Cryocoolers

Part 1: Fundamenta ls
THE INTERNATIONAL CRYOGENICS MONOGRAPH SERIES

General Editors K. D. Timmerhaus, Engineering Research Center

University of Colorado, Boulder, Colorado

Alan F. Clark, National Bureau of Standards

U.S. Department of Commerce, Boulder, Colorado

J. L. Olsen, Laboratorium fiir Festkorperphysik

Eidgendssische Technische Hochschule, Zurich, Switzerland

Founding Editor K. Mendelssohn, F.R.S. (deceased)

H. J. Goldsmid
Thermoelectric Refrigeration
G. T. Meaden
Electrical Resistance of Metals
E. S. R. Gopal
Specific Heats at Low Temperatures
M.G. Zabetakis
Safety with Cryogenic Fluids
D. H. Parkinson and B. E. Mulhall
The Generation of High Magnetic Fields
W. E. Keller
Helium-3 and Helium-4
A. J. Croft
Cryogenic Laboratory Equipment
A. U. Smith
Current Trends in Cryobiology
C. A. Bailey
Advanced Cryogenics
D. A. Wigley
Mechanical Properties of Materials
at Low Temperatures
C. M. Hurd
The Hall Effect in Metals and Alloys
E. M. Savitskii, V. V. Baron, Yu. V. Efimov,
M. I. Bychkova, and L. F. Myzenkova
Superconducting Materials
W. Frost
Heat Transfer at Low Temperature
I. Dietrich
Superconducting Electron-Optic Devices
V. A. Al'tov, V. B. Zenkevich, M. G. Kremlev, and
V. V. Sychev
Stabilization of Superconducting Magnetic Systems
G. Walker
Cryocoolers, Part 1: Fundamentals
Cryocoolers, Part 2: Applications
 Cryocoolers
Part 1: Fundamentals

Graham Walker
The University of Calgary
Calgary, Alberta, Canada

Springer Science+Business Media, LLC

 Library of Congress Cataloging in Publication Data
Walker, G. (Graham), 1930-
Cryocoolers.
(The International cryogenics monograph series)
Includes bibliographical references and indexes.
Contents: pt. 1. Fundamentals-pt. 2. Applications. 1. Low temperature engi-
neering. I. Title. II. Series.
TP482.W34 1983 621.5'9 83-2166

ISBN 978-1-4899-5288-2

ISBN 978-1-4899-5288-2 ISBN 978-1-4899-5286-8 (eBook)

DOI 10.1007/978-1-4899-5286-8

© 1983 Springer Science+Business Media New York

Originally published by Plenum Press, New York in 1983
Softcover reprint of the hardcover 1st edition 1983
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
High-capacity, three-stage Vuilleumier cycle cryocooler.
(Courtesy of Hughes Aircraft Co., Los Angeles.)
Foreword

The rapidly expanding use of very low temperatures in research and high
technology during the last several decades and the concurrent high degree
of activity in cryogenic engineering have mutually supported each other,
each improvement in refrigeration technique making possible wider oppor-
tunities for research and each new scientific discovery creating a need for
a refrigerator with special features. In this book, Professor Walker has
provided us with an excellent exposition of the achievements of this
period, the fundamental principles involved, and a critical examination of
the many different cryogenic systems which have led to a new era of
low-level refrigeration.
I feel fortunate to have had a part in the developments discussed in
this book. During the early 1930s I constructed several rotary engines
using leather vanes. Their performance was not good, but I was able to
liquefy air. I had been impressed by the usefulness of leather cups in tire
pumps and in Claude-type engines for air liquefaction. I was trying to find
a way to avoid that part of the friction generated by a leather cup as a
result of the radial force of the working gas on the cylindrical part of the
cup. During the 1950s I built two efficient helium liquefiers in which
essentially leather pistons were used. A steel core was encased in a stack
of leather rings separated by thin steel rings of slightly smaller diameter,
the stack of alternate leather and steel rings being compressed to provide
a rigid piston which could be machined to fit the cylinder. The wearing
quality was excellent. One of these liquefiers provided thousands of liters
of liquid helium during about two years of use. Instead of becoming
smaller in diameter because of wear, the diameter of the pistons actually
increased by absorption of water. The necessity of controlling the water
content was a disadvantage.
I was also intrigued by Heylandt's crowned piston concept during the
1930s but ruled it out because of the sealing problem at the warm end of
the piston. I had seen some of them in action in oxygen plants. A slight
vii
viii Foreword

air leakage was of no importance but a helium leak of such magnitude

could not be tolerated. Little did I realize then that by 1960 I would find
an adequate solution of the leakage problem by the use of rubber 0-rings.
During the late 1930s I experimented with free piston expanders and
compressors and with diaphragm engines. When the war began in 1941 I
was able to secure substantial funds for the development of oxygen gen-
erators and decided to see if Kapitza's engine could be useful. At least two
attempts had been made to reproduce Kapitza's liquefier in the U.S. One
was a complete failure and the other almost so.
I had tried several combinations of metals for the piston and cylinder.
The problem was to produce a fit between piston and cylinder close enough
to reduce leakage to a reasonable value but not close enough to encourage
seizure. Of the four combinations I tried, Kapitza's choice of bronze and
stainless steel was the least desirable. Kapitza found it necessary to employ
a relatively wide annular gap and contrived a way for the piston to travel
very rapidly during the power stroke so as to minimize leakage. By making
both piston and cylinder of nitrided alloy steel (hardness= 90 RC) I was
able to reduce the radial clearance by an order of magnitude and thus
reduce leakage to a negligibly small value. It became possible, therefore,
to use conventional gear for controlling the motion of piston and valves.
Several dozen engines with nitrided pistons and cylinders were
manufactured for oxygen production during the war. After the war this
type of engine was incorporated in helium liquefiers and was manufactured
by Arthur D. Little and its successor, CTI-Cryogenics. Between 1947 and
1970, 365 units were marketed.
I started using oil-lubricated rubber 0-rings to seal piston rods and
valve stems in nitrogen and oxygen plants in 1950 and in helium liquefiers
a few years later. Finally, I returned to the Heylandt type of piston, first
in nitrogen plants, and in 1960 in a helium liquefier. The piston was made
from a rod of laminated phenolic plastic (Micarta) about 2 ft long machined
to fit loosely in a stainless steel cylinder sealed at the room-temperature
end by a single 0-ring. The advantages over the earlier engine proved to
be enormous. The pistons and valves could be removed easily for inspection
without disturbing the insulating vacuum or breaking any piping. There
was no friction between piston and cylinder except when massive amounts
of air or water gained access to the working fluid. The reliability was very
high.
The loss of efficiency from the flow of gas to and fro in the annular
space between the hot and cold ends was very slight and mostly from the
mismatch of temperature between cylinder wall and piston surface. The
thin (0.002-0.006-in.) annular passageway became an effective regenerator.
The inefficiency, as in other reciprocating engines, was mostly the
result of irreversible heat transfer between the working gas and the walls
Foreword ix

of the expansion chamber in response to the relatively great temperature

change the gas undergoes. Engines working at lower temperature levels
tended to be more efficient because the drop in temperature of adiabatic
expansion was smaller. In fact, the efficiency of the two-phase engine was
apparently above 90%. Not only was the change in temperature very small,
but also the vanishingly small heat capacity of cylinder and piston made
the expansion truly adiabatic.
When I retired from MIT in 1964, I joined the engineering staff of A.
D. Little and soon thereafter put together an experimental model of their
standard liquefier equipped with Micarta pistons sealed by 0-rings. A
considerable number were produced over the next few years. Concurrently,
I developed a slightly larger machine, again with Micarta pistons, which
led to the 1400 series, completely replacing the earlier series. There are
200 of these units now in use.
A two-phase engine has replaced the Joule-Thomson valve in all of
the recent machines I have built and in some that CTI has manufactured.
The size of plant on which this change has been made ranges from 5 to
1000 1/hr. The gain in liquid production has been 25% to 33%.
One of the Naval Research Laboratory liquefiers has an inverted
engine, that is the cold end of the cylinders is at the top. It was astonishing
that the two-phase engine behaved normally and gave the expected
improvement over the Joule-Thomson. It seemed strange to have liquid
helium formed on top of the piston without appreciable loss by flowing
down the annular gap toward the hot end of the piston. Apparently
convection is very ineffective in thin layers of the order of 0.004 in. when
so little time (a period of one stroke) is available.
There is a wealth of information in this book. The comprehensive
Bibliography and Guide to Cryogenic Engineering Literature (Appendix
II of Part 2) is indicative of the prodigious effort which Professor Walker
has expended in preparation for writing this book.
Professor Walker and those who contributed certain chapters are to
be congratulated on this scholarly and timely treatise on a very interesting
subject.

S. C. Collins
Preface

The kindest thing to be said about this book is that it is like the curate's
egg-good in parts. Many times in its composition I have felt I would be
better engaged in reading rather than writing a book on cryocoolers.
Perhaps my modest effort will stimulate others better qualified than I to
do the job.
Completion is due entirely to Klaus Timmerhaus, Associate Dean
of Engineering at the University of Colorado, and to Dr. Alan Clark of
the National Bureau of Standards, Boulder, Colorado, joint editors of the
International Cryogenics Monograph Series published by Plenum Press.
They saw a draft of my first book on Stirling-cycle machines and flattered
me with an invitation to write another on cryocoolers. With the starlight
in my eyes I signed the contract that has been their rod to beat me with
since. Now at the end I am most grateful for their unquenchable interest
and enthusiasm despite the discouragements with which I have confronted
them. They have proved the burr in my saddle that only a completed
manuscript will remove.
Luster has been added to my effort by the substantial contributions
of others. I am indebted first to Samuel Collins for the Foreword. Ray
Radebaugh of the Thermophysical Properties Division of the National
Bureau of Standards, Boulder, Colorado, contributed Chapter 11, dealing
with his interest and expertise in the fundamentals of cryocooling and
Chapter 12, dealing with the systems and techniques for achieving very
low temperatures.
Fred Chellis, senior Applications Engineer of Cryogenic Technology
Inc., Waltham, Massachusetts, has contributed Chapter 10, dealing with
the practical problems of cryocooler design and operation. Fred was highly
qualified to do this, having been around cryocoolers longer than he cared
to remember and really should have been the author of this book. He died
unexpectedly during the production of this volume; his death was a great
loss to the cryocooler community.
xi
xii Preface

Evgeny Mikulin, a Professor of Cryogenic Engineering at the Baumann

Institute in Moscow, contributed Chapter 13. This review of the state of
development in the USSR of cryogenic cooling engines was a most valuable
contribution to offset my predominantly North American view. Mikulin
spent two months visiting the University of Calgary in 1977 and we came
to know him well. In a similar way Yoshihiro Ishizaki of the University of
Tokyo contributed Chapter 14, a review of cryocooler development in
Japan.
In Chapter 4, I have reproduced, with permission, a great deal of an
unclassified report on Vuilleumier cooling engines by Ronald White of the
U.S. Air Force Flight Dynamics Laboratory of the Wright-Patterson Air
Force Base, Ohio.
Professor Kurti of the University of Oxford convinced me to use the
descriptor cryocoolers rather than the more prosaic cryogenic cooling engines
that came naturally to me.
I am most grateful to all these men.
The book is dedicated to a trio who in different ways, contributed
much to my interest in this field. Aubrey Burstall, Professor Emeritus of
the University of Newcastle, perceived with great foresight, the future
significance of Stirling engines and initiated research at Newcastle on the
subject soon after World War II. With much persistence and charm he
secured delivery of the first Philips Stirling engine liquefier in Great Britain.
It was my good fortune to install and research this machine under Burstall's
guidance.
Jan Kohler led the team from Philips in their brilliant development
of a commercial Stirling cryocooler. He is recognized as a principal figure
in modern regenerative cryocooler development. I am most grateful to
Kohler for the insight and inspiration I gained from his stimulating public
presentations and technical writings.
Samuel Collins completes the trio to whom this book is dedicated.
While serving as professor at the Massachusetts Institute of Technology,
Collins invented a machine permitting the virtual routine production of
liquid helium and thus unlocked the way to widespread superconducting
research and development. When the ramifications are worked through
and superconducting systems are commonplace, Collins' place in technology
will surely be recognized as the equal of James Watt and the condenser:
the development that prefaced the industrial revolution.
My own debt to Collins lies principally in his book Expansion Machines
for Low Temperatures. By good fortune this became available simul-
taneously with the Philips engine and my need of it at Newcastle. I gained
much insight and encouragement from his book.
Since those days, over twenty years ago, many have helped me to
understand things better. My old friends Ted Finkelstein and William Beale
Preface xiii

come quickly to mind. More recently I have benefited much from mutual
exchanges with Bill Martini, Costa Rallis, and E. H. Cooke-Yarborough.
My colleagues at the University of Calgary withstand, good-humoredly,
my preoccupation with regenerative machines and my insistence on discuss-
ing with them matters I suspect are not their principal interest. In particular,
John Kentfield is a worthy foil off which I have bounced many ideas. The
Head of the Department, Dr. Peter Glockner, has helped in his encourage-
ment and advice.
I wrote this text as my principal activity on a six-month sabbatical
leave from the University of Calgary. I am most grateful to the University
for the opportunity to devote myself to the task unhampered by my normal
academic duties.
Many people helped in the production of the book. Bert Unterberger
and his assistants worked tirelessly on the diagrams and illustrations con-
tained herein. Karen Undseth, Edie Schulz, and Pamela Appleton did a
good job transforming my chicken-scratching info a readable text. Marlene
Stewart and Karen Undseth labored indefatigably in their customary fash-
ion on the index, corrections to the original text, and the galley proofs
thereby relieving me entirely of this onerous last lap. I am most grateful
for all their help.
My children Josephine and Christopher have been deprived of their
rightful allocation of my time but seem to have survived in good order.
Finally I owe my greatest thanks to my wife Ann. For over twenty
years now she has heard all about Stirling and other engines and still
manages a credible show of interest. With unfailing good humor she
somehow, in the end, makes it all worthwhile.

G. Walker
Calgary, Alberta
Contents of Part 1

Chapter 1 Introduction
Definition . . . . . . . . 1
Classification of Cryocoolers 1
Temperature Level . . 1
Refrigeration Capacity 1
Rotary Machines . . . 3
Reciprocating Machines 3
Mixed Units . 4
Large Systems 4
Small Systems 4
Heat Exchangers 5
Flow Regulation 5
Applications of this Text 6
Historical Development 6
Recent Development 10
Status Surveys 20
References . . . . . 26

Chapter 2 Elementary Thermodynamics of Cryocoolers

Introduction . . . . . . . . . . . . . . . . . 29
Part I: Thermodynamics Review
State Properties 29
Volume . . 30
Pressure 30
Temperature 31
Internal Energy 31
Enthalpy 32
Entropy . . . . 32
Equations of State and Thermodynamic Tables 32
The Temperature-Entropy (T-S) Plane 33
The T-S Diagram . . . . . . . . . . . . . 35
The Pressure-Volume (P- V) Diagram 36
The First and Second Laws of Thermodynamics 38
The Carnot Cycle 39
Coefficient of Performance . . . . . . . . . 42
XV
xvi Contents of Part 1

Part II: Regenerative Cycles

The Reitlinger Cycle 44
The Stirling Cycle 45
The Ericsson Cycle . . . . 48
The Stirling Cycle as a Prime Mover, Heat Pump, and
Pressure Generator . . . . 48
A Versatile Demonstration Engine . 50
The Rallis Cycle . . . . . . . . . 53
Rallis Isothermal Regenerative Cycle 53
Ideal Stirling Cycle . . . . . . . 58
Ideal Ericsson Cycle . . . . . . 59
Ideal Cycle with Constant-Volume Regenerative
Cooling and Constant-Pressure Regenerative
Heating . . . . . . . . . . . . . . . . 60
Ideal Cycle with Constant-Pressure Regenerative
Cooling and Constant-Volume Regenerative
Heating . . . . . . . . . . 61
Rallis Adiabatic Regenerative Cycle 62
Thermal Regeneration 65
Pseudo-Ericsson Cycle 67
Pseudo-Stirling Cycle . 68
Parametric Effects 71
Other Regenerative Cycles 74
Stirling Cycle 75
Vuilleumier Cycle 75
Solvay Cycle . . 76
Gifford-McMahon Cycle 77
Part III: Recuperative Cycles
Cryocooler with Recuperative Heat Exchanger 78
Joule-Thomson Expansion 79
Joule-Brayton Cycle 82
Siemens Cycle . . . . . 83
Linde-Hampson Cycle 83
Linde Dual-Pressure Cycle 84
Claude Cycle . . . . . 84
Analysis of Recuperative Systems 85
Compressors and Expanders 87
Compressors . . . . . 87
Multistage Compression . 90
Expanders . . . . . . . 91
Isothermal vs Isentropic Expansion 93
References . . . . . . . . . . . 94

Chapter 3 Stirling Cryocoolers

Introduction: Historical Review 95
Classification . . . . . . . 104
Single-Acting Stirling Engines 105
Double-Acting Stirling Engines 108
Multiple-Element Cooling Systems 116
Contents of Part 1 xvii

Piston and Displacer Motion 117

Two- Piston Stirling Engine 117
Integral Stirling Engine 118
Split Stirling Engine 119
Double-Acting Stirling Engine 119
Practical Regenerative Cycle 123
Mass Distribution in Ideal Cycle 124
Harmonic Piston Motion 124
Realistic Work Diagrams 124
Temperature-Entropy Diagrams 126
Schmidt Cycle . . . . . . . . 126
Nonisothermal Compression and Expansion 126
Aerodynamic Friction Loss 127
Heat Exchanger Thermal Potential 128
Regenerator Contamination 130
Regenerator Thermal Saturation 130
Theoretical Analysis of Stirling Engines 131
Introduction . . . 131
Ideal Stirling Cycle . . . . . . . . 132
The Schmidt Cycle . . . . . . . . 134
Principal Assumptions of the Schmidt Cycle 134
Nomenclature 135
Basic Equations . . . . . . . . 136
Mean Cycle Pressure . . . . . . 138
Heat Transferred and Work Done 138
Expansion Space . . . . . . . 138
Compression Space . . . . . . . 140
Mass Distribution in the Machine 141
Heat Extracted and Engine Output in Dimensionless
Units . . . . . . . . . . 142
The Finkelstein Adiabatic Cycle 142
Nodal Analysis . . . . . 145
Finkelstein Nodal Analysis 147
Urieli Nodal Analysis . . 147
Sunpower Nodal Analysis 148
Other Nodal Analyses 150
Summary 151
Practical Design 152
Design Parameters 152
Optimization of Design Parameters 153
Consolidated Design Chart 157
Alternative Basis for Optimization 159
Machine Design Procedure 159
Types of Stirling Engines 161
Multiple-Expansion 161
Free Piston 164
Free-Displacer Split-Stirling 169
Stirling Cryocoolers of Intermediate Capacity 172
Existing Intermediate Capacity Cryocoolers 172
The Werkspoor Cryocooler 176
Future Developments . . . . . . . . . . 179
xviii Contents of Part 1

Large-Capacity Stirling Cryocoolers 180

References . . . . . . . . 181

Chapter 4 Vuilleumier Cryocoolers

Introduction . . . . . . . . . . . . . . 185
Vuilleumier and Duplex Stirling Engines 185
Advantages . . . . . 186
Thermodynamic Aspects 187
Comparative Data 188
Historical Review 188
Hughes Vuilleumier Cryocoolers 191
Philips Vuilleumier Cryocoolers 195
AiResearch Vuilleumier Cryocoolers 199
R.C.A. Vuilleumier Cryocoolers 202
Cycle of Operation . . . . . . . . . 206
Inherent Thermodynamic and Heat Transfer Losses 220
Shuttle Loss . . . . . . . . . 220
Pumping Loss . . . . . . . . . . 221
Heat Transfer through Displacer . . 222
Heat Transfer through Cylinder Wall 222
Heat Generated by Friction between Displacer and
Cylinder . . . . . . . . . . . . . . 223
Regenerator Losses . . . . . . . . . . . 223
Heat Load due to Friction in the Regenerator 223
Heat Load due to Limiting Value of Film
Coefficient in Regenerator 224
Net Refrigeration 225
Total Heater Power Input 225
VM Cooler Variations 225
Multistage VM Coolers 225
Phase Angle . . . . . 227
Similar Cycles 227
VM Accessories and Components 228
Hot End Temperature Controller 228
Heaters . . 229
Motors 230
Regenerators 230
Heat Rejection 231
Split-Vuilleumier Cryocoolers 231
References . . . . . . . . . 233

Chapter 5 Gifford-McMahon, Solvay, and Postle Cryocoolers

Introduction . . . . . . . . 23 7
The Solvay Cycle . . . . . . . . . . . . . 240
The Gifford-McMahon Cycle . . . . . . . . 245
Multiple-Expansion Gifford-McMahon Cycle 247
Combination Gifford-McMahon Engine and Joule-
Thomson Expander . . . . . . . . . . . 250
Contents of Part 1 x ix

Fluidic-Driven Displacer . . . . . . . . . . . . 251

Machine Design and System Optimization . . . . . 255
Advantages and Disadvantages of Gifford-McMahon
Cryocoolers 257
Heat Balance Analysis 259
The Postle Engine 261
References . . . . . . . 263

Chapter 6 Joule-Thomson Cryocoolers

Introduction . . . . . . . . 265
Theoretical Aspects of Joule-Thomson Expansion 266
Isenthalpic Expansion 266
The Inversion Curve . . . . . 267
Maximum Inversion Temperature 267
Joule-Thomson Coefficient 268
Cryogenic Cooling and Gas Liquefaction Systems Using
Joule-Thomson Expansion . . . . . . . .. 271
Linde-Hampson Systems . . . . . . . . . . . . 271
Start-Up and Cool-Down of a Simple Linde-Hampson
System . . . . . . . . . . 274
Precooled Linde-Hampson System 275
Cascade System . . . . . . . 277
Linde Dual-Pressure System . . . 280
The Rietdijk Expansion-Ejector 282
Performance Comparison of Gas Liquefaction Systems 283
Figure of Merit . . . . . . . . . . . . . . . . 283
Ideal Liquefaction System . . . . . . . . . . . . 283
Performance of Linde-Hampson System with Different
Fluids . . . . . . . . . . . . . . . 284
Comparison of Liquefaction Systems for Air 285
Miniature Joule-Thomson Coolers . . . . . 288
Hymatic Coolers . . . . . . . . . . . . 288
Performance of Miniature Joule-Thomson Coolers 290
Self-Regulating Joule-Thomson Coolers 291
Complete Refrigeration System . . . . . . . 292
Air Products Coolers . . . . . . . . . . . . 293
Joule-Thomson Expansion Combined with Cooling
Engines 294
References . . . . . . . . . . . . . . . . . . 294

Chapter 7 Claude and Joule-Brayton Systems

Introduction . . . . . . . . . . . . 297
Isentropic versus Isenthalpic Expansion 297
Claude Cycle Systems 302
Claude Cycle . . . . . . . . 302
Analysis of Claude Cycles . . . 305
Optimum Recirculation Fraction 306
Heylandt Cycle . . . . . . . 307
XX Contents of Part 1

Low-Pressure Air Liquefiers 0 0 0 309

Collins Low-Pressure Air Liquefier 309
Precooled Claude System 312
Dual-Pressure Claude Cycle System 314
Comparison of Claude Cycle Liquefaction systems for
Air 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 314
Multiple-Expansion Engines 0 0 0 0 0 0 0 0 316
Claude Stepped Piston Two-Stage Expander 316
Collins Multiple-Expansion Engine 0 0 0 0 318
Arrangements in Multiple-Expansion Systems 318
Reciprocating Expansion Engines 322
Claude Expansion Engines 0 0 0 0 0 0 322
Heylandt Crowned Piston 0 0 0 0 0 0 0 325
Kapitza Hydrodynamic Lubricated Piston 326
Collins Expansion Engine 0 0 0 0 0 0 326
Doll-Eder Valveless Expansion Engines 334
Bellows Expansion Engines 340
Rotary Stroking Engine 0 0 0 344
Rotary Expansion Engines 344
Miniature Claude Cycle Systems 345
Terbot Mixed Refrigerant Cycle 346
Compressors for Claude Cycle Systems 34 7
Reciprocating Compressors 348
Rotary Compressors 3 50
References 0 351

Name Index 355

Subject Index 357
Contents of Part 2

Chapter 8 Heat Exchangers in Cryocoolers

Introduction . . . . . . . . . 1
Nomenclature . . . . . . . . 1
Heat Exchangers in Cryocoolers 1
Types of Heat Exchangers Used in Cryocoolers 4
Recuperative Exchangers 4
Tubular Exchanges . . . . 4
Plate-Fin Exchangers . . . 9
Perforated Plate Exchangers 11
Fundamentals of Recuperative Theory 13
Exchanger Effectiveness . . . . . . 17
Design of Recuperative Heat Exchangers 18
Maldistribution of Flow 19
Axial Heat Conduction . . . . . . . 22
Friction Effects . . . . . . . 24
Transient Response of Heat Exchangers 25
Oscillatory Flow Systems 26
Enhanced Heat Transfer Surfaces 30
Regenerative Heat Exchangers 30
Dynamic Regenerative Exchangers 30
Static Regenerative Exchangers 32
Common Theory for Static and Dynamic Types 32
Advantages and Disadvantages of Regenerative
Exchangers . . . . . . . . . . . . . . . 32
A Low-Temperature Problem: The Regenerator
Material Heat Capacity 33
Ideal Regenerator . . . . . . . . . . . . . 36
Hausen Regenerator . . . . . . . . . . . . 36
Presentation of Performance Data: Reduced-Length-
Reduced-Period Method . . . . . . 39
NTU-Effectiveness Method . . . . . . . . 41
Application of Theory to Regeneration in
Stirling-Type Engines with Oscillatory Flow 41
Regenerator Design for Stirling Engines 43
Experimental Performance . . . . . . . . 46
xxi
xxii Contents of Part 2

Heat-Transfer and Fluid-Friction Characteristics of

Dense-Mesh Wire Screens 47
Regenerative Annulus . . . . . . . 49
Heat Pipe . . . . . . . . . . . . . . . 54
Heat Exchangers for Very Low Temperature 55
References . . . . . . . . . . . . . . . 55

Chapter 9 Some Aspects of Design

Introduction 59
Target Definition 59
Reliability . . . 60
Coldfinger Design 62
Conduction Heat Leakage 64
Cooldown . . . . . 65
Regenerator . . . . 66
Shuttle Heat Transfer 67
Balancing . . . . . . 68
Case 1: Single Revolving Mass 68
Case 2: Several Masses Revolving in the Same Plane 70
Case 3: Several Masses Rotating in Several Planes 71
Reciprocating Masses . . 71
Partial Primary Balancing 72
Secondary Inertia Forces 74
Multiple Reciprocating Forces 74
Design Guidelines for Engine Balancing 76
Perfect Dynamic Balance 77
Bearings . . . . . . . . . 79
Fluid-Lubricated Bearings 80
Oil or Gas Lubrication 81
Grease-Lubricated Bearings 81
Ball and Roller Bearings 81
Gas Bearings . . . . . . 82
Gas-Lubricated Pistons 83
Gas Bearings on Shafts and Flat Surfaces 84
Dry-Rubbing Bearings 87
Seals . . . . . 89
Static Seals 89
Dynamic Seals . . 91
Piston Side Thrust 93
Hermetic Seals . . 95
Close Tolerance Seals 96
Materials . . . . . . 98
Significant Properties 99
Mechanical Properties 99
Physical Properties 100
Fluorocarbons 102
Closure . . . . . . 102
Contents of Part 2 xxiii

Cooling . . . . 103
Air Cooling 104
Water Cooling 105
Spacecraft Radiative Cooling 105
Electrial and Electronic Systems 106
Drive Motors 106
Brushless de Motors 108
Electric Controls 109
References . . . . . . 110

Chapter 10 Practical Problems in Cryocooler Design and

Operation
F. F. Chellis
Introduction . . . . . . . . . 113
Comparison of Cryocooler Types 114
Integral Stirling 115
Split-Stirling . . . . . . 116
Integral Vuilleumier (VM) 116
Split-Vuilleumier . . . . 116
Gifford-McMahon (GM) 117
Design Considerations and System Trade-Offs 117
Heat Rejection 117
Microphonics . . . . . . . . . . . . . 118
Thermophonics . . . . . . . . . . . . 118
Special Problems Related to Cryocooler Operation 119
Gas Contamination . . 119
Helium Gas Retention 120
Rubber 0-Ring Seals 121
Casting Leaks 121
Metal Porosity 121
Weld Joints 121
Seal Problems 121
References . . . 128

Chapter 11 Fundamentals of Alternate Cooling Systems

Ray Radebaugh
Introduction . . . . . . . . . 129
Thermodynamic Considerations 130
Ideal Refrigeration Cycles . . 130
Interaction of Force with System 134
Entropies and Refrigeration Principles of Various Systems 138
Helium-4 and CC1 2 F 2 • • • • • 139
Other Gas-Liquid-Solid Systems 143
Electron Systems . . 145
Phonon Systems 152
Other Solid Systems 153
Magnetocaloric Systems 155
xxiv Contents of Part 2

Electrocaloric Systems 160

Chemical Systems 163
Mixtures ..... 165
Photon Systems 168
How Much Entropy is Enough? 171
Alternate Means to Eliminating Mechanical Parts 172
References . . . . . . . . . . . . . . . . . 173

Chapter 12 Very-Low-Temperature Cooling Systems

Ray Radebaugh
Introduction . . . 177
He 3 Refrigerators 179
Properties of He 3 • • • • • 179
Single-Cycle He 3 Refrigerators 181
Continuous He 3 Refrigerators 185
He 3 -He 4 Dilution Refrigerators 187
Properties of Liquid He 3 -He 4 Mixtures 187
Principles of Dilution Refrigerators 190
Examples of Dilution Refrigerators . . 202
Multiple Mixing Chambers . . . . . 206
He 4 Circulating Dilution Refrigerators 207
Pomeranchuk Cooling . . . . . . . . 211
Properties of He 3 on the Melting Curve 211
Examples of Pomeranchuk Cooling 213
Magnetic Refrigerators 213
Electron-Spin Systems . . . . . 216
Nuclear-Spin Systems . . . . . 226
Hyperfine Enhanced Nuclear-Spin Systems 240
Combined Systems . . . . . 244
Dynamic Nuclear Polarization 247
References . . . . . . . . . . 251

Chapter 13 Cryogenic Engineering and Cryocooler

Development in the USSR
Evgeny Ivanovich Mikulin
Introduction . . . . . . . . . . . . . . . 257
Cryogenic Research Centers in the USSR . . 262
Principal Cryogenic Publications in the USSR 262
Properties of Substances at Low Temperatures 264
Computing and Analyzing Cryogenic Processes and Cycles 266
Entropy Method . . . . . 270
Exergy Method . . . . . 271
Expanders and Stirling Engines 272
Expansion Engines 272
Turboexpanders 277
Stirling Engines 278
Contents of Part 2 xxv

Heat Transfer Processes and Heat Exchangers 279

Boiling of Cryogenic Liquids 281
Convective Heat Transfers 282
Heat Exchangers and Regenerators 283
Cryogenic Insulation 0 0 0 0 0 0 284
Cryogenic Cooling Systems and Some Types of Cryogenic
Equipment 0 0 0 0 0 0 0 0 0 0 0 0 0 285
Helium Liquefiers and Cooling Systems 0 0 0 286
Cryogenic Vessels and Associated Apparatus 288
References 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 291

Chapter 14 Cryogenic Engineering and Cryocooler

Development in Japan
Yoshihiro Ishizaki
Helium Liquefiers and Refrigerators 293
Component Development 0 295
Dry Helium Compressor 295
Reciprocating Expanders 298
Final Liquefaction Process and Efficiency 299
Instability in Forced Cooling Systems 0 0 0 302
Superconducting Magnetic Levitation of Trains (Maglev) 0 306
Pulsed Refrigeration System 308
Conceptual Design 311
Appendix 1401 313
References 0 0 0 0 0 313

Bibliography
General Reading 0 0 315
Government Reports 350
Relevant Conferences 367
Patents 367
Appendix I Glossary of Terms for Cryocoolers and List of
Organizations 0 0 0 0 0 0 0 0 0 o o o o o 375
Appendix II Organizations Having Substantial Interest in
Cryocoolers and Cryocooler Manufacturers 385
Organizations 385
Manufacturers 385

Appendix Ill Guide to the Cryogenic Engineering Literature

Introduction 0 0 0 0 387
Government Reports 387
NTISearches 0 0 388
Superintendent of Documents (SupDocs) 388
The Cryogenic Data Center 0 0 0 0 0 0 389
xxvi Contents of Part 2

Conference Proceedings . . . 389

Foreign Government Sources 390
Open Literature Sources 390
Advances in Cryogenic Engineering 390
Cryogenics . . . . . . . . . . . 390
International Cryogenic Engineering Conference 391
Applications of Cryogenic Technology 391
International Institute of Refrigeration 391
Low-Temperature Physics . . . . . . 392
Books, Monographs, and Course Notes 393
House Journals . . . . . . . . . . 394

Namelndex . . 395
Subject Index 357
Chapter 1

Introduction

DEFINITION

A cryocooler is a device or ensemble of equipment for producing refriger-

ation at temperatures less than 120 K (216 R). The "quality'"' or worth of a
unit of such refrigeration depends on the temperature at which the refriger-
ation is available. This is illustrated in Fig. 1.1. The figure shows the
theoretical ideal work required to generate a unit of refrigeration as a
function of the refrigeration temperature.
At 0.5 K, 600 W of work are necessary to generate 1 W of cooling.
At 4 K this has declined to 7 4 W of work input. At 10 K, 29 Ware necessary,
and at 120 K, only 1.5 Ware required theoretically per watt of refrigeration.
These values were calculated for an ideal concept based on the Carnot
thermodynamic cycle with a "sink" temperature of 300 K. Unfortunately,
these ideal values cannot be attained in practice. Historically, the actual
work input has never been less than twice the Carnot value. In most cases,
it is many times the ideal, perhaps ten or even a hundred times.
Nevertheless, the example shown in Fig. 1.1 illustrates the point that
the refrigeration temperature is as important a criterion as the actual
quantity of refrigeration.

CLASSIFICATION OF CRYOCOOLERS

Temperature Level

For classification purposes it is expedient to divide the temperature

range, 0-120 K, into six levels:
i. Levell: 60-120 K iv. Level4: 4-10 K
n. Level 2: 20-60 K v. Level 5: 1-4 K
iii. Level3: 10-20 K v1. Level 6: less than 1 K
1
2 Chapter 1

LEVEL
6 5 4 3 2
700

;;;
~
~ 600 30
•
~
NOTE CHANGE
Of SCALE -........
z
'?
~
500
'""

\
~

~.
<r

"<r
~

\ \
w 400 20
<r
~
0
~
~

• 300
~

<r
w
~

0
1\ \
\
200 10
<r

"
~
0
<r
<
<r
100 Fig. 1.1. Theoretical ideal work to
generate a unit of refrigeration as a
• ~
0

""' 10 25 60
0
120
function of the refrigeration tem-
perature (assumed to be the Carnot
cycle with a sink temperature of
REFRIGERATION TEMPERATURE (K)
300K).

Refrigeration Capacity
For the purposes of classification, another important parameter is
refrigeration capacity. It is necessary to specify not only the refrigeration
capacity, but also the temperature at which it is associated. None of the
cryocoolers known at present can operate effectively over the whole cryo-
genic temperature range. Therefore, for classification and reference, bench-
mark temperature levels of 1 K, 4 K, 20 K, 80 K, and 120 K are proposed.
These correspond approximately to the liquefaction temperatures at normal
pressure of: helium (4 K), hydrogen (20 K), nitrogen (80 K), and methane
(120 K).
A possible classification scheme for cryocoolers is shown in Table 1.1.
In this classification five ranges of cooling capacity are prescribed: micro-
miniature, miniature, small, intermediate, and large.
The principal means of achieving cryogenic refrigeration in the level
1, 2, 3, 4 temperature range (120-4 K) is the expansion of compressed gas
in some form of engine. There are many varieties of cooling engines. They
may be broadly classified as
i. Rotary machines.
ii. Reciprocating machines.
Introduction 3

Table 1.1. Proposed Classification of Cryogenic Cooling Systems Based on Cooling

Capacity and Refrigeration Temperature.

Cooling capacity

Designation lK 4K 20K 80K 120 K

(a) Microminiature (very low <0.25W <1 w <l.SW

capacity for electronic
applications)
(b) Miniature (low capacity <O.SW <2W <8W <12W
electronic and instrument
applications)
(c) Small (general purpose <1 w <lOW <lOOW <0.8kW <1.2 kW
laboratory and
commercial systems)
(d) Intermediate (small-scale <25W <lOOW <lkW <lSkW <25kW
gas liquefaction and
industrial superconducting
systems)
(e) Large (large-scale gas >25W >lOOW >lkW >15kW >25kW
liquefaction systems,
LNG production, tonnage
oxygen, helium recovery, etc.)

Rotary Machines
Rotary machines, turbines, and other forms of rotary expander, are
characterized by small size and weight, high speed, and high fluid flows at
relatively low compression ratios (maximum pressure/minimum pressure).
Rotary machines are relatively inefficient in small sizes but improve with
increases in size and speed. They are relatively inflexible for operation
away from the design condition and so are best suited for continuous,
steady-state operation. There are few bearings and no rubbing surfaces in
contact so that reliability and life expectancy can be very high. Rotary
machines are, inherently, in near perfect dynamic balance.

Reciprocating Machines
With reciprocating machines the effect of size is much less dramatic.
To a first approximation they are as efficient in small sizes as in large. Of
course the relative significance of the mechanical and thermofluid losses
tends to decrease as the size of the machine increases.
Reciprocating engines can operate over a very wide pressure range
and operate quite well even when conditions are reasonably removed from
4 Chapter 1

the design condition. They are relatively large, heavy machines and, to
minimize friction and wear, must operate at low speeds, often an order of
magnitude less than that of rotary systems. The increased number of moving
parts and various rubbing seals and sliding bearings involved decrease the
prospects for high reliability and long life so that relatively high maintenance
costs might be anticipated for reciprocating systems. The out-of-balance
reciprocating forces must be balanced in some way or absorbed in relatively
complicated foundation or mounting installations.

Mixed Units
A cryocooler generally comprises a compression unit and an expansion
unit. Both can be rotary or reciprocating or can be mixed. It is not
uncommon to find in intermediate and the smaller large-scale cooling
systems a combination of reciprocating compressor and turbine expander.
This is because the process of expanding a gas in a turbine is immeasurably
easier to achieve than compressing one. In a turbine expander the gas
readily decreases to a lower energy level. In a rotary compressor the gas
must be 'coaxed" by ingenious aerodynamic design to ascend in pressure
against all natural inclination. The "coaxing" to higher pressure can be
accomplished in more brutal fashion with a reciprocating compressor and
much greater compression ratios can be achieved.

Large Systems
Despite the above difficulties, the advantages of reduced size, weight,
and maintenance of rotary compressors compel their adoption in large
systems to an increasing degree. It is fundamentally a question of develop-
ment effort. Given the appropriate level of commercial application there
is little doubt that satisfactory rotary compressors could be developed for
all intermediate and large-scale applications. Substantial improvements
have been made and work continues with a variety of compression equip-
ment of the centrifugal, axial flow, and screw compressor variety.

Small Systems
Miniature, small, and intermediate cryogenic cooling systems almost
invariably utilize reciprocating systems. Valiant efforts have been made
with small turbosystems, but these remain of interest only in aircraft,
Introduction 5

spacecraft, and instrument applications, where size, weight, and dynamic

balance are of primary significance.

Heat Exchangers
Heat exchangers are vital elements of all cryocoolers. Two principal
types of heat exchanger are used: (i) recuperative and (ii) regenerative. A
recuperative heat exchanger is a device where separate flow passages are
provided for the hot and cold fluids. The fluids are separated by a solid
wall and heat is transferred by conduction across the wall. The fluids may
flow continuously or periodically. A regenerative heat exchanger has a
single set of flow passages through which the hot and cold fluids flow
alternately and periodically. The regenerative matrix, often a porous finely
divided mass (of metal wire or spheres), may be thought of as a thermo-
dynamic sponge alternately accepting or rejecting heat as the hot or cold
fluid flows through it.
Both types of exchange may be used in counterflow or in parallel-flow
operation. In counterflow operation the hot and cold fluids flow in opposite
directions. In parallel flow operation the fluids flow in the same direction.
Counterflow operation is much more effective than parallel flow operation
and should always be used except in special circumstances.
The type of heat exchanger used, either regenerative or recuperative,
is a key feature in identifying the particular family of machine. Within this
family various machines may be further separated and classified by the
thermodynamic cycle on which they operate or by some other feature.
Cryocoolers with regenerative heat exchangers include Solvay and
Gifford-McMahon, Stirling, Ericsson, and Vuilleumier engines. Cryo-
coolers with recuperative heat exchangers include Linde, Hampson, Claude,
Collins, and Joule-Thomson engines.

Flow Regulation
Another key feature is the manner in which the flow of working fluid
is controlled, namely, by valves or by volume variations. All recuperative
engines have valves. Therefore, classification by means of flow regulation
is specific to regenerative engines. Here an arbitrary distinction is adopted:
i. Stirling engines where the flow is controlled by volume changes;
ii. Ericsson engines where the flow is controlled by valves.
This distinction between engines with and without valves is important and
has a profound impact on the design and operation of regenerative
machinery.
6 Chapter 1

Application of This Text

This natural division of application, reciprocating engines to small
coolers and rotary engines to large coolers, made the task of defining the
limits of this presentation relatively simple. It is devoted to small and
intermediate capacity systems, and so we have eliminated all discussion of
rotary machines except for passing reference to the principal source
materials. Attention will be concentrated wholly on reciprocating systems.
A companion volume on rotary systems is contemplated.
The above definitions and classifications are not widely established,
but are included as an attempt to bring some order to a confused field.

HISTORICAL DEVELOPMENT

Refrigeration as an aid to improving the quality of life has been known

from early times. There are many Biblical references to "cool waters."
The Romans had well-developed systems for hauling ice from distant Alpine
regions and storing it underground in caverns insulated with straw. In
England, Charles II was delighted by the ice cream confections of his
Italian chef. All this depended on the natural refrigeration induced by
environmental change.
It was not until the 19th century that serious development of
mechanical refrigeration was attempted. Invention of the reciprocating
steam engine made available the necessary power. It also triggered the
development of improved metals and materials, better construction and
manufacturing processes with consequent growth of populations and their
expectations.
In 1815, Robert Stirling, a minister in the Church of Scotland, con-
ceived the regenerative heat exchanger and the first engin~ now known as
the Stirling engine. This was used for the production of power from furnace
heat. However, in 1834, John Herschel, in a letter to "The Athenaeum"
clearly anticipates application of the Stirling engine for making ice. It was
not until 30 years later that Alexander Kirk (1874)* succeeded, about
1861, in reducing the concept to practice.
Collins and Cannaday (1958) have presented a good historical review
of early refrigerative engines commencing about 1828 with the Cornish
mine engineer, Richard Trevethick, better known for his work on steam
power systems and his adventures in the silver mines of Peru. Collins makes
the point that refrigeration expansion engines were not developed as

* References listed at the end of chapter.

Introduction 7

well-defined independently built pieces of machinery. Instead they always

appeared as components of integrated systems including compressors and
the associated heat exchange equipment. Nearly all the early machines
used air as the working fluid with the notable exception of the engine
invented by the Australian, Davy Postle, in 1873. This machine used
hydrogen and is further remarkable as the earliest known "free piston"
engine and the forerunner of the class of cryocooler presently called
Gifford-McMahon machines.
The first machines were used for making ice and for process refriger-
ation. The great impetus for large-scale refrigeration followed the opening
of the Suez Canal and the development of the meat trade between Australia,
New Zealand, and Great Britain. Lloyds Register for 1899 I 1900 records
285 ships with refrigerating facilities, 133 with air, 76 with ammonia, and
76 with carbonic acid as the working fluid. Twelve years later the corre-
sponding figures were 66 with air, 241 with ammonia and 463 carbonic
acid. This illustrates the growth of the meat and fruit trades and the decline
of air cycle machines in favor of condensing cycles.
The development of cryocoolers followed much later. Collins (1958)
writes:

The equipment required for liquefying a gas, namely a compressor, a heat

exchanger and an engine for expansion was fully appreciated as early as
1860. The lack of a compelling interest in the achievement of low temperature
seems to have caused the many competent scientists and engineers of that
day to neglect the subject.

Air was first liquefied in 1877 by Cailletet and, simultaneously, by

Pictet. Working independently, both expanded high-pressure precooled air
and oxygen in a nozzle to obtain the liquefied gas droplets.
The original engine expressly intended for the liquefaction of air was
described by E. Solvay in a patent of 1887. The engine used hydrogen as
the working fluid and air or other gases were expected to condense on the
outer surface of the expansion cylinder. Solvay's engine failed to attain a
sufficient reduction in temperature to achieve liquefaction of air. However,
Collins attributes Solvay with "a real understanding of the problems of gas
liquefaction in general" and refers to him "as the first to specify broadly
the apparatus needed."
Karl von Linde, a professor at the Munich Polytechnical Institute in
Germany, was, in 1895, the first to achieve the liquefaction of air in quantity.
Linde's apparatus comprised a compressor, counterflow heat exchanger,
and Joule-Thomson expansion valve. His work laid the foundation for the
German Linde Company, now renowned in the design and construction
of low-temperature equipment. Similar work was carried out in England
8 Chapter 1

contemporaneously by Hampson and led eventually to the British

Oxygen Company, equally well known for innovative work in the
cryogenic field.
The first use of counterflow heat exchangers was in 1849, by the
American, John Gorrie of Apalachicola, Florida, in an air refrigerator for
making ice. Later, in England, in 1854, Charles Siemens, in a provisional
patent application, and elsewhere in a report on tests made with a British
version of the Gorrie engine (Siemens, 1882), discussed the use of
counterflow heat exchangers and clearly anticipated their use for the pro-
duction of very low temperatures. In view of this early appreciation, Collins
found it remarkable that more than 30 years were to elapse before Linde
achieved the liquefaction of air in quantity.
In 1896, Kamerlingh Onnes of the Leiden Laboratory outlined
schemes for the liquefaction of hydrogen by means of a heat exchange
system employing liquid oxygen in the first stage and a hydrogen expansion
engine for the final stage of cooling. The cycle he proposed was quite sound
and, much later, was used for the liquefaction of hydrogen. However,
Onnes failed in his early attempts and gave up. Later, of course, Onnes
secured his niche in history with the first liquefaction of helium and the
initial observation of the phenomenon of superconductivity.
In France, the resourceful Claude completed his first expansion engine
in 1899. This was no better than the Solvay engine of twelve years earlier,
but, Claude was made of sterner stuff and his perseverance and careful
analysis eventually led to successful operation. For the first time a reason-
ably efficient expansion engine and heat exchanger were combined with
compressed air free of water and carbon monoxide. When used in conjunc-
tion with an auxiliary liquefaction and separation element, the eagerly
awaited continuous stream of liquid air was finally achieved in May, 1902.
Claude's expansion engine development was accompanied by many
improvements related to liquid air production and separation. His work
led to the French industrial company L' Air Liquide, nowadays concerned
with the construction and operation of plants for the production of industrial
gases.
A noteworthy design feature for expansion engines introduced in 1912
by Heylandt, was the use of a "crowned" piston. A long cap of low-
conductivity material was mounted on the piston to separate the cold
expanding gas from the piston rings used to seal the high-pressure gas in
the cylinder. This allows the seals to work at ambient temperatures and,
more importantly, isolates the thermal energy resulting from frictional
degradation of the seals thus preventing dissipation of the low-temperature
refrigeration arising from expansion of gas in the cylinder. It is now a
common feature of gas liquefaction machinery.
Introduction 9

The first expansion engine for hydrogen was constructed by Claude

in 1920 to separate hydrogen from coke oven gas. The hydrogen engine
was similar in design to the earlier air engines. In his book Collins (1958)
includes photographs of Claude expansion engines and an account of similar
engines used in the United States by duPont de Nemours.
Helium was liquefied in an apparatus incorporating a helium expansion
engine described, in 1934, by Peter Kapitza, then a professor at the
University of Cambridge. Kapitza, long resident in Russia, was recently
(1978) awarded a Nobel prize for his extensive cryogenics activity. In the
Kapitza engine, helium at 30 atmospheres pressure expanded to 2.2 atmos-
pheres and achieved a temperature reduction from 20 to 10 K. The engine
operated at about 100 strokes a minute and was operated by a hydraulic
drive system. It was notable as the first use of a controlled leak of the
pressurized working fluid to lubricate the piston during operation in the
cylinder.
About the same time (1938) Collins completed a diaphragm expansion
engine which operated sufficiently well on compressed air to justify incor-
porating the unit in a helium liquefaction system. The unit is described in
detail by Collins in his book. Temperatures of 10 K were achieved in late
1939 but no liquid helium was produced because of leakage in the vacuum
circuit. The war in Europe caused a redirection of effort to oxygen gener-
ation and work on the diaphragm engine was never resumed.
Turning now to a brief consideration of turbine systems, we return to
the close of the 19th Century. This was the period when pioneer work on
steam power turbines was in progress by Sir Charles Parsons and others.
Possibilities for the use of turbines for gas liquefaction were not overlooked.
An early discussion of the prospect was given in a letter written in 1898
by Lord Rayleigh to the Journal, Nature, wherein he suggested a turbine
would obviate many of the practical difficulties of reciprocating engines.
The same year, Edgar Thrupp, a British engineer, patented a liquefaction
system using an expansion turbine. About the same time an American,
Joseph Johnson, patented a liquefaction system incorporating a single-stage
De Laval impulse turbine. U.S. patents for turbine expanders were granted
in 1914 to the American inventors Charles and Orin Crummett. Later on,
the American engineer Harvey Davis patented a compound expansion
turbine for gas liquefaction in 1922.
However, it was not until the early 1930s that expansion turbines were
applied commercially for gas liquefaction by the Linde Company. Initially,
the turbine was an axial flow single-stage impulse machine but was sub-
sequently replaced by an inward radial flow turbine invented by the Italian,
Guido Zerkowitz. Other German refrigerative turbines of this era were
described by Swearingen (1947) for both high-pressure and low-pressure
10 Chapter 1

applications. Few details of these machines are known but Collins (1958)
describes them as "a commercial success and certainly matched the
efficiency of any low temperature reciprocating expander available at the
time."
An extensive theoretical treatment of refrigerative turbines was pub-
lished in 1939 by Kapitza. He concluded that a low-pressure turbine
liquefier was preferable to a high-pressure liquefier with reciprocating
expander. Not only was the unit thermodynamically superior but it was
also cheaper to build. Kapitza compared inward flow radial and axial flow
impulse turbines and concluded the former was the preferred machine for
gas liquefaction. Experimental results were presented for an inward flow
radial expander capable of handling 570 kg of air per hour.
The above account of the early history of cryocoolers is a much
abbreviated review of the historical survey given by Collins and Cannaday
(1958). This excellent book will be found in any reasonably equipped
library. Reference to it is strongly recommended.

RECENT DEVELOPMENT

Two important developments in cryocoolers occurred soon after the

Second World War. Both developments have already had a major impact,
but the principal ramifications are still developing and will eventually affect
the course of human social development as we enter what people of vision
already call "the coming age of superconductivity."
The first of these events was the development, in 1946, of a helium
liquefier by Samuel Collins at the Massachusetts Institute of Technology.
With two expansion cylinders and a compact contraflow heat exchanger it
was possible to liquefy any gas without the need for supplementary refriger-
ation. Later, a large helium-hydrogen liquefier was fabricated in 1956 at
MIT with special emphasis on a heavy-duty machine that would require
minimum maintenance. Descriptions of the Collins engine are given in
Collin's book (1958), and in his various other papers listed in the Bibliogra-
phy. The Collins helium liquefier was developed for commercial application
by the A. D. Little Co. of Cambridge, Massachusetts and was sold commer-
cially by an associated company, Cryogenic Technology, Inc., of Waltham,
Massachusetts.
Commercial development of the Collins Cryostat put the possibility
for liquid helium research in the hands of any reasonably well equipped
university or industrial laboratory. Moderate capital funds and will were
the only prerequisites. This transformed a situation previously existing
where helium research was confined to a zealous few prepared to put up
Introduction 11

with considerable difficulties and expense in their work. The transformation

coincided more or less with an explosive growth in cryogenic engineering
in the late 1950s. This arose as a consequence of long-range, rocket missile
development, and the "Sputnik" era. Funds were virtually unlimited for
military, education, and space-related activities in the United States at that
time and all through the 1960s.
One consequence was the growth of interest about phenomena at
helium temperatures, particularly in the field of superconductivity. Now,
as a result, there exist or can be foreseen manifold military and civil
applications for superconducting systems. These embrace the application
of Josephson tunnel diodes to electronic computers, the use of SQUIDs
(superconducting quantum interference devices) as sensitive mag-
netometers for all manner of dynamic measurements varying from mag-
netocardiograms of the heart to the detection of nuclear submarines.
Superconducting magnet systems will be used in fusion and magnetohy-
drodynamic power systems, in magnetic levitation rapid transit systems, in
high-speed ship propulsion, and in industrial electric motors of high power-
to-weight ratios.
The other important development after the Second World War was
the production of Stirling engines operating as refrigerators by the Philips
Company in Holland. In the late 1930s, Philips started work on small
combustion heated Stirling engines, to generate power for radios in areas
without electricity. This development continued during the war. By 1946,
an engine of one horsepower had been developed. The invention of transis-
tors and improvements to storage batteries and tubes eliminated the need
for the engine. However, work continued at Philips and has led to
Stirling engines of 5-W output for artificial hearts to large 2000-kW engines
burning solid industrial, municipal, and agricultural wastes. Engines of
150 kW have been installed in vehicles and boats with advantages in terms
of efficiency, noise, multifuel capacity, and reduced environmental pollution
(Walker, 1980).
Stirling engines can also be used as refrigerators. This was appreciated
by John Herschel as early as 1834. Later, in 1874, Alexander Kirk described
a Stirling cooling engine that had been used for ten years. However, little
development of Stirling refrigerators had occurred until 1946, when one
of the new Philips engines was driven with no heating. The unit operated
so well as a refrigerator that the researchers were astonished to attain
temperatures of the cylinder head so low that air liquefied on it. Thereafter,
research on Stirling engine refrigerators was undertaken by a separate team
led by Jan Kohler (1954). Under his direction, research proceeded apace
and the machine was ready for commercial sale by 1956. Initially, the
machine was used for air liquefaction. Shortly thereafter, a separation
12 Chapter 1

column was introduced to produce nitrogen. Other developments followed

including the invention of multiple expansion machines to attain tem-
peratures as low as 7 K incorporated in hydrogen and helium liquefiers.
Larger machines of intermediate capacity rating were developed for
industrial liquefaction. Smaller machines of miniature and microminiature
capacity were developed for military and electronic applications. Impetus
for this work arose in the 1950s and 1960s from the development of military
infrared night vision equipment and heat-seeking infrared missile guidance
systems. Sensitive infrared devices were developed with appropriate optical
characteristics at temperatures corresponding to liquid nitrogen (80 K) and
liquid hydrogen (20 K). Their use in field applications required the develop-
ment of miniature cooling engines capable of attaining these temperature
levels with moderate electrical or thermal power input to obviate the need
for liquid cryogens.
Much of the military interest and funding for research on infrared
applications was centred in the United States. As a consequence, Philips
research on small cooling engines was carried out by the North American
Philips Laboratories at Briarcliff Manor, New Y ark. Small cryocoolers were
produced at the North American Philips factory at Ashton, Rhode Island.
This concentration by North American Philips on small cryocoolers is
thought to be the consequence of various "Buy American" acts of the U.S.
Congress rather than a free choice by the Philips Company.
The manufacturing facilities at Ashton were not a commercial success
and in the early 1970s were sold to a competitor, Cryogenic Technology
Inc. (CTI), of Waltham, Massachusetts, an outgrowth of A. D. Little, Inc.,
Cambridge, Massachusetts, responsible for the commercial development
of the Collins Cryostat. With the facilities at Rhode Island went the rights
and privileges in North America of all Philips patents on Stirling cryo-
coolers. The sale included the agency for handling in the United States all
the intermediate and large Stirling cooling engines made by the Philips
company in Eindhoven. This consolidation established CTI as the leading
cryogenic engineering company in North America and therefore the world.
The United States Government is never comfortable with a single
source of supply for its requirements. It distributes knowledge of develop-
ments in technology to several suppliers and allocates purchases on
considerations other than a straightforward commercial basis. One con-
sequence has been the development of activity in small Stirling cryocoolers
since the early 1960s.
Infrared night vision equipment was in the early stages of development
at the end of the war. Missiles, notably the German V2, were at a similar
stage. The obvious attractions of night vision equipment and heat-seeking
missile guidance systems resulted in substantial funding for infrared (IR)
Introduction 13

equipment development. Good progress was made with IR sensors operat-

ing at cryogenic temperatures and by the early 1960s the real difficulty
became the development of small, light-weight, low-power, low-interfer-
ence cryocoolers capable of attaining and maintaining an IR sensor at 80 K
and, later, 20 K. Sensors with favorable optical characteristics at these
temperatures were developed in various laboratories using liquid-nitrogen
and liquid-hydrogen cooling.
From the early 1960s substantial research funds for cryocooler
development have been provided by the U.S. Government, principally the
Air Force in a program from the Flight Dynamics Laboratories of the
Wright-Patterson Air Force Base, Dayton, Ohio. Other substantial cooler
development has been supported by the U.S. Army Night Vision and
Electro-Optics Laboratory at Fort Belvoir, Virginia. The U.S. Navy also
has interests in cryocooler development and various programs have been
supported by the Naval Engineering Laboratory, Washington, D.C., the
Office of Naval Research, Washington, and the Naval Weapons Research
Laboratory at China Lake, California.
The National Aeronautics and Space Administration (NASA) is
interested in IR systems and spacecraft cryocoolers. They have supported
relatively small programs for development through the NASA Goddard
Space Flight Center, Greenbelt, Maryland.
The principal United States Government participant in this work is
the National Bureau of Standards Cryogenic Engineering Laboratory,
Boulder, Colorado, recently reconstituted as the Thermophysical Properties
Division of the National Engineering Laboratory.
Under the unique American system of easy access and wide dissemina-
tion of research information most of the achievements have been reported
either in the open literature or in reports available through the National
Technical Information Service (NTIS), Springfield, Virginia, or the Defense
Documentation Center (DDC), Cameron Island, Virginia. A list of the
available reports is contained in the Bibliography. Reports and patents are
so numerous as to justify separate listing from the open literature.
A variety of exceptionally able and competent research and develop-
ment teams for small cryocoolers have been assembled in the United States
by the contractor companies. Most of the sponsoring agencies also maintain
a small "in-house" research effort working in the frontier areas, to provide
the technical capability to supervise the work of contractors. Among these,
Philips Laboratories in Briarcliff Manor, New York, supported by the
resources of the Philips Laboratories at Eindhoven, has to be one of
the most experienced. They have a continuous program extending back
to the earliest days of the pioneer Stirling engine work by Kohler at
Eindhoven.
14 Chapter 1

Similarly, Cryogenic Technology Inc. has a broad experience in both

regenerative and recuperative cooling engines extending back through
A. D. Little Inc., to the pioneer work of Samuel Collins at the Massachusetts
Institute of Technology.
Another pioneer company in the Stirling cryocooler field in the 1960s
was the Malaker Corporation of High Bridge, New Jersey. Under the
technical leadership of John Daunt, now a professor at Stevens Institute,
Hoboken, New Jersey, the company made many excellent, innovative
cryocoolers, but failed commercially about the late 1960s/early 1970s.
Work on Stirling and Vuilleumier cycle cooling engines for missile
infrared detector and guidance systems has been in progress since the mid
1960s at the Hughes Santa Barbara research center, and from about 1967
at the Hughes Aircraft Co., in Culver City, California. Production is
concentrated at the Hughes plant at Torrance, California. The company
has made a large number (literally thousands) of very advanced Stirling
and Vuilleumier cryocoolers for missiles, far more than any other company.
They have a policy of never publishing their work in the open literature.
The NASA report by Russo (1976) is readily available from NTIS and is
a good starting point for those interested in the Hughes activity.
Another California company active in the cryocooler field is Submarine
Systems, Inc., now called Kinergetics Inc. of Tarzana. The company, led
by Ken Cowans, formerly at Hughes, has done excellent work in the field
of Vuilleumier and Stirling cryocoolers and in the development of super-
compact recuperative heat exchangers.
More recent entrants are Texas Instruments Inc. of Dallas, (Anon,
Texas Instruments, 1974), and the Martin-Marietta Corp. of Orlando,
Florida. Both are active in the development and production of the small
Stirling cryocooler like the one shown in Fig. 1.2. This is the standardized
interservice cryocooler intended for forward-looking infrared (FLIR) sys-
tems. The same unit is also in production at CTI and Magnavox, the Dutch
Philips Company subsidiary in North America, and in Europe at the
German Company AEG-Telefunken A.G.
Woodard et al. (1978) of Martin-Marietta have reported a program
of manufacturing methods for split-cycle Stirling cooling engines for the
U.S. Army. A split-Stirling device was also featured in the Texas Instrument
report cited above.
William Gifford, recently deceased, a coinventor at A. D. Little of the
Gifford-McMahon cryocooler, contributed much to the cryocooler field.
Gifford, a distinguished professor of engineering at the University of
Syracuse, New York, founded Cryomech Inc. of Jamestown, New York.
Cryomech manufactures Gifford-McMahon cryocoolers in a range of sizes
for civil and military applications. Gifford-McMahon cryocoolers are rela-
Introduction 15

COMPRESSOR CYLINDER HEAD

PISTON

HELIUM PRESSURE
TRANSFER

EXPANSION CHAMBER

COLD CAP SPRING LOADED

AGAINST DE lECTOR

PINCH TUBE
IHELIUM FILL]

L- CRANKCASE

Fig. 1.2. Small integral Stirling cooling engine for forward looking infrared applications.

tively large, heavy machines with separate compressor and expander units
characterized by exceptional durability and long life (Higa et al., 1977).
Gifford was a prolific writer and the Bibliography contains many entries
indicative of his exceptional activity.
Gifford-McMahon machines are manufactured by CTI, Waltham,
Massachusetts, and by Air Products and Chemicals Inc., Allentown,
Pennsylvania, (APCI). Ralph Longsworth of APCI, a former graduate
16 Chapter 1

student and favored protege of Professor Gifford, calls the APCI machine
a "modified Solvay engine." It is indistinguishable from the Gifford-
McMahon engine. Those wishing to clarify the distinction should refer to
Chapter 2-Elementary Thermodynamics of Cryocoolers.
The exceptional sensitivity of infrared detectors and superconducting
devices requires that the cold region of the cooling engine be isolated from
mechanical shock and vibration, thermal variation, and electromagnetic
interference. For systems mounted in gimbals to facilitate rotary or swivel
action it is important to keep the mounted mass as small as possible. Much
ingenuity has been exercised in efforts to satisfy these demands. Early
success was achieved by condensing liquid nitrogen on a cold cylinder head
and transferring the droplets along a tube to the cold region by the
evaporative boiling process known as "Leidenfrost boiling." However, this
technique appears to have been supplanted with the development of cooling
engines comprised of separate or "split" expander and compressor systems.
Gifford-McMahon cryocoolers were always of this type. They have
valves, two or four in number, with two connecting lines, one supply and
one return between the expander and compressor units. Miniature split-
Stirling cryocoolers are a comparatively recent development, although the
arrangement whereby a displacer operates in a separate cylinder which is
independently driven dates from the Stirling brothers' second engine in
1827. This was a standard beam engine intended for steam drive but
modified to operate as a hot-air engine. Many later hot-air engines had
separate cylinders and drive arrangements including the Robinson and
Heinrici engines (see Walker, 1980). The Kirk engines of 1860 were also
of this type, as was the free-displacer engine invented by Postle (1873)
and by Ringbom (1905). However, it does not appear to have been widely
used for cooling engines until quite recent times. Credit for the reinvention
of free displacer cooling engines belongs to Walter Higa of the Jet Propul-
sion Laboratory, Pasadena (Higa, 1965). Similar concepts appear to have
occurred to Stuart Horn at the Night Vision Laboratory and to Fred Chellis
at CTI about the same time.
The great advantage of the split-Stirling cryocooler is that it has no
valves and only one connecting tube. There are disadvantages, principally
in moving the displacer in properly synchronized relation to the compressor
piston. Despite this, the unit is sufficiently attractive to justify the present
concentrated development effort.
Another company active in Vuilleumier cryocooler development was
AiResearch Manufacturing Company of Phoenix, Arizona working on
contract for the NASA Goddard Space Flight Center, Greenbelt, Maryland.
Browning et al. (AiResearch, 1974) gave a comprehensive summary report
on this program.
Introduction 17

Military and spacecraft applications have severe weight and volume

requirements. To attain these, the price must be paid in terms of regular
maintenance and relatively limited life. There are many civil applications
for cryogenic coolers of low capacity where the reverse is true. The impor-
tant parameters are reliable operation with no maintenance and long
lifetimes at low cost. There are virtually no restraints on size, weight, or
efficiency. Applications include SQUID magnetometers for, say, routine
magnetocardiograms in the offices of a physician and application of Joseph-
son tunnel diodes in computers.
A cryocooler that has the potential to fulfil this function was described
by Zimmerman and Radebaugh (1977). The unit is shown in Fig. 1.3. It
was constructed of a readily available, low-cost fiberglass filament epoxy
tube with a solid nylon displacer. The machine was capable of achieving
a minimum temperature of 8.5 K. A subsequent special version of the
engine achieved a minimum value of 3.5 K. This memorable achievement
coincided with the visit by the author to the U.S. National Bureau of
Standards Boulder laboratory of Zimmerman. At that low temperature,
the helium working fluid (at a pressure substantially less than 1 atm)

Glass Reinforced
Plastic Cylinder

Vacuum Space

Radiation Shields
Fig. 1.3. Diagrammatic cross section of
low-speed, low-cost, long-life cryogenic
cooling engine (after Zimmerman and
Radebaugh, 1977).
18 Chapter 1

liquefied in the cylinder. It was the first time a two-phase single-component

working fluid regenerative cryocooler had been successfully operated.
Such low temperatures are not necessary for the operation of supercon-
ducting devices. Transition to the superconducting state of niobium metal
occurs at 9 K. Research on new superconducting materials is pushing the
transition temperature towards 20 K.
The development of superconducting electronic devices has stimulated
the interest of many companies with no previous history in the cryogenic
field. There is every possibility that new concepts for low-cost maintenance-
free cooling machines for civil applications will emerge from these
companies untrammeled by the constraints imposed by military or space
considerations.
Another important emerging civil field for cryogenic refrigeration is
that concerned with superconducting power. The U.S. Navy has, for many
years, been investigating and assessing the prospects for superconducting
power for ship propulsion. There is increasing interest in electric power
transmission by superconducting cable. This will entail a virtual pipeline
of liquid or vapor phase helium with a liquid-nitrogen or natural gas shroud
and having terminal and intermediate on-line cooling facilities.
Other superconducting power prospects with concomitant cryogenic
cooling requirements include the intense magnetic fluxes required for fusion
power and the magnetic levitation of trains and rapid transit systems.
The present explosive growth in the liquid natural gas (LNG) industry
is likely to become a significant contributor to future needs for cryocoolers
of intermediate capacity at relatively high temperatures. As LNG storage
and pipeline transmission facilities develop, the need for intermediate
cooling and reliquefaction systems will be established.
So far, attention has been concentrated principally on refrigerating
cryocoolers. These are capable of attaining 80 K in a single stage of
expansion. Multiple expansion machines of two, three, or four stages can
attain temperatures of 20, 10, and 6-7 K, respectively. Most applications
in the future will lie in this range. However, there will always be a demand
for lower temperatures and so a brief outline of very low temperature
cooling systems is included.
The best known system is the Joule-Thomson (JT) cryocooler. When
a gas expands from a high pressure to a low pressure through a restricted
orifice, the expansion occurs at constant enthalpy. If the initial temperature
is below a critical value called the inversion temperature, the expansion
will result in a decrease in temperature.
Air, and its components oxygen and nitrogen, can be readily liquefied
by expansion in a JT system without precooling, although, of course, this
helps. Hydrogen and helium can also be liquefied, but less readily, and
Introduction 19

need to be precooled (perhaps by an expansion engine) to a temperature

below the inversion temperature before expanding. Many gases are still
liquefied in commercial quantities by systems incorporating JT coolers,
complex contraftow heat exchangers and reciprocating or turboexpansion
engines.
Considerable effort has been invested in the development of miniature
and microminiature capacity coolers for military infrared systems for air-
craft, and for hand-carried portable and vehicle-mounted Army use. The
systems have virtually instantaneous cool-down characteristics and the cold
region is small, lightweight, and without vibration, and mechanical or
thermal, and electromagnetic interference. Despite these advantages, JT
coolers have some drawbacks that have inhibited their widespread use and
which appear to be intractable. The first is that very low gas flow rates
must be maintained since very small refrigerating effects are required (less
than 1 W at 80 K). This involves use of an expansion orifice of minute
dimension which can become blocked by small solids, either dust, debris,
or frozen water and carbon dioxide. The expansion gas must, therefore,
be ultraclean and free of contaminants. This eliminates use of atmospheric
air and involves the use of nitrogen in compressed gas bottles. Compressed
gas bottles are heavy and suitable for only short-term operation. The
logistical and replacement problems are simply not practical for field use.
Furthermore, in battle situations the high-pressure bottles can be dangerous
if ruptured by a bullet or shrapnel. The British company, Hymatic Engineer-
ing Ltd. is well known in the field of portable JT coolers.
For temperatures below the liquefaction temperature of helium (4.2 K)
there are three principal methods of refrigeration:
i. Dilution refrigerators.
ii. Magnetocaloric refrigeration.
iii. Electrocaloric refrigeration.
A review of these methods of very low temperature refrigeration can
be found in Chapter 12, and in excellent books by Lounasmaa (1974) and
Betts (1974), and in the survey papers by Steyart (1978).
This historical review is far from complete. We have entirely neglected
the history and development of large-scale cryogenic refrigeration for gas
liquefaction, tonnage oxygen production, and the like. This properly
belongs to a future companion volume directed specifically to large-scale
systems and turbines. We have also neglected many contributions to the
development of cryogenic technology made in Europe, by the industrial
organizations Linde Aktiengesellschaft Wiesbaden in Germany, L'Air
Liquide Ltee in Paris, the British Oxygen Company in London, and cryo-
genic laboratories at the Universities of Oxford, Delft, and elsewhere.
20 Chapter 1

There is a long tradition of cryogenic engineering in Russia evident

by the recent (1978) award of a Nobel Prize toP. Kapitza. We have sought
to fill this omission by inclusion of Chapter 13 contributed by Professor
Ivan Evgeny Mikulin of the Moscow Baumann Technical High School.
There is also an appreciable level of cryogenic activity in Japan. This is
reviewed in Chapter 14, contributed by Yoshiro Ishizaki of the University
of Tokyo.
In our review of the cryogenic literature, Chapter 16, sources of
information are identified for those seeking to rectify other deficiencies.
In conclusion, mention must be made of other research not specifically
concerned with cryocoolers, but so pertinent to it as to be of substantial
interest. Over the past decade research on artificial hearts, using radio-
isotope heating and regenerative engines to produce work, has been funded
by the U.S. National Institutes of Health. The problems encountered by
companies engaged in this work are similar to those encountered with small
regenerative cooling engines. The artificial heart program is described in
a recent book on Stirling engines by Walker (1980).
Free-piston Stirling engines are of substantial research interest both
as prime movers and as refrigerating machines. Chapter 11 in the above
book on Stirling engines was contributed by William Beale, the inventor
of the free-piston Stirling engine. A companion volume devoted entirely
to free-piston Stirling engines is in course of production (Walker and Senft,
1983). Philips at Eindhoven already have in commercial production free-
piston Stirling cryocoolers. ERG Inc., of San Francisco, and Sunpower Inc.,
Athens, Ohio, were recently awarded development contracts by the NASA
Goddard Space Flight Center, Greenbelt, Maryland.

STATUS SURVEYS

Various surveys of cryocooler systems have been attempted. One of

the most extensive and comprehensive compilations was made by Stro-
bridge (1974) of the National Bureau of Standards Cryogenic Engineering
Laboratory. He solicited data from all possible sources and compiled these
in the form of charts reproduced herein. A total of 144 refrigerators and
liquefiers are included, 21 from Europe, the remainder from the U.S.
He divided the systems into three groups on the basis of refrigeration
temperature:
Group A-1.8 to 9 K
Group B-10 to 30 K
Group C-30 to 90 K
Introduction 21

He further distinguished the various systems as follows:

1. Old, those included in a previous survey (1969);
u. New, systems not included in the previous survey;
iii. Under development, as the name implies, systems not available
commercially but at an advanced stage of development.
To maintain his guarantee of confidentiality to those providing the data,
Strobridge offers no information permitting the identification of any of the
systems shown, nor was the attempt made to compare the various types
of system or distinguish between Stirling, Vuilleumier, Gifford-McMahon,
etc.
The Strobridge report contains the four charts reproduced in Figs.
1.4-1.7. These show the efficiency, mass, and volume of over 100 cryo-
coolers as a function of the refrigerating capacity ranging from less than a
watt to 1 MW. Mean characteristics have been drawn on all four charts.
Costs are presented as a function of input power rather than refrigerating
capacity.
The efficiency is presented as a percentage of the Carnot value defined
as 71 =actual COP/Carnot COP (COP=coefficient of performance). The
volume and mass are believed to be of a complete operational system. The
cost is the quoted price (in 1974 U.S. dollars) for low-volume purchases,
i.e., one to five units. Strobridge recommends that costs be estimated by
the equation C = 6000P 0 ·7 , where C is the cost in dollars and P is the
installed power. An allowance of 10% per annum for inflation would raise

+t
10 2
- f-- .
--
--
f--+-
•
~ fi' "'
'""'- f-~ i

•• •
l>
0 r.· •0 '
~ '
1-lftj 0
I
NEW OLD -
f-- 0 ~ •·18·9K-
-I- l> 9 A - 10 • 30 K _
l>
O¢

-+-II
• · 30-90 K
Lunder Development _

,I, l I 1,1,1,1 I 1,1,1,1 I

ld 102 103 104 105 10°
CAPACITY, WATTS

Fig. 1.4. Efficiency of cryogenic cooling systems as a function of refrigerating capacity (after
Strobridge, 1974). (Efficiency=actual COP/Carnot COP.)
22 Chapter 1

Fig. 1.5. Volume of cryogenic cooling

K>1 10t 10 3 oo• system as a function of refrigerating
CAPACITY ,WATTS capacity (after Strobridge, 1974).

the value of the constant from 6000 in 1974 to 10,000 in 1980, 17,000 in
1985, and 25,000 in 1990.
Figures 1.5 and 1.6 can be used to determine the volume and mass
of cryocoolers as a function of refrigeration temperature as shown in Fig.
1.8. In this figure, characteristics are sketched in for refrigerating capacities
of 1, 10, 100, and 1000 W.
Charts such as these have obvious attraction for planning purposes.
It must be recognized their value is rather limited. At first glance there
appears to be a surprisingly close similarity for all the systems. However,
the charts have logarithmic scales and small chart separations can be
differences of several times. For example, consider Fig. 1.4. At 1 W capacity
the range of variation in efficiency is from a low value of 0.4% to a high
value of 10%-a range of variation of 25 times. Similarly, at 1000 W
capacity the range of variation is between high and low values of 50% and
7%, a range of 7 times.
A most important factor not addressed in the Strobridge charts is the
useful life or the maintenance requirements of the different systems. A
cryocooler for aircraft or space use would likely be small, lightweight,
highly pressurized, fast acting, and consequently have a short life or operat-
ing period between maintenance. The same equation relating life, strain,
size, weight, and cost applies inexorably to cooling engines as to all other
machines or biological systems. Only the exact terms of the equation are
unknown and must be defined by experience.
Introduction 23

Fig. 1.6. Mass of cryogenic cooling sys-

tems as a function of refrigerating capac-
ity (after Strobridge, 1974). CAPACITY, WATTS

An extensive study of cryocoolers for spacecraft to provide long-term

storage of cryogenic fluids was carried out by Jensen et al. (1971) for the
NASA Manned Spacecraft Center. The four volumes of the Jensen study
contain a wealth of information. The study is now somewhat dated but few
fundamental changes have occurred so that the material remains a valid
reference.

-gto'~t=~~~~~~~~~=1iit=~ftrt-~~~
u

101
INPUT POWER, kW

Fig. 1.7. Cost of cryogenic cooling system as a function of input power (after Strobridge, 1974).
24 Chapter 1

~ 10°r-~~~~~~~~~-+----~~
:::l
..J

~ 10- 1 r-~"d--~~----="""+-__:::,~----~~

'o-•L-------'-----L____::..______r:-_____L___j
10°r----,----,-----,----.----.-~

-,;
~

;
V>
,o•r-~~~~4-~~r---~~~4-~
"'~ IO'r-----t--~~--~~---f~--4-~
Fig. 1.8. Volume and mass of cryogenic
100 0~--~20,---~--~~---.~0~--~,0~0~ cooling systems as a function of refriger-
ation temperature (prepared from Figs. 1.5
TEMPERATURE (K) and 1.6) (after Strobridge, 1974).

Their findings are summarized on Figs. 1.9 and 1.10, reproduced from
the Jensen report. These show the specific weight (kilograms per watt) and
the coefficient of performance of a variety of cryogenic cooling systems as
a function of refrigerating capacity at 77 K.
The cryocoolers considered include the following:
i. A group of Ericsson cycle systems, Gifford-McMahon, Taconis,
and Solvay;
ii. Joule-Thomson systems;
iii. Brayton cycle systems;
iv. Vuilleumier cycle engines;
v. Stirling cycle engines.
In their report, Jensen et al. presented similar curves for cryocoolers
providing refrigeration at 20 and 4.2 K. They paid particular attention to
the problems of heat rejection by radiator, and assumed this to be included
in the total system.
A feature of Figs. 1.9 and 1.10 is the relatively low specific weight
and the high coefficient of performance of the Stirling cryocoolers. Another
interesting comparison is the order-of-magnitude difference between
Introduction 25

.."
"
0.

'"
..
(I)
(I)

:I;

10 100
CAPACITY (WATTS)

Fig. 1.9. Specific weight of cryogenic cooling systems as a function of capacity at 80 K (after
Jensen et al., 1971).

.."'
0
z
2
a:
...a:
0

"'
...
Cl.

0
,_
z
"'0
...... 001~----~~~~~~----------~~~~~~~~----1

"'0
0

CAPACITY (WATTS)

Fig. 1.1 0. Coefficient of performance of cryogenic cooling systems as a function of capacity

at 80 K (after Jensen eta/., 1971).
26 Chapter 1

ground-based and flight weight Gifford-McMahon, Taconis, and Solvay

group of machines.
Earlier, Daunt and Goree (1969) carried out a survey of miniature
cryogenic refrigerators for the U.S. Navy Office of Naval Research. This
survey contains a large compendium of data for cooling engines then extant.
The report is still of substantial interest but the information contained
therein is now more than a decade old and could not be regarded as a
contemporary review.
Another very comprehensive survey of cryogenic cooling techniques
for space-borne systems was carried out by Donabedian (1972). This
included an evaluation of mechanical closed-cycle cryocoolers, open-cycle
expendable systems, passive radiation concepts utilizing the low tem-
perature of deep space, and, finally, thermoelectric devices. This report is
highly recommended. A survey of cryogenic coolers for space systems,
carried out by Vance (1974), is not as comprehensive as the Jensen or
Donabedian reports cited above. Cox (1974) reviewed potential cooling
methods for an underwater naval detection system using a cryogenic
SQUID. He concluded that the choice of cryocooler would lie between
Claude or Stirling cycle engines in conjunction with a JT valve.
Morgan (1971) of Hughes Aircraft carried out a study for the U.S.
Air Force of airborne air liquefiers. Basically, the study involved a com-
parison of Stirling and Vuilleumier cycle liquefiers capable of producing
2.3 kg per hour of liquid air for use in flight. Various parametric studies
were completed and preliminary designs prepared for two systems. The
Stirling system was found to be more efficient, lighter in weight, smaller
in size, and cheaper to make than the Vuilleumier machine. The only
advantage offered for the Vuilleumier machine was a presumed longer life
due to reduced mechanical loading.
A very comprehensive survey of cryogenic refrigerators for supercon-
ducting transmission cables was made by Kadi and Longsworth (1976).
This survey is of particular interest as it was directed to intermediate and
large capacity systems involving both reciprocating and rotary systems.
Much technical information of existing or proposed systems in Europe and
North America is given.

REFERENCES

AiResearch Manufacturing Co. (1974). Fractional Watt Vuilleumier Cryogenic Refrigerator

Program Engineering Notebook, Vol. 1, Thermal analysis. Contract NAS 5-21715, God-
dard Space Flight Center, Greenbelt, Maryland, May.
Introduction 27

Anon-Texas Instrument Equipment Group. (1974). "Final Engineering Report on the

Design and Development of Two Miniature Cryogenic Refrigerators." Contract No.
DAAK02-73-0495, Night Vision Lab., Ft. Belvoir, Virginia, NTIS Report No. AD
784436.
Betts, D. S. (1974). Refrigeration and Thermometry Below 1 K. Sussex Univ. Press, Falmer,
Sussex (in the U.S., Crane Russak and Co. Inc., 347 Madison Avenue, New York).
Collins, S.C., and Cannaday, R. L. (1958). Expansion Machines for Low Temperature Processes.
Oxford University Press, Oxford.
Cox, J. E. (1974). "Potential Cooling Methods for an ELF SQUID." Naval Res. Lab., NRL
Memorandum Report 2899.
Daunt, J. G., and Goree, W. S. (1969). "Miniature Cryogenic Coolers." Office of Naval
Research, NTIS AD 860866.
Donabedian, M. (1972). "Survey of Cryogenic Cooling Techniques." U.S. Air Force, SAMSO
Report No. SAMSO-TR-73-74, AD No. 755 780, NTIS, Springfield, Virginia.
Higa, W. N. (1965). "A Practical Philips Cycle for Low Temperature Refrigeration." Cryogenic
Techno/. 8, 203-209, (July-August).
Higa, W., and Wiebe, E. (1977). "One Million Hours at 4.5 Kelvin. Appl. of Closed Cycle
Cryocooler to Small Superconducting Devices", Proc. of Conf. NBS Boulder, Oct., pp.
99-108 (issued as NBS Spec. Pub. 508, April1978).
Jensen, H. L., Nast, T. C., Glassford, A. P. M., Vernon, R. M., and Ekern, W. F. (1971).
"Investigation of External Refrigeration Systems for Long-Term Cryogenic Storage."
Summary Report, Report No. NASA CR 115192, NASA Manned Spacecraft Center,
NTIS Accession No. N71-37107 (Lockheed Missiles and Space Co., Sunnyvale, Califor-
nia) (February 1971), See also, same authors, same title: (a) System Review Report,
NASA CR 115191, NTIS N71-37105 (May 1970); (b) Final Report, NASA CR 114920,
NTIS N71-20279 (February 1971).
Kadi, F. J., and Longsworth, R. C. (1976). "Assessment and Study of Existing Concepts and
Methods of Cryogenic Refrigeration of Superconducting Transmission Cables." E.R.D.A.
Report No. C00-2552-6, NTIS, Springfield, Virginia.
Kirk, A. (1874). "On the Mechanical Production of Cold," Proc. Inst. Civil Eng. 37, 244-315
(London).
Kohler, J. W. L., and Jonkers, C. 0. (1954). "Fundamentals of the Gas Refrigeration
Machine." Philips Tech. Rev.16(3), 69-78; See also "Construction of the Gas Refrigerat-
ing Machine," Philips Tech. Rev. 16(5), 105-115.
Lounasmaa, 0. V. (1974). Experimental Principles and Methods below 1 K. Academic Press,
New York.
Morgan, N. E. (1971). "Analysis and Preliminary Design of Airborne Air Liquefiers."
AFFDL-TR-71-171, WPAFB, Dayton, Ohio, NTIS AD 892500 (Hughes Aircraft Co.,
Electro-Optical Division, Culver City, California).
Russo, S. C. (1976). "Study of a Vuilleumier Cycle Cryogenic Refrigerator for Detector
Cooling on the Limb Scanning Infrared Radiometer." Report No. NASA-CR-145078,
NTIS N77-11211/8ST.
Siemens, C. W. (1882). Proc. Inst. Civil Eng. 68, 179-186.
Solvay, E. (1887). Deutches Reichspatent No. 39280.
Steyart, W. A. (1978). "Magnetic Refrigerators for Use at Room Temperatures and Below."
Report No. LA-UR 78-1764, Los Alamos Scientific Laboratory, New Mexico.
Strobridge, T. R. (1974). "Cryogenic Refrigerators-An Updated Survey." National Bureau
of Standards Tech. Note 655 (Supt. Documents, U.S. Govt. Printing Off.) (12 pp.).
Swearingen, J. S. (1947). "Expansion Turbines for Low Temperature Processing." Trans.
Amer. Inst. Chern. Eng. 43, 83-90.
Vance, R. W. (1974). "Cryogenic Coolers for Space Systems." Report No. SAMSO-TR-74-
180, Space and Missile Systems Organization, A.F. Systems Command, Los Angeles,
A.F. Station, NTIS AD-785-083 (Aerospace Corp., El Segundo, California).
Walker, G. (1980). Stirling Engines. Oxford University Press, Oxford.
Walker, G., and Senft., J. (1983). Free-Piston Stirling Engines (in preparation).
28 Chapter 1

Woodard, R. S., Welch, P. H., and Jansson, R. M. (1978). "Manufacturing Methods and
Technology for the Establishment of Production Techniques for a Split-Cycle Stirling
Cryogenic Cooler," Report No. 15181, Contract DAA B07-77-C-0631, U.S. Army
Elec. R. and D. Comm., Fort Monmouth, New Jersey (Martin-Marietta Corp., Orlando,
Florida).
Zimmerman, J. E., and Radebaugh, R. (1977). "Operation of a SQUID in a Very Low-Power
Cryocooler. App. of Closed Cycle Cryocoolers to Small Superconducting Devices," Proc.
of Conf. NBS, Boulder, October 1977, pp. 59-66 (Issued as NBS Spec. Pub. 508, April
1978).
Chapter 2

Elementary Thermodynamics of
Cryocoolers

INTRODUCTION

The objective of this chapter is to review the various operating cycles of

cryocoolers. For this to be meaningful it is necessary to be acquainted with
the elements of thermodynamic analysis. For the benefit of those without
such acquaintance we include a brief thermodynamics review. To the purists
we offer apologies for liberties taken with the subject in the interest of
brevity. A more rigorous treatment may be found in university engineering
texts (see Rogers and Mayhew, 1967).

Part I: Thermodynamics Review

Matter can exist in three phases: gas, liquid, and solid. It can be
changed from one phase to another by the transfer of energy to or from
the system (the particular element of matter under consideration). It is a
matter of common experience that to melt ice (water in the solid phase)
we supply heat, whereupon the ice is converted to liquid. Further heating
causes the liquid to boil and evaporate to vapor or gas. The difference
between vapor and gas is that in one case (vapor) the temperature is below
and in the other case (gas) is above the critical temperature.

STATE PROPERTIES

The state of a substance can be defined by the state properties. These

are various properties of the substance that characterize it at any given
condition. There are six state properties of primary interest in thermody-
29
30 Chapter 2

namics and others of less interest for specialized applications. The combina-
tion of any two properties is sufficient to completely specify the state of a
pure (single component) substance. The six properties of general interest
are: pressure, volume, temperature, internal energy, enthalpy, and entropy.
Consider now the simple closed system shown in Fig. 2.1 consisting
of a piston and cylinder closed at one end. Assume the piston has no mass
and is free to slide without friction in the cylinder. We assume further
there are no leaks of the working fluid past the piston seal. Finally we
assume that some known mass of working fluid is contained in the cylinder,
say 1 kg. A weight of mass M placed on the piston will cause it to descend
until it comes to rest as shown in Fig. 2.1, a distance L from the bottom
of the cylinder.

Volume
For the system shown in Fig. 2.1 the state property volume, V, is the
total volume of the working fluid contained in the cylinder with diameter,
D, below the piston, i.e., V = (7r/4)D 2 L. We have unit mass of working
fluid in the cylinder so the volume measured is the specific volume.

Pressure
Pressure is the force exerted on the piston to keep it "floating" in the
cylinder a distance L from the bottom expressed as the force exerted per
unit area of the piston. In the system shown in Fig. 2.1 the total downward
force is the gravitational force acting on the mass M on the piston.
Elementary Thermodynamics of Cryocoolers 31

From Newton's law, force= mass times acceleration, we know the

downward force, F, is
F=Mg (2.1)
The downward force is resisted (and balanced exactly) by the upward force
exerted on the piston by the working fluid acting uniformly over the whole
area of the piston, A:
pA=F (2.2)
the pressure can therefore be calculated simply by dividing the known
downward force by the area of the piston so that
p=F/A=Mg/A (2.3)

Temperature
The temperature, T, indicates the energy level of the working fluid.
Using the concepts of the kinetic theory of gases one can assume the
working fluid to be composed of molecules consisting of "billiard balls" of
matter either singly or in groups of two or three moving about in random
fashion as well as vibrating and rotating. The temperature is the degree of
activity of the billiard balls. Application of heat causes the billiard balls to
become more agitated and move faster with increased random collisions.
This is perceived as an increase in the temperature.
Temperature is measured on four scales: Centigrade, Fahrenheit,
Kelvin, and Rankine. Two useful temperatures are the ice-point and boiling
point of water at atmospheric pressures. These correspond to ooc and 100°C
for the Centigrade scale or 32°F and 212°F for the Fahrenheit scale.
The lowest temperature that can exist is -273.15°C. This is taken as
the starting point or zero datum of the Kelvin scale. The Centigrade unit
is the same as the Kelvin unit of temperature change. The absolute
minimum temperature (-273.15°C) corresponds to -459.67°F. This is
taken as the zero datum of the Rankine scale. The Fahrenheit unit is the
same as the Rankine unit of temperature change.

Internal Energy
Internal energy, U, is similar to temperature as a measure of the energy
level of the working fluid. According to Joule's law, the internal energy of
a perfect gas varies only with the temperature. In the special case of a
perfect gas the change in internal energy is the change in temperature
multiplied by the specific heat at constant volume, Cv, i.e.,
dU=CvdT (2.4)
32 Chapter 2

Enthalpy
The enthalpy, H, is property of state defined by three other state
properties in combination:
H=U+pV (2.5)
It is a useful measure to facilitate calculations for continuous flow systems
where combinations of the internal energy and the "flow work", p V, occur
frequently.
For a perfect gas the change in enthalpy is simply
dH=CpdT (2.6)
where CP is the specific heat at constant pressure.*

Entropy
Entropy, S, is a property of state useful in calculating the quantities
of energy with respect to temperature exchanged in changing from one
state to another. It is defined by the equation
dS = J(dQ)/T (2.7)
The change in entropy is the integral sum of the quantity of heat transferred,
Q, divided by the temperature at which it is transferred.
Books have been written about entropy and students become confused
by it. Everything that happens causes an increase in the entropy. It becomes
difficult to understand how this incessant aggregation does not result in
the whole world getting top-heavy with entropy. To the newcomer not of
philosophical bent, entropy is best regarded simply as a handy way to
calculate heat transfers.

Equations of State and Thermodynamic Tables

The above six properties of state (temperature, pressure, volume,
internal energy, enthalpy, and entropy) are interrelated, and any two, say
the pressure and temperature, completely define the state of the substance.
In practice gases never are perfect, but many approximate this condition.
A perfect gas is one that obeys the simple equation of state:
pV=mRT (2.8)
where m is the number of moles of gas present and R is a constant that
*Since pV=RT then dH=CvdT+pdV=CvdT+RdT. Since (Cv+R)=Cp then dH=
CpdT.
Elementary Thermodynamics of Cryocoolers 33

is characteristic of the gas and equal to CR/ M), where R is the universal
gas constant and M the molecular weight of the fluid.
Many gases follow an equation of state:
pV=ZmRT (2.9)
Z is the coefficient of compressibility, a factor which varies widely as the
condition of the gas changes. It is determined experimentally. More compli-
cated equations of state exist, or the properties may be determined experi-
mentally and summarized in tables of thermodynamic data.
Properties of state for various fluids have been measured over many
levels of temperature and pressure. The results are best presented in tabular
form called thermodynamic tables. The tables are usually in two parts. The
first part contains properties of the saturated fluid, the second part contains
properties of the superheated fluid.

THE TEMPERATURE-ENTROPY (T-S) PLANE

The distinction between saturation and superheated fluid is shown in

Fig. 2.2, where the temperature, T, is plotted as function of the entropy,
S. It is called the T-S plane. Assume the simple piston/cylinder system,
shown in Fig. 2.1, is at such a low temperature that the working fluid is
all liquid. We can represent this by point A on Fig. 2.2. If the cylinder is
heated, the temperature will rise and the working fluid will expand and
cause the piston to move up the cylinder. The pressure will remain the
same because the mass on the piston is unchanged. This will continue until,
at point B, the liquid working fluid is saturated (saturated liquid). Further
addition of heat will cause evaporation of the working fluid. This evaporated
vapor will be at the same tempearture and pressure as the liquid but will
have radically different values of specific volume, entropy, enthalpy, and

ISOTHERM

~ \
T
D C

DEW-POINT

BUBBLE -POINT
Fig. 2.2. Heating and evaporation process
displayed on the temperature-entropy
plane. s
34 Chapter 2

internal energy than the liquid. Let that state be represented by point C.
As more heat is added an increasing fraction of the working fluid will
evaporate. The temperature and pressure remain the same but the piston
will move a long way up the cylinder. The volume of the evaporated state
is much greater than the liquid state. The heat added to cause evaporation
of the working fluid is called the latent heat (or enthalpy) of vaporization.
In the process of vaporizing we can think of the fluid moving from
point B to point C. At some intermediate point, say D, we can talk about
the fluid having a "vapor fraction" x. This is the fraction of fluid at the
saturated vapor state C. It may be calculated simply as

. length ED Sv -SB
vapor fraction x = C (2.10)
length B Sc- SB

The fluid is never actually at any intermediate state for it can only exist at
either state B or state C. It is impossible for any particular element of the
fluid to be at state D. The intermediate state is simply a concept to help
visualize the fraction of fluid that has evaporated. The vapor fraction x is
simply the fraction of the total mass that has been evaporated from state
B to state C. Eventually all the liquid evaporates to vapor at state C called
saturated vapor. Further heat supplied to the cylinder causes the fluid to
increase in temperature along the line CE. It is then said to be superheated.
The line ABDCE is a line of constant pressure called an isobar. A
line of constant entropy, say FG, is called an isentrope and a line of constant
temperature, say HJ, is called an isotherm. If necessary, we could also show
lines of constant volume (isometric) or constant enthalpy (isenthalp) and
constant internal energy.
Point B is called the bubble point, where vapor first appears when
ascending the isobar from A to E. Point C is called the dew point where
liquid droplets first appear when descending the isobar from E to A.
If a larger mass were to be placed on the piston the pressure would
increase. If the above experiment were repeated, the same process of
evaporation would occur but at different values of temperature and entropy.
This is illustrated in Fig. 2.3. The new high-pressure isobar KLMN is shown
along with the original isobar ABDCE. Evaporation at the higher pressure
occurs at a higher temperature and starts, at point L, at an increased value
of the entropy. The increase in entropy during evaporation (SM -Sd is less
than the increase (Sc- SB) at the lower pressure.
Further isobars could be obtained for pressures greater and smaller
than the two considered so far. As the pressure increases, vaporization
occurs at higher temperatures. The loci of the bubble points BL and the
dew points CM may be drawn to obtain the envelope BLPMC. These loci
Elementary Thermodynamics of Cryocoolers 35

CRITICAL- POINT

~ GAS

7-----~A~R
:::--L~~;•_;.o-:~t--M
-------- .
T / •
.. E

. -------.------·.
.
A
e
D

LIQUID ... VAPOR

Co
•
. •

Fig. 2.3. Elements of the temperature-

entropy diagram for a fluid in the liquid-
vapor region. s

are called the saturation curves. BLP is the saturated liquid line. CMP is
the saturated vapor line. The area encompassed by the saturation curves
is called the liquid-vapor envelope. Within the envelope, fluid exists partly
as saturated liquid and partly as saturated gas. Fluid at any state to the
right side of the envelope exists as a superheated vapor or gas. Fluid on
the left side of the envelope exists only as liquid.
The apex of the liquid-vapor envelope, point P, is called the critical
point. Properties of state at that condition are called the critical properties,
critical pressure, critical temperature, etc. At the critical point no distinction
can be made between a saturated liquid and a saturated vapor-they are
the same fluid.
Above the critical temperature, no liquid can exist whatever the
pressure exerted on it. The critical temperature is therefore taken as the
breakpoint distinguishing gas from vapor. The critical isotherm QPR is
shown on Fig. 2.3. Fluid above this temperature is said to be gaseous. Fluid
in the evaporated state below this temperature is said to be vapor.

THE T-S DIAGRAM

Temperature-entropy diagrams of T-S planes have been produced

for most known fluids. On such charts it is customary to reproduce lines
of constant enthalpy and volume as well as isobars.
T-S diagrams are widely used in the analysis of engines and thermody-
namic processes and will be referred to many times in the ensuing discussion.
They are particularly useful for estimating quantities of energy or heat
transferred when a fluid changes from one condition to another. For
example, consider the change in state from condtion A to condition B
represented on Fig. 2.4. There are an infinite number of process paths
36 Chapter 2

F Fig. 2.4. Representation of thermodynamic processes on

the T-S diagram.

which the system may change from state A to state B. Three such paths
are shown. The most direct ,route between A and B is the straight line
process ACE. An alternative involving some high temperatures is the path
ADB. Another alternative involving low temperatures is the process AEB.
Whatever the path followed the state properties at B will always be the
same. The final pressure, temperature, volume, or entropy is independent
of the process. However, the energy involved in accomplishing the change
depends greatly on the process.
It is useful to recall that entropy was defined as dS = JdQ/T. From
this it can be readily seen that rearrangement of the equation results in
dQ = JT dS, the heat transfer is equal to the integrated area of the T-S
diagram. Thus, the energy transfer of the process can be estimated by
computing the area enclosed on the T-S diagram. In Fig. 2.4 the energy
transfer associated with the process AEB is the shaded area enclosed by
AEBFG. This is the least energetic process of the three shown. Process
ACE would encompass a greater area and process ADB a still greater
area with increased energy flows.

THE PRESSURE-VOLUME (P-V) DIAGRAM

Another graph of exceptional importance to thermodynamic analysis

is the pressure-volume diagram or P- V plane frequently called the work
diagram. Areas on the P- V plane represent work done just as areas on
the T-S plane represent heat transfers.
Consider again the piston/ cylinder system shown in Fig. 2.1. At the
start let the temperature be low so the volume is small. This could be
represented on the P-V and T-S planes shown in Fig. 2.5 by point A.
Now let the cylinder be heated. This will cause the temperature to increase
or some liquid to evaporate. The fluid will expand and the piston will move
Elementary Thermodynamics of Cryocoolers 37

p A B
T

Fig. 2.5. Constant pressure expansion on the pressure-volume and temperature-entropy

diagram.

up the cylinder to a new state B. The pressure remains constant because

the mass M on the cylinder is unchanged. We can represent that constant
pressure process on the P- V and T-S diagrams by the line A-B. We know
that the amount of heat added is the shaded area on the T-S diagram
under the line AB. In the same way it can be shown that the work done
by the piston is the shaded area on the P- V diagram under the line AB.
The area can be readily calculated as
area of P-V diagram= Pconst(VB- VA) (2.11)
The piston moved a distance dL so the change in volume is
(2.12)
and
W=pA(dL) (2.13)
The mass M was elevated by dL as the piston moved. The work done was
therefore
W=M(dL) (2.14)
Now the pressure of the working fluid is p = M/ A. Then
W=pAdL=MdL (2.15)
The work done is therefore equal to the area of the P- V diagram. It can
be expressed mathematically as

(a) dW=fpdV (for a nonflow process) (2.16)

and

(b) dW=J Vdp (for a flow process) (2.17)

38 Chapter 2

THE FIRST AND SECOND LAWS OF THERMODYNAMICS

There are two important laws of thermodynamics controlling the

operation of all thermal processes and machines whether they act as prime
movers converting high-temperature heat to work or as refrigerators pump-
ing heat from a low temperature to a higher temperature. The first law is
simply a statement of the conservation of energy: Energy can be neither
created nor destroyed, only changed from one form to another.
By way of illustration consider heating of the piston/cylinder system.
This increases the temperature of the working fluid causing it to expand
and lift up the piston and the mass upon it, thereby doing some work. The
heat supplied was equal to the sum of the work done and the change in
internal energy:

dQ=dW+dU (2.18)

This is the mathematical statement of the first law for a nonflow system.
The first law is important to refrigerator operation. To produce
refrigeration, work must be supplied to drive the system. According to the
first law the amount of energy to be dissipated at the upper temperature
is the sum of the work supplied and the refrigeration generated.
The first law is also important in refrigeration where rubbing friction
takes place in the low-temperature region. The rubbing friction will require
input of additional work to keep the machine working and that work will
be converted (degraded) by frictional action into heat energy. Furthermore
the heat is produced in the very region one is trying to refrigerate. The
additional friction work degrades the refrigeration generated and increases
the heat to be dissipated.
The second law of thermodynamics is more subtle. The simplest
statement particularly apt for refrigerators is that "heat will not flow from
a low-temperature area to a high-temperature area without the application
of work." Enunciation of the second law in this form is attributed to
Clausius. It sounds so commonplace as to appear trivial: everyone knows
that heat will flow only from hot to cold and that to produce refrigeration,
work must be done; but one must never be deceived by the apparent
triviality of the second law.
Concepts for machines which contravene the two laws of thermody-
namics are called perpetual motion machines-perpetua mobilae. Those
which contravene the first law are called perpetual motion machines of the
first class. Those which contravene the second law are of the second class.
Angrist (1968) in a brilliant review of perpetual motion machines provides
examples of exceedingly subtle contraventions of the second law.
Elementary Thermodynamics of Cryocoolers 39

Unfortunately, there is no way to demonstrate the inevitability of the

first and second laws. Equally however, there is no record in all human
experience of a single contravention of either law. Belief in the first and
second laws is therefore like religious faith. If one has it, all can be explained.
If one lacks faith, nothing can be explained. With faith, we continue to
look at cooling cycles!

THE CARNOT CYCLE

The operation of thermal engines is best analyzed by comparison with

ideal engines. These are assumed to work so that a succession of idealized
thermodynamic processes occur within the system. The succession is end-
lessly repeated so the system is in a state of quasisteady operation. The
succession of thermodynamic processes is called a cycle. All the processes
are assumed to be reversible, i.e., all the fluid involved is assumed to be
instantaneously at the same state so there are never any pressure, tem-
perature, or velocity gradients, never any local heat transfers, and never
any frictional effects. One cycle that has the highest performance between
given temperature limits, is the Carnot cycle. Sadi Carnot, a young French
Army artillery cadet first expounded the cycle in 1841 without the slightest
idea of the niche in history he was thereby carving for himself.
The cycle consists of four thermodynamic processes, two isothermal
and two isentropic (or adiabatic-without heat transfer). Operation of the
Carnot cycle is presented here for a refrigerating machine and for both
condensing and noncondensing working fluids. The cycle is illustrated in
Fig. 2.6. The simple piston/cylinder system used for our earlier discussion
is now "improved" by the addition of a piston rod coupled to a mechanism
that can absorb or provide work as required. The end of the cylinder, the
cylinder head, can be infinitely thermal conducting or infinitely thermal
insulating as required. The cylinder walls are thermally insulated and the
piston moves in the cylinder without friction or leaks.
To start the cycle assume the piston is at the outer dead point (ODP)
so that the volume contained is a maximum. The pressure and temperature
of the working fluid are at their minimum values, represented on Fig. 2.6a
by the point 1. In Fig. 2.6 the P- V and T-S planes for a gaseous working
fluid are drawn on the left-hand side. Corresponding P- V and T-S planes
for the condensing working fluid are drawn on the right-hand side.
Let the piston move towards the cylinder head, assumed to be perfectly
insulating. The compression process (1-2) is therefore isentropic (adiabatic)
since no heat is transferred to or from the cylinder. Now let the cylinder
head become perfectly conducting so that the further compression
40 Chapter 2

v s
:r\
PI:
II
~4I
·I I
I: I
v

Fig. 2.6. The Carnot cycle. The series of diagrams on the left side are the P-V and T-S
diagrams for operation with a gaseous working fluid. The diagrams on the right side are the
P-V and T-S diagrams for operation with a condensing working fluid. (a) Process 1-2,
isentropic compression; (b) process 2-3, isothermal compression; (c) process 3-4, isentropic
expansion; (d) process 2-3, isothermal expansion; (e) combined diagram for the cycle.

(2-3), occurs isothermally at temperature Tc. Heat Oc is rejected from

the working fluid to the surroundings. At point (3) the head changes
to the perfectly insulatjng state and as the piston moves on the outstroke
the expansion process (3-4) occurs isentropically. The temperature of the
working fluid decreases to TE. Finally the cylinder head becomes perfectly
conducting and the expansion continues. Process (4-1) occurs isothermally
Elementary Thermodynamics of Cryocoolers 41

and heat QE is abstracted from the surroundings of the cylinder to maintain

the temperature at TE. This is the refrigerating effect of the cycle and the
useful product.
The four separate processes are combined in Fig. 2.6e and the closed
cycle 1-2-3-4 is obtained. The shaded area of the P-V diagram is the
work supplied to operate the cycle. The shaded area on the T-S diagram
is also the work necessary to drive the cycle. It is the difference in the heat
rejected, Oc, in process (2-3) and the heat supplied, QE, in process (4-1),
i.e.,
(2.19)
The refrigerating effect QE is the area 1-4-5-6 on the T-S plane. The heat
to be rejected from the engine is the sum of the work done and the heat
lifted or refrigeration effect Oc = W + QE.
The work diagrams for gaseous and condensing working fluids are
much different in shape and area. The diagram for the gaseous engine has
a small area even though the range of pressures (Pmax/Pmin) is greater than
for the condensing engine. The small area is due to the fact that, with a
gaseous working fluid, there is little difference in the slope of an isothermal
process and an adiabatic process. Although the pressure range and volume
compression ratios are high the actual amounts of heat transfer and work
done are small. The cycle is very efficient but the amount of refrigeration,
area 1-4-5-6, is limited because the expansion process 4-1 is constrained
by the need to keep the maximum pressures and volumes within tolerable
bounds.
The condensing system has no such difficulty. There the isothermal
processes occur at constant pressure during condensation and evapor-
ation. As a consequence, the work diagram has a large area and large
quantities of heat are transferred in phase change processes (4-1) and (3-2).
The condensing Carnot cycle is the basis for all refrigeration plants
for food preservation, air conditioning, domestic refrigeration, etc. In
practical machines the adiabatic expansion process (3-4) is replaced by an
isenthalpic expansion, except in very large plants, because it is cheaper
and easier to accomplish. The effect is a reduction in performance with
both a loss in the refrigeration and an increase in the work required.
Carnot engines cannot be contemplated for cryocoolers. With gaseous
working fluids the large temperature range requires impractically high
pressures and volumes for the amount of refrigeration produced. Similarly,
no known fluid has the characteristics for the working medium in a condens-
ing Carnot cryocooler.
It is evident from Fig. 2.6e that the cycle work increases as the
difference in temperature widens. T c is the temperature of the cooling air
42 Chapter 2

or water available for heat rejection from the cycle. In practice Tc would
be greater than the atmospheric ambient temperature to accomplish the
heat transfer. Temperature TE is the refrigeration temperature and is fixed
by the application. In practice TE would be less than the refrigeration
temperature actually required so as to induce heat to tranfer to the cylinder.
It is thermodynamically (and therefore economically) wasteful to operate
the engine at lower refrigeration or higher rejection temperatures than the
application requires or the ambient conditions allow.
Reflection on Fig. 2.6e will lead to the very important conclusion that
when operating between given temperature limits the cycle with the
maximum refrigerating effect per unit of work done is that with isothermal
heat addition and rejection to the cycle. Given upper and lower temperatures
of operation the best refrigeration performance always comes from the
cycle with isothermal heat transfers to and from the cycle.

COEFFICIENT OF PERFORMANCE

The performance of a refrigerator is expressed in terms of the

coefficient of performance, COP. This is the ratio of the heat extracted to
the work done:
COP= QE/W = QE/(Oc- QE) (2.20)
If we have a perfect gas the heat transfers can be represented as
QE = CPTE and Oc = CpTc (2.21)
Therefore the coefficient of performance is
COP= TE/(Tc- TE) (2.22)
This is called the Carnot coefficient of performance and is the maximum
theoretical coefficient of performance attainable by a refrigerator operating
between given temperature limits. The second law of thermodynamics
restrains the maximum theoretical COP to the Carnot value.
The Carnot value is the reference index for all other refrigerators. A
refrigerator "efficiency" is sometimes quoted as the ratio of the actual COP
to the Carnot COP, i.e.,
7Jref =(COP actuat)/(COP carnot) (2.23)
Practical refrigerators never approach the Carnot value. Figure 2. 7 is a
comparison of various refrigerating systems published by Kohler (1965),
and attributed to Professor Nesselman. The highest efficiencies are attained
by vapor compression machines at the relatively high temperatures for
Elementary Thermodynamics of Cryocoolers 43

/ [\
50
v
,/EVAPORA T IDN
I

f.----
/
I/ ~
0
~

PHILIPS
' ~
'
~
STIRLING

''
u

"
z
0 30

~ 1-
20
c~
1--..... ~
"'"
Fig. 2.7. Comparison of the efficiency of JOUL THOMSON

fractical refrigerating systems as a func- 0

tion of refrigeration temperature.

Efficiency is the ratio actual COP /Carnot
COP (after Kohler, 1965).
0
0 so wo
'" 200 '''0
~
TEMP['RATURE iDFGREES KELVIN

domestic refrigeration, air conditioning, and the like. These are basically
Carnot systems with a condensing working fluid and are modified to replace
the expansion engine process.
The work per unit refrigeration increases dramatically as the refriger-
ation temperature decreases. Values calculated for a Carnot cycle system
with Tc of 300 K are quoted in Figure 1.1 (see Chapter 1). The work per
unit of refrigeration at 100 K is 2 watts per watt. At 10 K the ratio is 29
watts per watt and at 1 K it is 299 watts per watt.
Exactly the same cycle may be used as a heat pump to abstract heat
at ambient temperatures and to reject it from the system (say for heating
a building) at higher temperatures. It is customary to quote the coefficient
of performance of a heat pump as the ratio of heat rejected to work done:
(2.24)
This coefficient is greater, by unity, than the coefficient of performance
defined as a refrigerator. This is true even though the same machine may
be working over the same temperature limits, i.e.,
(COPheatpump = Oc/(Oc- QE) (2.25)
(COP) refrigerator= QEj(Qc- QE) (2.26)
so that
(COPheat pump= 1 +(COP) refrigerator (2.27)
It is fashionable to talk about refrigerators as heat pumps even though
they are used as refrigerators. In such unfortunate circumstances it is
necessary to be pedantic about the nomenclature to avoid the evident
problems that will arise without a precise definition of terminology.
44 Chapter 2

Part II: Regenerative Cycles

THE REITLINGER CYCLE

The coefficient of performance for the Carnot engine was found above
to be COP= TE/(Tc- TE). It is the maximum value that can be achieved
by a refrigerating engine operating between temperature limits TE and Tc.
However, it is less well known that other thermodynamic cycles can attain
the same performance. In fact, there are an infinite number of thermody-
namic cycles that can achieve the Carnot value. All these cycles incorporate
regenerative heat transfer processes in place of the adiabatic compression
and expansion of the Carnot cycle. The high-performance regenerative
cycles all embrace isothermal processes of heat addition and rejection from
the cycle. These are the necessary key features of high-performance ther-
modynamic cycles.
The most general form of idealized thermodynamic cycle, the Reitlin-
ger cycle, shown in Fig. 2.8, consists of two isothermal processes where
the heat is supplied to and drawn from the cycle and two polytropic
regenerative processes. It is essential in the ideal reversible engine that
external heat exchange occurs only at the upper and lower temperature
limits of the cycle. Heat reception and heat rejection during the polytropic
phase must be provided by thermal storage in a regenerator. This can be
conceived as a thermodynamic sponge, absorbing and releasing heat as
required.
The generalized regenerative cycle was first analyzed by Reitlinger
(1876) a century ago. It can exist in an infinite variety of forms distinguished
by the character of the polytropic regenerative phases. The Carnot cycle
is simply a special case of the Reitlinger cycle in which the regenerative
polytropic phases are replaced by adiabatic, isentropic phases. Two other
special cases have been given names: the Stirling and Ericsson cycles where

[]'
Tc

p T

TE 3 4

v s
Fig. 2.8. The generalized Reidinger isothermal cycle.
Elementary Thermodynamics of Cryocoolers 45

the regenerative processes occur at constant volume and constant pressure,

respectively. An infinite number of variations may be postulated in which
the regenerative processes are not constant volume, constant pressure, or
isentropic. All these cycles are characterized by the same thermodynamic
refrigeration performance:
TE
COP= T. T (2.28)
c- E

THE STIRLING CYCLE

The Stirling cycle illustrated in Fig. 2.9 is similar to the Carnot cycle.
Operation is as follows. Consider a cylinder containing two opposed pistons,
with a regenerator between them. The regenerator may be thought of as
a thermodynamic sponge, alternatively releasing and absorbing heat. In
practice this is accomplished with a matrix of finely divided metal in the

~~.
·~ <:;J.
Tc 2

v (oJ s
REGENERATOR

(2)

(3)

(4)!
_!!J 1 DISPLACEMENT

"' / "' I
I
(I)

v
(2)
/ i/ (3)

~~ (4)

Fig. 2.9. The Stirling cycle with conceptual engine.

(a) P-V and T-S diagrams. (b) Piston arrangement
at the terminal points of the cycle. (c) Time-displace-
ment diagram.
~ ~ (c J
(I)
46 Chapter 2

form of wires or strips. One of the two volumes between the regenerator
and the pistons is called the expansion space, and is maintained at low
temperature TE. The other volume is called the compression space, and is
maintained at ambient temperature T c- There is a temperature gradient
(Tc- TE) across the transverse faces of the regenerator and it is assumed
there is no thermal conduction in the longitudinal direction. The pistons
are assumed to move without friction or leakage of the working fluid
enclosed between them.
Assume that the compression-space piston is at the outer dead point,
and the expansion-space piston is at the inner dead point, close to the face
of the regenerator. All the working fluid is then in the compression space
at ambient temperature. The volume is a maximum, and the pressure and
temperature are represented by (1) on the P-V and T-S diagrams, shown
in Fig. 2.9. During compression (process (1-2), the compression piston
moves towards the inner dead point, and the expansion-space piston
remains stationary. The working fluid is compressed in the compression
space, and the pressure increases. The temperature is maintained constant
because heat Oc is abstracted from the compression-space cylinder.
In the transfer process (2-3), both pistons move simultaneously, the
compression piston towards, and the expansion piston away from, the
regenerator. The volume between them remains constant. The working
fluid is transferred through the porous metallic matrix of the regenerator
to the expansion space. In passing through the regenerator, the working
fluid is cooled from Tc to TE, by heat transfer to the matrix, and emerges
from the regenerator into the expansion space at temperature TE. The
decrease in temperature passing through the matrix at constant volume
causes a decrease in pressure.
In the expansion process (3-4 ), the expansion piston moves to the
outer dead point; the compression piston remains stationary at the inner
dead point, adjacent to the regenerator. As the expansion proceeds, the
pressure decreases as the volume increases. The temperature remains
constant because heat QE is abstracted from the cylinder surroundings. This
is the useful refrigeration of the cycle.
Finally both pistons move simultaneously to transfer the working fluid
(at constant volume) back through the regenerative matrix from the
expansion space to the compression space. In passing through the matrix
heat is transferred from the matrix so the working fluid increases in
temperature and emerges at Tc into the compression space.
The cycle is therefore composed of four heat-transfer processes:
Process (1-2)-Isothermal compression; heat transfer from the work-
ing fluid to the external sink at ambient temperature,
Tc.
Elementary Thermodynamics of Cryocoolers 47

Process (2-3)-Constant volume; heat transfer from the working fluid

to the regenerative matrix.
Process (3-4)-Isothermal expansion; heat transfer to the working
fluid from an external source at the refrigerating tem-
perature TE.
Process (4-1)-Constant volume; heat transfer to the working fluid
from the regenerative matrix.
The heat transferred in process (2-3) is the same as in process (4-1). The
only heat transfer process between the engine and its surroundings are (a)
heat supply at TE and (b) heat rejected at Tc. Heat supply and rejection
at constant temperature satisfies the requirement of the second law of
thermodynamics for maximum thermodynamic performance. The
coefficient of performance of the Stirling cycle is the same as for the Carnot
cycle:
(2.29)
The advantage of the Stirling cycle over the Carnot cycle lies in the
replacement of two isentropic processes by two constant-volume regenerative
processes. This greatly increases the area of the P-V diagram and the
quantities of heat transfer to and from the system so that it is not necessary
to resort to very high pressures and swept volumes.
The P- V and T-S diagrams of Carnot and Stirling cycles between
given limits of pressure, volume, and temperature are compared in Fig.
2.10. The shaded areas (4-1-5) and (2-3-6) represent the additional work
made available by substituting constant-volume processes for isentropic
processes. The isothermal processes (2-5) and (6-4) of the Carnot cycle

s
Fig. 2.10. Stirling and Carnot cycles superimposed with common values for the minimum
and maximum temperatures, pressures, and volumes. Cross-hatched areas on the P- V plane
represent the increased work output of the Stirling cycle. Cross-hatched areas on the T-S
plane represent increased heat transfer of the Stirling cycle.
48 Chapter 2

Fig. 2.11. Ericsson and Carnot cycles superimposed with common values for the maximum
and minimum values of temperature, pressure, and volume. Cross-hatched areas represent
the increased work output and heat transfer of the Ericsson cycle.

are extended to processes (1-2) and (3-4), respectively. The quantities of

heat transfer for the Stirling cycle are increased by the same proportion
as the work. The coefficient of performance (heat extracted per unit of
work) is the same for both.

THE ERICSSON CYCLE

For the Ericsson cycle, the processes of constant-volume regenerative

heat transfer are replaced by constant-pressure regenerative processes.
This leads to the P- V and T-S diagrams shown in Fig. 2.11. The coefficient
of performance of the cycle is the same as that of the Carnot cycle but for
given limits of pressure, volume, and temperature the work and heat transfer
are much greater.

THE STIRLING CYCLE AS A PRIME MOVER,

HEAT PUMP, AND PRESSURE GENERATOR

When the Stirling engine is used as a prime mover, heat at high

temperature T max is supplied to the cycle. In the expansion process (3-4)
(Fig. 2.9) part of the heat is converted to work, and part is rejected, as
heat, at ambient temperature Tc. The machinery is precisely the same as
a refrigerator. Comparison of the cycles for operation as a prime mover
and refrigerating machine is given in Fig. 2.12. In operating as a prime
mover the pressure and temperature are greater in expansion than in
Elementary Thermodynamics of Cryocoolers 49

TmoK

PRIME
MOVER~.

TE 3._?""-_ __.4

COOLING
ENGINE

s
Fig. 2.12. Stirling cycle as cooling engine and as prime mover. In both applications the
compression process occurs at ambient temperature Tc. The expansion process occurs at a
low temperature TE in the cooling engine and at a high temperature T max in the prime mover.
In both cases heat is absorbed during expansion and rejected during compression.

compression. In operating as a refrigerator the reverse is true: the pressure

and temperature are both less in expansion than in compression.
As a heat pump the Stirling cycle operates exactly as the refrigerator.
The temperature of the expansion space TE is less than the compression
space Tc- The difference between operation as a heat pump and refrigerator
is that both TE and Tc are increased. In the prime mover and refrigerator
application, T c is the ambient atmospheric temperature at which cooling
water is available. For a heat pump, T c is the temperature at which heat
is rejected from the system, and is the useful output for heating a concert
hall or office building. Therefore T c is above the ambient atmospheric
temperature. Heat is supplied to the cycle (at TE), from atmospheric air
or river water, at approximately the ambient atmospheric temperature.
A comparison of the Stirling cycle as a heat pump and a refrigerator
is drawn in Fig. 2.13. In both cases, work from an external source is required
to drive the cycle, and is equivalent to area (1-2-3-4). In the case of the
heat pump, the useful product is the heat rejected at temperature T c, and
the performance of a heat pump is therefore assessed as

(COP)HP =heat rejected= T c (2.30)

work done Tc- TE
This is the inverse of the thermal efficiency. The coefficient of the perform-
ance of a refrigerator
(2.31)
is not the inverse of thermal efficiency.
Stirling engine systems have been used to pump a fluid and increase
its pressure. When the fluid to be pumped was a liquid or gas, separated
50 Chapter 2

----u
Tatmos

3
2

Fig. 2.13. Stirling cycle as a heat pump and

(a) REFRIGERATING MACHINE
as a cooling engine. When operating as a
cooling engine or as a heat pump the
expansion process occurs at a temperature
less than that of the compression process
and an input of work to operate the cycle
is required. The cooling engine absorbs heat
during expansion at temperatures less than
the ambient and rejects heat at near
ambient temperatures. The heat pump
absorbs heat during expansion at ambient
v temperatures and rejects heat at tem-
(b) HEAT PUMP peratures greater than ambient.

by a diaphragm or piston from the working fluid in the Stirling cycle device,
the system can be classified as a Stirling cycle working as a prime mover.
In other instances, where the working fluid itself was the fluid to be
compressed and pumped, there are, invariably, valves or other flow-
controlling devices. According to the defintion given in Chapter 1 these
systems cannot be classified as Stirling engines but rather as Ericsson
engines. Nevertheless, when discussed in the literature they are frequently
called Stirling engines. In most cases, fluid is added to the cycle when the
pressure is low and withdrawn at a higher pressure. Heat is supplied at
high temperature and rejected at ambient temperature. Work during
expansion is greater than work during compression by an amount equivalent
to the "pump" work of the compressed fluid.

A VERSATILE DEMONSTRATION ENGINE

In the late 1950s when the Philips Stirling engine liquefiers were
introduced it was common for them to demonstrate the versatility of the
Stirling engine for operation as a prime mover, cooling engine, or a heat
pump. A simplified diagram of the engine used for the demonstration is
shown in Fig. 2.14. The engine was of a type known as the piston-displacer
machine rather than the opposed piston version used above to describe
the cycle.
Elementary Thermodynamics of Cryocoolers 51

EXPANSION
SPACE

FREEZER

REGENERATOR

DIS PLACER

COOLER
,_-,_,__COMPRESSION
SPACE
PISTON

Fig. 2.14. Diagram of the versatile prime

mover/refrigerator/heat pump Stirling engine
demonstrated by Philips.

In this engine the space above the dis placer, below the cylinder head,
is the expansion space. Between the underside of the displacer and the top
of the piston is the compression space. The two spaces were in communica-
tion through a water-cooled heat exchanger, regenerative heat exchanger,
and a finned freezer heat exchanger. The machine was mounted on a
bed-plate and coupled to an electric dynamometer capable of operating as
a motor or as a generator in either direction of rotation. Helium or hydrogen
gas was used as the working fluid. A thermocouple was attached to the
cylinder head and a large temperature scale was projected onto the wall
of the lecture room with a moving light spot to indicate the cylinder head
temperature.
The engine would be started initially as a refrigerator with the
dynamometer working as a motor driving the machine in the forward
direction. The cylinder head around the expansion space would cool rapidly
and on the temperature scale one could observe the decrease in tem-
perature. At about 80 K air condensed on the cylinder head and ran into
a collector at the base of the head. After a short time power to the motor
would be switched off and in less than a revolution the machine would
stop and self-start in the reverse direction.
It was now operating as a prime mover generating power. With the
reversal of rotation the expansion and compression spaces switched loca-
tions so the expansion space was between the displacer and piston. Heat
was drawn from the cooling water to drive the engine and rejected from
the cylinder head to the pool of liquid air serving as the engine coolant.
The reservoir of liquid air was soon dissipated and soon the temperature
of the cylinder head started increasing. The engine continued to run but
with progressively less vigor as the temperature increased until finally it
slowed to a very low speed.
At that point power would once more be supplied to the dynamometer
causing it to run again as a motor but preserving the reverse direction of
52 Chapter 2

rotation. With this power input the temperature of the cylinder head
continued to increase until eventually it attained and then exceeded the
ambient temperature. The machine was now operating as a heat pump
drawing heat from the "cooling" water and rejecting it at higher tem-
perature to the surroundings of the cylinder head now acting as the com-
pression space. The temperature of the head would increase rapidly,
whereupon the lighting would be lowered and the cylinder head could be
observed glowing cherry red. At that point Jan Kohler would break his
discourse, reach into his pocket, withdraw a cigar and casually lean down
to light it on the hot cylinder head. It was beautifully executed and done
so matter of factly. Following this, power to the motor would be switched
off and again, immediately, the engine would stop and starting running in
the opposite direction.
With this second reversal the direction of rotation was now as it had
been initially. The expansion and compression spaces were restored to the
configuration shown in Fig. 2.14. The engine now ran again as a prime
mover, withdrawing heat from the hot cylinder head, rejecting it to the
cooling water. After some time the cylinder head cooled and, eventually,
there was insufficient temperature difference between the expansion and
compression spaces and the engine stopped.
It was a remarkable demonstration, beyond the capability of any other
machine. The versatility for the same unit to operate as a prime mover,
refrigerator, and heat pump has practical implication for the future. A
Stirling refrigerator driven by a Stirling prime mover has already been
demonstrated by William Beale at Sun power Inc. Use of a common working
fluid has outstanding attractions to overcome seal problems. The combina-
tion may be a solar, isotope, or combustion heated cryocooler. More likely
in the near future it will be developed as a gas-fired air conditioning
refrigerator. The combination or duplex Stirling engine should not be
confused with the Vuilleumier cycle machines we shall consider later (see
Chapter 4).
Another concept to utilize the versatility of the Stirling engine is to
use it as a prime mover extracting heat from a thermal battery (lithium
fluoride) as an automotive propulsion engine. For downhill operation, or
when braking, the engine could be converted to function as a heat pump
feeding thermal energy into the thermal battery. In practice it is not
necessary to physically reverse the engine. The same effect can be gained
in some forms of engine by simply switching fluid connections (see Walker
1979). The combined regenerative braking/propulsion concept is attractive
in mining applications where trucks are used on loaded downhaul operation
and for underground mines. It will become attractive for road vehicles as
the fossil fuel supplies dwindle and the world moves into the "electric
Elementary Thermodynamics of Cryocoolers 53

economy." The Stirling engine with thermal storage will likely become the
preferred automotive propulsion unit. Commuters will plug in their car
for an electric (thermal) recharge on arrival at work or on arrival at home.
This consumer behavior prototype already exists in Canada, where, in
winter, it is customary when parking to "plug in" an engine block heater.
A. J. Oxley of the U.K. Rutherford Laboratory has explored use of
liquid nitrogen as the coolant for Stirling engines. Engines could be ener-
gized from an ambient temperature source to operate as a prime mover
for underground mining applications with no emission of noxious fumes.
Other possibilities for large-scale power generation exist using the tem-
perature difference between sea water and liquid natural gas now coming
into the United States, Japan, and Europe in very large quantiites.

THE RALLIS CYCLE

Generalized reference cycles for regenerative power engines were

developed by Rallis et al. (1977). The basic cycle was developed in two
forms characterized by
i. Isothermal processes of compression and expansion (called the
Rallis isothermal regenerative cycle);
ii. Isentropic processes of compression and expansion (called the Rallis
adiabatic regenerative cycle).
In both forms the regenerative processes occur partly at constant pressure
and partly at constant volume in any given combination.
The Stirling and Ericsson cycles are both special cases of the Rallis
isothermal regenerative cycle and many other special cases can be defined
which have no identifying names. The cycles were developed by Rallis for
prime movers, but the same approach appears well suited for application
to refrigeration applications.
Analytical expressions for the general form of the Rallis cycle are
developed below for both the isothermal and adiabatic versions. They are
then reduced to various special cases and, finally, a comparison is drawn
with some computed results for the ideal Stirling cycle and the pseudo-
Stirling cycle with adiabatic compression and expansion processes.

RALLIS ISOTHERMAL REGENERATIVE CYCLE

A conceptual arrangement of an engine capable of operating on the

Rallis isothermal regenerative cycle is shown in Fig. 2.15. The P- V and
54 Chapter 2

REGENERATION
CONSTANT CONSTANT
EXPANSION COMPRESSION (1·2)
VOLUME (2~3) PRESSURE (2·2')
(3-4)

~~~-~
7
_ TE

Q IN AT
I' I I=I :I:::-+-:1I~r:·
REGENERATION
Q OUT AT
TE (REFRIGERATION) CONSTANT
VOLUME (4~ I)

Fig. 2.15. Conceptual arrangement of an engine for operating on the Rallis isothermal
regenerative cycle.

T-S planes for the cycle are shown in Fig. 2.16. The cycle is defined by
two isothermals, temperatures Tc and TE, bounded by constant volume
and constant pressure regenerative processes at the two extremes as shown
above. Tc is the ambient temperature at which cooling is available, TE is
the refrigeration temperature.
Processes (1-2)-Isothermal compression, heat is rejected from the
system to the surroundings.
Process (2-2')-Constant pressure regenerative cooling.
Process (2'-3)-Constant volume regenerative cooling.
Process (3-4) -Isothermal expansion, heat is extracted from the sur-
roundings, (this is the useful refrigerating effect of the
cycle).
Process (4-4')-Regenerative heating at constant pressure.
Process (4'-1)-Regenerative heating at constant volume.
For convenience in analysis we define the following nondimensional
parameters:
T = Tc/TE
rc = Vd V2 , volume compression ratio
re = V 4 / V 3 , expansion compression ratio
a= Tz·/Tz, regenerative cooling constant pressure ratio
b = T 3 /T2·, regenerative cooling constant volume ratio
d = T4·/T4, regenerative heating constant pressure ratio
e = TdT4·, regenerative heating constant volume ratio
e = (Ts- Tz)/(T3- Tz), regenerator effectiveness
COP= coefficient of performance defined as the ratio Qj W
W =specific external work required per cycle
Q = specific external heat extracted at the low temperature
(refrigeration)
Pm = W/swept volume, mean effective pressure
P = Pm!Pmax• mean effective pressure ratio of the cycle
Pmax = Pz = Pz·, maximum cycle pressure
Elementary Thermodynamics of Cryocoolers 55

Pmax

Pmin 1----+---"-4-----'4,

Vmin v Vmox

Fig. 2.16. Rallis isothermal regenerative cycle.

Now

ab = (Tz·/Tz)(T3/Tz·) = (TE/Tc) = 1/T (2.32)

de= (T4.jT4)(TtfT4·) = (TE/Tc) = 1/T (2.33)
(ref a)= (dre)
therefore

and

(2.34)

The external work required per cycle is [from Eq. (2.19)]

W = pz(Vz- Vz·) + RTzln rc -p4(V4,- V4)- RT4ln re

= RTz(1-a) + RTzln rc- RT4(d -1)- RT4ln re

=R[(1-a +ln rc)Tz-(d -1 + ln re)T4]
= RTc[(l-a + ln rc)- (1/T)(d -1 + ln re)]
RTc
= -{T[l- a+ ln (adre)]- (d -1 + ln re)}
T

= RTE{T[l-a + ln (adre)]- (d -1 + ln re)} (2.35)

The heat transfer to the working fluid during the expansion (3-4) is
the refrigerating effect of the cycle. With imperfect regeneration the
56 Chapter 2

refrigerating effect is partially or wholly consumed in cooling the working

fluid from T 5 to T3.
There are two cases to consider:

and

For case (a) the external refrigerating effect is

and for case (b) the external refrigerating effect is

Now for case (a) T 3 :::;; T 5 :::;; T 2 ·, so that

and for case (b) T 2 ·:::;; T 5 :::;; T 2 , so that

a :::;; -1)
1 - e ( -'T' T -

Thus for case (a), the refrigeration is (where 'Y = Cv!Cv):

Q =RTE In re -CvTc[ l-ee~ 1) -; J

=RTc[lnre _ _l_ r-e(r-1)-1]
'T (y-1) 'T

RTc
= {(y -1) In re -[r -e(r -1)-1]}
r(y-1)

RTE
= - -{('}' - 1) In re - [ 'T - e ('T - 1) - 1]} (2.36)
( 'Y -1)
Elementary Thermodynamics of Cryocoolers 57

The coefficient of performance (Q / W) is

RTE
- 1 {( 'Y -1) In re- [T - e (r -1) -1]}
COP=-y~----------------------­
RTE{r[1- a+ n (adre)]- (d -1 +In re)}
1 (y-1)lnre-[r-e(r-1)-1]
(2.37)
('Y - 1) {T [ 1 - a + In (adre)] - (d - 1 + In re)}

The mean effective pressure

Pm = W/stroke volume
RTE{r[1- a+ In (adre)- (d -1 +In re)]}
v4,- v3
RTE{r[1-a +In (adre)- (d -1 +In re)]}
(2.38)
(dre -1) v3

and in dimensionless form with respect to the maximum pressure

P= Pm =RTE[r(1-a+ln(adre)-(d-1+1nre)]
Pmax (dre -1) v3 X Pz'

but Pmax = P2' = bp3 and since P3 v3 = RTE then

P= r[1-a+ln(adre)-(d-1+lnre)]
(2.39)
b(dre -1)

Similarly for case (b) where T 2, :s; Ts :s; T 2 the external refrigerating
effect is

=RTE In re -CvTE[ (D J
-1 -CpTc[ 1-eC~ 1 ) -a J

=RTE{ In re- ('Y ~ 1) C~b) -r('Y: 1) [ 1-eC~ 1) -a]}

= RTE {c'Y -1) In re + (b - 1) -y[r -e(T -1)-ar
(y-1) b
J} (2.40)
58 Chapter 2

The coefficient of performance is

R~E { (y -1) In re + ( b - 1) -y[r- t:(r -1)- aT J}

COP=" 1 b
RTE{r[1- a+ In (adre)]- (d -1 +In re)}

{ (y -1) In re + (T) -y[r-e(r -1)-ar]}

(2.41)
(y -1){r[1-a +In (adre)]-(d -1 +In re)}
The mean effective pressure and the specific power are not dependent on
the regenerator effectiveness and so are the same as case (a). Now consider
some special cases.

Ideal Stirling Cycle

In this cycle the regenerative processes (2-3) and (4-1) occur wholly
at constant volume as shown in Fig. 2.17. Therefore a = d = 1, since T2· = T2
and T 4 • = T 4 , b = e = r, re = rc =rand T 3 ::::; T 5 ::::; T 2 only. Then the coefficient
of performance is given by [from Eq. (2.37)]
1 {(y-1)lnr-[r-e(r-1)-1]}
COP=--
(y-1) r[1-1+ln(r)]-(1-1+lnr)
= _1_{(y -1) ln r -[r -e(r -1) -1]}
y -1 (T -1) ln r
= _1_{(y -1) In r -(r -1)(1-e)}
( y -1) (r -1) In r

_1_[(y -1) In r -(r -1)(1-e)J

(2.42)
(y - 1) ( r - 1) In r

and if e = 1
1 TE
COP=--= the Carnot value (2.43)
r-1 Te-TE

2/2'

p
.~-
~
'

414 1
Fig. 2.17. Ideal Stirling cycle. v s
Elementary Thermodynamics of Cryocoolers 59

The mean effective pressure of the Stirling cycle is [from Eq. (2.38)]
_ T[ 1 - 1 + In (1 X 1 X r)] - ( 1 - 1 + In r)]
Pm=RTE (r- 1)V3

RTE (T -1) In r
(2.44)
v3 (r-1)

and the pressure ratio P = PmiPmax

P=C~1)C~~) (2.45)

Ideal Ericsson Cycle

In this cycle the regenerative processes (2-3) and (4-1) occur wholly
at constant pressure as shown in Fig. 2.18. Therefore b = e = 1 since T 2• = T 3
and T4' = Tb 1I a = d = T, re = rc = r and T3 ::5 T5 ::5 Tz only.
Then the coefficient of performance is:

(y- 1) In r 1 -~-1 ) -
+ (- y[ T - s (T - 1) ~7 J
COP=----~----------~--------~
(y - 1){[1 - 1 +In (1 x 1 x r)]- (T - 1 +In r)}
(y -1) In r+y(T -1)(1-s)
( y - 1) (T - 1) In r

and when s = 1

the Carnot value (2.46)

The mean effective pressure of the Ericsson cycle is

RTE { T [ 1 -;+In (; r)] -( T - 1 +In r)}

Pm = ---------------
V4'- v3

.'~ 4 1/41
T
D
2'13
2

E
4
114'

Fig. 2.18. Ideal Ericsson cycle. v s

60 Chapter 2

V4·- V3 = V3(Tr -1)

Pm = RTE IT(T -1 +TIn r)/[T- (T -1 +In r)]l
V3 (Tr-1)
RTE (T-1)
=-----lnr (2.47)
v3 (n-1)
and the pressure ratio P = Pm!Pmax is
p-(T-1)~ (2.48)
T (n-1)

Ideal Cycle with Constant-Volume Regenerative Cooling and Constant-

Pressure Regenerative Heating
In this cycle the regenerative process of cooling, (2-3), occurs at
constant volume, and of heating, (4-1'), occurs at constant pressure, as
shown in Fig. 2.19. (This cycle has no given name so far as is known.)
Therefore a= e = 1 since T 2• = T 2 , and T 4 • = T1. 1/ b = d = T, and T 3 -5. T 5 -5.
T 2 only. The coefficient of performance is
COP =-1-1 (y -1) In re -[T -e(T -1)-1] I
(y-1) 7[1-1 +In (1 Xd Xre)]-(d -1+ln re)
1 I('Y -1) In re- [T- e(T -1) -1] I (2.49)
= (y -1) (TIn dre)-ln (re)-d + 1)
The mean effective pressure Pm of this cycle is
_ RTE{T[1-1+ln(1XTXre)]-(T-1+lnre)}
Pm =

and since V4·- V3 = (V4- V3) = V3[TV4j(V3 -1)]

_ RTE TIn (ne)-ln (re)-T + 1
(2.50)
Pm= v3 (Tre-1)

212'

D
212' Tc 1/4'

~
3
4
1/4
'
T

3 4
Fig. 2.19. Cycle with constant-
volume regenerative cooling and
constant-pressure regenerative
v s heating.
Elementary Thermodynamics of Cryocoolers 61

and the pressure ratio P = PmiPmax is

p =[TIn (Tr.)-ln (r.)-T+ 1]
(2.51)
T(Tr.-1)

Ideal Cycle with Constant-Pressure Regenerative Cooling and Constant-

Volume Regenerative Heating
In this cycle the regenerative process of cooling, (2-3), occurs at
constant pressure, and of heating, (4' -1), occurs at constant volume as
shown in Fig. 2.20. (This cycle also has no given name as far as is known.)
In this cycle b = d = 1 since T2 = T 3 and T 4 • = T 4 ; (1/a) = e = T and T 2·S
Ts s T 2 • The coefficient of performance is [from Eq. (2.41)]

(y-1)lnr. [ -(1-1)]
1-
T]
-y [ T-e(T-1)-;
COP=I 1 I
(y -1) { T[ 1-;+ In (r,/T)]- (1-1 +In r,)}

[(y -1) In r. -y(T -1)(1-e)]

(2.52)
(y -1)[T -1-T In (r,/T) -In r,]

The mean effective pressure is [from Eq. (2.38)]

RTE [T -1 +In (r,/T)-ln r,]

(2.53)
= v3 (r. -1)

and the pressure ratio P = PmlPmax is

p -1 +In....:....;::'
= T___ (r./T) ---'-_-In____:r._ (2.54)
r. -1

p
2~
~I T
n
2 13~
1 Fig. 2.20. Cycle with constant-
414
414 1 pressure regenerative cooling
and constant-volume regenera-
v s tive heating.
62 Chapter 2

v s
Fig. 2.21. Rallis adiabatic regenerative cycle.

RALLIS ADIABATIC REGENERATIVE CYCLE

The pressure volume and temperature entropy planes for the Rallis
adiabatic cycle are shown in Fig. 2.21. The cycle is defined by two adiabatic
processes having initial values of temperature defined by given temperatures
Tc and TE and bounded by constant-volume and constant-pressure regen-
erative processes and external heat transfers to and from the system. Tc
is close to the ambient temperature at which cooling is available. TE is the
refrigeration temperature.
Process (1-2)-Adiabatic compression, no heat transfer to or from
the working fluid.
Process (2-6)-Constant-pressure cooling. Temperature is reduced
from T 2 to T 6 by heat rejection from the system to
the surroundings.
Process (6-2')-Constant pressure regenerative cooling.
Process (2'-3)-Constant-volume regenerative cooling.
Process (3-4 )-Adiabatic expansion, no heat transfer to or from the
working fluid.
Process (4-7)-Constant-pressure heating. Temperature increased
from T 4 to T 7 by heat abstracted from the surround-
ings. This is the useful refrigerating effect of the cycle.
Process (7-4')-Constant-pressure regenerative heating.
Process (4'-1)-Constant-volume regenerative heating.
For convenience in the analysis we define the following nondimensional
parameters different from those in the Rallis isothermal regenerative cycle:
a= T2·/T6, regenerative cooling constant-pressure ratio
d = T4·/T7 , regenerative heating constant-pressure ratio
Elementary Thermodynamics of Cryocoolers 63

f = T6/T2 = 1/r~y-1), constant-pressure cooling ratio

g = T1/T4 = r~y-1), constant-pressure heating ratio
(the refrigerating effect)

regenerator effectiveness so that

Ts = T6+s(T3- T6) = T6-sc ~ 1)r6

COP is the coefficient of performance (0/W), where Q is the specific
external heat extracted in the low-temperature process (4-7) less the
heat necessary to reduce the working fluid temperature from T 5 to T 3
(due to imperfect regeneration). The specific external work input
required per cycle, W, the mean effective pressure, Pm. and the mean
effective pressure ratio of the cycle, P, are defined as before.
Now
ab = (T2.fT6)(T3/T2') = (TE/Tc) = 1/r
de= (T4'/T7)(TdT4') = (Tc/TE) = T

T2 = r~y-1)

T6 = fT2 = fr~y- 1 lT1 = Tc = T1

therefore
fr~y- 1 ) =1
or
!=1/r~y- 1 )

T2' = aT6 = aT1

T3 = bT2' = abT1 = Tdr
T3 T1
T4= (y-1)=~
re Tre

therefore
g/r~y- 1 l = 1 or g = r~y-1)
T4' = dT1 = dT3 = (d/r)T1
but since T4' = Tde then 1/e = d/r or ed = r.
64 Chapter 2

The external work required per cycle is

RT1
=-
'Y -1
d
- [ (yrc<'Y - 1> -ya +a -1)- ( y--y-----=-rr--+-
d
T
1
Tr~'Y T
1)]
T

=
RT1)[
(y- 1 y(rc
(y-1) 1 1 ) ) +;(d-1)
y ( d-,~y-
-a)+(a-1)-; 1 J

=RT1
- - [ y(r/
< _0
-a)+(a-1)- 'Y ) (dr.'Y
(y- 1
< -1> 1
-1)+-(d-1) J
y-1 rr. T

= T~~\) [ Ty(r~y- 1 )- a)+ T(a -1)- r~'Y'Y_ 1 > (dr~y- 1 ) -1) + (d -1) J
(2.55)

= RTE [Ty(r~y- 1 )- a)+ T(a -1)- (y'Y_ 1 > (dr~y- 1 ) -1) + (d -1)]
(y-1) '·
(2.56)

The mean effective pressure is

- RTE
p =
m (y-1)(V4·- V3)

x [ Ty(r~y- 1 )- a)+ T(a -1)- ,~'Y'Y_ 1 > (dr~y- 1 ) -1) + (d -1) J

but
Elementary Thermodynamics of Cryocoolers 65

so that
- RTE
p =
m ( 'Y -1)(d -1) v3

x [ ry(r~y- 1 ) -a)+ r(a -1)- f~Y'Y_ll (dr~y- 1 ) -1) + (d -1)] (2.57)

In dimensionless form with respect to the maximum pressure, Pmax =

p 2 = bp 3 and since RTE/V3p 3 = 1 then

P = b ('Y _ 1~(d _ 1) [ ry(r~y- 1 ) -a)+ r(a -1)- r~">''Y_ 1 l (dr~y- 1 ) -1) + (d -1)]
(2.58)

Thermal Regeneration
With perfect regeneration 0<7-4'l + 04' = 0(6-2') + 0(2'-3l so that Ts, the
temperature of the gas at the end of regenerative cooling, is Tc or TE.
With less than perfect regeneration, e < 1 (due to heat leaks or friction
effects) or with less than perfect external system cooling (so that T 6 > Tc
at the start of regenerative cooling) then T 5 , temperature at the end of
regeneration, will be greater than T 3 • The energy to reduce the temperature
of the gas from T 5 to T 3 will therefore have to be drawn from the
constant-pressure refrigeration process 04-7·
The heat extracted or available refrigeration will therefore be reduced
to
0 = 0(4-7)- 0(5-3)

There are two cases to consider:

For case (a)

1 -e (T- r-1)-
;-s, ~a
66 Chapter 2

For case (b)

For case (a)

0 = Q(4-7)-Q(5-3)

= Cv(T?- T4) -Cv(Ts- T3)

= cp(T?- r4)-cv{ r6[ 1-eC ~ 1) - r3 ]}

=CvTE( 1- r~y~l))-CvTE{T[ 1-eC~ 1 )-1 ]}

= rEt~~!1) (r~y- 1 ) -1)- Cv[T- e (r -1) -1]}

= rE[r~~!l) (r~y-1) -1)- Cv(T -1)(1- e) J

The coefficient of performance (Q/W) is, with W from Eq. (2.56)

TE[ (~!l)(r~y- 1 )-1)-Cv(T-1)(1-e-)]

COP= r,
(:~~)[,-y(r~y- 1 )- a)+ T(a -1)- r~}-o (dr~T- 1 ) -1) + (d -1)]
and since Cv = [ y/( y -l)]R and Cv = [R/( y -1)], then

(y~ 1 ) (r~y- 1 ) -1) + (T -1)(1- e)

COP= r' (2.59)
,-y(r~y- 1 )- a)+ (a -1)- (y~l) (dr~y- 1 ) -1) + (d -1)
r.
For case (b)
Q = 04-7- 02'-3- Os-2'
= Cv(T?- T4)- Cv(T2·- T3)- Cv(Ts- T2·)
= Cp(T?- T4- Ts+ T2·)-Cv(T2·_ T3)
Elementary Thermodynamics of Cryocoolers 67

and since CP = [y/(y -1)]R and Cv = R/(y -1)

Q=(y- 1
1 1 )-r+s(r-1)+;-
RTE ) { y [ 1-,~y- a] (a -r)}
-T-

Therefore the coefficient of performance is

COP=
( ~~~){
'
y[ 1- ,cY~ 1 l -r +s(r -1)+;]- (a ;r)}
e

(:~~) [ ry(r~y- 1 )- a)+ r(a -1)- ,~}_ 1 ) (dr~y- 1 ) -1) + (d -1) J

y[ 1- r~y~l) -r +e(r -1)+_;] -(a ;r)

(2.60)

Pseudo-Ericsson Cycle
Now consider the special case where the regenerative processes occur
only at constant pressure so that rc = re = r, b = e = 1, (1/ a)= d = T and
T 2 , s T 5 s T 6 only. The P- V and T-S planes are shown in Fig. 2.22. This
cycle may be called (after Rallis et al., 1977) the reversed pseudo-Ericsson
cycle. From Eq. (2.60)

Y[ 1- ,c!-1) -r+e(r-1)+;]-(;-1)
COP=------~--------------------------

ry ( r (y-1) - ;
1) +r (1; - 1) - ,cy-
'Y ( rr (y-1) - 1) + (r- 1)
1)

y[ 2- ,CY1_1) -r +s(r -1) J

y ( rr (y-1) - 1) -,cy-1)
-'Y- (rr (y- 1 ) - 1)

2- ( -1 -) - + e (T - 1)
,cy-1) T

(2.61)
(1-y)
and when s = 1, perfect regeneration,
(y-1) 1 )
COP= (,c:-1l (1-=- y) (2.62)
68 Chapter 2

1/4'

2'/3

v s
Fig. 2.22. Pseudo-Ericsson cycle.

The mean effective pressure is

Pm = (y- )~~1) v2' {Ty(r(y- 1) -~) + T[ ~- r(y'Y_1) (Tr(y- 1) -1) + (T -1) J}

RTE
(y-1)(T-1)V2'

(2.63)

and the mean effective pressure rate becomes

y(r(y-O -1)(Tr(y- 1) -1)

P= < o (y-1)(T-1) (2.64)
Tr'Y

Pseudo-Stirling Cycle
Another special case arises when the regenerative processes occur
wholly at constant volume. This may be thought of as the reversed pseudo-
Stirling cycle. In such a case, however, it is most unlikely that the heat
transfers to and from the system Q< 2 _ 6 l and Q< 4 _ 7 l would in fact occur at
constant pressure. Rather, it is more likely they would occur at constant
volume and therefore the equations derived above would not be applicable.
Elementary Thermodynamics of Cryocoolers 69

s
Fig. 2.23. Pseudo-Stirling cycle.

Let us consider the special case of the pseudo-Stirling engine with

constant volume processes of heat transfers to and from the system. The
T-S and P-V planes for this cycle are shown in Fig. 2.23. In this cycle let
rc = re =rand the other ratios are as defined above with a= d = 1, (1/b) =
e = T and T 3 s T 5 s T 6 only. The external work input to the cycle, W, is
P2 V2-P1 v1 P3 v3-P4 v4
W="---~--
(y-1) (y-1)

R
= - - [(T2- T1)- (T3- T4)]
(y -1)

RT3 [ (y-
= (y-
1) ( 1 )]
1) T(r -1)- 1- ,cy-1)

= ( :~~) [ (r(y-1) -1)( T - ,cy~ 1 l)]

= RT3 [( (y- 1 ) -1)( (y-ll -1)] (2.65)
(Y _ 1)(r(y 1l _ 1) r Tr

The mean effective pressure is

_ RT3 [( (y-1) 1)( (y-1) 1)] (2.66)

Pm = (y- 1)(/Y 1l)(r- 1)V3 r - Tr -

and since Pmax = T,Cy- 1 lP3 and RT3jp3 V3 = 1 the dimensionless mean
effective pressure ratio becomes
Pm (r(y- 1 ) -1)(T,Cy- 1 ) -1)
p = -- = --.,.27(-y------.-1')- - - - - (2.67)
Pmax Tr (y-1)(r-1)
70 Chapter 2

The refrigerating effect is

Q = Cv(T7- T4)-Cv(Ts- T3)
It is assumed that the gas is cooled from T 2 to T 6 by heat transfer from
the system to the surroundings prior to thermal regenerative cooling from
T 6 to T 5 • With perfect regeneration T 5 = T 3. With imperfect regeneration
T 5 2: T 3. Therefore if E = (Ts- T6)j(T3- T6) then

Ts = T6 + E (T3- T6) = [ 1- s(" ~ 1) JT6

Then the heat Q extracted is

Q = Cv(T?- T4)-cv[ 1-s(" ~ 1)r6- T3 J

but T1=TE, T4=TE/r<.,-o, and T6=TE/r, Cv=R/(y-1) and Cv=
-yR/(-y-1). Therefore

Q =( :~~) { ( 1- ,<-r1-ll)- 'Y[ r( 1- ~ 1 -1 J}

E 'T

RrE { cr<y-1) -1) }

=(-y- 1) (/"' 1)) -y[r-s(r-1)-1]

RTE [,<y- 1) -1
=(-y- 1) ,<-r 1) -y(r-1)(1-s)
J (2.68)

The coefficient of performance COP is

RTE ['(y-
( _ 1) ,(-r 1)
1 ) -1
-y(r-1)(1-s)
J
COP=~'Y--~------------------
RTE [( <-r-1) _ 1)(rr(y-1) _: 1)]
-y- 1 (,<-r-1)) r
(r(-r- 1 ) -1)- yr(-r-o (r -1)(1- s)
COP= (r 1 _ 1)(r 1_ 1) (2.69)

and if s = 1, ideal regeneration, then

,<-r- 1 ) -1 1
COP= (r (-r-ll - 1)( rr (-r-n - 1) = rr (-r- 1 l - 1 (2.70)

Many other special cases with different combinations of constant

pressure and constant volume regeneration and adiabatic heat transfers to
and from the system may be devised. Three are shown in Fig. 2.24. None
of these cycles are known to be identified with a particular name of an
inventor.
Elementary Thermodynamics of Cryocoolers 71

p T
3

4' 4

v (a) s 2
2'/3 5 6 2

p T 2'13

v (b) s 2
2

p T 3

v (c) s

Fig. 2.24. Three adiabatic cycles with various combinations of constant-pressure and constant-
volume regenerative processes.

PARAMETRIC EFFECTS

The above equations provide a basis for useful and interesting studies
of the effects of various parameters of the cycle. By way of example, results
computed for the ideal Stirling cycle are presented in Fig. 2.25. The
compression temperature Tc was assumed to be 320 K and the refrigeration
temperature, TE, 80 K. The figure shows the coefficient of performance as
a function of the volume compression ratio with different values of regen-
erator effectiveness. It is interesting to note the marked deterioration in
coefficient of performance with even the slightest departure from perfect
regeneration, particularly at low volume compression ratios. The volume
compression ratio in actual Stirling cryocoolers rarely exceeds 2. At this
value of volume compression ratio in Fig. 2.25 the COP with a regenerator
effectiveness of 0.9 is only 0.11, a third of the Carnot value (0.33 with
perfect regeneration).
72 Chapter 2

0.4

..,
REGENERlTOR
EFFECTIVENESS

..ob--
--- -
0.3

.ov -
--
1/ /
..oy
I . /
v
1/
I
1" /
~
f..--

I ; / v v
0 .I

I / 1/ ..
Fig. 2.25. Coefficient of performance
of the ideal Stirling cycle as a function
~ f..-- of the cycle volume compression ratio
4 6 8 10 12 with different values of regenerator
VOLUME COMPRESSION RATIO effectiveness.

Similar results calculated for the Rallis adiabatic cycle are presented
in Fig. 2.26. This shows the coefficient of performance as a function of the
volume compression ratio at different values of the regenerator effective-
ness. Temperatures of the working fluid at the beginning of compression
and expansion were assumed to be 320 K and 80 K as before.
An appreciation of the drastic consequence of replacing isothermal
processes by adiabatic processes may be gained by comparing Figs. 2.25
and 2.26. With adiabatic expansion the available refrigeration is substan-
tially consumed in completing the regenerative cooling process and the
effect is progressively cumulative as the volume compression ratio is
increased. It is sobering to realize that in practice, with cryocoolers operat-
ing at normal speeds (say 10 to 20Hz), the compression and expansion
are more likely to be adiabatic than isentropic processes.
In a cryocooler the maximum pressure is important for it determines
the strength requirement of the cylinder wall and the forces acting on the
drive mechanism. To a first approximation the maximum pressure is indica-
tive of the weight of the engine. The mean effective pressure is indicative
of the cycle power input per unit of piston displacement and hence, to a
first approximation, of the refrigeration capacity per unit volume of
machine. The ratio of mean effective pressure to maximum cycle pressure
is therefore an important guide to the refrigeration output per unit weight,
the specific capacity of the cryocooler.
Elementary Thermodynamics of Cryocoolers 73

I l
IDEA~OTHERMAL
I
STIRLING CYCLE
I

----
0.3 e--. 1
.."'
u
z
:IE
n
I

~
a: REGENERATOR
"'...a.. 0.2
EFFECTIVENESS

0 \' € '"1.0
....
z £•0.99

"'
u
...iL " \\ - <•0.975-
1- •. 0.95

~ _..... ~ v
1- <•0925
8
k------- l----
r:::g<;~~
<•0.9

v
<•085 -
0. I
17 <•0.8

I .--

-
1/

II
2
/ 4 6
./
8 10 12
VOLUME COMPRESSION RATIO

Fig. 2.26. Coefficient of performance of the pseudo-Stirling cycle as a function of the cycle
volume compression ratio with different values of regenerator effectiveness.

In Fig. 2.27 the pressure ratio (flm!Pmax) is compared for the adiabatic
and the isothermal reference cycles as a function of the volume compression
ratio. It can be seen from this figure there is a marked and progressive
decline in the output per unit weight for both cycles with increase in the
volume compression ratio. The isothermal cycle is superior to the adiabatic
cycle.

\
0.7

0.6

. 0.5
~
a
£
j
\~
\'
0.4

i 0. 3
~ ISOTHERMAL CYCLE

~~ ..........
-
-
2

~
Fig. 2.27. Comparison of the pressure ratio: 0 .I
ADIABATIC CYCLE
>-----
mean effective pressure/maximum cycle
pressure (flm/Pmaxl for the ideal and the
pseudo-Stirling cycles as a function of the 0
0 4 6 10 12
volume compression ratio. VOLUME COMPRESSION RATIO
74 Chapter 2

OTHER REGENERATIVE CYCLES

Figure 2.28 shows four different diagrammatic arrangements for cryo-

coolers using regenerative heat exchangers to separate the ambient and
low-temperature regions of the machine. This diagram was reproduced
directly from Radebaugh (1977). There are slight but important differences
in terminology compared with our previous usage where quantities of work,
heat transfer, and temperature at low temperature were referred to without
a subscript. In Radebaugh's diagrams the subscript 0 is used for ambient
temperatures. At high temperatures the subscript h is used.
The four systems shown are representative of the two principal classes
of regenerative engine characterized by the method of fluid flow control,
namely, by volume changes or by valves. Machines of the type where flows
are controlled by volume changes may be classified conveniently as Stirling
engines (note the word cycle is not used). Similarly, machines of the type

Wo Wo W

O.T
(b) Stirling Cycle

Fig. 2.28. Composite diagram of four different regenerative engines (after Radebaugh, 1977)
where 0 represents ambient temperature and h high temperature.
Elementary Thermodynamics of Cryocoolers 75

where fluid flows are controlled by valves may be classified as Ericsson

engines.
In Fig. 2.28 there are two Ericsson engines (with valves), namely, the
Solvay machine, diagram (a) and the Gifford-McMahon machine, diagram
(c). There are also two Stirling engines (without valves), namely, the Stirling
machine, diagram (b), and the Vuilleumier machine, diagram (d).

Stirling Cycle
On the Radebaugh diagram the Stirling cycle is shown in Fig. 2.28b.
It consists of the following:
i. A piston receiving work, W 0 , from an external source during the
piston upstroke and generating work, W, during the piston down-
stroke. When operating as a refrigerator the net cycle input work
is (W0 - W).
ii. A cooler rejecting heat 0 0 at temperature T 0 from the working
fluid to the surroundings.
iii. A displacer mechanically coupled to the piston but operating out
of phase with it so that fluid is displaced between the ambient
temperature compression space and the low-temperature
expansion space.
iv. A regenerative heat exchanger acting as a thermodynamic sponge
and receiving heat from the working fluid when the fluid is passing
from the ambient to low-temperature region. It releases heat to
the working fluid when the fluid is passing back from the low to
ambient temperature region.
v. A freezer receiving heat, Q, or abstracting it from the surroundings
at low temperature, T. This is the useful refrigeration product of
the system.

Vuilleumier Cycle
Another regenerative cycle related to the Stirling cycle is the Vuil-
leumier (pronounced Viamay) cycle shown in Fig. 2.28d. This is is some-
times described as the Stirling cycle with a thermal compressor rather than
a mechanical compressor.
Comparison of the diagrams for the Stirling and Vuilleumier cycles
shows the lower halves for both systems (the low-temperature region) to
be identical. The principal difference is that the Vuilleumier cycle has no
mechanical compressor but has another displacer and regenerator arrange-
ment with an additional heat exchanger called the heater.
76 Chapter 2

The heater receives heat, Qh, at high temperature, Th. Motion

of the displacer causes fluid to pass through the upper regenerator from
the heated space above the displacer to the cold space below the displacer.
This transfer of the fluid at constant total volume between a heated and
cooled space causes the pressure to change. The pressure is greatest when
the fluid is concentrated in the hot space and least when the fluid is
concentrated in the cold space. The lower displacer assembly is directly
connected to the upper displacer assembly through the cooler. Therefore
pressure changes in the upper assembly are also experienced in the lower
assembly.
The cyclic pressure change will produce a refrigerating effect as in the
Stirling engine cycle provided that the fluid is concentrated in the cold
expansion space at the time when the pressure is decreasing from the
maximum to minimum value. The two displacers are therefore coupled as
indicated by the broken line connecting them in Fig. 2.28d. In practice the
displacers are not synchronized as the figure suggests but operate about
90° out of phase, with the low-temperature displacer leading the high-
temperature displacer.
Vuilleumier cycle engines can be made to operate with no input of
work from an external source. However, it is customary to provide a small
work input to facilitate starting and maintain the proper frequency of
operation. Vuilleumier cycle systems are well suited to produce refriger-
ation from a high-temperature thermal input derived from isotope degrada-
tion, concentrated solar energy, exhaust heat or combustion heating of
fossil fuels and solid wastes.

Solvay Cycle
The Solvay cycle engine shown in Fig. 2.28a consists of a compressor,
operating at ambient temperatures, compressing gas and receiving work
input W0 from an external source. The heat of compression Oo is rejected
from the system in the cooler at T 0 •
The compressed fluid is expanded in an expansion engine to produce
work output, W, and a low temperature of the fluid. The low temperatures
produced by expansion induce a flow of heat Q to the working fluid at
temperature T, in the freezer. A regenerative heat exchanger is interposed
between the ambient temperature compressor and the low-temperature
expander. The regenerator facilitates the attainment of very low tem-
peratures in the expansion space.
The Solvay engine is virtually identical to the opposed piston Stirling
engine shown in Fig. 2.9 used for the detail description of the Stirling cycle.
The only difference between that engine and the Solvay machine in Fig.
2.28a is the presence of valves to control flow in the Solvay machine. The
Elementary Thermodynamics of Cryocoolers 77

use of valves to isolate the compressor and the expander is an important

factor. It allows the frequency of operation of the compressor and the
expander to be selected independently and for the possibility of higher
pressure ratios (pmax!Pmin) than can be achieved in Stirling engines.
The use of a slow-speed expander with consequent low noise and
vibration levels is particularly important in the miniature cryocoolers used
for electronic applications at cryogenic temperatures. Some of the detector
elements used in these applications are incredibly sensitive and cannot
function reliably in the presence of the slightest mechanical vibration or
electromagnetic interference.
The isolated compressor can operate at high speed with all the heat,
vibration, and noise necessary without affecting the sensitive detectors used
in the cold region. It is customary to use standard commercial air or gas
compressors suitably modified for the compressor units of both Solvay and
the Gifford-McMahon cryocoolers. Standard compressors are available at
comparatively low cost and their use allows development effort to be
concentrated on the cryogenic region of the engine.

Gifford-McMahon Cycle
The Gifford-McMahon cycle system shown in Fig. 2.28c is similar to
the Solvay unit with the difference that a displacer is used instead of the
expansion engine of the Solvay system. The Gifford-McMahon engine is
very like the piston/displacer Stirling engine shown in Fig. 2.28b but has
valves isolating the compressor from the regenerator and displacer unit.
Use of valves requires the provision of two connecting ducts rather than
the single connecting duct of a Stirling engine. Separation of the compressor
unit and the use of valves to control the flow gives the Gifford-McMahon
engine the same advantage of independent compressor I expander operation
and high-pressure ratios as the Solvay engine.
The Gifford-McMahon engine has an advantage over the Solvay
engine. It uses a displacer rather than a piston. A displacer requires only
a low-pressure seal. A piston requires the use of a high-pressure seal
whereas this apparently trivial distinction is the biggest single reason why
Gifford-McMahon engines are used very widely for cryogenic cooling
engines whereas not a single Solvay engine is in production. Air Products
and Chemicals, Inc. sometimes refer to their cryocoolers as modified Solvay
engines. They are indistinguishable from Gifford-McMahon engines and
do not include the low-temperature expansion engine characteristic of
Solvay machines.
Parenthetically, it may be noted that all regenerative cryocoolers
include a recuperative heat exchanger for cooling and sometimes to improve
heat transfer at the expansion space.
78 Chapter 2

Part Ill: Recuperative Cycles

CRYOCOOLER WITH RECUPERATIVE HEAT EXCHANGER

Many cryocoolers utilize recuperative heat exchangers rather than the

regenerative exchangers considered above. In a recuperative exchanger
separate flow passages are provided for both the hot and cold fluid streams.
Usually the fluids flow continuously at steady state. The exchanger is used
more effectively if the fluids flow in opposite directions (counterflow) rather
than in the same direction (parallel flow).
The teminology of recuperative cryogenic cooling systems is unfortu-
nately as confused as that for regenerative systems. Here we shall adopt
terms in general usage although it should be understood these are not
accepted universally.
A cryocooler utilizing counterflow recuperative heat exchangers is
shown in Fig. 2.29. The corresponding P-V and T-S planes are shown in
Fig. 2.30. The system consists of the following:
i. A compressor operating at ambient temperature. We assume com-
pression to be the isentropic process (1-2) shown on the P-V and
T-S planes.
ii. A water-cooled heat exchanger for cooling the gas to ambient
temperature in the constant-pressure cooling process (2-3).
iii. A counterflow recuperative heat exchanger cooling the gas to a
low temperature in the constant-pressure cooling process (3-4).
iv. An expansion engine operating at low temperature wherein the
gas expands, process (4-5). The process is assumed isentropic and
the gas at the exit from the engine is at temperature T 5 , the
minimum temperature attained in the cycle.

EXPANDER
COMPRESSOR

Fig. 2.29. General arrangement for a Joule-Brayton cycle refrigerating engine with recupera-
tive heat exchanger.
Elementary Thermodynamics of Cryocoolers 79

v
COMPRESSOR T
Pmax

v s
EXPANDER

Fig. 2.30. Pressure-volume and temperature-entropy diagrams for a Joule-Brayton cycle

refrigerating engine with recuperative heat exchanger.

v. A "freezer" heat exchanger wherein the cold gas leaving the engine
is heated in the constant-pressure heating process (5-6). The heat
is abstracted from the surroundings of the freezer and is the useful
refrigeration of the cycle.
v1. The low-pressure fluid returns through the low-pressure flow
passages of the counterflow recuperative heat exchanger and is
heated in the constant pressure process (6-1).
It is possible for the expander and cooler to be coupled so the output work
of the expander helps to drive the compressor. This reduces the work
required by the compressor. However, in many systems it is not convenient
for the expansion engine and the compression engine to be coupled. They
may be physically separated. The machines may even be of different types.
It is not unusual to find a high-speed expansion turbine supplied with fluid
compressed in low-speed reciprocating compressors. Furthermore, the
expansion work may be negligible compared with a compression work. It
is simply dissipated in a friction brake or used to drive a small generating
set or pump.

JOULE-THOMSON EXPANSION

Frequently no expansion engine is used. Rather the fluid is expanded

from a high pressure to a low pressure in a throttled expansion. This occurs
when fluid expands through a small orifice with high frictional degradation.
It is called a "Joule-Thomson expansion" and occurs at constant enthalpy.
80 Chapter 2

It is discussed in detail in Chapter 6. Here it is sufficient to indicate that

a JT expansion shown in Fig. 2.31 results in a decrease in pressure from
p 1 to p 2, a decrease in temperature from T1 to T2, and an increase in
entropy from S 1 to S 2 • A decrease in temperature occurs when the expansion
takes place near the liquid/vapor envelope of the T-S diagram. At higher
temperatures the expansion occurs without change in temperature or may
even result in an increase in the temperature.
Joule-Thomson expanders are used in many liquefaction systems, as
shown in Fig. 2.31b. The pressure decreases from p 1 to p 2 along the constant
enthalpy line (a-b) and the fluid becomes a two-phase mixture of saturated
liquid and saturated vapor. At the end of the expansion, point (b), the fluid
exists in the two saturated states (c) and (d). The fraction of fluid existing
as saturated vapor is the vapor quality:
(2.71)
where .::lS(c-b) is the change in entropy per unit mass between (c) and (b)
and .::lS(c-ctl is the change in entropy between the saturated vapor and liquid
states (d) and (c), respectively.
The density of the saturated liquid is greater than the vapor so the
two phases readily separate by natural gravitational action. The liquid is
drawn off as the useful product for refrigerative cooling. The saturated
vapor is returned to the compressor through a recuperative heat exchanger
for recycling. With a liquefier it is necessary to add new fluid at low pressure
and ambient temperature equal to that drawn off as saturated liquid.
It is possible to accomplish the liquefaction directly in an expansion
engine. This is also shown in Fig. 2.31b by the isentropic process (a-f).
The "yield" of liquid product is very much higher than in the constant
enthalpy (JT) process (a-b). This is because in the isentropic process there
is a decrease in the enthalpy.

s s
(a) GASEOUS PROCESS {b) LIQUEFACTION PROCESS

Fig. 2.31. Joule-Thomson expansion.

Elementary Thermodynamics of Cryocoolers 81

In practice it is undesirable to have a two-phase system present in the

engine. In turbine systems the liquid droplets rapidly erode the turbine
blades. In reciprocating engines the liquid tends to accumulate on the
cylinder walls and piston crown. When a fresh charge is admitted the change
in temperature from T 2 to T 1 (Fig. 2.3lb) causes the liquid droplets
remaining to evaporate and reduce the intake of fresh charge.
The preferred alternative to direct liquefaction in the engine is shown
in Fig. 2.32. This utilizes an expansion engine for gaseous isentropic
expansion to a low temperature and a JT valve to accomplish the liquefac-
tion process.
This combined system operates as follows:
i. Low-pressure gas enters the compressor at state (1). The gas
stream comes from the high-temperature recuperative exchanger
(A) and is supplemented with new gas entering at state (0).
ii. The compressed gas at state (2) is cooled in a water cooler to
state (3) and in the high-temperature recuperator to state (4).
iii. Part of the cooled gas expands in the expansion engine to state
(5). Here we assume the process is isentropic.
iv. The remaining fraction of high-pressure gas is cooled in the
low-temperature recuperator (B) to state (6).

Fig. 2.32. Claude cycle refrigeration

and gas liquefaction system.
82 Chapter 2

v. The low-temperature gas at state (6) expands in a JT cooler to

state (7), a two-phase fluid comprised of saturated liquid at state
(9) and saturated vapor at state (8).
VI. The two-phase fluid separates in the separator and saturated
liquid is drawn off (state 9) at low pressure as the useful product
of the cycle.
vii. The saturated vapor at state (8) combines with the engine exhaust
stream at state (5) to form a mixture at state (10).
viii. The low-pressure fluid at (10) enters the low-temperature
recuperator (B) and is heated to state (11).
ix. Finally the low-pressure stream enters the high-temperature
recuperator (A) and is heated to ambient conditions at state (11).

JOULE-BRAYTON CYCLE

The system shown in Figs. 2.31 and 2.32 operates on a thermodynamic

reference cycle known variously as the Joule cycle (in England) and the
Brayton cycle (in North America). Here we shall not adjudicate the merits
of this historical controversy but will recognize common practice to accord
the credit equally and use the term Joule-Brayton cycle.* The cycle was
originally conceived as the reference cycle for power systems and is widely
used for gas turbines. In that application the gas is compressed at ambient
temperature from a low pressure to a high pressure [process (5-4) in Fig.
2.30], heated in a combustion chamber to the maximum temperature
[process (4-2) in Fig. 2.30] and then expanded in the turbine [process (2-1)
in Fig. 2.30]. This is the simple Joule-Brayton cycle. It is possible to utilize
heat contained in the exhaust gas to heat the compressed gas and thus
improve the engine thermal efficiency. When this is done the reference
cycle is called the regenerative Joule-Brayton cycle. The term regenerative
is widely used even though the actual heat exchanger may be either a
recuperative or regenerator.
It is correct to use the term Joule-Brayton cycle for refrigerating
applications as well as power systems. The cycle is made up of reversible
processes without regard to the direction of heat flow. However, the cycle
is never used in refrigerating applications without a recuperative or regen-
*The matter is beyond dispute for Joule's paper of 1851 ("On the Air Engine," Phil. Trans.
1852, 65-77, London) clearly precedes Brayton's work of 1873 ("Combustion Engine
Processes," Lichty L. C. 1967, McGraw Hill, New York; also "The Brayton Ready Motor
or Hydrocarbon Engine," Sc. Am. 34, 1876). Actually both Joule and Brayton were well
behind John Barber, whose English patent of 1791 clearly describes a turbo expander
driving a reciprocating air compressor, operating on what is now known as the Joule or
Brayton cycle. Barber's work is virtually unknown but is described briefly by Norman Davey
in "The Gas Turbine," 1914, D. Van Nostrand Co. Inc., Princeton, New Jersey.
Elementary Thermodynamics of Cryocoolers 83

erative heat exchanger. The term Joule-Brayton cycle used here should
therefore be understood as the abbreviated form of "Joule-Brayton cycle
with recuperative constant pressure heat exchange."
When the cycle is used with a regenerative heat exchanger it is the
cycle we have earlier called the pseudo-Ericsson cycle (see Part II, Chapter
2).

SIEMENS CYCLE
The Siemens cycle is a reference cycle similar to the Joule-Brayton
cycle but with an isothermal compression (process 1-3 in Fig. 2.30). The
cycle was named after Sir Charles Siemens who was granted a patent in
1857 for a refrigerator working approximately on the reference cycle. The
isothermal work of compression is substantially less than the isentropic
work. Rotary machines are always assumed to operate with isentropic
processes but reciprocating compressors run slower than rotary systems
and have better means of cooling. Sometimes it is more appropriate to
assume their ideal compression process to be isothermal rather than isen-
tropic. We could also assume the expansion process to be isothermal
[process (4-6) in Fig. 2.30] rather than adiabatic [process (4-5) in Fig.
2.30]. This cycle with isothermal compression and expansion and constant
pressure recuperative heat exchange has not been named. With regenerative
heat exchange the cycle is of course the ideal Ericsson cycle.

LINDE-HAMPSON CYCLE
A cryocooler utilizing a Joule-Thomson expansion is shown in Fig.
2.33. This is known as the Hampson, the Linde-Hampson, the Linde, or

~~~ Compressor
Compressor
0 0 .T 0 Valves

J.T. Valve

O.T

Fig. 2.33. Hampson liquefaction cycle (after Radebaugh, ja) Hampson Cycle
1977). (Joule- Thomson Cycle)
84 Chapter 2

simply the Joule-Thomson cycle. It consists essentially of a compressor, a

water-cooled cooler, a recuperative heat exchanger, a Joule-Thomson
expansion valve, and a freezer heat exchanger or a liquid/vapor separator
depending whether or not the fluid condenses.
The preferred terminology is to call the system shown in Fig. 2.33 the
Linde-Hampson cycle if there is simply a single stage of expansion through
a single JT valve as shown.

LINDE DUAL-PRESSURE CYCLE

More efficient systems can be made with two or more stages of

expansion in multiple JT valves. These are called Linde dual-pressure cycle
systems. The essential difference between Linde-Hampson and Linde dual-
pressure cycles is that the Hampson cycle has only one JT valve and a
single stage of expansion. The Linde dual-pressure cycle uses several JT
valves and at least two levels of pressure are used.
A system where a single compressor serves a multiplicity of JT valves
all expanding over the same stage of expansion is a compound Linde-
Hampson cycle rather than a Linde dual-pressure arrangement. To mini-
mize confusion it is recommended that the term Joule-Thomson be restric-
ted to the actual expansion valve rather than the complete ensemble of
equipment in the Linde-Hampson cooler.

CLAUDE CYCLE

The addition to a Linde-Hampson cycle of an expansion engine and

additional stage of recuperative exchanger converts the system to a Claude
cycle. This was the system discussed above with reference to Fig. 2.32 and
arranged for a condensing working fluid downstream of the JT valve. Figure
2.34 shows a Claude system operating as a refrigerator with no condensation
downstream of the JT valve. A Claude cycle without the JT expansion
reverts to the Joule-Brayton system.
Multiple stages of expansion engines are used in Claude cycle systems.
The various expansions may occur in several individual engines or a long
P.iston of successively smaller diameters may operate in a long cylinder
with appropriate diameters to match the piston. This latter tandem com-
pound arrangement was used by Collins in the helium liquefier engines
operating in the three- or four-stage Claude cycle developed at the
Massachusetts Institute of Technology after the Second World War.
Elementary Thermodynamics of Cryocoolers 85

I
Q,T
Fig. 2.34. Claude cycle refrigeration system with no liquefaction of
the working fluid (after Radebaugh, 1977). (b) Claude Cycle

ANALYSIS OF RECUPERATIVE SYSTEMS

The elementary analysis of recuperative systems with continuous,

steady flow of the working fluid is relatively simple compared with regenera-
tive systems with continuously variable conditions.
Recuperative system analysis is best approached by use of the general
energy equation for steady flow. Consider the thermodynamic "black box"
shown in Fig. 2.35 with a variety of energy inputs to the box, including
kinetic energy, internal energy, flow work, and heat transfer. There are
also a variety of energy outputs from the box. When the system is in steady
state operation, the energy input must be equal to the energy output. If
this equality were not so, an accumulation or depletion of energy would
occur and the system would not be in steady state.

INTERNAL ENERGY HEAT TRANSFER

FLOW~ CD
+----;
INTERNAL
ENERGY

/
~

KINETIC/
ENERGY
1----+

WORK

!
Fig. 2.35. Steady flow thermodynamic black box.
86 Chapter 2

Summarizing the energy inputs and equating these to the energy

outputs we can express the general steady flow energy equation in mathe-
matical terms as
mZ1 +mCf+mU1 +p1 V1 +0 = mZz+mC~+rhUz+pz Vz+ W
(2.72)
where m is the mass rate of flow of the fluid, Z is the height of fluid above
some prescribed datum, C is the velocity of the fluid, U is the internal
energy of the fluid, p is the pressure of the fluid, V is the volume of fluid
flow, 0 is the net heat transfer to the fluid, and W is the net work output.
Frequently the terms (mU +pV) can be combined to the new term mH
where H is the enthalpy of the fluid. In many cases the change in the
datum heat [m(Zl-Zz)] is negligibly small.
Then Eq. (2.72) can be restated as

m(Hl-Hz)=m(C~-Ci)+ W -0 (2.73)

The change in enthalpy from state 1 to state 2 is therefore the sum of the
change in kinetic energy, the work done, and the heat transferred.
Elementary analysis of recuperative cryocoolers can be approached
by considering enthalpy values at the different locations of the system.
Consider the Joule-Brayton cycle system shown in Fig. 2.30. The change
in enthalpy per unit mass flow (Hz- H1) during compression is the work
of compression that must be supplied to the cycle. This assumes the velocity
of the gas at 2 is about the same as at 1. The heat transfer 0 during
compression is zero because the process is assumed to be isentropic (i.e.,
A.S = A.O/T = 0).
Similarly, the heat transfer 0<3 -z> from the high-pressure fluid in the
water-cooled heat exchanger is (Hz- H3) and in the recuperative heat
exchanger 0<3-4) = (H3- H4). The work produced in the expander W4-s is
(H4 - Hs) and may or may not be applied to partially offset the work of
compression. Finally the useful refrigeration of the cycle is 0<6 -s> =
(H6 - H 5 ) and the heat transfer in the recuperative exchanger is 0<1-6) =
(Hl-H6).
The net work input to the cycle, Wno (assuming the expander work
output is used to reduce the compression work) is

(2.74)

The heat extracted in refrigeration is

(2.75)
Elementary Thermodynamics of Cryocoolers 87

Therefore the coefficient of performance is

(2.76)

COMPRESSORS AND EXPANDERS

Compression and expansion processes have so far been assumed to

be isothermal or isentropic. In practice they are neither of these but occur
in a manner that is characteristic of the machine used following what is
termed generally a polytropic process.

Compressors
Figure 2.36 illustrates on the P- V and T-S planes, the different
processes for compression.
1. Process (1-2) is an isothermal compression represented by PV =
const.
ii. Process (1-4) is isentropic represented by p V, = const. y is the
ratio of specific heats of the gas, y = CP/ Cv.
m. Process (1-3) is polytropic between an isothermal and an isentropic
process. The compression can be represented by p V" = const,
where (1 :5 n :5 y ).
IV. Process (1-5) is polytropic represented by the equation pVm =
const, where m ~ y.
The ideal reference cycle for reciprocating compressors is the cycle with
isothermal compression. For rotary systems the isentropic compression

4
P maa

PV=C

I.
Pv"=c
T
p
I I
pvY.c
I,( I
P mon 7 PV •C

v s
Fig. 2.36. Compression processes on the P-V and T-S planes.
88 Chapter 2

cycle is the ideal reference cycle. In a rotary machine at high speed the
gas particles are moving so fast that there is simply no time for cooling to
take place. In the reciprocating compressor the gas passes through the
machine at a slower pace and so some cooling is possible. To achieve
isothermal compression would require either infinite rates of heat transfer
or an impracticably low speed of operation. In practice reciprocating
compressors follow the intermediate polytropic path, process (1-3) on Fig.
2.36. Rotary compressors follow the polytropic process (1-5).
The isothermal efficiency for a compressor (usually reciprocating) is
defined as
isothermal compression work area 1-2-6-7 Hz-Hz
T/(iso) =
actua l compression
.
war k area 1-3-6-7 H3-Hl
(2.77)
The isentropic efficiency for a compressor (usually rotary but sometimes
also used for reciprocating) is defined as

adiabatic compression work area 1-4-6-7 H4-Hl

T/(isen) =
actual compressiOn
. wor k area 1-5-6-7 H 5 -Hl
(2.78)
Areas are measured on the P- V work diagram. Enthalpies, H, can be
calculated or simply read from the appropriate point on the T-S diagram.
The compression work measured by areas on the P- V plane or changes
in the enthalpy, is called the indicated work. This is considerably less than
the work actually required to drive the compressor. Additional work is
necessary to overcome mechanical and aerodynamic friction losses, pneu-
matic losses in charging [process (7-1) in Fig. 2.36] and discharging [process
(6-2)] and for driving accessories.
The ratio of indicated work to actual shaft input is called the mechanical
efficiency:
T/(mechl =indicated work/brake work (2.79)

This usually ranges from 60% to 80% for reciprocating engines and up to
90% for rotary systems.
Reciprocating compressors have crankshafts, with main bearings, con-
necting rods with big and small end bearings, pistons with piston guide
rings and sealing rings. They may also be equipped with water cooling
requiring some pump work, with forced lubrication requiring oil pump
work and a fan for air cooling. The mechanical and hydraulic losses tend
to be appreciable compared with the pneumatic effects arising from the
relatively slow moving gas being compressed.
Elementary Thermodynamics of Cryocoolers 89

Rotary systems have only main shaft bearings and rotary shaft seals
which consume little power. The principal losses are pneumatic arising
from friction and windage effects of the high-speed rotation of the disks,
drums, and blades within the compression space.
Frequently the isentropic efficiency is quoted for reciprocating com-
pressors rather than isothermal efficiency. This was adopted by compressor
manufacturers to put their machines in a better light since the ratio of the
isentropic work to the actual shaft input is larger than the ratio of isothermal
work to actual shaft input. Both are arbitrary ratios and the actual work
required remains the same regardless of the efficiency used. When discussing
compressor efficiencies it is important to establish if the efficiency being
quoted is isentropic, isothermal or mechanical.
Yet another efficiency is much used, the volumetric efficiency, defined
as
volume of gas actually pumped
Tl(vol) = d'1sp1acement vo1ume of compressor (2.80)

The volume of gas passing through the compressor is always less than the
piston displacement volume. The reduction is due to a number of causes
illustrated by the work diagram shown in Fig. 2.37. This shows the ideal
work diagram 1-2-3-4 superimposed on an actual compressor indicator
diagram. The effects have been exaggerated for the sake of clarity. Due
to pressure losses in the intake duct (perhaps including a filter) the pressure
at the start of compression (5) is less than the theoretical value (1).
The mass of gas in the cylinder can be calculated as
PsVs
me= RTs (2.81)

where me is the mass of gas in the cylinder, ps, Vs, and T 5 are the pressure,
volume, and temperature at point (5), and R is the characteristic gas

7 6

Fig. 2.37. Comparison of actual and ideal com-

pressor work diagrams. v
90 Chapter 2

constant. Clearly a reduction in the pressure will mean a reduction in the

mass of gas contained.
The actual compression process follows the path (5-6) rather than the
ideal path (1-2). It is necessary to compress the gas to a higher final pressure
(6) than theoretically required (2) so as to overcome pressure losses in the
delivery valve and ductwork including the aftercooler. This has no effect
on volumetric efficiency but requires extra work and reduces isothermal
or isentropic efficiency.
The principal effect on volumetric efficiency arises from residual gas.
It is not possible for all the compressed fluid to be discharged from the
compressor cylinder. A small amount of residual fluid (2% to 5%) will
remain in valve cavities and clearance spaces when the piston reaches top
dead center position. Now as the piston descends the compressed residual
gas expands from (7) to (8). No fresh charge will be inhaled to the cylinder
until the pressure falls to (8) and the inlet opens to admit new gas. A
sizable fraction of the piston stroke is consumed in this re-expansion of the
compressed gas and is the biggest factor reducing the volumetric efficiency.
Expansion of the clearance gas does work on the piston. Work input to
the compressor is therefore proportionately decreased as the mass flow
entering the compressor is decreased.
A reduction in inlet pressure or increase in inlet temperature will
reduce the mass flow to the compressor. When comparing one compressor
with another on the basis of volumetric efficiency or the "free air delivery
rating," it is important to specify the same reference pressure and tem-
perature. For comparative purposes it is best to reduce the available data
to terms of mass rate of flow at the standard temperature and pressure of
ooc and 101.3 kPa.

Multistage Compression
When gases are compressed over a large pressure range it is advan-
tageous to perform the compression in two or more stages with intercooling
between stages and aftercooling the final stage. In Fig. 2.38 the P- V and
T-S diagrams for two-stage compression are superimposed on those for a
single stage.
The single-stage compression diagram is represented by process 1-11-
6-10. The two-stage compression process is represented by the low-
pressure stage, process 1-2-8-9, and the high-pressure stage, process
3-4-6-7. At the intermediate pressure the fluid is assumed to be cooled
to the original temperature. Process 1-3-5 is the isotherm for the tem-
perature at state 1.
Elementary Thermodynamics of Cryocoolers 91

12
I
I
I
I
I
I
I
\ I
T \ /~13
\2/ l
I
I
I
II Pmin
I

v
Fig. 2.38. Multistage compression.

The advantage of multistage compression is twofold. It increases the

volumetric efficiency and also the isothermal (or isentropic) efficiency.
Increase in the volumetric efficiency comes about because in a two-stage
machine the clearance gas expands only to (9) rather than the larger volume,
(10), with single-stage compression. The low-pressure stage can admit a
greater volume of new gas. The isothermal or isentropic efficiency increases
because of the reduced compression work required with intercooling back
to the initial temperature. The saving in work is indicated by the shaded
area, 2-11-4-3 in Fig. 2.38.
The disadvantage of multistage compression is that a more complicated
and expensive machine is required with more moving parts, bearings,
pistons, seals, valves, and intermediate pressure heat exchangers and inter-
coolers. The benefits of additional stages of compression occur at a progress-
ively diminishing rate. It is rarely worthwhile exceeding four stages even
for the highest pressure ratios.
To a first approximation the optimum division of pressure staging for
equal compression work in each stage is:
(2.82)
where p, is the pressure ratio per stage, n is the number of stages, Pmax is
the maximum pressure (at end of compression), and Pmin is the minimum
pressure (at start of compression).

Expanders
Processes for expansion are shown in Fig. 2.39. From an initial high
pressure the fluid may expand:
i. Isothermally, process (1-2) represented by pV =canst.
92 Chapter 2

LATE CUTOFF

Pmax
I-------,
""""6
PV•C ',

Pv" .. c ' ', Pmox

I
........ 7
pyY •C T
p
PVm·CI
I
I
Pmin !"'8_L_.L._~:....L~~'--L..~

v s
Fig. 2.39. Expansion processes on the P-V and T-S diagrams.

ii. Isentropically, process (1-4) represented by p V "Y = const, where y

is the ratio of specific heat at constant pressure and constant
temperature (y = Cp/Cv).
iii. Polytropically, process (1-3) represented by pV" =const where
1 :::; n :::; y.
iv. Polytropically, process (1-5) represented by the equation pVm =
const where m :;;;:y.
The process of expansion is never isothermal or isentropic but in both
reciprocating or rotary machines lies somewhere between the two [process
(1-3)]. It cannot be isothermal for this would require infinite rates of heat
transfer between the gas and its surrounds. Similarly it cannot be isentropic
for this would require perfect insulation. In a turbine aerodynamic friction
causes a higher exit temperature, T 3 , than with isentropic expansion. This
frictional degradation causes an increase in the entropy of the fluid during
expansion and is in accord with the second law of thermodynamics.
The expansion process (1-5) cannot be attained naturally for it requires
the withdrawal of heat during expansion with consequent reduction of
entropy. It would occur in an expander equipped to refrigerate the
expansion process and is of little interest. Our main purpose is to use the
expansion process to generate refrigeration rather than consume it.
In reciprocating expanders point 1 on the P- V diagram is called the
cutoff point, where admission of working fluid is terminated and expansion
begins. After expansion the exhaust valve opens and fluid is expelled from
the cylinder.
The complete cycle for an expansion engine is the P- V diagram
1-2-8-9. The area of the P-V diagram is the work output of the engine.
The refrigeration generated is
i. dQ = dU +dW for a nonflow system (reciprocating expander);
ii. dQ = dH for a flow system (rotary expander).
Elementary Thermodynamics of Cryocoolers 93

To increase the refrigeration the cutoff may be delayed to 6. The fluid

expands to the end of the piston stroke at 7, the exhaust valve opens and
the fluid "blows down" to low pressure 2. Delay in the cutoff point increases
the size of the P-V work diagram and hence the refrigeration effect but
at the expanse of increased use of compressed gas. The process of free
expansion (7-2) in the exhaust blowdown occurs at a Joule-Thomson
expansion (constant enthalpy) so the fluid is used less effectively than by
controlled expansion in the engine. The refrigeration produced by the
engine is increased but the refrigeration per unit mass flow is reduced.
An expander with variable cutoff can be used where a large refriger-
ation capacity is required for starting and cool-down purposes. Later the
cutoff point can be adjusted to provide maintenance levels of refrigeration.

ISOTHERMAL VS. ISENTROPIC EXPANSION

An interesting point was discussed by Radebaugh (1977) in connection

with the Siemens cycle and the use of isentropic or isothermal expansion.
Consider the Siemens cycle shown in Fig. 2.40. It consists of isothermal
compression (1-2), recuperative cooling at constant pressure (2-3), isen-
tropic expansion (3-4), refrigeration (4-5), and recuperative heating at
constant pressure (5-1). The heat extracted in the refrigeration process is
the area 4-5-7-6 on the T-S diagram. Now consider an isothermal
expansion (3-5) instead of the adiabatic expansion (3-4). With isothermal
expansion the refrigeration effect is greater by the area 4-3-5. However,
with adiabatic expansion the minimum temperature is much lower than

Fig. 2.40. Isothermal and isentropic expansion pro-

cesses. s
94 Chapter 2

with isothermal expansion. In situations where the heat leak from surround-
ings dominates the internal heat generation (usually the case with small
electronic or superconducting devices) it is advantageous to absorb heat
over a range of temperatures. In these situations a better comparison might
be to increase the temperature at the start of the adiabatic expansion to
generate the same quantity of refrigeration (at variable temperature) with
adiabatic expansion as with isothermal expansion. This modified adiabatic
expansion is represented by the process (8-9). Refrigeration is available
at variable temperatures in the process (9-10). The amount of refrigeration
is area 9-10-12-11 and has the same area as the isothermal process
3-5-7-6. With refrigeration at variable temperatures the duty and hence
the size of the recuperative heat exchanger is decreased since the tem-
perature is only reduced to (8) rather than (3).

REFERENCES

Angrist, S. W. (1968). "Perpetual Motion Machines." Sci. Am. 218, 114-22.

Kohler, J. W. L. (1965). "The Stirling Refrigeration Cycle." Sci. Am. 212(4), 119-127.
Radebaugh, R. (1977). "Refrigeration Fundamentals: A View toward New Refrigeration
Systems. Applications of Closed Cycle Cryocoolers to Small Superconducting Applica-
tions." NBS Special Pub. 508, Supt. of Doc. U.S. Govt. Printing Office, Washington, DC.
Rallis, C. J., Uurieli, 1., and Berchowitz, D. M. (1977). A New Mathematical Model for
Stirling Cycle Machines. Proc. 12th I.E.C.E.C., Paper No. 779254, Washington, D.C.,
August 28-September 2.
Reitlinger, J. (1876). "Uber Kreisprozesse zwischen zwei Isothermen." Z. Ost. Ing. Arch. Ver.
Rogers, G. F. C., and Mayhew, Y. R. (1967). Engineering Thermodynamics Work and Heat
Transfer. 2nd Edition. S.l. Units, Longmans Group Ltd., London.
Chapter 3

Stirling Cryocoolers

INTRODUCTION: HISTORICAL REVIEW

A Stirling engine operates on a closed thermodynamic regenerative cycle

with the same working fluid repeatedly compressed and expanded at
different temperature levels so there is a net conversion of heat to work
or vice versa. It can be used as a cooling engine, prime mover, heat pump,
or pressure generator.
As a prime mover, the Stirling engine has a long history dating back
to 1817 and the invention of the first regenerative engine by the Reverend
Robert Stirling. It is currently receiving an unprecedented level of research
and development interest for applications ranging from miniature engines
for artificial hearts to automotive engines and large stationary power
generators using combustible wastes. Walker (1980) has recently reviewed
this work in detail.
As a cooling machine, it extends back nearly as far. In 1834 John
Herschel conceived the closed cycle regenerative cooling engine for making
ice. The concept was not reduced to practice for 30 years. Alexander Kirk
(1874) described a Stirling cooling engine in use for 10 years. Robert
Stirling and his engineer brother James, did no work on Stirling engines
as cooling machines. Nor did the hundreds of others busy developing the
many variations of hot-air engines made in large quantities in the 19th
Century in Britain, Europe, and North America.
The first Stirling cooling engines were made by Alexander Kirk, a
Scottish engineer working at the oil works at Bathgate, Scotland. He
constructed cooling engines for a variety of applications both in Great
Britain and overseas. In addition, he established the cascade cycle now
widely used for cryogenic and low-temperature gas liquefaction and pro-
cessing. Collins (1958) credits him with the accolade of "scientific thinker
as well as talented machine designer." Despite this promising base, the
95
96 Chapter 3

Kirk machines were never made in larger numbers or used to achieve

cryogenic temperatures.
The next significant development occurred with the start of the Philips
cooling engine research about 1946. It followed a decade of effort already
invested in small Stirling engine prime movers. The Philips program was
the first time that substantial effort was invested simultaneously in Stirling
engines for power and cooling applications. Although there were separate
development teams, they were located in the same research building and
cooperated well together with members common to both groups.
Kohler (1965), the leader of the cooler development group, has pro-
vided an intensely interesting account of the early program at Philips. He
recalls that, in 1945, H. Rinia and F. K. du Pre operated a one horsepower
engine as a refrigerator and obtained temperatures in the liquid air range
of 80 K. Thereafter, research on the Stirling cooling engines was assigned
to a separate group with Kohler in charge. Introducing him to the group
Rinia observed:
... our hot-air engines cooled marvelously, whereas when we built machines
expressly designed as refrigerators they were unsatisfactory.
Kohler writes:
I designed my first models along the lines of the hot-air engines and as a
result they never failed as refrigerators.
It was in the early 1950s that Stirling engines were recognized as a
specific generic class of machine. Up to then they had been generally
classified as "hot-air engines" with the name of a particular researcher
attached. This was no longer appropriate when hydrogen or helium sup-
planted air as the working fluid in high performance prime movers then
under development at Philips in Eindhoven.
Since that time "Stirling engine" has been widely accepted as the
generic title for closed cycle regenerative thermal systems whether used as
prime movers or cooling engines. This is preferred to the practice of some
who refer to the Kirk cycle when describing Stirling engines working as
refrigerators. It is worth separating the two principal classes of regenerative
machines into
i. Stirling engines (without valves),
ii. Ericsson engines (with valves),
depending on whether the internal fluid flows are controlled by volume
changes or valves.
Development of Stirling cryocoolers was brilliantly successful. Com-
mercial versions were in production by 1953; a remarkable achievement
Stirling Cryocoolers 97

compared with the parallel development of Stirling engine prime movers

where development continued over the past 40 years with no commercial
product so far. There are now clear signs this sustained effort will be
rewarded by successful commercial products (domestic heat pumps) in the
near future.
The manufacture of Stirling cryocoolers has been profitable for the
Philips Company. The first engine shown in Figs. 3.1 and 3.2 was an air
liquefier capable of producing 5 or 6 liters an hour within 15 minutes of
being switched on. It required no special expertise to install and maintain,
was small and sufficiently low priced to be within the means of any university
physics, chemistry, or engineering laboratory. In the quarter century since
its introduction, several thousand models of this engine and derivatives
have been made.
It is available in single-cylinder or four-cylinder versions (see Fig.
3.48) and may be equipped with a fractionating column to separate oxygen
from liquid air and produce liquid nitrogen. For many applications, this is
attractive because of the reduced safety hazard. A version of the engine

Fig. 3.1. Early Philips Stirling air liquefier.

98 Chapter 3

Fig. 3.2. Early Philips Stirling air liquefier.

with double expansion was also developed which can achieve temperatures
of 20 K and has been used for research and hydrogen, neon, and helium
liquefaction.
Derivatives of the early engine have been incorporated in many cry-
ogenic systems involving cooling or gas liquefaction and reliquefaction.
These include standard cabinets with cold air recirculation for metallurgical
processing, freeze drying, food processing, and a whole range of research
applications requiring refrigeration at precisely controlled temperatures in
the range from ambient to 20 K. It is used for reliquefaction of "flash" gas
in liquid oxygen or other cryogenic reservoirs.
In addition to the basic Stirling cryocooler, much other work has been
carried out in miniature engines and in the intermediate size coolers shown
in Figs. 3.3 and 3.4. No significant application for the intermediate size
 Cryoc oolers
Stirlil!g 99
100 Chapter 3

cooler has yet developed but will come with the growth of LNG transporta-
tion and storage systems and the eventual introduction of superconducting
systems.
Military interest in infrared thermal imaging equipment for night vision
and heat-seeking missile guidance focused research attention on miniature
cryocoolers in the 1960s. This is now sustained by continued military
interest, by spacecraft instrument applications, and by incipient applications
of superconducting electronic and electrical systems on a broad front .
The principal military interest in miniature cryocoolers was manifest
in the United States. Policies of the U.S. Government led Philips to
concentrate their development and production of miniature, militarily
supported cryocoolers at the associate company North American Philips.
The research facilities of the company are located in Briarcliff Manor, New
York and the production facilities were at Ashton, Rhode Island. In the
early 1970s, the production facilities were sold to a competitor, Cryogenic
Technology Incorporated (CTI), of Waltham, Massachusetts. CTI also
distributes in the United States the larger Philips cryogenic equipment
made in Eindhoven.
The research laboratory continues at Briarcliff Manor principally
engaged in research and prototype development of Stirling cryocoolers and

Fig. 3.4. Intermediate capacity Stirling cryocooler.

Stirling Cryocoolers 101

prime movers supported by various agencies of the United States govern-

ment. There is a close relationship between the Philips Company of Eind-
hoven and Philips Laboratories, although the companies are independent.
Philips Laboratories also has an uneasy cooperative, but yet competitive,
association with CTI and, perhaps more significantly for the future, much
common interest with the Magnavox Corp., the United States affiliate of
Eindhoven-based Philips Company.
The Philips Laboratory has assembled a remarkably competent
resourceful and innovative scientific engineering team under the leadership
of A. Daniels and F. K. du Pre. In the Bibliography many reports and
papers originating from this team will be found under the authorship of
Daniels, du Pre, Pitcher, Lehrfeld, Balas, Lindale, and Storace.
An excellent comprehensive summary of recent work on miniature
Stirling engines for cryogenic cooling can be found in the unclassified report

Fig. 3.5 . Miniature Stirling cryocooler with rhombic drive (after Lindale, 1978).
102 Chapter 3

by Lindale (1978). The development is described of six long-life low-power

cooling engines for gamma ray detectors in a satellite experiment. A
photograph of one of the miniature cryogenic cooling engines is shown in
Fig. 3.5 and in cross section in Fig. 3.6.
Research on Stirling engines continues at Eindhoven and has resulted
in the development and recent commercial introduction of an ultracompact
80 K cooler operating as a free piston Stirling engine with a linear motor
drive. The military version of this unit is shown in Fig. 3.7 (Haarhuis, 1978;
de Jonge, 1979).

-1----- 2nd STAGE

D ISPLACE A/REGENERATOR
COLD FINGER

1st EXPANSION STAGE

FLANGE

WORKING SPACE n.&---1stSTAGE

D ISPLACE A/RE G EN ERATOR
FILL VALVE

Fig. 3.6. Cross section of miniature Stirling cryocooler with rhombic drive (after Lindale, 1978).
Stirling Cryocoolers 103

Fig. 3. 7. Ultracompact free-piston Stirling cryocooler with linear motor drive (Philips Research
Labs).

The availability of research funds induced other companies to work

on Stirling cryocoolers. In the mid 1960s the Malaker Corporation, High
Bridge, New Jersey was active in miniature cryocooler development. With
the unfortunate failure of the company, the Malaker cryocoolers have
disappeared from the scene.
About the same time, Hughes Aircraft Co., Culver City, California,
developed small Stirling cryocoolers for heat-seaking infrared missile guid-
ance. Thousands of these engines have been made. The Martin-Marietta
Corp., Orlando, Florida, Texas Instrument Co., Dallas, Texas, and CTI,
Waltham, Massachusetts are other U.S. defense contractors with significant
research and development experience in miniature Stirling cryocoolers.
Much of their work has been sponsored by the U.S. Army Night Vision
Laboratory, Fort Belvoir for forward-looking infrared (FLIR) coolers for
night vision equipment. Figure 3.2 is an integral Stirling FUR cooler in
current production. Other work has been done on split-Stirling systems.
104 Chapter 3

Significant research programs directed to the engineering development

and eventual production of Stirling cooling engines have been variously
reported from both Japanese and Russian sources. Details of these develop-
ments are given in Chapters 13 and 14.

CLASSIFICATION

The essential elements of a Stirling engine are two spaces, at different

temperatures, whose volumes can be varied cyclically and which are con-
nected by ducts containing heat exchangers. These simple elements can be
combined in a surprisingly wide range of mechanical arrangements. Many
of these have been identified (by Finkelstein, 1959) with the name of the
originator. Many variations used in the 19th Century have been adopted
or reinvented for application in contemporary developments and new
arrangements are being devised.
Figure 3.8 is a family tree of fundamental characteristics on which the
following classification scheme is based. The key identifying feature of any
regenerative system is whether or not valves are used to control the flow
of the working fluid. If valves are used, the machine is an Ericsson engine.
If valves are not used the machine is a Stirling engine.

Fig. 3.8. Classification of cryo-

coolers.
Stirling Cryocoolers 105

The use of valves for flow regulation has the advantage of increased
flexibility in flow control and timing with the possibility of virtually unre-
stricted pressure ratios (Pmax/ Pmin). Valves add to the mechanical com-
plexity of the system and provide sources of noise and additional points
of wear so the prospect for long life with low maintenance is reduced.
Stirling cryocoolers, machines operating on a regenerative cycle
without valves, can be classified into two principal groups, namely, machines
with mechanically induced compression and machines with thermally
induced compression. In this review, machines with thermal compression
are treated separately as Vuilleumier Engines in Chapter 4.
Machines with mechanical compressors may be further subdivided into
two families identified as single-acting and double-acting engines.

Single-Acting Stirling Engines

A single-acting Stirling engine is an ensemble of expansion space,
compression space, and associated heat exchangers in one or two cylinders
with two reciprocating elements, one of which must be a piston. The other
may be a piston or a displacer. * Every ensemble constitutes a complete
system operating independently of other single acting systems coupled on
a common crackshaft or other kinematic mechanisms.
All existing arrangements of single-acting Stirling engines may be
subdivided into two groups as shown in Figure 3.9; two-piston machines
and piston/displacer machines.
A further distinction can be made in the piston/displacer group
between (a) machines where the piston and displacer operate in a single
cylinder called an integral-Stirling machine, and (b) machines where separ-
ate cylinders are provided for the piston (compressors) and the displacer
(expander), an arrangement called the split-Stirling machine.
In all these machines the reciprocating elements, the pistons or dis-
placers, may or may not be connected to a kinematic drive mechanism
(cranks and connecting rods, Scotch yoke or rhombic drive mechanism).
Machines with a kinematic drive mechanism are called disciplined piston
(and displacer) engines. Machines with no kinematic drive mechanism are
called free piston engines. Another variation has one of the reciprocating
elements (usually the piston) coupled to a kinematic mechanism and the
* There are important differences between a piston and a displacer. A piston sustains a high
pressure difference but no appreciable temperature difference between the upper and lower
faces. A piston must, therefore, have a good fluid seal to prevent leaks and act as a structural
element with high-pressure forces acting on it. A displacer is an element with a low pressure
difference and a high temperaure difference between the upper and lower faces. It can,
therefore, be a lightweight structural element with a minimal seal requirement but designed
with long thermal conduction paths to minimize heat leaks.
106 Chapter 3

SINGLE ACTING STIRLING

COOLING ENGINES

MAY HAVE EITHER PISTONS OR

FREE PISTONS (A ND OISPLACERS)

INTEGRAL-STIRLING SPLIT ST IRLING

PiSTON ANO OISPlACER PISTON ANO DISPlACER

IN THE SAME CY'LINOEA IN SEPARATE CYUN0£RS

A- PISTON
9 - OISPLACER
C- EXPANSION SPACE
0-COMPRESSION SPACE
E- REGENERATOR
F- FREEZER
G-COOLER

Fig. 3.9. Classification of single-acting Stirling cryocoolers.

other operating as free piston (or displacer). This has been called the
hybrid crank controlled (or disciplined) piston/free displacer engine. It is
widely used for split-Stirling coolers and can also be used for integral-
Stirling machines (see Walker, 1973).
A distinction also has to be made between machines with a separate
stationary regenerator and those where the regenerator is incorporated in
the displacer and moves with it.
Design variants for integral-Stirling engines are shown in Fig. 3.10
and for split-Stirling engines in Fig. 3.11. Design variants for two-piston
machines are shown in Fig. 3.12. The originator of a particular configuration
is given on these figures where it is known. In most cases, the system was
conceived for use as a prime mover and many of these cases have never
been made as cooling engines. This does not preclude future development
of new and different arrangements.
The spectrum of possibilities for two-piston arrangements is very wide.
It includes the conventional piston/cylinder systems shown in Fig. 3.12
and many others employing metal bellows and diaphragms, rotary assem-
blies and free-piston devices where motion is induced by fluidic forces
alone, or in combination with mechanical springs. The only two-piston
machine available commercially (from Philips) is the intermediate capacity
cooler shown in Fig. 3.4. This interesting machine utilizes a Joy crompressor
Stirling Cryocoolers 107

FREE-PISTON HYBRID CRANK DISCIPLINED

AND DISPLACER CONTROLLED PISTON/ PISTON AND
FREE-DISPLACER DIS PLACER

~ ~
KINEMATIC DRIVE
a) RHOMBIC
b) SCOTCH YOKE
c) CRANK/
CONNECTING ROO

{BEALE)

(WALKER)

OPTION REGENERATOR MAY BE INCORPORATED

WITH DISPLACER OR MAY BE SEPARATE
FIXED UNIT WITH EXTERNAL DUCT
Fig. 3.10. Design variants of integral
Stirling cryocoolers. (STIRLit.;G)

crankcase and pistons to generate the hydraulic drive for the two opposed
pistons of the upper, central Stirling engine cooler. It is a difficult machine
to classify. The Stirling system pistons are activated by fluidic forces only,
pneumatic on one side, hydraulic on the other, to make it a free piston

SINGLE-ACTING 1
SPLIT STIRLING

r ENGINE

FREE PISTON
I
HYBRID CRANK DISCIPLINED PISTON
AND DISPLACER CONTROLLED PISTON AND DISPLACER

Yg
/FREE DISPLACER
KINEMATIC DRIVE

~ (HIGA)

I
PARALLEL-CYLINDER
T
PARALLEL-CYLINDER
I

'\ i
TWIN-EXPANSION CYLINDERS AT RIGHT
REGENERATIVE EXTERNAL CYUNDER ANGLES REGENERATIVE
DISPLACE A REGENERATOR

F
DISPLACER

~
(LAUBEREAU
SCHWARTZKOPH) (HEINRIC!)
{RAINBOW)
(ROBINSON)

Fig. 3.11. Design variants of split-Stirling cryocoolers.

108 Chapter 3

MOVING CYLINDER f-------+1 STATIONARY CYLINDER

PARALLEL CYLINDER

,-
~ (RIDER)

V CYLINDER

~
(FINKELSTEIN)

r--

OPPOSED PISTON

~
L___

Fig. 3.12. Design variants of two-

piston Stirling cryocoolers.

engine. Yet, it has a kinematic vee cylinder crank connecting rod mechanism
through which the input work is supplied to the engine to generate the
hydraulic forces inducing the Stirling piston motion.
Regenerative Stirling engines using bellows are not uncommon in the
power systems field. The artificial heart engines of the McDonnel
Douglas/Richland Energy Laboratory and the Thermo-Electron Corpor-
ation incorporate a variety of bellows and diaphragms. The radio-
isotope/propane-fueled power generator developed by Cooke-Yarborough
at the U.K. Harwell Atomic Energy Establishment uses a diaphragm
(Walker, 1980).

Double-Acting Stirling Engines

Double-acting Stirling engines, shown in Fig. 3.13, are ensembles of
multiple cylinders arranged so the expansion space of one cylinder is
connected through associated heat exchangers to the compression space
in an adjacent cylinder. There is, therefore, the same number of Stirling
engine systems as there are cylinders, and only one reciprocating element
per cylinder. It may be called a piston-displacer, for it has both a large
Stirling Cryocoolers 109

act1ng I
I Double-
p1ston engmes J

I
Two- cylinder tw1n
system Compress1on
1n one cylmder and
expans1on 1n the
other for both
systems

(Franchot)

lnterconnecllng adjacent cyls

conslltute one system

Fig. 3.13. Classification of double-acting Stirl-

ing engines.
~ (Siemens)

pressure and a large temperature difference between the top and bottom
faces.
The great advantage of double-acting engines is that the number of
reciprocating elements is half the number required in multiple arrange-
ments of single-acting engines. This can lead to major simplifications of
the kinematic drive arrangements and to reduced costs.
The principal disadvantage is the limited flexibility in design, and to
a lesser extent, operation. Moreover, for prototype development, it is
necessary to proceed with the entire multiple cylinder engine rather than
an experimental single-cylinder unit which can then be reproduced in
multiple cylinder versions. Nevertheless, the elimination of half the moving
parts is so significant an economy, that the major interest for prime mover
Stirling engines greater than 15 kW is concentrated on double-acting
arrangements.
Double-acting engines were invented in the 19th Century. Babcock
(1885) attributes the invention of the two-cylinder, twin system to the
Frenchman, Charles Louis Franchot in 1853. It was reinvented a century
later by Finkelstein (1959) for the prototype unit of a domestic refrigerator
and was later incorporated by him in the concept of a multiple-cylinder
free-piston unit for power generation.
The general form of double-acting engines with three or more cylinders
is ascribed by Babcock to Sir William Siemens who, in 1863, conceived
 .....
.....
Q

Fig. 3.14. Four-cylinder double-acting Stirling engine invented by Sir William Siemens in 1863 (after Babcock, 1885).

(")
='
~
..;;-
~
Stirling Cryocoolers 111

the four-cylinder engine shown in Fig. 3.14. The engine had four cylinders
each containing a single piston and which were coupled in a square arrange-
ment through the associated heat exchangers and regenerators. It was
activated by a wobble-plate drive mechanism.
So far as is known, the Siemens engine was not taken beyond the
conceptual stage until it was reinvented, complete with wobble-plate, 80
years later by van Weenan in course of the Philips program. Following
some early success at Philips, the Siemens engine was abandoned in favor
of the single-acting "rhombic drive" engine. Later, in 1965, one of the
Philips licensees, General Motors, in the United States, resumed work on
double-acting engines and by 1970, Philips and the other licensees had
concentrated their efforts on double-acting engines for automotive engine
applications. The Philips unit, shown in Fig. 3.15, is a four-cylinder version
with a swash-plate drive* and exists in prototype 150- and 75-kW forms.
Until recently (1978), the engine was under development for automotive
application by Philips and the U.S. Ford Motor Co. in a 7-year $180 million
program funded by the U.S. Department of Energy. Ford abandoned its
development effort on Stirling engines in 1978.
An alternative kinematic mechanism has been adopted by the Philips
licensee United Stirling of Sweden using twin cranks and connecting rods
as shown in Fig. 3 .16. Another version with the cylinders arranged in a
shallow vee form was earlier favored. This used a single crankshaft but the
inclination of the cylinders increases both production and maintenance
costs. The parallel cylinder version is the unit intended for preproduction
evaluation in 40-, 75-, and 150-kW versions for automotive use, in mining
applications, for boats and stationary power generation (see Walker, 1980).
The United Stirling engine is the power unit for development of a Stirling
automotive unit in a program sponsored by the U.S. Department of Energy.
Mechanical Technology Inc. of Latham, New York and American Motors
are participating in the program.
Double-acting engines have not been used for cryogenic refrigerating
purposes. This is due entirely to the absence of need for cryogenic coolers
of intermediate or large capacity. Significant change is now at hand with
the development of relatively large-scale superconducting systems and
liquid natural gas transportation and reliquefaction systems. These develop-
ments will require new cryocooler concepts. Double-acting Stirling cryo-
coolers of intermediate and large-scale capacity will likely be prominent.
Already one major U.S. corporation is known to be actively developing a

*A distinction between a wobble-plate and a swash-plate should be made but is sometimes

difficult to appreciate. Maki et at (1971) define a swash plate as an inclined plate rigidly
attached to a rotating shaft. A wobble plate is not rigidly attached to a rotating shaft but
is mounted so that the plate nutates as the shaft rotates.
112 Chapter 3

BURNER UH.AVSl OUTLET

Fig. 3.15 . Philips double-acting Stirling engine with swash-plate drive (after Walker, 1979).

double-acting free piston Stirling cooling engine along lines enunciated by

Finkelstein (1978).
The first known application of double-acting cooling engines for
refrigeration was a Franchot twin-cylinder engine reinvented in England
by Finkelstein (1959). A practical form of engine was made and investigated
as the potential replacement for domestic refrigeration units. Minor funds
were provided and no significant attempts were implemented to optimize
Stirling Cryocoolers 113

Fig. 3.16. United Stirling double-acting Stirling engine with crank-connecting rod drive.

the design or incorporate the most suitable materials. Nevertheless, the

unit functioned well with a coefficient of performance that was somewhat
less than the conventional vapor compression unit in the normal domestic
refrigeration temperature range. At lower temperatures, the performance
of the Finkelstein unit was superior to the vapor compression machine.
The performance characteristics crossed in favor of the Stirling unit at
about - 20°C.
So far as is known, the earliest conceptual application of the Siemens
engine with swash-plate drive as a cooling engine was in an outline design
by Walker (1965), for a leading British vehicle components manufacturer.
This unit was intended as the air conditioning cooler for an armored fighting
114 Chapter 3

vehicle. The design was compact with a mass and volume half the conven-
tional vapor compression unit. Later on, other designs were evaluated for
railcar air conditioning systems and for cockpit cooling of fighter aircraft
and crop-spraying aircraft. In all cases, promising designs emerged but did
not proceed to prototype development.
The growth of interest and capability in single-acting free-piston
Stirling engines in the 1970s resulted in the development of concepts for
multiple-cylinder double-acting free-piston Stirling engine coolers and
prime movers. One engineering prototype development along the lines
proposed by Finkelstein (1978) is under way and further hardware develop-
ment is expected in the future.
One version of a three-cylinder free-piston Stirling cooling engine of
intermediate capacity is shown in Fig. 3.17. Three cylinders are used to
take advantage of the 120° phase difference in conventional3-phase electric
power supplies to energize the linear motor drive. Ingenious electrical
design is required to devise a motor capable of accepting high-frequency
(50 or 60 Hz) conventional power supplies and operating with a frequency
half to one third this value.

SPACE

"Ol'Ofl ORIVE
CONTAINED WITHIN
RECIPROCATOR

Fig. 3.17. D esign concept for three-cylinder free-pisto n double-acting Stirling cryocooler.
Stirling Cryocoolers 115

The optimum swept volumes of the expansion and compression spaces

of a cryocooler are markedly different. The ratio of swept volume
V (comp)/ V (expl is greater than unity and increases as the temperature
difference between the space becomes greater. At cryogenic temperatures,
the difference is as much as five times. To accommodate this difference in
swept volume, various possibilities are illustrated in Fig. 3.18. In case (a),
a stepped piston is used, with the diameter of the expansion piston being
smaller than the compression piston. The annular space above the com-
pression space can be utilized as a pneumatic spring to reduce the loading
on the piston drive mechanism during the downward compression stroke.
An alternative arrangement shown in Fig. 3 .18(b) has a double-ended,
stepped piston with central linear motor actuation to driye the piston. This
would be one cylinder of a multiple-cylinder arrangement with the spaces
all connected in appropriate fashion to ensure proper phasing between the
volume variations in one system and another. Yet another variant is shown
in Fig. 3.18c. The cylinder is equipped with a center post of substantial
diameter projecting into the displacer interior. A pneumatic spring cavity
is located below the center post. It is possible to engineer the system so
the center post is equipped with linear gas bearings at the upper and lower
ends to stabilize the location of the displacer in the cylinder. This accurate
location allows the use of close tolerance seals between the displacer and
cylinder enclosure so that rubbing seals are eliminated from the system.
Attachment of the armature of a linear motor to the underside of the
displacer and containment of the motor stator within the pressure cylinder

UHEAR G.AS
BEARING
SPRING
CAVITY

SPRING CAVITY

UNEAtt MOTOR

(b) (c)
(o)

Fig. 3.18. Alternative design features to accommodate the difference in swept volumes in the
expansion and compression spaces of double-acting Stirling cryocoolers.
116 Chapter 3

completes the arrangement for one cylinder of an advanced multiple-

cylinder free-piston double-acting Stirling system.

Multiple-Element Cooling Systems

Some military applications require multiple-element cryogenic spot
cooling. For example, multiple infrared detectors are used in the forward
and rear scanning missile detector systems of some manned bombers.
Future computer applications for superconducting elements will likely
involve multi-element cryogenic spot cooling. There are various ways to
achieve this. Figure 3.19 shows a three-cylinder system. The two outer
cylinders are identical and contain displacers A and B which may be
activated by a kinematic drive or by fluid forces only. The central cylinder
contains a piston which is double-acting. The underside of the piston is
connected to the left-hand displacer cylinder. The top side of the piston is
connected to the right-hand displacer cylinder.
Such an arrangement can be recognized simply as two single-acting
Stirling systems with a common double-acting compressor. This was first
applied as early as 1827 by the brothers Stirling, Robert and James, in a
25 horsepower Stirling engine used in a Dundee foundry. It ran at 45
revolutions per minute and was successful but suffered repeated failures
of the displacer cylinder hot ends due to the poor materials then available.
An alternative arrangement, shown in Fig. 3 .20, has multiple expansion
cylinders supplied from a common (single-acting) compressor cylinder. This
was identified by Walker and Burn (1976) as the Rainbow B configuration.
The Rainbow A configuration with a split compression space and a single
expander was adopted by a British engine designer, Horace Rainbow of
Bristol, in the design of a Stirling boat engine. Cryogenic cooling engines
of this type have been made by CTI, Waltham, Massachusetts for the U.S.

COMPR£SSION
S..ACE

Fig. 3.19. Two single-acting Stirling systems with

STlfH.tNG 11827) common compressor cylinder.
Stirling Cryocoolers 117

COMPRESSION SPACE

n
II
II
II
- - - -----II
----~ IL~

I:Jl]·
L_
L___J
Fig. 3.20. Multielement cryogenic spot RAINBOW I CHELLIS (1970)
cooling.

Air Force (Chellis et al., 1972). Up to 12 cooling heads may be supplied

from a single compressor with lead distances (length of duct between
compressor and expander) up to 60 in. In split-Stirling machines of this
type, the displacers are not coupled to a kinematic drive but are activated
only by fluidic forces, a pneumatic spring on one side and a variable pressure
due to the compressor on the other.

PISTON AND DISPLACER MOTION

In Chapter 2, the ideal Stirling cycle was observed to be a thermo-

dynamic cycle of two isothermal and two constant volume regenerative
processes having the pressure-volume and temperature-entropy diagrams
shown in Fig. 2.9.
The elements of a Stirling engine consist of two spaces of variable
volume at different temperatures, coupled together through the regenera-
tive heat exchanger and the freezer and cooler heat exchangers. These
elements can be arranged in a wide variety of configurations and can be
broadly classified into single- and double-acting engines. The single-acting
machines are either two-piston or piston-displacer systems. These latter
can be further divided into integral- or split-Stirling engines. Now it is of
interest to consider the motion of reciprocating elements in all these
variations and show the ideal cycle is the same, regardless of the
configuration.

Two-Piston Stirling Engine

Consider first the two-piston Stirling engine shown in Fig. 2.9. This
was used in Chapter 2 to describe the operation in detail of an ideal Stirling
engine. The system consists of a cold expansion space at temperature TE
118 Chapter 3

and a compression space at the ambient temperature T c. The regenerative

heat exchanger is interposed between the two spaces with supplementary
recuperative heat exchangers, the "freezer" on the cold side and the
"cooler" on the ambient temperature side.
The cycle consists of four processes:
(1-2) Isothermal compression in the compression space at ambient
temperature. Heat is rejected to the surroundings of the com-
pression space.
(2-3) Constant volume regenerative cooling where heat is transferred
from the working fluid to the regenerative matrix. The reduction
in temperature at constant volume causes a reduction in the
pressure.
(3-4) Isothermal expansion in the expansion space at the refrigeration
temperature TE. Heat is absorbed from the surroundings of the
expansion space.
(4-1) Constant volume regenerative heating where heat is transferred
from the regenerative matrix to the working fluid. The increase
in temperature at constant volume causes an increase in the
pressure.
The disposition of the two pistons in the expansion and compression
spaces at the four terminal points of the cycle is shown in Fig. 2.9. The
time-displacement diagram for the two pistons is also shown. It is of interest
to note that volume variations in the expansion space lead those in the
compression space, i.e., the expansion space volume is zero in process
(1-2), whereas the compression space volume does not become zero until
later [in process (3-4 )].

Integral Stirling Engine

Now consider the integral Stirling machine shown in Fig. 3.21. This
consists of a displacer and a piston operating in a single cylinder. The space
above the displacer is the expansion space at temperature TE. The space
between the piston and displacer is the compression space at ambient
temperature Tc- The two variable volumes are connected by a duct contain-
ing the regenerator and the two recuperative exchangers, the freezer above
the regenerator and the cooler below the regenerator.
The cycle of operation of the engine is exactly the same as with the
two-piston machine above. The four processes are (1-2) isothermal com-
pression, (2-3) regenerative cooling at constant volume, (3-4) isothermal
expansion, and (4-1) regenerative heating at constant volume. The
disposition of the reciprocating elements at the four terminal points of the
Stirling Cryocoolers 119

Cylinder Head

Oispl a cer

Regenerator

Compression Space (Tel

Piston

Fig. 3.21. Integral Stirling engine. Oisplocer Rod

cycle and the time-displacement diagram for the reciprocating elements

are shown in Fig. 3.22.

Split Stirling Engine

A similar set of disposition and time-displacement diagrams is drawn
in Fig. 3.23 for the split-Stirling arrangement. In this case, the displacer
and piston operate in separate cylinders. The cycle and the four processes
are exactly the same as described above. In this arrangement, the compres-
sion space is divided between the compressor cylinder and the displacer
cylinder. This is frequently overlooked in discussions of split Stirling systems
where the displacer cylinder is often referred to as the "expander" and
the piston cylinder as the "compressor."
For all the above engines, the P-V and T-S planes shown in Fig. 2.9
are applicable with the numbers 1-2-3-4 corresponding in all diagrams.

Double-Acting Stirling Engine

The double-acting Stirling engine is a little more complicated to under-
stand. Consider the four-cylinder double-acting Stirling engine arrange-
ment shown in Fig. 3.24. The four cylinders (a), (b), (c), and (d) each
contain a single reciprocating element, a piston-displacer. The space above
each element is the expansion space at low temperature TE and the space
below is the compresion space at ambient temperature T c - The expansion
space of one cylinder is connected to the compression space of the adjacent
120 Chapter 3

(a )

(I)
Compte'SsiOn Regenero1 1'-'e E~~:panston ReQeneroltve
Cooli ng Heottng

(b )

Fig. 3.22. (a) Piston and displacer disposition, and (b) displacement-time diagrams for the
integral Stirling engine operating on the ideal Stirling cycle.

cylinder through the appropriate heat exchangers so that four separate

Stirling engine systems are formed, A, B, C, and D.
System A combines the compression space of cylinder (a), (Vc)a with
the expansion space of cylinder (d), (VE)d. Similarly system B is made up
of (Vc)b and (VE)a and so on. There are of course four work diagrams, one
for each system and the numbers at the terminal points do not correspond
Stirling Cryocoolers 121

Fig. 3.23 . Piston and displacer disposition and the displacement-time diagrams for the
split-Stirling engine operating on the ideal Stirling cycle.

B c 0 A

~ (ol (b)
(I)
(c) (d)
A

~
EXPANSION
(2)
B
v

~
COMPRESSION (3)
c
v

Fig. 3.24. Disposition of the re-

ciprocating elements of the four-
cylinder double-acting Stirling
engine.
~ (4)
0
v
122 Chapter 3

with those given on Fig. 2.9, except for system A. Therefore, the four P- V
diagrams for the double-acting engine are also included in Fig. 3.24.
At station 1 the reciprocator in cylinder (a) is at the top dead center
position. The reciprocator in cylinder (b) is at the midstroke and descending.
In cylinder (c) the reciprocator is at the bottom dead center position and
in cylinder (d) the reciprocator is at midstroke and ascending. There is,
therefore, a phase difference of 90° between the reciprocator positions in
adjacent cylinders. The phase difference depends on the number of cylin-
ders, i.e., a= 360/n, where a is the phase difference and n is the number
of cylinders. Here n can be 3, 4, 5, or 6 for satisfactory operation.
At station 2 all the pistons have moved by one half stroke but the
phase relationship has been preserved. In cylinder (a) the reciprocator has
descended to midstroke, in cylinder (b) it has descended to the bottom
dead center position, in (c) it has ascended to the midstroke position, and
in (d) it has ascended to the top dead center position.
Consider what has happened in this process (1-2). System A, made
up of ( V c )a and ( V E )d, has decreased in volume to accomplish the isothermal
compression 1-2 shown on the P-V diagram for system A. Simultaneously
system B made up of (Vc)b and (VE)a maintained the total system volume
constant but pushed the working fluid from (Vc)b to (VE)a, that is the
constant volume regenerative cooling process (1-2) shown in the P- V
diagram for system B. While this was going on, system C was experiencing
an isothermal expansion and system D was experiencing a constant volume
regenerative heating process. If the pistons are now allowed to progress
another half stroke to the disposition shown in station 3, the analysis of
events will show the process (2-3), shown on the four P- V diagrams, will
occur. Similarly, we can continue through the processes (3-4) and the final
process (4-1).
To facilitate understanding, the time-displacement diagrams for the
four Stirling systems have been drawn in Fig. 3.25. These diagrams combine
the appropriate expansion and compression spaces constituting each system.
Thus, the upper diagram for system A, combines the motion of the
reciprocator in cylinder (d) (top half of the diagram) with the motion of
the reciprocator in cylinder (a) (bottom half of the diagram). The shaded
area in the top half represents the volume variation in the expansion space
of system A (located in cylinder (d) above the piston/displacer). The shaded
area in the lower half of the figure represents the volume variation of the
compression space of system A [located in cylinder (a) below the piston/dis-
placer]. At any instant, the sum of the two shaded areas represents the
total "live" working space in system A. The constant volume of the
connecting ducts containing the regenerator and associated heat exchangers
must be added to obtain the total working space. The same pattern of
Stirling Cryocoolers 123

Fig. 3.25. Time-displacement dia-

grams for the reciprocating elements
of the four-cylinder double-acting DISCONTINUOUS HARMONIC
Stirling engine. PISTON MOTION PISTON MOTION

presentation is repeated for the displacement-time diagrams for systems

B, C, and D included in Fig. 3.25.
Newcomers to the field will be perplexed by the complicated interplay
of events in double-acting engines. Let them be assured the perplexity is
shared by this author and others. If the subject of their interest is intermedi-
ate or large-scale refrigerating capacity, the closest study of double-acting
systems is warranted. The Koh-i-noor of large-capacity regenerative
cryocoolers has yet to be found and most surely lies in the double-acting
free piston field.

PRACTICAL REGENERATIVE CYCLE

So far it has been assumed the Stirling engine works on the ideal
Stirling cycle of two isothermal and two constant-volume regenerative
processes. Now it is time to consider the difficulties arising if attempts
are made to make cooling engines work on the ideal cycle. Having assessed
that problem, we can define a more realistic model with which to compare
the performance of actual engines.
124 Chapter 3

Mass Distribution in Ideal Cycle

The ideal Stirling cycle assumes much that is beyond the bounds of
practical possibility. At the four terminal points of the cycle, all the working
fluid is assumed to be concentrated in either the compression or expansion
space at temperature Tc or TE. This would require the regenerative matrix
and recuperative freezer and cooler heat exchangers to have zero void
volumes. Since fluid must pass through the heat exchangers, they have
some flow passages and a finite void volume. Therefore, it is never possible
to have all the working fluid instantaneously at the same state (except of
course at the start when the whole machine is at ambient temperature).
Some working fluid will always be in the void spaces of the heat exchangers
at intermediate temperatures.
The void volume will reduce the volume compression ratio ( V max/ V min)
with the consequence of a reduced pressure compression ratio (Pmax!Pmin)
and reduction in the pressure-volume (work) diagram for the cycle.

Harmonic Piston Motion

The ideal cycle demands the reciprocating elements move with discon-
tinuous motion. It is not impossible to achieve the required motion but
fairly elaborate kinematic or electrohydraulic mechanisms are required. It
is easier to produce the simple harmonic motion customarily found in
reciprocating engines with cranks and connecting rods or pneumatic springs.
A reasonable approximation to the ideal piston motion can be attained
with harmonic motion. This is illustrated by the simple harmonic curves
given along with the discontinuous time-displacement curves in Figs. 3.22
and 3.25.

Realistic Work Diagrams

The change in piston motion from discontinuous to continuous har-
monic has profound effects on the operation of Stirling engines. With
harmonic motion, the work diagram assumes the kidney shape figure shown
in Fig. 3.26 rather than the familiar four-cornered figure.
The working fluid is distributed between the various spaces and three
work diagrams may be drawn, one for the expansion space, one for the
compression space, and one for the total working space, including the
"dead space," the total volume of the internal voids in the heat exchangers
and the duct connecting the two "live" spaces. Figure 3.27 shows a typical
set of work diagrams. The diagram for the expansion space is traversed in
a clockwise direction, whereas the diagrams for the compression and
Stirling Cryocoolers 125

..,a:
=>
..,a:
(f)
(f)

a.

Fig. 3.26. Work diagram for a Stirling engine with harmonic

piston motion.

expansion spaces are traversed in the reverse, anticlockwise direction. This

is because work is produced by gas expanding in the expansion space,
whereas work must be done on the gas in the compression space.
The area of the work diagrams is the work done by or on the gas.
The difference in area of the pressure-volume diagrams for the compression
and expansion spaces is the net work of the cycle. In the case of a
refrigerator, the compression space diagram is larger than for the expansion
space and work must be supplied to the cycle. The amount required is
equal to the area of the diagram for the total working space or the difference
in area of the compressor and expansion space diagrams. With a prime
mover, the expansion space diagram is larger than the compression space
diagram. As before, the difference in the area of the diagrams is equal to
the area for the total working space, but in this case, the total working
space diagram is traversed in the clockwise direction and represents the
net cycle output of the engine.
When compression and expansion occur isothermally, the areas of the
work diagrams also represent the quantities of heat transfer to the expansion
cylinder (and freezer) and from the compression space (and cooler).

I
~ \""'
...
~
::l
...
II:
0.4 '---t--l
I
\
"'-...
r-
~
)
~
~
"" ~~
a. ozt---+---1
EXPANSION
I
IOMPRESSION TOTAL WORKING T • 2 , K = I 86
SPACE SPACE SPACE a;. 179. X•l

02 04 0 02 0.4 0.6 02 04 06 08 10
(V./VW) (Vc/VW) (V/Vwl
VOLUME

Fig. 3.27. Typical work diagrams for the expansion, compression, and total working spaces
of a Stirling cryocooler.
126 Chapter 3

The validity of the equality Q = f p dv with isothermal conditions and

ideal gas is widely accepted for a closed system as a straightforward
application of the first law of thermodynamics enunciated for a closed
system: dQ = du +dw. It is less familiar for the nonclosed, variable mass,
unsteady flow situations such as exist here. However, by means of thermo-
dynamic argument too involved for inclusion here, it can be shown that
the equality is obtainable just as in the simpler, closed system provided
the conditions of ideal gas behavior and constant temperature conditions
are met.

Temperature-Entropy Diagrams
The introduction of continuous harmonic piston motion and finite void
volume in the heat exchangers, eliminates conditions inherent in the ideal
cycle that all the working fluid is instantaneously at the same state and
proceeds as a lumped mass around the cycle. With continuous harmonic
piston motion, the fluid is distributed about the machine in various spaces
at various temperatures.
Any particular element of mass inhabits a region of the engine and
when the engine is operating, the element will not move entirely from the
compression to expansion space, but will simply commute to and from,
say, the expansion space to the freezer. Another particle may inhabit the
regenerator and cooler and so on. It is not possible to draw meaningful
temperature-entropy diagrams for all the working fluid. Rather, individual
diagrams for particles at different regions of the engine can be constructed.
The individual envelopes drawn on the T-S plane can then be wholly
enclosed in boundaries that probably represent some kind of temperature-
entropy diagram.
Such representations have only limited use and are rarely drawn in
practice. Areas of these composite diagrams do not represent simple heat
transfer as in the idealized cycle.

Schmidt Cycle
The case discussed above with harmonic piston motion and isothermal
compression and expansion was analyzed by Gustav Schmidt (1871). It is
convenient to identify this case as the Schmidt cycle. The analysis is
presented in detail later in this chapter.

Nonisothermal Compression and Expansion

In practice, the condition of isothermality can never be achieved. It
would require either infinite rates of heat transfer in the compression and
Stirling Cryocoolers 127

expansion spaces or the engine running at very low speeds. In real engines
running at realistic speed (15 to 50 Hz), conditions in the cylinders are
closer to adiabatic (no heat transfer) than isothermal (infinite heat transfer).
They may be close to isothermal in the freezer and cooler since these
components are designed to maximize heat transfer.
The departure from isothermal conditions in the compression and
expansion spaces causes a marked redistribution of working fluid in the
machine resulting in a decrease of refrigerating capacity and an increase
in the input work, thereby reducing the coefficient of performance. The
reduction may be as much as one half the ideal Carnot value.
Analysis of the Stirling engine with adiabatic processes is somewhat
more complicated than the isothermal Schmidt cycle. The theory has been
developed, and will be discussed later. For reference purposes, the adiabatic
case is called the Finkelstein adiabatic cycle.

Aerodynamic Friction Loss

Another important effect in real engines is the loss of refrigerating
capacity due to aerodynamic pressure losses. The effect is illustrated in
Fig. 3.28. The upper diagram shows the pressure-time diagrams for the
compression and expansion spaces. Differences in these two curves are the
pressure losses arising from flow through the regenerator, freezer, and
cooler heat exchangers. The effect of pressure drop on refrigerating capacity
can be gauged by reference to the work diagrams for the expansion and
compression spaces. The shaded areas are the reduction in area of the

VOLUME V8 VOLUME Vc
Fig. 3.28. Regenerator pressure drop and the effect on EXPANSION COMPRESSION
expansion cylinder work. SPACE SPACE
128 Chapter 3

expansion space diagram due to regenerator pressure drop. The refrigerat-

ing capacity is decreased approximately by the same proportion as the area
of the work diagram is decreased.
The regenerator pressure drop has a direct and immediate effect on
the refrigerating capacity and should be kept to the minimum. The pressure
drop is a function of the square of the velocity of the fluid. For minimum
pressure drop, the velocity should be low. This can be achieved with a
large, open, regenerative matrix and large fluid ducts in the freezer and
cooler passages. Unfortunately, this increases the dead space relative to
the "live" swept volumes and has deleterious effects on refrigerating capac-
ity. The dead space must be as small as possible. This dichotomy between
the smallest possible pressure drop and the lowest possible dead space
provides a challenge to ingenious mechanical designers.

Heat Exchanger Thermal Potential

The function of the cooler and the freezer heat exchangers is to
maximize heat transfer to or from the working fluid. In both cases, a
temperature difference between the working fluid and the heat exchanger
walls is necessary to establish the energy transfer.
In the cooler, energy is transferred by convection from the working
fluid to the walls of the cooler, hence by conduction through the walls of
the cooler and is carried off by convective heat transfer from the walls of
the cooler to water or air at ambient temperature. Water is a much more
effective coolant (by an order of magnitude) than air. If water is used in a
closed cooling circuit, a circulating pump and a separate air-cooled convec-
tive recuperative heat exchanger (in automotive terminology, the
"radiator") will be necessary and this addition will make the system large,
heavy, and expensive.
Direct air cooling is attractive, but to improve effectiveness, a large
finned area is required to increase the area for heat transfer. A fan may
be necessary to blow cooling air through the fin complex.
The temperature difference between the average temperature of the
working fluid and the coolant may be assumed to be somewhere in the
range of 20 to 50 °C. To assure heat flow in the right direction, the gas
must be hotter than the coolant.
Similar considerations apply at the low-temperature region of the
machine. When a machine is operating in steady state, the average tem-
perature of the working fluid in the cold region will be that temperature
at which the refrigeration available exactly balances the thermal load. The
thermal load is made up of several constituents:
Stirling Cryocoolers 129

i. The external refrigeration effect in cooling or condensing a steady

flow of gas or liquid circulating through the freezer at constant
temperature and pressure,
ii. Various heat "leaks" by conduction, convection, and radiation from
the surroundings at higher temperature.
Careful attention to design, the use of structural materials having a low
thermal conductivity and high performance thermal insulating materials
can reduce the heat leaks to a surprising degree. Nevertheless, whatever
measures are taken, some residual heat leak must remain. Indeed, in many
cryogenic coolers, particularly the miniature and microminiature engines
used for electronic and infrared sensing applications, the external refrigerat-
ing component is minimal, a nominal fraction of a watt. The sole objective
of the machine is to achieve and maintain a cryogenic temperature at the
cold finger. The entire refrigerating effect is dissipated by external heat
leaks to the cold region. The electronic device or infrared sensor to be
cooled may be mounted directly on the cold finger and connected by fine
metallic conductors to the electronics associated with it. In most cases, the
presence of the sensor has little effect on the temperature achieved in the
cold region of the engine.
The temperature difference between the average temperature of the
gas in the cold region and the surroundings of the cold region, might be 2
to 20°C. For energy flow in the right direction, the gas must be cooler than
the cylinder surrounds.
To assess the effects of these temperature potentials, consider a cryo-
cooler producing refrigeration at 80 K and supplied with coolant at ambient
temperature, 290 K. The ideal coefficient of performance of a refrigerator
operating between these temperatures is
TE 80
COP <carnotl = T. T 280-90
0.38
c- E

Now include the heat exchanger thermal potentials of, say, 50 K in the
cooler and 20 K in the freezer. The thermodynamic temperature range of
operation is, therefore, increased to (290 +50) = 340 K and (80- 20) =
60 K. The coefficient of performance under these conditions is, therefore,
60
340-60 = 0 ' 214

The effect of the temperature potentials is evidently profound. In the case

chosen, the COP, with the admittedly exaggerated temperature potentials,
decreased to a value only 56% of the COP calculated ignoring the heat
exchanger thermal potentials.
130 Chapter 3

Regenerator Contamination
The importance of the regenerator effectiveness to the ideal Stirling
cycle was demonstrated in Chapter 2 (see Fig. 2.21). Remarkably effective
regenerators (s > 0.95) are routinely used in Stirling cryogenic coolers with
matrices of finely divided materials in the form of wires or balls. This finely
divided material is also an excellent filter and effectively "strains out" any
contaminants contained in the working fluid passing through. The con-
taminants include lubricating oil leaking from the crank case, wear debris
from dry rubbing piston rings or seals, and condensed liquids or solids of
water and carbon dioxide from the working fluid. Blockage of regenerator
flow passages increases the fluid friction in the matrix (i.e., a larger pressure
drop) with all the unfortunate consequences to refrigerating capacity dis-
cussed above.

Regenerator Thermal Saturation

Another important phenomenon associated with the regenerator is
"thermal saturation." In moving through the matrix from one end to the
other, there is a progressive decrease in the temperature of the matrix
although at any point the temperature of the matrix material is assumed
to remain constant with ideal regeneration. This assumption implies the
thermal heat capacity (mass times specific heat) of the matrix is infinitely
large compared with the gas passing through it.
When the heat capacity of gas becomes significant compared with the
matrix, there is increasing departure from ideality and the regenerator
approaches the state of thermal saturation. This can arise in two
ways; the first is by increasing the frequency of operation or the pressure
of the working fluid (and hence the density). To a first approximation, the
refrigerating capacity is a linear function of the machine speed or
the pressure of the working fluid. Increasing either or both increases the
refrigerating capacity but at a progressively diminishing rate until a
maximum is attained. Further increase in speed or density actually decreases
the refrigerating output. The reduced rate of increase arises from the
combined action of increased pressure drop and thermal heat capacity of
the gas (incipient thermal saturation of the matrix).
The second cause of thermal saturation becomes significant at low
temperature levels (below 20 K). A decrease in temperature results in
decreased specific heat of solids and liquids but increased specific heat of
gases. At very low temperatures, helium is the only gas not liquefied and
so must be used as the working fluid for cryocoolers. Helium has a high
specific heat compared with many other gases and this already high value
Stirling Cryocoolers 131

increases sharply at low temperatures. At the same time, the specific heat
of conventional matrix materials (copper and bronze wire, etc.) decreases
so that thermal saturation prevents effective operation of the matrix.
Change of the matrix material to lead produces an improvement. Lead
is not available in the form of very fine wires but can be obtained as fine
powder or "shot." Further improvement can be gained by the use of "rare
earth" materials (europium sulfide) with advantageous discontinuities in
the specific heat temperature characteristic. Eventually, though, thermal
saturation of the matrix prevails as the helium working fluid nears the
liquefaction temperature. The ultimate temperature achievable is rarely
less than 6 K. Zimmerman (see Chapter 1) has achieved 3.5 Kin a special
expansion machine with a "regenerative annulus," a pressure of less than
1 atmosphere of the helium working fluid and precooling with liquid helium
to 10 K.

THEORETICAL ANALYSIS OF STIRLING ENGINES

Introduction
Theoretical analyses of Stirling engines acting as prime movers or
cooling engines have been developed with varying degrees of sophistication.
The most simple analysis is that for the ideal Stirling cycle, the thermo-
dynamic cycle, comprised of two isothermal and two constant-volume
regenerative processes. As we have seen, this involves such gross ideali-
zation as to be suitable only for the most elementary design calculations.
A more realistic analysis devised by Gustav Schmidt (1871) has become
the classical analysis of the cycle and is believed to give a reasonable
approximation of machine performance. The analysis is highly idealized,
and in practice, the indicated performance will likely be no better than
50% of the Schmidt cycle performance and often a good deal less. This is
because the Schmidt cycle assumes compression and expansion to be
isothermal. In practical engines running at 15Hz or more, the processes
are more nearly adiabatic. This apparently minor difference causes a redis-
tribution of the mass working fluid and has a critical effect on engine
performance.
Following introduction of the Schmidt analysis in 1871, nearly a
century was to elapse before Finkelstein (1960) devised the generalized
analysis of engines with processes of compression and expansion other than
isothermal. The Schmidt cycle with isothermal compression and expansion
then became a special case of the generalized Finkelstein analysis. Another
special case is the cycle with adiabatic processes in the engine cylinder (the
132 Chapter 3

Finkelstein adiabatic cycle). The generalized theory permits consideration

of other cases of limited heat transfer in the engine cylinders.
Further development of the adiabatic model allowing for secondary
effects, primarily aerodynamic flow losses, and thermal effects, is possible
to make the adiabatic model the preferred level of analysis for routine
performance predictions. The analysis is sufficiently complex to require the
use of digital computers, but only to a moderate extent, involving little
computer time and reasonable expenditure.
More advanced theoretical models for regenerative engine analysis
were devised by Finkelstein (1975) culminating in a nodal analysis program
installed on a commercial computer network and available for general use
on payment of a royalty fee. In the closing years of the 1970s substantial
efforts were devoted to Stirling engine simulation by other workers and a
variety of advanced engine simulation programs have become available.
A comprehensive discussion of the generalized Finkelstein analysis
and subsequent nodal analysis programs is beyond the scope of this work.
Here, with particular regard to cooling engines, we shall
i. summarize equations for the ideal Stirling cycle,
n. present in reasonably comprehensive detail, the Schmidt isothermal
analysis,
iii. outline the Finkelstein adiabatic analysis,
iv. compare and comment briefly on the advanced level analyses for
Stirling engines making reference to the source documents for
those who wish to investigate further.

Ideal Stirling Cycle

Equations for analysis of the ideal Stirling cycle are summarized below.
The reader is advised that the gross idealization of the Stirling cycle
precludes the use of these equations for anything other than the most
elementary investigations.
The principal deficiency of the ideal cycle is that all the working fluid
is assumed to be, instantaneously, at the same condition in either the
compression or expansion spaces. This implies the internal void volume of
the regenerator is zero and requires the pistons or other reciprocating
elements to move with a discontinuous motion. The processes of expansion
and compression are assumed to be isothermal and the effects of imperfect
regeneration and aerodynamic pressure drops are not considered.
A more complete treatment of the ideal cycle, based on the Rallis
regenerative reference cycle, is given in Chapter 2. With reference to Fig.
2.6 and the discussion of the ideal cycle given in Chapter 2, the equations
for the ideal cycle may be summarized as follows:
Stirling Cryocoolers 133

Required Data.
1. Some reference temperature and pressure, or volume, say, condi-
tions, at state 1.
ii. Temperature ratio T = Tc/TE.
iii. Volume ratio r = V max/ V min·
For unit mass of working fluid, assumed to be a perfect gas, then V1 =
RTd Pt, from the characteristic gas equation.
The following data are required for each of the four cycle processes:
(a) Isothermal Compression Process (1-2). In this process, heat is
abstracted from the working fluid and rejected from the cycle at the
maximum cycle temperature Tc. Work is done on the working fluid equal
in magnitude to the heat rejected from the cycle. There is no change in
internal energy, and there is a decrease in entropy.

Heat transfer (Q) =work done (W) =P 1 V 1 ln (1/r) =RT1 ln (1/r). Change
in entropy (S 2 - S1 ) = R In (1/r ).
(b) Constant- Volume Regenerative Cooling Transfer Process (2-3).
In this process heat is transferred to the regenerative matrix from the
working fluid, decreasing the temperature from Tc toTE. No work is done,
and there is a decrease in the entropy, and internal energy, of the working
fluid.

Heat transfer (Q) = Cv(T3 - T 2 ). Work done (W) = 0. Change in entropy

(S3 -Sz) = Cv In (1/-r).
(c) Isothermal Expansion Process (3-4) In this process, heat is sup-
plied to the cycle at a low temperature TE, during expansion of the working
fluid. Work is done, by the working fluid, equal in magnitude to the heat
supplied. There is no change in the internal energy, but an increase in the
entropy of the working fluid:

Heat transfer (Q) =work done (W) = p 3 V3 ln r = RT3 ln r. Change in

entropy (S4- S 3 ) = R In r.
(d) Constant- Volume Regenerative Heating Transfer Process (4-1).
In this process, heat is transferred to the working fluid from the regenerative
matrix increasing the temperature of the working fluid from TE to Tc. No
work is done, and there is an increase in the internal energy and entropy
134 Chapter 3

of the working fluid.

Heat transfer (Q) = Cv(T1 - T 4 ). Change in entropy (S 1 -S4 ) = Cv In 'T.

In the regenerative processes, the heat transferred to the matrix from
the working fluid in process (2-3) is restored to the working fluid from the
matrix in process (4-1). There is no net gain or loss of heat by the working
fluid or the matrix. Therefore
the total heat extracted (at TE) = RT3 In r
the total heat rejected (at T c)= RT1 In r
and
. RT3 In r
coefficient of performance= T 1 RT 1
R 1 nr- 3 nr

= 1/('T -1)
This value corresponds to the Carnot value between the same temperature
limits.

The Schmidt Cycle

The classical analysis of the operation of Stirling engines is due to
Schmidt (1861). The theory provides for harmonic motion of the reciprocat-
ing elements, but retains the major assumptions of isothermal compression
and expansion and of perfect regeneration. It, thus, remains highly ideal-
ized, but is certainly more realistic than the ideal Stirling cycle. Provided
a reasonable level of caution is exercised in interpretation, the predictions
of the Schmidt theory can be a useful tool for engine design.
Principal Assumptions of the Schmidt Cycle.
1. The regenerative process is perfect.
2. The instantaneous pressure is the same throughout the system.
3. The working fluid obeys the characteristic gas equation, p V = RT.
4. There is no leakage, and the mass of working fluid remains
constant.
5. The volume variations in the working space occur sinusoidally.
6. There are no temperature gradients in the heat-exchangers.
7. The cylinder wall, and piston, temperatures are constant.
8. There is perfect mixing of the cylinder contents.
Stirling Cryocoolers 135

9. The temperature of the working fluid in the ancillary spaces is

constant.
10. The speed of the machine is constant.
11. Steady state conditions are established.
Nomenclature.*
A= a factor (r 2 + 2-rK cos a+ K2 ) 112 .
B =a factor (r+K +2S).
K =const.
M =total mass of working fluid.
N = machine speed.
p =instantaneous cycle-pressure.
Pmax =maximum cycle-pressure.
Pmean =mean cycle-pressure.
Pmin =minimum cycle-pressure.
P =engine output.
p mass =pI RTc dimensionless power parameter based on the mass of
working fluid.
p max= pI (Pmax vT ), dimensionless power parameter' based on the
maximum cycle-pressure and combined swept volume.
Q = heat transferred to the working fluid in the expansion space, the
heat extracted.
Omass = QIRTc the dimensionless cooling parameter, based on the mass
of working fluid.
Omax = 0 I (Pmax V r ), the dimensionless heat, based on the maximum cycle
pressure.
R =characteristic gas constant of the working fluid.
S = (2Xr)l(r + 1), reduced dead volume.
Tc =temperature of the working fluid in the compression space, gen-
erally assumed to be 300 K.
Tv =temperature of the working fluid in the dead space.
TE =temperature of the working fluid in the expansion space.
V c =swept volume in the compression space.
VE =swept volume in the expansion space.
Vv =total internal volume of heat-exchangers, volume of regenerator,
and associated ducts and ports.
Vr = (Vc + VE) = (1 +K)VE, combined swept volume.
V w = ~ VE(1 +cos c/J) + ~ V c[1 +cos (c/J- a)]+ V v, volume of total work-
ing space.
Vwmax =maximum volume of total working space.
*Note: Lower case subscripts indicate instantaneous values of temperature, pressure, volume,
and mass. Upper case subscripts indicate maximum (or constant) values. Thus: E ore refers
to expansion space, C or c refers to compression space, D or d refers to dead space.
136 Chapter 3

X= Vv/VE, dead-volume ratio.

a = angle by which volume variations in the expansion space lead
those in the compression space (in fractions of 1T radians, or
degrees).
8 = (r 2 + K2 +2TK COS a) 112 /(r + K +2S).
8 =tan - l [(K sin a)/(r + K cos a)].
K = Vc/VE, swept-volume ratio.
T = Tc/TE, temperature ratio.
c/J =crank angle.
Basic Equations. Volume of expansion space:
Ve =~VE(l+coscjJ). (3.1)
Volume of compression space:
Vc = !Vc[l +cos (c/J -a)] (3.2)

=!KVE[l +cos (c/J -a)] (3.3)

Volume of dead space, being that constant volume of the total working
space not included in the volumes of the expansion or compression space:
Vv=XVE (3.4)

Mass of working fluid in expansion space:

Me= (Pe Ve)/(RTe)
Mass of working fluid in compression space:

Mass of working fluid in dead space:

Since the total mass of the working fluid remains constant

If the instantaneous pressure is the same throughout the system, and equal
top, say, and if Te and Tc are constant at TE and Tc then, substituting for
the volumes, eliminating R, and rearranging

K/p = (Tc/TE)(l +cos c/J) +K[l +cos (c/J -a)]+ (2 V 0 Tc)/(VETv) (3.6)

If the temperature variation in the dead space is linear in the axial direction,
then the mean temperature
Stirling Cryocoolers 137

and, since Tc/TE = T then, from Eq. (3.6)

K/p = T(1 +cos <P)+K[1 +cos (<P -a)]+2S (3.7)

where S (the reduced dead volume)= 2XT/(T + 1).

To simplify Eq. (3.7) consider first
y = x cos <P + z sin <P (3.8)

then y = .Jr 2 cos (<P -(3), where tan (3 = z/x, z = r sin (3, and x = r cos (3
since
.Jr 2 cos (<P - (3) = .Jr 2 (cos <P cos (3 +sin <P sin (3)
= r cos <P cos (3 + r sin <P sin (3
= x cos <P + z sin <P

Equation (3.8) is similar in form to Eq. (3.7) therefore, by analogy

K/p = [(T +K cos a) 2 +(K sin a) 2 ] 112 cos (<P- 0) +T +K +2S

= (T 2 +2TK COS a+ K2 ) 112 COS (<{J- 0) +T + K + 2S (3.9)

where
tanO=(K sina)/(T+K cos a)
Let
A= (T 2 +2TK cos a +K 2 ) 112 , B =T+K+2S, and 8 =A/B
then
K/p =A cos (<P -O)+B
and
p =K/{B[1+8 cos (<(J -0)]}
The instantaneous pressure p is
(a) a minimum, when <P = 0, i.e., (<P- 0) = 0,
(b) a maximum, when <P = (0 +1r), i.e. (<P -0) = 1r.
Therefore, Pmin =K/[B(1 +8)], and Pmax =K/[B(1-8)].
Thus
P = Pmax(1-8)/[1 +8 COS (<{J- O)] (3.10a)

= Pmin(1 + 8)/[1 + 8 COS (<{J- O)] (3.10b)

and the pressure ratio p, = Pmax/Pmin = (1 +8)/(1-8). (3.11)
138 Chapter 3

Mean Cycle Pressure. The mean cycle pressure is given by

r21T
Pmean=(hr) Jo pd(<{J-(})

r21T
=(h) Jo {Pmax(1- 8)/[1 + 8 COS (</J- O)]} d(<{J- (}) (3.12)

which can be resolved to

Pmean = Pmax[(1-8)/(1 +8)] 112 (3.13)

Heat Transferred and Work Done. Since the processes of expansion

and compression take place isothermally the heat transferred Q is equal
to the work done P, therefore

Q =P= JpdV

dV = -!VE sin </J d<{J (3.14)

and, if
P = Pmean[1- ~COS (</J- O)] (3.15)
approximately, where

then
r21T
Q = -! Jo {Pmean V£[1- ~cos (</J- 0)] sin </J} d<{J

r21T
= -!Pmean VE Jo [sin <P- ~(cos <P cos(} sin <P +sin(} sin 2 <P)] d<{J

<P] 21T
=- !Pmean VE [ -~sin(} 2 0

= -!7TPmean VE~ sin(} (3.16)

Expansion Space. The variation in volume in the expansion space
follows the equation
Stirling Cryocoolers 139

which conforms to the required equation [Eq. (3.14)], therefore the heat
transferred in the expansion space, from Eq. (3.16), is given by
Q = 7rPmean VE8 sin 8/[1 + (1- 8 2)112 ] (3.17)
There is no need to invoke the approximate expression for p in Eq.
(3.15). Instead, Eq. (3.10b) is a suitable starting point, and can be combined
with (3.14) to calculate the heat transfer per cycle, Q, as follows:*

!
Q = ,( dV = ,( -pmin( 1 + 8 ) . VB sin <P d<P
jP T1+8cos(<{>-8) 2
= -VEPmin(1+8)I
2
where
I _ ,( sin <P d<P
j 1 +8 cos(</> -8)
let <P- 8 = t/1, then t/f/2 = t: then d<f> x dt/1 = 2dt/(1 + t 2) and

8!= f{ 8 sin tjJ cos 8 8 cos 8 + sin 8

1+8costjJ
+
1+8costjJ
sin 8 } d
1+8costjJ
t/1

8I = [-cos 8 log (1 + 8 cos t/1 )]~':o +sin 8[t/l ]~': 0 - sin 8 ,( 2 2 d: 2)

j 1+t +8 1-r

2 . 2 . ,( dt
= 7rSm8- sm8j (1 + 8 )+( 1 _ 8 )t2

. 2sin8,((1-8) 112 ,( [(1+8)/(1-8)] 112 dt

=27rsm8- 1-8 j 1-8 j {[(1+8)/(1-8)]1/2}2+t2

. 2 sin 8 I 1 1 2 12
=27rsm8-( 1 _ 82 )172 tan- {t/[(1+8)/(1-8)] 1 } o/J~o
11"

In evaluating the definite integral note that t = 0 at both limits;

tan - 1 (0) = 0, 7r etc. It is clear from (a), if 8 ~ 0, that the cyclic integral
must take the value -271" sin 8 for its third term. Thus
8I = 271" sin 8 -271" sin 8/(1-8 2)112 = 271" sin 8[1-1/(1-8 2/ 12 ]
and

*I am indebted to Dr. A. J. Oxley of the United Kingdom Rutherford Laboratory for the
alternative derivation given here-G. W.
140 Chapter 3

Substituting for Pmin gives

0 = VEPmean87T sin 8/[(1-8 2 / 12 + 1]
which is identical to Eq. (3.17).
Compression Space. The variation in volume of the compression space
follows the equation
(3.18)
and, by a process similar to that above, we can obtain expressions for the
pressure and volume in the required form, so that heat transferred in the
compression space is given by
Oc = [7TPmean VEK8 sin (8 -a)]/[1 + (1-8 2 ) 112 ] (3.19)
Dividing Eq. (3.19) by Eq. (3.17),
Oc/0 = [K sin (8 -a)]/sin 8
= K(sin 8 cos a -cos 8 sin a)/sin 8
= K[(cos a -sin a)/tan 8]
but tan 8 = K sin a I (T + K cos a) and, therefore, 0 cl Q = -T.
The heat transferred in the expansion space is of opposite sign to the
heat transferred in the compression space, and is numerically different by
the temperature ratio T. By analogy, the work done in the two spaces has
the same relationship, P c = -TPE, and the net power is P = PE + P c =
(1-T)Q.
In the case of the machine acting as a prime mover TE > Tc, i.e., T < 1,
and the thermal efficiency
77 =(heat supplied- heat rejected)/(heat supplied)
= (0 -T0)/0 = 1-T = (TE- Tc)/TE
This corresponds to the Carnot efficiency.
When the machine acts as a refrigerator, Tc > TE, i.e., T > 1, and
coefficient of performance= heat extracted/work done
= 0/(0- Oc) = 1/(1-T)
= TE/(TE- Tc)
For a heat pump, Tc > TE, i.e., T > 1, and
coefficient of performance= heat rejected/work done
= Oc/(0- Oc) = T/(1-T) = Tc/(TE- Tc)
This corresponds to the inverse thermal efficiency.
Stirling Cryocoolers 141

Mass Distribution in the Machine. From the characteristic gas

equation,
M=pV/RT
where

(a) Expansion Space v. = !VE(1 +cos <P ). The instantaneous mass of

working fluid in the expansion space is given by
M = !VEPmean(1-8 2 ) 112 (1 +cos <P)
(3.20)
e [RTE(1 + 8 cos (<fo- 8)]
The rate of change of mass of working fluid in the expansion space is
dMe VEPmean(1- 8 2 ) 112{8[sin (<P- 8) -sin 8] -sin <P}
d<fo 2RTE[1 +8 cos (<fo -8)f
(b) Compression Space Vc = !KVd1 +cos (<P -a)]. The instantaneous
mass of working fluid in the compression space is given by
M = !{K VEPmean(1- 8 2 / 12 [1 +cos (<P- a)]}
(3.21)
c 2RTc [1 + 8 cos (<fo- 8)]
The rate of change of mass of working fluid in the compression space is
dMc K VEPmean(1- 8 2 ) 112{8[sin (<P- 8) +sin (a - 8)- sin (8- a)]}
d<P = 2RTc[1 + 8 cos (<P- 8)] 2
(c) Dead Space (Vv =XVE, constant). The instantaneous mass of the
working fluid in the dead space is given by
Md = [XVEPmean(1-8 2 ) 112]
(3.22)
RTv[1 +8 cos (<P- 8)]
The rate of change of mass of working fluid in the dead space is
dMd [XVEPmean(1-8 2 ) 112 8 sin (<fo -8)]
d<P RTv[1 +8 cos (<P -8)] 2
Now dMc +dNc +dMd = 0, so that the total mass of working fluid Mr
is constant. Now
Mr = VEPmean(1-8 2 ) 112 {r(1 +COS <P) + K[1 +COS (<P -a)]+ 2S}
2RTc[1 + 8 cos (<P- 8)]
and when <P =0
Mr = VEPmean(1- 8 2 ) 112 [r +S + (K/2)(1 +COS a)] (3.23)
RTc(1 +8 cos 8)
142 Chapter 3

Heat Extracted and Engine Output in Dimensionless Units. (a) The

heat lifted per unit mass of working fluid, combining Eqs. (4.17) and (4.23)
is given by
Q m5 sin 8(1 +cos 8)
Omass = RTc = {(1- 8 2 ) 112 [1 + (1- 8 2 ) 112] [ 'T + (K/2)(1 +COS a)+ S]}
(3.24)
Similarly, the net engine-output per unit mass of working fluid is given by
P mass= P/ RTc = ('T -1)Qmass (3.25)

(b) Nondimensional expressions, in terms of characteristic pressures

and volumes, may be devised as follows. The combined swept volume is
given by
Vr = (VE + Vc) = (1 +K)VE

Combining this with Eqs. (3.1) and (3.13), then

Omax = Qj(Pmax Vr)
= [ 1r(l- 8) 112 8 sin 8]/{(1 + K )(1 + 8) 112 [1 + (1- 8 2 ) 112 ]} (3.26)
and
Pmax = (T -1)Qmax (3.27)

The Finkelstein Adiabatic Cycle

Finkelstein (1960) devised a generalized thermodynamic analysis of
Stirling engines in which the processes of compression and expansion were
not confined to isothermal conditions. In this generalized analysis the
processes of compression and expansion in the engine cylinders could be
specified to occur anywhere between the two limiting cases: (a) isothermal
(infinite rates of heat transfer between the working gas and the cylinder
walls); (b) adiabatic (zero rates of heat transfer between the working gas
and the cylinder wall).
The model assumed that, in the freezer and cooler, infinite heat transfer
isothermal conditions prevailed so that fluid in the heat exchangers was
always at the lower temperature TE or the upper temperature Tc. The
temperature of the working fluid in the cylinders varied during the cycle
and could be greater or less than TE (in the expansion space) or Tc (in the
compression space).
Finkelstein's generalized analysis retained all the other significant
assumptions of the Schmidt cycle so that it remained highly idealized.
Stirling Cryocoolers 143

Nevertheless, the possibility for nonisothermal processes represented the

most significant theoretical development in nearly a century.
The limiting isothermal case of Finkelstein's theory corresponded
exactly to the Schmidt cycle and the equations for this case reduced to a
form identical with the Schmidt cycle equations given above. For the
adiabatic limiting case, closed form solutions were not obtained nor for
other intermediate cases, but the theory was readily amenable to numerical
analysis by standard methods. Significant simplifications were introduced
by the assumption of adiabatic conditions.
In his presentation, Finkelstein included only one set of numerical
results. This referred to a cooling engine with the temperature T = Tc/TE =
2. The coefficient of performance of 1.0 with isothermal processes was
reduced to 0.543 with adiabatic processes. Similarly, Stoddart (1960) found
that a Stirling engine prime mover having a Schmidt cycle (Carnot) efficiency
of 50% with isothermal compression and expansion processes had an
efficiency of only 34% with adiabatic processes.
Later, Khan (1962) investigated the effect of variation in the principal
design parameters and obtained numerical results for a large number of
different cases with adiabatic compression and expansion processes. Some
of these results were summarized by Walker and Khan (1965).
In the adiabatic cycle, the thermal efficiency and coefficient of perfor-
mance become functions, not only of temperature (as in the isothermal
cycle) but also the swept volume ratio, K, the phase angle, a, and the dead
volume ratio, X. The power and heat extracted are functions of all these
parameters both in the isothermal and adiabatic cases. The results presented
by Walker and Khan provided some indications of the effects on efficiency
of design parameters other than temperature. One unexpected result was
that increase in dead space increased the thermal efficiency, even though
the actual power output declined.
More recently, Lee (1976), has reconsidered the Finkelstein adiabatic
cycle for both cooling engines and prime movers. This included the prepar-
ation of a fast, well-behaved computer program in FORTRAN IV language.
The program, containing extensive documentation to facilitate understand-
ing is reproduced in entirety as an appendix to Lee's thesis.*
Figure 3.29, reproduced from Lee (1976), shows the cyclic temperature
variation of the working fluid in an adiabatic cycle Stirling cooling engine,
having temperatures of 100 Kin the freezer (TE), and 300 Kin the cooler
(Tc ), so that T = 3. The phase angle a = 105°, swept volume ratio K =
Vel VE = 1.0, and the dead volume ratio X= Vv/ VE = 1.0. It is interesting
* Copies of the thesis by Lee may be obtained through the usual channels or by direct
application to Dr. G. Walker, Department of Mechanical Engineering, University of Calgary,
Alberta, Canada. Payment of a reproduction and handling charge will be required.
144 Chapter 3

CRANK ANGLE

Fig. 3.29. Cyclic temperature variation in the compression and expansion space of a Stirling
cryocooler operating on the Finkelstein adiabatic cycle (after Lee, 1976).

to note the temperature of fluid in the expansion space was less than the
freezer temperature for much of the time. The mean temperature in the
expansion space was 90 K, a difference of 10 K below the freezer tem-
perature, TE. In the compression space, the mean temperature of the fluid
coincided more or less with the cooler temperature of 300 K.
Significant work using the adiabatic cycle approach was reported by
Qvale and Smith (1969) and by Rios and Smith (1969). They consider a
basic cycle with adiabatic compression and expansion and then separately
assess the effects of irreversibilities. This approach allows for independent
study of individual heat exchange components in a series of successive
approximations that can be extended to the required degree of complexity.
Little has been published about theoretical work on Stirling engines
at the Philips Company or their licensees. Mention is made in the various
papers about the extensive use of computers and the ability to predict
engine performance to one or two percent of test bed performance. Odd
straws in the wind gathered over the years have led to the understanding
that a family of computer programs have been developed by Philips. The
ensemble of programs are thought to include some overall cycle simulation
programs for the complete engine at various levels of sophistication and
complexity, whereas others are available for the detail simulation and
design of single components or subassemblies in the engine. The basis of
the Philips thermodynamic analysis simulation program is thought to be
similar to that developed by Smith, basically the Finkelstein adiabatic cycle
with successive approximations and corrections. The adherence by the
Philips Company over many years to similar design configurations has
Stirling Cryocoolers 145

permitted the accumulation of extensive practical experience. On the basis

of the experience, "fudge" factors of appropriate magnitude can be judi-
ciously applied to the analytical results to provide close and realistic
simulation.

Nodal Analysis
Nodal analysis of Stirling engines was pioneered by Finkelstein (1975)
but other nodal analysis programs were prepared independently. Develop-
ment was principally directed to prime movers, although the programs are
equally applicable to cooling engines. In nodal analysis programs, the
attempt was made to model the simultaneous energy and fluid flows
occurring in the engine and thus simulate exactly the engine cycle and
performance. This was achieved by writing and solving equations for the
conservation of mass, momentum, and energy for particular nodes, cells,
or elements of the engine. The equations were too complex for general
analytical solution and were solved numerically in terms of small incre-
mental time steps. The equations were invariably reduced to one-
dimensional form with additional simplifications. A comparative discussion
of the various programs has been given by Urieli (1979).
All the nodal analysis programs are basically similar in their general
approach. The design of engine to be simulated must be known in exact
detail to the extent that the mechanical arrangement, cylinder wall thickness
and material, heat exchanger tube diameter, fin dimensions or matrix pore
sizes, are all specified. This design is then broken down, as experience
dictates, into a number of nodes, elements, or control volumes. Some
operating conditions must be specified such as charge, pressure, and tem-
perature of the energy source and sink.
Differential equations for the conservation of mass, momentum, and
energy must be developed and generally converted to difference equations.
Empirical formulas for the aerodynamic friction and heat transfer effects
must be included as well as an equation of state for the working fluid. A
mathematically stable method must then be found for numerical solution
of the difference equations to resolve the pressure, temperature, and mass
distribution in the engine at the end of a particular time step, given the
conditions at the beginning of the step.
The usual procedure for solution is to start with some initial arbitrary
assumed conditions and then proceed through several engine cycles until
quasisteady state is achieved, when the instantaneous cyclic values of
pressure, temperature, and mass distribution are not significantly different
from the preceding cycle. The cyclic pressures and volumes are integrated
to calculate work transfer, heat flows are estimated, and the overall thermal
146 Chapter 3

efficiency or coefficient of performance is computed. In most cases, an

isothermal Schmidt cycle-type calculation is included at the beginning or
end of the nodal analysis program.
The results are presented in a variety of formats. Typically the input
data in terms of engine geometry and operating parameters are restated
for comparison purposes. This is followed by tabulations at selected cyclic
crank intervals of pressures, temperatures, velocities, mass content, and
mass velocities at representation stations in the engine. Finally, consider-
ations of the overall cycle are presented, usually in terms of energy flows,
work terms, heat transferred, or ratios of these such as the thermal efficiency
and coefficient of performance.
Most programs now incorporate automatic plot routines so that, if
required, the data can be presented pictorially. Remarkable three-
dimensional pictures can be produced in this way. Figures 3.30 and 3.31
were given by Schock (1978) and are typical of the interesting and informa-
tive representation achieved by computer plot routines. These particular
diagrams show the variation in temperature with node position or volume
as a function of the cyclic angle for a prime mover.

w
a:
,_
:>

"a:~
::;
7: ~
c --

Fig. 3.30. Three-dimensional representation of the gas temperature versus node position and
cyclic angle in a Stirling engine prime mover (after Schock, 1978).
Stirlin2 Cryocoolers 147

Fig. 3.31. Three-dimensional representation of the gas temperatures versus node volumes
and cyclic angle in a Stirling engine prime mover (after Schock, 1978).

Finkelstein Nodal Analysis. After development of the adiabatic

analysis, Finkelstein worked on increasingly sophisticated computer simula-
tion of Stirling engines. This culminated in the late 1960s in adaptation of
the NASA Thermal Analysis Program (TAP) to Stirling cycle simulation
and the creation of the first nodal analysis program.
Its first practical application was at the University of Calgary in
1968/69 in support of the development of miniature cryocoolers for the
British Ministry of Technology. Figure 3.32 shows a simplified cross section
through the cooling engine. Figure 3.33 shows the two-dimensional rep-
resentation of the engine used for the nodal analysis simulation. This work
was briefly discussed by Finkelstein et at. (1970).
Following further refinement and development, Finkelstein (1975)
presented the theoretical basis for the analysis and described the numeric
differencing technique for solution of the partial differential equations
obtained. The Finkelstein nodal analysis program is installed on a commer-
cial computer network and available for general use on payment of a royalty
fee. Potential users may obtain further details from Dr. T. Finkelstein.*
Urieli Nodal Analysis. Israel Urieli, working under the supervision
of Professor Rallis at the University of Witwatersrand, provided the most

* TCA Inc., P.O. Box 643, Beverly Hills, California.

148 Chapter 3

Fig. 3.32. Simplified cross section of Stirling cryocooler

(after Finkelstein eta/., 1971).

complete discussion (Urieli, 1977) of Stirling engine simulation by nodal

analysis. His thesis is recommended for all those professionally interested
in the field of Stirling engine computer simulation. A good outline of the
approach, more readily available than copies of the thesis, was provided
by Urieli et a/. (1977) and Urieli (1979). The thesis contains a complete
program listing in FORTRAN of the Urieli program. The program is likely
to be applicable to engines other than the model assumed for investigation
by Urieli, but a very substantial effort would be required to install a fully
operational version of the Urieli program.
Work on the nodal analysis field continues at the University of Wit-
watersrand under the direction of Professor Rallis with experimental work
to validate the computer model and investigation of alternative theoretical
models (Berchowitz et a/., 1977). Further information may be obtained
from I. Urieli* and Professor C. Rallis. t
Sunpower Nodal Analysis. Gedeon (1978) outlined the techniques of
Stirling engine simulation in use at Sunpower of Athens, Ohio, manufac-
turers of Beale free-piston Stirling engines. Numerical simulation is an
* Ormat Turbines, P.O . Box 68, Yavne, Israel.
t Department of Mechanical Engineering, University of Witwatersrand, South Africa.
 Stirling Cryocoolers 149

T=w 1
C = 1/K 1 p,

·~""7r-;::;:-
;; 7~ p,/v, t~~~ ,,..~.....,j/1
...--
~0
T=w 1

10

Crankcase

Fig. 3.33. Two-dimensional representation of Stirling cryocooler used for nodal analysis (after
Finkelstein eta/., 1971).
150 Chapter 3

integral part of the development approach at Sunpower. Engine refinement

and program validation proceeds hand in hand with routine daily exchanges
between the staff engaged in laboratory and theoretical work.
The Sunpower program is fast and fully validated. It is a complex of
programs capable of cycle simulation at different levels of sophistication
as selected by the program controller. The program is self-optimizing to
the degree specified in the program input. In the field of Stirling engines,
Sunpower has accumulated the greatest body of practical engine develop-
ment and operating experience, supported by computer simulation, outside
of Philips and their licensees. It is, therefore, gratifying that the Sunpower
simulation programs are available for general use on a contract basis. Active
moves are afoot (Beale, 1979) to install the program on a commercial
computer network and organize educational courses in the use of the
programs. Further details of Sunpower Stirling engine development may
be gained from William Beale.*
Other Nodal Analyses. Schock (1978) described the Stirling Nodal
Analysis Program (SNAP) prepared by Fairchild Industries under contract
to the U.S. Department of Energy in support of a Beale free-piston Stirling
engine development. The Schock program is closely similar to the Urieli
program. It is well described in a draft report (Schock, 1978a). An abbrevi-
ated description may be found in the paper (Schock, 1978b) contained in
the proceedings of the 1978 Intersociety Energy Conversion Engineering
Conference. Details of the status of the program may be obtained from
Alfred Schock. t
Roy Tew and others at the NASA Lewis Research Center, Cleveland,
Ohio, have developed a Stirling engine nodal analysis program as part of
the DOE/NASA Stirling Engine Automotive Program. The description of
the Lewis program was given by Martini (1978). Experimental studies are
being carried out at the Lewis Research Center (Cairelli et al., 1977). The
final answer to computer simulation and experimental validation will event-
ually be placed in the public domain by the team at NASA Lewis Research
Center. Readers are urged to consult the current literature relating to the
DOE/NASA Stirling Automotive Engine. Further details may be obtained
on application to Robert Ragsdale.t
Finegold and Vanderbrug (1977) outlined a nodal analysis program
for Stirling engine simulation developed at the Jet Propulsion Labroaratory,
Pasadena, California. The program is explained and completely listed in a
42 page appendix to the final report discussing Stirling engines for under-

* President, Sunpower Inc., Athens, Ohio.

t Energy Systems Department, Fairchild Space and Electronics Co., Germantown, Maryland.
~ Stirling Engine Project Office, NASA Lewis Research Center, 21000 Brookpark Road,
Cleveland, Ohio.
Stirling Cryocoolers 151

water applications. Present status and potential applications for the com-
puter program are not known.
Dr. Allen Organ, formerly of Kings College, University of London,
now at the School of Engineering, University of Cambridge, has recently
published papers indicative of a powerful nodal analysis program in forma-
tion, but not yet fully developed. Readers are advised to contact Dr. Organ
directly for up-to-date information.
Several papers have hinted at the existence and use of nodal analysis
type programs at Philips and their licensees. No details have been given.
It is understood that for general design work, adiabatic cycle programs are
preferred over the more sophisticated nodal analysis programs.

Summary
Stirling engine simulation by nodal analysis is an expensive, time-
consuming activity to be reserved for those professionally engaged in
Stirling engine development or academics engaged in the training of
engineers at the graduate level.
At present, no comprehensive evaluation is available for the various
Stirling engine simulation programs described above. Martini (1978)
attempted the preparation of a design manual including a comparison of
the various theoretical procedures, but concentrated his efforts principally
on isothermal type analyses with only passing reference to the more sophisti-
cated nodal analyses. He later extended this work to higher-level simulation
programs (1980).
The program of Urieli and Finegold-Vanderbrug are completely listed
in the references cited above. However, neither of these programs has been
extensively validated by experimental data. The programs by Schock and
by Tew are neither listed nor have been validated experimentally. Neither
program is available commercially. Finkelstein's program is available com-
mercially on payment of moderate royalties, but no documentation of
validity with reference to practical engines is available. The Sunpower
program is available commercially and has been validated extensively
during the development of Beale-type free-piston engines. The applicability
of the program to engines with crank mechanisms (disciplined-piston
engines) has also been established recently.
For those wishing to obtain a simulation or optimization of a Stirlihg
engine concept, the choice is between Sunpower and Finkelstein. Another
possiblity is that Philips could offer consultancy services to evaluate new
concepts using their undoubted expertise and experience in the field.
It is simply out of the question to "dabble" in nodal analysis even
using the existing programs. The few companies and individuals wishing
152 Chapter 3

to operate this high level of simulation must be prepared to invest substantial

time and money to establish the program procedure and to accomplish the
work. Obtaining results is easy. Understanding them and appreciating the
implications is beyond the level of dilettanish efforts.
For those unable to indulge in nodal analysis, close study of the
excellent manual by Martini (1978) and its successor is recommended.
Martini offers a commercial Stirling engine simulation service using pro-
grams at various levels of sophistication. Further information may be
obtained on application to Dr. William Martini, 2303 Harris, Richland,
Washington.
The adiabatic cycle with the appropriate thermo/fluid corrections is
suitable for most design and optimization of Stirling engines. The adiabatic
cycle requires use of digital computers, but not to an unconscionable degree,
and these are so readily available that their use is routine to most engineers.
The adiabatic cycle is a good deal better than the isothermal case and
requires only minor computing cost and time compared with nodal analysis.
Readers are referred to the thesis by Lee (1976), for a program listing,
and to the works of Rios, Qvale, and Smith referenced below. At yet earlier
stages in design, the isothermal analyses are adequate. Martini (1978) has
developed programs suitable for analysis using pocket calculators.
At the start of a project, it is usually sufficient to carry out simple
Schmidt-cycle calculations. To attain the likely performance of a practical
engine, one simply divides the coefficient of performance and heat extracted
by 2 (if an optimist) or 3 (if a realist).

PRACTICAL DESIGN

Design Parameters
The principal independently chosen design parameters for Stirling
cooling engines are as follows:
1. The temperature ratio T = Tc/TE, the ratio of temperatures in the
compression and expansion spaces.
2. The swept volume ratio K = Vc!VE, the ratio of swept volumes
in the compression and expansion spaces.
3. The dead volume ratio X= Vv!VE, total internal volume of heat
exchangers and associated ducts expressed as a fraction of the
swept volume in the expansion space.
4. The phase angle a by which volume variations in the expansion
space lead those in the compression space.
Stirling Cryocoolers 153

5. The pressure of the working fluid expressed as the maximum or

mean pressure, Pmax or Pmean·
6. The speed of the engine N in cycles per unit time.
7. The bore and stroke of the reciprocating member in the expansion
and compression sp.aces.
8. The number of Stirling engine systems in the total machine, i.e.,
single or multiple systems.
9. For multiple systems the unit may be a series of single-acting
Stirling engines operating independently or a double-acting
arrangement with all systems coupled.
10. The method of activating the piston and displacer action, i.e., free
piston or kinematic drive.
11. The design configuration, integral Stirling, split Stirling, two-piston
machines, fixed or moving regenerator, type of kinematic drive,
crank-connecting rod, Scotch yoke, rhombic drive, swash-plate,
wobble-plate, etc.
The Schmidt cycle is useful to explore the effects of variation in the
first six parameters. The remainder are decisions in the bailiwick of the
mechanical designer. Selections are made on the basis of familiarity, past
experience, constraints imposed by the sponsor (including weight, volume,
cost, vibration), cooling system, and material considerations; or progressive
optimization involving sophisticated nodal analysis of alternative outline
designs.

Optimization of Design Parameters

In the Schmidt analysis presented above, the heat extracted in the
expansion space per cycle was found to be [Eq. (3.17)]
QE 'lT(l-8) 1128 sin 8
Pmax Vr =(1+K)(1+8) 112 [1+(1-8 2 ) 112 ]
where
8 =tan - 1 [(K sin a )/('r + K + 2S)]
8 = (r 2 +K 2 +2rK cos a) 112 /(r +K +2S)
S = (2Xr)/(r + 1)
Further
Oc =rOE
W=(r-1)QE
COP= 1/(r -1)
154 Chapter 3

From this it is easy to see the refrigerating capacity is a linear function

of (a) the engine operating frequency N, (b) the maximum pressure of the
working fluid Pmax• and (c) the size of the engine expressed in terms of
characteristic volume VE or the combined swept volumes VT = (Vc + VE).
According to the Schmidt theory, to double the refrigerating capacity one
simply doubles the frequency, the maximum cycle pressure, or the combined
swept volumes, i.e., uses a bigger machine. Alternatively, a combination
of increased speed, pressure, and size might be used.
In practice, the results are not as straightforward. Increase in the
pressure and speed will certainly increase the refrigerating capacity, but
the improvement occurs at a progressively diminishing rate because of
increasing friction losses and thermal saturation of the matrix. Increase in
speed or pressure to excess actually decreases the available refrigeration.
The effect on performance of the four principal design parameters (T,
K, a, and x) is less obvious. It is not clear which combinations should be
used to achieve the optimum performance. This is important, for these
parameters are determined at the design stage and, except for temperature,
T, cannot be readily changed without structural change to the engine.
The temperature ratio T is fixed by the refrigeration duty (TE) and the
temperature of the cooling available (Tc). The refrigeration temperature
TE is the temperature achieved when the refrigeration load, including the
various heat leaks to the engine, exactly balance the refrigerating effect
generated. Lower temperatures can be achieved if necessary by increasing
the pressure and speed of the engine, thereby increasing the available
refrigerating effect.
Figure 3.34 shows the effect of variation in the expansion space
temperature on the heat extracted, (QE!Pmax VT ), and the input work per
cycle (P/ Pmax V T) calculated using the Schmidt theory for a case where all

TEMPERATURE RATIO T: I TeiTEl

CO 10 6 4 3 2 I5

0~1> 03 \
~£
g
w \WORK DON~
~ 02 \ --1-
lS
""'
0 \
HE~ LIFTED

-
v- ~
"" ~
0 ~I€£ 0
0 It---
........ --

/
v 100 K
EJ. I

200 K
EXPANSION SPACE TEMPERATURE
~ !'......
TE
300 K
Fig. 3.34. Effect of variation in expansion
space temperature TE on heat extracted,
input power and coefficient of performance
with constant values of K, X, and a.
Stirling Cryocoolers 155

~I,;
I 0
:: ~ 0 2 ,-----,-----,--...-----,

'"z
g
"'
<r
0
;!:

:5 011---fr----t--1-----j

Fig. 3.35. Effect of variation in the swept volume ratio 20

K on the heat extracted (QEI Pmax V T) with the constant SWEPT VOLUME RATIO Vc I VE

values of -r, X, and a.

the other design parameters were maintained constant, i.e.,X = 1, a = 105°,

K = 2 and with T c = 300 K. As the expansion space temperature decreases,
the heat lifted decreases, the input work increases and the coefficient of
performance decreases. This is perfectly in line with our expectations from
previous experience with the Carnot cycle. The Schmidt cycle has the same
thermodynamic performance as the Carnot cycle since it has the same
isothermal heat transfer to and from the cycle.
Figures 3.35, 3.36, and 3.37 show the effect on heat lifted (QE/Pmax VT)
and the work done (P/ Pmax V T) as a result of independent variation of one
of the three parameters K, a, and X with the other two maintained constant
and with r maintained at 3 (TE = 100 K, Tc = 300 K) for all cases. The
coefficient of performance is constant since it depends only on r.
Figure 3.35 shows the effect of variation in the swept volume ratio k.
For a given combination of r, X, and a, there is a definite optimum value

~03

0]~e
-
1
0

!:~

'"0z
WO'K DONE
............
0

K
~ 02
;!:

"'
0 ~

--
..............

HEAT LIFTED
,"""-
~
r---

Fig. 3.36. Effect of variation in the dead space ratio X

on the heat extracted (QEI Pmax VT) with constant values
of r, a, and K.
0
m 2
0

10
DEAD SPACE RATIO V0 / VE
20
156 Chapter 3

Fig. 3.37. Effect of variation in the phase angle a on

PHASE ANGLE C! the heat extracted (OEI Pmax V T) with constant values of
T, X, and K.

of the swept volume ratio Vel VE. Change in the temperature ratio 'T
produces a change in the refrigeration capacity and in the optimum swept
volume ratio. Increase in the value of T (decrease in the expansion space
temperature) produce a decrease in the heat extracted and an increase in
the optimum swept volume ratio.
Figure 3.36 shows the effect of variation in the dead space ratio, X.
Increase in the dead space causes a decrease in the heat lifted so the dead
space volume must be minimized for high refrigerating effects. In practical
engines, the situation is more complicated with increased dead space. A
positive improvement in refrigeration performance can result, provided
the increased dead space is effectively utilized as increased void volume
in the heat exchangers. This results in increased thermal capacity or bigger
flow passages with reduced fluid friction losses. It is vital that the dead
volume be limited to effective space and not wasted in oversized ducts,
large ports, and excessive clearance space.
Theoretical deterioration of refrigerating capacity with increase in
dead volume can be understood by considering the effect on the volume
compression ratio ( V max! V min). Increases in dead space decreases the
volume compression ratio and the pressure compression ratio (Pmax!Pmin).
Reduction in the pressure ratio results in decreased area of the expansion
space work diagram and a smaller refrigerating capacity.
Figure 3.37 shows the effect of variation in the phase angle a. The
refrigerating capacity is remarkably insensitive to changes in the phase
angle from 60° to 120° of crank rotation, but does exhibit an optimum
value around 90° at which the heat lifted is a maximum.
For any given combination of temperature ratio, 'T and dead space,
X, three-dimensional surfaces may be drawn as shown in Fig. 3.38. This
illustrates the heat lifted (OE! Pmax V T) as a function of swept volume ratio
K and the phase angle a. Different combinations of 'T and X result in the
Stirling Cryocoolers 157

Fig. 3.38. Three-dimensional repre-

sentation of refrigerating capacity
(QEIPmax Vr) as a function of swept
volume ratio K and phase angle a for a
given combination of temperature ratio
(T = 3) and dead volume ratio (X= 1).

generation of different surfaces, sitting one upon the other like onion skins.
The apex of the surface in Fig. 3.38 is the maximum refrigeration capacity
achievable with the given combination of r and X. The values of K and a
corresponding the apex of the surface are the optimum values necessary
to achieve the maximum refrigeration.

Consolidated Design Chart

Construction of surfaces such as Fig. 3.38 provides a visual demonstra-
tion of the interdependence of the refrigeration capacity on various design
parameters. It would be tedious indeed to resolve optimum combinations
of parameters by such diagrams. The advent of computers made the task
of optimization straightforward. Optimal combinations for several hundred
cases were determined (Walker, 1962) and consolidated design charts
prepared for both prime movers and refrigerating machines. The chart for
refrigerating machines is reproduced in Fig. 3.39. The chart for engines is
reproduced elsewhere (Walker, 1980).
The design chart is in three sections, having a common abscissa of
expansion space temperature TE (Tc is assumed to be constant at 300 K).
The corresponding scale for the temperature ratio r is shown at the top
of the diagram. The upper diagram shows the refrigerating capacity as a
function of expansion space temperature TE for different values of the dead
space, X. The center diagram shows the optimum phase angle a (units of
the a ordinate are rr radians) as a function of expansion space temperature
TE for the same several dead space ratios. The lower diagrams shows the
158 Chapter 3

I #. v
#. '/ /
0·12
/-~ 0 / v
A~ ~.....-: Y!
8

~~~
v I
~ ....- ~
,/ P"
I
1\ i\ I
0·04
r-"
1-----
X""0•1,0·25, 0·5, 1·0, 1·5, 2·0

I I I I I

X"' 0·1, 0·25, 0·5, 1·0, 1·5, 2·0

I \I
7

1/
II
6

5
I I I
I

0-.

" r--..
!'.:: .......

"~ 1'-.
I')._
~
~
:-
1 I T '::::::::: ::§
X"'0·1, 0·25, 0-5, 1·0, 1·5, 2·0

40 50 60 75 100 150 00 30
Fig. 3.39. Consolidated design chart for Stirl-
30
Expansion space temperatura (K) ing cryocoolers based on Schmidt theory.

optimum swept volume ratio K as a function of expansion space tem-

perature, TE for the different values of the dead space, X.
The maximum refrigeration decreases linearly with expansion space
temperature and is a strong function of the dead space ratio, X, falling
progressively as X increases. The optimum phase angle is independent of
both dead space and temperature effects. The optimum swept volume ratio
K is more or less independent of the dead space but is a strong function

of the expansion space temperature. As the temperature decreases, the

optimum swept volume increases at a progressively increasing rate from
approximately unity near ambient temperature to 5 or so at cryogenic
temperatures. This shows that at lower temperatures, a bigger compressor
with increased power input is necessary to achieve a smaller refrigerating
capacity.
Stirling Cryocoolers 159

To use the chart, one simply determines the expansion space tem-
perature at which refrigeration is required. Through this point, a vertical
line is drawn crossing the three diagrams. For different values of the dead
space, X, one can read off values of the maximum refrigeration capacity,
the optimum phase angle and swept volume ratio to achieve capacity. The
actual refrigerating capacity will be 30% to 50% of the indicated capacity.

Alternative Basis for Optimization

The design chart, Fig. 3.39, was prepared for the optimum design in
terms of heat lifted QE expressed in dimensionless units based on the
maximum pressure (Pmax) and the combined swept volume (VT ). This is
the preferred basis for optimization because the maximum pressure is
indicative of the weight of an engine and the combined swept volume (VT)
is indicative of the size of an engine. Thus, optimization generates an engine
design having the maximum refrigeration capacity for a given size and
weight.
There are alternative bases for optimization. The heat extracted can
be expressed per unit mass of working fluid (QE/mRT) [Eq. (3.25)], where
m is the mass of working fluid. Thermodynamic cycle calculations are often
calculated on the basis of unit mass of working fluid so the parameter
appears as a familiar and suitable basis for optimization.
However, when applied to Stirling engines, an optimization based on
unit mass of working fluid results in an engine design where the working
fluid is limited. The design is grossly distorted to achieve the maximum
refrigerating effect from a small mass of working fluid. It is clearly not the
proper basis for optimization. There are other bases, equally false, explored
in some depth by Walker (1962).
Once the combination T, X, a, and K has been established, it is just as
valid to use (QE/mRy) as (QE/PmaxVT). They produce identical values of
QE. The danger of a distorted design only arises when optimization studies
are in progress.

Machine Design Procedure

The consolidated design chart was prepared using Schmidt theory
based on an idealized thermodynamic cycle not attainable in practice. Use
of the chart should be confined to the preliminary layout stages of design.
It will provide a handy basis for starting the design and the calculation of
approximate performance.
Various aspects of mechanical and thermal design are discussed in
Chapter 9. If these guidelines are followed, a first estimate of actual
160 Chapter 3

indicated refrigerating capacity might be obtained by dividing the chart

value (QE/Pmax VT) by 3 (this decrease largely arising from adiabatic rather
than isothermal processes). The corresponding indicated input cycle work
might be estimated by assuming W = 1.2T(QE!Pmax VT ).
Now the various thermal losses can be estimated and can be debited
against the initial estimate of the indicated refrigerating value to provide
a first guess at the net available refrigerating capacity. For a machine
operating in thermal balance, the equality of thermal losses to indicated
refrigerating capacity will allow a preliminary estimate of the operational
refrigerating temperature.
The indicated cycle work must be increased to allow for mechanical
and fluid friction losses in the kinematic drive and rubbing friction of seals
and guides on the pistons or crossheads. The shaft input is likely to be
anywhere from 20% to 80% greater than the indicated cycle work depend-
ing on the size of the machine, the ingenuity and skill of the design engineer,
and the quality of the final manufactured product.
Another important factor is the efficiency of the input power drive
motor. Electric drive motors are used, both alternating and direct current,
with and without brushes. In large sizes, the efficiency of these motors may
approach 90%. In the smaller sizes used for low capacity coolers, the
efficiency rarely exceeds 60% and is often substantially less.
Various design layouts and machine configurations can be quickly
evaluated and preliminary performance estimates made. When a suitable
design has evolved which conforms to the constraints imposed by the
application or client, one can proceed to the next level of analysis.
The second stage of design would likely involve analysis according to
the Finkelstein adiabatic cycle using digital computers, but at a relatively
low level. Supplementary corrections for fluid flow and thermal losses can
be made using the methods of Martini or Smith, Rios, Qvale, and Rea
mentioned above. Various revisions to the outline design may be made at
this stage to optimize the system according to the application.
Finally, a detail design may be prepared and analyzed using the nodal
analysis programs discussed above. A relatively coarse network may be
specified initially to minimize computer processing time with progressive
elevation of the level of sophistication as the design approaches the ultimate
stage.
This is an idealized procedure for machine design rarely followed in
practice. Time and budget do not allow comprehensive exploration of
design alternatives. Many machine designers have "blinkers" which inhibit
their disinterested assessment of new and better ways to go. Sponsors rarely
have an adequate basis of technical expertise to critically assess their
contractors' work. Many times they naively assume that, even though the
Stirling Cryocoolers 161

contractor is a recent entrant to the field, he will go straight to the optimum

design. There is a great divide between computer and hardware specialists.
Too often, computer simulation becomes an end in itself providing endless
delight to those pursuing it but contributing little to the machine design.
Designers cannot understand the acronymic sublanguage of the computer
specialist and system analysts often have unrealistic expectations of
engineering materials and production processes.
Nothing impresses a sponsor so much as working hardware. Frequently,
contract renewals depend on the existence and demonstration of a pro-
totype. Therefore, the project manager is anxious to start cutting metal at
the earliest possible stage and is impatient with the design and simulation
types who, in his terms, "frig about" unable to make up their minds.
Eventually he insists the design be frozen at some arbitrary level and
production started. Thereafter, various small changes are possible with all
the attendant red tape and paperwork but substantial design modifications
cannot be made. Frequently, man-years of effort are devoted to the
engineering development of an inferior design or improperly evaluated
concept.

TYPES OF STIRLING ENGINES

Multiple-Expansion
The possiblity of an engine with multiple-expansion spaces coupled
to a single compressor was discussed above. The expansion process in all
spaces occurs simultaneously and at the same pressure. In electrical
engineering terms, this is "parallel expansion." There also exists the possi-
bility of "series expansion" at different temperature levels in a double
(triple or quadruple) expansion engine.
The principle of the double-expansion engine is illustrated in Fig. 3.40.
This shows the cross section of an integral Stirling single-acting piston-
displacer engine. The regenerator is contained within the displacer and
moves with it. The displacer-regenerator is arranged in two sections having
different diameters, with the top section substantially smaller than the lower
section. The stepped displacer operates in a cylinder which has correspond-
ing changes in diameter to fit the displacer.
Two expansion spaces are thereby created, namely, the customary
expansion space above the displacer, and an intermediate or "extra"
expansion space as shown in Fig. 3.40. The volume variations of the two
expansion spaces are in phase with each other. The two spaces are mutually
coupled through the porous regenerators and with the compression space
162 Chapter 3

20 K

60 K

300 K

Figure 3.40. Double-expansion Stirling engine

(after Lindale, 1978).

so the pressure in all spaces is the same, apart from slight differences due
to regenerator pressure drop.
When the machine operates, it produces refrigeration at two tem-
perature levels, shown in Fig. 3.40 as 60 K in the intermediate expansion
space and 20 Kin the upper expansion space. The generation of intermedi-
ate temperature refrigeration creates a "thermal buffer" to absorb heat
leaks from ambient temperature sources and thereby protect the refriger-
ation generated at the lower temperature. The heat leak to the upper
expansion space is now much reduced for the temperature potential
between the buffer and low-temperature expansion space is only 60-20 K.
The drain of low-temperature refrigeration is correspondingly reduced.
Space limitations preclude a full discussion here of thermodynamic aspects
of multiple-expansion engines, but the same principle of cascade refriger-
ation at progressively lower temperatures is widely used in large-scale
recuperative gaseous liquefaction systems.
The pioneer work in double-expansion Stirling engines was carried
out at Philips in the early 1960s (Prast, 1963) using a modified air liquefier
of the type shown in fig. 3.1 with additional expansion space. With this
machine, temperatures of 20 K were readily achieved. It has been incorpor-
ated in a variety of hydrogen and helium liquefiers and other low-
temperature refrigeration systems.
Later on, Daniels and du Pre (1971) described a miniature triple-
expansion Stirling engine which attained a temperature of 7.8 K. A
diagrammatic cross section of the engine is shown in Fig. 3.41 and the
multi-element displacer in Fig. 3.42. Figure 3.43 is a view of the prototype
unit with the actual cylinder head. The tangle of tubes and wires around
the cylinder head are part of the apparatus for measuring the expansion
Stirling Cryocoolers 163

Fig. 3.41. Diagrammatic cross section of a miniature triple-expansion

Stirling engine (after Daniels and du Pre, 1971 ).

space temperature using a hydrogen vapor-pressure technique with copper

constantan thermocouples for the intermediate temperatures.
The cool-down characteristic of the engine is shown in Fig. 3.44. This
shows the temperatures of the three expansion spaces as a function of time.
The engine was relatively slow to cool down (about 1 hr) because of the
low pressure and speed. The relatively large refrigerating capacity of the
first stage enabled it to pull down the temperature of is expansion space
very quickly (about 20 min to reach 60 K). The smaller second stage lagged
behind in temperature quite considerably but was still substantially ahead
of the small third stage most of the time. Eventually in "hare and tortoise"
fashion the third stage passed both the other spaces to reach the lowest
temperature. Part of the reason for the slow cool-down was the relatively
large mass at the low-temperature end associated with the instrumentation.
A production machine would be cleaner and have a much reduced cool-
down period.
164 Chapter 3

Fig. 3.42. Multielement regenerative displacer

of the triple-expansion Stirling engine (after
Daniels and du Pre, 1971).

Multiple-expansion systems have been used with other regenerative

engines, i.e., Vuilleumier and Gifford-McMahon machines, with much the
same result. Zimmerman and Radebaugh (1977) described a multiple-
expansion engine, shown in Fig. 1.3, constructed of cheap, readily available
materials that may be the forerunner of a low-cost, maintenance-free,
cryocooler for widespread use in superconducting electronic applications.

Free-Piston
This already long chapter could be greatly extended by proper con-
sideration of free-piston Stirling engines. In these machines, the pistons
Stirling Cryocoolers 165

Fig. 3.43. Laboratory prototype of miniature triple-expansion Stirling engine (after Daniels
and du Pre, 1971 ).

and displacers are not coupled through a kinematic chain (crank and
connecting rod) but are activated solely by fluidic forces acting on them.
Free-piston Stirling engines are in rapid development as prime movers
and, to a lesser extent, as cryocoolers (Haarhuis, 1978; de Jonge, 1979).
A comprehensive discourse on Stirling free-piston engines is contained in
 166 Chapter 3

300

INITIAL WORKING PRESSURE 75 psiq

OPERATING SPEED: 600 RPM

w
a: 160
.....,::>
a:
w
140 \' \

\
0. \
::.
~ 120
\
•oo I \

80
'
\ \ / - lsi STAGE
'
60
"---~---,
40
''
- - - 2nd STAGE
>O
- 3rd STAGE Fig. 3.44. Cooldown characteristics of
10 20 30 40 !)Q 60 70
a miniature triple-expansion Stirling
TIME (MIN) FROM START engine (after Daniels and du Pre, 1971).

Chapter 11 (contributed by William Beale) in Stirling Engines (Walker,

1980). An increasing number of technical papers about Stirling free-piston
engines are appearing in the technical literature. Enough material exists
for a monograph on the topic planned to succeed this present work.
One type of free-piston engine, called the Beale free-piston Stirling
engine, after the originator William Beale of Sur..power, Athens, Ohio, is
shown in Fig. 3.45. The engine consists essentially of a lightweight displacer
and a heavy, massive, piston contained in a single cylinder in the arrange-
ment classified above as a single-acting integral Stirling system. The
expansion space is above the displacer and the compression space is between
the piston and displacer. The two spaces are coupled through a regenerator.
In a cryocooler, work input is necessary to drive the engine. This may
come from a pulsating hydraulic actuator or, as in Fig. 3.45, from a linear
motor providing an electromagnetic force on the piston.
The space below the piston, called the "bounce space," acts as a
pneumatic spring and contains the same gaseous working fluid used in the
Stirling engine above the piston. Application of an electromagnetic force
on the piston by the linear motor will cause the piston to descend, thereby
compressing gas in the bounce space. Release of electromagnetic force will
allow the compressed gas in the bounce space to return to the piston.
Stirling Cryocoolers 167

:,,
tVe l ,..,. IVe)mo• tVelmln

I
Pmln IP~,
(V(lmu IVe:Jmln

.
IVtlmln
6 •

Fig. 3.45. Beale free-piston inte-

•
TOTAL WORKING SPACE
EXPANSION
gral Stirling engine. v v

Application of the force at the resonant natural frequency of the pis-

ton/bounce space will cause the piston to oscillate with simple harmonic
motion in the same way a cork bobs up and down in water or a weight
suspended from a spring.
Oscillations of the piston in the cylinder will cause the pressure to
vary in harmonic fasion. If the diplacer were not included, the space above
the piston woudd act simply as another bounce space. The pressure variation
above and below the piston would be exactly out of phase by one half
cycle (180°) and vary at the same frequency as the piston oscillation.
In order to produce refrigeration, it is necessary to include the displacer
and regenerative heat exchanger as shown in Fig. 3.45. The displacer is
equipped with a large-diameter displacer rod passing right through the
piston and subject to the pressure of fluid in the bounce space across
the whole area of the displacer rod. An equivalent area at the top of the
displacer is subject to the pressure of the working fluid in the expansion
space. The net fluidic force acting on the displacer is equal to the difference
between the pressure in the bounce space and the working space.
168 Chapter 3

The mass ratio of piston to displacer and the area ratio of displacer
rod area to cylinder cross section area are critical to successful operation.
When these are in proper proportion, the displacer will be found to oscillate
in the cylinder as shown in Fig. 3.45, at the same frequency as the piston
but with a phase difference about 80° in advance of the piston. The motion
of the displacer in the cylinder causes the working fluid to move back and
forth between the expansion and compression spaces. The pressure is
decreasing when the displacer is closest to the piston. Thus, there is a
concentration of fluid in the expansion space and heat is abstracted from
the surroundings of the expansion space at low temperatures. Similarly,
the fluid is concentrated in the compression space when the pressure is
increasing; the displacer is widely separated from the piston and the piston
is rising on the compression stroke.
Figure 3.45 contains displacement-time curves for the piston and
displacer and pressure-time curves for the working space and bound space.
Typical pressure-volume diagrams are also included for the expansion,
compression, and total working spaces. On Fig. 3.45, the numbers 0 through
8 correspond to critical points of the cycle of operation in accordance with
the following key:

0-Minimum compression space volume V C(min)· The piston and dis-

placer are in contact.
1-Maximum expansion space volume V E(max)· The displacer is at the
bottom of its stroke.
2-Minimum pressure in the working space Pmin·
3-Maximum system volume V(max)· The piston is at the bottom of its
stroke and the pressure in the bounce space is a maximum.
4-Maximum compression space volume V c(max)· The displacer and
piston are at the point of maximum separation.
5-Minimum compression space volume V E(min)· The displacer is at
the top of its stroke.
6-Maximum pressure in the working space, Pmax·
7-Minimum system volume V(min)· The piston is at the top of its
stroke and the pressure in the bounce space is a minimum.
8-Minimum compression space volume V c(min)· This is a return to
the same condition as 0 and the completion of the cycle.

Many alternative arrangements and configurations are possible. Some of

these are briefly considered by Beale in the quoted reference (Walker,
1980).
This free-piston family also includes the resonant liquid-piston engines
exemplified by the Harwell Fluidyne engine (West, 1971). A number of
Stirling Cryocoolers 169

research papers on these engines were presented at the 14th Intersociety

Energy Conversion Engineering Conference, Boston, August 1979. So far
as is known, no practical work with liquid-piston Stirling cryocoolers has
been attempted, but any number of diverting concepts exist which sooner
or later will be the subject of investigation.
A liquid-piston engine overcomes the problem of piston seals and
their dry-rubbing friction. Furthermore, it is interesting to reflect on the
possibilities for cryocoolers where the expansion space piston is actually
condensed working fluid. Successive condensation and vaporization of the
working fluid in the cold space would allow substantial transfers of heat
and large changes in pressure with minimal actual fluid movement. The
prospect of a two-phase two-component reactive liquid-piston Stirling
cryocooler is highly diverting. The condensed fraction of the unreacted
two-component mixture would be the cold end piston with dissociated
vapor working fluid and condensed liquid of perhaps one fraction as the
ambient temperature piston. No fluids with the proper combination of
properties come immediately to mind and no exploratory work along this
avenue is known to be in progress.

Free· Displacer Split-Stirling

An important class of cryocooler much favored for infrared sensor
and electronic applications is the split Stirling engine with a free displacer
in the expander element. A cryocooler of this type manufactured by CTI
is shown in Fig. 3.46. It consists of a compressor element and an expander
element coupled by a single long tube. The compressor piston could be a
free piston device energized to operate at resonant frequency with pneu-
matic or mechanical springs, but in all known machines, the compressor
piston is activated by a conventional crank connecting rod mechanism
driven by an electric motor.
A diagrammatic cross section of the expander is shown in Fig. 3.47.
The expander contains a regenerative displacer with a displacer rod added
at the ambient temperature end extending through the seal into the bounce
space. The cold end of the expander cylinder is the expansion space and
the ambient temperature end contains part of the compression space. The
other part of the compression space is in the compressor above the piston.
When the compressor operates, the pressure of the working fluid Pc
fluctuates periodically at the same frequency as the compressor. The pres-
sure in the bounce space remains substantially constant at the mean cylinder
pressure (pb = Pmean). There is, therefore, a net fluidic force acting on the
dis placer equal to the difference in pressure, Llp = Pc - pb, times the area
of the the displacer rod.
170 Chapter 3

Fig. 3.46. Free displacer split-Stirling cryocooler, Model No. CM-3 80 K manufactured by
Cryogenic Technology Inc., Waltham, Massachusetts.

Assume first that the displacer has a low mass (therefore a low inertia)
and is completely unrestrained to motion in the cylinder within the confines
of the permissible stroke. The net fluidic force acting on the displacer will
cause the displacer to shuttle back and forth in the cylinder as shown in
the P-V plane, Fig. 3.47b, for both the compression and expansion spaces
in the cylinder. If the displacer had no mass, the motion would be the
straight line A-B-C. However, the displacer must have some mass and so
will follow the curved path A-E-B-F-C. It will return along the same path
and consequently there will be no refrigeration effect, for the area of the
work diagram on the return stroke will be exactly the same as on the
outward stroke. The net area of cycle work will be zero.
Now let the displacer be partially restrained by the sliding seals at the
warm end of the displacer, separting the working space and bounce space.
Static friction will prevent the displacer moving until a sufficient fluidic
Stirling Cryocoolers 171

(b)

·~·b_ v
(c)
v

·~~
Fig. 3.47. Diagrammatic cross sec-
tion of split-Stirling expander ele-
ment, including the pressure-time
and work diagrams for the com pre- {a) EXPANDER ELEMENT
sion and expansion spaces of the (d )
expander element. EXPA"lSION SPACE

force is acting to overcome the friction. The displacer will then move, but
dynamic friction will require some fluidic forces to be acting on the displacer.
Therefore, the line of motion followed on the P-V plane on the outstroke
will be different from that followed on the instroke and a definite network
diagram will be created. Diagrams arising from both a light and heavy seal
restraint are shown by Fig. 3.47d. Increase in the restraint increases the
separation of the outstroke and instroke paths, thereby increasing the cycle
work and the refrigerating effect.
This ingenious free-displacer split-Stirling concept was originated by
Walter Riga of the California Jet Propulsion Laboratory (Riga, 1965) and,
independently shortly thereafter, by Fred Chellis at Cryogenic Technology
Inc., Waltham, Massachusetts and Stuart Horn of the U.S. Army Night
Vision Laboratory, Fort Belvoir, Virginia.
The principal difficulty is the need to maintain the seal friction restraint
constant throughout the life of the unit. It is difficult to reproduce and
maintain constant the friction conditions of presently available dry rubbing
materials in a helium atmosphere. One solution (Walker, 1978) is to
incorporate a small auxiliary compressor feeding a separate duct to the
bounce space to provide a reproducible well-regulated fluidic drive force
on the regenerative displacer.
Horn et al. (1973) have presented an excellent theoretical analysis of
the free-displacer split-Stirling engine operation. No validation of the
analysis by comparison with experimental data has been published in the
open literature.
172 Chapter 3

STIRLING CRYOCOOLERS OF INTERMEDIATE CAPACITY

The great majority of Stirling cooling engines in use are small systems
of low capacity used for local gas liquefaction or miniature electronic
instrument or infrared thermal sensors. Larger machines have been con-
structed, but presently there is no large demand for intermediate capacity
cooling systems. Future developments anticipated for the relatively near
term may change the situation. One development is the on-board
reliquefaction of vaporized liquid natural gas for marine LNG tankers. As
the "energy crisis" becomes more acute the marine transport of LNG is
increasing dramatically. It is possible to drive the propulsion engines of
the ship with the vaporized gas but it is difficult to arrange the rate of
boil-off to correspond exactly with the rate of consumption and some
reliquefaction may be required to eliminate the emission of combustible
vapors. Similar considerations apply to land-based peak-saving LNG
storage systems or tidewater terminal facilities.
Another area requiring cryocoolers of intermediate capacity is foreseen
for future superconducting applications. These include the superconducting
magnets of ship propulsion engines, magnetohydrodynamic power
generators, magnetic levitation of high-speed rapid transit systems, atomic
physics particle accelerators, and the magnetic plasma containment systems
of fusion power reactors. Superconducting electric cable transmission (likely
in conjunction with a liquid hydrogen/liquid natural gas cryogenic pipeline)
is another area where major requirements for intermediate cryogenic
cooling systems are required.
The future for these machines is, therefore, bright and it is timely to
look briefly at presently available machines and to speculate on future
development.

EXISTING INTERMEDIATE CAPACITY CRYOCOOLERS

A four-cylinder version of the original Philips Stirling gas liquefier is

shown in Fig. 3.48. The unit can be equipped with a condensation pressure
head operating at up to 20 atmospheres pressures (Model No. PPG-400)
or with a cooling pressure head (Model No. PPS-401), able to provide
cooling at any temperature and pressure in the range from 60 to 300 K
and 1 to 30 atmospheres.
The refrigeration characteristics of the four cylinder engine are shown
in Fig. 3.49. This shows the refrigeration capacity and power requirement
as a function of the condensing temperature TE for normal conditions of
operation. The speed of the engine is 1450 revolutions per minute (24Hz)
Stirling Cryocoolers 173

Fig. 3.48. Four-cylinder, single-acting integral-Stirling cryocooler, Philips Model No. PPG-
400.

and the mean pressure of the helium working fluid is 25 atmospheres. The
engine cool-down time is 10 min to attain rated capacity from a warm start
and a duty cycle of up to 300 hr of continuous operation is prescribed. The
unit requires 3 m 3 hr of cooling water at ambient temperatures. The weight
of the system including the bedplate and drive motor is 1000 kg, excluding
the concrete foundation. The installation envelope is 1m highx0.75 m
wide x 1.8 m long, again excluding the concrete foundation which adds
0.9 m to the height.
A machine of large capacity, again developed and manufactured by
Philips (Model No. PPG-2500) was shown earlier in Figs. 3.3 and 3.4. The
development of this machine has been described in detail by Dros (1965).
It is an interesting and unusual machine consisting of four single-acting
Stirling engine systems of the two-piston type in an opposed configuration.
174 Chapter 3

MODEL No PPG·400 .
SPEED
MEAN
- 1450rpm
~
0
PRESSURE - 25 otmos. u
COOLING l) 310• C
WATER 2)3m 3 /hr.

~
..
~
;!!;
3or----t----~----r----+----t----;

"'
.."'
;o
0

0
~

Fig. 3.49. Refrigeration charac-

teristics of the four-cylinder,
single-acting integral Stirling cryo-
0 100 200 300
00
cooler, Philips Model No. PPG-
CONDENSING HEAT TEMPERATURE (TEl(K) 400.

The unit embodies the hermetic rolling seal developed by Philips for their
Stirling prime-mover engines. It is driven by an electric motor supplying
shaft work to a vee-piston crank and connecting rod assembly (actually a
Joy air compressor unit). The pistons activate a hydraulic fluid connection
which operates on the compression and expansion space pistons causing
them to oscillate in the proper sequence. Massive external tie rods, easily
seen in Fig. 3.3, are used to contain the cylinder end caps. When the
machine is operating, these elastic members can be seen to be visibly
stretching and relaxing. It takes some getting used to this motion as normal
for machine operation.
The refrigeration characteristics for the engine, presented by Dros,
are reproduced in Fig. 3.50. The characteristics show the refrigeration
capacity (designated PE ), the input power (designated PM) and the coefficient
of performance relative to the Carnot value (designated as TilT/c) as a
function of the cold space temperature, TE. The discontinuity in the charac-
teristics, an increase in the slope of the refrigeration capacity curve and
the conversion to constant input power, PM, arises as a consequence of the
control system. When the power input to the drive motor approaches the
rated limit, 134 kW, working fluid is released from the cylinder to reduce
the pressure and so maintain the power input constant as the expansion
space temperature decreases further. Reduction in pressure of the working
Stirling Cryocoolers 175

80% 140/<W 80% uo

- -
0 ~ 12 0
\ ~

0 .~ c
0
~
0 / ~~ 0 0 r- -!J!.nc ~
I ~~ ~~
30 0 ~60
20
VI}
/
/ ~
0 20

A
~ K' iO

0 / 1 / 0
/ H2 1/ He

0 0 0
0 50 100 150 200 250"1< 50 100 160 200 250"1<
--<>1£ -1£

Fig. 3.50. Refrigeration characteristics of the four-cylinder, single-acting opposed-piston

Stirling cryocooler, Philips Model No. PPG-2500.

fluid decreases the refrigeration capacity of the engine as shown by the

change in slope of the capacity characteristic.
Dros presented refrigeration characteristics for the engine operating
with both hydrogen and helium as the working fluid. It can be shown
(Walker, 1980) that the ratio of heat transfer to pumping loss is a function
of the molecular weight of the transfer fluid. The highest ratios are achieved
with fluids having the lowest molecular weight. For this reason, hydrogen
is the preferred working fluid for Stirling engines. Helium is the preferred
alternative in many cases on grounds of safety, for helium is inert and
hydrogen is very flammable. The superiority of hydrogen at the higher
expansion space temperature can be clearly seen by comparing the charac-
teristics presented in Fig. 3.50 for hydrogen (left-hand diagram) and helium
(right-hand diagram). At 100 K, the coefficient of performance relative to
the Carnot value is about 46% for hydrogen and about 42% for helium.
At lower temperatures, hydrogen loses the advantage over helium
because helium deviates less from ideal gas behavior than hydrogen as it
approaches the saturated vapor state. Hydrogen liquefies at higher temper-
tures than helium (20 K compared with 4 K at 1 atmosphere) so for
low-temperature cryocoolers, helium is the only possible fluid in operation
when a single-phase fluid desired.
The model PPG-2500 cooling engine operates at 730 revolutions per
minute at a mean pressure of 40 atmospheres. It requires 20 m 3 per hour
of cooling water at ambient temperatures. The weight of the engine includ-
176 Chapter 3

ing the drive motor is 6500 kg excluding the concrete foundation. Overall
dimensions excluding the foundation are 1.95 m high, 1.6 m wide, and
2.97 m long.

The Werkspoor Cryocooler

Another Stirling cooling engine of intermediate capacity is shown in
Fig. 3.51. This was developed in the 1950s and 60s by the Dutch marine
and general engineering company Werkspoor N.V. of Amsterdam. It was
marketed in the United States by the CVI Corporation of Columbus, Ohio
as the principal refrigeration component for small-scale liquefaction equip-
ment for natural gas. Vogelhuber and Parish (1968) outlined an integrated
LNG system incorporating the Werkspoor engine and Joule-Thomson
expanders capable of liquefying up to 2000 mscfd of natural gas.
A commercial brochure isssued by CVI entitled "Technical Description
of an 80 MCFD LNG Liquefaction Facility," March 1968, contained a
reproduction of an unfamiliar reference* that is the most complete descrip-
tion of the engine known. In Fig. 3.51, it can be seen that the engine was
a single-acting integral Stirling machine with a rhombic drive mechanism
for activating the piston and displacer in the proper phase relationship.
The unit was driven by an electric motor. The engine and drive motor were
skid-mounted on a bedplate with separate control console. With a rhombic
drive, perfect dynamic balance could be readily achieved so that substantial
foundations were not required.
Most descriptions of the engine feature a single cylinder version but
include mention of the possiblity of two-, three-, and four-cylinder
machines. It is not known if multiple cylinder versions were actually
constructed.
The VMF Review article included the following technical description
of the single-cylinder version with electric drive motor:
Total mass of skid mounted unit: 2000 kg,
Mean pressure of hydrogen working fluid at full load: 35 atm,
Cooling water consumption: 9m 3 hr,
Cooling water temperature increase: l0°C,
Cooling water pressure drop: 1.5 atm,
Cooling water pump power input: +0.6 kW,
Refrigeration temperature range: +75 to 180 K,
Installation envelope: 1.9 m high, 1.03 m wide, 2.6 m long,
Control console dimensions: 0.98 m high, 0.95 m wide, 0.76 m long.

*"The Werkspoor Cold Gas Refrigerator," Anon. VMF Review, 111(23), 125-131 (date
unknown).
Stirling Cryocoolers 177

DIS PLACER

EXPANSION
SPACE

.!1----- 'vi'----- FREEZER

- - - I T - - - - REGENERATOR

COMPRESSION
SPACE

PI STON

Fig. 3.51. The Werkspoor Stirling crycooler of intermedia te capacity.

178 Chapter 3

.
o-
IZO
~
COEFF11CIENT OF PERFORMANCE
RELATIVE TO THE CARNOT ~
i
-
I

"''"

\\
o- 100 04 o-
i?
;!' -~ 0
o-

.",.
0 80
t'-.. 03~
'">

..
0
z
WERKSPOOR

MEAN
ENGINE
..~
/ ~ - 35 atmos

.'"
PRESSURE
>- 60
o- SPEED - 1!500rpm 02~

~ ~TEA
CYCLE INPUT

.
u COOLING - 25°C ~
.
~ 0
u ~

40
z
!0?
o-
r-- 0 ::'
~
0
~ REFRIGERATION CAPACITY o-

"''"
~ z

' 1.-----
20
'"
..~
----
u
::: Fig. 3.52. Refrigeration charac-
8 teristics of the single-cylinder
0 50 100 150 200 250 300 rhombic drive, integral Stirling
COLO SPACE TEMPERATURE (TEl(K) Werkspoor cryocooler.

The refrigerating characteristics of the Werkspoor single-cylinder Stirling

cooling engine are shown in Fig. 3.52. The mean pressure was 35 atm and
the cooling water outlet temperature was 25°C. The peak value of the ratio
of actual coefficient of performance to the Carnot value was a little over
0.4. The peak value for the Dros hydraulic opposed piston engine was
given as 0.47 (see Fig. 3.50). The reasons for this significant difference are
not known, and so far as is known, have never been compared in the
literature before. Perhaps they arise as a result of the difference in speed
(730 rpm for the Philips engine, 1500 rpm for the Werkspoor machine). It
is hard to believe that a properly designed rhombic drive engine would be
so very much less efficient than the hydraulic opposed piston engine.
Perhaps considerable potential for development remained inherent in the
W erkspoor engine.
Other interesting characteristics included in the VMF Review article,
reproduced in Fig. 3.53, show the effect on refrigerating capacity, QE, and
shaft input power, P, as a function of the mean pressure of the working
fluid, Pmean· The characteristics were measured at constant condensing and
cooling water outlet temperatures of 75 K and 25°C, respectively. The
engine speed was 1500 revolutions per minute.
In their paper, Vogelhuber and Parish (1968) indicated that 18 Werk-
spoor engines had been installed, seven on board ships for "no-loss storage
of ethylene and methane," one at an LNG facility for modular upgrading
of facility capacity, and 10 in other cryogenic applications. The unit was
said to be capable of operating 2000 hours before replacement of the piston
rod rolling seals was necessary (to prevent ingress of lubricant to the working
Stirling Cryocoolers 179

120

y
I
~ 100

..
1-
:0

:;;
SHAFT INPUT
POWER P ~
:;:

lv
80

"~ I
0:
0
60
I
~
.
u
"u
"
~FULL LOAD

/
/
40
";::z
"
I
I
0:

"'~

k +
REFRIGERATION
20 /
~
"' / Fig. 3.53. Refrigeration capacity
and power input of the Werkspoor
-
0:

cryocooler as a function of the

o 1o 20 30 40 50 60 mean pressure of the hydrogen
MEAN PRESSURE OF HYDROGEN WORKING FLUID (pmeon)(otmos) WOrking fluid.

space and contaminating the regenerator, and egress of working fluid from
the working space). Replacement of the seals occupied four man-hours of
effort.
Vogelhuber and Parish presented data on installed and operating costs
of the Werkspoor engine liquefaction system and conventional Claude
cycle and cascade cycle systems. They concluded the cost for the Stirling
based system was generally favorable in liquefaction capacities up to
1.4 mmcfd. Such an installation would be a hefty liquefaction plant necessi-
tating duplicate four-cylinder Werkspoor engine with two-stage Joule-
Thompson expansion stages, precooling equipment, and all the associated
purification and storage facilities.
Nothing further about the Werkspoor engine was published. It is not
known if the engine is in production, if CVI still handles the engine in the
United States, and how many of the units were built.
The rhombic drive for Stirling engines was invented by Rolf Meijer
at Philips about 1952/53. To those familiar with Stirling engines, it is clear
from Fig. 3.51 that the Werkspoor engine owed much to contemporary
Philips engine prime movers. Werkspoor does much of the heavy machine
work and fitting of the Dros hydraulic engine and probably was familiar
with the earlier rhombic prime movers constructed by Philips.

Future Developments
There appear to be excellent future prospects for applications of
intermediate capacity engines as reviewed earlier. It is unlikely that any
180 Chapter 3

of the machines discussed above will enter volume production. They are
too complicated and expensive. There have to be better and easier ways
to accomplish the task.
There is every prospect that double-acting Stirling engines will come
to dominate the intermediate capacity field as inevitably as in the prime
mover automotive field. Double-acting engines have half the moving parts
of single-acting systems, so the economic motivation is overwhelming.
Double-acting engines can use a variety of kinematic drives, wobble-plates,
swash-plates, Scotch-yoke, crank and connecting rods. All these have been
discussed in more detail elsewhere (Walker, 1980). Already, multistage
four-cylinder wobble-plate helium compressors can be seen on prototype
development trials at Cryogenic Technology Inc., Waltham, Massachusetts.
With little imagination, one can foresee their adaptation to the Philips
swash-plate engine configuration shown in Fig. 3.15. Such developments
will relegate the Dros and Werkspoor engines to the status of historical
curiosities.
It is more than likely that double-acting free-piston Stirling engines
with hydraulic or linear electric power input will come to dominate the
field. In large capacities, these could be made so compact, so reliable
(elimination of seals, use of gas bearings, etc.), and so simple as to be
virtually unassailable as a commercial product. Already the literature
contains many references to free-piston Stirling engines as prime movers,
and Philips have in production a miniature double-acting free-piston Stirling
cooling engine (see Fig. 3.7). More significantly, a large U.S. manufacturer
of large-scale cryogenic equipment is presently working hard on a prototype
double-acting Stirling free-piston engine along the lines enumerated by
Finkelstein (1978).

LARGE-CAPACITY STIRLING CRYOCOOLERS

No large-capacity Stirling cooling engines have been developed or

seriously addressed as concepts in the literature. The breakpoint for
economic superiority of one system compared with another is difficult to
determine. Vogelhuber and Parish (1968) attempted to provide a com-
parison, but their data, now a decade old, are obsolete and not sufficiently
comprehensive for general conclusions to be drawn.
Crystal ball gazing leads to the feeling that large Stirling engines could
never be competitive with cascade cycles or Claude systems for base-load
gas liquefaction. However, recalling with humility the times we have been
wrong in previous forcasting, the opinion is ventured with much trepidation.
Stirling Cryocoolers 181

REFERENCES
Babcock, G. H. (1885). "Substitutes for Steam." ASME Trans. 7, 680-741.
Beale, W. (1979). Private Communication. Sunpower Inc., Athens, Ohio.
Berchowitz, D. M., Rallis, C. J., and Urieli, I. (1977). "A New Mathematical Model for
Stirling Cycle Machine." Proc. 12th I.E.C.E.C., pp. 1522-1527, Washington, D.C.,
August 28-September 2.
Cairelli, J. E., and Thieme, L. G. (1977). "Initial Test Results with a Single Cylinder Rhombic
Drive Stirling Engine." Proc. ERDA, Highway Veh. Syst. Cont. Coord. Mtg., Dearborn,
Michigan, October 4-6.
Chellis, F., and Stewart, R. W. (1972). "Multiple Cold Finger Refrigeration System for
Cooling Infra-Red Detectors." AFFDL-TR-72-84, AD No. 903-939L (Cryogenic Tech-
nology Inc., Latham, Massachusetts).
Collins, S.C., and Cannaday, R. L. (1958). Expansion Machines for Low Temperature Processes.
Oxford University Press, Oxford.
Daniels, A., and du Pre, F. K. (1971). "Triple-Expansion Stirling Cycle Refrigerator." Adv.
Cryog. Eng. 16, 178-184.
de Jonge, A. K. (1979). "A Small Free-Piston Stirling Refrigerator." Paper No. 799245,
pp. 1136-1141, Proc. 14th I.E.C.E.C., Boston, Massachusetts.
Dros, A. A. (1965). "An Industrial Gas Refrigerating Machine with Hydraulic Piston Drive."
Philips Tech. Rev. 26(10), 297-306.
Finegold, J. G., and Vanderbrug, T. G. (1977). Stirling Engines for Undersea Vehicles."
Final Report No. 5030-63, J.P.L., March.
Finkelstein, T. (1959). "Air Engines." Engineer 207,492-497, 522-527, 568-571, 720-723.
Finkelstein, T. (1959). "Development and Testing of a Stirling Cycle Machine with Charac-
teristics Suitable for Domestic Refrigeration." English Electric Report W /M/3A.
Finkelstein, T. (1960). "Generalized Thermodynamic Analysis of Stirling Engines." S.A.E.
Paper No. 118B, January.
Finkelstein, T., Walker, G., and Joshi, T. (1970). "Design Optimization of Stirling-Cycle
Cryogenic Cooling Engines by Digital Simulation." Cryog. Eng. Conf., Paper K4, Boul-
der, Colorado, June.
Finkelstein, T. (1975). "Computer Analysis of Stirling Engines." Adv. Cryog. Eng. 20,
269-282.
Finkelstein, T. (1978). "Modeling Design and Optimization of Stirling Engines." Proc. 13th
I.E.C.E.C., San Diego, California, August 20-25.
Finkelstein, T. (1978). Pressure Compounding of Stirling Engines. Proc. 13th I.E.C.E.C.,
San Diego, California.
Gedeon, D. (1978). "Optimization of Stirling Cycle Machines." Proc. 13th I.E.C.E.C., San
Diego, California.
Goldwater, B., and Morrow, R. B. (1977). "Demonstration of a Free Piston Stirling Linear
Alternator Power Conversion System." Proc. 12th I.E.C.E.C., Paper No. 779249, pp.
1488-1495, Washington, D. C., August 28-September 2.
Haarhuiss, G. J. (1978). "The MC 80-A Magnetically Driven Stirling Refrigerator." Proc.
7th Int. Cryog. Eng. Conf., London, IPC Business Press, Guildford, U.K.
Higa, W. N. (1965). "A Practical Philips Cycle for Low Temperature Refrigeration." Cryog.
Techno!. 8, 203-209, July-August.
Horn, S. B., Lumpkin, M. L., Walters, B. T., and Acord, T. T. (1973). "Miniature Cryogenic
Cooler for TOW Night Sight." Proc. Closed Cycle Cryocooler Tech. and Appl., Vol. 1,
pp. 55-72, AFFDL-TR-73-149, WPAFB, Ohio, AD No. 918234.
Khan, M. I. (1962). "The Application of Computer Techniques to the General Analysis of
the Stirling Cycle." M.Sc. thesis, Univ. of Calgary.
Kirk, A. (1874). "On the Mechanical Production of Cold." Proc. Ins. Civil Eng. 37, 244-315,
London.
Kohler, J. W. L. (1965). The Stirling Refrigeration Cycle. Sci. Am. 212(4), 119-127.
182 Chapter 3

Lee, K. (1976). "The Stirling Cycle with Adiabatic Compression and Expansion." M.Sc.
thesis, Univ. of Calgary.
Lindale, E. (1978). "Stirling Cycle Refrigerators for Gamma Ray Detector." Report No.
PL-42-Cr78-0713, Johns Hopkins Univ., Applied Physics Lab., Laurel, Maryland (Philips
Laboratories).
Maki, E. R., and de Hart, A. 0. (1971). "A New Look at Swash-Plate Drive Mechanisms."
S.A.E. Trans. (Truck, Powerplant Fuels and Lubricants Mtg.), Vol. 80, Paper No. 710829,
St. Louis, Missouri, October.
Martini, W. (1978). Design Manual for Stirling Engines. NASA CR 135382, DOE/NASA
Contractor Report (NTIS, Springfield, Virginia). See also: Stirling Engine Design and
Feasibility for Automotive Use. (1979). (ed. M. J. Collie), Noyes Data Corp., Park Ridge,
New Jersey.
Prast, G. (1963). "A Philips Gas Refrigerating Machine for 20 K." Cryogenics 3, 156-160,
September.
Prast, G. (1965). "A Gas Refrigerating Machine for Temperatures down to 20 K and Lower."
Philips Tech. Rev. 26(1), 1-11, January.
Qvale, E. B. (1963). "An Analytical Model of Stirling-Type Engines." D.Sc. thesis, MIT.
Qvale, E. B., and Smith, J. L. Jr. (1969). "An Approximate Solution for the Thermal
Performance of a Stirling Engine Regenerator." Trans. A.S.M.E., I. Power 91, 109-112,
April.
Rios, P. A., Qvale, E. B., and Smith, J. L., Jr. (1968). "An Analysis of the Stirling Cycle
Refrigerator." Adv. Cryog. Eng. 13, 332-342.
Rios, P. A., and Smith, J. L. Jr. (1969). An Analytical and Experimental Evaluation of the
Pressure-Drop Losses in the Stirling Cycle. A.S.M.E. Paper No. 69-WA/Ener-8, Vol.
92, Ser. A., No.2, pp. 182-188, April.
Schmidt, G. (1861). "Theorie der Geschlossenen Calorischen Maschine Von Laubroy und
Schwartzkopff in Berlin." Z. Ver. Oster. Ing., p. 79.
Schmidt, G. (1871). "Theorie der Lehmannschen Calorischen Maschine." Z. Ver. dt. Ing.
15(1).
Schock, A. (1978). "Nodal Analysis of Stirling Cycle Devices." (a) Draft Report DOE; (b)
Proc. 13th I.E.C.E.C., San Diego, California.
Stoddart, D. (1960). "Generalized Thermodynamic Analysis of Stirling Engines." B.Sc. thesis,
Durham Univ.
Urieli, I. (1977). "A Computer Simulation of Stirling Cycle Machines." Ph.D. thesis, Univ.
of Witwatersrand, Johannesburg, South Africa, February.
Urieli, 1., Rallis, C. J., and Berchowitz, D. M. (1977). "Computer Simulation of Stirling Cycle
Machines." Proc. 12th I.E. C. E. C., Paper No. 779252, Washington, D.C., pp. 1512-1521,
August 28-September 2.
Urieli, I. (1979). A Review of Stirling Cycle Machine Analysis. Paper No. 799236, pp.
1086-1090, Proc. 14th I.E.C.E.C., Boston, Massachusetts.
Vogelhuber, W. W., and Parish, H. C. (1968). Compact LNG System Using Large Stirling
Cycle Cold Gas Refrigerator. Paper No. 28, Session 5, Proc. First Int. L.N.G. Conf.,
Chicago, Illinois (Institute of Gas Technology, Chicago, Illinois).
Walker, G. (1962). An Optimization of the Principal Design Parameters of Stirling Cycle
Machines. J. Mech. Eng. Sci. 4(3).
Walker, G., And Khan, M. (1965). "The Theoretical Performance of Stirling Cycle Machines."
S.A.E. Paper No. 949A, Int. Auto. Eng. Congress, Detroit, Michigan, January.
Walker, G. (1973). Stirling Cycle Machines. Oxford University Press, Oxford, England.
Walker, G., and Burn, K. (1976). "Exploratory Study of the Rainbow Variant Stirling Engine."
Proc. 11th I.E.C.E.C., Lake Tahoe, New York, September.
Walker, G. (1978). "Split-Stirling Cooling Engine with Fluidic Control of Displacer Motion."
Invention Disclosure, Calgary, Alberta, December.
Walker, G. (1980). Stirling Engines. Oxford University Press, Oxford, England.
West, C. (1971). "The Fluidyne Heat Engine." Report No. AERE-R-6776, Atomic Energy
Research Establishment, Harwell, Berks, England.
Stirling Cryocoolers 183

Zimmerman, J. E., and Radebaugh, R. (1977). "Operation of a SQUID in a Very Low Power
Cryocooler." Proc. of Conf. on Appl. of Closed Cycle Cryocoolers to Small Super-
conducting Devices, pp. 59-66, Boulder, Colorado, October (issued as NBS Spec. Pub.
508, April 1978-U.S. Govt. Printing Office).
Chapter 4

Vuilleumier Cryocoolers

INTRODUCTION

A Vuilleumier or VM cryocooler uses thermal energy at high temperatures

to generate cyclic pressure fluctuations in the closed engine. The pressure
fluctuations are used to produce refrigeration at low temperatures. The
heat absorbed by the engine at high temperatures and at low temperatures
is rejected at ambient temperatures. Figure 4.1 is a schematic ·diagram
showing the essential elements of the engine.
The pressure fluctuations are produced by shuttling gas periodically
from an ambient temperature region to a high-temperature region by the
action of a reciprocating displacer in the hot cylinder. There is another
displacer in the cold cylinder. The hot and cold displacers are coupled by
some form of kinematic mechanism to maintain proper phasing and
synchronous operation.
A little work is required to operate the displacers and overcome the
fluid and mechanical friction of seals and bearings. In most cases the work
input is supplied by an electric motor. Ingenious self-acting systems have
been made which use some of the heat input to generate enough work to
drive the displacers. Difficulties with seals limit the usefulness of this
attractive feature.

Vuilleumier and Duplex Stirling Engines

A Vuilleumier engine can be thought of as a Stirling engine working
as a refrigerator with a thermal compressor instead of a mechanical com-
pressor. They should not be confused with the Stirling-Stirling duplex
arrangement described in Chapter 3 in which one Stirling engine acting as
a prime mover receiving heat at high temperature produces work to drive
another Stirling engine operating as a refrigerator accepting heat at low
temperature (the useful refrigeration effect). Both reject heat at ambient
temperature.
185
186 Chapter 4

COLD
PRODUCTION
IW
77K

Fig. 4.1. Essential elements of a Vuilleumier cryocooler (after Pitcher and du Pre, 1970).

Advantages
The principal advantage claimed for Vuilleumier cryocoolers is the
potential for long life made possible by the low mechanical forces acting
on the drive mechanism, bearings, and seals. These low forces arise as a
consequence of the low pressure ratios (1.1 to 1.2) in Vuilleumier engines
compared with Stirling engines (2 to 2.5). The pressure ratios are low in
Vuilleumier engines for they are generated solely as the result of changes
in temperature, not both temperature and volume as in a Stirling engine.
Stirling cryocoolers are used where space and weight are at a premium.
The design is constrained to a machine operating at high pressure and
speed with maximum possible compression ratios to obtain the highest
refrigerating effect per unit mass or volume. The price to be paid for this
is a short operating life. It is often overlooked that it would be simple to
relax the output requirements imposed on Stirling engines and obtain the
same low speed, low compression ratio, low refrigeration output charac-
teristic of Vuilleumier engines. This would endow Stirling engines with the
same long life claimed as the principal attraction of Vuilleumier engines.
Disinterested comparative studies have shown the Stirling cryocooler
as more efficient and lower in cost, volume, and weight than the equivalent
Vuilleumier unit. The long-life characteristic of Vuilleumier engines has,
in fact, proved elusive despite a decade of intensive development. One
Vuilleumier Cryocoolers 187

particular difficulty has been the high wear rates of the hot rider, the sliding
guide ring, at the high-temperature end of the hot displacer.
The real advantage of the Vuilleumier cryocooler is the ability to use
a thermal energy input to produce refrigeration. This can be supplied by
concentrated solar energy, radioisotopes, combustion of fossil fuels or waste
materials, and finally, waste thermal streams of industrial plants or power
systems.
Solar and isotope energy made the Vuilleumier engine a strong candi-
date for cryogenic cooling in long-duration space exploration. This was
recognized by NASA, and exploratory developments have been undertaken
by the Goddard Spaceflight Center. Waste thermal streams or combustion
of waste products can be combined with Vuilleumier engines to produce
refrigeration at both cryogenic and higher temperatures for food processing
and air conditioning. Terrestrial applications of Vuilleumier engines for
air conditioning with solar energy are very attractive. In many locations,
air conditioning refrigeration is required only when the sun is shining. This
is a happier circumstance than the use of solar energy for power production,
where the main requirement for power arises at night.
In all these potential applications for Vuilleumier engines, similar
considerations apply to duplex Stirling systems, but so far as is known, no
comparisons have been made. Self-starting and complete independence of
an auxiliary electric power supply would weigh heavily in favor of duplex
Stirling systems. There are good prospects that an ingenious designer could
instill the same desirable characteristics in the Vuilleumier engine.

Thermodynamic Aspects
Most Vuilleumier cryocoolers have used electric resistance heating to
provide thermal input to the hot displacer cylinder and a motor to operate
the engine. Resistance heating for Vuilleumier engines is thermodynami-
cally inefficient. The second law of thermodynamics limits the fraction of
heat supplied that can be converted into work, the Carnot penalty. There-
fore, a high price is paid to produce electric power. Frequently, only 10%
to 30% of the energy supplied is transformed to electricity. The electric
power can be transformed in electric motors to useful work at relatively
high efficiency without payment of further thermodynamic penalties. The
work produced eventually becomes converted to low grade heat by way
of frictional dissipation.
The use of electricity in resistance heating produces no useful work.
All the energy is converted (degraded) to heat directly. If now that heat
is used in an engine, the Carnot penalty must be paid again. If the work
is used to drive a refrigerator, the refrigerator itself sustains a further
188 Chapter 4

Carnot penalty. The use of electric resistance heating in Vuilleumier cryo-

coolers requires payment of an extra Carnot penalty over and above the
two exacted if electric motors are used to provide the work to drive a
refrigerating machine.
Clearly, energy utilization in a resistance heated Vuilleumier cryo-
cooler has to be less effective than the equivalent Stirling engine driven
by an electric motor. The electric resistance-heated duplex Stirling engine
suffers the same thermodynamic penalty. Electric resistance heating should
never even be contemplated where efficiency is a prime factor.
Most of the resistance heated Vuilleumier cryocoolers constructed to
date have been miniature engines for infrared night vision and heat-seeking
guidance applications. In these sizes, efficiency is of little significance for
the power levels are relatively low. Electric power can then be justified on
grounds of convenience and instant availability in the military vehicles,
aircraft, and missiles where such units are found.

Comparative Data
Chellis (1973, 1977) has given comparative data, reproduced in Tables
4.1 and 4.2 for a number of cryocoolers of different sizes and types. These
clearly show the weight, size, and efficiency penalties of the Vuilleumier
engines in comparison with equivalent Stirling machines.

HISTORICAL REVIEW

The earliest description of this type of machine was given by Rudolph

Vuilleumier (1918) in a U.S. patent. Vuilleumier is described as a citizen
of the United States, resident in New York. No other information is known
about him or his work. A search of the literature* has failed to produce a
single reference to a working model or any subsequent development.
The engine or a minor variant seems to have been reinvented indepen-
dently 20 years later by Vannevar Bush (1938), then a professor at the
Massachusetts Institute of Technology. Again, little is known of his work
on the concept. Later on, Professor K. W. Taconis (1951) of the Cryogenic
Laboratory at the University of Leiden in Holland described, in a U.S.
Patent, yet another form of heat-powered refrigerator with regenerative
heat exchangers. These systems and their potential for large-scale cryogenic

* I would be very glad to receive any information readers may have about the work, life, or
machines of Rudolph Vuilleumier, or about similar machines invented earlier in other parts
of the world.-G.W.
 ~
;=
i
Table 4.1. Comparison of Closed-Cycle Cryocoolersa
i:
--
Cooling Compressor Expander
capacity Input
Cryocooler power Size Wtb Size Wt MTBF
type Watts K (Watts) (in.) (Ib) (in.) (Ib) (hr) Application
Gifford-McMahon 3 20 2,100 28xl6x16 175 18x 10x6 22 20,000 Ground based
par amp
Gifford-McMahon 1 26 790 lOxlOxlO 20 11 x6x2.6 5 1,000 Airborne IR
Integral Stirling 1 26 475 14x 10x6 26 - - 250 Airborne IR
Integral Stirling 1 77 100 11 x4.5 x3 3.5 - - 250 Airborne IR
Split Stirling 1 77 135 10 X 5.5 X 2.5 5 3x2.5x2 0.3 1,000 Airborne IR
Split Vuilleumier 1 77 350 10x4x3 lOS 3 x2.5 x2 0.3 2,200 Airborne IR
a After Chellis (1977).
• Includes fan.
' Includes motor inverter and electronic temperature controller for hot end.

...
~
 ....
~

Table 4.2. Comparison of Small 77 K Refrigerators"

Smaller units Larger units

1 to 2 watts 3 to 5 watts
-
Power source 400Hz ac 28 V de 400Hz ac
Cycleh ST ST VM VM ST VM VM ST ST GM
Capacity at 77 K (W) 0.9 1.5 2.0 1.0 1.0 lc 1.0 3.5 3 5.0
Power input (W) 90 120 340 125 42 120 90 180 530 575
System weight (lb) 3.25 4.0 14 6.3 7.7 7.0 6.2 6.5 12 15
Operating speed (RPM) 1750 1750 600 600 1620 - 450 1750 - 333
Watts in/watt at 77 K 100 80 170 125 42 120 90 51 176 115
Lb weight/watt at 77 K 3.6 2.6 7 6.3 7.7 7.0 6.2 1.85 4 3

"After Chellis (1973).

h ST, Stirling; VM, Vuilleumier; GM, Gifford-McMahon.
'1 W at 85 K.

('l
=-
i
~
""
Vuilleumier Cryocoolers 191

cooling were reviewed by Yendall (1960). He concluded they were not

attractive economically for large-scale cryogenic applications.
The next significant event was the patent application by Hogan (1964)
and a paper by Chellis and Hogan (1964) (of A. D. Little, Cambridge,
Massachusetts), describing a low-temperature heat-powered Vuilleumier-
type refrigerator, for cooling in the 10-20 K range. It absorbed heat at
ambient temperature and heat was rejected from the system at 77 K to
liquid nitrogen acting as the coolant.

Hughes Vuilleumier Cryocoolers

Hughes Aircraft Co., Culver City, California, commenced work on
Vuilleumier cycle engines in the mid 1960s. A patent outlining
modifications to the Vuilleumier engine to drive its own displacers and
even produce useful work was granted to Ken Cowans (1968), then of
Hughes Aircraft. Shortly thereafter, came a voluminous report (Magee and
Doering, 1968) detailing Hughes' work on the theoretical development,
historical aspects, and laboratory evaluation of "bread board" Vuilleumier
cycle engines. It was this report that established the claim of the Vuilleumier
engine for very high reliability. They declared the engine to have the
potential for long operating life, and this reputation has been perpetuated
thereafter despite variously reported failures of the engine to actually
achieve long life (Harkless, 1973).
Other reports containing much important information about experi-
mental performance and materials used in Vuilleumier cooling engines
have been prepared by Hughes for various sponsors, particularly the U.S.
Air Force. Because no papers about the Hughes Aircraft cryocooler
developments are known to the author, it is necessary to resort to contractor
reports prepared by Hughes. Enough work has been done for others that
a good appreciation of the state of development can be gained.
Following the Magee-Doering report, Leo (1970) prepared an interest-
ing summary of Vuilleumier cycle characteristics, and later Leo (1971)
gave an expanded and updated version of the earlier Magee report. Doody
(1971) described a double-expansion Vuilleumier cycle engine for low-
temperature infrared applications. This is of interest as an early, if not the
original, application of double-expansion techniques in Vuilleumier cryo-
coolers to achieve temperatures lower than were possible with single-
expansion engines. The first use in Stirling engines of double-expansion
techniques occurred a decade earlier (Prast, 1963). Globe et al. (1971) and
Rhia (1971) presented informative material on Vuilleumier cryocoolers
for infrared missile guidance systems. Later, Rhia (1971), in three consecu-
tive reports, included design and test data for a long-life, high-capacity
Vuilleumier engine for space applications.
192 Chapter 4

About the same time Morgan (1971) reported on a comparative

study of Stirling and Vuilleumier cryocoolers to liquefy 1 to 12 kg liquid
air per hour for infrared and other sensor applications aboard bomber
planes. This document is notable as a serious attempt to compare Stirling
and Vuilleumier machines for the same cryogenic application. The Stirling
configuration was an enlarged version of the two-piston arrangement
favored by Hughes for their small Stirling cooling engines. This would not
be the "optimum" Stirling engine if, say, Philips were to design it for this
specific application. Nevertheless, in the comparison by Morgan, the Stirling
cryocooler emerged the winner on grounds of cost, size, weight, and
efficiency.
Despite this, work continued on Vuilleumier engines at Hughes. In
1973, Berry contributed to the U.S. Army Night Vision Laboratory study
(Horn, Cowans, and Berry, 1973) of low-production-cost Vuilleumier
cooling engines. Simultaneously, Berry (1973) described a new family of
modular split-Stirling and split-Vuilleumier cryocoolers. At the same time,
Renyer (1973) outlined progress on the high-capacity, long-life Vuilleumier
refrigerator shown in Fig. 4.2 and, in cross-section, in Fig. 4.3. This is a
three-stage unit capable of 12 W at 75 K, 10 W at 33 K, 3 W at 12 K with
a total power input of 2700 W. The outline design for the engine was given
by Rhia (1971) in the reports referenced above. Work on the engine over
the five year period 1970/75 was summarized by Doody (1975) with
extensive test data on the limited test accomplished.
At the other end of the scale, Berry (1974) described the design and
development of an "ultraminiature" split-Vuilleumier cooling engine
designed for a capacity of 0.75 W at 75 K of minimal size and weight
in the cold region to permit mounting in a gimbaled sensor.
Work began on a large-capacity Vuilleumier cooling engine for satellite
sensors and preliminary design aspects of this machine were reported by
Morgan (1975). An unclassified, unrestricted report for NASA was pre-
pared by Russo (1976) describing a Vuilleumier cryocooler for detector
cooling on the limb-scanning infrared radiometer. This report is
recommended for readers interested in Hughes work but unable to
gain access to the restricted Air Force reports.
One Air Force report available without restriction is that of White (1976).
It is of exceptional value to beginners. White has been involved in the Air
Force Vuilleumier program virtually from the start and is, therefore, able
to write meaningfully over the whole range of material. Substantial extracts
of this report are contained in this chapter. Elsewhere, White (1973)
has outlined a digital simulation program for the three-stage Hi-Cap
Vuilleumier cooling engine. This was developed for predicting the off-
Vuilleumier Cryocoolers 193

l.,..__

ELECTRONICS
INTERFACE UNIT

Fig. 4.2. Triple-expansion high-capacity Vuilleumier cryocooler (after Renyer, 1973).

design performance of the Hi-Cap cooler. White gave various results to

show the effect of variation on the design and operating parameters of the
engine. More recently, White (1976) gave full details of the computer
program, including a program listing, but the report is, unfortunately,
restricted in circulation. White (1976) has also reported on the preliminary
tests carried out on the high-capacity Vuilleumier cooling engines
developed by Hughes and Philips Laboratories.
 ....
'f

HOT CVUN'Dt:A lASE

OOOLIM<i.....CIUT

fiiiFfiiiGlAATOfl
ATT~MUfT
('")
=-
Fig. 4.3 . Triple-expansion high capacity Vuilleumier cryocooler (after Renyer, 1973). i...
~
Vuilleumier Cryocoolers 195

Philips Vuilleumier Cryocoolers

The development of a high-capacity Vuilleumier cooler was carried
out by the Philips Laboratories, Briarcliff Manor, New York, in parallel
with the Hi-Cap program of Hughes Aircraft Co. discussed above. The
sponsor was again the U.S. Air Force Flight Dynamics Laboratory of the
Wright-Patterson Air Force Base, Dayton, Ohio, under the "Advanced
Surveillance Technology" program.
The Philips development was reported in detail by Pitcher (1975) with
interim technical reports presented earlier (Pitcher, 1971, 1972). General
review material was presented at the 1973 Cryocooler Conference by
Lehrfeld and Pitcher (1973) and was supplemented by an excellent review
paper (Pitcher, 1973) on life limiting components and systems for spacecraft
cooling. Following construction and evaluation of a prototype high-capacity
cooler, the development of a flight design unit was undertaken. This was
reported by Pitcher (1977).
The Philips Vuilleumier engine was constructed to the same technical
specification as the Hughes Aircraft Hi-Cap cooler. This included three
stages of refrigeration, 0.3 W at 11.5 K, 10 W at 33 K, 12 W at 75 K with
a maximum power input of 2700 W, a life of 20,000 hr unattended oper-
ation, a total weight of 205 lb (including electronics but excluding the heat
rejection system) and with liquid cooling having inlet cooling temperatures
of 10 to 45°C. The Philips Vuilleumier cryocooler conforming to this
specification is shown in Fig. 4.4 and schematically in Fig. 4.5. The Philips
engine was markedly different from that of the Hughes Aircraft Hi-Cap
cooler shown in Figs. 4.2 and 4.3.
A notable feature of interest in the Philips engine was the use of the
rhombic drive mechanism (see Chapter 3), a Philips invention associated
with the development of Stirling engines. This permitted the opposed-piston
configuration, shown in Fig. 4.4, to be perfectly balanced and eliminated
the "side thrust" forces on cylinder walls found in other kinematic drives.
Another Philips Stirling engine development, the roll-sock seal, was incor-
porated in the design to hermetically seal the working spaces from the
crankcase. The hermetic seal allowed the use of oil lubrication in the
crankcase and with it an assurance of the drive mechanism to endure
20,000 hr. The low frequency of operation (7Hz) supported the expectation
of long life from the seal and drive mechanism.
Problems of oil-helium management were foreseen in long-term oper-
ation in the zero gravity environment, for the flexible diaphragms of the
rolling seals were slightly permeable to helium. Some prospect was foreseen
of a gas bubble causing the lubrication oil pump to lose suction. To prevent
this, the crankcase was entirely filled with oil, and to compensate for volume
196 Chapter 4

Fig. 4.4. Philips Laboratories triple-expansion high capacity Vuilleumier cryocooler (after
Pitcher, 1975).

changes due to the ingress and egress of displacer rods, a bellows device
was incorporated in the crankcase system.
The rigidity of the rhombic drive mechanism permitted the use of
close tolerance regenerator seals (operating at ambient temperatures) in
both the hot and cold cylinders. Rubbing surfaces were entirely eliminated
in the "dry" helium-filled spaces of the engine to further enhance the
potential for extremely long lifetime.
On reading the Philips reports, one cannot help admiring the high
level of engineering competence displayed therein. At the same time, they
convey the impression they could have done a better job with a Stirling
cryocooler. Replacement of the Vuilleumier thermal compressor (heated
electrically) by a mechanical compressor (driven electrically) would have
converted the system to a Stirling cryocooler. There is little doubt it would
have been more efficient, lighter, smaller, and cheaper, with no sacrifice-in
fact, an enhanced potential-of long life.
Vuilleumier Cryocoolers 197

3rd COLD STATION, Tc

t:li __-!1~-- 3rel EXP VOLUME

2nd COLO REGENERATOR --IRS(l~

==;!\jll---ltl COLD STATION, T 1

CE--- 1st EXP VOLUME

Itt COLO REGENERATOR ------ft~ COLO

DISPLACE A

HEAT REJECTION, T0

HOT REGENERATOR

HEAT INPUT, Th

Fig. 4.5. Schematic arrangement of Philips Laboratories triple-expansion high-capacity

Vuilleumier cryocooler (after Pitcher, 1975).

Support for this comes from the detailed failure analysis given by
Pitcher (1975). He concluded that the potential for a mechanical failure
rate was 6. 72 per 106 hours of operation and an electrical/ electronic failure
rate was 26 per 106 hours of operation. Of the total electrical/electronic
failures, 9.6 were attributed to the heater control. Elimination of the heater
would, therefore, reduce the electrical/electronic failure to 16 per 106
hours of operation. The use of a large motor for the increased work input
would reduce the prospect of motor controller failures (estimated at 9 per
106 hours of operation). At the same time, the mechanical failure rate
might increase to a balanced electrical and mechanical failure rate of 10
per 106 hours of operation.
Evidence of the more effective use of electric power in a Stirling engine
is contained in results presented by Pitcher (1975) as an appendix to the
main report. The results are reproduced in Table 4.3. It contains estimates
198 Chapter 4

Table 4.3. Estimate of Power Requirements for Multistage Stirling and Vuilleumier
Cooling Engines with Different Temperature Levels and Cooling Capacities"

Inlet
Cold Power coolant
Temperature (K) production (W) input temp.
Entry
no. 2 3 2 3 (W) (oC) Comments

1 110 30 13 5.0 5.0 0.5 2000 25 VM

2 110 30 13 5.0 5.0 0.5 1200 25 Stirling
3 112 30 11.5 5.0 8.0 0.3 2200 25 VM
4 87 30 11.5 5.0 8.0 0.3 2350 25 VM
5 75 30 11.5 5.0 10.0 0.3 2500 25 VM
6 75 30 11.5 12.0 10.0 0.3 2700 25 VM
7 75 30 11.5 20.0 10.0 0.3 2900 25 VM
8 55 12 11.2 0.9 1606 25 Stirling
9 39 12 11.2 1.0 1500 25 Stirling
10 110 55 12 13.0 7.8 0.9 1900 25 Stirling
11 110 39 12 13.0 7.8 1.0 1800 25 Stirling

"After Pitcher (1975).

made by Philips of eleven cooling engines, of Stirling and Vuilleumier

types, having varying cooling capacities at different temperature levels.
Entry numbers 1 and 2 are directly comparable in terms of temperature
and cold production. The power input to the Stirling engine (entry No. 2)
is 1200 W. The power input to the Vuilleumier engine is 66% greater at
2000W.
Entry No. 6 is the prototype Vuilleumier engine described by Pitcher
in his reports. Other important data, unfortunately not included in Table
4.3, include the pressure level of operation and the speed, volume, and
mass of the machines. So far as is known, these data have not been disclosed
by Philips.
The high-capacity cooling engine described above was by no means
the only Vuilleumier cryocooler made by Philips. Daniels and du Pre (1971)
briefly discussed the miniature engine shown in Figs. 4.6 and 4.7. This was
earlier discussed by Pitcher and du Pre (1970) and in more detail by Pitcher
(1970), prepared for the Night Vision Laboratory, Fort Belvoir, Maryland.
This report is unclassified, unrestricted, and highly recommended for
anyone interested in small Vuilleumier cooling engines.
Another interesting task performed by Philips for the Night Vision
Laboratory was reported by Cohen and Daniels (1973). This was the design
and development of a split-Vuilleumier cooling engine of 2 W cooling
capacity at 75 K in an ambient environment at a temperature of 320 K.
The engine was intended to be the self -acting type capable of producing
sufficient work to drive the displacers (patent by Cowans, 1968).
Vuilleumier Cryocoolers 199

Fig. 4 .6. Miniature prototype Vuilleumier cryocooler (after Pitcher, 1970).

Difficulty was experienced in achieving the mass and self-acting goals

for the hot side of the engine. The development did not proceed beyond
the "bread board" stage. The report is notable for the comprehensive
discussion of both the analytical and experimental aspects, the inclusion
of a computer program listing for the analytical work, the inclusion of test
results, and, above all, a remarkably competent exposition of the problem
of the cold-finger seal that besets all split-Stirling or Vuilleumier cryo-
coolers.

AiResearch Vuilleumier Cryocoolers

A Vuilleumier cryocooler was developed by the AiResearch Manufac-
turing Co., Torrance, California for the NASA-Goddard Space Flight
Center (GSFC), Greenbelt, Maryland. The objective was to achieve a long
200 Chapter 4

INCH[$
0 I
L....L_J
SCALE

Fig. 4.7. Cross section of miniature Vuilleumier cryocooler (after Pitcher, 1970).

life (2 to 5 year reliability) for a small cryocooler having a capacity of 5 W

at 75 K. Later, a miniature engine was required having a capacity of l W
at 65 K. The prototypes were heated electrically but the ultimate thermal
source was to be isotope or concentrated solar energy.
The GSFC-AiResearch development was brilliantly executed by a
competent research development team familiar with working at the frontiers
of technology. They were without a large body of previous experience in
the field and, as a consequence, were able to develop innovative approaches
without the constraints imposed by past successes. Their work has been
reported in the reports and papers referenced below.
Figure 4.8 is a cross section of one of the AiResearch engines. It is
essentially the same general configuration as the Philips opposed piston
Vuilleumier Cryocoolers 201

machine discussed earlier. However, AiResearch chose to use a crank-

connecting rod drive instead of the Philips rhombic drive. They used hard,
dry rubbing, inorganic materials for the crankshaft bearings and the linear
crosshead bearings. Noncontacting close tolerance seals were used on the
displacers.
Much effort was devoted to experimental studies of dry-rubbing
materials for bearings and seals. The bearing material selected was the
proprietary Boeing compact 6-84-1 sliding against Inconel 718 flame
sprayed with chrome carbide or tungsten carbide. Browning et at. (1973)
advise that

... selection of the Boeing material was based on wear rates, the predictability
of this material over extended period of operation and, to a lesser extent, a
low coefficient of friction. Flame-sprayed material was selected for the mating
surface because of wear rates, coefficient of friction and fabrication consider-
ations, primarily the ease of flame spraying the desired bearing mating
material.
Chrome carbide is used at the hot end of the refrigerator which may
operate at temperatures as high as 1200°F. Tungsten carbide is used for the
remainder of the bearings which operate at temperatures below 200°F.
Tungsten carbide cannot be used at the elevated temperatures of the hot end
due to the difference in thermal expansion between the tungsten carbide and
the base material Inconel 718.

AiResearch paid close attention to the development of highly effective

heat exchangers. These include the recuperative heater, cooler, and freezer
heat exchangers and the two regenerative exchangers, one for the hot side,
one for the cold side. Some of the heat exchanger configurations adopted
or evaluated are shown in Figs. 4.9 and 4.10 for the freezer, Fig. 4.11 for
the ambient temperature cooler, and Fig. 4.12 for the high-temperature
heater. Results measured on a prototype 5-W AiResearch Vuilleumier
cooling engine were given by Browning et at. (1973) and are reproduced
in Fig. 4.13. The review paper by Browning et at. (1973) discusses the
AiResearch-NASA program from the aspects of long life, thermal design,
and cold end movement.
Earlier, Gasser, Yoshikawa, and Browning (1972) presented a similar
general review of the NASA-AiResearch program. Detailed reports of
various aspects of the work were presented by Yoshikawa (1970), Browning
and Potter (1972), and Browning, Miller, and Potter (1972). The design
of a miniature Vuilleumier engine for NASA by AiResearch was described
by Miller and Potter (1973). AI Sherman (1972) of NASA GSFC discussed
the Vuilleumier cooling engines from the aspect of selected performance
characteristics, and earlier, Sherman (1971) presented a comprehensive
mathematical analysis of a Vuilleumier refrigerator.
202 Chapter 4

Sherman (1978) has presented a contemporary review of the NASA

development program relating to spacecraft cryocoolers. A progress report
of this current activity was given by Sherman et al. (1979). It is of interest
to note that further work on Vuilleumier cooling engines has been
abandoned in favor of Stirling engine developments.

RCA Vuilleumier Cryocoolers

NASA interest in Vuilleumier engines appears to have started in the
late 1960s. The earliest known manifestation of this was the ICICLE
(integrated cryogenic cooled isotope engine) study carried out for NASA
by the R.C.A. Corp. (Anon.-RCA, 1970).
Later on, Crouthamel and Shelpuk (1972) reviewed research on a
combustion-heated Vuilleumier refrigerator for Army field-portable
infrared night vision systems. The machine was designed to be self sustaining

Fig. 4.8a. AiResearch Vuilleumie r cryocooler.

 ~
f
§.
<I>
..
(")

MAGNETIC COUPLING
.:!Q
NNECTI NC.~ LINEAR BEARINGS HOT REGENERAT OR g
Q
FOR HOT DISPLACER
.."';-
COLD DIS PI A~FR
COLD REGENERATOR
COLD END HEAT
EXCHANGER

EVACUATED
SUPERINSULATION HOT D I SP LA~ F R
SUMP HEAT HOT END
COLD REGENERATOR SEAL
EXCHANGER EXCHANGER
LINEAR BEARINGS AMBIENT HEAT
FOR COLD DISPLACER PIPE ASSEMBLY

Fig. 4.8b. Cross section of AiResearch Vuilleumier cryocooler, 5 W capacity at 77 K, for NASA-Goddard Space Flight Center (after Browning
eta/., 1973).

.,.
(.>
=
204 Chapter 4

EGENERATOR
COLD EXPANSION VOLUME

a. Original Design - T = 111 K, Zero Heat Load

DISPLACER/REGENERATOR COLD EXPANSION VOLUME

b. Modification 1 - T • 74 K, 1 watt Heat Load

COLD EXPANSION VOLUME

c. Modification 2 - T 76.7 K, 3.0 watt Heat Load

Fig. 4.9 . Different freezer configurations for the AiResearch/ NASA Vuilleumier cryocooler
(after Browning eta/., 1973).

with no requirements for a drive motor. A feasibility model was constructed

and tested. The thermocatalytic combustion of 40 g/hr of propane provided
the sole energy input to produce
i. 1.6 W of refrigeration at 77 K,
ii. 16 W of internal power to operate the refrigerator,
iii. 2 W of external power to operate the external heat fan.
The combustion heated system had appreciable weight advantages com-
pared with the electrical heated system (including storage battery for say
a 12-hr mission). Total system weight for the electrical heated system was
3.1 kg for the engine plus 16.8 kg for storage batteries. The propane fuel
and container for the same mission would weigh only 0.9 kg and the burner
Vuilleumier Cryocoolers 205

CHANNEL SECTION OF HEAT EXCHANGER

OUTER COPPER PLATE
.,..,---,-r-.--T"""T---r-r4r---,-d,---,--,-,L,--,-"---,-.,-, OUTER PRESSURE WALL

~~;:;~r:;,~~~~~~~~-::FPERFORATED PLATE
FLOW DISTRIBUTOR
COLD REGENERATOR

INNER WALL COLD

; ; DOME SECTION REGENERATOR
OF HEAT
EXCHANGER

TYP 16 FLOW CHANNELS

Fig. 4.10. Preferred freezer arrangement for the AiResearch/NASA Vuilleumier cryocooler
(after Browning eta/., 1973).

system might increase the mass of the engine by 0.5 to 3.6 kg. The saving
in mass would, therefore, be (19.9 -4.5) = 15.4 kg.
Ackermann (1971) made a theoretical analysis and parametric optimi-
zation of Vuilleumier cooling engines for the Night Vision Laboratory, Fort
Belvoir, Virginia. An analytical paper for steady state operation of an
idealized Vuilleumier cooling engine was contributed by Rule and Qvale
(1969). Qvale, then an assistant professor at Purdue University and now
at the Technical University of Copenhagen, had earlier worked with
Professor Smith at the Massachusetts Institute of Technology in the
development of a simulation program for Stirling engines.
EDITOR's NoTE: The following sections-Cycle of Operation, Inher-
ent Thermodynamic and Heat Transfer Losses, VM Cooler Variations,
and VM Accessories and Components-are reproduced, with minor
editorial changes, from the unclassified United States Air Force report
by Ronald White (1976). The report was prepared by White to satisfy
206 Chapter 4

HOT lt(C(N£AATOA.
SPAC[
PR£SSUM( Vf SSH
\I'AU.

k•V IIIII li•\T

INPlll ..,.#

Fig. 4.11. Cooler arrangement for the AiResearch/ NASA Vuilleumier cryocooler (after
Browning et al., 1973).

the requirements for a Master of Mechanical Engineering degree at

the University of Dayton, Ohio.
I am most appreciative of the contribution to the public domain by
Mr. White of the material in this report and for his permission for
its reproduction here. For those wishing to study Vuilleumier engines
beyond the level presented herein, this report is recommended as
the best place to start. Considerations of space precluded reproduc-
tion of the entire report, and recourse to the original material is highly
recommended.-G.W.

CYCLE OF OPERATION

The continuous production of refrigeration at low temperature requires

the expenditure of a certain amount of energy. In most refrigeration cycles
this is supplied as mechanical energy by an electric motor. In the Vuilleumier
(VM) cycle the energy is supplied in the form of heat. Externally, the VM
refrigerator is viewed as a machine which absorbs heat at both a high and
Vuilleumier Cryocoolers 207

low temperature and then rejects this heat at some intermediate tem-
perature. The heat input at the hot end of the machine provides the energy
required to produce the cooling effect at the cold end, while the heat from
both the hot and cold ends is rejected at the ambient (intermediate) end.
The VM refrigerator, in its simplest form, is composed of two cylinders
with displacers, two thermal regenerators, three heat exchange areas,
connecting passages, and a mechanism to drive the displacers in the proper
sequence.
The displacer is a long loose-fitting pistonlike device whose function
is to move (displace) gas from one end of the cylinder to the other (see
Fig. 4.14). The cylinder-displacer combination is arranged so that there is
a gas passage connecting the two ends of the cylinder. With this connection,
the effect of moving the displacer is to displace the gas from one end of
the cylinder to the other. Note that the total volume of the gas remains
constant (assuming an infinitely thin push rod). This is in contrast to a
piston-cylinder combination that would change the gas volume. Since the
COOLING COILS OR HEAT PIPES
FLOW DISTRIBUTION PASSAGE

ALUMINUM SUMP
SUMP PRESSURE
COOLING COLLAR
VESSEL WALL

PORT TO BACKSIDE
DISTRIBUTION OF HOT DISPLACER
FLOW PATH \-----_SUMP FILLER
BLOCK

B. EXPANDED VIEW OF PLATE FIN SURFACE (SECTION AA ABOVE)

Fig. 4.12. Heater arrangement for the AiResearch/NASA Vuilleumier cryocooler (after
Browning et al., 1973).
 208 Chapter 4

75

70 /
~
/
w65
i"' / /
~ 60
"'...
/
/ TEST CONDITIONS

55
v PRESSURE 800 PSIG PEAK -
SUMP TEMPERATURE 14o" F
HOT END TEMPERATURE 1000°F Fig. 4.13. Performance characteristics
SPIEED 365 RiM of the prototype AiResearch/NASA
50
I Vuilleumier cryocooler (after Brown-
0 7

THERMAL LOAD, WATTS ing et at., 1973).

only VM that utilizes a piston is the special-case self driven VM, the term
"displacer" will be used throughout this report as a reminder that the total
gas volume of the system is not affected by the movement of the displacers.
The pressure on each end of the displacer is theoretically the same (in real
refrigerators it differs by the pressure drop through the regenerator and
connecting passages). Therefore, the energy needed to move the displacer
is very low, especially when compared to a refrigerator using pistons. Since
the two ends of the cylinder will be at significantly different tempera-
tures the displacer is long with thin walls and packed with insulation to
reduce the end-to-end heat conduction.
Next, a thermal regenerator is arranged to connect the two ends of
the cylinder so that gas passing from one end of the cylinder to the other
must pass through the regenerator. The thermal regenerator is an energy
storage device whose heat capacity greatly exceeds that of the gas. In
operation, gas passing through a regenerator is heated or cooled depending
on flow direction. For example, gas flowing from the ambient-temperature
end to the high-temperature end of a cylinder is heated by the hot regen-
erator. The energy added to the gas was stored in the regenerator packing,

DIS PLACER
GAS PASSAGE

CYLINDER

Fig. 4.14. Displacer and cylinder (after

White, 1976).
Vuilleumier Cryocoolers 209

or matrix, by gas flow in the reverse direction during a previous part of

the cycle. The regenerator matrix consists of a porous material with a high
ratio of surface area to volume, a high heat capacity, and low heat conduc-
tivity in the direction in which the gas flows through it. Examples of
regenerator matrix materials used in VM refrigerators are balls, stacked
screens, and "wool" made of (depending on temperature) stainless steel,
monel, bronze, copper, or lead. At each point in the regenerator there is
a very high heat transfer coefficient between the matrix and the gas. The
regenerator connects two regions of different temperatures, and therefore
has a temperature gradient along its matrix. At one end, the matrix has
the temperature of one region and at the other end the temperature of the
second region. Because of the presence of the regenerator, the gas can
flow back and forth between the two volumes without serious loss or gain
of heat. This thermal isolation (although not strictly true) tends to maintain
each of the volumes at nearly constant, but different, temperature levels.
The energy storage capability of the regenerator is generally great enough
that its temperature profile is only slightly altered by providing energy to
the gas.
To better understand how these components function together
to produce refrigeration, the power input section will be discussed first,
followed by the refrigeration section, and then the total refrigerator.
The power input section, the first cylinder-displacer, is composed of
the hot cylinder, hot displacer, hot regenerator, hot heat exchanger, and
part of the ambient temperature heat exchanger. Theoretically if it were
separated from the remainder of the refrigerator it would look like Fig.
4.15. For this theoretical discussion, the regenerator is assumed to be 100%

CYLINDER

REGENERATOR
MATRIX

DISPLACER
SEAL

Fig. 4.15. Power input section of Vuil-

leumier cryocooler (after White,
1976).
210 Chapter 4

efficient with no significant gas void spaces. Therefore the gas in the system
is distributed between the two (hot and ambient) active volumes. The term
"active volume" is used to describe the volume swept by the displacer.
When the displacer is in the top position the hot active volume is nearly
zero and all of the gas is in the ambient active volume. The ambient volume
is held at an ambient temperature (significantly below the hot temperature)
by the ambient heat exchanger. Using the perfect gas law, the gas is at low
temperature (volume is constant), therefore it is at a low pressure. When
the displacer is in the bottom position the ambient active volume is nearly
zero and all of the gas is in the hot active volume. Since this volume is
maintained at a high temperature (usually about 1200°F, 920 K) by addition
of heat through the hot end heat exchanger, the gas is hot and therefore
at high pressure. Since there are no valves in the system, all of the gas
throughout the system including any at the other end of the cylinder (or
in any dead volumes to be discussed later) is at the same pressure. Therefore,
it can be seen that by moving the hot displacer, the system pressure can
be varied from a maximum to a minimum in a cyclic way and that heat is
supplied to the hot end and rejected at the ambient end. The hot section
acts as a thermal compressor (whose effect on pressure is similar to a piston
in a dead ended cylinder). A similar thermal compressor was patented by
V. Bush (1939).
The manner in which the cold section, the second cylinder-displacer,
of the VM refrigerator uses the changing pressure to produce a cooling
effect is similar to that of the cold section of the Stirling cycle refrigerator,
since the method by which the cyclic pressure is produced (thermal
compressor or mechanical compressor) has little effect on the cold section.
Referring to Fig. 4.16, if the dis placer is in the top position and pressure
from the compressor is increasing, gas will flow into the ambient active
volume (the cold active volume is zero) and increase in pressure. This
compression causes the temperature of the ambient volume gas to increase,
but this heat is rejected by the ambient heat exchanger so that at the end
of the compression process the ambient active volume is filled with high-
pressure ambient temperature gas. The cold displacer is then moved to its
lower position and the gas is displaced from the ambient end, through the
cold regenerator to the cold end. As the gas passes through the cold
regenerator, it deposits its heat in the regenerator matrix and emerges at
the cold end as cold high-pressure gas (any decrease in volume of the gas
is made up by the compressor). After all the gas had been displaced to the
cold active volume, the compressor displacer is moved so that the pressure
decreases to the minimum. This decreases the pressure throughout the
refrigerator including the cold section. The gas in the cold active volume
undergoes an expansion (some of it goes back through the regenerator
Vuilleumier Cryocoolers 211

VOLUME

REGENERATOR
MATRIX
CYLINDER

GAS SUPPLIED FROM

COMPRESSOR
AMBIENT
ACTIVE VOLUME AMBIENT HEAT EXCHANGER

Fig. 4.16. Cold section of Vuilleumier --VERY THIN

cryocooler (after White, 1976). PUSH ROD

toward the compressor) becoming colder low-pressure gas. This colder gas
then has the ability to absorb heat from the device to be cooled (refriger-
ation). After the gas has absorbed the heat, the cold displacer is moved to
the upper position, displacing the gas from the cold volume through the
regenerator where it picks up the heat that was deposited on the previous
half cycle and emerges at the ambient end where it will reject the heat
picked up from the refrigeration load. After this displacement is completed,
the compressor starts to increase the pressure and the cycle begins again.
From this description it can be seen that by combining two cylinder-
displacer-regenerator assemblies and phasing their motion properly, useful
refrigeration can be obtained.
To this point, nothing has been said about what sort of mechanism
drives the displacer push rods. In the previous discussion a stop-start square
wave motion was implied to simplify the discussion of the events in each
section. In almost all of the real VM refrigerators built to date, the displacers
have been driven harmonically. Harmonic drives (such as a crankshaft with
connecting rods) are easy to fabricate and avoid the problem of high
acceleration at the end of the stroke. However, harmonic drives do compli-
cate the discussion of the VM cycle. A schematic of a harmonically
(sinusoidal) driven VM refrigerator is shown in Fig. 4.17.
The cycle operates through the use of displacers moving the gas from
one section to another without the requirement to compress the gas in a
closed volume. Therefore, the pressures throughout the system are nearly
equal at any moment. The seals shown in the schematic are to force all of
the displaced gas through the regenerators. Since the pressure drop across
the regenerators in real refrigerators is only a few pounds per square inch,
the loading on the seals and displacer drive bearings is minimal, which
212 Chapter 4

GAS-FILLED
WORKING VOLUME HEAT EXCHANGER
(TYPICALLY HELIUM ABSORBS HEAT
AT HIGH PRESSURE) FROM LOAD AT LOW
TEMPERATURE

COOLING
CYLINDER

HEAT EXCHANGER
REJECTS HEAT
TO AMBIENT

(HOT)
POWER
Q DISPLACER

CRANKSHAFT

CRANK THROW
CRANKCASE
SOURCE OF
HEAT AT
HIGH TEMPERATURE 1 - - - - - '
(-1200 "F)

Fig. 4.17. Schematic diagram of Vuilleumier cryocooler (after White, 1976).

contributes significantly to the long life of VM refrigerators. Only a small

(few watt) timing motor is needed to drive the mechanism, since the forces
are so small.
The equations to be presented in this report assume isothermal oper-
ation in the refrigerator expansion and compression volumes since such an
assumption makes it possible to derive a set of thermodynamic equations
that are relatively simple and that are fairly representative of the thermo-
dynamics of the VM refrigerator. Isothermal operation is not achieved in
an actual refrigerator, but can be approached by careful design.
As shown in Fig. 4.17, the idealized VM refrigerator has two cylinders
fitted with displacers which separate the refrigerator into three volumes
VH, VA, and Vc. The subscript H is the hot temperature level, A is the
ambient temperature level, and C is the cold temperature level. VA actually
consists of the summation of the two active volumes at the ambient end
of the two cylinders. The crankcase is fitted with filler blocks so that the
volume in the crankcase region is only the ambient active volume. It is
VuiUeumier Cryocoolers 213

Fig. 4.18. Theoretical temperature-entropy diagram for

Vuilleumier cryocooler (after White, 1976). ENTROPY- S

assumed that there are no pressure drops within the refrigerator in the
idealized model. The thermal regenerators are assumed to be perfect; i.e.,
no temperature difference is required between gas flows in each direction
in order to transfer heat, and therefore no heat flows through the regen-
erators over a complete cycle. The regenerators are assumed to have an
infinite heat capacity and therefore the temperature of the regenerator is
invariant with time. In the idealized refrigerator it is assumed that there
are no dead volumes, that is, all the volume inside the refrigerator is active
volume (VH, VA, and Vc). This means the gas volume in the regenerators,
heat exchangers, around the sides of the displacers, around crankcase parts,
at the ends of cylinders, etc. is assumed negligible (this will be modified
later). The heat conduction axially through the regenerator matrix, and
along the walls of the displacers and cylinders, is assumed negligible.
Attempts have been made to describe the cycle using a temperature-
entropy (T-S) diagram such as Fig. 4.18. It is apparent that the diagram
describes more than one unit mass of gas and the diagram is somewhat
artificial since a great many simultaneous operations occur. For an example,
let us look at three masses of gas that cycle between the ambient end and
the hot end. The mass of gas in the hot end nearest the hot regenerator
entrance will be the last mass in and (if no mixing occurs) the first mass
out. Its T-S diagram would look like Fig. 4.19a. The second mass (between
the first and last) would look like Fig. 4.19b. The third mass of gas is the
first one in the hot end and the last one out. It would have a T-S diagram
that looks like Fig. 4.19c. It can be seen that in order to use the T-S
diagram to describe this cycle a large number of individual T-S diagrams
for both ends of the refrigerator would need to be summed (mixing could
further complicate the analysis).
However, the pressure-volume (P- V) diagram for each volume
provides a clearer and more accurate picture of the operation than does
the T-S diagram.
The P- V or indicator diagrams for the three sections of a well-designed
V -M refrigerator (P 1 = P 2 = P max and P 3 = P 4 = P min) are shown in Figs.
4.20a-4.20c. Note that the ambient volume is the sum of the ambient end
volumes of the hot and cold cylinder.
214 Chapter 4

TH ) ; ,' 1,' I /
.... I 1 I
I I I I
W I I

~ //
=
1- I
I
~ I
~ TA / / I I
.... 'z ',

ENTROPY- S ENT .. OPY- t

Fig. 4.19. Temperature-entropy diagram for the mass of gas in the hot space that is (a) last
to enter, first to leave the hot space; (b) arbitrarily intermediate the first and last elements
of gas to enter or leave the hot space; and (c) first to enter, last to leave the hot space (after
White, 1976).

At each volume for a complete cycle, the first law reduced to

Q = (hz- h 1) + W
where Q and W represent energy per unit mass of gas flowing into or out
of the active volume. The temperature of the gas crossing the boundaries
does not vary with time (since ideal regenerators were assumed), so h 2 = h 1
and Q = W or

The heat flow is equal to the work at that volume. There is, of course,
no net work resulting from the idealized refrigerator because no pressure
differences exist within the machine at any given time. Since the pressure
at each end of each displacer is equal and by elementary geometry the
change of volume at one end of a cylinder is exactly equal to the negative
of the volume change at the other end, we have

f PA dVA = f PHdVH +f PcdVc

Since the displacer motion has been assumed harmonic, the pressure
changes are assumed to occur isothermally, and if the phase separation
angle is assumed to be 90°, the volume at each of the ends can be described

(a)PMAX
1~
12
~ I
I I
I!' I
ijl I
"'~ '
I
PMIN - - ,41 - 3

VOLUME-V ---+-

Fig. 4.20. Pressure-volume (indicator) diagram for the (a) cold volume; (b) hot volume, and
(c) ambient volume of a Vuilleumier cryocooler (after White, 1976).
Vuilleumier Cryocoolers 215

by the maximum active hot and cold volumes:

VH = tvHM(1-cos 8) (4.1)
Vc = iVcM(1-sin 8) (4.2)
VA= tvHM(1 +cos 8)+iVcM(1 +sin 8) (4.3)
Here V HM is the maximum swept volume of the hot end and V eM is the
maximum swept volume of the cold end. ·
If the gas is assumed to be ideal, n = pV/RT and
p{ VHM(1-cos 8) VcM(1-sin 8) VHM(1 +cos 8)
n =- + + ---"'=-'-----'-
R 2TH 2Tc 2TA
VcM(1 +sin 8)}
(4.4)
+ 2TA
rearranging, the equation becomes

n= ~ [ VHM( 2 ~H + 2 ~J + Vc~ 2 ~c + 2 ~J
+ VH~ 2 ~A- 2 ~H) COS 8+ Vc~ 2 ~A- 2 ~J sin 8]
multiplying numerator and denominator by 2TA the equation becomes

n = 2 ~R [ VH~~;+1)+ VcM(~;+1)
+ V H~ 1 - ~;) cos () + V c~ 1 - ~;) sin 8J (4.5)

This equation can be simplified by grouping the parameters, using the

symbols a, b, and c defined by

a= VH~ 1+~;)+vc~ 1+~;) (4.6)

b = VHM( 1- ~;) (4.7)

c=Vc~1-~;) (4.8)

Where a reflects the charge of gas in the refrigerator, b reflects the effect
of the movement of the hot end displacer on the total pressure, and c
reflects the effect of the cold-end displacer on the pressure.
216 Chapter 4

The pressure, P, at any instant and at every point in the working

medium is 2nRTA
p = (4.9)
a + b cos 0 + c sin 0

To obtain the angle 0 at which P is maximum or minimum, differentiate

the equation and set dP/dO =0, and then tan 0 =c/b. Using Eq. (4.9) to
calculate the area inside the P- V diagrams:
2
QH=,(PdVH=nRTAVHMf b sinO . dO1T (4.10)
j 0 a+ cos 0 +c sm 0

Oc f
= P dVc = nRTA V CM f2 1T

o a
b
cos 0
.
+ cos 0 + c sm 0
dO (4.11)

where QH is the heat input to the hot end and Oc is the heat input to the
cold end.
These equations can be integrated using formula 2.558-2. from Table
of Integrals Series and Products by I. S. Gradshteyn and I. M. Ryshik. Note
that a 2 >(b 2 +c 2) and that JdO/(a +b cos 0 +c sin 0) must be integrated
from 0 to 7T and from 7T to 27T with -'TT/2 :s; arctan 0 :s; 7T/2. The result is

OH
27TcnR
= b2+c2
[ a
VHMTA 1- (a2-b2-c2)1/2
J (4.12)

27TbnR [
Oc =- b2+c2 VcMTA 1
a
(a2-b2-c2)1/2
J (4.13)

These are the equations for the heat input to the hot end and the heat
absorbed by the cold end (per cycle) for an ideal VM refrigerator.
The relationship between the maximum hot active volume and the
maximum cold active volume is
VHM TH(TA- Tc)
(4.14)
VcM Tc(TH-TA)
This equation is derived from the ideal work and the ideal refrigeration
that takes place in the machine during a cycle, when the pressure at crank
position 1 is assumed equal to the pressure at crank position 2 as indicated
in Fig. 4.17.
The coefficient of performance of this ideal refrigerator is

(4.15)
Vuilleumier Cryocoolers 217

which is the same as a Carnot engine, driving a Carnot refrigerator,

therefore the figure of merit (FOM) is unity.
The maximum pressure ratio is obtained by differentiating the pressure
equation:
Prnax a +(b 2 +c 2 ) 112
(4.16)
Prnin =a -(b 2 +c 2 ) 112
In order to better understand this ideal refrigerator an example was
calculated and plotted in Fig. 4.21. The assumptions were as follows:
Temperature of hot end= 812 K
Temperature of ambient end= 366 K
Temperature of cold end= 77 K
VHM is 87.23% of the total volume
VcM is 12.77% of the total volume
Phase angle is 90°
() = 0 at hot displacer top dead center
The volumes VH, VA, and Vc were plotted to show how they vary with
crank position, e. The percentage of the total number of molecules NH,
NA, Nc, in each of the volumes vs. crank position was also plotted, as was
the resulting pressure (in arbitrary units).
From Fig. 4.21, it can be seen that although the maximum cold volume
is only 12.77% of the total, at one point of the cycle 52% of the molecules
of the system are in the cold end. Also the maximum percentage of
molecules ever in the hot end is 55%. The maximum pressure occurs 45°
before the maximum hot volume occurs and the minimum pressure occurs
45° before the minimum hot volume occurs. Noting the shape of the pressure
curve in the region of 180° to 270°, it can be seen that the pressure curve
is not quite sinusoidal.
All of the above equations and discussion assumed an ideal
refrigerator; now the various nonideal factors will be added to these
equations until a real refrigerator is discussed.
There are a number of void volumes in a real refrigerator. These
include clearances around the displacers to prevent scraping, clearances at
the ends of the cylinders to prevent hitting the end and to allow for thermal
expansion, gas flow spaces in the regenerators, heat exchangers, gas transfer
passages, and clearances in and around the mechanism. The void volumes
must undergo the pressure variations of the cycle but do not contribute to
the cooling effect. The void volumes decrease the maximum pressure ratio
produced and, therefore, the amount of cooling produced.
To illustrate the magnitude of some of these void volumes, the cold
regenerator flow passage void volume is usually 1.5 to 3.5 times the active
218 Chapter 4

9- DEGREES

.
400:

..
300 ~

zoo ~
.
0

'00

Fig. 4.21. Pressure-volume interactions in a

ti-DEGIHf.S Vuilleumier cryocooler (after White, 1976).

cold volume, while the displacer side clearance void volume is usually about
0.17 times the active cold volume. The effect of void volumes is related to
the temperature of the void volume. The void volumes at the coldest
temperatures have the greatest effect on the refrigerator performance while
the void volumes at the hot end have the least effect on refrigerator
performance. The effect of void volumes is more pronounced in miniature
VM refrigerators, due in part to ordinary manufacturing tolerances. In
miniature VM refrigerators, decreasing the void volume is both essential
and expensive. Reduction of void volume requires close tolerances and
unusually shaped parts (especially in the crankcase). Void volume reduction
techniques used in current refrigerators even include the use of epoxy to
fill the screw slots in the internal screws.
The effect of void volume can be accounted for by adding terms to
Eq. (4.4) of the form:

where Vvl is a void volume and Tv 1 is the temperature of that void volume,
Vuilleumier Cryocoolers 219

or Vv!Tv where Vv is the sum of void volumes and

The addition of void volumes will change Eq. (4.5) to

a= VHM{ 1+ ~:}+ VcM{ 1+ ~;}+2Vv{~:} (4.17)

The other equations (4.9)-(4.16) remain the same except the numerical
value of a has changed. Typical values of the reduced void volume ratio
(reduced void volume to cold end volume) for small VM refrigerators built
to date are in the range of 1.5 to 3.7, where reduced void volume is defined
as
Tc
Vv reduced = Vv actual X -T-- (4.18)
v actual

Another correction needed for the basic VM equations is the compress-

ibility factor for the working fluid, which is helium. Helium has been used
as the working fluid in all VM refrigerators built to date since it behaves
as a nearly perfect gas and has the lowest temperature capability. The
effect of compressibility of helium is slight until temperatures below 70 K
are reached. Adding the compressibility factor (Z) to Eq. (4.4) gives
P [VHM(1-cos 8) VcM(1-sin 8) VHM(l +cos 8)
n =- + + ----=-=-'------'-
2R THZH TcZc TAZA
VcM(1 +sin 8) 2 Vv]
+ +-- (4.19)
TAZA TvZv
which changes Eqs. (4.6) or (4.17), (4.7), (4.8), and (4.14) to

a= VHM(ZH + TA) + VcM(Zc + TA) + 2 VvTA (4.20)

ZH ZA TH Zc ZA Tc ZvTv

b = VHM(ZH_ TA) (4.21)

ZH ZA TH

c = VcM(Zc _ TA) (4.22)

Zc ZA Tc

VHM = THZH (TAZA- TcZc) (4.23)

VcM THZH- TAZA TcZc

(4.24)
220 Chapter 4

Equations (4.9)-(4.13) and (4.16) are still valid, however, the values of a,
b, and c within these equations have changed slightly.
To summarize the theoretical VM section, if compressibility effects
and void volumes are to be accounted for, the equations to use are
(4.9)-(4.13), (4.16), and (4.18)-(4.24). Equation (4.13), Oc, is the gross
refrigeration produced by the refrigerator. To obtain the net refrigeration
produced by the refrigerator, all of the various cold end losses must be
subtracted from the gross refrigeration. Equation (4.12), QH, is the P- V
heat input to the hot cylinder. To obtain the actual heat (power) input
needed, the various hot end losses must be added to the P- V input.

INHERENT THERMODYN AMIC AND

HEAT TRANSFER LOSSES

Since it is impossible to build perfect regenerators and to eliminate

all undesirable heat transfer processes in the refrigerator, these losses must
be subtracted from the gross refrigeration to determine the net refrigeration
available at the cold end. Similar losses must be added to the heat input
to the gas to determine the required heat input to the hot end.

Shuttle Loss
Shuttle loss is caused by the mismatch of thermal gradients between
the displacer and the cylinder. The cold cylinder wall is at ambient tem-
perature at one end and at cryogenic temperature at the other. It has a
fixed length and a gradient from warm to cold that is approximately linear.
The displacer is shorter than the cylinder wall by the length of the stroke,
however, it has the same temperature extremes: warm at one end and cold
at the other. When the displacer is at one extreme of its travel the tem-
perature gradients are somewhat mismatched. As it passes through its
stroke and reaches the other extreme the gradients are again mismatched,
but now they are mismatched in the opposite direction so that the displacer
picks up heat from the cylinder when it is at the warm end and it gives off
heat to the cylinder when it is at the cold end of its stroke. Hence, there
is a picking up of heat at the warm end, a shuttling of the displacer to the
cold end, where it drops off the energy to the cylinder. It is thus termed
a shuttle loss.
If the motion of the displacer is approximately harmonic and if the
thermal time lag of the cylinder and displacer materials is small compared
with the reciprocating time, the shuttle heat transfer can be computed from
Vuilleumier Cryocoolers 221

the following equation (Leo, 1971):

Q =0.186Y2Ckg (Tw-Tc) (4.25)

SH S Ley
where OsH is the shuttle heat loss, Y is the stroke, C is the wetted
perimeter (7T x diameter), S is the radial clearance between displacer and
cylinder, Tw is the temperature of the warm end of the cylinder, Tc is the
temperature of cool end of cylinder, Ley is the length of cylinder, kg is the
thermal conductivity of the gas (helium). The terms "warm" and "cool"
are used since this equation is used to compute the shuttle losses of both
the hot and the cold ends of VM refrigerators.

Pumping Loss
The pumping loss is due to the fixed clearance volume that must exist
between the displacer and the cylinder so that the displacer can move
without rubbing. This volume is bounded on one end (usually the ambient
end) by a displacer seal and is open on the other end. Because this volume
around the displacer allowed for running clearance is fixed, the mass of
gas in this volume at any one time is proportional to the (cycling) pressure.
At the minimum pressure point there is a minimum mass of gas in this
volume. As the pressure increases toward the maximum pressure point,
gas flows into this volume proportional to the increasing pressure. There-
fore, using the cold cylinder as an example, cold gas flows into this volume
and down into the warm areas where it picks up energy. Then as the gas
pressure falls from the maximum back to the minimum pressure, some
warm gas flows out of this volume into the cold regions. This causes a loss
of refrigeration called the pumping loss. This loss occurs on both the cold
end and the hot end of VM refrigerators when displacer seals are used
to force gas flow through the regenerators. The following equation for
pumping loss was derived by Leo (1971):

2(7TDc) 0 "6 Ley(Pmax- Pmin)1. 6N1. 6 C~· 6 (Tw- Tc)S 2 "6

(4.26)
Qpu = 1.5ZR 1. 6 k~· 6 [(Tw + Tc)/2t 6

where Qpu is the pumping loss, De is the diameter of the cylinder, Ley is
the length of the cylinder' p max is the maximum pressure, p min is the
minimum pressure, N is the cycle speed, Cv is the specific heat of the gas,
T w is the temperature at the warm end, Tc is the temperature at the cool
end, S is the radial clearance between displacer and cylinder, Z is
the compressibility factor for gas, R is the gas constant, and kg is the
conductivity of the gas.
222 Chapter 4

By examining Eqs. (4.25) and (4.26) it can be seen that there is an

optimum gap (clearance) for best performance (lowest losses). If the gap
is too small the shuttle loss will be high and if the gap is too large the
pumping loss will be high. Hence, for a fixed diameter and stroke, there
is an optimum gap for least losses and there is much to be gained by
maintaining the gap accurately. To accomplish this the materials of the
cylinder and displacer are selected so that their coefficients of thermal
expansion are matched.

Heat Transfer through Displacer

This loss is the heat transfer through the displacer due to the difference
in the temperatures of the ends of the displacer. It is expressed as

Q = koAv (T _T ) (4.27)
D Lv w c

or if the displacer is hollow

Q _7Tko(D~v-Dio)(Tw-Tc) (4.28)
v- 4Lo

where 0 0 is the conduction heat loss through the displacer, k0 is the

conductivity of the displacer material, Av is the area of the end of the
displacer, L 0 is the length of the displacer, Tw is the temperature of the
warm end, Tc is the temperature of the cool end, D 00 is the outside
diameter of the displacer, and Dm is the inside diameter of the displacer.

Heat Transfer through Cylinder Wall

This loss is the heat transfer due to the difference in the temperatures
of the ends of the cylinder. If the inside and outside diameters of the
cylinder are known Eq. (4.28), with the appropriate conductivity and
dimensions, can be used. If the refrigerator is being designed and the
thickness of the cylinder wall is not known the following equations can be
used:
Q _ 7TkcyPmaxD~y(Tw- Tc)
(4.29)
ey- 2uLcy[(1- Pmax)/20']

where OeY is the conduction heat loss through the cylinder, key is the
conductivity of the cylinder material, Pmax is the maximum pressure, Dey
is the inside diameter of the cylinder, Ley is the length of the cylinder,
and u is the allowable stress.
Vuilleumier Cryocoolers 223

Heat Generated by Friction between Displacer and Cylinder

This loss is due to friction between the cylinder wall and the displacer
riders and seals:

where F 1 is the friction force in the direction of travel and Y is the stroke.
It is rather difficult to evaluate this loss since much must be known about
the seal forces, rider loads, and the locations at which this loss is generated.

Regenerator Losses
The regenerators of the VM refrigerator must be very efficient if the
refrigerator is to be a practical device for producing refrigeration at tem-
peratures below 100 K. In fact, the overall performance of the machine
depends directly on the efficiency of the regenerators. The regenerator
losses can be divided into two types:
Heat Load Due to Friction in the Regenerator. This refrigeration loss
is caused by aerodynamic heating when the working medium flows through
the regenerator:

Q _ !!.Preg(Pmax+ P min) V CMN

(4.30)
RF- pJZRTc
where:

(4.31)

or
_ cPmax+Pmin) 3 V~MN 3[LR
Q RF- 3 3 3 2 2 (4.32)
2Z R T cgcAregP D.J
where QRF is the loss due to aerodynamic friction in the regenerator, Pmax
is the maximum system pressure, P min is the minimum system pressure,
VcM is the maximum displaced cold volume (if a hot regenerator is being
calculated this would be maximum displaced hot volume), Tc is the tem-
perature of the displaced volume (hot or cold), N is the cycle speed, f is
the friction factor, LR is the length of the regenerator, Z is the compressibil-
ity factor for the working fluid, R is the gas constant for the working fluid,
Areg is the cross-sectional area of the regenerator, p is the density of the
working fluid, D. is the equivalent hydraulic diameter, gc is the gravitational
conversion factor, and J is the conversion factor.
224 Chapter 4

Heat Load Due to Limiting Value of Film Coefficient in Regenerator.

The limiting value of the film coefficient in the regenerator prevents
sufficient cooling of the working fluid as it flows from the hot to the cold
end of the regenerator. Leo (1971) discussed in general the derivation of
the following equations that evaluate the amount of energy that remains
in the working fluid because of this limiting value of film coefficient:

Q =( 1 -N )C (T -T)(Pmax+Pmin)VcMN (4.33)
RH R P w c 2RTcZ

where
h'h"L~
(4.34)

Where QRH is the loss due to the limiting value of the film coefficient in
the regenerator, NR is the regenerator efficiency, Cv is the specific heat of
the working fluid, (Tw- Tc) is the temperature difference across the regen-
erator, VcM is the maximum displaced cold volume (if a hot regenerator
is being calculated this would be the maximum displaced hot volume), Tc
is the temperatare of the displaced volume (hot or cold), N is the cycle
speed, R is the gas constant of the working fluid, Z is the compressibility
factor, h' is the heat transfer coefficient-flow out, h" is the heat transfer
coefficient-flow in, LR is the length of the regenerator, G is the mass
velocity (mass flow rate per unit cross-sectional area), and rh is the hydraulic
radius. The value of NR for cold regenerators is usually in the range
from 0.995 to almost 1.0 and for hot regenerators is in the range of 0.9
to 1.0.
These equations for the regenerator losses are only approximate since
they are based on average mass flow rates rather than instantaneous mass
flow rates. Another way to compute the performance of the VM refrigerator
was partially outlined by Globe et al. (1971). It involves cutting the VM
refrigerator into a large number of control volumes and determining the
mass flow rate into and out of each of these control volumes as a function
of crank position. The losses, especially the regenerator and pumping losses,
can be calculated and their effect on the pressure can be iterated with the
mass flow equations so that a more accurate description of the refrigeration
at every crank position is obtained. This can be done for an existing
refrigerator design with the aid of a computer but is extremely difficult to
do when optimizing a new design since the control volumes themselves are
being changed during the optimization process. Therefore, the equations
presented above or similar equations are usually used for optimizing new
designs.
Vuilleumier Cryocoolers 225

Net Refrigeration
To obtain the net refrigeration, the cold end losses for each stage are
summed and subtracted from the gross refrigeration for that stage. In
multistage refrigerators some of the losses from colder stages appear as
increased refrigeration at warmer refrigeration stages. These include con-
duction losses, regeneration heat loss, shuttle loss, and pumping loss. These
losses from a colder stage are called interstage heat flow and should be
added to the gross refrigeration of the warmer stage. Other cold end losses
that were not discussed here but might be applicable depending on
refrigerator design are regenerator conduction, insulation, seal leakage
losses, and heat leaks down instrumentation leads.

Total Heater Power Input

To calculate the heater power required at the hot end, the hot end
losses are summed and added to the heat input to the gas. Losses not
discussed above that should be included depending on the actual design
are the hot end insulation loss, conduction losses down heater and
instrumentation leads, and heat leaks through the insulation inside the
hollow hot displacer.

VM COOLER VARIATIONS

Multistage VM Coolers
In many applications there is a need to produce useful refrigeration
at more than one temperature level at the same time. An example of this
is cooling an electronic device to a very low temperature while cooling a
Dewar heat shield surrounding the device to an intermediate temperature
to intercept the Dewar heat leaks. This saves considerable power since the
majority of the heat is removed at the higher temperature for much less
input power (and refrigerator size) than if all the heat had been removed
at the lower temperature.
Two mechanisms have been used to add additional cold stages to VM
refrigerators. The parallel cylinder system (Fig. 4.22) adds another smaller
diameter cold cylinder parallel to the first cold cylinder. In this configuration
the displacer seals are at the ambient temperature region of the displacers
(an advantage) but fabrication of the cylinder assembly is more difficult.
Keeping the two cylinders straight and parallel from brazing temperature
down through cryogenic temperature is difficult and expensive.
226 Chapter 4

FIRST STAGE ACTIVE

YO LUliE

FIRST STAGE HEAT

EXCHANGER

AMBIENT HEAT EXCHANGER

Fig. 4.22. Multiple stage Vuilleumier cryocooler arrangement with parallel cylinders (after
White, 1976).

The other configuration is the series cold cylinder configuration (Fig.

4.23). This configuration is easier to fabricate but requires a displacer seal
at the base of the second stage displacer that seals properly at the first
stage (cryogenic) temperature.
Adding additional cold stages affects the previously developed
equations in the following ways. Additional cold stage factors (second term)
are added to Eq. (4.20). Additional terms are added to Eq. (4.22) and
additional equations similar to Eq. (4.13) are written to describe the gross
refrigeration of the additional stages. The ratio between the cold active

2nd STAGE HEAT EXCtiANGER

2nd STAGE ~
ACTIVE vot.lME

2nd STAGE ----..l.l-

OISPLACER

11 t STAGE
REGENERATOR
tlf STAGE
OISPLACER

I ----AMBIENT HEAT EXCHANGER

II Fig. 4.23. Multiple-stage Vuilleumier

I
cryocooler arrangement with tandem
I - PUSH ROO
cylinders (after White, 1976).
Vuilleumier Cryocoolers 227

volumes of a two-stage refrigerator is obtained by taking the ratio of the

gross refrigeration equations [similar to Eq. (4.13)]. The result is
Oc, VcM,
--=-- (4.35)

Phase Angle
Phase angles other than 90° were investigated by E. B. Qvale and T.
T. Rule (1968) and by B. Leo (1971). Phase angles other than 90° complicate
the Eqs. (4.1)-(4.13) considerably and make the fabrication of parts more
difficult. The optimum phase angle is a function of the active swept volumes
of the hot and cold cylinders and of the thermal boundary conditions. The
investigations showed that for refrigerators with heat rejection tem-
peratures near room temperature, the optimum phase angle would be in
the range from 90° to 102° depending on cold end temperature (the lower
the temperature the greater the angle). Multistage refrigerators further
complicate these relationships and reduce the range of optimum phase
angles. Since the gross refrigeration is changing very slowly with respect
to phase angle near the optimum, most manufacturers are using a 90° phase
angle.

Similar Cycles
There are other heat powered refrigeration cycles similar to the VM
cycle. One by Bush (1938) is quite similar to the VM except that the two
ambient volumes (one at the ambient end of the cold cylinder and one at
the ambient end of the hot cylinder) are separated by a thermal regenerator
and reject heat to heat sinks at different temperatures. Another heat-
powered refrigerator was patented by Taconis (1951). It differs from the
VM in the timing of the movements (three instead of four motions) of the
displacers. Another heat-powered refrigerator was patented by Hogan
(1964). It produced cooling in the 10-20 K range, while the hot end
absorbed heat at room temperature and the heat rejection was at 77 K
(the heat was rejected to liquid nitrogen). A patent by Cowans (1968)
describes a modification to the VM refrigerator that allows it to drive its
own dis placers and produce useful shaft power. This is done by increasing
the cross-sectional area of either the hot displacer connecting rod or both
connecting rods, so that with the addition of connecting rod seals, and by
lowering the crankcase pressure below the minimum pressure in the VM
cycle, a net force can be created to drive the refrigerator. This has the
advantage that the small timing motor used on most VM refrigerators is
228 Chapter 4

not needed (but something must give it a shove to get it started). However,
this adds the life-limiting problem of dynamic connecting rod seals that
must be able to seal against the full cycle pressure (several hundred pounds
per square inch). This type of sealing problem is avoided by most VM
refrigerators since in a "pure" VM cycle the only dynamic seals in the
system are the displacer seals. Displacer seals usually experience very small
pressure differences of 5 to 15 psi which contributes to their very long life.

VM ACCESSORIES AND COMPONENTS

Hot End Temperature Controller

One important accessory required by electrically heated VM
refrigerators is the hot end temperature controller. By examining the
VM theoretical equations, it can be seen that the higher the temperature
of the hot end, the higher the efficiency of the refrigerator. VM refrigerators
are usually designed to operate at the highest temperature possible,
consistent with metallurgical limits. The most popular hot end material
is Inconel 718. The strength of this material falls off quite rapidly
above 1250°F, so VM refrigerators are usually designed to operate at
about 1200°F. However, there are several problems with trying to operate
at l200°F. A change of input voltage to the heater can change the heater
power and the hot end temperature. Aircraft power supply voltages can
vary as much as ±~ of the mean voltage. In addition, the ambient heat
rejection temperature aboard an aircraft can vary as much as 200°F, which
will affect the power requirement and therefore the hot end temperature.
Cold end load changes also have an effect on the power required and the
hot end temperature. Refrigerator malfunctions such as loss of working
fluid or a stalled motor prevent the working fluid from absorbing sufficient
heat and cause the hot end to overheat. The heater is usually sized to
supply the correct power at the minimum voltage and maximum ambient
temperature. To prevent hot end overheating problems, a hot end tem-
perature controller is used.
Proportional controllers are usually used since maintaining the heater
at a nearly uniform temperature reduces the heater stresses and improves
the life of this component. Frequently a simple on-off controller, set at a
higher temperature, backs up the primary controller as an additional safety
measure. A variety of controller concepts have been used. One of the most
popular is pulse width modulation, due to its high efficiency. However, this
type controller requires considerable filtering and shielding to prevent the
electromagnetic interference it creates from affecting nearby equipment.
Other concepts include linear proportional control, zero voltage (ac) switch-
Vuilleumier Cryocoolers 229

ing, slow on-off switching, and mechanical devices such as Curie point
switches and vapor bulb thermometers. On large VM refrigerators the
controller problem is reduced by calculating both the minimum power and
the maximum power required. The minimum power is then supplied by a
large heater with a simple on-off controller for malfunction protection
only. The difference between the minimum and maximum power is supplied
by a smaller heater with a proportional controller and necessary shielding.
This arrangement is more efficient and reduces the size and weight of the
controller and electronic filters.

Heaters
Two types of electric hot end heaters have been used in VM
refrigerators. The furnace type is a ribbon of heater wire wrapped on a
ceramic mandrel and held in place with cement. The ceramic furnace
surrounds the hot end of the hot cylinder and transfers heat by radiation.
The Calrod-type sheathed heater has been the most popular. Heat is
transferred either by radiation (Yoshikawa, 1970) or by brazing the heater
sheath directly to the hot cylinder. Since the watt density of the heaters
required by most VM designs is very high (for this type of heater) the
heater must be properly heat sunk to the hot cylinder or burnouts will
occur. Care also must be taken to be sure that the active (heat-producing)
portion of the heater terminates while still thermally connected to the hot
cylinder. The larger-diameter low-resistance lead-in wire that runs between
the heater wire and the terminal (inside the sheath) must be of a material
that does not embrittle or corrode when exposed to insulation or atmos-
pheric contaminants. Single-ended straight wire heaters have caused
numerous failures and have been largely abandoned in favor of two-ended
helically wound single wire heaters. Straight wire heaters are available in
smaller sheath diameters but must use smaller-diameter heater wire since
the total length of the heater wire is less. This, coupled with the possible
nonuniform reduction of heater wire diameter during the swaging of the
heater sheath and higher stresses during heater cycling have contributed
to numerous straight wire heater failures. The helical single-wire heaters
have larger heater sheaths (less convenient for the refrigerator designer)
but have a larger-diameter longer heater wire that does not change cross-
section (the helix angle changes) during the swaging of the heater sheath.
The heater wire is closer to the sheath (less temperature drop) and is less
sensitive to thermal cycling. These heaters are much more reliable than
straight wire heaters. In critical applications additional redundant heaters
are added to avoid scrapping an expensive hot cylinder assembly due to a
burned out heater.
230 Chapter 4

Motors
As mentioned earlier, ac induction motors with the rotor inside the
helium space and the stator with its windings outside the helium space
have been successfully used in a large number of applications. These motors
are either two or three phase and are purchased with a matching inverter.
Total efficiency for the motor and inverter is about 25%. In small VM
refrigerators, the motor power is a small fraction of the total power,
therefore cleanliness and reliability are more important than efficiency. In
a few applications an inverter could not be used (due to space or ambient
temperature problems) and since life was less critical a de brush-type motor
was used. Special brush materials were used along with special commutator
coatings. The motor windings were potted to reduce the generation of
contaminates. Very few data have been gathered on this motor, so its
limitations are still unknown. A brushless de motor is being used on one
refrigerator. The motor efficiency is expected to be at least 55%. It will
be in the oil-filled crankcase of the refrigerator and should present no
contamination problem.

Regenerators
The cold regenerators have been previously discussed and are discussed
in considerable detail in all of the references. The term "internal regen-
erator" is used for regenerators inside the displacer and the term "external
regenerator" is used for regenerators attached to or a part of the cylinder.
Cold regenerator matrix materials are usually screens of 100 mesh to 500
mesh in copper alloys or stainless steel and balls of monel or lead in sizes
down to 0.002 in. Lead balls are usually used for temperatures below 50 K
since lead is one of the few materials with appreciable specific heat at these
temperatures.
The hot regenerators are described as internal or external als0,
however, the forms of this matrix are more varied. Stacked screens, balls,
tubes, and the annulus have been used for the hot regenerator matrix. The
internal annular regenerator is composed of the walls of the displacer and
the cylinder. In this configuration the gas flows between the displacer and
the cylinder. A displacer seal is not used. This eliminates one wearing part
(the seal) and eliminates the pumping loss, but makes the radial location
of the displacer within the cylinder very critical, which in turn makes hot
rider ring wear extremely critical. As an example, if the regenerator is
designed with a 0.007-in. radial gap between the displacer and cylinder,
and if rider wear allows the displacer to be out of concentricity by 0.002 in.,
15% of the gas flow is on the narrow side of the regenerator while 85%
Vuilleumier Cryocoolers 231

of the flow is on the wide side. This causes the regenerator loss to be
doubled (Leo, 1971). Since lubrication and wear of the hot rider is a serious
problem the internal annular regenerator is rarely used anymore. The
external annular regenerator is composed of the walls of the cylinder and
one or more liner sleeves. The liner sleeves must be very thin and
concentric with the cylinder. A displacer seal riding on the sleeve assures
gas flow through the regenerator. Concentricity is a problem with this
regenerator also.
The tubular regenerator is composed of small-diameter thin wall tubes
constrained (usually in a single layer) between the cylinder (or displacer)
and a cylindrical liner with the tube axis in the direction of the cylinder
axis. The gas flows in the axial direction either through the tubes or in the
triangular-shaped spaces between the tubes and the liner (or cylinder). This
type of hot regenerator is easy to fabricate and is being used in several
VM refrigerators. The screen and ball regenerators are similar to the cold
regenerators except matrix elements are larger and of materials such as
monel and stainless steel.

Heat Rejection
Heat rejection has been accomplished in several ways. These include
rejection to forced ambient air (Globe et at., 1971), rejection to a pumped
liquid which in turn rejects the heat to air or a radiator (Doody, 1971),
and rejection by heat pipes (Yoshikawa, 1970; Browning and Potter, 1972;
AiResearch Mfg. Co., 1972, 197 4; Browning et at., 1972; Miller and Potter,
1973, 1974).

SPLIT-VUILLEUMIER CRYOCOOLER

In the above discussion of Vuilleumier cryocoolers, White considered

only integral units, particularly high-capacity units intended for spacecraft.
The omission of split-Vuilleumier cryocoolers was surprising for they are
an important class of miniature cooler much used in military applications.
Figure 4.24 shows a split-Vuilleumier cryocooler manufactured by
Cryogenic Technology Inc. Table 4.4 shows the specifications of the same
unit designated Model No. VM-1 and described as having the following
advantages:
i. The U.S. Air Force standard cryogenic cooler;
ii. Highest MTBF (mean time before failure) of existing coolers of
similar size and capacity;
232 Chapter 4

Fig. 4.24. Photograph of split-Vuilleumier cryocooler, Model No. VM-1. Capacity 1 W at

77 K (courtesy Cryogenic Technology, Inc.).

iii. Shock and vibration resistance ratings higher than any other
existing coolers;
iv. Split configuration for remote cooling;
v. Low operating cost over a long in-service life;
vi. Simplified maintenance and service requirements.
Split coolers are attractive for applications where minimum vibration and
mass are important as in gimbaled, swiveling infrared sensors, or missile
sights. The compressor and expander units are coupled by a thin flexible
tube permitting independent motion of the expander cylinder. The void
volume of the connecting tube increases the dead volume and thereby
Vuilleumier Cryocoolers 233

Table 4.4. Specifications for CTI Model No. VM-1 Split-Vuilleumier Cooling
Enginesa

Refrigeration types Split-Vuilleumier cycle

Refrigeration capacity 1.0 W @ 77 K (25°C ambient)
0.7 W@ 77 K (55°C ambient)
Cooldown time to 85 K <10 min (55°C ambient)
Weight 10.0 lb (4.54 kg)
Power requirements 115 Vac, 3c/J, 400 Hz, 370 W
Ambient temperature limits Operating: -54°C to 55°C
Storage: -54°C to 71°C
Scheduled maintenance 2200 operating hours
Military specifications Meets applicable U.S. Military specifications

" From CTI Technical Specification Sheet.

decreases the amplitude of pressure fluctuations with deleterious effects

on the cooling capacity.
The regenerative displacer of the expander may be activated by an
electric motor or solenoid or by fluidic control. Electric drive increases the
size and weight of the expander unit but is a reliable system for maintaining
proper synchronization of the compressor and expander displacers. Fluidic
control is attractive to eliminate the complexities of the additional
electronics and reduce size and weight. However, the same problems of
maintaining displacer synchronization exist in Vuilleumier cryocoolers
as were discussed earlier (Chapter 3) in connection with split-Stirling
systems. Split-Vuilleumier cryocoolers were described by Berry (1974),
Cohen and Daniels (1973), and Hornet al. (1973).

REFERENCES

Ackermann, R. (1971). "Vuilleumier Refrigerator Analysis, Low Temperature Refrigerator

Optimization." Contract No. DAAK02-70-C-0291, USAECOM, Fort Belvoir, Virginia,
AD 724770 (Cryomech Inc., Jamesville, New York).
AiResearch Manufacturing Co. (1972). "Vuilleumier Program Engineering Notebook."
Contract NAS 5-21096, Goddard Space Flight Center, Greenbelt, Maryland, August.
AiResearch Manufacturing Co. (1974). "Fractional Watt Vuilleumier Cryogenic Refrigerator
Program Engineering Notebook," Vol. 1, Thermal Analysis. Contract NAS 5-21715,
Goddard Space Flight Center, Greenbelt, Maryland, May.
Anon.-RCA (1970). "ICICLE Feasibility Study Final Report." NASA CR 109880,
AD-N70-34012 (Radio Corp. of America, Camden, New Jersey).
Berry, R. L. (1973). "Modular Cryogenic Refrigerators." Closed Cycle Cryogenic Cooler
Technology and Applications, pp. 119-136, AFFDL-TR-73-149, Vol. 1, WPAFB, Day-
ton, Ohio.
Berry, R. L. (1974). "Ultraminiature Vuilleumier Refrigeration System." AFFDL-TR-74-16,
WPAFB Dayton, Ohio (Hughes Aircraft Co., Culver City, California).
234 Chapter 4

Block, R. F. (1973). "The Design, Development and Test of a Heat Pipe Interface for the
Hot Cylinder of a V-M Refrigerator." Closed Cycle Cryogenic Cooler Technology and
Applications, pp. 271-288, AFFDL-TR-73-149, Vol. 1, WPAFB, Dayton, Ohio.
Browning, C. W., Miller, W. S., and Potter, V. L. (1972). "75 K Vuilleumier Cryogenic
Refrigerator Task IV, Model Fabrication." Report No. 72-8687, NASA GSFC,
Greenbelt, Maryland (AiResearch Manufacturing Co., Torrance, California).
Browning, C. W., and Potter, V. L. (1972). "75 K Vuilleumier Cryogenic Refrigerator. Task
II, Analytical and Test Program," Report No. 72-8497, NASA GSFC Greenbelt,
Maryland (AiResearch Manufacturing Co., Torrance, California).
Browning, C. W., and Potter, V. L. (1972). "75 K Vuilleumier Cryogenic Refrigerator, Final
Report for Task II-Analytical and Test Program." Contract NAS 5-21096, Goddard
Space Flight Center, Greenbelt, Maryland, August.
Browning, C. W., Miller, W. S., and Potter, V. L. (1972). "75 K Vuilleumier Cryogenic
Refrigerator-Final Report for Task IV Model Fabrication." Contract NAS 5-21096,
Goddard Space Flight Center, Greenbelt, Maryland, November.
Browning, C. W., Potter, V. L., Miller, W. S., and Gasser, M. G. (1973). "Developments
Toward Achievements of Long-Life Cryogenic Vuilleumier Refrigeration Systems."
Closed Cycle Cryogenic Cooler Technology and Applications, pp. 225-244, AFFDL-TR-
73-149, Vol. 1, WPAFB, Dayton, Ohio.
Bush, V. (1938). "Apparatus for Heat Transfer." U.S. Patent No.2, 127,286, August 16.
Bush, V. (1939). "Apparatus for Compressing Gases." U.S. Patent No. 2,157,229, May 9.
Chellis, F., and Hogan, W. H. (1964). "A Liquid-Nitrogen-Operated Refrigerator for
Temperatures Below 77 K." Adv. Cryog. Eng. 9, 545-551.
Chellis, F. (1973). "The Future of the Stirling Cycle Refrigerator in Airborne 1-R
Applications." Closed Cycle Cryogenic Cooler Technology and Applications, pp. 11-118,
AFFDL-TR-73-149, Vol. 1, WPAFB, Dayton, Ohio.
Chellis, F. (1977). "Design Compromises in the Selection of Closed Cycle Cryocoolers."
Proc. Appl. of Closed Cycle Cryocoolers to Small Superconducting Devices, Boulder,
Colorado, pp. 109-122, October (NBS Spec. Pub. 508).
Cohen, B., and Daniels, A. (1973). "Design and Development of a Free Displacer/Thermal
Compressor Cryogenic Refrigeration System." Final Report, Contract No. DAAK 02-71-
C-0421, N.V.L. USAECOM Fort Belvoir, Virginia, AD 920028 (Philips Laboratories).
Cowans, K. K. (1968). "Heat Powered Engine." U.S. Patent No. 3,379,026, April23.
Crouthamel, M. S., and Shelpuk, B. (1973). "A Combustion-Heated Thermally Activated
Vuilleumier Refrigerator." Adv. Cryog. Eng. 18, 339-351.
Daniels, A., and du Pre, F. K. (1971). "Miniature Refrigerators for Electronic Devices."
Philips Tech. Rev. 32(2), 49-56.
Doody, R. D. (1971). "Two-Stage Vuilleumier Cycle Cryogenic Refrigerator System for
Advanced Forward Looking Infrared (AFLIR) Applications." AFFDL-TR-71-17, DDC
Number AD886822, August.
Doody, R. D. (1975). "Long Life High Capacity Vuilleumier Refrigerator for Space Applica-
tions." AFFDL-TR-75-108, WPAFB, Dayton, Ohio (Hughes Aircraft Co., Culver City,
California).
Gasser, M.G., Yoshikawa, D. K., and Browning, C. W. (1972). "An Approach to Long-Life
VM Cryogenic Refrigerators for Space Applications." Appl. Cryog. Techno/. 4, 416-434.
Globe, J. B., Rhia, F. J. III, Gainey, R. T., Aske, H. D., and Doody, R. D. (1971). "Vuilleumier
Cycle Cryogenic Refrigeration System for Infrared Scanner Applications." AFFDL-TR-
71-18, WPAFB, Dayton, Ohio (Hughes Aircraft Co.).
Harkless, L. B. (1973). "Reliability Test Results on V-M Coolers." Closed Cycle Cryogenic
Cooler Technology and Applications," pp. 93-110, AFFDL-TR-73-149, Vol. 1, WPAFB,
Dayton, Ohio.
Hogan, W. H. (1964). "Closed-Cycle Cryogenic Refrigerator and Apparatus Embodying
Same." U.S. Patent No. 3,151,446, Oct. 6.
Horn, S. B., Cowan, K. C., and Berry, R. L. (1973). "Low Production Cost VM Coolers."
Closed Cycle Cryogenic Cooler Technology and Applications, pp. 73-82, AFFDL-TR-73-
149, Vol. 1, WPAFB, Dayton, Ohio.
Vuilleumier Cryocoolers 235

Horn, S. B., Lumpkin, M. L., Walters, B. T., and Acord, T. T. (1973). "Miniature Cryogenic
Cooler for TOW Night Sight." Closed Cycle Cryogenic Cooler Technology and Applica-
tions, pp. 52-70, AFFDL-TR-73-149, Vol. 1.
Lehrfeld, D., and Pitcher, G. K. (1973). "An Oil-Lubricated Triple Expansion VM Cooler
for Long Duration Space Missions." Closed Cycle Cryogenic Cooler Technology and
Applications, pp. 189-210, AFFDL-TR-73-149, Vol. 1, WPAFB, Dayton, Ohio.
Leo, B. (1970). "Designer's Handbook for Spaceborne Two-Stage Vuilleumier Cryogenic
Refrigerators." AFFDL-TR-70-54, Dayton, Ohio (Hughes Aircraft Co., Culver City,
California).
Leo, B. (1971). "Vuilleumier Cycle Cryogenic Refrigeration System Technology Report."
AFFDL-TR-71-85, WPAFB, Dayton, Ohio (Hughes Aircraft Co., Culver City,
California).
Magee, F. N., and Doering, R. D. (1968). "Vuilleumier Cycle Cryogenic Refrigerator
Development." AFFDL-TR-68-2, WPAFB (Hughes Aircraft Co., Ltd.).
Miller, W. S., and Potter, V. L. (1973). "Fractional Watt Vuilleumier Cryogenic Refrigerator-
Final Report for Task I Preliminary Design." Report No. 72-8846, NASA GSFC,
Greenbelt, Maryland (AiResearch Manufacturing Co., Torrance, California).
Miller, W. S., and Potter, V. L. (1974). "Fractional Watt Vuilleumier Cryogenic Refrigerator-
Final Report for Task II Analytical and Test Programs." Contract N AS 5-21715, Goddard
Space Flight Center, Greenbelt, Maryland, March.
Morgan, N. E. (1971). "Analysis and Preliminary Design of Airborne Air Liquefiers."
AFFDL-TR-71-171, WPAFB, Dayton, Ohio, AD 892500 (Hughes Aircraft Co., Culver
City, California).
Morgan, N. E. (1975). "Preliminary Design of a Cryocooler for Satellite Sensors (CCSS)."
AFFDL-TR-75-154, Vol. 1, WPAFB, Dayton, Ohio (Hughes Aircraft Co., Culver City,
California).
Pitcher, G. K. (1970). "Design and Development of Laboratory Model Vuilleumier
Cycle Cryogenic Refrigerator." Report No. 243, DA Prof. No. 663704P05201,
Contract No. DAAK02-69-C-0364, N.V.L. USAECOM, Fort Belvoir, Virginia (Philips
Laboratories).
Pitcher, G. K., aod du Pre, F. K. (1970). "Miniature Vuilleumier Cycle Refrigerator." Adv.
Cryog. Eng. 15, 447-451.
Pitcher, G. K. (1971). "Development of Spacecraft Vuilleumier Cryogenic Refrigerators-
Part !-Thermodynamic Design and Study of Life Limiting Components." AFFDL-TR-
71-147, Part 1, WPAFB (Philips Laboratories).
Pitcher, G. K. (1972). "Development of Spacecraft Vuilleumier Cryogenic Refrigerators Part
11-Development and Fabrication." AFFDL-TR-71-147, Part II, WPAFB (Philips
Laboratories).
Pitcher, G. K. (1973). "Mechanical Life of Space Cryocoolers." Closed Cycle Cryogenic Cooler
Technology and Applications," pp. 211-224, AFFDL-TR-73-149, Vol. 1, WPAFB,
Dayton, Ohio.
Pitcher, G. K. (1975). "Spacecraft Vuilleumier Cryogenic Refrigerator Development-Final
Report." AFFDL-TR-75-114, WPAFB (Philips Laboratories).
Pitcher, G. K. (1977). "Development of a Flight Design for an Oil Lubricated VM
Cryocooler." AFFDL-TR-77-95, WPAFB (Philips Laboratories).
Prast, G. (1963). "A Philips Gas Refrigerating Machine for 20 K." Cryogenics 3, 156-160,
September.
Renyer, B. L. (1973). "High Capacity Long-Life Vuilleumier Cycle Refrigerator." Closed
Cycle Cryogenic Cooler Technology and Applications, pp. 14 7-162, AFFDL-TR-73-149,
Vol. 1, WPAFB, Dayton, Ohio.
Rhia, F. J., Ill. (1971). "Development of Long Life High Capacity Vuilleumier Refrigeration
System for Space Applications." AFFDL-TR-71-92: Part I-AD 887399L. Part 11-
"Description of Preliminary Refrigerator Design and Results of Thermal Analyses and
Life Studies," AD 890601L. Part 111-"Refrigerator Design and Thermal Analyses,"
AD 901235L. Part IV-"Design and Mechanical Analyses of Refrigerator and Interface
Unit," AD 920 255.
236 Chapter 4

Richter, R., and Mahefkey, E. T. (1973). "The Applicability of a Solar Collector Thermal
Power System to a Vuilleumier Cooler." Closed Cycle Cryogenic Cooler Technology and
Applications, pp. 245-270, AFFDL-TR-73-149, Vol. 1, WPAFB, Dayton, Ohio.
Rule, T. T., and Qvale, E. B. (1969). "Steady State Operation of the Idealized Vuilleumier
Refrigerator." Adv. Cryog. Eng. 14, 343-352.
Russo, S. C. (1976). "Study of a Vuilleumier Cycle Cryogenic Refrigerator for Detector
Cooling in the Limb Scanning Infrared Radiometer." NASA CR 145078 AD/N77-1211
(Hughes Aircraft Co., Culver City, California).
Sherman, A. (1971). "Mathematical Analysis of a Vuilleumier Refrigerator." NASA TM-X-
65534, Ace. No. N-71-25812.
Sherman, A. (1972). "Selected Vuilleumier Refrigerator Performance Characteristics." Adv.
Cryog. Eng. 18, 352.
Sherman, A. (1978). "Cryogenic Cooling for Spacecraft Sensors, Instruments, and Experi-
ments." Astronaut. Aeronaut., November.
Sherman, A., Gasser, M., Goldowsky, M., Benson, G., and McCormick, J. (1979). "Progress
on the Development of a 3-5 year Lifetime Stirling Cycle Refrigerator for Space."
Cryogenic Eng. Conf., Madison, Wisconsin, August.
Taconis, K. W. (1951). "Process of and Apparatus for Heat Pumping." U.S. Patent No.
2,567 ,454, September 11.
Vuilleumier, Rudolph. (1918). "Method and Apparatus for Inducing Heat Changes." U.S.
Patent No. 1,275,507, August 13.
White, R. (1973). "Program for Predicting V-M Cooler Off-Design Performance. Closed
Cycle Cryogenic Cooler Technology and Applications," pp. 163-188, AFFDL-TR-73-
149, Vol. 1, WPAFB, Dayton, Ohio.
White, R. (1976). "Computer Program for Optimizing Three-Stage Vuilleumier Cycle
Cryogenic Refrigerators." AFFDL-TR-76-27, WPAFB, Dayton, Ohio.
White, R. (1976). "Vuilleumier Cycle Cryogenic Refrigeration." AFFDL-TR-76-17,
WPAFB, Dayton, Ohio, AD/A-027-055.
Yendall, E. F. (1960). "A Novel Refrigerating Machine." Adv. Cryog. Eng. 2, 188-196.
Yoshikawa, D. K. (1970). "75 K Miniature Vuilleumier Cryogenic Engine-Task 1." Report
No. 70-6854, NASA GSFC, Greenbelt, Maryland (AiResearch Manufacturing Co.,
Torrance, California).
Chapter 5

Gifford-McMahon, Solvay, and

Postle Cryocoolers

INTRODUCTION

This chapter is concerned with cryocoolers having regenerative heat

exchangers and valves to control the flow of working fluid, systems defined
earlier as Ericsson engines.
Within the generic family, there are two main branches: (a) Solvay
and (b) Gifford-McMahon or Postle engines. The essential difference
between them is that Solvay engines incorporate a piston as the reciprocat-
ing element in the expansion space. Work is extracted from the gas during
expansion. The Gifford-McMahon or Postle engines incorporate a displacer
as the reciprocating element in the expansion space. No work is extracted
from the gas during expansion.
The Solvay engine was invented in Germany about 1887 (Solvay,
1887), but according to Collins and Cannaday (1958), the prototype was
not properly executed and never attained cryogenic temperatures. It was
subsequently neglected until reinvented in the late 1950s by William Gifford
and Howard McMahon of A. D. Little, Inc., Cambridge, Massachusetts
(1960).
The Postle engine invented by Davy Postle of Stanthorpe, Queensland,
Australia was described in a British patent of 1873 (Thomas, 1873). Postle
was apparently not eligible to apply for the patent, as it was issued to
William Henry Thomas of New Cross, England.
The Postle engine was never intended to be a cryocooler, but rather
a marine refrigerator for the emerging meat trade between England and
Australia, following the opening of the Suez Canal. It was a very advanced
machine using hydrogen as the working fluid, regenerative free-piston
displacers, and fluidic control of the inlet/outlet valves and displacer
237
238 Chapter 5

motion. It is not known if actual engines were constructed and put into
service.*
The reinvention of the Solvay and Postle engines appears to have
occurred simultaneously. Some development of Solvay engines occurred
at A. D. Little Inc. but effort was concentrated principally on the Postle
displacer concept. Successful prototypes were developed and put into
production. They are widely known as Gifford-McMahon cryocoolers.
No Solvay engines are known to be in production anywhere. Cooling
engines made by Air Products and Chemicals Inc. of Allentown, Pennsyl-
vania are described as "modified Solvay" engines but have displacers rather
than expansion engines and are, therefore, Gifford-McMahon engines. The
lack of development effort given to Solvay engines is hard to understand.
It presumably arises from the difficult problem of a dynamic seal on the
piston to contain the working fluid in the system. The same problem exists
with Stirling and Claude cryocoolers and the seal technology developed
there could be applied to Solvay machines. Possibilities include the Philips
rolling seal, welded metal bellows, or the hydrodynamic lubricated, close
tolerance pistons used in Kaptiza, Collins, and Doll-Eder Claude cycle
expanders. There are many variants of Solvay machines. It is likely these
will receive development attention with the increasing need for intermedi-
ate-capacity cooling systems as superconducting applications materialize.
The Gifford-McMahon cryocooler was conceived at A. D. Little in
the mid-1950s and rapidly became a commercial product. The engine is
characterized as a large, heavy, slow-running system requiring minimal
maintenance. Higa et al. (1978) recently described an installation of 30
Gifford-McMahon cryocoolers with a cumulative total of one million hours
of operation over the preceding decade.
The durability of the Gifford-McMahon cryocooler makes it the system
of choice where size, weight, and efficiency are not of prime importance.
The coolers come in a variety of capacities and configurations. Table 5.1
summarizes the technical specification of Gifford-McMahon cryocoolers
available from Cryomech Inc., Jamesville, New York.
Initial production versions of Gifford-McMahon cryocoolers were
introduced by A. D. Little Inc. in the early 1960s and have remained in
production there and at the successor company Cryogenic Technology,
Inc., Waltham, Massachusetts. Figure 5.1 is a general view of a CTI Model
No. 0120 Gifford-McMahon cryocooler. This is a two-stage machine cap-
able of generating 1 W of refrigeration at 26 K.
The competing range of commercial cryocoolers was introduced by
Cryomech, Inc. of Jamesville, New York, a company founded by William

* The author would appreciate any information readers could supply on this point.
 C'l
Q
=
I
.=-
,.,==
Table 5.1. Technical Specifications for Gifford-McMahon Cryocoolers"
10
==
::r
Q
Type of refrigerator l'l
riJ
Q
AL01 AL02 D101 GB02 GB05 GB12 ~
10
~
Output 10

Cooling capacity 1@ 25 K 40@ 40K 0.5@ 25 K 1@ 9.5 K 1@ 15K 4@ 13K =-=

(W) 3@ 27K 80@ 80K 1@ 27K 2@ llK 8@ 20K - Q
-=
20@ 80K 4@ 14.5 K - - "'
~
Lowest temperature 23K 23K 23K 7.5K 9K ("l
Power required (kW) 0.9 3.0 0.6 3.0 0.9 3.0 ~
Q
Refrigeration ,.,
Q
temperature 23 K-80K 23 K-80K 23 K-80K 7.5 K-20K 12 K-20K 9K-25 K Q

Expander Concentric Dual Concentric Dual Concentric Dual ;-

.
single stage single stage single stage 2 stage 2 stage 2 stage "'
Applications Solid state Air liquefaction Solid state Superconducting High-temperature Target
detector cooling Cryopumping detector cooling Nb to Nb 3 Sn superconductors hydrogen
Air liquefaction 1. Receiver (RF) recondensing
Cryopumping 2. Detectors (sub
Thermal shielding mil-infrared)
3. Magnetometers
Weight (!b)
Cold head 5 25 4 25 6 25
Compressor 125 175 35 175 75 175

" All require no cooling, are Gifford-McMahon cycles with heat exchangers, operate at 144 rpm, and are rated for 3000 hr endurance. Courtesy Cryomech, Inc.,
Jamesville, New York.

N
~
-=
240 Chapter 5

Fig. 5.1. Gifford-McMahon cryocooler CTI Model No. 0120 (courtesy Cryogenic
Technology Inc.).

Gifford on leaving A. D. Little. As a professor at Syracuse University,

Gifford pursued extensive studies of cryocoolers (see the extensive listing
for Gifford in the Bibliography). He accomplished this with the aid of a
number of able graduate students. One of the most capable, Dr. Ralph
Longsworth, led Air Products and Chemicals, Inc. of Allentown, Pennsyl-
vania, into the manufacture of small regenerative cryocoolers of both the
Gifford- McMahon and Stirling varieties.

THE SOLVAY CYCLE

Figure 5.2 is a schematic diagram for a Solvay expansion engine. It

consists of a cylinder. with a piston connected to a crank-connecting rod
assembly, a regenerative heat exchanger, inlet and outlet control valves,
and a gas compressor. The engine operates in the inverted position with
the crankshaft above the top end of the cylinder. The cold part is at the
bottom end. The piston and cylinder are made long and thin and of low
conductivity material. The bottom end operates at cryogenic temperatures
whereas the top end is at the ambient atmospheric temperature. A dynamic
Gilford-McMahon, Solvay, and Postle Cryocoolers 241

COMPRESSOR
OISPLACER

OUTLET
VALVE

Fig. 5.2. Schematic diagram of Solvay

cryocooler.

(sliding) gas seal is provided on the piston to stop leakage of gas from the
cylinder. The seal is located near the top of the piston and operates at
ambient temperatures. Location of the seal at the bottom of the piston
would reduce the available refrigeration generated because of thermal
degradation of the frictional work of the seal.
The regenerative heat exchanger consists of a porous matrix of finely
divided material, usually metallic, contained in a cylinder of low-conduc-
tivity material. The function of the regenerator is to act as a thermal sponge
abstracting heat from the gas when it is passing downwards through the
matrix and releasing heat to the gas when it is passing upwards through
the matrix.
Inlet and outlet valves control the flow of gas into and out of the
expansion engine system. The valves are coupled to the crankshaft so their
operation is synchronized with the piston position. The inlet and outlet
valves are connected through a loop containing a gas compressor and
adequate low-pressure and high-pressure reservoirs. A water cooler is
included in the high-pressure side delivery duct of the compressor to remove
the heat of compression.
The operation of the engine is as follows:
(1) Pressure Buildup. With the piston at the bottom of its stroke, the
inlet valve opens and high-pressure fluid enters, compressing the fluid
already in the engine from pressure p 1 to pressure p 2 •
(2) Intake Stroke. With inlet valve open, the piston is raised from the
bottom dead center position. Gas is admitted through the inlet valve and
regenerator into the bottom of the cylinder. Passing through the regen-
erator, the gas cools by heat transfer to the matrix material of the
242 Chapter 5

regenerator. The reduction in gas temperature causes a reduction in

the gas pressure so that further charge is drawn into the cylinder through
the inlet valve.
(3) Expansion Stroke. The inlet valve now closes and the piston con-
tinues to ascend in the cylinder expanding the gas which experiences a
reduction in temperature. Some gas flows from the matrix to the cylinder
and the pressure decreases. Expansion continues until the piston reaches
top dead center position. The gas has expanded only partly and the pressure
is still appreciably above the minimum of the cycle.
(4) Exhaust Stroke. When the piston reaches the top of its stroke the
exhaust valve opens and fluid flows to exhaust until the pressure falls to
the minimum value. The exhaust valve is kept open as the piston descends
to the bottom of the stroke and gas is forced from the cylinder through
the regenerator and exhaust valve to the gas compressor. In passing through
the regenerator, the gas is heated by heat transfer from the matrix. This
cooling of the matrix prepares it to cool the gas passing through it en route
to the cylinder in the next cycle. When the piston reaches the bottom of
the cylinder, most of the low-pressure gas has been exhausted from the
engine. The exhaust valve closes, the inlet valve opens, and high-pressure
gas enters the engine to repeat the cycle.
A work diagram for the Solvay engine is shown in Fig. 5.3. The area
of the P- V diagram represents the net decrease in enthalpy which, for an
ideal gas, is equivalent to the available refrigeration.
The void volume of the regenerator and connecting ducts are shown
on the P- V diagram as the "clearance volume," V,. Admission of high-
pressure charge occurs until the volume of gas in the cylinder is V1 , at
which point the inlet valve is closed and the gas begins to expand. Normally,
in a closed piston and cylinder, expansion of gas follows a curved path on
the P- V plane, similar to that shown by the broken line in Fig. 5 .3, and
may be an isothermal, adiabatic, or a polytropic process. In the Solvay
engine, the process is markedly different because as the expansion proceeds
REGENERATOR VOLUME

Fig. 5.3. Work diagram for Solvay

v cryocooler.
Gifford-McMahon, Solvay, and Postle Cryocoolers 243

the mass of gas in the cylinder increases as fluid flows to the cylinder from
the regenerator. In Fig. 5.3, the expansion process is arbitrarily represented
by a straight line. The shaded area represents additional refrigeration
attributable to expansion in the cylinder of gas in the regenerator. This
partly compensates for the gas used to fill the regenerator void space which
appears to expand in the system to no useful purpose.
The gas required to fill the regenerator is not so great as might be
expected from a simple consideration of the void volume. The temperature
of the regenerator varies from the low refrigeration temperature at the
bottom end to ambient temperature at the top. The effective average value
of temperature may, therefore, be twice that of the expansion cylinder so
the mass required to fill the matrix is only half that required if it were all
at the expansion space temperature.
For best performance, the piston should have a discontinuous motion
with brief pauses at the end of each stroke. However, Gifford has indicated
that it is possible to use harmonic piston motion for simplicity without
significant loss of performance. Other constructional features are that the
piston may be quite loose fitting in the cylinder with up to 0.5 mm (0.02 in.)
clearance except for the gas tight seal at the warm end. To minimize heat
conduction, the cylinder must be made of thin-wall stainless steel tube and
the piston of low-conductivity, plastic material. The regenerator might be
a stack of dense mesh wire screen of 0.03-mm (0.001-in.) diameter copper
or phosphor bronze wire having 80 to 160 strands per em (200 to 400
strands per in.) contained in a thin-wall stainless steel tube.
A miniature Solvay cooling engine described by Gifford et al. (1960)
is shown in Fig. 5.4. The engine, intended for infrared detector applications,
was capable of attaining 55 Kin a single stage. The cylinder was a stainless
steel tube 5 mm (0.2 in.) in diameter with a wall thickness of 0.13 mm
(0.005 in.). The piston, made of linear base laminated phenolic, was desig-
ned to incorporate the regenerator consisting of fine wire mesh screen.
The inlet and outlet pressure of the helium working fluid was 1725 and
345 kPa (250 and 50 psi) with a consumption of 6 to 9 x 10-4 m 3 /sec
(0.2 = 0.3 fe /min).
The unit was equipped with a subassembly at ambient temperature
containing the inlet and exhaust valves, the piston seal, and cam shafts for
the valve and piston actuation. A 400-Hz, 115-V motor was used to drive
the crankshaft through a reduction gear at 500 rpm. The overall weight of
the cooling unit was 0.14 kg (10 oz) including the timing motor but excluding
the external gas compressor or compressed gas bottles.
The cooling characteristic for the device is shown in Fig. 5.5. It was
possible to cool to operating temperatures within 2-3 min and to maintain
temperatures in the 50-80 K range, indefinitely.
244 Chapter 5

o· 0
0

0 0

::
Fig. 5.4. Experimental miniature Solvay
cryocooler (after Gifford eta/., 1960).

130

120

t ItO

~ \
\
~ 100
a:
"'
0.

l
... 90
0
z
"'9

\\
80
8
70

~
60
........... .....__ Fig. 5.5 . Cooldown characteristic of the experi-
50
0 4 mental miniature Solvay cryocooler (after Gifford
ELAPSED RUNNING TIME - (mon ) eta/., 1960).
Gifford-McMahon, Solvay, and Poslle Cryocoolers 245

THE GIFFORD-McMAHON CYCLE

The Gifford-McMahon cycle, superficially similar to the Solvay cycle,

is shown schematically in Fig. 5.6. It consists of a cylinder, closed at both
ends, and containing a displacer having a length about three quarters of
the cylinder. The displacer is connected to a light drive mechanism so that
it can be moved up and down in the cylinder. The two volumes, one above
and one below the displacer, can be varied from zero to maximum but the
total volume remains constant.
The two spaces are connected through a thermal regenerator and to
a gas supply. The gas supply system consists of inlet and outlet valves, a
gas compressor, and high- and low-pressure reservoirs. The valves are
coupled to the drive mechanism so their operation is synchronized with
the position of the displacer. A heat exchanger is included downstream of
the gas compressor to cool the gas to ambient temperature after com-
pression.
The pressure above and below the displacer is the same except for
small pressure drops across the regenerator when gas is flowing through
it. Virtually no work is required to move the displacer in the cylinder. No
work is done on the gas and the gas does no work on the displacer. Pressure
in the system is increased or decreased by operation of the inlet or outlet
valves.
The displacer is a loose fit in the cylinder except at the top and where
it is equipped with a dynamic (sliding) seal to prevent leakage past the
displacer from one space to the other. The pressure is virtually the same
in both spaces so the seal problem is not severe and, furthermore, some
leakage can be tolerated between the two spaces.

Fig. 5.6. Schematic diagram of Gifford-

McMahon cryocooler.
246 ChapterS

The regenerator consists of finely divided metallic material which cools

the gas when passing downwards to the cold space and heats the gas when
passing upwards from the cold space.
The engine operates as follows:
(1) Pressure Buildup. With the displacer at the bottom of the cylinder
and exhaust valve closed the inlet valve is opened, increasing the pressure
level in the system. Fluid enters through the inlet valve to fill the regenerator
and the space above the displacer, volume 1.
(2) Intake Stroke. With the inlet valve open, the displacer is moved
from the bottom of the cylinder to the top. This displaces high-pressure
fluid from the space above the displacer, volume 1, through the regenerator,
to the space below the displacer, volume 2. In passing through the regen-
erator, the gas cools, causing the pressure to decrease and further gas to
enter the system to maintain the maximum cycle pressure.
(3) Pressure Release and Expansion. With the displacer at the top of
the cylinder, the inlet valve closes and exhaust opens. Fluid escapes and
the pressure falls. The drop in pressure causes a reduction in the gas
temperature level. The decrease in temperature of gas in the bottom
cylinder space, volume 2, is the useful refrigeration process of the cycle.
(4) Exhaust Stroke. With exhaust valve open, the displacer moves
from the top to bottom of the cylinder, displacing fluid from volume 2,
below, to volume 1, above the displacer. As it flows through the regenerator,
the fluid is heated by the matrix to near ambient temperatures. This process
cools the matrix ready for the gas entering in the succeeding cycle.
During pressure buildup and expansion, the regenerator is most
effective if gas can be made to flow through it at a constant rate. This will
be achieved by progressive opening of the inlet and exhaust valves with
the displacer stationary at the bottom and top of the cylinder, respectively.
The inlet valve opening should be extended over about 75° and the exhaust
valve about 110° of crank rotation.
Figure 5.7 shows the work diagrams for volumes 1 and 2 and Fig. 5.8
shows the temperature/time characteristic for gas in the engine. At the
beginning of a cycle, volume 1 is maximum and volume 2 is minimum.
The pressure is low, represented on Figs. 5.7 and 5.8 by state (a). The inlet
valve opens and fluid enters volume 1 and the regenerator. The fluid
contained therein is compressed, and its temperature is, consequently,
increased. It mixes with incoming fluid to a final temperature of mixed
fluid shown by state (b) on Fig. 5.7. Now the displacer moves from the
bottom to top of the cylinder and fluid is displaced from volume 1 to
volume 2 through the regenerator. In passing through the regenerator, the
fluid cools to temperature (c) and more fluid enters through the inlet valve
to maintain the pressure constant.
Gifford-McMahon, Solvay, and Postle Cryocoolers 247

VOLUME I (AMBIENT TEMPERATURE SPACE) VOLUME 2 (COLD SPACE I

Fig. 5.7. Work diagrams for the Gifford-McMahon cryocooler.

At (c) the inlet valve closes quickly and the exhaust valve opens slowly
and progressively. The pressure falls as fluid leaves the engine through the
exhaust valve. Fluid in volume 2 expands .as the pressure decreases and
the temperature of gas in volume 2 decreases. Heat transfer from the
cylinder walls increases the gas temperature to (d) in Fig. 5.7. Finally, the
displacer moves from the top to bottom of the cylinder causing fluid to
return back through the regenerator to volume 1 or exhaust. In passing
through the regenerator, the temperature of the gas increases to a value
slightly less than the mixture, following admission of high pressure gas. The
gas leaving the exhaust is, therefore, warmer than the intake feed gas. Heat
removed in this way is equal to the total refrigeration achieved. The net
result of investing the compressor work is that heat is pumped from a low
temperature to above the ambient level. The useful product is the refriger-
ation effect generated during expansion.
The work done by the gas on the displacer is shown by the work
diagram for volume 2. The work done by the displacer on the gas is shown
by the work diagram for volume 1. The pressures are the same in both
spaces so the areas of the work diagrams are identical (neglecting the
displacer rod area). For an ideal gas and adiabatic conditions, the area of
the diagram for volume 2 is equal to the change in enthalpy of the gas in
volume 2. It represents the net refrigeration available. Similarly, the area
of the diagram for volume 1 represents the change in enthalpy of a perfect
gas under adiabatic conditions.

Multiple-Expansion Gifford-McMahon Cycle

Additional expansion spaces to operate at different temperature levels
can be added with very little increase in overall unit complexity. Figure
5.9 shows a schematic diagram for a multistage heat pump with three
expansion spaces giving refrigeration at three different levels, 80, 35, and
14 K. Figure 5.10 shows a temperature/time history for the triple stage
248 ChapterS

REGENERATOR
6T

...
"'
~ 200

."'......"'
<(

1- 100

Fig. 5.8. Temperature/time history of gas in the Gifford-McMahon cryocooler.

device. The unit requires additional displacers, cylinders, and regenerators,

but the same valves operating at room temperature may be used and the
three displacers are all operated by the same actuator.
The system is equivalent to three expansion engines operating at
different temperatures. The only low-temperature parts are closed circuits

Fig. 5.9. Schematic diagram of multiple stage

Gifford-McMahon cryocooler (after Gifford and
Hoffman, 1961).
Gifford-McMahon, Solvay, and Postle Cryocoolers 249

w
"'::>
!;;( 200
~
31-

Fig. 5.10. Temperature/time history of gas in a multiple-stage Gifford-McMahon cryocooler.

and loosely fitting dis placer I cylinder sets with little mechanical stress. The
problems of construction are minimal. The additional stages operate exactly
like the single-stage device with gas supplied at the temperature of the
stage above. The expansion spaces describe P- V diagrams representing
the useful refrigerating output of the stage and the regenerative heat
exchangers prevent the loss of refrigeration to the stage above.
Multistaging is attractive because it provides a thermodynamically
efficient process to attain cryogenic refrigeration temperatures. For a given
refrigeration requirement, there is a decreased power requirement.
Expansion over given pressure and volume ratios produces the same
refrigeration regardless of temperature. However, the mass of gas involved
is inversely proportional to the temperature. A unit of refrigeration at 20 K
requires four times the gas and consequently four times the power, as at
80 K. Most cryocooler applications require the removal of heat at different
temperature levels, such as precooling before liquefaction, or the use of
cooled radiation shields to reduce heat leaks to the system at the lowest
temperatures. It is, therefore, thermodynamically advantageous for
refrigeration to be available at different temperature levels.
Furthermore, multistaging relieves the operating regimen of regen-
erators and heat exchangers. Consider the "refrigeration loss" Q,:
Q, =MCp fiT, (5.1)

where M is the mass flow of gas, CP is the heat capacity of the gas, and
fiT, is the temperature difference between the gas entering and leaving the
250 ChapterS

regenerator at the cold end. In expanding, the gas changes in temperature,

flT•. Now flT. must be greater than flT, otherwise no refrigeration would
be achieved. flT. is proportional to the temperature T at which the
expansion occurs, typically equal to 0.3 T. Thus, with expansion at 10 K,
flT. = 3 K. For there to be some useful refrigeration achieved, flT, would
need to be less than 3 K, say 1-1.5 K. To achieve this in a single stage
between 10 and 300 K requires an impractically high heat exchanger
effectiveness. In a three-stage system with an intermediate expansion at
40 K, a flT, of 1 K is possible for the regenerator operates only in the 10
to 40 K range.

Combination Gifford-McMahon Cryocooler and Joule-Thomson

Expander
The application of a multistage Gifford-McMahon cryocooler to
helium liquefaction was described by Gifford and Hoffman (1961). A
schematic diagram of the unit is shown in Fig. 5 .11. It consisted of the
three-stage heat pump described above with additional recuperative heat
exchangers below the three regenerators. The recuperative heat exchangers,

Fig. 5 .11. Schematic diagram of

multiple stage Gifford-McMahon
cryocooler combined with a Joule-
Thomson expander for helium
liquefaction (after Gifford and
Hoffman, 1961).
Gifford-McMahon, Solvay, and Postle Cryocoolers 251

at progressively decreasing temperatures, cooled a stream of high-pressure

helium gas below the inversion temperature for expansion in a Joule-
Thomson (JT) valve (see Chapter 6). Additional counterflow heat
exchangers were included in the JT circuit to maximize thermal performance
between the high- and low-pressure gas streams.
Two compressors were incorporated to compress helium from the
low-pressure JT circuit at 103 kPa (15 psi) to 550 kPa (80 psi) and compress
helium from 550 to 1920 kPa (80 to 280 psi). The pressure range for
operation of the multistage heat pump was 550 to 1920 kPa (80 to 280 psi)
fe
with a flow rate of 1.35 x 10- 2 m 3 /s (45 /min) and expansion at tem-
peratures of 80, 35, and 14 K. The JT circuit operated over a pressure
range of 1920 to 103 kPa (280 to 15 psi) with a flow rate of 1.2 x 10- 3 m 3 /s
fe
(4 /min). The net useful refrigeration available was 7000 mW at 4.2 K.
The cool-down time was about five hours. The refrigerator unit was 0.38 x
0.3 x 0.71 m (15 x 12 x 28 in.) and weighed about 41 kg (90 lb) (exclusive
of vacuum pumps) and required about 100 W of power. The compressor
unit was 0.55 x 0.6 x 1.9 m (22 x 24 x 75 in.), weighed about 318 kg (700 lb ),
and required 3 kW of power and free access of air for cooling.

Fluidic-Driven Displacer
A refinement to the original Gifford-McMahon cryocooler occurred
with application of a technique known as "cryomatic gas balancing." The
displacer mechanical drive was eliminated and a free displacer used instead.
Movement of the displacer was caused by differential fluidic (gas pressure)
forces acting on the displacer arising from the inlet and exhaust of gas
through a simple rotary cored valve.
A schematic diagram of a single-stage Gifford-McMahon cryocooler
with fluidic-driven free displacer is shown in Fig. 5.12. It consists of two
connected coaxial cylinders of different length and diameter containing the
free displacer. The displacer is about three quarters the length of the larger
cylinder and has a coaxial extension to fit the smaller cylinder. This creates
three chambers of variable volume depending on the position of the
displacer. Chamber 1 is the volume between the top of the displacer
extension and the small cylinder head. Chamber 2 is the volume between
the top of the displacer and the large cylinder. Chamber 3 is the volume
between the bottom of the displacer and the large cylinder. Gas seals are
provided on both the displacer and displacer extension to prevent the
leakage of gas past the displacer from one chamber to another. Chambers
2 and 3 are interconnected by a thermal regenerator of finely divided
metallic material. The pressure in both is, therefore, the same, apart from
small differences due to the pressure drop in the regenerator matrix when
252 ChapterS

ROTARY VALVE

RECIPROCATI G
OISPLACER

Fig. 5.12. Schematic diagram of a

Gifford- McMahon cryocooler with a
free, fluidic-driven displacer.

gas is flowing through. Gas is admitted and exhausted from chamber 1 and
the regenerator conduit through a rotary cored valve. The same valve
regulates both the inlet of high-pressure gas from the compressor and the
exhaust of low-pressure gas.
The sequence of operations illustrated in Fig. 5.13 is as follows:
(1) Pressure Buildup. At the beginning of the cycle the displacer is at
the bottom of the cylinder so the volume of chamber 3 is zero and the
volumes of chambers 1 and 2 are at their maximum value. The rotary valve
is in such a position that chamber 1 is full of gas at a high pressure and
chambers 2 and 3 are connected to exhaust and at low pressure. This is
represented in Fig. 5.13a. When the valve has turned to connect chambers
2 and 3 to the inlet port the pressure increases as represented in diagram
(b). Fluid enters through the inlet port to fill chamber 2, the regenerator
and the clearance space in chamber 3.
(2) Intake Stroke. Further rotation of the valve opens chamber 1 to
exhaust and the pressure decreases. The upward force on the displacer due
to pressure in chamber 3 now exceeds the downward force due to pressure
in chambers 1 and 2, by an amount equal to l:l.p times the area of chamber
1. As a result, the displacer moves to the top of the cylindei (Fig. 5.12c),
displacing fluid from chamber 2 through the regenerator to chamber 3. In
passing through the regenerator, the fluid cools and additional fluid enters
to fill chamber 3 and the regenerator.
Gifford-McMahon, Solvay, and Postle Cryocoolers 253

ROTARY VALVE

11l!JINOICATES HIGH PRESSURE GAS

CliNOICATES LOW PRESSURE GAS
Ia I (b) (C) (d) Ia I

ll
L VOLUME, CHAMBER 3

Fig. 5.13. Cycle of operation of the Gifford-McMahon cryocooler with free , fluidic-driven
displacer.

(3) Pressure Release or Expansion. Further rotation of the valve opens

chambers 2 and 3 to exhaust and the pressure decreases. Fluid flows out
of chamber 3 through the regenerator during the expansion and useful
refrigerating effect is developed in chamber 3. The situation is represented
in Fig. 5.13d.
(4) Exhaust Stroke. The same low pressure now exists in all three
chambers so the displacer will fall under gravity to the bottom of the
cylinder. Should seal friction prevent this, the valve soon opens chamber
1 to the high-pressure gas supply. The pressure force acting downwards
then moves the displacer to the bottom of the cylinder, displacing fluid
from chamber 3 to chamber 2. In passing through the regenerator, the
fluid is heated by the matrix back to near ambient temperature. This cools
the matrix in preparation for the gas entering in the next cycle. The increase
in temperature decreases the density of the gas and much of the fluid
leaving chamber 3 passes to exhaust rather than remaining in chamber 2.
The work diagrams for the unit are similar to those discussed earlier
for the engine with a mechanically driven displacer. The small diagram for
chamber 1 represents the pneumatic work done in driving the displacer up
and down the cylinder. The diagram for chamber 3 represents the change
254 ChapterS

in enthalpy and, for adiabatic conditions with an ideal gas, the refrigeration
effect of the process per cycle. The diagram for chamber 2 is virtually
identical but reversed in direction to that for chamber 3. A temperature-
time history of gas in the engine would be the same as that presented in
Fig. 5.8.
In a description of the fluidic-drive displacer engine, Gifford (1966)
does not discuss valve operation. It is virtually certain that the same careful
attention to programmed valve opening would be required as with a
mechanically driven displacer. This gains improved effectiveness of the
regenerator resulting from steady flow during the pressure buildup and
expansion processes. A good deal of expertise is concealed by Gifford's
disarming comment:
... timing between pressurizing and depressurizing the chambers may be set
by proper selection of the components connecting the chambers with the
valve .. .
Fluidic-driven displacers have been used in multistage units, where
refrigeration is generated at different temperature levels. This requires an
additional displacer /cylinder unit and regenerator per stage. A schematic
digram of a two-stage cryomatic gas balancing refrigerator is shown in Fig.
5.14. An experimental unit of this type mentioned by Gifford (1966) had

Fig. 5.14 . Schematic diagram of two-stage

Gifford-McMahon cryocooler with free,
fluidic-driven displacer.
Gilford-McMahon, Solvay, and Postle Cryocoolers 255

refrigeration capacities of 5 W at 15 K and 18 W at 50 K and was incorpor-

ated into a liquid helium 4.2 K refrigerator. He indicated that temperatures
as low as 23 K had been reached with a single-stage machine having a
32-mm (1.25-in.) diameter displacer operating in the speed range 1 to 3Hz
(50-150 cycles per min).

Machine Design and System Optimization

Some of the interesting heat transfer, fluid friction and mechanical
design problems of Gifford-McMahon refrigerating engines have been
discussed by Gifford and Hoffman (1961). An important factor is the valve
action and motion of the displacer. For best performance of the regenerator,
the gas flow should be constant one way for half the cycle and constant
the other way for the remainder of the cycle. Slight variations are permiss-
ible but very large surges of fluid through the matrix seriously affect
regenerator effectiveness.
The requirements are not compatible with the flow patterns necessary
to maximize the refrigeration effect in the expansion space for a displacer
moving with continuous harmonic motion. Here, sudden changes in pres-
sure from one extreme to the other are required at the time the displacer
is stationary at the ends of its stroke. To maximize the area of the work
diagram for the expansion space, a rapid increase in pressure is required
when the displacer is at the bottom of the cylinder and a sudden decrease
when the displacer is at the top of the cylinder. Efforts to achieve this
would result in rapid surges of gas through the system and lead to a serious
decline in regenerator performance.
An effective solution to the problem was achieved by (1) causing the
displacer to move in a discontinuous manner so that there are lengthy
pauses at each end of the stroke, and (2) increasing and decreasing the
pressure through valves which open slowly to bleed gas in and out at an
approximately constant rate. A composite diagram of the cyclic displacer
motion, valve action, and mass flow for a typical case is shown in Fig. 5.15.
Another important problem is the design and construction of the
regenerator. The regenerator should have the smallest possible void volume
and pressure drop yet have high thermal effectiveness. Regenerators made
from 100 to 200 mesh wire screen punchings contained in a thin wall
cylinder have been found satisfactory. Several hundred screens may be
required to obtain a regenerator a few inches long. Effective regenerator
matrices have also been made from small spheres or balls simply poured
into a small cylinder.
The matrix material must be selected with the operating temperature
in mind. The heat capacity of all solids decreases at low temperatures
256 ChapterS

..
0
>=
u UP

."'"'
u
~DOWN
90 180 270 360
c;

IN

Fig. 5.15. Cyclic history of displacer motion,

valve motion, and mass flow to and from the
cylinder of a Gifford-McMahon cryocooler
(after McMahon et al., 1960).

whereas the heat capacity of helium increases. The operation of a regen-

erator depends on the matrix material having a high heat capacity compared
with the gas flowing through. The problem of decreased matrix heat capacity
becomes severe at temperatures of 50 K or less with the normal matrix
materials of iron, steel, brass, and copper. For lower temperatures, lead is
the preferred matrix material because of its relatively high heat capacity.
The lead must be prepared as fine shot, powder, or some other finely
divided form.
To optimize design, one must establish the best compromise of a
number of factors:
(1) Refrigeration. The cold capacity of the machine is equal to the
area of the P- V diagram for the expansion space. Pressure drops in the
regenerator decrease the pressure excursion in the expansion space and so
reduce the refrigeration.
(2) Regenerator Pressure Drop. This may be minimized by utilizing a
short, fat matrix loosely filled with large diameter filling.
(3) Regenerator !:.T Loss. Due to imperfect operation of the regen-
erator, there is a difference in the temperature at which gas enters and
leaves the regenerator which reduces the refrigeration available. This may
be minimized by utilizing a large matrix of finely divided densely packed
solids to give maximum area for heat transfer and a high matrix heat
capacity.
(4) Regenerator Volume Loss. For each cycle, the regenerator void
volume must be filled with high-pressure gas. In the Solvay engine, a
fraction of the potential refrigeration in the void volume gas is gained, but
not in the Gifford-McMahon engine. The gas in the void volume has to
be compressed so its presence in the regenerator represents a direct loss
Gifford-McMahon, Solvay, and Postle Cryocoolers 257

of potential refrigeration. The loss may be minimized by a small densely

packed matrix.
(5) System Heat Leaks. Thermal leaks to the cold regions of the
machine may occur by conduction, convection, and radiation. Care must
be taken to minimize: (a) the convective losses by the use of vacuum
enclosures or insulation, (b) the radiation losses by radiation shields or
spectrally selective surfaces, and (c) conduction losses by the use of long
thin wall enclosures of low-conductivity material.

Advantages and Disadvantages of Gifford-McMahon Cryocoolers

Gifford-McMahon cryocoolers have an important advantage over con-
ventional expansion engines in Claude cycle systems. All the valves and
sliding seals operate at room temperature. The operation of rapidly opening
and closing valves for long periods without failure and with adequate
lubrication at low temperatures is a difficult problem. Furthermore, the
valves and seals must prevent gas leakage, otherwise there will be a decrease
in refrigeration effectiveness. At low temperatures, the problem of leakage
is aggravated by the decrease in gas viscosity. Operation at ambient tem-
perature permits the use of the elastomeric materials for valves and seals
and the use of standard lubricants to reduce wear.
Preventing gas leakage during the operation of a piston or displacer
at low temperatures is difficult without lubrication. Clearances must be
reduced to very small values and special dry sealing rings must be used. If
the seals operate at room temperature, a wide choice of materials and
lubricants is available and piston clearances below the seal may be relaxed,
virtually without limit.
Another advantage of Gifford-McMahon cryocoolers is their use of
regenerative heat exchangers. Regenerators are cheap, simple, easy to
build, and highly effective in small sizes. They will operate with appreciable
amounts of condensable contaminants in the working fluid with no serious
impairment of cycle efficiency or blocking of the flow passages. Conden-
sables deposited during the inlet flow are reevaporated and swept out
during the exhaust flow. In systems with recuperative counterflow heat
exchangers, condensable constituents, i.e., water, oil vapor, etc., must be
eliminated from the working fluid; otherwise they will accumulate in the
low-temperature regions and block the flow passages. Reversing heat
exchangers are used to periodically exchange intake and exhaust channels
to free the flow passages from accumulated condensed deposits. Regen-
erators automatically achieve the effect of reversing heat exchangers.
Solvay engines are inherently more efficient than Gifford-McMahon
engines because the fluid expands more completely. However, the piston
258 Chapter 5

seal of the Solvay engine has to withstand the full pressure difference
between the cylinder and the "crankcase." In the Gifford-McMahon
engine, the displacer seal has to sustain only the difference due to pressure
drop in the regenerator. Moreover, the difference reverses direction and
tends to cancel any leakage. A moderate leakage can be tolerated. The
displacer drive system is not required to sustain the large forces of a piston
drive and may be structurally simple and lightweight with fewer vibration
problems.
The principal disadvantage of the Gifford-McMahon cycle is the high
work input per unit of refrigeration (W/Q) characteristic. This arises from
the incomplete expansion and the mass of compressed gas consumed in
the regenerator void volume with little or no return of refrigeration.
The ideal work per unit of refrigeration for various cycles is shown in
Fig. 5.16. This was prepared by Gifford (1966) for refrigeration at 80 K
with helium as the working fluid, heat rejection at 300 K and a range of
pressure ratios p 2 /p 1 up to 5. The W/Q characteristic is drawn for the
Gifford-McMahon cycle, the Joule-Brayton cycle, and the ideal Stirling
and Carnot cycles. In discussing the figure, Gifford suggests that because
of the speed of operation, the reference cycle for Stirling cryocoolers should
be the Brayton cycle, the ideal reference cycle for small reciprocating and
turbine expanders. While allowing that the Gifford-McMahon system does
require more work for a given level of refrigeration, he suggests the
difference is not great at the small compression ratios of 1.5 to 2.5 used
in practice. Furthermore, the practical advantages outweigh the additional
work. The principal advantage is high reliability because the units are
simple, slow moving, and unstressed. The gas seal on a typical machine

100

90
/- GIFFORD McMAHON

80 / CYCLE

0 / .Y
BRAYTON CYCLE
(NO WORK RECOVERY)

/
04 v .---
~
0
...- BRAYTON CYCLE

4
oy /
0
TIRLING 8 CARNOT
0

0 Fig. 5 .16. Comparison of the work

input to heat lifted ratio for various
0
I 4 ideal thermodynamic cycles (after
Gifford, 1966).
Gifford-McMahon, Solvay, and Postle Cryocoolers 259

may move at the rate of 0.05 to 0.15 m/s (2 to 5 in./s) compared with the
0.75 to 1.25 m/s (30 to 50 in./s) of small Stirling cryocoolers. Continuous
operating times of 5 to 10,000 hr are not uncommon and Gifford's cryo-
coolers are noted for their high reliability and extended maintenance-free
operation.

Heat-Balance Analysis
Ackerman and Gifford (1971) presented data on the experimental
performance of a Gifford-McMahon engine with a fluidic-drive free dis-
placer. The experimental refrigerator is shown in Fig. 5.17. The machine
was a single stage expansion engine with a displacer 19 mm (0.75 in.) in
diameter, 0.1 m (4 in.) long and with a stroke of 12 mm (0.5 in.). With
helium as the working fluid and pressures ranging from 689 to 1960 kPa
(100 to 285 psi) it was found that the best operating speed was 3Hz
(190 cycles/min).

CAPACITANCE
DISPLACEMEN
TRANSDUCER FLOW REVERSING
~ VALVE

~ • SEALS
~
c
...z
~

...
:ll
~

Fig. 5.17. Cross section of experimental Gifford-McMahon cryocooler (after Ackermann et

al., 1971).
260 ChapterS

The ideal refrigeration of 14.5 W, was calculated from

Q = (pz-pl)Vv xN (5.2)
where Vv is the displacer swept volume and N is the number of cycles per
minute. The actual pressure-volume diagram was quite different from the
ideal case. Specimen pressure displacement diagrams for expansion space
temperatures of 50 and 300 K are shown in Fig. 5.18. At 77 K the indicated
diagram refrigeration was 6.5 W, and the actual refrigeration was measured
as 3.9 W, this being the heat supplied electrically to the expansion space
heat exchanger to maintain constant temperature.
The difference between the ideal and actual diagram refrigeration,
i.e., 14.5 and 6.5 W, respectively, was attributed to losses due to pressure
drop, void volume, and valve losses which degraded the pressure-displace-
ment diagram. The difference between the indicated diagram and actual
refrigeration was attributed to thermal losses including motional heat leak
(shuttle heat transfer or the bucket brigade loss), thermal conduction, and
imperfect regeneration. Estimates of the thermal conduction losses gave a
total of 1.58 W. Loss due to imperfect regeneration was estimated as 0.45 W
and to motional heat leak as 1.07 W. The total thermal loss estimate was
3.11 W compared with 2.6 W determined experimentally. The seal blow-by

300f-----

:::.---- =---~
1
( 250osi~ ;;e--T•50°K
W• 5.8wath
INPUT CONDITIONS;

..,
z
Ci
\ P"• 3DO polg
PL • 9!5 pliQ
RPM•I90

"'"'
0::
\ wideal· 14. 5 wath

.,.,"':::>
0::

"'0::
Q.

0::
"'<.>
.,:::>z
0

:>-
1112.5 poig
1001------ -======-=-===~_jiOO paig
Fig. 5.18. Comparison of ideal and
1----- .SOOin.--------+1 actual work diagrams for the experi-
mental Gifford-McMahon cryocooler
DISPLACEMENT (after Ackermann et al., 1971).
Gifford-McMahon, Solvay, and Postle Cryocoolers 261

losses were found to be small and were ignored. No account was taken of
radiation losses.
Much of the large difference, 8 W, between the ideal and indicated
diagram refrigeration was attributed to the large void volume in the cold
region necessary to accommodate the instrumentation. In commercial
machines, the regenerative heat exchangers are frequently included in the
displacer to reduce the cold end voids and conduction heat loss because
the regenerator tube is eliminated. Motional heat leak then remains the
prime thermal loss.

THE POSTLE ENGINE

The refrigerator invented by Davy Postle in 1873 was not intended

as a cryocooler, indeed it is not definitely known that the concept was ever
reduced to practice. A short description is included here to stimulate future
development and because the patent is not readily accessible (Thomas,
1873).
A diagram of the engine is shown in Fig. 5.19. It consisted of a
double-acting compressor cylinder and a separate cylinder containing two
identical displacer and valve assemblies in opposed form, separated by a
central diaphragm provided with flow passages for the circulation of
refrigerating brine. Each displacer cylinder of large diameter was provided
at the ambient temperature extremities with a coaxial smaller-diameter
cylinder. The displacer, fitting closely in the displacer cylinder, was similarly
provided with a coaxial extension fitting the smaller-diameter cylinder.
The hollow dis placer was equipped with a number of fine-bore tubes serving
as regenerative ducts connecting the expansion space, adjacent to the
central diaphragm, with the ambient temperature space at the cylinder end.
The end of the cylinder was provided with ducts for the cooling water flow.
Inlet and outlet valve seats were provided in the displacer extension
to correspond with a double-ended self-acting poppet valve mounted on
a valve stem sliding freely in a tube contained within the displacer along
the cylinder axis. The valve stem was slightly longer than the displacer
overall.
The engine operated as follows. With the element disposed in the
arbitrary positions shown, fluid at an increasing pressure entered the valve
chamber from the compressor cylinder. The flow of fluid caused the poppet
valve to slide to the right closing the inlet valve to the cylinder. The high
pressure acting on the displacer extension then moved the displacer inward
towards the diaphragm. Near completion of the inward stroke, the end of
the valve stem contacted the diaphragm causing the inlet valve to separate
262 ChapterS

DOUBLE -ACTING
- - - - - - - COMPRESSOR -CYLI DER___ -

COOLI G'
WATER
REGENERATIVE
DUCTS

DAVY POSTLES COOLING ENGINE

Fig. 5.19. Diagram of the Postle refrigerator.

from the valve seat moving inwards with the displacer. With the inlet valve
open, high-pressure fluid filled the cylinder and was cooled by the cooling
water heat exchanger on the cylinder end plate.
As the compressor piston reached the end of its stroke and started to
return, the pressure began to fall and fluid flowed from the dis placer cylinder
towards the compressor. The flow acting on the valve caused it to move
to the right, thereby closing the exhaust valve. Now the pressure in the
cylinder exceeded the decreasing pressure in the valve chamber, thereby
inducing the displacer to move from the inward to the outward position.
In this movement, fluid was transferred through the regenerative ducts to
the central expansion space.
Gifford-McMahon, Solvay, and Postle Cryocoolers 263

As the displacer approached the end of its outward stroke, the valve
stem impacted the cylinder cover plate causing the valve to separate from
the exhaust valve seat. Now with the exhaust valve open, fluid passed back
to the compressor cylinder all the while expanding in the expansion space
and generating refrigeration. When the compressor piston reached the end
of its expansion stroke and started to return, the flow of fluid reversed and
the operation of the system was repeated as described above.
The novel features of the self-acting valve arrangement of the Postle
engine endow the concept with the potential for development as a regenera-
tive cryocooler in the future. An alternative could be equipped with a gas
spring to operate at resonant frequency and be supplied with constant
pressure fluid from a reservoir.

REFERENCES

Ackermann, R. A., and Gifford, W. E. (1971). "A Heat Balance Analysis of a Gifford-
McMahon Cryorefrigerator." Adv. Cryog. Eng. 16, 221-29.
Collins, S.C., and Cannaday, R. L. (1958). Expansion Machines for Low Temperature Processes.
Oxford University Press, Oxford, England.
Gifford, W. E., and McMahon, H. 0. (1960). "A New Low Temperature Gas Expansion
Cycle-Part II." Adv. Cryog. Eng. 5, 368-372.
Gifford, W. E., and Hoffman, T. E. (1961). "A New Refrigeration System for 4.2 K." Adv.
Cryog. Eng. 6, 82-94.
Gifford, W. E. (1966). "The Gifford-McMahon Cycle." Adv. Cryog. Eng. 11, 152-159.
Higa, W. H., and Wiebe, E. (1978). "One Million Hours at 4.5 Kelvin." Proc. Appl. of
Closed-Cycle Cryocoolers to Small Superconducting Devices, pp. 99-108, Boulder,
Colorado, October 1977 (Published as NBS Spec. Pub. 508).
McMahon, H. 0., and Gifford, W. E. (1960). "A New Low Temperature Gas Expansion
Cycle-Part 1." Adv. Cryog. Eng. 5, 354-366.
Solvay, E. (1887). Deutsches Reichspatent No. 39,280.
Thomas, W. H. (1873). British Patent No. 709, "Producing Cold for Preserving Animal Food,
etc." (Available from Science Museum Library, London).
Chapter 6

Joule-Thomson Cooling Systems

INTRODUCTION

In a Joule-Thomson cryocooler gas expands from high pressure to low

pressure at constant enthalpy (often called a throttled or Joule-Thomson
expansion) and thereby experiences a decrease in temperature. The system
is widely used for gas liquefaction. It may be a simple Linde-Hampson
cryocooler or associated with more complex systems involving closed cycle
expansion engines to precool the gas before throttled expansion occurs.
Initial experiments on the isenthalpic expansion of a gas were carried
out in the early 1850s at Manchester, England by J. P. Joule and William
Thomson (later to become Lord Kelvin). The initial experiments were
made with gas passing through a nozzle (isentropic expansion). Later a
porous plug replaced the nozzle and Joule and Thomson investigated
isenthalpic expansion with different gases. They found (Joule, 1852), with
initial temperatures near ambient, that many gases (air, C0 2 , Oz, N 2 )
experienced a decrease in temperature but that hydrogen experienced an
increase in temperature.
The first use of isenthalpic expansion in a refrigerating system is
attributed, by Daunt (1956), to Perkins as early as 1834 in the British
patent 6662. Sulfuric ether [(C 2 H 5 h0] was used by Perkins as the working
fluid in his evaporative cooling system and again by Harrison in 1857 for
an improved system. Later, in 1876, Karl von Linde in Germany used
ammonia as a working fluid for refrigeration. About the same time Pictet
in Switzerland used sulfur dioxide as a refrigerant.
The first use of Joule-Thomson isenthalpic expansion for cryocooling
was in air liquefiers devised independently but simultaneously by Linde
(1897) in Germany and Hampson (1895) in England. Both systems incor-
porated the recuperative heat exchanger principle invented by Siemens in
1857. Hampson used a close coiled cross flow exchanger of the type now
identified as a Hampson heat exchanger. Isenthalpic expansion valves were
265
266 Cbapter6

also incorporated by Claude in the gas liquefaction systems he developed.

Nowadays JT expansion valves are used routinely in nearly all modern gas
liquefaction and vapor-compression refrigeration systems.

THEORETICAL ASPECTS OF JOULE-THOMSON EXPANSION

Isenthalpic Expansion
Consider the system shown in Fig. 6.1. It consists simply of a duct
containing a porous plug, a valve that may be partly "cracked" open, or
a metal plate with very fine holes drilled through it. Gas at high pressure
enters at station 1 and expands through the constriction to a low pressure
at station 2. Fluid passing through the plug experiences high fluid friction
which degrades the pressure but no work is done. For the purpose of
illustration it is assumed the duct is thermally insulated so that no heat is
transferred to or from the fluid. The fluid flows so slowly that the change
in kinetic energy in expansion is negligible. The duct is horizontal with no
change in potential energy.
Applying the steady flow energy equation to the system shown in Fig.
6.1 and equating changes in kinetic energy, potential energy, work, and
heat transfer to zero leaves
H1=H2
The actual flow in the valve is thermodynamically irreversible and the fluid
is not homogeneous. The fluid downstream at 2 can be reasonably assumed
to be mixed and homogeneous. It will be found to have the same enthalpy
as the inlet fluid.
Outlet conditions of temperature and pressure following isenthalpic
expansion from given inlet conditions are plotted on the temperature-
pressure plane shown in Fig. 6.2. The curves obtained are lines of constant
enthalpy. Starting at the right-hand, high-pressure, side of the figure and

®
Fig. 6.1. Duct with porous plug for Joule-Thomson expansion of gas.
Joule-Thomson Cooling Systems 267

MAXIMUM INVERSION LINES OF CONSTANT

TEMPERATURE ENTHALPY

Fig. 6.2. Generalized constant enthalpy

curves for a gas displayed on the press-
ure-temperature plane. PRESSURE

moving along the constant enthalpy curve, A-B, the temperature initially
increases as the pressure decreases. The increase in temperature reaches
maximum at point C. Further isenthalpic expansion causes a decrease in
temperature.

The Inversion Curve

Curves similar to A-C-B are shown at higher and lower initial tem-
peratures of the gas. The loci of the apex points, D-E-C-F-G-H are
connected to establish the inversion curve for the gas. An initial condition
of pressure and temperature to the left of the inversion curve will result
in the gas experiencing a decrease in temperature on isenthalpic expansion
to a low pressure. Conversely, from an initial condition of pressure and
temperature to the right of the inversion curve the gas will experience an
increase in temperature on isenthalpic expansion. If the initial conditions
are on the right of the inversion curve and the final pressure is to the left
the final temperature may be greater, equal, or less than the initial tem-
perature.

Maximum Inversion Temperature

Point D on the inversion curve is called the maximum inversion
temperature. It occurs where the pressure is zero. Maximum inversion
temperatures for different gases are given in Table 6.1.
Gases can be divided into two groups: (a) those for which the maximum
inversion temperature is above ambient temperatures, i.e., carbon dioxide,
268 Chapter6

Table 6.1. Maximum Inversion Temperature for

Different Gasesa

Maximum inversion
Gas temperature, (K)

Above ambient
Carbon Dioxide 1500
Oxygen 760
Argon 722
Nitrogen 623
Air 602

Below ambient
Neon 250
Hydrogen 202
Helium 40

" After Barron (1966).

oxygen, argon, nitrogen, air, etc., and (b) those for which the maximum
inversion temperature is below ambient temperature, i.e., neon, hydrogen,
and helium. This difference in maximum inversion temperature, either
above or below the ambient value, explains why Joule and Thomson
observed the curious phenomena that on expansion through a porous plug
hydrogen experienced an increase in temperature whereas all the other
gases experienced a decrease.
Gases with low values of the maximum inversion temperature must
be precooled, by an expansion engine or other means, to a temperature
within the inversion envelope to ensure a decrease in temperature on
isenthalpic expansion.

Joule-Thomson Coefficient
The change in temperature for an isenthalpic pressure drop is defined
by the Joule-Thomson coefficient for a fluid: f.J.H = (aTjap)H the change in
temperature due to a change in pressure at constant enthalpy. The Joule-
Thomson coefficient is the slope of the isenthalpic curves shown in Fig.
6.2. The coefficient is zero along the inversion curve (at the peak of the
isenthalps) since the slopes of the isenthalps are zero. To the right of the
inversion curve the Joule-Thomson coefficient is negative and the tem-
perature increases on expansion. To the left the Joule-Thomson coefficient
(slope of the isenthalp) is positive and the temperature decreases on
expansion. Figure 6.3 shows the Joule-Thomson coefficient for air as a
function of temperature at several different pressures (reproduced from
Daunt, 1956).
Joule-Thomson Cooling Systems 269

08

06
~~~,
I .,., ~
04

I ~~
~~
); .------
-..;::

p•200 otm

Fig. 6.3. Joule-Thomson coefficient for

--- _.//

air as a function of temperature at several

-0
' •o 100 zoo zoo
different pressures (after Daunt, 1960). TEMPERATURE (K)

The Joule-Thomson coefficient can be written*

(6.1)

and from elementary thermodynamics

(6.2)

and

dH = CpdT+ [v- T(:;) Jdp (6.3)

A comparison of the coefficients in Eqs. (6.2) and (6.3) indicates that

c p
=(aH)
aT p
(6.4)

and

(aH)
ap
- v - T(~)
T aT p
(6.5)

Combining Eq. (6.1), (6.4), and (6.5) we may write

(6.6)

*The following is a summary of the excellent treatment of this topic given by Barron (1966).
270 Chapter6

For a perfect gas: v = RT/p and

v
(6.7)
T
so that from Eq. (6.6) for a perfect gas

(6.8)

A perfect gas would experience no change in temperature in a throttled

or Joule-Thomson expansion. Therefore a positive or negative Joule-
Thomson coefficient must arise from the departure of real gases from
perfect gas behavior.
Recall that enthalpy is defined as
H=u+pv (6.9)
Substituting this in Eq. (6.1) results in

(6.10)

The first term in Eq. (6.1 0) is indicative of the departure of a real gas from
perfect gas behavior in terms of Joule's law. This states that the internal
energy of a gas is a function only of temperature. For real gases the first
term is always negative and contributes to a decrease in temperature on
expansion. This arises because the molecules of the gas become increasingly
separated as expansion proceeds and their microscopic potential energy is
increased. No external work or heat transfer occurs, so the increase in
potential energy must be offset by a decrease in the microscopic kinetic
energy of the gas molecules and is manifested by a decrease in temperature.
The second term in Eq. (6.10) is indicative of the departure of a real
gas from perfect gas behavior in terms of Boyle's law. This states that the
product of pressure and volume of a perfect gas is a function only of
temperature and is illustrated in Fig. 6.4 by the horizontal line. This shows
the product of pressure and volume (an isothermal) to be independent of
the pressure.
The behavior of real gases is also illustrated in Fig. 6.4. At low pressures
and at temperatures near saturation conditions gases are more compressible
than Boyle's law predicts since attractive forces are acting to condense the
gas. This means the second term in Eq. (6.10) is negative and contributes
to the production of a temperature decrease. At high pressures the
molecules are closer together and repulsive forces are brought into action.
The gas is therefore less compressible than Boyle's law predicts, and the
Joule-Thomson Cooling Systems 271

Fig. 6.4. Product of the pressure and

volume of a gas as a function of pressure
at several different temperatures. PRESSURE p

second terms in Eq. (6.10) is positive. This contributes to a temperature

increase on expansion.
Increase or decrease in gas temperature on expansion depends upon
the relative magnitude of the two terms in Eq. (6.10). Along the inversion
curve shown in Fig. 6.2 the two terms are equal in magnitude (and opposite
in sign). The second term dominates to the right of the inversion curve
and the first term to the left of the inversion curve.
Those wishing to pursue the thermodynamic aspects of Joule-Thomson
expansion beyond the elementary level presented here are referred to the
paper by Professor Baehr (1963).

CRYOGENIC COOLING AND GAS LIQUEFACTION SYSTEMS

USING JOULE-THOMSON EXPANSION

Linde-Hampson Systems
A schematic diagram of an elementary Linde-Hampson gas liquefac-
tion system is shown in Fig. 6.5. The system consists of a compressor, a
recuperative heat exchanger, a Joule-Thomson valve, and a liquid/gas
separator and liquid reservoir.
The system operates as follows:
(a) Compression (process 1-2"). Gas enters the compressor at condi-
tion (1) and is compressed to state (2). Work W is supplied and heat Q
rejected during compression of the gas. For the purpose of analysis the
compression may be assumed isothermal or isentropic. In practice it would
be a polytropic process (1-2") between the isothermal process 1-2 and the
isentropic process 1-2'.
272 Chapter6

Fig. 6.5. Schematic diagram of simple

Linde-Hampson gas liquefaction system.

(b) Aftercooling (process 2"-2). The compressed gas is cooled to

ambient temperature in a water or air cooled heat exchanger.
(c) Recuperative cooling (process 2-3). The compressed gas enters a
contraflow recuperative heat exchanger and is cooled to condition (3).
(d) Joule-Thomson expansion (process 3-4). The compressed gas
expands to low pressure at constant enthalpy. If the gas at (3) is below the
inversion temperature it experiences a decrease in temperature. The tem-
perature at (3) may be so low that the final condition (4) is located in the
liquid vapor envelope on the T-S diagram and the fluid will partially
condense during expansion. At (4) it will exist partly as saturated liquid at
(5) and partly as saturated vapor at (6). If the temperature at (3') is not
low enough the final condition (4') will be located outside the liquid-vapor
envelope and no saturated liquid will be produced.
(e) Vapor-liquid separation. The saturated liquid at (5) is collected
as the useful product of the system and the cold, saturated vapor at (6) is
returned to the compressor.
(f) Recuperative heating (process 6-7). Low-pressure saturated vapor
at (6) passes through the recuperative heat exchanger en route to the
compressor and experiences recuperative heating to condition (7). If liquid
has been produced the mass of gas flowing in the return line will be less
than the mass of compressed gas flowing to the JT valve. The reduced mass
flow stream (6-7), will experience a higher temperature differential across
the recuperator than the high-pressure stream (2-3) and the exit tem-
perature (7) will approach the compressor inlet temperature (1). To illus-
Joule-Thomson Cooling Systems 273

trate the make-up/mixing process (7) is shown substantially less than the
compressor inlet (1).
(g) Make-up and mixing (process 8/7-1). As liquid gas is produced
and leaves at condition (5), new make-up gas must be admitted at the
compressor inlet to maintain constant mass flow and continuous operation.
The make-up gas is admitted at ambient temperature, (8), and mixes with
the return effluent (7) for admission to the compressors, (1).
For the purpose of elementary analysis it is assumed that:
i. the flows take place with no pressure drops,
n. there are no external heat leaks,
iii. the recuperative heat exchanger is perfect, and
iv. the compression process is isothermal.
Applying the steady-flow energy equation for mass, m, and enthalpy, h,
to the recuperative heat exchanger, the expansion valve and the separator
we obtain

where

The fraction of gas liquefied:

(6.11)
m2 h?-hs
and in the idealized system h 7 = h 1 • The fraction of gas liquefied therefore
depends only on the pressure and temperature at (1) and the pressure ratio
p 2 /Pl· Many systems for air liquefaction use ambient pressure and tem-
perature for condition (1). This fixes h 1 and h 5 so the only variable is the
isothermal compression ratio p 2 /p 1 which determines h 2 •
The maximum fraction of gas will be liquefied when the optimum
pressure p 2 is that at which the enthalpy h 2 is a minimum. Maxima and
minima occur when the rates of change or slope of a function are zero. In
this case h 2 is a minimum when the rate of enthalpy change as a function
of pressure (at constant temperature) is zero, i.e.,

(ah) _0 (6.12)
ap T=Tt

It can be seen from Eq. (6.1) and Eq. (6.4) that Eq. (6.12) is equivalent to:

(ah)
ap T=Tt
= o= (p.,H · Cph=T, (6.13)
274 Chapter6

The pressure p 2 at which the enthalpy h 2 is a minimum occurs at the

apex of the isenthalp, i.e., point C along the enthalp A-B in Fig. 6.2.
For air at ambient temperatures of 15 to 20 C the optimum pressure on
the inversion curve (point C) is about 41.3 MPa (6000 psi). Unfortunately
this is much too high for economic operation, and lower pressures, about
half or less of the above figures, are generally used.
It is interesting to note from Eq. (6.11) that the fraction of gas liquefied
is negative so long as h 7 ( = h 1 ) is less than h 2 • This is the case with hydrogen,
helium, and neon at atmospheric ambient temperatures and shows that
liquids of these gases could never be produced in a simple Linde-Hampson
system. Not enough energy can be exchanged in the recuperator to reduce
the compressed gas temperature to the point where liquid could be obtained
on isenthalpic expansion. The greater limitation with these gases is that
the system could never produce any cooling at all from an initial start at
ambient temperatures.

Start-Up and Cool-Down of a Simple Linde-Hampson System

The system shown in Fig. 6.5 was assumed to be in a state of steady,
continuous flow operation. It is of interest now to consider the conditions
prevailing during start-up and cool-down with the whole system initially
at the ambient temperature, T 1 •
Consider first a fluid having a maximum inversion temperature well
above the ambient temperature. The process of cool-down for such a gas
is illustrated in the series of T-S diagrams given in Fig. 6.6. The initial
operation at time t = 0 is shown in Fig. 6.6a. After compression (process
1-2") the gas is cooled in the aftercooler (process 2"-2) and then passes to
the recuperative heat exchanger. No cooling is accomplished because the
heat exchanger is all at ambient temperature. The gas expands isenthalpi-
cally in the Joule-Thomson valve from ambient temperature and high
pressure. The temperature at (4') is lower than the ambient temperature
but of course no liquid is produced.
Gas returning to the compressor through the recuperator will be
reheated to the initial temperature by the transfer of heat from, and hence
cooling, the high-pressure fluid passing to the expansion valve. The system
is therefore able to "boot-strap" its way through the initial cool-down
process. The situation at some intermediate time is represented in Fig.
6.6b and yet later in Fig. 6.6c when the system is in steady-state operation
and producing liquid.
Now consider the other group of fluids given in Table 6.1 where the
maximum inversion temperature is below ambient temperature. Such a gas
used in the simple Linde-Hampson system would never be able to produce
Joule-Thomson Cooling Systems 275

T T

s
Ia) (b) (c)

Fig. 6.6. Start-up and cool-down of a simple Linde-Hampson system where the gas has a
maximum inversion temperature above ambient.

refrigeration, liquefaction, or cryogenic cooling. This is illustrated in Fig.

6.7.
For the sake of clarity we assume the compression process is isothermal
(process 1-2). As before on inital startup there is no effect as the gas passes
through the recuperator so that it reaches the Joule-Thomson valve at
ambient temperature T 1 . On expansion to the lower pressure the gas
actually experiences an increase in temperature to (4'). Therefore in return-
ing to the compressor through the recuperator the hot gas will heat rather
than cool the incoming high-pressure stream en route to the expansion
valve. The system could never cool itself down and would simply continue
operating at increasingly elevated temperature levels until failure.

Precooled Linde-Hampson System

Precooling is incorporated in the Linde-Hampson system to overcome
the problem discussed above in the liquefaction of hydrogen, helium, and
neon. One possibility. is to cool-down the whole liquefaction circuit to a

Fig. 6. 7. Start-up of a simple Linde-Hampson system where

the gas has a maximum inversion temperature less than
ambient (hydrogen, helium, and neon).
276 Chapter6

temperature below the inversion temperature but this is not advantageous

thermodynamically. Furthermore, it is not easy to operate compression
equipment at temperatures much below ambient. The preferred alternative
is to incorporate the precooling at a location downstream of the compressor
as shown in the schematic diagram, Fig. 6.8.
The precooler is a cold liquid bath through which the pressurized gas
duct passes en route to the expansion valve. The precooler can be located
downstream of the recuperative heat exchanger and immediately upstream
of the Joule-Thomson expansion valve. However, this is less advantageous
than the arrangement shown where the recuperative exchanger is divided
into two, a high-temperature and low-temperature section. It is most
advantageous thermodynamically to arrange the division so the pressurized
gas stream enters the precooler heat exchanger at a temperature not much
greater than that at which the precooler liqiuid coolant is available.
The precooler liquid coolant can be any liquid with a normal boiling
point (at atmospheric pressure) below the maximum inversion temperature
of the gas to be liquefied. For neon and hydrogen (maximum inversion
temperatures are 250 and 202 K, respectively) the precoolant liquid could

Fig. 6.8. Precooled Linde-Hampson system.

Joule-Thomson Cooling Systems 277

be fluorine (normal boiling point 85 K), oxygen (90 K), air (78 K), methane
(113 K), argon (87 K) or nitrogen (77 K). In practice for reasons of economy
and safety nitrogen is always used except in unusual circumstances.
Helium has a maximum inversion temperature of only 40 K. Therefore
the only liquids suitable for helium precooling are hydrogen (normal boiling
point 20 K) or neon (27 K). On grounds of safety neon is the preferred
fluid but economic considerations dictate that hydrogen is almost always
used.

Cascade System
A helium liquefier operating on the precooled Linde-Hampson cycle
consists of a series of separate systems having different fluids at different
minimum temperature levels. This is known as a "cascade system." It was
used originally for the liquefaction of air. The schematic arrangement of
the compound system is shown in Fig. 6.9. The useful product, helium, is
available at 4 K. The helium is precooled by a hydrogen bath at 20 K. The
hydrogen is precooled in a nitrogen bath at 77 K.
A system of this type was described by Mann et al. (1960). The liquefier
was 0.5 m diameter and 2m high and was constructed at the Cryogenic
Engineering Laboratory of the National Bureau of Standards, Boulder,
Colorado. Mann reports the following performance of the helium liquefier:
i. Liquid helium: 2.4 kg per hour
ii. Helium flow rate: 12.6 kg per hour
iii. Liquid yield (a/b)= 0.19
iv. Liquid hydrogen evaporated: 1.2 kg per hour
v. Liquid nitrogen evaporated: 10 kg per hour.
Geist and Lashmet (1960) described a lower capacity (750 mW useful
refrigeration at 4.4 K) Joule-Thomson cascade helium liquefier using
nitrogen and hydrogen precooler circuits. The system was constructed by
Air Products and Chemicals Inc., Allentown, Pennsylvania as a prototype
for application with masers, cryotrons, superconducting magnets, and other
electronic devices requiring refrigeration at liquid temperatures. The unit
was notable for the use of Corblin* diaphragm compressors to eliminate
the need for fluid purification equipment. The nitrogen and hydrogen
circuit used two-stage compressors having inlet, intermediate and discharge
pressures of 0.1, 1.96 (1.69 for Hz) and 24.8 MPa [15, 285 (245 for Hz),
and 3600 psi], respectively. The helium compressor operated over the
pressure range 0.1 and 1.86 MPa (15 and 270 psi).

* Societe des Compresseurs a Membrane Corblin, Paris, France.

278 Chapter6

NITROGEN COMPRESSOR

Fig. 6. 9. Schematic diagram for helium

liquefaction using the external cascade
system.

The advantages of precooling and the cascade system are not confined
to gases with subambient maximum inversion temperatures (neon, hydro-
gen, and helium) but apply equally to gases of the other group of gases to
improve the liquid yield fraction. Figure 6.10, reproduced from Barron
(1966), shows the liquid yield fraction (y) (the fraction of total mass
expanded that is saturated liquid after expansion) as a function of the
temperature of compression (assumed to be isothermal). The figure refers
to a simple idealized Linde-Hampson system using nitrogen. It is evident
that the amount of liquid produced is a strong function of the compressor
temperature: the lower the compressor temperature the higher the fraction
of gas converted to saturated liquid after expansion.
A similar improvement can be obtained using the precooled Linde-
Hampson system. An auxiliary refrigerating system cools the compressed
gas between the compressor aftercooler and the main recuperative heat
Joule-Thomson Cooling Systems 279

"'
;;:
5"' 0.2f----l--+--\--++------r----l
::::;

;;
(f)

lL O.lf---------,f-------'rt---'<------"tc---1
0
Fig. 6.10. Fraction of gas liquefied in a simple
Linde-Hampson nitrogen liquefier as a func- ~
tion of the gas temperature at the end of
isothermal compression at several different
~
0:: 0 100
lL
compression ratios. COMPRESSION TEMPERATURE (K)

exchanger. The auxiliary refrigerator can use the same working fluid but
this is not essential since the auxiliary circuit is entirely independent. It is
often advantageous to use alternative fluids with more favorable thermody-
namic or thermofluid characteristics.
Barron (1966) considered a precooled Linde-Hampson cycle for
nitrogen liquefaction with Freon 12 as the refrigerant in the auxiliary
precooling circuit. For his study he assumed the nitrogen system to operate
between pressures of 0.1 and 20 MPa (15 and 3000 psi) and temperatures
of 77 and 294 K (138 and 530 R). The Freon 12 system was assumed to
operate between pressure of 103 and 590 kPa (15 and 86 psi) and tem-
peratures of 243 and 469 K (438 and 844 R). He found the addition of
the Freon 12 precooler increased the liquid yield by 40% compared with
the simple Linde-Hampson nitrogen liquefaction system. At the same time
the total work requirement increased only marginally so the work require-
ment per unit mass liquefied decreased by almost 40% compared with the
simple cycle. This improvement results from the closer approximation to
an ideal reversible system. It can be carried further with multiple refriger-
ation systems with continuing improvement but at a progressively diminish-
ing rate.
The cascade system was first applied for the liquefaction of air. One
system suggested by Keesom (1933) utilized auxiliary refrigeration systems
(in order of decreasing temperatures) of (a) ammonia used to liquefy, (b)
ethylene used to liquefy, (c) methane used to liquefy nitrogen. The system
is known as the "outer cascade cycle."
A later development, known as the "incorporated cascade cycle" uses
a compound working fluid having several components which are mixed
and processed together in a single compressor. The mixture circulates
through a series of heat exchangers and separators so that various fractions
280 Chapter6

of the gas become liquid and are separated at different temperature levels.
The various fractions are further expanded and subcooled to provide
refrigeration to cool and condense different components of the mixture.
Detailed consideration of this complex system is beyond the level of
this book. Suffice to say the incorporated cascade system is used commer-
cially in large installations for the liquefaction of natural gas. It is also used
in smaller plants of low cooling capacity. The principal advantage is the
high flexibility. Depending on the choice of components and the composi-
tion of the gas the system can be used to achieve refrigeration at any
temperature level between ambient and cryogenic temperatures. Carbonell
et a/. (1972) have described a development of the incorporated cascade
system using a complex working fluid and regenerative rather than
recuperative heat exchangers. The regenerators are switched periodically
to reverse the direction of fluid flow.

Linde Dual-Pressure System

The Linde dual-pressure system is an alternative to the precooled
Linde-Hampson gas liquefier. With the dual-pressure system the work per
unit mass of gas liquefied is reduced as in the precooled system but the
fraction of gas liquefied decreases. The dual-pressure cycle is shown
schematically in Fig. 6.11. It consists of high-pressure and low-pressure
compressors, a three-way recuperative heat exchanger, intermediate-
pressure Joule-Thomson expansion valve and separator, and a low-pressure
Joule-Thomson expansion valve and separator.
The cycle operates as follows. The low-pressure return stream and
fresh make-up gas combine and enter the low-pressure compressor at (1).
After compression to the intermediate pressure the gas passes through the
low-pressure aftercooler, mixes with the intermediate-pressure return
stream, and enters the high-pressure compressor at (2). After compression
the gas passes through the aftercooler, the three-way recuperative heat
exchanger and then expands to the intermediate pressure, (5). The vapor,
at (7), and intermediate pressure return back through the recuperator as
the return stream forming part of the inlet to the high-pressure compressor
at (2). The saturated liquid at intermediate pressure, (6), expands through
a second Joule-Thomson valve to (8) and, in the low-pressure separator,
is separated to the useful product liquid stream at (9), and the low-pressure
return stream at (10).
The savings in work gained by the dual-pressure Linde cycle is shown
in Fig. 6.12, adapted from Barron (1966). It shows the work to produce
a unit mass of liquid as a fraction of the work to produce the same mass
in a simple Linde-Hampson cycle. Two curves are given for different ratios
Joule-Thomson Cooling Systems 281

LOW PRESSIJRE COMPRESSOR MAKE-UP CHARGE

E~TRDPY

LOW PRESSURE
SEI'jiRATOR ~----•1

LIQUID PRODUCT

Fig. 6.11. Schematic diagram of gaseous liquefier using the Linde dual-pressure system.

Fig. 6.12. Ratio of the work required per FRACTION OF GAS RECIRCULATED
unit mass liquefied in the Linde dual- AT INTERMEDIATE PRESSURE

pressure cycle and the simple Linde-

Hampson cycle. The ratio is shown as a O 2. MAXIMUM
MINIMUM
function of the intermediate pressure for INIT IAL TEMPERATURE

two different values of the fraction of gas FLUID AIR

recirculated at the intermediate pressure 0 8 16 20

(abstracted from Barron, 1966). INTERMEDIATE PRESSURE (MPo)

282 Chapter6

of gas recirculated at the intermediate pressure. Barron assumed an air

liquefaction system operating between pressures of 0.1 and 20 MPa (15 to
3000 psi). With 80% recirculation at the intermediate pressure of 7 MPa
over half the work of liquefaction can be saved by use of a dual pressure
cycle. Under these conditions the liquid yield per unit mass of gas com-
pressed to the maximum cycle pressure is reduced to about 7 5% of the
yield attained per unit mass of gas compressed in a simple cycle.

The Rietdijk Expansion-Ejector

An interesting variant of the Linde dual-pressure system was explored
by Rietdijk (1966) using a single high-pressure compressor and an
expansion-ejector to accomplish compression of the low-pressure fluid to
the intermediate pressure. A schematic diagram of a liquefaction system
incorporating the Rietdijk expansion-ejector is given in Fig. 6.13 repro-
duced from Haarhuis (1967). The system operates as follows. Fluid enters
the compressor at intermediate pressure P 2 and is compressed to P 1 • It
then enters the recuperative heat exchanger equipped with precooling
systems if necessary and is cooled to a low temperature. The fluid expands
in a throttled expansion in the expansion-ejector device to the intermediate
pressure, P 2 , and experiences a reduction in temperature to T 2 • The

Prl!cool,ng
stations

Fig. 6.13. Schematic diagram of a

gaseous liquefaction system incor-
porating a Rietdijk expansion-
ejector (after Haarhuis, 1967).
Joule-Thomson Cooling Systems 283

reduction in temperature may be sufficient to cause the fluid to partially

condense and collect in the separator identified in Fig. 6.13 as vessel C1•
The saturated liquid at pressure P 2 passes through a low-temperature
recuperative heat exchanger and expands through a Joule-Thomson valve
(identified as E.V. in Fig. 6.13) to the minimum cycle pressure P 3 and
temperature T 3 • The saturated liquid gas accumulating in vessel Cn at these
conditions is the useful product of the cycle. The saturated vapor at pressure
P 3 returns through the low-temperature recuperative heat exchanger and
is compressed in the expansion-ejector to the intermediate pressure P 2 •
Haarhuis (1967) has indicated that a Rietdijk expansion-ejector system
results in reduction of the compressor volume of 2.5 times, saving in the
size of heat exchangers and the compressor power consumption compared
with a conventional Joule-Thomson liquefier.
Haisma and Roozendaal (1967) have given further theoretical and
experimental aspects of expansion-ejector performance. Prast (1968)
described a helium liquefaction system utilizing a double expansion A-20
Philips Stirling cooling engine and the Rietdijk expansion-ejector.

PERFORMANCE COMPARISON OF GAS

LIQUEFACTION SYSTEMS

Figure of Merit
To provide the means of comparison of gas liquefaction systems the
figure of merit is sometimes used. This is defined as the theoretical minimum
work required to liquefy unit mass of gas compared with the actual work
requirement of a system: FOM = Wid/ Wac· The figure of merit can range
from zero to unity and indicates how closely a system approaches ideal
operation.

Ideal Liquefaction System

Theoretical minimum work to liquefy a gas is estimated assuming the
gas is liquefied according to the process shown in Fig. 6.14. This consists
of a compressor receiving gas at ambient pressure and temperature, condi-
tion (1). The gas is compressed isothermally to (2) and the heat of compress-
ion Oc is rejected from the system. The compressed gas expands isentropi-
cally in an engine to (3) where it exists as saturated liquid.
The liquefaction process is the first two processes in the Carnot cycle
discussed in Chapter 2. The process of liquefaction is ideal thermodynami-
cally but not suitable for practical use. The principal reason is that the
pressure at the end of the isothermal compression is impossibly high (of
284 Chapter6

CHARGE

ENTROPY

Fig. 6.14. Schematic diagram of

ideal gas liquefaction system.

the order of 104 MPa for air), and far exceeds the capacity of any compress-
ion equipment existing or in prospect. Nevertheless the process provides
a useful basis on which to compare theoretical and practical liquefaction
systems.
Using the above concept for gas liquefaction Barron (1966) calculated
the ideal minimum work requirement for different fluids. His results are
reproduced in Table 6.2.

Performance of Linde-Hampson System with Different Fluids

Barron (1966) calculated the performance details of Linde-Hampson
gas liquefaction systems for a number of different gases. To provide a
common basis of comparison all the cycles were assumed to operate over
the pressure range 0.1 to 20 MPa (15 to 3000 psi) with isothermal compres-
sion at 294 K (70°F) and no pressure drop or loss in recuperator effectiveness.
The results are reproduced in Table 6.3 arranged in order of success-
ively increasing temperatures. The data include the critical pressure and
temperature, the mass of gas liquefied per unit mass of gas compressed
the ("liquid yield"), the work required per unit mass of gas compressed
and per unit mass of gas liquefied, and the figure of merit for the particular
conditions assumed; the figure of merit varied from 0.074 for fluorine to
a maximum of 0.600 with carbon tetrafluoride, CF 4 • Hydrogen, helium,
and neon are not included in Table 6.3 because they cannot be liquefied
in the simple Linde-Hampson cycle.
Joule-Thomson Cooling Systems 285

Table 6.2. Ideal-Work Requirements for Liquefaction of Gases

Beginning at 294 K and Atmospheric Pressurea

Ideal work of
Normal boiling liquefaction
Gas point (K) (kJ/kg)

Helium, He 4 4.2 6650

Hydrogen, H 2 20.3 11,594
Neon, Ne 27.1 1291
Nitrogen, N 2 77.3 742
Air 78.7 710
Carbon monoxide, CO 81.6 742
Fluorine, F 2 85.3 479
Argon, A 87.2 465
Oxygen, 0 2 90.2 613
Methane, CH 4 111.6 1044
Carbon tetrafluoride, CF4 145 171
Ethane, C2H6 184.5 330
Ammonia, NH 3 239.1 232

"After Barron (1966).

Comparison of Liquefaction Systems for Air

Barron (1966) assembled an interesting tabulation of ideal, theoretical,
and observed liquefaction systems for air. Some of the material has been
reproduced in Table 6.4. For all the cases the minimum pressure was
assumed to be atmospheric (0.1 MPa) and the ambient temperature 294 K
(70 F). For the theoretical cases the maximum pressure was taken to be
20 MPa (3000 psi). In some cases the compression process was assumed
to be isothermal and the recuperative process to be 100% effective. In
other cases, the compressor efficiency was assumed to be 70% and the
recuperator effectiveness to be 95%. Conditions for the observed systems
were not included by Barron but likely correspond reasonably closely to
the conditions of operation of the theoretical systems.
The first entry in Table 6.4 is the ideal reversible liquefaction process
discussed above. The second group involves three entries for simple Linde-
Hampson systems. The first entry assumes isothermal compression and
perfect recuperation. The second assumes nonisothermal compression and
nonperfect recuperation. The third gives the observed results of an actual
Linde-Hampson air liquefier. The excellent correlation between the
theoretical result and the observed result should be noted. The next group
of entries refers to the precooled Linde-Hampson cycle. The precooling
temperature was assumed to be 227 K. As before the entries included an
idealized cycle and another where allowance was made for compressor
 N
QC
~

Table 6.3. Performance of the Linde-Hampson System Using Different Fluids P1 =0.1 MPa, T 1 = 294 K, p 2 = 20 MPa, T2 =
294 K; Heat-Exchanger Effectiveness, E =1.00; Compressor Overall Efficiency Ylc =100% a

Normal Work required Work required

boiling Critical Critical Fraction of per unit mass per unit mass Figure of
point temperature pressure gas liquefied compressed liquefied merit
Fluid (K) (K) (MPa) (liquid yield) (kJ/kg) (kJ/kg) FOM = W;d/ Wac

Nz 77.3 126 3.35 0.087 85.1 976 0.141

Air 78.7 133 3.87 0.086 83.0 967 0.137
co 81.6 132 3.45 0.092 85.5 931 0.148
Fz 85.3 118 5.5 0.053 63.2 1195 0.074
A 87.2 150 4.83 0.124 58.9 474 0.182
Oz 90.2 154.6 5.01 0.110 73.5 670 0.197
CH4 111.6 190.7 4.58 0.209 141.4 675 0.288
CF4 145 227.5 3.69 0.303 16.0 53 0.600
CzH6 184.5 305.5 4.85 0.505 61.7 122 0.502

a After Barron (1966).

n
=-
i
~
~
Joule-Thomson Cooling Systems 287

Table 6.4. Comparison of Liquefaction Systems Using Air as the Working Fluid
T 1 = 294 K and p 1 =Atmospheric Pressure, Ylc =Overall Compressor Efficiency,
E = Heat Exchanger Effectiveness, i = Fraction of Mass Flow Recirculated a

Fraction of Work per unit Figure of

gas liquefied mass liquefied merit
Air liquefaction system (liquid yield) (kJ/kg) W;d/Wac

1 Ideal reversible system 1.000 132.4 1.000

2 Linde-Hampson system, p 2 =
200 atm, 1Jc = 100%, e = 1.0 0.086 967.5 0.137
3 Linde-Hampson system, Pz =
200 atm, 1Jc = 70%, e = 0.95 0.061 1965 0.068
4 Linde-Hampson system, observed 1909 0.070
5 Precooled Linde-Hampson system
p 2 = 200 atm, T 3 =-50 F, 1Jc =
100%, e = 1.00 0.179 414 0.320
6 Precooled Linde-Hampson system
Pz = 200 atm, T3 =-50 F, 1Jc =
70%, e =0.95 0.158 683 0.194
7 Precooled Linde-Hampson system
observed 1032 0.129
8 Linde dual-pressure system p3 =
200 atm, p 2 = 60 atm, i = 0.8,
1Jc = 100%, e = 1.00 0.060 507 0.261
9 Linde dual-pressure system, p 3 =
200 atm, p 2 = 60 atm, i = 0.8,
1Jc = 70%, e = 0.95 0.032 1479 0.090
10 Linde dual-pressure system,
observed 1173 0.113
11 Linde dual-pressure system
precooled to -50 F, observed 662 0.201
12 Cascade system, observed 602 0.221

"After Barron (1966).

efficiency and recuperator ineffectiveness. Agreement between the calcu-

lated and observed results was not as good as the case of the precooled cycle.
The next group refers to the Linde dual-pressure cycle. Again three
entries are included, arranged as before. There was poor correlation
between the observed and calculated results.
The final entries present the observed results in terms of liquid yield
and work per unit mass liquefied for a Linde dual-pressure cycle with
precooling to 227 K, and a cascade, presumably of the "outer cascade
cycle" type although the details of operation and the fluids used were not
included by Barron. It is interesting to compare the observed figures of
merit for the various systems. The simple Linde-Hampson system had the
worst performance with a figure of merit of only 0.070. With precooling
thi~ improved to 0.129. The Linde dual-pressure system had a figure of
merit of 0.113. The combination of precooling and dual pressure improved
288 Chapter6

the figure of merit to 0.201. This compared favorably with the cascade
system having the best figure of merit, 0.221. The simple Linde-Hampson
system therefore consumed about three times the work per unit mass
liquefied of the thermodynamically superior but very much more compli-
cated cascade or precooled, dual-pressure Linde cycle.

MINIATURE JOULE-THOMSON COOLERS

Hymatic Coolers
Miniature Joule-Thomson cooling systems have been developed for
electronic applications, particularly infrared thermal detectors. The
development was carried out in Great Britain by the Hymatic Engineering
Co. Ltd., Redditch, Worcester with Government sponsorship through the
Royal Radar Establishment, Malvern, Worcester. Accounts of this work
have been given by Mcinroy (1967), Parkinson (1959, 1967), and Nicholds
(1968, 1970).
Mcinroy described the construction and gave the performance data for
both liquid air and liquid hydrogen/liquid nitrogen Joule-Thomson
miniature liquefiers. Figure 6.15 is a cross section of the miniature air
liquefier described by Mcinroy as follows:
The miniature liquefier is enclosed by a representative cell for producing
temperatures of 79 K (air) or 77 K (N 2 ) in the region of the detecting element
of the cell. Liquid is produced in the collecting chamber by the Joule-Thomson
effect from high pressure air (60 to 400 atm) applied to the inlet.
The weight of the liquefier is 5.3 g and has a heat exchanger of 7.2 mm
diam x 25 mm long mounted on a hollow mandrel, or former, of Monel the
lower end of which is closed by a brass cap to which the nozzle assembly is
fitted. The cavity thus produced is filled with expanded perlite to reduce heat
leaks by convection. A sintered bronze filter of 2 micron pore size is placed
in the inlet cavity of the former to exclude harmful dust particles.
The original prototype liquefier built at the Royal Radar Establishment
had no nozzle assembly or inlet filter. Instead the end of the finned tube
heat exchanger was squeezed together to a fish-tail shape which restricted
the flow to the desired volume. This operated satisfactorily but required
very pure air for its operation as the fine slit became easily blocked.
Contaminant gases such as C0 2 at low vapor pressure sublimed at the cold
end to form minute particles which blocked the exist orifice. The addition
of an inlet filter and the porous nozzle effected a substantial improvement
in reliability, but the requirement for a very high degree of gas purity
remained. Indeed the problem remains today as perhaps the single greatest
impediment to the wider use of miniature Joule-Thomson coolers.
Joule-Thomson Cooling Systems 289

INLET

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0 FINNED
DEWAR 0
FLASK 0 >£AT
0 HAl«
0 COILS
0
0
0
0
0
,0
0
0
0
0

HOLLCW
Fig. 6.15. Cross section of miniature Joule- SCREW
>-.OULE
Thompson air or nitrogen liquefier (after
Mcinroy, 1967).

Mcinroy described one successful arrangement for cleaning the gas

compressed to 40 MPa (6000 psi) in reciprocating, oil-lubricated compress-
ors as follows:
1. Air compressor, oil lubricated crankcase and upper cylinders, 400 atm,
water cooled cylinders and intercoolers (max. cooling water temperature
50°C), maximum compression ratio in any one stage 4.9, minimum distilla-
tion temperature of lubricating oil 230°C.
2. Mechanical oil and water trap for collection of oil and water droplets.
3. Soda-lime bed for chemical absorption of C0 2 and S0 2 traces. This must
be first sorbent bed because water is formed.
4. Activated charcoal for bulk hydrocarbon removal in wet state.
5. Activated alumina for water adsorption. Preferably arranged to be reacti-
vated in situ.
290 Chapter6

6. Activated charcoal for remaining hydrocarbon removal in dry state.

7. Molecular sieves, 13X for remnants of H 2 0, C0 2 , S0 2 and hydrocarbons.
8. 5 micron filter.
Air produced by this process has been called "Pure Air" and conforms
to the following specification:
Maximum allowable Minimum obtainable

Water content 1 wpm (Dew Point 0.1 wpm (Dew point

minus 77°C) minus 90°C)
C0 2 content 1.2 wpm 0.3wpm
Hydrocarbon content 0.045 wpm 0.03 wpm

Performance of Miniature Joule-Thomson Coolers

Typical performance characteristics for both air and hydrogen
liquefiers were given by Mcinroy and are reproduced in Figs. 6.16 and
6.17. Each series of three diagrams per figure shows the gas flow, time to
achieve liquefaction, and the gross cooling power achieved. The heat leak
to the system must be subtracted to obtain the net available cooling power.
Mcinroy gave the heat leak in the configuration shown as 0.4 W at 22 C
ambient temperature for the air liquefier, and 0.13 W for the hydrogen
liquefier when the nitrogen/air precooler was operating.
When argon was used rather than air as the working fluid in the system
shown in Fig. 6.15 an appreciable improvement in performance was noted.
At an ambient temperature of 22°C the time to achieve liquefaction
decreased to 0.7 of the time with air. The cooling power was 50% greater
than with air. With an increase in the ambient temperature to 75°C the
improvement in cooling power was even greater to 2.2 times the cooling
power with air. Such a result as this might of course be anticipated from
Table 6.3, which clearly shows the advantage of argon compared with air
as a working fluid in Joule-Thomson systems.
g, 6
..
~ !00

.
'9
/
. 3
2 •
\
2 7

....
u

/
~
w
I .....
....
(.)

v
~

·e ~·
"...z
v ~ !SO
0

\
/
:::;
~... g. 1-- L s•
z

... v
'
I

.
2
m
I
(.)

1'-
..
....0
" 0 /
!w
~

10 20 30 40 (,!) 0 10 20 30 40 10 20 30 40

AIR INLET PRESSURE (MPo)

Fig. 6.16. Performance of miniature Joule-Thomson nitrogen liquefier (after Mcinroy, 1967).
Joule-Thomson Cooling Systems 291

2 o r--c=F=:t::;::r=-r-, "20
o;;
160

!.,.:;:~ /
AIR

•
I
v
!
• ot CONSTANT 20MPo
1
~ I 51---t--+--t-f---;t'l ~1!5
J 120
I
0.10
HYO~OGEf>~V .!
a:
"'~10 I .
u
~

80
TIME TO H2
t IQUEFY
FROM TIME OF
PRODUCTION OF
LIQUID AIR
:~

",g Q. II \
31:
~ 0 51---t--+--t-+----l
"z
~0!5 0
1\_
0
..."' u

" "'0"'
0 04 08 12 ~ 0 04 0.8 12 0 04 08 12

HYDROGEN INLET PRESSURE ( MPa)

Fig. 6.17. Performance of miniature Joule-Thomson hydrogen liquefier (after Mcinroy, 1967).

Self-Regulating Miniature Joule-Thomson Coolers

The development of fixed orifice miniature coolers of the type
described by Mcinroy above was followed by the introduction of an impor-
tant development, the self-regulating miniature Joule-Thomson cooler. In
the self-regulating system the expansion orifice is throttled when sufficient
liquid has been generated to provide a stable pool of liquid to cool
the detector. With a cooler supplied from compressed gas bottles rather
than directly from a compressor the self-regulating system gave an
order of magnitude reduction in the total system weight because of reduced
gas flow.
The self-regulating cooler can cool down more rapidly since the initial
gas flow can be very much higher than normal. Another important advan-
tage is that the problem of contamination is much reduced. Solid con-
taminants at the expansion nozzle tending to obstruct the flow simply cause
the self-regulating device to open so as to maintain the flow necessary to
achieve the desired cooling. The self-regulating cooler has significantly less
noise than the fixed orifice Joule-Thomson coolers which are themselves
greatly advantageous in this regard compared with cooling engines of the
integral Stirling type.
Nicholds (1970) gave performance details and the test protocol for
self-regulating Joule-Thomson coolers. Stephens (1970) has given further
details of the self-regulating system including the operation of the regulatory
flow orifice needle valve activated by a vapor pressure thermometer and
metal bellows arrangement. Buller (1971) has presented interesting data
on the performance of similar self-regulating Joule-Thomson miniature
coolers constructed at the Hughes Santa Barbara Research Center, Goleta,
California.
292 Chapter6

l TOTAL HEAT LOAD IW ot 78K

RESERVOIR• I x 10-Sm:S
PRESSURE· 40 MPo
MODEL No. MAC 222-200

3 I"'-
"'-...
z
0
2 ~
~ ~
"'
:::>
0 ~
I
..... Fig. 6.18. Operating time for miniature
Joule-Thomson self-regulating cooler
(Hymatic Model No. MAC 222-200) as
260 280 300 320 340 360 380 a function of ambient temperature (after
OPERATING I STORAGE TEMPERATURE Stephens, 1970).

Complete Refrigeration System

Little work on miniature Joule-Thomson coolers has been reported
from United States sources but one paper by Stephens (1970) is of interest.
In this he disclosed that the Bendix Corporation of Davenport, Iowa had
a licensing agreement with the British Hymatic Engineering Company to
manufacture Hymatic coolers in North America. One interesting charac-
teristic given by Stephens (1970) is reproduced in Fig. 6.18. This shows
the duration of operation of a nominal1 W at 77 K self-regulating Hymatic
Joule-Thomson cooler (Model No. MAC 222-200), energized from a
compressed gas bottle having a capacity of 1 x 10-3 m 3 and containing
nitrogen compressed, at the start, to 40 MPa (6000 psi). The duration of
operation is shown as a function of the ambient operating temperature. A

::
COOLER/CELL ASSEMBLY - . .

I
g-~-o '

I
AIR CJ SOLENOIDVAL\.~• CLEANER
I

AVERAGE FREE
DURING
CHANGING:
FREE AIR
THROUGHPUT
BETWEEN CJ GASCLEA~=WI.1. I
AIR DELIVERY
99 dm'/min
REACTIVATIONS
283m' ~ SELF SEALING COUPLING AIRCRAFT SHOE I MISSILE .I
(3.5ft'/min). (10,000ft') ~ ' PYLON
i.e., 7 BOTTLES i.e., 400
PEA HOUR BOTTLES

STOP VALVE AND BURSTING DISC Q T O R A G E BOTTLE

VOLUME 1.64 dm 3 (1 00 in 3 )
MAXIMUM PRESSURE 422 atu (6000 lbf!in')
AT zoac: 1 COMPRESSOR HOUR = 29.75 MISSILE HOURS PER WATT 119 x 1 o• MISSILE
HOURS PER WATT PER CLEANING UNIT REACTIVATION

Fig. 6.19. Schematic arrangement for complete refrigerating system to provide miniature
Joule-Thomson cooler system for heat-seeking air-to-air missiles aboard aircraft (after
Stephens, 1970).
Joule-Thomson Cooling Systems 293

reduction of 40 K from 360 to 320 K doubled the duration from 1 to 2 hr.

Further reduction of 30 to 290 K increased the duration to 3 hr. This result
might be anticipated from our previous acquaintance with the precooled
Linde-Hampson cycle.
Stephens (1970) included an interesting description of a complete
refrigeration system based on miniature Joule-Thomson coolers suitable
for heat-seeking, air-to-air missiles on board aircraft. A schematic arrange-
ment of the system in Fig., 6.19 is reproduced from Stephens (1970). The
following operational advantages were listed:
Operating temperature -40°C to +70°C,
Rapid cooldown, less than 30 s,
Minimal microphony,
Minimal mass on guidance gimbals,
Long standby time,
Instant readiness,
Short logistic supply train,
Simple servicing,
Transportable support equipment,
Simple power supplies.

Air Products Coolers

Geist and Lashmet (1960, 1961) have presented data on the design
and performance of miniature Joule-Thomson coolers developed by Air
Products Inc. of Allentown, Pennsylvania. The systems described included
open cycle systems with hydrogen to provide t to 2 W of useful refrigeration
in the range 15 to 60 K. Refrigeration to cool the hydrogen below the
inversion temperature was obtained by an auxiliary Joule-Thomson system
using compressed nitrogen or by liquid nitrogen. Closed cycle hydrogen
and neon systems were also described. These were intended for relatively
long-term operation. The closed cycle systems incorporated a two-stage
nonlubricated enclosed compressor with electric drive.
Later Currie (1967) described a small open cycle Joule-Thomson
cooler using helium as the working fluid to generate a useful refrigerating
capacity of 0.5 W at 4 K. The unit was a cascade system using gaseous
hydrogen Joule-Thomson expansion to liquefy hydrogen for precooling
the helium. Liquid nitrogen was used to precool both the high-pressure
hydrogen and helium gases en route to the expansion nozzles through
compound recuperative heat exchangers. Kessler (1970) of Deutsches
Elektrones-Synchrotron, Hamburg, West Germay described a refriger-
ation system incorporating an Air Products Joule-Thomson unit to provide
refrigeration for a small hydrogen "target" for use with nuclear research.
294 Chapter 6

JOULE-THOMSON EXPANSION COMBINED WITH

COOLING ENGINES

In all of the above we have explored various ways and means to utilize
Joule-Thomson expansion alone to produce low-temperature refrigeration
and gas liquefaction. Many other systems exist which incorporate both
cooling engines and a Joule-Thomson expansion system. One such combi-
nation mentioned in discussion of the Reitdijk expansion-ejector system
was described by Prast (1968). This combines a double-expansion Stirling
engine as the two-stage precooler of a helium liquefaction system. Haarhuis
(1967) described a similar system incorporating two Philips double-
expansion Stirling engines.
Chapter 5 includes the description of a helium liquefier built by A.
D. Little Inc. combining a three-stage Gifford-McMahon cooling engine
with a Joule-Thomson expansion process. Further development of that
concept was described by Stuart and Hogan (1965).
Chapter 7 describes many Claude cycle cryocoolers that include Joule-
Thomson expansion valves particularly where the fluid is to be liquefied.
It is customary for the fluid to be expanded partially in the engine and
partially in the Joule-Thomson valve. Expansion in the engine occurs
isentropically and achieves a greater temperature difference than isenthalpic
(JT) expansion over a given pressure ratio. However, expansion of two-
phase fluids in engines presents problems, so that the preferred combination
is partial expansion in an engine to approach the saturated vapor condition
and then Joule-Thomson expansion to achieve liquefaction.

REFERENCES

Baehr, H. D. (1963). "On the Thermodynamics of the Cold-Air Cycle with Throttling." Proc.
XI Int. Congr. of Ref./Prog. in Ref. Sci. Techno/. 1, 319-328, Munich, Germany (Pubs.
Pergamon Press, Oxford and Vertag Muller, Karlsruhe).
Barron, R. (1966). Cryogenic Systems. McGraw-Hill Book Co., Toronto.
Buller, J. S. (1971). "A Miniature Self-Regulating Rapid-Cooling Joule-Thompson Cryostat."
Adv. Cryog. Eng. 16, 205-213.
Carbonnell, E., Chovet, P., Johannes, C., Marinet, D., and Solente, P. (1972). "Refrigerator
without Moving Parts at Low Temperature Able to Cool Down to 90-100 K." Proc. 4th
Int. Cryo. Eng. Conf., Eindhoven, pp. 68-70.
Currie, R. B. (1967). "A Joule-Thomson Laboratory Expander." Adv. Cryo. Eng.12, 557-563.
Daunt, J. G. (1956). The Production of Low Temperatures down to Hydrogen Temperatures.
Handbuch der Physik Bd IV, Springer Verlag, Berlin.
Geist, J. M., and Lashmet, P. K. (1960). "Miniature Joule-Thomson Refrigeration Systems."
Adv. Cryog. Eng. 5.
Geist, J. M., and Lashmet, P. K. (1961). "Compact Joule-Thomson Refrigeration Systems
15-60 K." Adv. Cryog. Eng. 6.
Joule-Thomson Cooling Systems 295

Haarhuis, G. J. (1967). New Type Helium Liquefier." Proc. XII Int. Cong. of Ref./Prog., in
Ref. Sci. Techno/. 1, 121, Madrid (llR Paris).
Haisma, J., and Roozendaal, K. (1967). "Investigation of the Behaviour of an Expansion-
Ejector in the Low Temperature Region beyond the A- Transition of Helium." Proc. XII
Int. Cong. of Ref./Prog., in Ref. Sci. Techno/. 1, 111-210, Madrid (llR Paris).
Hampson, W. (1895). British Patent 10165.
Joule, J.P. (1852). Sci. Pap. 2, p. 216.
Joule, J.P., and Thomson, W. (1852). Phil. Mag. 4, 481.
Keesom, W. H. (1933). Commun. Phys. Lab. Univ. Leiden, suppl. 76a.
Kessler, G. (1970). "Joule-Thomson Hydrogen Refrigerator Target." Adv. Cryog. Eng. 15,
443-446.
Linde, H. V. (1897). German Patent 88824 [see also Z. ges Kiilteind 4, 23 (1897); and the
Engineer, Nov. 13 and 20 (1896)].
Mann, D. B., Bjorkland, W. R., Macinko, J., and Hiza, M. J. (1960). "Design, Construction
and Performance of a Laboratory Size Helium Liquefier." Adv. Cryog. Eng. 5, 346-353.
Mcinroy, J. (1967). Miniature Joule-Thomson Coolers." Proc. XII Int. Cong. of Ref./Prog.
in Ref. Sci. Techno/. 1, 59-68, Madrid (llR, Paris).
Nicholds, K. E. (1968). "Low Temperature Devices for Laboratory Operation." Proc. 2nd
Int. Cryo. Eng. Conf., pp. 65-66, Brighton Iliffe, Sci. and Tech. Pub!., Guildford, U.K.
Nicholds, K. E. (1970). (a) "Performance of Self-Regulating Joule-Thomson Minicoolers,"
pp. 277-282. (b) "Miniature Cryogenic Cooling Systems for an Upper Atmosphere
Infrared Research Programme," pp. 283-286, Proc. 3rd Int. Cryog. Eng. Conf., Berlin,
Iliffe Sci. and Tech. Pub!., Guildford, U.K.
Parkinson, D. H. (1959). "Some Problems in the Design of Helium Liquefiers Based on the
Joule-Thomson Effect." Proc. X Int. Cong. of Ref./Prog. in Ref. Sci. Techno/. 1, 53-57
(Pub!. Pergamon Press, Oxford).
Parkinson, D. H. (1967). "Miniature Refrigeration Systems-A Review." Proc. XII Int. Cong.
of Ref./Prog. in Ref. Sci. and Tech., Vol. 1, 69-77, Madrid (11R, Paris).
Prast, G. (1968). "A 3.5 K Refrigerator Based on the Three-Space Stirling Refrigerator."
Proc. 2nd Int. Cryo. Eng. Conf., Brighton, Iliffe Sci. and Tech. Pubs. Ltd., Guildford, U.K.
Rietdijk, J. A. (1966). "The Expansion-Ejector-A New Device for Liquefaction and
Refrigeration at 4 K and Lower." Included in Liquid Helium Technology, Bull. IIR,
Annexe 1966-65, p. 241.
Stephens, S. (1970). "A Self-Regulating Miniature Joule-Thomson Refrigerator." Appl.
Cryog. Techno/. 3.
Stuart, R. W., and Hogan, W. H. (1965). "A Small Helium Liquefier." Adv. Cryog. Eng. 10,
62-68.
Chapter 7

Claude and Joule-Brayton Systems

INTRODUCTION

Here we consider a family of cryocoolers using recuperative heat exchangers

and achieving cryogenic temperatures by expanding gas in an engine thereby
producing work. When the gaseous working fluid is to be liquefied it is
customary for the final stage of expansion to be done with a Joule-Thomson
expansion valve. Systems with expansion engines and a final Joule-
Thomson expansion are called Claude cycle liquefiers. When the system is
used as a refrigerator rather than a liquefier the final Joule-Thomson stage
is not included and all the expansion occurs in expansion engines. These
are called Joule-Brayton cycle refrigerators.
This family exists i'1 endless variety. It is widely used for gas liquefac-
tion, separation, dnd processing systems and less widely for simple refriger-
ation applications where regenerative systems are preferred. Proper con-
sideration of Claude systems has been deferred to a companion volume
devoted to systems for gaseous l;quefaction, separation, processing and
refrigeration in large and intermediate capacities involving turbomachinery.
Here the field of interest has been limited to cryocoolers of intermedi-
ate or less capacity usir.g reciprocating expanders. The important areas of
gaseous separation and processing associated with large-scale gaseous
liquefaction systems have been ignored. Such systems are fundamental to
the production of industrial gases, principally oxygen for welding, steel-
making, and hospital use and nitrogen for food-processing. The liquefaction
and processing of natural gas is necessary to the separation of helium and
the intercontinental transport of liquid natural gas as an energy fuel.

ISENTROPIC VS. ISENTHALPIC EXPANSION

Practically all cryocoolers involve expansion of the working fluid from
a high pressure to a low pressure. The working fluid is usually a gas but
297
298 Chapter7

T T

s s
(a) ISOTHERMAL {bl ISENTHALPIC

/
8

s s
(c) ISENTROPIC (d I POLYTROPIC

Fig. 7.1. Expansion processes represented on the T-S plane.

can be a liquid (gas previously liquefied) or wet vapor (gas cooled below
the critical temperature with liquid droplets entrained in it). The expansion
can approximate various idealized thermodynamic processes, namely,
isothermal, isenthalpic, isentropic and polytropic. These are represented
on the temperature-entropy diagram (T-S plane) as shown in Fig. 7.1.
Assume first the gas is at some high pressure and low temperature
TA. The gas expands to a low pressure and heat is supplied to bring the
temperature back to the initial temperature TA. An apparatus to carry out
such comparative work is shown in Fig. 7.2. It consists of a piston reciprocat-

CYLINDER HEAD
I VE

Fig. 7 .2. Idealized expansion engine.

Claude and Joule-Brayton Systems 299

ing in a cylinder closed at one end. A valve is connected to the cylinder

through which fluid is admitted or expelled from the cylinder. The cylinder
head is assumed to be perfectly insulating or perfectly heat conducting as
required. The walls of the cylinder are perfectly insulated and the piston
is assumed to be without mass and to experience no friction in moving
about in the cylinder.
For the isothermal process represented in Fig. 7.1a assume that the
cylinder is filled with gas at condition A. The filling process occurs when
the piston is at the top of its stroke (volume between the piston and cylinder
head is least). When the piston moves tlown the cylinder the gas expands
to a low pressure. The cylinder head is assumed to be perfectly conducting
and as the gas expands the temperature tends to decrease so that heat is
transferred through the cylinder surroundings to maintain the temperature
constant. This heat abstracted, the refrigeration effect, is represented by
the shaded area A-B-C-D in Fig. 7.1a. Work is done on the piston by
the gas. The amount of work is exactly equal to the heat abstracted from
the surroundings.
For the isenthalpic process represented in Fig. 7.1b assume that, with
the cylinder empty of gas, the piston is moved to the bottom of the stroke
so the volume between the piston and the cylinder head is a maximum.
The inlet valve is opened slightly and working fluid slowly enters the
cylinder. The pressure increases and eventually reaches the final required
pressure, point E, Fig. 7)b. Gas passing the inlet valve is said to be
throttled. A throttled expansion occurs at constant enthalpy and is categor-
ized as a Joule-Thomson expansion. If the temperature TA before
expansion is below the inversion temperature (see Chapter 6) then on
expansion a decrease in temperature (toTE, Fig. 7.1b) will occur. Assume
the cylinder head becomes perfectly conducting and the fluid is heated to
the final condition, TB, by heat transfer through the cylinder head from
the surroundings. The heat abstracted is represented by the area E-B-C-D
on Fig. 7.lb, which is less than area A-B-C-D on Fig. 7.1a. For gas
experiencing Joule-Thomson expansion the refrigerating effect is about
one third of that with isothermal expansion even though the fluid started
and finished at the same states. In the isenthalpic process no work was
done; the piston remained stationary at the lower dead-point and the
cylinder volume was constant.
For the isentropic process, Fig. 7 .lc, we assume the same sequence as
the isothermal case. The cylinder is filled with gas at state A and expanded
to a low pressure. The only difference is that the cylinder head is perfectly
insulated during expansion and no heat is transferred during expansion
so the entropy remains constant (LlS = JLiQ/T = 0). Work is done on the
piston and the temperature is reduced as shown by the process A-E in
300 Chapter7

Fig. 7.1c. The cylinder head then becomes perfectly conducting and heat
is transferred from the cylinder surroundings to heat the gas to the final
temperature, TB. The shaded area E-B-C-D in Fig. 7.1c is larger than
the isenthalpic process and smaller than the isothermal process.
The polytropic process of Fig. 7.1d follows a similar pattern to the
isentropic with the difference that the temperature at the end of the
expansion Tk, is greater than at the end of isentropic expansion TE. This
difference may arise from two causes, limited heat transfer during the
expansion (usually the case with reciprocating engines but not turbines),
and irreversibilities (principally friction) which occur during expansion
(particularly in turbines) to increase entropy and, hence, the fluid tem-
peratures. The effect of limited heat transfer and irreversibilities during
expansion is to decrease the potential refrigerating effect compared with
the isentropic process. The potential refrigerating effect is the shaded area
E'-B-C' -D in Fig. 7 .1d, smaller than the area for the isentropic process
by area E-E'-C'-C.
Figure 7.1 shows clearly that in terms of refrigerating effect the order
of preference of the four processes is isothermal, isentropic, polytropic,
and isenthalpic.
In practice isothermal expansion cannot be achieved. It would require
infinite rates of heat transfer or the expansion to proceed infinitely slowly.
It is customary to adopt an isentropic process as the reference for expansion
engines. This cannot be achieved in practice and realistic cycle analyses
adopt an arbitrary value (60% to 85%) for the isentropic efficiency to
account for the polytropic process.
The same work is necessary to compress the gas regardless of the
subsequent expansion. Since the Joule-Thomson process produces so little
refrigerating effect compared with isentropic expansion it is to be expected
that Joule-Thomson cryocoolers will be less efficient than those incorporat-
ing expansion engines. The principal reason for using Joule-Thomson
expansion is simplicity. It is much easier and cheaper to provide a simple
throttle valve than an expansion engine with means to absorb the work
produced. All small domestic refrigeration systems operating on the vapor-
compression cycle have JT expansion valves.
In larger refrigerating and cryogenic cooling systems the additional
expense of expansion engines can be justified by the improved efficiencies.
Even there Joule-Thomson expansion systems are invariably used for the
final stage of expansion when there is partial liquefaction of the working
fluid. Two-phase working fluids are difficult to handle in expansion engines.
Daunt (1956) discussed isenthalpic and isentropic expansion in
more formal thermodynamic terms than the intuitive-descriptive method
followed above. The following is abstracted from Daunt's presentation:
Claude and Joule-Brayton Systems 301

... Writing the First and Second Laws of thermodynamics as:

TdS = C"dT-T(:;)" dp (7.1)

we obtain for the case of isentropic expansion (dS = 0) an immediate evalu-

ation for the coefficient of cooling,

as=
aT)
(ap T ( av)
s = C" aT "
(7.2)

This compares with the Joule-Thomson coefficient of cooling:

(7.3)

whence

(7.4)

Asp is increased or T decreased, thereby diminishing v the Joule-Thomson

coefficient aH approaches the reversible isentropic cooling coefficient as.
Moreover as the critical temperature is approached C" tends to an infinite
value; hence in regions near critical, isenthalpic expansions also must be
highly efficient. ...

The favorable environment for Joule-Thomson expansion near the

critical point is evident on reference to a temperature-entropy diagram.
Figure 7.3, a T-S plane for helium, shows two constant-pressure lines, the
liquid vapor envelope, the critical point C, and three isenthalpic curves
A-B, D-E, and F-G. The isenthalp A-B, starting at the lowest initial
temperature, TA, has a high slope not too significantly different from the
isentrope A-H. At higher temperatures expansion over the same pressure

Fig. 7.3. Comparative slope of isenthalps on the

T-S plane. ENTROPY S
302 Chapter7

Fig. 7 .4. Ratio of the isenthalpic-isen-

tropic coefficients of cooling as a function
of pressure at the start of expansion (for
PRESSURE (MP<l) air) (after Daunt, 1956).

range results in a much smaller temperature decrease (isenthalpe D-E).

If the initial temperature, Tp, is above the inversion curve the temperature
will actually increase on isenthalpic expansion.
The relationship between o:H and as can be written [from Eqs. (7.2)
and (7.3)] as

v
(o:H/as)=1-T
(aT)
a;; P (7.5)

This relationship for air as a function of pressure for different values

of temperature T (start of expansion) is given in Fig. 7.4 (after Daunt,
1956). As the temperature reduces, (o:H/as) increases and at 100 K
approaches unity. At higher temperatures the ratio (o:H/as) is a strong
function of pressure, increasing progressively with increase in pressure.
This helps to explain why the Linde dual-pressure cycle is so much more
efficient than the simple Linde-Hampson cycle. In the dual-pressure cycle
the principal mass of fluid is maintained at the intermediate pressure (about
4 to 5 kPa).

CLAUDE CYCLE SYSTEMS

Claude Cycle
A Claude cycle system for gas liquefaction, showninFig. 7.5, includes
(a) compressor equipped with an aftercooler, air or water cooled, (b)
recuperative heat exchangers, designated as high- (ambient), intermediate-,
and low-temperature exchangers, or alternatively upper-, intermediate-,
Claude and Joule-Brayton Systems 303

MAKE-UP
CHARGE
COMPRESSOR

------------

ENTROPY

Fig. 7.5. Schematic diagram of

typical Claude cycle cryocooler
and liquefier.

and lower-stage exchangers, (c) expansion engine, (d) Joule-Thomson

expansion valve, and (e) separator and liquid product reservoir. The system
operates as follows:
i. Recirculated vapor at condition (11) mixes with fresh charge at
condition (12) for the compressor inlet charge at condition (1).
ii. The vapor is compressed to condition (2"). Frequently the process
is assumed to be isothermal [process (1-2)] or isentropic [process
(1-2')]. In practice the process would correspond to the process
(1-2") with reciprocating compressors, or the process (1-2"') with
a rotary compressor. The compression process may be arranged
to occur in two or more stages with intercooling back to ambient
temperature between stages. In Fig. 7.5, the compression process
is shown in two stages, process (1-A) and process (B-C) with
intercooling, process (A-B) between stages. Multistage compres-
sion with intercooling reduces the compression work. The work
of compression decreases as the number of stages increases but
at a progressively diminishing rate. It is rare to find more than
304 Chapter7

four stages of compression in reciprocating systems. Multistage

compression is used for rotary and reciprocating compressors.
iii. Fluid leaving the compressor is cooled to ambient temperature
in the watercooled aftercooler and then passes to:
IV. The high- (ambient) temperature recuperator and is cooled to
condition (3).
v. A fraction of the fluid then enters the expansion engine and
expands. The process of expansion may be isothermal (3-10),
isentropic (3-9), or an intermediate, polytropic process (3-9').
vi. The remaining fluid at high pressure passes through the intermedi-
ate recuperator, is cooled to condition (4), and then enters:
vii. The low-temperature recuperator and is further cooled to condi-
tion (5).
vm. Finally, the fluid expands through the Joule-Thomson valve and
expands isenthalpically to condition (6), a two-phase, saturated
liquid, saturated vapor mixture.
ix. The mixture is separated into the fluid and vapor phases in the
separator.
x. The saturated liquid remains in the separator reservoir and is
drawn off at condition (7) as the liquid product of the system.
Sometimes the saturated liquid, perhaps liquid air or nitrogen, is
not drawn off but remains in the separator reservoir as the pre-
cooler bath for a second Claude, or Joule-Thomson system far
helium, hydrogen, or neon liquefaction.
xi. Saturated vapor leaves the separator at state (8) and enters the
low-temperature recuperator and is warmed to condition (9) by
heat exchange with the downcoming high-pressure stream.
xii. The effluent stream mixes with the exhaust from the expansion
engine to condition (9'). The temperature of the fluid leaving the
low-temperature recuperator will not be exactly equal to that of
the exhaust stream from the expansion engine. Condition (9') is
the state of the mixed streams.
xiii. The combined low-pressure streams now pass through the two
recuperative heat exchangers being heated to a final temperature,
condition (11), and subsequently mix with incoming make-up
charge to provide inlet gas to the compressor.

Strictly speaking, the two separate streams: (a) process (3-9'), and (b)
process (3-4-5-6-8-9') cannot be represented on the T-s plane because
the diagram is constructed for unit mass of the working fluid. It is useful
to represent the processes as followed here, providing it is always recalled
that per unit mass of working fluid compressed, fraction (1- x) passes
Claude and Joule-Brayton Systems 305

through the expansion engine and only x is involved in the actual liquefac-
tion process.

Analysis of Claude Cycles

Claude systems are analyzed by assuming steady state continuous
operation and applying the steady flow energy equation to determine the
fluid enthalpy at critical points in the cycle. Per unit mass of working fluid
entering the compressor:
(1) The compressor work is
We= (H; -Hl) (7.6)
(2) The heat transferred in the aftercooler is
(H; -H2) (7.7)
(3) The heat transferred in the high- (ambient) temperature
recuperator is
(7.8)
where y is the fraction of gas liquefied per unit mass of gas compressed.
(4) The heat transferred in the intermediate temperature recuperator
is
x(H3- H4) = (1-y )(H10- Hn) (7.9)
(5) The work done by the gas in the expansion engine is
WE=(1-x)(H3-H9) Ci.lO)
(6) The heat transferred in the low temperature recuperator is
x(H4 -Hs) = (x -y)(H9'- Hs) (7.11)
(7) The enthalpy after expansion in the JT valve, H 6 is equal to the
enthalpy before expansion H 5 •
(8) The fraction of gas liqqefied [condition (7)] is

y =x(H6-H7) (7.12)
Hs-H?
(9) The fraction of gas not liquefied but existing at state (8) is
(x -y) (7.13)
The work done by the expansion engine may be used to partially offset
the work of compression. In that case the net cycle work is
(We- WE)= (H;- Hl)-(1-x)(H3-H9) (7.14)
306 Chapter7

The work generated by the expansion engine is relatively small (about 10%
of the compression work). Accordingly the expansion engine and com-
pressor are generally not coupled. The expansion work can be absorbed
in useful ways to drive pumps or generators or simply dissipated in hydraulic
or electrical dynamometers.

Optimum Recirculation Fraction

It is of interest to determine how the pressurized stream should be
split at condition (3) between the fraction (x) going to the Joule-Thomson
expansion valve and the remainder (1- x) expanding to provide refriger-
ation to cool the x stream from T 3 to T 4 • Resolution of the optimum
division is a complex problem best approached by numerical optimization
procedures. Daunt (1956) reproduced data given in the literature for air
liquefiers. The calculations, based on typical heat exchanger effectiveness
parameters, thermal losses, and irreversible effects in expansion, approxi-
mate practical conditions.
Figure 7.6 (after Daunt, 1956) shows the compressor power input per
unit mass of liquid air produced (kW hr/kg liquid) as a function of the
fraction of gas compressed passing to the Joule-Thomson expansion valve.
Several curves are included for different pressures at the end of compres-
sion. The important points to note are that (a) the different curves have
well-defined minima representing the optimum fraction (x) for most

30

8
~0 PRESSURE AT END OF
:::; COMPRESSION (MPa)

~
2.0
"'~
"'
J::
'
~
1-

,.[:5
"'"'50
1.0

"'"'
"'
0
a.
Fig. 7.6. Power requirement per unit mass
FLUID AIR
of gas liquefied as a function of the frac-
0 02 04 0.6 08 IO tion of gas expanded in the Joule-
FRACTION OF GAS EXPANDING IN Thomson valve (for air) (after Daunt,
JOULE-THOMSON VALVE 1956).
Claude and Joule-Brayton Systems 307

efficient operation, and (b) as the pressure increases the maximum efficiency
and the fraction to be expanded in the Joule-Thomson expansion increase.
The conditions of maximum efficiency for operation of an air liquefier
were provided by Davies (1949) in the data reproduced in Table 7.1. This
shows, for several different compressor pressures, the fraction of gas passing
through the expansion engine, i.e., (1- x ), the optimum temperature for
gas to enter the expansion engine, the theoretical power required in kW hr
per liter of liquid produced and, finally, the actual power required (kW hr
per liter of liquid produced). These data confirm, as in Fig. 7 .6, that the
optimum fraction of gas expanded in the engine decreases and the system
efficiency improves as the pressure increases. The temperature for
admission of the gas to the expansion engine increases as the pressure
increases. At a maximum cycle pressure of 200 MPa the temperature has
increased to 293 K, approximately the ambient temperature.

Heylandt Cycle
Heylandt was the first to appreciate that the optimum temperature
for admission to the expansion engine increased as the cycle pressure and
system efficiency increased. Heylandt introduced a modified Claude cycle
by eliminating the upper-stage recuperative heat exchanger of Fig. 7.5 and
elevated pressure level (for an air liquefier) from 4 MPa (used by Claude)
to 20 MPa. He further increased the fraction of mass flow passing to the
Joule-Thomson valve from 0.2 (at 4 MPa) to 0.4 (at 20 MPa).
As evident from Table 7.1 there is little difference in the theoretical
power requirement for the Heylandt-Claude cycle but there are substantial
practical advantages. The actual power required per unit mass liquefied
with the Heylandt system is only 75% of the power with the original Claude
system. Air is admitted to the expansion engine at ambient temperature
immediately following the aftercooling process on leaving the compressor.
About 60% of the air passes through the expansion engine and exits from
the engine at a temperature and pressure of about 150 K and 0.1 MPa
(condition 9' in Fig. 7.5). The relatively high temperature of operation of
the expansion engine has important practical advantages. The high-
temperature recuperator is eliminated and appreciable savings gained
thereby. The lubrication problems of the piston in the cylinder and the
thermal insulation or isolation problems are greatly minimized. The
efficiency of the Heylandt cycle is the highest of all simple Claude systems
and was much favored for air liquefaction systems particularly in the United
States.
Gardner and Smith (1960) have given a comprehensive summary of
the thermodynamics of the Heylandt cycle from the aspect of power
 ~

Table 7 .1. Conditions for Operation of Claude Cycle Air Liquefier at Maximum Efficiencya

Turbo
expansion Original
machine Claude Heylandt
(Kapitza) system cycle

Maximum pressure (MPa) 0.65 2 3 4 4 10 20

Fraction (1 - x) expanding in the
expansion engine 0.88 0.84 0.80 0.75 0.68 0.60
Temperature at inlet to expansion
engine (K) 115 155 173 191 215 241 293
Power required (theoretical)
(kW hr per liter liquid) - - 0.33 - - 0.32
Power required (practically)
(kW hr per liter liquid) 1.5 1.08 1.00 0.90 0.83 0.78 0.70

"After Davies (1949).

n
=-
?.
~
....
Claude and Joule-Brayton Systems 309

consumption and thermodynamic reversibility associated with gas liquefac-

tion and separation.

Low-Pressure Air Liquefiers

The Heylandt cycle represents one limiting case of the Claude system,
when the cycle pressure is so high that the optimum temperature for
admission to the expansion engine has increased to ambient temperature.
The fraction of fluid to be recirculated through the engine is a minimum
60%.
The other limiting case occurs when the maximum cycle pressure is
so low that a very high fraction of the gas must be recirculated through
the expansion engine (>90% ). This system is used for the large-scale
production of liquid air or industrial gases. The combination of low pressure
and high flow rate favors the use of turbines and rotary compressors and
permits appreciable savings in plant costs with increased reliability. Another
advantage is that the use of regenerative rather than recuperative heat
exchangers facilitates the removal of impurities (water vapor and carbon
dioxide) in the air to be liquefied.
The low-pressure turbo air liquefier was introduced in the early 1930s
using the reversing regenerator concept described somewhat earlier by
Frankl. Hence the term Linde-Frankl is frequently used to describe the
whole system for gaseous separation rather than simply liquefaction. The
Russian cryogenics pioneer P. Kapitza is also frequently associated with
the low-pressure turbo air liquefier. In 1939 he provided the first detailed
description of the system including the results achieved in a small experi-
mental prototype unit. The applicability of the low-pressure system
to large-scale processing with turbo equipment puts the Linde-Frankl
system beyond our field of interest here except to note in passing. Daunt
(1956) has given a good account of the system and references the source
documents.

Collins Low-Pressure Air Liquefier

A small scale air liquefier involving relatively low-pressure and
reciprocating machinery was described by Collins eta!. (1958). The system
is of interest because it involves the use of reversing recuperative heat
exchangers to achieve the same result as the Frankl regenerative heat
exchangers with periodic flow switching, that is, the eliminating of special
equipment to remove water vapor and carbon dioxide from the air to be
liquefied.
310 Chapter7

Fig. 7.7. Collins low-pressure air liquefier.

A schematic of the flow system for the Collins air liquefier is shown
in Fig. 7.7. The system operates as follows:
(i) Inlet air (at atmospheric conditions) enters the compressor at condi-
tion (1).
(ii) The air is compressed to 13 MPa and passes to a watercooled
aftercooler.
(iii) The compressed fluid then passes through the reversing valves,
the high- and intermediate-temperature recuperators and the low tem-
perature reversing valves. At condition (6), downstream of the low-
temperature reversing valve, the pressure and temperature is 13 MPa and
125 K. By the time this temperature is achieved all the significant water
vapor, carbon dioxide, and other miscellaneous impurities in the inlet air
have precipitated on the walls of the recuperative heat exchangers. If
operated continuously the flow passages in the recuperators would soon
be blocked by accumulations of these precipitated solids. However, the
reversing valves operate at approximately 3-min intervals to cause the
return flow to pass through the passages containing the precipitated con-
taminants. As the return flow warms in passing through the recuperators
Claude and Joule-Brayton Systems 311

the contaminants reevaporate from the walls of the flow passages, are
entrained in the return flow and exhaust from the system at approximately
atmospheric conditions. In this system there is no recirculation of the return
effluent as in the other cases considered earlier.
(iv) At condition (6), pressure 13 MPa, temperature 125 K, the flow
separates into two streams. The minor stream enters the condenser/re-
cuperator where it condenses under pressure to the liquid phase, condition
(7), and hence passes to the liquid reservoir and subsequently to a
rectification column for separation into the liquid-nitrogen useful product
and gaseous oxygen for release to the atmosphere.
(v) The major stream at condition (6) passes to the expansion engine
where it expands to a final pressure and temperature of approximately
0.15 MPa and 87 K.
(vi) The low-pressure, low-temperature effluent enters the con-
denser/recuperator to accomplish the condensation of the liquid air in (d)
above and hence passes to exhaust through the upper stage recuperators.
In passing the effluent reevaporates and absorbs the contaminants precipi-
tated in the course of the initial cooling.
(vii) Prior to fluid actually entering the expansion engine the oppor-
tunity is taken to provide auxiliary cooling to the intermediate stage
recuperator. In passage to the expansion engine the fluid at condition (6),
pressure 1.3 MPa, temperature 125 K, is further subdivided. A small
amount of the gas is fed through a third set of flow passages in the
intermediate stage recuperator and reheated to about 200 K, condition
(12). The major stream is partially throttled to about 1.25 MPa, condition
(13), to provide sufficient pressure potential to induce flow through the
intermediate recuperator.
(viii) The two streams at condition (12) and (13) are mixed to provide
inlet pressure and temperature, condition (14), of 1.25 MPa and 146 K to
the expansion engine.
Auxiliary cooling in the intermediate recuperator is necessary to cool
the pressurized incoming stream to 125 K (at condition (5)), where all the
carbon dioxide will have precipitated out.
Collins (1956) conceived the use of reversing recuperative exchangers
as an alternative to Frankl regenerative exchangers for relatively low-
pressure systems. Application in systems with high-pressure streams would
pose formidable sealing problems in the low-temperature reversing valves.
On flow reversal the high-pressure fluid contained in the void volume of
the heat exchangers simply passes to exhaust and serves no useful purpose.
This so-called "carry-over loss" would incur a serious decline in the system
efficiency at high pressures.
312 Chapter7

Precooled Claude System

Figure 7.8 is a schematic diagram of a Claude cycle system incorporat-
ing a precooler heat exchanger. The precooler is shown downstream of the
upper stage recuperator but could also be located elsewhere depending on
the temperature regime. If necessary two stages of precooling at different
temperatures might be employed, one downstream of the aftercooler and
the other as shown.
Precooling improves the performance of Claude cycle systems to the
same degree as we saw in Chapter 6 with Linde-Hampson systems. Gases
with low inversion temperature-neon, hydrogen, and helium-need to be
precooled in Linde-Hampson systems for liquefaction to be achieved. The
same need for precooling does not exist with Claude systems since the
cooling effect arises from the more effective isentropic expansion as well
as Joule-Thomson isenthalpic expansion. All gases, even helium can be

MAKE UP
CHARGE

LIQUID
PRE COOLER

Fig. 7.8. Schematic diagram of Claude cycle with liquid

LIQUID PRODUCT precooler.
Claude and Joule-Brayton Systems 313

liquefied in Claude cycle systems with no precooling. Nevertheless, precool-

ing confers advantageous benefits to the cooling performance and should
always be considered. For air the precooler can be a cold liquid refrigerant,
Freon, propane, ammonia, etc., generated by a vapor compression refriger-
ation system. For gases with a low inversion temperature-neon, hydrogen,
or helium-the precooler will probably contain liquid nitrogen. Barron
(1966) asserts that with liquid-nitrogen precooling a Claude system for
hydrogen production has a figure of merit (see Chapter 6) SO% to 75%
higher than that of the Linde-Hampson system.
More sophisticated systems for hydrogen liquefaction have been
developed. Figure 7.9 is a schematic diagram for a Claude cycle hydrogen
liquefaction plant incorporating a liquid-nitrogen precooler and a further
stage of low-temperature precooling generated by an auxiliary refrigeration
system using helium as the working fluid. The helium system is itself
precooled in the liquid nitrogen bath and remains in the gaseous state
throughout. The helium system does not include a Joule-Thomson
NITROGEN
VAPOR
VENT
HYDROGEN HELIUM
CIRCUIT CIRCUIT

Fig. 7 .9. Schematic diagram of hydrogen liquefier

operating on the Claude cycle with a helium
precooler circuit operating on the Joule-Brayton
cycle.
314 Chapter7

expansion downstream of the expansion engine and may be characterized

as a precooled Joule-Brayton cycle.
A preferred arrangement to that shown in Fig. 7.9 would be to
eliminate the hydrogen expansion engine and Joule-Thomson expansion
valve so as to produce liquid hydrogen by direct condensation in a low-
temperature precooler using the refrigeration produced by expansion in
the helium expansion engine. This allows the use of relatively low pressures,
for the hydrogen need be compressed only to the extent required to maintain
the flow through the heat exchangers (less than 1 MPa). It relaxes both
the compressor work requirement and the difficult sealing problems with
high-pressure hydrogen, a highly inflammable gas.
The helium gas-refrigerated, hydrogen liquefaction system is relatively
insensitive to helium pressures permitting the use of relatively low values.
Barron (1966) claims that for helium gas pressures of 1 MPa a system
figure of merit of 0.11 can be achieved.

Dual-Pressure Claude Cycle System

Another modification to the basic Claude cycle is the dual-pressure
system shown schematically in Fig. 7.10. In this cycle the principal mass
of working fluid circulates in the low to intermediate pressure fluid circuit
embracing the low-pressure compressor and aftercooler, the expansion
engine, and the upper and intermediate stage recuperators. The cycle is
encompassed with the state points 1-2'-2-12-10-11-1 shown on the T-S
plane included in Fig. 7.10.
A small fraction of the fluid at condition (2), the intermediate pressure,
is further compressed to high pressure and cooled in recuperators to
condition (6). It is then expanded in the Joule-Thomson valve to the system
minimum pressure and passed to the liquid/vapor separator so that satur-
ated liquid [condition (8)] and saturated vapor [condition (9)] are produced.
In a brief description of the dual-pressure cycle, Barron (1966) claims the
work requirement per unit mass of gas liquefied is reduced compared with
the basic Claude cycle system, but this does not accord with other data
given by Barron, reproduced in Table 7.2.

Comparison of Claude Cycle Liquefaction Systems for Air

The performance of different Claude cycle systems for air liquefaction
was compared by Barron (1966). His data are reproduced in Table 7.2.
The table contains a brief specification of the system, the liquid yield per
unit mass of gas compressed, the cycle work input required per unit mass
of gas liquefied, and the figure of merit. The table is comparable with a
Claude and Joule-Brayton Systems 315

CONSTANT
PRESSURE . . __

ENTROPY S

Fig. 7.10. Schematic diagram of the dual-pressure Claude cycle cooling system.

similar table reproduced in Chapter 6 for variants of the Linde-Hampson

cycle. In Table 7.2 the first three entries refer to basic Claude cycle systems.
The first entry is idealized by the assumption of compressor and expander
isentropic efficiencies of 100% and perfect recuperation. The second entry
refers to a more realistic theoretical cycle with assumed values for machine
efficiencies and with imperfect recuperation. The third entry is a record of
an actual Claude system in operation. It is of interest to note the observed
figure of merit (0.201) is about half the predicted "realistic" cycle of the
second entry (0.356) and one quarter the predicted idealized cycle (0.808).
The next three entries in the same format as above refer to Heylandt
systems where the cycle pressure is increased from 4 MPa to 20 MPa. All
three entries have figures of merit somewhat better than the corresponding
entry for the basic Claude cycle.
The final two entries in the table refer to dual-pressure Claude cycle
systems. Barron did not include observed values for the dual-pressure
Claude cycle system. The figure of merit for the dual-pressure Claude cycle
was substantially less than that of the corresponding entry for the basic
Claude cycle.
316 Chapter7

Table 7.2. Comparison of Liquefaction Systems for Air as the Working Fluida

Work per
Liquid unit mass
yield, liquefied Figure of
y kJ/kg merit

1. Claude system, p 2 = 4 MPa, (1 - x) = 0. 7,

Tic= Tie= 100%, e = 1.0 0.260 164.2 0.808
2. Claude system, p 2 = 4 MPa, (1- x) = 0. 7,
Tic = 70%, 'Yie = 73, e = 0.95 0.189 372.7 0.356
3. Claude system, observed 662.08 0.201
4. HeyIandt system, p 2 = 20 MPa,
(1-x)=0.6, Tic =Tie= 100, e = 1.0 0.373 157.7 0.821
5. HeyIandt system, p 2 = 20 MPa,
(1-x)=0.6, Tic =70%, Tie =72%, e =0.95 0.301 345.2 0.384
6. Heylandt system, observed 614.8 0.216
7. Dual-pressure Claude system,
p 2 = 0.65 MPa, p 3 = 4 MPa, (1-x) = 0.7,
Tic= Tie= 100%, E = 1.0 0.119 219.7 0.604
8. Dual-pressure Claude system,
Pz = 0.65 MPa, p 3 =4 MPa, (1-x) = 0.7,
Tic= 70%, Tie= 72%, e = 0.95 0.073 503.0 0.264

"T1 = 294 K, p 1 = 0.1 MPa, 11< is the compressor overall efficiency, 11. is the expander overall efficiency, e
is the heat exchanger effectiveness (after Barron, 1966).

MULTIPLE-EXPANSION ENGINES

Claude Stepped Piston Two-Stage Expander

So far we have considered only systems with a single stage of expansion.
Significant advantage can be gained by multistage expansion, usually
arranged to occur at progressively decreasing temperatures and pressures.
The earliest use of multiple expansion was described by Claude (1913)
with reference to an air liquefier. The system is shown schematically in
Fig. 7 .11. A notable feature was the use of a single stepped piston with the
high-pressure and low-pressure cylinders arranged coaxially and operating
as follows: Air compressed to about 4 MPa passed through the watercooled
aftercooler and a column containing potassium hydroxide to dry the air
and remove the carbon dioxide. The clean compressed air passed through
the upper stage recuperator and was cooled to 140 K. Thereafter the flow
divided into two streams. The principal stream passed through the high
pressure expander emerging at some intermediate pressure and tem-
perature. It then passed through the intermediate-stage recuperator and
was reheated before final expansion in the low-pressure expander to the
Claude and Joule-Brayton Systems 317

MAKE-UP

AFTER COOLER

UPPER-STAGE
RECUPERATOR

INTERMEDIATE
RECUPERATOR

RECUPERATOR/
CONDENSER

RESERVOIR

Fig. 7 .11. Schematic diagram of the Claude

multiple-expansion, stepped-piston, air liquefier. LIQUID PRODUCT

lowest cycle temperature. The low-pressure, low-temperature exhaust then

entered the third-stage recuperative condenser and thence to the upper-
stage recuperator back to the compressor.
The minor stream continued at progressively diminishing temperature
through the intermediate- and lower-stage recuperators. The flow con-
densed in the lower-stage recuperator and the liquid product collected
under pressure in the reservoir for subsequent separation into oxygen and
nitrogen.
The use of stepped pistons for multiple-stage expansion in Claude
systems was discussed by Collins et al. (1968) with regard to his helium
liquefier. About the same time Kapitza and Danilov (1968) advocated the
use of stepped pistons to facilitate sealing of the low-temperature stage in
small expansion engines. Later Danilov and Kovatchev (1972) discussed
the construction and performance of a small helium expander with a
two-stage reciprocating piston having a relatively high adiabatic efficiency.
Stepped pistons, sometimes called differential pistons, are of course used
elsewhere. They are found in reciprocating compressors, and in multiple-
stage Gifford-McMahon cooling engines. Zimmerman used a multiple-step
piston in the split Stirling system shown in Fig. 1.2.
318 Chapter7

Collins Mnltiple-Expansion Engine

One of the most significant developments of cryogenic cooling systems
occurred in the late 1940s when Collins (1947) made a multiple-expansion
Claude cycle system capable of liquefying helium. Claude cycle helium
liquefiers had been constructed earlier. The first to use a reciprocating
engine was described by Kapitza (1934). This was a single-stage, precooled
Claude cycle system. The Kapitza engine was further notable as the first
to use a hydrodynamic gas lubricated piston to minimize frictional heat
generation at low temperature.
Collins incorporated some of the ideas exemplified in the Kapitza
engine and introduced other innovative features including the use of
multiple-expansion engines. The system developed was capable of liquefy-
ing any gas, air, hydrogen, or helium without the benefit of precooling
although performance was improved by precooling.
A schematic diagram of the Collins helium liquefier is shown in
Fig. 7.12. For the sake of simplicity, only two engines are shown, but,
depending on the helium inlet pressure, systems with up to five expansion
engines have been used. The expansion engines can be physically separated
but coupled to a common crankshaft or can be the stepped (differential)
piston type discussed above. A Joule-Thomson valve is always incorporated
as the final stage of expansion in the liquefier flow circuit.
Typical performance details for a two-stage Collins liquefier were given
by Weinstock (1968). Approximately 25% of the helium flow is expanded
in the first expansion engine to provide cooling in the 30-60 K range.
About half the helium flow is expanded in the second expansion engine to
provide cooling in the 8-15 K range. The remainder is expanded in the
Joule-Thomson valve. Liquid nitrogen precooling of the helium before
expansion in the engines increases the fraction of helium liquefied by two
or even three times the production with no precooling. Some important
parametric effects of multiexpander systems were explored by Class et al.
(1960) with particular reference to large-scale hydrogen liquefaction.

Arrangements in Mnltiple-Expansion Systems

Systems with multiple expansion can be in parallel or series arrange-
ment of the expansion engines. The two alternatives are shown in
Fig. 7.13. In both cases the expansions take place at significantly different
temperatures separated by a recuperative heat exchanger.*
* This should not be confused with thealternative situation where an expansion proceeds at
more or less the same temperature, either (a) in parallel where gas expands over the same
pressure range; The cylinders are supplied from a common source and may exhaust to a
common duct; or (b) in series where the gas expands in a number of stages at successively
lower pressure as for example in a multistage turbine or reciprocating engine.
Claude and Joule-Brayton Systems 319

Fig. 7.12. Schematic diagram of the Collins-Claude

cycle multiple-expansion gas liquefier.

There is perhaps an elegant thermodynamic argument to demonstrate

the superiority of the parallel arrangement, but the argument is not obvious.
However, it is difficult to avoid a strong intuitive feeling that the parallel
arrangement is best. The question, basically, is which is most advantageous:
(a) to expand two small masses over the whole pressure range, or (b) to
expand a single mass (the sum of the two masses) in two partial expansions.
The dilemma is illustrated in Fig. 7.14. Idealized work diagrams are
shown for the upper-stage and lower-stage expansions of the parallel
arrangement [diagrams (a) and (b) on the left side of the figure] and the
series arrangement [diagrams (c) and (d) on the right side of the figure].
On all four diagrams the isotherms are drawn, (a-b) for temperature T 1
and (c-d) for temperature T 2 • It is assumed these temperatures are the gas
temperatures at the start of the expansion process, T 1 for the upper stage,
T 2 for the lower stage.
The isotherms are located on the P-V planes so that on diagrams (a)
and (b) the isotherm for T 1 passes through the point where the volume is
320 Chapter7

al PARALLEL b l SERIES
ARRANGEMENT ARRANGEMENT

Fig. 7 .13 . Parallel and series arrange-

ment of multiple-expansion engines in
Claude cryocoolers.

unity and the pressure 3, in arbitrary units. The isotherm is drawn for a
perfect gas according to the equation PV = canst. The isotherm for T2
passes through the point where P = 3 and V = 0.5 units. This is equivalent
to specifying the temperature as T 2 is half the temperature at T 1 • Assume
an equal mass of gas expands in both the upper and lower expansion stages.
Then the volume of a given mass of gas at temperature T 2 will be half the
volume of the same mass of gas at temperature T 1 •
Diagrams (a) and (b) have been drawn for the above situation. In both
cases the expansion process was assumed to be an arbitrary polytropic
process so that the temperature is somewhat less at the end of expansion
than at the beginning. To a first approximation the work produced in the
upper stage engine is twice the work produced in the lower stage engine
(comparing the areas of the two diagrams).
Now consider the series arrangement represented in the diagrams (c)
and (d) of Fig. 7.14. In this case the sum of the two separate masses
experiences the expansion but in two stages. The isotherms for T 1 and T 2
have been drawn for the situation where double the mass of gas was present
for the diagrams (c) and (d) than was assumed for diagrams (a) and (b).
Claude and Joule-Brayton Systems 321

UPPER STAGE EXPANSION

"'
SERIES ARRANGEMENT
0:
iil
<J)

"'
0:
Q_

2
(b) (d)
VOLUME

Fig. 7 .14. Work diagrams for the upper and lower stage expansion engines in parallel (diagrams
a and b) and series arrangements (diagrams c and d).

This simply means that the isotherms pass through points where, when
P = 3, the volume is 2 for T 1 and unity for T2 •
The arbitrary decision was made to select the expansion ratio to be
equal in both stages, i.e., Pr(upper) = PrOower) = .J3 = 1.73. The work diagrams
then correspond to those shown in diagrams (c) and (d) of Fig. 7.14. As
before an arbitrary polytropic expansion was assumed where the tem-
perature at the end of the expansion was somewhat less than the tem-
perature at the start.
The sums of the areas of diagrams (a) and (b) are greater by approxi-
mately 10% than the areas of the diagrams for (c) and (d). This indicates
that for this arbitrary idealized example the parallel arrangement will
produce more refrigeration (evidenced by the greater work output) than
the series arrangement. Confirmation of this view was provided by
Carbonnell (1974) in a discussion of the relative merits of parallel and
series arrangements. He confirmed the advantage of the parallel arrange-
ment in terms of a 10% increased refrigeration capacity for given constant
adiabatic efficiency and pressure expansion ratio.
Carbonn ell was concerned with large-scale refrigeration using turbines
and pointed out that the series arrangement was characterized by high flow
rates over low expansion ratios. These factors contribute to increased
turbine adiabatic efficiency great enough to favor series coupling for tur-
bines. Reciprocating engines, on the other hand, can handle high expansion
322 Chapter7

ratios at relatively low mass flow rates. This favors the parallel arrangement
where reciprocating expanders are used. Confirmation was provided by
the eminent Russian cryogenic engineer I. B. Danilov in the discussion to
a paper concerned with multiple-expansion helium liquefiers (Kapitza and
Danilov, 1968).

RECIPROCATING EXPANSION ENGINES

Claude Expansion Engines

The expansion engines used by Claude in his original air liquefier
were, in the words of Collins (1958), "simply a judicious adaptation of the
steam engine of the day to the task in hand." Figure 7.15 reproduced from
Collins (1958) is a cross-section view of a Claude-type engine. The
resemblance to a steam engine is striking. Claude's first intent was to induce
liquefaction of air directly to the engine cylinder to overcome the problem
of low-temperature lubrication of the piston by using the liquid air as a
lubricant. On initial startup a light petroleum distillate was used as the
lubricant.
Claude met with early success in his experimental work. In his book
Claude (1913) relates in entertaining fashion his early struggles. It was ...
a cold snowy night in November (1898) within the somewhat forbidding
tramway depot of the Bastille-Charenton Tramways, a totality of conditions
of doubtful comfort ...
that a temperature thought to be less than -150°C was attained. It was not
until years later that Claude realized there was an instrumentation error and
that the true temperature was about -100°C, scarcely better
than had been achieved by Solvay in his expansion engine a dozen years
earlier. In ignorance of this but with much enthusiasm and incredible
endurance, Claude persisted with his experiments, moving in 1900 with
" ... new apparatus installed at the compressed air works of the omnibus
company at La Villette."
Further disappointments awaited him until one Sunday in May 1901,
Claude conceived the notion of liquefying the air under pressure outside
the cylinder. This was at the eleventh hour, for the shareholders, their
patience exhausted, had arranged a meeting the following day to receive
his proposals to terminate the work. In desperate straits Claude feverishly
modified his apparatus that day and was rewarded:
... at last to see a thin thread of liquid escaping from the cock and filling
little by little one of the three litre liquid air holders which had so long
awaited its appearance.
Claude and Joule-Brayton Systems 323

Fig. 7.15. Sectional view of Claude-type expansion engine.

Thereafter was laid the foundation of the great French industrial syndicate
L' Air Liquide, Ltee.
Claude was much concerned about piston lubrication with metal rings
but eventually (1912) adopted the leather cup as a piston seal. It is surprising
this took so long, for leather had long been used on air expansion engines
for ordinary refrigeration and Claude must have been aware of this.
In 1920, Claude constructed hydrogen expansion engines to obtain
temperatures below the boiling point of liquid nitrogen. These engines
were used to separate hydrogen from coke oven gas. Collins (1958)
described hydrogen expansion engines "currently in use" constructed in
324 Chapter7

the United States by E. I. duPont de Nemours and Co. to Claude design as

... single cylinder, double-acting machines of 10 inch bore and 12 inch
stroke. The pistons are fitted with "bakelite" rings with bronze expanders.
Segmental carbon-"bakelite" piston rod packing is employed. Wear on cylin-
der and rings is said to be inappreciable in a year of essentially continuous
operation. The inlet valve is balanced and is mechanically actuated; the
discharge valve, also mechanically actuated, is of the poppet or mushroom
type.
In normal operation for recovery of hydrogen from coke oven gas, the
engine inlet pressure is about 25 atm. and the discharge pressure about
1.6 atm. The inlet temperature is about 113 K and the discharge temperature
about 68 K. Much lower temperatures could be obtained easily but are not
wanted because of the resultant solidification of methane and other impurities.
The energy output of the hydrogen engines is used for compressing air.
Land (1960) has given a brief but useful review of contemporary expansion
engine practice, including piston and turbine engines from the aspect of
lubrication, efficiency losses, power absorption, and control.
Many modern reciprocating expansion engines favor the hydrodynamic
gas lubricated piston concept introduced by Kapitza (1934) and other
innovations introduced by Collins (1958) including tensile connecting rods.
Some designs still prefer the more traditional approach. Witter (1966) of
the Union Carbide Corp., Linde Division, in the description of a small
closed cycle helium liquefier operating on the Claude cycle with two stages
of expansion, gave an interesting critique of different design features. He
was in favor of the compressive connecting rod and carbon-filled fluorocar-
bon piston rings. The principal advantage seen by Witter was elimination
of the seals around the connecting rod and valve rods. An interesting
feature mentioned by Witter was the use of a metal bellows seal around
the valve rods operating at ambient temperatures.
The use of dry-rubbing piston rings as an alternative to hydrodynamic
gas lubrication appears to be popular for larger plants, although a clear
trend to increasing use of turboexpanders is evident. Aberle and Westbrook
(1963) described large helium and nitrogen liquefaction systems with dry-
rubbing reciprocating expansion engines. These were constructed by the
Linde Company, Union Carbide Corp. of Tonawanda, New York for space
simulators at the U.S. Air Force Arnold Engineering Development Center,
Tullahoma, Tennessee and at the NASA Lewis Research Center, Cleve-
land, Ohio. Similarly, Sato (1968) has confirmed a preference for expansion
engines with piston rings (of filled PTFE material) in the discussion of
large-scale helium liquefiers constructed by the Tokyo Oxygen Co. of Japan.
An important and comprehensive discussion of a large reciprocating
expansion engine for tonnage hydrogen liquefaction was given by Morain
et al. (1963). Various aspects relating to the mechanical design of the engine
Claude and Joule-Brayton Systems 325

were given with estimates of the various losses due to heat leaks and ring
friction. Comparative data were presented for the predicted and actual
engine performance on test.

Heylandt Crowned Piston

A significant improvement in expander technology occurred with the
introduction by Heylandt (1912) of the crowned piston concept shown in
Fig. 7 .16. The normal piston carried a long extension of low thermal
conductivity and the cylinder was similarly extended, again with low-
conductivity material. The piston extension or piston crown was a clearance
fit in the cylinder and moved without friction. The main piston was equipped
with sealing rings or other device to prevent leakage from the cylinder past
the piston. These operated at substantially ambient temperatures with a
water cooled heat exchanger provided around the cylinder to carry off the
frictional heat generated. Low-temperature expansion of the gas occurs in
the cylinder above the piston crown thermally isolated by the low-conduc-
tivity materials of the piston crown and cylinder extension from the sealing
device on the piston operating at ambient temperatures.
It will be recalled that Heylandt was also responsible for a modification
of the Claude cycle where gas is admitted to the expander at ambient
temperature downstream of the compressor aftercooler and at high
pressure. Engines of the Heylandt type have been extensively used for
the manufacture of industrial gases involving liquefaction and separation
processes.

Fig. 7 .16. Heylandt crowned piston.

326 Chapter7

Kapitza Hydrodynamic Lubricated Piston

Kapitza (1934), then a professor at the University of Cambridge,
described an apparatus for the liquefaction of helium, essentially a pre-
cooled Claude cycle system with a reciprocating expansion engine. Figure
7.17 is a reproduction of the sectional view of the system given by Kapitza.
The engine included some novel features, principally the provision of
hydrodynamic gas lubrication of the piston. The piston of stainless steel
with no sealing rings or other sealing device was made to be a loose fit in
the cylinder of bronze. In operation the leakage past the piston was high
during the early part of the power stroke when the pressure in the cylinder
was maximum. Therefore a hydraulic work-absorbing system was used
instead of a conventional crank-connecting rod system coupled to a
dynamometer. The hydraulic mechanism permitted the power (expansion)
stroke to be accomplished rapidly to minimize leakage past the piston with
a relatively slow return. The piston diameter was 3.12 em (1.25 in.) and
the stroke varied from 3.5 to 5.0 em (1.4 to 2.0 in.). Clearance of the piston
in the cylinder was 0.05 mm (0.002 in.). It was found necessary to cut
grooves in the piston along its length at intervals of 6.25 mm (0.25 in.) to
prevent seizure of the engine. The grooves were 0.25 mm (0.01 in.) deep
and wide.*
The engine operated at 2Hz (120 strokes per minute) with an inlet
pressure of 3 MPa and a discharge pressure of 0.22 MPa. The temperature
change of helium in the engine was from about 20 K to 10 K. Interesting
details of the subsequent development of this device were given by Malkov
et al. (1968) and by Kapitza et al. (1968).

Collins Expansion Engine

The Collins system has become the virtual standard machine for use
in the Western world. In a recent technical memorandum (CT1-1978) the
manufacturers claimed that over 125 of the smaller model 1410 liquefier
were in use in over 20 countries. A decade earlier Weinstock (1968),
quoting a private communication, indicated that 500 Collins liquefiers were
sold by A. D. Little in the twenty years since the prototype development
in 1946. Whatever the number it is certainly true that up to 1950 the
number of laboratories dealing with liquid helium was limited to a handful
of specialist operations. The development of the Collins machine put the

*The piston would not be naturally centered in the cylinder as one might expect. Instead
the high-pressure gas flowing past the piston from the high-pressure to the low-pressure
region causes the piston to move to one side of the cylinder and provide the flow path with
least restraint.
Claude and Joule-Bray ton Systems 327

I
Z4 ·

"*15I
I
i
I 13
I
I
i :::!8

37
38 /16
34

20cm
10 .._.._.,..-.ll
0,..H...._..._..,-ll....._,
Scale

Fig. 7.17. Sectional view of the Kapitza helium liquefier including an expansion
engine with hydrodyna mic lubrication (after Kapitza, 1934).
328 Chapter7

production of liquid helium onto a routine push-button basis and thereby

transformed the capabilities and potential of low-temperature research.
Collins evolved the design of the helium liquefier through a variety of
systems. In his book Collins (1958) describes in detail an early diaphragm
engine completed in 1938 capable of attaining temperatures of 10 K. The
engine experienced minor leakage problems which had a highly deleterious
effect on the vacuum insulation surrounding the engine. The Second World

Fig. 7 .18. Section diagram of low-pressure Collins expansion

engine (after Collins, 1958).
Claude and Joule-Brayton Systems 329

War intervened and Collins concentrated his attention on the development

of the low pressure air liquefier (shown in Fig. 7. 7) with reversing recupera-
tive heat exchangers used for oxygen production. The system incorporated
expansion engines for relatively low inlet pressure of 0.6 to 1.5 MPa.
After the war Collins did not resume work on the diaphragm engines
but incorporated his wartime experience in a multiple-expansion helium
liquefier using reciprocating expansion engines. Figure 7.18 is a section
view of a Collins expansion engine cylinder and valve chamber. Particular
features that led to successful operation include the use of a very close
fitting piston (of soft material over a steel core) sliding in a cylinder with
hardened surface, a highly efficient continuous recuperative heat exchanger
(see Chapter 8), and the design innovation of putting the piston and valve
rods in tension rather than the customary compression mode. This latter
feature permitted the use of thin flexible rods reducing the conductive heat
leak and facilitating free movement of the piston in the cylinder.
Another important feature was to contain the engines and heat
exchangers in a stainless steel Dewar vessel containing the liquid-helium
product. The external environment of the engines and heat exchangers was
low-pressure helium, therefore minor leaks from the pressurized engine/re-
cuperator system were of no significance compared with a similar leak to
a vacuum enclosure.
Further construction and operating details of the system were given
by Collins (1947, 1952, 1956, 1958, and 1966). The following is a descrip-
tion of the chief characteristics of the helium liquefier by Collins (1958):

Rate of liquefaction 25-321/hr

Power required 45kW
Helium circulated 215 g mol/min (185 cfm)
Operating pressure 12.5 atm
Refrigeration Expansion engine plus liquid
nitrogen
Liquid nitrogen (If no liquid nitrogen is used, rate
of helium liquefaction is
10 1/hr)
Heat exchanger Hampson type
Actual work expended (N 2 plant
included) 3.1kWhr/l
Computed requirement (actual
liquefier but with N 2 plant and
helium compressor assumed
reversible) 0.87 kWhr/1
Computed requirement (entire pro-
cess reversible) 0.24 kWhr/1
330 Chapter7

Description of Liquefier
A flow diagram of the actual cycle is given in Figure [7.19]. It differs
from the idealized cycle, not only because the heat exchangers and expansion
engines are necessarily imperfect, but also because practical considerations
have influenced the choice of apparatus and procedures. For the sake of
compactness of the liquefier and greater production of liquid from available
compressed helium, liquid nitrogen is employed to the extent of its utility.
The transfer of heat from gaseous helium to liquid nitrogen evaporating at
a constant temperature is irreversible, of course, and a net increase of entropy
is incurred. A final difference lies in the substitution of a throttle valve for
the fourth engine of the cycle.
In the flow diagram shown [Figure 7.19], compressed helium (about
12 atm) from the compressor is treated for entrained oil in an oil trap and
for vaporized oil in a refrigerated heat exchanger 3. Thereafter the stream
divides, about 8 percent going to heat exchanger 4, in which it is cooled to
80 K by means of liquid nitrogen and then expanded in engine E 1. the
remainder going to the principal heat exchanger 5. The temperature of the
gas in heat exchanger 5 ranges from room temperature at the upper end to
15 K at the bottom. At the zone of exchanger 5 where the temperature is
40-45 K, a second fraction (about 15 percent of the whole) of the compressed
helium is led off for expansion in the second engine, E 2 • At the lower end

w
CD
._::>
a::
w
u._
(f)
z
<l
._
a::

0
..J
w
I
(f)

z
._0
<l
0
<l
a::

4-STAGE
COMPRESSOR

Fig. 7 .19. Flow diagram of relatively large-scale Collins helium liquefier (after Collins, 1952).
Claude and Joule-Brayton Systems 331

of exchanger 5 a final division of the stream occurs. About 52 percent is

used in expansion engine E 2 and 25 percent flows through the small
exchangers 6 and 7 to the expansion valve D.
All the helium that enters exchanger 4 passes through engine El> the
pressure falling to substantially one atmosphere and the temperature falling
from 80 K to about 45 K. The expanded helium joins the low pressure stream
in exchanger 5. All three engines are served by a single crankshaft and are
identical in size. The bore is 2 inches, and the stroke is 2 inches. Each engine
is comprised of a single cylinder. Although the three engines receive com-
pressed helium from the same supply line and discharge their spent gas into
the same low-pressure conduit, they operate at different temperature levels
and embrace different sections of the heat-exchange system. Helium enters
E 2 at about 45 K and is discharged at about 25 K. The inlet temperature of
E 3 is about 15 K and the outlet temperature is estimated to be about 9 K.
Liquid helium forming at the throttle valve D drops into the bottom of
the Dewar vessel in which the heat exchangers and engines are suspended.
Space for 30 liters is provided.
Certain features of the earlier helium cryostat have been retained. The
heat exchangers and engines, which hang from a steel plate, are surrounded
by an atmosphere of helium rather than by the insulating vacuum. Minor
leaks from the high-pressure stream can be tolerated. The helium atmosphere
is contained by a large metal Dewar vessel, the vacuum jacket of which is
continuously pumped. The lower half of the inner wall of the vacuum jacket
is enclosed by a nitrogen-cooled radiation shield.
The engine cylinders and pistons are made of nitrided nitralloy and are
so closely fitted that piston rings are unnecessary. The piston rods are
relatively long and slender, and are made to operate in tension to promote
perfection of alignment of the piston within the cylinder. There exists a
thermal gradient in the helium atmosphere surrounding the heat exchangers
and engines, the temperature being approximately 295 Kat the top and 4.2 K
at the bottom. As far as practicable, the cylinders of the engines are located
at the proper elevation for matching temperatures inside with outside in
order to reduce convection to a minimum. The stuffing boxes for the piston
rods and valve pull rods and the running gear of the engines are placed on
top of the lid of the Dewar so that heat generated in these parts can be kept
out of the cold region.
The piston rods and valve pull rods are attached to the ends of horizontal
walking beams, as in ancient steam engines. The beams are 2 feet in length,
are pivoted at one end, and at the point immediately above the crankshaft
are fitted with ball bearings to act as cam followers. With a stroke of only 2
inches the end of the piston rod travelling in an arc of 2-foot radius is not
pulled away from the vertical by an appreciable angle. Speed control is
achieved by centrifugal action of a split flywheel within a brake drum. The
two halves of the flywheel are fitted with brake shoes. Speed is adjusted by
changing the compression of a spring. This adjustment may be made while
the engine is running.
The Collins engine became the prototype for a family of helium
liquefier systems in series production up to the present time. Figure 7.20
shows the model 1430 helium liquefier system currently in production at
332 Chapter7

Fig. 7.20. Collins-Claude cycle helium liquefier (CTI Model No. 1430).

CTI, Waltham, Massachusetts. The system has the capacity to produce 5

to 40 liters per hour of liquid helium or 19 to 100 W of refrigeration at
4.6 K depending on the compressor option and mode of operation chosen.
For maximum output the helium refrigerator module requires 40 liters per
hour of liquid nitrogen and 2.5 kW of electric power. The liquefier module
weighs 770 kg and is contained in an envelope 1.27 x 0.375 x 0.69 m. The
associated helium compressor module requires 100 kW of electric power
and 70 kg per min of cooling water at 24°C and 0.3 MPa. The mass of the
compressor module is 1455 kg and is contained in an envelope 0.415 x
0.475 x 0.715 m. The constituents and general arrangement of a typical
CTI/Collins helium liquefier are shown in Fig. 7.21.
Following the pioneer work by Collins at the Massachusetts Institute
of Technology and the initial development of the commercial version at
A. D. Little successive improvements and variants appeared. Pastuhov
Claude and Joule-Brayton Systems 333

HEliUM STORAGE TANK

TRANSFER TUBE -.__
!OPTIONAL} .... r- _....
~

REGULATOR !BY CUSTOMER) I

PURE HELIUM (BY CTI-CRYOGENICS) CAPACITY 650 SCF
GAS STORAGE AT 16 ATM-MINIMUM
(BY CUSTOMER)

( ·-'1,/
I. .I
b

-----'f---.>--./ HELIUM
IMPURE HELIUM GAS GAS BAG
FROM COLLECTING STATIONS (OPTIONAL)

Fig. 7.21. Typical liquid helium installation utilizing Collins-Claude cryocooler.

(1960) of A. D. Little Inc. discussed application of a Joule-Brayton cycle

variant of the Collins-Claude engine for refrigeration over the temperature
range 20 to 100 K. Later Collins and Streeter (1968) described a variant
which by the addition of a very low-pressure Joule-Thomson stage was
able to generate refrigeration at 1.8 K. Robinson (1965) described a large-
capacity (120 1/h) Collins helium liquefier constructed by A. D. Little.
Further details of a large-scale helium refrigerator installed at the Brook-
haven National Laboratory to cool a superconducting magnet were given
by Robinson and Hogan (1968).
Subsequently, Johnson, Collins, and Smith (1971) described a
modification of the large-scale liquefier where the Joule-Thomson
expansion valve was replaced by an expansion engine with condensation
of the helium during the expansion stroke. The engine was driven by a
hydraulic plunger mechanism independent of the upper stage expansion
engine. The valves were also operated by a hydraulic-pneumatic system.
Use of the expansion engine instead of a Joule-Thomson valve increased
the liquid yield by 3 3%.
Development of Collins engines stimulated the construction of similar
machines of different sizes and capacities. Hood et at. (1964) discussed the
performance of Joule-Brayton helium refrigerators for operation in the
10-30 K range that have been clearly inspired by Collins work. These units
334 Chapter7

were constructed by Cryovac Inc. of Columbus, Ohio for the U.S. Air
Force Arnold Engineering Development Center.
Zeitz (1963) described a closed cycle helium refrigerator for 2.5 K
constructed by Air Products and Chemicals Inc., Allentown, Pennsylvania.
This unit was a compound device consisting of a high-pressure (0.2-2 MPa)
double expander Joule-Brayton refrigerator generating refrigeration at
10 K to precool a low-pressure (0.009-0.5 MPa) Linde-Hampson system
generating refrigeration at 2.5 K. The expansion engines of the high-
pressure system are acknowledged to be "similar to the machines developed
by Collins and Kapitza."
Furihara et al. (1968) described what they whimsically called "a new
type of closed cycle helium refrigerator for laser applications." From the
description given it appeared to be a carbon copy of the Collins-Claude
cycle double expander with final-stage Joule-Thomson expansion. A
Collins-type expansion engine was also incorporated for helium expansion
in a compound helium-hydrogen-neon refrigerating system described by
Zeitz et al. (1965). It was constructed by Air Products and Chemicals Inc.
for the NASA Lewis Research Center to cool a cryogenic magnet used for
plasma heating and acceleration experiments.
Another large-scale (1.4 kW at 4.5 K) helium refrigerator incorporat-
ing two-stage Collins expansion engines was described by Winters and
Snow (1966). The unit was constructed by Air Products and Chemicals
Inc., Allentown, Pennsylvania for the NASA Langley Research Center.
The same company also constructed the lower-capacity Claude cycle system
using a single Collins expansion engine with no precooler described by
Meier and Currie (1968). These units were built to provide 1 W of useful
refrigeration at 4 K with additional refrigeration at 30 K. The units were
designed for long-term reliable automatic operation with redundant com-
pressor and expansion engines provided with automatic switches following
component failure. At the time of reporting (1968) 10 systems in field use
had accumulated 32,000 operating hours with only one machinery failure.

Doii-Eder Valveless Expansion Engines

Two researchers at the Bavarian Academy of Sciences in Munich,
West Germany, R. Doll and F. X. Eder (1964) introduced an important
new concept in expansion engine technology: the valveless engine with
hydrodynamic gas lubricated piston. A simplified cross section of the engine
is shown in Fig. 7.22 reproduced from Kneuer et al. (1968). The engine
cylinder (1) has two annular channels (2) and (3) for the admission of
high-pressure gas and the exhaust of low-pressure gas. The piston (4) has
a central bore (5) connected via radial channels 6 to a ring groove (7) on
Claude and Joule-Brayton Systems 335

Fig. 7 .22. Simplified cross section of Doll-Eder valveless,

gas lubricated piston expansion engine.

the piston surface. Work is transferred via the piston rod to the crank-gear
and hence to a dynamometer.
Gas flow and the resultant work diagram for the engine are illustrated
in Fig. 7.23 (after Kneuer et al., 1968). At condition 1, with the piston in
the bottom dead center position, the ring groove on the piston is aligned
with the inlet annular channel in the cylinder. High-pressure gas enters
the engine filling the clearance space, volume V1. principally the central
bore of the piston. As the piston rises the piston ring groove and the
cylinder inlet annular channel move out of alignment thereby closing off
admission of further fluid at high inlet pressure. The fluid contained in the
cylinder expands to a lower pressure and work is produced.
At condition (2) the lower edge of the piston clears the annular exhaust
channel in the cylinder releasing expanded fluid to exhaust. The pressure
drops rapidly to the exhaust pressure [condition (3)] as the piston ascends
to top dead center. Finally the piston descends closing off the exhaust
annular channel and compressing the gas retained in the cylinder to condi-
tion (4). Near the piston top dead center position the inlet port is again
uncovered and fresh charge is admitted to the cylinder to repeat the cycle.
There is a leakage of high-pressure gas to exhaust in the annulus
formed by the piston in the cylinder. The leakage is indicated, by the arrows
on Fig. 7.23, as a flow from (a) the inlet to the exhaust annular channel,
(b) the underside of the piston to the exhaust channel, and (c) the inlet
channel to the space above the piston. This leakage flow is the hydrodynamic
gas lubricant for the piston. Kneuer et al. (1968) identify the flow as a loss
that is unavoidable with valveless engines and give the following equation
for the magnitude of the flow loss:
v = (pt-p~)a 3 U/24p1L{3 (7.15)
where v is the volumetric rate of flow, P1 is the inlet pressure, P2 is the
336 Chapter?

,.,
1(P 1 )

Fig. 7.23. Operational cycle and work diagram for

{b)
Doll-Eder valveless engine.

exhaust pressure, a is half the clearance of the piston in the cylinder, U

is the circumference of the piston, L is the length of the leakage path, and
{3 is the viscosity of the working fluid. The flow is a function of the cube
of the clearance in the cylinder and compels close tolerance clearance fitting
of the piston in the cylinder. Kneuer et at. (1968) mention a clearance of
less than 2 and 3 I.L with piston diameters of 28 and 38 mm.
Becker et at. (1968) in their discussion of the leakage loss suggest that
with 2 I.L clearance the loss is 5% to 10% of the total gas flow with pressures
in the range 0.3 to 3 MPa. This flow is said to be sufficient to stabilize the
piston and provide complete gas lubrication:

... by means of a large number of grooves on the piston surface. The clearance
between successive grooves has a proper profile and achieves high stabilizing
forces.

Doll-Eder valveless expansion engines have been used extensively by

the German company Linde Aktiengesellschaft of Munich. In 1968 Kneuer
Claude and Joule-Brayton Systems 337

et al. wrote that over 50 valveless engines had been incorporated in helium
liquefiers in the previous four years. Baldus and Sellmaier (1968) described
a complicated system to generate temperatures of 1.8 K in a helium II
refrigerator utilizing a Doll-Eder valveless expansion engine in a Claude
cycle system with a low-pressure Joule-Thomson expansion circuit generat-
ing 30 W of refrigeration at 1.8 K.
Publication of the original paper in 1964 stimulated considerable
research on valveless expanders. Leger et al. (1971) described, in fascinating
detail, the design and development of a Claude cycle helium liquefier with
two valveless expansion engines, in a parallel arrangement. One interesting
feature of the Leger engine was the use of different materials for the
cylinder and the piston. The piston was made of material having a lower
thermal expansion coefficient than the cylinder. On starting, the clearance
of the piston in the cylinder was as high as 11 f.1. and reduced, due to
differential contraction, to 2.5 f.1. at the normal running stage. The use of
different materials is contrary to conventional practice where the same
materials are used for piston and cylinder to maintain a constant clearance
during cooldown. In early engines with the piston and cylinder made of
the same steel, Leger found the piston cooled more slowly than the cylinder
and resulted in premature wear contributing to reduced performance at
low temperatures. In some cases the differential rate of cooling caused the
engine to seize and resulted in deformation of the cylinder.
Matsubara et al. (1968) have described a hybrid expander with the
inlet high pressure fluid entering through valveless ports as in the Doll-Eder
engine but exhausting through a crank-operated valve of conventional
design. This allows independent selection of exhaust valve timing but has
no other advantage over the Doll-Eder engine and introduces the com-
plexities of mechanically operated valves.
Another valveless, gas-lubricated expansion engine of exceptional
interest was described, albeit briefly, by Clarke et al. (1968). The engine
was a free-piston machine taking the form of a double-ended Doll-Eder
engine with a single reciprocating element, the piston, operating in both
cylinders. One end operated as a conventional Doll-Eder expansion engine
at low temperature. The other end operated as an auxiliary compressor
absorbing the work of the expansion engine to compress helium. The engine
had a bore and stroke of 31.75 mm (1~in.) and operated at a frequency
of 45Hz with an inlet pressure and temperature of 2.5 MPa and 24 K. It
was said that

... many hundreds of hours of development running carried out on prototype

engines of this type confirm the original concept of simplicity and reliability.
Depending on the design and operating requirements the efficiency on isen-
tropic heat drop basis can be well over 70 percent ...
338 Chapter7

Piston

Distance piece
30mm 10
31 mm OD
300 mm long

Sliding ring

Stroke
42 mm
Fig. 7 .24. Sketch of expansion engine with valves
operated by piston motion (after Claudet et a/.,
1972).

Claudet and Verdier (1972) described an interesting prototype expansion

engine for Claude cycle systems with valves operated by the piston motion.
Figure 7.24 is a sketch of the engine (after Claudet et al., 1972) and
Fig. 7.25 shows the relative piston and valve positions during operation.
The cycle of events in the engine is as follows:
(1) Exhaust. With the piston at bottom dead center and beginning its
ascent the inlet valve is kept closed by the high pressure of the inlet fluid.
The outlet valve located in the piston is opened by a spring F located
under the valve. Gas is released from the cylinder to exhaust.
(2) Compression. Exhaust gas leaves the cylinder as the piston ascends
towards top dead center. As the piston approaches the end of the upward
stroke the spring F above the exhaust valve is compressed and closes the
exhaust valve. Shortly thereafter the spring F is compressed sufficiently to
cause the inlet valve to lift from its seat (alternatively an extension of the
exhaust valve impinges on the underside of the inlet valve causing it to
lift) and new charge is admitted to the engine.
(3) Induction. The piston passes the top dead center position with the
inlet valve open and the exhaust valve closed. As the piston descends the
Claude and Joule-Brayton Systems 339

c)

Fig. 7.25. Piston and valve positions in Claudet-Verdier engine operation (after Claudet et
al., 1972).

spring force F holding the inlet valve open decreases and the inlet valve
closes.
(4) Expansion. The piston descends with both valves closed. The
pressure decreases as the expansion proceeds.
The performance of the prototype Claudet/Verdier engine was most
encouraging. With inlet-outlet pressures of 0.9 and 0.14 MPa they observed
isentropic efficiencies in the range 50% to 70%. The temperature difference
across the engine ranged from 5 K with an inlet temperature of 20 K to
26 K with an inlet temperature of 76 K.
Russian work on valveless piston expansion engines was described by
Dobrov and Orlov (1974). Their engine utilized a six-phase gas distribution
cycle in an attempt to overcome some of the deficiencies of the conventional
four-phase Doll-Eder eng1ne. The Dobrov-Orlov concept required mutual
but out-of-phase reciprocation of both the piston and cylinder, but it is
not clear how this was accomplished. One's feeling is that the added
complexity obviates the principal attraction of simplicity in the Doll-Eder
valveless engine.
Parulekar and Narayankhedkar (1972) and (1976) discussed the design
and performance of a valveless expansion engine with novel features
including a regenerative heat exchanger contained within the piston. The
piston was activated by alternating fluid pressure in the "warm" and cold,
expansion, part of the cylinder. Fluid pressure in the fluid drive region was
regulated by solenoid valves activated by the piston motion. There seems
every possibility this system could be made to operate at resonant frequency
as a free piston machine of the same family as that described by Clarke et
al. (1968). It could also be classified as an Ericsson engine of the types
described in Chapter 5.
340 Chapter7

~ HELIUM

~
~ BRASS

~
% STAINLESS STEEL

II P.T.F.E .

Fig. 7.26. Sectional view of formed bellows engine (after Long and Simon, 1954).

Bellows Expansion Engines

Replacement of the conventional cylinder and piston of an expansion
engine by a flexible metal bellows was suggested by Simon as early as 1928.
The earliest known account of a practical engine of this type was given by
Long and Simon (1954). A simplified cross section of their engine is given
in Fig. 7 .26. The internal diameter of the bellows was 4 em and the stroke
of the engine was 2.7 em with 15 bellows convolutions. The inlet and outlet
pressures were 0.55 and 0.13 MPa and the frequency of operation was 2
to 3Hz. Figure 7.26 shows the admission and discharge valves were con-
tained in a central valve block within the bellows cavity. The valve rods
operated in tension to open the valves against the force of stainless steel
springs attempting to close the valves. Expansion of the bellows was
restrained by (a) a thin-wall stainless-steel tube, in compression, supporting
the valve block, and (b) a connecting rod in tension attached to the bottom
flange of the bellows. Long and Simon reported they were able to liquefy
helium with the bellows engine.
Claude and Joule-Brayton Systems 341

DISTANCE TUBE

I l ET

- - -- V ALV E TUBE

Fig. 7.27. Sectional view of welded bellows engine (after Smith, 1967).

The advantage of a bellows engine is elimination of the piston leakage

and friction losses of conventional engines. In combination with a dia-
phragm compressor a completely sealed system free of all contaminants
can be obtained. The principal disadvantage was expected to be the limited
life of the bellows. The operating experience showed at low temperatures
the fatigue life improved so much that by limiting the pressure a virtually
unlimited life could be obtained.
Long and Simon used formed bronze bellows. They have a fairly high
volume of retained gas when the bellows are in the compressed state. This
342 Chapter7

~--CRANK CASE

LIOUIO NITROGEN
PRECOOLING
TIMING BELT

HEAT EXCHANGER

0 I STANCE TUBE

ELECTR IC
LIOU IO NITROGEN
MOTOR
RADIATION SHIELD

SURGE CHAMBERS

TIMING BELT

Fig. 7.28. Bellows engine test apparatus (after Smith, 1967).

corresponds to a conventional expander with a high clearance volume

involving substantial recompression of expanded gas.
A later development of a bellows expansion engine was reported by
Smith (1967). This work was carried out at the Massachusetts Institute of
Technology for the U.S. Air Force. A cross section of Smith's bellows
engine is shown in Fig. 7.27 and the engine test apparatus in Fig. 7.28.
Smith utilized welded metal bellows of stainless steel in place of the formed
bronze bellows of the earlier engine. The welded metal bellows were
superior from two aspects. First, the bellows may be more tightly collapsed.
They were formed by edge welding rings of the sheet metal so that only
the weld beads prevent the convolutions from closing tight. Second, the
bellows were made of high-tensile-strength stainless steel free from the
effects of the severe deformations required to produce a formed bellows.
 ("'l

f
lr
Q,
=
"'
H[AO H(AT EXCHANG(R LINEAR ROlARY MOTOR ROTOR
....
STATOR ~
;r
EXPANSION CHAMBER
COMPRESSOR H[ AO I!=
EXPANO[R PISTON
i=
i"'
..
a
"'
-- -- --

GAS eEARING

GAS BEARING GAS BEAR I NG SPRING ASS£ MBLY

Fig. 7.29. Diagrammatic view of rotary-stroking engine (after Schulte et al., 1965).

~
"""
344 Chapter7

Smith used bellows having an outer diameter of 3.17 em (1.25 in.) and
an inner diameter of 1.27 em (0.5 in.). Nineteen convolutions of 0.1 mm
(0.004 in.) thickness provided a stroke of 1.9 em (0.75 in.) and a displace-
ment of 1.1 x 10-4 m 3 (0.44 in?). The inlet and discharge pressures were
0. 51 and 0.11 MPa (7 5 and 16 psi). Two bellows to this design failed at
1.3 x 106 cycles (43 hrs) and 1.15 x 10 6 (38 hrs). The design was then
modified as detailed by Smith to a bellows of 35 convolutions and a
diaphragm thickness of 0.2 mm (0.008 in.) with the stress at the inner weld
always in compression. At the time of reporting this assembly had operated
for 7.7 x 106 cycles (258 hrs) without failure at temperatures of 28 and
18 Kat inlet and exhaust, respectively.

Rotary Stroking Engine

Schulte et al. (1965) described a novel Joule-Brayton engine with
reciprocating compressor and expander tuned to operate at a relatively
high resonant frequency (60Hz). The system is shown in Fig. 7.29 (repro-
duced from Schulte et al., 1965). In addition to the reciprocating motion
the dynamic elements were also driven with rotary motion so as to generate
the hydrodynamic gas lubrication to maintain the piston "floating" in the
cylinders.
Later Breckenridge (1969) described a development of the Schulte
engine now expanded to a Claude cycle system with two stages of compres-
sion, two stages of engine expansion and a Joule-Thomson expansion stage.
Subsequent developments of the rotary stroking engine have been reported
regularly by Breckenridge. Many references under this name may be found
listed in the Bibliography in both the open literature and the list of
government reports.
Free-piston Stirling engines for power generation have been developed
and are in series production (see Walker, 1980, Chapter 11). These have
demonstrated the feasibility of resonant frequency systems with reciprocat-
ing elements and self-induced hydrodynamic gas lubrication. Elimination
of the need to rotate as well as reciprocate the compressor-expander
elements must lead to a simpler, and therefore more reliable, system. Given
this parallel development it is hard to justify a sanguine view for the future
of the Schulte rotary stroking engine.

ROTARY EXPANSION ENGINES

It is tempting and appropriate at this stage to venture forth into a

review of rotary expanders (turbines) applied to Claude cycle systems. We
Claude and Joule-Brayton Systems 345

sqall resist the temptation. The subject is simply too big for this already
over-long chapter to contain. Instead we consign turbines to a separate
folio, companion to this volume, concerned primarily with large-scale
refrigeration, liquefaction, and gas processing. Such division is in the natural
order of things. Rotary machines, turbines, and compressors, are charac-
teristically best suited for high flow rate, low-pressure ratio systems as
compared with reciprocating devices which favor low-flow, high-pressure
ratio systems.
Much work has been done on miniature turbo expanders and compress-
ors for both Claude and Joule-Brayton cycle systems. It is appropriate to
briefly review that work below since our interest is limited to small- and
intermediate-capacity systems. However, a detailed consideration has been
left to the companion volume. Many papers on miniature turbine systems
may be found in the extensive Bibliography included here under the names
of Colyer and Sixsmith. Some of these are briefly discussed below.

MINIATURE CLAUDE CYCLE SYSTEMS

There are few references to miniature Claude cycle systems. Most

work has been done with miniature refrigerators for flight systems operating
on aircraft or spacecraft. In most cases these operate without a Joule-
Thomson valve for the final stage expansion and so are Joule-Brayton
rather than Claude cycle systems. A miniature Claude or Joule-Brayton
system was the rotary stroking engine reviewed above and described by
Schulte et al. (1965) and later by Breckenridge (1969).
Much work has been done for the Air Force by the General Electric
Co. Schenectady, New York on miniature rotary Joule-Brayton and Claude
cycle systems. Gessner and Colyer (1968) have reviewed the early progress
of this work and many subsequent reports have been made by Colyer (see
Bibliography). Other miniature turbomachinery Claude and Joule-Brayton
cycle systems have been described by Maddocks (1968). This work was
done by the AiResearch Manufacturing Co., Los Angeles for the U.S.
Army Satellite Communications Agency. Campbell (1967) reported on the
development and performance of a "miniature Claude cycle" refrigerator.
The system included a reciprocating engine but no Joule-Thomson
expansion valve, so it was a Joule-Brayton rather than a Claude system.
It is not difficult to understand the paucity of literature. Claude systems
are not well suited for miniaturization due to valves and recuperative heat
exchangers. A better choice for miniature applications with recuperative
heat exchangers is the Linde-Hampson system, basically a Claude cycle
system with the expansion engine (a bulky, expensive unit) removed. The
346 Chapter7

principal objective of miniaturization-an ultracompact design-is more

readily achieved, and in small capacities there is no difficulty in using
reciprocating compressors to achieve the high pressures characteristic of
Linde-Hampson systems. Despite these advantages for miniaturized appli-
cations little development of Linde-Hampson systems has occurred. The
reason is that the low flow rates of a miniature system require extremely
fine passages in the recuperative heat exchangers and the Joule-Thomson
valve. They are susceptible to contamination and blockage for the flow is
continuous and one way so accretion of contaminants is continuous and
cumulative. The need for uncontaminated fluid requires filtering and
absorbing devices that make it difficult to achieve the low mass and volume
objectives of miniaturized equipment.
Regenerative heat exchangers are the preferred alternative to
recuperative heat exchangers. They tend to be self-cleaning and less suscep-
tible to blockage by contaminants in the working fluid. They are easy to
fabricate at low cost compared with recuperative heat exchangers and can
be made very small and of low mass. Regenerative systems can be pressur-
ized to a high level so as to keep the system volume and mass within
acceptable bounds. They exist in enough variations, Stirling (integral or
split), Vuilleumier, and Gifford-McMahon, to meet demand limitations
better than any miniature Claude system yet in prospect.

TERBOT MIXED REFRIGERANT CYCLE

An interesting modification to the basic Claude cycle was described
by Terbot (1976). The flow diagram shown in Fig. 7.30 is reproduced from
Terbot (1976). In the low-temperature region the cycle utilizes helium (or
any other gas) and is identical to the conventional Claude cycle with two
stages of expansion engine and Joule-Thomson expansion valve.
The difference between the Terbot and conventional Claude cycle is
that at the inlet to the compressor another refrigerant (Terbot suggests
dichlorodifluoro-methane, refrigerant 12) is mixed with the helium and
compressed with it in the compressor. After passing through the aftercooler
and the upper-stage recuperative heat exchanger the mixture, cooled by
this time to 120 K, consists of liquid R12 and gaseous H 2 • The liquid R12
is separated from the helium and recirculated back through the upper stage
recuperator mixed with helium en route to the compressor inlet. Addition
of the refrigerant R12 to helium increases the molecular weight of the
mixture to be compressed and allows use of "compact, reliable inexpensive
standard centrifugal compressor[s]."
Terbot presented comparative data for conventional Claude cycle
systems of intermediate and large capacity using reciprocating and oil-
Claude and Joule-Brayton Systems 347

MIXTURE
He&R-12
1034 k Pa
300K

ADSORBER

HELIUM
~ ~g ~Par---+-----f----4----,
~+-----,

648 k Pa
300K

Fig. 7.30. Flow diagram for Terbot

mixed refrigerant Claude cycle sys- ....._----------~ ~2~2~g~~:A~~~~~I~G
tem (after Terbot, 1976). 5K ~ 667 W

flooded screw compressors and for the new mixed refrigerant cycle. The
Claude cycles included liquid-nitrogen precooled cycles in some cases. The
mixed refrigerant cycle emerged from this comparison indicating appreci-
able (15%) savings in the capital and operating cost.
No studies are known of the Terbot mixed refrigerant cycle in smaller
systems using reciprocating compressors or with auxiliary refrigerants other
than Rl2.

COMPRESSORS FOR CLAUDE CYCLE SYSTEMS

Compressors are a necessary component of all Claude, Joule-Brayton

and Joule-Thomson systems. In some systems (Joule-Thomson) very high
348 Chapter7

pressures (up to 40 MPa) are used to achieve an effective compact system.

In other cases (Claude and Joule-Brayton}, more moderate pressures can
be utilized. In all, the compressor is vital and frequently a significant fraction
of the overall cost. The main input of energy is power to drive the com-
pressor-normally, the prime operating cost. The compressor is the element
most likely to require maintenance attention and is the key component
affecting reliability.
Improved reliability is often gained by the use of redundant (spare)
units. In many cases the provision of a redundant compressor rather than
duplication of the whole system dramatically improves the reliability of a
system. Provision can be made for automatic start-up of the auxiliary or
redundant compressor in the event of an unplanned shut-down of the
principal compressor. Compressors can generally be classified into two
types: (a) reciprocating compressors, and (b) rotary compressors.

Reciprocating Compressors
Reciprocating compressors are massive, bulky, slow running (up to
50 Hz) machines having a high isentropic efficiency able to handle a virtually
unrestricted range of pressure ratios but comparatively small mass flow.
They are easy to start and to control. They are insensitive to sudden and
large changes in load. They have a high capital cost and, because they
require relatively frequent maintenance, have a high operating cost.
Their efficiency can be improved with multistage compression pro-
cesses having interstage cooling in air or water cooled heat exchangers.
Aftercoolers downstream of the compressor cool the compressed gas to
near ambient temperatures. Two stages of compression are commonly used
for the low-pressure Claude cycles (e.g. pressure ratios of up to 8 or 10).
Above this pressure ratio an increased number of stages may be used. It
is rarely worth going beyond four stages for the degree of improvement
occurs at a progressively diminishing rate.
Reciprocating compressors are found in a variety of shapes, sizes, and
arrangements. They may be double or single-acting crank-connecting rod
machines with automatic spring control or cam regulated valves. They may
be air-cooled or water-cooled, with a variety of drive systems, electric
motors, steam turbines, diesel engines. They exist in many mechanical
arrangements: inline multiple cylinder, vee-cylinder, radial cylinder, or
parallel cylinder with swash-plate drive.
An important aspect of compressors is the method of lubricating the
rubbing surfaces, the reciprocating piston rings, the rotating crankshaft
main bearings and the oscillatory big and little end bearings of the connect-
ing rods. Rubbing surfaces properly lubricated with oil have superior
Claude and Joule-Brayton Systems 349

friction and wear characteristics to any other known combination. The

science of establishing and maintaining the hydrodynamic oil film of a
self-energized metal piston ring sliding in a metal cylinder, developed to
a high level for combustion engines, can be directly carried over to gas
compressors. The act of providing an oil film on the cylinder wall must
inevitably lead to contamination of the compressed gas by oil vapor and
droplets. While of no concern to an internal combustion engine it is the
principal drawback to oil lubrication in refrigeration compressors. Provision
must be made downstream of the compressor to absorb or freeze out the
entrained oil and other contaminants, such as water and carbon dioxide,
that may be present in the compressed gas. Collins (1956) states that
alternate layers of fine steel wool and fibre-glass have been found to be
"especially effective in removing atomized oil."
To avoid the contamination problem many compressors are "dry-
lubricated." No oil is provided for lubrication. Instead dry piston rings rub
on cylinder walls and shafts run in bearings that are completely dry. This
has long been done for air compressors using carbon piston rings and
bearings, sometimes filled with a metal powder reinforcement. When the
gases are very dry, carbon is unsuitable; apparently, a trace of water is an
active participant in the local surface chemistry of a carbon bearing. A
vapor pressure of - 30°C is sometimes quoted as the criterion for use of a
"dry-rubbing" carbon bearing. Below this the carbon behaves as an abrasive
and the rubbing friction is much increased.
In these cases and many where carbon was previously used, another
material is available, the plastic polytetrafl.uoroethylene (PTFE, Teflon,
Fluon). It has a low coefficient of friction when rubbed on another surface
coated with the same material. When a piston ring or bearing of PTFE
material is rubbed on a metal cylinder wall or shaft the frictional resistance
and wear are initially high but very quickly the metal wall is "plated" with
an extremely thin layer of the wear debris torn from the PTFE component.
Thereafter the friction and wear is reduced so low that many compressors
now have PTFE-based piston rings. PTFE journal bearings are also used
but to a lesser extent for prelubricated sealed roller or ball bearings can
be used in these applications with very high reliability.
To improve the wear characteristics of PTFE a hard filler is added of
metal powders, brass, copper, silver, etc., chopped glass, or carbon powder.
The glass-filled PTFE material known as Rulon-A, dark red in color, enjoys
a high reputation.
Successful designs of compressors could be engineered with hydrody-
namic gas lubrication of pistons in the same way used by Kapitza and
Collins in expansion engines. Gas-lubricated pistons are used routinely in
free-piston Stirling engines (see Walker, 1980) which operate in both the
350 Chapter7

compressive and expansive mode. The only compressor known to have

gas-lubricated pistons is that included in the Joule-Brayton rotary stroking
engine discussed above (Schulte et at., 1965). This machine was complicated
by rotating the reciprocating elements to energize the hydrodynamic bear-
ing film. This is not acceptable for widespread application and has been
demonstrated to be unnecessary by contemporary free-piston Stirling
developments.
Reciprocating compressors are not particularly sensitive to the physical
characteristics of the gas to be compressed. Moreover, the commercial
demand for cryogenic systems has been relatively limited. As a consequence
most of the reciprocating compressors found in cryogenic service are not
recognizably different from the gas compressors found in other applications.
Many compressors used are simply air compressors suitably modified for
the gas to be compressed. Modification usually takes the form of an effort
to reduce leakage and contamination by air or lubricant. For single-acting
compressors the entire crankcase is sometimes made airtight. For double-
acting compressors the provision of a specially packed piston rod gland is
all that is necessary. With oil-lubricated machines a lubricant having a low
vapor pressure is selected. Modifications to the intercooler and aftercooler
may be necessary and supplementary filters and absorbers will likely be
added.

Rotary Compressors
Rotary compressors may be characterized as small, light-weight, fast-
running machines having a relatively low isentropic efficiency able to handle
a limited range of pressure ratio with very high rates of mass flow. They
are easy to start and control but have a limited range of operation and are
acutely sensitive to sudden large changes in load. They are comparatively
inexpensive and require little maintenance with an excellent record of
reliability. There are three principal types of rotary compressor: axial flow,
centrifugal flow, and rotary screw.
Centrifugal flow compressors are the simplest type with a single rotating
disc carrying vanes. They are low cost and reliabile but have a relatively
low efficiency and are capable of handling only a limited pressure ratio.
Axial flow compressors comprise several stages of fixed and moving
blades in which the gas is successfully compressed. The compression ratio
per stage is very small but there may be many stages and a high overall
compression may be gained. Axial flow machines are expensive but can
have a higher efficiency than centrifugal machines. They are sensitive to
the physical characteristics of the fluid being compressed.
Claude and Joule-Brayton Systems 351

Rotary screw compressors appear to be finding increased application

in cryogenic systems. They consist of two or more cylindrical rotors with
multiple helical lobes or grooves meshing so that as the rotors contrarotate
the space between the lobes decreases and the fluid trapped therein is
compressed. Oil is sprayed into the rotors at the low pressure end and
hence the term "oil-flooded screw compressor" is frequently used to
describe the system.
Gas-lubricated bearings are used increasingly with the turbomachinery,
compressors and expanders, in cryogenic systems over the entire range of
refrigeration capacity from miniature systems to large-scale commercial
gas liquefaction.

REFERENCES

Aberle, J. L., and Westbook, A. J. (1963). "Liquid Helium and Nitrogen Supply Systems for
Space Simulators." Adv. Cryog. Eng. 8, 190-198.
Anon.-CTI (1978). "Standard Helium Liquefier Model No. 1410." Tech. Mem. 91477,
CTI, Waltham, Massachusetts.
Baldus, W., and Sellmaier, A. (1968). "A Continuous Helium II Refrigerator." Adv. Cryog.
Eng. 18, 434-440.
Barron, R. (1966). Cryogenic Systems. McGraw-Hill Book Co., Toronto.
Becker, H., Doll, R., and Eder, F. X. (1968). "Measurement of the Efficiency of the Valveless
Doli-Eder Expansion Engine." Proc. Second Int. Cryog. Eng. Con[., Iliffe Sci. and Tech.
Pub., Guildford, U.K., pp. 9-11.
Breckenridge, R. W. (1969). "A 3.6 K Reciprocating Refrigerator." Adv. Cryog. Eng. 14,
387-393.
Campbell, D. N. (1967). "Miniature Closed Circuit Single Stage Claude Cycle Refrigerator
for 28 K." Proc. First Cryog. Eng. Con[., Tokyo, pp. 197-198, Heywood-Temple Ind.
Pub., London.
Carbonnell, E. (1974). "Large Scale Refrigerators with Turbines." Proc. Fifth Int. Cryog.
Eng. Con[., Kyoto, pp. 353-356, IPC Sci. and Tech., Guildford, U.K.
Clarke, M. E., and Gardner, J. B. (1968). "New Developments in Expansion Machinery for
Low Temperature Refrigerators." Proc. First Cryog. Eng. Con[., Tokyo, pp. 273-275,
Heywood-Temple Ind. Pub., London.
Class, C. R., Spero, R. P., and Mcintosh, G. E. (1960). "Efficient Utilization of Ortho-Para
Catalyst." Adv. Cryog. Eng. 3, 64-72.
Claude, G. (1913). Liquid Air Oxygen and Nitrogen. P. Blakiston Son and Co., Phila,
Pennsylvania.
Claudet, G., and Verdier, J. (1972). "Simplified Cryogenic Reciprocating Expansion Engine."
Proc. Fourth Int. Cryog. Con[., pp. 80-82, Eindhoven, IPC Sci. and Tech. Press Ltd.,
Guildford, U.K.
Collins, S. C. (1947). "A Helium Cryostat." Rev. Sci. Instrum. 18, 157.
Collins, S.C. (1952). "Helium Liquefiers." Science 116, 289.
Collins, S.C. (1956). Helium Liquefiers and Carriers. Handbuch der Physik Vol. XIV, 113-135.
Collins, S.C. (1966). "Helium Refrigerator and Liquefier." Adv. Cryog. Eng. 11, 11-15.
Collins, S.C., and Cannaday, R. L. (1958). Expansion Machines for Low Temperature Processes.
Oxford University Press, Oxford.
Collins, S.C., and Streeter, M. H. (1968). "Refrigerator for 1.8 K." Proc. First Cryog. Eng.
Con{., Tokyo, pp. 215-217, Heywood-Temple Ind. Pub., London.
352 Chapter7

Cowans, K. W., and Walsh, P. J. (1965). "Continuous Cryogenic Refrigeration for Three to
Five Micron Infrared Systems." Adv. Cryog. Eng. 10, 468-476.
Danilov, I. B., and Kovatchev, V. T. (1972). "Two-Stage Expansion Engine with Differential
Piston." Proc. Fourth Int. Cryog. Eng. Conf., pp. 87-89, IPC Sci. and Tech. Press Ltd.,
Guildford, U.K.
Daunt, J. G. (1956). The Production of Low Temperatures down to Hydrogen Temperatures.
Handbuch der Physik Vol. IV, Springer-Verlag, Berlin.
Davies, M. (1949). The Physical Principles of Gas Liquefaction and Low Temperature
Rectification. Longmans Green and Co. Ltd., London.
Dobrov, V. M., and Orlov, A. V. (1974). "The Valveless Piston Expansion Engine with
Six-Phase Gas Distribution Cycle." Proc. Fifth Cryog. Eng. Con{., Kyoto, IPC Sci. and
Tech. Press Ltd., Guildford, U.K.
Doll, R., and Eder, F. X. (1964). "Gas Lubricated Low Temperature Piston Expansion
Engine without Control Valves." Adv. Cryog. Eng. 9, 561-564.
Furihara, A., Kikuchi, K., and Nagano, H. (1968). "Closed Cycle Helium Refrigerator for
Laser Applications." Proc. First Int. Cryog. Eng. Con{., Tokyo, pp. 213-214, Heywood-
Temple Ind. Pub., London.
Gardner, J. B., and Smith, K. C. (1960). "Power Consumption and Thermodynamic Reversi-
bility in Low Temperature Refrigeration and Separation Processes." Adv. Cryog. Eng.
5, 32-46.
Gessner, R. L., and Colyer, D. B. (1968). "Miniature Claude and Reverse Brayton Cycle
Turbomachinery Refrigerators." Adv. Cryog. Eng. 13, 474-484. See also: Colyer,
D. B., and Gessner, R. L. (1968). "Miniature Cryogenic Refrigerator Turbomachinery."
Adv. Cryog. Eng. 13, 485-493.
Hood, C. B., Vogelhuber, W. W., and Barnes, C. B. (1964). "Helium Refrigerators for
Operation in the 10-30 K Range." Adv. Cryog. Eng. 9, 496-506.
Johnson, R. W., Collins, S. C., and Smith, J. L. (1971). "Hydraulically Operated Two-Phase
Helium Expansion Engine." Adv. Cryog. Eng. 16, 171-177.
Kapitza, P. (1934). "The Liquefaction of Helium by an Adiabatic Method." Proc. R. Soc.
London Ser. A 147, 189.
Kapitza, P. L., and Danilov, I. B. (1968). "Cascade Helium Liquefiers with Piston Type
Engines." Proc. First Int. Cryog. Eng. Con{., Tokyo, pp. 228-231, Heywood-Temple Ind.
Pub., London.
Kneuer, R., and Turnwald, E. (1968). "Measurements and Experience with Valveless Piston
Expansion Engines." Proc. Second Inst. Cryog. Eng. Con{., Brighton, U.K., pp. 12-15,
Iliffe Sci. and Tech. Pub. Ltd., Guildford, U.K.
Land, M. L. (1960). "Expansion Turbines and Engines for Low Temperature Processing."
Adv. Cryog. Eng. 5, 250-260.
Leger, P., Thomas, P., and Zemati, C. (1971). "Design of Valveless-Type Piston Expanders
with Variable Clearance for a 1.5 W, 4.5 K Helium Refrigerator." Proc. Third Int. Cryog.
Eng. Con{., Berlin, pp. 259-263, Iliffe Sci. and Tech. Pub., Guildford, U.K.
Long, H. M., and Simon, F. E. (1954). Appl. Sci. Res. 4, 237.
Maddocks, F. E. (1968). "Applications of Turbomachinery to Small-Capacity Closed Cycle
Cryogenic Systems." Adv. Cryog. Eng. 13, 463-473.
Malkov, M.P., and Danilov, I. B. (1968). "Cryogenic Equipment of the Institute for Physical
Problems." Proc. First Int. Cryog. Eng. Con{., Tokyo, pp. 225-228, Heywood-Temple
Ind. Pub., London.
Matsubara, Y., Ishizaki, Y., and Oshima, K. (1968). "New Type Expansion Engine for
Refrigeration." Proc. First Cryog. Eng. Con{., Tokyo, pp. 210-211, Heywood-Temple
Ind. Pub., London.
Meier, R. H., and Currie, R. B. (1968). "A 4 K Single Engine CyCle Helium Refrigerator."
Adv. Cryog. Eng. 13, 441-449.
Morain, W. A., and Homes, J. W. (1963). "An Analysis of the Performance of Large
Reciprocating Expansion Engines with the Aid of a Computer and Laboratory
Prototype." Adv. Cryog. Eng. 8, 228-235.
Claude and Joule-Brayton Systems 353

Parulekar, B. B., and Narayankhedkar, K. G. (1972). "Small-Capacity Valveless Piston

Expansion Engine Type Configuration." Proc. Fourth Int. Cryog. Eng. Con{., Eindhoven,
pp. 90-92, IPC Sci. and Tech. Press Ltd., Guildford, U.K.
Parulekar, B. B., and Narayankhedkar, K. G. (1976). "Performance of the Valveless Low
Temperature Heat Pump." Proc. Sixth Int. Cryog. Eng. Con{., pp. 77-79, IPC Sci. and
Tech. Press, Guildford, U.K.
Pastuhov, A. (1960). "Helium Refrigeration." Adv. Cryog. Eng. 1, 41-43.
Robinson, G. Y. (1965). "Large Capacity Helium Liquefier." Adv. Cryog. Eng. 10, 22-26.
Robinson, G. Y., and Hogan, W. H. (1968). "Large Scale Helium Refrigeration System for
the Range 2-20 K." Proc. First Cryog. Eng. Conf., Tokyo, pp. 219-220, Heywood-Temple
Ind. Pubs., London.
Sato, T. (1968). "Some Experiments on the Reciprocating Expander for Large Helium
Liquefiers." Proc. First Cryog. Eng. Con{., Tokyo, pp. 217-218, Heywood-Temple Ind.
Pubs., London.
Schulte, C. A., Fowle, A. A., Huechling, T. P., and Kronauer, R. E. (1965). "A Cryogenic
Refrigerator for Long-Life Applications in Satellites." Adv. Cryog. Eng. 10, 477-485.
Smith, J. L. (1967). "A Metal Bellows Expansion Engine." Adv. Cryog. Eng. 12, 595-601.
Terbot, J. W. (1976). "A New Helium Refrigerator for Superconducting Cable Systems."
Adv. Cryog. Eng. 21, 190-196.
Walker, G. (1980). Stirling Engines. Oxford University Press, Oxford.
Weinstock, H. (1968). "A Review of Cryogenic History and Principles of Refrigeration."
App/. Cryog. Techno/. 1.
Winters, A. R., and Snow, W. A. (1966). "Capacity and Economic Performance of a Large
5 K Helium Refrigerator." Adv. Cryog. Eng. 11, 116-125.
Witter, C. E. (1966). "Design of a Closed Cycle Helium Temperature Liquefier." Adv. Cryog.
Eng. 11, 107-115.
Zeitz, K., Geist, J. M., and Lashmet, P. K. (1965). "Methods for Producing Low Temperature
Refrigeration." U.S. Patent 3,199,304.
Zeitz, K., and Woolfenden, B. K. (1963). "A Closed Cycle Helium Refrigerator for 2.5 K."
Adv. Cryog. Eng. 8, 206-212.
Name Index

A. D. Little Company, [I] 10, 12, 14,237-240 Energy Research and Generation
AEG-Telefunken A. G., [I] 14 Incorporated (ERG), [I] 20
AiResearch Manufacturing Company, [I]
16, 199-202; [2] 387
Air Products and Chemical Incorporated
Fairchild Space and Electronics Company,
(APCI), [I] 15, 77, 238, 240, 293
[I] 150
American Motors, [I] Ill
Finkelstein, Theodore, [I] 109, 131, 142, 147
Flight Dynamics Laboratories, [I) 13
Ford Motor Company, [11 Ill
Baumann Institute of the Moscow High
Franchot, Charles Louis, [I] 109
Technological School, [2) 262
Beale, William, [I) 52, 166
British Oxygen Company, [I) 8, 19
Bush, Vannevar, [I) 188 General Motors, [I) Ill
Gifford, William, [I) 14, 237-263
Gorrie, John, [I] 8
Carnot, Sadi, [I] 39
Chellis, Fred, [I] 171
Claude, Georges, [I) 8, 9, 322, 323; [2] 375
Hampson, W, [I) 8
Collins, Samuel, [I] 9, 10, 326-329; [2] 376
Harwell Atomic Energy Establishment, [I]
Cowans, Ken, [I) 14
108
Crummett, Charles, [I) 9
Herschel, John, [I) 6, 95
Crummett, Orin, [I] 9
Heylandt, D., [I) 8
Cryogenic Data Center, U.S., [2] 389
Higa, Walter,[!) 16,171
Cryogenic Society of America, [2] 391
Horn. Stuart, [I] 171
Cryogenic Technology Incorporated (CTI),
Hughes Aircraft Company, [I] 14, 103,
[!) 10, 12, 14, 15, 100, 103, 238; [2] 387
191-193; [2) 387
Cryomech Incorporated, [I] 14, 238
Hughes Santa Barbara Research Center, [I]
14
Hymatic Engineering Limited, [I) 19,
Daniels, A., [I] 101
288-290
Daunt, John, [I] 14
Davis, Harvey, [I] 9
Defense Documentation Center (DDC),
U.S., [2] 388 International Institute of Refrigeration, [2]
du Pre, F. K., [I] 96, 101 391-392

3SS
356 Name Index

Japanese National Railways, [2] 306 Parsons, Sir Charles, [1]9

Jet Propulsion Laboratory, [1]16, 150 Perkins, C., [1]265
Johnson, Joseph, [1]9 Philips Company, [I] I I, 12, 20,96-101, 106,
Joule, J. P., [2] 379 I 44, I 72-176, I 79; [2] 43
Pictet, R., [I] 7
Postle, Davy, [1]7, 237-238, 261-263; [2]380
Kapitza, Peter, [I] 9, 20, 326; [2] 379
Kinergetics Incorporated, [I] 14
Kirk, Alexander, [I] 6, 95 Rayleigh, Lord, [I] 9
Kohler, Jan W. L., [I] I I, 52, 96; [2]43 Reitlinger, J ., [I] 44
Rinia, H., [1]96

L'Air Liquide Limitee, [1]8, 19,323

Landau, L., [2] 258 Schmidt, Gustav, [I] 126, 131
Linde, Karl von, [I] 7; [2]379 Siemens, Sir Charles William (also
Linde Company, [1]7, 9, 19; [2]91 Karl Wilhelm Siemens), [I] 8; [2] 382
Longsworth, Ralph, [I] 15, 240
Siemens, Sir William, [I] I 10; [2] I
Solvay, E., [I] 7
Stirling, Robert, [1]6, 95; [2] I
Magnavox Company, [I] 14
Submarine Systems, Incorporated, [I] 14
Malaker Corporation, [I] 14, 103
Martini, William, [I] 152 Sunpower Incorporated, [I] 148-150
Martin-Marietta Corporation, [I] 14, 103;
[2] 387
Taconis, T. W., [I] I 88
McDonnell Douglas/ Richland Energy
Texas Instruments Incorporated, [I] 14, 103;
Laboratory, [I] 108
McMahon, Howard, [1]237 [2] 387
Thermo-Electron Corporation, [I] 108
Mechanical Technology Incorporated, [I]
Thomson, William, [2] 382
Ill
Thrupp, Edgar, [I] 9
Meijer, Rolf, [I] 179
Trevethick, Richard, [I] 6

NPO Cryogenmash, [2] 262

National Aeronautics and Space United Stirling, [I] I I I
Administration (NASA), [I] 13, 16, 20, Urieli, Israel, [I] 147-148
I 87, I 99; [2] 226 U.S. Air Force, [I] 13, 191, 195; [2]387
NASA Lewis Research Center, [I] I 5 I; [2] U.S. Army Night Vision and Electro-Optics
387 Laboratory,[1]13, 103,192, 198;[2]387
National Bureau of Standards Cryogenic U.S. Department of Energy (DOE), [I] I I I,
Engineering Laboratory, [I] I 3 I 50; [2] 387
National Engineering Laboratory, [I] I 3 U.S. National Institutes of Health, [1]20
National Technical Information Service
(NTIS), U.S., [2] 388
Naval Engineering Laboratory, [I] I 3 van Weenan, F. L., [I] II I
North American Philips Laboratories, [I] I 2, Vuilleumier, Rudolph, [I] 188; [2] 383
13, 101, 195-199; [2]387

Werkspoor, N. V.,[J] 176-179

Office of Naval Research, [I] I 3
Onnes, Kamerlingh, [1]8
Oxley, A. J., [1]53 Zerkowitz, Guido, [ 1]9
Subject Index

Active volume, [I] 210 Blockage, [I] 345

Adiabatic compression, [2] 375 Blow period, [2) 38, 41-42
Adiabatic-cycle simulation programs, [I] Boiling heat transfer, [2] 281
142-145, 151, 152 Bounce space, [I] 166
Adiabatic demagnetization rotating frame Boyles law, [I] 270
(ADRF), [2) 247-249 Brayton cycle, [I] 258; [2] 262, 375, 379
Adiabatic expansion, [2] 375 Brushless direct-current motors, [2] 108-109
Adsorption pumping, [2] 183-185 Bubble point, [I] 34
Advanced Surveillance Technology Bucket brigade loss: see Shuttle heat transfer
program, [I] 195 (loss)
Aftercooler, [2] 375
Air compressors, [I) 349
Air cooling, [2] 104-105 Calrod-type sheathed heater, [I) 229
Air liquefaction, [I] 7, 8, II, 12, 18, 19, 26, Carbon-filled fluorocarbon piston rings, [I]
309-311, 314; [2] 2 324
Air Products Co. coolers, [I) 293 Carnot cooling engines, [I) 41
Air separation, [2) 259, 277 Carnot efficiency, [2] 132
AiResearch Vuilleumier cryocoolers, [I] Carnot thermodynamic cycle, [I) I, 39-42,
199-202 44, 47, 238; [2) 103-104, 130-134
Applications, [I] 2, 3, 10-12, 16, 17, 24, 26, Carry over loss, [I] 311; [2) 30, 33
52, 53, Ill, 187-188, 191,201 Cascade-cycle refrigeration,[!] 162,179, 180;
Artificial hearts, [I) 20; [2] 52 [2) 30
Automotive applications, [I] 52, 53, Ill Cascade system, [I] 277-280
Auxiliary refrigerating system, [I] 278-279 Casting leaks, [2] 121
Axial flow compressors, [I] 350; [2] 375 Centrifugal flow compressors, [I) 350; [2) 375
Axial flow dynamic heat exchangers, [2] Charcoal, [2] 35
30-32 Charcoal adsorption pumps, [2) 183-184
Axial heat conduction, [2] 22-24 Check valving, [2] 124
Chemical cooling system, [2] 163-165
Claude cooling engines, [I] 5, 179, 180, 257,
Balancing, [2] 68-69 297-350
Ball and roller bearings, [2] 81-82 Claude cycle, [I] 84, 257, 302-307; [2] 286,
Bearing rings, [2) 80 375
Bearings, [I] 200-20 I; [2] 79-88 Claude stepped piston two-stage expander,
Bellows, [2] 96 [I] 316-318
Bellows expansion engines, [I] 339-342 Clearance, [2] 376

357
358 Subject Index

Clearance space, [2] 376 Crowned piston, [I] 8

Close tolera nee seals, [2] 96-98 Cryocoolers, [I] I; [2] 377
Closed-cycle refrigeration, [2] 130 Cryogenerator, [2] 377
Cluster heat exchanger, [2] 5, 19 Cryogenic engineering, [I] I
Coefficient of performance (COP), [I] 42, 43; Cryogenics, [I] I; [2] 392-393
[2] 376 Cryomatic gas balancing, [I] 251-255
Coiled foil regenerator, [2] 53 Cycle simulation, [I] 145-146; [2] 266-272
Coiled tubular exchangers, [2] 6-9, 283-284 Cylinder walls, [2] 93, 99
Coldfinger, [2] 62-68 heat loss through, [I] 222
Cold finger seal, [I] 199; [2] 376
Cold string: see Cold finger
Collins cooling engines, [I] 5, 309-311 de brush-type motor, [I] 230
Collins helium liquefier (or cryostat), [I] 10, Dead space, [2] 44
318, 326-344; [2] 376 Dead volume ratio, [I] 156; [2] 377
Collins low-pressure air liquefier, [I] Dense mesh wire screens, [2] 47
309-311' 329 Design, machine: see Machine design
Combination Stirling engine: see Duplex Design charts, [I] 157-159
Stirling engine Dew point, [I] 34
Composite regenerative heat exchanger, [2] Diamagnetic materials, [2] 220
34-36 Diaphragm compressor, [I] 339
Compound working fluid, [2] 376 Diaphragm expansion engine, [I] 9
Compression space, [I] 46, 140; [2] 376 Diaphragms, [2] 96
Compressors, [I] 87-90, 344, 347-350; [2] Dichlorodifluoromethane (CCJ,F,), [2] 141
295-298, 376 Dielectric materials, [2] 160
Computer simulation programs, [I] 143, Differential pistons: see Stepped pistons
144-152,161, 192;[2]26,28,43,266-272 Dilution refrigerators, [I] 19; [2] 55, 144-145,
Concentricity, [2] 51-52 146, 179, 187-211,244-247, 259
Condensing cooling engine, [I] 41 Direct heating, [2] 377
Conduction, [I] 257 Discontinuous piston motion, [2] 377
Conduction heat leakage, [2] 64-65 Displacer, [I] 105, 208; [2] 377
Contamination, [I] 345, 348-349; [2] 82-83, Displacer motion, [I] 117-123,208,255
119-120 Displacer seal, [2] 125
Continuous He' refrigeration, [2] 185-187 Displacive materials, [2] 161
Continuous magnetic refrigerators, [2] DOE/NASA Stirling Engine Automotive
185-187 Program, [I] !50
Continuous system simulation language cssL Doll-Eder valveless expansion engines, [I]
IV, [2] 26 334-339
Convection, [I] 257; [2] 282-283 Double-acting free-piston Stirling engines,
Cooldown characteristics, [I] 274-275; [2] 26, [I] 180
65-66, 119 Double-acting Stirling engines, [I] 108-116,
Cooler, [2] 376 119-123, 180; [2] 377
Cooling, [2] I 03-106 Double-bundle nuclear refrigerator, [2]
Copper powder, [2] 196-198 235
Corblin diaphragm compressors, [I] 277 Double-circulation dilution refrigerator, [2]
Cost, [2] 114 210-211
Counterflow, [2] 15-16 Double-expansionengines, [I] 161-164, 191
Counterflow heat exchangers, [I] 7 Drive motors, [2] 106-108
Crankcase, [2] 107 Dry heli urn compressor, [2] 295-298
Crank drive, [2] 377 Dry-rubbing bearings, [2] 87-88
Critical point, [I] 35 Dry-rubbing materials, [I] 201, 348-349
Cross-flow, [2] 15 Dry-rubbing piston rings, [I] 324
Subject Index 359

Dual-pressure Claude cycle system, [I] Finegold-Vanderbrug nodal analysis, [I]

314-315; [2) 377 150-151
Ductile-brittle transformation, [2] 100 Finkelstein adiabatic cycle, [I] 131-132,
Duplex Stirling engines, [I] 50-53, 185-187, 142-145, 160; [2] 378
222; [2] 377 Finkelstein nodal analysis, [I] 147, 151
Dynamic nuclear polarization, [2] 247-251 First Law of thermodynamics, [I] 38-39
Dynamic regenerative heat exchangers, [2] Flame-trap construction (matrices), [2) 32,
30-33 46
Dynamic seals, [2] 91-93 Flash loss, [2] 30 I, 302
Flexibility, [I] 280
Flow maldistribution, [2] 19-22
Flow regulation, [I] 5, 246
Effectiveness, exchanger, [2] 17-19,21,47 Fluid friction loss, [2] 43
Efficiencies, regenerator, [2] 47
Fluid-lubricated bearings, [2] 80-81
Electric resistance heating, [I] 187-188
Fluidic-driven displacer, [I] 251-255
Electrical and electronic systems, [2] I06-110
Fluon, [I] 349
Electrocaloric refrigeration systems, [I] 19; Fluorocarbons, [2] I02
[2] 160-162
Free-displacer split-Stirling engine, [I]
Electromagnetic inlet valve, [2] 275
169-171; [2] 378
Electron spin systems, [2) 216-226 Free-piston Stirling engines, [I] 20, 164-169,
Electronic applications, [I] 169 344
Electronic controls, [2] 109-110 Freezer, [2] 378
Electrons, [2) 145-152
Friction, [I] 38, 127-128, 169, 171, 223, 339;
Enhancing heat transfer, [2] 30 [2] 24-25, 66, 81-82, I 02
Enthalpy, [I] 30, 32
aerodynamic, [I] 127-128
Entropy, [I] 30, 32; [2] 138-172, 270 Furnace-type heater, [I] 229
Entropy analysis method, [2) 270
Ericsson cooling engines, [I] 5, 75, 237-263;
[2] 28
Gap regeneration, [2] 52-53
Ericsson thermodynamic cycle, [I] 44, 48, 53;
Gas liquefaction, [I] 2, 180,271-288,294
[2] 160, 377
Gas-lubricated bearings, [I] 350; [2) 82-83,
ideal, [I] 59-60
84-87
pseudo, [I] 67-68, 83
Gas-lubricated pistons, [2] 83-84
Ettingshauser effect, [2] 151
Gaseous cooling engine, [I] 41
Europium sulfide, [2] 33
Generalized Finkelstein analysis, [I] 131-132
Exergy, [2] 271-272
Giaque-Hampson exchangers, [2] 7
Exhaust, [I] 247
Gifford- Me Mahon cooling engines, [I] 5, 15,
Expanders, [I] 91-93
16, 77,237-240, 245-261; [2) 117,378
Expansion, [I] 247
Gifford-McMahon cycle, [I] 77, 245-261
isentropic, [I] 93-94
GLAG theory, [2] 258
isothermal, [I] 93-94
Grease-lubricated bearings, [2] 81
Expansion engines, [I] 3, 4, 333; [2] 272-276
Guide rings, [2] 52, 66, 79
Expansion space, [I] 46, 138-140; [2) 377
External annular regenerator, [I] 231
External regenerator, [I] 230-231
Hampson cooling engines, [I] 5
Hampson heat exchanger, [I] 128-131, 265;
[2] 6-7, 15
Fanning friction factor, [2) 49 Harmonic drives, [I] 211
Ferroelectrics, [2] 161 Harmonic piston motion, [2] 378
Figure of merit, [I] 283; [2] 119 Harwell Fluidyne engine, [I] 168
Film coefficient, [I] 224 Hausen regenerator, [2] 36-39
360 Subject Index

He 3 refrigerators, [2] 179-187 Ideal regenerator, [2] 36

He4 -circulating dilution refrigerators, [2] Ideal Stirling cycle: see Stirling
207-211 thermodynamic cycle
He 3-He 4 dilution refrigerator, [2] 55, Ideal thermodynamic cycles, [I] 39-41, 44; [2]
144-145, 146, 187-211,259 130-134
Heat-balance analysis, [I] 259-261 Inconel 718, [I] 228
Heat exchanger thermal potential, [I] Incorporated cascade cycle, [I] 279-280
128-129 Indirect heating, [2] 378
Heat exchangers, [I] 5; [2] 1-55, 279-280, Infrared Astronomical Satellite (IRAS)
283-284 Program, [2] 225-226
Heat pipes, [2] 54-55, 106, 378 Infrared night vision equipment, [I] 12, 14,
Heat pump, [I] 48-49; [2] 378 16, 19, 169, 188, 192, 202-205, 232; [2]
Heat rejection, [I] 231; [2] 117-118 61, 92, 108, 114, 179, 184-185
Heat transfer, [2] I, 19, 279-283 Infrared telescope, [2] 225
Heat transfer losses, [I] 220-225 Inherent thermodynamic and heat transfer
Heater power input, [I] 225 losses, [I] 220-225
Heaters, [I] 229; [2] 378 Inhibited heat transfer, [2] 282-283
Helically wound wire heaters, [I] 229 Input power drive motor, [I] 160
Helium, [I] 175; [2] 82, 120-121, 139-141, Insulation, [I] 284-285
144-145, 265 Integral Stirling engine, [I] 118-119; [2] 115
Helium', [2] 179-180, 187-190,211-215 Integrated cryogenic cooled isotope engine
Helium4 , [2] 187-190 (ICICLE) program, [I] 202
Helium expansion engine, [I] 9 Intercooler, [2] 378
Helium-hydrogen liquefier, [I] 10 Intermediate cryocoolers, [I] 2, 4, 26, 109,
Helium liquefaction, [I] 10, 18,250,277, 172-180, 238; [2] 378
326-334; [2] 257-258, 286-288, 293-294, Internal annular regenerator, [I] 231
299-302 Internal energy, [I] 30,31
Hermetic sealing, [I] 195; [2] 95-96 Internal regenerator, [I] 230-231
Heylandt crowned piston, [I] 325; [2] 378 Interstage heat flow, [I] 225
Heylandt cycle, [I] 307, 309, 314-315 Inversion curve, [I] 267
Hi-Cap Vuilleumier cooling engine, [I] Inverters, [I] 230
192-193 lsenthalpic expansion, [I] 297-302; [2] 300
Historical background, [I] 6-10 Isentrope, [I] 34
Hot-end temperature controller, [I] 225-226 Isentropic efficiency, [I] 88
Hot-rider ring wear, [I] 230 Isentropic expansion, [I] 265-267, 297-302
Hughes Vuilleumier cryocooler, [I] 191-193 Isentropic process, [2] 378
Hybrid free-displacer-crank-controlled Isobar, [I] 34
piston engine, [2] 378 Isotherm, [I] 34
Hydraulic seal, [2] 96 Isothermal analysis, [I] 152
Hydraulic work-absorbing system, [I] 326 Isothermal efficiency, [I] 88
Hydrodynamic fluid lubrication, [I] 324, Isothermal process, [I] 299-300; [2] 379
326, 334-339, 342, 349; [2] 80-87 Isothermality, [I] 126-127
Hydrogen, [I] 175; [2] 82, 143
Hydrogen expansion engine, [I] 8, 9
Hydrogen liquefaction, [I] 8, 18 Japanese cryocooler development, [2]
Hymatic coolers, [I] 288-290 293-313
Hyperfine enhanced nuclear-spin systems, Josephson tunnel diodes, [I] II
[2] 240-244 Joule-Brayton cycle liquefiers, [I] 82-83,
297, 333, 342-345, 347-350; [2] 379
Joule-Thomson coefficient, [I] 268-271
Ideal Ericsson cycle: see Ericsson Joule-Thomson cooling engines, [I] 5, 18,
thermodynamic cycle 19, 265-294, 347
Subject Index 361

Joule-Thomson expander, [I] 250-251 Martini design manual, [I] 152

Joule-Thomson expansion, [I] 79-83, Mass distribution, [I] 124, 141
250-251,265-294,297,299,300,333;[2 ] Material properties, [2] 264-265
379 Materials, [I] 243; [2] 32, 98-103, 119-120
Joule's law, [I] 270 cold-regenerator matrix, [I] 230
regenerator matrix, [I] 256
Matrix: see Regenerative matrix
Kapitza hydrodynamic lubricated piston, [I] Maximum inversion temperature, [I]
326 267-268
Kapitza resistance, [2] 195-201, 206-211 Mean cycle pressure, [I] 137
Kinematic drive mechanism, [I] 105; [2] 379 Mechanical efficiency, [I] 88
Kirk cycle, [I] 95 Metal bellows seal, [I] 324
Krytox-AB, [2) 120 Metallurgical limit, [2] 379
Methane, [2] 143
Microminiature cryocoolers, [I] 2, 3; [2] 379
Labyrinth seals, [2] 97 Microphonics, [2] 118
Large cryocoolers, [I] 2, 4, 26, 109, Ill, 180, Millikelvin temperature cooling systems, [2]
188-192; [2) 379 171-251
Leakage (gas), [I] 257, 335; [2] 33, 97, Miniature cryocoolers, [I] 2, 3, 4, 24, I03,
120-I21 188, 198, 288-293, 344-345; [2] 113-128,
Leidenfrost boiling, [I) 16 379
Linde co-axial heat exchanger, [2] 15 Mixed refrigerant cycle, [I] 346-347
Linde cooling engines, [I] 5 Mixed units, [I] 4
Linde dual-pressure cycle, [I] 84, 280-282, Mixtures, [2] 165-168, 187-190
287-288 Motional heat leak, [I] 260
Linde-Frane) system, [I] 309 Motors, [I) 230
Linde-Hampson cycle, [I] 83-84, 265, Multiple-element cooling systems, [I]
271-280, 284-288, 345; [2] 261' 379 116-117
Linear bearings, [2] 79, 80 Multiple-expansion engines, [I] 12, 161-164,
Liquefaction process, [2] 299-302 247-250, 316-322; [2] 33
Liquid natural gas production, [I] 280 Multiple-expansion Gifford-McMahon
storage, [I] 18, 172 cycle, [I] 247-250
transport, [I] 172 Multiple mixing chambers, [2) 206-207
Liquid-piston engines, [I] 168-169 Multiple reciprocating masses, [2] 74-76
Literature, cryogenic engineering, [2] Multistage compression, [I] 90-91, 348
387-394 Multistage refrigerators, [2) !50
Load changing, [2] 26 Multistage Vuilleumier coolers, [I] 225-227
Long engine-operation life, [I] 186-187, 191; M ultistaging, [I] 249
[2] 113-114, 117
Low-pressure air liquefiers, [I] 309-311
Lubrication, [I] 231, 348-349; [2] 79-87, NASA Thermal Analysis Program (TAP),
96-98 [I] 147
Net refrigeration, [I] 225
Nitrogen, [2] 143
Machine design, [I] 159-161, 255-257 Nodal analysis, [I] 145-152, 160; [2] 28,43
Magnetic refrigerators, [2] 216-251 Nonisothermal compression and expansion,
Magnetically-levitated (MAGLEV) vehicles, [I] 126-127
[2) 306 NTU (number of transfer units), [2] 20,
Magnetocaloric effect, [2) 160 21-22, 29-30, 41
Magnetocaloric refrigeration systems, [I] 19; NTU-effectiveness method, [2] 41
[2] 155-160, 161 Nuclear-spin systems, [2] 226-240
MAN/ MWM nodal analysis, [I] 151 Nusselt number, [2) 48
362 Subject Index

Oil flooding, [2] 380 Precooling, [I] 275-278, 312-314; [2] 380
Open-cycle refrigeration, [2] 130 Pressure, [I] 30-31
Optimization of design parameters, [I] Pressure drop, [I] 127-128; [2] 5, 24, 380
153-157, 255-257; [2] 266-272 Pressure excursion, [2] 380
Optimum recirculation fraction, [I] 306-307 Pressure generator, [I] 49-50
Organ nodal analysis, [I] 151 Pressure oscillation, [2] 305
Orthohydrogen, [2] 143 Pressure ratio, [I] 186; [2] 380
Oscillatory bearings, [2] 79, 80 Pressure-volume ( P- V) diagram, [I] 36-37
Oscillatory flow, [2] 4, 26-30, 41-43, 302-305 Prime mover, [I] 48-49; [2] 380
Oscillatory temperature, [2] 302-305 Publications, cryogenic engineering, [2]
Outer cascade cycle, [I] 279 262-264, 387-394
Overheating, [I] 228 Pulse-width modulation, [1]228
Pulsed refrigeration system, [2] 308-313
Pump work, [2] 5
Parahydrogen, [2] 143 Pumping loss, [1]221-222
Parallel flow, [2] 15
Parallel multiple-expansion engine
arrangement, [I] 318-322 Radiation, [I] 257
Paramagnetic materials, [2] 218-226 Rallis adiabatic regenerative cycle, [1]53,
Parametric effects, [I] 71-73 62-70
Peltier heat, [2] 146 Rallis isothermal regenerative cycle, [1]53-58
Pentaerythrityl fluoride, [2] 154 Rallis thermodynamic cycle, [I] 53; [2] 380
Perfect dynamic balance, [2] 77-79 RCA Vuilleumier cryocoolers, [I] 202-206
Perforated plate exchanger, [2] 4, 11-12 Reciprocating compressors, [1]347-349; [2]
Phase angle, [I] 156, 227; [2] 380 91
Phase equilibrium, [2] 136 Reciprocating cooling machines, [I] 2, 3, 4;
Philips nodal analysis, [I] 151 [2] 380
Philips Vuilleumier cryocoolers, [I] 195-199 Reciprocating expansion engines, [I]
Phonon drag effect, [2] 152-153 322-324 [2) 298-299
Phonons, [2] 137, 147, 152-153 Reciprocating masses, [2] 71-79
Photon cooling systems, [2] 168-171 Recuperative cycles, [I] 78-94
Piston crosshead system, [2] 94 Recuperative heat exchangers, [ 1]5, 78, 257,
Piston-displacer engine, [I] 50-53 346; [2] 1-30, 32, 380
Piston-displacer single-acting Stirling Recuperative system analysis, [1]85-87
engines, [I] I 06 Reduced length, [2] 39-41
Piston leakage, [I] 339 Reduced period, [2] 39-41
Piston motion, [I] 117-124, 243 Redundant units, [I] 347
Piston rings, [I] 324; [2] 92, 274 Refrigerant, [2] 139-143
Piston seals, [I] 169 Refrigeration, [I] 256
Piston side thrust, [2] 93-95 dilution refrigeration, [I] 19
Pistons, [I] I 05, 325-331; [2] 380 electrocaloric refrigeration, [1]19
Plate-fin exchangers, [2] 4, 9-11 magnetocaloric refrigeration, [1]19
Plated tube heat exchanger, [2] 23-24 Refrigeration capacity, [1]2, 128, 154,
Polytetrafluoroethylene (PTFE), [I] 749; [2] 157-159; [2] 380
87-88, 92 Refrigeration load, [2] 380
Polytropic process, [I] 87, 298-300 Refrigeration loss, [I] 250
Pomeranchuk refrigeration, [2] 144, 211-216 Refrigeration quality, [IJ I
Porosity, [2] 121, 380 Regenerative annulus, [I] 131; [2] 381
Postle cryocoolers, [I] 237-238, 261-263; [2] Regenerative cycles, [ 1]44-77; [2] 381
380 Regenerative displacer, [I] 233
Practical regenerative cycle, [I] 123-131 Regenerative heat exchangers, [I] 5, 6, 257;
Precooler heat exchanger, [I] 275-277 [2] I, 3, 30-54, 381
Subject Index 363

Regenerative matrix, [I] 128, 130-131, Series multiple-expansion engine

256-257; [2] 32, 43-4 7, 381 arrangement, [I] 318-322
Regenerator contamination, [I] 130 Shuttle heat transfer (loss), [I] 220-221, 260;
Regenerator heat transfer losses, [I] 223-224 [2] 67-68, 381
Regenerator pressure drop, [I] 128, 256 Siemens cycle, [I] 83; [2] 382
Regenerators, [I] 230-231, 256-257; [2] Siemens double-acting four-cylinder engine,
66-67, 283-284 [I] I 09-11 I
Regenerator thermal saturation, [I] 130-131 Silica gel, [2] 35
Reitlinger cycle, [I] 44-45; [2] 381 Silver powder heat exchangers, [2] 199-200,
Reliability, [I] 258, 347; [2] 62, 150 204
Residence time, [2] 41 Simulation programs: see Computer
Reversal period, [2] 38 simulation programs
Reversible mixing, [2] 166-168 Single-acting piston-displacer Stirling
Reversing recuperative exchangers, [I] 311 engine, [I] 161
Reynolds number, [2] 48 Single-acting "rhombic drive" engine, [I] Ill
Rhombic drive mechanism, [I] 179, 195; [2] Single-acting Stirling engines, [I] 105-108
7, 8, 94-95, 381 Single-bow transient technique, [2] 40
Rietdijk expansion ejector, [I] 282-283 Single-cycle He 3 refrigerators, [2] 181-185
Roll-sock seal, [I] 195; [2] 381 Sintered powder heat exchanger, [2] 196-198,
Rolling diaphragm seals, [2] 95, 96 203
Rotary bearings, [2] 79-80 Sintering, [2] 45
Rotary compressors, [I] 4, 344, 347, 350; [2] Sinusoidal oscillatory flows, [2] 26-29
298 Size, [2] 114
Rotary cooling machines, [I] 2, 3; [2] 381 Small cryocoolers, [I] 3, 4, 13, 240; [2] 382
Rotary expansion engines, [I] 344 Solvay cooling engines, [I] 5, 76-77, 237-243,
Rotary screw compressors, [I] 350 257
Rotary stroking engine, [I] 342-344 Solvay cycle, [I] 76-77, 240-243
Rubber seals, [2] 121 Space applications, [I] 26, 187, 191, 202; [2]
Rulon- A, [I] 349; [2] 88, 92 I 05-106, 394
Russian cryocooler development, [2] 257-291 Spacecraft radiative cooling, [2] 105-106
Spacers, [2] 4, II
Specific heat, [2] 33-36, I00-102
Saturated liquid, [I] 33 Speed, engine, [2] 76
Saturated vapor, [I] 34 Split-Stirling cryocoolers, [I] 16, 119, 192
Saturation curves, [I] 35 Split-Vuilleumier cryocoolers, [I] 192,
Schmidt isothermal cycle, [I] 126, 131, 198-199,231-233
134-142, 153; [2] 381 State of cyclic operation, [2] 36
Schock nodal analysis, [I] 150, 151 State properties, [I] 29-32
Schulte rotary stroking engine, [I] 344 Static regenerative heat exchangers, [2] 30,
Scotch-yoke system, [2] 94 32
Screw compressor, [2] 381 Static seals, [2] 89-91
Seal rings, [2] 9 Status surveys, [I] 20-26
Seals, [I] 195,201,227,238,257, 324; [2] 30, Steady enthalpy flow method, [2] 41
89-98, 121-128 Stepped pistons, [I] 316-317
Second Law of thermodynamics, [I] 38-39, Stirling cooling engines, [I] 5, II, 14, 75,
42 95-180, 185, 192, 201; [2] 26, 28,41-47,
Self-acting valve arrangement, [I] 261-263 I 15, 272, 278-279
Self-cleaning ability of regenerative heat design parameters, [I] 152-157
exchangers, [2] 33 Stirling cycle liquefiers, [I] 26
Self-regulating Joule-Thomson coolers, [I] Stirling engine theoretical analysis, [I]
291 131-152
Semipermeable membranes, [2] 166 Stirling engines, [I] 185
364 Subject Index

Stirling Nodal Analysis Program (SNAP), Thermal leaks, [I] 257

[I] 150 Thermal load, [I] 128
Stirling thermodynamic cycle, [I] 44-50, 53, Thermal losses, [I] 160
75 Thermal regeneration, [I] 65-67
ideal, [I] 58-59, 132-134 Thermal regenerator, [I] 208-209
pseudo, [I] 68-71 Thermal resistance, [2] 230, 238
Storage battery, [2] 165 Thermal saturation, [I] 130-131;{2] 34
Storage of liquefied gas, [I] 18, 19; [2] 284, Thermal storage, [I] 52-53
288-291 Thermal wheels, [2] 32
Straight-wire heaters, [I] 229 Thermodynamic analysis, [2] 266-272
Straw regenerator, [2] 33 Thermodynamics of cryocoolers, [I] 29-94
Strength, material, [2] 99-100 Thermoelectric refrigeration, [2] 145-152,
Sunpower nodal analysis, [I] 148-151 168
Superexpress trains, [2] 306 Thermomagnetic effects, [2] 151
Superconducting electric cable Thermophonics, [2] 118-119
transmission, [I] 172 Throttled expansion: see Joule-Thomson
Superconducting electronic devices, [I] 18, expansion
164, 172, 238; [2] 389 Tidal flow. [2] 4
Superconducting heat switches, [2] 224-225 Toughness, material, [2] 99-100
Supercond ucting magnets, [2] 306, 308-312 Transient response, [2] 25-26
Superconducting quantum interference Transportation of liquefied gas, [2] 284,
device (SQUID), [I] II, 26 288-291
Superconductivity, [I] II Triple-expansion Stirling engine, [I] 162
Superconductors, [2] 139 Tubular exchangers, [2] 4-9
Superheated fluid, [I] 35-36 Tubular regenerator, [I] 231
Superleak, [2] 208 Turbines, [I] 3, 9, 10,231.
Swash-plate, [I] Ill; [2] 382 Turbo compressors, [2] 382
Swept volume ratio, [I] 155; [2] 382 Turbo expanders, [2] 272, 277-278, 382
Two-phase single-component working fluid
cryocooler, [I] 18
Teflon, [I] 349; [2] 87, 88 Two-piston single-acting Stirling engines, [I]
See also Polytetrafluoroethylene (PTFE) 105-106
Temperature, [I] 29, 31 Two-piston Stirling engine, [I] 117-118, 192
Temperature-entropy ( T-S) diagram, [I]
33-36, 126
Temperature oscillation or instability, [2] Ultralow-temperature cooling systems, [2]
302-305 177-251
Temperature ratio, [I] 154; [2] 382 Underwater applications, [I] 26
Terbot mixed-refrigerant cycle, [I] 346-347 Uniflow expansion engine, [2] 272-276
Tew-Valentine nodal analysis, [I] 150-151 United Stirling nodal analysis, [I] 151
Thermal Analysis Program: see NASA Urieli nodal analysis, [I] 147-148, 151
Thermal Analysis Program USSR, cryocooler development in the, [2]
Thermal buffer, [I] 162 257-291
Thermal capacity, [I] 35-36 Utilization factor, [2] 39
Thermal conductivity, [2] 100-101, 197-198
Thermal contraction coefficient, [2] 100-102
Thermal design, [2] 17-18 Valveless expansion engines, [I] 334-339
Thermal efficiency, [2] 382 Valves, [I] 5, 243-247, 255-257; [2] 275,
Thermal energy input, [I] 187 298-299
Thermal fatigue, [2] 31 Van Vleck arrangements, [2] 241-244
Thermal isolation, [2] 64 Vapor compression machines, [I] 42
Subject Index 365

Void volumes, [I] 217-218; [2] 383 Vuilleumier cycle liquefiers, [I] 26
Volume, [1]30
Volume compression ratio, [2] 383
Volume variations, [1]5 Wall effect, [2] 18
Volumetric efficiency, [1]89 Water cooling, [2] I 05
Vuilleumier cooling engines, [1]5, 14, Wear, [2] 66, 93
185-233; [2] 26, 383 Weld joints, [2] 121
accessories and components, [1]228-229 Werkspoor cryocooler, [1]176-179
power input section, [I] 209-210 Wobble-plate, [I] Ill
refrigeration section, [I] 210-211 Work diagrams, [I] 124-126
Vuilleumier cycle, [ 1]75-76, 191, 206-220; [2] Working fluid, [1]175; [2]96-98, 383
116 Working space, [2] 383