International Cryogenics Monograph Series

Series Editors: Steven W. Van Sciver · Sangkwon Jeong

J.G. Weisend II Editor

Cryostat
Design
Case Studies, Principles and Engineering
International Cryogenics Monograph Series

Series editors
Steven W. Van Sciver, Florida State University, Tallahassee, FL, USA
Sangkwon Jeong, KAIST, Daejeon, Korea, Republic of (South Korea)
The International Eryogenics Monograph Series was established in the early 1960s
to present an opportunity for active researchers in various areas associated with
cryogenic engineering to cover their area of expertise by thoroughly covering its
past development and its present status. These high level reviews assist young
researchers to initiate research programs of their own in these key areas of
cryogenic engineering without an extensive search of literature.

More information about this series at http://www.springer.com/series/6086

J.G. Weisend II
Editor

Cryostat Design
Case Studies, Principles and Engineering

123
Editor
J.G. Weisend II
European Spallation Source (ESS), ERIC
Lund University
Lund
Sweden

ISSN 0538-7051 ISSN 2199-3084 (electronic)

International Cryogenics Monograph Series
ISBN 978-3-319-31148-7 ISBN 978-3-319-31150-0 (eBook)
DOI 10.1007/978-3-319-31150-0

Library of Congress Control Number: 2016940359

© Springer International Publishing Switzerland 2016

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG Switzerland
For Shari, Rachel, Alex
and Nick—As Always
Preface

Cryostats are technical systems that maintain equipment or cryogenic liquids at

cryogenic temperatures. As such, they are one of the fundamental building blocks
of cryogenic systems. Examples of cryostats include the magnet cryostats that
comprise the majority of the Large Hadron Collider (LHC) at CERN, spaceborne
cryostats containing sensors operating below 1 K, MRI cryostats found in most
large hospitals, and large cryogenic liquid storage vessels. Cryostats that contain
superconducting radio frequency cavities are frequently referred to as cryomodules,
while cryostats whose principal function is to store cryogenic fluids are also referred
to as dewars. Cryomodules and dewars are also covered in this work.
The proper design of cryostats requires the knowledge of many disciplines
including cryogenic properties of materials, heat transfer and thermal insulation,
instrumentation, safety, structures, and seals. One of the best ways to learn about
cryostat design is to study the design choices and resulting performance of previous
designs.
This book provides such a review. It starts with an introductory chapter on the
principles of cryostat design including practical data and equations. This chapter is
followed by a series of case studies on existing cryostats. The studies describe the
cryostat and the design choices made along with the resulting performance of the
cryostat. The cryostat examples used in the studies are chosen to cover the wide
range of cryostat applications and the authors of each case are leading experts in the
field, all of whom participated in the design of the cryostats being described.
Chapters 2 and 3 are case studies involving superconducting magnets for large
particle accelerators. Due to the large numbers of magnets required in these cases,
low heat leak, reliability, and cost are key requirements. Chapter 4 describes a one
of a kind spaceborne dewar system whose requirements are very different than that
of accelerator cryostats. Chapters 5 and 6 describe cryomodules that contain
superconducting RF cavities in particle accelerators. As will be seen there are two
broad families of these cryomodules (segmented and continuous) with different
design drivers and approaches. Taken together these chapters describe a total of six
different cryomodules and the evolution of cryomodule design from the 1980s to

vii
viii Preface

the 2010s can be seen. Chapter 7 presents special topics in cryostat design. These
topics are of particular importance for MRI magnet cryostats; which provide the
examples but are broadly valuable for all cryostats. A cryostat design for very low
(50 mK) temperatures is described in Chap. 8. In addition to the lower temperatures,
this cryostat has unique material requirements due to the need to keep the
radioactive background of the associated experiment as low as possible. Transfer
lines connect cryostats and are a type of cryostat themselves. Transfer line features,
an overview of major transfer line systems, and a detailed case study of a transfer
line are found in Chap. 9. The final chapter provides a summary by listing
guidelines for a successful cryostat design. Extensive references throughout provide
sources of further information.

Lund, Sweden J.G. Weisend II

June 2016
Acknowledgments

This book would not have been possible without the contributions from the chapter
authors, all of whom are involved in ongoing projects with very demanding
schedules. I wish to express my thanks to my co-authors for their hard work. Tom
Spicer, Cindy Zitter, and the team at Springer have been very supportive and
helpful in the development and production of this book. I am honored that this book
is being published as part of the long-running International Cryogenic Monograph
Series. Professor Steven Van Sciver, in addition to being one of the series editors,
was also my thesis advisor at Wisconsin and I would like to thank him for making
this book possible in many ways than one. My colleagues at the European
Spallation Source and Lund University have also been a great help.
As always, the help and support of my family has been the key to the completion
of this book. For Shari, Rachel, Alex, and Nick—my continuing gratitude.

Lund, Sweden J.G. Weisend II

June 2016

ix
Contents

1 Principles of Cryostat Design . . . . . . . . . . . . . . . . . . . . ........ 1

J.G. Weisend II
1.1 Cryostat Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Cryogenic Properties of Materials . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Thermal Contraction . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 Thermal Conductivity . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.3 Heat Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.4 Material Strength. . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3 Thermal Insulation and Heat Transfer . . . . . . . . . . . . . . . . . . . 11
1.3.1 Reducing Conduction Heat Transfer . . . . . . . . . . . . . 11
1.3.2 Reducing Convection Heat Transfer. . . . . . . . . . . . . . 11
1.3.3 Reducing Radiation Heat Transfer . . . . . . . . . . . . . . . 12
1.3.4 Other Insulation Approaches . . . . . . . . . . . . . . . . . . . 15
1.4 Structural Supports for Cryostats . . . . . . . . . . . . . . . . . . . . . . 17
1.4.1 Alignment Approaches. . . . . . . . . . . . . . . . . . . . . . . 18
1.4.2 Suspension of Components from a Room
Temperature Top Flange . . . . . . . . . . . . . . . . . . . . . 19
1.4.3 Space Frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.4.4 Support Posts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.4.5 Supports in Space Cryogenics . . . . . . . . . . . . . . . . . . 23
1.5 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.5.1 Temperature Measurement . . . . . . . . . . . . . . . . . . . . 24
1.5.2 Pressure Measurement . . . . . . . . . . . . . . . . . . . . . . . 27
1.5.3 Flow Measurement . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.5.4 Level Measurement . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.5.5 Installation, Wiring, Heat Sinking
and Feedthroughs . . . . . . . . . . . . . . . . . . ........ 28
1.5.6 Commercial Availability of
Instrumentation Systems. . . . . . . . . . . . . . ........ 30
1.5.7 Best Practices for Cryostat Instrumentation . ........ 31

xi
xii Contents

1.6 Seals and Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

1.7 Transfer Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.8 Safety. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.9 Thermoacoustic Oscillations . . . . . . . . . . . . . . . . . . . . . . . . . 40
1.10 Prototyping and Series Testing . . . . . . . . . . . . . . . . . . . . . . . 42
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2 SSC Collider Dipole Cryostat . . .......................... 47
Thomas H. Nicol
2.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.2 Vacuum Vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.3 Thermal Radiation Shields . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.4 Multilayer Insulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.5 Cryogenic Piping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.6 Suspension System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.7 Interconnect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.8 Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3 Twenty-Three Kilometres of Superfluid Helium
Cryostats for the Superconducting Magnets
of the Large Hadron Collider (LHC) . . . . . . . . .............. 67
Philippe Lebrun
3.1 The LHC and Its Cryogenic System . . . . . . .............. 67
3.2 Feasibility of a Large Distributed Superfluid
Helium System . . . . . . . . . . . . . . . . . . . . .............. 70
3.3 Prototype Cryostats and String Tests . . . . . .............. 75
3.4 Industrial Series Production, Installation
and Commissioning . . . . . . . . . . . . . . . . . .............. 83
3.5 Concluding Remarks . . . . . . . . . . . . . . . . .............. 91
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . .............. 92
4 The Superfluid Helium On-Orbit Transfer (SHOOT)
Flight Demonstration . . . . . . . . . . . . . . . . . . . . . . . . .......... 95
Michael DiPirro
4.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.2 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.2.1 Structural Requirements . . . . . . . . . . . . . . . . . . . . . 98
4.3 Dewar and Cryostat Details . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.3.1 Dewar Fabrication Details . . . . . . . . . . . . . . . . . . . . 99
4.3.2 Cryostat Details . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.4 Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.4.1 Development Notes . . . . . . . . . . . . . . . . . . . . . . . . 103
4.4.2 Phase Separation . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Contents xiii

4.4.3 Liquid Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . 105

4.4.4 Thermomechanical (Fountain Effect) Pumps . . . . . . . 107
4.4.5 Cryogenic Stepper-Motor Valves . . . . . . . . . . . . . . . 109
4.4.6 Cryogenic Relief Valves . . . . . . . . . . . . . . . . . . . . . 110
4.5 Safety. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.6 Working with SHOOT on the Ground . . . . . . . . . . . . . . . . . . 113
4.7 On-Orbit Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
4.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5 TESLA & ILC Cryomodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
T.J. Peterson and J.G. Weisend II
5.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.2 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.3 Functional Requirements Summary . . . . . . . . . . . . . . . . . . . . 120
5.4 Cryomodule Mechanical Design. . . . . . . . . . . . . . . . . . . . . . . 120
5.4.1 Cryomodule Major Components and Features . . . . . . 120
5.4.2 Cryomodule Weight . . . . . . . . . . . . . . . . . . . . . . . . 129
5.4.3 Major Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.5 Cryomodule Vacuum Design and Vacuum Vessel . . . . . . . . . . 129
5.6 Cryomodule Thermal Design and Helium Flow Design. . . . . . . 131
5.6.1 Major Thermal Design Features . . . . . . . . . . . . . . . . 131
5.6.2 Design for Large 2 K Heat Transport
and Helium Flow . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.6.3 Pressure Drop Analyses . . . . . . . . . . . . . . . . . . . . . 135
5.6.4 Typical TESLA-Style Cryomodule Maximum
Allowable Working Pressures . . . . . . . . . . . . . . . . . 136
5.6.5 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
5.6.6 Cryomodule Test Requirements . . . . . . . . . . . . . . . . 137
5.6.7 Pressure Stability at the 2 K Level . . . . . . . . . . . . . . 137
5.7 Cryomodule Helium Inventory. . . . . . . . . . . . . . . . . . . . . . . . 139
5.8 Early Results from the TESLA Cryomodules. . . . . . . . . . . . . . 139
5.9 Modifications for CW Operation in the LCLS-II Linac . . . . . . . 143
5.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6 Segmented SRF Cryomodules. . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
E. Daly, Thomas H. Nicol and J. Preble
6.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.2 C20 Cryomodule Design for CEBAF . . . . . . . . . . . . . . . . . . . 149
6.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.2.2 Modularity and Segmentation. . . . . . . . . . . . . . . . . . 149
6.2.3 Requirements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.2.4 Design Description and Choices . . . . . . . . . . . . . . . . 150
xiv Contents

6.2.5 Cryogenic System Interfaces . . . . . . . . . . . . . . . . . . 153

6.2.6 Vacuum Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.2.7 Heat Load Estimates . . . . . . . . . . . . . . . . . . . . . . . . 154
6.2.8 Cavity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.2.9 Cavity Pair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.2.10 Tuner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.2.11 Helium Vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.2.12 Input Coupler . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
6.2.13 HOM Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
6.2.14 Magnetic Shields—Inner and Outer . . . . . . . . . . . . . 157
6.2.15 Thermal Shield and Multilayer Insulation . . . . . . . . . 157
6.2.16 Vacuum Vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.2.17 Cryounit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.2.18 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
6.2.19 Final Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
6.2.20 Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
6.3 The Spallation Neutron Source (SNS) Cryomodule . . . . . . . . . 161
6.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.3.2 Cavity String . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
6.3.3 Cryomodule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
6.3.4 Cryomodule Heat Loads and Thermal Design . . . . . . 166
6.3.5 Thermal Performance of the SNS Cryomodule . . . . . . 173
6.4 The CEBAF C100 Energy Upgrade Cryomodule . . . . . . . . . . . 178
6.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
6.4.2 Lessons Learned from C20 Experience . . . . . . . . . . . 178
6.4.3 Cavity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
6.4.4 Cavity Frequency Tuner . . . . . . . . . . . . . . . . . . . . . 181
6.4.5 Cold Mass and Space frame . . . . . . . . . . . . . . . . . . 181
6.4.6 Vacuum Vessel . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
6.4.7 End Cans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
6.4.8 C100 Performance . . . . . . . . . . . . . . . . . . . . . . . . . 185
6.5 SSR1 Cryomodule Design for PXIE. . . . . . . . . . . . . . . . . . . . 185
6.5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
6.5.2 Cryomodule Design . . . . . . . . . . . . . . . . . . . . . . . . 186
6.5.3 Final Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
6.5.4 Status and Plans . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
7 Special Topics in Cryostat Design . . . . . . . . . . . . . . . . . . . . . . . . . 195
Wolfgang Stautner
7.1 Boil off Minimization for Cryostats Without a Cryocooler. . . . . 195
7.1.1 Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
7.1.2 Pitfalls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Contents xv

7.2 Cryocooler Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

7.2.1 Cryocooler Integration—Options Overview . . . . . . . . 201
7.2.2 Cryocooler Integration Examples . . . . . . . . . . . . . . . 202
7.2.3 Schematics and Options of Cryocooler
Integration—Overview . . . . . . . . . . . . . . . . . . . . . . 203
7.2.4 Cryocooler Integration Techniques
for Special Applications . . . . . . . . . . . . . . . . . . . . . 207
7.3 Designing with Inclined Tubes in Cryogenic Systems . . . . . . . . 208
7.3.1 Pitfalls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
7.4 Cryogenics for Cryostats: Pressure Rise . . . . . . . . . . . . . . . . . 212
7.4.1 Quench Pressure Rise in Cryostats and Quench
Duct Sizing—A Modeling Example . . . . . . . . . . . . . 213
7.5 Advanced Cryostat Cryogenics—Carbon
Footprint Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
8 Design and Operation of a Large, Low Background,
50 mK Cryostat for the Cryogenic Dark Matter Search. . . . . . . . . 219
Richard L. Schmitt
8.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
8.2 Physics Detectors and Towers . . . . . . . . . . . . . . . . . . . . . . . . 219
8.3 Cryogenic System General Description . . . . . . . . . . . . . . . . . . 221
8.4 Dilution Refrigerator Introduction . . . . . . . . . . . . . . . . . . . . . 222
8.5 Icebox General Description . . . . . . . . . . . . . . . . . . . . . . . . . . 222
8.5.1 Icebox Cans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
8.5.2 Suspension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
8.5.3 C-Stems and Tails . . . . . . . . . . . . . . . . . . . . . . . . . 225
8.6 E-Stem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
8.6.1 Thermal Contraction . . . . . . . . . . . . . . . . . . . . . . . . 226
8.6.2 Materials, Radiopurity. . . . . . . . . . . . . . . . . . . . . . . 227
8.6.3 Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
8.6.4 Underground Assembly. . . . . . . . . . . . . . . . . . . . . . 228
8.7 Thermal Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
8.7.1 Thermal Conductivity . . . . . . . . . . . . . . . . . . . . . . . 229
8.7.2 Joint Conductance . . . . . . . . . . . . . . . . . . . . . . . . . 230
8.8 Heat Load. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
8.9 Detector Signal Feedthrough . . . . . . . . . . . . . . . . . . . . . . . . . 231
8.10 Dilution Refrigerator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
8.11 Liquid Transfer Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
8.12 Liquefier Addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
8.13 External Cold Trap. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
8.14 E-Stem Cryocooler. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
8.15 Insulating Vacuum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
xvi Contents

8.16 Automation and Control . . . . . . . . . . ... ...... . . . . . . . . . 236

8.17 Cryogenic Operation . . . . . . . . . . . . ... ...... . . . . . . . . . 237
8.18 Lessons Learned . . . . . . . . . . . . . . . ... ...... . . . . . . . . . 237
8.18.1 Cryogenic System Assembly and Testing . . . . . . . . . 237
8.18.2 Wiring . . . . . . . . . . . . . . . ... ...... . . . . . . . . . 237
8.18.3 Mixture Purification . . . . . . ... ...... . . . . . . . . . 237
8.18.4 Micro Vibrations . . . . . . . . ... ...... . . . . . . . . . 238
8.18.5 Can Supports . . . . . . . . . . . ... ...... . . . . . . . . . 238
8.18.6 Inner Vacuum, Yes or No? . ... ...... . . . . . . . . . 238
References . . . . . . . . . . . . . . . . . . . . . . . . ... ...... . . . . . . . . . 239
9 Cryogenic Transfer Lines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Jaroslaw Fydrych
9.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
9.2 Cryoline Routing and Modularization . . . . . . . . . . . . . . . . . . . 247
9.3 Cryoline Cross-Section Arrangements . . . . . . . . . . . . . . . . . . . 249
9.4 Supporting Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
9.5 Thermal Contraction Compensation . . . . . . . . . . . . . . . . . . . . 256
9.6 Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
9.7 Manufacturing and Installation. . . . . . . . . . . . . . . . . . . . . . . . 258
9.8 Case Study: XFEL/AMTF Cryogenic Transfer Line . . . . . . . . . 259
9.8.1 Technical Requirements . . . . . . . . . . . . . . . . . . . . . 260
9.8.2 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
9.8.3 Manufacturing the Cryoline Modules . . . . . . . . . . . . 266
9.8.4 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
9.8.5 Commissioning and Performance . . . . . . . . . . . . . . . 272
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
10 Guidelines for Successful Cryostat Design . . . . . . . . . . . . . . . . . . . 275
J.G. Weisend II
10.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
10.2 Guidelines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
10.3 A Final Comment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Contributors

E. Daly Thomas Jefferson National Laboratory, Newport News, VA, USA

Michael DiPirro Code 552, NASA Goddard Space Flight Center, Greenbelt, MD,
USA
Jaroslaw Fydrych European Spallation Source ERIC, Lund, Sweden
Philippe Lebrun CERN, European Organization for Nuclear Research, Geneva,
Switzerland
Thomas H. Nicol SRF Development Department, Technical Division, Fermi
National Accelerator Laboratory, Batavia, IL, USA
T.J. Peterson Cryomodule Engineering, SRF Development Department,
Technical Division, Fermi National Accelerator Laboratory, Batavia, IL, USA
J. Preble Thomas Jefferson National Laboratory, Newport News, VA, USA
Richard L. Schmitt Fermi National Accelerator Lab, Batavia, IL, USA
Wolfgang Stautner GE Global Research, Electromagnetics and
Superconductivity Lab, One Research Circle, Niskayuna, NY, USA
J.G. Weisend II European Spallation Source ERIC, Lund University, Lund,
Sweden

xvii
Chapter 1
Principles of Cryostat Design

J.G. Weisend II

Abstract This chapter provides an introduction to the engineering and design of

cryostats. It reviews cryostat requirements and provides detailed technical infor-
mation on topics relevant to cryostats. These topics include: materials, heat transfer
and thermal insulation systems, structural supports, safety, instrumentation, seals
and connections, transfer lines and thermoacoustic oscillations. The role of proto-
typing and series testing in cryostat development is also discussed. Numerous
tables, figures and equations provide useful information for the cryostat designer.

1.1 Cryostat Requirements

Successful engineering design depends upon a thorough understanding of the

requirements of the component being designed. Rushing into design before fully
understanding the requirements can lead to poorly optimized designs and a loss of
time and resources.
Cryostats typically have a set of common requirements, not all of which can be
simultaneously met to the same degree. Explicitly determining the requirements of
a cryostat and determining which ones are most important is the key first step to a
successful cryostat design. The following are requirements that should be consid-
ered in cryostat design.
1. Operating Temperatures and Allowable Heat Load
Since the purpose of a cryostat is to maintain systems at cryogenic temperatures, the
definition of the operating temperatures and the allowed heat load at those tem-
peratures are generally some of the first requirements defined. These requirements
drive a number of aspects of the resulting design including: material selection,
design of the thermal insulation system, cost, complexity and size. There are a

J.G. Weisend II (&)

European Spallation Source ERIC, Lund University,
P.O. Box 176, 22100 Lund, Sweden
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 1

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_1
2 J.G. Weisend II

number of subtleties in these requirements. An important distinction is that between

static heat leak and dynamic heat load. Static heat leak refers to heat entering the
cryostat from the outside environment. In many cryostats, this is the sole or prin-
cipal source of heat deposited at cryogenic temperatures. However, in other cryo-
stats there may be an additional heat load to cryogenic temperatures that stems from
the operation of the cryogenic equipment itself. Such heat loads typically vary with
time and are thus dynamic. Examples of such heat loads include heating of the
cavity walls in superconducting radiofrequency systems and heat deposited by
particle beams in accelerator or target cryostats. Dynamic heat loads can be the
dominant heat loads to the cryostat, which may affect the design choices made in
the cryostat’s thermal insulation system.
The temperatures at which heat is deposited in the cryostat are determined by a
number of factors. The required operating temperature of the cryogenic equipment
is one requirement but temperatures of intermediate shields or heat sinks may be
driven by the optimal performance of small cryocoolers or large cryogenic refrig-
eration plants.
Cryostats are frequently part of a larger cryogenic system. Allowable heat loads
and operating temperatures must be consistent with the larger system needs and
overall system optimization may be possible [1]. The allowable heat loads in a
cryostat may also be related to the number of cryostats in the system. A system with
a large number of cryostats, such as LHC or International Linear Collider
(ILC) may require a smaller heat load per cryostat to keep the overall cryogenic
system a reasonable size. Such systems put additional effort into reducing the heat
load of each cryostat (see case studies in the following chapters).
2. Alignment and Vibration
Depending on the application, some cryostats have requirements on the alignment
of the cryogenic equipment they contain or have limits on their allowable vibra-
tions. Issues to consider when responding to these requirements are the impact of
repeated thermal cycling on the alignment and the need, if any, to adjust the
alignment while the cryostat is cold or under vacuum Alignment requirements can
be quite tight. In the case of the TESLA cryomodules [2] the magnets had to
maintain an alignment to ±0.25 mm relative to an ideal reference.
3. Safety
Safety requirements including maximum credible accidents should be determined at
the start of the design and any design implications factored in from the beginning.
Retrofitting a design to meet safety requirements can be very expensive and time
consuming. In particular, issues due to overpressure and venting as a result of
vacuum or pipe failures can require significant design effort [3].
4. Allowable Size and Weight
These are particularly important for space borne cryostats but are also relevant for
cryostats that must be transported.
1 Principles of Cryostat Design 3

5. Amount and Type of Required Instrumentation

This may well vary between prototype and mass produced cryostats.
6. Number of Feedthroughs for Instrumentation, Power, External Manipulators,
Optical Windows and Cryogenic Lines.
7. Ease of Access to Cryostat Components
8. Existence of Ionizing Radiation or Magnetic Fields
This will influence material and instrumentation choices. The performance of
superconducting RF cavities, for example, is strongly affected by local magnetic
fields. This results in a need for supplemental magnetic shielding in the cryostat as
well as a limit on magnetic material near the cavities.
9. Additional Material Requirements
Some cryostats have additional unique material requirements. For example the
Cryogenic Dark Matter Search cryostat (Chap. 8) requires the use of materials with
no radioactivity to avoid false signals. This severely constrains the material choices.
10. Existing Regulatory Code Requirements
These might include pressure vessels codes or codes associated with flammable
materials.
11. Expected Lifetime of the Cryostat
12. Number of Cryostats Required
Is this a single unique cryostat or will it be mass-produced? Different design
choices may be made in either case.
13. Cost and Schedule Limits
These should be known and considered at the beginning of the design.
All designs involve compromise and it’s extremely unlikely that all cryostat
requirements can be met equally. The trick is to properly prioritize the requirements
in order to create an optimal design. Figure 1.1 shows an example of this approach.
The figure shows the E158 liquid hydrogen target cryostat operated at the SLAC
National Accelerator Lab [4]. This cryostat contained a loop of circulating LH2
operating at 20 K. A high energy electron beam was scattered by the LH2 as part of
a fundamental physics experiment. The presence of the beam had two important
effects: the cryostat was exposed to a extremely high ionizing radiation environment
and the beam deposited roughly 900 W of heat into the 20 K LH2. This resulted in
two important design choices. First, materials such as kapton based multilayer
insulation and ceramic based electrical insulation that survive high radiation doses
were chosen for the cryostat. Second, given the high dynamic heat loads from the
beam, expensive techniques to reduce the static heat load, such as actively cooled
thermal radiation shields were not used.
4 J.G. Weisend II

Fig. 1.1 View of the upper portion of the SLAC E158 liquid hydrogen target cryostat [4]

1.2 Cryogenic Properties of Materials

Selection of the proper materials is an important aspect of cryostat design. Many

materials are unsuitable for use at cryogenic temperatures and the properties of
materials can change greatly between room temperature and cryogenic tempera-
tures. Failures in cryostats can frequently be traced to the use of inappropriate
materials or to not accounting for the change in material properties as a function of
temperature.
The following materials are suitable for use at cryogenic temperatures:
• Austenitic stainless steels e.g. 304, 304L, 316, 321
• Aluminum alloys e.g. 6061, 6063, 1100
• Copper e.g. OFHC, ETP and phosphorous deoxidized
• Brass
• Fiber reinforced plastics such as G-10 and G-11
• Teflon (depending on the application)
• Niobium and Titanium (frequently used in superconducting RF systems)
• Invar (Ni/Fe alloy) useful in making washers due to its lower coefficient of
expansion
• Indium (used as an O ring material)
• Kapton and Mylar (used in Multilayer Insulation and as electrical insulation
• Quartz (used in windows)
1 Principles of Cryostat Design 5

Some materials unsuitable for use at cryogenic temperatures include:

• Martensitic stainless steels—undergoes ductile to brittle transition when cooled
down.
• Cast Iron—also becomes brittle
• Carbon steels—also becomes brittle. Sometimes used in 300 K vacuum vessels
but care must be taken that breaks in cryogenic lines do not cause the vacuum
vessels to cool down and fail.
• Rubber and most plastics.
One should never use materials in cryostats unless their suitability for use at
cryogenic temperatures is verified and their behavior as a function of temperature is
well understood. Given the large, and frequently nonlinear, changes in material
properties as a function of temperature, simple extrapolation of room temperature
properties to cryogenic temperatures should never be done. References [5–9] are
good starting points for finding the cryogenic properties of materials. Searching in
the proceedings of the International Cryogenics Materials Conference series may
also be useful. There are also commercially available computer codes that produce
material properties at cryogenic temperature [10, 11]. Aspects of some key cryo-
genic properties of materials are given below.

1.2.1 Thermal Contraction

Almost all materials change size (most shrink) when cooled to cryogenic temper-
atures. This effect can be significant and can cause a number of issues in cryostats
including the development of interferences or gaps between adjacent parts upon
cooling, adverse impact on alignment and possible failure upon cool down in over
constrained components or wires.
Thermal contraction should always be considered in cryostat design. Guidelines
for design include:
• Allow components room to contract through the use of sliding joints, bellows or
only fixing at a single point
• Install sufficient slack into wire and cable assemblies to allow for contraction
• Overlap thermal shields and multilayer insulation to avoid gaps upon cool down
and contraction
• Analyze the effect of thermal contraction on system alignment and on the
possibility of interferences between developing between adjacent components
• Play particular attention to the impact of thermal contraction on different
materials connected to or adjacent to each other
Thermal contraction is generally described by the property of thermal expan-
sivity [12] which is the change in volume or length of a material with temperature.
This property, which itself is a function of temperature, is of limited use in cryostat
6 J.G. Weisend II

Table 1.1 Integrated thermal Material ΔL/L (300–100 K) ΔL/L (100–4 K)

contraction for practical
engineering materials Stainless steel 296 × 10−5 35 × 10−5
between 300 and 100 K and Copper 326 × 10−5 44 × 10−5
between 100 and 4 K Aluminum 415 × 10−5 47 × 10−5
Iron 198 × 10−5 18 × 10−5
Invar 40 × 10−5 –
Brass 340 × 10−5 57 × 10−5
Epoxy/fiberglass 279 × 10−5 47 × 10−5
Titanium 134 × 10−5 17 × 10−5

design. More useful is the integrated change in length of a material between two
end temperatures. This is shown in Table 1.1 for a number of practical engineering
materials.
In order to find the change in length per unit length from 300 to 4 K, simply add
the two columns together. Thus, on average, a 1 m length of stainless steel will
contract 331 × 10−5 m or 3.3 mm. Notice that Invar, a Iron–Nickel alloy contracts
very little as a function of temperature. Invar has been used in cryostats (see Chap. 5)
where very little length contraction is desired.

1.2.2 Thermal Conductivity

The thermal conductivity of materials, particularly metals, varies strongly with

temperature between 300 and 0 K. This variability must always be taken into
account in cryostat design. Figure 1.2 shows the thermal conductivity of a number
of materials as a function of temperature. In many cases, the analysis may be
simplified by the use of thermal conductivity integrals [8]. In this approach, the
conduction heat transfer in one dimension is given by:

Q ¼ Gðh1  h2 Þ ð1:1Þ

where Q is the total heat transferred by conduction and the Geometry Factor (G) is
given by:

1
G ¼ R x2 dx ð1:2Þ
x1 Að xÞ

Here x is the position along the conduction path and A is the cross sectional area.
Note that in the case of a component with length L and constant cross sectional area
A, Eq. 1.2 becomes simply A/L.
1 Principles of Cryostat Design 7

Fig. 1.2 Thermal conductivity as a function of temperature for a variety of engineering materials
[8]

The thermal conductivity integral θι is given by:

ZTi
hi ¼ k ðT ÞdT ð1:3Þ
0
8 J.G. Weisend II

Fig. 1.3 Thermal conductivity integrals of a variety of engineering materials [8]

where T is the temperature and k is the thermal conductivity. Knowing the

appropriate thermal conductivity integral for a given material and temperature
allows quick calculation of the heat transfer between known temperatures while still
taking into account the variation of thermal conductivity with temperature.
Figures 1.3 and 1.4 show thermal conductivity integrals for a range of materials.
1 Principles of Cryostat Design 9

Fig. 1.4 Thermal conductivity integrals of a variety of engineering materials—Courtesy of

Lakeshore Cryotronics

1.2.3 Heat Capacity

The heat capacity or specific heat of materials generally decreases as a function of

temperature. This has two principal effects in cryostats. First, for a constant rate of
cooling, the drop in temperature over time increases as the cryostat becomes colder.
This effect is easily seen in larger systems which cool down more rapidly at lower
temperatures. Second, at cryogenic temperatures, the lower heat capacity means that
small heat leaks can cause significant temperature rises. This can be particularly
important for system operating below 1 K. Figure 1.5 shows the specific heat as a
function of temperature for a number of engineering materials.
10 J.G. Weisend II

Fig. 1.5 Specific heat as a function of temperature for a variety of engineering materials [9]

1.2.4 Material Strength

Material strength properties such as yield strength and ultimate tensile strength also
change with temperature. In many metals, those that don’t become brittle, the strength
actually increases with decreasing temperatures. However in this case, most cryostats
are designed using the lower, more conservative values at 300 K. The reason for this
is two-fold. First, all cryostats start out at room temperature and second, due to
operational issues even a cold cryostat may return suddenly to room temperature.
There are some cases in which the higher values at lower temperatures may be
taken into account. In the case of Superconducting RF cavities made from niobium,
the increased strength of the niobium at cryogenic temperatures may be used when
analyzing the response of the cavity to accidents that only occur when the cavity is
at cryogenic temperatures. An example of such an accident is the loss of beam tube
vacuum and the subsequent condensation of air on the surface of the cavity. This
condensation deposits heat and rapidly boils away the helium surrounding the
cavity causing large pressure rises in the helium space. Such a scenario can’t occur
if the cavity is at room temperature and thus the cavity’s stronger, lower temper-
ature strength properties can be used in determining the cavity’s response to the
sudden pressure rise.
As always, the material strength properties should always be checked at the
temperatures of interest to the cryostat being designed.
1 Principles of Cryostat Design 11

1.3 Thermal Insulation and Heat Transfer

A primary function of a cryostat, of course, is to maintain the equipment inside at

cryogenic temperatures. This is accomplished by developing thermal insulation
systems that block heat from the outside environment from reaching the cryogenic
equipment. While a perfect insulation system doesn’t exist, properly designed
cryostats can reduce the heat leak to very low, even μW, levels. This is accom-
plished by interrupting the three mechanisms of heat transfer within the cryostat:
conduction, convention and radiation.

1.3.1 Reducing Conduction Heat Transfer

The Fourier heat conduction equation (Eq. 1.1) shows that the heat transferred by
conduction is proportional to the temperature difference, the thermal conductivity
and the cross sectional area of the conduction path and is inversely proportional to
the length of the conduction path. Thus, the conduction heat leak can be reduced by
using low thermal conductivity materials, reducing the cross sectional area of the
conduction path and increasing the length of the conduction path. As you read
through the case studies later in this book, you will notice again and again, the
presence of long, thin conduction paths made from low conductivity material.
Another way to reduce the conduction heat leak to the lowest temperature is to
intercept the heat at an intermediate temperature. This reduces the temperature
difference and thus the heat to the lowest temperature point. The heat conducted
from higher temperatures to the intermediate heat sink still has to be removed but it
is thermodynamically more efficient to remove the heat at a higher temperature.
Chapter 7 gives an example of the optimization of such heat sinks.
Figure 1.6 shows the support post first developed for the SSC magnet cryostat
and shown here for use on the TESLA (now ILC) cryomodule [13]. The design of
this post reduces the conduction heat leak via the principles described above. The
conduction path in the post is made from low conductivity G-10 cylinders with a
small cross sectional area. In addition, active heat sinks are installed at the 77 K and
4 K levels. As a result, the post transfers only 0.03 W of heat between room
temperature and 2 K over a distance of 140 mm while still being able to support a
load of 50 kN.

1.3.2 Reducing Convection Heat Transfer

Convection heat transfer is that caused by a moving fluid. In many cryostats, this
mode of heat transfer is eliminated by using a vacuum space to separate the cold
and warm portions of the cryostat. This removes any fluid that can transfer heat via
12 J.G. Weisend II

Fig. 1.6 The support post for the ILC cryomodule—Courtesy of Nicol [13]

convection. The use of such an isolation vacuum was a key feature of James
Dewar’s original liquid hydrogen cryostat.
In general, a pressure of 10−6 millibar or less is sufficient vacuum to eliminate all
convective heat transfer. The cold wall of the vacuum space acts as a cryopump for
residual gases. Cryostats, particularly those operating below 77 K may have iso-
lation vacuums of 10−8 millibar or less when cold. It is not good practice to depend
upon cryopumping only but rather to use mechanical pumps to reduce the pressure
to near 10−6 millibar prior to cooling the cryostat down. In some cases, the use of
isolation vacuum spaces is not possible or practical and other techniques such as
foam insulation are used; see Sect. 1.3.4.

1.3.3 Reducing Radiation Heat Transfer

All surfaces emit heat via electromagnetic radiation. This thermal radiation is fre-
quently the largest source of heat leak into a cryostat. There are a number of
techniques for reducing the radiation heat leak into a cryostat. These techniques are
driven by the applicable governing equation for idealized (infinite parallel plates,
surface emissivities equal and ≪1) radiation heat transfer [8]:
  
q¼ r TH4  TL4 ð1:4Þ
ð N þ 1Þ  2
1 Principles of Cryostat Design 13

where q is the heat transferred in W/m2, ε is the emissivity of the surfaces, σ is the
Stefan-Boltzmann constant (5.67 × 10−8 W/(K4m2)) and N is the number of
uncooled radiation shields between the high temperature surface (TH) and the low
temperature surface (TL). Equation 1.4 can also be shown to apply for the case of
long concentric cylinders where the gap between the cylinders is small compared to
the diameter of the inner cylinder [14]. Based on Eq. 1.4, the heat leak can be
reduced by decreasing the emissivity of the surface, adding uncooled radiation
shields between the high temperature and low temperature surfaces and reducing
the temperature of the high temperature surface.
The temperature of the high temperature surface can be reduced by inserting a
lower temperature actively cooled radiation shield between the room temperature
(300 K) surface and the lowest temperature surface. The effect of this can be
significant. If we use Eq. 1.4 assuming an emissivity of 0.2 and no uncooled shields
(N = 0), the heat leak between a 300 and 4.2 K surface is 46 W/m2. If the warmest
surface facing the 4.2 K surface is now 77 K then the heat leak is reduced to
0.2 W/m2. The use of actively cooled thermal radiation shields is very common and
is seen in most of the case studies that follow.
Actively cooled shields may cooled by cold vapor boiling off from a cryogenic
reservoir, cooling flows from a cryogenic refrigeration plant and conductive cooling
from a small cryocooler attached to the shield. In many laboratory dewars, a liquid
nitrogen bath serves as the thermal radiation shield for the inner liquid helium
space.
Figure 1.7 shows emissivities of typical material used in cryostats. When cal-
culating radiation heat leak take care not to underestimate the value of the emis-
sivity; allow for the possibility of tarnishing or oxidizing of the surface that might
result in a higher emissivity. Silver plating a surface or coating the surface with
highly reflective tape will help reduce the emissivity but can be costly and is
generally done only where the absolute minimum radiation heat leak is desired [15]
or where there is no room for multilayer insulation.
The reduction of thermal radiation heat leak by the use of uncooled shields to
block the line of sight between the cold and warm surfaces can be significant.
Multilayer insulation (MLI) also sometimes called superinsulation is the logical
extension of this approach. MLI consists of many layers of thin aluminized Mylar
or sometimes aluminized Kapton placed into the vacuum space between the warm
and cold surfaces. A design feature of MLI is to reduce or eliminate the thermal
conduction between the successive layers. This is typically done by placing a
nonconductive mesh or paper between the aluminized Mylar sheets or in some
cases by only aluminizing only one side of the Mylar and then crinkling the sheets
so there are only point contacts between them. Since MLI makes the term N in
Eq. 1.4 large, its impact on the thermal radiation heat leak is impressive. As such,
most of the case studies that follow employ MLI in their design. Figure 1.8 shows
an example of the MLI from the Large Hadron Collider (LHC) magnet cryostat
[16].
There are some cautions with the use of MLI. Care much be taken not to pack
the MLI too tightly in the vacuum space. Doing so causes the conduction heat load
14 J.G. Weisend II

Fig. 1.7 Emissivities of

typical cryostat materials [14]

Fig. 1.8 The MLI system for the large hadron collider (LHC) magnet cryostats. Reproduced with
permission from Poncet and Parma [16]. Copyright 2008, AIP Publishing LLC
1 Principles of Cryostat Design 15

Fig. 1.9 Examples of the proper installation of MLI for corners (a), steps (b), penetrations (c) and
overlaps (d) [18]

to increase, reducing the benefit of the MLI. This effect was seen in the SSC
Collider Dipole thermal measurement model (see Chap. 2). A good guideline is to
keep the density of the MLI in the vacuum space to around 0.5 layers per mm [14].
Another hazard with MLI is associated with seams, ends, penetrations and corners
[17]. If these aspects of the installation are not handled properly, their effect can
dominate the heat leak. The basic rule here is to avoid connecting the warmest outer
layer of the MLI with one of the colder inner layers. This would, in effect, short
circuit the MLI and reduces its effectiveness. It is also important to overlap the
layers of the MLI at seams so that when the MLI shrinks during cool down, a gap
does not open up that allows the warm surface to directly see the cold surface below
the MLI. Figure 1.9 illustrates the proper way to address these issues. Detailed
analysis and measurements have been made of specific MLI designs and additional
information may be found in Refs. [8, 18–20].

1.3.4 Other Insulation Approaches

The general approach of using vacuum insulated cryostats combined with actively
cooled thermal shields and MLI in the vacuum space is not used in all cryostats. In
some cases, such as cryogenic tanks for rocket propellants, the weight penalty
caused by the presence of a second wall is too high. In other systems, the cost or
complexity of vacuum spaces, MLI and thermal shields out weighs the benefit of a
16 J.G. Weisend II

reduced heat leak. This is frequently seen in large storage and transport tanks,
particularly for those operating at or above liquid nitrogen temperatures.
One alternative is to remove the vacuum space entirely and coat the cryogenic
surface with insulating foam, typically a closed cell foam. This approach was taken
with the cryogenic propellant tanks used in the Saturn 5, Space Shuttle and Ariane 5
launchers. This approach is also seen in vent lines and intermittently used liquid
nitrogen fill lines mainly to avoid condensing water or air on the cold line surfaces.
In the more high performance applications of cryogenic propellant tank insulation,
the foam systems can be quite sophisticated, employing water proof barriers and
helium gas purging to prevent condensation of water or air. Figure 1.10 shows the
foam system for the Saturn 5 s stage liquid hydrogen tank [21]. The overall heat
transfer from this insulation system (including conduction, convection and radia-
tion) was 0.86–1.1 mW/(cm K) at an average temperature of 144 K.
Alternative insulation approaches also include: using glass microspheres [22] or
Perlite powder in a annular space (either filled with inert gas or vacuum) sur-
rounding the cryogenic vessel. Opacified powder in which metallic flakes are mixed
with the insulating powder can provide higher performance due to its ability to
reflect back more of the thermal radiation. Recent advances in aerogel [23, 24] have
led to its use in thermal blankets for cryogenic systems.
A nice comparison for scaling purposes of various insulation systems is given in
Table 1.2 [25]. The heat flux in all cases includes conduction, convection and
radiation. Notice that between 300 and 77 K the use of Perlite powder in a vacuum
is more effective than just a high vacuum space. This results from the Perlite
intercepting some of the thermal radiation heat leak. References [21, 25] provide
detailed descriptions of many of these alternative insulation system.

Fig. 1.10 The foam

insulation system of the
Saturn 5 s stage LH2 tank
[21]
1 Principles of Cryostat Design 17

Table 1.2 Comparison of thermal insulation approaches (6 in. thick insulation in all cases)
Type of insulation Total heat flux (W/m2)
300–77 K 77–20 K
Polystyrene foam 48.3 5.6
Gas filled Perlite powder 184.3 21.8
(5–6 lb/ft3 filled with He)
Perlite powder in vacuum 1.6 0.07
(5–6 lb/ft3)
High vacuum 9 0.04
(10−6 torr, ε = 0.02)
Opaciated powder 0.3 –
(Cu flakes in Santocel)
MLI 0.03 0.007
Reference [25]

1.4 Structural Supports for Cryostats

Structural supports for cryostats serve many purposes. An example is the stand or
frame that connects the room temperature outer wall of the cryostat to the outside
environment (floor, ceiling etc.). Since these devices connect two room temperature
components, they are designed using standard structural engineering techniques.
Cryostat stands aren’t trivial, they may have to support very large loads or react
against strong forces (for example from unbalanced pressure loads or magnetic
fields) they may need to allow precision movement or alignment of the cryostats.
More interesting for this book are the supports that connect the cryogenic parts
of the cryostat with its room temperature outer wall. There are a number of chal-
lenges for these supports.
1. They must support the weight of the components plus any applied loads (e.g.
pressure, magnetic field, transport, space launch) while at the same time mini-
mizing the conduction heat leak (see Sect. 1.3.1).
2. They must allow for thermal shrinkage of both the components and the supports
during cool down.
3. If required, they must achieve the proper alignment of the cryogenic compo-
nents upon cool down.
4. The supports and the cryogenic components that are attached must be designed
to avoid any unwanted resonant frequencies.
The structural supports in the case studies that follow are designed to meet these
challenges.
This section first addresses the issue of alignment and then discusses various
design solutions for cryostat structures.
18 J.G. Weisend II

1.4.1 Alignment Approaches

Some cryostats have little to no alignment requirements. Examples of these include

many small laboratory systems, cold boxes for refrigeration plants, transport and
liquid storage Dewars. In these cases, the support system must allow the compo-
nents to shrink and move without causing damage. As mentioned in Sect. 1.2.1 not
just supports but all components such as pipes and wiring must be designed to
allow for thermal contraction. Over constraining a component by rigidly fixing it so
that it can not move may well lead to failure upon cool down.
Many other cryostats do have tight alignment requirements. In these cases, the
position of certain components must be in the correct location once at cryogenic
temperatures. Examples of these cryostats include: superconducting magnets,
superconducting cavities and other components used in particle accelerator beam
lines, telescopes and other optical systems and space based systems. As will be seen
in the case studies, the alignment requirements can be quite strict with allowable
tolerances as small as a fraction of a millimeter. There are three general approaches
to meeting alignment requirements.
Maintain the alignment upon cool down. In this approach, the structure is
designed so that the cool down is symmetric. Those components with tight align-
ment requirements are properly aligned while warm and keep this alignment when
cooled down. In some cases the alignment of the components changes but stays
within the allowable alignment tolerance. Examples of this approach include the
ESS cryomodules [26].
Allow the alignment to change in a predictable way. Here the components are
allowed to change position during cool down, but do so in a predictable way that is
repeated in each cool down. With these designs, the component is installed so that
its out of alignment while at room temperature but moves into alignment when
cooled. An example of this is the ILC cryomodule (Chap. 5). In this design, it was
calculated that the center of the cavity beam tube would shrink upwards by 1.8 mm
upon cool down. The cryomodules were aligned into the accelerator beam line so
that the center of the beam tube was 1.8 mm lower than the ideal beam line. Upon
cool down, the cavity beam tube was properly aligned with the rest of the system.
Allow realignment once cold. In this approach, components with tight alignment
tolerances are realigned into the proper position after they have reached their final
cryogenic temperature. This can be accomplished via screws or rods that can adjust
the position of the cryogenic components from room temperature or by the use of
remotely controlled positioners or motors within the cryostat. However, there are a
number of disadvantages with this choice. The screws, rods, motors or positioners
may add heat leak, may be expensive and may reduce system reliability. While
cryostats have been designed in this way, it’s generally best to use approach as the
last resort or as a back up to the other two approaches.
1 Principles of Cryostat Design 19

1.4.2 Suspension of Components from a Room

Temperature Top Flange

This is a very common choice and is used in many laboratory cryostats as well in
liquid storage Dewars and refrigerator cold boxes. An example of this approach is
shown in Fig. 1.11. All the components inside the innermost part of the cryostat are
suspended from the room temperature top flange and not fixed anywhere else. As

Fig. 1.11 Cutaway view of a

vertical test cryostat at the
SLAC National Acceleratory
Laboratory
20 J.G. Weisend II

the system is cooled down, the components are allowed to shrink upwards towards
the top flange. This design choice does not tend to guarantee the alignment of
components and functions best where alignment tolerances are very loose or
nonexistent. The suspension of the components also means that changes in the
cryostat’s orientation or motion during transport or operation may cause the com-
ponents to oscillate and become damaged. The use of stops to limit the motion or
shipping restraints may be required. This approach has the virtues of being simple,
low cost and may result in low conduction heat leaks to the cryogenic components.
Figure 1.12 shows this design applied to two different refrigerator cold boxes.
When applied to liquid storage Dewars, this design choice results in everything,
including the inner vessel and thermal shields, being suspended from the neck of
the Dewar. Again, stops or shipping restraints may be required.
This approach can also be taken with accelerator beam line components if the
alignment requirements aren’t too stringent. Figure 1.13 shows the inner compo-
nents of the ATLAS Upgrade cryomodule built at Argonne National Laboratory
[27]. The Superconducting cavities, magnets, shields and cryogenic lines are all
fixed to the upper flange. The movement of the components up as they cool down is
within the alignment tolerances.

1.4.3 Space Frames

In this approach, the cryogenic components are attached to a, typically cylindrical,

frame that is thermally and structurally attached to the inside of the room tem-
perature vacuum vessel of the cryostat. The frame itself is kept at or near to 300 K
and the cryogenic components are connected by long low conductivity metal rods
made from stainless steel, titanium etc.
Figures 1.14 and 1.15 show views of the design of the European Spallation
Source (ESS) Elliptical Cavity Cryomodule [26]. Note the use of the hanging rods
that attach the cavity string, thermal shields and cryogenic pipes to the space frame.
The components can all be assembled onto the space frame and then the entire
complete assembly can be inserted into the vacuum vessel, aligned properly and then
locked into place. The long cylindrical axis of such systems results in all the com-
ponents cooling symmetrically towards the center with the result that the cavities
maintain their original room temperature alignment at cryogenic temperatures. Space
frames are very popular in Superconducting Radiofrequency cryomodules and have
also been used in the Spallation Neutron Source and the Jefferson Lab 12 GeV
Upgrade projects among others (see Chap. 6).
1 Principles of Cryostat Design 21
 22 J.G. Weisend II

b Fig. 1.12 Views of two cryogenic refrigerator cold boxes showing components suspended from a
top plate. a Courtesy by Linde Kryotechnik AG and b the CTI-4000 cold box at the SLAC
National Accelerator Laboratory

Fig. 1.13 Components suspended from the upper plate of the ALTAS upgrade cryomodule [27]

Fig. 1.14 Cut away view of the ESS elliptical cavity cryomodule. Courtesy of Gilles Olivier
(CNRS-IPN Orsay)
1 Principles of Cryostat Design 23

Fig. 1.15 Cross sectional view of the ESS elliptical cavity cryomodule. Courtesy of Gilles Olivier
(CNRS-IPN Orsay)

1.4.4 Support Posts

Cryogenic components can also be connected to the 300 K vacuum vessel via
support posts. The components can either sit on top of the support posts or be
suspended from the support posts. The posts themselves are designed to minimize
the heat transfer between 300 K and cryogenic temperatures (see Sect. 1.3.1). An
advantage of this approach is that the posts can be designed to support very large
loads and are thus frequently used in magnet systems. Since the cryogenic com-
ponents are not supported in a symmetric manner the position of the components
will change during cool down and this effect must be allowed for in the design.
Examples of cryostats using support posts include: the LHC magnet cryostat
(Chap. 3), the ILC cryostat (Chap. 5) and the SSC magnet cryostat (Chap. 2).

1.4.5 Supports in Space Cryogenics

Space cryogenic systems have additional challenges in cryostat supports. Not only
must they minimize heat lead leak via the supports but the large forces that develop
during the rocket launch require (at least temporarily) strong supports. Creative
approaches have been developed to solve this problem.
Since the orientation of the launch loads is well known, one approach is to
optimize the design by making the supports which resist the launch loads stronger
while minimizing the strength of the supports in other orientations. This approach
was taken with the AXAF cryostat [15].
24 J.G. Weisend II

Another technique is the use of Passive Orbital Disconnect Struts (PODS). These
supports are designed in such a way that during orbit or while stationary on Earth,
the cryogenic components are supported on thin wall, low heat leak tubes.
However, during launch, the forces are taken up by much stronger, thick wall tubes
that are parallel to the thin wall tubes. Once the spacecraft is in orbit, the thick wall
tubes passively disconnect and the lower heat leak thin wall tubes again provide the
support. One advantage of this approach is that the use of the thin wall tubes while
on Earth reduces the heat load during any long launch delays and permits verifi-
cation of the expected heat load in space. These devices have gone through a
number of generations and have been extensively tested and studied [28]. They
were a key component of the Gravity Probe B cryostat [29].

1.5 Instrumentation

The correct measurement of properties such as temperature, pressure, flow, level

and vacuum in cryostats is a key factor in the success of the cryostats.
Measurements allow us to understand if our cryogenic components are working
properly, enable us to control them and permit the collection of scientific data.
There are many subtleties in the selection and installation of cryogenic instru-
mentation. Poor sensor selection and installation can result in wildly inaccurate
readings or sensor failure.
A frequent mistake is to fix one’s attention only on the sensor making the
measurement itself. By far, the better approach is to think of each measuring point
as a complete system including: the sensor, sensor calibration, wiring, feed through,
data acquisition hardware and software. The cost of such a complete system can
easily exceed $1000 per measurement point. Additional material on cryogenic
instrumentation may be found in [8, 30].

1.5.1 Temperature Measurement

Temperature in cryogenics is determined by measuring a physical quantity such as

voltage, resistance or pressure that changes in a known repeatable way with tem-
perature. A wide range of options for temperature measurement exists for cryo-
genics. The existence of reliable commercially available temperature sensors for use
in cryogenics has been a major advance in the field. The need to use home made
temperature sensors has almost been completely eliminated. The use of home made
temperature sensors should be avoided in lieu of commercial devices that are
repeatable, reliable, accurate and whose behavior is well understood. The perfor-
mance of temperature sensors depends on the operating temperature. Some devices,
such a silicon diodes, are responsive (though with varying sensitivity) from room
temperature all the way down to 1 K. Others, such as platinum resistors, become
1 Principles of Cryostat Design 25

Fig. 1.16 Dimensionless sensitivity of cryogenic temperature sensors [8]

essentially unresponsive below a given temperature (typically 30 K). Choosing the

correct temperature sensor for the temperature range of interest is key. In some
cases, a combination of sensors will be required. Figures 1.16 and 1.17 show the
sensitivity and temperature resolution of a wide range of temperature sensors.
Self Heating. An issue with temperature sensors, particularly those operating in a
vacuum space is that the excitation current applied to the sensor can cause the
sensor itself to heat up and thus provide a false measurement. This problem is
avoided by always following the manufacturer’s recommendations regarding
excitation.
Environmental Conditions. The performance of temperature sensors is greatly
influenced by the environment in which they operate. Some sensors such as, silicon
diodes or germanium resistors perform poorly in magnetic fields or ionizing radi-
ation environments while others such as carbon glass or Cernox sensors work well
26 J.G. Weisend II

Fig. 1.17 Relative

temperature resolution of
cryogenic temperature sensors
[8]

under these conditions. Care must be taken to choose the correct sensor for the
expected operating conditions. Table 1.3 shows the impact of magnetic and ion-
izing radiation on common cryogenic temperature sensors. More details may be
found in [8].
Calibration. Some types of temperature sensors, such as platinum resistors or
silicon diodes have general calibration curves that apply for all sensors of a given
type. Others such as cernox or germanium need to be individually calibrated for

Table 1.3 Impact of magnetic field and ionizing radiation on common cryogenic temperature
sensors
Sensor type Suitability for use in Suitability for use in ionizing
magnetic fields radiation
Silicon diode Poor Poor
Platinum resistor Poor Good
Germainium resistor Poor Poor
Carbon glass resistor Fair Good
Ruthenium oxide Good Good
Cernox Good Good
Thermocouple Good TBD
Cryogenic linear TBD Good
temperature sensors
Data from Ref. [8]
1 Principles of Cryostat Design 27

best accuracy. In addition, the calibrations of some sensors will drift over time and
manufacturers will have recommendations on the need for periodic recalibration of
their sensors. In many cases, the calibration drift is not significant, but in situations
requiring high accuracy, the cryostat has to be designed to either allow in situ
recalibration of the sensors or removal of the sensors for recalibration.
Redundancy. Even the best installed sensors can fail. Once installed inside of
cryostats, temperature sensors can be very difficult to remove. In situations where
proper temperature measurement is vital to the success of the cryostat, either
redundant sensors should be installed at each measuring point or the sensors should
be installed in such a way that they can be replaced. This can be accomplished by
placing the sensors on removable probes or by installing access ports near the
sensors to allow their replacement in the case of in accessible cryostats such as
space cryostats, redundant sensors are the answer.

1.5.2 Pressure Measurement

Pressure measurements in cryogenic systems are generally carried out by room

temperature pressure sensors. There are a variety of high precision room temper-
ature pressure sensors commercially available. Room temperature pressure trans-
ducers have the advantage of being easily replaceable. They are connected to the
measuring point via small capillary tubes. These tubes have some disadvantages.
First, if improperly designed, they can lead to thermoacoustic oscillations (see
Sect. 1.9) and experiments [31] have shown that the capillaries can damp out and
delay high speed pressure signals. In those situations where either space require-
ments or the need for accurate high speed pressure signals exist, there are some
pressure transducers that will operate at cryogenic temperatures. Another option is
to use non-cryogenic pressure transducers but locate them near the point of mea-
surement in an insulated and heated enclosure to maintain them at room tempera-
ture. An example of this approach is given in [32]. Many pressure transducers, both
warm and cryogenic, are susceptible to damage from ionizing radiation. Such
devices may need local shielding when operated in a radiation environment.

1.5.3 Flow Measurement

As in the case of pressure measurement, flow measurement in cryostat systems is

best carried out at 300 K. This increases the amount of flow meters available and
allows easier repair or replacement. In cases where flow measurement at cryogenic
temperatures is required, many typical flow metering approaches e.g. Venturi
meters, orifice plate meters, turbine flow meters etc. can be applied at cryogenic
temperatures. Care must be taken to ensure that the meters are properly calibrated
for the properties such as density, viscosity, pressure, temperature of the cryogenic
28 J.G. Weisend II

fluids. Additional installation requirements such as a certain amount of straight pipe

upstream and downstream of the meter must also be met. Many cryogenic flow
meters function by measuring a pressure difference and thus the comments on
pressure transducers also apply here. Additional information on cryogenic flow
metering can be found in [8] with some specific examples given in [33, 34].

1.5.4 Level Measurement

Measuring the level of a cryogenic liquid bath is important for proper control and
operation of cryogenic systems. There are a number of ways to accomplish this
measurement. In liquid helium systems, the use of superconducting level gauges is
common. In these devices a probe containing a superconductor is the placed into the
bath. The superconducting probe is designed so that the wire is superconducting
when in contact with liquid helium and normal conducting when in contact with the
corresponding helium vapor. The total voltage drop across the wire can then related
to the liquid level. Work [35, 36] has been done to extend this approach to warmer
fluids such as hydrogen, nitrogen and oxygen using high temperature supercon-
ductors; but alternative methods such as capacitive or pressure differential tech-
niques are more common for these fluids.
In the capacitive approach [37, 38], the differing dielectric constants between the
liquid and vapor phases of the fluid are used to determine the fluid level. These are
commercial devices and work with a number of cryogenic fluids. As they do not
require superconductors to operate they are also useful for fluids warmer than liquid
helium.
The differential pressure approach essentially measures the weight of the liquid
level of the fluid from which the level can be calculated. Again, these don’t require
superconductivity.
A final technique, that provides discrete rather than continuous liquid level
measurements is the use of Liquid Vapor Detectors (LVDs). These devices can
either be discrete superconductors that function similarly to superconducting liquid
level probes or more commonly temperature sensors that are powered so that they
self heat and provide a very different response in the vapor phase than the liquid
phase. Reference [39] provides an example of such devices. This technique was
tested in the SHOOT Project (see Chap. 4).

1.5.5 Installation, Wiring, Heat Sinking and Feedthroughs

Cryogenic instrumentation systems can easily give inaccurate results if improperly

installed. There are a number of issues to take into account.
Keep in mind that wires will generally contract significantly when cooled (see
Sect. 1.2.1). If the instrumentation wires are installed with no slack or loops
1 Principles of Cryostat Design 29

Table 1.4 Required heat sinking lengths for wire to achieve a measurement within 1 mK of the
actual value
Heat sinking length (mm)
Material Th[K] Tl[K] 0.21 mm2 0.032 mm2 0.031 mm2 0.005 mm2
(24 AWG) (32 AWG) (36 AWG) (40 AWG)
Copper 300 80 160 57 33 19
300 4 688 233 138 80
Phosphor-Bronze 300 80 32 11 6 4
300 4 38 13 7 4
Maganin 300 80 21 4 4 2
300 4 20 7 4 2
304 SS 300 80 17 6 3 2
300 4 14 5 3 2
Results assume that the wire and sensors are in vacuum and that the thermal conductivity of the
adhesive is given by the fit to the thermal conductivity of GE 7031 varnish [8]

between fixed points; upon cool down the wires will shrink and may break or come
loose from a solder connection. Always allow sufficient slack in wire installation to
allow for this shrinkage.
Instrumentation wires provide a thermal conduction path and thus heat leak
between room and cryogenic temperatures. Thus, the number and diameter of the
wires along with the type of material in the wires should be chosen to reduce this
heat leak. Experience with an early prototype of the ILC Cryomodule (see Chap. 5)
illustrates what can happen if this conduction path is not properly considered. Heat
sinking the wires at intermediate temperatures (again allowing for shrinkage) will
also reduce the heat leak.
An important aspect of heat sinking comes in when using cryogenic temperature
sensors in a vacuum space. This is a common installation in which temperature
sensors are attached to the outside of a pipe or vessel containing cryogenic fluid.
The sensor itself while attached to the cold surface sits inside a vacuum space. Heat
leak down the wires, even from an intermediate temperature heat sink can cause the
temperature sensor to read an erroneously high temperature despite itself being in
good thermal contact with the pipe or vessel. The solution is to always heat sink the
wire to the measurement point immediately adjacent to the sensor. Table 1.4 shows
the length of wire required to be heat sunk adjacent to the sensor to result in a
temperature measurement of within 1 mK of the actual value. As can be seen, the
impact of wire choice can be quite impressive. If one were to use 36 gauge
Phosphor-Bronze wire, a common choice in cryostat systems, only 7 mm of wire
would need to be heat sunk near a sensor operating at a 4 K surface. However, if
one were to use 24 gauge copper wire in the same scenario, more than half a meter
of wire would need to be heat sunk near the sensor to get a proper reading.
A final installation comment involves the use of feed throughs. If a sensor is
immersed in a cryogenic fluid there are basically two options for connecting the
sensor via wires to the outside environment. The most common and by far
30 J.G. Weisend II

preferable option is to feed the wire through a tube that links the cryogenic fluid
environment to the room temperature environment. The room temperature side of
the tube is topped with a room temperature electrical feed through that passes the
signals to the outside world while maintaining a barrier between the cryogenic
space and the outside environment. There are some disadvantages to this approach.
The required tube can be hard to route between the two spaces and provides a
parallel thermal path from 300 K to the cryogenic space. Such a tube may also be
susceptible to thermoacoustic oscillations (see Sect. 1.9).
The second approach, which can be tempting, is to use a cryogenic feed through
that connects the cryogenic space to the vacuum space and then run just wires in the
vacuum space up to a room temperature connector. This eliminates the tube and
along with it, the possibility for thermoacoustic oscillations and the parallel con-
duction path. It also allows greater flexibility in routing the wires out of the cryostat.
Despite its apparent advantages, the use of cold electrical feedthroughs should be
avoided if at all possible. The reason is that these cold feedthroughs tend to have
poor reliability and can leak. If a leak occurs then there is a leak directly between
the cryogenic fluid and the cryostat vacuum space. Such a leak may well spoil the
vacuum and stop operation of the cryostat until fixed. Such a repair can be quite
time consuming. It’s better to use a warm feed through approach. In some cases,
such as the SHOOT Helium Dewars (Chap. 4), the use of cold feed throughs is
necessary but in theses case extensive testing and prototyping is required to ensure
high reliability.

1.5.6 Commercial Availability of Instrumentation Systems

During the past 25 years, the commercial availability of sensors and related
instrumentation systems suitable for cryostats and cryogenic systems has signifi-
cantly increased. This is particularly true for temperature sensors but also applies
for other measuring systems. Whenever possible, the best practice is to use com-
mercial solutions. This allows the cryostat designer to benefit from the extensive
research, development and testing that have gone into these devices and to take
advantage of the customer support provided by these companies. A good source for
finding cryogenic instrumentation is the Cryogenic Society of America’s annual
Buyers Guide [40].
Some instrumentation problems can not be solved by existing devices and
research into cryogenic instrumentation continues. Good resources for recent
research in cryogenic instrumentation include Advances in Cryogenic Engineering
[41], Proceedings of the International Cryogenic Engineering Conference [42] and
the journal Cryogenics [43].
1 Principles of Cryostat Design 31

1.5.7 Best Practices for Cryostat Instrumentation

Summing up this section, when considering instrumentation in cryostat design, one

should keep in mind:
1. Treat instrumentation as a complete system: sensor, wiring, feed through,
calibration and data acquisition hardware and software.
2. Don’t use more accuracy and precision than required.
3. Use commercially produced sensors whenever possible—there is a lot
available.
4. When possible, mount sensors outside cryostat at 300 K (e.g. pressure trans-
ducers, flow meters).
5. Play close attention to the installation of wiring and to the proper heat sinking
of both wiring and sensors.
6. Consider the possibility of thermoacoustic oscillations. See Sect. 1.9 of this
chapter.
7. For critical devices inside of cryostats, install redundant sensors whenever
feasible.
8. Be sure to consider how to recalibrate sensors.
9. If at all possible avoid, cold instrumentation feedthroughs.
10. Once prototyping is complete, minimize number of sensors in series production
of cryostats.

1.6 Seals and Connections

A common feature of cryostats is the need to connect together components such as

pipes, vessels, flanges and feedthroughs. In many cases, these connection have to
provide vacuum tight seals at cryogenic temperatures; separating a cryogenic fluid
from a vacuum space.
Whether the connection will operate at room temperature or cryogenic tem-
perature, the most reliable approach is to weld the components together. Good
quality welds, particularly in stainless steel, are the simplest way to create reliable,
vacuum leak tight connections. A common temptation is to include many
demountable joints, such as flanged connections, in the cryostat design to allow
easy assembly and disassembly. This temptation should be avoided. While reliable,
demountable, connections do exist; the difficulty in finding and fixing leaks in
cryogenic systems is so large that welding and then later cutting systems apart is the
better approach. This is particularly true in cryostats that are produced in large
numbers such those for the LHC and SSC magnets (Chaps. 2 and 3) or those for the
ILC cryomodules (Chap. 5).
Other metal joining techniques include brazing (sometime used in copper sys-
tems) and soldering. Soldering of tubes and pipes should be avoided, except
32 J.G. Weisend II

Fig. 1.18 Example of a

ConFlat flange using a soft
copper gasket—Courtesy
MDC Vacuum Products, LLC

possibly in the case of small laboratory experiments where reliability is less

important than the ability to make quick changes. Even in this case, flanged con-
nections are superior if space permits.
When welding dissimilar metals such as aluminum, stainless steel and niobium,
transition joints are needed. Reliable transition joints based on brazing, explosive
welding or diffusion bonding of the dissimilar metals together have been developed
for use at cryogenic temperatures [44, 45] and are available commercially [46].
There are situations where demountable cryogenic connections are required.
Two common approaches for this problem are flanged assemblies and bayonets.
Flanged assemblies require a gasket or O-ring to create the final seal between the
mating surfaces. At cryogenic temperatures, polymer or rubber O-rings will become
brittle and won’t work. However, soft metal gaskets (such as copper) will work at
cryogenic temperatures. Figure 1.18 illustrates such a system. These devices are
reliable and can be obtained commercially [47]. Another flanged option is the use of
indium wire for the O-ring. This approach is illustrated in Fig. 1.19. Note the relative
areas of the indium wire and Vee grove indicated in the figure. Indium O-ring flanges
are generally made in house as opposed to purchased commercially. Flanged con-
nection can be used in all orientations and can separate vacuum spaces, fluids from
vacuum and fluids from each other. Care must be taken in flanged connections that
differential thermal contraction upon cooling does not result in the bolts holding the
flanges together shrinking less than the flanges themselves. This may result in a
lessening of the force holding the flanges together resulting in a leaky joint. One
solution to this problem is to use Invar washers in the bolt assembly. Invar (see
Table 1.1) contracts very little between 300 and 4 K. Properly designed into a
1 Principles of Cryostat Design 33

Fig. 1.19 View of an Indium O-Ring flange [48]

flanged assembly, these washers can serve to keep or increase the force on the flange
joint during cool down, preventing leaks. Figure 1.20 and Eq. 1.5, both from Ref.
[48] show how to design such a joint assembly using Invar washers.

X ¼ ½LA ðaA  aS Þ þ eðLA þ LS Þ=ðaS  e  aI Þ ð1:5Þ

where
X required thickness of the Invar washer
αA temperature expansion coefficient for aluminum
αS temperature expansion coefficient for stainless steel
αI temperature expansion coefficient for Invar
LA thickness of aluminum flange
LS thickness of stainless steel flange
ε unit strain of bolts due to cool down (m/m).
Another common approach to making demountable cryogenic piping connec-
tions is to employ bayonets. Bayonets are a set of nested pipes (one “male” and one
“female”) connected together by a room temperature seal. Figure 1.21 illustrates a
typical bayonet pair. Note that the bayonets employ some of the thermal insulation
techniques described above; for example, long thin walls connect the room
34 J.G. Weisend II

Fig. 1.20 Flange assembly using Ivar washers [48]

Fig. 1.21 Example of a cryogenic bayonet pair [48]

temperature and cryogenic portions of the bayonets. The small gap between the
female and male bayonets is designed to impede the development of convective
heat transfer cells between the cryogenic and room temperature portions thus fur-
ther reducing heat leak.
Bayonets have a number of advantages, they are commercially available, reliable
and can be easily disconnected to decouple different cryogenic components. There
are disadvantages as well to bayonets; they are expensive, add additional heat leak
1 Principles of Cryostat Design 35

Fig. 1.22 Cutaway view of the SNS medium beta cryomodule—Courtesy E. Daly-Jefferson Lab,
see Chap. 6). Connection to the cryogenic system are done via bayonets on the end boxes

to the system, are not as reliable as welded assemblies and can be hard to disconnect
if the pipe diameter becomes too large. Additionally, unlike welds and flanged
assemblies, bayonets can only be used in vertical or near vertical orientations. That
is, the warm end of the bayonet must be above the cold end. If this is not done, then
there is the possibility that the cryogenic fluid will flow into the annual space
between the bayonets causing significant heat leak and cooling and possible failure
of the room temperature seal.
Bayonets are a valuable tool in cryostat design but are best used only in situa-
tions where the regular disconnection of cryogenic components is expected.
Bayonet connections are common in segmented cryomodule designs (Chap. 6)
where they allow rapid disconnection of individual cryomodules. Figure 1.22
shows an example of bayonets used in a cryomodule design.

1.7 Transfer Lines

In large scale systems, cryostats are typically connected to each other, to cryogenic
refrigeration plants or to other equipment via insulated transfer lines. These transfer
lines move cryogenic fluids between the various components. Transfer lines can be
quite complicated and their proper design is an important aspect of a successful
cryogenic system. Complicated transfer line systems are also referred to as cryo-
genic distribution systems. In effect, transfer lines are simply another type of
cryostat. Chapter 9 presents more details and examples of cryogenic transfer lines.
Figure 1.23 [49] shows the cross section of a typical multiple line cryogenic
transfer line. Notice that it follows the basic principles of thermal insulation. The
internal cryogenic pipes are enclosed within an outer vacuum jacket and surrounded
by both a actively cooled radiation and an MLI blanket. A low conductivity support
separates the pipes and shield and connects them to the outer vacuum jacket.
36 J.G. Weisend II

Fig. 1.23 Cross section of a typical multiline cryogenic transfer line—Courtesy J. Fydrych, ESS
(see Chap. 9)

Transfer lines have particular requirements based on their typical geometry. Due
to their length, there is significant thermal contraction of the lines during cool down.
This contraction has to be dealt with on the design. The common approach is to
have some of the internal supports fixed to both the internal pipes and to the outer
vacuum jacket with the remainder of the internal supports allowing relative
movement between the pipes and the outer vacuum jacket. Bellows are then
installed on the pipes between the fixed supports to allow relative contraction of the
pipes. An example of this approach is shown in Fig. 1.24 [50]. Note the fixed
support and the support that can move relative to the outer vacuum jacket. An
alternative approach is to fix all the supports to the vacuum jacket but design some
of them so that the pipes can move freely through them upon cool down.
There are other possibilities rather than the use of bellows in the design of
transfer lines to allow for thermal contraction. If the transfer line contains enough
bends or elbows of sufficiently large bend radius then the motion of the pipes with
in the bends relative to the vacuum jacket may be enough to compensate for thermal
contraction. Another option is to construct the cryogenic pipes from Invar. As
shown in Sect. 1.2.1, this material contracts very little upon cool down.
Constructing pipes from Invar greatly reduces the amount of thermal contract that
needs to be allowed for in the design, potentially simplifying the design. This
approach is described in Refs. [51, 52].
A final issue that must be allowed for in transfer lines is the effect of the line
pressure on closed ends of the line. Given the size of many cryogenic pipes, this
pressure can result in a very large unbalanced force that must be transferred to the
vacuum jacket and from the jacket via support structures to the building or earth.
Improper allowance of this effect can result in damage, motion or loss of alignment
of the transfer line.
1 Principles of Cryostat Design 37

Fig. 1.24 Example of fixed and moving spacers in a cryogenic transfer design. Reproduced with
permission from Parente et al. [50]. Copyright 2006, AIP Publishing LLC

Fig. 1.25 The ITER cryodistribution system. Reproduced with permission from Serio [53].
Copyright 2010, AIP Publishing LLC

Figure 1.25 [53] is an example of a complex transfer line or distribution system.

While most transfer line are custom made, there exist commercial options for
simpler systems [54].
38 J.G. Weisend II

1.8 Safety

Safety is a broad topic throughout the field of cryogenics. References [55–60]

provide a good overview of cryogenic safety. There are safety issues specific to
cryostats and these issues should be planned for at the very earliest stage of the
design. Altering the design later to remove safety issues can be very time con-
suming and expensive.
One of the most common mistakes in cryostat design that lead to safety hazards
is the use of inappropriate materials. Materials (see Sect. 1.2) that are inappropriate
for cryogenic use can become brittle and suddenly fail. The solution here is to
always use materials proven for cryogenic service or conduct tests of the material to
show that it is suitable for cryogenic temperatures. One should always keep in mind
that materials nominally expected to operate at room temperature, may though some
other failure, become cooled to cryogenic temperatures and then themselves fail.
Such scenarios should be considered in the design of the cryostat.
A hazard that always has to be addressed in cryostat design is over pressur-
ization. This hazard stems from the very large volume difference between a cryo-
genic liquid at its normal boiling point, i.e. at 1 Bar and the gas phase at room
temperature and pressure. Table 1.5 shows this volume ratio for number of cryo-
genic fluids.
As can be seen, the ratios are very high. Thus, if a cryogenic liquid is in a closed
volume and converted to a warmer gas, very high pressures will result. These
pressures can easily result in explosive failure of the closed volume causing death,
injury and property damage.
The design solution to this problem is to always install properly sized relief
valves in any cryostat so that pressures never go above those for which the cryostat
is designed.
The details of properly sizing relief valves depend on local requirements,
including the applicability of pressure vessel codes. In most locations, it is a
requirement that pressure vessels be designed, built and certified in accordance with
these codes such as the AMSE Pressure Vessel Code [61] or the European Pressure
Directive [62]. Since pressure vessels are frequently defined as vessels in which the
maximum allowable working pressure is greater than 1.5 Bar (absolute) then many
cryostat vessels fall under these codes. In some cases, such as cryostats for space

Table 1.5 Volume ratios Fluid Vgas/VLiquid

between cryogenic fluids at
their normal boiling point and Helium 701
at 300 K and 1 Bar Parahydrogen 788
Neon 1341
Nitrogen 646
Argon 779
CO2 762
Oxygen 797
1 Principles of Cryostat Design 39

applications, weight restrictions prevent full implementations of pressure vessel

codes. In this case local authorities generally required demonstration of an equiv-
alent level of safety.
While details on pressure vessel requirements and relief valve sizing vary
between localities, general guidelines for pressure safety in cryostats include:
1. Always use relief valves that are certified under the applicable pressure code.
2. Never place a shut off valve between the relief valve and the space it is
protecting.
3. Consider how any recalibration requirements of the relief valves will be met.
4. Perform a risk analysis of the system. Ask “what if” questions and then install
relief valves to cope with resulting hazards. In particular, look for volumes that
may not be pressure relieved under certain conditions. Keep in mind that process
valves may be operated incorrectly or may leak; that cryogenic systems may
warm up suddenly and that vacuum insulation systems may fail.
5. While isolation vacuum jackets typically don’t see pressure, they will if the
cryogenic system they contain leak. Allow for this eventuality by always
installing appropriately sized pressure relief devices on vacuum jackets.
6. Never disable or remove pressure relief devices and never introduce cryogenic
fluids into systems without suitable pressure reliefs.
7. Have the cryostat’s pressure relief system reviewed by in-house or external
experts.
When considering possible failure modes for sizing relief valves, a frequently
used worse case scenario is the sudden loss of insulation vacuum. It key to
remember here that the issue here is not the sudden appearance of convective heat
transfer (Sect. 1.3.2) but rather that the in rush of air will condense on the cryogenic
surfaces depositing large amounts of heat and boiling off the cryogenic fluids
contained in the vessels and piping. This effect is much more significant than heat
deposited by convective heat transfer. The addition of multilayer insulation wrap-
ped around the surface of the cryogenic vessel or pipes helps reduce this effect.
Further information is given in [63]. In the case of accelerators with supercon-
ducting magnets or superconducting RF cavities, the same effect is seen in the
sudden loss of beam tube vacuum [32].
Additional details of pressure rises in magnet cryostats are provided in Chap. 7
of this book. A useful reference for the sizing of relief valves is found in [64].
Cryostats containing oxygen or flammable liquids like hydrogen and LNG have
additional unique hazards; always involve experts in these areas when working with
such fluids.
As mentioned at the start of this section, safety in cryogenics extends beyond
issues associated solely with cryostats. Significant hazards including Oxygen
Deficiency Hazards need to be considered. References [55–60] should be consulted
for the broader issue of cryogenic safety.
40 J.G. Weisend II

1.9 Thermoacoustic Oscillations

Thermoacoustic oscillations (TAOs) are spontaneously occurring pressure oscilla-

tions that are frequently seen in cryogenic systems. They generally occur in
geometries in which a tube connects room temperature with a cryogenic fluid and is
closed at the 300 K end. The tube thus has a strong temperature gradient between
the warm and cold ends. Unfortunately, this geometry describes many common
cryostat design solutions including: tubes connecting instrumentation wires to room
temperature feed throughs, a capillary tubes connecting a cryogenic bath to a room
temperature pressure transducer, pressure relief lines and a closed bayonet
connections.
“TAOs begin when the temperature gradient causes cold gas in the tube to warm
and expand, thus increasing in pressure. This increased pressure then pushes the gas
into the colder end of the tube, causing the pressure in the warmer end to fall. The
gas then moves back to the warmer end to occupy this now lower pressure space.
Under the proper conditions of tube size and temperature gradient, sustained
pressure oscillations can be set up” [65]. Thermoacoustic oscillations can result is
large pressure swings which can damage equipment and also are very efficient at
moving heat between the warm and cold ends of the tube greatly increasing the heat
load to the cryogenic fluid.
Ideally, one would like to design systems to avoid TAOs. While this is not
completely possible, there have been studies on idealized systems that provide
some guidance. The physics behind TAOS indicate that there should be two zones
of stability where TAOS won’t occur. If the tube is small enough, then the friction
of the oscillating gas on the tube wall will damp out the oscillations preventing the
TAOS from starting. This is known as viscous damping. The other area of stability
occurs in large tubes or vessels where the mass of the fluid is too big for the fluid to
be moved significantly by the heat being transferred form the warm tube wall and
thus oscillations will not start. This is know as inertial damping.
Experimental studies by Gu [66] on the idealized case of tubes with linear
temperature gradients have shown exactly these regions of stability. Figure 1.26
shows these results. For each set of parameters, the left hand region of stability
represents the viscous damping while the right hand region of stability represents
the viscous damping. While idealized these curves can at least give some guidance.
The parameters on the plots are defined as:

TH
a¼ ð1:6Þ
TC

LH
n¼ ð1:7Þ
LC

where
TH the warm temperature
1 Principles of Cryostat Design 41

Fig. 1.26 Stability maps for thermoacoustic oscillations [67]

TC the cold end temperature

LH the warm section length
LC the cold section length
The limit between the cold and warm lengths of the tube is defined as the point
in the linear temperature gradient where

T ¼ ðTH þ TC Þ=2 ð1:8Þ

The studies were done with a 1 m long tube. In order to apply the results to other
geometries, use the reduced radius r′ defined as:
pffiffiffi
r 0 ¼ r= L ð1:9Þ

The studies described above used a very ideal case and thus even if you are in
the stable regions you may find TAOs occurring. There are mitigations that can be
taken in the event that TAOs are present. These include:
• Add volume to the warm end or design the warm end with fixtures to allow the
addition of this volume
• Install a check valve between the warm and cold end (near boundary between
the two)—this converts the problem to a closed cold tube with no TAOs
• Heat sink the closed end (thus changing TH/TC)
• Allow flow through the warm end.
Further details on TAOs may be found in Refs. [67–70].
42 J.G. Weisend II

1.10 Prototyping and Series Testing

A common aspect in the creation of successful cryostats is the building and testing
of a prototype. This process allows the performance of the design to be tested and
compared against requirements. It is not uncommon for the results of the prototype
testing to lead to changes, sometimes major changes, in the final cryostat design.
Prototyping is found in almost all of the case studies that follow in this book. Full
scale prototypes, tested under the expected final operating conditions provide the
most value but in some cases, testing of prototype subassemblies is sufficient.
For systems in which a number of identical cryostats are required, series testing
of the final production cryostats is also recommended. Testing the performance of
completed cryostats under their final operating conditions provides the best evi-
dence that the cryostats are being properly manufactured. Due to the complexity of
many cryostats, small changes in the manufacturing process can lead to significant
problems such as vacuum leaks, electrical shorts or higher than designed heat loads.
Series testing of the completed cryostats as they are produced can identify these
problems in time to correct them in the remaining cryostats under production. Series
testing was carried out for the LHC magnet cryostats [71], the SRF cryomodules for
XFEL [72] and 12 GeV Upgrade [73]. Such testing is also planned for the LCLS II
[74] and ESS [75] cryomodules.
The cost and schedule impact of both prototyping and series testing can be
significant; requiring expensive test stands and months of additional time. There is
always a temptation to skip these steps and move directly to a test of the final
integrated system that uses the cryostats. Doing so, however, leads to a significant
risk that the original design may not meet specifications or that manufacturing
errors have resulted in poor performance. If this occurs, the cost and schedule
impact of repairing the cryostats can easily be more than the initial impact of any
forgone prototyping and series testing. Generally speaking, it is advisable to allow
for prototyping and series testing in the initial cryostat project plan.

References

1. V. Ganni, P. Knudsen, Helium refrigeration considerations for cryomodule design. Adv. Cryo.
Eng. 59B (2014)
2. J.G. Weisend II, C. Pagani, R. Bandelmann, D. Barni, A. Bosotti, G. Gygiel, R. Lange,
P. Pierini, D. Sellman, S. Wolff, The TESLA test facility (TTF) cryomodule: a summary of
work to date. Adv. Cryo. Eng. 45A (2000)
3. F. Haug, T. Peterson, J.G. Weisend II, Cryogenic Safety—A Guide to Best Practice in the Lab
and Workplace (Springer, USA, in Preparation)
4. J.G. Weisend II, R. Boyce, W. Burgess, A. Candia, R. Carr, J. Gao, K. Gustafsson, C. Jones,
W. Kaminskas, G. Oxoby, R. McKeown, H. Quack, A. Scott, T. Weber, The cryogenic system
for the SLAC E158 experiment. Adv. Cryo. Eng. 47A (2002)
5. R.P. Reed, A.F. Clark (eds.), Materials at Low Temperatures (American Society of Metals,
1983)
1 Principles of Cryostat Design 43

6. J.G. Weisend II, V. Flynn, E. Thompson, R.P. Reed, A reference guide for cryogenic
properties of materials. SLAC-TN-03–023, web information available at http://www.slac.
stanford.edu/cgi-wrap/getdoc/slac-tn-03-023.pdf. Accessed 7 Jan 2015
7. T.F. Durham, R.M. McClintock, R.P. Reed, Cryogenic Materials Data Handbook (US Dept.
of Commerce, National Bureau of Standards, Washington, DC, 1961)
8. J.G. Weisend II (ed.), The Handbook of Cryogenic Engineering (Taylor & Francis, New York,
1998)
9. G. Ventura, M. Perfetti, Thermal Properties of Solids at Room and Cryogenic Temperatures
(Springer, New York, 2014)
10. Cryocomp, Eckels Engineering. http://www.eckelsengineering.com/
11. METALPAK, CPAK and EXPAK from Cryodata. http://www.htess.com/software.htm
12. H.M. Rosenburg, Low Temperature Solid State Physics (Oxford University Press, Oxford,
1963)
13. T.H. Nicol, TESLA test cell cryostat support post thermal and structural analysis.
Fermilab-TM-1794 (1992)
14. S.W. Van Sciver, Helium Cryogenics (Springer, New York, 2012)
15. S.R. Breon, R.A. Hopkins, S.J. Nieczkoski, The X-ray spectrometer—a cryogenic instrument
on the advanced X-ray astrophysics facility. Adv. Cryo. Eng. 37B (1992)
16. A. Poncet, V. Parma, Series-produced helium cryostats for the LHC magnets: technical
choices, industrialisation, costs. Adv. Cryo. Eng. 53A (2008)
17. W.L. Johnson, A.O. Kelley, J.E. Fesmire, Thermal degradation of multilayer insulation due to
the presence of penetrations. Adv. Cryo. Eng. 59A (2014)
18. J.G. Weisend II, Cryogenic engineering, in Mechanical Engineer’s Handbook, ed. by M. Kutz
(Wiley, New York, 2015)
19. J. Fesmire, S. Augustynomicz, B.E. Scholtens, Robust multilayer insulation for cryogenic
systems. Adv. Cryo. Eng. 53B (2008)
20. J. Polinski, M. Chorowski, A. Choudhury, T. Datta, Synthesis of the multilayer cryogenic
insulation modeling and measurements. Adv. Cryo. Eng. 53B (2008)
21. T.M. Flynn, Cryogenic Engineering (Marcel Dekker, New York, 1997)
22. R.G. Baumgartner, E.A. Myers, J.E. Fesmire, D.L. Morris, E.R. Sokalski, Demonstration of
microsphere insulation in cryogenic vessels. Adv. Cryo. Eng. 51B (2006)
23. S. White, J. Demko, A. Tomich, Flexible aerogel as a superior thermal insulation for high
temperature superconductors. Adv. Cryo. Eng. 55A (2010)
24. J.E. Fesmire, S. White, G. Gould, S. Augustynowicz, Aerogel blanket insulation materials for
cryogenic applications. Adv. Cryo. Eng. 55B (2010)
25. R. Barron, Cryogenic Systems (McGraw-Hill, 1966)
26. G. Oliver, J.P. Thermeau, P. Bosland, G. Devanz, F. Lesigneur, C. Darve, ESS Cryomodule
for elliptical cavities, in Proceedings of the 16th International Conference on RF
Superconductivity (2013)
27. J.D. Fuerst, S.M. Gerbick, M.P. Kelly, M. Kedzie, et al. Assembly, installation and
commissioning of the ATLAS upgrade cryomodule. Adv. Cryo. Eng. 55A (2010)
28. R.T. Parmley, Passive orbital disconnect strut (PODS-III), structural and thermal test program.
NASA CR 166473 (March 1983)
29. R.T. Parmley, Unique cryogenic features of the gravity probe B experiment. Adv. Cryo. Eng.
33 (1988)
30. J.W. Ekin, Experimental Techniques in Low-Temperature Measurements (Oxford University
Press, Oxford, 2006)
31. F. Haug, A. McInturff, Measurement of pressure transmission in long capillaries. Adv. Cryo.
Eng. 35 (1990)
32. R.C. Dhuley, S.W. Van Sciver, Sudden vacuum loss in long liquid helium cooled tubes. IEEE
Trans. Appl. Supercond. 25, #3 (2015)
33. S.W. Van Sciver, D.S. Holmes, X. Huang, J.G. Weisend II, He II flowmetering. Cryogenics
31, 75–86 (1991)
44 J.G. Weisend II

34. T. de Jonge et al., Development of a mass flowmeter based on the Coriolis acceleration for
liquid, supercritical and superfluid helium. Adv. Cryo. Eng. 39 (1994)
35. K. Kajikawa et al., Fundamental investigation of a superconducting level sensor for liquid
hydrogen with MgB2 wire. J. Phys. Conf. Ser. 97 (2008)
36. R. Kurunanithi el al., Calibration of an HTS based LOX 400 mm level sensor, in Proceedings
of ICEC 25: Physics Procedia (2015), p. 67
37. I.V. Velichkov, V.M. Drobin, Capacitive level meters for cryogenic fluids with continuous
read-out. Cryogenics 30 (June 1990)
38. R. Sawada et al., Capacitive level meter for liquid rare gases. Cryogenics 43, 449–450 (2003)
39. M.J. DiPirro, A.T. Serlemitsos, Discrete liquid/vapor detectors for use in liquid helium. Adv.
Cryo. Eng. 35 (1990)
40. 2015 Buyer’s Guide (Cryogenic Society of America, 2015). http://www.cryogenicsociety.org/
buyers_guide/
41. Advances in Cryogenic Engineering 1–62 (1955–2016)
42. Proceedings of the 25th ICEC, Physics Procedia 67 (2015). http://www.sciencedirect.com/
science/journal/18753892/67
43. Cryogenics. http://www.journals.elsevier.com/cryogenics/
44. J.D. Fuerst et al., Niobium to stainless steel braze transition development, in Proceedings of
the 11th Workshop on RF Superconductivity
45. W. Soyars et al., Superfluid helium testing of a stainless steel to titanium piping transition
joint. Adv. Cryo. Eng. 55A (2010)
46. An example of commercially available transition joints can be found at: http://www.rbdh.com/
Bimetallic-transition-joints,164.html
47. See, for example, ConFlat flanges from various manufacturers
48. G. McIntosh, Cryostat design, in The Handbook of Cryogenic Engineering, ed. by J.G.
Weisend II (Taylor & Francis, New York, 1998)
49. E. Pyata et al., XFEL injector-1 cryogenic equipment. Phys. Procedia 67, 868–873 (2015)
50. C. Parente et al., The local helium compound transfer lines for the large hadron collider
cryogenic system. Adv. Cryo. Eng. 51 (2006)
51. S. Claudet et al., Two 100 m invar transfer lines at CERN: design principles and operating
experience for helium refrigeration, in Proceedings of ICEC 20 (2005)
52. A. Thakar et al., Design and analysis of a bellows free cryogenic transfer line, in Proceedings
of International Conference on Current Trends in Technology (2011)
53. L. Serio, Challenges for ITER cryogenics. Adv. Cryo. Eng. (2010)
54. An example of commercially available transfer lines may be found at: http://www.nexans.de/
eservice/Germany-en/navigate_315900/Cryogenic_Systems.html
55. F.J. Edeskuty, W.F Stewart, Safety in the Handling of Cryogenic Fluids (Springer, 1996)
56. M.G. Zabetakis, Safety with Cryogenic Fluids (Plenum Press, 1967)
57. F.J. Edeskuty, M. Daugherty, Safety, in The Handbook of Cryogenic Engineering, ed. by J.G.
Weisend II (Taylor & Francis, 1998)
58. T.M. Flynn, Safety with cryogenic systems (Chap. 10), in Cryogenic Engineering (Marcel
Dekker, 1997)
59. Cryogenic Safety Manual: A Guide to Good Practices (British Cryogenics Council, 1991)
60. J.G. Weisend II (ed.), Cryogenic Safety—A Guide to Best Practice in the Lab and Workplace
(Springer, in preparation)
61. Boiler & Pressure Vessel Code 2015 (ASME, 2015)
62. Pressure Equipment Directive, PED 2014/68/EU
63. C. Heidt, S. Grohmann, M. Sußer, Modeling of the pressure increase in liquid helium cryostats
after failure of the insulating vacuum Adv. Cryo. Eng. 59B (2014)
64. EN 13648-1:2008, Cryogenic vessels. Safety devices for protection against excessive pressure.
Safety valves for cryogenic service
65. J.G. Weisend II, Thermoacoustic oscillations, in Cold Facts Winter 2013 (2013)
66. Y. Gu, Thermal Acoustic Oscillations in Cryogenic Systems, PhD Thesis, University of
Colorado, 1993
1 Principles of Cryostat Design 45

67. Y. Gu, K.D. Timmerhaus, Experimental verification of stability characteristics for thermal
acoustic oscillations in a liquid helium system. Adv. Cryo. Eng. 39 (1994)
68. T. Yazaki et al., Stability limit for thermally driven acoustic oscillation. Cryogenics (July
1979)
69. T.J. Miller, Y. Gu, Elimination of thermal acoustic oscillations in cryogenic pumps. Adv.
Cryo. Eng. 51 (2006)
70. N. Dittmar et al., Onset of thermoacoustic oscillations in flexible transfer lines for liquid
helium, in Proceedings of ICEC 25—Physics Procedia 67 (2015)
71. V. Chohan et al., LHC magnet tests: operational techniques and empowerment for successful
completion
72. M. Wiencek et al., Tests of the accelerating cryomodules for the European X-ray free electron
laser, in Proc. of SRF 2013 (2013)
73. J. Hogan et al., 12 GeV upgrade project—cryomodule production, in Proceedings of IPAC
2012 (2012)
74. LCLS-II Final Design Report (LCLSII-1.1-DR-0251-R0), 22 Nov 2015
75. W. Hees et al., The ESS cryomodule test stand, in Proceedings of ICEC 25—Physics
Procedia, vol. 67 (2015)
Chapter 2
SSC Collider Dipole Cryostat

Thomas H. Nicol

Abstract The Superconducting SuperCollider (SSC) project required almost 8000

dipole magnet cryostats. These cryostats had design requirements that included low
heat leak, structural stability and cost minimization. Detailed designs were devel-
oped and numerous prototype magnets were built and tested. While the SSC project
did not continue, many aspects of the SSC dipole magnet cryostats were used in
later projects such as the LHC and ILC. This chapter describes the SSC dipole
magnet cryostat design including design criteria, thermal insulation systems,
structural supports and interconnects. Test results from prototypes, particularly in
the area of heat leak are also presented.

2.1 Introduction

The design of the Superconducting Super Collider (SSC) required the construction
of 7680 superconducting dipole or bending magnets, each 17 m long, 1356
quadrupoles or focusing magnets, each 3.3 m long as well as several special
magnets, spool pieces, and other hardware. The entire accelerator was to be housed
in an 83 km-long underground tunnel [1].
Fermilab was involved in the design and production of cryostats for all prototype
SSC dipole magnets built beginning in 1986. Late in 1989 a fundamental change in
the dipole coil design; increasing the physical aperture from 40 to 50 mm, resulted
in the need to redesign the dipole cryostat. The larger aperture gave rise to an
increase in diameter of the overall cold mass assembly of approximately 25 % and a
weight increase of approximately 50 % both of which impact the cryostat design
significantly.
Cryostat designs for superconducting magnets are largely driven by thermal and
structural considerations. Designers must continually be cognizant of the heat load

T.H. Nicol (&)

SRF Development Department, Technical Division, Fermi National Accelerator Laboratory,
P.O. Box 500, Batavia, IL 60510, USA
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 47

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_2
48 T.H. Nicol

to the helium system and of the structural loads imposed on the cryostat systems
from static weight, shipping and handling, quench loads, and ambient ground
motion. These two considerations are generally at odds with one another. Low heat
load implies a minimum of structural material conducting heat from the environ-
ment. Sound structural design implies material with sufficient strength to resist both
static and dynamic forces.
This chapter summarizes the results of the 50 mm collider dipole cryostat design
effort. Thermal and structural aspects of the design of each of the major cryostat
systems will be described in detail. The 40 mm dipole cryostat purposely served as
the starting point for this design work. Wherever possible, development work on
that and other similar cryostats was borrowed in order to take advantage of proven
technology. References to the 40 mm cryostat will be made for comparison
wherever appropriate. Each of the cryostat systems will be addressed in turn;
vacuum vessel, 80 and 20 K thermal radiation shields, multilayer insulation (MLI),
cryogenic piping, suspension system, and magnet interconnect.
To begin the design process, one needs a good handle on the pertinent allowable
heat loads to each thermal station as well as a good definition of the structural
environment to which the magnet will be subjected. Given the large number of
magnets, even a small heat load, especially to 4.5 K, could result in a significant
increase in the overall capacity of the cryogenic plants needed. Table 2.1 sum-
marizes the thermal and structural design criteria for the 50 mm dipole that affect
the cryostat design. Figure 2.1 illustrates a cross section through the 50 mm
cryostat [2, 3].

Table 2.1 Thermal and structural design criteria

4.5 K 20 K 80 K
Static heat loads
Infrared (W) 0.053 2.335 19.1
Support conduction (W) 0.160 2.400 15.8
Interconnect (W) 0.150 0.320 2.1
Total static (W) 0.363 5.055 37.0
Dynamic heat loads
Synchrotron radiation (W) 2.169
Splice heating (W) 0.140
Beam microwave (W) 0.195
Beam gas (W) 0.136
Total dynamic (W) 2.640
Total dipole (W) 3.003 5.055 37.0
Structural load summary
Cold mass weight 11,360 kg
Shipping and handling 2.0 g Vertical
1.5 g Axial
1.0 g Lateral
2 SSC Collider Dipole Cryostat 49

Fig. 2.1 SSC 50 mm collider dipole cryostat—cross section

2.2 Vacuum Vessel

The vacuum vessel was the outermost cryostat component and served to contain the
insulating vacuum. In addition, it functioned as the major structural element to
which all other systems were ultimately attached to the accelerator tunnel floor.
Furthermore, it served as a pressure containment vessel in the event of a failure in
an internal cryogen line. The vessel was a carbon steel cylinder 685.5 mm OD,
7.92 mm wall with an overall length of 14,922.5 mm.
The vessel had five penetrations at which the internal cold mass supports were
attached. The internal supports were positioned to minimize the deflection of the
cold assembly due to its own weight and were spaced on 3179.76 mm centers
starting at the center of the magnet. Two external supports served to tie the vessel to
the tunnel floor. The external supports were spaced to minimize the deflection of the
vacuum vessel due to its own weight and to the weight of the internal components.
They were located 4154.17 mm on either side of center.
The 7.92 mm wall of the vessel was not sufficient to transfer suspension system
loads to ground immediately around the internal support locations. At these posi-
tions a 19.05 mm thick, 559 mm long reinforcement transferred internally gener-
ated loads, primarily from shipping and handling, from the magnet support system
to the vacuum vessel. Figure 2.2 illustrates the major features of the vacuum vessel
and notes applicable dimensions. The vacuum vessel material was low carbon steel.
Ordinarily one would like to use a material with greater fracture toughness in the
event of a major cryogen spill inside the cryostat, however, cost mandated a more
readily available material. In addition, failure analysis indicated that a high fracture
toughness material was not required [2, 3].
50 T.H. Nicol

Fig. 2.2 SSC 50 mm collider dipole vacuum vessel layout

Fig. 2.3 SSC 50 mm collider dipole complete magnet

Referring only to drawings, it’s easy to lose sight of the scale of these magnets.
Figure 2.3 shows a complete dipole magnet being transported within the Fermilab
site on a full-length tractor-trailer.

2.3 Thermal Radiation Shields

In order to minimize thermal radiation heat load to the 4.5 K magnet, the cryostat
used two thermal radiation shields. The outer shield was cooled by LN2 and
operated nominally at 80 K. The inner shield was cooled by helium gas returning to
2 SSC Collider Dipole Cryostat 51

Fig. 2.4 Lower half of the 20 K thermal shield assembly

the refrigerator and operated nominally at 20 K. Each shield was cooled by discrete
pipes connected to formed shells which totally encompassed the internal cold
assembly. The 20 K shield used a single extrusion attached to the shell. The 80 K
shield used two; one serving as the shield supply, the other as the return. This
scheme required that the shells have high thermal conductivity to minimize thermal
gradients around their circumference. Copper and aluminum are the materials of
choice for this application. Thermal analyses on both shields were performed using
both materials. The results from these analyses indicated there is no significant
difference in the thermal performance of either shield between these two materials.
Material cost dictated the ultimate material choice. Copper and aluminum are
approximately the same cost per unit weight, however, with a density over three
times that of aluminum, the use of copper as the shield material would have resulted
in an assembly cost three times higher than an aluminum shield. The selected
material for both 80 and 20 K shields was 6061-T6 aluminum, 1.59 mm thick. The
shields were segmented along their length to minimize the effect of bowing during
cool down caused by the asymmetry of the cooling tubes with respect to the shield
shells. The thermal shields are shown in Fig. 2.1. The lower half of a 20 K thermal
shield assembly is shown in Fig. 2.4 [2, 3].
52 T.H. Nicol

2.4 Multilayer Insulation

The design requirements for the thermal insulation system in the SSC dipole
cryostat dictated that heat load from thermal radiation and residual gas conduction
be limited to the values listed in Table 2.1. Essential to meeting these requirements
was an insulation system design that addressed transient conditions through high
layer density for improved gas conduction shielding, had sufficient mass and heat
capacity to reduce the effects of thermal transients, and was comprised of materials
suitable for extended use in a high radiation environment. Finally, the system
design was such that fabrication and installation techniques guaranteed consistent
thermal performance throughout the entire accelerator.
The insulation system needed to have a mean apparent thermal conductivity of
0.76 × 10−6 W/cm-K in order to meet the design heat load budget. This was
achieved by using a multilayer insulation (MLI) system comprised of reflective
layers of aluminized polyester separated by layers of spun-bonded polyester spacer.
The reflective layers consisted of flat polyester film aluminized on both sides to a
nominal thickness not less than 350 Å. The spacer layers consisted of
randomly-oriented spun-bonded polyester fiber mats. The mean apparent thermal
conductivity of an MLI blanket comprised of these materials was measured to be
0.52 × 10−6 W/cm-K [6, 7]. The MLI system for the SSC 50 mm collider dipole
cryostat consisted of full cryostat-length assemblies of MLI fabricated and installed
as blankets on the 80 and 20 K shields [2–4].
The MLI system for the 80 K thermal shield consisted of two 32-reflective-layer
blanket assemblies, for a total of 64 reflective layers. The stack height of each
32-layer blanket was 8.86 mm, with a mean layer density of 3.61 layers per mm.
The blanket design incorporated 32 reflective layers of double-aluminized poly-
ethylene terephthalate (PET) film. The reflective layers were separated by single
spacer layers of 0.10 mm spun-bonded PET material. Single layers of 0.23 mm
spun-bonded PET covered the blanket top and bottom and served to position the
polyester hook and loop fasteners at the blanket edges. The fasteners were affixed to
the cover layers by sewing. A third 0.23 mm PET layer was located midway
through the blanket assembly and separated the upper and lower 16 reflective layers
of MLI. The multiple blanket layers were sewn together as an assembly along both
edges of the blanket. Non-lubricated polyester thread was used in all sewing
operations.
The MLI system for the 20 K thermal shield consisted of ten reflective layers of
PET film aluminized with a nominal coating of 600 Å per side. The reflective layers
were separated by three spacer layers of 0.1 mm spun-bonded PET material. Single
layers of 0.23 mm spun-bonded PET covered the blanket top and bottom and
served to position the polyester hook and loop fasteners at the blanket edges. The
fasteners were attached to the cover layers by sewing. The multiple blanket layers
were sewn together as an assembly along both edges of the blanket.
At each blanket edge, the upper MLI layers were sewn together from the upper
cover layer through to the middle three layers with thread terminated in the three
2 SSC Collider Dipole Cryostat 53

Fig. 2.5 Cross section through the 80 K MLI blanket

layers. The seam location was then incremented 7.62 cm laterally along the
mid-layer and the lower MLI layers were sewn together from the middle three
layers through to the lower cover layer. Figure 2.5 shows a cross section of the
80 K MLI blanket.
To facilitate assembly and all the blankets required to support the SSC
pre-production program, a system was devised to wrap and sew complete blanket
assemblies. It consisted of a wheel with a width equal to the perimeter of the
blankets and a circumference equal to their length. The reflector and spacer
materials were fed off supply reels onto the wheel and the edges were sewn after
wrapping the required number of layers was complete. Figure 2.6 shows the
assembly wheel during the winding process.

2.5 Cryogenic Piping

In addition to providing the necessary structural support and thermal insulation for
the cold mass assembly, the cryostat served to contain the piping for all of the
cryogenic services required for magnet and magnet system operation. All
the cryogenic pipes are shown in Fig. 2.1. Table 2.2 provides a summary of the
pressure and flow parameters for each of these services. Table 2.3 lists the pipe
sizes for each of the cryostat pipes. Each pipe was anchored at the center of
the magnet assembly, and was free to slide axially at the remaining support points
to allow for thermal contraction during cool down.
54 T.H. Nicol

Fig. 2.6 MLI blanket winding and sewing apparatus

Table 2.2 Cryogenic system pressure and flow parameters

System P(max) Fluid P(oper) v T (K) Flow ΔP
(atm) (atm) (cm3/g) (g/s) (psi)
20 K shield 18 He 2.5 163.4 20 100 0.016
80 K shield 18 LN2 5.0 1.286 84 750 0.012
Single phase 18 He 4.0 7.255 4.25 100 0.010
LHe return 18 He 3.0 7.609 4.425 85.6 0.010
GHe return 9 He 0.9 73.07 4.3 26.8 0.00053
Vac vessel 2 He 0.0 na 300 na na
2 SSC Collider Dipole Cryostat 55

Table 2.3 Cryostat pipe sizes

System Pipe ID (mm) Pipe OD (mm) Flow area (mm2) Material
20 K shield 82.55 88.90 5352 Aluminum
80 K shield 57.15 63.50 2565 × 2 Aluminum
LHe return 45.14 47.62 1600 Stainless steel
GHe return 86.41 88.90 5864 Stainless steel

2.6 Suspension System

The suspension system in any superconducting magnet serves as the structural

attachment for all cryostat systems to the vacuum vessel, which in turn anchors
them to the accelerator tunnel floor. Conventional suspension systems were effec-
tively reinvented during early SSC magnet development. The intention in devel-
opment of the 50 mm dipole cryostat was to take advantage of that earlier work
tailoring it only for the revised thermal and structural design parameters. Schedule
requirements for the redesign work did not allow continued suspension R&D and,
in fact, the thermal and structural performance of the 40 mm design did not indicate
it to be necessary [2, 3, 5–7].
Using the thermal and structural parameters outlined in Table 2.1 the suspension
system for the new design was extrapolated from its 40 mm counterpart. The
emphasis was on meeting the allowed suspension system conduction heat load,
satisfying the structural requirements, and maximizing the lateral suspension stiff-
ness. This latter constraint is not explicitly defined in the design criteria but arose
out of concerns during testing of 40 mm prototypes that the lateral natural fre-
quency was too closely coupled with resonances found in an over the road shipping
environment.
As shown in Fig. 2.7, the suspension system consisted of two major components;
reentrant style support posts and anchor tie bars. There were five support posts
located along the length of an SSC cold mass assembly spaced 3179.76 mm apart
centered about the middle of the cold mass. Five being the number that limited the
cold mass sag due to self-weight to the allowed 0.25 mm. The support posts resisted
vertical and lateral loads imposed during shipping and handling. The cold mass was
allowed to slide axially with respect to all but the center post to allow for thermal
contraction of the cold assembly during cool down. This implied that, given no other
restraint, the center support would need to resist axial loads. A single support was not
capable of resisting this potential 1.5 g load. Using a special support at the center
that could handle this load would have imposed inordinately high heat loads on the
refrigeration system. Rather, a means was developed to distribute axial loads to all
five supports without impacting the conduction heat load. Axial tie bars were used to
connect the top of each support post to its neighbor(s). In this scheme, an axial load
acted first at the center support and then distributed to all five supports.
The tie bars needed to be dimensionally stable when cooled from their assembly
temperature of 300 K to their operating temperature of 4.5 K, otherwise their
56 T.H. Nicol

Fig. 2.7 SSC collider dipole cryostat suspension system components

Fig. 2.8 SSC collider dipole cryostat re-entrant support post

contraction would impose high bending loads on the support posts. Uniaxial gra-
phite fibers in an epoxy matrix provided the solution. This material is extremely stiff
and yet exhibits virtually no change in length during cool down. An elastic modulus
of 82.5 GPa can be achieved in such a composite even with relatively low modulus
fibers.
The configuration of the support posts for the 50 mm cryostat is shown in
Fig. 2.8. The outer composite tube was 2.16 mm thick and filament wound using
S-glass in an epoxy matrix. The inner tube was 3.18 mm thick and filament wound
using graphite fibers in an epoxy matrix. The ability of this design to use two
2 SSC Collider Dipole Cryostat 57

Table 2.4 Re-entrant support post design parameters

Design lateral load 32960 N
Design vertical load 22240 N
Overall support height 177.80 mm
Outer composite tube 241.30 mm OD, 2.16 mm wall
90.55 mm long, glass/epoxy
Design stress: 132 MPa
Inner composite tube 177.80 mm OD, 3.18 mm wall
150.88 mm long, graphite/epoxy
Design stress: 154 MPa
Theoretical lateral natural frequency 9.41 Hz
Heat load to 80 K 3.133 W
Heat load to 20 K 0.492 W
Heat load to 4.5 K 0.033 W

different materials allowed each material to be used in its optimum temperature

range. Table 2.4 provides a summary of the critical design and performance
parameters of the 50 mm cryostat supports.
Figure 2.9 shows the cold mass and a cradle and support post assembly.

2.7 Interconnect

As its name implies the magnet interconnect served as the connection area between
magnets at which each of the cold mass and cryostat pipes between magnets were
connected. Each pipe was anchored axially at the center of the magnet which means
they contracted between 25 and 32 mm depending on whether they were stainless
steel or aluminum. The relative contraction at the interconnect was then 50–64 mm,
i.e. twice the single magnet value. Bellows were required on each cold mass, shield,
and cryogenic pipe to allow for this expansion. All bellows were hydro-formed
stainless steel. For the aluminum extrusions on the 80 and 20 K shield pipes an
aluminum to stainless steel transition joint was required for the bellows connection.
Joints using diffusion bonding or brazing between these two materials had been
successfully employed in 40 mm prototypes. Lateral instability was a concern for
interconnect bellows and required the use of internal squirm protectors on each
bellows assembly. During individual magnet testing, the bellows assemblies
employed metal seals and mechanical connections, but for tunnel installation, all
bellows and pipe connections were welded [2, 3].
Just like for the magnet as a whole, radiative heat transfer needed to be mini-
mized in the interconnect area. This was accomplished by shield bridges that
spanned the 80 and 20 K shield gap between adjacent magnets. These bridges were
essentially extensions to the magnet shields, modified to contain their respective
58 T.H. Nicol

Fig. 2.9 Cold mass and cradle and support post assembly

bellow OD’s if necessary. A sliding joint between bridge sections on adjacent

magnets accommodated contraction during cool down. Each was covered with the
same MLI scheme used throughout the body of the magnet. These shield bridges
also contained pressure reliefs for each shield to prevent pressure buildup within
either shield in the event of an internal piping failure. The reliefs were located in the
upper half of the shield sections in order to prevent liquid spills from impinging
directly onto the vacuum vessel wall. Figure 2.10 illustrates a typical magnet
interconnection. The external vacuum bellows and thermal shield bridges have been
removed for clarity.
2 SSC Collider Dipole Cryostat 59

Fig. 2.10 SSC 50 mm collider dipole magnet interconnect

2.8 Test Results

Early in the 50 mm magnet development program a full scale thermal model was
constructed and tested in a series of tests at Fermilab. The objectives of the thermal
model program were to evaluate the cryostat design from a production standpoint,
to gain experience in magnet handling and transportation, to monitor the transient
thermal and structural responses of the cryostat, and to measure heat leaks to the
cold mass and thermal shields. The thermal model was identical to later magnetic
models except that the cold mass assembly contained a simulated collared coil
assembly and the model ends were reconfigured for open cycle heat leak mea-
surements. The model was instrumented to evaluate thermal and structural per-
formance and included temperature sensors to monitor cool down, warm-up, and
steady state conditions and strain gauges to monitor the performance of the sus-
pension system during onsite transit, cool down, and operation. Construction of the
thermal model utilized the same fabrication facility, assembly procedures, and
components as the magnetic models [8, 9].
The measurements were performed in an open cycle mode with supply reser-
voirs at each end of the cold mass. Figure 2.11 shows an overview of the mea-
surement system. The inner shield was supplied from an external dewar. The inner
20 K shield temperature was an operational variable which was controlled by
varying its flow rate. The outer 80 K shield was supplied by LN2 reservoirs at each
end of the measurement system. The external piping was equipped to measure flow
rates of the cold mass and inner and outer shield liquid and gas streams.
After cool down, initial filling, and stabilization of the three systems, the tem-
perature of the inner shield was regulated to a selected test value. After regulation,
temperatures and heat leaks were monitored to establish equilibrium, often taking
several days. Once at equilibrium, steady state heat leak data was taken.
60 T.H. Nicol

Fig. 2.11 Thermal model measurement schematic

The cold mass heat leak was determined by boil-off measurements. Since the
measurement included the heat leak to the cold mass and the reservoirs at each end
of the magnet, a separate measurement called the zero-magnet test was conducted
using just the helium reservoirs, which was then subtracted from the total system
heat leak. The inner shield heat leak was computed from the shield flow rate and the
temperature rise in the shield gas stream. Like the cold mass, the heat leak to the
80 K shield was determined using the liquid boil-off rate. As with the cold mass,
the zero-magnet test results were subtracted from the total.
The measurement program took place over approximately a three month period
and involved 23 data-taking runs as identified by the inner shield temperature. The
following results have been screened to exclude transient periods, upset conditions,
and operational problems.
Cool down was gradual due to the open cycle nature of the operation. The cold
mass was initially cooled and filled with liquid nitrogen, evacuated, and then filled
with liquid helium. The total time required to cool and fill the cold mass with liquid
helium was 294 h.
The support post bending loads due to differential axial thermal contraction were
low with the exception of the downstream end post which indicated a load of
4800 N. A probable cause for such a load is a non-operational cold mass slide, most
likely the result of binding in the slide material itself. The support post temperature
profiles agreed well with those of an identical post measured in a heat leak test
facility.
Random mechanical noises occurred during cool down and throughout the
measurement program likely the result of relative motion between the cold mass,
2 SSC Collider Dipole Cryostat 61

Fig. 2.12 Cold mass heat

leak versus inner shield
temperature

thermal shields, and support posts and the vacuum vessel due to ambient temper-
ature variations and settling of the test pad.
The subtractive heat leak contribution of the end reservoirs was measured with
the inner shield cooling tube filled with liquid helium to eliminate conduction and
thermal radiation to the cold mass end to include end shine thermal radiation. Under
these conditions, the shield temperature was 7.6 K and the measured background
was 865 mW. The background was not strongly dependent on reservoir liquid
level.
Calibration heaters in the cold mass were employed to evaluate the accuracy of
the measurement system. The heaters were energized with the inner shield at
approximately 20 K. The increase in the measured cold mass heat leak corre-
sponding to heater power levels of 200 and 398 mW were 197 and 376 mW,
respectively.
The cold mass heat leak at the cryostat design point with the inner shield at
20 K, was 140 ± 40 mW as compared to the predicted 128 mW and budgeted
300 mW. The cold mass heat leak versus inner shield temperature is shown in
Fig. 2.12.
Considerable differences exist between measured and predicted heat leaks at
several experimental points. Factors that can contribute to these differences are as
follow:
• The center section heat leak (140 mW) is small relative to the balance of the
apparatus (865 mW). This unbalance amplifies end vessel effects.
• Thermal communication exists between the inner shield cooling circuit and the
end vessel helium reservoirs as a result of the inner shield supply and return
piping passing through the cold mass helium reservoirs. Even though the piping
was insulated, transients in inner shield supply conditions could be seen to affect
the reservoirs.
62 T.H. Nicol

Fig. 2.13 Inner shield heat

leak versus inner shield
temperature

• Level instability (sloshing) occurred occasionally in the end vessels. The

sloshing was most often associated with changes in the inner shield circuit
operation.
• The predicted sensitivity to insulating vacuum is high. At the design inner shield
operating temperature of 20 K, the residual gas (helium) conduction at 10−6 torr
was 43 mW and at 10−5 torr was 430 mW. The insulating vacuum at the vac-
uum vessel mid-span port ranged from 1.3 × 10−6 to 2.9 × 10−6 torr during the
span of the measurements. The location and installation of the vacuum gauge
was demonstrated by diagnostic measurements to inaccurately relate changes in
the insulating vacuum that could correspond to changes in the outgassing rate of
the mild steel vacuum vessel due to changes in ambient temperature.
Consequently, vacuum was not monitored frequently during most of the data
taking.
• Variations in atmospheric pressure result in temperature changes of the liquid
which influence the apparent heat leak. Atmospheric pressure was not monitored
frequently during most of the data taking. Where possible, the heat leak was
corrected for changes of liquid temperature and pressure with time. The cold
mass end to end temperature variation during operation was characteristically
about 30 mK.
The inner shield heat leak at the cryostat design point was 5.0 ± 0.4 W com-
pared to the predicted 2.77 W and budgeted 2.5 W. The inner shield heat leak
versus inner shield temperature is given by Fig. 2.13.
A corroboration of the inner shield heat leak measurement was provided by
performing inner shield boil-off measurements during the cold mass background
measurement. With the shield cooling tube filled with liquid helium, the heat leak as
measured by boil-off was 5.84 W.
2 SSC Collider Dipole Cryostat 63

Fig. 2.14 Outer shield heat

leak versus inner shield
temperature

The factor of two difference between measured and predicted heat leaks was felt
to be due to thermal shorts between the 20 and 80 K systems, by locally com-
pressed insulation between the shields and by an insufficient number of layers of
insulation. Shield and support temperature monitors indicate the possibility of
shorts. An autopsy of the thermal model to investigate the existence of such thermal
shorts was not conducted for programmatic reasons.
The subtractive heat leak contribution of the end reservoirs was made with the
center section removed and the reservoirs connected together. The nominal back-
ground heat leak was 24 W. The background was found to vary with level. The
outer shield heat leak at the cryostat design point was 19 ± 2 W compared to the
predicted 23.3 W and the budgeted 25 W. The outer shield heat leak versus inner
shield temperature data is given by Fig. 2.14. The less than predicted measured heat
leaks were felt to be associated with the thermal shorts that are suspected to exist
between the inner and outer shields and their connections. The scatter in the
measured heat leak was felt to be associated with the sensitivity of the insulation
systems to insulating vacuum in the higher pressure ranges, i.e. greater than 10−4
torr. As noted earlier, the insulating vacuum, while not accurately measured,
showed changes with ambient temperature.
Following the completion of several prototype 40 and 50 mm aperture dipole
magnets, similar, more extensive tests were conducted on strings of magnets at the
SSC Laboratory in Waxahachie, TX. Similar discrepancies between measured,
predicted, and budgeted heat leaks were documented in those tests as well, but the
project was discontinued before more exhaustive tests and analyses could be
conducted.
Figure 2.15 shows a completed SSC collider dipole magnet on the test stand in
the Fermilab Magnet Test Facility.
64 T.H. Nicol

Fig. 2.15 Complete SSC dipole assembly on the Fermilab test stand

2.9 Summary

The SSC development program afforded us the opportunity to extend the design of
cryostats for superconducting magnets far beyond the state of the art present at the
end of the Fermilab Tevatron program. Advances in new materials technology
opened up options for cryostat designers in both thermal and structural materials.
Strict limits on allowable heat load forced us to develop new mechanisms for
structural support and thermal shielding. The end result was a cryostat design which
met the demands of the SSC and which served as the starting point for the
development of other magnet systems far into the future.
It is gratifying to see projects like the LHC (Chap. 3), TESLA, LCLS-II (Chap. 5),
and many others around the world, taking advantage of developments in multilayer
insulation technology and fabrication techniques, thermal shield design, cryogenic
support structures, and interconnect design that were made during the R&D activ-
ities, prototyping, and production of SSC collider dipole magnets.
2 SSC Collider Dipole Cryostat 65

References

1. SSC Central Design Group (1986), Superconducting super collider conceptual design.
SSC-SR-2020, March 1986
2. T.H. Nicol, Design development for the 50 mm superconducting super collider dipole cryostat.
Supercollider 3 (Plenum Press, New York, 1991), pp. 1029–1036
3. T.H. Nicol, SSC 50 mm collider dipole cryostat design. in Advances in Cryogenic Engineering,
vol. 37A (Plenum Press, New York, 1991), pp. 525–533
4. J.D. Gonczy, W.N. Boroski, R.C. Niemann, J.G. Otavka, M.K. Ruschman, C.J. Schoo, A
blanket design, apparatus, and fabrication techniques for the mass production of multilayer
insulation blankets for the superconducting super collider. in Advances in Cryogenic
Engineering, vol. 35A (Plenum Press, New York, 1990), pp. 507–516
5. T.H. Nicol, R.C. Niemann, J.D. Gonczy, SSC magnet cryostat suspension system design. in
Advances in Cryogenic Engineering, vol. 33 (Plenum Press, New York, 1988), pp. 227–234
6. T.H. Nicol, R.C. Niemann, J.D. Gonczy, A suspension system for superconducting super
collider magnets. in Proceedings of the Eleventh International Cryogenic Engineering
Conference (Butterworth & Co., Surrey, UK, 1986), pp. 533–538
7. E.T. Larson, J.A. Carson, T.H. Nicol, R.C. Niemann, Improved design for a SSC coil assembly
suspension connection. in Advances in Cryogenic Engineering, vol. 33 (Plenum Press, New
York, 1988), pp. 235–241
8. R.C. Niemann, W.N. Boroski, J.D. Gonczy, T.H. Nicol, J.G. Otavka, M.K. Ruschman, SSC
dipole magnet cryostat thermal model measurement results. in Advances in Cryogenic
Engineering, vol. 33 (Plenum Press, New York, 1988), pp. 251–258
9. J.D. Gonczy, W.N. Boroski, T.H. Nicol, R.C. Niemann, J.G. Otavka, M.K. Ruschman,
Advances in Cryogenic Engineering, vol. 33 (Plenum Press, New York, 1988), pp. 243–250
Chapter 3
Twenty-Three Kilometres of Superfluid
Helium Cryostats for the Superconducting
Magnets of the Large Hadron Collider
(LHC)

Philippe Lebrun

Abstract The Large Hadron Collider (LHC) at CERN is the world’s largest sci-
entific instrument. The 1600 “high-field” superconducting magnets that make up
the 23 km circumference accelerator ring represent the largest use of superfluid
helium (He II) to date. This chapter describes the design evolution of the LHC
magnet cryostats with particular emphasis on the He II cooling system, thermal
insulation system and structural supports. Prototype testing, series production,
installation and commissioning of these cryostats is also discussed. Numerous
figures and tables illustrate the cryostat and present performance results.

3.1 The LHC and Its Cryogenic System

The Large Hadron Collider (LHC) at CERN, the European Organization for
Nuclear Research near Geneva, is the largest research instrument ever built, and the
most advanced tool in elementary particle physics (Fig. 3.1): its two combined
synchrotrons, 26.7 km in circumference, accelerate and bring into collision intense
counter-rotating beams of protons and ions at high energy, to probe the structure of
matter and study the forces of nature at the unprecedented scale of tera-electron-volt
(TeV) per elementary constituent [1]. This energy, about thousand times the mass
of the proton, corresponds to the temperature conditions in the early universe
10−10 s after the Big Bang: the LHC therefore constitutes a powerful time machine,
in particular to explore the mechanism of electro-weak symmetry breaking which
occurred in the very early universe and to search for possible constituents of dark
matter, relics from this period. In terms of length scale, 1 TeV is equivalent to
10−18 m, and the LHC may thus also be seen as an ultra-microscope resolving
dimensions thousand times smaller than the proton. Exploration of this terra
incognita has already led to the discovery of the long-sought Higgs boson in 2012,

P. Lebrun (&)
CERN, European Organization for Nuclear Research, Geneva, Switzerland
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 67

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_3
68 P. Lebrun

Fig. 3.1 Outline of the Large Hadron Collider in the Geneva area (CERN photo)

and is continuing with precision measurement of its properties as well as with

search for new physics, “beyond the Standard Model”.
To guide and focus its rigid beams along the accelerator tunnel, the LHC uses
more than 1600 high-field superconducting magnets operating in 80 t of superfluid
helium at 1.9 K (Fig. 3.2). In addition to enhancing the performance of the Nb-Ti
superconductor by lower temperature operation, the LHC magnet cooling scheme
exploits the unique thermo-physical properties of superfluid helium for stabilization
against thermal disturbances and for heat transport. The superconducting magnets
operate in baths of “pressurized” helium II—i.e. above saturation pressure, in fact
sub-cooled liquid—close to atmospheric pressure. The large, but finite thermal
conductivity of helium II for heat fluxes of technical interest ensures that each bath
is quasi-isothermal, but is however insufficient to transport heat over long distances
in the accelerator tunnel. The 106.9 m long magnet strings are therefore cooled via a
heat exchanger tube threading its way in the cold mass, inside which a small
two-phase flow of saturated helium II gradually vaporizes as it absorbs heat from the
pressurized helium II baths (Fig. 3.3). This efficient cooling scheme presents several
benefits [2], among which those of avoiding the need for circulation pumps and of
minimizing the need for flow space in the transverse cross-section of the magnet.
The large size of the collider ring, and its implantation some 100 m underground,
limiting access to eight points around its perimeter, result in sector lengths of
3.3 km individually serviced as concerns accelerator and technical systems.
3 Twenty-Three Kilometres of Superfluid Helium … 69

Fig. 3.2 The LHC cryomagnets installed in the tunnel (CERN photo)

Saturated He II, flowing Heat exchanger tube

Pressurized He II, static

SC magnet Helium vessel SC bus-bar connection

Fig. 3.3 Principle of the superfluid helium cooling scheme of the LHC magnets

The LHC ring is cooled by eight large-capacity cryogenic plants, each normally
cooling a 3.3 km sector, but able to serve two adjacent sectors and thus provide
redundancy at partial load. This requires distributing and recovering helium flows at
different temperatures in the access shafts and along the machine tunnel over lengths
of up to 3.3 km, by means of a compound cryogenic distribution line running along
70 P. Lebrun

the magnets and feeding each 106.9 m string in parallel. Each cryogenic plant
provides a mix of liquefaction and refrigeration duties at 50–75 K and 4.5–20 K,
amounting to an equivalent entropic capacity of 18 kW at 4.5 K. It is complemented
at its cold end by a 1.8 K stage producing up to 2.4 kW of refrigeration power, by
means of multi-stage sub-atmospheric compression using cold hydrodynamic and
room-temperature volumetric compressors. Precooling the 37,500 t cold mass down
to about 100 K is achieved by vaporization of some 10,000 t liquid nitrogen. Liquid
nitrogen is otherwise not used in the system. 21,300 sensors monitor the machine
cryogenics, while 4700 analog control loops ensure its operation. The cryogenic
system of the LHC is described in Ref. [3].
In the following we tell the LHC cryostat story, approximately following
chronological order, from the first studies and experimental work which enabled to
establish the feasibility of such a large, novel superfluid helium system, to the
several generations of prototypes meeting the evolution of machine configuration,
their industrialization, series production and assembly, installation and intercon-
nection in the tunnel, commissioning and operational experience. Focus is on the
eight, 2818 m long continuous cold strings of the LHC arcs, containing in particular
the 1232 dipole magnet cryostats (each 15 m in length) which constitute 85 % of
the cold length and therefore drive the design, construction and industrialization
choices, as well as the overall performance and cost of the system.

3.2 Feasibility of a Large Distributed

Superfluid Helium System

The first approach to a large hadron collider to be installed in the existing tunnel
housing the LEP collider, was based on twin-aperture dipole magnets using “cos h”
coils made of Nb3Sn superconductor operated in baths of normal helium at 4.5 K,
to produce a bending field around 10 T [4]. The study included the conceptual
design of the corresponding cryogenic system, including cryostat simplified
cross-section. It identified the main requirements and boundary conditions of
large-capacity refrigerators at 4.5 K, thermal shielding at around 80 K, cryogenic
fluid distribution along the 3.3 km sectors, cryogen storage, cooldown of the large
magnet mass, quench handling and recovery and underground safety.
Soon after this it became clear that, in view of the limited availability, high cost
and technical difficulty of implementing Nb3Sn superconductors in accelerator
magnets, the alternative of operating the more conventional Nb-Ti superconducting
alloy at lower temperature in superfluid helium—to boost its current-carrying
capacity at high field—should also be explored. This technique, pioneered at CEA
Grenoble and first applied to cool high-field magnets for condensed-matter physics
(the Grenoble High-Field Magnet) and nuclear fusion research (the TORE SUPRA
tokamak at CEA Cadarache) [5], had however never been applied to the magnets
and distributed configuration of a large circular accelerator. It clearly appeared that
3 Twenty-Three Kilometres of Superfluid Helium … 71

the prime feasibility issues to be investigated were the management of heat loads
(and thus the design of a cryostat meeting the required thermal performance) and
the cryogenic distribution scheme—the latter also impacting the cryostat design to
incorporate the cryogenic pipelines. A preliminary study of a superfluid helium
cryogenic system for the LHC was then conducted [6].
A novel feature of the LHC with respect to previous superconducting acceler-
ators is that the “static” heat loads, i.e. the heat in-leaks to the cold mass and other
cryogenic components of the machine, represent only a fraction of the total. The
large excitation currents of the magnets—12 kA for the LHC dipoles—induce
resistive heating in the numerous non-superconducting joints between the winding
layers, between the coils and between the magnets in a string: this additional
“dynamic” heat load necessarily ends up into the helium bath at the lowest tem-
perature level and must therefore be contained by a tight specification on the
maximum joint resistance (typically lower than 1 nX per joint). Moreover, the
circulation of intense particle beams in the accelerator rings also induces dynamic
heat loads, through several processes such as dissipation of beam image currents in
the resistive wall of the beam pipes, synchrotron radiation from the beams (in the
UV range for LHC protons) and particle losses from the beam halo. While the latter
produce cascades of secondary particles eventually depositing their energy into the
magnet cold mass, the power of the other two processes may be intercepted at
higher temperature in order to reduce its thermodynamic impact on the refrigeration
system. This is the primary function devoted to the beam screens, the concept of
which was introduced for the first time in the study. Today, beam screens cooled by
supercritical helium between 5 and 20 K equip the LHC magnets and limit the heat
load to the 1.9 K level to acceptable values; they also perform several other
essential functions [7], in particular that of a distributed cryopump maintaining a
good level of dynamic vacuum in the beam pipes, and their elaborate design is
therefore the optimized result of a complex process involving several domains of
accelerator physics and technology (Fig. 3.4). Pre-design of the cryostat featuring a
90 K thermal shield, non-metallic supports of the cold mass with heat intercepts at
90 and 4.5 K, multilayer reflective insulation around the cold mass, and integrating
cryogenic distribution pipelines, gave an estimated heat in-leak at 1.9 K of
0.15 W/m. Containing the dissipation in the resistive joints to 0.05 W/m and
introducing the beam screens to intercept most of the beam-induced heat loads
yielded a total steady-state budget of 0.2 W/m at 1.9 K, to be compared with
0.9 W/m at 4.5 K for a normal helium cryostat in absence of beam screen. The
conclusion of this preliminary study was that a superfluid helium cryogenic system
for the LHC, including magnet cryostat and distribution scheme, appeared possible,
though more complex than one at normal helium temperature.
The following years saw the refinement of such studies, particularly concerning
the cryogenic distribution scheme and the assessment of dynamic heat loads, but the
cryostat design concept (Fig. 3.5) and estimated performance of 0.15 W/m at 1.9 K
were essentially confirmed [8]. The addition of a complete shield at 5 K sur-
rounding the cold mass allowed—in principle—to reduce the residual gas con-
duction to the cold mass and thus the dependence of total heat load on the quality of
72 P. Lebrun

FUNCTION PROCESS DESIGN FEATURE

Low-conduction supports
Limit residual heat load
Reduce beam-induced to cold mass
cryogenic loads High-conductivity copper
Intercept synchrotron plating
radiation
Increase development
time of transverse Cooling at low
resistive-wall instability Limit resistive wall temperature
impedance
Austenitic stainless steel
Resist eddy-current structure
forces at magnet quench Structural material with
high resistivity
Pumping slots
Preserve field quality in Low-permeability
magnet aperture materials
Avoid temperatures
favoring desorption of
Maintain good beam Provide pumping from common gas species
vacuum shielded cold surface
Sawtooth absorber
Limit development of Limit reflectivity and SEY
electron cloud of beam screen surface Beam scrubbing

Fig. 3.4 Functional design map of beam screen for particle accelerator

Fig. 3.5 Transverse

cross-section of conceptual
cryostat housing LHC
twin-aperture dipole magnet
at 1.9 K. 1 magnet coils;
2 magnet collars; 3 magnet
yoke; 4 magnet shrinking
cylinder and helium vessel;
5 beam screen; 6 shield at
5 K; 7 MLI insulated shield at
80 K; 8 non-metallic support
post; 9 vacuum vessel;
A helium II heat exchanger
tube; B–F cryogenic
distribution and recovery lines
at 1.9, 5 and 80 K
3 Twenty-Three Kilometres of Superfluid Helium … 73

the insulation vacuum. Assessment of the practical benefit of this additional feature
led to several iterations discussed in the following.
The early conceptual studies of LHC cryostats benefitted, at least down to
normal helium temperature, from the extensive design and development work
performed in the USA, particularly at Fermi National Accelerator Laboratory in
preparation for the SSC project [9]. Still, experimental validation of the LHC
cryostat design proper was required, first by measurement of critical components
and eventually on a full-scale prototype. A dedicated test set-up was built to per-
form precision heat in-leak measurements at 80, 4.2 and 1.8 K in the laboratory,
combining several independent methods (cryogen boil-off, “heat-meters” i.e. cali-
brated thermal impedances, and transient temperature rise in a superfluid helium
bath). Enabling the measurement of heat loads down to the mW range at 1.8 K, this
experimental set-up was used inter alia to qualify support posts made of
non-metallic composite materials (glass-fibre/epoxy and carbon-fibre/epoxy) and to
optimize position and practical realization of their heat intercepts [10]. Another
critical source of heat in-leak in the LHC cryostat stems from the 50,000 m2 lateral
surface area of the cold mass—the equivalent of seven soccer fields—receiving heat
by radiation and residual gas conduction from the 80 K thermal shield in case a 5 K
shield is not used or difficult to thermalize at its nominal temperature. It was
therefore necessary to compare, on samples of practical geometry and size, the
performance of several reflective single-layer and multilayer systems at low
boundary temperature and in different conditions of insulation vacuum. This was
done on a dedicated set-up housing a 3 m2 cylindrical sample of the insulation
system to be tested, with possibility of varying the warm and cold boundary
temperatures [11]. It confirmed that very low heat fluxes can be achieved by pas-
sive, reflective systems in good vacuum, and that multilayer systems show a clear
advantage to limit heat flux runaway in case of degraded insulation vacuum.
An opportunity to build rapidly a quasi-full-scale, twin-aperture prototype dipole
was provided by the termination of series production of superconducting magnets
for the HERA proton ring. Two sets of spare Nb-Ti dipole coils from the HERA
production line were integrated into a common yoke to build a 9.15 m long
twin-aperture prototype (TAP), to be operated in superfluid helium in order to
simulate a LHC dipole: it was estimated that a field of 7.5 T could be produced with
the coils powered at 8610 A. The TAP magnet needed a dedicated cryostat to be
operated and tested in superfluid helium. Conversely, this provided a unique
opportunity for experimenting, for the first time and almost in full scale, the design
of the LHC cryostats and for studying some of the thermal and mechanical prob-
lems associated with their cryogenic operation.
The TAP cryostat [12] was designed at CERN along the principles of the
conceptual LHC cryostats (Fig. 3.6) and built by industry. In addition to the
magnet, it housed cryogenic distribution pipelines—some of which were used for
heat interception—and featured three glassfibre/epoxy posts supporting the magnet,
the central one fixed longitudinally and the other two allowed to follow the thermal
contraction of the cold mass via rollers on tracks fitted to the vacuum vessel. These
rollers and tracks were also used during assembly to insert the cold mass into the
74 P. Lebrun

Fig. 3.6 Transverse

cross-section of the TAP
magnet and cryostat.
1 HERA-type
superconducting coils;
2 magnet collars; 3 magnet
yoke; 4 magnet shrinking
cylinder; 5 cold bore tube;
6 helium vessel; 7 screen at
5 K; 8 thermal shield at 80 K
with MLI; 9 vacuum vessel;
10 support post; A & B 1.8 K
helium pipes; C & D 4.5 K
helium pipes; E liquid
nitrogen pipes

vacuum vessel. Demountable shipping restraints were provided to take transport

acceleration loads. A liquid-nitrogen cooled thermal shield wrapped with 30 layers
of MLI intercepted most of the radiative heat in-leak from room temperature, and a
second screen surrounded the cold mass and pipeline assembly, actively cooled
from 4.5 K pipes. The all-welded, austenitic stainless steel (AISI 304L) helium
enclosure had a design pressure of 2 MPa to allow fast cooldown/warmup and resist
resistive transitions of the magnet. The vacuum vessel was made of unalloyed
construction steel (with austenitic stainless steel flanges), for reasons of magnetic
shielding of the stray flux and cost in view of a large series production. Its inner
surface was protected by low-outgassing epoxy paint compatible with high-vacuum
operation. The cylindrical shell was provided with reinforcement rings at the lon-
gitudinal positions of the cold mass supports.
The TAP magnet was integrated into its cryostat by industry and shipped to CEA
Saclay for cold tests. The magnet and cryostat performed very well, with maximum
field of 8.25 T reached after seven resistive transitions, excellent leak-tightness of
the superfluid helium enclosure, heat in-leak below 1 W at 1.8 K and about 4.6 W
at 4.5 K [13]. Following resistive transitions of the magnet, high values of peak
pressure, up to 3 MPa, were measured in the annular space between coils and cold
bore tube at mid-length, thus raising concern about the choice of design pressure for
the helium enclosure. A dedicated study based on a hydrodynamic model using
thermo-physical properties of helium [14] confirmed the measurements and gave
indications to limit this peak pressure, particularly when considering longer mag-
nets. From the cryostat point of view, the TAP project established the soundness of
the design principles and boundary conditions, confirmed performance expectations
from design calculations and measurements on partial components and demon-
strated the feasibility of industrial production at market price—including quality
assurance, licencing procedures and operational safety.
3 Twenty-Three Kilometres of Superfluid Helium … 75

3.3 Prototype Cryostats and String Tests

Based on the previous experience, a small series of cryostats were designed at

CERN and procured from industry to house the first generation of 10 m long
prototype dipole magnets, to be tested individually and later connected and oper-
ated in a string. The cryostats also had to accommodate the cryogenic distribution
pipework sized for feeding a full LHC sector. Design requirements included heat
in-leaks meeting the allowed thermal budget at the three temperature levels or
ranges in the LHC cryogenic system (50–75 K, 4.5–20 K and 1.9 K), provision of
a stable, precise and reproducible mechanical support for the magnet, withstanding
of thermal and mechanical transients associated with rapid cooldown and magnet
resistive transitions, as well as all possible failure modes. Technical solutions
adapted to large series production were favoured.
The resulting design [15], inspired by the TAP cryostat but with a number of
variants, is shown in Fig. 3.7. The cold mass support system still used three
glassfibre/epoxy posts with two levels of heat interception, but the functions of
assembly and of accommodation of longitudinal contractions were now handled
separately, the former provided by a removable assembly trolley and the latter by
sliding plates incorporating low-friction material. The support posts proper were
developed and their construction optimized based on mechanical and thermal
performance measurements [16]. The thermal shield at 80 K, made in commercially
pure aluminium (AA 1100) to ensure good azimuthal temperature uniformity,
consisted of a rigid lower tray incorporating a welded extruded aluminium profile
providing the cooling channel, onto which a roughly cylindrical upper shell was
tack-welded in four longitudinal sections. This compound construction provided
sufficient mechanical rigidity and azimuthal thermal conduction to limit deflections
of the shield during thermal transients, and thus avoid damage or thermal shorts.

Fig. 3.7 Transverse

cross-section of
first-generation prototype
dipole and cryostat. 1 cold
bore; 2 magnet coils;
3 magnet collars; 4 magnet
yoke; 5 magnet shrinking
cylinder and helium vessel;
6 bus bars; 7 He II heat
exchanger bore; 8 cold mass
multilayer insulation;
9 thermal shield with
multilayer insulation;
10 vacuum vessel; 11 support
post; 12 1.8 K GHe line;
13 10 K GHe line; 14 4.5 K
He line; 15 2.2 K LHe line;
16 50 K GHe supply;
17 70 K GHe return
76 P. Lebrun

The actively cooled 4.5 K screen surrounding the cold mass of the TAP cryostat
was replaced here by a 10-layer blanket of MLI, partially thermalized onto the
4.5 K lines but insulated from the cold mass and 2 K lines by polyester spacers.
Another, important novelty with respect to the TAP cryostat was the first engi-
neering study and optimization of magnet and cryostat interconnection. To maxi-
mize the magnet filling factor of the machine and thus the beam energy reach, the
interconnection space must be kept as short as possible, while allowing to reliably
perform, inspect and test some 14 orbital welds joining 11 pipes. Space for auto-
matic orbital welding and cutting machines, as well as for “clam shells” for local
leak-testing of welds, was taken into account. The interconnection also includes all
bellows accommodating construction and alignment tolerances, as well differential
thermal contractions and displacements occurring in all possible phases of opera-
tion, including failure modes (as a reminder, the LHC cold mass shrinks by some
80 m in perimeter upon cooldown). Bellows specifications, including travel,
misalignments, maximum rigidity and fatigue lifetime were established, pointing to
hydro-formed multiply construction with external shells against buckling where
needed. Samples were procured from several bellows manufacturers and exten-
sively tested, thus providing a qualified industrial basis for future series
procurement.
The design of the prototype dipole cryostat was adapted to house the main
quadrupole, chromaticity sextupole and corrector magnets, beam position monitors
and local cryogenic equipment as well as a vacuum barrier segmenting the insu-
lation vacuum, all included in a so-called short-straight-section cryostat [17], and a
prototype was built and tested.
In parallel with these activities, preparation was made to perform cryogenic tests
of the magnets and cryostats. A 7200 m2 test hall already equipped with a 6 kW at
4.5 K cryogenic refrigerator was further provided with a low-pressure helium
pumping unit capable of 6 g/s at 1 kPa, upgradeable to 18 g/s with the addition of a
cold booster compressor. Cooldown and warmup units, each providing up to
120 kW down to 100 K by vaporization of liquid nitrogen [18] were installed. The
test station was planned to develop in a modular way following the ramp-up of
series production, with up to twelve parallel test benches, the first two of which
were commissioned to test prototypes [19]. Each bench, equipped with 18 kA
current leads, could be operated independently of the others, only sharing—when
and as much as needed—the common facilities such as normal liquid helium
supply, pumping capacity at 1 kPa and gaseous helium flow for cooldown and
warmup.
At the same time it was noted that precision thermal measurements on cryostats
mounted on magnet test benches would be difficult to perform, due to parasitic heat
loads (in particular the comparatively high heat in-leak coming from the
anti-cryostats installed in the magnet apertures to perform magnetic measurements)
as well as to schedule interference (reaching thermal steady-state requires long
stabilization time incompatible with magnet turnaround on test benches). It was
therefore decided to build a full-scale thermal model [20] from a 10 m prototype
cryostat equipped with a dummy cold mass and finely instrumented. The cryostat
3 Twenty-Three Kilometres of Superfluid Helium … 77

thermal model could be operated independently of the magnet tests station,

allowing to vary parameters such as insulation vacuum and temperature levels, and
providing precision of ±0.03 W at 1.8 K, ±0.1 W at 5–20 K and ±3 W at 50–
75 K. Heat load measurements in nominal operating conditions on the thermal
model showed excellent agreement with calculated values [21]. Moreover, several
off-design runs enabled to measure the sensitivity of the cryostat heat in-leaks upon
the temperature of heat intercepts and thermal shield, as well as on the residual
pressure in the vacuum insulation space, also in good agreement with the calcu-
lations. The thermal performance of the prototype cryostats was therefore well
understood in nominal and quasi-nominal conditions, and a mathematical model
benchmarked on the experimental measurements could be used reliably to quan-
titatively assess its degradation in off-design conditions [22]. The pending question
of the inner screen surrounding the cold mass, either “floating” or tentatively
thermalized on the 5 K line, was then revisited [23], using realistic values of
impedance for the thermalization contacts and net-type insulating spacers in MLI:
the study indicated potential improvement for a thermalized system, from
60 mW/m2 down to 30 mW/m2 in good insulation vacuum (10−4 Pa), at the cost of
increased complexity in cryostat assembly.
Another pending issue was that of using normal construction steel for the
cryostat outer vessel, normally operating at room temperature but possibly subject
to being cooled below the ductile-to-brittle transition temperature in case of
catastrophic loss of insulation vacuum. This was tested on a full scale cryostat
equipped with a dummy cold mass initially cooled at 80 K [24]. Upon breach of the
vacuum with gaseous helium, heat transfer to the cold mass induced water con-
densation and frost formation on the outer vessel, but its minimum temperature
reached at 260 K remained above the acceptable limit. Moreover, temperature
evolution was slow enough to allow detection of the accident and implementation
of remedial action, e.g. by bringing additional pumping speed to restore, at least
partially, the insulation vacuum.
The prototype magnets in their cryostats were then assembled stepwise into a test
string [25], a full-scale working model of the elementary cell of the machine lattice
and local cryogenics, installed on a 1.4 % ramp corresponding to the maximum
slope of the LHC tunnel, with a final length of 50 m [26]. The test string (Fig. 3.8)
enabled us to practice assembly, leak-tightness and alignment procedures, then
mechanical stability and vacuum measurements upon controlled cooldown of the
65-t cold mass, and finally cryogenic operation and control, magnet powering and
discharge, handling of resistive transitions and recovery [27]. Vacuum and thermal
transients were investigated, in particular as concerns non-standard operation and
failure modes [28]. In spite of the absence of beam, the test string also served as an
excellent training ground for young recruits, who eventually had to operate from a
dedicated control room a compound system integrating several accelerator tech-
nologies, with many measurements and control loops. Finally, it constituted an
important demonstrator which contributed to the approval of LHC construction by
the CERN Council.
78 P. Lebrun

Fig. 3.8 First-generation LHC Test String (CERN photo)

In the last round of studies before project approval, several important changes
were implemented in the configuration of the LHC and its cryogenic system. To
minimize their number and the associated manufacturing and testing costs, the LHC
dipole magnets had their length extended from the previous 10 m to the maximum
3 Twenty-Three Kilometres of Superfluid Helium … 79

which could be practically transported in European roads (40-t limit) and conve-
niently handled for assembly, testing and installation, i.e. 15 m, thus requiring new,
longer cryostats. In order not to lose in useful aperture, the cold mass had to be
curved, with a horizontal sagitta of 9 mm, but the cryostats could remain straight.
As concerns cryogenic system architecture, grouping all refrigeration equipment in
the five points around the ring (instead of eight) already equipped with technical
infrastructure led to significant cost savings, and allowed for redundancy at
part-load, but cryogenic flows had then to be distributed over the full 3.3 km length
of the sectors (instead of half-sectors). This required larger-diameter pipes which
became difficult to integrate in the cryostat transverse cross-section. Consequently,
all distribution pipework and local equipment was moved to a compound cryogenic
line running in the tunnel along the magnets, and the remaining pipework in the
cryostats was then limited to that needed for operation of the cryostat proper.
Another important consequence is that the cryogenic line could be installed,
commissioned and tested independently of the progress in magnet installation, thus
bringing valuable schedule flexibility in the last phase of project construction [29].
Moreover, it had been observed that in view of the large size of the LHC ring,
“passive” cryogenics, i.e. magnet cryostats and cryogenic distribution represented
the largest share of the cost breakdown, in fact larger than “active” cryogenics, i.e.
the refrigeration plants. Consequently, any simplification of the cryostats and
cryogenic distribution scheme, even though it may result in increasing the refrig-
eration duty, would concur to lowering the total cost. A simplified cryogenic
scheme was then implemented [30], with local cooling loops extending over the
length of a full lattice cell (106.9 m) instead of the previous half-cell, implemen-
tation of local 4.5–2.2 K sub-cooling heat exchangers in each local cooling loop
allowing the suppression of one pipe inside the distribution line, and a large overall
reduction in the number of cryogenic components.
A second generation of prototype cryostats for dipoles [31] and
short-straight-sections [32] was then designed and built, based on the experience
acquired with the first-generation prototypes and incorporating the changes in LHC
configuration described above. In the dipole cryostat (Fig. 3.9) the smaller outer
diameter (914 mm) allowed the shell of the vacuum vessel to be made of standard,
helically welded elements for pipelines, with a wall thickness of 12 mm. To prevent
any risk of embrittlement, the construction steel used (DIN GS-21 Mn5) was
however specified to have a minimum energy absorption of 28 J/cm2 as measured
in an ISO-standard Charpy test at 240 K. The shells were equipped with AISI 304L
end flanges and reinforced, at the longitudinal positions of the magnet supports, by
welded rings and forged brackets for lifting points and alignment fiducials. The
longer, heavier cold mass was still supported on three posts, the longitudinal
positions of which were compromised between minimizing vertical sagitta (Bessel
conditions) and keeping the ends of the cold mass horizontal to ease interconnect
assembly (Airy conditions). The vertical off-axis between magnet and cryostat axes
was reduced to 80 mm, thus requiring shorter support post of 210 mm total height,
redesigned to take higher combined loading of 125 kN in axial compression and
40 kN in shear [33]. The longer thermal shield, a quasi-cylinder cooled from one
80 P. Lebrun

Fig. 3.9 Second-generation prototype cryostats assembled into second Test String (CERN photo)

generatrix, was prone to larger deformations upon thermal transients, with the risk
of damage and unwanted thermal shorts [34]. The technical solution experimented
in the first-generation prototypes, i.e. separating the functions of mechanical rigidity
and azimuthal temperature homogeneity, was further refined towards industrial-
ization: the lower tray was made of a full-length extrusion of A6060 aluminium
3 Twenty-Three Kilometres of Superfluid Helium … 81

Fig. 3.10 Extruded aluminium alloy (A6060) bottom tray of cryostat thermal shield (CERN
photo)

alloy, integrating two 80 mm diameter channels, one used for the cooling, the other
left open (Fig. 3.10). The cooling channel was fitted at the ends with
aluminium-to-stainless steel pipe transitions for interconnection. The upper shell of
the thermal shield made of commercially pure aluminium A1100, was segmented in
five sections (Fig. 3.11). A semi-rigid screen, thermalized to the 5 K pipes pro-
viding heat intercepts on the support posts was tentatively re-introduced in these
cryostats with the aim of further reducing the heat in-leak on the cold mass. The
cryostat interconnect was also redesigned to accommodate installation misalign-
ments of up to ±4 mm in any direction and thermal contractions of up to 48 mm
between the longitudinally fixed central posts of neighbouring dipoles. The
short-straight section cryostat (Fig. 3.12), designed in collaboration with IPN Orsay
(France) incorporated many design features of the dipole cryostat, and showed a
similar transverse cross-section. It was equipped with a vacuum barrier for seg-
menting the insulation vacuum, and a cryogenic service module connected to the
cryogenic distribution line by a multi-pipe “jumper” connection. The service
module also contained beam position monitors, conduction-cooled current leads for
the independently-powered orbit corrector magnets, and instrumentation feed-
throughs. However, all local cryogenic equipment was now moved to the cryogenic
line on the other side of the jumper connection.
82 P. Lebrun

Fig. 3.11 Aluminium A1100 shell of cryostat thermal shield (CERN photo)

Fig. 3.12 Longitudinal cross-section of the second-generation short-straight-section cryostat

3 Twenty-Three Kilometres of Superfluid Helium … 83

Five second-generation dipole cryostats, and two short-straight section cryostats

were procured from industry and assembled around their magnets for cold testing
and installation into a 107 m long second-generation test string [35], the last
large-scale experimental validation of the LHC accelerator systems before entering
construction. A first systematic risk analysis of the LHC cryogenic system [36]
established all possible failure modes and analysed their consequences, thus con-
firming the cryostat design parameters and safety features.

3.4 Industrial Series Production, Installation

and Commissioning

The design of the series cryostats (Fig. 3.13) derived directly from the
second-generation prototypes, the measured thermal performance of which was
used to establish reliably a reference thermal budget for the purpose of sizing the
cryogenic refrigeration plants and distribution system. A “heat-load working group”
involving all stakeholders was established to collect all calculated and measured
data, validate the estimates (Table 3.1) and track possible future evolutions.
Changes of a few manufacturing aspects streamlined the design towards
industrialization in view of series production [37]. Trading moderate improvement

Fig. 3.13 Transverse cross-section of series dipole cryostat

84 P. Lebrun

Table 3.1 Heat in-leaks of 106.9 m long standard cell (6 dipoles and 2 short straight sections)
50–75 K 4.5–20 K 1.9 K
Magnet support posts (W) 157 9.78 1.07
Thermal shield (W) 301 – –
Cold mass MLI (W) – 0.85 11.3
Beam screen (W) – – 1.62
Instrumentation feedthrough (W) – – 4.25
Beam vacuum feedthrough (W) 2.40 – 0.42
Orbit corrector current leads (W) 10.4 2.44 0.53
Beam position monitor (W) – 0.93 0.60
Cryogenic service module (W) – 0.01 0.21
Vacuum barrier (W) 11.6 0.03 0.42
Total heat in-leak (W) 482 14.0 20.4
Average linear heat in-leak (W/m) 4.51 0.13 0.19

Fig. 3.14 Glassfibre-reinforced epoxy support post (CERN photo)

of thermal performance against complexity and cost, the 5 K thermalization of the

screen was finally not retained, and the cold mass was eventually covered with a
10-layer blanket of MLI. The complete rationale for this choice is summarized in
Ref. [38]. Following completion of welding work, the series vacuum vessels were
stress-relieved by vibration to ensure mechanical stability over time. Instead of the
previously used low-outgassing epoxy paint, their inner surface was sandblasted
and simply cleaned to high-vacuum standards, with no other surface treatment.
Magnet support posts (Fig. 3.14) were composed of monolithic 4 mm thick
3 Twenty-Three Kilometres of Superfluid Helium … 85

columns of glass-fibre reinforced epoxy, manufactured by resin-transfer moulding,

a highly automated technique ensuring low production cost and reproducible
quality for series of 4700 units [39]. At the heat intercept locations, external flanges
of aluminium alloy and internal stainless-steel reinforcement rings are glued to the
columns. Aluminium anti-radiation disks and reflective coating on the lower,
warmer section exposed to radiation from the room-temperature outer vessel,
complete the assembly. The 30-layer MLI around the thermal shield and 10-layer
MLI around the cold mass were made of pre-fabricated blankets of 6 mm thick PET
film, double-aluminized with 400 Å, interleaved by a very low weight polyester
spacer. The developed perimeter of each blanket was adjusted to compensate for
their differential thermal contraction with the stainless-steel and aluminium shells
supporting them. Besides more precise manufacturing, pre-fabrication enabled the
multilayer insulation systems to be assembled by non-specialized personnel, using
Velcro™ fasteners, and avoided installation errors, thus ensuring future thermal
performance of the multilayer insulation systems in the LHC ring.
For series production, it was decided to get the cryostats assembled by industry
on the CERN premises, with CERN procuring all components and providing the
contractors with a build-to-print specification, the use of two assembly halls
totalling some 10,000 m2 floor space, lifting and transport facilities, heavy and
specialized tooling, detailed procedures and quality management tools, as well as
with initial training of the execution personnel on the pre-series units, co-assembled
with CERN experts (Figs. 3.15 and 3.16). Assembly of the 1232 dipole and 474
short-straight-section cryostats spanned five years, representing some 850,000 h of
work with personnel numbers—operators and technicians—peaking at 145 [40].

Fig. 3.15 Dipole cryostat assembly hall (CERN photo)

86 P. Lebrun

Fig. 3.16 Short-straight-section cryostat assembly hall (CERN photo)

Following assembly, each cryo-magnet was moved to the test hall using a
specially designed vehicle used for both lifting and transport (Fig. 3.17), connected
to one of the twelve test benches, pumped down, cooled down and cryogenically
tested, with the magnet powered up to 10 % above nominal (Fig. 3.18). Magnetic
measurements in cold conditions were only performed on a fraction of the pro-
duction. After removal from the test bench each cryo-magnet was allocated a
position in the ring based on test results, prepared for installation and lowered to the
tunnel level via a large, elliptical shaft equipped with a 40-t gantry crane with a
hook travel of 54 m. The major axis of 18 m of the elliptical shaft is able to
accommodate the full assembled length of a cryo-dipole, with protection end
covers, horizontally (Fig. 3.19).
Cryostat assembly and cryogenic tests followed the production rate of magnets
in industry (Fig. 3.20), while installation in the tunnel started later and proceeded at
a faster rate. As a consequence, a large buffer storage of assembled and tested
cryo-magnets was available, enabling the optimization of their final installation
positions in the accelerator tunnel in order to reduce dispersion in the magnetic and
geometrical properties of assembled sectors. The cryo-magnets were stored in the
open air for several months, with their ends capped and insulation space kept under
dry nitrogen.
3 Twenty-Three Kilometres of Superfluid Helium … 87

Fig. 3.17 Completed dipole cryo-magnet transported to test hall (CERN photo)

Fig. 3.18 Cryogenic tests of magnets at CERN (CERN photo)

Transport in the tunnel from the shaft to the installation position was done via an
optically guided, electrically powered trolley at low velocity (3 km/h) to minimize
inertial forces in view of the large masses transported. On location, the cryo-magnet
weight was loaded to a transverse transfer platform, and then transferred onto its
pre-positioned final support jacks (Fig. 3.21). Electrical magnet interconnections
and hydraulic cryostat interconnections were then performed. This was done by
88 P. Lebrun

Fig. 3.19 Lowering of

cryo-magnet to LHC tunnel
via elliptical shaft (CERN
photo)

industry in the framework of a build-to-print contract. CERN developed the

methods, special tooling and quality assurance procedures to be used by the con-
tractor, and provided initial training of its personnel [41]. Some 40,000 cryogenic
pipe welds were reliably executed using automatic TIG orbital welders (Fig. 3.22).
The inner part of a completed cryo-magnet interconnection, with the MLI blankets,
thermal shield and outer vacuum sleeve removed, is shown in Fig. 3.23. The domed
ends of the helium vessels, as well as the reflective multilayer insulation blankets
covering the cold mass and the thermal shield, are clearly visible on either side of
the interconnection.
Following global pressure and leak-tightness tests, each 3.3 km sector of the
LHC was cooled down and powered. Total sector heat in-leaks (Fig. 3.24) were
measured within the estimated values [42], a final validation of the quality of the
design, construction and installation of the cryostats.
3 Twenty-Three Kilometres of Superfluid Helium … 89

Fig. 3.20 Series production of LHC dipole cryo-magnets

Fig. 3.21 Transverse positioning of cryo-magnet on final location (CERN photo)

90 P. Lebrun

Fig. 3.22 Cryogenic piping interconnection in LHC tunnel by orbital TIG welding (CERN photo)

Fig. 3.23 Cryostat interconnection inner piping (CERN photo)

3 Twenty-Three Kilometres of Superfluid Helium … 91

Continuous cryostat IT 1 IT 2
800
700
600
500
[W] 400 On average,
300 measured values
200 20 % lower than
100
0 design estimates
DR Meas. DR Meas. DR Meas. DR Meas. DR Meas.

Sector 1-2 Sector 4-5 Sector 5-6 Sector 6-7 Sector 8-1

Fig. 3.24 Measured heat in-leaks of LHC sectors at 1.9 K (DR design report estimates; IT inner
quadrupole triplet)

3.5 Concluding Remarks

The LHC contains the largest system of advanced, high-performance cryostats

housing superconducting accelerator devices, and may thus be considered, mutatis
mutandis, as reference for future, similar projects. Several lessons can be drawn
from this experience.
The technology of superfluid helium can be reliably applied to very large sys-
tems built by industry. In particular, the design principles, construction techniques
and quality assurance methods used in normal helium cryostats are fully adequate to
ensure leak-tightness of superfluid helium enclosures.
Cryostats of superconducting accelerator devices, magnets or RF cavities, must
meet complex and often conflicting geometrical, mechanical and thermal require-
ments, at the interface between accelerator physics and technology requirements,
industrial manufacturing constraints and cost. Their design must therefore follow
closely the changes in configuration of the project and sometimes drive them. It is
therefore essential for the cryostat designers to be fully integrated with the project
team and in close contact with the magnet, RF, cryogenics and vacuum groups.
Within the cryogenic system, the total cost of the cryostats and distribution lines
often exceeds that of the refrigeration plants. Global optimization of the complete
system is therefore necessary, which may lead to simplify cryostat construction at
the cost of loss in thermal performance, rather than designing for lowest possible
heat-in-leak.
The large amount of thorough design and development work performed at CERN
and in partner laboratories on the LHC cryostats has contributed to significantly
increase knowledge and confidence in the design, engineering and quality assurance
methods for this type of equipment. It is interesting to note that this work has already
found an application in another project. Cryostats similar to those of the LHC dipoles
are housing the superconducting magnets of the beam line for the T2K long-baseline
neutrino experiment at the J-PARC laboratory in Tokai, Japan [43]. On the training
and education side, a tutorial on cryostat design inspired from the development of
LHC cryostats was first prepared for the CERN Accelerator School
92 P. Lebrun

“Superconductivity and Cryogenics for Accelerators and Detectors” in 2002 [44]

and is now part of the curriculum of the yearly European Course of Cryogenics
commonly organized by the Technical University of Dresden, the Wroclaw
University of Technology and the Norwegian University of Science and Technology
Trondheim. This has been further developed as a course in a more recent venue of
the CERN Accelerator School on “Superconductivity for Accelerators” [45].

Acknowledgments Having been involved in this work from the onset, the author has tried to
convey his strong conviction that the success of LHC cryostats is the result of teamwork by many
contributors at CERN, in partner laboratories and in industry. A number of them appear as
co-authors in the reference list below. The expertise, dedication and hard work of all should be
acknowledged.

References

1. LHC Design Report, vol. I, The LHC Main Ring, CERN-2004-003 (2004). ISBN
92-9083-224-0
2. P. Lebrun et al., Cooling strings of superconducting devices below 2 K: the helium II bayonet
heat exchanger. Adv. Cryo. Eng. 43A, 419–426 (1998)
3. P. Lebrun, Cryogenics for the Large Hadron Collider. IEEE Trans. Appl. Supercond. 10,
1500–1506 (2000)
4. A. Asner et al., A feasibility study of possible machine options, in Proceedings of
ECFA-CERN Workshop on Large Hadron Collider in the LEP Tunnel, Lausanne,
CERN-84-10-V-1 (1984), pp. 49–142
5. G. Claudet, R. Aymar, Tore Supra and helium II cooling of large high-field magnets. Adv.
Cryo. Eng. 35A, 55–67 (1990)
6. G. Claudet et al., Preliminary study of a superfluid helium cryogenic system for the Large
Hadron Collider, in Proceedings of Workshop on Superconducting Magnets and Cryogenics,
Brookhaven BNL 52006 (1986), pp. 270–275
7. V. Baglin et al., Cryogenic beam screens for high-energy particle accelerators, in Proceedings
of ICEC24-ICMC 2012 Fukuoka, Cryogenics and Superconductivity Society of Japan (2013),
pp. 629–634
8. G. Claudet et al., Conceptual study of the superfluid helium cryogenic system for the CERN
Large Hadron Collider, in Proceedings of ICEC12 Southampton, Butterworth (1988),
pp. 497–504
9. T. Nicol, SSC Collider Dipole Cryostat, Chap. 2, this book
10. H. Danielsson et al., Precision heat inleak measurements on cryogenic components at 80 K,
4.2 K and 1.8 K. Cryogenics 32(ICEC Supplement), 215–218 (1992)
11. P. Lebrun et al., Investigation and qualification of thermal insulation systems between 80 K
and 4.2 K. Cryogenics 32(ICMC Supplement), 44–47 (1992)
12. M. Granier et al., Design and construction of a superfluid-helium cryostat for a ten-meter long
high-field superconducting dipole magnet. Cryogenics 30(September Supplement), 98–102
(1990)
13. M. Granier et al., Performance of the twin-aperture dipole for the CERN LHC, in Proceedings
of EPAC’92, JACoW (1992), pp. 1414–1416
14. P. Lebrun et al., Investigation of quench pressure transients in the LHC superconducting
magnets. Cryogenics 34(ICEC Supplement), 705–708 (1994)
15. J.-C. Brunet et al., Design of LHC prototype dipole cryostats. Cryogenics 32(ICEC
Supplement), 191–194 (1992)
3 Twenty-Three Kilometres of Superfluid Helium … 93

16. M. Blin et al., Design, construction and performance of superconducting magnet support posts
for the Large Hadron Collider. Adv. Cryo. Eng. 39A, 671–678 (1994)
17. W. Cameron et al., Design and construction of a prototype superfluid helium cryostat for the
Short Straight Sections of the CERN Large Hadron Collider. Adv. Cryo. Eng. 39, 657–662
(1994)
18. V. Benda et al., Cryogenic infrastructure for superfluid helium testing of LHC prototype
superconducting magnets. Adv. Cryo. Eng. 39, 641–648 (1994)
19. V. Benda et al., Cryogenic benches for superfluid helium testing of full-scale prototype
superconducting magnets for the CERN LHC project. Cryogenics 34(ICEC Supplement),
733–736 (1994)
20. L. Dufay et al., A full-scale thermal model of a prototype dipole cryomagnet for the
CERN LHC project. Cryogenics 34(ICEC Supplement), 693–696 (1994)
21. V. Benda et al., Measurement and analysis of thermal performance of LHC prototype
cryostats. Adv. Cryo. Eng. 41, 785–792 (1996)
22. G. Riddone, Theoretical Modelling and Experimental Investigation of the Thermal
Performance of LHC Lattice Cryostats. Doctoral Thesis, Politecnico di Torino (1996)
23. G. Ferlin et al., Comparison of floating and thermalized multilayer insulation systems at low
boundary temperature, in Proceedings of ICEC16-ICMC 1996, North-Holland (1997),
pp. 443–446
24. P. Lebrun et al., Experimental investigation of accidental loss of insulation vacuum in a LHC
prototype dipole cryostat. Adv. Cryo. Eng. 41, 799–804 (1996)
25. A. Bézaguet et al., The superfluid helium cryogenic system for the LHC test string: design,
construction and first operation. Adv. Cryo. Eng. 41, 777–784 (1996)
26. A. Bézaguet et al., Cryogenic operation and testing of the extended LHC prototype magnet
string, in Proceedings of ICEC16-ICMC 1996, North-Holland (1997), pp. 91–94
27. M. Chorowski et al., Thermohydraulics of quenches and helium recovery in the LHC
prototype magnet strings. Cryogenics 38, 533–543 (1998)
28. P. Cruikshank et al., Investigation of thermal and vacuum transients on the LHC prototype
magnet string, in Proceedings of ICEC16-ICMC 1996, North-Holland (1997), pp. 681–684
29. V. Benda et al., Conceptual design of the cryogenic system for the Large Hadron Collider, in
Proceedings of EPAC’96 Sitges, JACoW (1996), pp. 361–363
30. M. Chorowski et al., A simplified cryogenic distribution scheme for the Large Hadron
Collider. Adv. Cryo. Eng. 43A, 395–402 (1998)
31. J.-C. Brunet et al., Design of the second-series 15 m LHC prototype dipole magnet cryostats.
Adv. Cryo. Eng. 43A, 435–441 (1998)
32. W. Cameron et al., The new superfluid helium cryostats for the Short Straight Sections of the
CERN Large Hadron Collider. Adv. Cryo. Eng. 43A, 411–418 (1998)
33. M. Mathieu et al., Supporting systems from 293 K to 1.9 K for the Large Hadron Collider.
Adv. Cryo. Eng. 43A, 427–434 (1998)
34. G. Peon, G. Riddone, L.R. Williams, Analytical model to calculate the transient
thermo-mechanical behaviour of long thin structures cooled from a pipe: application to the
LHC dipole thermal shield, in Proceedings of ICEC16-ICMC 1996, North-Holland (1997),
pp. 477–480
35. E. Blanco et al., Experimental validation and operation of the LHC Test String 2 cryogenic
system. Adv. Cryo. Eng. 49, AIP Conf. Proc. 710, 233–240 (2004)
36. M. Chorowski, P. Lebrun, G. Riddone, Preliminary risk analysis of the LHC cryogenic
system. Adv. Cryo. Eng. 45B, 1309–1316 (2000)
37. N. Bourcey et al., Final design and experimental validation of the thermal performance of the
LHC lattice cryostats. Adv. Cryo. Eng. 49, AIP Conf. Proc. 710, 487–493 (2004)
38. P. Lebrun, V. Parma, L. Tavian, Does one need a 4.5 K screen in cryostats of superconducting
accelerator devices operating in superfluid helium? Lessons from the LHC. Adv. Cryo. Eng.
59A, AIP Conf. Proc. 1573, 245–252 (2014)
94 P. Lebrun

39. V. Parma et al., The LHC cryomagnet supports in glass-fibre reinforced epoxy: a large-scale
industrial production with high reproducibility in performance. Adv. Cryo. Eng. 54, AIP Conf.
Proc. 986, 211–218 (2008)
40. A. Poncet, V. Parma, Series-produced helium II cryostats for the LHC magnets: technical
choices, industrialization, costs. Adv. Cryo. Eng. 53A, AIP Conf. Proc. 985, 739–746 (2008)
41. J-Ph Tock et al., The interconnections of the LHC cryomagnets at CERN: strategy applied and
first results of the industrialization process. IEEE Trans. Appl. Supercond. 18(2), 116–120
(2008)
42. S. Claudet at al., Cryogenic heat load and refrigeration capacity management at the Large
Hadron Collider, in Proceedings of ICEC22—ICMC 2008, KIASC (2009), pp. 835–840
43. T. Nakamoto et al., Construction of superconducting magnet system for the J-PARC neutrino
beam line. IEEE Trans. Appl. Supercond. 20(3), 208–213 (2010)
44. P. Lebrun, Design of a cryostat for superconducting accelerator magnets: the LHC main
dipole case, in Proceedings of CAS Superconductivity and Cryogenics for Accelerator and
Detectors, CERN-2004-008, ISBN 92-9083-230-4 (2004), pp. 348–362
45. V. Parma, Cryostat design, in Proceedings of CAS Superconductivity for Accelerators,
CERN-2014-005 (2014), pp. 353–399. ISBN 978-92-9083-405-2
Chapter 4
The Superfluid Helium On-Orbit Transfer
(SHOOT) Flight Demonstration

Michael DiPirro

Abstract Cryostats for space applications have unique requirements including

significant weight and volume limits as well as the need to take into account zero
gravity when designing fluid systems. This chapter describes the design, testing and
space operations of the Superfluid Helium On-Orbit Transfer (SHOOT) flight
demonstration. Topics covered in the design of the SHOOT cryostats include:
structural and thermal insulation design, phase separation and liquid acquisition
systems, the use of thermomechanical pumps, instrumentation and safety.
Operations on both the ground and on orbit are also discussed.

4.1 Introduction

In the 1980s NASA recognized that there would be an increasing demand for
cryogenic instruments in space. In particular, cooling to below 4 K would become
routinely required for advanced astronomical instruments. In that period cry-
ocoolers for this temperature range were in their infancy; none provided long life,
high reliability, and low input power required for space flight. This left stored
cryogens, uniquely liquid helium, to be required for many future missions.
Refueling helium dewars in orbit was being studied to extend these instrument’s
lifetimes in the early days of the space shuttle and in the concept phase of the
International Space Station [1]. The Superfluid Helium On-Orbit Transfer
(SHOOT) Flight Demonstration was proposed in 1984 to carry out a series of tests
of the technology and physics required to store and move liquid helium from a large
depot into individual dewars aboard various satellites.
The experimental components for SHOOT, which drove the requirement for an
on-orbit test, were the liquid acquisition devices (LADs). These LADs could only
be thoroughly tested in the proper acceleration conditions, i.e., in a very low
acceleration for an extended period of time, in orbit. Of course one must demon-

M. DiPirro (&)
Code 552, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 95

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_4
96 M. DiPirro

Fig. 4.1 Schematic of one of TRANSFER LINE

the SHOOT dewars. A TM TRANSFER LINE COUPLER TO OTHER DEWAR
pump; B liquid acquisition EMERGENCY HE VENT
device; C HeI/HeII phase VENT LINES SHROUD
separator (also known as the HE VENT E
low-flow phase separator); SHROUD
E
D high-flow phase separator;
HIGH
E anti-thermoacoustic LOW FLOW
FLOW
VENT LINE
oscillation volumes; VENT
F venturi; G fill line LEGEND LINE

PRESSURE
CHECK/
RELIEF VALVE

BURST
DISC
BELLOWS

1/2 INCH D
VALVES C F

3/4 INCH B
VALVES
MANUAL
G
VALVES

BALL A
VALVE

PRESSURE
GAUGE

strate the ability to control the fluid position during a number of different states of
the system: precooling a tank, filling or draining a tank, and during adverse
accelerations. To provide these conditions it is therefore necessary to implement an
entire system demonstration in orbit. In the design of the space demonstration
consideration was given to scaling of the system to larger sizes, hold times of the
liquid helium tank before launch, and many other practical limitations. There were
also several opportunities that presented themselves in a technology-only mission,
namely obtaining engineering data as a primary goal. SHOOT also served as a test
bed for intelligent machine control of a complex process in orbit.
The SHOOT payload consisted of two dewars connected by a flexible transfer
line. The two dewars, labeled “port” and “starboard”, contained removable cryostat
inserts (“cryostats”) on which the experiments’ components, the valves and
plumbing were mounted. See Fig. 4.1 for the schematic of one of the SHOOT
dewars. The two dewars were identical, but the cryostats contained different LADs
(Fig. 4.2) and liquid/vapor detectors to measure the liquid position on orbit.
Superfluid helium was pumped between the two dewars using the thermome-
chanical, or fountain, effect that allowed heat to be directly converted to a pressure
differential across a material with fine pores.
During the SHOOT design phase the shuttle Challenger suffered a catastrophic
failure during launch. This led to a fundamental rethinking of the way the remaining
4 The Superfluid Helium On-Orbit Transfer … 97

(a) 56 cm (b)

VANES
INTERFACE SPONGE
TO TM PUMP

TM PUMP

TRANSFER
LINE
89 cm
SCREENED
CHANNELS

VANES

Fig. 4.2 Arrangement of a the screen channels in the starboard helium tank and b the vanes and
sponge reservoir in the port helium tank. The vanes are fewer in number and extend out to the wall
of the helium tank contrary to the schematic illustration

shuttles would be used and to an emphasis on safety. Up until this point the plan
was to use SHOOT on two missions: the first to explore the fundamental aspects of
helium management in orbit, and the second (SHOOT-II) to explore the human
interfaces in a real dewar-to-dewar helium refill. When the shuttle manifest con-
tracted, some of SHOOT-II’s goals were folded into SHOOT. A manual, astronaut-
Extra Vehicular Activity (EVA) compatible connector that astronauts would
exercise during an EVA on SHOOT-II was dropped due to cost and schedule, but
an expert system computer program was planned along with an astronaut operated
helium transfer from within the shuttle cabin.
In the ensuing years the reality of the cost of space flight and liquid-helium
servicing also became apparent—science missions would not be able to afford to
resupply on orbit. In most cases the bill was higher than the original science
spacecraft cost. The raison d’etre for SHOOT would now be flight-proving the
cryogenic components already developed or under development and demonstrating
the technology of management of a cryogenic liquid on orbit.
98 M. DiPirro

4.2 Design Considerations

The SHOOT system was designed for ground testing as well as to function in space.
The dewars had to be able to vent while horizontal or vertical. This involved careful
positioning of the pumps and liquid/vapor detectors, and adequately sized ground
support equipment vacuum pumps. The emergency vent system had to work when
surrounded by air or in vacuum, and as far as possible, the operation of the system
had to be demonstrated on the ground as well as on orbit.
SHOOT was a short duration space mission (days rather than months or years),
which allowed some compromises on heat leaks compared to the extreme thermal
isolation required for long life missions. The SHOOT tank was structurally sup-
ported from the outer vacuum shell by S-glass composite straps that were sized for
infinite fatigue life rather than for the short duration required for testing and launch.
This permitted a relatively simple suspension of the two vapor-cooled shields
(VCSs) by attaching clamps to the straps. In fact, two VCSs were all that were
required rather than the more standard three VCSs, which, again sacrificed thermal
performance in favor of simplicity and size of the helium tank. In this configuration
with a room temperature outer vacuum shell, the heat leak to the liquid helium was
about 180 mW.

4.2.1 Structural Requirements

The dewar structural requirements for inertial, thermal and pressure loads are shown
in Table 4.1. The largest inertial load requirement was to survive a very hard
landing! (Fig. 4.3).
In addition the dewars were designed as fracture critical hardware, with a
fracture control plan and provisions for non-destructive testing. The non-destructive
testing included ultrasonic inspection of the raw material per MIL-STD-2154,
special dye penetrant inspection per MIL-STD-6866 Type I, Method C or D, and
radiography of welds per MIL-STD-453.
The dewars were designed for two launches and landing and a complete qual-
ification test program with a safe-life safety factor of 4.
Bolted hardware was lock-wired to assure that a positive torque was maintained.

Table 4.1 Inertial loads used in the SHOOT design

Axis Preliminary coupled load Updated coupled load (G) Verification load (G)
(Gravity, G)
X ±15.7 ±12.0 ±5.7
Y ±4.7 ±3.8 ±3.1
Z ±8.4 ±6.0 ±5.5
4 The Superfluid Helium On-Orbit Transfer … 99

Fig. 4.3 The space shuttle Endeavor landing a Kennedy space center. The “slapdown” of the nose
wheel in a worst case landing is the most severe inertial load for the design

4.3 Dewar and Cryostat Details

4.3.1 Dewar Fabrication Details

The SHOOT Payload was mounted to a cross-bay carrier on the space shuttle (see
Fig. 4.4).
Each dewar consisted of five main components (Fig. 4.1): the cryostat (which is
discussed in the next section), fiberglass support straps, helium tank, inner and outer
vapor cooled shields (IVCS/OVCS), and main shell. The dewars were designed to
meet the shuttle safety requirements and the cryostats were readily serviceable. The
dewar design loads were based on a combination of inertial and thermal loads, the
maximum internal pressure of 410 kPa, and the support strap preload. Each load
(Table 4.1) was multiplied by an appropriate factor of safety and then applied to a
payload model in a NASTRAN finite element structural analysis. In general,
highest element loads were extracted for each component so that the component
was designed based on those loads.
An uncertainty factor of 1.25 was applied to these loads. The SHOOT dewars
and components were designed to the preliminary loads. The SHOOT support
structure was designed to the updated loads. The verification load was applied
during vibration testing.
Each dewar had three straps on each end that were located 120° apart. The angle
of the straps to the tank was chosen to minimize the tank to vacuum shell differ-
ential contraction effect on the straps’ tension. The straps were fabricated by
Structural Composites Industries from Owens-Corning Fiberglass Type S-2 high
100 M. DiPirro

Fig. 4.4 SHOOT at Kennedy space center. Upper left SHOOT on the cross bay bridge. Upper
right SHOOT mounted in STS-57. Lower center SHOOT is mounted near the top of the payload
bay just behind the Spacehab module

strength roving, bonded with SCI REZ 081 epoxy resin. The fiberglass was wound
around two bobbins with a center-to-center length of 133 mm. The width of each
strap was 20 mm and the thickness of each leg was 1.3 mm. The straps could only
carry tensile loads, thus a preload of 11,000 N, based on inertial, thermal, and
pressure loads, was applied to ensure that the straps were always in tension. The
straps were instrumented with two strain gages to accurately measure the applied
preload. The straps underwent a maximum of 12,000 cycles during ground test and
flight, with a maximum load of 20,000 N. Several straps were tested from the
production lot and showed an ultimate strength of greater than 90,000 N and a
fatigue life (to the maximum load) of greater than 250,000 cycles. Additional
testing showed that there was only a 10 % reduction in ultimate strength after
12,000 cycles [2].
The helium tanks had a 206 L helium capacity. The cold plate of the cryostat
was mounted to the helium tank on the forward end with an indium seal and a
closeout cover was mounted on the aft end also with an indium seal. The barrel of
4 The Superfluid Helium On-Orbit Transfer … 101

the helium tank was formed from a rolled cylinder of 3.18 mm thick 2219 T37
aluminum sheet. The two heads were machined from a 15 cm thick plate of 2219
T37 aluminum and welded to each end of the cylinder. The tank then was heat
treated to the T87 condition, machined to final dimensions, and qualified by
pressurizing to 735 kPa. The stresses that developed from this pressure, envelope
the maximum stress due to inertial, thermal, internal pressure, and strap preload
which would be seen during flight. Additional units were proof-tested to 575 kPa.
During dewar assembly each tank was covered by a five layer multilayer insulation
(MLI) blanket, consisting of double aluminized Mylar with Dacron net spacers.
The VCS used the cold, vented gas from the helium tank to intercept heat
coming into the dewar. Each SHOOT dewar had two VCSs made from 0.5 mm
sheets of 1100 series H14 aluminum. Each VCS was anchored at each strap through
an aluminum block that was clamped between the legs of the strap. Epoxy bonding
was attempted at first, but peeling of the outer fiberglass layer occurred at the bond
edge. The forward anchors allowed flexibility in the axial direction of the tank to
prevent buckling of the shields during cooldown. The forward and aft cones of the
VCS’s were removable to allow access to the cryostat and helium tank. During
dewar assembly the inner VCS was covered by a 15 layer MLI blanket, and the
outer VCS was covered with a 35 layer MLI blanket.
The main shell contained the cryostat, helium tank, and VCSs and provided
attachment points between the dewar and the flight support structure. The barrel of
the main shell was a rolled cylinder of 3.18 mm thick 2219 T37 aluminum sheet.
The hard points of the main shell, which served as the interface to the flight support
structure and the anchor for the support straps were two girth rings. The girth rings
were machined from a 15 cm thick plate of 2219 T37 aluminum and welded to each
end of the cylinder. The main shell was then heat treated to the T87 condition and
machined to final dimensions. The forward and aft heads were spun domes from
aluminum 2219 and 6061 respectively. After spinning they were heat treated to the
T6 condition and then machined to final dimensions.

4.3.2 Cryostat Details

The dewar/cryostat system was designed for the easy removal of various compo-
nents. The majority of these components were in the cryostat. The port and star-
board cryostats were nearly identical to allow easy-change out of parts or even the
entire assembly. To facilitate the removal of the cryostat, the dewar plumbing and
valves were located in a small volume. Thermal attachment of the gas vent lines to
the vapor cooled shields was made over a relatively short distance at the top of the
dewar. Thermal conduction in the vapor-cooled shield material provided a nearly
isothermal environment. For relatively small dewars, as in SHOOT, this scheme led
to an acceptable temperature difference along the shields of about 13 K in the outer
shield and 2 K in the inner shield.
102 M. DiPirro

In the jargon of space helium dewar design, SHOOT had an “instrument

dominated” heat load as opposed to a “parasitic dominated” heat load. That is, the
heat generated within the helium tank was large enough, on average, that opti-
mization of vapor cooling to reduce parasitic heat did not much increase its lifetime.
Coupled with the mission duration of about one week, this allowed the plumbing
design to be optimized for the transfer of helium, rather than configured as a
long-term storage dewar.
The cryostat plumbing lines consisted of a low-flow vent, high-flow vent,
transfer line, and two emergency vents (refer to the schematic in Fig. 4.1). The
low-flow vent led from the liquid/gas phase separator within the cryogen tank
through a warm valve and out of the dewar. This was the normal vent path out of
the dewar and was the primary means of cooling the VCSs. It was made of
12.7 mm diameter stainless steel tubing with 0.5 mm thick walls. The tube was
interrupted by thin-walled bellows between the cold plate, IVCS, OVCS, and warm
plate to allow for some misalignment and provide structural isolation between
dewar components. The total length of the vent was approximately 1.2 m.
A cold valve could be used to bypass the phase separator for chill down of the
dewar. The high-flow vent was a much shorter (0.6 m), straighter, and wider
(18 mm i.d.) tube which led from the high-flow phase separator out of the warm
plate to a large diameter Ball valve. It provided a low impedance vent path at the
expense of vapor cooling efficiency for those times when large amounts of heat
were being dissipated in the liquid such as during a transfer. The high-flow vent line
was only weakly coupled thermally to the VCSs to minimize warming of the gas
vented and thus decrease the pressure drop in the line. This, however, also led to an
increased heat input to the cryogen tank of about 50 mW in the standby mode. The
transfer line left the cryogen tank through two valves, running through the shields to
the external coupler and external transfer line. This was similar tubing to the
low-flow vent, and was thermally isolated from the shields as was the high-flow
vent. The dual emergency vents provided the means of releasing fluid within the
cryogen tank in a controlled, safe way following a catastrophic loss of the dewar
guard vacuum. See the section on safety for more details.
Over 300 electrical leads traveled from hermetic connectors on the warm plate to
the cold region of the cryostat. 90 % of the leads were made of manganin wire, with
the remainder of 0.13 mm diameter copper. The leads were heat stationed at each
VCS through a connector potted in Stycast 2850. This has proved to be an effective
heat sinking method. Approximately 180 leads were fed into the liquid helium tank
through 32 pin hermetic connectors. In our experience, these are the largest com-
mercial connectors that can continue to be hermetic at low temperatures.
The main attachment of the cryostat to the cryogen tank was the cold plate at the
top of the tank. The cold plate was machined from an explosion welded
aluminum-stainless steel sheet [3].
4 The Superfluid Helium On-Orbit Transfer … 103

4.4 Components

4.4.1 Development Notes

Each individual component developed for or used in SHOOT went through a

thorough process of testing and qualification before integration into the flight
dewar/cryostat system. Several of the components’ performance could be visually
observed by using glass dewars. We were able to see the phase separation per-
formance of the high-flow phase separator, the low-flow phase separator, the
liquid/vapor detectors, and part of the screen channel liquid acquisition system.

4.4.2 Phase Separation

The “natural state” of liquid helium in a dewar on the ground is 4.2 K at one
atmosphere (4 K at the reduced pressure in Boulder, CO!). By the same token, in
orbit where the pressure is practically zero the equilibrium temperature will be less
than 2 K (i.e., superfluid). The equilibrium temperature of the liquid will result from
a balance between the pressure drop in the vent line plus the phase separator and the
boil off rate.
Without a significant acceleration aiding liquid settling, a method to allow
boiling to maintain a low temperature and pressure while not exhausting liquid
directly must be employed. For superfluid helium, the fountain effect can be used
for this purpose. The fountain effect directly converts a temperature gradient into a
pressure gradient in the same direction. In the two-fluid model, superfluid 4He is
composed of two interpenetrating components: a “super” component that flows
without viscosity and a “normal” component that has viscosity and behaves as a
normal Newtonian fluid. The fountain effect is realized in a porous material whose
pores are small enough to significantly restrict the flow of the normal component
while allowing the super component to flow freely. A temperature gradient across
the porous medium then leads to a gradient in the proportion of normal and super
component which gives rise to an osmostic-like pressure gradient. This pressure
gradient is in the same direction as the temperature gradient and is about 12 times
larger than the vapor pressure gradient in the opposite direction.
The device typically used is called a porous plug which consists of a disk of
porous material (usually stainless steel) with small pores (*a few microns in
diameter) through which a temperature and pressure gradient can be developed
while venting out one side. Bulk superfluid helium is prevented from passing out of
the plug by the fountain effect across the pores. The pores are sized at about 4
microns in effective diameter; large enough to allow some heat conduction out of
the tank to maintain a relatively small temperature drop, while small enough to
obtain a relatively large fountain effect (up to about 14 kPa). Trade-offs between
ultimate temperature in the tank and large-enough fountain pressures usually end
104 M. DiPirro

Fig. 4.5 Picture of two

high-flow phase separators.
The cylinders are made of
porous sintered stainless steel
in a reentrant geometry to
maximize the flow area and
minimize pressure drop across
this device

with a porous plug impedance that is on the same order as the impedance of the vent
line downstream of the porous plug.
SHOOT used two types of phase separators: a high- and a low-flow phase
separator, schematically shown in Fig. 4.1 as D and C, respectively.
The high-flow phase separator had commonly-used porous stainless steel but in a
large area and with pores of effective diameter of 8 microns rather than the standard
4 microns (see Fig. 4.5). The 8 microns and large area were necessary to maintain a
low temperature in the dewars during superfluid transfer. When transfers were not
taking place, SHOOT’s temperature was 1.06 K—the coldest thing in space!
As a passenger on the space shuttle, SHOOT had to abide by the access
restrictions common to attached payloads. One of those restrictions was that the last
servicing needed to be performed about 65 h prior to launch. For superfluid helium
payloads on expendable launch vehicles (IRAS, COBE, IRTS, etc.) the last ser-
vicing was 12 h before launch. Waiting this extra time on the launch pad and
maintaining the liquid below the lambda point would require an on-board vacuum
pump [4] or a normal liquid helium guard tank [5]. The former had proven unre-
liable on two previous shuttle missions and the latter required a large fraction of the
space within the vacuum shell. For SHOOT a new approach was implemented:
launch with normal liquid helium (He-I, T > 2.17 K) and convert to superfluid
(He-II, T < 2.17 K) on orbit. The added benefit to this approach was that
vapor-cooling could be used while on the launch pad even when the liquid was
below 4.2 K due to stratification.
The low-flow phase separator was designed to phase separate He-I from its
vapor in addition to He-II from its vapor. The flow of He-I out the vent was
throttled by closely spaced (*6 microns), parallel, high conductivity, copper
washers. The washers were hollow in the center and helium flowed radially through
4 The Superfluid Helium On-Orbit Transfer … 105

Fig. 4.6 The SHOOT

low-flow phase separator.
Square and circular disks
alternate in the stack with tiny
Kevlar fiber spacers. The bolt
at the center top was sued to
adjust the clearance between
washers

the gaps to the vent line. The gap width in the stack of washers was controlled by an
adjustable screw and pressure plate at the top of the device (see Fig. 4.6). The high
conductivity copper carried the heat of vaporization of the helium back to the liquid
in the center hole. We discovered that even at 2.17 K the Kapitza resistance at this
inner boundary was significant. We then crenelated the inner surface to increase the
surface area. This decreased the thermal boundary resistance to an acceptable value
and this phase separator performed perfectly at pressures up to 110 kPa. As the
dewar vented and temperature decreased the flow rate decreased, reaching a min-
imum at around the lambda point. During on-orbit operation the high-flow phase
separator was used in parallel from slightly above the lambda point to continue
lowering the temperature. For more details see Refs. [6, 7].

4.4.3 Liquid Acquisition

The LADs were the heart of the SHOOT demonstration. The screen channels
(Fig. 4.7) could only be demonstrated on a small scale in one g and the vanes in the
port dewar (Fig. 4.8) could not be demonstrated on the ground at all. Their function
was to gather the liquid from various places within the tank and feed it to the pump
by using surface tension. These are described elsewhere [8].
106 M. DiPirro

Fig. 4.7 One of 4 screen channels in the starboard dewar. The fine mesh screen (lower half) ran
along the helium tank wall to scavenge as much helium out of the dewar as possible. A row of
liquid-vapor detectors can be seen attached to the sheet metal top of the screen channel

Fig. 4.8 The SHOOT port

dewar cryostat insert. The
vanes extend to the outer edge
of the inside of the tank and
feed fluid to the center. At the
top of the cryostat is the low
flow phase separator housing
(center) and the high flow
phase separator housing
(right)
4 The Superfluid Helium On-Orbit Transfer … 107

Fig. 4.9 The SHOOT thermomechanical pump. The item on the left is a manganin wire-wound
50 Ω heater which is inserted into the ceramic cup on the right. Together they comprise a pump
capable of producing a flow of 1 m3 per hour of superfluid helium

4.4.4 Thermomechanical (Fountain Effect) Pumps

SHOOT used the very simple and reliable fountain effect pumping method. The
pumps were simply constructed from a mullite cup having pores with 0.4 micron
effective diameter. These pumps were capable of flow of over one m3 per hour and
pressures up to 60 kPa [9] (see Fig. 4.9).
Motor driven mechanical pumps were tested with superfluid helium but had
difficulty with cavitation at the inlet when a hydrostatic head was not present [10].

4.4.4.1 Instrumentation

Knowing roughly where the liquid was during the mission was a key requirement.
To solve this problem discreet liquid/vapor detectors were developed.
Implementation was very quick (concept to working model in 2 weeks, see
Figs. 4.10 and 4.11) and effective [11]. LVDs could be read out by injecting a
constant current through a string of Si chips and reading their individual voltages.
The response to a change of state was milliseconds and their position accuracy in
terms of the liquid-vapor interface was 10s of microns.
The key to understanding the performance of any cryogenic system is ther-
mometry. SHOOT developed a state-of-the-art flight electronics system measured
germanium resistance thermometers to 16 bit resolution and accuracy with a 17 Hz
excitation. Thermometers were selected for relatively low resistance (<5000 Ω) to
108 M. DiPirro

Fig. 4.10 Test model of the SHOOT liquid/vapor detectors. The tiny dots near the top of the
picture space 10 mm apart are the 0.25 mm cube Si detectors mounted on a horizontal 0.05 mm
diameter stainless steel wire

Fig. 4.11 Close-up picture of one of the liquid-vapor detectors inside the screen channel before
the screen is welded in place

mitigate AC parasitic effects. This thermometry enabled high-resolution heat pulse

mass gauging as described below.
In low gravity, where the liquid-vapor interface is not well known a priori, the
amount of superfluid helium can be simply determined by applying a heat pulse and
watching the temperature rise. This technique was used successfully on SHOOT to
an accuracy of better than 2 %. It turns out that I learned a bit of forgotten ther-
modynamics at the same time because most of the 2 % error was due to calculating
the fluid amount through enthalpy rather than properly using internal energy!
SHOOT measured the flow rate between the dewars in two ways: by the heating
and upstream temperature of the TM pump [9] and by using a venturi flow meter.
The venturi was read through variable reluctance pressure transducers and matched
the TM pump inferred flow rate to better than 2 %. Integrating the two flow meters
provided a check on the heat pulse mass gauging [12].
4 The Superfluid Helium On-Orbit Transfer … 109

Strain gauge-based pressure transducers were also used to measure absolute

pressure in each cryogen tank. These pressure transducers were bought from
Teledyne Taber and used a Wheatstone bridge arrange for the strain gauges which
eliminated temperature effects to first order.

4.4.5 Cryogenic Stepper-Motor Valves

Stepper-motor driven cryogenic valves for SHOOT were developed and qualified
by Ralph Haycock at Utah State University [13]. Two valve sizes were developed,
namely a nominal ½ inch valve and nominal ¾ inch valve. The finished valve
assembly is shown in Fig. 4.12. The basic valve assembly consists of four major
elements: a four-phase stepper motor, a recirculating ball screw assembly, a valve
stem load cell, and the valve body. The stepper motor is a four-phase reluctance
type motor. The ball bearings in the motors were replaced by dry-film lubricated
bearings with increased dimensional tolerance to allow for thermally induced dif-
ferential contraction. A redundant set of micro-switches at each end position pro-
vides positive indication of the valve status and control for the motor. The
recirculating ball screw assembly consists of the transmission gears, the recircu-
lating ball screw, and two rigid links that connect the recirculating ball nut to the
valve stem through a load cell. The load cell assembly consists of the valve stem
and an adjustable load cell consisting of Belleville spring washers stacked in

Fig. 4.12 The SHOOT stepper-motor-driven cryogenic valve

110 M. DiPirro

parallel. The spring washers maintained the stem tip force at cryogenic temperatures
despite the differential thermal contraction of the valve elements.
The valves could be operated at any temperature from 300 K down. They
required 1 A (1 W at low T) for 30 s generating a total heat input of about
30 J. Two valve sizes were developed, a ¾ inch throughput (based on a NuproTM ¾
inch valve) and a ½ inch throughput. The larger valve generated a seating force of
202 N. This force caused micro-cracks in the original copper stem when pressed
against the stainless steel seat. The copper was replaced with TorlonTM in the flight
valves. When sealing against superfluid helium these valves were tested to be leak
tight (<10−10 Pa m3/s) after more than 100 cycles. The micro-switch position
sensors were somewhat unreliable at low temperature. Although they were adjus-
table at room temperature, they were very sensitive to small changes in position
caused by differential contraction.

4.4.6 Cryogenic Relief Valves

Each SHOOT dewar had three volumes potentially where trapped liquid could exist
between closed valves (Fig. 4.1). Pressure building up in these volumes could result
in a vacuum failure. To prevent this, cold relief valves were used. A cold relief
valve was developed for SHOOT. A spring held a stainless steel seat against a
conical VespelTM valve head until upstream pressure built in the bellows, moving
the stem out. The spring provided a seating force of about 200 N on a seat of
3.8 mm diameter. The cracking pressure was approximately 150 kPa. Tests showed
the pressure necessary to achieve full flow was 165 kPa. A number of open-close
and thermal cycles were performed, both at room temperature and 4.2 K. In all
cases the leak rate at 4.2 K in the closed position was less than 1 × 10−6 Pa m3/s. at
a pressure difference of 105 kPa, rising to 8 × 10−6 Pa m3/s. at 140 kPa.
This valve was not designed to replace burst discs in an emergency venting
situation. It had a low throughput designed to handle pressurization of liquid from a
nominal heat input of a few watts.

4.5 Safety

In 1984, at the beginning of the SHOOT project, we were determined to comply

with all of the space shuttle safety requirements to the letter, without asking for
waivers. Two-fault tolerance for any failure with major cost or safety impacts was
required. To that end, we designed dewars with two dedicated vent lines from the
cryogen tank to the outside of the dewar. The dewar cryogen tank was also
designed as a “leak-before-burst” vessel. Cold burst disks were located on the
cryogen tank and were connected to two warm burst disks on the vacuum shell of
each dewar. The outlet of the cold burst disk was hermetically connected to the inlet
4 The Superfluid Helium On-Orbit Transfer … 111

of the warm burst disk by 19 mm outer diameter (18 mm inner diameter) stainless
steel tubes. Thus, there would be no worry about multilayer insulation restricting
the flow path through the vacuum space. These dedicated emergency vent lines did
add some extra parasitic heat, but since SHOOT’s lifetime was to be short on orbit
and dominated by heat generated in the transfers, it was an acceptable trade.
The burst disks used for SHOOT were similar to the ones used for COBE [14],
but had indium sealed burst diaphragms rather than welded ones. This made it
possible for the burst disk activation pressure to be measured and diaphragm
punctured before being refurbished for flight.
The burst disks and vent lines were sized to prevent the dewar tank pressure
from rising above 413 kPa. An emergency venting analysis was performed to show
that the maximum design pressure of the dewar would not be exceeded in the event
of a large rupture of the cryogen tank. See Figs. 4.13 and 4.14. Such a large rupture
would only be possible if fork-lift tines were rammed through the main vacuum
vessel of a dewar.

Fig. 4.13 The predicted

blow down pressure in the
event of a catastrophic loss of
dewar guard vacuum. 60 psia
is roughly 4 atmospheres, the
pressure at which the cryogen
tank burst disks rupture

Fig. 4.14 Two predictions of

the predicted helium flow rate
versus time out of one
emergency vent line after a
burst disk rupture
112 M. DiPirro

Fig. 4.15 Sketch of the SHOOT dewar in the orientation in the shuttle bay. The location of any
liquid air formation is shown

Each dewar’s exhaust helium, for both normal operations and for a sudden loss
of vacuum, was channeled into a duct ending in a diffuser that would disperse the
cold effluent. See Fig. 4.15. However, in the event of vacuum loss during ground
operations, liquid air could form on the outside of the vent duct and diffuser. We
tested this possibility by transferring liquid helium into the duct and diffuser at a
rate of 400 L per hour for 30 min. Liquid air formation on the outside lower edge of
the diffuser was collected and measured. A total of 0.4 L of liquid air was collected
over 30 min. To prevent this liquid air from dropping to the floor or into the shuttle
bay, a 1 L capacity drip pan was constructed from double aluminized Kapton, a
material that is compatible with liquid air. This pan was fastened to the outside of
the dewar immediately below the dripping area.
In the final review the Flight Safety panel at Johnson Space Center described our
safety design and test effort as “a conspiracy to meet the requirements” with which
we happily agreed!
4 The Superfluid Helium On-Orbit Transfer … 113

4.6 Working with SHOOT on the Ground

The two dewars contained many fill and vent ports which led to an interesting effect
that we noticed on the ground. 4.2 K gaseous helium has a density about 1/7th that
of liquid so it is relatively heavy. When we had a dewar tipped on its side with one
of the vents above the other, the lower vent had He gas outflow, while the upper
vent, only a few cm higher was sucking in air!
Near the end of the ground test program one of the external burst disks on one of
the dewars developed a small leak. Checking for leaks involved opening a valve
that could potentially leak itself. Therefore a leak check cold was not done. After a
couple of weeks the dewar was warm as planned and cool down before shipment to
the launch site was planned. At this point the burst disks and emergency vent lines
were checked. The one emergency vent line could not be evacuated because the
cold burst disk on the tank had been ruptured. The rupture was not in the normal
direction out of the tank, but in the opposite direction! That is, the leaking external
burst disk had allowed enough air into the emergency vent line to freeze a solid ice
plug into the line while the dewar was cold. The solid ice plug expanded upon
warming, rupturing the burst diaphragm into (instead of out of) the helium tank.
This was a very important lesson on the need to leak check thoroughly and as often
as possible. This problem cost us 5 weeks of overtime for repairs before shipping to
the launch site, but we still arrived in time for our original launch date.
The rubber o-rings that formed a seal into the vacuum space were used in areas
of the dewar that were not cold. We did not consider the possibility of diffusion
through these o-rings because we carefully ducted vented helium gas away from the
dewars. However, given enough time even a few hundred ppm of helium gas in the
surrounding atmosphere will contaminate the guard vacuum and lead to a higher
heat load. We noticed that over several months without pumping the guard vacuum
the heat load increased from 180 to 220 mW.
Another point of discovery was how our low temperature top-off procedure
evolved over time. To achieve the maximum fill level on orbit the dewars needed to
be topped-off with liquid helium near the superfluid transition. This required
pumping down the SHOOT dewars and the supply dewar, well below atmospheric
pressure. Similar procedures had been used by several other flight dewars; however
SHOOT had differences. SHOOT used two vents on each dewar. We also filled the
second SHOOT dewar from the first through the flight transfer line. This meant that
we had several flow paths and several vent paths through the top-off. The procedure
was rehearsed many times before flight, but each time it required a modification for
a new issue: sometimes a vent would produce liquid air, sometimes a vent would be
plugged by solid air, and so forth. This continued even on the last servicing on the
launch pad. The final servicing concluded successfully, but only by making one
final on-the-spot redline to the procedure.
114 M. DiPirro

4.7 On-Orbit Operations

This case study will not go into many of the details of the SHOOT on orbit results,
but a sampling of the measurements made is given here. First, as an experiment, it
was a huge advantage for SHOOT to be riding in the space shuttle. Beneficial and
adverse accelerations were possible, the communication, power, and data storage
were provided, and mission support from the Johnson Space Center and Kennedy
Space Center contributed greatly to the success of this Goddard Space Flight
Center-led mission.
Second, a lesson about unintended consequences was learned on orbit. SHOOT
was launched containing He-I and needed to be pumped down over about 24 h to
reach its operating temperature. The plan for this pump down was to use the
low-flow phase separator to get close to the superfluid transition, then open the
valve to the high-flow phase separator to complete the pump down. This had been
tested on the ground to the extent possible, but the high-flow phase separator was
difficult to test with a significant hydrostatic head. Thermometers on each phase
separator indicated its proper operation: higher temperature on the upstream side of
the separator and lower downstream. The port dewar containing the vane system
pumped down more quickly than the starboard dewar. Everything appeared normal
when the valve upstream of the port dewar’s high-flow phase separator was opened,
but when the valve on the starboard dewar was opened an alarm rang on the shuttle.
The astronauts sleep with the shuttle’s nose pointed toward Earth, a gravity-
gradient, stable attitude because attitude control thrusters were noisy. The alarm
woke the astronauts because the shuttle attitude had swung by 5° from vertical, then
swung back to the other side 6°. We learned later that this was caused by a sudden
venting of 10s of liters of liquid helium out the vent. Although the flow out the vent
was diffuse, the gas impinged on an open shuttle bay that concentrated the
momentum of the outflowing gas in the opposite direction. At the time we thought
that the high flow phase separator on the starboard tank was broken allowing liquid
to escape. What we found out later is that we had such a large pressure gradient and
flow rate out of the high-flow phase separator that the upstream liquid was sub-
cooled significantly from the liquid in the remainder of the tank [15]. This was the
result of a late change to the design that thickened the thin-walled copper heat
exchanger around the high-flow phase separator so that the heat flow was restricted.
Even superfluid helium can have significant gradients! Discovery of this was
hampered by the thermometer wiring being switched on the starboard phase sep-
arator, and the different behavior of the port LAD (vanes) which prevented most of
the liquid from leaving the port dewar.
After this loss of helium event we had a little over half the amount of liquid we
expected to have, but still managed to complete all of the types of transfers planned
for the mission. These transfers were: port to starboard high and low rate, starboard
to port high and low rate, transfer during adverse accelerations, chill down and
transfer into an empty and cold and empty and warm tank, intelligent system
controlled transfer, etc.
4 The Superfluid Helium On-Orbit Transfer … 115

One of the more interesting side benefits to the expected data were some indi-
cations of how slosh in superfluid helium behaved [16]. After a fluid settling
acceleration the superfluid motion damped out very quickly with only a couple of
large amplitude oscillations.
SHOOT also obtained interesting results on stratification of He-I [17, 18].

4.8 Summary

Many components demonstrated for the first time in space, or anywhere for some,
came to be used for other projects:
• Stepper-motor valves (XRS, XRS2, SXS and other liquid helium missions)
• Venturi flow meter (to be used on RRM3)
• Mass gauging (ISO and to be used on SXS)
• Fountain pumps (balloon payloads ARCADE and Super-ARCADE)
• Liquid/vapor detectors (shown to also work for other cryogens including liquid
nitrogen)
• Cryogenic/ambient burst disks (XRS, XRS2, and SXS)
The SHOOT mission was fully successful in satisfying the original objectives.
Along the way several interesting observations were made and lessons learned. The
original purpose of helium resupply had disappeared but SHOOT became a
demonstration platform for many useful cryogenic components and techniques.
SHOOT is still the most extensive cryogenic fluid management experiment con-
ducted in space.

References

1. P. Kittel, Orbital resupply of liquid helium. J. Spacecraft Rockets, 23, 391 (1986)
2. K.F. Weintz, M.I. Basci, J.M. Uber, SHOOT Dewar support strap design and performance.
Cryogenics 34, 357–360 (1994)
3. Explosive Fabricators, Incorporated, Louisville, Colorado
4. D. Petrac, U.E. Israelsson, T.S. Luchik, The lambda point experiment: helium cryostat,
cryo-servicing, functions, and performance. Adv. Cry. Eng. 39, 137–144 (1994)
5. R.T. Parmley, J. Goodman, M. Regelbrugge, S. Yuan, Gravity Probe B dewar/probe concept.
Proc. of the SPIE 619, 126 (1986)
6. P.J. Shirron, J.L. Zahniser, M. J. DiPirro (eds.), A liquid/gas phase separator for He-I and
He-II. Adv. Cryog. Eng. 37, 105 (1992)
7. J.G. Tuttle, M.J. DiPirro, P.J. Shirron, Liquid/gas phase separators for the superfluid helium
On-orbit transfer (SHOOT) project. Adv. Cryog. Eng. 39, 121 (1994)
8. P.J. Shirron, M.J. DiPirro, J.G. Tuttle, Flight performance of the SHOOT liquid acquisition
devices. Cryogenics 34, 361 (1994)
116 M. DiPirro

9. M.J. DiPirro, E.R. Quinn and R.F. Boyle, Tests of a nearly ideal, high rate thermomechanical
pump, in Proceedings of 12th International Cryogenic Engineering Conference (Butterworths,
London, 1988), p. 646
10. P.R. Ludtke, D.R. Daney, Cavitation characteristics of a small centrifugal pump in He I and
He II. Cryogenics 28, 96 (1988)
11. P.J. Shirron, J.G. Tuttle, M.J. DiPirro, Performance of discrete liquid helium/vapor and
He-I/He-II discriminators. Adv. Cryog. Eng. 39, 1105 (1994)
12. M.J. DiPirro, P.J. Shirron and J.G. Tuttle, Mass gauging and thermometry on the superfluid
helium on-orbit transfer flight demonstration. Adv. Cryo. Eng. 39, 129 (1994)
13. These are now a commercial product of ATK/Mission Research
14. The COBE burst disks were built by Ametek Straza which became Katema. Currently a
version is marketed by Hydrodyne
15. M.J. DiPirro, P.J. Shirron, J.G. Tuttle, Superfluid helium transfer in space. Cryogenics 34
(ICEC Supplement), 267 (1994)
16. M.J. DiPirro, P.J. Shirron, J.G. Tuttle, On orbit superfluid transfer: preliminary results of the
SHOOT Flight Demonstration. Cryogenics 34, 349 (1994)
17. P.J. Shirron, M.J. DiPirro, Low gravity thermal stratification of liquid helium on SHOOT.
Cryogenics 32, 159 (1992)
18. J.G. Tuttle, M.J. DiPirro, P.J. Shirron, Thermal stratification of liquid helium in the SHOOT
dewars. Cryogenics 34, 369 (1994)
Chapter 5
TESLA & ILC Cryomodules

T.J. Peterson and J.G. Weisend II

Abstract The Superconducting Radio Frequency (SRF) cryomodules for the

TESLA and International Linear Collider (ILC) projects have undergone significant
design evolution and prototype testing. Given the large number of cryomodules
required for these projects, key requirements include low cost and low heat leak.
This chapter provides details of the of the ILC cryomodule design and traces its
evolution. Topics include: requirements, He II cooling of the SRF cavities,
mechanical (including alignment) and thermal design, fluid flow, safety and vac-
uum systems. The results of early prototype testing and their impact on the design
are also described. Changes in the design to enable its use in the LCLS II accel-
erator project are also described.

5.1 Introduction

The Deutsches Electronen Synchrotron (DESY) in Hamburg, Germany, led the early
electron-positron linear collider collaboration which developed the concept called
TESLA (Tera-Electronvolt Superconducting Linear Accelerator) [1] in the late 1980s
and early 1990s. The TESLA collaboration desired cryomodules with a lower cost
per unit length than had been developed up to that time. Among the cost-reducing
features would be to connect cryomodules in continuous, long strings similar to
cryostats for long strings of superconducting magnets (see for example Chaps. 2 and
3). The elimination of the external cryogenic transfer line by placing all cryogenic
supply and return services in the cryomodule could also reduce costs, not only
directly for the cryogenic components, but also by reducing tunnel space required.

T.J. Peterson (&)

Cryomodule Engineering, SRF Development Department, Technical Division,
Fermi National Accelerator Laboratory, P.O. Box 500, Batavia, IL 60510, USA
e-mail: [email protected]
J.G. Weisend II
European Spallation Source ERIC, Lund University, P.O. Box 176, 22100 Lund, Sweden
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 117

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_5
118 T.J. Peterson and J.G. Weisend II

This approach, driven mainly by the number of cryomodules required, as many as

2000 in some accelerator concepts [1–3], is in contrast to the segmented cryomodules
described in Chap. 6. Segmenting the cryomodules and separating them by warm
sections containing magnets, vacuum systems and beam instrumentation tends to be
the preferred solution in systems in which fewer than 100 cryomodules are required.
The warm sections allow easier access to various accelerator systems. However, the
numerous warm—cold transitions along the length of the accelerator result in higher
cost and heat leak than can be tolerated in systems the size of TESLA.
Prototype TESLA cryomodules were incorporated into a test linac called the
TESLA Test Facility (TTF) at DESY [4]. TTF was modified and expanded into a
light source facility called the Free-electron LASer in Hamburg (FLASH). TTF and
FLASH demonstrated successful operation of TESLA-style cryomodules in a linac
with beam.
Various projects subsequently adopted the TESLA cryomodule design concept
with some modifications depending on specific requirements and innovations.
These projects include the International Linear Collider (ILC) [2, 3], X-ray Free
Electron Laser (XFEL) at DESY [5], and the Linac Coherent Light Source II
(LCLS-II) [6] at SLAC. LCLS-II includes more extensive modifications than the
other projects described here due to LCLS-II being a continuous wave (CW) linac
with higher dynamic heat loads than the other pulsed linac designs. This chapter
describes the TESLA-style cryomodule design in general, with examples from
TESLA prototype cryomodules and the various subsequent projects. Figure 5.1
illustrates one 12 m long cryomodule mechanical assembly.

5.2 Definitions

See Table 5.1.

Fig. 5.1 ILC cryomodule assembly showing vacuum vessel, eight RF power input couplers, three
internal supports on top, and open ports for instrumentation flanges and for power leads providing
current to a centrally located superconducting magnet package
Table 5.1 Definitions
Term Definition
SRF Superconducting radio frequency
CW Continuous wave operating mode
HOM Higher order mode
TESLA Tera-electronvolt superconducting linear accelerator (TESLA)—the predecessor to the ILC concept, developed collaboratively by DESY,
Saclay, INFN, Fermilab, Jefferson Lab, Cornell, and other labs and universities under the leadership of DESY. The TESLA collaboration
developed cryomodules with RF cavities in close-fitting helium vessels suspended from a large helium return pipe, cryogenic piping all within
the cryostat. Hence, cryomodules having that configuration are referred to here as TESLA-style
5 TESLA & ILC Cryomodules

ILC International linear collider

XFEL X-ray free electron laser, an SRF linac at DESY in Hamburg, Germany, which includes TESLA-style cryomodules
LCLS-II Linac coherent light source II, which includes a 4 GeV superconducting linac at SLAC utilizing TESLA/ILC-style cryomodules modified for
CW operation in the existing SLAC tunnel
2K Nominally 2 K temperature level, typically 1.8–2.1 K
5K Nominally 5 K temperature level, which includes temperatures approximately from 4.5 to 7.5 K
70 K Nominally 70 K temperature level, which ranges from approximately 35–85 K. (Note that cryogenic circuits, thermal shields, and thermal
intercepts will be *80 K when LN2 is the coolant. Lower temperatures are provided by gaseous helium.)
MAWP Maximum Allowable Working Pressure, a term that is used to define the safe pressure rating of a component or test system
Cold Those portions of the cryomodule within the vacuum vessel which are cooler than room-temperature—RF cavities, piping, thermal shield, etc.
mass
HGRP Helium gas return pipe, the 300 mm diameter helium pipe which also serves as the structural backbone of the cold mass in TESLA-style
cryomodules
MLI Multilayer insulation, or “superinsulation”, aluminized mylar wrapped in layers alternately with a conductively insulating material in the
insulating vacuum space to block thermal radiation
Q0 Cavity quality factor, of interest for cryomodule design since power dissipated by the RF cavity to the nominally 2 K temperature level is
inversely proportional to Q0
119
120 T.J. Peterson and J.G. Weisend II

5.3 Functional Requirements Summary

The TESLA and ILC main linac cryomodules contain superconducting cavities
together with cryogenic distribution lines, thermal shielding, magnets, and instru-
mentation. Most fundamentally, the cryomodule provides insulating vacuum,
cooling, RF power, and RF cavity support and alignment for the proper functioning
of the SRF cavities in the linac.
Strings without breaks between individual cryomodules include up to 12 cry-
omodules. In addition to niobium superconducting RF cavities, cryomodules may
include a combined focusing and steering magnet, a beam position monitor, and
gate valves on the beamline for vacuum isolation during assembly. Beam line
higher-order-mode (HOM) absorbers are located in the cold beamline interconnects
between cryomodules. The cavity tuners, HOM couplers, and fundamental power
couplers attach to each cavity. Cryogenic transfer lines are required only where
bypasses for warm sections of the linac must be provided.
TESLA-style cryomodules have been fabricated with four 3.9 GHz RF cavities
as well as with eight 1.3 GHz RF cavities. The ILC concept included some
eight-cavity cryomodules with one magnet package (the configuration illustrated in
Fig. 5.1) and some nine-cavity cryomodules without a magnet. The final TESLA
cryomodule design, as described in the TESLA TDR [1], contained 12 RF cavities
in a further attempt to improve the packing factor of RF cavities along the linac.
A mechanical tuner with motor-driven and piezo-driven components provides
active control of the RF cavity resonant frequency. RF cavities are independently
powered through a fundamental power coupler on each cavity connected via
air-filled waveguide to an RF source (klystron or solid-state amplifier).
Cryogenic circuits provide 2 K liquid helium to the cavities with a valve for
liquid supply in each string of cryomodules or, in the case of LCLS-II, via a valve in
each cryomodule. Managing a liquid helium bath of up to 100 m or so in length is
one factor in determining string length. The cryostat includes a nominally 5 K
thermal intercept circuit, which may include a thermal radiation shield, and a
nominally 70 K thermal radiation shield and thermal intercept circuit. Dynamic RF
heating of the cavities at 2 K is a major contributor to overall cryogenic refrigeration
required.
The Table 5.2 provides a summary of the key functional requirements for a
TESLA/ILC-style cryomodule that may drive the design. These are topics, which
require particular attention during design.

5.4 Cryomodule Mechanical Design

5.4.1 Cryomodule Major Components and Features

Cryomodules consist of various complex subassemblies, which we describe in this

section, starting from the inside with the niobium RF cavity and working out to the
Table 5.2 Key functional requirements for the cryomodules
Key requirement Description
Series configuration Cryomodules with insulating vacuum open at each end to the next, so connected insulating vacuums, and with cold beam
pipe through the interconnect. This configuration implies considerations for the in situ connection of not only the accelerator
beamline but also cryogenic pipes and insulating vacuum
No external parallel Cryomodules include all cryogenic piping within the cryostat, as opposed to having a parallel external cryogenic transfer line
transfer line to supply cooling to cryomodules
Microphonics As for any SRF cavity support structure, minimize cavity vibration and coupling of external sources to cavities. This is
addressed by means of providing a stiff support system and stiffening of elements such as the thermal shield
5 TESLA & ILC Cryomodules

Alignment Provide good cavity, quadrupole, and BPM alignment (typically <0.5 mm RMS)
Vibration Minimize quadrupole magnet vibration for gradient stability of typically 0.01 % RMS or better
Seismic design In many locations (especially Japan and California), one must design in accordance with local seismic requirements
Thermal efficiency Intercept significant heat loads at intermediate temperatures above 2.0 K to the extent possible. Intercepts include not only
thermal radiation shield or shields and support structures, but may also include intercepts for large cables, magnet current
leads, and RF components such as HOM couplers and absorbers
Pressure safety Cryomodules and components must comply with the requirements of the institution and/or locality in which they will operate.
For accelerator laboratories in the US, this means the Department of Energy regulation 10CFR851—a level of safety
equivalent to pressure code safety level. Designers must protect the helium and vacuum spaces including the RF cavity from
exceeding MAWP
Magnetic shielding Provide excellent magnetic shielding and very low residual field (for example, ≤5 mG) at the niobium cavities to preserve
high Q0 for less power lost as heat to 2 K
Thermal performance For RF cavities operating in CW mode, one may have to allow removal of over 20 W per cavity and/or 150 W at 2 K per
cryomodule
Cool-down Provide, to the extent possible given the cryomodule string configuration, cool-down conditions which retain high cavity Q0.
Recent recognition of the role of thermal gradient on the niobium cavity to “sweep out” magnetic flux as the Nb passes
through the superconducting transition during cool-down may create new design goals for management of cool-down, such as
the addition of a cool-down valve in each cryomodule where retaining high Q0 is particularly important
121
122 T.J. Peterson and J.G. Weisend II

Fig. 5.2 Helium vessel concept (not to scale) for an ILC 9-cell cavity

vacuum vessel. Figure 5.2 shows a sketch (not to scale) of the titanium helium
vessel assembly surrounding a 1.3 GHz niobium cavity consisting of multiple
elliptical cells. Liquid helium completely fills the helium vessel, surrounding the RF
cavity with a liquid level into the 2-phase pipe.
Heat flows via helium II heat transport through the liquid helium to the liquid
surface in the 2-phase pipe where the helium evaporates. Thus, cooling below the
lambda point involves no bubbling, an advantage in reducing microphonics dis-
turbance of the RF cavity.
Figures 5.3 and 5.4 illustrate the dressed cavity and a portion of the RF cavity
string. Each of the cavities is encased in a titanium helium vessel, supported from

Fig. 5.3 Dressed cavity. The titanium helium vessel is shown as transparent for a view of the
9-cell niobium RF cavity inside. This image includes an original design version of the Blade Tuner
and shows the 2-phase pipe connection on top of the helium vessel. The connection from the
helium vessel to the 2-phase pipe is called the nozzle or “chimney”
5 TESLA & ILC Cryomodules 123

Fig. 5.4 Dressed cavity as part of the cavity string within a cryomodule for LCLS-II. Other
features shown include the RF power coupler for one cavity, an end lever tuner for an LCLS-II
dressed cavity, support arms for the dressed cavities, and the invar rod just below the HGRP which
provides the axial position anchor for the dressed cavities

the HGRP by means of stainless steel brackets connected to four titanium pads on
the helium vessel itself. Each bracket is equipped with a longitudinal sliding
mechanism and adjusting screws and pushers for alignment. A mechanical, coaxial
(blade) tuner and a piezo-electric tuner are mounted to the vessel.
Dressed cavities are connected to one another via the beam tube connections and
the 2-phase pipe connections. An ILC cryomodule includes a string of eight or nine
such cavities connected in series. In the case of eight cavities, the ninth element in
the string is a corrector magnet package, also cooled at nominally 2 K. With the
cavities and magnet package each being about 1.3 m in length, the entire cry-
omodule assembly is about 12 m long.
The inter-cavity spacing—which accommodates RF- and HOM-couplers and a
flanged interconnecting bellows—amounts to approximately 300 mm, depending
on mechanical and RF details. For example, desire to operate beam in either
direction may dictate even-wavelength spacing. Conversely, desire to disrupt the
flow of “dark current” (electrons from field emission) may dictate non-integer
wavelength spacing. Manually operated valves required by the clean-room
assembly terminate the beam pipe at both module ends. The valves are fitted
with simple RF shields (Fig. 5.5).
Magnetic shielding surrounds the dressed cavity in order to limit imposed
magnetic field on the niobium superconductor. One must watch also for internal
sources of magnetic fields. No component of the cryomodule should impose more
than a specified low magnetic field (such as 5 milligauss, the specification for
LCLS-II) on the niobium cavity during cool-down through the 9.2 K transition
temperature, since trapped flux then reduces the efficiency of the RF cavity,
124 T.J. Peterson and J.G. Weisend II

Fig. 5.5 Cooling scheme for an ILC cavity string

resulting in higher heat loads at 2 K and/or limiting peak operating gradient of the
cavity.
The 300 mm diameter helium gas return pipe (HGRP), prominently visible
above the cavities in Figs. 5.6 and 5.7 serves both as the “backbone” support for the
RF cavity string in each cryomodule and also as the low pressure, 2 K, saturated
vapor return line to the cryogenic distribution box. The HGRP is support by three
epoxy-fiberglass posts [7–9]. To accommodate the HGRP thermal contraction when
cold relative to the vacuum vessel, the two side post brackets can slide over the top
flanges while the central post bracket is locked in position.
The support posts (see Figs. 5.7 and 5.8) consist of a fiberglass tube terminated
by two shrink-fit stainless steel flanges. Two additional shrink-fit aluminum flanges
are provided to allow intermediate thermal intercept connections to the 5–8 K and

Fig. 5.6 Some major features of a TESLA/ILC-style cryomodule. This particular image is of an
LCLS-II 1.3 GHz cryomodule assembly, but the features shown are typical of TESLA/ILC-style
cryomodules (credit for figure—Fermilab)
5 TESLA & ILC Cryomodules 125

Fig. 5.7 Cross-section of an ILC cryomodule showing major features which are typical of a
TESLA-style cryomodule

Fig. 5.8 Cold mass support and alignment system (credit for figure—Fermilab)

nominally 70 K temperature levels; the exact location of these flanges has been
optimized to minimize the heat leakage.
RF cavities are supported from the HGRP by means of four lugs on the helium
vessel. Stainless steel arms welded to the HGRP extend down to C-shaped holders
containing needle bearings, shown in Fig. 5.9. The cavities are anchored in position
126 T.J. Peterson and J.G. Weisend II

Fig. 5.9 Support system of cavity helium tank to the HGRP and invar rod (credit for figure—
Fermilab)

longitudinally via a clamp to an invar rod. The needle bearings allow shrinkage of
the stainless HGRP relative to the RF cavities and magnet package.
Cavity positions relative to fiducials on the vacuum vessel are set during
assembly with no requirement for later internal adjustment of cavity position within
the cryomodule after assembly. Alignment needs to be maintained with cryomodule
transport, thermal cycling, and pressure cycling. (For example, return to position
within 0.5 mm RMS tolerance for LCLS-II.) Final alignment requires positioning
of the vacuum vessel assembly with reference to the external fiducials, which were
in turn referenced to the cavity string. Table 5.3 summarizes approximate allocation
of alignment tolerances, illustrating that for each source of misalignment, due to
their additive nature, tolerances must be tighter than the overall requirement.
Linac beam operational requirements determine cryomodule lattice dimensions
and intracavity spacing. Since the linac is operated cold, but the cryomodule is
assembled warm, careful analysis of thermal contractions is required for the design
of warm assembly dimensions. Key dimensions include main coupler
cavity-to-cavity distance and cryomodule slot length.
RF power travels via waveguide and connections to the cryomodule input
couplers. Various other electrical connectors are located on round cover plates
mounted on the vacuum shell, the connector plates sealed with O-ring seals. A pair
5 TESLA & ILC Cryomodules 127

Table 5.3 Typical alignment tolerances for RF cavities within a cryomodule

Subassembly Tolerances Total envelope
(RMS) (mm)
Cryomodule Cavity and helium ±0.1 Positioning of the cavity with respect to
assembly vessel external reference ±0.5 mm
Supporting system ±0.2
Vacuum vessel ±0.2
construction
Action
Transport, Transport and ±0.2 Reproducibility and stability of the
testing, and handling (±0.5 g in cavity position with respect to external
operation any direction) reference ±0.5 mm
Vacuum pumping ±0.3
Cool-down
RF tests
Warm-up
Thermal cycles

of instrumentation flanges is also associated with each input coupler to reduce wire
lengths and risks associated with long runs of wires inside the cryomodule.
Figures 5.10 and 5.11 illustrate the region in the cryomodule around the
superconducting magnet and one example (for a Type 3+ cryomodule) of a magnet
package. Like for the RF cavities, the 300 mm diameter HGRP supports the
superconducting magnet package, consisting of a quadrupole and two correction
dipoles. Also like for the RF cavities, the invar (see Chap. 1) rod maintains the
magnet package axial position, while the HGRP supports the magnet mass and
provides lateral and vertical positioning. A beam position monitor is pinned to the
magnet package for precise relative alignment of the two. Current leads shown in
Figs. 5.10 and 5.11 are conductively cooled, based on the design developed by
CERN for the LHC corrector magnet current leads [10]. Conductive cooling has the
advantage of not requiring helium gas flow, which in these cryomodules one could
not provide from the 2-Kelvin magnet coolant due to the low (nominally 30 mbar)
pressure. Relatively low currents (100 A or less) enable the use of conductive
cooling without too much heat input to the low temperature levels of the
cryomodule.
The cryostat for TESLA, ILC, and XFEL includes two aluminum radiation
shields operating in the temperature range of 5–8 K and 40–80 K respectively. For
LCLS-II, the nominally 5 K thermal shield is eliminated although the helium circuit
is retained to provide 5 K thermal intercepts for RF power input couplers, magnet
current leads, support posts, and some RF cables. Thermal radiation shield bridges
are provided through each cryomodule interconnect.
Blankets of multilayer insulation (MLI) are placed on the outside of the thermal
shields and other cold surfaces. At the nominally 70 K level, MLI blankets consist
of typically 30 layers. At the 2 and 5 K levels, MLI blankets consist of 5–10 layers
128 T.J. Peterson and J.G. Weisend II

Fig. 5.10 A “Type 3+” cryomodule, very similar to those for XFEL, showing from the left below
the 300 mm pipe: the last dressed cavity in the string, the magnet package, BPM, gate valve, HOM
absorber, and another vacuum gate valve (credit for image: XFEL TDR [5])

Fig. 5.11 A Type 3+

cryomodule magnet package,
showing current leads, helium
vessel containing the magnet,
beam position monitor
attached to magnet end, and
2-phase pipe section
associated with the magnet
(credit for image: XFEL TDR
[5])

while the nominally 70 K blanket contains 30 layers. Cavity and quadrupole helium
vessels, gas return pipe and 5–8 K pipes are wrapped with 5 to 10 layers of MLI to
reduce heat transfer in the event of a vacuum failure.
Thermal shields are constructed from a stiff upper part (divided into two halves),
and multiple lower sections (according to the number of the cold active compo-
nents, e.g. cavities, magnets). Intermediate flanges on the fiberglass posts support
the thermal shields from the top. Thermal shields are screwed to the center post but
can axially slide on the other two posts, to which they are still thermally connected,
allowing for thermal contraction relative to the vacuum vessel. Slots cut into the
5 TESLA & ILC Cryomodules 129

Table 5.4 Approximate typical cryomodule mass (kg)

Cryomodule component Mass
(kg)
Cold mass (cavity string, thermal shield, HGRP, other cold pipes) 3300
Vacuum vessel 3300
End flanges, reinforcing rings 300
HGRP supports (posts, brackets, support covers, 3 sets) 900
Other items (vacuum bellows, external support base, power couplers, 800
instrumentation, etc.)
Total 8600

thermal shields near cooling pipe attachments reduce bowing due to differential
thermal contraction during cool-down and warm-up.

5.4.2 Cryomodule Weight

Table 5.4, lists cryomodule component and total weights, in particular for the
LCLS-II cryomodule. ILC would be very similar. The bottom line is 8600 kg for
the cryomodule assembly.

5.4.3 Major Interfaces

Major interfaces from the cryomodule to other linac components include a wave
guide connection to the RF power input coupler, instrumentation connectors,
magnet power lead connections, the anchors from the adjustable supports to the
floor, pump connections to the insulating vacuum pump-out ports, and may also
include guard helium vacuum connections for helium-shrouded subatmospheric
connections.

5.5 Cryomodule Vacuum Design and Vacuum Vessel

The cryostat outer vacuum vessel is constructed from carbon steel and, for these
TESLA-style cryomodules, has a standard diameter of 38 inches. Adjacent vacuum
vessels are connected to each other by means of a cylindrical sleeve with a bellows,
sealed to each cryomodule end by an O-ring vacuum seal. In the event of accidental
spills of liquid helium from the cavity vessels, a relief valve on the sleeve together
with venting holes on the shields prevent excessive pressure build-up in the vacuum
vessel.
130 T.J. Peterson and J.G. Weisend II

Wires and cables of each module pass through the vacuum shell via metallic
sealed flanges with vacuum tight connectors. The insulating vacuum system is
pumped during normal operation by permanent pump stations located at appropriate
intervals. Additional pumping ports are available for movable pump carts, which
are used for initial pump down, and in the event of a helium leak. Due to isolation
of the cavity vacuum at the RF power coupler by means of a ceramic window, the
RF power coupler typically needs an additional vacuum system on its room tem-
perature side; this is provided by a common pump line for all couplers in a module,
which is equipped with an ion getter and a titanium sublimation pump.
A new feature for the LCLS-II cryomodules in particular is access ports for the
tuners, serving to provide access to the piezo tuners within the cryomodules. This
feature eliminates the need to pull the entire cold mass string out of the vacuum
vessel in order to access tuner components.
The insulating vacuum is protected from over pressurization by means of a
spring-loaded lift plate. A single worst case piping rupture internal to the insulating
vacuum must be analyzed to determine lift plate size. Provisions are provided to
allow free passage of the helium out past thermal shield and MLI to the lift plate,
which is typically 200–300 mm in diameter.
The cavity (beam) vacuum must be kept very clean and particle free. The
vacuum gate valves shown in Fig. 5.10 provide beam vacuum isolation during
cryostat assembly and interconnect assembly. In the linac, beam vacuum pumping
and readouts are provided at warm beamline sections. Warm beamline vacuum
components adjacent to the cryomodule cold beam vacuum must be clean and
particle free to avoid contamination of the cavity vacuum. Pumping may be pro-
vided by titanium sublimation pumps, with these at the ends of cryomodule strings
providing “guard” pumping to reduce cryopumping of the warm beamline into the
cold RF cavity vacuum. Ion pumps may provide vacuum readout

Fig. 5.12 LCLS-II cryomodule assembly cross-section

5 TESLA & ILC Cryomodules 131

5.6 Cryomodule Thermal Design and Helium Flow Design

5.6.1 Major Thermal Design Features

The RF cavities are maintained at nominally 2 K by means of a stagnant bath of

saturated liquid helium. Temperatures of 1.8–2.1 K are feasible in a large system,
and the RF cavity helium vessel and piping design accommodate any temperature
within this range. The 2 K helium is supplied by a pressurized, nominally 2 K line,
to the cryomodule or string liquid level control valves.
A nominally 5 K cooling circuit will typically incorporate pressures above the
helium critical pressure (2.27 bar) so as to avoid 2-phase flow. Although no phase
change takes place as the supercritical helium warms, in the temperature range of 5–
8 K as much as 30 J/g may be absorbed, comparable to that of a phase transition.
A 5 K circuit provides low temperature thermal intercept for the support posts,
magnet current leads, RF power coupler, HOM absorbers, and instrumentation
wires. Unlike the concept for ILC and the design of XFEL, CW cryomodules
typically have no thermal radiation shield at this temperature level since RF losses
already provide 10s of Watts or more per cryomodule at the 2 K level.
The highest temperature level will be helium in the range 35–85 K or perhaps
liquid nitrogen at around 80 K. This temperature level provides not only conductive
thermal intercepts but also a thermal radiation shield. At many institutions, there
will be no liquid nitrogen in the tunnel, implying the use of a gaseous helium
circuit. However, often for test purposes in various test cryostats and facilities, as
well as in some accelerators, the “70 K” thermal shield may be cooled with liquid
nitrogen at approximately 80 K. This higher temperature will have some impact on
thermal loads at the lower temperatures which should be assessed.
The thermal shield must be designed such that introduction of cold (process
temperature) helium into the thermal shield piping when the thermal shield is warm,
resulting in a very fast cool-down, does not damage the thermal shield or other parts
of the cryomodule. (The issues are warping and associated forces, thermal stresses,
etc.) Slots cut into the thermal shield perpendicular to the trace piping allow thermal
contraction of the pipe ahead of thermal shield cooling. Thermal shield trace piping
is arranged such that counterflow heat transfer does not inhibit cool-down of the
thermal shield.
Heat loads for early TESLA-style cryomodules are summarized in Sect. 5.8,
below. Special considerations for the high heat loads at 2 K with CW operation are
described in Sect. 5.6.2 of this document.
Evacuated multilayer insulation (MLI) is used within the cryomodule on the
thermal radiation shield, piping, and helium vessels. MLI on colder piping and
vessels under the thermal radiation shield, while not very effective in terms of
reducing overall heat load, greatly reduces boiloff rates from loss of vacuum
incidents, in turn reducing emergency venting pipe and valve size requirements.
Each helium vessel includes an electric heater for 2 K flow and pressure control.
In order to avoid cold feed-throughs from superfluid helium to insulating vacuum,
132 T.J. Peterson and J.G. Weisend II

these heaters are installed on the outsides of the helium vessels. The presence of a
steady-state pressure drop results in a pressure change at the cryomodule with a
change in flow rate (e.g. due to heat load change or liquid level control valve position
change), even with constant cold compressor inlet pressure (perfect cryoplant pres-
sure regulation). Heaters distributed within the cryomodules are required to com-
pensate for heat load changes so as to control subsequent flow and pressure changes.
Cables are thermally intercepted at the 70 K level. Special attention is given to
thermal intercepting of the piezo actuator wires and housing so as to assure piezo
temperatures remain below 80 K, to improve lifetime and performance. Input
coupler thermal intercepts are provided at 5 K and at 70 K.
Figure 5.12 illustrates the cryomodule cross-section and cryogenic circuit labels.
Figure 5.13 shows the cryomodule flow scheme. The large-scale cryogenic system
concept for TESLA is described in [11, 12].
Line A, the 2.2 K supply line, provides helium above the critical pressure (above
2.2 bar), so single-phase flow at all temperatures, to the liquid level control valves
which maintain the saturated 2 K liquid level in the 2-phase pipe over the RF
cavities.
Lines C and D are the low temperature helium thermal intercept supply and
return lines. Lines E and F, similarly, are the higher temperature thermal intercept
(and thermal shield) supply and return lines.
Line H, the warm-up/cool-down line, provides helium to capillary tubes which
supply the bottom of each helium vessel, allowing parallel warm-up or cool-down
of the string of helium vessels.

Fig. 5.13 LCLS-II 1.3 GHz cryomodule schematic. For ILC, the 2-phase pipes and cool-down
pipes are connected in series through a string of cryomodules
5 TESLA & ILC Cryomodules 133

Lines passing through cryomodule interconnects are all welded, the only
mechanical seals being the vacuum O-rings and all-metal beam tube vacuum seals.

5.6.2 Design for Large 2 K Heat Transport and Helium

Flow

Heat from the outside surface of the niobium RF cavity, and heat entering via
conduction from the beam pipe at the RF cavity ends, is carried through stagnant
saturated Helium II to the liquid helium surface in the 2-phase pipe via superfluid
heat transport. For heat transport through saturated superfluid helium around 2.0 K,
1 W/cm2. is a conservative rule for a vertical pipe [13, 14].
The critical heat flux for a non-vertical pipe connection from the helium vessel to
the 2-phase pipe may be considerably less than 1 W/cm2 (Fig. 5.14). Configurations
other than vertical require analysis to verify that the anticipated heat flux is less than
the critical heat flux. Also, temperatures above 2.0 K result in a lower critical heat
flux due to reduced superfluid heat transport. For LCLS-II, these considerations have
resulted in our increasing the inner diameter of the nozzle (or “chimney”) from
the helium vessel to the 2-phase pipe from 55 mm (the TESLA/XFEL size) to
95 mm.
End flange to cavity including support tabs must have sufficient helium
cross-sectional area to carry heat from the cavity ends and beam tube into the bulk
helium in the helium vessel (Fig. 5.15). Arrows indicate conduction through end of
power coupler outer conductor and beam pipe followed by heat transport through a
thin layer of liquid helium at the end of the first cavity cell.)

Fig. 5.14 Maximum heat flux in saturated superfluid at 2.0 K

134 T.J. Peterson and J.G. Weisend II

Fig. 5.15 Heat transport path

from input coupler and beam
tube region to 2 K helium

Evaporation from the surface of the saturated helium liquid results in vapor flow
within the 2-phase pipe over the liquid surface to an exit port connecting to the
HGRP. Studies done at CEA Grenoble for CERN’s LHC magnet cooling [15]
provided the result that a 5 m/s vapor “speed limit” over liquid is a conservative
“rule of thumb” not to entrain liquid droplets into the vapor. For LCLS-II, the
closure of each 2-phase pipe in each cryomodule, limiting the 2-phase pipe vapor
flow to that generated by one cryomodule, keeps the vapor flow rate quite low, less
than 2 m/s. Nevertheless, the 0.5 % slope (Fig. 5.16) combined with the fact that
the connecting nozzle from the helium vessel is 100 mm OD results in a 100 mm
OD 2-phase pipe, larger than the 69 mm for XFEL. The result is that the pressure
drop within the 2-phase pipe is negligibly small (Fig. 5.17).

Fig. 5.16 Liquid level illustration for single LCLS-II cryomodule (credit for figure—Fermilab)
5 TESLA & ILC Cryomodules 135

Fig. 5.17 Pressure drop and vapor velocity in 2-phase pipe in a single cryomodule. The vapor
exits from the 2-phase pipe at the 5700 mm central location. Liquid helium is added, with some
vapor also, at the 7500 mm location. The asymmetry of vapor speed and pressure drop is due both
to the off-center input of the 2-phase flow and a 0.5 % upward slope from left to right in the figure
(credit for analysis and plot—Joshua Kaluzny, Fermilab, and Joel Fuerst, Argonne National
Laboratory)

5.6.3 Pressure Drop Analyses

Pressure drops within and through the cryomodule must be analyzed in combination
with the helium distribution system. Pipes are sized for the worst case among
steady-state, peak flow rates, upset, cool-down, warm-up, and venting and condi-
tions. Pressure drops must be analyzed for each helium flow path to ensure that
steady-state operation matches system design and that non-steady conditions
(cool-down, emergency venting, warm-up) are properly handled. Input variables for
these analyses include line size, allowable temperature rise, allowable pressure drop,
and heat load. (Temperature rise and heat load determine mass flow for supercritical
pressure helium or helium gas in the thermal intercept pipes and 2 K supply pipe.)
For the HGRP, pressure drop is particularly important and limited, since this
pressure drop determines the helium temperature at the helium vessel, hence RF
cavity temperature. We design for a pressure drop <10 % of total pressure in normal
operation.
Maximum allowable pressure for emergency venting, combined with distances
to relief devices may also influence line sizes [16]. Helium piping and vessels vent
136 T.J. Peterson and J.G. Weisend II

Table 5.5 Cryomodule pipe size comparison

Pipe function Pipe name TTF inner XFEL inner ILC inner LCLS-II
(Fig. 5.12) diameter diameter diameter inner
(mm) (mm) (mm) diameter
(mm)
2.2 K subcooled A 45.2 45.2 60.2 54.8
supply
Gas helium return B 300 300 300 300
header, structural
support
5 K shield and C 54 54 56.1 54.8
intercept supply
5 K shield and D 50 65 69.9 50.8
intercept return
High temperature E 54 65 72.0 54.8
shield and intercept
supply
High temperature F 50 65 79.4 52.5
shield and intercept
return
2-phase pipe G 72.1 >72.1 69.0 97.4
Helium vessel to 54.9 54.9 54.9 95
2-phase pipe nozzle
(“chimney”)
Warm-up/cool-down H 38.9 38.9
line

into the adjacent cryomodules and out to the distribution system, allowing place-
ment of all process relief valves in the distribution system.
Loss of vacuum venting: pressure in the helium vessel of the dressed cavity less
than the cold maximum allowable working pressure (MAWP) of the helium vessel
and dressed cavity. Venting path includes nozzle from helium vessel, 2-phase pipe,
may include gas return pipe, and also includes any external vent lines. Worst-case heat
flux to liquid helium temperature metal surfaces with loss of vacuum to air is assumed
to be 4.0 W/cm2. Worst-case heat flux to liquid helium temperature surfaces covered
by at least 5 layers of multilayer insulation (MLI) is assumed to be 0.6 W/cm2.
Finally, we match cryomodule and cryogenic distribution system design to the
cryogenic plant in terms of providing flow rates, temperatures, and pressures
consistent with cryogenic plant requirements (Table 5.5).

5.6.4 Typical TESLA-Style Cryomodule Maximum

Allowable Working Pressures

See Table 5.6.

5 TESLA & ILC Cryomodules 137

Table 5.6 Maximum allowable working pressures for typical TESLA-style cryomodules
Region Warm MAWP (bar) Cold MAWP (bar)
2 K, low pressure space 2.05 4.1
2 K, positive pressure piping 20.0 20.0
(separated by valves from
low P space)
5 K piping 20.0 20.0
45 K piping 20.0 20.0
Insulating vacuum space 1 atm external with full
vacuum inside 0.5 positive
differential internal
Cavity vacuum 2.05 bar external with full 4.1 bar external with full
vacuum inside 0.5 positive vacuum inside 0.5 positive
differential internal differential internal
Beam pipe vacuum outside 1 atm external with full 1 atm external with full
of cavities vacuum inside 0.5 positive vacuum inside 0.5 positive
differential internal differential internal

5.6.5 Instrumentation

The cryomodule must be instrumented with liquid level probe (or probes) for the
2-phase helium II system when level control is incorporated into each cryomodule,
thermometry for cool-down and monitoring of critical input coupler, HOM coupler,
and current lead temperatures, and other instrumentation. Table 5.7 lists possible
instrumentation for a CW electron linac cryomodule.

5.6.6 Cryomodule Test Requirements

At least some fraction of cryomodules are typically tested prior to installation for
the following:
• Leak and pressure tests for quality assurance and pressure code compliance.
• Temperature profiles
• Approximate heat loads
• RF cavity performance
• Tuner performance
• Instrumentation

5.6.7 Pressure Stability at the 2 K Level

It is possible to generate pressure pulses within a cryomodule, for example via heat
input from the warm end of a closed pipe. Hence, one should try to avoid
138 T.J. Peterson and J.G. Weisend II

Table 5.7 Example of cryomodule instrumentation for a CW electron linac cryomodule

(LCLS-II). Production cryomodules may have a reduced set of necessary instrumentation
Sensor description Quantity and location
Piezo actuators (Fast Tuners) Typically 2 or 4 per tuner
Stepper motors One stepper motor per tuner
Embedded platinum temperature sensors on One for each stepper motor
the stepper motors
Cavity field probes—transmitted power 1 field probe per cavity
HOM field probes—transmitted power 1 on each HOM × 2 HOMs per cavity
Coupler Electron Pick-ups 2 Pick-ups per cavity (1 is mounted on the cavity
side of the coupler and 1 is mounted on the wave
guide side of the coupler).
Helium vessel cernox RTDs 1 RTD at each cavity
Main coupler platinum RTDs 2 RTDs at each cavity (1 at 12 o’clock & 1 at 10
o’clock).
Beam position monitor (BPM) 1 BPM per cryomodule—4 pick-ups
Coupler tuner motors 1 stepper motor, 2 limit switches, and 1
potentiometers for each main coupler (if remote
coupler tuning is incorporated)
Cool down temperature sensors 2 at the bottom of the 5 K shield and 3 inside the
300 mm pipe at each end
Cool down temperature sensors 2 RTDs—mounted on the bottom of the 70 K
HOM coupler RTDs 2 RTDs on each HOM coupler
BPM Pick-Up Temperature Platinum RTDs 2 platinum RTDs—1 primary and 1 redundant
Quadrupole and corrector coil voltage taps 6 VT’s (3 Pairs)—2 VT’s for the quad, 2 Vt’s
for coil quench protection for each of 2 correctors
Quadrupole and corrector coil cold mass 1 RTD for each coil package × 4 coil packages
temperature sensors
Quadrupole and corrector coil wire heater 1 wire heater—1 heater for the entire coil
package
Quadrupole and corrector coil voltage taps 2 voltage taps/current lead × 2 leads/coil
for power lead quench or current protection
Quadrupole and corrector coil lead RTDs 2 at 4 K region, 2 on 4 K intercept, 1 on
4 k-70 K region, and 1 at 70 K intercept
Helium vessel heaters 1 electric heater per helium vessel
Magnetic sensors Optional, typically for R&D or prototype
Beam loss monitor Depends on linac requirements
Liquid level sensors If cryomodules have individual liquid levels

“dead-headed” lines which can warm up, for example, the line terminating at the
cool-down valve after it is closed. To avoid a warm valve providing such a warm
termination on the closed pipe, one solution is to locate the valve lower than the
supply pipe such that cold helium sits on the valve.
The relatively large vapor volume contained in the HGRP combined with the use
of helium II result in an inherently stable pressure, since pressure fluctuations are
5 TESLA & ILC Cryomodules 139

not only damped by the large vapor volume, but helium II heat transport supports
liquid-vapor equilibrium at saturation temperature. An advantage of helium II, aside
from the tremendous heat transport capability, is this inherent pressure stability,
which in turn reduces the problem of helium pressure fluctuations affecting cavity
resonant frequency.

5.7 Cryomodule Helium Inventory

Table 5.8 lists cryomodule helium inventory for the various lines and volumes. Note
that most of the helium mass is contained in the helium vessels, the liquid around the
RF cavities. Depending on how liquid level is managed, there may also be additional
liquid helium in the 2-phase pipe. The large volume of the 300 mm pipe, filled with
low-pressure helium vapor, provides a nice buffer to damp pressure fluctuations and
to receive helium gas with sudden boiling due to some off-normal condition.

5.8 Early Results from the TESLA Cryomodules

The ILC cryomodule design evolved from the cryomodules designed for the
TESLA (TeV Superconducting Linear Accelerator) project in the 1990s. [1] Early
experimental results from these cryomodules led to design improvements seen in
the ILC cryomodules and also illustrate several important features of cryostat
design.
Table 5.9 shows the predicted and measured static heat leak for first two TESLA
cryomodules built and tested [17].
Notice that the measured heat leak of cryomodule #1 significantly exceeds the
predicted static at all three temperature levels with the 2 K heat load being more
than twice the predicted value. The explanation of a significant part of this dis-
crepancy can be seen in Fig. 5.18. Look first at the upper part of the figure. This
shows the measured temperatures at various points on both the 70 and 5 K thermal
shields and the support posts. Notice on the right side of the figure that sensors on
the innermost (5 K) thermal shield indicate temperature of 13–17 K rather 6–7 K as
shown on the rest of the shield. This result is indicative of an increased heat load to
that shield in that location. Cryomodule #1 was instrumented with a stretched wire
measurement system that allowed real time monitoring of the position of the
superconducting cavities so that the support system could be verified to meet the
alignment requirements (see below). This measurement system included large
numbers of instrumentation wire connecting the 2 K system to room temperature.
In cryomodule #1, there was inadequate heat sinking of the wires at the 70 K level
and thus much more heat than predicted was deposited at the 5 K shield level and at
the 2 K level. There was also found to be issues with the proper treatment of the
MLI penetrations (see Chap. 1) at the power coupler ports; which also added to the
static heat leak.
 140

Table 5.8 ILC cryomodule line sizes, helium conditions, and helium inventory
Location State Temp (K) Pressure (bar) Pipe ID (mm) Density (kg/m3) Volume (L) Mass (kg) Equiv liquid (L)
One helium vessel Liquid 2.00 0.031 145.700 23.0 3.35 26.8
9 helium vessels Liquid 2.00 0.031 145.700 207.0 30.16 241.3
2-phase pipe Vapor 2.00 0.031 69.0 0.830 47.3 0.04 0.3
2 K supply pipe Subcooled 2.40 1.200 60.2 147.800 36.0 5.32 42.6
300 mm pipe Vapor 2.00 0.031 300.0 0.830 893.9 0.74 5.9
5 K supply Supercritical 5.00 5.000 56.1 129.000 31.3 4.03 32.3
8 K return Supercritical 8.00 4.000 69.9 30.760 48.5 1.49 11.9
40 K supply Gas 40.00 16.000 72.0 18.460 51.5 0.95 7.6
80 K return Gas 80.00 14.000 79.4 8.220 62.6 0.51 4.1
Warmup line Vapor 2.00 0.03 38.9 0.830 15.0 0.01 0.1
Totals
One module 1393 43 346
Short string 9 modules 12,538 389 3115
Avg cryo unit 21 short string (system on one cryogenic plant) 263,288 8177 65,414
T.J. Peterson and J.G. Weisend II
5 TESLA & ILC Cryomodules 141

Table 5.9 Predicted and measured static heat leak for TESLA cryomodules #1 and #2 [17]
Temperature Predicted Measured heat Measured heat Measured heat
level (K) heat leak leak leak leak
(W) (W) cryomodule (W) cryomodule (W) cryomodule
#1 (alone) #1 (with #2) #2
70 76.8 90 81.5 77.9
4.5 13.9 23 15.9 13
2 2.8 6 5 4

Prior to the later test of cryomodule #1 with cryomodule #2, some of the issues
above were addressed and as a result the heat leak performance of Cryomodule #1
improved somewhat. Cryomodule #2 was constructed with far fewer instrumenta-
tion wires associated with the stretched wire system and those that remained were
better heat sunk. Examining the lower part of Fig. 5.18, which shows the shield
temperature distribution of cryomodule #2 shows the impact of these changes. The
temperatures at the right side of the 5 K thermal shield have been reduced and are
much closer to the nominal 5 K level. This result is also reflected in Table 5.9. The
measured heat leak for cryomodule #2 is within predictions for the 4.5 K level and
much closer to predictions for the 70 K and 2 K levels. Based on these experiences,
later ILC cryomodules have been built that meet the predicted static heat loads [18]
The story above illustrates the importance, as described in Chap. 1, of careful
design and heat sinking of instrumentation wires and of the need for proper
installation of MLI in penetrations.
The stretched wire system [19] that caused early heat leak problems did how-
ever, provide very useful data on the performance of the support system. Figure 21
shows the results of the displacement of the wire position monitors (WPMs) of the
stretched wire systems between room temperature and cryogenic temperatures for
cryomodule #2.
The various WPM positions are correlated to various cavity and magnet loca-
tions wthin the cryomodule. WPM positions 1–10 represent SRF cavity positions
while positions 11 and 12 represent the superconducting magnet assembly in the
cryomodule. Notice from Fig. 5.19 that the maximum displacement occurs at the
end of the cryomodule. This is a result of unbalanced forces on and resultant
bending of the 300 mm gas return pipe which acts as the structural backbone of the
cryomodule at to which the cavities and magnet assembly are fixed.
The allowed displacement for the cavities from their ideal location when cold is
±0.5 mm. Note that the measurements show that all the cavities, even the first one
are within this specification. However, for the superconducting magnet assembly,
the result is different. The magnet package is required to be within ±0.25 mm of its
ideal location when cold. The data in Fig. 5.19 shows that magnet assembly located
at the end of the cryomodule does not meet this requirement. As a result of these
measurements, changes have been made in the later XFEL and ILC cryomodules.
The support post at the magnet end of the cryomodule has been moved closer to the
magnet location. This has the effect of stiffening the 300 mm pipe near the magnet
142 T.J. Peterson and J.G. Weisend II

Fig. 5.18 Temperature measurements of the thermal shields and support posts for TESLA
cryomodule #1 & cryomodule 2 [17]

0.4
Horizontal displacement [mm]

0.3
0.2
0.1
0 1/11/98 11.44
-0.1 1/2/99 11.42
-0.2
22/3/99 10.40
-0.3
-0.4
-0.5
-0.6
0 1 2 3 4 5 6 7 8 9 10 11 12

WPM position [m]

Fig. 5.19 WPM displacement between room temperature (1/11/98) and 2 K for two separate cool
downs (1/2/99 and 22/3/99) [17]

assembly and reduces the displacement of the magnet assembly to a limit suitable
for use in XFEL and LCLS II. In the case of the much larger ILC machine, which
has tighter alignment requirements, the final design change will likely be to move
the magnet assembly to the middle of the cryomodule where the displacement is
minimized.
5 TESLA & ILC Cryomodules 143

In addition to illustrating important aspects of cryostat design, the early expe-

riences with the TESLA cryomodules also shows the importance of prototype tests
in cryostat design. The construction and testing of a prototype cryostat and the
subsequent altering of the design based on the prototype results, is an important part
of a successful cryostat project. Large systems such as the SSC magnets (Chap. 2),
LHC magnets (Chap. 3), SHOOT (Chap. 4), ILC (Chap. 5) and the 12 GeV
Upgrade (Chap. 6) all take this approach. It is important to allow sufficient time in
any project schedule for the prototyping process.

5.9 Modifications for CW Operation in the LCLS-II Linac

Modifications of a TESLA-style cryomodule for LCLS-II illustrate design features

for CW operation and design for relatively larger heat loads than a pulsed linac
would typically have. Similar considerations for high heat loads are described for
CW operation in Energy Recovery Linacs (ERL) by Kugeler et al. [14] LCLS-II
adopted the most recent 8-cavity TESLA-style as developed for XFEL, with magnet
at the end (unlike the ILC concept, which moved the magnet to the center), but
modified for CW operation and for the SLAC tunnel 0.5 % slope. Some major
differences from the TESLA-style design include:
• Modified power input coupler for CW power,
• Larger diameter 2-phase 2 K helium piping for increased heat and vapor
transport and for liquid helium management with the tunnel slope,
• Closure of each end of the 2-phase pipe with the addition of a liquid helium
supply valve in the cryomodule,
• Closure of each end of the cool-down/warm-up pipe manifold for individual
cryomodule cooling to facilitate Q0 preservation, with the addition of a
cool-down valve in each cryomodule,
• New end lever tuner design for increased stiffness and fine control,
• The addition of tuner access ports to facilitate inspections or small repairs of
tuner components,
• Removal of the 5 K thermal shield due to its marginal benefit relative to
dynamic heat loads and the advantages of simplification of design,
• And enhanced magnetic shielding for Q0 preservation.

5.10 Summary

The TESLA collaboration developed a unique variant of SRF cryomodule designs,

the chief feature being use of the large, low pressure helium vapor return pipe as the
structural support backbone of the cryomodule. Additional innovative features
include all cryogenic piping within the cryomodule (no parallel external cryogenic
144 T.J. Peterson and J.G. Weisend II

transfer line), long strings of RF cavities within a single cryomodule, and cry-
omodules connected in series. Several projects, including FLASH and XFEL at
DESY, LCLS-II at SLAC, and the ILC technical design have adopted this general
design concept.
Advantages include saving space by eliminating the external transfer line, rel-
atively tight packing of RF cavities along the beamline due to fewer warm-cold
transitions, and potentially lower costs. However, a primary disadvantage is the
relative lack of independence for warm-up, replacement, and cool-down of indi-
vidual cryomodules.

Acknowledgments We thank the scientists, engineers, and designers in the TESLA, ILC, and
LCLS-II collaborations, and in the TESLA Technology Collaboration for the use of figures from
3-D cryomodule design models and other information about TESLA-style cryomodules.

References

1. TESLA, Technical design report, part II: the accelerator, published by DESY, Notkestrasse 85,
22607 Hamburg, Mar 2001
2. International linear collider reference design report—ILC global design effort and world wide
study, in Volume 3: Accelerator, eds. by N. Phinney, N. Toge, N. Walker (2007)
3. International linear collider technical design report—ILC global design effort and world wide
study, in Part II: The ILC Baseline Reference, eds. by J. Carwardine, N. Phinney, N. Toge,
P. Burrows (2012)
4. TESLA Test Facility Linac Design Report, Editor: Don Edwards, March 1995
5. The European X-ray free-electron laser—technical design report. DESY XFEL project group,
European XFEL project team, Deutsches Elektronen-Synchrotron, Member of the Helmholtz
Association, Notkestrasse 85, 22607 Hamburg, Germany, July 2007. http://www.xfel.net
6. LCLS-II Final Design Report (LCLSII-1.1-DR-0251-R0) November 22, 2015
7. T.H. Nicol, R.C. Niemann, J.D. Gonczy, SSC magnet cryostat suspension system design, in
Advances in Cryogenic Engineering, vol. 33 (Plenum Press, New York, 19880, pp. 227–234
8. T.H. Nicol, R.C. Niemann, J.D. Gonczy, A suspension system for superconducting super
collider magnets, in Proceedings of the Eleventh International Cryogenic Engineering
Conference (Butterworth & Co., Surrey, UK, 1986), pp. 533–538
9. T.H. Nicol, TESLA test cell cryostat support post thermal and structural analysis, Fermilab
TM-1794, Aug 1992
10. A. Ballarino, Conduction-cooled, 60 a resistive current leads for LHC dipole correctors, LHC
project report 691, European laboratory for particle physics. CERN CH—1211 Geneva 23
Switzerland, 5 Mar 2004
11. G. Horlitz, T. Peterson, D. Trines, The TESLA 500 cryogenic system layout, in Advances in
Cryogenic Engineering 41A, ed. by P. Kittel, et al. (Plenum, New York, 1996), pp. 911–920
12. S. Wolff, H. Lierl,B. Petersen,J. Weisend II, The cryogenic system of TESLA 500—An
Update, in Proceedings of the International Cryogenic Engineering Conference ICEC 17
(Bournemouth, U.K., 1998), pp. 879–882
13. T. Peterson, Fermilab, Notes about the limits of heat transport from a TESLA helium vessel
with a nearly closed saturated bath of helium II”, TESLA report #94-18 (June 1994)
14. O. Kugeler, A. Neumann, W. Anders, J. Knobloch, Adapting TESLA technology for future
CW light sources using HoBiCaT. Rev. Sci. Instrum. 81, 074701 (2010)
(Helmholtz-Zentrum-Berlin (HZB), Berlin, Germany, p. 12489)
5 TESLA & ILC Cryomodules 145

15. B. Rousset, A. Gauthier, L. Grimaud, R. van Weelderen, Latest developments on He II

co-current two-phase flow studies, in Advances in Cryogenic Engineering, Cryogenic
Engineering Conference, vol. 43B , (1997), pp. 1441–1448
16. Petersen, B., Wolf, S., Numerical simulations of possible fault conditions in the cryogenic
operation of the TTF/FEL—and Tesla linear accelerator, in Proceedings of the 18th
International Cryogenic Engineering Conference (ICEC18), Mumbai, India (2000)
17. C. Pagani et al., Construction, commissioning and cryogenic performances of the first TESLA
test facility (TTF) cryomodule, Adv. Cryo. Eng. 43A (1998)
18. N. Ohuchi et al., Thermal performance of the S1-Global cryomodule for the ILC, in
Proceedings of IPAC2011 (2011)
19. D. Giove, A wire position monitor (WPM) system to control the cold mass movements inside
the TTF cryomodule, in Proceedings of the 1997 Particle Accelerator Conference, vol.
3 (1997)
Chapter 6
Segmented SRF Cryomodules

E. Daly, Thomas H. Nicol and J. Preble

Abstract Linear accelerators based on Superconducting Radio Frequency

(SRF) cavities have become increasingly important and wide spread. This chapter
reviews the design and operation of four different SRF segmented cryomodules.
These cryomodules were used in projects from the 1980s to the 2010s and as such
this chapter provides a good overview of the development of this type of cryostat.
Topics described include: requirements, cooling of SRF cavities, tuners, power
couplers, structural and thermal insulation systems, magnetic shields, vacuum
systems and instrumentation. Performance results from both prototypes and series
operation are presented.

6.1 Introduction

An important application of cryogenics in particle accelerators is the cooling of

Superconducting Radiofrequency (SRF) cavities. These cavities provide accelera-
tion to the charged particle beam by storing radiofrequency (RF) energy in a
resonant structure (or cavity). The system is designed such that the electrical field
produced in the SRF cavity accelerates the charged particles as they pass through
the cavity. The cavities are made from pure niobium operated in its superconducting
state to minimize the wall losses of the RF energy. Thus, more of the input RF
energy goes into accelerating the beam itself.

E. Daly (&)
J. Preble

Thomas Jefferson National Laboratory, 12000 Jefferson Avenue,
Newport News, VA 23606, USA
e-mail: [email protected]
J. Preble
e-mail: [email protected]
T.H. Nicol
Fermi National Accelerator Laboratory, P.O. Box 500, Batavia
IL 60510, USA
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 147

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_6
148 E. Daly et al.

Since this application involves RF energy and is thus an alternating current

application of superconductivity, there are always going to be wall losses and
heating within the superconducting niobium. These losses are a function of a
number of parameters including the RF frequency and the temperature. The optimal
temperature for most SRF applications is between 1.8 and 2.1 K.
The SRF cavities are built into cryostats known as cryomodules. These are
described below and in Chap. 5. In addition to the cavities themselves, SRF cry-
omodules contain other components. Power couplers (or couplers) bring the RF
energy from the room temperature power supplies to the SRF cavities. Tuners
adjust the resonant frequency of the cavities by physically changing the cavities
shape. Higher order mode (HOM) couplers remove unwanted frequencies generated
by the beam moving through the SRF cavity. A complementary approach is to
absorb the HOM frequencies in a dissipative material to prevent its propagation.
Such materials are known as HOM absorbers or HOM loads. In addition, the
cryomodules typically contain magnetic shielding to protect the cavities from stray
magnetic fields (including the earth’s) which can degrade cavity performance.
Requirements for cryomodules include: low heat leak, low vibration, tight
alignment tolerances on the cavities and a high vacuum level in the beam tube
vacuum for optimum cavity performance.
There are two ways that cryogenic and vacuum systems are distributed to
individual modules in superconducting magnets or cavity strings. The first, referred
to as fine segmentation or segmented, refers to systems in which the insulating
vacuum and the cryogenic circuits are confined to an individual cryomodule or
cryostat with the only connection between modules being the beam tube. This
approach has a number of advantages; it may allow for individual warm up and cool
down of cryomodules, the separate cryostat isolation vacuums prevents a failure in
one isolation vacuum space from propagating to other cryostats and the space in
between the segmented cryomodules may be filled with magnets, beam instru-
mentation and vacuum systems that are more easily operated at 300 K. Fine seg-
mentation is the most common approach used in SRF cryomodules and is seen in
the examples given in this chapter as well as in the European Spallation Source
(ESS) [1] and Facility for Rare Isotope Beams (FRIB) [2] projects.
There are disadvantages to fine segmentation. The biggest is that the many warm
to cold transitions add additional heat leak and are quite expensive. Thus, for
systems involving large numbers of cryostats or cryomodules, the second approach
sometimes called coarse segmentation or continuous is used. Course segmentation
refers to systems in which the cryogenic circuits and insulating vacuum inside
individual cryostats are more or less continuous for long lengths, at least over the
length of several cryomodules. Accelerator magnet systems such as the SSC
(Chap. 2) and the LHC (Chap. 3) are configured this way as are the SRF cry-
omodules envisioned for the ILC, LCLS II and the XFEL at DESY (Chap. 5).
Generally speaking, systems involving less than 100 cryostats or cryomodules
use the segmented approach while those involving more than 100 components use
the continuous approach.
6 Segmented SRF Cryomodules 149

This chapter describes the design and experience of four different segmented
cryomodules. The cryomodules are described in rough chronological order of their
design and were designed between the 1980s and the 2010s. Since these cry-
omodules have many similar requirements, their design solutions are similar.
However, an evolution of design can be also seen over the course of the devel-
opment of these cryomodules. Due to the historical nature of these descriptions, the
units used as those in place at the time of the original designs. Thus, a mixture of SI
and English units will be seen.

6.2 C20 Cryomodule Design for CEBAF

6.2.1 Introduction

The Continuous Electron Beam Accelerator Facility (CEBAF) was built as a

4 GeV, 200 µA, electron accelerating facility [3]. The C20 cryomodule is the
original design for the CEBAF accelerator and was developed in the late 1980’s.
CEBAF was the world’s largest installation of Superconducting Radio Frequency
(SRF) accelerating cavities. CEBAF had 338 identical SRF cavities housed in 43
cryostats. The cavities were assembled in pairs which became the natural design
segment for the cryomodules. Four of these pairs were housed in identical cryostats
that would be assembled together along with end cans to form a single cryomodule
(see Fig. 6.1). The exception to this was the first cryostat, a quarter cryomodule,
which included a single cavity pair cryostat with end caps.

6.2.2 Modularity and Segmentation

The CEBAF accelerator (see Fig. 6.2) includes two antiparallel linacs, an injector,
and three experimental halls. Each linac contains 25 repeating zones. The first 20 of
these zones contain one C20 cryomodule each. A zone is 10 m long and contains a
cryomodule and a warm beamline girder which are 8.35 and 1.65 m long respec-
tively. Each cryomodule has two helium supplies and returns, a primary circuit
operating at 2 K and a shield circuit operating at 50 K, eight SRF cavities with
5 kW radio frequency (RF) power feeds, and various electrical connections for RF

Fig. 6.1 C20 CEBAF cryomodule

150 E. Daly et al.

Fig. 6.2 CEBAF accelerator

signals, instrumentation, and motor controls for the helium supply valve and cavity
tuner stepper motors. A cryomodule can be removed and replaced with another
cryomodule or drift tube in about a week minimizing accelerator operational down
time. This has been valuable over the life of the machine as eleven cryomodules
have been removed and refurbished over the first 23 years of operation.

6.2.3 Requirements

The cryomodule requirements were detailed in a design handbook that covered the
entire CEBAF accelerator ensuring appropriate system integration. The top level
cryomodule requirements include an operating voltage of 20 MV, capable of
accelerating 200 µA of beam recirculated five times, a nominal operating heat load
of 68 and 140 W for the primary and shield circuits respectively. The eight SRF
cavities contained in a cryomodule are required to be aligned relative to the nominal
beam trajectory with a root-mean-square angular position of 2 mR and centered on
the beam axis to ±0.02 in. The detailed requirements contained in the design
handbook are the extension of these high level requirements to individual systems
or components. Table 6.1 shows a summary of these requirements.

6.2.4 Design Description and Choices

The original CEBAF cryomodule was the first large scale installation of SRF
technology in the US. The adoption of this new technology was a risk and many
design choices were made to allow for testing and verification of the cavity
6 Segmented SRF Cryomodules 151

Table 6.1 Summary of CEBAF C20 cryomodule requirements from CEBAF design handbook
(1991 revision)
Item Value Units
Number of cryomodules 42.25
Number of such sections used in injector 2.25
Number of such sections used in linac 40
Length of section, including magnetic elements 9.60 m
Number of cryounits per cryomodule 4
Number of cavities per cryomodule 8
Cavities
Type Superconducting
Duty cycle CW
Operating frequency 1497, ±0.00002 MHz
Accelerating gradient, active length ≥5 MV/m
Intrinsic limitation of accelerating gradient ≥20 MV/m
Power coupled into beam, per cavity 2.5 kW
Total linac-associated power consumption ≤10 MW
HOM impedances, for most important modes 500, 170,000
Regenerative beam breakup threshold, per beamlet >200 Actual µA
Regenerative beam breakup threshold, per beamlet >2000 Computed µA
Current transport capability 1000 µA
Number of cells per cavity 5
Operating mode π
Cell shape Elliptical
Fundamental power coupler, waveguide Beamline
HOM coupler Beamline
Output waveguide cutoff frequency 1900 MHz
Number of output waveguides 2
Angle between waveguides 90
Elastic pressure sensitivity <60 Hz/torr
Microphonic tolerance, cryomodule surface 10 microns/sec
Cryostat (cryounit)
Thermal intercept shield 1
Magnetic shield 2
Liquid Level probe 1
Superinsulation blankets 2
Support rods 10
Heater (100 W in liquid bath) 1
Cryogenic thermometers (minimum) 2
Bridge piece set
Bridging ring, insulating vacuum 1
Thermal shield closure with braided hose 1
Superinsulation blankets 2
Magnetic shield 1
Beampipe with pump out 1
Helium pumping line, with bellows 1
152 E. Daly et al.

performance during assembly and to minimize potential for compromising the

demonstrated performance by subsequent processes. Two SRF cavities were
assembled into a hermetically sealed Cavity Pair (CP), see Fig. 6.3, which under-
goes performance testing in a Vertical Test Area (VTA) [4] and are never vented
after testing to avoid contamination.
The cavities have frequency tuners installed and this assembly is mounted in a
316L stainless steel helium vessel. The helium vessel is assembled into a cryostat
that includes cold and warm layers of magnetic shielding, an intermediate tem-
perature thermal shield, multilayer insulation and a vacuum vessel. Together this
assembly is a Cryounit (CU) see Fig. 6.4. Four of these CUs are assembled with a
supply and return cryogenic end can to make a Cryomodule (CM). The cryomodule
is the smallest assembly that is transported to and installed in the accelerator.

Fig. 6.3 C20 cavity pair

Fig. 6.4 Bridging area between two cryounits

6 Segmented SRF Cryomodules 153

6.2.5 Cryogenic System Interfaces

Helium is supplied from a large central helium cryogenic plant [5] by a vacuum
jacketed distribution system [6]. At each cryomodule there are two supply and two
return bayonet connections. A cryomodule has two mating bayonet connections in
each end can. A typical connection in the cryomodule is a female bayonet with a
ball valve for isolation and a safety relief valve. All volumes that can have cold
helium trapped inside include safety relief valves. The supply end can includes a
primary and shield 3.8 cm bayonets. The primary supply carries 3 bar, 2.2 K
helium to a J-T valve. The J-T valve is controlled to maintain a set liquid level in
the cryomodule of 2 K, 0.031 bar liquid helium. The shield supply line is a 3 atm
40 K helium line with no valves internal to the cryomodule. The return end can
includes two bayonets and a primary circuit relief stack. The primary supply
bayonet is a larger 7.9 cm bayonet to support pumping the required mass flow at the
lower return pressure. The shield return bayonet is the same size as the supply. The
relief stack is designed for high flow rates due to the large liquid inventory and
potential for very high heat loads in the event of a loss of vacuum in the cryostat.
There is a two stage relief setting. A lower pressure, 1.2 atm, relief is sized for the
maximum mass flow through the inlet J-T valve and a higher pressure, 4.1 atm
differential, relief valve is sized for the flow associated with a loss of vacuum [7].
Both relief valves are reseating designs. All cryogenic connection between air and
sub-atmospheric helium include a double o-ring seal with vacuum pumping
between the o-rings to ensure air is not allowed to leak into the cryogenic circuit.

6.2.6 Vacuum Interfaces

The CEBAF linacs are separated from the warm beam transport vacuum chambers
as they include cold components that will cryopump all gasses and accumulate
contamination over time. The separation of the vacuum spaces is done with a
differential vacuum pump, an electrostatic precipitator, and a fast gate valve. The
combination of these elements is designed to protect against off normal vacuum
events in the warm beam transport chambers. The cryomodule has three separate
vacuum systems, the beamline which include one 30 L per second ion pump,
insulating vacuum, and the RF power coupler which includes one 20 L per second
ion pump each. The cryomodule beamline interface to the warm girder is an
electro-pneumatic controlled 70 mm viton sealed gate valve. The insulating vac-
uum has a single manually actuated 152 mm viton sealed gate valve that is con-
nected to a pumping station for initial pump down and maintenance as needed.
The RF power couplers are configured with a manifold allowing two couplers to be
pumped by a single ion pump. The manifold has a 70 mm manual right angle viton
o-ring sealed valve used initial pump down and maintenance operations.
154 E. Daly et al.

6.2.7 Heat Load Estimates

The CEBAF cryomodule is dominated by dynamic RF heat loads. The cavity

specification allows for a maximum dynamic heat load of 5.4 W per cavity. Due to
improved gradient performance of the cavities over the specification the typical
operating conditions for the cavities are at *150 % of the design values [8]. Along
with a dynamic waveguide heat load this accounts for the *100 W of dynamic 2 K
heat load. The static heat load budget for each cryomodule is 18 W and has been
measured to be closer to 14 W on average. The static heat load has several com-
ponents including the bayonets, beamline transitions from 2 to 300 K, and the RF
power couplers. The shield static heat load is measured at *150 W and comes
from the heat stationing of the bayonets, beamlines transitions and the RF power
couplers.

6.2.8 Cavity

The LE5 SRF cavity developed at Cornell’s University’s Newman Laboratory, has
five cells resonant at 1497 MHz, one fundamental power coupler (FPC) located at
one end of the cavity and two higher order mode couplers (HOM) located at the
opposite end. The FPC and HOM are waveguide couplers sized for specific RF
cutoff frequencies. The cavity is fabricated from high RRR sheet metal niobium.
The cavity is operated submerged in a 2 K helium bath. The high RRR is needed
for the improved thermal conductivity in order to minimize the temperature gradient
across the sheet metal keeping the RF current carrying surface at the lowest tem-
perature possible. The cavity was designed for high current operations and the
resulting superior HOM damping performance was an important factor in selecting
the design for CEBAF.

6.2.9 Cavity Pair

Two cavities are assembled together into a cavity pair (see Fig. 6.3) with the FPCs
together at the center of the pair [9]. The cavity pair includes an inner adapter
between the cavity flanges, two end adapters at the ends of the cavities, end dishes
for interfacing to the helium vessel, vacuum valves, and two PFC extensions
between the cavity FPCs and the helium vessel feedthrough plate. All components
inside the end adapters are made from niobium. All components are joined using
bolted connections with indium wire seals. The center of the cavity pair is fixed and
bellows in the end dish assemblies allow for the differential thermal contraction
between the niobium cavity pair and the stainless steel helium vessel.
6 Segmented SRF Cryomodules 155

6.2.10 Tuner

The SRF cavities are required to be tuned to 1497 MHz ± 50 Hz during 2 K

operation. After cooldown of the cavity the frequency is required to be tuned
±200 kHz to ensure operations at 1497 MHz. The frequency is controlled by a
mechanical tuner (see Fig. 6.5) that can stretch or compress the cavity longitudinally
changing the resonant frequency. The tuner resolution requirement is 50 Hz. With the
cavity frequency sensitivity of *200 kHz per millimeter, the resulting mechanical
resolution requirement is 25 nm. The tuner attaches to the first and fifth cell of the
cavity and is submerged in liquid helium during operation. The tuner is driven by a
stepper motor mounted external to the cryomodule that transfers torque through two
rotary feedthroughs. The rotary feedthrough connects to a right angle driveshaft. The
driveshaft ends in a ball screw that acts to change the length of the tuner active leg.
The two cell holders are attached on one side to the active tuner leg and the opposing
tuner dead leg. The tuner includes limit switches and hard stops to limit the tuner
range and prevent damage to the cavity. Major disadvantages of this design include
the requirement for the tuner to be submerged in the helium bath, required rotary
feedthroughs on the vacuum and helium vessels, and the backlash as the tuner goes
from compression to tension (in later implementation of the tuner the backlash was
eliminated by biasing the tune of the cavity and operating the tuner in tension only).
Early accelerated life testing was done [10] to qualify the mechanical systems.

6.2.11 Helium Vessel

The helium vessel has several major functions in the CM design. The first is to house
the cavity pair in a bath of liquid helium. This requires the vessel to be leak tight and

Fig. 6.5 C20 cavity frequency tuner

156 E. Daly et al.

operate over a range of temperature from 350 to 2 K. The second is to provide the
required feedthroughs for RF power, instrumentation, and tuner mechanical drive to
the cavity pair. Additionally, the helium vessel provides the support structure for the
cavity pair inside the cryostat. The body of the helium vessel is a 24 in outer diameter
cylinder with a massive feedthrough plate. The entire assembly is made from 316L
stainless steel. The feedthrough plate is a thick plate to provide a stable interface for
the waveguide indium seals, instrumentation metal gaskets seals, and welded rotary
feedthrough connections. After inserting the cavity pair inside the helium vessel
body the waveguide indium seals are made up and the end dish assemblies are
welded to dished heads that bridge the end dishes to the helium vessel body closing
the assembly. The dished head assemblies include mounting supports and a 4 in.
upper and 0.75 in. lower helium connection. The mounting supports accommodate
four nitronic rod on each end that spa from the vacuum vessel to the helium vessel.
The large helium connection provides the cross sectional area required for heat and
mass transport from one helium vessel to the next and the lower connections
maintains constant liquid level throughout the four helium vessels in the CM.

6.2.12 Input Coupler

The CEBAF cryomodule uses a waveguide RF fundamental Power Coupler (FPC).

The FPC consists of a warm RF vacuum window that provides a hermetic seal
between the RF distribution system air filled waveguide and the cryomodule vacuum
waveguide, a warm waveguide assembly, a warm to cold waveguide transition [11]
and a cold RF window that provides a hermetic seal between the cryomodule
waveguide vacuum and the cavity vacuum. The waveguide vacuum space between
the warm and cold RF windows is monitored for vacuum, arcing, and heating. The
warm to cold transition waveguide assembly is a thin stainless steel rectangular
waveguide with two bellows, stiffening rings for the vacuum load, a shield heat
station, flanges on each end, and is copper plated on the inside. The assembly is
optimized for a nominal 1500 MHz RF power of 5 kW. This determines the
thickness of the copper plating and the location of the thermal intercept. As the RF
power increases the optimum position for the intercept moves toward the cold end of
the waveguide. The bellows are three convolutions made of thin stainless steel and
located close to the ends of the waveguide. The bellows makes up for any slight
misalignment between the helium vessel feedthrough plate to the sheet metal cav-
ities. Additionally, the thin walled bellows create a thermal break in the waveguide
wall. The FPC contributes 0.8 and 5.5 W to the 2 and 40 K heat loads respectively.

6.2.13 HOM Loads

The LE5 cavity selection for use in CEBAF was partly due to the excellent HOM
damping of the cavity and HOM loads. The HOM loads are assembled on the end
6 Segmented SRF Cryomodules 157

of a superconducting waveguide that is mounted on the cavity pair and remains

inside the helium vessel. For the CEBAF beam currents the dissipated power in a
HOM absorber is a fraction of a watt making it preferable to dissipate the HOM
power in the helium bath rather than bringing the heat to the shield or room
temperature. The size, both cross section and length, of the waveguide is such that
the higher frequency HOM RF power is transmitted to the loads while the funda-
mental power is cutoff and does not propagate to the HOM load. The simple design
of a cutoff waveguide has many advantages for fabrication and assembly.
The HOM load material is a particular challenge and was the subject of an extensive
development program [12].

6.2.14 Magnetic Shields—Inner and Outer

There are two passive magnetic shields in the cryomodule. One located on the
outside of the helium vessel which operates close to 2 K and a warm shield just
inside the vacuum vessel that operates at room temperature. Both shields are made
from the same high permeability shielding material *0.01–0.02 in thick. Efforts are
made to minimize the holes, joints, or other features that may allow magnetic field
into the cavity location. Joints are made with overlaps and taped to ensure minimum
leakage in the shield. Considerable effort is given to the details of the mechanical
design as any cold working of the material during assembly can degrade the per-
formance. The shield around the exterior of each helium vessel is joined with
bridging material extending to the next helium vessel forming a continuous cylinder
down the length of the cryomodule that is capped at each end at the end can end
plate. Subsequent designs have used segment shields that have end caps after each
cavity. These segmented designs provide improved shielding performance.
Penetrations from the warm to cold parts of the cryomodule require holes in the
shielding. The waveguide penetrations to the cavities are a particularly difficult area
to shield as the waveguides themselves have a significant cross sectional area.
Extensions can be used to help attenuate the field leakage around a hole in the
shielding and should be used where possible. Significant effort was made to
eliminate any magnetic material inside the magnetic shielding. The use of 316L
stainless steel was required inside the vessel and 304 stainless steel was permitted
outside of the magnetic shielding.

6.2.15 Thermal Shield and Multilayer Insulation

Multilayer insulation (MLI) is comprised of alternating layers of thin aluminized

Mylar and a low thermal conductive spacer material. MLI is used between the
helium vessel and thermal shield and between the thermal shield and vacuum
vessel. A nominal 15–30 and 45–60 layers are used respectively. MLI is used to
158 E. Daly et al.

reduce the radiative heat load to the cryogenic circuits and to reduce the heat flux to
the 2 K circuit in the case of a catastrophic loss of insulating vacuum. A copper
sheet metal thermal shield is located between the 2 K and room temperature sur-
faces. The shield is cooled by a single straight line running the length of the
cryostat. The thermal shield acts to intercept radiative heat load and is connected to
all warm to cold transitions to intercept conduction heat loads.

6.2.16 Vacuum Vessel

The vacuum vessel serves multiple functions in the cryomodule. The 304 stainless
steel vessel is the boundary of the insulating vacuum, the support structure for the
helium vessels, and provides penetrations for the RF power coupler waveguides and
instrumentation feedthroughs. The vacuum vessel is a cylinder with weld flanges on
both ends and a large access port aligned with the helium vessel feedthrough plate.
Mounted on each weld flange are supports for the nitronic rods that connect the
vacuum vessel and helium vessel as a low thermal conducting structural support.
The access port is closed with a Top Hat assembly that has provisions for the RF
power waveguides, cavity tuner mechanical drives, RF signals, and instrumentation
feedthroughs. The Top Hat area contains almost all of the penetrations through the
insulating vacuum shell.

6.2.17 Cryounit

The CEBAF Cryounit (CU) is the assembly of the helium vessel, power couplers,
instrumentation and vacuum vessel around a cavity pair (see Fig. 6.6). A layer of
magnetic shielding and MLI are assembled onto the outside of the vessel. This is
inserted into the vacuum vessel which has the thermal shield and MLI preinstalled.
The helium vessel is support by four nitronic rods on each end configured in a
double x pattern for alignment purposes. Two additional longitudinal nitronic
restraint rods are installed to complete the installation. The nitronic rods are used to
rough align the cavity flanges to the center of the vacuum vessel and tensioned to
set the load on the rods for maintaining the cavity position during and after cool-
down. The Vacuum vessel access port is used to install the FPC warm to cold
waveguide extensions along with thermal straps and instrumentation wiring.
A aluminum “Top Hat” closes the access port and has warm waveguide and
window assemblies installed. The warm waveguides are connected to a vacuum
manifold allowing a single ion pump to support two FPCs. This assembly maintains
the hermetic seal of the cavity pair and is the building block for the cryomodules.
Four CU are assembled together in a cryomodule.
6 Segmented SRF Cryomodules 159

Fig. 6.6 CEBAF C20 cryounit

6.2.18 Instrumentation

The cryomodule cryogenic instrumentation includes temperature sensors, helium

liquid level probes, helium bath pressure transducers, vacuum gauges and pumps,
and a J-T actuator that is driven with a 24 VDC motor and the valve position is
monitored with LVDTs. The temperature sensors are silicon diodes with *1/4
degree accuracy. In operation the absolute accuracy of the temperature sensors has
not been important as these are primarily used for controlling cooldown where rate
of change is important for identifying off-normal conditions like a degraded insu-
lating vacuum that produce large changes in temperature. The liquid level probes
used are a superconducting wire type. It is important to adjust the excitation current
160 E. Daly et al.

to the operating conditions as the helium pressure changes the cooling on the
superconducting wire. Warm capacitance manometer pressure transducers are used
for pressure measurement. The connection of the pressure transducer to the helium
bath should keep the 2 K heat load small while trying to maximize the conductance
from the bath to the transducer. In the C20 design the annual space of the return
bayonet is used for this connection. This adds no heat load to the system but does
limit the conductance and associated bandwidth of the measurement made possible
with the connection. The waveguide vacuum space is monitored for RF heating
using an infrared sensor mounted inside the waveguide vacuum and isolated from
the RF by a cutoff tube, RF discharge with a photomultiplier tube mounted outside
the vacuum space through a transparent window mounted on a metal seal flange,
and vacuum discharge using a 20 L per second ion pump. Additional vacuum
monitoring is done for the insulating vacuum using a cold cathode gauge and the
beamline using a 30 L per second ion pump. The waveguide instrumentation is
used in the RF interlock chain while the beamline ion pump is used in the RF and
beamline valve interlock chains.

6.2.19 Final Assembly

The cryomodule is made up from six major components, four CUs and two end
cans. The CUs and end cans are assembled onto a rail where they are aligned in x
and y as well as longitudinally. The location of the end cans is tightly controlled.
This along with the careful control of the cryogenic distribution system allows for
the use of standard size u-tubes avoiding the need for custom fabrications. The CUs
and end cans are assembled with a set of “bridging” components. Each bridging
section between the components has the helium circuits connected with welded
connections and the MIL, thermal shield, and magnetic shields are extended, and
insulating bridging ring is positioned and welded in place. The final alignment of
the cavities is done by adjusting the nitronic rods which are accessed through small
o-ring sealed ports located on each bridging ring.

6.2.20 Status

The first C20 cryomodule was installed in the CEBAF tunnel in 1990 and the
machine was completed in 1993. Since that time the cryomodules have performed
well above their design parameters. The original 5 pass energy of 4 GeV was
exceeded and 6 GeV operations was achieved with the original cryomodules. Since
that time there has been a slow degradation in gradient performance of some
cryomodules. As a result a refurbishment program was initiated. The progress with
cavity processing in the intervening time allowed the performance specification for
cavity gradient to be increased to 12.5 MV/m. This results in a “C50” cryomodule
6 Segmented SRF Cryomodules 161

with 50 meV energy gain per cryomodule per pass. Along with the increase in
cavity performance some designs were improved during this work. The FPC
waveguides had a newly designed “dogleg” section installed eliminating flashover
on the ceramic RF window and the tuner mechanical coupling was modified to
eliminate twisting and resulting backlash in the drive train. To date eleven cry-
omodules have been refurbished and the twelfth is in process.

6.3 The Spallation Neutron Source (SNS) Cryomodule

6.3.1 Introduction

The SNS is a 1 GeV negative hydrogen ion accelerator with up to 2 MW power

producing a source of neutrons for materials research. The initial portion of the
acceleration is achieved via a conventional negative proton injector, a drift tube
linac (DTL) and a coupled cavity linac (CCL), that provide a nominal energy gain
up to 185 MeV. The machine was changed in December 1999 from a warm
temperature to a cold temperature linac to improve overall machine performance.
The super-conducting linac (SCL), discussed here, contains 11 medium beta
cryomodules capable of 345 MeV and 12–21 high beta cryomodules capable of up
to 1300 MeV. After passing through an accumulator ring the beam goes to a
mercury target where a neutron beam of as much power as 2 MW is produced.
The cryomodule (CM) is based on the CEBAF CM with improvements bor-
rowed from LHC, TESLA, and the JLab 12 GeV upgrade and uses the frequency
scaled KEK fundamental power coupler (FPC). Figure 6.7 is the elevation view of
the high beta CM, while Fig. 6.8 is the flow schematic. The FPC requires a 4.5 K
lead flow to cool the outer conductor; therefore the LHC concept of producing the
2 K in the CM rather than in the refrigerator is utilized.
The refrigerator produces a 3 bar, 4.5 K stream, which feeds two Joule-
Thomson (JT) valves in parallel. The first supplies a small sub-cooler in the CM and
then cools the cavity. The second feeds the power coupler outer conductor. The CM

Fig. 6.7 High beta cryomodule schematic

162 E. Daly et al.

Fig. 6.8 Flow schematic

shield is cooled by a 4 bar, 35 K stream, which first cools the supply transfer line
(TL) shield, then the CM shield, and finally the return TL shield before returning to
the refrigerator at 52 K. The bayonet design permits replacement of a CM in less
than a day if needed without warming up the entire linac. In the nine years since the
initial CEBAF cooldown, the linacs have never been warmed up and only four CMs
have been replaced during scheduled accelerator shutdowns.
The relevant parameters for both medium and high beta CMs are given in
Table 6.2 and the relevant refrigeration capacities are given in Table 6.3.

6.3.2 Cavity String

The design, manufacture, and performance of the SRF cavities are reported else-
where [13]. The medium beta CM consists of three six-cell cavities while there are
four six-cell cavities for the high beta CMs. The SNS cavities operate at 805 MHz.
During initial cold tests, the cavities have met the design requirement for acceler-
ating gradient, which is an Epeak of 27.5 MV/m at a design quality factor, Q0, of
6 × 109 at 2.1 K. Power dissipation per cavity at 7 % duty cycle is 2 and 3.5 W for
the medium and high beta cavities respectively. The relationship of Q0 with tem-
perature follows BCS theory. Accordingly, the Q0 of 14 × 109 at 2 K decreases to
6.9 × 109 at 2.3 K. The two different beta cavities are elliptical in shape,
6 Segmented SRF Cryomodules 163

Table 6.2 SNS cryomodule parameters

Medium High
Slot length 5.839 m 7.891 m
CM length (bore tube) 4.239 m 6.291 m
CM diameter 1.22 m (*48″)
2 K Heat load (static/dynamic) 25/14 W 28/20 W
Maximum coupler flow 0.075 g/s
Shield heat load including Transfer Line 170 W 200 W
Tunnel H × W 10 × 14 ft
Control valves per CM 5
Bayonets per CM 4
Radiation hardness 108 rads
Pressure rating 2 K System Warm 3 atm
Cold 5 atm
Shield and 4.5 K systems 20 atm

Table 6.3 Refrigeration capacities

He temp (K) Capacity (W) Pressure (atm) Flow (g/sec)
Linac shields 35–52 8300 4.0 90
Linac cavities 2.1 2400 0.041 120
Secondary 4.5 3.0 0.15

manufactured from 4 mm thick niobium and have stiffening rings at 80 mm to

minimize microphonics. The performance of the first prototype medium beta cavity,
shown in Fig. 6.9, exceeded the required performance.
The cavities are housed in a titanium helium vessel, which matches the coeffi-
cient of thermal expansion of the cavity. The cavity is maintained at operating
frequency through a TESLA-style tuner mounted on one end of the helium vessel,
which operates through a bellows. The tuner, manufactured out of stainless steel, is
actuated through a cold stepping motor through a harmonic drive. Power is brought
into the cavity through the coaxial FPC [14] at one end of the cavity. Higher order
mode extraction filters are attached at each end of the cavity to damp some
potentially dangerous longitudinal modes. The string of cavities in the helium
vessel with couplers, HOM damping filters and hermetic valves are assembled in
JLab Class 100 clean room to minimize contamination. Figure 6.10 shows the
arrangement of the cavity string including the cavity, coupler, helium vessel and
tuner.
164 E. Daly et al.

6 cells b=0.61 cavity 6SNS61-1

Q0 vs. E acc

T=1.94K T=1.96K T=1.96K -20' RF processing

1.00E+11
Test #3
Q0

1.00E+10

1.00E+09
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0
Eacc [MV/m]

Fig. 6.9 Medium beta cavity performance

Fig. 6.10 Cavity string components inside the space frame situated on tooling

6.3.3 Cryomodule

A high beta CM consists of two end cans, a vacuum tank, a space frame, a thermal
shield, two magnetic shields and a hermetically sealed string of four cavities in the
helium vessel each with FPCs, a field probe, two HOM filters, bellows between
cavities and seal valves. A medium beta CM is similar in construction but houses a
string of three hermetically sealed cavities instead of four. The general arrangement
of these components is shown in Figs. 6.11 and 6.12. The CMs used in CEBAF, the
6 Segmented SRF Cryomodules 165

Fig. 6.11 High beta cut-away

Fig. 6.12 Medium beta cut-away

largest use of SRF in the US, employs a similar construction arrangement. What
makes the SNS design unique is that the CMs are assembled at JLab and shipped to
the SNS site at Oak Ridge National Laboratory (ORNL). This led to the devel-
opment of a structure that can handle the over-the-road shipment of 500 miles while
maintaining alignment of the cavities in a low loss cryostat [15]. The space frame
was developed at JLab for the next generation JLab CM to facilitate installation of
long strings of cavities efficiently at a relatively low cost. The space frame was
adapted and strengthened to handle the relatively high transportation loads incurred
during shipment to ORNL. The vacuum tank provides a support structure for the
space frame and minimizes the gaseous conduction to the cold surfaces on the
cavity string. The string is supported off of the space frame through nitronic rods,
which limit the solid conduction between room temperature and the cryogenic
surfaces. These rods are also able to take the shipping loads and maintain the
cavities in alignment. A thermal shield, operating at 50 K, surrounded with
multilayer insulation (MLI) provides a radiation barrier between the cavities and the
outside world. There are two magnetic shields, one outside the space frame and the
166 E. Daly et al.

other at the helium vessel, which in concert reduce the earth’s and stray magnetic
fields by a factor of *100 to minimize the effect on the cavity Q0. L shaped end
cans, a design developed for CEBAF to save space, close off the cavity string in the
vacuum tank and provides the interface for the helium to cool the cavities, the
couplers and the thermal shields. Between each CM is a 1.6 m warm space that
contains quadrupole magnets and diagnostics including beam position monitors,
current transformers and wire scanners.

6.3.4 Cryomodule Heat Loads and Thermal Design

The design parameters for the CM were developed in conjunction with the design
parameters for both the cryogenic system and the SC portion of the linac and are
given in Table 6.2. Specifically, the heat load budgets were developed considering
operational experience with the CEBAF cryogenic system, where applicable. The
heat load estimates and the budget for the CM are given in Table 6.4. This section
contains a summary of the heat loads as well as detailed descriptions of the con-
tributions of each subsystem.

6.3.4.1 Cavity, Helium Vessel and Tuner

The SNS helium vessel design, described in [16], supports the cavity during all
phases of operation, facilitates cavity tuning and contains the cryogens with
appropriate plumbing. Each 0.61 m diameter titanium helium vessel contains a
single niobium cavity whose six cells are immersed in *150 L of He-II during
normal operation. The ends of the cavities protrude through the helium vessel heads
into the insulating vacuum to allow beamline assembly as well as attachment of
FPCs, Higher Order Mode (HOM) filters and field probes. These end groups are
subject to radiation heat transfer from the 50 K shield and the FPC, solid heat
conduction from instrumentation leads and the FPC as well as power generated
within the HOMs. Extraction of HOM signals to room temperature is planned,
requiring a 3.6 mm semi-rigid coaxial cable running from the cavity to the vacuum
tank exterior. Based on finite element analysis, the heat loads incident on the HOMs
must not exceed 0.25 W in order to maintain the HOMs well below the niobium
critical temperature. Additionally, instrument wiring for temperature diodes, liquid
level sensors, and heaters are routed from the helium vessel to the vacuum tank
exterior. The helium vessel heads possess attachment points for Nitronic-50™
stainless steel support rods and the cavity tuner. The calculated heat load inter-
cepted by the cavity and helium vessel assembly is given in Table 6.4.
The dynamic heat load for each cavity design has been calculated for
Epeak = 27.5 MV/m. It is desirable to increase the accelerating gradient such that
Epeak = 35 MV/m, thereby reducing the number of high beta cryomodules required
6 Segmented SRF Cryomodules 167

Table 6.4 Calculated heat loads (W) for the medium and high β cryomodules
Estimate summary MBCM HBCM
Qty 2.1 K 50 K Qty 2.1 K 50 K
Static—(see itemized list) 1 9.7 131 1 11.5 161
U-Tube Allotment 1 10.0 24 1 10.0 30
Dynamic (see itemized list) 1 8.3 1 17.1
Total heat load per CM 28.0 155 38.6 191
Budget per CM 39.0 170 48.0 200
Dynamic contributions
Cavity 3 6.0 4 14.0
Power coupler 3 0.6 4 0.8
Bellows 2 0.2 3 0.3
HOM (2 per cavity) 3 1.5 4 2
Total dynamic 8.3 17.1
Static contributions
Radiation—HV and bellows 3 1.1 41.7 4 1.8 65.3
Power coupler (radiation) 3 2.1 4 2.8
Tuner 3 0.75 4 1
He vessel supports 3 0.2 18 4 0.3 24
Warm beam tube conduction 2 0.1 2 2 0.1 2.5
Warm beam tube radiation 2 0.9 0.9 2 0.9 0.9
Cables (3 per cavity) 1 0.5 1.8 1 0.5 1.8
Supply bayonets 2 1.0 12 2 1.0 12
Radiation 1 0.04 9.0 1 0.04 9.0
PC J-T valve 1 0.25 2 1 0.25 2
Subcooler J-T valve 1 0.25 2 1 0.25 2
Shield relief 2 0.0 4 2 0.0 4
5 K transfer line 1 0.1 1 0.1
50 K transfer line 1 3 1 3
Return bayonets 2 1.5 6 2 1.5 6
Radiation 1 0.1 11.2 1 0.1 11.2
Cooldown/PC return 1 0.25 10 1 0.25 10
Shield relief 1 0.3 2 1 0.3 2
Cooldown valve 1 0.25 2 1 0.25 2
5 K transfer line 1 0.0 1 0.0
50 K transfer line 1 3 1 3
Total static 9.7 131 11.5 161

to reach a given linac output energy. The dynamic heat load increases at least
quadratically with increasing peak gradients or higher if operated in the field
emission region. These increased losses would be handled by some portion of the
available margin in the primary circuit of the refrigeration system.
168 E. Daly et al.

Fig. 6.13 Nitronic-50™

support rods attach the helium
vessel assembly to the space Vacuum Vessel
frame

Support Rod

Space Frame

Fundamental
Power Coupler

6.3.4.2 Vacuum Vessel and Space Frame

The vacuum vessel and space frame, described in [17], serve to locate the cavities
accurately within the CM via the Nitronic-50™ support rods, provide the structural
links to the external supports and remain at room temperature (300 K) during
normal operation (Fig. 6.13). The 0.99 m diameter vacuum vessel contains the
insulating vacuum and provides pressure containment in the unlikely event of
cryogenic piping failure. The insulating vacuum is cryo-pumped during accelerator
operation to less than 1.3 × 10−9 atm, an adequate vacuum level to minimize
residual gas conduction. If a cryogenic piping failure occurs, two spring-loaded
parallel plate pressure reliefs, identical to those used in the CEBAF CM, open at
1.2 atm on the vacuum vessel. In addition, there are many flanged penetrations in
the vacuum vessel: ports in the bottom of the vacuum vessel through which the FPC
provides RF power to the cavities, instrumentation ports containing electrical
feed-throughs for diodes, cavity heaters, tuner power and limit switches, field
probes, and connections for the warm helium coupler exhaust.

6.3.4.3 Thermal Radiation Shield

As in CEBAF, a copper shield, 2.37 mm thick, operating between 35 and 50 K is

used to intercept thermal radiation (Fig. 6.14). The shield is cooled by supercritical
helium gas (4 atm, 35 K) through a single pipe, 22.2 mm outer diameter by 1 mm
thick wall. Thermal analysis of both single-pass and three-pass shield cooling
showed *2 % increase in the total static heat load to 2 K with a single-pass
cooling scheme. Cost ultimately dictated the choice of single pass versus three-pass
piping.
The shield is divided into segments that cover the helium vessel assemblies and
bridges that cover both the cavity-cavity and cavity-end-can interconnect regions.
The bridge sections allow access to the cavity string for assembly, alignment,
instrumentation wire routing and tuner maintenance. There are three segments in a
6 Segmented SRF Cryomodules 169

Fig. 6.14 MBCM thermal shield assembly (left) and piping strain-relief (right)

medium beta cryomodule and four segments in a high beta cryomodule.

A hydro-formed stainless steel bellows, soldered with EasyFlo-45™, connects the
copper piping between segments. In addition to the bellows, a series of tabs 76 mm
long × 76 mm wide, where the pipe is soldered with BCuP to the shield, provides
strain relief from differential thermal contraction of the piping and shield during
cool down and operation.
The horizontal and vertical helium vessel Nitronic-50™ support rods are
heat-stationed to the shield segments by copper straps, 1 mm × 25 mm × 125 mm
long. All wiring and cabling routed to 2 K is heat stationed at 50 K on the nearest
shield segment as well. Due to the long length (>600 mm), small cross-sectional
area (20 mm2) and difficulty during assembly, the two axial support rods are not
heat stationed. The additional heat load to 2 K is negligible (<1 mW/support).
Each bridge is fixed at only one end to an adjoining segment with PEM fasteners
and allowed to slide over the other neighboring segment. Each segment-bridge
assembly weighing *45 kg is supported by eight G10 straps, 38 mm × 4.8
mm × 508 mm long, terminated with stainless steel pipe clamps affixed to the
space frame support tubes. The four horizontally oriented supports react lateral
shipping and handling loads, the four vertically oriented supports react both gravity
and transportation loads and all eight react any axial loads.

6.3.4.4 Multilayer Insulation (MLI)

The insulation system design is identical in principle to that employed in the

original CEBAF CM and is expected to achieve the same thermal performance. The
insulation scheme reduces the radiative heat load to the cold surfaces, provides
ample mass and heat capacity to mitigate thermal transients and is comprised of
materials suitable for use in a high radiation environment.
Blankets are constructed of alternating layers of double-aluminized Mylar™
(DAM), 25 μm thick, and Reemay™ #2250, a spun-bonded polyester 75 μm thick.
The maximum emissivity allowed by emissometer measurement is 0.035 at room
170 E. Daly et al.

temperature. The surfaces of the helium vessel are covered with two 12-layer
blankets, the beamline components and piping are spiral-wrapped with 15 layers at
a 50 % overlap, and the 50 K surfaces are covered with four 15-layer blankets. The
joints and seams, staggered both axially and around the circumference by
25–50 mm, are closed with 25.4 mm wide aluminized Mylar™ tape.
Effective thermal conductance through the blanket layers is calculated consid-
ering three mechanisms: radiative heat transfer—governed by emissivity (ε) and
geometry, solid conduction—governed by effective thermal conductivity (keff) and
geometry, and residual gas conduction—governed by average local pressure.
A Fortran program, TRANSAM, written originally for transfer line design was used
to estimate the expected heat flux to the 50 and 2 K surfaces. Using a worst-case
scenario with average pressure of 10−4 torr, ε = 0.2 and keff = 1.5 mW/cm K, the
heat flux to the 50 K thermal shield was calculated conservatively as 2.5 W/m2.
Using the same average pressure, ε = 0.06 (lowered since emissivity typically
decreases with decreasing temperature) and keff = 1.5 mW/cm K, the heat flux to
the 2 K surfaces was calculated to be 94 mW/m2. While these values are conser-
vative, proper assembly and installation of the blankets is required for ideal per-
formance. Taped joints, insufficient overlap and compaction of blanket layers can
reduce the effectiveness of the MLI thereby increasing significantly the incident
heat load. Combining the realities of assembly with the fact that nine of twenty-two
measured CEBAF CMs, as reported in [18], exceeded the 50 K static heat load
design criteria, it was decided to increase the estimated radiation heat load to 200 %
of the calculated values.

6.3.4.5 End Cans, Heat Exchanger and Cryogenic Piping

The end cans route helium to and from the CM, provide controls for the primary
and secondary circuits, provide beamline vacuum connections and contain pressure
relief valves used in the event of loss of vacuum (Figs. 6.11 and 6.12). The end
cans use bayonet connections that are identical to those used in the CEBAF CM
[19]. In addition to the primary and shield supply bayonet connections, two J-T
valves similar to those used in CEBAF are installed in the supply end can: the
primary J-T has a Cv = 0.3 and the secondary J-T has a Cv = 0.05. The return end
can contains a sub-atmospheric primary return bayonet, the shield return bayonet, a
Circle Seal™ pressure relief valve plumbed in parallel with a parallel plate relief
valve, a cool-down valve with a Cv = 3.0 and a helium-helium heat exchanger
(HX).
The counter flow HX (Fig. 6.15) is a plate-fin type core constructed of aluminum
with a pressure rating of 12 atm. Stainless-to-aluminum joints transition from the
aluminum body to the stainless steel process piping. The design parameters are
given in Table 6.5. Silicon diodes are installed in the assembly to measure HX
terminal temperatures in order to assess the HX effectiveness.
The stainless piping within the CM is strain-relieved from the end can con-
nections with flexible metal hose. All of the shield circuits, the fill-lines between
6 Segmented SRF Cryomodules 171

Fig. 6.15 Helium-helium heat exchanger

Table 6.5 Heat exchanger design parameters

High pressure side Low pressure side
Core length 600 mm –
Core cross section 100 mm × 100 mm –
Maximum allowable 1 × 10−9 mbar l/s external –
helium leak rate 1 × 10−4 mbar l/s cross-passage
Pipe size 26.7 mm OD × 2.1 mm wall 60.3 mm OD × 2.8 mm wall
(¾″ IPS Schedule 10) (2″ IPS Schedule 10)
Maximum flow rate 6.0 g/s 6.0 g/s
Fluid inlet pressure 2.95 atm 0.040 atm
Maximum pressure drop 0.020 atm 0.001 atm
Inlet temperature 5.0 K 2.1 K
Outlet temperature 2.2 K 3.96 K
Capacity 60 W –
NTU (integrated) 4.37 –
Effectiveness 0.97 –
UA (integrated) 93.2 W/K and 20 % –
and UA margin

helium vessels and the primary circuit plumbing through the HX high-pressure side
are 26.7 mm OD × 2.1 mm wall. The primary return piping through the HX and in
the relief stack is 60.3 mm OD × 2.7 mm wall, sized for a catastrophic loss of
beamline vacuum. The helium vessel return headers are 88.9 mm OD × 3.0 mm
wall, and along with the helium vessel provide 10 % ullage. The secondary circuit
consists of 6.35 mm outer diameter × 1.3 mm wall stainless steel tubing, sized to
reduce helium inventory in that circuit, and a 1 L surge tank which functions to
damp potential flow oscillations that may arise during operation.
172 E. Daly et al.

6.3.4.6 Fundamental Power Coupler

Due to operating frequency and space constraints, a coaxial coupler with inner and
outer conductors was selected instead of a waveguide coupler to provide funda-
mental RF power to the cavities. Besides delivering RF power, the coupler must not
adversely affect the electromagnetic performance of the cavity or the thermal per-
formance of the cryomodule. Cavity cleanliness procedures require the coupler to
be inserted into the cavity at the six o’clock position.
To handle static and dynamic heat loads in the coupler and cavity, the outer
conductor is cooled by a nominal stream of 3 atm, 5 K supercritical helium flowing
at 0.038 g/s per coupler and the inner conductor is conduction cooled by a 1 L per
minute water flow. The ceramic RF window is maintained at 300 K during oper-
ation by a heater-thermocouple control loop.
The outer conductor is a machined and welded stainless steel assembly with
flanged connections to the cavity and the warm window. The internal vacuum
surface is copper-plated with a nominal thickness of 15 μm and an estimated
RRR = 10 to reduce resistive wall losses induced by the RF surface currents.
A single helical, square-grooved flow passage, 2.3 mm wide × 1.6 mm deep
(Fig. 6.14), with a pitch of 2.6 turns/cm is machined into a thick-walled stainless
steel tube, 75.75 mm internal diameter. A thin-walled stainless steel tube (1.6 mm
O.D.) is then shrunk-fit over the outer diameter of the flow passages and welded
leak tight onto the ends nearest the flanges. Due to the small hydraulic diameter of
2.16 mm, flow velocities and Reynolds number are kept high to mitigate potential
deleterious effects of buoyancy, re-circulation and poor heat transfer in the helium
stream.
Conduction heat load estimates (Table 6.3) were calculated for the outer con-
ductor using an Excel model that included combined conduction and convection
heat transfer as well as temperature-dependent thermal properties for stainless,
copper and supercritical helium gas (Fig. 6.16). A separate estimate of the static
radiation heat load emitted from the FPC to the beam pipe was calculated.

Power Coupler Outer Conductor Thermal Profile

Parameters:
Mass Flow = 0.0375 g/s
300 Power = 550 kW;
Duty = 8.7%
RRR = 10
250
Temperature (K)

Results:
200 QHV = 0.18 W
Q300K = 36.8 W
Qloss = 2.92 W
150 Qexhaust = 39.5 W
He Gas Exit Temp = 210 K

100

50 Surface Temp (K)

Gas Temp (K)
0
0 5 10 15 20
Distance from Cold End (cm)

Fig. 6.16 He gas cooling passages (right) and FPC outer conductor assembly (left). The helium
gas flow rate of 0.038 g/s produces an exhaust temperature of 210 K (center)
6 Segmented SRF Cryomodules 173

The values were used as input to a finite element model of the cavity ends to verify
that they remain below the niobium critical temperature.
In an effort to determine if thermo-acoustic oscillations or flow instabilities exist
in the secondary circuit, a test rig was been designed and built.

6.3.5 Thermal Performance of the SNS Cryomodule

6.3.5.1 Budgeted Versus Estimated Heat Loads

For the primary circuit maintained at 2.1 K, the heat loads are divided into three
categories: (1) a static contribution consisting of solid conduction, residual gas
conduction and radiation, (2) a dynamic contribution consisting of RF heating in the
cavity and coupler and (3) a distribution system allotment that consists of u-tube
and transfer line heat loads. There is no dynamic contribution to the 50-K shield and
its cooling circuit. Details of the estimated heat loads are given in Table 6.3.
Budgeted heat loads in contrast, are those that each component are allowed to have
and are used in the sizing of the cryogenic refrigeration system. A comparison of
the budgeted and estimated heat loads is provided in Table 6.6. For the medium
beta cryomodule, the total estimated heat load to the primary circuit is 72 % of the
budget per cryomodule, and for the shield circuit is 91 % of the budget per cry-
omodule. The primary static heat load for the medium beta cryomodule is estimated
as 9.7 W and includes contributions from bayonets, valves, supports, wiring,
cabling and radiation. The dynamic heat load is estimated as 8.3 W and is divided
evenly between the three cavities and FPCs housed within the cryostat.

6.3.5.2 Measured 2 K Static and RF Heat Loads

The following measurement technique, similar to that employed during

CEBAF CM acceptance testing [5], is used to quantify the static heat load and RF
heat load. With the primary J-T supply valve and u-tube return valve closed, the

Table 6.6 Comparison of heat load budget versus estimate for both SNS cryomodules
β = 0.61 cryomodule (3 cavities) β = 0.81 cryomodule (4 cavities)
2.1 K 50 K 2.1 K 50 K
Description Budget Estimate Budget Estimate Budget Estimate Budget Estimate
Static 15 9.7 146 131 18 11.5 170 161
Dynamic 16 8.3 N/A N/A 28 17.1 N/A N/A
U-Tubes and 10 <10 24 <24 10 <10 30 <30
distribution
Total 41 <28 170 <155 66 38.6 200 <191
174 E. Daly et al.

helium bath pressure rate-of-rise is measured during three conditions: with the
helium bath heater on (HEATER), with only RF power on (RF) and with no heater
or RF power (STATIC). Five separate 30-second measurements are made in the
following order—STATIC, HEATER, STATIC, RF and then STATIC. Using a
known heater power (Qheater) and the average rate-of-rise (dP/dt) for each condition,
the static and RF heat loads are then calculated from the following equations:


Qstatic ¼ Qheater  ðdP=dtjstatic Þ= dP=dtjheater þ static dP=dtjstatic ð6:1Þ



QRF ¼ Qheater  dP=dtjRF þ static  dP=dtjstatic = dP=dtjheater þ static dP=dtjstatic : ð6:2Þ

Note that during this measurement, the flow to the secondary circuit for the FPC
outer conductors ideally should be maintained constant. Actually the inlet tem-
perature to the CM increases during the two-and-one-half-minute data acquisition
cycle. Typically several measurements are taken before opening valves to recover
the pressure. Over that time span the inlet temperature can increase by as much as
5–7 K. There is a corresponding rise in the FPC cold flange temperature. This
transient temperature more closely reflects the fluid temperature rather than the cold
flange temperature since measurements have been made at relatively high gradients
without indication of quenches. The helium bath pressure initially is approximately
3 kPa (0.030 atm) and may rise to as high as 4.1 kPa (0.040 atm) during the series
of measurements. The actual total pressure rise is primarily dependent on the cavity
gradients selected for each measurement.
The measurements are accurate to ±½ W due in part to the accuracy of the
capacitance manometer pressure transducer as well as the measured voltage and
current on the cavity bath heater. Effects from valve leakage and reduced flow in the
supply bayonet have been neglected, which result in conservative heat load esti-
mates. Overall, the static heat load measurements agree well with the estimates; the
average static load is approximately 10 W (Fig. 6.17).

25.0
Static 2K Heat Load (W)

20.0
Measurement - 2K Static Heat
Load (W)
15.0
Budget - 2K Static Heat Load
(W)
10.0
Estimate - 2K Static Heat Load
(W)
5.0

0.0
M0 M1 M2 M3
CM Serial Number

Fig. 6.17 Measured primary static heat load versus CM serial number
6 Segmented SRF Cryomodules 175

4.0
3.5
3.0
RF Heat Load (W)

2.5 Measured RF Heat Load @

10.1 MV/m (pulsed)
2.0
Budget - RF Heat Load Per
1.5 Cavity
1.0
0.5
0.0

CM-Cavity Position

Fig. 6.18 Measured RF heat load versus CM-cavity position at 10.1 MV/m

The measured RF heat load at the required gradient of 10.1 MV/m is approxi-
mately equivalent to the estimate for the prototype cavities, 2.7 W, and is less than
1 W for each of the cavities in the three production CMs (Fig. 6.18). One cavity,
M3-3, was not measured due to schedule constraints.
The power dissipation scales quadratically with accelerating gradient and is
inversely proportional to a geometric factor. For the β = 0.81 cavities, the required
accelerating gradient is 15.5 MV/m and the geometric factor is reduced by 12 %.
Assuming no degradation due to field emission, the power dissipation will be
approximately two times greater for the higher beta cavities. Initial vertical tests on
these cavities have yielded acceptable results.

6.3.5.3 Measured 50 K Shield Heat Load

The measurements of the 50-K shield heat load are in good agreement with pre-
dictions (Fig. 6.19). The measurement procedure included the following steps: first
for a given valve position, monitor the gaseous helium inlet and outlet temperatures
and the mass flow rate, then calculate the heat load using gaseous helium properties
and time-averaged temperatures and flow rates. This procedure was repeated at
several different flow rates. The accuracy of this type of measurement, approxi-
mately ±10 %, is driven by the helium mass flow meter.
The heat load measured in the prototype CM (M0) is 129 % higher compared
with measurements on the production CMs. This is most likely caused by increased
residual gas conduction in the insulating vacuum space due to the large helium leak
rate that had been detected in the prototype shield process piping after cooldown.
The prototype required active turbo-pumping during the entire test period to
maintain the vacuum in the range of 10−6 torr. In addition, changes in the
multilayer insulation blanket design and improved blanket installation techniques
contributed to the lower overall heat load in the production CMs.
176 E. Daly et al.

50K Shield Heat Load (W) 160.0

150.0
Shield Heat Load (W)
140.0

Budget - 50K Static Heat

130.0 Load (W)

120.0 Estimate - 50K Static Heat

Load (W)

110.0

100.0
M0 M1 M2 M3
CM Serial Number

Fig. 6.19 Measured shield heat load versus CM serial number

6.3.5.4 Fundamental Power Coupler Thermal Performance

The primary requirement for the FPC is to transmit a maximum of 550 kW of

RF power at a 7 % duty cycle. Adding a 10 % engineering margin, the design goal
for the average power handling capability became 53 kW. Another important
requirement is to insert the coupler from the bottom to minimize particulate
introduction and generation in the cavities. The outer conductor has counterflowing
helium-gas cooling to intercept the static and RF heat loads. The maximum flow
rate is 0.075 g per second per coupler. The flow passages are sized to keep the flow
velocity high enough to overcome buoyancy forces and potential flow instabilities.
To verify the design and confirm the thermal calculations, the FPC installed on
the M0-2 cavity was instrumented to measure the temperature profile along the
outer conductor. In addition, the exhaust helium mass flow rate was measured
during this particular test to determine the maximum power handling capability of
the FPC. With a stub tuner installed, the FPC transmitted 9 kW of CW power to the
cavity. The resonant condition created by the stub tuner settings resulted in a
transmitted power equivalent to 100 kW within the FPC, nearly double the design
goal.
The thermal model used the cold and warm end temperatures, the helium mass
flow rate and the equivalent transmitted RF power during the test as inputs. The
results from the thermal model agree well with the experimental temperatures
(Fig. 6.20). The calculated conductive heat load transmitted to the cavity from the
outer conductor during the higher power operation was estimated as 1.1 W—the
expected conductive heat load at nominal operating conditions is less than 0.25 W.

6.3.5.5 Helium Heat Exchanger Performance

The HX is located in the CM return end can to increase overall cryogenic system
efficiency. This enables the SNS supply transfer lines to contain 5-K supercritical
6 Segmented SRF Cryomodules 177

300
Calculated Cu-Plated
250 Inner Surface Temp
(K)
Temperature (K)

200

150 Calculated Helium Gas

Temp (K)
100

50 Sleeve Exterior
Measured Temp. (K)
0
0 5 10 15 20
Distance from Cold End (cm)

Fig. 6.20 Comparison of analytical and measured temperatures on the M0-2 FPC outer conductor

helium as compared with *3 K for the CEBAF machine. The HX, a brazed-
aluminum plate-fin construction, pre-cools the helium upstream of primary JT valve
to increase liquid yield. The effectiveness is specified as greater than 90 % for the
design capacity of 60 W. The overall thermal conductance (UA) and number of
transfer units (NTU) are 93.2 W/K and 4.37, respectively. The assembly is required
to be leak tight to 1 × 10−9 mbar l/s.
Initial measurements indicate that the performance is consistent with estimates.
The measured inlet and outlet temperatures on the low-pressure-side of the HX are
2.0 and 3.5 K respectively, compared with estimates of 2.1 and 3.96 K. Inlet and
outlet temperatures for both streams are required to calculate the effectiveness of the
counter-flow HX. Since the high-pressure-side inlet temperature indication is not
reliable due to the placement of the silicon-diode temperature sensor, the effec-
tiveness has not been calculated.

6.3.5.6 Conclusions on the Thermal Performance

Measurements of thermal performance have been completed on the prototype

and three production CMs. Results obtained for the static, RF and shield heat loads
are in good agreement with estimates and are within budgeted heat loads.
Measurements of the outer conductor thermal profile are in good agreement with
calculations and indicate that the heat load from the FPC is acceptable. The FPC
and its outer conductor can transmit approximately two times more than the
required maximum RF power expected from the SNS klystrons. Initial measure-
ments of the available heat exchanger terminal temperatures agree with predictions.
The heat exchanger performance assessment is incomplete; more study is required
to quantify the heat exchanger effectiveness.
178 E. Daly et al.

In addition, better CM and cavity construction techniques and procedures have

incrementally improved the thermal performance.

6.4 The CEBAF C100 Energy Upgrade Cryomodule

6.4.1 Introduction

In 2008 construction began on the energy upgrade of the CEBAF accelerator to

12 GeV [20]. To achieve the 12 GeV energy five new cryomodules were added to
each of the two existing linacs. Each cryomodule provides 100 MV of accelerating
voltage with eight 1500 MHz SRF cavities. Each cavity contains seven 1500 MHz
cells, two hook style HOM couplers, and a FPC [21]. Each cavity has 0.7 m of
active length and is required to operate at 19.2 MV/m to provide the 100 MV of
required voltage and have sufficient voltage overhead for robust operations. A new
low loss cell geometry [22] is used to reduce the 2 K heat load. The cavities,
cryomodule, and helium refrigerator are designed for a total of 300 W 2 K heat
load for each cryomodule. The cavity and cryomodule alignment requirements
remain unchanged from the original CEBAF specification [23]. The C100 cry-
omodules (see Fig. 6.21) are installed in existing empty linac zones and are
required to maintain the same cryogenic circuits and interfaces as the C20 cry-
omodules (see Fig. 6.22).

6.4.2 Lessons Learned from C20 Experience

The C100 cryomodule design benefited from the production and decade of opera-
tions of the C20 cryomodules as well as advances in the SRF technology and

Fig. 6.21 C100 cryomodule assembly

6 Segmented SRF Cryomodules 179

Fig. 6.22 12 GeV CEBAF with ten C100 cryomodules

Fig. 6.23 C100 cavity string

facilities. The largest change in the C100 cryomodule was the move away from a
cavity pair assembly used as the basis for the C20 cryomodule to an eight cavity
string (see Fig. 6.23) assembled in the clean room. The cavity string assembly is too
large to be tested prior to assembly into a cryomodule. This requires each cavity to be
vented after being tested individually prior to being assembled into the cavity string.
This use of a cavity string assembly is based on a high level of confidence that the
cavity performance will be maintained after the string assembly process. This
confidence comes from the cavity design that eliminates all indium seals, improved
cleaning techniques including high pressure water rinsing, improved cleanrooms,
assembly tooling, and procedures as well as demonstrated performance.
Another significant change to the design was moving all the cavity frequency
tuner friction bearing components outside the helium and vacuum space. This
eliminates motors and gears operating at cryogenic temperatures and in a vacuum.
Important features retained from the C20 cryomodule include using all welded
180 E. Daly et al.

Fig. 6.24 C100 FPC assembly with the cavity and helium circuits

stainless steel cryogenic circuits including the helium vessel, a double RF window
FPC, and individual heaters for each cavity that are inside the helium vessels. The
stainless steel cryogenic circuit construction provides a high reliability fabrication
avoiding leaks from flanged connections at a minimum cost. Some instrumentation
flanges are required to route temperature, level, and heater signals out of the
cryogenic circuits. These are kept to a minimum and are done with smaller knife
edge seal flanges that undergo rigorous QC checks including cold shocking and
leak checking prior to use. The double RF window on the FPC provides redundancy
against vacuum leaks that can be a problem on high power RF transmission lines.
The space between the two windows creates an intermediate vacuum space that can
be monitored for off normal conditions and shutdown RF to avoid catastrophic
failures (see Fig. 6.24).
The individual heaters are important for balancing the dissipated power on a
cavity by cavity basis on a high power cryomodule and provide for high resolution
cavity performance measurements [24] when used in a segmented cryomodule with
isolation valves like the CEBAF cryomodules.

6.4.3 Cavity

The SRF cavity is a 1500 MHz seven cell cavity with a close fitting, 10 in. OD,
stainless steel helium vessel (see Fig. 6.25). The transition from the Niobium cavity
to the stainless steel is done with a brazed assembly incorporated into the beamline
just outside of the cavities [25]. The FPC, HOM couplers, and RF pickup probe are
located outside of the helium vessel. The cavity assembly includes flanged inter-
faces for each of these and the two beamline connections. Unlike the original
CEBAF cavity where the FPC and HOM and RF pickup probes are located inside
the helium vessel, on the C100 cavity these elements are located in the insulating
6 Segmented SRF Cryomodules 181

Fig. 6.25 C100 cavity cut

assembly with helium vessel

vacuum space. This requires careful attention to the thermal management [26] as all
heat must be conductively removed to the helium inside the helium vessel. The
resulting temperature gradient can increase the beamline temperature enough to
make them normal conducting. Most often this results in a thermal quench of the
cavity. Early on this was a limit in the C100 cavities. The niobium HOM pickup
probes would heat from the fundamental mode RF currents and above a certain
operating gradient would transition to the normal conducting state and quench the
cavity. JLab developed very high thermal conducting RF feedthroughs [27] to shunt
the heat from the HOM pickup probe away from the cavity.

6.4.4 Cavity Frequency Tuner

The cavity frequency tuner design (see Fig. 6.26) was a radical change from the
existing CEBAF tuner [28]. Based on lessons learned from the C20 experience all
friction bearing components were placed outside the insulating vacuum space and
can be replaced with the cryomodule cold. The tuner actuator has a stepper motor
coupled to a harmonic drive reducer for coarse tuning and a piezo actuator for fine
tuning. The tuner has a minimum range of 400 kHz and a resolution of 1 Hz. The
tuning motion of the drive system enters the cryostat through two thin-wall con-
centric tubes. The tubes are connected to the upper and lower arms of a scissors
type jack. The axial motion of the tubes translates into a linear stroke parallel to the
cavity center line. To eliminate friction, motion within the vacuum is done with
flexure plates.

6.4.5 Cold Mass and Space frame

The C100 cryomodule uses a cold mass subassembly made up of the cavity string
(Fig. 6.27), cold magnetic shielding, cavity frequency tuners, cryogenic piping, and
associated instrumentation [29]. The cold mass assembly length is constrained to fit
into the existing slot length of the original cryomodule while increasing the active
182 E. Daly et al.

Fig. 6.26 C100 tuner

Fig. 6.27 Cavity string

anchoring at the center of the
space frame using nitronic
rods

cavity length by 40 %. The increased cavity length is accommodated by the

elimination of all bellows between the cavities. The cavity string is anchored in the
center of the cryomodule and all cooldown and cavity tuning effects must be
managed over the entire length of the four cavity half string extending out from the
center to the end of the cryomodule.
There are vacuum boundaries that extend from the cold mass to the warm
vacuum vessel at each cavity for the FPC and one at each end of the cavity string
(see Fig. 6.28). The relative movement at the cavity string ends is provided for by
the use of two multi-convulsion stainless steel bellows in series integrated into the
warm to cold beamline transition. There is a shield intercept point between the two
bellows minimizing the heat load to the shield and primary cryogenic circuit.
6 Segmented SRF Cryomodules 183

Fig. 6.28 Cryomodule cut

away showing the cold mass
and space frame

The FPC is rectangular section mounted normal to the direction of movement which
has two 3 convulsion bellows along its length separated as much as possible
resulting in a dogleg motion producing the required relative movement between the
cavity and vacuum vessel.
The cold mass is supported by the space frame (see Fig. 6.27). The space frame
provides advantages over supporting from the vacuum vessel [30]. The first is that it
allows all the assembly to take place outside the vacuum vessel (where things are
more readily accessible). The space frame with all the cryounits attached, is then
rolled into the vacuum vessel and locked down. The second advantage is the
requirement for fewer penetrations in the vacuum vessel, thus reducing cost and
possibility of vacuum leaks.
The space frame also supplies points for aligning the beam line while inside the
vacuum vessel. The space frame consists of a series of five welded tube and ring
sections of 304 stainless steel bolted together. The tubes are 2.5 in. outer diameter
with a 0.25 in. wall thickness. The rings are 29.75 in. outer diameter, 23.75 inner
diameter by 0.75 in. thick. The tubes are fitted to the rings in a socket joint and welded.
The cavities are suspended from the space frame support tubes using nitronic
rods. There are eight 0.198 in. diameter rods connected to each unit. The rods have
a silicon bronze swaged, threaded fitting attached to each end. The rods are
extended to the opposite side of the He vessel in order to increase the length and
keep the static heat load to a minimum.
All the mass of the space frame and attachments is supported by 4 stainless steel
wheels at the “quarter” points of the frame. Any axial loading (2 g transport loads)
is supported by two brackets at one of the quarter points.

6.4.6 Vacuum Vessel

The vacuum vessel defines the air to insulating vacuum interface. The space frame
with the cold mass is inserted axially into the vessel. End caps and tophats are
184 E. Daly et al.

Fig. 6.29 Vacuum vessel

installed to complete the cryomodule. The vacuum Vessel consists of 0.25 in. thick
304 stainless steel rolled to a 32 in. outer diameter cylinder. Five sections
approximately 61 in. long are welded together using a support ring at each joint and
the ends. The support ring is 1 in. thick with a 35 in. outer diameter. There are six
15.5 in. openings and one 17.5 in. opening for waveguide attachment. Also, there
are eight 4.5 in. openings for the tuners. Various other small openings for alignment
and lockdown access are also included (Fig. 6.29). The vacuum vessel is supported
by two saddle stands, one at each “quarter point”. The space frame (wheels) is
supported by the vacuum vessel at the same points, thus only adding compression
stresses into local areas of the wall under normal loading. The space frame is also
locked down axially to the vessel via a plate protruding from the space frame,
which is trapped in a bracket and welded to the vacuum vessel. There is a spring
relief plate located at both ends to limit the internal pressure to less than 2 atm.

6.4.7 End Cans

The End cans for the C100 cryomodule are adapted from the original CEBAF C20
end cans. The primary and shield circuits are connected to the distribution system
using removable u-tubes with bayonet connections. Each end can bayonet has an
isolation valve allowing u-tubes to be removed and installed while the cryomodule
and distribution system are cold. Every volume that can be isolated with valves
includes a relief valve to avoid over pressurization in case cold gas is allowed to
warm in the potential trapped volume. The supply and return end can piping
consists of the primary circuit, which provides cooling fluid to the superconducting
cavities, and the shield circuit that cools the nominal 50 K thermal shield. The
gas/fluid circuits consist of pipes, tubes, fittings, and valves. The design pressure for
the primary circuit is 5.0 and 20.0 atm for the shield circuit [31]. The end plate has
a pass through for the beampipe with an O-ring sealing surface as well as a shield
thermal strap to minimize the conduction heat load to the 2 K circuit.
6 Segmented SRF Cryomodules 185

6.4.8 C100 Performance

Ten C100 cryomodules have been installed and operated in the CEBAF accelerator.
Early commissioning of the first cryomodules included high current operations at
108 MV accelerating voltage meeting the design goals. Commissioning results
show that these cryomodules will deliver an average energy gain of 110 meV
which exceeds the design goal of 108 MV. The C100 cavities are able to operate at
an average maximum operating gradient of 19.6 MV/m [32]. Early operations
identified challenges to stable operations due to microphonics. As a result the tuner
end plates were made stiffer significantly reducing the transmission of vibrations to
the cavities.

6.5 SSR1 Cryomodule Design for PXIE

6.5.1 Introduction

Fermilab is in the process of designing a Project X Injector Experiment (PXIE), a

CW linac, to validate the Project X concept, reduce technical risks, and obtain
experience in the design and operation of a superconducting proton linac. The
overall facility will include an ion source, low and medium-energy beam transport
sections, an RF quadrupole, and two superconducting cavity cryomodules. One will
contain eight half-wave resonators operating at 162.5 MHz and eight supercon-
ducting solenoids. The second will contain eight single spoke resonators (SSR1)
operating at 325 MHz and four superconducting solenoids. The design of the
cryomodule being developed to house the 325 MHz single spoke resonators and all
related systems and services is described here.
The SSR1 cryomodule will operate with continuous wave (CW) RF power and
support peak currents of 5 mA chopped with arbitrary patterns to yield an average
beam current of 1 mA. The RF coupler design employed should support a future
upgrade path with average currents as high as 5 mA. The RF power per cavity at
1 mA average current and 2.2 MV accelerating voltage (β = 0.22) should not
exceed 4 KW with an overhead reserved for microphonics control. The RMS
normalized bunch emittance at the CM exit should not exceed 0.25 mm mrad for
each of 3 planes.
The current beam optics design for Project X requires that the SSR1 cryomodule
contains eight cavities (C) and four solenoids (S) in the following order: C–S–C–C–
S–C–C–S–C–C–S–C. Horizontal and vertical dipole corrector are located inside
each solenoid. A four-electrode beam position monitor is located at each solenoid.
The intent is that this cryomodule will have all external connections to the
cryogenic, RF, and instrumentation systems made at removable junctions at the
cryomodule itself. The only connection to the beamline is the beam pipe itself
which will be terminated by “particle free” valves at both ends. Minimizing mean
186 E. Daly et al.

Fig. 6.30 PXIE layout

time between failure and repair and in situ repair of some internal systems are
important design considerations in the cryomodule design. Figure 6.30 shows the
linac layout including the location of the SSR1 cryomodule. Table 6.7 lists perti-
nent operational requirements and parameters of the cryomodule assembly.

6.5.2 Cryomodule Design

Eventually, Project X will require several different cryomodule designs for cavities
operating at 162.5, 325, 650, and 1300 MHz. The SSR1 for PXIE is the first of
these being developed at Fermilab. Some details of individual cryomodule com-
ponents are described in the following sections.

6.5.2.1 Cryogenic Systems and Vacuum Interfaces

Fine segmentation is the configuration choice for Project X and PXIE cryomodules.
Each individual vacuum vessel will be closed at both ends and the cryogenic
circuits will be fed through bayonet connections at each cryomodule. Each cry-
omodule will have its own connection to the insulating vacuum pumping system.
Also, each cryomodule will have its own 2 K heat exchanger and pressure relief
line exiting near the middle of the module. This configuration provides flexibility in
terms of cryomodule replacement, and cooldown and warm-up times at the expense
of requiring more individual cryogenic connections, cold-to-warm transitions at
each end of each cryomodule, and extra space at each interconnect to close the
beam tube.

6.5.2.2 Vacuum Vessel

The vacuum vessel serves to house all the cryomodule components in their
as-installed positions, to provide a secure anchor to the tunnel floor, to insulate all
cryogenic components in order to minimize heat load to 80, 4.5, and 2 K, as well as
maintain the insulating vacuum. It is 1.219 m (48 in.) in diameter and manufac-
tured from carbon steel and is shown as the outermost shell in Fig. 6.37.
6 Segmented SRF Cryomodules 187

Table 6.7 SSR1 functional requirements specifications (subject to change)

General
Physical beam aperture, mm 30
Overall length (flange-to-flange), m ≤5.4
Overall width, m ≤1.6
Beamline height from the floor, m 1.3
Cryomodule height (from floor), m ≤2.00
Ceiling height in the tunnel, m 3.20
Maximum allowed heat load to 70 K, W 250
Maximum allowed heat load to 5 K, W 80
Maximum allowed heat load to 2 K, W 50
Maximum number of lifetime thermal cycles 50
Intermediate thermal shield temperature, K 45–80
Thermal intercept temperatures, K 5 and 45–80
Cryo-system pressure stability at 2 K (RMS), mbar *0.1
Environmental contribution to internal field 15 mG
Transverse cavity alignment error, mm RMS <1
Angular cavity alignment error, mrad RMS ≤10
Transverse solenoid alignment error, mm RMS <0.5
Angular solenoid alignment error, mrad RMS <1
Cavities
Number, total 8
Frequency, MHz 325
β geometric 0.22
Operating temperature, K 2
Operating mode CW
Operating energy gain at β = 0.22, MV/cavity 2
Coupler type—standard coaxial with impedance, Ω 105
Coupler power rating, KW >20
Solenoids
Number, total 4
Operating temperature, K 2
Current at maximum strength, A ≤100
∫B2dL, T2m 4.0
Each solenoid has independent powering
Correctors
Number, total 8
Number, per solenoid package 2
Current, A ≤50
Strength, T-m 0.0025
Beam position monitors
Number, total 4
Number of plates 4
Electrical center accuracy compared to geometric center, mm ≤±0.5
188 E. Daly et al.

6.5.2.3 Magnetic Shield

Just inside the vacuum vessel, virtually in contact with the inner wall, is a magnetic
shield to shield the cavities from the earth’s magnetic field. Preliminary tests show
that a 1.5 mm-thick mu-metal shield at room temperature reduces the residual field
inside the cryostat to less than 10 μT. It is likely that separate magnetic shields will
be installed around individual magnetic elements to further reduce the potential for
trapped fields in the superconducting cavity structures.

6.5.2.4 Thermal Shield and Multilayer Insulation

Each cryomodule will have a single thermal shield cooled with helium gas, nom-
inally at 45–80 K. It is currently envisioned to be made from 6000-series aluminum
with cooling channels on both sides. Two 15-layer blankets of multilayer insulation,
between the vacuum vessel and thermal shield will reduce the radiation heat load
from the room temperature vacuum vessel to approximately 1.5 W/m2. A 5 K
circuit will be available to intercept heat on the input couplers and current leads, but
there is no plan to install a full 5 K thermal shield.

6.5.2.5 Support System

All of the cavities and solenoids will be mounted on individual support posts which
are in turn mounted to a full-length strongback located between the vacuum vessel
and thermal shield. This enables the entire cavity string to be assembled and aligned
as a unit then inserted into the vacuum vessel during final assembly. The strongback
is aluminum to provide a uniform temperature base. Maintaining the strongback at
room temperature helps minimize axial movement of the cold elements during
cooldown, reducing displacement of couplers, current leads, and many of the
internal piping components.
The support posts are similar to supports utilized in SSC collider dipole magnets
(Chap. 2) and ILC and XFEL 1.3 GHz cavity cryomodules (Chap. 5). The main
structural element is a glass and epoxy composite tube. The tube ends and any
intermediate thermal intercepts are all assembled using conventional shrink-fit
assembly techniques in which the composite tube is sandwiched between an outer
metal ring and inner metal disk. The strongback and support posts are shown in
Fig. 6.31. All of the cavities and magnetic elements are mounted to the support
posts using adjustable positioning mechanisms [33].

6.5.2.6 Cavity and Tuner

The cryomodule contains eight single spoke, β = 0.22, 325 MHz cavities operating
in CW mode at 2 K in stainless steel helium vessels. Each has an integral coarse
6 Segmented SRF Cryomodules 189

Fig. 6.31 Strongback with supports

Fig. 6.32 Spoke cavity,

helium vessel, and tuner

and fine tuner that operates through a lever system and pushes on the cavity end
wall. For ease of maintenance, tuner access covers are incorporated into the helium
vessel design. The cavity and tuner system is shown in Fig. 6.32 and described
more completely in [34].

6.5.2.7 Input Coupler

The input coupler is a 105-ohm coaxial design that supplies approximately 2 kW

CW to each cavity in PXIE and ultimately up to 18 kW CW in Project X. The
coupler contains a single warm ceramic window that provides separation of the
warm and cold coupler sections. During cryomodule fabrication, the cold section
can be installed on the cavity in the cleanroom prior to assembly of the string. The
warm section can then be installed from outside the vacuum vessel during final
assembly. The inner conductor is solid copper with copper bellows to accommodate
motion due to misalignment and thermal contraction. The cold end of the outer
conductor is 316L-stainless steel. The warm end is copper with copper bellows.
190 E. Daly et al.

Fig. 6.33 Input coupler

Heat load estimates don’t suggest a significant penalty for not copper plating the
outer conductor. A forced-air cooling tube is inserted into the inner conductor after
assembly that supplies air to cool the coupler tip. Figure 6.33 shows the current
coupler design. The input coupler design is described more thoroughly in [35].

6.5.2.8 Current Leads

Each focusing element package contains up to three magnet coils, the main sole-
noid, operating nominally at 100 A and two steering correctors each operating
nominally at 50 A. A conduction cooled current lead design modeled after similar
leads installed in the LHC at CERN is being developed for use in the SSR1
cryomodule [36]. Figure 6.34 illustrates the design for the lead assembly. Thermal
intercepts at 45–80 K and at 5 K help reduce the heat load to 2 K, nonetheless,
these current leads represent a significant source of heat at the low temperature end.
There will be one lead assembly for each magnetic element.

6.5.2.9 Solenoid and Beam Position Monitor

The four magnet packages in the cryomodule each contain a focusing solenoid and
two dipole correctors all operating in a helium bath at 2 K. The Project X lattice,
especially the low-beta section, provides little room along the beamline for beam
diagnostics either inside individual cryomodules or between adjacent modules. In
order to conserve axial space along the beamline a button-type beam position
monitor (BPM) was chosen for installation in the SSR cryomodules. A total of four
will be installed in the cryomodule and tested in PXIE, one at each magnetic
element. These devices are compact and lend themselves well to incorporation right
into the solenoid magnet package as shown below in Fig. 6.35. The bellows in
either end of the beam tube allow independent adjustment of each magnet.
6 Segmented SRF Cryomodules 191

Fig. 6.34 Conduction cooled current lead assembly

Fig. 6.35 Solenoid and BPM assembly

6.5.3 Final Assembly

The final assembly of the SSR1 cryomodule for PXIE is shown in Figs. 6.36 and
6.37. Figure 6.36 shows the cavity string consisting of the cavities, solenoids, beam
position monitors, and internal piping mounted on support posts which are in turn
mounted to the strongback. Figure 6.37 shows the entire cryomodule assembly.
192 E. Daly et al.

Fig. 6.36 Cavity string assembly

Fig. 6.37 Cryomodule assembly

6.5.4 Status and Plans

Fermilab has the PXIE SSR1 vacuum vessel, cavities, strongback, support posts,
and many other cryomodule components in-house. Work is in the process to build
up production facilities, develop assembly procedures for the cavity string, and to
tooling for insertion of the cavity string into the cryomodule. Installation of the
cryomodule in the PXIE facility is expected in 2017 [37].
6 Segmented SRF Cryomodules 193

References

1. G. Oliver, J.P. Thermeau, P. Bosland, G. Devanz, F. Lesigneur, C. Darve, ESS Cryomodule

for Elliptical Cavities, Proceedings of the 16th International Conference on RF
Superconductivity (2013)
2. M. Johnson et al., Design of the FRIB Cryomodule, IPAC 2012, New Orleans (2012)
3. H.A. Grunder et al., in The Continuous Electron Beam Accelerator Facility, ed. by
E.R. Lindsrom, L.S. Taylor. Proceedings of the 1987 IEEE Particle Accelerator Conference,
vol. 1 (Washington, D.C., 1987), pp. 13–18
4. C. Reece et al., in Production Vertical Cavity Pair Testing at CEBAF, Proceedings of the
Particle Accelerator Conference (1993), pp. 650–658
5. C.H. Rode et al., 2.0 K CEBAF CRYOGENICS, CEBAF PR-89-029
6. D. Kashy et al., CEBAF transfer line system, CEBAF PR-91-024 June 1991
7. M. Wiseman et al., CEBAF cryounit loss of vacuum experiment. Appl. Cryog. Technol. 10,
287–303
8. W. Schneider et al., Thermal performance of the CEBAF superconducting linac cryomodule.
Adv. Cryog. Eng. 39, 589–596
9. J. Benesch et al., CEBAF’s SRF cavity manufacturing experience. Adv. Cryog. Eng. 39,
597–604
10. M. Wiseman, et al, Cryobench—Apparatus for testing cryogenic subcomponents, Appl.
Cryog. Technol. 10, 271–285
11. J.P. Kelley et al., Thermal design and evaluation of the CEBAF superconducting RF cavity’s
prototype waveguide. Adv. Cryog. Eng. 35, 675–682 (1990)
12. I.E. Campisi et al., Higher-order-mode damping and microwave absorption at 2 K, EPAC 92,
vol. 2
13. G. Ciovati et al., Superconducting prototype cavities for the spallation neutron source
(SNS) project. PAC2001, Chicago, IL, June 2001
14. I.E. Campisi et al., The fundamental power coupler for the spallation neutron source
(SNS) project. PAC2001, Chicago, IL, June 2001
15. T. Whitlatch et al., Shipping and alignment for the SNS cryomodule. PAC2001, Chicago, IL,
June 2001
16. J. Hogan et al., Design of the SNS Cavity Support Structure, PAC2001, Chicago, IL, June
2001
17. T. Whitlatch et al., Shipping and alignment for the SNS Cryomodule. PAC2001, Chicago, IL,
June 2001
18. W. Schneider et al., Thermal Performance of the CEBAF Superconducting Linac Cryomodule.
Advances in Cryogenic Engineering, vol. 39 (Plenum Press, New York, 1994), pp. 589–597
19. D. Kashy et al., CEBAF Transfer Line Systems. Advances in Cryogenic Engineering, vol.
37 (Plenum Press, New York, 1992), pp. 577–586
20. L. Cardman, L. Harwood, in The JLab 12 GeV Energy Upgrade of CEBAF for QCD and
Hadronic Physics, Proceedings of PAC07, pp. 58–62
21. C.E. Reece et al., in Optimization of the SRF cavity Design for the CEBAF 12 GeV Upgrade,
Proceeding of the 13th International Workshop on RF Superconductivity
22. J. Sekutowicz et al., in Cavities for JLab’s 12 GeV Upgrade, Proccedings of the 2003 Particle
Accelerator Conference
23. D. Douglas, J. Preble, JLAB-TN-98-022
24. M. Drury et al., Commissioning of the CEBAF Cryomodules, Proceedings of the 1993 Particle
Accelerator Conference, pp. 841–843
25. R. Hicks, JLAB-TN-07-037
26. R. Hicks, E. Daly, JLAB-TN-02-040, Thermal Analysis of a 13 kW Waveguide for the 12 GeV
Upgrade Cryomodule
27. C. Reece et al., High Thermal Conductivity Cryogenic RF Feedthroughs for Higher Order
Mode Couplers, Proceedings of 2005 Particle Accelerator Conference
194 E. Daly et al.

28. K. Davis et al., Development and Testing of a Prototype Tuner for The CEBAF Upgrade
Cryomodule, Proceedings of the 2001 Particle Accelerator Conference, pp. 1149–1151
29. J. Hogan et al., Design of the CEBAF Energy Upgrade Cryomodule Cold Mass. Proceedings
of the 2001 Particle Accelerator Conference, pp. 1595–1597
30. T. Whitlatch, Space frame structural analysis for the CEBAF upgrade cryomodule,
JLAB-TN-00-002
31. G. Chang, E. Daly, C100 cryomodule end can piping design pe ASME B31.3,
JLAB-TN-07-056
32. M. Drury et al., CEBAF upgrade: cryomodule performance and lessons learned, Proceedings
of SRF2013, pp. 836–843
33. T.H. Nicol, R.C. Niemann, J.D. Gonczy, Design and Analysis of the SSC Dipole Magnet
Suspension System, Supercollider 1 (Plenum Press, New York, 1989), pp. 637–649
34. L. Ristori, et al., Design of single spoke resonators for PXIE. Presented at IPAC 2012, paper
ID: 2689-WEPPC057
35. S. Kazakov, et al., Main couplers design for project X. Presented at IPAC 2012, paper ID:
2523-WEPPC050
36. A. Ballarino, Conduction-cooled 60 a resistive current leads for LHC dipole correctors, LHC
Project Report 691, 2004
37. T.H. Nicol, G. Lanfranco, L. Ristori, High intensity neutrino source superconducting spoke
resonator and test cryostat design and status. IEEE Trans. Appl. Supercond. 19(3), 1432–1435
(2009)
Chapter 7
Special Topics in Cryostat Design

Wolfgang Stautner

Abstract This chapter describes a series of special topics that, while coming from
the development of Magnetic Resonance Imaging cryostats, are applicable to many
other cryostat designs as well. The topics are: boil off minimization, cryocooler
integration, designing with inclined tubes and pressure relief and venting. This
chapter contains many figures, tables, equations, design algorithms and references
useful to cryostat designers.

7.1 Boil off Minimization for Cryostats Without

a Cryocooler

This section gives an example calculation for finding the optimal thermal position
and length required for linking a thermal shield to a neck tube that supports the
helium vessel for minimizing helium boil off (BOFF) [1]. Figure 7.1 shows a
typical neck design and Fig. 7.2 shows the model being examined.
The heat balance equation at node T1 is then written as follows:

At node T1: R31
R10
C11 þ C21
C00 þ C02 = 0 ð7:1Þ

The positional variability is indicated by the double arrow at node T1 and X.

Note that T3 refers to the temperature of a liquid nitrogen reservoir, usually 80 K.
Where in Fig. 7.2 and Eq. 7.1:
T1 Node temperature of thermal shield at e.g. 40 K
T2 Contact temperature of LN2 reservoir or similar with helium vessel
suspension tube
T3 Temperature of liquid nitrogen vessel or thermal shield with temperature T3
(can be higher or lower than T2 to simulate e.g. contact resistance at T2)

W. Stautner (&)
GE Global Research, Electromagnetics and Superconductivity Lab,
One Research Circle, Niskayuna, NY 12309, USA
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 195

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_7
196 W. Stautner

Fig. 7.1 Typical neck tube

with copper links using Top flange: 300 K
ultrasonic soldering for
consistent thermal contact
resistance quality of the
copper/steel interfaces

Thermal link 1: 80 K

Thermal link 2: X K

Bottom flange: 4 K

Fig. 7.2 Modeling heat

transport and required contact
length to achieve minimum
BOFF

Flow
4K

T1 = x

80 K
300 K
7 Special Topics in Cryostat Design 197

Fig. 7.3 Tube layout as

shown in the example

e.g. 170 mm (distance

to next thermal link or
3
thermal shield)

33 mm (thermal contact
2 length, soldered copper)

120 mm (distance to
1
helium reservoir top
plate)
X

T4 Temperature of vacuum vessel surrounding liquid nitrogen vessel

R31 Thermal radiation from 80 K shield to the intermediate shield temperature
R10 Thermal radiation from the intermediate shield temperature to helium
reservoir
C22 Thermal conduction from liquid nitrogen vessel or thermal shield at this
location
C21 Thermal conduction at tube bottom into node T1
C11 Thermal conduction leaving location T1 entering tube connected to helium
vessel
C10 Thermal conduction at tube bottom connected to helium vessel
C00 Conductive heat loade.g. heat sink to external node leaving thermal
boundary at node T1
C01 Conduction term due to residual gas (vacuum) to helium vessel (can also be
used to simulate contact resistance etc.)
C02 Conduction term due to residual gas (vacuum) to intermediate thermal shield
(can also be used to simulate contact resistance etc.)
Qadd Extra heat term to simulate thermal conduction heat load (e.g. suspensions)

Variables:
kHelium and kSteel = variable thermal conductivities for helium and steel
cpHelium = constant average value between T0 and T1 or T1 and T2
R31 6¼ R10 = thermal radiation
Nu1 6¼ Nu2 6¼ Nu3 = Nusselt number for each tube section
T3 6¼ T2 = Temperatures can differ
e tube 1 6¼ e tube 2 = tube emissivity coefficient can differ
198 W. Stautner

There are different ways of solving the heat balance equations. A particular
useful exercise is the finite difference equation (FDE) approach since this allows
you to change the tube structure for each node along the tube length. The equations
for conduction, convection, radiation and gas flow as written down by Dusinberre
[2, 3] and Croft [4]. The software Kryom Version 3.3 also gives you a good initial
estimate on the thermal link position and boil off [5] (Table 7.2).
The energy balance equations solved with FDE for each node are conveniently
written as:

C21
C11 þ R21
R10 ¼ 0; ð7:2Þ

C10 þ Cadd þ R10 ¼ H1v ; ð7:3Þ

C11
C10 ¼ H10 ð7:4Þ

C22
C11 ¼ H21 : ð7:5Þ

(with conduction and thermal radiation terms as shown in Fig. 7.2 and the increase
in enthalpy H10 and H21 of helium gas and with Hlv as the product of latent heat and
boiling off helium and considering boil off in ullage space).
We now iteratively search for the best TS link position that results in minimum
boil off. Furthermore, we obtain information on thermal shield temperature and the
profile along the neck tube (see Fig. 7.4). In addition we also get quantitative results
on the individual heat flow parameters for the cryogenic design.

Table 7.1 Typical neck tube assembly input parameters as shown in Fig. 7.3
He − TS − LN2 (K) 4 42 (start value) 80 Temperatures
Nu1 − Nu2 − Nu3 (–) 4 4 4 Nusselt number
dx1 − dx2 − dx3 (mm) 1 1 1 FDE length
Positions (1/2/3) 120 33 170 Tube length
Ra/Ri (mm) 15/14.5 Tube radii
Emissivity coeff e (–) 0.07/0.065 Thermal radiation
He − TS − LN2 (m2) 2.45 1.95 1.76 Cryogenic surfaces

Table 7.2 Results with Thermal radiation (R31) (mW) 0.138

Table 7.1 parameters
Thermal radiation (R10) (mW) 0.0139
Thermal conduction (C10) (mW) 1.02
Thermal conduction (C22) (mW) 127
Thermal conduction (C21) (mW) 18
Thermal conduction (Cadd) (mW) 1
Optimal TS temperature (K) 44.6
Optimal contact location (mm) 120
Optimal TS/Cu contact length (mm) 33
Minimum boil off per tube (ml/h) 19
7 Special Topics in Cryostat Design 199

Fig. 7.4 Optimum thermal link position and required contact length for 290 mm SS tube,
25.4 mm

One can modify the FDE nodes to include annular tubes, corrugated tubes,
multiple thermal shields or even forced flow.

7.1.1 Discussion

The process of optimizing boil off of a helium vessel can be explained as follows.
Once a thermal link from the TS to the neck tube is made, the heat balance around
T1 changes correspondingly until equilibrium between all heat sources is obtained
and a continuous boil off rate is established.
Figure 7.5 shows how the parameters within this heat balance node T1 change if
the TS link position is moved. If the link, for example, is made higher up the neck
tube (Figs. 7.2 and 7.3), the amount of heat conducted down from T2 to T1
increases. As a result, the thermal shield temperature increases as well but since the
tube length from T1 to T0 increased boil off will be lower. The designer therefore
tries to find the minimum as shown in the curves in Fig. 7.5.
200 W. Stautner

Fig. 7.5 Optimal thermal shield temperature, link position and minimum BOFF

7.1.2 Pitfalls

Minimizing boil off is even more important for multiple tubes in contact with a
helium vessel. The problem here is to design the tube arrangement in way that all
tubes will be receiving the same mass flow rate. In case of flow restrictions at the
neck tube bottom (for example a siphon entry tube or a current lead receptacle) the
flow along the tubing will not be evenly distributed. The flow restriction can occur
in the tube itself in the form of wire looms or copper baffles. Experiments confirmed
the existence of large temperature differences between wall and fluid in this case.
For small boil off flow the Nusselt number is almost always 4 [6] which relates to
perfect heat transfer between gas and wall. For imperfect heat transfer and
non-optimized neck tubes higher boil off rates will be obtained and the temperature
difference between wall and gas can become quite large calling for a longer copper
link to “lift” up the gas temperature.
The design for minimum boil-off results in efficient cryogen recondensation and
re-liquefaction. In times of high cost of helium; re-liquefaction of helium gas, even
for smaller cryogenic systems, is convenient and often advisable.
7 Special Topics in Cryostat Design 201

7.2 Cryocooler Integration

Since their commercial availability in the 1980s, cryocoolers have become excellent
and versatile tools for a diverse range of cryogenic applications, from heat sinking a
cold mass, to distributed cooling and even for liquefying or dedicated component
cooling. At the cryogenic design stage, the cryogenic engineer needs to know in
which way the chosen cooler: Gifford-McMahon (GM), Pulse Tube Refrigerator
(PTR) or Stirling type can be integrated in a cryogenic structure and in particular
how the thermal links to the cryocooler cold stages should be made.
Generally, the cryocooler interface cost is driven up significantly when
demanding minimum temperature difference and by associated requirements, e.g.
low vibration transmission, or when it is necessary to retract or remove the cooler.
Cryogenic designs have to balance this contact quality effort and make a trade-off
against what is needed and is feasible for assembly and cooler maintenance. The
most common design solutions for dedicated applications are therefore listed below
and illustrated in the following examples.

7.2.1 Cryocooler Integration—Options Overview

Cryocooler attachments for thermal mass cooling

• Fixed bolt on, braided, internal cryocooler parts removable
• Sleeved sleeve bolt-on or braided, cryocooler removable from
sleeve/serviceable
• Un-sleeved cryocooler exposed to cryogenic atmosphere, e.g. GM/PTR cooler
in helium gas
• Retractable temporarily engage/disengage from cold mass, e.g. after initial
cool-down, sleeved
Cryocooler attachments for specialty applications, distributed cooling and
component cooling
• Liquefying coolers (fins machined onto cooler cold stage/recondensing fins in
standalone cup)
• Heat-sinked heat pipes
• Heat-sinked current leads
• With heat exchanger (HX) and fan-assisted coolant flow
• Remote cold finger
202 W. Stautner

7.2.2 Cryocooler Integration Examples

A good example for directly embedding a pulse tube cooler (PTR) in the coil former
of the superconducting magnet cold mass itself is the head imaging prototype
system built by GE Global Research back in 1995 (see Fig. 7.6). Technically, this is
a bold approach since the cooler cannot be removed, but due to the absence of a
moving piston no vibrations are transmitted to the superconducting magnet struc-
ture. Furthermore, this cooler does not have to be serviced but can be purged of
contaminants when warmed up to room temperature, if necessary.
Direct cryocooler cold mass bolt-on is also an interesting option for a number of
small-scale MRI magnet systems using medium or high-temperature supercon-
ductors (HTS), see also (1) in Fig. 7.14.
Another early, interventional Helium-free LTS MRI magnet system [8] (see
Fig. 7.7) built by GE also used fixed installed cryocoolers. Other than PTRs, GM
type coolers tend to transmit their piston vibrations to any connected structure. To
mitigate this effect flexible braids are introduced between the interfaces to avoid any
ghosting appearing in MRI images.
Both coolers in Fig. 7.7 are sleeved to allow the removal of the cooler from the
cold plate during service. The cold plate or heat bus extends from the left to the
right of both Helmholtz style superconducting Nb3Sn coils, operating at 10–11 K.
Note that un-sleeved GM coolers can also be serviced if required, by carefully
removing the internal piston assembly from the cooler shell tubing. In this case a
warm swap to room temperature may be required which requires a cold mass warm
up, unless the cooler envelope can be heated to a high enough temperature to allow
servicing or refurbishment.

Early Cryomech Pulse Tube Cooler

that achieved a no-load base
temperature of 2.2 K.

Fig. 7.6 GE 0.5 T conduction-cooled head imaging system [7]

7 Special Topics in Cryostat Design 203

Fig. 7.7 a Cryocoolers (2) (dotted line) in cold box and in contact with SC coils via heat bus.
b Surgeons in the operating theater

7.2.3 Schematics and Options of Cryocooler Integration—

Overview

Figure 7.8 shows the most simple and straightforward way of achieving a good
thermal contact across the interfaces. The cryocooler comes with bolt holes on the
stages for bolting on to a thermal shield or a cold mass. Since the thermal shrinkage
at the stages needs to be considered, bolts with Belleville washers are advisable in
case of frequent warm up/cooldown.
A typical cryocooler bolt-on configuration for vibration isolation is shown
schematically in Fig. 7.9.
For a description of other common flexible conductive links: braids, foils etc.,
see also [9].
Figure 7.10 shows the race track test coil cooling [10] supported on Heim
columns [11, 12] and with attached current leads.
Recently a similar approach has been proposed by Sun [13] for testing race track
coils for wind power applications as shown in Fig. 7.11.
All direct bolt-on cryocoolers may be designed sleeved or non-sleeved,
depending on maintenance requirements. The same can be said for coolers that are
used to recondense cryogens, e.g. helium gas. See Figs. 7.12 and 7.13.
There are downsides to both designs. A non-sleeved design requires a good 1st
stage pressed contact onto the cooler interface and a work around for any con-
vective parasitic heat loads between 1st and 2nd stage that may arise (in the absence
of stratification). One also has to take precautions against air ingress that later could
block internal tubing to an immersed magnet making any recondensation impos-
sible. For the sleeved system, a good, repeatable thermal contact at both cryocooler
stages is required.
204 W. Stautner

(a) (b)
PTR/
GM

1st stage Thermal shield 1st stage

Interfaces
2nd stage Cold mass
2nd stage

Fig. 7.8 a Cryocooler bolted onto thermal shield and thermal mass, b SHI 4 K cryocooler RDK
408 D2 with bolt holes at first and second stages (courtesy of SHI cryogenics)

(a)
GM
(b)

1st stage Thermal shield contact

2nd stage Cold mass thermal contact

Cold plate /
heat bus

Fig. 7.9 a GM or PTR type cryocooler bolted onto cold plate with highly conductive copper or
aluminum braids to cold mass (GM type), see also (1) in Fig. 7.14. b Example of a foil welded
copper link

Cryocooler

Thermal shield part

Thermal shield
Cryocooler 2nd stage with
braided copper interfaces

Vacuum chamber wall

Fig. 7.10 Bolt-on design (vacuum chamber open for clarity), directly to cold plate and race track
cold mass with lead in/out
7 Special Topics in Cryostat Design 205

Fig. 7.11 Race-track coil for wind power applications (test bed)

Fig. 7.12 Sleeved cryocooler and with thermal shield contact with liquefaction cup fitted to the
second stage

Fig. 7.13 Sleeveless cryocooler, in direct contact with helium gas and link to thermal shield
206 W. Stautner

Preferred contact means across the interface are either Indium or the less costly
Apiezon N® grease.
A cooler can also be integrated and made detachable from its contact stages for a
number of reasons. For example, in case of a sudden, unplanned power outage and
following compressor shut down resulting in a loss of cooling power, a thermal
short would be created that leads to high heat loads to the cooling components (e.g.
thermal shield, current leads, cold mass). For large thermal mass cooling (e.g.
during cool down) a shielded drive mechanism can be installed. The drive mech-
anism moves the cooler cold end to the cold mass interface maintaining a specified
contact pressure that compensates for the cold mass shrinkage. To avoid a per-
manent parasitic heat load on the cold mass that can be substantial for larger
coolers, the cooler is then retracted automatically [14].
Figure 7.14 shows MRI applications using high temperature superconductors
and cooling approaches as discussed in this section.
The cryocooler for the so-called open C-magnet is rigidly and sleeveless bolted
onto a cold plate. From the cold plate, braids are routed to the copper sheets that are
sandwiched between the individual pancake coils. (2) and (3) show various

(1) (2) (3)

Cold mass with MLI in Cold mass after Cold mass after
vacuum case winding winding

Fig. 7.14 Cryocooler integration methods for small-scale HTS MRI. 1 Cooling technique:
conduction cooling cryogen: n/a cooler: single-stage GM [15]. 2 Cooling technique: thermosiphon
cryogen: neon, cooler: single-stage GM [16, 17]. 3 Cooling technique: Thermosiphon, dual
cryogens: helium or hydrogen, cooler: dual-stage GM [18]
7 Special Topics in Cryostat Design 207

solutions for extremity scanners for different superconducting wires using ther-
mosiphon technology and cryogen liquefaction with zero boil-off.

7.2.4 Cryocooler Integration Techniques for Special

Applications

Heat pipes can be heat-sunk to a first stage of a cryocooler, preferably for sleeved
versions. The cold end of the heat pipe extends to the component that needs to be
cooled. Heat pipes generally require no maintenance and, depending on the
application, are closed system components that give the opportunity of deeply
embedding them in a cryogenic structure.
A further common practice is to heat sink the resistive parts of a current lead or
the warm end of a HTS lead to a cryocooler. One little complication here is that one
needs to work around the conundrum of requiring good thermal contact and elec-
trical isolation that normally exclude each other.
Last but not least several novel applications have surfaced recently [19].
Cryozone for example attaches a heat exchanger onto the cold stage of a cryocooler,
mainly for steady state forced-flow HTS application cooling, e.g. superconducting
cables etc. [20] (see Fig. 7.15).
It is also possible to run actively cooled circulation loops with hybrid pulse tube
refrigerators combined with circulators and cold flow rectifiers as demonstrated by
Feller, Salerno, Nellis et al. [21].
Thus, cryocoolers are playing a key role whenever distributed cooling of large
surfaces is required. Gas or liquid circulation through cooling loops is possible
using micro pumps, blowers or impellers. For HTS applications, the compressed
gas can exit, return to the cooler and become part of the distributed cooling circuit.

Fig. 7.15 a Cryocooler with attached heat exchanger enabling cryogen flow circulation and
b cryofans (courtesy of Cryozone)
208 W. Stautner

Fig. 7.16 Cryocooler cold finger cooling with cold helium transfer line (courtesy of ICC Press)

Some designs for specific applications require operating the cryocooler at a

distance from the object to be cooled. One reason could be the better serviceability
for the cooler or if vibration transmission to the object should be minimized which
otherwise would interfere with a measurement. This kind of “elongation of the
cryocooler cold stage” has recently made possible with a novel approach taken by
Cryomech [22] shown in Fig. 7.16. This “cold finger” solution delivers very effi-
cient remote cooling of electronic and other components.

7.3 Designing with Inclined Tubes in Cryogenic Systems

When designing cryostats, the engineer occasionally may be tempted to introduce

tubing, penetrations or gas filled sleeves in an angled orientation. There are a
number of reasons for implementing those, either for getting better accessibility to
the internal cryostat structure from the top, when patient room height limitations
need to be considered or simply for better liquid draining.
7 Special Topics in Cryostat Design 209

In other designs where the cryostat envelope space is tight, gas filled tubes
and/or operating under high pressure can only be fitted inclined. Furthermore,
cryocoolers themselves are gas filled tubes under pressure, in particular the pulse
tube cooler, with its absence of any solid moving piston. The moving piston here is
the oscillating gas column itself. Not surprisingly there are operational limits due to
inclination for this cooler, as we will see.
So why should the cryogenic designer be careful when routing inclined tubes or
a network of pipes? Inclined tubes generally introduce a parasitic heat flow driven
by the density difference in the working medium due to the difference in the length
of the wall conduction heat path and the resulting heat load that sets up fluid motion
within these tubes. If there is only a small deviation required from the vertical it is
almost always better to keep the tubing vertical. The answer really depends on how
critical it is to maintain the cryostat heat balance and whether any cost increase that
may result is acceptable.
Inclined tubes can be classified as follows:
• Inclined tubes (neck tubes/turrets) at low or high operating pressure
(open1/closed)
• Inclined tubes with pulsating flow at high/low frequency
• Inclined tubes for thermal mass cool-down
• Inclined heat pipes
• Inclined thermosiphons
When designing cryostats we often notice that the heat balance deviates from
expected, calculated values. It is then we look out for parasitic heat loads that may
have been missed in the design or at last minute changes in the assembly. As
Scurlock noted in 1977 [23], “the boil-off is always 50–100 % greater than the
design figures”. This prompted Islam and Scurlock to conduct vapor column flow
visualization experiments for vertical and inclined tubes. The team was the first to
show in which way the gas column flow changes when tubes are tilted. Tests were
performed on wide open tubes as well as on ones with small diameters that are open
to atmosphere. It was possible to differentiate between 6 distinct regions in which
the flow changes within the column.
Figure 7.17 shows a typical tubing that gives easy access to a liquid reservoir, e.g.
for current lead insertion or for filling the reservoir or for using it as a boil-off tube.
This tilt angle was identified as a source of heat leak into the cryogenic liquid
storage vessel. Further research detail is given in [24–28]. With the starting interest
in pulse tube cooler technology researchers quickly noticed the orientation
dependency of the tubing. A typical curve is shown by Kasturirengan [29], (see
Fig. 7.18a) for a pulse tube cooler running at 15 bar at a frequency of 1.6 Hz. The
inclination angle is shown in Fig. 7.18b. Figure 7.18a shows the convective, or

1
Open under high pressure in this case means the open end of the tube is connected to a larger gas
reservoir under high pressure.
210 W. Stautner

290 K

265 K

77 K

Fig. 7.17 Tilted, open access tube for a cryostat [23]

Fig. 7.18 a Typical PTR cooling power loss, b at different inclination angles [29]

parasitic heat loss that reduces the cooling power at different inclination angles. For
high frequency operation this effect is reduced as Thummes reports [30].
The recent research of Langebach and Haberstroh [31] and others is of great
value when inserting tubes in cryostats that work under high pressure and are closed
or open (closed with reservoirs) at the ends. Why is that important for cryostat
designs? Cryostats sometimes are not only used as storage container or for
experiments but for housing a cold mass and in particular a superconducting
7 Special Topics in Cryostat Design 211

Fig. 7.19 Tilted pipe under pressure—convective behavior (courtesy of Langebach)

magnet. The cool-down of a cold mass is often facilitated and accelerated using a
heat pipe, a typical, early configuration is given by Begulilowa [32]. In cryogenic
designs, very rarely has one the comfort of routing the heat pipe vertically and they
most likely will always be inclined. For those instances the computer program
developed by Langebach [33] (based on his experimental correlations) calculates
the performance loss due to the inclination effect before the heat pipe starts to
liquefy at the cold end that is, e.g. during cool-down from room temperature to its
operating temperature. Once liquid is formed within the tube and the pressure
212 W. Stautner

within the tube is reduced to its designed operating pressure other curves apply that
show the degradation in performance of an unwicked heat pipe at tilted angles.
Referring to Fig. 7.19 −90° defines the tube with its warm end up (Tw) whereas
+90° means the cold end is up (Tk). Input parameters are operating pressure in bar,
temperature and average thermal conductivity of the tube wall, length, diameter and
thickness. Further information is required on whether the tube is closed at both ends
or only at one end and what cryogen is used (helium, hydrogen, nitrogen). In his
thesis Langebach examines different tubes at different boundary conditions and
Raleigh numbers and gives a range of methods to prevent convection in inclined
tubes from happening. With one of his solutions—the “snake” tube—parasitic heat
loads were considerably reduced.
As for the particular use of an inclined thermosiphon tubing, this highlights an
interesting aspect that needs further investigation. For some applications, boundary
conditions and operating conditions it may be beneficial to incline a thermosiphon
tube whenever a high quality thermal short is required, e.g. during an initial
cool-down.

7.3.1 Pitfalls

A horizontal thermosiphon tubing or a horizontal manifold is known not to work

and should be avoided as it stops the bubble movement (gas bubble entrapment)
and causes flow instabilities. For bubbles to move towards a recondensing surface, a
horizontal inclination angle of at least 5° is required.

7.4 Cryogenics for Cryostats: Pressure Rise

All cryostats, such as those used for accelerator magnets, SC generators or large
scale storage vessels require design against overpressure. In the event of a super-
conducting magnet “going normal” or by a preceding rupture of the vacuum vessel,
the stored magnet energy of several MJ is quickly dissipated to the helium bath as
well as to the magnet former itself (Fig. 7.20).
The amount of heat transferred to the bath is a function of the magnet energy
content, the current decay time constant, the transient heat transfer to liquid helium
(the obtainable heat flux density (see Table 7.3) in W/cm2), vessel volume and fill
level, the set pressure for the relief valve or burst disk, the magnet former surface
structure, as well as type and orientation of the latter. Each liter of liquid helium at
4 K creates 700 L of gaseous helium that need to be discharged safely; for a 3 T
magnet cryomodule this would amount to 1.4 million gas liters. External quench
gas capture is therefore difficult. Using a cryocooler for reclaiming the quench gas
and associated cooldown usually is not feasible within a reasonable timeframe for
standard bath-cooled cryomodules. If possible, transfer liquid helium back into
7 Special Topics in Cryostat Design 213

Fig. 7.20 Quench gas release

through quench duct after
decommissioning a
cryomodule for MRI (MRI
magnet installed in hospital
basement)

Table 7.3 Typical heat flux Designs—typical configuration Flux

densities in W/cm2 for
cryostats (collapse and/or Superinsulated bath cryostat 0.6
quench) [34] Liquid helium transport can with gas-cooled shields 2.0
Helium vessel without insulation 3.8
Usually for quench calculations 4–5

dewars rather releasing helium to outer space when decommissioning complete

cryomodule systems.

7.4.1 Quench Pressure Rise in Cryostats and Quench Duct

Sizing—A Modeling Example

Predicting quench pressure rises in cryomodules is generally a difficult task, but

nevertheless necessary. The possible increase in 4 K magnet temperature to 40–
50 K after a quench—depends on the chosen coil structure, the coil former material
and mass as well as on the amount and type of wire and coil configuration (e.g. race
track coil or solenoid). When designing the cryostat the important operating con-
ditions have to be fulfilled, with the magnet quench posing all greatest challenge.
Peak mass flow rates of up to 5 kg/s leaving the cryostat at a temperature of 10 K,
partly accompanied by liquid helium expelled at the turret exit (see position 2 in
Fig. 7.21) are not unheard of.
Quench pressure rise modeling has been attempted previously with forced flow
fusion type magnets [35, 36], accelerator quadrupole magnets [37] and smaller
solenoids [38], but only very little so far has been done on many magnet systems,
see for example in [39]. An excellent attempt in how to solve the similar cryostat
214 W. Stautner

Fig. 7.21 Basic quench parameters for state-of-the-art cryostats and energy transfer

vacuum collapse problem has been pointed out by the late Walker [40] on a
19,000 L helium storage tank at Fermi lab.
For a magnet system, pressure P, volume V, temperature T, enthalpy H and
internal energy I at the initial state 0 are given. Once the magnet dissipates its
energy to coil former and helium bath, the pressure rise can be described as iso-
choric, followed by an isentropic expansion [41] through the valve system
(Fig. 7.21: position 2) once the valve opens, and for reasons of simplicity the flow
can be regarded as adiabatic through the quench duct. The stepwise iterative cal-
culation with real gas properties is rather complex involving compressible friction
flow close to the speed of sound simultaneously satisfying gas dynamic equations
for Helium (Fanno line) with the associated transient heat transfer. Figure 7.22b
shows a theoretical quench pressure calculation result. Green line (bar) = quench
pressure decay (dip after burst disk opens), black (kg/s) = volume flow out of
cryostat, red (K) = temperature rise of the helium vessel over 6 s. Safe discharging
of quench gas requires careful design of the turret diameter and providing and
maintaining an unblocked gas passage. The gas transfer mechanism from the
cryostat to ambient atmosphere requires a reliable safety valve, a burst disk and a
quench gas line of sufficient diameter, the latter depending on the customer siting
requirements. Figure 7.22a shows typical quench rates in state-of-the-art
high-energy magnets in cryomodules. Note the sudden rise in pressure from 0 to
5 psi within 2 s. The entire duration of the gas release depends mainly on the
7 Special Topics in Cryostat Design 215

Fig. 7.22 a Typical examples of quench test results for cryo-modules cryostats, b modeling

opening diameter of the quench turret and on the discharging rate of the cryostat
valve. In the example in Fig. 7.22a, curve 1, the total magnet quench time is
approx. 5 s; after that time, the pressure and mass flow starts to decay
exponentially.
Curve 1 represents the course of the quench pressure when selecting a relief valve
that could cause potential problems for several reasons: although the quench valve
opens and the pressure drops immediately, the quench pressure still rises again 5 s
after the relief valve opens and remains constant. Three problems are highlighted
with this example. Firstly, should the gas flow see a further restriction, e.g. due to ice
formation, the pressure increase could make the burst disk blow which is undesirable
since then the magnet is exposed to ingress of air and further turret ice-up. Secondly,
in case the flow restriction in the safety valve increases even further and, if the burst
disk is chosen with a too small a diameter the quench pressure itself can exceed the
design pressure of the cryostat and invalidate the coding requirements for this
cryostat vessel. Lastly, the long quench duration itself leads to icing up of further
vent line components and faults. Design guidelines are given in [42, 43]. Curves 2
and 3 show the course of the quench pressure curve for a burst disk opening pressure
of 4 psi (red line). Curve 3 safely releases the helium gas with sufficient discharge
rate and a high enough kV-value, whereas curve 2 compromises the pressure vessel
design by considerably exceeding the design pressure with the danger of rupturing
the helium vessel or fitted components. It is also interesting to note that the opening
diameter initially helps to quickly release gas into the quench duct, hence the dip in
pressure. There is however a sharp increase one needs to watch out for when fitting
long quench ducts. Curve 2 is also a typical example where the quench duct is too
long leading to a back pressure in the cryostat. In this case the diameter has to be
increased or the number of pipe bends reduced.
216 W. Stautner

7.5 Advanced Cryostat Cryogenics—Carbon Footprint

Considerations

“Advanced cryogenics” [44] aims to reduce the liquid helium inventory by intro-
ducing heat pipe technology, creating a basically “near-dry” type,
conduction-cooled magnet as introduced by the accelerator magnet industry in the
1970s. One of the downsides of a reduced cryogenic volume is the longer time
needed to recool the magnet back to operating temperature. On a more positive note
any loss of vacuum in the cryostat triggering a quench for example, no longer
requires satisfying the high safety ordinance efforts as required for a 3000 L helium
vessel.

References

1. W. Stautner, Quantitative energy balance analysis in cryostats, in Physicist’s Conference, Oral

presentation, Münster, 1984
2. G. M. Dusinberre, Heat Transfer Calculation by Finite Differences (International Text book
Company, 1960)
3. G.M. Dusinberre, Numerical analysis of heat flow (McGraw Hill, 1949)
4. D. Croft, D. Lilley, Heat Transfer Calculations Using Finite Difference Equations (Applied
Science Publishers, 1977)
5. P. Hanzelka, I. Vejchoda, Academy of sciences of the Czech Republic, Institute of Scientific
Instruments (2003)
6. G. Arahonian, L.G. Hyman, L. Roberts, Behavior of power leads for superconducting
magnets. Cryogenics 21, 145 (1981)
7. J.M. van Oort, E.T. Laskaris, P.S. Thompson, B. Dorri, K.G. Herd, A cryogen-free 0.5
Tesla MRI magnet for head imaging. Adv. Cryog. Eng. 43, 139–147 (1998)
8. K.G. Herd, E.T. Laskaris, P.S. Thompson, A dual refrigerator assembly for cryogen-free
superconducting magnet applications. IEEE Trans. Appl. Supercond. 5, 185–188 (1995)
9. J R. Ross, M. Donabedian, Spacecraft thermal control handbook. Cryogenics II (2003)
10. W. Stautner, K. Sivasubramaniam, E.T. Laskaris, S. Mine, J. Rochford, E. Budesheim, K.
Amm, A cryo-free 10 T high-field magnet system for a novel superconducting application.
IEEE Trans. Appl. Supercond. 21, 2225–2228 (2011)
11. J.R. Heim, The heim column, National accelerator laboratory, report TM.334A (1971), pp. 1–
21
12. G. Hartwig, Support elements with extremely negative thermal expansion. Cryogenics 35,
717–718 (1995)
13. J. Sun, S. Sanz, H. Neumann, Conceptual design and thermal analysis of a modular cryostat
for one single coil of a 10 MW offshore superconducting wind turbine, in IOP Conference
Series, vol. 101 (2015) p. 012088
14. W. Stautner, Remote actuated cryocooler for SC generator and method of assembly the same,
US20140100113A1
15. F.J. Davies, W. Stautner, A.F. Byrne, M. Wilson An HTS magnet for whole-body MRI,
EUCAS’99
16. W. Stautner, K. Amm, E.T. Laskaris, M. Xu, X. Huang, A new cooling technology for the
cooling of HTS magnets. IEEE Trans. Appl. Supercond. 17, 2200–2203 (2007)
7 Special Topics in Cryostat Design 217

17. W. Stautner, M. Xu, E.T. Laskaris, G. Conte, P.S. Thompson, C. van Epps, K. Amm, The
cryogenics of a thermosiphon-cooled HTS MRI magnet—assembly and component testing.
IEEE Trans. Appl. Supercond. 21, 2096–2098 (2011)
18. W. Stautner, M. Xu, S. Mine, K. Amm, Hydrogen cooling options for MgB2-based
superconducting systems, in AIP Conference Proceedings, vol. 1573 (2014), p. 82
19. W. Stautner. K. Amm, M. Xu, Cooling systems for HTS applications—overview and critical
assessment, IWC-HTS plenary talk 1, Matsue-Shi (2015)
20. H. Vermeulen, Cryogenic circulators: the solution for cooling problems? Cold Facts 29(2),
49–48 (2013)
21. K.R. Feller, L.J. Salerno, A. Kashani, B.P. Helvensteijn, J.R. Maddocks, G.F. Nellis, Y.
B. Gianchandani, Technologies for cooling of large distributed loads, AIAAA, 092497 (2008)
22. C. Wang, E. Brown, A. Friebel, A compact cold helium circulation system with GM
cryocooler, in 18th International cryocooler conference ICC, Syracuse (2014)
23. M.S. Islam, R.G. Scurlock, Qualitative details of the complex flow in cryogenic vapor
columns. Cryogenics 655–680 (1977)
24. P. Lnyam, A.M. Mustafa, W. Proctor, R.G. Scurlock, Reduction of the heat flux into liquid
helium in wide necked metal dewars. Cryogenics 242–247 (1969)
25. M.S. Islam, R.G. Scurlock, Analysis of solid vapor heat transfer in helium vapor columns at
low temperatures. Cryogenics 323–328 (1978)
26. M.S. Islam, D.J. Richards, R.G. Scurlock, The influences of thermal stratification and flow
interaction on the enhanced natural convective heat transfer at low temperatures. Cryogenics
319–325 (1978)
27. J. Boarman, P. Lynam, R.G. Scurlock, Complex flow in vapor columns over boiling liquids.
Cryogenics 520–523 (1973)
28. P. Lnyam, W. Proctor, R.G. Scurlock, Reduction of the evaporation rate of liquid helium in
wide necked dewars, in Heat Flow Below 100 K, no. 2 (International Institute of Refrigeration,
Paris, France, 1965), pp 351–247
29. S. Kasturirengan, S. Jacob et al., Experimental studies of convection in a single stage pulse
tube refrigerator. Adv. Cryog. Eng. 49, 1474–1481 (2003)
30. G. Thummes et al., Convective heat losses in pulse tube coolers: effect of pulse tube
inclination. Cryocoolers 9, 393–402 (1997)
31. R. Langebach, C. Haberstroh, Natural convection in inclined pipes—a new correlation for heat
transfer estimations, in AIP Conference Proceedings, vol. 1573 (2014), pp. 1504–1511
32. R. Bewilogua et al., Application of the thermosiphon for precooling apparatus. Cryogenics 6,
34–36 (1966)
33. R. Langebach, Wärmeeintrag durch geneigte Rohrleitungen in kryogene Speicherbehälter,
Dissertation, TUD press, 2013, URL: http://books.google.de/books?id=Cez5nAEACAAJ
34. W. Lehmann, Internal report, Safety aspects LHe cryostats and LHe transport containers,
Research Center Karlsruhe, Report 08.01.01P04B (1978)
35. J.R. Miller, ORNL, Pressure rise during the quench of a superconducting magnet using
internally cooled conductors (1980), pp. 321–329
36. P.H. Eberhard, et al, Lawrence Berkeley Lab, Quenches in large superconducting magnets, in
Proceedings of 6th international Conference on Magazine Technology (MT 6), Paper 75,
Bratislava (1977), pp. 654–662
37. K. N. Henrichsen, et al, Analysis of some resistive transitions in the ISR super-conducting
quadrupole magnets. Adv. Cryog. Eng. 27, 245–256 (1982)
38. V. Kadambi, B. Dorri, Current decay and temperatures during superconducting magnet coil
quench. Cryogenics 157–164 (1986)
39. B. Seeber, Handbook of applied superconductivity, in Pressure increase during a quench,
Vol. 2, Figure G2.2.24 (Stautner) (1998), p. 1235
40. R.J. Walker, Calculation of the pressure rise in the Fermilab 19000 l helium dewar, Adv.
Cryog. Eng. 29, 777–784
41. R.J. Walker, private communication (1985)
218 W. Stautner

42. G. Bozóki, Überdrucksicherungen für Behälter und Rohrleitungen. Verlag TÜV Rheinland
(1977)
43. W. Lehmann, Sicherheitsauflagen beim Engineering von LHe- und LN2-Apparaten und –
Anlagen, Research Center Karlsruhe, Report 03.05.01P02A (1982)
44. Y. Lvovsky, W. Stautner, Novel technologies and configurations of superconducting magnets
for MRI. Supercond. Sci. Technol. 26(9), article id. 093001 (2013)
Chapter 8
Design and Operation of a Large,
Low Background, 50 mK Cryostat
for the Cryogenic Dark Matter Search

Richard L. Schmitt

Abstract Cryostats that operate below 1 K have additional requirements involving

the need for extremely small heat leaks and alternative cooling methods. This
chapter describes the design and operation of a cryostat operating at 50 mK for the
Cryogenic Dark Matter Search experiment. Included are descriptions of the cooling
and thermal insulation system, seal design, fabrication and operations. Valuable
data is provided on the thermal conductivity of Kevlar and the calculation of joint
conductance. The particular issue of using only radiopure materials in the cryostat
construction is also covered. A list of lessons learned from the cryostat operation is
provided.

8.1 Introduction

To support the Cryogenic Dark Matter Search (CDMS) experiment, a sub Kelvin
cryostat and support system was built and operated at Soudan MN. In addition to
the challenges of milliKelvin operation, the cryostat was made with low back-
ground material, and was located in an RF shielded clean room one-half mile below
the surface at Soudan Underground State Park.

8.2 Physics Detectors and Towers

The goal of CDMS was to detect dark matter, which has been detected through
large-scale gravitational interactions. CDMS used cryogenic germanium and silicon
detectors, which are capable of detecting weakly interactive dark matter (WIMPs).
WIMPs are detected through their interactions with the nuclei in the germanium.
When a nucleus is hit, it recoils, causing the whole germanium crystal to vibrate.

R.L. Schmitt (&)

Fermi National Accelerator Lab, P.O. Box 500, Batavia, IL 60510, USA
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 219

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_8
220 R.L. Schmitt

These vibrations, or phonons, propagate to the surface of the crystal where they heat
sensors consisting of thin aluminum traps connected by tungsten meanders. The
tungsten is kept at its Tc (critical temperature) and so is an ultra-sensitive ther-
mometer, one capable of sensing temperature changes a fraction of a milliKelvin.
The extra energy—as little as 5 keV—the phonons bring raises the tungsten’s
resistance by a minute amount. When the phonons reach the aluminum, they excite
quasi-particle states, which propagate to the tungsten and heat it up.
When the temperature of the tungsten rises so does the resistance of the circuit.
This causes the bias current to decrease since the voltage across the tungsten is held
constant. The resulting pulse is picked up by SQUID (Superconducting QUantum
Interference Device) amplifiers.
There are other particles that go through the detectors besides WIMPs.
The CDMS detectors were shielded to minimize the number of these other particles.
The detectors are capable of discriminating between most of them and the WIMP
signal [3, 17, 22].
The detectors ran in the Soudan Mine in Minnesota, a half-mile underground.
The deep site is chosen to shield the detectors from cosmic rays.
Six detector crystals were mounted in each tower assembly (see Fig. 8.1). The
towers provided structural support for the detectors and thermal heat sinking at the
base temperature. The tower features included heat sinks at 200 mK, 1 and 5 K for
the structural supports and signal wiring. Wiring and structural elements in the
tower were a major portion of the total heat load.

Fig. 8.1 Detector tower with one crystal. Completed tower supports six crystals [17]
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 221

8.3 Cryogenic System General Description

Science requirements were for a cryostat with very low radioactive background,
shielded against gamma and neutron radiation. Detector operation below the
tungsten Tc requires multiple stages of heat shielding. The cryostat had to be large
enough to contain seven detector towers. The low background requirement strictly
limited materials that could be used.
The CDMS ICEBOX provided the mounting, thermal shielding and cooling for
the CDMS detectors. The ICEBOX includes six nested cans with top access, a stem
(E-stem) to carry out the signal cables, a stem (C-stem) for heat conduction, and
aramid rope suspension. The inner can held 0.022 m3 volume. The CDMS Icebox
at Soudan operated for twelve years including multiple runs over one year in
duration.
Figure 8.2 shows the overall layout of the experiment, showing the detector
space surrounded by Icebox, a magnetic shield, lead and polyethylene shielding.
The assembly was installed in an RF shielded, class 10,000 clean room 2300 ft.
below the surface at Soudan Underground State Park [18] in Soudan, MN
(Fig. 8.3).

Fig. 8.2 Overall cryogenic layout, dilution refrigerator on the left, ICEBOX cans in the middle.
C-stem provides conductive cooling from the ICEBOX to the fridge. The signal feedthrough is on
the right [17]
222 R.L. Schmitt

Fig. 8.3 Icebox top, showing tops of the OVC, shield can, IVC can with five towers [17]

8.4 Dilution Refrigerator Introduction

The next sections refer to dilution refrigerator cooling. For the reader unfamiliar
with this type of refrigerator the description of a ‘wet’ can be found in general
literature, for example at [7].

8.5 Icebox General Description

The CDMS ICEBOX provided the mounting, thermal shielding and cooling for the
CDMS detectors. The ICEBOX includes six nested cans with top access, a stem
(E-stem) to carry out the signal cables, a stem (C-stem) for heat conduction, and
aramid rope suspension. The stem length was needed to pass through the thick-
nesses of the lead and polyethylene shielding. Thermal models have confirmed that
this number of layers is appropriate for the large milliKelvin detector operation.
For CDMS the temperatures chosen were suited to the operating stages of the wet
dilution refrigerator. Future large detectors may use dry fridges and may not be
limited by helium or nitrogen boiling temperatures.

8.5.1 Icebox Cans

Each of the cans provides heat shielding and mounting for the next inner can.
Figure 8.4 shows the top view of the can assembly. The cans are referred to
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 223

Fig. 8.4 ICEBOX cans, top view. The support adjustment screws for the outer layers can be seen
at the bottom of the view. The signal cables exit through the E-stem on the left [3]

(moving from the outermost inward) as the Outer Vacuum Can (OVC), the Shield
can (SH), the Inner Vacuum Can (IVC) can, the Still (ST) can, the Cold Plate
(CP) can, and the Mixing Chamber (MC) can. The names refer to the associated
dilution refrigerator stage.
The MC lid seen in Fig. 8.4 shows the seven hexagonal covers where towers can
be installed. The purge tubes on the right are removed before closure.
The OVC was the outer vacuum enclosure. It was a vertical right cylinder
22.5 in. diameter and 27 in. tall. The walls were 1/8 in. thick and the top and
bottom lids were 5/8 in. thick. The insulating vacuum is continuous with the outer
vacuum of the dilution refrigerator. The removable lid and stem connections were
sealed with Viton O-rings. The bottom of the OVC rested on the internal poly-
ethylene shielding.
The Shield can operated at 80 K. It was wrapped with multilayer insulation
(MLI) to reduce thermal radiation from the OVC. The inside was gold plated over
nickel flash to reduce thermal emissivity. It is conduction cooled through the
C-stem back to the liquid nitrogen reservoir in the dilution refrigerator. The Shield
Can is not vacuum leak tight, but does provide a complete thermal radiation barrier
and support heat intercept for the IVC. The shield can was 20 in. in diameter and
24.4 in. tall.
224 R.L. Schmitt

The IVC can operated at 5 K. The inside and outside were gold plated to reduce
thermal emissivity. It was conduction cooled through the C-stem back to the liquid
helium bath in the dilution refrigerator. The IVC separated the inner and outer
vacuums. The inner vacuum was continuous with the dilution refrigerator inner
vacuum and with the electronic feedthrough box. This can contained two lids; and
inner lid that provided termination and final heat sink for the signal cables, and an
outer lid that served as a vacuum tight cover.
All seals inside the background shielding were made with gaskets of annealed
C101 copper sheets pressed between two bull nose raised surfaces with a 0.044 in.
radius. Alignment pins on the outer lid and the stem connections ensured that
opposing bull nose ridges matched each other. Fastening screws applied enough
force to yield the annealed gaskets. Figure 8.5 shows the bull nose design. Seals
outside the background shielding were made with indium wire. The IVC was
designed to withstand full vacuum or atmospheric pressure with or without vacuum
in the OVC.
The main purpose of the IVC was for detector cool down. The inner layers are
thermally isolated from one another both in the cans and the dilution refrigerator,
good for normal operation but not for cool down. By adding helium gas at a
pressure of 0.1 atmosphere inside the IVC, inner cans and detectors become ther-
mally well connected. This gas remains in place during cool down until a tem-
perature of about 7 K is reached, at which time it is evacuated. The OVC remains
fully evacuated during cool down, maintaining the MLI performance.

Fig. 8.5 IVC seal design [11]

8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 225

The Still can operated around 1 K. It was conduction cooled through the C-stem
back to the dilution refrigerator’s still stage. The Still can is not vacuum tight but
does provide a thermal radiation barrier and heat intercept for the CP layer supports.
The Still can is 16 in. diameter and 16 in. tall.
The CP can operated around 230 mK. It was conduction cooled through the
C-stem back to the dilution refrigerator. The refrigerator connection was at a
midpoint in the counterflow heat exchangers between the still and the mixing
chamber. The CP can is not vacuum tight be does provide a heat intercept for the
MC layer supports. The CP can is 14.3 in. diameter and 14.3 in. tall.
The MC is the innermost can and operated at 56 mK. It was conduction cooled
through the C-stem back to the dilution refrigerator. This can was connected to the
dilution refrigerator mixing chamber. It provided heat sinking for the detectors and
supported the weight of the detectors and towers. Both of those functions were
carried out through the lid, which was made of a flat plate with hexagonal openings.
It was 12 in. diameter and 12 in. tall.

8.5.2 Suspension

The OVC can rested on the inner polyethylene shielding. The five inner cans were
each suspended on aramid fiber strings looped over pulleys attached to the top of
the next warmer can. Aramid ropes have very low thermal conductivity and low
elasticity. The loops were attached at the top of the warmer can and the bottom of
the colder can. The adjustment screws and rope arrangement for each can allowed
height adjustment and leveling. The innermost MC can had three equally spaced
suspension loops. A flexible cable in the C-stem allowed some motion between the
fridge and the MC can. The next four inner loops were suspended on two loops
diametrically opposed and perpendicular to the stem attachment. For these four cans
the dilution refrigerator provided the third support point, balancing the cans. The
side loops were carefully adjusted to match the tail flanges with the fridge bottom
flanges.

8.5.3 C-Stems and Tails

The Fridge Tails and Cold Stem connected the dilution refrigerator with the
ICEBOX. All six layers were continuous to the associated layers in the fridge,
which provided all cooling.
Changing the stem direction from horizontal to vertical was a challenging task.
This was accomplished with five nested tees and a bolted conduction rod connected
to the bottom of the fridge. The fridge bottom connection is shown in Fig. 8.6.
226 R.L. Schmitt

Fig. 8.6 Bottom of dilution

refrigerator [17]

8.6 E-Stem

The E-stem carried the signal cables from the top of the towers at 5 K to the room
temperature feedthrough box. It extended the OVC which terminated at the outer
end of the E-stem. The LN shield layer extended though the E-stem and terminated
with a stainless steel bellows at the outer end. A signal cable heat intercept at 90 K
was located at the outer end of the stem. The LN layer in the E-stem was gold plated
for low emissivity.
The IVC also extended through the E-stem. It carried the signal cables with a
5 K heat intercept near the can end. The IVC vacuum extends through the E-stem
and into the feedthrough box. The E-stem inside diameter was 2 in., large enough to
handle the maximum number of signal cables and their heat sinks at 90 and 5 K.
The inner, narrow portion of the E-stem passed through the gamma and neutron
shielding. The larger, outer portion contained the bellows isolating the inner and
outer vacuums and temperature layers.

8.6.1 Thermal Contraction

Aramid loops have a slight thermal expansion, but the copper cans shorten when
cooled, effectively lowering each can relative to the OVC. And since the dilution
refrigerator is internally suspended from the top, it raises the balancing supports
when cold. Therefore the ICEBOX cans lower and tilt when cooled down. The
nested stems fit very closely together and the heights were adjusted so that they
would have the maximum clearance cold. This meant that some adjacent stems
were touching at room temperature but separated during cool down. Flexible
connections in the E-stem allowed the LN and IVC to bend neat the can slightly
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 227

during cool down, and bellows at the outer end of the E-stem allowed contraction
along the stems.

8.6.2 Materials, Radiopurity

The materials used in the ICEBOX were strictly controlled to reduce the risk of
radioactive contamination. Many common materials used in cryostat fabrication
contain small amounts of radioactive contaminants or isotopes, so consequently
they could not be used. Copper alloy C101 was the primary cryostat material, the
highest purity available commercially. Although higher purity copper can be made
with lower oxygen content, it is not available in the required quantities. One of the
fabrication goals was to minimize cosmogenic activation which occurs after elec-
trostatic refining when the material is above ground. The material was purchased
with the most recent refining date reasonably achievable. Then the copper was
stored underground when not needed and brought to the surface for fabrication.
Fasteners were primarily socket head brass screws and brass alignment pins.
These were custom made from free machining brass rods, again for material con-
trol. Temperature sensors were RuO2 or platinum resistors, connected with phos-
phor bronze wiring, and either polyimide or Teflon wire insulation. Millmax pins
and counted solder were used to terminate sensor wiring inside the ICEBOX.
Samples of all materials used were tested in a low background radiation counter
before fabrication.

8.6.3 Fabrication

All machining was done with carefully cleaned work areas and new tools to reduce
the risk of embedding foreign material. All Icebox welding was done with electron
beam process so that no filler metal was needed. Following fabrication, the copper
assemblies were cleaned in an acid bath to reduce radon daughters and other surface
contaminants acquired during fabrication, rinsed with water and wrapped for
cleanliness. Gold plating over nickel flash will be applied where needed for thermal
emissivity. Following cleaning all handling was done with powder free clean room
gloves. One enhancement that could have improved thermal conductance of the
joints would have been gold plating of all joint surfaces.
Tapped holes were made with rolling taps. This reduces the risk of embedded
tool particles from cutting taps. It has the additional benefit of work hardening the
threads, increasing their strength.
The conductive tubes and rods in the C-stem were annealed to achieve high
thermal conductivity. The annealing was held at 410 °C in vacuum for one hour
and slowly cooled. This process followed a procedure described by Fickett [6] to
achieve a high RRR in oxygen free copper (see Fig. 8.7).
228 R.L. Schmitt

Fig. 8.7 Anneal temperature

for maximum RRR [6]

8.6.4 Underground Assembly

All parts were again wiped with alcohol before introduction into the clean room. All
handling was done with using cleanroom gloves, lint free cleanroom wipes, cleaned
fixtures and tools. The nesting of the cans, stems and dilution refrigerator attach-
ment required a specific assembly sequence. Icebox cans and the associated C-stem
and E-stem were assembled from the outside in. The OVC, outer vacuum container,
was mounted and leveled on the lead shielding. The OVC C-stem and E-stem were
attached, temporary covers applied and the joints were leak tested. All seals on the
OVC were made with Viton O-rings.
The gold plated joints were cleaned with alcohol wipes. All bare copper joints
were prepared for assembly by scrubbing the stem connections with counted Scotch
BriteTM and cleaning with alcohol wipes. This removed surface corrosion imme-
diately before assembly and the wipes removed any particles from the cleaning. The
aramid loops were attached to the can and a lifting fixture was used to lower the LN
can into the OVC can. The C-stem and E-stem were inserted into the OVC stems
and connected from the inside. All screws were tightened following a pattern using
a torque screwdriver. All screws were assembled dry and tightened to 90 % of their
breaking torque. Tests had shown that the breaking torque was repeatable for both
brass and stainless machine screws. And even stainless screws would break before
damaging the rolled threads. Stainless screws were only used outside the lead and
polyethylene shielding.
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 229

8.7 Thermal Model

Thermal modeling of detector and cryostat elements is crucial, but with the paucity
of consistent low temperature data and joint conductance information can be a
challenge. Material properties above 4 K are readily available from NIST
Cryogenic Properties Database (NIST) and the older Brookhaven National
Laboratory selected cryogenic data notebook [9].

8.7.1 Thermal Conductivity

Conduction through solids below 4 K requires data from various sources.

For CDMS the two materials of consequence were Kevlar (aramid fiber) and
copper. Plotting test results from several reports shown in Fig. 8.8, the NIST
(NIST) and Ventura [26] data appeared to be consistent over a wide temperature
range. And with higher conductivity than other reports they are the appropriate
choices for conservative design.
Copper thermal conductivity at low temperature is highly dependent on purity,
heat treatment and operating temperature. Extremely high thermal conductivity can
be obtained with 5 or 6 nines copper treated to remove oxygen and annealed.
However this is not practical for the large conduction tubes needed for a cryostat
such as the CDMS Icebox.
When large conductors were needed, the highest purity commercial alloy was used.
C101, aka oxygen free copper, is readily available in industrial sized sheets, tubes and
bars. Annealed copper following the procedure recommended by Fickett can achieve
reasonably high conductivity (see Fig. 8.7) [6]. Risegari reported test data for
annealed oxygen free copper between 30 and 150 mK. Risegari [19] Extrapolating the
NIST data for RRR = 150 below 4 K, assuming a linear relationship with temp
between 5 and 10 K does not line up very well with the Risegari equation.

Fig. 8.8 Low temperature Kevlar Thermal Conductivity

thermal conductivity of 10
aramid fibers (NIST) [5, 8, 12,
14, 26] 1

0.1 NIST
W/m-K

Duband
0.01 Zhang
Hust
0.001 Ventura

0.0001

0.00001
0.01 1 100
Kelvin
230 R.L. Schmitt

8.7.2 Joint Conductance

Bolted joints are required for assembly of most cryostats but can add a significant
conductance resistance in the millikelvin range. Copper surfaces readily corrode,
and even a small amount of corrosion can impose a significant decrease in joint
conductance. For CDMS the major can and stem connections were rigorously
scrubbed and cleaned immediately before assembly and performed reasonably well.
But it early operation it was clear that joints assembled with less rigor performed
poorly.
For SuperCDMS a study of over twenty-five publications regarding joint thermal
conductance was performed. Unfortunately, a large majority of the results were not
directly applicable to the SuperCDMS experiment because:
• Joint conductance results relied on electrical resistivity measurements and the
use of the Wiedemann-Franz Law, whose use across joints was questioned in
Nilles [13] and Didschuns [4]
• Joint clamping force was unknown
• Use of foreign materials such as grease between the joint surfaces. Use of
interfacial materials is to be avoided in SuperCDMS to maintain experimental
radiopurity and general cleanliness.
Upon review of the pertinent publications; Nilles [13], Kittel [10], Didschuns
[4], and Woodcraft [29], a design basis was chosen. Parameters for surface finish,
plating and clamping force were chosen that could be readily achieved in a project
with dozens of conductive joints and hundreds of bolts. Subsequent tests from
60 mK to 14 K in three refrigerators confirmed the validity of this design equation
(Eq. 8.1) [24]. The design basis and test results are shown in Fig. 8.9.

Fig. 8.9 Gold plated copper joint conductance, design basis and test results [24]
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 231

K ¼ 0:0624 T  0:00023 ð8:1Þ

where:
K Joint Conductance [W/K]
T Temperature [K]

8.8 Heat Load

Heat load calculations between stages were dominated by the detector tower
structure, signal wiring and wiring insulation. The thermal conductivity varies
widely as a function of temperature and a list of relevant papers for design below
1 K may be most useful for the readers
Nb-Ti, [15]
SS316, [1]
Kapton HN, [2]
Kapton HN, [20]
Graphite CFRP, [21]
Ti15333, [27]
POCO AXM-5Q, [28]
Manganin 4 % NI, [16]
Manganin 4 % Ni, [25].

8.9 Detector Signal Feedthrough

Vacuum feedthroughs are readily available for a variety of standard connector

types. But when hundreds or thousands of signal wires must be accommodated the
cost and space requirements of standard commercial feedthroughs can become
prohibitive. A variety of designs have been used. CDMS constructed a feedthrough
box from an ISO 400 six way cross. Two sides of the cross were fitted with vacuum
pumping or cryostat connections. Eleven D50 connectors were welded into blank
ISO 400 flanges on the other four sides. Strain relief was provided by brackets on
the outside of the flanges.
Another approach is to press a printed circuit board between vacuum flanges.
Connectors in the middle of the flange face inward. The signals traces can be single
or multilayer and extend radially through the seal area of the board. Connectors
outside the sealing surface face outward. Choice of seal type and number of board
layers is mainly driven by space considerations.
232 R.L. Schmitt

8.10 Dilution Refrigerator

CDMS at Soudan used a wet (consumes liquid nitrogen and liquid helium cryo-
gens) dilution fridge for cooling at all stages. Later modifications added cryocoolers
for heat interception in the E-stem, nitrogen and helium re-condensation.
After startup difficulties were resolved, the dilution refrigerator was very reliable,
with multiple runs over one year duration. The longest run was nineteen months,
and these later runs were terminated for scheduling reasons, not fridge problems.
Several enhancements improved the reliability of the dilution refrigerator.
Larger, separate liquid nitrogen reservoirs replaced the original, single reservoir
for the 80 K cold traps. This allowed temperature and level measurements to be
added for automatic nitrogen filling. The original 80 K cold traps were used for the
entire project.
The circulation pump with mechanical shaft seals was replaced with a magnetic
drive pump, eliminating possible shaft seal leaks and seal maintenance. The pump
oil was replaced with Fomblin SV for low oil vapor carry over. After draining and
filling with the replacement oil, it was again drained and heated in a beaker, from
which the original oil was skimmed off. The reliability of this pump has been
excellent.
The dilution refrigerator was designed with a cold trap inserted in the bath.
A second, external 4 K cold trap was added upstream and in series with the fridge
internal cold trap. The external cold trap resided in a separate liquid helium bath.
This external cold trap was initially filled manually from liquid helium supply
dewars, but soon was equipped with a cryocooler helium liquefier, which main-
tained the liquid helium without further transfers. Very small amounts of helium
leakage were made up with gas from high purity gas cylinders.
Regular, monthly cold trap regeneration was the key to long term reliable
operation.
Most new experiments utilizing dilution refrigerators choose ‘dry’ models,
which use a cryocooler(s) rather than liquids nitrogen and helium to provide the
upper stage cooling. The dry fridges avoid the cost of liquid helium, the effort of
dewar handling and liquid transfers. The latter is especially valuable underground
or where access is limited.

8.11 Liquid Transfer Systems

The dilution refrigerator at Soudan was a wet dilution refrigerator of a type typically
installed at the time. With the heat loads from the Icebox, it required daily transfers
of liquids nitrogen and helium. Common practice for many refrigerators at that time
was to manually transfer cryogenic liquids. But the Soudan underground location
with limited access required an automated solution. It was also important to use
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 233

liquid cryogens efficiently to minimize cost and effort to transport dewars

underground.
Dual 160 L nitrogen supply dewars were connected to a vacuum insulated
manifold outside the RF room. During transfers, liquid was drawn from the lead
supply dewar. If it ran dry during the course of a transfer the backup dewar would
complete the transfer. The backup would then become the lead for the next transfer,
allowing time for the empty to be replaced. Inventory was tracked by scales located
under each dewar.
160 L nitrogen dewars normally have pressure building regulators and vent
valves. But those regulators often provide imprecise control and/or leak through
wasting liquid nitrogen. More precise pressure control contributes to faster transfers
without risking an overpressure in the fridge. When the liquid withdrawal line was
connected to a dewar, a gas management line was also connected. The gas man-
agement line was equipped with pressure measurement, gas makeup and vent
valves. During standby the PLC operated solenoid valve vented nitrogen if the
pressure was too high. During transfers pressure was maintained with another
solenoid valve supplied from high pressure cylinders.
When a transfer was initiated, a gas line pressurized the supply dewar and a cool
down valve opened. When the transfer line was cooled down, the valve(s) filling the
dilution refrigerator or 80 K cold traps opened and the cool down valve closed.
A valve sequence after filling was used to ensure transfer efficiency and prevent
trapped volumes in the piping.
Dual 350 L liquid helium dewars were connected to a vacuum insulated man-
ifold outside the RF room. This was the largest helium dewar size that would fit on
the mine shaft elevator. Dewar selection, pressurization, inventory and transfer
procedures were similar to the nitrogen system. The helium transfer valves were
close to the dilution refrigerator inside the RF room and therefore were pneumat-
ically controlled from outside.
To minimize detector down time the helium and nitrogen transfers could be
automatically synchronized. Transfers could be initiated by time-of-day, by low
nitrogen or helium level in the fridge, or manually.

8.12 Liquefier Addition

The initial operation of the dilution refrigerator required daily transfers of both
liquids. During transfers the physics data taking was stopped, losing an hour each
day of experiment run time. Transporting dewars underground was typically done
weekly, taking several hours with the elevator doors removed, technicians at the top
and the bottom of the shaft. Normal deliveries to the remote site were weekly
although special shipments could be made.
As the commercial development of cryocoolers progressed it became feasible to
add liquefiers to the system. For this project all cryogenic equipment had to be
installed outside the RF room, complicating the installation.
234 R.L. Schmitt

Fig. 8.10 Low heat load transfer line between liquefiers and dilution refrigerator [23]

A transfer line was needed between the top of the dilution refrigerator to a
bayonet box outside the RF room. The RF wall penetration, the thermo siphon
design and the low heat load requirements prevented the use of a U-tube separation.
Instead the transfer line had an S shape from the liquefier connection to the fridge.
The line included an actively cooled 78 K heat shield, which extended into the male
bayonet into the fridge. Assembly was made at the midpoint with a retractable
bellows and VCR connections. A spare port on top of the fridge was modified to
accept a custom male bayonet. The transfer line total length was 100 in. The
transfer line is shown in Fig. 8.10.
The nitrogen liquefier was a GM type cryocooler was fitted with a
desuperheating/condensing heat exchanger and installed in a custom vacuum ves-
sel. An electric heater embedded in the heat exchanger was used to remove excess
refrigeration capacity. The cryocooler vacuum vessel was mounted on rubber iso-
lators to reduce the vibration forces carried into the support structure. During
holding operation the heater was controlled by the nitrogen reservoir pressure.
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 235

Small losses were replenished by automatically adding gas from the liquid nitrogen
supply dewars. The nitrogen liquefier system was very reliable, typically operating
with about 120 W of excess capacity.
The dilution refrigerator was designed to use the helium boil off gas to intercept
heat through the bath neck. Superheating helium gas also provided some cooling
the nitrogen reservoir. Consequently adding a helium condenser would not have
been feasible. The steady state heat load from the Icebox and dilution refrigerator
consumed about 20 liquid liters/day. This was marginally within the capacity of two
Cryomech PT-415 liquefiers, but with the transfer line heat load they were not quite
enough capacity. With a third liquefier the cooling capacity was more than enough
for the system. Electric heaters in the liquefiers were used to compensate for the
excess cooling power. Heater power control was based on helium bath pressure.
Liquid helium from the refrigerator bath fed the 1 K pot. Helium from the pot
was not recovered, so that loss plus leakage required regular makeup. Makeup gas
was taken from helium supply dewar boil off. Although this system could run for
months at a time, eventually trace amounts of contaminants would build up in the
transfer line and stop the thermo siphon. Recovery involved disassembly of the
transfer line, removal from the dilution fridge and a warmup. This could normally
be done in one day without disturbing the fridge base temperature operation.
A description of the liquefier system was published in Advances in Cryogenic
Engineering [23]. The source of contamination was not known, but an easier means
of removing and warming up the transfer line would have been an improvement.

8.13 External Cold Trap

Small amounts of contaminants in the mix would gradually plug the internal cold
trap. To improve the operating reliability a 4 K external cold trap was installed
downstream of the 80 K traps. This trap dewar was initially filled with liquid
helium transfers, but was fitted with a helium liquefier. The liquefier could handle
the total heat load and slowly fill the cold trap dewar from high pressure gas
cylinders. The liquefier heater was controlled by dewar pressure, which could also
add or vent helium gas when required. This system was very reliable.

8.14 E-Stem Cryocooler

During initial operation with fewer detectors it became apparent that the E-stem
conduction and heat sinks would not be adequate for more detectors. A two stage
SHI cryocooler was fitted to the E-stem to intercept signal cable heat. To minimize
vibration it was independently suspended and equipped with vibration isolation.
Extra thin, edge welded bellows isolated the vacuum shell from the ICEBOX.
Braided cables isolated the first stage and bundled fine wires isolated the second
236 R.L. Schmitt

stage. Even with this support and isolation system noise from the GM cryocooler
was troublesome for the detectors.

8.15 Insulating Vacuum

The insulating vacuum inside the system must be good to reduce the heat load, and
it is very important to limit the helium that could enter the IVC through leaks or
permeation. Dry roughing and turbo molecular pumps were used to evacuate the
ICEBOX. They were equipped with automatic valves that close in case of power
loss or pump failure.
Standard dilution refrigerators have evacuation tubes sized for typical small
applications. These tubes are not large enough to reasonably evacuate the larger
volumes in the ICEBOX. To improve the pumping speed for the OVC an adaptor
and custom radiation baffle was added to the bottom of the fridge tails. It was the
equivalent of a 3 in. diameter pumping line and had no noticeable effect on the
helium bath heat load. Pumping lines were added to the signal feedthrough box for
the IVC.

8.16 Automation and Control

The cryogenic system was monitored and controlled with an industrial PLC (pro-
grammable logic controller) and HMI (human machine interface). They were both
current industry products during installation and most of the operating period. The
automation and control system was backup up by UPS (uninterruptable power
supply) and standby generator.
This system provided local and remote monitoring, local and remote control,
historical cryogenic data collection and fully automated operation. Automatic
cryogen transfers, recovery from power outages, liquefier heater control and so on
were fully automated.
The millikelvin systems at Soudan did not have temperature control, generally
running as cold as possible. Helium and nitrogen liquefiers on systems on the other
hand must be controlled to avoid subatmospheric pressure and freezing the nitro-
gen. Resistance heaters in the liquefiers, with zero crossing SCR’s were controlled
using conventional PID loops with cryostat pressure as the process variables.
Pressure measurement is sensitive and small measurement errors will not cause the
system to operate below atmospheric pressure.
Much more complicated are the sequences, abnormal procedures and alarms.
Some examples from Soudan included conditional alarms, automatic set point
adjustment, restart sequences after power outages, backup cooling water start,
cryogen transfer sequences, etc. A modern industrial PLC with IEC 61131-3
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 237

programming languages can be configured to carry out these automation tasks in

graphical form, without resorting to complicated scripts.

8.17 Cryogenic Operation

The remote location made commissioning and early operation more difficult.
During the early years there were several operational problems that required the
fridge to be removed from the ICEBOX. Once settled, however the fridge operation
became very reliable. There were multiple runs at 56 mK over a year in length with
the longest at 19 months. These runs were usually stopped for reasons unrelated to
the dilution refrigerator. The dilution refrigerator and ICEBOX were in service for
twelve years.

8.18 Lessons Learned

8.18.1 Cryogenic System Assembly and Testing

After successful shop tests there was an assumption that after shipping and
reassembly the dilution refrigerator would operate successfully. But several cool
down attempts were required to commission the fridge and support equipment,
repair leaks and damage, etc. Repairs were made more difficult by the underground,
remote location. Future experiments in remote locations should consider a full
operational test of all cryogenic systems at a convenient location.

8.18.2 Wiring

Thermometer and touch sensor wiring was routed through the dilution refrigerator,
which was equipped with cryogenic rated micro D connectors and heat sinks at
each stage. While this was thermally and functionally successful, routing these
cables along the narrow C-stem was difficult. A better solution would be to include
these signal wires along with the detector signal cables.

8.18.3 Mixture Purification

The standard fridge came equipped with a 4 K cold trap immersed in the helium
bath. Anything that gets past this trap can cause plugging at the condenser
238 R.L. Schmitt

impedance and requires a subsequent warmup. The installation of an additional 4 K

cold trap reduces the risk of trap plugging during long runs.

8.18.4 Micro Vibrations

The initial installation did not utilize cryocoolers, but their installation caused noise
pickup in the experiment detectors. Even with typical mechanical isolation methods
the transmitted noise was significant. It is thought that when excited, the SNOBOX
cans vibrated at their natural frequencies, then transmitted the vibration through the
suspension to the adjacent inner cans. Transmissibility could have been reduced by
installing springs on the hanger supports, especially at higher frequencies. Future
projects should consider transmissibility of impulse from outside the cryostat or
even the negative spring constant concept.

8.18.5 Can Supports

The aramid fiber loops effectively supported the cans with low heat load. However
they were constructed of sixteen loops each. The inelasticity of Kevlar string
required the loops to be identical in length and consequently a more difficult
assembly procedure. More recent projects have used a single loop made of a larger
braided rope, simplifying assembly.

8.18.6 Inner Vacuum, Yes or No?

CDMS II at Soudan has a sealed Inner Vacuum Chamber (IVC), a separate vacuum
space from the Outer Vacuum system. During cool down, above 10 K, this IVC
space is filled with a small amount of helium exchange gas to thermally link the
inner cans to the IVC. While successful, this approach requires a leak-tight chamber
at 5 K and the ability to evacuate helium gas at 10 K before proceeding below
10 K. For CDMS the cooldown was accomplished with several steps: Introduce
liquid nitrogen into the helium bath and nitrogen shield reservoir, purge all liquid
and evacuate the nitrogen from the helium bath, slowly introduce liquid helium,
evacuate the exchange gas. Then start the dilution refrigerator.
Since a dry fridge does not have liquid nitrogen or helium reservoirs, and since
the typical two stage cryocooler does not have enough capacity to reasonably cool
down a large cryostat, it is more appropriate to dispense with the IVC and use
cooling tubes fixed to each stage for cool down. For a system of this type heat
exchangers attached to each stage would be linked in series using stainless steel
bellows. Helium gas coolant is circulated through these heat exchangers and out to
8 Design and Operation of a Large, Low Background, 50 mK Cryostat … 239

cryocoolers. This allows a large single stage cryocooler to start the cooldown at
room temperature, and a two stage cryocooler can take over below 40 K. Once the
layers are cold the helium gas will be evacuated from this cool down circuit. The
long conduction path provided by the bellows shape and the poor thermal con-
ductivity of stainless steel limits the heat load between layers to an acceptable value
once the helium is evacuated form this circuit.

References

1. M. Barruci, Measurement of thermal conductivity of the supports of CUORE cryostat.

Cryogenics 48, 166–168 (2008)
2. M. Barucci, Low temperature thermal conductivity of Kapton and Upilex. Cryogenics 40,
145–147 (2000)
3. CDMS at Fermilab (2003). Retrieved from http://ppd.fnal.gov/experiments/cdms/. http://titus.
stanford.edu/public/photos/cdms_icebox.jpg
4. I. Didschuns, Thermal conductance measurements of bolted copper to copper joints at
sub-Kelvin temperatures. Cryogenics 44, 293–299 (1992)
5. L. Duband, Thermal isolation of large loads at low temperature using Kevlar rope. Cryogenics,
643–647 (1993)
6. F. Fickett, Oxygen free copper at 4 K: resistance and magnetoresistance. IEEE Trans. Magn.
19(3) (1993)
7. G. Frossati, Experimental techniques: methods for cooling below 300 mK. J. Low
Temp. Phys. 87(3/4) (1992)
8. J. Hust, Low-temperature thermal conductivity of two fibre-epoxy composites. Cryogenics,
126–128 (1975)
9. J. Jensen, Brookhaven National Laboratory Selected Cryogenic Data Notebook (1980)
10. P. Kittel, Thermal conductance of gold plated metallic contacts at liquid helium temperatures.
Adv. Cryog. Eng. 37, 241–248 (1992)
11. LBL. (1993). IVC Seal Design
12. E. Marquardt, Cryogenic Material Properties Database. 11th International Cryocooler
Conference (Keystone, Co, 2000)
13. M. Nilles, Effects of oxidation and roughness on Cu contact resistance from 4 to 290 K. Adv.
Cryog. Eng. 34 (1987)
14. NIST (n.d.), NIST Materials Measurement Lab, Cryogenic Technologies Group, Material
Properties. Retrieved from http://cryogenics.nist.gov/MPropsMAY/materialproperties.htm
15. J. Olson, Thermal conductivity of some common cryostat materials between 0.05 and 2 K.
Cryogenics 33 (1993)
16. I. Peroni, Thermal conductivity of manganin below 1 K. Nucl. Phys. B 78, 573–575 (1999)
17. D. Bauer, PAC Meeting, March 30, (2007). Retrieved from https://www.fnal.gov/directorate/
program_planning/March2007PACPublic/BauerPAC03_07.pdf
18. M.D. Resources, Lake Vermilion-Soudan Underground Mine State Park (2015). Retrieved
from http://www.dnr.state.mn.us/state_parks/lake_vermilion_soudan/index.html
19. V. Risegari, The Art of Cryogenics: Low Temperature Experimental Techniques (Elsevier,
Oxford, UK, 2008)
20. D. Rule, Thermal conductivity of polypyromellitimide film with alumina filler from 4.2 to
300 K. Cryogenics 36, 283–290 (1996)
21. M. Runyan, Thermal conductivity of thermally-isolating polymeric and composite structural
support materials between 0.3 and 4 K. Cryogenics 48, 448–454 (2008)
22. J. Sander, Fermilab Today (2015). Retrieved from http://www.fnal.gov/pub/today/archive/
archive_2015/today15-02-06.html
240 R.L. Schmitt

23. R. Schmitt, Application of cryocoolers to a vintage dilution refrigerator. Adv. Cryog. Eng.
(2011)
24. R. Schmitt, Thermal conductance measurements of bolted copper joints for SuperCDMS.
Cryogenics (2015)
25. Y. Touloukian, Thermophysical Properties of Matter: Thermal Conductivity, Metallic
Elements and Alloys, vol. 1 (IFI/Plenum, New York, 1970)
26. G. Ventura, Low temperature conductivity of Kevlar. Crygenics, 489–491 (2000)
27. P. Wikus, The electrical resistivity and thermal conductivity of Ti 15V–3Cr–3Sn–3Al at
cryogenic temperatures. Cryogenics 51, 41–44 (2011)
28. A. Woodcraft, Thermal conductivity measurements of pitch-bonded graphite at millikelvin
temperatures. Cryogenics 49, 159–164 (2009a)
29. A. Woodcraft, Thermal design and performance of the SCUBA-2 instrument 1-K and mK
systems. Cryogenics 49, 504–513 (2009b)
Chapter 9
Cryogenic Transfer Lines

Jaroslaw Fydrych

Abstract Transfer lines are common in cryogenic systems and are a form of
cryostat. This chapter describes the requirements of transfer lines, surveys existing
transfer lines and discusses issues such as modularization, routing, supports, ther-
mal contraction, piping arrangement, materials, manufacturing and installation. The
chapter concludes with a detailed case study of the design, manufacturing and
performance of the XFEL/AMTF transfer line.

9.1 Introduction

Cryogenic transfer lines are typical components of almost all cryogenic systems.
They are intended for transferring cryogenic fluids between two cryogenic devices
[1]. Since the value of the cryogenic fluids is essentially in the thermodynamic
states of their molecules, the transferring should not cause significant changes in the
thermodynamic states of the transferred cryogens. It means that either the tem-
perature increase or, in case of liquids, the vapor quality and also pressure changes
should be negligibly small.
The simplest cryogenic transfer line is a vacuum jacketed pipe connecting two
nitrogen dewars as shown in Fig. 9.1. In this example, the line is used for transferring
liquid nitrogen (LIN). If the distance between the two dewars is short, the flowing
nitrogen stays in the line for a short period of time. Then, the flowing nitrogen
absorbs little heat and only a tiny portion of the nitrogen evaporates during the flow
and the rest keeps its initial thermodynamic state. The relatively high value of the
latent heat of nitrogen is obviously an advantage. However, if the line is very long the
nitrogen stays in the line much longer. Then the flowing nitrogen absorbs much more
heat. As a result, the vapor fraction of the nitrogen reaching dewar 2 is much larger.
The inflowing nitrogen vapor obviously does not stay in the dewar but flows out; it
counts as losses. The higher hydraulic resistance of the longer process line leads to a

J. Fydrych (&)
European Spallation Source ERIC, P.O. Box 176, 22100 Lund, Sweden
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 241

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_9
242 J. Fydrych

Fig. 9.1 A simple pipe-in-pipe cryogenic transfer line connecting two dewars

higher difference between pressure p2 and p1, which increases the difference between
the initial and final thermodynamic states in the transferred nitrogen.
The pressure, temperature and density of the flowing cryogen have certain
impacts on the mechanical design of the process line, vacuum jacket envelope and
their supports. The higher the pressure and density, the thicker the pipe walls and
the stronger the supports.
This simple example of a pipe-in-pipe cryogenic transfer line shows that
designers of these lines have to take into consideration a number of parameters. The
most important are:
– geometrical restrictions (total distance, possible routings and available space),
– required mass flow rate and its time characteristic,
– minimum and maximum temperature,
– minimum and maximum pressure,
– thermodynamic properties of chosen cryogenic fluid,
– mechanical properties of chosen pipe materials.
The design of pipe-in-pipe cryogenic transfer lines is extensively described in
[2–4]. All the above considerations and a number of design guidelines provided in
[4] are also valid in the process of designing much more complicated cryogenic
transfer lines, in which the vacuum jacket houses several process lines transferring
cryogenic fluids at different temperatures and pressures. Such multichannel lines are
usually used for transferring cooling power between a cryogenic plant and cryo-
genic users in so-called large scientific facilities. Typical cryogenic users are cry-
omodules with superconducting cavities and cryostats with superconducting
magnets, bus bars or cryogenic vacuum pumps. These facilities usually require
complex cryogenic systems that are capable of providing very high cooling power
at very low temperature levels and distribute this cooling power among the users.
Figure 9.2 shows a simplified schematic flow scheme of such a complex cryogenic
system. Here, a cryogenic device, let’s say a chain of cryomodules housing
superconducting cavities made of pure niobium, needs to be cooled down to a
temperature of 2 K. The required cooling power is produced by a cryogenic plant
9 Cryogenic Transfer Lines 243

Fig. 9.2 Schematic of a complex cryogenic system

and transferred by means of a constant flow of cold helium to the device via the
helium supply line. In this example, the cryogenic plant supplies supercritical
helium at a temperature and pressure of 4.5 K and 3 bar absolute, respectively.
Then the 2 K helium is produced at the cryogenic device, where the supercritical
helium is first precooled to 2.2 K in a counterflow heat exchanger (HX) and then
throttled in a Joule-Thomson valve down to a pressure of 31.3 mbar absolute.
This solution takes advantage of a significantly low value of the helium critical
pressure (2.25 bar absolute). Transferring helium at higher pressure eliminates all
the problems related to two-phase flow phenomena. Usually a pressure around
3 bar absolute is perfectly adequate. Then, the cryoplant compressors have to
compress helium to a pressure not higher then 20 bar absolute, which is not too
complicated from technical point of view.
Since helium is a very expensive cryogen it must be recovered and recirculated
to the cryogenic system. Therefore in the above example the helium vapor is
recovered from the cryogenic device and transferred back to the cryogenic plant via
the vapor low-pressure line (VLP). The vapor flow is driven by a vacuum pump at
the cryogenic plant, which generates the subatmospheric pressure required for
reaching 2 K in the helium vessel inside the cryogenic device. In the case of very
large systems, which require significantly high cold helium flows, vacuum pumps
are replaced by sets of cold and warm compressors.
All the cold elements of the cryogenic device are surrounded by a thermal shield.
The shield is actively cooled by an additional helium circuit at a temperature of
40–60 K. The thermal shield circuit is composed of the TS supply and return lines
connected to the high pressure helium line in the cryoplant cold box, downstream
the first set of its heat exchangers.
In the above example the cryogenic transfer line is one of the main components
of the cryogenic system. Apart from all the four process lines (He supply line, VLP
return line, TS supply line and TS return line) it includes also thermal shields,
vacuum jackets, vacuum barriers, supporting structures, thermal contraction
244 J. Fydrych

Fig. 9.3 Schematic illustration of a multichannel cryogenic line

compensation elements, pressure safety devices, vacuum pumping ports and some
instrumentation (temperature sensors and pressure transducers).
The cryogenic transfer lines are almost always manufactured and preassembled
at manufacturer sites that are located far away from their final locations. Therefore,
the lines have a modular structure that facilitates production, transportation and
installation, as well as some repair work, if needed. This strongly affects design
choices such as a number and lengths of sections (modules) and interconnection
arrangements. Figure 9.3 shows a schematic illustration of a straight section of a
multichannel cryogenic transfer line.
This section is composed of two straight modules. Each of them consists of process
pipes (1), thermal shields (2) and an external envelope (3). The two sections of the
external envelope are connected by an interconnecting sleeve (4). The cryoline sec-
tion ends with vacuum barriers (5) that separate the cryoline insulation vacuum from
those of the adjacent vacuum sectors. All forces resulting from the dead weights,
pressure loads and thermal contractions of the process lines are transferred to the
vacuum jacket via a process line fixed support (6), sliding supports (7) and the vacuum
barriers as well, which also act as fixed supports. The external envelope transfers these
forces to conventional facility structures (anchors or foundations at the building floor
and walls) via the fixed supports (8) and sliding supports (9) of the external envelope.
Both the process pipes and the external envelope are equipped with bellows.
Internal bellows (10) compensate the thermal shrinkage of process pipes, whilst the
external bellows (11) are to cope with all the thermal expansions and contractions
of the vacuum jacket. The external envelope is also equipped with a vacuum
pumping port (12) and pressure relief device (13).
The most well known large scientific facilities that use multichannel cryogenic
lines are listed in Table 9.1. In almost all these facilities, the cryogenic transfer lines
are part of cryogenic distribution systems, which in addition to cryogenic transfer
lines include a number of distribution boxes. For example the LHC cryogenic
distributions system, which is the largest system in the world, is composed of eight
separated cryogenic distribution lines. Each distribution line is 3.1–3.3 km long and
consists of a number of cryogenic transfer line sections that connect 38 service
modules (distribution boxes) [5, 6]. Another example is the RHIC cryogenic dis-
tribution system that includes 12 valve boxes and 25 sections of multichannel
cryogenic transfer lines [7].
Table 9.1 Examples of large scientific facilities using multichannel cryogenic transfer lines
No. Scientific Institution (location) Machine type Cryogenic devices Cryogenic transfer line
facility (operating temperature) Total Cold Temperatures (cryogen)
length line no.
1 ESSa European Spallation Proton linac and RF cavities (2 K) 450 m 4 4–53 K (helium)
Source ERIC (Lund, neutron source
Sweden)
2 FAIR/SIS100a GSI (Darmstadt, Heavy-ion storage Magnets (4.3 K) 900 m 4 or 5 4.3–100 K (helium)
9 Cryogenic Transfer Lines

Germany) ring
3 FAIR/ GSI (Darmstadt, Fragment separator Magnets (4.3 K) 500 m 4 4.5–100 K (helium)
SuperFRSa Germany)
4 FLASH DESY (Hamburg, Electron linac with RF cavities (2 K) 200 m 4 4.5–80 K (helium)
Germany) free electron laser
5 FRIBa MSU (Michigan, USA) Proton linac and RF cavities (2 K) 500 m 5 4–55 K (helium)
fragment separator
6 HERA DESY (Hamburg, Hadron-electron Magnets and RF cavities 6.3 km 4 3.7–80 K (helium and
Germany) storage ring and (4 K) nitrogen)
collider
7 ISR CERN (Genewa, Hadron storage ring Magnets (4.2 K) 400 m 2 4.2–100 K (helium)
Switzerland) and collider
8 ITERa ITER Organization Tokamak Tokamak magnets (4.3K), 3.5 km 4–6 4.3–100 K (helium)
(Cadarache, France) divertor cryopumps (4.5
K)
9 JT60SAa JAEA (Naka, Japan) Tokamak Tokamak magnets (4.4K), 100 m 5 or 6 4.4–100 K (helium)
divertor cryopumps (3.7
K)
10 KEKB KEK (Tsukuba, Japan) Electron-positron Crab cavities (2 K) 1.1 km 4 4.5–80 K (helium and
storage ring and nitrogen)
collider
(continued)
245
Table 9.1 (continued)
246

No. Scientific Institution (location) Machine type Cryogenic devices Cryogenic transfer line
facility (operating temperature) Total Cold Temperatures (cryogen)
length line no.
11 KATRIN TLK (Karlsruhe, Tritium-neutrino Magnets (4.5 K) 40 m 6 4.5–117 K (helium and
Germany) experiment nitrogen)
12 LHC CERN (Genewa, Hadron storage ring Magnets (1.9 K) 26 km 4 and 5 1.8–80 K (helium)
Switzerland) and collider
13 RHIC BNL (Brookhaven, Heavy-ion storage Magnets (4.6 K) 5.2 km 5 4.6–70 K (helium)
USA) ring and collider
14 SNS ORNL (Oak Ridge, Proton linac and RF cavities (2.1 K) 600 m 2 2.1–50 K (helium)
USA) neutron source
15 Tevatron FERMILAB (Batavia, Proton storage ring Magnets (5 K) 6.7 km 2 4.6–80 K (helium and
USA) and collider nitrogen)
16 TORE SUPRA CEA (Cadarache, Tokamak Tokamak magnets (1.8 K) 100 m 8 1.7 K–80 K (helium and
France) nitrogen)
17 TRISTAN KEK (Tsukuba, Japan) Electron-positron RF cavities and magnets 330 m 4 4.5–80 K (helium and
storage ring and (4.2 K) nitrogen)
collider
18 XFEL/AMTF DESY (Hamburg, Cryomodule test RF cavities (2 K) 170 m 4 4.5–80 K (helium)
Germany) stand
19 XFEL/LINACa DESY (Hamburg, Electron linac with RF cavities (2 K) 340 m 6 and 7 2–80 K (helium)
Germany) free electron laser
a
under design or construction in Feb. 2016
J. Fydrych
9 Cryogenic Transfer Lines 247

9.2 Cryoline Routing and Modularization

The routing of a cryogenic transfer line defines its detailed location in the scientific
facility site. In order to minimize both capital and operational costs the routing
should be simple and the total length of the line should be as short as possible.
However, in a great extent it is the geometrical features of the site infrastructure that
determines space available for the line. An example of the most complicated cry-
oline routings is the ITER cryogenic distribution system [8, 9], which 3 D model
representation is shown in Fig. 1.25 (Chap. 1). This system is to be located around
the machine in the tokamak building.
Figure 9.4 shows a 3D model of the ESS cryogenic distribution line which runs
from the cold box of the ESS accelerator cryogenic plant to the ESS linac tunnel in
an underground gallery [10]. Since the gallery has chicanes, which are required for
limiting the propagation of radiation from the tunnel to the cold box building, the
transfer line is significantly elongated and has additional elbows.
Due to a complex internal design requiring a lot of precise assembly works the
cryogenic transfer lines are usually produced in manufacturer workshops, which are
located far away from the scientific facilities. In order to facilitate production,
transportation and installation of cryogenic transfer lines, the lines are designed to be
composed of a defined number of modules [11]. Each module is a short section of the
cryoline that is connected to its adjacent sections via special interconnections.
Producing and assembling cryoline modules at manufacturer sites reduces produc-
tion costs and helps to meet high standards of production works. The prefabricated
modules are transported to the site and linked together with the interconnections. The
lengths and shapes of modules depend mainly on cryoline routings, transportation
method and fixed support locations. Usually the lengths of modules follow the single

Fig. 9.4 Schematic view of the ESS linac cryogenic transfer line
248 J. Fydrych

Straight module

Elbow module

Zed module

Tee module

Angular module

Fig. 9.5 Typical cryoline modules

or double random lengths of pipes (the most commonly stocked lengths) and are not
longer than the length of a typical semi-trailer (ca. 13 m).
Figure 9.5 shows some typical cryoline modules. The most frequently used are
straight and elbow modules. The Zed module is used where the cryoline routing has
a significantly small step, while Tee module is dedicated for distributing cryogenic
fluids to two cryogenic devices (or two groups of devices).
9 Cryogenic Transfer Lines 249

Fig. 9.6 Typical design of an interconnection between two cryoline modules

During the installation of a cryogenic transfer line at its final location the cry-
oline modules are connected to each other with so-called cryoline interconnections.
Figure 9.6 shows an example design of a cryoline interconnection.
The external envelope sections of two adjacent cryoline modules are ended with
a welding ring that the interconnection sleeve is welded to. The sleeve is made of a
pipe section which is slightly higher in size than the cryoline module external
envelope. Such a sleeve can be moved aside on one of the modules in order to get
an access to the internal parts of the interconnection. Some of interconnection
sleeves are equipped with bellows allowing for required axial or lateral displace-
ments. In this case, the internal process lines should also be equipped with axial or
lateral compensators.

9.3 Cryoline Cross-Section Arrangements

The arrangement of process lines in a cryoline cross section can impact the
thermo-mechanical behavior and thermal feature of the cryogenic transfer line.
Therefore the detailed locations of process lines and thermal shield inside the
vacuum jacket (external envelope) should be defined at one of the earliest design
stages [11]. Proper arrangement of the lines should help to:
• minimize the external envelope size,
• minimize heat fluxes among process lines,
• reduce insulation vacuum space,
• avoid any unwanted thermal bridges among internal components,
• link thermally the thermal shield components with their cooling line,
• provide space for the supporting and thermal shrinkage compensation system
elements (supports, spacers, bellows, metal hoses, etc.),
250 J. Fydrych

Fig. 9.7 Example cross

section of a four channel
cryogenic transfer line

• keep enough room required for assembly (welding process lines and wrapping
radiation foils in interconnections, etc.),
• keep balance in the distribution of process pipe and radiation shield dead weight
to avoid significant moments which can lead to unwanted torsions of process
pipes along the line.
Figure 9.7 shows a schematic arrangement of the cross section of a four-channel
cryogenic transfer line. Its all process lines are surrounded by a cylindrical thermal
shield wrapped with radiative foils (MLI). The thermal shield is connected with the
thermal shield return line, which works as a thermal sink for radiative heat loads
absorbed by the shield. Since the cold helium supply line has larger diameter than
that of the thermal shield supply line, there is a clear asymmetry in the locations of
these process lines in respect to the vertical center plane. This geometrical asym-
metry guarantees the symmetrical distribution of process line dead weights.
Figure 9.8 collects the cross-sections of cryogenic transfer lines of some well-
known scientific facilities. The cryogenic transfer lines of HERA, KEKB, LHC,
Tore Supra, TEVATRON, TRISTAN, RHIC and XFEL/AMTF were already built
and operated, whereas the cryolines of ESS, FRIB, ITER and XFEL/Linac injector
are currently being designed or constructed. The TEVATRON CTL is a single path
cryoline, what means that it is designed to transfer cryogenic fluids into one
direction only. In order to return helium back to the cryogenic plant the cryogenic
system needs to use two such lines (feed and return lines) or return the helium via a
warm process line. All the other cryogenic lines shown in Fig. 9.8 are dual path
lines. They consist of supply and return lines. The number of lines depends on the
related cryogenic device cooling process and varies from four (eg. TRISTAN and
HERA CTLs) to ten (Tore Supra CTL). The sizes of the process lines differ very
much from line to line, since they depend on cryogen mass flow rate, density and
allowable pressure drop. Process lines used for transferring cold helium vapour at
low pressure usually have the largest sizes. In case of the LHC/QRL CTL this line
is DN250 in size and it transfers helium vapour at a subatmospheric pressure of 16
mbar absolute and temperature of 4 K in a distance of 3.4 km. The diameters of the
cryoline external envelopes also vary from line to line. The largest is the LHC/QRT
CTL which is equal to 640 mm, while the smallest is the CTL of KEKB (156 mm).
9 Cryogenic Transfer Lines 251

600mm
80 K, 18 bar

80 K 4.5 K 1.8 K, 1.2 bar

5K (LN2) 0.8 bar
4.5 K, 18 bar
4.5 K
80 K 18 bar 4.5 K, 1.2 bar
(LN2)
1.7 K
12 mbar

4.4 K 80 K, 18 bar
TRISTAN CTL [12] Tore Supra CTL [13, 14]

4.5 K
5K

80 K 80 K 4.5 K to
18 bar 18 bar ca. 100 K
80 K

40 K 2K
4.5 K 4.5 K 4.3 K
3.5 bar 6 bar 6 bar

ITER prototype CTL [15] XFEL/Injector CTL [16]

55 K, 2.5 bar

35 K, 3 bar
5K
1.3 bar 4.5 K, 3 bar 4K
27 mbar

4K
30 mbar 40 K
19.5 bar

4.5 K, 3 bar 50 K, 19 bar

FRIB Linac CTL [17, 18] ESS Linac CTL[10]

Fig. 9.8 Cross-section arrangements of cryogenic transfer lines existing or under construction
252 J. Fydrych

600mm

80 K, 16.7 bar
4.5 K 4.5 K
80 K 3.5 bar 40 K, 17 bar
40 K

4.5 K
4.5 K
1.2 bar

HERA CTL [19] XFEL/AMTF CTL [20]

50 K
19 bar
40 K
4.6 K
3 bar
70 K 70 K

4K
20 K 16 mbar
1.3 bar
4K
4.5 K

75 K
18 bar

RHIC VJRR CTL [21] LHC/QRL CTL [4]

200mm

80 K
(LN2) 5K
80 K
(LN2)
80 K
(LN2)
4.4 K
4.6-5.3 K
3.3 bar

TEVATRON CTL [22] KEKB CTL [23, 24]

Fig. 9.8 (continued)

9 Cryogenic Transfer Lines 253

9.4 Supporting Structures

Due to high temperature and pressure differences, and significantly high dead
weights of cryogenic transfer line components, cryogenic transfer lines require
complex supporting elements to carry the weights, to direct and transfer static and
dynamic loads, as well as to control the movements of cryoline components. Forces
resulting from pressure, thermal and gravity loads acting on the process lines (and
in some cases on the thermal shields) have to be transferred to the external envelope
and further to the surrounding structures, such as building floors and walls, piping
bridges, etc. These forces can be in an order of magnitude of 10 kN or even higher.
For example a DN250 process line equipped with an axial expansion joint (of an
effective cross section Aef = 690 cm2), which maximum allowable pressure PS is
6 bar, at a test pressure of PT = 1.43
PS = 8.58 bar acts on its fixed support with
an longitudinal force Fx = Aef
PT = 59.2 kN. For a cryogenic transfer line that in
addition to the DN250 process line includes a DN65 cold helium line (PS = 6 bar)
and two DN50 thermal shield lines (PS = 25 bar) the total force that has to be
transferred by the fixed support can reach 90 kN.
The supporting system of any cryogenic transfer line is composed of internal and
external supporting structures. The internal supporting structures include fixed and
sliding supports and spacers of process pipes and thermal shield components. Since
these supports connect mechanically the cryoline components at highly different
temperatures they should be designed in a way that significantly minimizes all
unwanted heat transfers. For given temperature differences, Fourier’s law of ther-
mal conduction gives only three general ways for reducing heat flows: (1) elon-
gating the heat transfer distances, (2) decreasing the heat transfer cross sections and
(3) choosing materials with low heat transfer coefficients (see Chap. 1). All of these
methods are applied in designing process line supports, however, the mechanical
strength requirement limits the elongations, cross section decreases and applicable
materials.
Process line fixed supports are supporting elements that bind mechanically the
process lines to the external envelope and remove all six degrees of freedom of the
lines in respect to the vacuum jacket. These supports are usually made of stainless
steel in order to allow for welding their elements both to the process lines and
vacuum jacket. Figure 9.9 shows a typical design of a process line fixed support.
This support is composed of several welding rings and sleeves and a plate. Ring 1 is
for welding the fixed support to the inner surface of the vacuum jacket. At working
conditions the heat from the external environment flows via the vacuum jacket wall,
ring 1 and sleeve 1 to the fixed support plate. The TS return line is thermally
connected to the plate and works as a thermal sink that stabilizes thermally the plate
at a thermal shield temperature (usually between 50 and 80 K).
The cold process lines and the TS supply line are connected to the plate via a set
of two sleeves and two rings which elongate the heat flow paths. In the case of the
cold helium return line, heat flows from the plate via sleeve 2, ring 2, sleeve 3, ring
3 to the wall of the line. Process lines can be also connected to the plate with one
254 J. Fydrych

Fig. 9.9 Process line fixed support: isometric view (a) and its longitudinal cross section showing
elongation sleeves of cold process lines (b)

sleeve and one ring. Then the heat transfer path is shorter and more heat flows to the
transferred helium, but the fixed support has better mechanical properties, and since
its design is less complicated, its manufacturing is easier and less expensive. In
order to reduce heat leaks the cryoline designers can elongate the sleeves and make
them from tubes of thinner walls as long as the fixed support has appropriate
mechanical parameters. In case of long sleeves there is a risk of lateral deformations
of the sleeves when exposed to radial forces. It can lead to unwanted thermal
bridges which can significantly increase heat loads to the cold process lines. The
lengths of sleeves usually varies from 0.3 to 0.8 m, whist their wall thicknesses are
usually similar to the process line wall thicknesses (1.8–3.2 mm).
Process line sliding supports are supporting elements that direct pressure and
thermal loads to the fixed supports. They also carry the weight of a section of process
lines, equipment attached to the lines (thermal shield components, radiation foils,
getters, absorbers, instrumentation) and cryogenic fluids in the lines as well. Since
these support do not bear high mechanical loads they are typically made of com-
posite materials characterized by low thermal conduction coefficient. Figure 9.10
shows an example design of a process line sliding support. This support is composed

Fig. 9.10 Process line sliding support

9 Cryogenic Transfer Lines 255

Fig. 9.11 Process line spacer: a isometric view, b top view

of two plates connected to each other with three stiffening rods. This solution pro-
tects the plates against tilting and blocking process line movements. The dead weight
is transferred to the external envelope via a set of sliding batons. The batons have
spherical ends to reduce the area of contact with the vacuum jacket internal surface.
In order to minimize friction the batons can be replaced with wheels, rollers or balls.
Spacers are supporting structures of a special type. Their main function is to
keep the lines in proper locations in cryoline cross sections in order to avoid
unwanted thermal contacts. They do not bear significant loads from process lines
but only limit the movements of the lines in their radial directions. Figure 9.11
shows an example spacer design. The sizes of the holes for process lines are
distinctively larger then the line sizes. Since the spacer does not bear any significant
mechanical load it can be made of a material of low thermal conductivity (that
usually implies lower mechanical strength) and can have a form that elongates the
heat transfer paths, which additionally lessens its mechanical properties but does
not add any substantial heat loads to the cold lines.
The family of cryoline external envelope supports also includes fixed supports,
sliding supports and spacers. These supports carry the entire cryoline weight and
transfer all forces to the surrounding structures. The forces, apart of dynamic
and static forces from process lines, can include also forces resulting from thermal
and pressure loads acting on the external envelope itself. The vacuum jacket (ex-
ternal envelope) can be exposed to a significant variation of ambient temperature and
it can also get cold locally due to a cryogen leak from a cold process line in case of a
failure mode. Some additional pressure forces can appear if the set pressure of the
external envelope safety devices significantly exceeds the atmospheric pressure. The
total forces that the fixed supports of the external envelope have to bear can reach the
value of 10 kN. Figure 9.12 depicts example designs of external envelope fixed and
sliding supports. The fixed support has a hoop that increases the mechanical strength
of the support. The hoop is made of steel plates and is welded to the external
envelope through two steel patches and to the support base. The base of the support
is to be screwed to a dedicated component of the surrounding structures.
256 J. Fydrych

Fig. 9.12 External envelope supporting elements: a fixed support and b sliding support

The sliding support shown in Fig. 9.12b is designed for allowing some axial
movements of the cryoline. The support foot is welded to the vacuum jacket via a
steel patch. The support base is to be screwed to the surrounding structure but not
fixed to the support feet at all. The base blocks the lateral movements and rotations
of the support foot. Usually the expected movements of the sliding support feet are
not higher then a couple of centimeters.

9.5 Thermal Contraction Compensation

All the components of cryogenic transfer lines tend to contract due to temperature
variations. This pertains not only to components which are at cryogenic tempera-
tures (cold process lines and thermal shields) but also their vacuum jackets, since
they can be exposed to a variation of the ambient temperature as well as to a
significant temperature drop during a failure mode of a cold process line break.
Therefore, both the process lines and external envelope must be protected against
excessive stresses and forces that can result from thermal contraction [11]. Cryoline
thermal contraction compensation systems can employ axial expansion joints,
metallic flexible hoses and natural compensation loops.
Natural compensation loops require a lot of space, so they are used only if the
cryoline routing is adequately complex and have a number of elbow and angular
modules located close to each other. Taking this solution requires detailed and
complex thermo-mechanical analyses of cryoline behavior in the conditions of all
possible operation and safety modes. Metallic flexible hoses are usually used in
elbow or angular modules to compensate some rather small dislocations of process
lines. There are usually two hoses installed in the bent section of a process line.
Then the contractions of the line sections are absorbed by the lateral deformations
of the hoses.
9 Cryogenic Transfer Lines 257

Fig. 9.13 Typical axial expansion joints used for cryoline process lines: a with bellows subjected
to internal pressure, b with bellows subjected to internal pressure and equipped with a guiding
sleeve c with bellows subjected to external pressure

Axial expansion joints seems to be the most often used thermal-contraction

compensation components, even though they are considered as the main items that
can significantly lower the reliability of transfer lines [7, 25, 26]. The joints consist
of bellows made of one to a few plies whose thickness has to be as small as a
fraction of millimeter (usually from 0.2 to 0.5 mm). Since these very thin elements
are subjected to significantly high and cyclical thermo-mechanical loads, there is a
risk not only of leaks but also of loosing the mechanical stability of process lines.
Figure 9.13 depicts three typical types of axial expansion joints installed in the
process pipes of cryogenic transfer lines . The first one (Fig. 9.13a) is composed of
two pipe sections and bellows. It is the simplest joint but also very sensitive to
become unstable. If the pipe with such a joint is not properly guided and anchored or
the expansion joint has very low stiffness the bellows or even whole pipe segment
can buckle [27]. This can of course lead to unwanted thermal contacts increasing
heat loads and what is more dangerous, to leaks in the cold process lines.
Axial expansion joints equipped with special guiding sleeves, as shown
schematically in Fig. 9.13b, can mitigate the risk of buckling to a certain extend, if
the risk of local instability comes from insufficient stiffness of the bellows. However,
this configuration also requires a precise arrangement of anchors and guides.
Joints with bellows subjected to external pressure (see Fig. 9.13c) are the most
stable from a mechanical point of view. When such joints are pressurized they
actually become less sensitive to the sources of local instabilities. On the other hand
their production is significantly much more expensive.
258 J. Fydrych

9.6 Materials

Materials used for manufacturing cryogenic transfer lines are almost always the
same as for other type of cryostats. Process pipes and their strong fixed supports as
well as vacuum barriers are made of austenitic stainless steels adequate for tem-
peratures which lines are going to be cooled down to. External envelopes can be
manufactured of some low carbon steels, as long as the lowest expected temperature
of the vacuum jacket will not cause any brittle fractures. The ductile-brittle transition
of low carbon steels ends in a temperature of −50 °C, so the application of low
carbon steels is not recommended if the vacuum jacket temperature can drop below
this value. Then external envelope should also be made of austenitic stainless steel.
Since cryogenic transfer lines are dedicated for transferring fluids at pressure
usually higher than 0.5 bar gauge, their process lines are considered as pressure
equipment and have to be designed and constructed in respect to dedicated regu-
lations and standards applicable in the country where the lines are to be installed and
operated. Example documents are the ASME B31 of the USA and the European
metallic industrial piping code EN 13480, harmonized with the Pressure Equipment
Directive 97/23/EC (PED) of the European Union. All these standards include a list
of materials that are applicable for use at given cryogenic temperatures. Cold process
lines are usually made of the standard Ni-Cr low-carbon stainless steel 1.4306
(304L), due to its significantly low price and applicability in a whole cryogenic
temperature range. For external envelope usually the stainless steel 1.4301 (304) is
used, which is cheaper than grade 1.4306 but applicable down to 77 K only.
Cryoline thermal shield components are usually made of some aluminum alloys
of 6000 series. These alloys have good thermal conductivity, low outgassing rate
and the surfaces of their commercial sheets, pipes and profiles have significantly
low emissivity coefficient. They are also easily machined and highly weldable by
using gas tungsten arc welding technique. AA6061, which is a good medium-
strength all-purpose aluminum alloy, is used for thermal shields made of pipes of
sheets, whilst AA6063 alloy is used for thermal shield made of extruded shapes.
Sliding supports and spacers of the process lines are usually made of fiber-epoxy
composites, such as NEMA grades G-10, G-10-FR4 which is fire retardant, or
G-10CR that is characterized by much smaller variability of their mechanical and
thermal properties at cryogenic temperatures. In case of cryolines designed for
long-time operation in radiation environment their sliding supports can be made
ofG-11 or G-11CR composites, which are less sensitive to radiation and thus retain
their properties longer.

9.7 Manufacturing and Installation

Cryoline modules are assembled from the inner to the outer parts. Their fabrication
usually consists of the following works:
9 Cryogenic Transfer Lines 259

1. Welding of process line sections and external envelope sections (including all
required examinations of welds, such as visual examination, radiographic or
ultrasonic tests, etc.),
2. Cleaning the external and internal surfaces of the prepared sections,
3. Leak tightness tests of process line and external envelope sections,
4. Winding MLI on the process lines,
5. Installation of process line fixed and sliding supports,
6. Installation of thermal shield elements,
7. Winding MLI on the thermal shield,
8. Inserting the process line sections with their thermal shield into the external
envelope section,
9. Connection of the process line fixed support to the external envelope section,
10. Preparation of the cryoline module for transportation.
Cryoline interconnections are also assembled from their inner parts to outer
ones, and their assembling requires the following activities:
1. Sliding the interconnection sleeve over the end of one of the cryoline modules,
2. Welding the compensators or suitable pipe sections to the ends of the process
line sections,
3. Testing the process line welds (visual examination, radiographic or ultrasonic
tests, leak tightness tests, pressure test, if needed, etc.),
4. Winding MLI on the process lines,
5. Installation of thermal shield segments,
6. Winding MLI on the thermal shield,
7. Sliding the interconnection sleeve over the interconnection and welding it to the
welding rings of the cryoline modules,
8. Testing the sleeve welds (visual examination, leak tightness test, etc.).
Typically the installation of a cryogenic transfer line requires assembling of a
number of interconnections. In order to minimize the time and cost of the instal-
lation works the leak tightness test of the process pipe welds are performed in one
turn for all the interconnections. The leak tightness test of the sleeve welds are also
performed

9.8 Case Study: XFEL/AMTF Cryogenic Transfer Line

A good example of a multichannel cryogenic transfer line is the XFEL/AMTF

cryoline. This line is used for transferring cryogenic cooling power from the HERA
refrigerator to the Accelerator Module Test Facility (AMTF) in the national
research center of the Deutsches Electronen-Synchrotron (DESY). Figure 9.14
shows schematically the run of the cryoline in the DESY site. The AMTF is
intended for testing the superconducting cavities and cryomodules of the European
X-ray Free Electron Laser (XFEL) [28]. The test facility requires for its continuous
260 J. Fydrych

Pipeline bridge with

the XFEL/AMTF
cryoline
HERA
refrigerator
hall

AMTF hall

Fig. 9.14 Schematic run of the XFEL/AMTF cryogenic transfer line (courtesy DESY)

operation the cooling capacities of 0.8 kW at 2.0 K, 0.5 kW at 4.5 K and 3 kW at

40/80 K [20]. Since the 2 K helium is produced by isenthalpic expansion of the
4.5 K helium (previously subcooled to 2.2 K in counter flow heat exchangers) in
the facility cryostats and test boxes, the XFEL/AMTF cryogenic line houses only
4.5 and 40/80 K circuits.
The XFEL/AMTF cryoline project was executed by Wrocław Technology
Park AB under the Polish in-kind contribution to the XFEL project and coordinated
by National Centre for Nuclear Research, Poland [29]. All the requirements for the
cryoline were provided to the Polish partners in the comprehensive specification
prepared by DESY [30]. Based on this specification, Wrocław University of
Technology developed the detailed design of the line and supervised production
and participated in the acceptance tests. The line was manufactured and installed in
the DESY site by Kriosystem Ltd [20]. The execution of the cryoline project went
through the following phases:
Phase 0. Specification of requirements,
Phase 1. Cryoline design,
Phase 2. Cryoline module productions,
Phase 3. Transportation of the cryoline modules to the DESY site,
Phase 4. Installation of the cryoline modules in the site,
Phase 5. Cryoline commissioning.
The first cool down of the cryoline to the nominal temperatures was successfully
curried out in August 2012 and since December 2012 the line is in continuous
operation.

9.8.1 Technical Requirements

All the essential requirements for the XFEL/AMTF cryogenic transfer line were
prepared on the basis of DESY experience with cryogenic transfer lines for HERA
and FLASH. The specification defined the run of the cryoline and its interfaces to
9 Cryogenic Transfer Lines 261

Interface to Interface to
valve box subcooler box

Fig. 9.15 Routing of the XFEL/AMTF cryogenic transfer line on the dedicated pipeline bridge
(courtesy DESY)

the adjoining components [30]. The detailed routing of the line is shown in
Fig. 9.15. The line runs from a valve box located between the HERA refrigerator
hall and the HERA West building on a pipeline bridge in a height of 8 to 10 m
above the ground level. At the AMTF hall the line goes 8.4 m down in two steps
and ends at a subcooler box located inside the hall.
The line had to be designed, manufactured and installed in conformance with
Pressure Equipment Directive 97/23/EC. In order to reach the conformity with the
directive, the AD2000 Pressure Vessel Code [31] had to be applied. The specifi-
cation also described all the essential requirements for mechanical design, tests to
be carried out at manufacture workshop and DESY site and guarantee data. The
defined sizes and operating conditions of the process lines, thermal shield and
external envelope are given in Table 9.2.
The line had to be designed so as to ensure reliable and uninterruptible operation
for at least 5 years and withstand 200 cool down/warm up cycles without damage
and deterioration of quality. All pressure bearing components (process lines and
their supports) had to be designed for the maximum pressure of 20 bar against
vacuum and for the pressure of 0 bar against atmospheric pressure at all possible
operating temperatures. Similarly, all the components of external envelope and their

Table 9.2 Sizes and operating conditions of the process lines, thermal shield and vacuum jacket
Pipe Size Diameter and Design Operating Operating
thickness pressure pressure temperature
(bara) (bara) (K)
4.5 K supply DN50 60.3 mm × 2 mm 20 3.5 4.5–6
4.5 K return DN80 88.9 mm × 2.3 mm 20 1.2 4.5
40 K supply DN40 48.3 mm × 2 mm 20 17 40
80 K return DN40 48.3 mm × 2 mm 20 16.7 80
Thermal NA 300 mm × 4 mm NA NA 80
shield
Vacuum DN400 406.4 mm × 4.78 mm 1.5 −1 300
jacket
262 J. Fydrych

supports had to be designed for the internal pressure from 0 bar absolute to 1.5 bar
absolute against the atmospheric pressure. In addition, the process lines had to be
designed so as to withstand the maximal possible temperature difference between
the inlet and outlet. It should also allow for independent cool-down/warm-up cycles
of the different process circuits as well as for rapid cool-downs or warm-ups of any
circuit.
Requirements on the tightness of the process lines and external envelope defined
the values of allowable single and integral leak rates into the isolation vacuum. The
single leak rate from any component of process pipes and external envelope such as
welds, bellows and corrugated hoses should not be higher than 1 × 10−9 mbar dm3/
s at the design pressure and both room and operating temperatures, whilst the
integral leak rate to the insulation vacuum should not exceed 1 × 10−8 mbar dm3/s.
Maximum allowable heat loads to the thermal shield and 4.5 K process lines
should not be higher than 1.5 and 0.15 W per meter, respectively. The cryoline
should be aligned so as the connections to the valve and subcooler boxes are
achieved within the limits of lateral forces for the process pipes and the vacuum
shells of the connection ports of the boxes.
The design of the line also had to take into account the variations of ambient
temperature in the range from 260 to 310 K and all possible thermal reactions
resulting from the expansions and contractions of the bridge components.

9.8.2 Design

Due to a significantly high distance between a potential manufacturer workshop and

the DESY site it was decided that the sizes of the cryoline modules should not
exceed the dimensions of a standard European semitrailer (13.62 m × 2.75 m ×
2.48 m). So the maximum length and width of the modules were taken as 13 and
2.5 m, respectively [32]. The modularization of the cryogenic transfer line is shown
in Fig. 9.16. The line is composed of 11 straight modules, 3 elbow modules and 1
angular module with an angle of 150.5° [20]. The cross section of the line is shown
in Fig. 9.8.
The interconnections between the modules were designed as typical sleeve-
shaped interconnections holding the components of the process line thermal com-
pensation systems. Figure 9.17 shows the schematic layout of the process line

EM1
SM – straight module SM11 EM3
AM – angular module SM10
EM – elbow module SM9 EM2
Interface to SM8
valve box SM7
SM6
SM5 Interface to
SM1 SM4 subcooler box
SM2 SM3
AM1

Fig. 9.16 Modularization of the XFEL/AMTF cryoline

9 Cryogenic Transfer Lines 263

A - axial support
L - lateral support
H - flexible hose
CA - axial compensator

Fig. 9.17 Schematic layout of the process line support and compensation systems

supporting and thermal compensation systems [33]. These systems include axial
compensators, flexible hoses, fixed supports and sliding supports. The sliding
supports block lateral movements, whilst the fixed supports block both lateral and
axial movements. Due to accumulation of the pressure forces at the extremities of
the cryoline straight sections, each line requires only four strong fixed supports and
11 fixed supports of considerably smaller mechanical strength. The thermal con-
traction of each line is compensated by a set of 13 axial expansion joints and 2
flexible hoses. Since the three elbows of the process lines are very close to each
other this section works as a natural compensation loop and there is no need for any
expansion joints.
The external envelope acts on the bridge not only with its own pressure, thermal
and dead weight loads but it also transfers the loads from internal process lines. The
layout of its supporting and thermal compensation systems is schematically shown
in Fig. 9.18. This supporting system consists of two fixed supports which block
vertical, lateral and axial movements (VLA) in respect to the piping bridge. There
are also 21 sliding supports allowing for axial movements only (VL) and 2 sliding
supports blocking the line movements only in vertical directions (V). The thermal
expansions or shrinkages of the envelope are compensated by three expansion
joints. Each of the two long straight sections are equipped with axial compensators
(CA). The thermal deformations in the section of the three elbow modules are
compensated with a lateral expansion joint (CL), which is built in the intercon-
nection between module E90-3 and the Subcooler Box.

A - axial support
L - lateral support CA - axial compensator
V - vertical support CL - lateral compensator

Fig. 9.18 Schematic layout of the external envelope support and compensation systems
264 J. Fydrych

External
envelope Compensators
Internal normal External fixed
fixed support support

Process
Cryogenic Thermal lines
transfer line shield

Internal strong
fixed support Internal
sliding support

Pipeline bridge
structure (simplified)

Fig. 9.19 FEM model of the XFEL/AMTF cryogenic transfer line (courtesy Wrocław University
of Technology)

The thermal and mechanical behavior of the cryogenic transfer line with the
proposed support and compensation systems was analyzed numerically. For this
purpose a complex Finite Element Method model of the cryoline and its pipeline
bridge was built as shown in Fig. 9.19 [34]. The model consisted of 107,000
elements and included all the kinematic joints of the external envelope, thermal
shield and process line supports. It used shell elements for process lines, thermal
shield and external envelope, beam elements for sliding supports and pipeline
bridge structure, spring elements for all the compensators and finally membrane
elements for elastic hoses. The model took into account the variation of the material
properties in the specified temperature ranges. It was assumed that the process lines
are made of stainless steel of grade 1.4306 or 1.4541 (A5 = 40 %, Rp1,0 = 220 MPa
and Rm = 520 MPa) and the external envelope is made of SS1.4301 (A5 = 43 %,
Rp1,0 = 235 MPa and Rm = 540 MPa).
The thermo-mechanical strength analyses of the entire cryoline were carried out
for five cases [33]:
Case 1: Design conditions (process lines and at their design pressure and oper-
ating temperatures, thermal shield at 80 K and external envelope under
vacuum and at 285 K, which is an assumed assembling temperature),
Case 2: Pressure test conditions (process lines at their test pressure (28.6 bara)
and 260 K, which is the lowest possible temperature),
Case 3: Failure mode I conditions (the external envelope of SM2 and AM1 at
160 K and the other cryoline components at the design conditions),
Case 4: Failure mode II conditions (the external envelope of SM10 and SM11 at
160 K and the other cryoline components at the design conditions),
Case 5: Failure mode III conditions (the external envelope of EM1, EM2 and EM3
at 160 K and the other cryoline components at the design conditions).
9 Cryogenic Transfer Lines 265

max = 87.2 MPa

Fig. 9.20 Von Mises stress distribution on the EM1 process lines (courtesy Wrocław University
of Technology)

The thermo-mechanical strength analyses resulted in the distributions of von

Mises stresses and deformations as well as the forces acting on the supports. The
obtained maximum stresses on the external envelope and process lines were equal
to 80.4 and 87.2 MPa, respectively. Fig. 9.20 shows an example picture of the von
Mises stress distribution on the process line sections. The maximum allowable
stresses, which take into account a safety factor of 1.5, are equal to 156.7 MPa for
the external envelope and 146.7 MPa for the process lines. So, the numerical
modeling showed that the maximum stresses did not exceed 75 % of the allowable
stress values and therefore, according to [35], both the vacuum jacket and process
lines are in stress category II. It means that the lowest temperatures of the external
envelope and process lines can reach −255 and −273 °C, respectively, what
approves the selection of the materials.
The obtained values of deformations of process lines and forces acting on all the
external and internal supports were used as input for designing these supports and
selecting adequate expansion joints. The 3D models of the designed process lines
supports are shown in Fig. 9.21. The process line strong fixed support is made of
stainless steel of grade 1.4306 or 1.4541, whilst the normal fixed support and
sliding supports are made mainly of G10 (marked in green). The double sliding
support is made of two single supports fixed together with three metal rods and 3
G10 tubes. Such a support, apart of blocking lateral movements of the process lines
in respect to the thermal shield, protects also the bunch of these lines against
twisting. So each straight section is equipped at one end with a normal fixed support
and at the other with a double sliding support.
The G-10 plate of the normal fixed supports is fixed to the radiation shield,
which is made of ∅300 AA6060 pipe sections. The sliding supports are mechan-
ically fixed only to the 4.5 K return line (DN 80). The other pipes are not connected
to the support and are allowed to move axially, so all the process pipes can contract
and expand independently. The sliding support plates are equipped with stainless
steel rollers what allow them to move axially inside the radiation shield.
266 J. Fydrych

Fig. 9.21 3D models of the process line strong fixed support (a), normal fixed support (b), single
sliding support (c) and doubled sliding support (d) (courtesy Wrocław University of Technology)

9.8.3 Manufacturing the Cryoline Modules

All the 11 straight, 3 elbow and 1 angular modules of the XFEL/AMTF cryogenic
transfer line were manufactured by Kriosystem Ltd. in the period of 10 months. The
assembly works of the straight modules contained the following set of technical
operations:
1. Welding of process line sections and external envelope sections,
2. Test of the process line and external envelope sections,
3. Manufacturing components of the thermal shield,
4. Cleaning the surfaces of the process lines, external envelope and thermal shield,
5. Winding MLI on the 4.5 K and 40 K process lines,
6. Installation of the supports on the 4.5 K and 40 K process line sections,
7. Assembling the 80 K return line sections with the thermal shield components,
8. Inserting the 4.5 K and 40 K process lines into the thermal shield sections,
9. Fixation of thermal links to the thermal shield and the 80 K return line,
10. Winding MLI on the thermal shield sections,
11. Installation of the sliding supports to thermal shield sections,
12. Inserting the thermal shield sections in the external envelope sections,
13. Connection of the process line fixed supports to the external envelope sections.
9 Cryogenic Transfer Lines 267

Fig. 9.22 Sliding supports

on the XATL1 straight
module process lines
(courtesy Kriosystem Ltd.)

As soon as each of the cryoline modules was assembled it was packed and
prepared for its transportation to the DESY site. Figures 9.22, 9.23, 9.24 and 9.25
show the photos taken during the production of a straight module. Figure 9.22
shows process lines subassembly at technical operation 6 (double sliding support in
the foreground). Figures 9.23 and 9.24 show one of the straight module sub-
assemblies under technical operations 9 and 10, respectively. One of the fully
assembled straight modules is shown in Fig. 9.25. Its process lines are ended with
axial expansion joints enclosed in guiding sleeves (as schematically presented in
Figure 9.13b).

9.8.4 Installation

The modules of the XFEL/AMTF cryogenic transfer line just after their manu-
facturing were packed and later transported to the DESY site. The installation work
included the following steps:
268 J. Fydrych

Fig. 9.23 Aluminium thermal shield of the XATL1 straight module with its copper thermal links
to the thermal shield return line (courtesy Kriosystem Ltd.)

Fig. 9.24 MLI radiation foils wrapped on the XATL1 straight module thermal shield (courtesy
Kriosystem Ltd.)
9 Cryogenic Transfer Lines 269

Fig. 9.25 Straight module assembled (courtesy Kriosystem Ltd.)

1. Lifting the cryoline modules on the pipeline bridge,

2. Positioning of the cryoline modules,
3. Fixing the external supports to the piping bridge,
4. Welding the process line sections of all the cryoline modules and connecting
them to the process lines of the valve box and subcooler box,
5. Testing the welding seams of the process lines put during the installation,
6. Executing the pressure test of the entire process lines,
7. Wrapping MLI on the process line interconnection,
8. Installation of the thermal shield components in the cryoline interconnections,
9. Wrapping MLI on the thermal shield interconnections,
10. Closing the external envelope sleeves and welding them to the vacuum vessel
rings,
11. Testing the welding seams of the vacuum vessel put during the installation,
12. Performing the leak tightness test of the entire cryogenic line.
During the installation some inner parts of the cryoline could be exposed to
weather conditions. Therefore all the interconnections located on the bridge were
protected with special tents that were conditioned in order to keep the inner parts
dry and clean. Figure 9.26 shows a cryoline straight module being lifted on the
pipeline bridge (step 1) and Fig. 9.27 shows the tents. The installation of the
interconnection between the cryoline and subcooler box (step 4) is presented in
Fig. 9.28. Here, the process lines are being connected with flexible hoses (protected
with braids) and the lateral expansion joint is suspended above the interconnection.
When all the internal parts of the interconnection were installed, the lateral com-
pensator was moved down and welded to the dedicated welding rings on the
vacuum jacket of the cryoline and the subcooler. Figure 9.29 presents the XFEL/
AMTF cryogenic transfer line after its installation. In the foreground there is the
270 J. Fydrych

Fig. 9.26 Installation of one of the XFEL/AMTF cryoline straight modules on the pipeline bridge
(courtesy Kriosystem Ltd.)

Fig. 9.27 The XFEL/AMTF cryoline interconnections on the pipeline bridge protected with tents
(courtesy Kriosystem Ltd.)

angular module with its strong fixed support. The visible stubs on the straight
section, which are covered with yellow lids, are the safety devices of the external
envelope. In case of some failure modes of process line rupture these safety devices
will discharge the helium from the isolation vacuum space to the surroundings and
protect the external envelope against overpressurization.
9 Cryogenic Transfer Lines 271

Fig. 9.28 Connecting the XFEL/AMTF cryoline to the cold terminal of the subcooler box
(courtesy Kriosystem Ltd.)

Fig. 9.29 The XFEL/AMTF cryogenic transfer line installed on the pipe bridge (courtesy
Kriosystem Ltd.)
272 J. Fydrych

4.5 K supply
4.5 K return
40 K supply
80 K return

Fig. 9.30 The evolution of temperatures in the valve box process lines measured during the
commissioning of the XFEL/AMTF cryogenic transfer line (courtesy DESY)

9.8.5 Commissioning and Performance

The preliminary acceptance test of the XFEL/AMTF cryogenic transfer line was
carried out by DESY between 06.08.2012 and 18.08.2012 together with the cold
performance tests of the subcooler box [28]. The measured temperatures in the
valve box process lines are shown in Fig. 9.30. During the test the cryoline was
rapidly cooled down close to the nominal operating conditions for a period of four
days and then abruptly warmed up to ambient temperature. The data collected from
the commissioning allowed for checking the thermodynamic efficiency of the
cryoline by applying the Second Law analysis and Gouya-Stodola theorem. The
performed entropy generation analysis showed that the thermodynamic efficiency of
the cryoline was only 3 % different from the specified parameters [36]. In
December 2012 the line was cooled down again and since then it is in continuous
operation and serves the AMTF facility for the tests of the XFEL linac cryomodules
and their cavities [28].

References

1. J.G. Weisend II, Transfer lines, defining cryogenics, cold facts. Winter 2011 issue (2011)
2. N. Dittmar, C. Haberstroh, U. Hesse, M. Krzyzowski, Characterisation and optimization of
flexible transfer lines for liquid helium. part I: experimental results. Cryogenics 75, 6–12
(2016)
9 Cryogenic Transfer Lines 273

3. K. Kawano, K. Hamada, T. Kato, T. Honda, K. Nishida, K. Matsui, T. Hiyama, K. Ohtsu, S.

Sekiguchi, H. Tsujji, M. Ando, T. Hiyama, K. Ichige, Design and construction of long
cryogenic piping lines, in Proceedings of the Sixteenth International Cryogenic Engineering
Conference (Oxford, 1997), pp. 493–496
4. McIntosh, Vacuum-jacketed transfer line design, section 5-6, in Handbook of Cryogenic
Engineering, ed. by J.G. Weisend II (Taylor & Francis Ltd, 1998), pp. 269–278
5. W. Erdt, G. Riddone, R. Trant, The cryogenic distribution line for the LHC: functional
specification and conceptual design, in Advances in Cryogenic Engineering, Transactions of
the Cryogenic Engineering Conference—CEC, vol. 45 (2000), pp 1387–1394
6. B. Barbier, A. Bouillot, K. Brodzinski, M. Dupont, M. Fathallah, J-L. Fournel, E. Gitton,
S. Junker, H. Moussavi, C. Parente, G. Riddone, Manufacturing and installation of the
compound cryogenic distribution line for the large hadron collider, in Proceedings Twenty
First International Cryogenic Engineering Conference (Praha, Czech Republic, 2006),
pp. 31–34
7. M.A. Iarocci, D. Brown, J. Sondericker, K.C. Wu, J. Benson, Y. Farah, C. Lac, A. Morgillo,
A. Nicoletti, E. Quimby, J. Rank, M. Rehak, A. Werner, RHIC cryogenics. Nucl. Instrum.
Methods Phys. Res. A 499, 264–279 (2003)
8. V. Kalinin, E. Tada, F. Millet, N. Shatil, ITER cryogenic system. Fusion Eng. Des. 81(23–24),
2589–2592 (2006)
9. B. Sarkar, N. Shah, H. Vaghela, R. Bhattacharya, K. Choukekar, P. Patel, M. Chalifour, Value
engineering in system of cryoline and cryodistribution for ITER: in-kind contribution from
india. Paper presented in CEC-ICMC2015. Tucson, USA, 2015
10. J. Fydrych, P. Arnold, W. Hees, P. Tereszkowski, X.L. Wang, J.G. Weisend II, Cryogenic
distribution system for the ESS superconducting proton linac. Phys. Proc. 67, 828–833 (2015)
11. J. Fydrych, M. Chorowski, J. Polinski, J. Skrzypacz. Design methodology of long complex
helium cryogenic transfer lines, in Advances in Cryogenic Engineering, Transactions of the
Cryogenic Engineering Conference—CEC, vol. 55 (2010) pp. 1103–1110
12. K. Hosoyama, K. Hara, A. Kabe, Y. Kojima, T. Ogitsu, Y. Sakamoto, S. Kawamura,
K. Matsumoto, Cryogenic system For TRISRAN superconducting RF cavity. Fusion Eng.
Des. 20, 491–498 (1993)
13. G. Bon Mardion, B. Gravil, B. Pager, The tore supra, flexible and electrically insulating,
cryogenic connection line. Cryogenics 32(Supplement 1), 142–145 (1992)
14. G. Bon Mardion, Systeme cryo, les canalisations cryogeniques. Association Euroatom CEA
Note TS-71-93-02, Cadarache (1993), p. 30
15. B. Sarkar, S. Badgujar, H. Vaghela, N. Shah, R. Bhattacharya, C. Chakrapani, Design,
analysis and test concept for prototype cryoline of ITER, in Advances in Cryogenic
Engineering: Transactions of the Cryogenic Engineering Conference—CEC, vol. 53 (1998),
pp. 1716–1723
16. E. Pyata, L. Belova, T. Boeckmann, M. Kholopov, V. Konstantinov, V. Kulikov, D. Sellmann,
A. Zhirnov, N. Zolotukhina, XFEL injector-1 cryogenic equipment. Phys. Proc. 67, 868–873
(2015)
17. V. Ganni, K. Dixon, N. Laverdure, P. Knudsen, D. Arenius, M. Barrios, S. Jones, M. Johnson,
F. Casagrande, FRIB cryogenic distribution system, in Advances in Cryogenic Engineering:
Transactions of the Cryogenic Engineering Conference—CEC, vol. 59 (2014), pp. 880–886
18. V. Ganni, K. Dixon, N. Laverdure, S. Yang, T. Nellis, S. Jones, F. Casagrande, FRIB
cryogenic distribution system and status. Paper presented in CEC-ICMC2015, Tucson, USA,
2015
19. Liquid helium transport to the hera tunnel, the transfer line in DESY webpage: http://www-
mks2.desy.de/content/e3740/e4847/e5250/e6587/e6593/e6628/e6743/index_ger.html
20. J. Fydrych, M. Chorowski, P. Duda, P. Grzegory, A. Iluk, K. Malcher, G. Michalski,
E. Rusinski, G. Strychalski, Design, manufacturing and assembly of the cryogenic transfer line
for XFEL/AMTF, in Proceedings of the 12th IIR International Conference on Cryogenics,
Cryogenics 2012 (Dresden, Germany, 2012), pp. 28–33
274 J. Fydrych

21. E.C. Quimby, C.M. Lac, M. Iarocci, R. Sallash, A. Varghese, VJR/VJRR design, construction,
installation, and performance, in Advances in Cryogenic Engineering: Transactions of the
Cryogenic Engineering Conference—CEC, vol. 43 (1998), pp. 531–540
22. The fermilab tevatron cryogenic cooling system. Fermi National Accelerator Laboratory,
United States Department of Energy, Batavia, AMSE getmedia (1993)
23. K. Hosoyama, K. Hara, A. Kabe, Y. Kojima, Y. Morita, H. Nakai, T. Fujita, T. Kanekiyo,
K. Matsumoto, Development of a high performance transfer line system. Adv. Cryog. Eng. 45,
1395–1402 (2000)
24. J. Yoshida, K. Hosoyama, H. Nakai, K. Hara, Y. Kojima, K. Nakanishi, M. Noguchi,
T. Ichitani, S. Kaneda, S. Sakuma, K. Suzuki, T. Kanekiyo, Development of STF cryogenic
system in KEK, in Proceedings of PAC07, Albuquerque (New Mexico, USA, 2007),
pp. 2701–2703
25. K. Brodzinski, P. Cruikshank, J.L. Fournel, L. Tavian, N. Veillet, Failure mechanism and
consolidation of the compensation bellows of the LHC cryogenic distribution line. Physics
Procedia 67, 129–134 (2015)
26. A. Thakkar, M.I. Vyas, Design & analysis of bellows free cryogenic transfer line. Presented at
the international conference on current trends in technology ‘NUiCONE-2011’, Institute of
Technology, Nirma University, Ahmedabad, 2011
27. B. Skoczen, Stability, Fatigue and optimization of thin-walled structures under cryogenic
conditions—application in the structural design of colliders and cryogenic transfer lines.
CERN Yellow Report CERN-2001-001, p. 121 (2001)
28. B. Petersen, Full cryogenic operation of XFEL accelerator module test facility—the first year,
in Oral Contribution to the 6th International Workshop on Cryogenics Operation presented,
11 Nov. 2014. Electronic copy available at https://eventbooking.stfc.ac.uk/news-events/6th-
international-workshop-on-cryogenics-operations-218?agenda=1/
29. E. Pławski, J. Sekutowicz, W. Grabowski, K. Kosiński, J. Lorkiewicz, M. Wojciechowski,
Z. Gołębiewski, K. Meissner, G. Wrochna, M. Duda, M. Jeżabek, K. Kasprzak, A. Kotarba,
K. Krzysik, M. Stodulski, J. Świerblewski, M. Wiencek, M. Chorowski, E. Rusiński,
J. Fydrych, A. Iluk, K. Malcher, J. Poliński, P. Duda, J. Głowinkowski, P. Wilk, M.
Winkowski, P. Grzegory, G. Michalski, Polish in-kind contribution to European X-ray free
electron laser (XFEL)-status in summer 2014, chapter 3, in Vacuum Technique & Technology
(Monographs of Tele & Radio Research Institute, Poland, 2014), pp. 33–43
30. D. Sellmann, Y. Bozhko, B. Petersen, XFEL project. Technical Specification for the
Cryogenic Transferline XATL1, Deutsches Elektronen-Synchrotron, MKS-S0008.0 (2009),
p. 34
31. AD2000-Merkblatt B0, Design of pressure vessels, ICS 23.020.30, Nov 2008 edition
32. J. Fydrych, Functional description of the cryogenic transfer line XATL1, Technical document
WUT_ XATL1-TD_001-0111 Ver. 1, Wrocław University of Technology (2011), pp. 1–4
33. A. Iluk, Design calculation of the cryogenic transfer line XATL1, Technical document WUT_
XATL1-TD_002-0111 Ver. 3, Wrocław University of Technology (2011), pp. 1–32
34. E. Rusinski, M. Chorowski, A. Iluk, J. Fydrych, K. Malcher, Selected aspects related to the
calculations and design of a cryogenic transfer line. Arch. Civ. Mech. Eng. 14, 231–241
(2014)
35. AD2000-Merkblatt W10, Materials for pressure vessels, materials for low temperatures,
ferrous materials, ICS 23.020.30, Nov 2007 edition
36. P. Duda, M. Chorowski, J. Polinskim, Design, optimization and operational parameters of
multichannel cryogenic transfer line for XFEL AMTF, in Poster Contribution 367 to the 25th
International Cryogenic Engineering Conference, 9 July 2014. Electronic copy available at
https://indico.cern.ch/event/244641/session/35/contribution/367/
Chapter 10
Guidelines for Successful Cryostat Design

J.G. Weisend II

Abstract This chapter acts as a summary of the book by providing a set of

guidelines for successful cryostat design.

10.1 Introduction

The information in previous chapters along with the extensive reference lists pro-
vides a solid overview of cryostat design. As a way of summarizing this work,
listed below are some general guidelines to successful cryostat design.

10.2 Guidelines

1. Define and prioritize requirements first. Optimize the cryostat design based on
these priorities.
2. Design in safety features from the start of the project.
3. Only use materials shown to be appropriate for cryogenic temperatures.
4. Review literature and learn from previous efforts. Take advantage of existing
codes and standards if possible.
5. Use tested commercial solutions whenever possible.
6. Intercept heat at higher intermediate temperatures.
7. Allow for the effect of thermal contraction on cryostat alignment and design.
Do not over constrain the movement of cryostat components as they cool.
8. Avoid feed throughs and demountable seals at cryogenic temperatures.
9. Be sure to properly heat sink temperature sensor wires to ensure reduced heat
leak and an accurate reading.

J.G. Weisend II (&)

European Spallation Source ERIC, Lund University, Lund, Sweden
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 275

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0_10
276 J.G. Weisend II

10. Install sensors such a pressure transducers and flow meters at room temperature
when possible.
11. Analyze the design for possible thermoacoustic oscillations.
12. Conduct design reviews. These should include experts not directly involved in
the design under review. Ideally, there should be at least reviews at the pre-
liminary or conceptual level and again once the detailed design is complete.
Safety should be a part of these reviews or separate safety reviews should be
held.
13. Conduct prototype tests when required. Leave enough time in the design
process to benefit from the results of such tests.
14. In cases where a large number of cryostats are to be produced, carry out series
testing of the production cryostats in addition to any prototype testing. Allow
sufficient resources (time, facilities, funding) in the project plan to accomplish
these tests.

10.3 A Final Comment

It is hoped that by now the value of learning from cases studies in cryostat design
has been shown. Workers in this field can benefit the broader cryogenics com-
munity by publishing the details and results of their cryostat designs. This includes
what worked and perhaps more importantly what didn’t work. There are ample
venues for such information sharing including conferences, workshops and
journals.
Index

A Cool down, 5, 9, 15, 17, 18, 20, 23, 29, 33, 36,
Accelerator module test facility (AMTF), 259 51, 53, 55, 56, 58–60, 88, 113, 121,
Aerogel, 16 123, 127, 129, 131, 132, 135–138,
Alignment, 2, 5, 17, 18, 20, 36, 76, 77, 79, 102, 142–144, 169, 170, 201, 206, 209, 211,
120, 121, 123, 126, 139, 142 212, 224, 226, 233, 238, 260–262, 272
Aluminum, 4, 6, 32, 33, 51, 55, 57, 101, 102, Copper, 4, 6, 29, 31, 32, 51, 72, 102, 104, 105,
124, 158, 170, 177, 188, 220 110, 156, 158, 169, 172, 189, 196, 200,
Argon, 38 206, 224, 227–230
ASME codes, 258 Cryocoolers, 2, 201–203, 207, 209, 232, 233,
ATLAS heavy ion linac, 20 238, 239
integration and attachment to, 201
B Cryocooler systems, 202
Bayonets, 32–35, 153, 154, 163, 167, 173 Cryogenic dark matter search (CDMS), 3, 219
Beam loading, 3 Cryogenic tanks, 15
Beam screens, 71, 72, 84 Cryogens, 95, 115, 166, 203, 233, 241
Bellows, 5, 36, 57, 58, 76, 96, 102, 110, 123, Cryogen free, 53
129, 151, 154, 156, 163, 164, 167, 169, Cryomodule, 2, 11, 18, 20, 29, 31, 35, 42,
182, 189, 190, 226, 234, 235, 238, 244, 117–121, 123, 126, 127, 129–132, 134,
249, 257, 262 136, 138–141, 143, 148, 149–161, 164,
Bimetallic transition joints, 32 166, 173, 178, 185–192, 212, 214, 259,
Boil off, 60, 62, 73, 103, 195, 198–200, 207, 272
209, 235 Cryopumping, 12, 130
minimization, 195 Current leads, 76, 81, 84, 121, 127, 131, 188,
tests, 60, 61, 173, 174 190, 201, 203, 206
Burst disc, 96, 110
D
C Dewar, 12, 13, 18–20, 30, 59, 95–99, 101–106,
Capillary tube, 27, 40, 132 108, 110–114, 213, 232, 235, 241
Cavity. See Superconducting RF Dilution refrigerator, 221–226, 228, 232–238
Cold finger, 201, 208 Distribution systems. See Transfer lines
Cold mass, 47–49, 53, 55, 57–62, 68, 70–77,
79, 81, 84, 85, 88, 119, 125, 129, 130, E
138, 181–183, 201–204, 206, 210 Electrical insulation, 3, 4
Commissioning, 70, 83, 185, 213, 237, 260, Emissivity, 13, 169, 170, 197, 198, 223, 224,
272 226, 227, 258
Continuous electron beam accelerator facility European spallation source (ESS), 148, 245
(CEBAF), 149

© Springer International Publishing Switzerland 2016 277

J.G. Weisend II (ed.), Cryostat Design, International Cryogenics Monograph Series,
DOI 10.1007/978-3-319-31150-0
278 Index

F condensation, 10, 16
Facility for antiproton and ion research (FAIR), conduction, 6, 11, 166, 170
245 forced convection, 11, 12, 16, 172
Failure modes, 39, 75–77, 83, 270 free convection, 11, 16, 172, 212
Fatigue, 76, 98, 100 He II, 67, 75, 104, 166
Feedthroughs, 3, 28, 30, 31, 155, 156, 158, radiation, 11, 12, 166
181, 231 Heat transfer coefficient, 253
Flanges, 31, 32, 74, 79, 85, 118, 124, 125, Higher order mode (HOM) couplers, 148, 154
127–130, 154, 156, 158, 172, 180, 225,
231 I
Flow measurement, 27 International linear collider (ILC), 2, 117–119
Foam insulation, 12, 16 Inclined pipes, 209, 211
Fountain pump. See Thermomechanical pump Indium, 4, 32, 100, 111, 154, 156, 179, 206,
Facility for rare isotope beams (FRIB), 148 224
Instrumentation, 3, 24, 28–31, 40, 81, 84, 107,
G 118, 127, 129, 131, 137–139, 141, 150,
Germanium, 25, 26, 107, 219 156, 158–160, 166, 168, 180, 181, 185,
crystal, 219 244, 254
temperature sensor, 25, 26 Interconnects, 120, 133
Glass Insulation, 16 Invar, 4, 6, 32, 33, 36, 123, 126, 127
Guidelines, 5, 39, 215, 242, 275 International tokamak experimental reactor
for Instrumentation, 24, 26, 29–31, 107, (ITER), 37, 245, 247, 250, 251
137, 138, 159
for safety, 275 J
for successful cryostat design, 275 Jefferson Lab 12 GeV upgrade, 20
Joint thermal resistance, 230
H
He II, 67, 69, 75, 104, 117, 166 K
Helium, 10, 13, 16, 28, 30, 38, 48, 50, 60–62, Kevlar, 105, 219, 229, 238
67–77, 88, 91, 95, 97–108, 113, 119,
122, 127, 131, 133, 138–140, 150, 155, L
157, 159, 166, 170–172, 176, 195, 197, Large hadron collider (LHC), 13, 14, 67, 68, 70
198, 200–202, 205, 206, 208, 212–216, Level measurement, 28, 232
224, 235, 238, 243, 245, 246, 250, 260 Linac coherant light source II (LCLS-II), 64,
Helium vessel, 69, 72, 74, 75, 88, 119, 122, 118–120, 123, 124, 126, 127, 129, 130,
123, 128, 131–136, 138–140, 152, 132–134, 136, 138, 143, 144
154–158, 163, 164, 166, 168–170, 180, Liquid acquisition device (LAD), 95, 96
181, 189, 195, 197, 200, 213–215 Liquid level sensor, 28, 138
Heat capacity, 9, 169 Liquid vapor detector, 28, 96, 98, 103,
Heat exchangers, 68, 69, 72, 75, 79, 170, 171, 106–108, 115
176, 177, 186, 202, 207, 225, 234, 238, Liquefied natural gas (LNG), 39
243, 260 Loss of vacuum, 131, 136, 153, 170, 216
Heat leak, 2, 9, 11–13, 15–18, 20, 24, 29, 34, Low vibration, 148, 201
35, 59–63, 98, 118, 125, 139, 141, 148,
254 M
Heat load Magnetic field effects, 25, 26
Static, 2, 3, 48, 71, 141, 154, 168, 170, 173, Magnetic resonance imaging (MRI), 195
174, 176, 183 Magnetic shielding, 3, 121, 123, 143, 148, 152,
Dynamic, 3, 48, 71, 118, 143, 154, 166, 157, 158, 181
167, 172, 173 Maintenance, 153, 168, 189, 201, 203, 207,
Heat sinking of wires, 28, 29, 31, 139, 141 232
Heat station, 102, 154, 169 Materials, 3–6, 9, 11, 38, 51, 57, 64, 73, 148,
Heat switch, 168 161, 169, 221, 227, 229, 230, 242, 253,
Heat transfer 254, 258, 265, 275
Index 279

Microphonics, 121, 122, 163, 185 S

Multilayer insulation (MLI), 101, 119, 127, Safety, 2, 38, 39, 70, 74, 83, 98, 99, 102, 110,
128, 130, 139, 141, 157, 158, 170, 175, 121, 153, 214–216, 244, 255, 265, 270,
188, 223 275, 276
MLI Seals, 31, 57, 126, 133, 154, 156, 179, 224,
seams and penetrations, 15, 139, 141, 157 228, 232, 275
Modeling, 196, 213, 215, 229, 265 SNS, 35, 161, 162, 165, 173, 176, 246
Space cryogenics, 23, 95
N Space frame, 20, 164, 165, 168, 169, 181, 183,
Natural convection. See Free convection 184
Niobium, 4, 10, 32, 120–123, 133, 147, 148, Stainless steel, 4–6, 20, 31–33, 55, 57, 74, 81,
154, 163, 166, 173, 180, 181, 242 85, 102–104, 110, 123–125, 152, 154,
Niobium titanium, 4, 166 156, 157, 163, 166, 169–172, 180,
Nitrogen, 13, 16, 28, 38, 60, 70, 74, 76, 86, 182–184, 188, 226, 238, 258, 264, 265
131, 195, 197, 212, 222, 223, 232–235, Standards, 84, 247, 258, 275
238, 241, 242, 245, 246 String tests, 75, 78, 80
Strongback, 188, 189, 191, 192
O Structural loads, 48
Operating experience, 70, 166 Superconducting magnets, 18, 47, 64, 68, 73,
Optimization, 2, 11, 76, 86, 91 91, 117, 148
Oxygen deficiency hazard (ODH), 39 Superconducting RF, 3, 4, 10, 39, 120
Superconducting splices, 71
P Superconducting super collider (SSC), 47
Perlite, 16, 17 Superfluid Helium. See He II
Phase separator, 96, 102–106, 114 Superfluid Helium on orbit transfer (SHOOT),
Piping, 33, 39, 48, 53, 58, 59, 61, 119, 130, 28, 95–115, 143
131, 135, 137, 143, 168–170, 175, 181, Superinsulation. See Multilayer insulation
191, 233, 253, 258, 269 Supports, 17, 18, 23, 24, 36, 49, 55, 71, 74,
Platinum resistors, 24, 26, 227 118, 127, 129, 139, 156, 158, 166–169,
Porous media, 103 173, 220, 225, 238, 242, 244, 249,
Power coupler, 120, 123, 129–131, 133, 139, 253–255, 258, 259, 262, 263, 265, 266,
148, 153, 154, 156, 158, 161, 167, 172, 269
176 Support post, 11, 23, 55–57, 60, 61, 73, 75, 79,
Pressure control, 131, 233 81, 84, 124, 127, 131, 139, 141, 188,
Pressure drop, 102–104, 132, 135, 171, 215, 191, 192
250 Suspension system. See Supports
Pressure measurement, 27, 160, 233, 236
Pressure relief, 39, 40, 58, 168, 170, 186, 244 T
Pressure rise, 10, 39, 174, 212–214 Temperature sensors, 24–28, 30, 59, 138, 159,
Pressure vessels, 3, 38 227, 244
Project X injector experiment (PXIE), 185 Tera-electronvolt superconducting linear
Prototyping, 30, 31, 42, 64, 143 accelerator (TESLA), 117, 119
Purging, 16 Test facilities, 63, 118, 259
Thermal conductivity, 6–8, 11, 29, 51, 52, 154,
R 170, 212, 225, 227, 229, 231, 255, 258
Radiation effects, 3, 25, 26 Thermal conductivity integrals, 6, 8, 9
Radiopure materials, 219 Thermal contraction, 5, 6, 18, 32, 36, 53, 60,
Reliability, 18, 30, 32, 95, 180, 232, 235, 257 76, 81, 85, 110, 124, 126, 128, 129,
Requirements, 1–3, 18, 20, 28, 36, 38, 39, 42, 154, 169, 189, 226, 243, 244, 256, 275
52, 55, 70, 75, 91, 98, 110, 112, 118, Thermal insulation, 1, 2, 11, 33, 35, 52, 53, 95,
120, 126, 131, 137, 148, 150, 178, 187, 147, 219
201, 221, 231, 260, 275 Thermal model tests, 59, 76
280 Index

Thermal modeling, 229 Joule-Thomson, 243

Thermal radiation shields, 3, 13, 48, 50 relief, 38, 39, 110, 153, 170
Thermal stress, 131 Vapor cooled shields. See Thermal radiation
Thermoacoustic oscillation (TAO), 1, 27, 30, shields
31, 40, 276 Volume ratios, 38
Thermomechanical pump, 95, 107 Vibrations. See Low vibration
Thermosiphon, 206, 209, 212
Transfer lines, 1, 35, 36, 120, 176, 241, 244, W
247, 250, 253, 256–258, 260 Weight, 2, 15, 17, 28, 39, 47, 48, 49, 51, 85,
Tuner, 120, 122, 123, 130, 137, 138, 143, 148, 87, 129, 225, 244, 250, 253, 255, 263
150, 152, 155, 156, 158, 161, 163, Welding, 31, 32, 76, 84, 90, 227, 249, 250,
166–168, 176, 181, 184, 185, 188, 189 253, 259, 269
Wiring, 18, 24, 28, 31, 114, 158, 168, 169,
V 173, 220, 227, 231, 237
Vacuum barrier, 76, 81, 84, 243, 244, 258
Vacuum vessel, 5, 20, 23, 48–50, 55, 58, 61, X
62, 73, 74, 79, 84, 119, 122, 126–129, X-ray free electron LASER (XFEL), 119, 259
152, 156, 158, 168, 183, 186, 188, 192,
234, 269 Z
Vacuum systems, 118, 148, 153 Zero boil-off, 207
Valves