fields for Studying
the Structure and Dynamics of Ionic Liquids: A Critical Review of Recent Developments.
Chem. Phys. Chem. 2012, 13 (7), 1625-1637. DOI: 10.1002/cphc.201100997.
46. T. Pal and R. Biswas. Heterogeneity and viscosity decoupling in (acetamide + electrolyte)
molten mixtures: A model simulation study. Chem. Phys. Lett. 2011, 517 (4–6), 180-185.
DOI:10.1016/j.cplett.2011.11.002.
47. M. A. R. Martins, S. P. Pinho and J. A. P. Coutinho. Insights into the Nature of Eutectic
and Deep Eutectic Mixtures. J. Solution Chem. 2019, 48 (7), 962-982. DOI:
10.1007/s10953-018-0793-1.
48. A. González de Castilla, J. P. Bittner, S. Müller, S. Jakobtorweihen and I. Smirnova.
Thermodynamic and Transport Properties Modeling of Deep Eutectic Solvents: A Review
on gE-Models, Equations of State, and Molecular Dynamics. J. Chem. Eng. Data. 2019,
65 (3), 943-967. DOI: 10.1021/acs.jced.9b00548.
49. L. J. B. M. Kollau, M. Vis, A. van den Bruinhorst, A. C. C. Esteves and R. Tuinier.
Quantification of the liquid window of deep eutectic solvents. Chem. Commun. 2018, 54
(95), 13351-13354. DOI: 10.1039/C8CC05815F.
50. J. K. U. Ling and K. Hadinoto. Deep Eutectic Solvent as Green Solvent in Extraction of
Biological Macromolecules: A Review. Int. J. Mol. Sci. 2022, 23 (6). DOI:
10.3390/ijms23063381.
51. A. Kalyniukova, J. Holuša, D. Musiolek, J. Sedlakova-Kadukova, J. Płotka-Wasylka and
V. Andruch. Application of deep eutectic solvents for separation and determination of
bioactive compounds in medicinal plants. Ind. Crops Products 2021, 172. DOI:
10.1016/j.indcrop.2021.114047.
52. L. Lomba, C. B. García, M. P. Ribate, B. Giner and E. Zuriaga. Applications of deep
eutectic solvents related to health, synthesis, and extraction of natural based chemicals.
Appl. Sci. 2021, 11 (21). DOI: 10.3390/app112110156.

182
References

53. A. S. Dheyab, M. F. A. Bakar, M. Alomar, S. F. Sabran, A. F. M. Hanafi and A. Mohamad.

Deep eutectic solvents (DESs) as green extraction media of beneficial bioactive
phytochemicals. Separations 2021, 8 (10). DOI: 10.3390/SEPARATIONS8100176.
54. Y. H. Choi and R. Verpoorte. Green solvents for the extraction of bioactive compounds
from natural products using ionic liquids and deep eutectic solvents. Curr. Opin. Food
Sci. 2019, 26, 87-93. DOI: 10.1016/j.cofs.2019.04.003.
55. D. C. Murador, L. M. de Souza Mesquita, N. Vannuchi, A. R. C. Braga and V. V. de Rosso.
Bioavailability and biological effects of bioactive compounds extracted with natural deep
eutectic solvents and ionic liquids: advantages over conventional organic solvents. Curr.
Opin. Food Sci. 2019, 26, 25-34. DOI: 10.1016/j.cofs.2019.03.002 Scopus.
56. K. M. Jeong, J. Ko, J. Zhao, Y. Jin, D. E. Yoo, S. Y. Han and J. Lee. Multi-functioning
deep eutectic solvents as extraction and storage media for bioactive natural products
that are readily applicable to cosmetic products. J. Clean. Prod. 2017, 151, 87-95.
DOI:10.1016/j.jclepro.2017.03.038.
57. I. Wazeer, M. K. Hadj-Kali and I. M. Al-Nashef. Utilization of Deep Eutectic Solvents to
Reduce the Release of Hazardous Gases to the Atmosphere: A Critical Review.
Molecules 2021, 26 (1). DOI: 10.3390/molecules26010075.
58. F. Soltanmohammadi, A. Jouyban and A. Shayanfar. New aspects of deep eutectic
solvents: extraction, pharmaceutical applications, as catalyst and gas capture. Chem.
Pap. 2020, (75), 439-453. DOI: 10.1007/s11696-020-01316-w.
59. S. E. E. Warrag, C. J. Peters and M. C. Kroon. Deep eutectic solvents for highly efficient
separations in oil and gas industries. Curr. Opin. Green Sustain. Chem. 2017, 5, 55-60.
DOI:10.1016/j.cogsc.2017.03.013.
60. Y. Liu, Z. Dai, Z. Zhang, S. Zeng, F. Li, X. Zhang, Y. Nie, L. Zhang, S. Zhang and X. Ji.
Ionic liquids/deep eutectic solvents for CO2 capture: Reviewing and evaluating. Green
Energy Environ. 2021, 6 (3), 314-328. DOI:10.1016/j.gee.2020.11.024.
61. A. Krishnan, K. P. Gopinath, D.-V. N. Vo, R. Malolan, V. M. Nagarajan and J. Arun. Ionic
liquids, deep eutectic solvents and liquid polymers as green solvents in carbon capture
technologies: a review. Environ. Chem. Lett. 2020, 18 (6), 2031-2054. DOI:
10.1007/s10311-020-01057-y.
62. F. Bezold and M. Minceva. A water-free solvent system containing an L-menthol-based
deep eutectic solvent for centrifugal partition chromatography applications. J.
Chromatogr. A 2019, 1587, 166-171. DOI:10.1016/j.chroma.2018.11.083.

183
References

63. T. Cai and H. Qiu. Application of deep eutectic solvents in chromatography: A review.
TrAC, Trends Anal. Chem. 2019, 120, 115623. DOI:10.1016/j.trac.2019.115623.
64. J. Huang, X. Guo, T. Xu, L. Fan, X. Zhou and S. Wu. Ionic deep eutectic solvents for the
extraction and separation of natural products. J. Chromatogr. A 2019, 1598, 1-19.
DOI:10.1016/j.chroma.2019.03.046.
65. I. Pacheco-Fernández and V. Pino. Green solvents in analytical chemistry. Curr. Opin.
Green Sustain. Chem. 2019, 18, 42-50. DOI:10.1016/j.cogsc.2018.12.010.
66. F. Bezold, M. E. Weinberger and M. Minceva. Computational solvent system screening
for the separation of tocopherols with centrifugal partition chromatography using deep
eutectic solvent-based biphasic systems. J. Chromatogr. A 2017, 1491, 153-158.
DOI:10.1016/j.chroma.2017.02.059.
67. S. Roehrer, F. Bezold, E. M. Garcia and M. Minceva. Deep eutectic solvents in
countercurrent and centrifugal partition chromatography. J. Chromatogr. A 2016, 1434,
102-110. DOI: 10.1016/j.chroma.2016.01.024.
68. N. R. Rodríguez, A. S. B. González, P. M. A. Tijssen and M. C. Kroon. Low transition
temperature mixtures (LTTMs) as novel entrainers in extractive distillation. Fluid Phase
Equilib. 2015, 385, 72-78. DOI:10.1016/j.fluid.2014.10.044.
69. F. S. Oliveira, A. B. Pereiro, L. P. N. Rebelo and I. M. Marrucho. Deep eutectic solvents
as extraction media for azeotropic mixtures. Green Chem. 2013, 15 (5), 1326-1330. DOI:
10.1039/C3GC37030E.
70. L. Manyoni, B. Kabane and G. G. Redhi. Deep eutectic solvent as a possible entrainer for
industrial separation problems: Pre-screening tool for solvent selection. Fluid Phase
Equilib. 2022, 553, 113266. DOI:10.1016/j.fluid.2021.113266.
71. F. Bai, C. Hua, Y. Bai and M. Ma. Design Optimization of Deep Eutectic Solvent
Composition and Separation Performance of Cyclohexane and Benzene Mixtures with
Extractive Distillation. Processes 2021, 9 (10). DOI: 10.3390/pr9101706.
72. V. Hamilton, I. Andrusenko, J. Potticary, C. Hall, R. Stenner, E. Mugnaioli, A. E. Lanza, M.
Gemmi and S. R. Hall. Racemic Conglomerate Formation via Crystallization of
Metaxalone from Volatile Deep Eutectic Solvents. Cryst. Growth Des. 2020, 20 (7), 4731-
4739. DOI: 10.1021/acs.cgd.0c00497.
73. C. L. Hall, J. Potticary, V. Hamilton, S. Gaisford, A. Buanz and S. R. Hall. Metastable
crystalline phase formation in deep eutectic systems revealed by simultaneous

184
References

synchrotron XRD and DSC. Chem. Commun. 2020, 56 (73), 10726-10729. DOI:
10.1039/d0cc04696e.
74. F. Wolbert, C. Brandenbusch and G. Sadowski. Selecting Excipients Forming
Therapeutic Deep Eutectic Systems-A Mechanistic Approach. Mol. Pharm. 2019, 16 (7),
3091-3099. DOI: 10.1021/acs.molpharmaceut.9b00336.
75. A. Gutiérrez, S. Aparicio and M. Atilhan. Design of arginine-based therapeutic deep
eutectic solvents as drug solubilization vehicles for active pharmaceutical ingredients.
PCCP 2019, 21 (20), 10621-10634. DOI: 10.1039/C9CP01408J.
76. M. S. Álvarez and Y. Zhang. Sketching neoteric solvents for boosting drugs
bioavailability. J. Controlled Release 2019, 311-312, 225-232.
DOI:10.1016/j.jconrel.2019.09.008.
77. A. R. C. Duarte, A. S. D. Ferreira, S. Barreiros, E. Cabrita, R. L. Reis and A. Paiva. A
comparison between pure active pharmaceutical ingredients and therapeutic deep
eutectic solvents: Solubility and permeability studies. Eur. J. Pharm. Sci. 2017, 114, 296-
304. DOI:10.1016/j.ejpb.2017.02.003.
78. I. M. Aroso, J. C. Silva, F. Mano, A. S. D. Ferreira, M. Dionísio, I. Sá-Nogueira, S. Barreiros,
R. L. Reis, A. Paiva and A. R. C. Duarte. Dissolution enhancement of active
pharmaceutical ingredients by therapeutic deep eutectic systems. Eur. J. Pharm. Sci.
2016, 98, 57-66. DOI:10.1016/j.ejpb.2015.11.002.
79. I. M. Aroso, R. Craveiro, Â. Rocha, M. Dionísio, S. Barreiros, R. L. Reis, A. Paiva and A.
R. C. Duarte. Design of controlled release systems for THEDES—Therapeutic deep
eutectic solvents, using supercritical fluid technology. Int. J. Pharm. 2015, 492 (1–2), 73-
79. DOI:10.1016/j.ijpharm.2015.06.038.
80. H. Gamsjäger, W. Lorimer John, P. Scharlin and G. Shaw David, Glossary of terms related
to solubility (IUPAC Recommendations 2008), Pure Appl. Chem. 2008, 80 (2), 233-276.
DOI:10.1351/pac200880020233
81. J. Gmehling, M. Kleiber, B. Kolbe and J. Rarey, Chemical Thermodynamics for Process
Simulation, Wiley, 2019.
82. A. I. Kitaigorodskii, Mixed crystals, Springer-Verlag, Berlin, 1984.
83. L. Rycerz. Practical remarks concerning phase diagrams determination on the basis of
differential scanning calorimetry measurements. J. Therm. Anal. Calorim. 2013, 113 (1),
231-238. DOI: 10.1007/s10973-013-3097-0.

185
References

84. A. van den Bruinhorst, L. J. B. M. Kollau, M. C. Kroon, J. Meuldijk, R. Tuinier and A. C.

C. Esteves. A centrifuge method to determine the solid–liquid phase behavior of eutectic
mixtures. J. Chem. Phys. 2018, 149 (22), 224505. DOI: 10.1063/1.5051515.
85. L. P. Silva, M. A. R. Martins, J. H. F. Conceição, S. P. Pinho and J. A. P. Coutinho. Eutectic
Mixtures Based on Polyalcohols as Sustainable Solvents: Screening and
Characterization. ACS Sustain. Chem. Eng. 2020, 8 (40), 15317-15326. DOI:
10.1021/acssuschemeng.0c05518.
86. L. P. Silva, C. F. Araújo, D. O. Abranches, M. Melle-Franco, M. A. R. Martins, M. M.
Nolasco, P. J. A. Ribeiro-Claro, S. P. Pinho and J. A. P. Coutinho. What a difference a
methyl group makes – probing choline–urea molecular interactions through urea structure
modification. PCCP 2019, 21 (33), 18278-18289. DOI: 10.1039/C9CP03552D.
87. L. P. Silva, L. Fernandez, J. H. F. Conceição, M. A. R. Martins, A. Sosa, J. Ortega, S. P.
Pinho and J. A. P. Coutinho. Design and Characterization of Sugar-Based Deep Eutectic
Solvents Using Conductor-like Screening Model for Real Solvents. ACS Sustain. Chem.
Eng. 2018, 6 (8), 10724-10734. DOI: 10.1021/acssuschemeng.8b02042.
88. E. A. Crespo, L. P. Silva, M. A. R. Martins, M. Bülow, O. Ferreira, G. Sadowski, C. Held,
S. P. Pinho and J. A. P. Coutinho. The Role of Polyfunctionality in the Formation of [Ch]Cl-
Carboxylic Acid-Based Deep Eutectic Solvents. Ind. Eng. Chem. Res. 2018, 57 (32),
11195-11209. DOI: 10.1021/acs.iecr.8b01249.
89. E. A. Crespo, L. P. Silva, M. A. R. Martins, L. Fernandez, J. Ortega, O. Ferreira, G.
Sadowski, C. Held, S. P. Pinho and J. A. P. Coutinho. Characterization and Modeling of
the Liquid Phase of Deep Eutectic Solvents Based on Fatty Acids/Alcohols and Choline
Chloride. Ind. Eng. Chem. Res. 2017, 56 (42), 12192-12202. DOI:
10.1021/acs.iecr.7b02382.
90. L. Fernandez, L. P. Silva, M. A. R. Martins, O. Ferreira, J. Ortega, S. P. Pinho and J. A. P.
Coutinho. Indirect assessment of the fusion properties of choline chloride from solid-
liquid equilibria data. Fluid Phase Equilib. 2017, 448, 9-14. DOI:
10.1016/j.fluid.2017.03.015.
91. X. Meng, K. Ballerat-Busserolles, P. Husson and J.-M. Andanson. Impact of water on the
melting temperature of urea + choline chloride deep eutectic solvent. New J. Chem. 2016,
40 (5), 4492-4499. DOI: 10.1039/C5NJ02677F.

186
References

92. I. Chojnacka, L. Rycerz, M. Berkani and M. Gaune-Escard. Phase diagram and electrical
conductivity of the DyBr3–RbBr binary system. J. Therm. Anal. Calorim. 2012, 108 (2),
481-488. DOI: 10.1007/s10973-011-2017-4.
93. R. M. Saeed, J. Schlegel, C. Castano and R. Sawafta. Uncertainty of thermal
characterization of phase change material by differential scanning calorimetry analysis.
Int. J. Eng. Res. Technol 2016, 5 (1), 405-412.
94. Y. Corvis, P. Négrier, M. Lazerges, S. Massip, J.-M. Léger and P. Espeau. Lidocaine/l-
Menthol Binary System: Cocrystallization versus Solid-State Immiscibility. J. Phys. Chem.
B 2010, 114 (16), 5420-5426. DOI: 10.1021/jp101303j.
95. Y. Corvis, P. Négrier, S. Massip, J.-M. Leger and P. Espeau. Insights into the crystal
structure, polymorphism and thermal behavior of menthol optical isomers and racemates.
Cryst. Eng. Comm. 2012, 14 (20), 7055-7064. DOI: 10.1039/C2CE26025E.
96. A. van den Bruinhorst, L. J. B. M. Kollau, M. Vis, M. M. R. M. Hendrix, J. Meuldijk, R.
Tuinier and A. C. C. Esteves. From a eutectic mixture to a deep eutectic system via anion
selection: Glutaric acid + tetraethylammonium halides. J. Chem. Phys. 2021, 155 (1),
014502. DOI: 10.1063/5.0050533.
97. L. J. B. M. Kollau, R. Tuinier, J. Verhaak, J. den Doelder, I. A. W. Filot and M. Vis. Design
of Nonideal Eutectic Mixtures Based on Correlations with Molecular Properties. J. Phys.
Chem. B 2020, 124 (25), 5209-5219. DOI: 10.1021/acs.jpcb.0c01680.
98. L. J. B. M. Kollau, M. Vis, A. van den Bruinhorst, R. Tuinier and G. de With. Entropy
models for the description of the solid–liquid regime of deep eutectic solutions. J. Mol.
Liq. 2020, 302, 112155. DOI:10.1016/j.molliq.2019.112155.
99. M. Kick, P. Keil and A. König. Solid–liquid phase diagram of the two Ionic Liquids EMIMCl
and BMIMCl. Fluid Phase Equilib. 2013, 338, 172-178. DOI:10.1016/j.fluid.2012.11.007.
100. D. O. Abranches, M. A. R. Martins, L. P. Silva, N. Schaeffer, S. P. Pinho and J. A. P.
Coutinho. Phenolic hydrogen bond donors in the formation of non-ionic deep eutectic
solvents: the quest for type V DES. Chem. Commun. 2019, 55 (69), 10253-10256. DOI:
10.1039/c9cc04846d.
101. E. Gómez, N. Calvar, Á. Domínguez and E. A. Macedo. Thermal Analysis and Heat
Capacities of 1-Alkyl-3-methylimidazolium Ionic Liquids with NTf2–, TFO–, and DCA–
Anions. Ind. Eng. Chem. Res. 2013, 52 (5), 2103-2110. DOI: 10.1021/ie3012193.
102. J. M. Prausnitz, R. N. Lichtenthaler and E. G. d. Azevedo, Molecular Thermodynamics of
Fluid-Phase Equilibria, Prentice Hall PTR, Upper Saddle River, NJ, 1999.

187
References

103. A. Jain, G. Yang and S. H. Yalkowsky. Estimation of Melting Points of Organic

Compounds. Ind. Eng. Chem. Res. 2004, 43 (23), 7618-7621. DOI: 10.1021/ie049378m.
104. S. H. Yalkowsky. Carnelley's Rule and the Prediction of Melting Point. J. Pharm. Sci.
2014, 103 (9), 2629-2634. DOI: 10.1002/jps.24034.
105. B. Lian and S. H. Yalkowsky. Molecular Geometry and Melting Point Related Properties.
Ind. Eng. Chem. Res. 2012, 51 (51), 16750-16754. DOI: 10.1021/ie302574y.
106. J. S. Chickos, G. Nichols and P. Ruelle. The Estimation of Melting Points and Fusion
Enthalpies Using Experimental Solubilities, Estimated Total Phase Change Entropies, and
Mobile Order and Disorder Theory. J. Chem. Inf. Comput. Sci. 2002, 42 (2), 368-374. DOI:
10.1021/ci010341b.
107. J. S. Chickos, W. E. Acree Jr and J. F. Liebman. Estimating solid–liquid phase change
enthalpies and entropies. J. Phys. Chem. Ref. Data 1999, 28 (6), 1535-1673.
108. W. Acree and J. S. Chickos. Phase Transition Enthalpy Measurements of Organic and
Organometallic Compounds. Sublimation, Vaporization and Fusion Enthalpies From 1880
to 2015. Part 1. C1 C10. J. Phys. Chem. Ref. Data 2016, 45 (3), 033101. DOI:
10.1063/1.4948363.
109. A. Bondi. Physical properties of molecular crystals, liquids, and glasses. Wiley, 1968.
110. S. H. Yalkowsky. Estimation of Entropies of Fusion of Organic Compounds. Ind. Eng.
Chem. Fund. 1979, 18 (2), 108-111. DOI: 10.1021/i160070a003.
111. R.-M. Dannenfelser and S. H. Yalkowsky. Estimation of Entropy of Melting from Molecular
Structure: A Non-Group Contribution Method. Ind. Eng. Chem. Res. 1996, 35 (4), 1483-
1486. DOI: 10.1021/ie940581z.
112. A. Jain, G. Yang and S. H. Yalkowsky. Estimation of Total Entropy of Melting of Organic
Compounds. Ind. Eng. Chem. Res. 2004, 43 (15), 4376-4379. DOI: 10.1021/ie0497745.
113. P. Walden. Über die Schmelzwärme, spezifische Kohäsion und Molekulargrösse bei der
Schmelztemperatur. Zeitschrift für Elektrochemie und angewandte physikalische Chemie
1908, 14 (43), 713-724. DOI: 10.1002/bbpc.19080144302.
114. B. Lian and S. H. Yalkowsky. Molecular geometry and boiling related thermodynamic
properties. J. Chem. Thermodyn. 2012, 54, 250-253. DOI:10.1016/j.jct.2012.04.009.
115. D. S. Mishra and S. H. Yalkowsky. Ideal Solubility of a Solid Solute: Effect of Heat
Capacity Assumptions. Pharm. Res. 1992, 9 (7), 958-959. DOI:
10.1023/A:1015873521067.

188
References

116. D. J. W. Grant, M. Mehdizadeh, A. H. L. Chow and J. E. Fairbrother. Non-linear van't Hoff

solubility-temperature plots and their pharmaceutical interpretation. Int. J. Pharm. 1984,
18 (1), 25-38. DOI:10.1016/0378-5173(84)90104-2.
117. H. T. Do, Y. Z. Chua, J. Habicht, M. Klinksiek, M. Hallermann, D. Zaitsau, C. Schick and
C. Held. Melting properties of peptides and their solubility in water. Part 1: dipeptides
based on glycine or alanine. RSC Advances 2019, 9 (56), 32722-32734. DOI:
10.1039/C9RA05730G.
118. S. H. Yalkowsky and M. Wu. Estimation of the ideal solubility (crystal–liquid fugacity ratio)
of organic compounds. J. Pharm. Sci. 2010, 99 (3), 1100-1106. DOI:10.1002/jps.21897.
119. G. D. Pappa, E. C. Voutsas, K. Magoulas and D. P. Tassios. Estimation of the differential
molar heat capacities of organic compounds at their melting point. Ind. Eng. Chem. Res.
2005, 44 (10), 3799-3806. DOI: 10.1021/ie048916s.
120. J. Hildebrand, J. Prausnitz and R. Scott. Regular and related solutions: the solubility of
gases. Liquids, and Solids, Van Nostrand Reinhold Co., New York, 1970.
121. M. Wu and S. Yalkowsky. Estimation of the molar heat capacity change on melting of
organic compounds. Ind. Eng. Chem. Res. 2009, 48 (2), 1063-1066. DOI:
10.1021/ie801587c.
122. S. H. Neau, S. V. Bhandarkar and E. W. Hellmuth. Differential molar heat capacities to
test ideal solubility estimations. Pharm. Res. 1997, 14 (5), 601-605. DOI:
10.1023/A:1012148910975.
123. R. Shaw. Heat capacities of liquids. Estimation of heat capacity at constant pressure and
25.deg., using additivity rules. J. Chem. Eng. Data. 1969, 14 (4), 461-465. DOI:
10.1021/je60043a036.
124. V. Růžička Jr and E. S. Domalski. Estimation of the heat capacities of organic liquids as
a function of temperature using group additivity. I. Hydrocarbon compounds. J. Phys.
Chem. Ref. Data 1993, 22 (3), 597-618. DOI: 10.1063/1.555923.
125. Z. Kolská, J. Kukal, M. Zábranský and V. Růžicka. Estimation of the heat capacity of
organic liquids as a function of temperature by a three-level group contribution method.
Ind. Eng. Chem. Res. 2008, 47 (6), 2075-2085. DOI: 10.1021/ie071228z.
126. L. J. B. M. Kollau, M. Vis, A. van den Bruinhorst, G. de With and R. Tuinier. Activity
modelling of the solid–liquid equilibrium of deep eutectic solvents. Pure Appl. Chem.
2019, 91 (8), 1341-1349. DOI:10.1515/pac-2018-1014.

189
References

127. G. M. Wilson. Vapor-liquid equilibrium. XI. A new expression for the excess free energy
of mixing. J. Am. Chem. Soc. 1964, 86 (2), 127-130.
128. H. Renon and J. M. Prausnitz. Local compositions in thermodynamic excess functions
for liquid mixtures. AlChE J. 1968, 14 (1), 135-144. DOI:10.1002/aic.690140124.
129. A. Klamt. The COSMO and COSMO-RS solvation models. WIREs Comput. Mol. Sci.
2011, 1 (5), 699-709. DOI: 10.1002/wcms.56.
130. F. Eckert and A. Klamt. Fast solvent screening via quantum chemistry: COSMO-RS
approach. AlChE J. 2004, 48 (2), 369-385. DOI: 10.1002/aic.690480220.
131. F. Eckert and A. Klamt. Fast solvent screening via quantum chemistry: COSMO-RS
approach. AlChE J. 2002, 48 (2), 369-385. DOI: 10.1002/aic.690480220.
132. A. Klamt, F. Eckert and M. Hornig. COSMO-RS: A novel view to physiological solvation
and partition questions. J. Comput. Aided Mol. Des. 2001, 15 (4), 355-365. DOI:
10.1023/A:1011111506388.
133. M. M. Abdallah, S. Müller, A. González de Castilla, P. Gurikov, A. A. Matias, M. D. Bronze
and N. Fernández. Physicochemical Characterization and Simulation of the Solid–Liquid
Equilibrium Phase Diagram of Terpene-Based Eutectic Solvent Systems. Molecules
2021, 26 (6). DOI: 10.3390/molecules26061801.
134. Z. Song, J. Wang and K. Sundmacher. Evaluation of COSMO-RS for solid–liquid equilibria
prediction of binary eutectic solvent systems. Green Energy Environ. 2021, 6 (3), 371-
379. DOI:10.1016/j.gee.2020.11.020.
135. D. O. Abranches, M. Larriba, L. P. Silva, M. Melle-Franco, J. F. Palomar, S. P. Pinho and
J. A. P. Coutinho. Using COSMO-RS to design choline chloride pharmaceutical eutectic
solvents. Fluid Phase Equilib. 2019, 497, 71-78. DOI: 10.1016/j.fluid.2019.06.005.
136. S. M. Vilas-Boas, D. O. Abranches, E. A. Crespo, O. Ferreira, J. A. P. Coutinho and S. P.
Pinho. Experimental solubility and density studies on aqueous solutions of quaternary
ammonium halides, and thermodynamic modelling for melting enthalpy estimations. J.
Mol. Liq. 2020, 300, 112281. DOI:10.1016/j.molliq.2019.112281.
137. N. R. Rodriguez, T. Gerlach, D. Scheepers, M. C. Kroon and I. Smirnova. Experimental
determination of the LLE data of systems consisting of {hexane + benzene + deep
eutectic solvent} and prediction using the Conductor-like Screening Model for Real
Solvents. J. Chem. Thermodyn. 2017, 104, 128-137. DOI:10.1016/j.jct.2016.09.021.

190
References

138. A. S. L. Gouveia, F. S. Oliveira, K. A. Kurnia and I. M. Marrucho. Deep Eutectic Solvents

as Azeotrope Breakers: Liquid–Liquid Extraction and COSMO-RS Prediction. ACS
Sustain. Chem. Eng. 2016, 4 (10), 5640-5650. DOI: 10.1021/acssuschemeng.6b01542.
139. F. Bezold, M. E. Weinberger and M. Minceva. Assessing solute partitioning in deep
eutectic solvent-based biphasic systems using the predictive thermodynamic model
COSMO-RS. Fluid Phase Equilib. 2017, 437, 23-33. DOI:10.1016/j.fluid.2017.01.001.
140. X. Liu, Y. Nie, Y. Liu, S. Zhang and A. L. Skov. Screening of Ionic Liquids for Keratin
Dissolution by Means of COSMO-RS and Experimental Verification. ACS Sustain. Chem.
Eng. 2018, 6 (12), 17314-17322. DOI: 10.1021/acssuschemeng.8b04830.
141. H. Qin, Z. Wang, T. Zhou and Z. Song. Comprehensive Evaluation of COSMO-RS for
Predicting Ternary and Binary Ionic Liquid-Containing Vapor–Liquid Equilibria. Ind. Eng.
Chem. Res. 2021, 60 (48), 17761-17777. DOI: 10.1021/acs.iecr.1c03940.
142. T. Aissaoui, I. M. AlNashef and Y. Benguerba. Dehydration of natural gas using choline
chloride based deep eutectic solvents: COSMO-RS prediction. J. Nat. Gas Sci. Eng.
2016, 30, 571-577. DOI: 10.1016/j.jngse.2016.02.007.
143. M. A. R. Martins, L. P. Silva, N. Schaeffer, D. O. Abranches, G. J. Maximo, S. P. Pinho
and J. A. P. Coutinho. Greener Terpene–Terpene Eutectic Mixtures as Hydrophobic
Solvents. ACS Sustain. Chem. Eng. 2019, 7 (20), 17414-17423. DOI:
10.1021/acssuschemeng.9b04614.
144. C.-C. Huang and Y.-P. Chen. Measurements and model prediction of the solid–liquid
equilibria of organic binary mixtures. Chem. Eng. Sci. 2000, 55 (16), 3175-3185. DOI:
10.1016/S0009-2509(99)00593-X.
145. J. A. P. Coutinho. Predictive local composition models: NRTL and UNIQUAC and their
application to model solid–liquid equilibrium of n-alkanes. Fluid Phase Equilib. 1999, 158-
160, 447-457. DOI:10.1016/S0378-3812(99)00085-0.
146. J. A. P. Coutinho. Predictive UNIQUAC: A New Model for the Description of Multiphase
Solid Liquid Equilibria in Complex Hydrocarbon Mixtures. Ind. Eng. Chem. Res. 1998,
37 (12), 4870-4875. DOI: 10.1021/ie980340h.
147. J. A. P. Coutinho and E. H. Stenby. Predictive Local Composition Models for Solid/Liquid
Equilibrium in n-Alkane Systems: Wilson Equation for Multicomponent Systems. Ind.
Eng. Chem. Res. 1996, 35 (3), 918-925. DOI: 10.1021/ie950447u.

191
References

148. A. Alhadid, L. Mokrushina and M. Minceva. Influence of the Molecular Structure of

Constituents and Liquid Phase Non-Ideality on the Viscosity of Deep Eutectic Solvents.
Molecules 2021, 26 (14). DOI: 10.3390/molecules26144208.
149. R. Srinivasa Gopalan, P. Kumaradhas, G. U. Kulkarni and C. N. R. Rao. An experimental
charge density study of aliphatic dicarboxylic acids. J. Mol. Struct. 2000, 521 (1), 97-106.
DOI:10.1016/S0022-2860(99)00293-8.
150. R. N. Rogers and L. W. Ortiz. Thermal characterization of the plastic crystalline phase.
Thermochim. Acta 1972, 3 (5), 379-381. DOI:10.1016/0040-6031(72)87051-5.
151. J. Timmermans. Plastic crystals: A historical review. J. Phys. Chem. Solids 1961, 18 (1),
1-8. DOI:10.1016/0022-3697(61)90076-2.
152. R. J. C. Brown and R. F. C. Brown. Melting Point and Molecular Symmetry. J. Chem.
Educ. 2000, 77 (6), 724. DOI: 10.1021/ed077p724.
153. S. Das, A. Mondal and C. M. Reddy. Harnessing molecular rotations in plastic crystals: a
holistic view for crystal engineering of adaptive soft materials. Chem. Soc. Rev. 2020, 49
(24), 8878-8896. DOI: 10.1039/D0CS00475H.
154. P.-J. Alarco, Y. Abu-Lebdeh, A. Abouimrane and M. Armand. The plastic-crystalline
phase of succinonitrile as a universal matrix for solid-state ionic conductors. Nature
Materials 2004, 3 (7), 476-481. DOI: 10.1038/nmat1158.
155. F. Tumakaka, I. V. Prikhodko and G. Sadowski. Modeling of solid–liquid equilibria for
systems with solid-complex phase formation. Fluid Phase Equilib. 2007, 260 (1), 98-104.
DOI:10.1016/j.fluid.2007.05.028.
156. L. Lange and G. Sadowski. Polymorphs, Hydrates, Cocrystals, and Cocrystal Hydrates:
Thermodynamic Modeling of Theophylline Systems. Cryst. Growth Des. 2016, 16 (8),
4439-4449. DOI: 10.1021/acs.cgd.6b00554.
157. L. Lange, S. Heisel and G. Sadowski. Predicting the Solubility of Pharmaceutical
Cocrystals in Solvent/Anti-Solvent Mixtures. Molecules 2016, 21 (5). DOI:
10.3390/molecules21050593.
158. L. Lange and G. Sadowski. Thermodynamic Modeling for Efficient Cocrystal Formation.
Cryst. Growth Des. 2015, 15 (9), 4406-4416. DOI: 10.1021/acs.cgd.5b00735.
159. C. Loschen and A. Klamt. Cocrystal Ternary Phase Diagrams from Density Functional
Theory and Solvation Thermodynamics. Cryst. Growth Des. 2018, 18 (9), 5600-5608. DOI:
10.1021/acs.cgd.8b00923.

192
References

160. T. Sachsenhauser, S. Rehfeldt, A. Klamt, F. Eckert and H. Klein. Consideration of

dimerization for property prediction with COSMO-RS-DARE. Fluid Phase Equilib. 2014,
382, 89-99. DOI: 10.1016/j.fluid.2014.08.030.

193