refrigeration for cryogenic sensors and electronic systems ... · electromagnetic technology...

Sponsored by International Institute of Refrigeration-Commission A 1 12

Office of Naval Research-Naval Research Laboratory Cryogenic Engineering Conference

National Bureau of Standards

NAL BUREAU OF STANDARDS

The National Bureau of Standards1 was established by an act of Congress on March 3, 1901. The Bureau's overall goal is to strengthen and advance the Nation's science and technology and facilitate their effective application for public benefit. To this end, the Bureau conducts research and provides: (1) a basis for the Nation's physical measurement system, (2) scientitic and technological services for industry and government, (3) a technical basis for equity in trade, and (4) technical services to promote public safety. The Bureau's technical work is performed by the National Measurement Laboratory, the National Engineering Laboratory, and the Institute for Computer Sciences and Technology.

THE NATIONAL MEASUREMENT LABORATORY provides the national system ot physical and chemical and materials measurement; coordinates the system with measurement systems of other nations and furnishes essential services leading to accurate and uniform physical and chemical measurement throughout the Nation's scientific community, industry, and commerce; conducts materials research leading to improved methods of measurement, standards, and data on the properties of materials needed by industry, commerce, educational institutions, and Government; provides advisory and research services to other Government agencies; develops, produces, and distributes Standard Reference Materials; and provides calibration services. The Laboratory consists of the following centers:

Absolute Physical Quantities2 - Radiation Research - Thermodynamics and Molecular Science - Analytical Chemistry - Materials Science.

THE NATIONAL ENGINEERING LABORATORY provides technology and technical services to the public and private sectors to address national needs and to solve national problems; conducts research in engineering and applied science in support of these efforts; builds and maintains competence in the necessary disciplines required to carry out this research and technical service; develops engineering data and measurement capabilities; provides engineering measurement traceability services; develops test methods and proposes engineering standards and code changes; develops and proposes new engineering practices; and develops and improves mechanisms to transfer results of its research to the ultimate user. The Laboratory consists of the following centers:

Applied Mathematics - Electronics and Electrical Engineering2 - Mechanical Engineering and Process Technology2 - Building Technology - Fire Research - Consumer Product Technology - Field Methods.

THE INSTITUTE FOR COMPUTER SCIENCES AND TECHNOLOGY conducts research and provides scientific and technical services to aid Federal agencies in the selection, acquisition, application, and use of computer technology to improve effectiveness and economy in Government operations in accordance with Public Law 89-306 (40 U.S.C. 759), relevant Executive Orders, and other directives; carries out this mission by managing the Federal Information Processing Standards Program, developing Federal A D P standards guidelines, and managing Federal participation in ADP voluntary standardization activities; provides scientific and technological advisory services and assistance to Federal agencies; and provides the technical foundation for computer-related policies of the Federal Government. The Institute consists of the following centers:

Programming Science and Technology - Computer Systems Engineering.

'Headquarters and Laboratories at Gaithersburg, MD, unless otherwise noted; mailing address Washington, DC 20234. 2Some divisions within the center are located at Boulder, CO 80303.

Refrigeration for Cryogenic Sensors and Electronic Systems

Proceedings of a Conference held at the National Bureau of Standards, Boulder, CO, October 6-7, 1980

Edited by:

J. E. Zimmerman, D. B. Sullivan, and S. E. McCarthy

Electromagnetic Technology Division Center for Electronics and Electrical Engineering National Engineering Laboratory National Bureau of Standards Boulder, CO 80303

Sponsored by: International Institute of Refrigeration - Commission A 112 Office of Naval Research - Naval Research Laboratory Cryogenic Engineering Conference National Bureau of Standards

U.S. DEPARTMENT OF COMMERCE, Malcolm Baldrige, Secretary NATIONAL BUREAU OF STANDARDS, Ernest Ambler, Director

Issued May 1981

Library of Congress Catalog Card Number: 8 1-600038

National Bureau of Standards Special Publication 607 Nat. Bur. Stand. (U.S.), Spec. Publ. 607, 223 pages (May 1981)

CODEN: XNBSAV

U.S. GOVERNMENT PRINTING OFFICE WASHINGTON: 1981

For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402 Price $6.50

(Add 25 percent for other than U.S. mailing)

ABSTRACT

This document contains the proceedings o f a meeting o f r e f r i g - e ra t i on spec i a l i s t s he ld a t the Nat ional Bureau o f Standards, Boulder, CO, on October 6 and 7, 1980. Pa r t i c i pa t i on included represen- t a t i v e s of indust ry , government, and academia. The purpose o f the meeting was t o discuss progress i n the development o f r e f r i ge ra t i on systems which have been spec ia l ized f o r use w i t h cryogenic sensors and e lec t ron ic systems. The meeting focused p r i m a r i l y on the temperature range below 20 K and coo l ing capaci ty below 10 W . The meeting was j o i n t l y sponsored by the In te rna t iona l l r ~ s t i t u t e o f Refrtgeration-Commission A 1 t2 , the Of f ice o f Naval Research, the Naval Research Laboratory, the Cryogenic Engineering Conference, and the Natfonal Bureau of Standards.

Key words: Cryocoolers; cry0geni.c sensors; he1 i um; re f r ige ra t ion ; superconducting devices.

DISCLAIMER

Except where a t t r i b u t e d t o NBS authors, the content o f i nd i v i dua l sect ions of t h i s volume has no t been reviewed o r ed i ted by the Nat ional Bureau o f Standards. NBS there- f o re accepts no responsi b i 1 i t y f o r qua1 i t y o f copy, comments, o r recommendations there in . The mention o f t rade names i n t h i s volume i s i n no sense an endorsement o r recommendation o f the Nat ional Bureau o f Standards.

TABLE OF CONTENTS

Page

Cover: Drawing o f t he f i r s t S t i r l i n g engine from t he o r i g i n a l 1816 patent. The process of regenerat ive heat exchange i n t h i s engine i s a key fea tu re o f many present day cryocoolers. From the book S t i r l ing-Cycle Machines by G. Wal ker (Clarendon Press, Oxford, 1973).

. . . . . . . . . . . . . . . . . . . . . . . . 1. I r i t roductory Remarks and Summary, 1

2. "A Preview o f Unconventional Suggested Cooling Processes and Some

Prac t i ca l Problems Related t o Small Ref r igerators ," . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . by P ie r r e M. Roubeau 3

3. Ref ri g e r a t i on Requirements fo r Superconducti ng Computers , " . . . . . . . . . . . . . . . . . . . . . . . . by R. W. Guernsey and E. B. F l i n t 15

4. "Development Approaches f o r Long-Life Cryo-Coolers,"

. . . . . . . . . . . . . . . . . . . . . . . by Ronald White and Wi l l iam Haskin 21

5. "Theoret ical Analysis of A 3-Stage S t i r l i ng Cycle Cryocooler ,I1

by S tuar t B. Horn and Mark S. Asher . . . . . . . . . . . . . . . . . . . . . . . 30

6. "Some Thermodynamic Considerations o f He1 ium Temperature Cryocoolers , " . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . by D. E. Daney 48

7. "Dynamic Analysis o f a Small Free-Piston Resonant Cryoref r igerator , "

byR.A.Ackermann . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

8. "Methods f o r the Measurement o f Regenerator Ineffect iveness,"

by Ray Radebaugh, Del Linenberger, and R. 0. Voth. . . . . . . . . . . . . . . . . 70

9. "Serviceable Re f r ige ra to r System f o r Small Superconducting Devices,"

by Ralph C. Longsworth. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

10. "Performance o f A 1 Watt 4K Cryosystem Su i tab le f o r A Superconducting Computer,"

by E. B. F l i n t , L. C. Jenkins, and R. W. Guernsey . . . . . . . . . . . . . . . . 93

11. "Developments Toward Achievement o f a 3-5 Year L i f e t ime S t i r l i n g Cycle

Re f r ige ra to r f o r Space Appl icat ions,"

by Max G. Gasser, A l l an Sherman, and Wi l l i am Beale. . . . . . . . . . . . . . . . 103

12. "Thor Cryocooler,"

by G. M. Benson and R. J. Vincent . . . . . . . . . . . . . . . . . . . . . . . . 116

v

TABLE OF CONTENTS, continued

"A Cryogenic System f o r the Small I n f r a red Telescope f o r Spacelab 2,"

by E. W. Urban, L. Katz, J. B. Hendricks, and G. R. Karr. . . . . . . . . . . . . 117

"Requirements f o r and Status o f a 4 . 2 ' ~ Adsorption Re f r ige ra to r Using Zeol i tes,"

b y W i l l i a m H . H a r t w i g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I 2 7

"Cooling o f SQUID Devices by Means o f LTquid Transfer Techniques f o r Reduced

He1 ium Consumption and Enhanced Temperature S tab i l i ty,"

by Eldon A. Byrd and R. G. Hansen . . . . . . . . . . . . . . . . . . . . . . . . 136

"A Contamination Free Compressor fo r Small Scale S t i r 1 Png Refr igerators,"

by J. G. Daunt and C. Heiden . . . . . . . . . . . . . . . . . . . . . . . . . . 141

"A B i -Dt rect tonal L jnear Flotork6enerator w i t h I n t eg ra l Magnetic Bearfngs

f o r Lang L t fe t ime S t f r l i,ng Cycle Refri;gerators ," by PR i l t p A. Studer and Max G . Gasser . . . . . . . . . . . . . . . . . . . . . . 146

"Desfgn C0nsi:deratl'ons fo r Mitromi.ntatnre Refr igerators Ustng Laminar Flow

Heat bchangers, " . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . by W. A . L i t t l e . 154

"Progress i n t he Development o f Mtcrominiature Re f r ige ra to rs Using

Photol i thographtc Fabr i ca t ion Techniques," . . . . . . . . . . . . . . . . . . . . . . . . . by R. Hollman and W. A. L i t t l e . 160

"Low Temperature Regenerators, " . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . by Lawrence A. Wade 164

"Measurement o f Thermal Proper t ies of Cryocooler Mater ials," . . . . . . . . by 3. E. Zimmerman, D. B. Sul l ivan, R. L. Kautz, and R. D. Hobbs. 173

"Gas Heat Switches,' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . by Emanuel Tward. 178

"Operation o f a Prac t i ca l SQUID Gradiometer i n a Low-Power S t i r l i n g Cryocooler," . . . . . . . . . . . . . . . by D. B. Sul l ivan, J. E. Zimmerman, and J. T. Ives 186

"Current Status o f High Temperature Josephson Device Techno1 ogy, " . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . by M. Nisenoff . 195

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Attendance L i s t 210 v i

INTRODUCTORY REMARKS AND SUMMARY

This document contains the proceedings o f the second of two conferences .he ld a t the Nat ional Bureau o f Standards Boulder Laborator ies on the subject o f c losed-cycle cryocoolers f o r small superconducting devices and e lec t ron ic systems.

The f i r s t conference, whose proceedings are ava i lab le from the Superintendent of Documents, U. S. Government P r i n t i n g Office, Washington, DC, as NBS Special Pub1 i c a t i o n 508, e n t i t l e d "Appl i ca t i ons o f Closed-Cycle Cryocool ers t o Small Superconducting Devices", was a meeting of about 40 i n v i t e d speakers and par t i c ipan ts . I t s purpose was t o review the s ta te-of - the-ar t o f small cryocoolers fo r temperatures below.about 20 K, and t o describe the needs o f var ious prospect ive users of cryocoolers f o r superconducting and other cryogenic instruments used i n biomagnetism, geophysics, magnetic anomaly detect ion, superconducting computers, Josephson vol tage standards, space appl icat ions, m i 11 imeter-wave and i n f r a red detect ion, and laboratory measurements. Also presented a t t h i s meeting were some new ideas f o r very-low-power low-interference cryocoolers, S t i r 1 i n g and Joule-Thomson types, p a r t i c u l a r l y su i ted t o many of the l i s t e d appl icat ions. A subject o f spec ia l i n t e r e s t was a review o f progress on high-T superconductors, s ince the cost, complexity, and d r i v e power of a cryocooler can be reduEed considerably i f the instrument can be operated a t temperatures above 6 o r 8 K ra ther than a t 4 K o r below. There are, o f course, some appl icat ions where lower temperatures are requi red i n any case.

I n short, the purpose o f the f i r s t conference was t o de f ine the problem. The present conference was organized q u i t e d i f f e r e n t l y . It was open t o a l l i n t e res ted par t i c ipan ts , and most papers were contr ibuted. I n v i t e d papers are those by Roubeau and by Nisenoff , the former being a r a t he r f ree-ranging discussion o f some unconventional approaches t o cryogenic technology, and the l a t t e r a comprehensive review o f recent developments i n high- T superconducting devices and mater ia ls . The cont r ibuted papers are on r e f r i g e r a t i o n con- cgpts, systems and components su i ted t o the support o f cryogenic sensors and e l ec t r on i c systems below 20 K. Techniques and components developed f o r higher temperatures t h a t are appl icab le t o low-temperature r e f r i g e r a t i o n are included.

I n reviewing the progress t h a t has been made dur ing the l a s t three years, i t i s apparent t h a t t he desired goals have no t been achieved. Nevertheless, the re have been some notable developments. Several funding agenices, i n p a r t i c u l a r t he O f f i c e o f Naval Research and Wright-Patterson A i r Force Base, are now support ing conceptual s tud ies and developments of low-power cryocoolers f o r superconducting instruments, and a number o f such p ro jec ts are underway. As described i n t h i s repor t , a helium l ique f ie r -c ryocoo le r f o r a superconduct ing computer has been b u i l t and i s being tested. Also i n t h i s r epo r t are papers dem- ons t ra t ing increasing i n t e r e s t and emphasis on gas r e f r i g e r a t o r s using resonant mechanical systems dr iven by l i n e a r rec ip roca t ing e l e c t r i c motors and supported by gas bearings o r magnetic suspensions. These systems o f f e r the po ten t i a l o f u l t ima te simp1 i c i ty, freedom from contamination, long l i f e , and (eventual ly ) low cost. A1 though the operat ion o f such systems may a t f t r s t glance appear somewhat subt le, we might p o i n t o u t t he obvious, namely t h a t the mathematical desc r ip t ion o f a resonant rec ip roca t ing cryocooler i s i d e n t i c a l (except f o r non-1 i n e a r i t i e s ) i n a1 1 essent ia ls t o t h a t o f coupled e l e c t r i c a l resonant c i r c u i t s w i t h a sinusoidal d r i v i n g force, a system as f a m i l i a r as sunrise t o most e l ec t r on i c engineers (Probably even the non-1 i nea r i t i e s are analogous t o a considerable extent ) . F i na l l y , a book by Prof. G. Walker, S t i r l i n g Machines, has j u s t been publ ished by Plenum Press, and a companion book, Cryocoolers, i s i n press. These books are a very comprehensive source o f in format ion and references on the whole technology.

Approximatley one-half (57) of t he pa r t i c i pan t s a t t h i s meeting returned a quest ion- na i re , d i s t r i b u t e d dur ing the meeting, w i t h answers t o three questions. To the quest ion "Has the conference been worthwhile?", 56 sa id yes and 1 sa id somewhat. To "Should i t be he ld again?", the answer was yes, unanimously. To "At what in te rva l? ' ' 11 sa id 1 year, 33 sa id 2 years, 12 sa id 3 years, and 1 sa id 4 years.

There were several useful comments and c r i t i c i sms . One was t o a l low more lead t ime and t o prov ide f o r d i s t r i b u t i o n of r e p r i n t s a t the meeting. Some comments concerning more i ndus t r y pa r t f c i pa t i on are we l l taken and we suggest t h a t an e a r l i e r planning meeting by an organiz ing committee made up o f an even mix of indust ry , un ive rs i t y , and government representat ives might lead i n the r i g h t d i rec t ion . One should note t h a t both government and i.ndustry papers were wtthdrawn because they could no t be c leared by the i - r o r an izat ions [ p r o p r b t a r y considerat ions w i l l c e r t a i n l y const ra in some i,ndustry par t ic ipat i 'onq. We Bel i3ye t h a t the conference wi:ll be more l i 'vely and usefu l i f the mix o f pa r t i c i pan t s provldes f o r dtscussfon o f both new concepts and we l l conceived and engineered systems.

There were several suggestions t h a t the papers should have been b e t t e r screened. This i s an o f t - r a i sed c r i t i c i s m w i t h every conference. I t i s c l ea r t h a t papers can be re jec ted f o r obvious technica l f laws o r f o r f a i l u r e t o f i t i n t o the bas ic l i m i t s o f the conference. However, judgments concerning the q u a l i t y o f a con t r i bu t i on are very d i f f i c u l t t o make a,nd many conferences have opted t o exercibe such j ud ment on ly i:n the publtshed conference proceedi:ngs. This h what we c b s e t o do i:n 19.80. !be next orgsnizi:ng committee w i l l have t o face the questfon again.

To conclude, and q u i t e apar t from the quest ion o f the usefulness o f t h i s meeting o r s i m i l a r meetings, i t might be usefu l t o repeat again t he growing conv ic t ion t h a t the u l t ima te f a t e o f superconducting and o ther cryogenic instruments w i l l depend heav i l y on the development o f re1 iab le , compatible, low-cost cryocoolers 1 i ke those envisioned by the aurhors o f the f o l l ow ing papers.

The Ed i to rs

Organizing Comml'ttee: J . E.. Zimmerman , National ~ u r e a u of Standards - Chairperson E. A. Edelsack, Offi'ce o f Naval Research R. W. Guernsey, In te rna t iona l Business Machines W. Haskin, A i r Force F l i g h t Dynamics Laboratory F1. Ni:senoff, Naval Research Laboratory J. L. Olsen, Eidg. Technische Hochschule A. Sherman, Nat ional Aeronautics & Space Admin is t ra t ion D. B. Su l l ivan, Nat ional Bureau o f Standards K. D. Timmerhaus, Un i ve r s i t y o f Colorado

Conference Secretary: S. E. McCarthy, Nat ional Bureau of Standards

A PREVIEW OF UNCONVENTIONAL SUGGESTED COOLING PROCESSES AND SOME PRACTICAL PROBLEMS RELATED TO SMALL REFRIGERATORS

Pierre M. Roubeau

DP~-G/PSRM - CEN Saclay B.P. ~ ' 2 - 91190 Gif-sur-Yvette, France

This paper describes two non-conventional means of refrigeration, perhaps out of reach of practical realization, whose purpose is to shed light on some properties of gases and adsorbants having great potential applications. An account is given of a variation of the Stirling machine and finally a sketch is made of a crossed cycles method independent of the refrigeration process used.

This descriptive part is sandwiched by overall considerations on heat losses, mainly by the cryostat necks and in general by any structures for which the non-heat conducting properties could, in fact, be an important criterion of success for cryocoolers. Concerning minicoolers which make use of a circulating coolant,the need for a dry and tight compressor is equally emphasized and some solutions are suggested.

Key words: Adsorption; cryocoolers; cryostats; helium; refrigeration; Stirling cycle.

In the following,the refrigeration process will be considered between essentially room and liquid helium temperatures. Transposition to another fluid and another temperature range would in general remain valid subject to proper numerical adjustments.

2. NO! to liquid helium

The straightforward means of cooling a system down to liquid helium temperature is, evidently ... to immerse it in liquid helium.

But.. . In the first place this is an expensive liquid. However, the purpose of the "cryocoolers",

topic of this meeting is not only to save money,as it is well known that the scale factor is dramatic for small installations. At the refrigerator itself the negative calorie will always

cheaper the bigger the plant.

The problem is

most of the users are not nearby big liquefiers,

liquid helium is costly to transport,

its conservation is expensive and, even for relatively short durations, it is practically impossible to conserve in small quantities,

finally only very competent and conscientious personnel are able to handle it without prohibitive spillage ; and when, by chance, such people are available it is more advantageous to assign them to more productive tasks.

Evfdently, other parameters come in balance, such as need of independence, weight, and volume considerations, available finance etc...

In every case the final conclusion has been :

use of liquid helium : excluded =+ cryocooler : mandatory.

3. Analysis of refrigeration needs

What is the objective of refrigeration ?

1) to keep a system at 4.2 K by compensating unavoidable losses QA due to various causes ; heat radiation, residual gas conduction in an insulating vacuum, conduction by supports.

2) to absorb a certain amount of heat Q produced in the 4.2 K region by the system during its normal operation.

B

In small installations, conduction by supports : usu?lly the neck of the cryostat calls for a detailed analysis. When, as is frequently the case,Q =O,conduction in the structure B of a cryocool,er, well insulated from thermal radiation, turns out to be a very important element of the thermal loss.

Let us consider a cryostat with a neck, the cross-section S(x) of which changes in such a way as to keep a constant thermal gradient between 300 and 4.2 K (practically this section would depart only slightly from the average-one. Hence the following can apply to a real cryostat). The heat flux down such a neck Q = k(T) .S(x).dT/dx decreases essentially as k(T) i.e. more or less linearly with T and reaches a value 70 times smaller at 4.2 K.khere it is identical to QA. ,A regularly decreasing heat flux means that some refrigeration has been provided all along the neck at each temperature level.

Without being exactly proportional, the input power requirements and the structural complexity of a crgocooler are strongly correlated to the mipimum work required by a Carnot cycle work between room temperature and a) 4.2 K concerning Q

b) temperatures distrguted between 300 and 4.2 K at which 69/70th of the heat flux entering through the neck at 300 K are a6sorbed.The work related to (iA is then WA = QA. (300-4.2)/4.2 = 70.4 4 while that related to the neck

300 A W, = ~~(114.2) J4.2 (,l/T) (300-T)dT = 234.5 iA = 3.33 WA.

Thus, the total required power for the refrigeration of an ideal cryostat reduced to a neck, is more than four times that which is necessary to aEsorb the residual hrat flow which reaches the 4.2 K level.

In practice cryostat necks display neither a perfectly constant thermal gradient, nor an exactly linear thermal conductivity.Thisratio of 4 is nevertheless a good order of magnitude afleast as a lower limit for real cryostats.

One can wonder why this considerable requirement of refrigerating power is sometimes unnoticed. There are two reasons : l).real cryosfats are not reduced to necks and the above ratios 70and4 have-to-be reduced by Q total / Q neck and ifsome extra power is dissipa-

A ted at 4.2 K by : (Q +q ),/Q~. 2) f inal$y and more importantly it happens that when one makes use of liquid helium! tge latter, in excess of its heat of vaporization : 80 jlmole: at 4.2 K, has a specific heat practically constant and equal to the perfectgas value : 21 J1mole.K between 4.2 and 300 K, and is thus capable of a total heat a6sorption of precisely 70 times the heat of vaporization of liquid helium at 4.2 K. The refrigeration power provided "free of charge" by the vapor between 4.2 and 300 K is the~efore sufficient to alJsorb nearly all the heat 70 QA entering through the neck except for Q which is precisely the source which generates these vapours. A

Practically, taking account of the unavoidable losses of other origin, the neck "costs" nothing in most of the cryostats which use liquid helium. In cryocoolers the neck will be resuscitated, then the choice of the refrigeration process should be made preferably among those offering several cooling levels and even, if ~ossible, a refrigerating power distributed

throughout the whole temperature span between 300 and 4.2 K. We will come back later onto the neck problem after having proposed a number of non-conventional refrigeration processes the practical realization of which does not seem easy in the actual state of the technique and in the author's imagination. These processes will perhaps be vitiated by trivial error or by a fundamental flaw but it is hoped that these exposition will help to stimulate thinking along certain channels which are the present time seem insufficiently explored.

4. C against C P v

The simplest and most immediate way of cooling which is actually the least feasible, technically speaking, consists of a counter-flow heat exchanger in which a gas cooled at constant volume is subsequently warmed at constant pressure : figure 1. The gain is Cp - Ck = R = 8.3 J1mole.K for a perfect gas. The exchanged heat between 300 and 4.2 K is :

at constant volume, under an initial pressure of e.g. 3 bars : 3600 J/mole at constant pressure under 3 bars : 6000 mole under a constant, lou pressure (e.g. p < 0.1 bar) : 5970 J/mole

FLgwre 1 : Re&Lgeha;toh "Cp vn. CVw. In (7) ;the gab coo& at com,iiznt volume, and in (2) AX Wdhmn at com,iizn;t yJhennune.

Thus one can see that in a system operating between these pressures a cooling effect will be observed as soon as the efficiency attains : (6000-5970)/(5970-3600) = 1.27 % of the theoretical maximum.

Practically any leak or gas friction transfers a part of the gas flux to a Joule-Thomson~ cycle which is thermally neutral for a perfect gas and only slightly detrimental for helium (above 50 K) .

Although this process has no immediate prospect one can anticipate some technological advances which would bring: it nearer to a uractical auulication :

u

1) availability of materials with a ratio E (elHstic limit) C (specific heat) sufficiently high, at room

temperature,

2) the discovery of rubber-like products preserving their elastic properties at low temperature,

3) the operation of a system in which the cooling takes place in a two dimensional space where one can realize constant "vo1umes" without walls (the surface of a sphere is an example!),

4) the introduction of the gas at room temperature under a very high pressure into a matrix from which it could escape only as a superfluid below 2.17 K,

5) if.. . (any suggestion to fill this 5th paragraph would be welcome). We will make some supplementary remarks on this process. We will first calcnlate 2.ts refriger-

5

ation efficiency in the case of a perfect gas. It we take the optimum case, in which the pressure at the end of the cooling is equal to the constant pressure at which the warming occurs, one has the elementary compression work at constant temperature T1(300 K) dW = -P dV = RT dP (for one mole). On the other hand, the cooling e fect dQ = (C -C )dT =

dp 1 T R dT = RT -;- from which the differential efficiency in the inrerval & (i.e. e) -)isV T2 Y .L -

n = - = - d9 ' . The average efficiency 8 = dW T1

R(T 1-T2) = ( I - /Log TI /T2 can be compared

W RTILogT /T 1 2 1

to the Carnot efficiency n = T#T~-T~). As an example for TI = 300 K, T = 4.2 K: n = 0.014; and on a purely theoretical ground the cycle "Cp vs. C," becomes more ahantageour than the simple stage Carnot refrigerating machine if both of the following assumptions are simultaneously realized :

1) T2 < T1/2.5 e.g. T2 < 120 K

2) one needs distributed cooling between T1 and T and the Carnot machine envisaged delivers cooling at only one temperature: T2

2

The "Cb VS. CVtt machine has two other peculiarities, when using a perfect gas : a) it proviaes no cooling at T2. The necessary cooling will have to be delivered by a process founded on a different cyc1e.b) it does not "know" if it should produce cold or heat; it needs to be started in a refrigeration mode otherwise some parasitic heat source will start it in a warming mode and it will probably finally transform work into heat with an efficiency 1, and this is certainly not a remarkable performance.

5. The gas cocktail adsorption refrigerator

In this system a mixture of gases with different boiling temperatures is introduced under pressure into a column containing a moving adsorbant ribbon with which it moves parallel while being progressively adsorbed onto the coldest point where the saturation of the adsorbant is completed by the most volatile gas. The adsorbing ribbon is then led into a returnchannel under low pressure where the gases desorb successively at temperatures lower than those at which they have been adsorbed. Thus one produces : 1) a transportation of the heat of adsorption(practical1y independent of the temperature) from the lower to the higher temperature while creating a large AT favourable to thermal exchanges. 2) a counter-flow cooling of the adsorbed gas of specific heat = R by the desorbed low pressure gas of specific heat = Cp = 512 R.

The net effect for a gas mixture is hardly calculable due to the number of parameters but it is probably very Important.

This cycle is akin to a multistage refrigerator using several refrigerating fluids

with distributed boiling temperatures in which:

1) all fluids are incorporated in the same gas flux and use one compressor.

2) two bridges have been established between 50 and 25 K and between 14 and 5 K where mother nature has not yet produced any body in the liquid state.

What should be underlined is that only a relatively low adsorbant speed is necessary in order to realize a noteworthy gas flux. As an example, a flow of 1 cm3 of activated charcoal per second displaces an amount of adsorbed gas of the order of 0.01 mole per second (at saturation).

The heat of adsorption is a function of the degree of saturation of the adsorber. It varies between a maximum (adsorbant degassed) and the heat of vaporization (adsorbant saturated), Table 1 gives approximate values of these heats relative to activated charcoal.[ll

The curves of figure 3 give, for instance, the cooling effect in joules/mole.K in the following application : adsorbant = activated charcoal, gas = He4, high pressure = 1 bar, low pressure = 1 millibar. The computation has been made using the formulas [1 1 : L = 8600 - 2920 log (V + 112) J/Mole and loglOP = [I53 log (V + 112) - 450].(1/T-.016) + 1.9 + logloV, with doin Torr, V in cm3RTp/gram of adsorbank: ,. T in K.

(RTP = room conditions : 300 K, 1 bar, 1 Mole = 24500 cm;') . 6

Heat of vaporiza- charcoal tion (sublimation)

J l g. of adsorbant. O K

1

Amount adnoabed a a ~ u n o t i u n ol: tem~erra;tuhe. P"= 7 but cmve D : P = 7 b m P = 7 W b m - E : P = 7 mi,&Xbm di&@unce beAween A and B - F : g loba l e d d ~ c t = Q-E

The curves A and B in figure 2 give the amount adsorbed as a function of T under pressures of 1 bar and 1 millibar. Curves D and E of fig. 3 give the heats of adsorption (desorption) per gram of adsorbant and degree K under the same_ pressures. CurveF,which is the difference between D and E, -gives the glpbal effect. Finally, curve H (fig. 4) gives the net effect corrected with the specific heat difference times the .difference -of the amounts adsorbed : ((A-B) (Av/24500) (5/2R-R)) . Besides this cooling effect which extends from 4.2 to 24 K one has the available heat of desorption at 4.2 K of 660-520 = 140 ~ m 3 / ~ . of adsorbant, which will be calculated from the integral curve Q(V) given in figure 5 : Q(660)-~(520) = 27.5 - 25.8 = 1.7 J/g. of adsorbant which means 63 J/mole circulated.

Thus, the desorbing helium provides cooling essentially in the first zone 5K-14K where no liquid is available. Hydrogen will provide the bridge between 25 and 50 K and, above the latter temperature a number of gases exist for going up to room temperature.

J lg. of adsorbant. O K J / g , o f adsorbant 3 0

'C

20.-

2 0 0 4 0 0 6 0 0 v

6. The Bi-Stirling refrigerator

The machines which operate following a Stirling cycle, the Philips machine for instance, suffer contradictory requirements,

1) the pistons along which there is a thermal gradient have to move rapidly and offer an important clearance for minimizing the thermal exchanges between parts which periodically face each other' when at different temperatures,

2) regenerators and heat exchangers must present temperature differences and pressure drops as small as possible and this demands a small gas speed,

3) at very low temperatures the specific heat of metals used in regenerators : copper, l e a d . . . becomes very small and lead to important volumes hence to important dead spaces.

In conclusion, the Stirling machine which is capable of reaching its full efficiency if compressions, expansions and exchanges areisohrmal, even with arbitrary dead spaces (reversible cycles), suffers an efficiency decrease when the dead space increase coincides with the thermal gradients in the working space of the pistons.

The proposed variant consists of LJ) replacing the pistons with bellows operated slowly which a) suppresses the shuttle effect and, b) allows isothermal compressions, expansions and thermal exchanges. The dead space of the bellows reduces to a certain point the specific power of refrigeration but has no effect on the thermodynamic efficiency, 2) replacing part or all of the regeneratorby a heat exchangerbetween two machines which are 180' out of phase. One compressor with a manifold of valves assumes compressions and expansions. A separate system acts on the bellows to operate the gas transfers. Such a system is represented schematically in fig. 6.

Figure 6. Schematic representation of the bi-Stirling refrigerator, in which the compressor- expander (1) and the valves (2, 3, 4, and 5) are operated in synchronlsrn but 90' out of phase wrth the bellows (6, 7, 8, and 9) by the mechanical drive (LO). The fluid undergoes a Stirling refrigeration cycle alternately in the left-hand set of bellows and in the right- hand set, with displacement through the counterflow heat exchanger (11). Heat is rejected at the warm (ambient) end through the heat-exchanger (121, and heat is absorbed at the cold end from the heat-exchanger (13). The outer shell (141 is the vacuum case.

7. The crossed cycles refrigerator

This proposed process combines a) an alternating flow of a caloric fluid in a fixed column of ideally zero specific heat, with b) a reversible local cycle external to each level of the column and operating at the same frequency with appropriate dephasing.

At each point of the column the external cycle induces a fluctuation of the temperature larger than the one of opposite sign, which would be due to the displacemnt of the fluid alone.

A molecule of this fluid follows the cycle represented in figure 7. The line AB would depict a cycle with zero dephasi~g. The elliptic curve suggests a real cycle obtained with a given dephasing. The arc NPR is related to the heating part of the external cycle, and the arc RQN to the absorption of heat by the latter.

The heat received by the fluid between xp and x, and given back to it between K, and XQ is represented in figure 8.

The overall result is an outgoing fluid at the warmer end at a temperature above room temperature and an outgoing fluid at the colder endat a temperature below T2.

A practical realization wouldnecessarily include a heat capacity effect related

9

Figme ti : h u t . exchanged i n one c y c l e .

to the walls or to the\,local cycles, which will make this cycle more or less a relative of the Stirling cycle.

Figure 9a shows a conical piston compressor which, subject to an alternating displacement of small amplitude, cycles the pressure of a gas locally without displacing it. A separate system forces the fluid to oscillate from the warm end to the cold end and conversely.

Figure 9b shows a derived variant in which the displacement system has been omitted and replaced in its dephasing part by the pressure drop between cylindrical sliding parts. A net flow will occur if one chooses a displaced cold volume larger than the dead space at the volume warm end.

8. The necks of the cryostats

Cryocooler necks require special attention. In order to shed light on this problem let us consider, at first, the losses in a cryostat neck bathed by the vapours of an evapora- ted cryogenic liquid. These losses are represented in fig. 10 : 4 represent the supplementary losses related to the neck-as a function of the.losses from orher sources (normalization is made at unity for the-vatue QN extrapolated for QN if q R * 0 ) . In most cases the thinnest ?ec! would give a ratio Q /Q >>O1, but one can see that Q is practically zero for M $19.~ > I. 15 and therefoft '0 onemobtains a better mechan~cal sturdiness by increasing the neck thickness within the ratio Q /Q > 1.5 and . . . one forgets the neck problem.

R No

In the case of cryocoolers one does not have this distributed source of refrigeration constituted by the vapours heat content. This would result in a heat loss due to the neck an order of magnitude higher which, in most cases, will overide the other heat loads. If one intends for example to dissipate at 4.2 K all the heat conducted along the neck from room temperature one will face a disastrous efficiency. Practically all proposed processes shall include an intermediate refrigeration level and it is more and more evident that those having the best performance will be 3 staged.

Let us take, for instance, the schematic cryostat of fig.11 which has 3 levels of refrigeration at temperatures arbitrarily fixed in geometrical progression i.e. below room

11

a : rn end dt 300 K b : l n t n&ge dt 72 K c : 2nd n&ge LE.t 17.4 K d : 3hd 4Xage dt 4.2 K

temperature (300 K) : 72 K, 17.4 K and 4.2 K. The characterlstics o f this hzgh performance cryostat and its calculated losses are given in table 2.

Characteristics: useful volume : 2 liter

surface exposed to

heat radiation : 1000 cm 2

& I + ? emissivity : - = 0.01

1 = 3 x 10 = 30 cm

neck e = 0.01 cm

0 = 1.5 cm

Table 2 : T y p i c d o r t y o n m charrac;tehin;ticb.

Zone

300-72 K

72-17.4 K

17.4-4.2 K

The latter calls for two remarks : 1) ehelosses due to the neck between the lower stages are 10 times more important than the radiant heat between 72K and 17.4 K and 100 times more important between 17.4 K and 4.2 K. This emphasizes the significance of progress in neck technology and generally inllsupports-technology", that will be demanded by progress in other

12

Q Rad.

500 mw

1.7 mW

6. pW

Q Neck

130 mW

13 mW

800 1.1w

W. Carnot

2. watt

.040 watt

.003 watt

related fields. 2) Every installation which will make use of liquid nitrogen for cooling of the first stage is, even with due regard to the scale factor, more than ten times smaller, ten times lightersand finally ten times cheaper than a system starting from room temperature.

9. The need of contamination-free compressors

In every refrigerator which makes use of circulated helium, gas purification after compression, to free,it from oil, water and other contaminants such as air or other gases is-a burdensome task which often leads to the need of pei.iodically stopping the system for clean-up. The use of dry and tight compressors would be a major improvement. But, even with many potential applications outside the field of cryogenics, it is surprising that a small contamination-free compressor, is not comercially available.

In fact we have used a two stage rubber diaphragm compressor with slightiy modified heads for a year of uninterrupted service which worked between 0.2 and 3.5 bar abs, Only a very small contamination has been observed, probably due to diffusion of air through the diaphragm and' the rilsan tubing which leads to monthly warmup and purging of impurities.

Two types of small compressors which should give satisfying results are represented in figures12 and 13 . In the first one a very thin metallic bellows compressed by a liquid properly degassed, compre&ed itself by another bellows by means of a mechanical device, offers the advantages of 1) permitting an isothermal compression; 2) being insensitive to thermal expansion of the liquid and hence requires neither pump nor value forthe liquid body and 3) capable of tolerating important depressions on the low pressure side.

The second system is a thermocompressor : a gas volume is warmed by a short ohmic heating of a wire placed in this volume, the resulting increase of pressure opens the exhaust valve and pushes out a part of the gas. The subsequent cooling induces a depression which opens the inlet valve and restores the initial conditions. This system has the advantage of extreme simplicity. It is silent and is readily apt, by adding stages, to produce arbitrarily high compression ratios. It can even be used at very low pressure (= 1 Torr), subject to the use of actuated valves.

13

References

[I] P. Roubeau, G. deNigohossian, and 0. Avenel, Adsorption de L'helium 4 par le charbon actif - Colloque international "Vide et Technicens", Grenoble, 1969. Supplement to "Le Vide" No. 143.

IBM 'l'homns J. Watson Research Center Yoridown Heights, New Yok 10598

High w e d , low power Josephson junction logic and memory chips densely packed into a structure immersed in liquid helium may constitute the high performance computer of the future. In this paper + discuss chip cooling, input/output cable requirements, computer cryo*tem design considerations, and refrigeration requirements.

Conventional gas expansion refrigeration should be suitable for medium to large &e superconducthg computers (0.5 to 10 watts). Small systems, such as a 25 mW future minicomputer,. will require innovative refrigerator design that achieves high reliability at low cost.

Key words: Cryogenic equipment; 4K refrigeration; Josephson junction devices; superconducting computers; thermal contact.

1. Perspective

Research is in progress on a wholly new integrated circuit technology that may eventually replace semicondw tors in high p e r f o m c e digital equipment [I]. The basic elements are Superconducting Quantum Interference Devices (SQUID) that can be switched from a non-resistive to a resistive state in roughly 10 ps by the magnetic field from a nearby control line. From these elements, one can build a very high-speed logic family and memory to match [2].

High-speed switches (or gates) alone, however, do not guarantee a high-speed computer. Intercircuit signals, typically travelling at half the speed of light, take 10 ps to go only 1.5mm. In order to avoid excessive signal propagation times, then, one must pack the circuits (and chips) closely while managing to get power in and the heat out. For example, in making an ultra-high performance computer one would want to package the logic and memory in a cube 9 cm on a side. Such a machine made with present semiconductor technology might dissipate as much as 20kW 121. The delivery of this amount of power and its removal as heat in such a small package would be impossible with today's technology. A comparable machine made with superconducting circuits might dissipate roughly 1OW 121 which would be easy to remove, even at 4.2K. The power distribution [3] would be via fine superconducting lhes. I t is the dombination of the high speed g& the low power of superconducting circuits that may put them in future computers.

An obvious consideration in judging the practicability of a superconducting computer is the refrigerator, and so there is interest in the state of this technology and in what improvement in reliability and reductions in cost could come in the next ten years. We present in this paper, as best we can predict, what a superconducting computer's cryosystem requirements might be. Let us take a more detailed look at these by starting at the heat source.

2. Heat transport'. circuit to sink

Most of the heat generated by the circuils travels through several thin film layers to the chip substrate (a 6.4 mm square of silicon) which is isothermal because of its high thermal conductivity (- 2 W/cmK). Most of this heat

then passes out the non-circuit side of the chip to a refrigerated heat sink such as a liquid helium bath Heat transport in the bath is by natural convection, and heat transfer at the chip surface is aided by the stirring action of local boiling.

Figure 1 shows the expected chip temperam rise as a function of the chip power [4], and also shows anticipated power levels for logic, fast memory (cache), and main memory chips. These are essentially the same whether or not computations are taking place. The somewhat poorer heat transfer obtained as the chip power is brought up initially probably is due to incompletely developed nucleate boiling. Note that the substrate temperature for the three kinds of chips should run from 7 mK to 240 mK higher than the bath temperature. Note also that the highest chip power is only about 4% of the nucleate boiling limit, and if this power were increased tenfold then the chip temperature would be less than 400 mK above the bath.

In a small computer, one might choose to dispense with the bath and make a greased thermal contact to the chips. At 4.2% the grease to silicon boundary resistance is negligible compared to the resistance of practical thicknesses of grease. The heat transfer characteristic for a 2pm thick, 6.4 mm square of silicone grease [5] is shown in Fig. 1.

For the in te r fe romete r s currently in use at IBM, the maximum desirable chip substrate temperature is 4.5K, and one can expect successful operation of a machine. whose chips vary in temperature from 4.0 to 4.5K. Then, for example, one could operate with a total spacial and temporal temperature variation from 4.OK to 4.2K, and so the variously loaded chips could run up to 0.3K warmer.

CHlP TEMPERATURE RISE (mK)

I I 1 1 1 1 1 I I I 1 1 1 1 1 1 I I I I I I I L

F igu re 1 : Chip Cool i ng C h a r a c t e r i s t i c s

100

10

END OF NUCLEATE-,e - T= 4.2 K BOILING = 0 9 6' - -

CHlP GREASED-/ 0 8 - - I TO SILICON,

- - (EST.) 0 - --LOGIC 0

0 - - 0

AJ

- :FAST ,-' ,o$/

MEMORY - -

- - - LIQUID HELIUM - - -

1 I 1 1 l l l l l I I I I I I I l

10 100 1000

3. Medium to large computer cryosystem

The logic and memory chips of a medium to large computer might be mounted on cards that plug into a box-like structure as shown in Fig. 2. This would be immersed in liquid helium, and the chips would be cooled by natural wnvection of the liquid up the vertical channels. A large wmputer might require a box structure 9 cm on a side and would dissipate approximately 10W of heat into the bath. A medium capacity wmputer might dissipate 0.5W to 2w.

Communication to room temperature support equipment would be via a multiple stripline input/ouput cable as shown in Figs. 2 and 3. Each copper line and associated groundplane would have a cross-sectional area of roughly 6.1 x c d , and an ultra-high performance computer might have 400 lines. If these were thermally grounded at 75K and 18K as shown in Fig. 3, with a 20 cm length between stages, then the heat leak to these grounds would be 1.2W and LOW, respectively, and the heat leak to the 4.2K bath would be 0.3W which is small compared to the computer's 10W. Perhaps 20 of these lines would supply the power and so would have noticeable resistive heating that wuld bring the total 1/0 heat loads for the large system to 4.7W, 2.OW and 0.8W at 75K, 18K and 4.2K respectively. One sees that the I/O lines will have little effect on the 4.2K refrigeration requirement.

helium Liquid t flow I 1 I/O cable to

environment

Figure 2: A Superconducting Computer Package Unit

i\ Interboard flexible

, connection

COMPRESSOR

REFRIGERATOR I --------

COMPUTER 260 P S I A DISPLACER T

. -

LtQUlD IELlUM BATH

Figure 3: A Cryogenic System for Medium (0.5W) to Large (10W) Superconducting Computers. JT = Joule-Thompson expansion valve, ADS = adsorber, TD = thermal diode.

There are three principal functions that we wish to achieve in a medium to large sized computer cryosystem:

(1) Uninterrupted computer operation during refrigerator senicing, (2) Uninterrupted refrigeration operation during computer removal, and (3) A semi-annual or longer service interval

An integral refrigerator/dewar cryosystem that should meet these requirements is shown in Fig. 3. If the refrigerator should fail or be shut down, then the thermal couplings (TDI, TD2, and condenser) isolate it from the bath which then provides temporary refrigeration by gradually boiling away. When the refrigerator is returned to operation, helium gas is liquefied to replenish the bath. The wmputer can be removed by pulling it up into a special room temperature s e ~ c e housing (not shown) that serves as an air lock. Heat from the 1/0 cable and from radiation and conduction in the dewar neck is intercepted by 75K and 18K heat stations that are part of the computer support structure. Each of these delivers its heat across a narrow annular gap to a refrigerated ring in the dewar neck wall. The compressor, refrigerator, and thermaI couplings as shown schematically in the figure have been designed and built for IBM by the Advanced Products Division of Air Products and Chemicals. Further discussion of the design and performance is given by R. Longsworth and E. Flint g & elsewhere in these pmedings.

An important feature of the refrigerator/dewar interface scheme is the simplicity of the connection between the refrigerator and the couplings. This consists of six small lines that carry the JT gas at various temperatures, and these can be made flexible (for vibraton isolation) or longer (for remote refrgerator placement). For example, in order to simplify dewar design or to make the system fit in tight quarters, one might place the refrigerator in a separate enclosure joined to the dewar by an evacuated link that carries the interconnecting tubing.

Good magnetic shielding is requried .for the logic and memory devices. This can be provided by ferromagnetic shields and/or a superconducting shield as shown in Fig. 3. Materials having low remanent magnetization such as silicon, glass, polymide, aluminum, selected beryllium and copper, and molybdenum must be used for structures within the shield.

4. Cryogenic challenges

Refrigerator, interface, and dewar technology suitable for commercial medium to large superconducting computers is available, and we look for continued improvement in reliability, efficiency, and cost. Obtaining high reliability is of greatest importance. Installations in aircraft or satellites will pose obvious additional problems that challenge present day cryogenic technology.

The commercial viability of small superconducting computers (power < O.SW) depends on both the reliability and the cost of the cryogenic system. In the extreme case of a future minicomputer, one would like to have a 25 to 50 mW 4K refrigerator for a few thousand dollars. This calls for real engineering innovation. One could think of simply using a large dewar, but a reasonably sized system is likely to need monthly fillings and the logistics of this for a machine destined for a diverse market look unattractive.

A schematic drawing of a hypothetical future minicomputer is given in Fig. 4. There would be no liquid helium bath and the chips would be cooled by greased contact to a silicon heat sink. The refrigerator would be capable of rapid cooldown (- 10 min with the computer off). I t might be a gas cycle machine depending on an external compressor, or it might be an innovative self-contained system where only exhaust heat and electrical or mechanical power pass through the interface'. While magnetic, para-electric, and other cooling techniques are possible, gas expansion may be the most practical. An isothermal expansion of 13 mg of helium gas at 4K from 1 cm3 to 2 cm3 in one second can, in principle, produce 80 mW of refrigeration. The challenge is in delivering the expansion work out of the system and in recycling the gas.

5. Summary

Superconducting computers requiring closed cycle 4K cryosystems of up to 10W capacity may be on the market in the future. The medium to large size machines will be cooled by convection in a refrigerated liquid helium bath. Current refrigerator, interface, and dewar technologies are adequate for this application, although improved reliability, efficiency, and cost would be desirable. The future commercial viability of small superconducting computers depends on the development of new low capacity, low cost, but highly reliable 4K refrigerators.

TER I / O

-MAGNETIC SHIELDS

- I / O CABLE

-CABLE CONNECTOR

-THERMAL BASE FOR CHIPS (SILICON)

25 TO 5 0 mW OF REFRIGERATION NEEDED AT 3rd STAGE

-VACUUM 8 RADIATION SHIELDS

Figure 4: A Hypothetical Superconducting "Future ~ in icompute r" . TG = thermal ground

6. References

[I] Matisoo, J., The Superconducting Computer, Sci. Am. 242, [S], 50-65 (May 1980).

121 Anacker, W. g aI., Josephson Computer Technology, IBM J. Res. Dev., 24,[2], 107-264 (March 1980).

[3] Amen, P. C and Herrell, D. J., Power Design for Gigabit Josephson Logic Systems, IEEE Trans. Microwave Theory Tech., MTT-28, [5], 500-508 (May 1980).

4 Flint, E. B., Jenkins, L. C, and Guernsey, R. W., to be published

[5] Kreitman, M. M. and Callahan, J. T., Low Temperature Thermal Conductivity of Several Greases, Rev. SC~. Instrum., 40, 1562-1565 (1969).

DEVELOPMENT APPROACHES FOR LONG-LIFE CRYO-COOLERS

Ronald White Will iam Haskin

F l i g h t Dynamics Laboratory Air Force Wright Aeronaut ica l Laborator ies

Wright-Patterson A i r Force Base, Ohio 45433

Development of cryogenic r e f r i g e r a t o r s (cryo-coolers) capable of ope ra t ing f o r over t h r e e yea r s wi thout maintenance i s being sponsored by t h e F l i g h t Dynamics Laboratory f o r s e v e r a l Air Force missions. Future cryo-coolers f o r superconducting devices w i l l need s i m i l a r h igh r e l i a b i l i t y and d u r a b i l i t y . Therefore , t h e gene ra l approaches used and t h e l e s sons s t i l l being learned w i l l be descr ibed a s they apply t o a l l l ong- l i f e cryo-coolers.

Highly r e l i a b l e and durable cryo-coolers a r e developed i n a c a r e f u l l y planned and con t ro l l ed process . Required ope ra t ing condi t ions w i l l i n f luence t h e s e l e c t i o n of t h e b a s i c type of coo le r , bu t t h e s e l e c t e d concept must have an inhe ren t long- l i f e p o t e n t i a l . The p r i n c i p a l long- l i f e cryo-coolers being developed by t h e F l i g h t Dynamics Laboratory a r e t h e Vuilleumier (VM), turbo-Brayton and r o t a r y r ec ip roca t ing r e f r i g e r a t o r s . These machines e i t h e r use gas f i l m bear ings t o avoid rubbing con tac t o r have l i g h t l y loaded bear ings and s e a l s wi th low wear r a t e s .

Careful development i s necessary t o r e a l i z e t h e long- l i f e p o t e n t i a l of t h e coo le r s . This inc ludes d e t a i l e d ana lys i s of s t r e s s l e v e l s f o r a l l p a r t s of t h e c o o l e r , adequate q u a l i t y c o n t r o l s , contamination c o n t r o l , and r e l i a b i l i t y a n a l y s i s . F i n a l v e r i f i c a t i o n of adequate d u r a b i l i t y , w i th acceptable cooler performance, i s demonstrated by r e l i a b i l i t y growth i n a t e s t , analyze,and c o r r e c t program.

Key words: Component development; contamination c o n t r o l ; cryogenics; helium r e t e n t i o n ; long- l i f e ; r e f r i g e r a t o r s ; r e l i a b i l i t y ; ro tary- r e c i p r o c a t i n g ; turbo-Br~.yton; Vuil leumier.

1. In t roduc t ion

Cryogenic r e f r i g e r a t o r s (cryo-coolers) capable of ope ra t ing f o r over t h r e e years with- ou t maintenance a r e being developed by t h e F l i g h t Dynamics Laboratory f o r s e v e r a l A i r Force missions. Other a p p l i c a t i o n s may permit occass ional maintenance, bu t good r e l i a b i l i t y and dependabi l i ty w i l l always be e s s e n t i a l . This i s e s p e c i a l l y t r u e f o r f u t u r e superconducting devices . The exper ience gained and the l e s sons s t i l l being learned i n t h e development of long- l i f e cryo-coolers w i l l b e n e f i t f u t u r e programs.

Active development of s e v e r a l types of long- l i f e cryo-coolers i s i n progress . D i f fe ren t des ign approaches a r e being pursued u n t i l s u f f i c i e n t success i s obtained t o s a t i s f y var ious requirements. A i r Force e f f o r t s inc lude Vuilleumier (VM) c y c l e , turbo-Brayton and ro ta ry - r e c i p r o c a t i n g ( ~ 3 ) Brayton cycle r e f r i g e r a t o r s . These coo le r s produce temperatures a s low as lOoK wi th cool ing c a p a c i t i e s ranging from 0.1W a t 10°K t o over 10W a t temperatures between 700 and 100°K. The turbo-Brayton and ro ta ry - rec ip roca t ing r e f r i g e r a t o r s can be designed t o produce temperatures down t o 40K o r lower, and component development i s planned f o r improved r e f r i g e r a t o r s t o provide cool ing a t 4OK. The cryo-coolers p r e s e n t l y being developed w i l l be used t o desc r ibe t h e approaches f o r ob ta in ing long- l i f e .

The VM cooler achieves internal pressure changes by gas displacement to different temperature volumes rather than by mechanical compression. The total gas volume in the cooler remains constant. Forces on bearings and seals are therefore sufficiently low to allow long operating times. Heated cylinders with domes maintained at 9600K are used for thermal compression, and gas expansion in the cold cylinder provides refrigeration. A small electric motor is used to overcome friction and control the motion of the displacers in the cooler. Thermal regenerators are used to exchange heat between incoming and exiting gas streams, to obtain adequate efficiency, for this reversing flow refrigeration cycle in much the same way as counterflow heat exchangers do in continuous flow refrigeration cycles. Figure 1 shows a schematic of a VM cooler which has two hot cylinders and three stages of expansion. This cooler [l] is being developed at the Hughes Aircraft .Company for long duration Air Force missions.

The turbo-Brayton cooler uses small, high speed turbine wheels supported on gas film bearings to achieve compression and expansion of the working gas in the machine. Since there are no rubbing parts, the cooler should be capable of long term operation. A two- stage development model turbo-Brayton cooler [2] is being assembled at the AiResearch Manufacturing Company of California. Figure 2 shows as.chematic for the arrangements of this cooler.

The concept of combining the efffciency of recpirocating refrigerators with the long- life capability of components suspended on gas film bearings provided the basic approach for rotary-reciprocating refrigerators (R~). This type of machine is characterized by pistons rotating on their central axes at 1200-1400 RPM as they reciprocate at a cycle rate twice their rotational speed. This motion provides porting action to control the flow of the working gas which is compressed and expanded in different sections of the machine to produce refrigeration. The pistons are driven by electromagnetic actuators. Counterflow heat exchangers are used to implement the reverse Brayton cycle. A two-stage unit of this type has been tested and components are being developed by Arthur D. Little, Inc. [3] for a three-stage unit. Figure 3 shows the basic arrangement for a R~ system.

The Stirling cycle is often used for comparison because it is potentially the most efficient. It is a refrigerative cycle with a cold section very similar to the VM cooler. However, the gas is compressed by a mechanical piston instead of heated cylinders. The forces on the bearings and piston seals are therefore significantly higher than the VM which has resulted in wear and shorter life. Recently NASA contracted with Philips Labora- tories[h] to design and build a long life linearly driven Stirling cooler utilizing non-contacting magnetic bearings and clearance seals. This cooler will be described in another paper at this conference.

As indicated above, there are reasons to expect each type of cooler to be capable of operating for long periods of time. However, careful development is necessary to determine the true long-life potential. Methods used in these developments are described in the next section.

2. Development approach

In planning the development of a long-life cryo-cooler, the first essential step is to choose a type of machine which has an inherent capability to operate for long periods of time under expected environmental conditions. Although this is obvious, the applicable technology may need to be examined in some detail before making a decision. For example, an oil lubricated reciprocating compressor is generally accepted as having better durability and reliability than a dry lubricated reciprocating compressor. However, use of the oil lubricated compressor depends on the ability to separate the oil from the working gas in the refrigerator. Periodic replacement of the oil separation filters will be necessary. Thus, the oil lubricated compressor is not acceptable for use in a location where equipment cannot be.maintained. The maintenance problem might be circumvented by use of devices such as rolling diaphram seals, but these add further complications which must be considered. These include a full-time oil support for the diaphram seals, a system to dispose of the helium that permeates into the oil through the diaphram seals, a helium makeup and control system, and devices to compensate for oil displacement and expansion.

HERMAL SHIELD COOLlNG

FOCAL PLANE COOLING

FIGURE 1. VM COOLER

FIGURE 2. TURBO-REFRIGERATOR SYSTEM SCHEMATIC

ARMATURE OF ROTARY MOTOR STATOR OF ROTARY MOTOR CLEARANCE SEAL ARMATURE OF LINEAR ACTUATOR

COOLANT CIRCUIT STATOR OF LINEAR ACTUATOR GAS SPRING CAVITY

FIGURE 3. ROTARY-RECIPROCATING REFRIGERATOR COMPRESSOR

Other factors which influence the choice of a basic concept for a long-life cryo-cooler include power consumption and weight restraints which may be imposed. Turbine machinery with gas film bearings should operate for very long periods of time, but the efficiency is low unless a high cooling capacity is needed. The rotary-reciprocating refrigerator has a potential for very long-life and high efficiency but with higher weight and more complicated control circuits than other coolers. The VM cooler does not have as great a life potential as the gas bearing coolers, but it will have significant life because of the light loads. Its efficiency will not be as high as the R ~ , but its weight is considerably less, the controls are simple, and the design more mature. Much better efficiency should be obtain- able with Stirling cycle machines if novel means can be implemented to achieve long-life. The relative importance of different factors must be evaluated in making a decision.

Cryogenic refrigerators being developed by the Flight Dynamics Laboratory for long term continuous operation generally use gas film bearings to avoid rubbing contact or use lightly loaded bearings and seals with low wear rates. The turbo-Brayton and rotary-reciprocating refrigerators with gas film bearings may be subject to long term gas erosion and contamination effects, but their good durability potential is generally recognized. Durability of VM coolers may be somewhat questionable, but thermal compression and the constant volume aspect of the VM cycle provide significant benefits. Bearing loads and operating speeds can be an order of magnitude lower than for conventional Stirling coolers with equivalent capacity. The low forces permit long-life to be achieved with dry lubricated bearings. A low pressure drop across the sliding seals allows innovative, long lasting seal designs. Careful material selection along with thorough component specifications to assure adequate quality are leading to significant progress in extending the demonstrated life of VM coolers.

Other concepts for long-life coolers are being pursued by other organizations in relation to their particular requirements. NBS has demonstrated a low speed, low capacity Stirling cycle cooler 15) that is well suited to laboratory or similar use. NASA is sponsoring the development of a Stirling cycle cooler [ 4 ] which employs magnetic bearings -to avoid rubbing contact. The combined efforts of all organizations should provide the technology to increase the potential for long life coolers in the next five years.

2.1 Component development

After the basic cryo-cooler concept is chosen,the long-life potential can be realized only by careful analysis and controlled fabrication of each part or component. Sufficient analyses must be performed to determine the mechanical and thermal stress levels for each part. The fatigue or wear limits for each part must then provide a margin of safety when compared to the durability requirements of the cooler. Those parts which may cause failures are identified for special emphasis in the development phase.

During the development of VM coolers, many early problems,such as heater life and mechanical interference forces,were eliminated by bench testing and analys is . A special material had to be found for the rider rings which guide the hot end of the displacers in the hot cylinders. Experience with sliding seals in the coolers indicated that simply finding a better material would not provide adequate life. Therefore, new designs are being evaluated along with new materials and better control of the surfaces on which the seals rub. These are examples of the detailed effort which is now showing a payoff for the total cooler.

In the case of the rotary-reciprocating refrigerator, technical issues include the design and fabrication of gas film bearings, the complexities of electrical drive circuits, and the fabrication of high effectiveness heat exchangers. Contamination control is also very important in relation to materials selection, assembly methods,and gas management. These problems have been identified, and solutions to them are being verified as development continues. Since the electrical control circuits for the rotary-reciprocating refrigerator are rather complicated, special emphasis has been placed on the basic design of these circuits to provide highly derated part operation with well established functions. Electrical problems with other coolers have shown that reliability and durability problems are not limited to mechanical devices. Technical issues for the turbo refrigerator include design and fabrication of compliant gas bearings, design and fabrication of high effectiveness heat exchangers, development of a high speed motor and controller, and the design and

26

fabrication of small turbine wheels. Quality control and contamination have been given a high level of attention. Included also, is a continuing effort to increase the efficiency of individual components such as heat exchanger effectiveness, compressor adiabatic efficiency, turbine adiabatic efficiency, and reduction of seal leakage and bearing losses.

2.2 Quality assurance

An adequate quality assurance program must be implemented during the development of a cryo-cooler. This program provides for verified identification of materials and complete inspection of all purchased and fabricated parts. This information is necessary to assure a reasonable probability of successful cooler operation and to help analyze the cause of difficulties which do occur. Part failures have been caused by materials which did not actually contain a specified constituent. In a similar manner, machined parts must be inspected for proper dimensions and quality of workmanship. Component assembly procedures must be reviewed for adequate process control, and the components should be inspected to assure that specified steps, such as heat treatments, were actually accomplished. Accept- ance of this extra quality control burden is essential to the final success of a long-life cryo-cooler.

2.3 Contamination control

One of the most significant problems in developing long-life coolers is the elimination and control of contamination. The problem is usually broken into two categories: particulate contamination and gaseous contamination. Particulate contamination can result from inadequate cleaning, manufacturing defects, flaking of coatings and wear debris from bearings, riders, and seals. The effect of particulate contaminates varies from the scratching of highly finished close tolerance bearing and seal surfaces to the partial or total plugging of heat exchangers, regenerators, and other small gas passages. Scratching of surfaces can result in accelerated seal wear, bearing failures, additional particle generation, and the jamming of the refrigerator mechanism.

Advanced design concepts, cleanliness standards,and particle traps are utilized to eliminate or reduce particulate contamination. Design techniques include the elimination of rubbing parts by the utilization of gas bearings and non-contacting seals. Examples of this technique are the turbo-Brayton and rotary-reciprocating refrigerators.

In coolers that do have rubbing parts, thequantityof particulate contamination can be reduced by reducing the load on the parts and by dry lubrication. An example of this is the Vuilleumier refrigerator which utilizes heat instead of mechanical energy as the primary input power. Inside the refrigerator,the pressure drops across the thermal regenerators are only a few pounds per square inch. This results in low loads on the dynamic (sliding) seals, displacer riders, and bearings which, coupled with their slow speed, results in low wear, low contaminate production,and long-life.

Other techniques to reduce particulate contamination are to keep trapped dirt sources, such as motor windings,out of the working fluid (helium) or to hermetically seal the contamination source. Successful techniques have been to place the pressure wall between the solid metal motor rotor (inside the cooler) and the motor stator with its windings outside the cooler or to "can" the motor stator.

Cleaning prior to and during assembly is required for all cryogenic coolers but is especially critical for the extremely close clearance gas bearing refrigerators. Clearances in these bearings are typically measured in millionths of an inch.

Once assembled, the continuous flow coolers utilize filters to trap particulate contamination missed by cleaning or generatedwithinthe cooler. Reversing flow refrigerators such as the VM and Stirling utilize thermal regenerators which, because of the reversing flow, tend to be somewhat self-cleaning. Particulate traps can be created by adding small, deep groves perpendicular to the working fluid flow that allow contaminate or wear partic- ulates to drop in and become trapped. The location and number of these type of particle traps must be carefully considered since they decrease refrigerator performance by increasing refrigerator dead volume and pressure drop.

Condensation or freezing of gaseous contamination is an even more severe problem than particulate contamination in cryogenic coolers since, at the very low temperature that some of these coolers produce, everything except helium is a liquid or a solid. Gaseous contamination can clop; Joule-Thomson valves, clog heat exchangers, partially clog thermal regen- erirtors, create thermal shorts from cryogenic regions to warmer regions of the refrigerator, and possibly even stall the refrigerator mechanism. Sources of gaseous contamination include organic materials such as seals, riders, bonding agents, plastic parts, motor wire inaulation,and the gaseous contaminates adhering to the surfaces of the metal parts during assembly, Gaseous contamination has also been obtained from untested dirty helium fill bottles. Due to the difference in partial pressures of the contaminates, between the pure hslium inside the refrigerator and the atmosphere outside, refrigerators with large static seal areas may experience additi~nal long-term gaseous contamination problems caused by permeation of contaminates past the seals. Design techniques to reduce gaseous contamination include: elimination oforganicswithin the refrigerator, hermetically sealing the motor windings or placing the windings outside the working fluid, and the use of welded joints whenever possible, Operational techniques include testing , certifying, or cold trapping the helium fill gas, careful purging of the cooler after assembly and'after break-in, and bakeout of the cooler parts and the assembled cooler, which is limited by the high temperature tolerance of the seals and epoxy or solder bonding agents.

2.5 Helium retention

All of the refrigerators 4f interest utilize gaseous helium as the working fluid. Since helium molecules are small single atomg, it is one of the most difficult gases to contain for long periods of time. Techniques to maintain adequate working fluid within the refrigerator are helium makeup and helium retention, Heli.um makeup has been confined to large re$rigerators and utililies a high pressure bottle, pressure reducing regulator, valves, pXeS6U$W sensors, and control logic. This adds parts to the refrigerator which in turn can affect system reliability. He1it.w retention has been accomplished by various means. For short life coolers, single or double "Q" rings have been used. Since rubber "0" rings are permeable to helium, longer life coolers may have. an "0'' ring to withstand the internal refrigerator presgure fluctuations backed by a soft metal cut ring seal or indium coated metal "K" seal for positive sealing. These seals require careful preparation of the sealing surfaces and ireplacement of the seal after the jdint is opened. The best way to seal joints 5s by welding, This can be considered after the refrigerator development phase has bean completed and there is ng longer a need to frequently disassemble the refrigerator. Clever weld flaqge designs allow the weld to'be ground off and opened and then rewelded one or more times without destroying the expensive parts.

The detection and location of very small leaks is always a challenge especially on refrigerators of complex geometry. Leak isolation is currently being investigated under the QM effort.

2.6 Reliability

~eliabilit~ can be enhanced if the refrigerator designer pays careful a'ttention to the environmenfs and conditions that the cooler will experience during developmental testing and its ultimata use, Design simplicity pith attendant low parts count enhances reliability. Since long-life coolers are not high production rate items, 100% inspection is possible and appropriate. Inspection records of wearing parts can be used to establish wear rates that may be extrapolated to indicate refrigerator life. The use of flight qualified electronic parts, although more expensive, appears to prevent numerous breakdowns later during refrigerator life testing or use.,

In selecting the type of cooler to be developed, it is sometimes necessary to choose between a coaler that has good,reliability associated with simplicity and a low number of parts but is limited by wearout of parts as compared to a more complicated machine that hgs lower reliability but better durability. Awareness of this conflict in relation to operational requirements can aid in selecting the best cryo-cooler for a particular application.

The quality of the support equipment should also be carefully considered before acquisition. Cooling cart (for heat rejection) breakdowns have been a continuing problem that

interrupts cooler testing and causes delays. Power supply and instrumentation malfunctions have resulted in many lost hours of operation and have confused the malfunction analyses. Adequate spares should be acquired to prevent long term breakdowns of the entire system.

Reliability and durability can be verified in several ways. Current Flight Dynamics Laboratory programs use the Test, Analyze, and Correct (TAC) method to demonstrate the life and increase the reliability of long-life refrigerators. This method provides reliability information on the failure modes and mechanisms of crit3cal components while undergoing development and life testing. Reliability improvement (growth) results when failure modes and mechanisms are discovered, identified, and their recurrence prevented through implementation of corrective action.

3. Conclusion

In summary, working across all refrigerator technical fronts, there is a high expecta- tion that the technology for five year coolers will be available within the next five years. Development of a highly reliable, long-life cryo-coolermustbe acarefully planned and controlled process. Required operating conditions will influence the selectionof the basic type of cooler, but the selected concept must have an inherent long-life potential. This potential is then realized through careful analysis of stress levels for each part and adequate quality control. Contamination control must also be implemented. Good reliability can be obtained by limiting the number of parts in a relatively simple cooler and by assur- ing adequate design margins for all parts. Final verification of adequate durability with acceptable cooler performance is demonstrated by reliability growth in a test, analyze, and correct program.

4. References

[l] Doody, R. D., The High Capacity Spaceborne Vuilleumier Refrigerator, Paper No. 245-15 presented at the SPIE Annual Technical Symposium, July 1980.

[21 Wapato, P. G. and Norman, R. H., Long Duration Cryogenic Cooling with the Reversed Brayton Turbo-Refrigerator, Paper No. 245-17 presented at the SPIE Annual Technical Symposium, July 1980.

[3] Breckenridge, R. W. , Refrigerators for Cooling Spaceborne Sensors, Paper No. 245-16 presented at the SPIE Annual Symposium, July 1980.

[ 4 ] Shqrman, A., Gasser, M., Goldowski, M., Benson, G., and McCormick, J., Progress on the Development of a 3-5 Year Lifetime Stirling Cycle Refrigerator for Space, Paper No. KC-1 presented at the Cryogenic Engineering Conference, August 1979.

151 Sullivan, D. B., and Zimmerman, J. E., Very Low-Power Stirling Cryocoolers using Plastic and Composite Materials, International Journal of Refrigeration, Volume 2, No. 6 pp 211-213, November 1979.

THEORETICAL ANALYSIS OF A 3-STAGE STIRLING CYCLE CRYOCOOLER

S t u a r t B. Horn Mark S. Asher

Far I n f r a r e d Team, Systems I n t e g r a t i o n D i v i s i o n N igh t V i s i o n & E lec t ro -Op t i cs Laboratory

F o r t B e l v o i r , V i r g i n i a

Previous analyses o f t h r e e stage coo le rs have neg lec ted t h e r e a l gas equat ion o f s t a t e o f helium. Those analyses have t h e r e f o r e neg lec ted t h e compressibility e f f e c t s t h a t occur below 20K. Thts paper de r i ves t h e thermodynamic performance o f t h r e e stage S t i r l i n g t ype coo le rs us ing t h e r e a l gas equat ion o f s ta te . An a n a l y s i s of a smal l , e f f i c i e n t t h r e e stage coo le r i s gfven w i t h t h e low temperature gas e f f e c t s i l l u s t r a t e d .

Key woyds: Cryocooler; cryogenic; he1 iurn; low temperature; r e f r i g e r a t o r ; S t i r 1 i n g cyc le .

The three stage Stirling Cryocooler (Fig. 1) consists of a motor driven reciprocating helium compressor and a cold finger containing three expansion chambers separated by porous regenerators. The displacer-regenerator assembly may be driven by a cam or slider crank mounted on the same shaft as the compressor motor. Alternatively, a linear solenoidal drive may be used t o drive the displacer assembly with no side thrust. The motion of the displacer could be easily changed with the solenoidal drive whereas a new lobe would have t o be ground if a cam were used t o drive the displacer.

Cooling is produced in the three expansion chambers by phasing the displacer motion with the pressure pulse generated by the compressor so that the working gas expands a t high pressure and contracts a t low pressure (Fig. 2). The regenerator? serve t o maintain the temperature difference between the three expansion spaces in the cold finger. As gas travels down toward the cold end of the cold finger i t is precooled by the porous regenerator matrices. The matrices a re recooled by the gas af ter expansion a s i t flows back t o the ambient end. The matrix materials should have high volumetric specific heats relative t o the compressed helium. If this is so, the axial temperature distribution in the regenerator matrices will be essentially constant in t ime and l i n e a r between t h e expansion spaces. Because t h e heat capac i t y o f t h e s o l i d drops so severely.. a t low temperature (below 30K) and t h e he l ium heat capac i t y i s pressure dependent i n t h e same range, t h i s assumption i s c l e a r l y n o t exact a t t h e lower end o f t h e machine.

In most cases, refrigeration will be required only at the cold end of the cold finger. The upper two stages should then be designed t o produce no net refrigeration. They should serve only t o eliminate the flow losses associated with the regenerators.

The computer program used t o predict the performance of the cryocooler is a modif ied version of the one described by Horn & Lumpkin (1) of the Army Night Vision Laboratories. This is a second-order isothermal analysis with decoupled losses. The original program was modified t o include a three stage cold finger and real gas equations of s t a t e in the cold end. The program contains an i terative procedure t o find the correct mass distribution in the machine and in i terative solution t o the real

FIG. 1. THREE STAGE STIRLING CRYOCOOLER

TI2

COOLER PERIOD

FIG. 2. NORMA~IZED EXPANSION SPACE PRESSURES, VOLUMES, AND VOLUME RATES

READ DATA

1 INITIALIZE MASS DISTRIBUTION BETWEEN WORKING AND CRANKCASE VOLUMES

INITIALIZE DLSPLACER AND COMPRESSOR POSITIONS FORt=O

AND FIND NEW SOLVE FOR THE SYSTEM DISPLACER AND PRESSURE FOR A PARTICULAR COMPRESSOR DISPLACER AND COMPRESSOR POSITIONS

AVERAGE CYCLE PRESSURES

ADJUST MASS FRACTION

AND SUBTRACT LOSSES

FIG. 3. PROGRAM FLOWCHART

gas equat ion o f s ta te . The non-zero Joule-Thomson c o e f f i c e i n t has n o t been completely inc luded i n t h i s model.

A thermodynamic model of the cryocooler is shown in Fig. 1. In order for there t o be no long term leakage of gas around the compressor piston seals, the t ime averaged pressure in the crankcase must be equal t o the cycle averaged pressure in the working volumes:

To satisfy this condition the program solves these integrals numerically and adjusts the amount of mass present in the working and crankcase volume. In order t o evaluate the cycle averaged pressure the program solves the real gas equation of s t a t e numerically for every t ime step. There a r e then three nested i terative procedures contained in the program a s shown in the flow chart (Fig. 3).

First, a rough estimate must be made of the mass distribution in the machine. The total amount of mass that should be present in the entire working volume can be found by summing the masses in the separate subvoiumes:

2 Vre 3 v c l + g + Vd l i ne R T3 T1 T3+Tamb Tamb ( 2 )

+ -

Vamb + - + - Tamb Tamb

It has been assumed that the machine begins i ts cycle with both the compressor and the displacer a t mid-stroke. Also, al l isothermal subvolumes a re assumed to already be a t their equilibrium temperatures. The subscript 1 refers t o the first mass distribution iteration. The mass in the crankcase can be found from the following relation:

PC Vcr Mcrl = - - R Tcr

Where the volume of the crankcase is assumed t o be essentially constant. The total mass of $he system is then:

MTOT = Mwl + Mcrl ( 4 )

W e now define the first est imate of the working gas mass fraction as:

In order t o proceed further we must find the pressure in the working space as a function of t ime for one complete cycle. During this f irst cycle we will assume that there is no leakage of mass from the compression space into the crankcase. For some general point in the 'cycle, then, we can relate the instantaneous pressure in the working space to the mass fraction by again summing up the mass contributions of the subvolumes:

Pw . Mtot Fj = TI V C ~ ~ + 2 Vreq3 Vdl i ne T3+Tamb

+ - Tamb (6)

Vambi + - Tamb + p r e g l Vreg l + p c l Vcli

Where the following assumptions have been made:

1) All subvolumes a r e isothermal.

2) All subvolumes in the working space a r e in pressure equilibrium (no pressure drops).

3) The gas i n the regenerators i s a t t he a r i thmet i c mean temperature o f the expansion spaces t h a t they connect. This i s a s imp l i f y i ng assumption. While the temperature i s a l i n e a r func t ion o f pos i t i on , the gas dens i t y va r ies l oga r i t hm ica l l y w i t h z.

4) The ideal gas law is correct in all subvolumes except the cold expansion space and the cold regenerator.

The subscript i in Eq.. (6) refer5 t o a particular instant in the cycle while the subscript j refers t o a mass fraction iteration. The motion of the compressor and the displacer can be either sinusoidal or matched constant acceleration parabolic sections. The displacer and compressor motions define the instantaneous volumes Vcli, VcZi, V C ~ ~ , Vamb, and Vzi. Solving Eq. (6) for the system pressure we have:

pwi tot Fj - ( p r e g l i b e g 1 + P C $ v c l i ) l R - Denom

(7) where :

Vc2i + 2Vreg2 Vc3i + 2 Vreg3 Vdl ine Vamb Vz. Denom= - + - + - + +1 T2 T2 tT3 T3 T3+Tamb Tamb Tamb Tamb

All terms in Eq. (7) a re known except for pregl ; and pcIi. An i terative procedure must be used t o find a pc l i and a pregl. that result in pressure equilibrium throughout the machine. The flowchart in Fig. 4a shows how this procedure works.

First, a rough estimate of the instantaneous pressure in the machine is made by using the ideal gas law t o find p c l and p reg1 in Eq. (7):

P W ~ = tot -( 2Pwi Vregl + Pwi Vcli) R R(T1 +T2) Kn 1 /e%m

Solving for Pwi we have:

Pwi (guess) = R Mtot F j [I +I ( 2 Vregl + %)I Denom Denom TI + T2 T2

I USING IDEAL GAS LAW: I I

4 GUESS COLD EXPANSION SPACE DENSITY:

+ CALCULATE INCREMENT FOR QCli:

I P (GUESS) ,A@ = 0.28 Wi- RTl

INCREMENTAL SEARCH

I

SOLVE FOR SYSTEM PRESSURE USING REAL GAS EQN. OF STATE:

Pwi =cbl@,,{GUESS),T,) - 4

SOLVE FOR COLD REGENERATOR DENSITY:

@REG ii =

4

FIG. 4a. FLOWCHART FOR FINDING SYSTEM PRESSURE

1 ,

EVALUATE FUNCTION y:

NEW = + '@C'I~N~UESS,' T1) - +(@f?EG 1,

1

OLD ' NEW

@ci,m~ess, = @cii(~uess~ + A@

YES 4

+ +

An initial guess'is then made for the instantaneous density in the cold expansion space:

pc1 j (guess) = 0.2 PWi (guess) - R TI

The factor of 0.2 is present t o insure that the initial guess of pcli is below the actual value. Once p c l i (guess) is known we can solve for a new pressure from a real gas equation of state.

The equation of state that we will use is presented by Mann (2). The equation contains seventeen terms and is contained in the Appendix for continuity. It is accurate from 4K to 300K and t o pressures of 100 atm. This equation is explicit in pressure; tha t is, pressure is found as a function Z$ of density and temperature:

We can now substitute our f irst guess for the density in the cold expansion space along with the temperature in this space in Eq. (1 1):

pwi = 4 ( P C ~ (guess) , TI)

Next Eq. (6) is solved fo t pregl in terms of PA and Pwi:

pregl i = lpwi R - Fj tot - p c l i (guess) VCI;] vreg1-'

We now define a function Jl that will be zero when the appropriate denstities are found:

Note that $ is really a function of p c l i only because P reg1 can be found from knowledge of p cli. We can now define an increment by which t o sequentially increase p cli until we have found an interval of values of P cli tha t contain a root t o Eq. (14):

We now increment p c l . successively by AP until t he function Vl changes sign. This defines an interval which contains a root of $ and we can quickly converge on this root by the bisection method. The bisection procedure is illustrated in the flowchart shown in Fig. 4b.

This process is continued until the pressures and densities a re known for each t ime s tep in a cycle of the machine. .The integrals for the t ime averaged pressure in t h e working space and crankcase in equation 1 can then be evaluated numerically:

1 " Pw = (1-Fj) Mto t R Tcr / Vcr n i5

+ 41 " 4~1 , GUESS

42 " &l, GUESS - *Q yi = NEW % Uz' WOLD

I 4

1

SOLVE FOR SYSTEM PRESSURE USING REAL GAS EON. OF STATE:

Pw, = O (PAVE, 11)

4 SOLVE FOR COLD REGENERATOR DENSITY:

Pwi DENOM

- F j M ~ O ~ - eAVE vela 'k: +

EVALUATE FUNCTION ty:

FIG. 4b. BISECTION SEARCH

where we have assumed that the crankcase volume is essentially constant. The term on the left represents the working space average pressure (&)'while the term on the right is the crankcase average pressure (k r ) . ,If the t ime averaged pressure is not the same in the working space we must adjust the mass fraction. This is done by relating the pressures in the (j+l)th iteration t o the pressures in the jth iteration as follows:

We can solve for Fj+l by substituting equations (17) and (18) into (16) , resulting in:

Fj+l = (1-Fj) Fw Fj Pcr

After adjusting the mass fraction the progam goes through another cycle and calculates the system pressures for each t ime step. This prbcess is repeated until t he t ime averaged pressures in the working space and the crailkcase a re within some specified error c'.

The gross cooling for each s tage is found by integrating the system pressure with respect t o the expansion space volume:

692 = A2 fig1 Ti-

These expressions a r e valid because the expansion spaces a r e assumed t o be isothermal; the cyclic integral of the work done by the gas is then equal t o the heat absorbed by the gas. ow ever, enthalpy i s a fundtion of pressure and since the ingoing and outgoing streams are a t different pressures, this leads to an incomplete account of the Joule-Thomson effect.

After finding the gross cooling for each stage the net cooling is found by subtracting off the various thermal losses presented in Table 1:

The thermodynamic work done by the compressor is found by numerically integrating the system pressure with respect t o the compression space volume:

The net work done by the compressor is found by adding the thermodynamic work t o the flow by adding the thermodynamic work t o the flow losses of the gas in t h e transfer line and regenerators:

k(net) = i c + FIOW l o s s e s (25)

TABLE I - Thermal Losses

CAUSE LOSS

Shuttle

Regenerator Heat Transfer

Temperature Swing

Pressure Drop

Conduction

Stage

Regenerator motion causes heat t o b e transferred from warm t o cold end.

Irreversible heat transfer between regenerator matrix and helium.

Temperature cycling of regenerator matrix

Viscous drag in flow passages causing reduction in gross cooling.

Irreversible heat transfer between warm end of displacer assembly and cold end.

TABLE I1 - Summary of Optimized Three-Stage Cooler Design

Temperature, OK, Swept Volume, i 5

Dead Volume, in R e e n e r a t o r Void Volume,

a. Data for Various Stages

1 2

in' Reynera to r Flow Area, 4.74

in Regenerator Matrix Material Lead-Antimony Nickle Spheres Nickle Spheres

Spheres Regenerator Length, in. 0.780 0.880 0.930

b. Ambient End Data

Charge Pressure - 500 Psia Ambient Temperature - 300K Cycle Speed - 25 Hz 3 Compressor Swept Volume - 0.131 in-3 . Compressor Dead Volume - 2.98 x 10 ~ n - ~ . Connecting Line Dead Volume - 5.65 x 10 In

RESULTS

Airborne SQUIDS require small, eff icient cryocoolers with small magnetic signatures. To achieve the small packaging requirement designs utilizing high charge pressures and cycle speeds to compensate fo r t he lack of physical s ize were examined.

The optimized three s tage cooler design is summarized in Table 2. The design was optimized by adjusting t h e length of t h e regenerators and t h e s t age temperatures until t he desired ne t cooling was obtained a t t he coldest stage. The performance of t he machine is shown in Table 3,

The program was re-run using the s ame data except t h a t t h e real gas equation of s t a t e was replaced with t h e ideal gas equation of state. The results a r e shown in Table 4. I t is clear t ha t t h e e f f e c t of compressibility in t h e cold s t age expansion 'space and regenerator substantially a l te red t h e performance of t h e machine. The pressure ra t io and gross cooling were reduced when compressibility was not included in t he analysis. Compresssibility e f f ec t s tend t o magnify t he swept volume in t he cold expansion space relat ive t o t h e other subvolumes in t he machine. The compressibility fac tor Z is defined by:

PV = ZMRT (26 )

The ef fec t ive tempera ture weighted swept volume of t h e cold expansion space is then given by:

A vL~ - Zcl n

where %I is t h e cyclic average compressibility factor . In t h e cold expansion space, similarly, t h e ef fec t ive tempera ture weighted dead volume of t he cold end regenerator is:

I Vregl = Vregl -

Zregl T I

As can be seen from Fig. 5, t h e compressibili$y f ac to r s f o r both t h e cold end expansion space and the cold end regenerator a r e always less than unity. However, t h e average compressibility fac tor for t h e cold end expansion space is less than tha t of t h e cold end regenerator (since t he regenerator is a t a higher temperature). This causes t h e t o t a l tempera ture weighted swept volume of t h e machine t o increase a t t h e expense of t he to ta l tempera ture weighted dead volume. The pressure ra t io and gross cooling a r e thus increased.

Several runs were also made using various displacer motions other than sinusoidal. Motions were found t h a t gave substantial increases in ne t cooling. The compressor swept volume was then reduced until t h e ne t cooling was close t o i t s original value, but because of t h e smaller compression space the input power was reduced. The performance characterist ics of this design a r e l is ted in Table 5 and the optimized displacer waveform is shown in Fig. 6. The optimum waveform has short dwells a t t h e ends of t h e stroke tha t tend t o round off t he work curve in t he cold expansion space. By using the optimized waveform, t he COP of t he cooler was increased by 9%.

TABLE 111 - Performance Characteristics of Three Stage Cryocooler

Stage 1 2 3

Gross Cooling, W Shuttle Loss, \Y Regenerator Heat Transfer

LOSS, W Regenerator Temperature

Swing, K Temperature Swing Loss, K Pressure Drop Loss, W Total Conduction Loss, W Pumping Loss, W Net Cooling, W

Pmax - 46.1 atm Pmin - 26.9 atm Pressure Ratio - 1.71 Motor Shaft Power - 54.4 W Flow Losses - 3.10 W

TABLE IV - Performance Characteristics of Three-Staee

Cryocooler Ignorring Compressibility

Stage 1 2

Gross Cooling, W Shuttle Loss, W Regenerator Heat Transfer Regenerator Temperature


Pmax - 44.0 a tm Pmin - 27.5 atm Pressure Ratio - 1.62 Motor Shaft Power - 52.2W Flow Losses - 3.77 W

TABLE V - Performance Characteristics of the Three-Stage Cryocooler

with Non-Sinusoidal Displacer Motion

Stage 1 2 3

Gross Cooling, W Shuttle Loss, W Regenerator Heat Transfer

Loss, W Regenerator Temperature


Pmax - 48.0 atm Pmin - 25.8 atm Pressure Ratio - 1.86 Motor Shaft Power - 48.6 W Flow Losses - 1.66 W

ITERATION NO.

FIG. 5. COMPRESSlBlLlTY FACTORS FOR COLD EXPANSION SPACE CVC1) AND COLD REGENERATOR (VREGI)

CONCLUSION

Tnis investigation demonstrates the feasibility of constructing a small, efficient three stage 10K Stirling Cryocooler. I Compressibility should be taken into account when writing computer simulations for cryocoolers operating in this temperature range. Also, the displacer motion can be optimized t o increase the performance of the machine.

NOMENCLATURE

2 Az - Swept area in compressor piston (m ) 2 A1,2,3 - Swept area of expansion spaces 1, 2, and 3 (m ) Fj - Mass fraction in working volume for jth mass fraction iteration Mcrj - Mass in crankcase volume in jth mass fraction iteration (kg) Mtot - Total mass of helium charge (kg) Mwj - Mass in working volume in jth mass.fraction iteration (kg) n - Number of discrete t ime steps per one cooler period P - Pressure (Pa) PC - Charge Pressure (Pa) Pcr - Crankcase Pressure (Pa) Pw. - Instantaneous working volume pressure (Pa) ~ ~ i , 2 , 3 - Cross cooling in stages 1, 2 and 3 (W) '

Qnet - Net cooling in a particular stage (W) Qloss - Cooling losses in a particular stage (W) R - Ideal gas constant for helium (Jlkg K) T - Temperature (K) Tambr- Ambient temperature (K) Tcr - Crankcase temperature (K) TI, T2, T3 - Tepperature in expansion spaces 1, 2, and 3 (K) V - Volume (m ) 3 V_amb. - Instantaneous ambient end volume (m ) 3 Vcl, 9c2, 8 3 - Midstroke y l u m e in expansion spaces 1, 2, and 3 (m ) Vcr - Crankcase volume (m ) 3 Vreg 1, 2, 3 - Void volume in regenerators+, 2, and 3 (m ) Vzi - Instantaneous compressor volume (m ) W c - Thermodynamic work done by compressor (W) Xi - Instantaneous displacer velocity (m/s) Zi - Instantaneous compressor velocity (m/s) Z - Compressibility factor V C ~ ~ , V C ~ ~ , V C ~ ~ - Instantaneous volume in expansion spaces 1,2, and 3

CREEK LETTERS

3 p - Density (kg/m .) 3 pregli - Instantaneous gas density in regenerator 1 (kg/m I3 pcl . - Instantaneous gas density in expansion space 1 (kg/m ) -r - Cooler period (sec)

SUBSCRIPTS

i - Associates a quantity with the ith t ime step j - Associates a quantity with the jth mass fraction iteration

APPENDIX: Equation of Sta te for ~e~

4 Mann (2) presents the following equation of s t a t e for He between 4.OK and 300K. This equation shows good correlation with experimental data for pressures up t o 100 atm:

2 4 2 P = RdT + (Rnl T + n2 + n3/T + n4/T + n5/T )d

3 + (Rn6T + n7)d + n8 d 4

2 + (n9/T + N ~ ~ / T ~ + nl I / ~ 4 ) d 5

2 + exp (-nI2 d /T)

2 3 4 5 + (n13/T + n14/T + n15/T Id

2 5 + exp(-n12 d IT) +n16 d +n17 d6

Where d is in gmollliter, T in K and

BIBLIOGRAPHY

(1) S.B. Horn and M.E. Lumpkin, Advances in Cryogenic Engineering, Vol. 14, Plenum Press, New York (19741, p. 221.

(2) D. B. Mann, NBS TN 154, 1962.

SOME THERMODYNAMIC CONSIDERATIONS OF HELIUM TEMPERATURE CRYOCOOLERS*

D. E. Daney

Thermophysical Properties D iv is ion Nat i onal Bureau o f Standards

Boulder, Colorado

Attempts t o reach 1 i qu id he1 ium temperatures wi th regenerative cryocoolers have been generally unsuccessful. The pr inc ipa l and we1 1 known cause o f t h l s f a i l u r e i s the decay of. the regenerator matr ix heat capacity a t low temperature. As the regenerator effectiveness decreases, more and more o f the re f r i ge ra t i on e f fec t goes to the regenerator losses rather than t o the load. A less well-recognized problem f o r some cryocoolers has been operation under adverse thermodynamic conditions. For example, operation a t high pressure (40 bar) can decrease the calculated e f f i c iency o f the lowest stage by a factor o f 10 or more - pr imar i l y because o f the strongly negative Joule-Thomson e f f e c t o f helium a t low temperature .and high pressure.

We f i r s t invest igate the e f f e c t o f the operating regime on the e f f i - ciency o f helium temperature cryocoo1ers. Next, we propose two schemes o f hybrid regenerative-counter flow cryocoolers as sol utfons to the problem o f low helium-temperature regenerator heat capacity. Final ly, we calculate the net mass flow through a typical hybrid regenerator-heat exchanger f o r one o f the schemes tha t we propose.

Key words: Cryocool er; hybr id cryocool ers; ref r igerat ion; regenerator; regenerative cooler; S t i r l i ng cooler; S t i r l ing cycle.

1. Int roduct ion

Much o f the i n te res t i n re l iab le , low power, and compact cryocoolers f o r cool i n g smaller superconducti ng devices fs focused on regeneratf ve devf ces such as the St f rl f ng machi ne. These coolers have a proven record o f performance a t temperatures above 20 K, and the i r re la t f ve sfmplf cf ty i s pa r t f cu la r l y appeal? ng because f t o f fe rs the potent ia l o f high re l fab f 1 f ty . However, attempts to reach l f q u f d helfum temper tures wfth regen- e erat fve cryocoolers re ject fng heat to ambfent temperature have f a i l e d . Although temperatures as low as 6.5 K have been achfeved, the e f f i c iency f a l l s rapid ly a t temperatures below 10 K C1,2,31 .

*Contribution o f the Natfonal Bureau o f Standards, not subject to copyrfght.

l~emperatures as low as 3.1 K have been achfeved by Zfnmerman and Sul l ivan [8] wi th heat re jec t ion a t 8.6 K. Thfs was achfeved by operaton a t low pressure - a very o r i g ina l approach t o a1 lev fa t fng both o f the thermodynamic problems df scussed i n thfs paper.

The prfncfple and well known cause of thfs faflure is the decay of the regenerator matrix heat capacity a t low temperature. An accompanying loss f n regenerator effectfveness results i n more and more of the refrigeration effect gofng fnto regenerator losses rather than the load. A less well understood problem for some cryocoolers has been operatfon under adverse thermodynamic condftfons, i .e., operatfon a t pressures and pressure rat ios where the efficiency f s poor.

T h f s paper proposes an alternative approach to the problem of dfmfnfshing low temperature regenerator heat capacity. Instead of searching for new regenerator materials, I suggest using hybrfd cycles w f t h helfum counterflow fn the ffnal stage of the refrigerator. Two schemes whfch embody t h f s prfncfple are given.

2. Regenerator heat capacf ty

The requirement of hfgh regenerator heat capacity has long been recognfzed, and we discuss f t here only i n sufffcient detafl to deffne the magnitude of the problem fo r helfum temperature regeneratfve coolers. An empfrfcal expressfon for the effect of heat capacfty on regenerator effectfveness is given by Kays and London 141 for rotary regenerators:

where E i s the regenerator effectiveness, EO the effectfveness for fnffni t e regenerator heat capacity and Cr f s the heat capacf ty of the regenerator dfvfded by heat capacfty of the gas passfng through in one direction during one cycle. Using th f s expressfon we find that C r must be greater than 3 for the degradation 'in E to be less than 0.01. By compar- fson, the experfmental results of Gffford and Acharya C51 give C r greater than 5 for a lead sphere regenerator operatfng between 80 K and 20 K. Whatever the exact values, both resul ts support the conclusfon that C r should be several tfmes greater than one if a sfg- nf ffcant deterf oratf on i n effectfveness f s not to occur.

Comparing the volumetrf c heat capacfty of helfum wf t h that of lead (the best common regenerator material 1, ffgure 1, we note that hfgh losses may be expected below about 10 K . A1 though much may be accompl f shed through system optfmfzatfon and materf a1 s research, ffgure 1 offers l f t t l e hope for e f f fc fen t regeneratfve cryocoolers operating a t 4 K. Thfs, of course, i s not a new opfnfon.

3. Thermodynamfc regfmen

Because of their great complexity, regeneratfve cycle calculatfons are usually made usf ng the sfmpl ifyfng assumptfon of an fdeal gas equatfon of state. So long as the operating temperature is well above, the c r f t i ca l pof n t , 1 i t t l e error f s generally i ntroduced by thfs assumption. As the operatfng temperature approaches the crf t f ca l temperature, however, the error introduced i n some cases may be large enough to make the calculatfons meanfngless.

In order to characterfze the effect of the operatfng conditions on the performance of the f f nal stages we analyze a ffnal Brayton cycle stage usfng real he1 fum property es and methods descrfbed previously C61. A1 though not a substitute for detaf led regeneratfve cycle calculations, the results for a 6 K operatfng temperature, given In figure 2 , clearly i l lus - t r a t e a strong effect of operatfng regfmen on performance.

He(5 bar)

5 10 15

TEMPERATURE, K

Figure 1. Volumetric heat capacity o f helium and lead.

PRESSURE, P 1 /P5

Figure 2. Performance of final refrigerator stage as a function of pressure and pressure ratio.

The dependence o f the idea l work on the pressure i s p r ima r i l y due t o the inc reas ing ly negative Joule-Thomson e f f e c t o f he1 f um w i t h increas ing pressure. The r e s u l t i s progres- s i ve l y l e ss coo l ing as the pressure increases. Low pressure r a t i o s are less e f f i c i e n t because a h igh f r a c t i o n o f the r e f r i g e r a t i o n i s l o s t i n the heat exchanger (regenerator) . For less e f f e c t i v e heat exchange, the e f f e c t o f pressure r a t i o i s even stronger. For the model used here, a t l eas t , r e l a t i v e l y low operat ing pressures and modest pressure r a t i o s are favored.

4. Hybri d regenerat ive coolers

As an a l t e rna t i ve t o the search f o r a high heat capacity regenerator mater ia l , we propose a hybr id machine w i t h counterf low heat exchange i n the f i n a l stage. Such an arrangement l a rge l y e l iminates the need f o r a high heat capacity regenerator because heat i s t rans fe r red d i r e c t l y from one helium stream t o another. Figure 3 shows two possible arrangements f o r achieving such counterflow. I n Scheme A, two regenerat ive coolers oper- a t i n g 180 degrees out o f phase share a common counterf low heat exchanger, o r hybr id heat exchanger-regenerator i n the f i n a l stage. The gas en te r ing one expansion space i s cooled i n counterf low by the gas f low ing from the second expansion space. The mass f low wave-form i s d i s t o r t ed from the sinusoidal , so t h a t some instantaneous l oca l f l ow imbalance occurs. Thus, a hybr id heat exchanger-regenerator (e.g. - a heat exchanger w i t h h igh heat capaci ty) i s desirable. A1 though a s ing le p a i r o f displacers are shown i n the f igure, any number could be coupled i n a s im i l a r fashion. For example, f o r three d isp lacers the phase angle would be 120 degrees and f o r fou r displacers the phase angle would be 90 degrees i n order t o f u r t h e r reduce any f low imbalance. Scheme A was suggested to me by Ray Radebaugh of NBS; i t s exact o r i g i n i s unknown t o us.

I n Scheme B the l a s t stage s f the d isp lacer i s t i g h t l y sealed from the r e s t o f the system so t h a t i t may a c t as a conventional expansion engine. During the h igh pressure p a r t o f the cycle, gas i s admitted to the plenum v i a valve 2. When the p is ton-d isp lacer reaches bottom dead center, valve 3 opens and admits gas t o the cy l i nde r u n t i l the p i s t o n r i s e s t o the cu t - o f f pos i t ion. As the p is ton continues upward, the gas cool s by expansion; and a po r t i on i s l i que f ied . At top dead center the exhaust va lve (4) opens, and two-phase hel ium i s discharged i n t o the load heat exchanger-plenum. Return o f the gas t o the main working volume occurs through valve 1 dur ing the low pressure p a r t o f the cycle. Counter- f l ow heat exchange between the h igh and low pressure streams occurs i n a somewhat conven- t i o n a l heat-exchanger. However, because the mass flow i n the two heat exchanger streams i s n o t instantaneously equal, one side ( t he low pressure side) should have a volume much greater than the other s ide t o provide add i t i ona l heat capacity. Because o f i t s complexity, Scheme B i s probably feas ib le only f o r coolers w i t h a capacity near one wa t t and above. Scheme B i s an extension o f a concept suggested by S. C. Co l l i n s some years ago C71.

I n order t o ca lcu la te the reduct ion i n the regenerator load accomplished by Scheme A, we ca lcu la ted the pressure and mass f low f o r a 3-stage S t i r l i n g cooler using rea l helium proper t ies and sinusoidal movement o f the d isp lacer and p is ton. The f o l lowing character- i s t i c s were assumed:

Hybrid Counter Flow /

Heat Exchanger-Regenerator

Scheme A. Displacers 180' out of phase.

Exhaust Valve

Scheme B.

Figure 3. Hybrid re f r igera tor schemes f o r the f i n a l stage o f regenerative cooler.

Table 1. Spl i t - S t i r 1 i n g cooler parameters f o r sample ca l cu l a t i on o f mass f low

Isothermal compressor temperature 300 K 1 s t expansion space temperature 100 K 2nd expansion space temperature 25 K 3 rd expansion space temperature 6 K Compressor volume 300 cm3 1 s t expansion space volume 17 cm3 2nd expansion space volume 4 cm3 3 rd expansion space volume 1 cm3 Regenerator vo id volume = 1 /2 expansion space volume.

F igure 4 shows t h a t the mass f low r a t e i s nonsfnusoidal w i t h the f low maximum and min i - mum separated by the compressor-displacer p.hase angle, 90 degrees i n t h i s case.

It i s the t o t a l c y c l i c mass f low through the regenerator, no t the instantaneous f low rate, t h a t determines the load on the regenerator heat capacity. Thus, i n Table 2 we com- pare the e f f e c t o f the number o f d isp lacers on the c y c l i c change i n the t h i r d stage expan- s ion space mass.

Table 2. Cyc l i c mass f low w i t h mu l t i p l e displacers

Compressor-displacer angle Number o f 90 degrees 120 degrees

D i spl acers Flow r a t i o 1/N x f l ow r a t i o Flow r a t i o 1/N x f l ow r a t i o

The column denoted f low r a t i o i s the c y c l i c change i n the t o t a l t h i r d stage mass compared t o the change f o r a s ing le displacer. Since mu l t i p l e d isp lacers would g ive N times the capaci ty o f a s ing le d isp lacer , a f u r t h e r reduct ion i n mass f low i s achieved by the reduc- t i o n i n the s i ze o f each d isp lacer when m u l t i p l e d isp lacers are used. Thus, the column denoted 1/N x f l ow r a t i o gives the t o t a l reduct ion c y c l i c mass f low f o r Scheme A.

A dramatic reduct ion i n mass f low occurs when the number o f d isp lacers gives a phase angle between d i spl acers equal t o the compressor-displ acer phase angle. Four d i s p l acers (w i t h a 90 degree compressor-displacer phase angle) g ive a f i f t y f o l d decrease i n the regenerator heat load.

5. Summary

I n summary, I have proposed an e f f i c i e n t helium temperature regenerat ive cryocooler. It i s a hybr id device us ing counterf low heat exchange i n the f i n a l stage. It should have a maximum cyc le pressure below 10 bar, and a pressure r a t i o above 2.

Figure 4. Pressure and mass f law waves f o r a S t i r l i n g cooler.

6. References

C11 Stuart , R. W. and Cohen, 8. M., Operation of a Three Stage Closed-Cycle Regenerative Re f r ige ra to r i n 6.5 K Region, i n Advances i n Cryogenic Engineering, Vol . 15, edf ted by K . D. Timerhaus (Plenum Press, New York, 1970).

[2] Daniels, A. and du Pre", F. K., Triple-Expansion S t i r 1 ing-Cycle Refr igerator , Advances i n Cryogenic Engineering, Vo1 . 16, ed i ted by K. D. Tfmmerhaus, 178-184 (Plenum Press, New York , 1971 1.

[3] Zimmerman, J. E. and Radebaugh, R., Operation o f a SQUID i n a Very Low-Power Cryo- cooler, NBS Special Pub l i ca t ion SP-508, ed i ted by J. E. Zimmerman and T. M. Flynn, Natfonal Bureau o f Standards (1 977).

C41 Kays, W. M. and London, R. L., i n Compact Heat Exchangers, 20 (McGraw-Hi11 Book Co., Inc., New York, New York, 1978).

[ 5 ] Gif ford, W. E. and Acharya, A., Low-Temperature Regenerator Test Apparatus, i n Advances i n Cryogeni c Engineering, Yo1 . 15, edi ted by K. D. Timmerhaus, 436-442 (Plenum Press, New York, New York, 1976).

C6l Daney, D. E., Low-Temperature Losses i n Superc r i t i ca l He1 ium Ref r igerators, i n Advances i n Cryogenic Engineering, Vol. 15, edi ted by K. D. Timmerhaus, 205-212 (Plenum Press, New York, New York, 1976).

C71 Daney, D. E., Design, Construction, and Test ing o f a Small Novel A i r L iquef ier , S. M. Thesis, Massachusetts I n s t i t u t e o f Technology, 1962.

C81 Zimnerman, J. E. and Sul l ivan, D. B., A m i l l i w a t t S t i r 1 i n g cryocooler f o r temperature below 4 K, Cryogenices - 19, 170-171 (1979).

DYNAMIC ANALYSIS OF A SMALL FREE-PISTON RESONANT CRYOREFRIGERATOR

R.A. Ackermann, Ph.D

Mechanical Technology Incorporated Latham, New York 12110

The development o f small, cryogenic r e f r i ge ra to r s i s rece iv ing i n - creased a t t en t i on through the use o f l i n e a r resonant machinery which has the po ten t i a l fo r increas ing the r e l i a b i l i t y and e f f i c i e n c y of these devices. The use of l i n e a r machinery, however, presents a complex second dimension t o the analys is o f the system's dynamics and thermodynamics which are i m p l i c i t l y t i e d together through the pressure wave produced i n the r e f r i g e r a t o r . This paper describes the dynamics o f a l i n e a r cryogenic r e f r i g e r a t o r and presents the ana l y t i ca l development o f the equations o f motion through the use o f phasor notat ion. Solu- t i ons f o r both the t r ans i en t and steady-state motions are presented, and the condi t ions o f s t a b i l i t y and se lec t ion o f key operat ing parame- t e r s such as frequency, p is ton-d isp lacer amplitudes and phase angles are discussed.

Key words: D i splacer and p i s t on dr ives; 1 inear resonant machinery; 1 inear f ree -d i splacer c ryo re f r ige ra to r ; 1 inear e l e c t r i c motors ; phasor notat ion; thermodynamic and dynamic analyses.

1.0 Descr ip t ion o f a 1 inear resonant cryogenic r e f r i g e r a t o r

A schematic f o r a f ree -p is ton l i n e a r resonant r e f r i g e r a t o r i s shown i n Figure 1. The machine cons is ts o f two dynamic components: a compressor p i s t on which i s electromagnet ical ly actuated a t resonance by a l i n e a r e l e c t r i c motor, and a d isp lacer which i s actuated a t resonance by the pressure wave developed i n the machine. Re f r ige ra t ion i s produced i n the device by the expansion and compression o f the working f l u i d a t d i f f e r e n t temperature l eve l s and by the con t ro l l e d f l ow o f t he working f l u i d between the temperature leve ls due t o volume changes i n the expansion and compression spaces. The operat ion o f the r e f r i g e r a t o r i s def ined by the f o l l ow ing se t o f equations.

1 .I Volume equations

The equations de f i n i ng the changes i n the expansion and compression volumes f o r sinu- so ida l motion of the p i s t on and d isp lacer are:

I V(expansion) = Ve = 7 (Ve)o ( I + s inu t ' )

1 V(compression) = Vc = 7 (Vc) (1 + s i n ( a t ' - 8 ) )

0

5 7

where :

( V ) = Swept volume i n the expansion space e o

(Vc)o = Swept volume i n t he compression space

w = Operating frequency (rad/sec)

e = Angle by which volume var ia t ions i n the compression space.

r- UNEAR MOTOR

MBIENT END Fa]

Fig. 1 MTI Single-Stage L inear Free- Displacer Cryore f r ige ra to r

i n the expansion space l ead those

Fig. 2 Displacer and Volumetric Phasor Diagram

1.2 Pressure equation

The pressure equation i s der ived from the mass c o n t i n u i t y equation f o r the system and i s given by

where :

F = Tota l mass i n the system = Pressure

T = Temperature V = Volume R = Gas constant

subscr ip ts :

ch = Charge pressure e = Expansion space c = Compression space v = Void volume To = I n i t i a l charge volume a = Ambient

The conversion of the absolute volume expressions t o phasor no ta t i on i s made through the transformations :

where :

V = Absolute volume va r i a t i ons -f ve = The phasor expansion space volume v a r i a t i o n due t o the d isp lacer motion

and -t

'c,d = Compression space volume v a r i a t i o n due t o the d isp lacer motion

-f v = Compression space v a r i a t i o n due t o t h e p i s t on motion

CYP

The advantage o f t h i s phasor no ta t ion i s t h a t the amplitude and phase re l a t i onsh ip of each o f t h e system operat ing parameters may now be represented on a phasor diagram t h a t provides phys ica l i n s i g h t i n t o the mathematics. The phasor diagram f o r a cryogenic r e f r i g e r a - t o r i s shown i n F igure 2. The development o f t h i s diagram proceeds from t he r e l a t i onsh ip between t he two pos i t i on phasors.

P is ton Motion

x = (x ) s inwt P P 0

Displacer Motion

xd = (xd) s i n (wt + y) 0

where :

(xdJo and (x ) = the amplitudes of the mption p 0

y = the angle by which t he d isp lacer motion leads the p i s t on motion.

From these motions, the expansion and compression volume phasors may be located on the d ia- gram (Figure 2) as fo l lows:

a The expansion volume v a r i a t i o n i s produced by t he d isp lacer and i s 180" ou t o f phase w i t h the d isp lacer motion.

5 = -Ad (~~)~sin(ot + Y) = Ad ( x ~ ) ~ sin ut'

Ad = the d isp lacer f ron ta l area

a The compression volume va r i a t i ons are produced by both t he d isp lacer and p i s t on motions.

-+ where vCyp i s the v a r i a t i o n produced by the p i s t on motion, which i s represented by a phasor t h a t i s 180" ou t o f phase w i t h t he p i s t on motion

+ v = - A ( x ) s inwt

C¶P P P o

Ap = the p i s t on f r on ta l area

and GqYd i s the v a r i a t i o n produced by t he d isp lacer , which i s i n phase w i t h the d ~ s p l a c e r motion

AROD = the d isp lacer r od area as shown i n Figure 1 . The func t ion of the d isp lacer rod w i l l be covered i n a l a t e r discussion.

The t h i r d dimension i n the analys is i s t he development of the pressure phasor, which i s the r esu l t an t of the pressure va r i a t i ons caused by the d isp lacer and p i s t on motions. This r e l a t i onsh ip i s shown i n F igure 3, and i s def ined as

The d isp l acer motion produces two pressure effect's. F i r s t , var iat ions i n the expansion space produce a pressure change because o f the temperature dif ference between the expansion and compression volumes and the shu t t l i ng o f the working f l u i d between these volumes by the displacer motion. For a cryogenic re f r igerator, where the expansion volume i s a t the co ld temperature, the expansion vol - ume component o f the pressure wave i s a phasor i n phase w i th the displacer mo- t ion . This pressure component i s represented by be,d. Secondly, the volume var iat ions produced by the displacer motion on e i t he r end o f the displacer d i f f e r by the volume occupied by the displacer rod. Thus, neglecting the temperature difference, the displacer motion also causes a pressure change 180' out o f phase w i th i t s motion. This pressure component i s represented by be,d.

r The piston motion produces a pressure change due t o volume changes o f the compression space. This motion produce$ a pressure change i n phase w i th the p is- ton motion, which i s represented by pcYp i n Figure 3.

The sum o f these three components produces the ideal pressure phasor. I n an actual machine a pressure drop w i l l occur across the regenerator, producing a phase lag between pressure changes i n the expansion and compression spaces. The pressure drop i s represented by A$ i n Figure 3, and the expansion and compression space pressure phasors are given by

* Pe = (Pe)o sin (at' + a)

4 p = (P ) sin (at1- (B + 8))

C C 0

where -t

a = angle by which the pressure. phasor (pe) leads the expansion volume phasor

B = angle by which the pressure phasor (PC) lags the compression volume phasor.

REFRIGERATOR COMPONENTS I MOTOR CIRCUIT

Fig. 3 Pressure Phasor Diagram Fig. 4 Dynamic Free-Body Diagram

Up to th is point in the analytical development, the equations apply for both a kinema- t i c or free-piston machine. In a kinematic machine the proper amplitude and phase separation of the piston and displacer are established by a system of mechanical linkages. In a resonant machine the mechanical linkages are replaced by an equivalent se t of dynamic components (springs, masses, dampers, and a displacer drive). Dynamicall y, the free-piston, 1 inear resonant cryogenic refrigerator may be represented by the spring, mass, damper system, shown in Figure 4. The link between the piston and displacer i s a pneumatic linkage created by the engine pressure wave; thus, the additional complexi ty in analyzing this device becomes evident. Analysis of a kinematic device i s less complicated because the linkage (drive mechanism) i s an independent variable controllable through the positioning of the linkage on a crank mechanism. In a free-piston machine the pneumatic 1 inkage i s a dependent variable established by the dynamics of the system. Thus,the thermodynamics and dynamics are tied together through the pressure phasor.

1.3 Thermal equations

Having established the pressure and volumetric phasor relationships, definition of the power input, refrigeration capacity, and equations of motion for a resonant machine now are possible. An overall energy balance of the system gives the power delivered to the machine as the difference between the heat rejected to ambient and the refrigeration capacity.

iIN = (IREJ - (IREF = ' iP where

FEN = the electromagnetic driving force

xp = piston velocity

Considering isothermal compression and expansion processes

(IREF = JPe dVe

By applying the following phasor relationships to these integrals,

where

Po = mean system operating pressure

and integrating over a full cycle leads directly to the following expressions for the re f r i - geration and heat rejection

where

f = the operating frequency i n cycles/se=

These two equations define the diff icwl ty in designing a resonant cryorefrigerator. First , the only independent variables are the frequency ( f ) , and the cross-sectional areas, Ad, A R ~ , and Ap. Second, the five variables ( x ~ ) ~ , (xplo, a, 8, and y are a l l dependent variab es that are functions of the dynamic components, i .e., the masses, spring components, and damping. Finally, the pressure i s a function of the temperatures and volumetric changes as defined by the mass continuity equation, Equation 1. Therefore, the performance of a 1 i near resonant cryorefrigerator i s defined by a mi nimm of 3 simultaneous equations . These are

1 . The thermodynamic equations defining the refrigeration ,capacity 2. The mass continuity equation 3. The equations of motion

1.4 Equations of motion

The equations of motion are derived from the dynamic free-body diagram presented in Figure 4. The equation of motion for the displacer i s given by

i n which Mdjid i s the inertia force on the displacer, cdid is the damping force which results from the viscous losses in the gas springs, Kdxd i s the conservative spring force, and Fd i s the displacer driving force that maintains the displacer motion. The equation defining the driving force i s found from a force balance on the displacer as shown i n Figure 5.

This balance i s

In terms of the phasors described in Figure 3, th is reduces to

where

Ref~rr izg again to the phasor diagram in Figure 3, i t may be seen that the pressure phasors pe, pc, and can be written as 1 inear combinations of the conservative (displacement) and nonconservative (velocity) components as shown i n Figure 6. The figure shows that the foll owing relationships exist .

HEAT EXCHANGE COMPONENTS

Fig. 6 Pressure Phasor Components

Fig. 5 Displacer Force Diagram

+ Ap = -cdid + c i

P P

Substitution of these expressions into the displacer equation of motion leads directly to

' in which the term cd*Ad-id represents a nonconservative restoring force on the displacer produced by the viscous drag in the heat exchanger components; and cp-Ad-Rp and kp*A~o~.xp represent the driving forces on th: displacer produced by the pressure drop across the displacer, i .e . , the heat exchanger ~ p , and the pressure force on the displacer rod.

The equation of motion for the piston i s given by

in which the inertia, damping, and spring forces are described as above. The pressure force ,on the piston i s

F = $ . A P C P

= Ep*xp + Ed.xd

and the electromagnetic driving force for the motor i s given by

FEN = a I , ( t )

where :

a = e/ ip

e = Voltage drop across the motor

Im( t ) = Current i n the motor c i rcui t shown i n Figure 4.

For our case, where the electromagnetic driving force i s produced by. a linear i.nduc- tqon motor, the current i s found from

or, by differentiating

where:

L = the motor inductance K = internal electrical resistance Eo = the amplitude of the applied voltage E(t) IA = the operating frequency T-,. = the phase angle between the applied voltage and the piston displacement

Then, the three equations defining the system dynamics are

The s o l u t i o n t o t h e dynamic equation, i n m a t r i x form, i s

i n which it represents t h e t r a n s i e n t so lu t i on , and xs represents t h e s teady-s ta te s o l u t i o n . A d iscuss ion o f these two s o l u t i o n s fo l l ows .

1.4.1 T rans ien t s o l u t i o n

The t r a n s i e n t equat ion i s

I n t h e m a t r i x form given, t h e t r a n s i e n t equat ion represents a t y p i c a l e igenvalue problem i n which t h e eigenvalues represent t h e p r i n c i p a l modes o f v i b r a t i o n , and t h e e igenvectors de- f i n e the ampl i tudes o f t h e f requency response curves. I n s o l v i n g t h i s s e t o f equat ions t h e e lec t romagnet ic f o r c e c o e f f i c i e n t , a, i n t h e mass and s p r i n g mat r ices prevents a c losed-form s o l u t i o n f rom be ing found and numerical techniques must be used t o f i n d t h e eigenvalues. I n order t o s i m p l i f y t h e problem and g a i n some i n s i g h t i n t o t h e dynamics o f t he mechanical system, we s h a l l assume t h a t t h e e lec t romagnet ic f o r c e may be represented as a f o r c i n g f u n c t i o n o f t h e form

where

q = t h e phase angle between t h e force and p i s t o n mot ion.

With t h i s assumption t h e t r a n s i e n t equat ion becomes

Also, by drawing t h e f o r c e diagram f o r $he d i s p l a c e r (F igure 7, which i s t h e phasor rep re - s e n t a t i o n o f Equat ion 2), t h e term cpAdxp can be expressed as

This substitution leads to a closed-form solution and the following eigenval ues

From necessary cr i ter ia for dynamic s tabi l i ty , the system i s stable i f the eigenvalue coefficients

satisfy the following conditions

If a j i s negative the solution i s unstable If a j i s positive the solution i s stable If b j i s positive the solution i s harmonic

F i g . 7 . Displacer Force Phasor Diagram

I f b j i s negat ive the so l u t i on i s no t harmonic, and I f b j i s negat ive and aj<bj the so l u t i on i s unstable.

Thus, we see t h a t i n order f o r the mechanical system t o be harmonic and stable, the necessary condi t ions are

Rd > ? s i ny + r( P P

which essen t i a l l y says t h a t the d isp lacer d r i v i n g forces due t o the pressure d i f f e r e n t i a l across the heat exchange components and the d isp lacer rod must no t exceed e i t h e r t h e conser- va t i ve spr ing o r nonconservative damping forces.

1.4.2 Steady-state so l u t i on

The steady-state so l u t i on i s found by subs t i t u t i ng f o r the motions

x = (xp) s inwt P 0

xd = (xd) s in (w t + y) 0

This subs t i t u t i on leads t o the steady-state equation

+ j"]xd)o s i n ( ~ t + yl =

'd p ( x ) s inwt P 0

The so lu t ions t o t h i s system o f equations are and the phase re l a t i onsh ip between the p i s t on r a t i o i s

given i n terms o f the p is ton-d isp lacer r a t i o and d isp lacer . The so l u t i on f o r the amplitude

where:

R w = = the na tu ra l frequency o f t he d isp lacer dd Md

damping c o e f f i c i e n t

And the so l u t i on f o r the phase angle i s

2.0 Conclusions

The phasor analys is described presents a usefu l t oo l f o r analyzing and understanding the operat ion o f small, f ree -p is ton S t i r l i n g r e f r i ge ra to r s . The usefulness o f the approach has been demonstrated by the i n s i g h t i t has given t o the operat ion and se lec t ion o f important dynamic components. However, several assumptions were made i n the analys is t ha t have a bearing on i t s accuracy. These were:

1. The motions were described by a se t o f l i n e a r d i f f e r e n t i a l equations i n which the conservat ive fo rce terms are l i n e a r funct ions o f displacement and noncon- se rva t i ve terms are l i n e a r funct ions o f the ve l oc i t y . Nonlinear e f f ec t s are assumed neg l i g i b l e .

3. Only the f i r s t harmonic o f the motions was used t o describe the p is ton and d i sp l acer motions. Higher order terms were considered negl i g i b l e.

3. To s i m p l i f y the analysis, i t was assumed t h a t the e l e c t r i c a l c i r c u i t could be represented by a s inusoida l f o r c i ng func t ion on t he p is ton. The e f f e c t o f t h i s assumption would be t o decouple the motor 's e l e c t r i c a l c i r c u i t from the machine dynamics, making the mathematics manageable. The operat ion o f a r e a l machine i s e f f ec ted by t h i s coupl ing where the i n t e rac t i on w i l l in f luence t he dynamics o f the p i s t on and d isp lacer .

The e f f e c t o f non l i nea r i t i e s on the equations o f motion i s the poss ib le in t roduc t ion o f ad- d i t i o n a l i n s t a b i l i t i e s i n t o the dynamic behavior; the e f f e c t o f h igher order terms i s t o introduce add i t i ona l terms i n t o t he steady-state so l u t i on t h a t may be o f the same order o f magnitude as those given. A complete and thorough analys is o f the device, therefore, must r e l y on numerical procedures, and these procedures must i n t eg ra te the dynamics and thermodynamics o f t he r e f r i g e r a t o r .

References

111 Walker, G . , S t i r l ing-Cycle Machines, Un ive rs i t y o f Calgary, Canada, 1978.

[2] De J~nge , A.K., "A Small Free-Piston S t i r l i ng Refr igerator," 14th I n t e r soc i e t y Energy Conversion Engineering Conference, American Chemical Society, Washington, D.C., 1979.

METHODS FOR THE MEASUREMENT OF REGENERATOR INEFFECTIVENESS*

Ray Radebaugh, Del Linenberger, and R. 0. Voth

Thehophysical Properties D iv is ion National Bureau of Standards

Boul der , Colorado 80303

The regenerator ineffectiveness i s usually the dominant heat 1 oss term i n regenerative-cycle re f r igera tors which operate below about 15 K. Systematic measurements of regenera to r ineffect iveness have never been done i n t h i s temperature range i n spi te of the fac t tha t theoret ical models are usually not adequate to f u l l y describe and predict the thermal behavior of regenerators. The paper introduces the proper clef i n i - t i on of regenerator ineffectiveness, which i s val i d fo r non-ideal gases and systems where a pressure wave exists. A l l parameters which influence the regenerator ineffectiveness are discussed. The concept o f a temperature-dependent ineffectveness i s proposed which s impl i f ies computations of regenerator ineffect iveness when the materi a1 properties are temperature dependent and which y ie lds the average ineffect iveness when it i s integrated between the lower and upper operating temperatures o f the regenerator. Previous methods used and proposed f o r the measurement of regenerator ineffectiveness a t cryogenic temperatures w i l l be reviewed and the i r 1 imi tat ions discussed. A new method w i l l be presented which fo r the f i r s t time w i l l al low measurements of regenerator ineffectiveness a t a l l phase angles between the mass f low ra te and the pressure wave, an important consideration with non-zero regenerator volumes. Temperatures considered are those beween 4 and 35 K.

Key words: Cryogenics; G i f ford-McMahon cryocool ers; heat capacity ; heat transfer; he1 ium; ref r igerators; regenerators; S t i r l i ng cryocoolers; thermal conductance; Vui l 1 eumier cryocool ers.

1. Introduction

Small cryocoolers fo r temperatures down to about 7 K generally use a regenerative cycle. The most c m o n examples are the S t i r l ing, Gifford-McMahon , and Vui l leumier cryocoolers. Their potential fo r long l i f e a t re la t i ve l y low capacities i s one of the main a t t rac t ions o f the regenerative-cycl e cryocool ers . They a1 so have high overal l ef f i c i enci es f o r temperatures down to about 15 K. A t lower temperatures the ef f ic iency drops rap id ly because o f the low specif ic heat o f the regenerator matr ix a t those low temperatures. Present1 y, the 1 owest no-1 oad temperature which can be achieved with these regenerati ve-cycl e re f r igera tors i s about 6 K, although Zimnerman and Sul l ivan [I]. have operated a low-pressure single-stage device a t 3.1 K wi th an upper temperature o f 8.6 K. Higher pressures are required when the upper end i s a t room temperature, and the resu l t ing higher heat capacity o f helium gas i n the system can become comparable to or greater than the matrix capacity for temperatures below about 10 K. Thus the overal l performance o f these re f r igera tors around 10 K and below i s determined pr imar i ly by the effectiveness of the regenerator i n transfer- r i n g heat from the incoming warm gas t o the outgoing cold gas.

* Work sponsored by F l i g h t Dynamics Laboratory9 Wright-Patterson A i r Force Base, Ohio.

The theoretical behavior of regenerators i s so complex that various simp1 i fying assumptions are made i n order to determine their behavior. Many of these assumptions are unreal- i s t i c for regenerators operating below about 15 K. Reliable experimental measurements of regenerator effectiveness have not been done a t temperatures below 20 K . Generally canpari- sons between different regenerators are done by substituting them i n an actual refrigerator and comparing the refrigerator performance. Such an approach i s not very satisfactory because ( 5 ) several variables may be changed simultaneously, ( i i the range of variables i s often limited, and ( i i i ) i t does not permit a separation of the various components which contribute to the regenerator ineffectiveness.

In this paper we describe a method for measuring regenerator ineffectiveness which allows the regenerator to be subjected to real is t ic operating conditions. These conditions can be varied independently i n order t o better understand the behavior of a regenerator. In addition, the method a1 lows for the separation of various components of regenerator inef fec- t i veness.

2. Definition and Theory of Regenerator Ineffectiveness

The effectiveness E of a heat exchanger i s usually defined as

d E = - = actual heat transfer rate maximum possible heat transfer rate'

The heat exchanger ineffectiveness i s given by

where dreg i s defined as (dm, - 0). The term dreg then represents the heat load on the region of interest (the cold expansion space) caused by the excess enthal py flow into th i s region. For a heat exchanger w i t h a given A we then find Qreg from

The introduction of the terms E and x are generally useful since they are either independent of the temperature difference fran one end of the heat exchanger to the other or a weak function of it . Thus the terms E and x are useful i n the characterization of a heat exchanger since they eliminate one of the variables. The effectiveness according to E q . (1) can be written as

where C i s the fluid specific heat a t constant pressure and i i s the mass flow rate. The subscripts h and c refer to the hot and cold streams, respectively, ' i n ' and 'out' refer to the two ends of the heat exchanger, and ( i C ) m i n refers to the minimum of $Cc and i h C h . When Chand Cc are temperature dependent, averaged values must be used, or the equation

can be used where H i s the fluid enthalpy.

For many regenerative heat exchangers the mass flow rate, temperature, and pressure are all functions of time. In t h a t case Eqs. ( 4 ) and ( 5 ) can s t i l l be used for the regenerator effectiveness b u t now E will be a function of time. In addition we must specify where ih and ic are to be measured since they will vary with position in the regenerator for the same instant of time, unless the gas volume i n the regenerator is negligible compared with the gas volume passing through i t . The most logical approach would be to choose 6 at the end of the regenerator next to the region of interest, i.e. the expansion space.

Usual1 y the instantaneous effectiveness is of l i t t l e intere$t. Instead a val ue averaged over a complete cycle is used to determine the average Qreg. The time averaged effectiveness is then

where v is the cycle frequency. In this expression nib = 0 during that part of the cycle when cold gas flows fran the expansion space through the regenerator. Likewise $in would be zero i n part of the cycle. An energy balance on the regenerator requires that the heat transferred fran the hot stream equals the heat transferred to the cold stream. Hence, the numerator in Eq. (6) could be replaced by the cold stream val ues.

We can simp1 ify E q . ( 6 ) whenever isothermal conditions exist for the gas inlets. In that case we let Th, i n = TU and Tc,in = TL, where TU and TL are independent of time. Equation ( 6 ) then becomes

and the ineffectiveness becomes

when the denominator refers to the hot stream. Additional terms arise in the numerator when the denominator refers .to the cold stream and when the specific heat in the two streams are not equal. The terms H(TU) and H(TL) are independent of time if (a) the pressure is constant, or ( b ) the gas is an ideal gas since H is independent of pressure for an ideal gas. If either of those conditions is true, then E q . (8) becomes

where m is the total mass of gas passing by the end of the regenerator. The AH product should be the minimum of %AHc and m@Hh, where

Except for the case where some of the gas is being withdrawn (e.g. for producing liquid). mh = mc.

For regenerators w i t h a finite gas volume and which are subjected to pressure variations, Eqs. ( 7 ) and (8) are st i l l n o t satisfactory. Consider the case where TU = TL in E q . (8). The denominator then becomes zero and we wou1.d expect that the numerator would

a1 5.0 become zero i n a manner such t ha t x remains f i n i t e as Tu--+TL. However as the gas pressure i n the regenerator changes it wi l ' l change temperature and therefore Th, out w i l l no t be the same as TL. Since the numerator i n Eq. (8) remains f i n i t e as Tu-+TL, the inef fect iveness approaches i: w . Thus i n such a regenerator there i s an addi t ional source o f enthalpy f l u x besides the normal f l u x due t o imperfect heat exchange i n a temperature gradient. This addi t ional enthalpy f l u x can be thought o f as another heat f low term', but i t s presence s t i l l ind icates imperfect heat exchange w i t h i n the regenerator. This addi t ional heat f low term can be understood by look ing a t the energy balance equations f o r the regener,- a t o r as shown below.

Gas Energy Balance Equation:

Heat Enthal py Energy t r a n s f e r change storage

h = Heat t rans fe r c o e f f i c i e n t OL = Surface area per u n i t length Tm = Mat r i x temperature Tg = Gas temperature

fi = Mass f low ra te H = Gas enthalpy

Ag = Cross-sectional area o f gas stream p = Gas density U = Gas in te rna l energy

Mat r i x Energy Balance Equation:

Heat Energy t r ans fe r storage

Am = Cross-sectional area o f matr ix P, = Density o f mat r i x Um = Mat r i x in te rna l energy Cm = Mat r i x spec i f i c heat

We note t h a t if the gas energy storage term were zero, i.e. A = 0, then a zero enthalpy grad ient and zero temperature d i f ference between Tg and T, sa?isfy both equations. How- ever when Ag # 0, the gas energy storage term w i l l be f i n i t e whenever the pressure i s a funct ion o f time. The f i n i t e gas energy storage term then forces a l l the other terms i n both e q u a t i 0 n s . b be f i n i t e even f o r Ty = TL. This e f f e c t becomes p a r t i c u l a r l y important a t low temperatures where the volumetr ic gas i n t e rna l energy, pU, i s s i g n i f i c a n t l y l a rge r than the matr ix vol umetric in te rna l energy, pmUm.

I n order t o proper ly u t i l i z e the previous de f i n i t i ons f o r E and 1, we must separate dreg i n t o two par ts

where dT i s the heat load on the expansion space r esu l t i ng from a temperature gradient i n 7 3

the regenerator and QU i s the heat load caused by the qas energy storage e f fec ts i n the regenerator, which occurs when pressure waves ex is t . Figure 1 shows the expected dependence o f Qreg on the quant i ty (&Imin AT, ,where AT i s the temperature d i f ference across the regenerator. We would expect t ha t QU would depend on the gbsolute.temperature and got on PT f o r small ATIT. Thus the in te rcep t i n Fig. 1 should give Qu, and QT would then be Qreg - Qu. We shal l then consider the regenerator ineffect iveness t o be t ha t associated only w i th QT. The average ineffect iveness ha i s j u s t the slope o f the s t ra i gh t l i n e passing through the p o i n t P i n Fig. 1. Note t ha t according t o Fig. 1, xa changes somewhat w i th AT. This behavior would occur when the thermal propert ies o f the gas and the matr ix are temperature dependent. I n addit ion, X a also depends on both TU and TL when TU - TL i s large. Thus it i s d i f f i c u l t t o character ize a regenerator i n a general sense w i th the term xa.

Figure 1. The dependence o f breg on the parameter (fiCIminhT. The in te rcep t gives buy the heat load on the expansion space due t o gas energy storage e f f ec t s i n the regenerator. The slope o f the dashed l i n e gives the average regenerator i ne f f ec t i ve - ness a t the po in t P.

We suggest tha t a regenerator i s be t te r characterized by using an incremental e f fec- t iveness E i or ineffect iveness i, which i s a funct ion only o f temperature. I n Fig. 1 the slope a t any po in t along the curve o f QT vs. (fiC)min AT i s then X i . For some f i xed TL, e.g. 6 K, X i i s then evaluated as a function of absol u te temperature and the average inef fect iveness f o r a regenerator operating between any TL and TU i s then

whenever Ch and Cc are constant. When Ch and Cc are not constant the X-axis i n Fig. 1 should be replaced by &CH(T~) - H(TL)] and the averaqe inef fect iveness becomes

A regenerator i s then best character ized by speci fy ing ~ u ( T ) and x i (T ) . O f course each o f these quan t i t i es w i l l a1 so be funct ions o f the heat capacity r a t i o , dead volume, pressure r a t i o , heat t rans fe r u n i t s , pressure, and phase angle.

3. Keasurement Methods

F igure 2 shows the schematics f o r three d i f f e r e n t methods o f measuring regenerator inef fect iveness. F igure 2( a) shows the dual regenerator technique developed by G i f f o rd [2,31. I n t h i s scheme the gas f i r s t f lows i n one d i r ec t i on through the regenerators and then reverses d i r ec t i on a f t e r a per iod o f t ime due to the ro ta ry valve. The pressure i s always constant i n such a scheme. The heat f l ow due t o inef fect iveness i n the k o i den t i ca l regenerators causes b o i l - o f f o f some cryogen. G i f f o r d used 1 i q u i d H2 fo r the cryogen and so the val ue f o r T was f i xed a t 20 K. For lower temperatures l i q u i d He can be used bu t then a semi-weak thermal coupl ing [41 must be placed between the heat exchanger and the hel ium bath so t h a t the heat exchanger temperature w i l l be h igh enough to prevent l i q u i f y i n g the working helium gas. The method used by G i f f o r d and i n t ha t proposed by Walsh e t a l . C41 f i xes the upper end o f the regenerators a t 77 K. The b o i l - o f f r a t e o f the helium gives the time averaged value o f the heat f low, which i s j u s t Qre , or the numerator o f Eq. (8). a measurement of Thyout would y i e l d the instantaneous A . T%S dual regenerator technique i s simple t o ca r ry out bu t i t s disadvantage i s t ha t the pressure i s constant, un l i ke t h a t which ex i s t s i n an actual r e f r i ge ra to r . Thus there i s no way to study the e f f ec t s due t o gas energy storage.

Valve

Test Regenerators

u H.E. & H.E.)

Surge Volume r-l Heater

Figure 2. Schematics o f three d i f f e r e n t methods o f measuri ng 'regenerator inef fect iveness: (a ) the dual regenerator method, (b) the surge volume method, and ( c ) the va r i - abl e expansion space method proposed here.

7 5

F i g u r e 2(b) shows a scheme used by Moore E51 whereby a s i n g l e regenerator i s c losed o f f by a surge volume. That approach al lows f o r a pressure wave w i t h i n t h e system b u t t he phase angle between the mass f l o w r a t e and the pressure wave i s f i x e d a t -90°. I n a rea l r e f r i g - e r a t o r the phase angle i s c lose t o 0". The gas energy storage c o n t r i b u t i o n to the e f fec - t i veness i s phase angle dependent and so the approach shown i n F ig . 2 (b ) s t i l l does no t a l l ow f o r a proper cha rac te r i za t i on o f regenerators w i t h f i n i t e gas volumes.

The method shown i n F ig . 2 ( c ) i s p r o p o s ~ d here as one which w i l l p rov ide a measurement o f the regenerator i ne f fec t i veness and QU under cond i t i ons which e x i s t i n an ac tua l r e f r i g e r a t o r . The proper phase angle between the pressure wave and t h e mass f l ow r a t e i s es tab l i shed by a v a r i a b l e expansion space, which i n t h i s case i s done w i t h a d i sp lace r . Heat supp l ied by the heater i s absorbed by the r e f r i g e r a t i o n which occurs w i t h proper phase angles. The regenerator i ne f fec t i veness causes a he l ium b o i l - o f f . Both heat exchangers (H.E.1. i n F i g . 2 (c) must be a t the same temperature b u t y e t they must be the rma l l y i s o l a t e d from each other. Th is approach al lows t h e r e f r i g e r a t i o n power t o be separated from t h e regenerator i ne f fec t i veness . F igu re 3 shows how t h i s separat ion occurs. I n t h i s f i g u r e t h e hea t exchangers are now i so the rma l i ze rs s ince they a l so serve t o b r i n g the gas passinp through them to the f i x e d temperature TL. The c y c l i c nature o f t he process means t h a t t he i so the rma l i ze rs must have a h igh heat cap.acity. I n order t o s imu la te S t i r l i n g and Vu i l l eumie r c ryocoo lers the continuous compressor and r o t a r y va l ve shown i n F ig . 2 ( c ) can be rep1 aced by a compressor which operates synchronously w i t h the d i s p l acer.

lsothermalizer Displacer

--1 - I I

I I I

h H - I I I I 1 I I I I I I I I I

Regenerator L - - - - - - J L, ------------ J

f t Qreg Qret

F i g u r e 3. System boundaries for t he energy balance equat ions which g i ve dreg and dref.

An energy balance on the l e f t i so the rma l i ze r shows t h a t

where h i s p o s i t i v e f o r gas f l ow ing to. t h e r i g h t . For r i p o s i t i v e T f TL because o f t h e regenerator i ne f fec t i veness . For m negat ive, T = TL f o r a p e r f e c t i so the rma l i ze r because the gas i s l e a v i n g the isothermal i z e r . Thus h[H(TL) - H ( T ) l i s zero when h i s neg- a t i v e , which mepns t h a t Ea..(16) i s the same as the numerator i n Eq. ( 9 ) i f we a l l o w f o r a s i g n change i n Qre . Thus Qreg i s a d i r e c t measure o f t he heat f l o w due t o regenerator i n e f f e c t i v e n e s s an% Oref i s a measure o f t he r e f r i g e r a t i o n ra te . The i n e f f e c t i v e n e s s i s then given as

An energy balance around the r i g h t isothermal i z e r and the displacer shows t h a t

where P i s the gas pressure a t the displacer

4. Descript ion

and dV i s the volume change.

of Present Method

A de ta i led schematic o f the low temperature par ts o f the proposed regenerator t e s t apparatus i s shown i n F ig. 4. Each isothermal izer i s connected to a helium bath v i a a va r i - able thermal l i n k , which w i l l be discussed i n more d e t a i l l a t e r . The var iab le thermal l i n k s al low the isothermal i ze r s to be maintained a t any temperature between about 5 K and 35 K w i t h varying amounts o f heat d iss ipa t ion (up to 2 W ) w i t h i n each isothermal izer. It would be possible to maintain isothermal izers #1 and #2 a t 77 K to provide a comparison w i t h r esu l t s from tes t s where the upper end i s a t 77 K. The heat d iss ipa t ion from isothermal izer

TO Compressor I

r;l vent

YY

Regenerator

Displacer 0

seal

To Main To Maln He Bath He Bath Rmrml Lhk u Ha Bath

Isolated He Bath

Figure 4. A schematic o f the low temperature art ?f the regenerator t e s t apparatus proposed here which shows the placement o f m, P, and T transducers.

77

#3 i s Qre and so improved reso lu t ion i s provided by using an i so la ted and shielded he1 ium bath to assorb the heat. A small amount o f heat suppl ied by a heater on each isothermal izer i s used for temperature p n t r o l . Isothermal izer #1 i s included so t h a t heat i npu t t o isothermal i z e r #2 gives Qre on the ho t s ide of the regenerator. The f a c i l i t y has the c a p a b i l i t y t o make (Tu - TL!/T~ << 1 i n order t o measure incremental inef fect iveness a t a l l temperatures between 5 and 35 K. I n add i t i on the temperature gradient can be reversed so t h a t Qreg i s measured by applying heat t o isothermal izer #3. A heat i npu t can be measured much more accurately than a helium b o i l - o f f r a t e caused by heas d iss ipa t ion . The change i n the sign o f dT/dx through zero also al lows us to determine Qu, the heat load due t o gas energy storage.

High speed sensors w i t h i n the gas stream a t several places a1 low us to measure the instantaneous T, P, and h f o r frequencies up to 20-30 Hz. Slow speed germanium thermometers on the isothermal izers are used to measure t h e i r temperature. Temperature cont ro l of the isothermal i zers i s c r i t i c a l t o the operat ion of the system. For measurements o f incremental i ne f fec t i veness at , for example, 10 K t he temperature d i f fe rence TU - TL may be about 1-2 K. For an ineffect iveness of 2% the d i f fe rence o f T - TL may be about 20-40 mK. Thus temperature cont ro l t o w i t h i n + 1 mK a t 10 K i s needed i n order t o measure the inef fect iveness t o w i t h i n 5%. Temperature cont ro l of Ssothermalizers #2 and #3 i s done i n the usual fashion using PID con t r o l l e r s and carbon thermometers as sensors. Isothermal izers #1 and #4 are t o be maintained a t the same temperature t o w i t h i n + 1 mK as t h e i r adjacent isothermal izer . However, the c a l i b r a t i o n accuracy o f germanium thermometers i s not good enough to y i e l d temperature d i f ferences less than 1 mK a t 10 K. We thus propose to use a d i f f e r e n t i a l gas thermometer. Each isothermal izer has a gas bulb o f nominally the same size. The c a p i l l a r y 1 ines from each p a i r o f gas bulbs go t o opposite sides o f a sens i t i ve d i f f e r e n t i a l pressure transducer. The room-temperature volume on one side o f each transducer i s made adjustable so t h a t the low-temperature volume to room-temperature volume r a t i o on each side o f the transducer can be made exact ly equal. The cor rec t volume i s determined by a c a l i b r a t i o n a t room temperature and a t l i q u i d He temperature. For a volume r a t i o o f 0.05 and a f i l l i n g pressure o f 1 atm a t roan temperature a 1 mK change i n temperature causes a pressure d i f f e r - ence o f about 2.7 Pa (0.02 t o r r ) . The room temperature volumes must be kept to w i t h i n about 2 O.Ol°C o f each other to keep d i f f e r e n t i a l pressure f l u c t ua t i ons less than the 2.7 Pa (0.02 t o r r ) . The e l e c t r i c a l outputs from the two d i f f e r e n t i a l pressure transducers are fed i n t o separate PID temperature con t ro l le rs , which i n t u rn con t ro l the heat inpu t to i sothermali- zers # 1 and #4.

The two vents shown i n Fig. 4 a l l ow the apparatus to measure the regenerator thermal conductance, heat capacity, and heat t rans fe r coe f f i c i en t . To measure the thermal conductance, heat i s appl ied to the r i g h t s ide o f the regenerator and the heat i s absorbed a t the 1 e f t s ide by the gas stream f lowing cont inuously through isothermal i z e r s # 1 and #2 and out through the vent. The heat capacity o f the regenerator ( inc lud ing t h a t due to adsorbed He and boundary l aye r He) i s measured by passing He gas cont inuously through the regenerator and out through the vent next t o the displacer. A f t e r the gas and regenerator are i n thermal equi l ibr ium, heat i s suddenly appl ied to the gas upstream f ran the regenerator to pro- v ide a quasi-step funct ion i n temperature vs. time. The gas temperature a t the regenerator i n l e t , Ti , and the ou t l e t , To, are then measured as a func t ion o f time. Figure 5 shows the typ ica l behavior f o r these two temperatures. The time l a g i s t ha t associated wi th t h e wave f r o n t passing through the regenerator. The t o t a l heat capacity o f the regenerator i s given by

m

where Cp i s the spec i f i c heat o f the gas a t constant pressure. The number of heat t r ans fe r un i t s , Ntu, i s def ined as

where h i s the heat t rans fe r c o e f f i c i e n t and o i s the heat t rans fe r surface area. The term Ntu can be determined from the temperature-time curves i n Fig. 5 from the r e l a t i o n

78

-

TIME

Figure 5. A temperature vs. time plot for the regenerator in le t and outlet temperatures af ter a quasi-step change i n temperature. The regenerator heat capacity and N t u are derived from the temperature behavior.

High values of N t u will be difficult to measure accurate1 y since aTo will tend to be smeared out i n practice. If a i s known, then h can be determined from E q . (20).

\

The two cri t ical parts of the regenerator t e s t apparatus are the variable thermal links and the isothermalizers. For the variable thermal links we propose to use gas gap heat switches like that described by Bywaters and Griffin L61, in which heat i s transferred between several concentric copper cylinders by helium gas in the gaps. Figure 6 shows the arrangement. For pressures below about 130 Pa (1 t o r r ) the thermal conductance across the helium gap i s proportional to the gas pressure. The variable pressure i s achieved by the use of a molecular sieve adsorbant which can be heated to some temperature above 4.2 K 171. When the molecular sieve i s held a t about 20 K by a heater, i t adsorbs l i t t l e He gas and so the pressure remains high. As the heat to the molecular sieve i s decreased, the temperature of the sieve decreases and the resultant adsorption causes the gas pressure to decrease. In1 the full 'off ' condition the only heat leak i s through the fiberglass-epoxy support tube. Below 35 K radiation i s negligible for polished surfaces. Measurements of the electrical conductance between the two ends would show whether the copper cylinders are touching each other, For long l i f e (i.e., time greater than a few months) it i s necessary to imbed a t h i n foil of Cu-Ni or stainless steel within the fiberglass-epoxy tube i n order to prevent the diffusion of He gas through the tube walls.

The isothermalizers are the other set of cr i t ical components. Their design i s not easy because of the following four requirements: a) high heat capacity, b) high Ntu, c ) small pressure drop, and d ) small gas volume. The f i r s t two and the l as t two requirements generally work in opposition to each other. The temperature difference between the gas and the isothermalizer a t the outlet, ATo, i s related the inlet temperature difference, AT^, by

- 4 Thus for AT0/ATi = 10 , we need N t u = 9.2. Wi th such values the gas outlet temperature i s w i t h i n 1 mK of the isothermal i zer even for aTo = 10 K. To calculate the required heat capacity of the isothermalizers we estimate the total heat to be stored i n each cycle.

79

Figure 6. Drawing of a gas-gap heat switch used for the variable thermal link. The helium gas pressure between the concentric copper cylinders i s control led by the temperature of the molecular sieve adsorbent.

For a refrigeration rate of dref the total heat to be stored i s d r e f ~ . For a total heat capacity of C i the temperature fluctuation 6T i n the isothermalizer 1s then given by

For a fixed (jref the largest C i i s needed a t the lowest v. For Gref = 1 W and C i = 50 J/K we get 6 T = 5 mK a t v = 4 Hz b u t 6T = 1 mK a t v = 20 Hz. Further reductions i n 6T can be achieved by staging. A total heat capacity of a l eas t 40 J/K between 5 K and 40 K can be 5 achieved w i t h the use of 50 cm3 of Pb and 100 cm of He gas a t a pressure of 2 MPa (20 am). The dimensions of the Pb and the He space must not be larger than the thermal pene- trat ion depth a t the working frequency range, which a t 10 Hz i s about 1 mm i n Pb a t 35 K and 50 v m i n He below 20 K.

Design c a l c u l a t i o n s show t h a t an Ntu va lue o f 9.2 can be achieved w i t h pressure drops genera l l y l e s s than 1% o f the pressure f o r pressures greater than about 1 MPa (10 atm) and temperatures up t o 35 K. A f l ow r a t e o f 1.5 g/s can be accomodated w i t h a gas volume i n t h e i so the rma l i ze r o f about 0.3 cm3. We propose t o use an annular gap o f about 28 pm thickness, 3 1 cm e f f e c t i v e width, and 3.5 an long t o prov ide the necessary surface area.

5. Concl usions

The regenerator t e s t apparatus proposed here should be capable o f measuring regenerator i n e f f e c t i v e n e s s i n the range 4-35 K w i t h v a r i a b l e phase angle between the mass f l ow r a t e and the pressure wave. Thus the cond i t i ons can be made t o match very c l o s e l y t h a t which occurs i n an ac tua l r e f r i g e r a t o r . I n add i t i on , t he apparatus would be capable o f separat ing o u t a l l o f t he sources c o n t r i b u t i n g t o the o v e r a l l i n e f f e c t i v e n e s s and show t h e i r dependence on ya r ious parameters.

6. References

C l l Zimmerman, J. E. and Su l l i van , D. B., A M i l l i w a t t S t i r l i n g Cryocooler f o r Temperatures Below 4 K, Cryogenics - 19, 170-171 (1979).

[21 G i f f o rd , W. E., Bas ic I n v e s t i g a t i o n o f Cryogenic R e f r i g e r a t i o n Methods, A i r Force F l i g h t Dynamics Laboratory, Wright-Patterson A i r Force Base, Ohio, 45433, AFFDL-TR-68-61, (1968).

[31 G i f f o r d , W. E. and Acharya, A., Low-Temperature Regenerator Test Apparatus, Adv. Cry. Eng. - 15, 436-442 (1970).

C41 Wal shy P. J . , Wade, L. A., and Chaudhary, T. M., Cryoregenerator Heat Capacity Enhancement, A i r Force F l i g h t Dynamics Laboratory, Wright-Patterson A i r Force Base, Ohio, 45433, AFFDL-TR-80-3034 (1980).

C51 Moore, R. W., Jr., An I n v e s t i g a t i o n o f Thermal Regenerators f o r t h e 4 t o 20 K Temperature Range, A i r Force F l i g h t Dynamics Laboratory, Wright-Paterson A i r Force Base, Ohio, 45433, AFFDL-TR-70-64 ( 1970).

C61 Bywaters, R. P. and G r i f f i n , R. A., A Gas-Gap Thermal Switch f o r Cryogenic App l ica t ions, Cryogenics - 13, 344-349 (1973).

C73 Cast1 es, S., p r i v a t e connnundcation (1980).

SERVICEABLE REFRIGERATOR SYSTEM FOR SMALL SUPERCONDUCTING DEVICES

Ralph C. Longsworth

A i r Products and Chemicals, I nc . APD Cryogenics

P.O. Box 2802 Al lentown, PA 18105

The design of a r e f r i g e r a t i o n system f o r c o o l i n g small superconduct- i n g devices such as Josephson Technology computers which can be se rv i ced w h i l e t h e superconducting device remains i n a c o l d ope ra t i ng s t a t e i s described. The re f r i g e r a t i o n s y s t ~ m inco rpo ra tes new (pass ive) thermal coup l ings which pe rm i t t h e Disp lex expander and JT c i r c u i t t o be warmed up, t h e expander serviced, and t h e JT c i r c u i t purged, w h i l e he l ium b o i l s of f a t a r e l a t i v e l y low r a t e f rom t h e ba th t h a t i s c o o l i n g t h e superconduc t i ng device.

When connected t o t h e 1.4 L t e s t pot , a warmup-purge-cooldown t e s t was run. Helium was i n i t i a l l y condensed i n t h e p o t by t h e r e f r i g e r a t o r a t t h e r a t e of 0.3 L l h r then the r e f r i g e r a t o r tu rned o f f . I t took 7 hours t o warm up and purge the r e f r i g e r a t o r , f o l l owed by 3 hours t o coo l back down t o 4K. Jus t a few minutes be fo re reach ing 4K, he l ium stopped ven t i ng and s t a r t e d t o be drawn i n t o t h e t e s t p o t i n d i c a t i n g t h a t s l i g h t l y more than 1.4 L o f he l ium was used d u r i n g t h e s e r v i c e cyc le .

Key Words: Cryogenic r e f r i g e r a t o r ; 1 i q u i d he1 ium; Josephson technology; serv iceab le ; superconducting device; thermal coup1 ing .

1. I n t r o d u c t i o n

As a p p l i c a t i o n s f o r small superconducting dev ices become commercial, t he re w i l l be a need f o r c losed-cyc le r e f r i g e r a t o r systems t h a t pe rm i t such a dev ice t o be main ta ined i n a c o l d ope ra t i ng s t a t e cont inuous ly , i n c l u d i n g per iods when t h e r e f r i g e r a t o r i s off and being serviced. A t t h e 1977 meeting o f t h i s group, van der Hoeven and Anaker [l] desc r i bed t h e r e f r i g e r a t o r requirements f o r a Josephson technology data process ing system w i t h p a r t i c u l a r emphasis on r e l i a b i l i t y and s e r v i c e a b i l i t y . A t t h e same meet ing Longsworth [2 ] descr ibed means o f r e f r i g e r a t i ng small superconduct i ng devices, whi 1 e p r o v i d i n g i so l a- t i o n f rom v i b r a t i o n and e lec t romagnet ic "noise".

Th i s paper descr ibes t h e design o f t h e r e f r i g e r a t o r system developed f o r t h i s app l i ca - t i o w which i nco rpo ra tes new thermal coup l ings t h a t pe rm i t t h e expander and JT c i r c u i t t o be warmed up, t h e .expander serviced, and t h e JT c i r c u i t purged, w h i l e he l ium b o i l s o f f a t a r e l a t i v e l y low r a t e from t h e ba th t h a t i s c o o l i n g the superconducting device. A companion paper t o t h i s r e p o r t s on t e s t r e s u l t s which show reasonably good agreement between c a l - c u l a ted values and ac tua l performance.

2 . System D e s c r i p t i o n

The bas i c r e f r i g e r a t o r system i s an APCI Model CS-308 two-stage ~ i s ~ l e x ~ r e f r i g e r a - t i o n system w i t h a JT l oop t o p rov ide r e f r i g e r a t i o n a t R/ 4.2K: R e f r i g e r a t i o n capac i t y r a t i n g s a r e 2.5 wa t t s a t 4.5K p lus 2 wat ts a t < 20K and 10 wat ts a t <77K. The o r i g i n a l JT c i r c u i t which was designed f o r c o o l i n g samples a t 4.5K has been mod i f i ed t o pe rm i t remote c o o l i n g o f a l i q u i d he l ium dewar us ing t h e f i r s t and second stage r e f r i g e r a t i o n t o i n t e r c e p t heat l e a k and 4K r e f r i g e r a t i o n t o recondense he l ium b o i l - o f f .

The compressor i s a three-stage o i l - l u b r i c a t e d rec ip roca t ing u n i t t h a t suppl ies gas a t 17.7 atm and receives gas from the expander a t 4.1 atm and from the JT loop a t 0.8 atm. I n add i t i on t o t he compressor, the standard CS-308 system includes a two-stage Disp lex expander, the three JT heat exchangers and adsorbers, manually ad justab le JT valve, a 4K heat s ta t ion , a hydrogen f i l l e d thermal swi tch connecting the second and 4K heat s t a t i ons ( f o r cooldown), a r a d i a t i o n s h i e l d attached t o the f i r s t stage, and the gas b a l l a s t tank.

F igure 1 i s a p i p i ng schematic o f the present system. The helium po t simulates the helium dewar i n the f i n a l system and serves the purpose o f enabl ing the performance o f the thermal coupl ing system and thermal losses dur ing se rv ic ing o f the r e f r i g e r a t o r t o be measured. The bas ic CS-308 system i s thus modi f ied by adding o r subs t i t u t i ng the f o l l i ng components :

a) thermal coupl ings 1 and 2 are added

b) a he1 ium condenser replaces t he standard 4K heat s t a t i o n

c ) gas l i n e s a re added on the h igh pressure JT l i n e t o t ranspor t r e f r i g e r a t i o n the separate thermal coupl ings and condenser

d) a bypass va lve f o r cooldown i s subs t i tu ted f o r the H2 thermal swi tch

e) a separate feed l i n e f o r precool ing and l i q u i f y i n g make-up helium i s added

OW-

0

f ) se l f - sea l ing gas coup1 ings on the JT c i r c u i t are added t o permi t i t t o be i so l a t ed

g) purge 1 ines on the JT c i r c u i t are added t o permi t i t t o be f lushed

h) add i t i ona l temperature sensors and heaters are added t o permi t con t ro l o f r a p i d warmup f o r serv ice and moni tor ing o f performance

i ) an absolute pressure JT bypass pressure regu la to r f o r automatic con t ro l o f the l i q u i d temperature i s added

j) an experimental 1.4 L helium po t connected t o a separate warm f lange by a support tube, having i t s own r a d i a t i o n shield, i s added.

Re f r ige ra t ion systems s i m i l a r t o t h i s are present ly being b u i l t f o r coo l ing superconduct ing magnets and spectroscopy samples i n which the r e f r i g e r a t o r and magnet o r spectroscopy sample holder are i n separate enclosures connected by a f l e x i b l e hose conta in ing the gas l i n e s .

An important design c r i t e r i a t h a t i s adhered t o i s t h a t the c i r c u l a t i n g r e f r i g e r a n t helium, which must be kept f ree of contaminants, i s never mixed w i t h the helium t h a t i s i n contact w i t h the superconducting device.

3. Re f r ige ra t ion

F igure 2 shows the f i r s t and second stage capaci ty o f the DE-208L Disp lex expander before adding the JT loop. The width of the capaci ty curve r e f l e c t s a small interdepen- dence o f the heat load-temperature r e l a t i o n f o r the two stages and manufac'turing var ia- t i ons from u n i t t o u n i t . When the JT loop i s added there are add i t i ona l heat loads imposed, p r i m a r i l y heat exchanger losses on the f i r s t stage, and both heat exchanger losses and enthalpy di f ference loading imposed on the second stage.

I n order t o determine the complex r e l a t i onsh ip between f i r s t , second, and JT heat s t a t i o n heat loads versus temperatures, ca lcu la t ions were made o f r e f r i g e r a t i o n losses i n the JT heat exchangers assuming the fo l lowing mat r i x o f operat ing condi t ions.

. JT pressure a t the compressor of 0.8 and 1 atm corresponding t o 0.23 and 0.29 g/s f 1 ow respec t i ve ly

. second stage temperatures o f 16 and 20K

. f i r s t stage temperatures of 50, 65 and 80K 83

ABSOLUTE PRESSURE REGULATOR

GAS BALLAST TANK

BY PASS - VALVE

HE - HEAT EXCHANGER T. - TEMP SENSOR H 7 HEATER P - PRESSURE GAUGE - - - --,I

RADIATION SHIELDS L BEFRIGERATOR I HELIUM POT tlELLUM POT 1 VACUUM

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

Figure 1 . Refrigerator/Dewar P i p i n g Schematic

SEC

ON

D S

TAG

E C

AP

AC

l TY

WA

TTS

FI

RST

STAG

E C

APA

CIT

Y -W

ATT

S

Heat exchanger and enthalpy losses ca lcu la ted a t these condi t ions are subtracted from the minimum capaci ty curve of Figure 2 t o ob ta in the performance map shown i n Figure 3. F igure 3 p l o t s f i r s t and second stage r e f r i ge ra t i on ava i l ab l e versus f i r s t stage tempera- t u r e f o r T2 = 16 and 20K and P = 0.8 and 1 atm. Re6r igerat ion produced a t the 4K heat s t a t i o n i s a lso l i s t e d . Use ofTthe curve i s i l l u s t r a t e d by assuming a 10 wat t heat load on the f i r s t stage, second stage temperature o f 16K, and 1 atm JT r e t u r n pressure. F i r s t stage temperature i s 56K, second stage capaci ty i s 2.5 watts, and 2.3 wat ts i s ava i lab le a t about 4.3K. There i s a small pressure drop loss i n the low pressure s ide of the JT loop, thus the sa tu ra t ion pressure a t the 4K heat s t a t i o n i s h igher than the r e t u rn pressure a t the compressor.

FIRST STAGE TEMPERATURE - K

Figure 3. CS-308 Ava i lab le Re f r ige ra t ion o f F i r s t and Second Stages Versus F i r s t Stage Temperature. Capacity o f the 4 K stage i s tabulated f o r second stage temperatures o f 16 K and 20 K and r e t u r n pressures o f 0.8 and 1.0 atmospheres. The dashed l i n e s a t 56 K and 2.5 W r e f e r t o the example given i n the t ex t .

Figure 4 p l o t s the r e f r i g e r a t i o n ava i lab le a t the second and 4K heat s ta t ions as a funct ion o f second stage temperature f o r 0.8 and 1 atm JT r e t u r n pressure assuming a 15 wa t t heat load on t he f i r s t stage. This p l o t shows t h a t i f there i s a 2 wa t t heat load on the second stage and i f the JT r e t u r n pressure i s 1 atm, then the second stage tempera- t u r e i s 16.0K and 2.3 wat ts i s ava i l ab l e a t the 4K heat s ta t ion .

These ca lcu la ted. capac i t i es neglect conduction and r ad i a t i on losses which can t y p i - c a l l y be kept r e l a t i v e l y small by ca re fu l design.

4. Cooldown

I t takes about 2-112 hours t o cool down the standard CS-308 r e f r i g e r a t o r and about 3 hours t o cool down t he r e f r i g e r a t o r and thermal couplings. I f add i t i ona l mass i s i n good thermal contact w i t h each o f the heat s ta t ions then the f i r s t stage can cool 2.3 Kg o f copper per hour and the second stage can cool ,9 Kg o f copper per hour. For the con f igu ra t ion shown i n Figure 1, there i s poor thermal coupl ing o f the helium po t w i t h the heat s ta t ions, thus the r e s t o f the system cools down much f a s t e r than the he1 ium pot.

I n a dewar w i t h a l a rge neck tube i t i s poss ib le t o provide means f o r the f i r s t and second stage thermal b a f f l e s t o cool down the he1 ium po t by convection. Another opt ion t o cool the hel ium po t f a s t e r i s t o feed hel ium i n through the l i que fac t ion c i r c u i t and vent i t through the top o f the neck tube. A small c i r c u l a t i n g pump could be used t o e l im ina te hel ium losses. Such an arrangement can be used as an a l t e rna te means t o cool the leads from a superconducting device such as a magnet.

5. Thermal Couplings

F igure 5 shows a cross sect ion view o f the f i r s t stage thermal coupl ing which trans- f e r s heat by convect ion from the base t o the f inned tube i n the top which i n t u rn i s cooled by f l ow from the JT loop. I n the event t h a t the r e f r i g e r a t o r shuts down, the thermal coupl ing stays co l d because gas c i r c u l a t i o n i n the JT loop stops. Helium w i l l t y p i c a l l y b o i l o f f a t a r e l a t i v e l y low r a t e during, t h i s time, i :e., 0.1 t o 0:5 L lh r .

The pr imary purpose of the convect ive thermal coupl ing i s t o enable the JT loop t o be warmed t o room temperature and purged w i thou t s ign i f i can t1 y increas ing the he1 ium bo i 1 - o f f r a t e . Th is i s poss ib le because the helium s t r a t i f i e s i n s i de the thermal coupl ing when the top i s warmer than t he bottom so heat i s on ly t rans fe r red by conduction i n the gas and housing from top t o bottom.

High heat t r ans fe r ra tes w i t h modest temperature d i f ferences and r e l a t i v e l y compact design are achieved by separat ing the ho t r i s i n g gas from the co l d down-flowing gas and l oca t i ng extended surface heat exchangers i n such a way as t o promote convect ive c i r cu l a - t i o n o f the gas. Since the convect ive heat t ransfer r a t e i s propor t ional t o pressure, i t i s best t o keep the pressure a t the h igh pressure s ide o f the system. I n o ther appl ica- t i o n s the thermal convect ive coupl ing can be used as a va r iab le thermal conductance device by ad jus t ing the gas pressure.

The hel ium condenser shown i n cross sect ion i n Figure 6 i s designed w i t h the JT stream coo l ing a f i nned tube i n the top chamber which i s separated from the main 1 i q u i d helium po t by s tandof f tubes. These standoff tubes are designed t o have a low heat leak t o the l i q u i d heli'um when the r e f r i g e r a t o r i s o f f and when the JT loop i s warmed up t o be purged.

HELIUM PRESSURIZATION

ImT - J T FW

(HEAT OUT)

FINNED TUBE GAS COOLER

- (HEAT IN)

Figure 5. Convective Thermal Csupling f o r F i r s t and Second Heat Statl'ons.

The JT loop gas coupling, the convective thermal couplings,and the helium condenser a re a l l (passive) thermal disconnect switches t h a t serve t o thermal ly i s o l a t e the r e f r i g e r - a t o r from the l i q u i d hel ium bath when the r e f r i g e r a t o r shuts down and when i t i s being serviced.

- FINNED TUBE HELIUM CONDENSER

FLANGE To HELIUM POT

Figure 6. Helium Condenser.

6. Service

When one reviews f a i l u res i n small c losed-cycle coolers, i t i s seen t h a t there are very few cases o f leaks developing i n the vacuum chamber i f a l l the j o i n t s are proper ly made and c a r e f u l l y leak checked. It i s thus s u f f i c i e n t t o leave the p i p i ng i n place and no t break vacuum if a) the expander can be disassembled f o r serv ice and, b) i f the JT loop can be warmed and purged o r poss ib ly f lushed.

S e r v i c i n g a c o l d expander i n l e s s than an hour has been demonstrated. Th is was done i n t h e f o l l o w i n g sequence:

a ) d isconnect gas l i n e s t o t h e expander and l e t gas o u t

b) remove va l ve motor ( t h e d i s p l a c e r s tays i n place)

c ) screw a spec ia l t o o l w i t h purge gas f l o w i n g through i t i n t o t h e t o p o f t he d i s - p l a c e r

d) w i t h purge gas f lowing p u l l t he d i s p l a c e r i n t o a sealed p l a s t i c bag

e ) warm d i s p l a c e r and s e r v i c e i t w h i l e keeping a cover on the c y l i n d e r w i t h a smal l purge f l o w o f he l ium

f ) reassemble t h e expander i n the reverse o rde r

g ) evacuate and purge t h e expander severa l t imes before reconnect ing t h e gas l i n e s .

The most common source o f t r o u b l e w i t h t h e JT loop i s blockage due t o f reeze-out o f contaminants. Se rv i c ing o f t h e JT loop thus u s u a l l y cons i s t s o f warming i t t o room temp- e r a t u r e and purg ing o r f l u s h i n g o u t contaminants. Th i s i s done i n t h e f o l l o w i n g sequence:

a) Disconnect t h e gas coup l ings t o i s o l a t e t h e JT loop.

b) Warm t h e hea t exchangers w i t h a back f l o w o f gas through the low pressure s ide o f t h e h e a t exchangers, ven t i ng o u t through the bypass l i n e .

c ) Simultaneously warm t h e r e f r i g e r a t o r and adsorbers by t u r n i n g on t h e heaters .

d) Warm t h e tops of t h e thermal coup1 ings and he1 ium condenser by f l o w i n g purge gas i n through the bypass l i n e and o u t through t h e h igh pressure l i n e s . A d j u s t t h e heaters on adsorber 3 and the r e f r i g e r a t o r w h i l e doing t h i s t o keep the"Keat exchangers a t room temperature.

e) Reconnect gas 1 ines a f t e r warmup and purg ing i s complete.

L i q u i d he l ium b o i l o f f i s minimized by keeping t h e tops of t he thermal coup l ings c o l d as l ong as p o s s i b l e - thus t h e r e f r i g e r a t o r , heat exchangers,and adsorbers a re warmed up f i r s t . Since t h e tops o f t h e thermal coup l ings a re r e l a t i v e l y l i g h t they can be warmed f a i r l y q u i c k l y .

When t h i s -procedure was f i r s t t r i e d on t h e u n i t w i t h 1.4 L o f l i q u i d i n t h e t e s t pot , i t took a l ong t ime t o a d j u s t t h e hel ium purge f low r a t e and s e t t h e heaters. The t o t a l t ime f rom shutdown t o r e t u r n t o 4.2K took 10 hours. F i v e hours were spent warming the r e f r i g e r a t o r t o room temperature. With the r e f r i g e r a t o r warm and t h e thermal coup l ings co ld , the b o i l - o f f r a t e was 0.13 l i q u i d L/hr . The nex t 2.1 hours were spent warming the tops o f t h e thermal coup l ings and c o l d adsorber du r ing which t ime t h e b o i l - o f f r a t e i n - creased t o 0.24 l i q u i d L l h r . Three hours were then requ i red t o coo l back down t o 4.2K.

Helium stopped ven t i ng and s t a r t e d t o be drawn i n t o the feed l i n e about 15 minutes before t h e system reached 4.2K i n d i c a t i n g t h a t t he e n t i r e warmup and purge consumed s l i g h t - l y more than 1.4 L o f l i q u i d helium. The u n i t then recondensed he l ium a t a r a t e i n excess o f 0.3 L l h r .

The warmup and purge p e r i o d can be reduced from 7 hours t o about 2.5 hours w i t h some small changes i n i ns t rumen ta t i on and p r a c t i c e .

7. Summary

Many thousands o f small cryogenic r e f r i g e r a t o r s t h a t operate a t about 10K on the G i f f o r d McMahon and modi f ied Solvay cyk les are i n use today. The most common small 4K (systems) i n use today use these 10K r e f r i g e r a t o r s as a base t o achieve 4K by adding a JT loop. While i t essen t i a l l y doubles the complexi ty o f the r e f r i ge ra to r , i t i s s t i l l poss ib le t o achieve very h i gh r e l i a b i l i t y , as described by Higa [3], i f care i s taken i n the construct ion, operation, and maintenance o f the un i t s .

The r e f r i g e r a t o r described i n t h i s paper extends the app l i ca t ions f o r t h i s type o f r e f r i g e r a t o r t o coo l ing small superconducting devices such as Josephson technology data processors, which i n add i t i on t o h igh r e l i a b i l i t y , requ i re an a b i l i t y t o serv ice the r e f r i g e r a t o r wh i l e the superconducting device continues t o operate. This has been accom- p l i shed by means o f thermal convect ive couplings which pass ive ly disconnect the r e f r i g e r a t o r from a he1 ium dewar when the r e f r i g e r a t o r shuts down and permi t the r e f r i g e r a t o r c i r c u i t t o be warmed and purged.

Wahup, purging, and cooldown back t o 4.2K has been demonstrated w i t h a t o t a l loss o f about 1.4 L o f 11qu1d helium.

References

[I] van der Hoeven, B. J. C. Jr., and Anacker, W. "Ref r igerator Requirements fo r Po ten t ia l Josephson Data Processing Systems", NBS Special Pub1 i c a t i o n 508, p. 213, A p r i l 1978.

[2] Longsworth, R. C., "Concepts f o r Cooling Small Superconducting Devices Using Closed- Cycle Regenerative Refr igerators" , NBS Special Pub l i ca t ion 508, p. 45, A p r i l 1978.

[3] Higa, W. H. and Wiebe, E., "One M i l l i o n Hours a t 4.5 Kelvin", NBS Special Pub l i ca t ion 508, p. 99, A p r i l ,1978.

PERFORMANCE OF A 1 watt 4K CRYOSYSTEM SUITABLE FOR A SUPERCONDUCTING COMPUTER

E. B. Flint, L. C. Jenkins, R. W. Guernsey

IBM Thomas J. Watson Research Center Yorktown Heights, New York 10598

Superconducting computers of the future will require closed cycle refrigeration systems capable of removing 0.5-10, watts at 4K. The refrigerators must be reliable and the interface to the cryostat must allow continued computer operation during refrigerator shutdown. Such a system has been designed by R. Longsworth and one unit has been built by Air Products. This cryocooler is being operated now in a test configuration that allows detailed study and analysis of the design. The results of these studies, including cool-down times, refrigeration capacity, and interface performance will be discussed and compared with the design calculations..

Key words: Convective heat transfer; cryocooler; heat transfer by evaporating and condensing helium; helium refrigerator; superconducting computer; thermal couplings; thermal isolation.

1. Introduction

1.1 Desired cryosystem characteristics

In the future, one may be able to buy an extremely high-speed computer whose logic and memory chips opente at 4.2K. This might be packaged in a cube 9 cm on a side and would be immelsed in liquid helium where it would dissipate about 10W of heat 111. At this time, it is of interest to evaluate cryosystems that might be appropriate for such a machine. A general view of the requirements, as anticipated at this time, is given elsewhere in these proceedings.

Medium to large size superconducting computers are likely to need from O.SW to 10W of refrigeration at 4.2K. The cryosystem should provide continued computer operation in the case of refrigerator shut down, and it also should allow one to service the computer without disturbing the refrigerator. A semi-annual service interval .and a mean time between failure in excess of two years would be desirable. High reliability and fool proof service procedures are paramount. This paper documents the performance of a refrigerator and thermal interface designed to support a lW, medium sized superconducting computer.

1.2 The interface concept

A study was done by R. Longsworth of Air Products, in which he developed a design for a complete cryogenic system using existing refrigeration technology which meets the requirements of a superconducting computer. The design he developed uses the Air Products CS-308, 4K refrigerator to which was added a thermal interface which couples the dewar and bath to the refrigerator. A schematic of the system is shown in Fig. 1. The interface provides good thermal coupling between the refrigerator and the dewar structure and processor during normal operation and good isolation when the refrigerator is warm for servicing. A companion paper by.R. Longsworth [2] describes the design in some detail. Following the study, Air Products [3] built the system in a test configuration that now is undergoing extensive tests at IBM. While this work is far from complete (for example, there are few reliability data) there are data characterizing the refrigerator capacity and interface performance. In this paper we

9 3

BY-PASS VALVE \

- 1 1 0 CABLES

LIQUID HELIUM BATH RADIATION? sti,.Ltl 1 L - IQl -

- SUPPERCONDUCTING MAGNETIC SHIELD

/COMPUTER LOGIC 8 MEMORY CARDS

Figure 1: Shown is a schematic representation of a superconducting computer cryosystem with a large liquid helium reservoir for thermal ballast. The cryosystem under test has a 1.4 liter pot and includes thermometers TI-T8 and heaters HI-H6.

describe the cryogenic system and instrumentation used in the experiments, the performance data,and the areas where we believe the data indicate more work, both documentation and engineering, is required.

2. Description of the cryogenic system

In this section we describe the system includilig the refrigerator, the thermal interface, and the instrumentation used in the experiments. Figure 1 shows the refrigerator and interface with a dewar which might be suitable for a processor. The large liquid helium bath provides thermal ballast during periods when the refrigerator is shut down. In the test configuration a 1.4 liter helium pot replaces the bath shown.

The 4K refrigerator includes the 2 stage expander which provides refrigeration at 60K and 15K and a Joule-Thompson (JT) loop which produces the refrigeration at 4K. There is a by-pass line around the low pressure side of the third JT heat exchanger which is used for the cool-down from room temperature. The JT valve can be adjusted from the top of the cryostat so as to compensate for different load conditions.

The thermal interface consists of two gas thermal diodes and a condenser. The refrigeration available at the first and second expanders is delivered to the tops of the diodes via the JT gas stream. The diodes conduct by gas convection from bottom to top. When the temperature gradient is reversed the gas stratifies, providing relatively good thermal isolation. The condenser is similar to the diodes except that it is a two phase device. In normal operation (top cold) the boil-off gas from the bath is condensed on fins cooled by the JT stream and then runs back into the ,bath. Thermal isolation is, once again, provided when the top is warm because the helium gas in the support tubes stratifies.

The system performance has been studied using instrumentation which includes thermometers and heaters at each of the refrigerator stages. At the first and second stages silicon diode thermometers (numbers: TI , T2, T5, T6, T7, T8) are used to monitor the refrigerator temperature, the heat station temperature at the base of the thermal diode and the temperature of JT gas as it returns from the diode. The thermometer, T2, was calibrated between 4K and 100K. All other silicon diode thermometers have four point calibrations (4.2K, 20.4K, 77.4K and 273.2K) which were used to scale the T2 temperature vs. voltage curve to obtain temperatures from the measured voltages. The relative accuracy of the thermometers is probably less than 1K at 70K and less than 0.1K at 20K. Thermome- ters T3 and T4 are calibrated germanium resistance thermometers located inside the JT stream. The manufacturer quoted accuracy of these thermometers is 0.005K. The temperature of the bath was measured by monitoring the vapor pressure and using the 1958 helium temperature scale.

There are six heaters (HI 5 60 watts, all others 5 20 watts) which were used to simulate heat loads and to warm the refrigerator for purging the JT loop and servicing the refrigerator.

3. Refrigerator performance

3.1 Capacity

The refrigerator was delivered with the thermal interface, test pot,and instrumentation in place. In order to determine the total cooling capacity at each stage, one must know the parasitic heat loads that result from these structures. We begin with a discussion of the methods used to measure these. At the f i s t and second stages heat leaks are intercepted at the refrigerator stage and the coupling station. Those for the two couplings were determined from the temperature increase in the JT gas stream and the JT flow rate, and these values were confirmed by calculations of heat leak from the support sFructures and from radiation. An additional check was made by extrapolating the heat transport characteristic, Q(AT), to AT = 0 for the couplings. Q is the.heat transported across the coupling and AT is the temperature difference between the JT gas (either entering or leaving the coupling) and base of the coupling. In the case of the first stage data, the parasitic loads are sufficiently large and the Q(AT) characteristic sufficiently non-linear that it is possible to conclude only that there were no inconsistencies in the data. For the second stage coupling the loads (parasitic and applied) are smaller and the three determinations of the parasitic load were in reasonable agreement.

The parasitic load to the JT stream from the test pot was measured by noting the boil-off rate with the JT and the pot thermally decoupled.. The result is in reasonable agreement with the calculated value. The parasitic load to the pot as indicated by the Q(AT) characteristic for the condenser extrapolated to AT = 0 was substantially higher and the difference is not understood. Offsets of - 0.035K and - 0.008K in the JT thermometers up and down stream of the condenser relative to the bath temperature would be required to explain the discrepancy. The parasitic loads at the three refrigerator stations are given in Table 1.

95

A complete map of the refrigeration capacity has not yet been made. The capacity a t 4K, however, has been measured with the first and second stages loaded as they would be in an actual computer cryostat. Table 1 shows the results of these tests. With 9.9 watts applied to the first stage (22W total) and 0.5 watts to the second stage (1W total) the minimum temperature in the pot was determined with applied loads of 1.OW and 1.5W to the bath. At 1.03W total dissipation in the bath the minimum temperature, Tmi, was 3.82K and at 1.53W total dissipation Tmi, was 4.24K. A summary of data is given in Table 1.

The refrigeration at 4.2K is 25% less than anticipated but the refrigeration at the other stages is near the design values.

Table 1

Parasitic Load (W) I Applied I Total I stage I

Station

Temperatures at

Couplings and Pot

1st 2nd JT

86K (est) 15.5 (est)

3.82 (.57 atm at compressor)

Refrig- erator1

84K (est) 17.8 (est)

4.24 (0.92 atm at compressor)

5.5 0.2 0

Notes: 1) heat load in this column are calculated; 2) heat loads in this column are measured and checked with calculation;

~ o u p l . ~

3 ) est. = estimated temperature. Temperature is determined from one measured temperature and the coupling heat transfer characteristic.

6.5 0.24 0.03

3.2 Reliability

Heat Load

It is important that a superconducting computer cryosystem be highly reliable and a maintenance interval of 6 months or longer is desired. Reliability studies will become a major part of our test program. When the refrigerator was delivered the compressor had run 1375 hours, and it has run an additional 1685 hours at IBM, bringing the total operation time to 3060 hours. Although this is only one third the scheduled maintenance interval, the refrigerator has been shut down and warmed to room temperature several times to correct problems, primarily with the instrumentation. Once the initial debug and performance mapping is completed, long term reliability testing will begin.

9.9 0.5 1.5

4. Interface performance

Heat Load (W)

The thermal interface between the refrigerator and dewar is a key element of the cryosystem. It must deliver the refrigeration efficiently in normal operation and provide good thermal isolation during periods when the refrigerator is warm for servicing. It also must provide for liquefaction of make-up helium. The thermal contact provided during cool-down from room temperature must be good so that the cooling time required is reasonable. In this section we present the test results for the thermal interface system and the JT temperature regulator.

Temp.

21.9 0 . 9 4 1.53

57 16.3

4.18 (est)

4.1 Thermal coupling heat transfer characteristics

The heat transfer characteristic, Q(AT), for the .first and second stage thermal diodes was measured by heating the base of the diode while monitoring the temperatures of the ingoing and outgoing JT gas and the diode base. For the second stage diode, thermometers T2, T7 and T8 were used to measure the temperatures. The results are shown in Fig. 2 where the heat transported through the diode is plotted as a function of the temperature drop across the diode. The circled points are taken fromR. Longsworth's design calculations and their locations below the measured curves indicate that the diode performs somewhat better than anticipated. The two sets of data for this diode were fit to a straight line which gave slopes (diode conductance) of 0.81 W/K and 1.85 W/K and intercepts of -0.24W and -0.31W. The intercepts should be the parasitic load and in fact are in approximate agreement with the calculated value and also with that determined from the measured enthalpy increase in the JT gas.

S e c o n d S t a g e D i o d e

F i g u r e 2 : H e a t t r a n s f e r c h a r a c t e r i s t i c f o r t h e s e c o n d s t a g e d i o d e .

Similar measurements and analysis were made for the first stage diode. There is some doubt as to the absolute accuracy of T6 which measures the diode temperature but it can be used to measure the temperature changes at the base of the diode. It therefore was not possible to obtain parasitic losses from these data. In addition, the heat transfer characteristic for this diode is non-linear which makes the analysis more difficult. The slopes obtained from fitting the data through Q = 15W total load to a straight line give a measure of the diode's effective thermal conductivity. For AT = Tdiode - TJT in the conductivity obtained is 0.69 W/K and for AT = Tdiode - TJT out the conductivity is 1.1 1 W/K. If mstead one fits the same data to a second order polynomial to allow for the curvature, the corresponding effective conductivities for 15W through the diode are 0.75 W/K and 1.22 W/K. The numbers obtained fromR. Longsworth's design calculations are 0.75 W/K and 1.9 W/K. One can conclude that the coupling performs substantially as designed.

The helium condenser used to couple the bath to the JT stream was tested similarly at several bath temperatures between 3.6K and 4.5K. The results for a nominal bath temperature of 4.OK are shown in Fig. 3. The squares represent the temperature difference between the bath and the JT loop upstream of the condenser, and the circles give the difference between the bath and the JT loop downstream of the condenser. R. Longsworth has calculated a temperature drop of 0.158K for 2 watts through the condenser. This is equivalent to a thermal conductivity of 12.7 watts/K. The values obtained from fits of the data to a straight line typically give 18-26 W/K with no discernable dependence on the operation temperature in the range of 3.6K < Tbath < 4.5K. The condenser is at least 50% more effective than anticipated.

4.2 System performance of the interface

The performance of the thermal interface in normal operation is shown in Table 1. The temperatures marked "estimated" have been determined from the heat transfer characteristic for the appropriate coupling. As the temperatures in the table indicate, the refrigeration is very effectively coupled to the heat stations and the bath.

A test of the thermal isolation provided by interface when the refrigerator is warm was conducted as follows. With the refrigerator in its normal operating mode the 1.4 liter test pot was filled. The refrigerator was then turned off and warmed to room temperature with the heaters on the two expander stages, the heater on the third adsorber, and a flow of 6 SL/M of helium gas through the low pressure side of the JT loop. The average boil-off during the 2.5 hours required to warm the refrigerator to room temperature was 0.055 liters/hr. At 2.15 hours 2.4 SL/M of the He gas flow in the JT was directed out the high pressure side. This warmed the tops of the two couplings and the condenser and increased the boil-off from the pot to 0.15 liter/hr. The refrigerator was restarted at the end of the purge, 3:15 hours after it had been shut down. The total helium loss during this period was 0.25 liters. After the refrigerator was started the boil-off increased to 0.24 liter/hr. It is likely that this increase indicates that the tops of the couplings had not been fully warmed to room temperature and a longer purge is required. Assuming normal operation 3 hours after the restart of the refrigerator, as demonstrated by Air Products, and a 0.20 liter/hr average boil-off during this time the total helium loss would be approximately 0.8-1.0 liters. Some adjustments in the purge procedure may make it possible to reduce the time and/or helium loss further. It is worthwhile to note that the 1 liter of liquid which is lost due to the cryogenic structures and the reverse heat leak of the thermal couplings is small compared to 8.4 liters which would be boiled by a 1 watt processor whose operation is maintained during this time.

4.3 Liquefaction

The liquid lost during a refrigerator shut down must be replaced by the liquefaction of He gas supplied from room temperature. .This is done by admitting the gas to the condenser via a tube that is thermally sunk at the first and second stages of the refrigerator. The liquefaction rate has been measured with and without additional heat loads applied to the three rpfrigerator stations. With no applied loads (parasitic losses only) the liquefaction rate is, 6.7 liter/day. With simultaneous loads of l o w , 0.56W, and 0.23W applied to the first and second stages of the refrigerator and the helium bath,respectively, the liquefaction rate dropped to 3.9 liter/day. Assuming ideal heat exchange at the first and second stages the 6.7 liter/day liquefaction rate represents a load of 11.8W, 2.3W, and 1.05W to the three stages of the refrigerator. The 1.05W load at 4.2K is roughly 50% less than the measured refrigeration capacity at this temperature and suggests that the heat exchange at the second stage is not complete. It may be possible to increase the liquefaction rate through better exchange at the first and second stages and/or thermally sinking the condensing line to the third JT heat exchanger. It is not essential, however, in this application that the liquefaction rate be optimized.

Figure 3: Heat transfer characteristic for the condenser. The squares are determined from the temperature difference between the bath and the JT temperature upstream of the condenser and the circles from the JT temperature downstream of the condenser.

4.4 Cool-down from room temperature

The time required to cool the helium test pot to 4.2K starting with the entire system at room temperature is approximately 10 hours. The f i s t and second stages cool more quickly, nearing their operating temperatures in ,3-4 hours. The difficulty is that the condenser is designed as a 2-phase device and provides only a very weak thermal link at higher temperatures. One could achieve a more rapid cool-down with the available refrigeration if, for example, a hydrogen gas thermal coupling were added betweeen the pot and the second stage. Another alternative would be to modify the condenser geometry to enhance gas convection. A model calculation, which assumes good contact between the He gas and the pot, indicates that such a modification could reduce the time required to cool the pot to 6 hours. The changes in the geometry of the condenser would not compromise its normal operation or the reverse heat leak. It is expected that the therml contact between the gas and pot would have to be improved with fins.

4.5 JT temperature regulator

A JT pressure/temperature regulator was supplied with the refrigerator. A segse pressure, referenced to vacuum, causes an increase or decrease in gas flow through the regulator from the expander return to the JT return which regulates the JT return pressure. In our tests the sense pressure was the saturated vapor pressure in the helium pot. Figure 4 shows the results of the tests. The horizontal axis is the heat applied to the liquid in the test pot and vertical axis shows the resulting temperature increase. The upper curve was taken with no regulation and represents the ability of the refrigerator to absorb the increased load. A linear fit gives a response characteristic of 0.039K/W. The lower curve gives a regulated response of 0.013K/W. This represents roughly a factor of . 3 improvement.

There is a mechanical hysteresis in the regulator of about 20 mm Hg which corresponds to a temperature difference of 0.028K at 4.2K. In fact, it was necessary to tap the regulator continuously to obtain the.data in the figure. Air Products had measured a substantially smaller hysteresis (8 mm Hg) prior to delivery. The cause of the hysteresis has yet to be studied.

The performance of the regulator has not yet been measured at lower bath temperatures. Since both the sense and the feedback response generated by the regulator depend on the slope of the saturated vapor pressure curve, dP/dT I ,,,, the gain of the system may be essentially unchanged at lower temperatues. Mechanical hysteretic effects, however, are likely to be more pronounced.

Summary

Tests of an Air Products refrigerator and thermal interface designed to support a superconducting computer have given the following results. In normal operation the interface provides good thermal contact between the dewar structure and the bath with a minimal loss of refrigeration. The liquid bath can be maintained automatically between 4.20 and 4.22K for applied heat loads varying from 0 to 1.6W. The interface also automatically isolates the dewar and bath in the case of refrigerator shut down.

The procedure and/or instrumentation used to warm the refrigerator for servicing and purging should be improved to reduce the time and liquid lost. The time required to cool the system can be and should be reduced by the addition of a thermal switch between the second stage and the helium pot. This will require a change in the hardware. The long-term reliability of the compressor package and refrigerator has yet to be studied.

JT Temperature Regulator

0 0 .4 0.8 1.2 1.6

POWER (W)

Figure 4: Increase in bath temperature, AT, as a function of the power applied to the bath with and without the JT regulator.

Acknowledgement

We would like to acknowledge the assistance given by J. Stasiak with the data acquisition.

References

] Anacker, W., &, "Josephson Computer Technology," IBM J. Res. Dev., 3 [2], 107-264 (March 198

[2] Longsworth, R., "Serviceable ~efr igerator System for a Small Superconducting Device," Proceedings of the Conference on Refrigeration for Cryogenic Sensors and Electronic Systems, NBS, Boulder, CO, Oct. 6 and 7, 1980.

[3] Air Products and Chemicals, Inc., APD Cryogenics, Allentown, PA, 18103.

DEVELOPMENTS TOWARD ACHIEVEMENT OF A 3-5 YEAR LIFETIME STIRLING CYCLE REFRIGERATOR

FOR SPACE APPLICATIONS

Max G. Gasser and

Allan Sherman

Cryogenics, Propulsion and Fluid Systems Branch Space Technology Division

Goddard Space Flight Center Greenbelt, Maryland 2077 1

and

William Beale

Sunpower, Inc. Athens, Ohio

A major effort is underway at the NASA/Goddard Space Flight Center to develop a long lifetime Stirling cycle cryogenic cooler for spaceflight. The initial objective is a 3-5 year, 5 watts at 6S°K, cooler to be followed by a 12°K cooler. The previous paper, "Progress on the Development of a 3-5 Year Lifetime Stirling Cycle Refrigerator for Space," presented at the 1979 Cryogenic Engineering Conference, outlined the approach [ 1 1. During the past year, the design of a refrigerator with linear magnetic bearings has changed. The displacer rod no longer goes through the piston, resulting in a machine that is comprised of a compressor section and an expansion section bolted together. Each section has its own magnetic bearings, linear motor, displacement transducer (LVDT), and displacement volume. The phase angle and displacements of the two moving parts are electronically controlled, utilizing the signals from the LVDT's. Delivery of the engineering model magnetic bearing cooler is scheduled for March, 198 1.

Work on the engineering model gas bearing cooler previously reported has ceased. However, an advanced concept for a gas bearing cooler is being studied.

Preliminary results indicate that the normal gas flow in a free displacer Stirling cycle cooler can be utilized to spin the displacer at a speed sufficient to form a self-acting hydrodynamic gas bearing with a clearance seal. The periodic gas flow from the cold end through the generator to the ambient end makes intermittent contact with a turbine located at the ambient end of the displacer. The induced spin is unidirectional.

Key words: Cryogenics; space application; long lifetime; Stirling cycle; linear magnetic suspension; clearance seals; hydrodynamic gas bearing.

1. Introduction

A wide variety of spacecraft instruments require cryogenic cooling to accomplish mission objectives. There are many cooler systems which have been developed or are being developed to satisfy these needs [2]. A particular critical'area is the requirement for a system capable of providing high cooling-load (5-10 watts) for 3-5 year missions. A major effort is, therefore, underway to develop a long-lifetime Stirling cycle cryogenic cooler for spacecraft.

The basic problem is to develop a closed cycle machine with moving parts that will reliably run for billions of cycles while unattended. Past attempts to solve this problem have used liquid or semi-liquid lubricants, dry lubricants, and "hard-on-hard" bearings with little or no lubricants [3] . Various seal designs were also tried. To-date, these approaches have not yielded a long-lifetime spaceborne cooler system.

The new approach initiated by NASAIGoddard Space Flight Center hopefully eliminates the problem-causing features of the past. The key elements of the approach are: (1) The reciprocating components are driven directly with linear motors. (2) While operating, there is no contact between the moving components and the machine housing and motor. The noncontact operation can be achieved by either magnetic or gas bearings and clearance seals.

The magnetic approach is presently being developed into an engineering model cooler for test and evaluation. The fluidic approach is still in the technology phase.

2. Cooler Utilizing Magnetic Bearings

2.1 System description

Figure 1 is the schematic of the North American Philips single expansion cryogenic cooler with linear magnetic suspension. There are three subassemblies: (1) the compressor section, (2) the expansion section, and (3) the balancing device. The displacer rod does not go through the piston, and at no time do the piston and displacer occupy the same space. Each subassembly has its own linear motor, bearings, and displacement transducer, The phase angle can be changed by the input signals to the motor. The linear motions will be sinusoidal. The displacer and piston are free to rotate, but are not intended to do so. There are no springs on the displacer; it is driven in both directions. The piston operates with a gas spring. There are two pressure transducers (not shown in the figures): (1) to measure the static pressure, and (2) to measure the dynamic pressure cycles. Temperature transducers (not shown) are mounted on the exterior of the pressure shell at the cold end and the ambient areas. A small electric heater (not shown) is located at the cold end to simulate a load.

Figure 2 shows the compressor section at mid-stroke. The motor is the moving magnet type with the magnets and coils sealed in metal cans to prevent outgassing products from reaching the working fluid. The piston is hollow and vented to the motor housing. The close clearance (0.0025 cm) between the piston and its cylinder forms the noncontact seal with the piston centered in the cylinder by the magnetic bearings in the attraction mode.

The bearings are co~pr ised of four nickelliron poles acting on a vanadium Permendur shaft. Two eddy current sensors, positioned 90 apart around the circumference, measure the gap. The error signal from these sensors is then used to adjust the current in the magnets to prevent contact between the moving member and the cylinder wall. A redundant set of eddy current sensors is also incorporated in the design.

Figure 3 shows the expansion section at mid-stroke. The cold end has a heavy copper cap, which is connected to the ambient flange by a thin titanium tube for low thermal conductivity. The cold heat exchanger is the gap between the displacer and the copper cap. The cold end of the displacer is hollow and contains a vacuum. Regeneration is by the phosphorbronze screen matrix and also in the gap between the displacer and the cylinder. (Again, in the seal area, this gap is 0.0025 cm.) The ambient temperature flange supports a vacuum dewar (not shown) around the cold end. Starting at the cold end, the helium working fluid flows through a central passage in the displacer in the location of the magnetic bearing. This is the part of the displacer which constitutes the clearance seal. The working fluid flows outward through radial passages, over the linear motor, between the rear magnetic bearing, around the LVDT, and into the piston compression volume. The displacer motor is the moving magnet type with the coils and magnets sealed in metal cans. The displacer bearings operate in exactly the same way as that described for the piston.

104

65' K COLD END

HELIUM Y F I L L AND VENT VALVE

COMPRESSOR LINEAR MOTOR - (MOVING MAGNET)

COMPRESSOR LINEAR POSITION SENSOR (LVDT)

I

I* COUNTER WEIGHT ON I CANTILEVER SUPPORT 8 SPRINGS

I I IN-HOUSE ACTIVE ' BALANCING SYSTEM.

LINEAR MOTOR WITH MAGNETIC BEARINGS. L ------ --- ----

Figure 1. schematic of the North American Philips single expansion cryogenic cooler with linear magnetic suspension.

105

PERMANENT POLE PIECES MAGNETS A VENT , HOLE

PRESSURE MOTOR COILS

COM PRESS1 ON SPACE

SENSORS / MAGNETIC BEARING

Figure 2. Philips Laboratories linear compression section of the Stirling cycle cooler shown in Figure 1.

All of the waste heat is removed by cooling jackets around the outside of both the piston and displacer housings.

Figures 4 and 5 show the passive counter-mass on cantilever support springs, which is one form of balancing device. The active balancing system, which utilizes accelerometers, a linear motor (moving iron type), and magnetic bearings, is described in reference [4]. The balancing device is bolted to the ambient end of the refrigerator. There is

POSITION

COOLING JACKET FOR WASTE HEAT REMOVAL AROUND THIS SECTION

EXPANSK SPACE

- SENSOR

COLD END

MAGNETS

HELIUM GAS N P E FLOW PASSAGES

.'... .:.:.:...:<. :.. ,,,f ...... %I-, ,

\ LVDT

Figure 3. Philips Laboratories expansion section of the Stirling cycle cooler shown in Figure 1. 106

LEAF SPRINGS

MASS

COIL SPRING

HOUSING

Figure 4. Cross-section of the counter-mass passive balance system for Figure 1.

no communication of the working fluid to the balancing device. The balancing could also be accomplished by two refrigerators back-to-back and operating in counter-phase.

The passive spring-mass system has been used on the Philips MC-80 (Einhoven) cryogenic refrigerator and is tuned to vibrate counter-phase to the piston.

Figure 5. End view of the counter-mass passive balance system for Figure 1.

107

The concept of the turbine spun, hydrodynamic displacer system having a bearing and turbine design compatible with a single-stage Stirling cryocooler of 5 watt heat lifting capacity at 65°K is described in the following para- graphs.

3.1 Hydrodynamic bearing system concept

Figure 9 shows a turbine spun displacer and an electrically driven piston which may either be spun or supported by hydrostatic bearings or magnetic bearings. This is the concept which we would like to develop into a proof-of-principle device.

3.2 Proposed spin bearing demonstration mo'del

Figure 10 shows the proposed proof-of-principle device.

The displacer has a mass of approximately 50 grams, and this can be sustained in a 1.5 g field with a slow speed gas bearing having a bearing gap of about 7 micrometers - a value typical of existing Sunpower seal clearances.

For purposes of turbine design, it would be desirable to spin the bearing more rapidly than the maximum stable 10 radianslsecond, but this, in theory at least, can lead to plain bearing instability (whirl) under low loading as would be in the case in space. The use of herringbone rather than plain bearing surfaces gives promise of elimination of whirl instability at low load and high speed.

The proposed design will be first tested with a plain bearing. If this proves unsatisfactory, a herringbone will be chemically etched on the plain bearing surface (the depth of the herringbone groove is only about 20 micrometers).

Figure 9. Sunpower, Inc. long life cooler concept.

111

RECIPROCATING ROTATING MOTOR

TURBINE ROTOR,

TURBINE AND BRAKE CONCEPT

/ J A J B SPINNING DISPLACER

ROTOR * SECTION A-A

FAN BRAKE

GAS

TURBINE STATOR - 12 BLADES TRANSDUCERS

SECTION 8-6

SPRING

Figure 10. Sunpower, Inc. proposed spin bearing demonstration model.

3.2.1 Turbine design

The basic problem of the turbine spun bearing system is that while the spinning bearing requires very low torque to maintain its load carrying capacity, its static friction starting torque is many times greater, at least as indicated by elementary static friction estimates. The high starting torque requires a high tangential gas velocity impinging on the turbine, which would impose an excessively high running speed, possibly inducing large out of balance forces. But, this is by nomeans an intractable problem. Many solutions are possible of which the least complex might be a centrifugal brake (fan) on the displacer. However, one is reminded of the many occurrences in the past in which great effort was expended to solve on paper what in the lab was found to be a nonexistent problem. a Just what the startup torque requirement is in a reciprocating spinning seal with a pressure difference across it is apparently not known, and it may be very much lower than the static friction torque.

The simplest approach is a plain bearing and a brute force turbine capable of overcoming full sliding friction and accelerating the bearing to a highqeed. Combined with the strong turbine, a geometrically similar fan brake is included, which is intended to limit displacer angular velocity to one tenth of the theoretical velocity capability of the turbine. Both the turbine and brake were designed for ease of experimental variation, with the assumption that some modification would inevitably be required.

The turbine was designed such that it is energized only about 118 of the cycle time. Its loss is in the range of tenths of watts at most, and little effect on the thermodynamic performance of the cooler is expected.

3.2.2 Instrumentation

The purpose of the demonstration model is to show hydrodynamic lubrication without contact as the displacer cycles in a manner compatible with operational cryocooler performance. In order to show this, instrumentation must be provided which shows presence or absence of contact.

112

The system uses two proximity sensors at 90" separation in two separate planes perpendicular to the displacer axis. These show a circle on an oscilloscope within which a dot representing the center line of the displacer runs, with contact represented by the dot reaching the inner edge of the circle. This display would show not only contact, but also eccentricity and whirl, and would be very useful in diagnostics.

3.2.3 Characteristics of proposed designs

The characteristics of the demonstration model are as follows:

Operating pressure Frequency Displacer diameter

Bearing length Mass Rod diameter Stroke Rotation rate

Piston diameter Stroke

Drive motor - Type - Power - Speed

13 bar (helium) 25 Hz 20 mm 40 mm 75 grams (for demonstration purposes) 5.3 mm 8 mm 300 radians/second (3500 RPM max) 52 mm 8 mm Dayton PM 42143, 12 volt 100 watts 1750 RPM

The displacer is rotated by a turbine, which has gas flow for only 118 of the total stroke. During the remainder of the displacer motion, the turbine stators are short-circuited by the normal warm heat exchanger ports. Volumes are sized so that the mass flow through the ports is the same as would be present if the cryocooler temperature existed. This was done by enlarging the volume of the demonstration model cold end by the temperature ratio, so that it represents a larger mass flow source and sink, as does the cold end of the cryocooler.

The displacer has a fan brake which will initially be blocked. If experiment shows problems with overspeed, the dam will be progressively removed, and the brake brought into greater effectiveness until the desired displacer rotating speed is met. At full effectiveness, the brake is expected to limit rotating speed below 300 radians per second.

Materials are anodized aluminum and chrome-plated aluminum ground and honed to true roundness and straightness. These have given good service in past applications to small Stirling heat engines.

3.3 Sequence of operations of the spin mechanism

Figure 11 shows the sequence of operations which causes the displacer to spin as it reciprocates in the normal mode of the free displacer driven by fluidic forces.

The displacer motion leads the piston by 90°, that is to say, at a time when the displacer is at the midpoint of its motion, going out (to the right), the piston is at the innermost point of its motion, just starting its outward motion (out).

When the displacer is at its outermost point, where the turbine has come into register with the turbine stator, the piston is at its midpoint and moving with maximum velocity outward. At this point, the maximum flow rate of gas occurs from the cold expansion space, through the regenerator, through the stator and turbine blades (since the warm port is blocked by the displacer), and then through the connecting tube to the piston space.

It is useful to note that this maximum mass flow occurs at a point (about 190") where the displacer is almost stopped, having just turned around and started in again. The flow occurs as a result of the influence of the motion of the piston, which is at its maximum rate. Cold gas in the expansion space moves through the regenerator, warming and expanding, then moving through the turbine toward the piston.

WARM PORT

GAS MASS FLOW RATE FROM WARM HEAT EXCHANGER PORT VS. DISPLACER POSITION

COMPRESSION SPACE

FLOW OUT OF

STA

90 180 270 360

DISPLACER AT 200' DEGREES AS ROTOR PORT FROM STATOR BEGINS TO CLOSE. WARM PORT BEGINS TO OPEN. DISPLACER

MOVING IN

Figure 1 1. Sequence of operations of Sunpower, Inc. proposed spin bearing demonstration model.

The turbine is positioned so as to intercept this maximum flow of gas at a brief interval during the outermost point of its travel, thus imparting a spin to the displacer.

As the displacer moves back in again, it covers the stator, and once again uncovers the warm port so that normal flow through it can occur. Flow outward at its midpoint, and the piston has almost reached its outermost travel. As the component motions continue, the flow through the warm port reverses and now flows through the port toward the cold space, increasing in magnitude under the influence of the piston. The piston reaches its midpoint and maximum inward velocity at about the time the displacer reaches its innermost point and reverses its direction to come out again and repeat the cycle.

During the entire cycle, the displacer is spinning, even though it receives a spin impulse only during the 45" of its travel, and as it spins, it pumps gas through the brake after the manner of a squirrel cage fan. This brake is designed to limit maximum rotating speed to reasonable values fixed by bearing performance. The brake is only effective at high speed and has no retarding effect on the displacer during startup, when maximum turbine torque is applied to begin the spinning action.

4. Conclusion

The major effort in the Goddard program is to develop the magnetic bearing cooler. The self-actuated spin bearing approach is still in the study phase; however, initial analysis shows that the approach is promising.

5. References

[ 11 Sherman, A., Gasser M., Goldowsky, M., Benson, G. and McCormick, J., "Progress on the development of a 3-5 year lifetime Stirling cycle refrigerator for space," in Advances in Cryogenic Engineering, Volume 25.

114

[2] Sherman, A., "Cryogenic cooling for spacecraft sensors, instruments, and experiments," in Astronauts and Aeronautics, November 1978.

[ 3 ] Gasser, et al., "An approach to long-life Vm cryogenic refrigeration for space applications," Applications of Cryogenic Technology, Volume 4, 197 1.

[4] Studer, P. and Gasser, M., "A bi-directional linear motor/generator with integral magnetic bearings for long lifetime Stirling cycle refrigerators," Topical Conference on refrigeration for cryogenic sensors and electronic systems, October 1980, at National Bureau of Standards.

THOR CRYOCOOLER

G. M. Benson, PhD and R . J . Vincent, PhD E R G , Inc.

Oakland, CA 94608

A resonant, free-piston St i r l ing-cycle cryocooler driven by l i nea r reciprocating synchronous e l e c t r i c motors and u t i l i z i ng gas bearings and gas springs has been designed, and selected components tes ted. This hermetically-sealed, dynamically balanced design i s based on E R G ' S previous work and has been engineered f o r long-l i fe through use of : 1 ) non-contacting l i n e a r gas bearings and narrow clearance reciprocating sea ls , 2 ) brush1 ess reciprocating e l e c t r i c motors with hermetically- sealed coi l and core, 3) s ta t ionary external f i r s t - s t a g e regenerator, and 4) sol i d - s t a t e cont ro l le r . The design has been efficiency-optimized by 1 ) t ransfer r ing heat nearly isothermally ( t o be t t e r approach Carnot e f f ic iency) , and 2 ) minimizing losses through use of advanced components of resonant free-pistons (which eliminate piston dr ive losses ) . The single-stage uni t , designed t o produce 10 watts a t 65 K , may be expanded t o three-stages, using a novel design, f o r producing 4.2 K. Designs and t e s t r e su l t s of selected components a r e described and compared with theoret ical and computer modeled r e su l t s .

Key words: Cryocooler; gas bearings; re f r igera t ion ; regenerator; resonant f r e e piston; S t i r l i ng cycle.

A CRYOGENIC SYSTEM FOR THE SMALL INFRARED TELESCOPE FOR SPACELAB 2

E. W . Urban and L. Katz Marsha l l Space F l i g h t Center H u n t s v i l l e , Alabama

J. B. Hendricks and G. R. Kar r U n i v e r s i t y o f Alabama i n H u n t s v i l l e Huntsv i 11 e, A1 abama

The cryogenic system f o r t h e small he1 ium cooled i n f r a r e d te lescope f o r Spacelab 2 i s descr ibed. The system cons is t s o f a 250 R storage dewar con ta in ing s u p e r f l u i d he l ium a t about 1.6K and a separate exper i - ment c r y o s t a t . The l i q u i d hel ium i s con f i ned w i t h i n the dewar i n zero-g by means o f a porous p lug. The c o l d vapor l eav ing the p lug sur face i s s p l i t i n t o two, n o t necessa r i l y equal, f lows. One f l o w path i s through heat exchangers which coo l t h e r a d i a t i o n s h i e l d s o f t h e s torage dewar and, t he re fo re , c o n t r o l t he l i q u i d s torage e f f i c i e n c y . The second f l o w pa th i s through a f l e x i b l e j o i n t t o t h e independent s e t o f heat exchangers f o r c o o l i n g the i n f r a r e d o p t i c a l apparatus and r a d i a t i o n sh ie lds i n t h e c r y o s t a t . By means o f commandable valves on t h e dewar and c r y o s t a t ven t l i n e s , t h e f r a c t i o n o f t o t a l he l ium b o i l o f f used f o r coo l i ng the dewar r a d i a t i o n s h i e l d s can be c o n t r o l l e d , and t h e c o o l i n g a v a i l a b l e a t t he c r y o s t a t can be ad jus ted f rom zero f l o w r a t e t o n e a r l y 50 mg/s o f hel ium gas a t 2K. A number o f unusual cryogenic problems which have been solved i n t h e development o f t h i s complex system a r e discussed.

Key words: Cryogenics, i n f r a r e d telescope, Space1 ab, super f 1 u i d he1 ium.


The I n f r a r e d Telescope (IRT) Experiment w i l l f i r s t be f l own on t h e Spacelab 2 miss ion i n l a t e 1983. I t i s a j o i n t p r o j e c t o f t h e Smithsonian Ast rophys ica l Observatory (SAO), t he U n i v e r s i t y o f Ar izona (UA), and t h e NASA Marsha l l Space F l i g h t Center (MSFC). Our r e s p o n s i b i l i t y a t MSFC i s t h e development o f t h e cryogen ic and mechanical systems of t he IRT. The U n i v e r s i t y o f Alabama i n H u n t s v i l l e (UAH) i s c o l l a b o r a t i n g on the cryogenic design, f a b r i c a t i o n o f spec ia l cryogenic apparatus, and cryogenic performance t e s t i n g . UA and SAO a r e respons ib le f o r t h e i n f r a r e d ins t rument and data ana lys i s . The NASA Ames Research Center i s c o l l a b o r a t i n g i n t h e ana lys i s o f contaminat ion-and zod iaca l l i g h t data. Th is paper descr ibes the cryogen ic apparatus, t h e s t a t u s o f assembly and t e s t i n g , and a number o f speci a1 cryogenic problems which we have faced du r ing development .

2. Cryogenic System

The IRT experiment employs a very unusual cryogenic design which r e s u l t s from both t h e requirements o f t h e i n f r a r e d science and a d e s i r e t o u t i l i z e commercial apparatus and f a b r i c a t i o n technology wherever poss ib le . Two views o f a sca le model o f t h e IRT a r e shown i n F igu re 1, and a schematic o f t h e cryogenic system i s shown i n F igu re 2. A d e s c r i p t i o n o f t h e mounting and opera t i on o f t h e 3.4m t a l l experiment i n t h e Spacelab 2 payload was p r e v i o u s l y g i ven [I 1 ,[2 3. The two major cryogenic components, t h e dewar subsystem and t h e c r y o s t a t subsystem, a re descr ibed be1 ow.

2.1 Dewar Subsystem

The dewar subsystem c o n s i s t s o f two par ts , a 250 l i t e r l i q u i d hel ium storage dewar and a t r a n s f e r assembly (TA), which form a c l o s e l y i n t e g r a t e d thermal systex. The r e s u l t i s e s s e n t i a l l y a s i n g l e complex s u p e r f l u i d he l ium dewar w i t h t h e l i q u i d containment vessel a t one end o f a neck s t r u c t u r e and a number of c o l d f low cont'rol components a t t h e o ther ; a l l a re surrounded by a common thermal p r o t e c t i o n system, discussed below.

BURST DISC FLOW METER HEATER LEVEL METER PRESSURE GAUGE OUARTZ CRYSTAL MICROBALANCE RELIEF VALVE SUN SENSOR TEMPERATURE SENSOR VALVE

*ACTIVATED BY GSE

ASSEMBLY - (ACTIVATED BY DEWAR FILL

UMBILICAL CONTROL BOX CONTROL BOX

(1 15V ACI)

.Figure 2. Schematic diagram of t h e i n f r a r e d telescope system.

RETRACTABLE VALVE OPERATOR (TYP)

rELECTR lCAL CONNECTOR

GHE VENT BAYONE - AND VALVE V73 .GHE VENT (TO FLOW CONTROL VALVES)

FlLL LlNE - RELIEF VALVE

V15 (RV4)

LHE F lLL BAYONET AND VALVE V6 I,'

TRANSFER ASSEMBLY

MULTILAYER INSULATION

VAPOR COOLED SHIELDS (TYP)

'250 LITER HELIUM DEWAR

VACUUM PUMP-OUT PORT WlTH INTEGRAL RELIEF VA V4-(RV6)

VENT LlNE RELIEF VALVE V16(RV3)

GAS LlNE HELIX

CRYOSTAT LlNE

LOW CONDUCTIVITY NECK TUBE (TYP)

VACUUM PUMP-OUT PORT WlTH INTEGRAL RELIEF VALVE

V8(RV-5)

MULTILAYER INSULATION

LIQUID WITHDRAWALNENT TUBE

LIQUID LEVEL

LIQUID HELIUM (1.6 K, 6 TORR)

/--MOUNTING RING

&/- /--FILL TUBE

LIQUID VESSEL

/OUTER SHELL

. SUPPORT STRAP (3 @ 1 2 0 ~ )

THERMOMETER HEATER

Figure 3 . Schematic diagram o f the IRT dewar subsystem.

120

When t h e experiment i s launched i n t h e Space S h u t t l e t h e dewar subsystem w i l l be h o r i z o n t a l , and l a r g e acce le ra t i ons w i l l be a p p l i e d t ransverse t o t h e dewar ax i s , To a i d i n suppor t ing t h e mass o f t h e 250 l i t e r s o f helium, t h e l i q u i d vessel and the dewar VCS aga ins t these loads, t h r e e f i be rg lass /epoxy s t raps a t t a c h the bottom o f t h e l i q u i d vessel r a d i a l l y t o t h e o u t e r s h e l l . Our dewar was f a b r i c a t e d by Cryogenic Associates, I nc . I t has been vacuum leak tes ted, b u t cryogenic t e s t i n g must awa i t t h e assembly o f t h e TA onto t h e dewar.

2.1.2 T rans fe r Assembly

The TA, designed and f a b r i c a t e d by UAH, i s c u r r e n t l y being assembled onto the dewar a t MSFC. W i t h i n t h e o u t e r case o f t h e TA i s a s e t o f t h ree nested heatexchanger tubs, shown i n F igu re 4, each w i t h a gas tube welded t o i t s upper c i rcumference. Each t u b w i l l be wrapped w i t h MLZ and b o l t e d t o a dewar s h i e l d extending up f rom t h e neck. As descr ibed below, a p o r t i o n o f t he dewar b o i l o f f gas l e a v i n g t h e porous plug, i s c i r c u l a t e d i n t u r n through t h e t h r e e heat exchangers and then t o ex te rna l va lves.

W i t h i n the i n n e r tube i s a demountable plumbing module which i s supported by t h e dewar neck. The module has f i v e tub ing connections, each u t i l i z i n g a Var ian m i n i c o n f l a t u l t r a h igh vacuum f lange. These connect ions a re t o t h e dewar f i l l and vent tubes, t he ex te rna l fill l i n e , t h e i n n e r heat exchanger tub,and t h e gas l ~ n e t o t h e c r y o s t a t . The module components, shown i n F i g u r e 3, opera te a t t h e temperature o f t h e l i q u i d helium, nomina l ly 1.6K. They a r e descr ibed below.

L i q u i d he l ium i s in t roduced i n t o t h e dewar subsystem through a standard bayonet and vacuum jacke ted va l ve ( ~ 6 ) ~ then through c o l d f i l l va l ve (V7) and t h e dewar f i l l tube. The co ld f i l l va l ve (V7) prevents l i q u i d hel ium from f l o w i n g o u t t h e f i l l l i n e t o the warm ou te r s h e l l when t h e experiment i s i n space. A b u r s t d i s k ( B l ) permi ts c o n t r o l l e d vent ing through t h e f i l l l i n e i n t h e event o f acc iden ta l ove rp ressu r i za t i on . I n space, l i q u i d f l ows i n t o the TA through t h e wi thdrawal l i n e and i s r e s t r a i n e d by t h e porous plug. The p r i n c i p l e and opera t i on o f porous plugs, which a re e s s e n t i a l l y l i q u i d l v a p o r phase separators f o r super- f l u i d hel ium a r e descr ibed elsewhere [3]. I n t h e event o f p lug blockage and overpresur- i z a t i o n a second bypass b u r s t d i s k (82) prov ides r e l i e f i n t o t h e ven t l i n e s . When the dewar i s being i n i t i a l l y f i l l e d o r when normal he1 ium (4.2K, 1 atmosphere vapor pressure) i s being converted t o s u p e r f l u i d (1.6K, .008 atmosphere pressure) , l a r g e gas f l o w r a t e s must be vented. A b y ~ a s s v a l v e (V5) i s opened i n p a r a l l e l w i t h t h e porous p lug t o pe rm i t h igh . f low r a t e s and avo id poss ib le contaminat ion o f t h e p l u g pores. A second bypass va lve (V17) i s opened when l i q u i d t o p o f f begins t o prevent i n i t i a l warm he1 ium from t h e warm t r a n s f e r 1 i n e f rom e n t e r i n g t h e l i q u i d vessel .

The c o l d valves, V5, V7, and V17 a re operated by vacuum-tight, r e t r a c t a b l e ac tua to r s h a f t s having l ow thermal c o n d u c t i v i t y and low thermal mass. When re t rac ted , t he ac tua to r s h a f t s w i l l be he ld i n p lace by a cover box which can be evacuated through va l ve (V20) t o prevent a i r leakage through t h e s h a f t r e t r a c t i o n sea ls du r ing the v i b r a t i o n s o f launch. To reduce wear on t h e v a l v e seats and min imize heat i n p u t when t h e va l ve ac tua to rs are engaged, t he va lves w i l l be opened a t t h e beginning o f c ryogen ic opera t ions and w i l l n o t be c losed except when t h e dewar subsystem must be turned on i t s s ide ( launch a t t i t u d e ) f o r t e s t i n g o r j u s t be fo re t h e ac tua l launch. Minor leakage p a s t t h e seats o f V5 and V17 can be to le ra ted , b u t V7 must remain sealed s u p e r f l u i d t i g h t even i n space t o avo id heat conduct ion down the f i l l l i n e f rom V6. To prevent t h e c o l d va lves f rom v i b r a t i n g open du r ing launch we have designed a sp r ing loaded d e t e n t l o c k operated by t h e va l ve ac tua to r .

The l i q u i d helium evaporates from the porous p lug a t a temperature o f abou't 1.6K and enters a mani fo ld from which i t i s d iv ided i n t o the dewar and c r yos ta t vent f lows discussed above. The cryostat. vent tube i s wrapped i n t o a he1 i c a l c o i l which f lexes w i t h the motion o f the c ryos ta t as the i n f r a r e d instrument i s scanned across t he sky. The insu la ted cryostat l i n e then passes through a f e r r o f l u i d i c seal which j o i n s the TA t o the cryostat,accommodates the scanning motion, and maintains the high vacuum w i t h i n the c r yos ta t and TA before the experiment i s ca r r i ed i n t o the vacuum o f space.

The dewar subsystem has been instrumented t o permi t accurate measurement and con t ro l o f i t s operat ion on the ground and on-orb i t . A number o f thermometers measure l i q u i d and VCS temperatures. Thermometers T5 and T6 on the porous p lug i nd i ca te the p lug performance and a commandable heater g ives us a means o f p lug f l ow con t ro l i n add i t i on t o the externa l valves (see below). A thermometer (T I ) and l i q u i d l e ve l probe are inser ted i n t o the dewar from the TA and can be replaced i n case o f mal funct ion. Several super f lu id t i g h t e l e c t r i c a l feedthrus are used t o b r ing instrumentat ion leads from the l i q u i d helium ou t i n t o the vacuum space.

2.1.3 Dewar Subsystem Thermal Performance

The dewar subsystem w i l l be tested by i t s e l f a t MSFC wh i le the i n f r a r e d instrument i s being i n s t a l l e d i n the c ryos ta t i n Arizona. I n i t i a l l y t he gas l i n e t o the c r yos ta t w i l l be terminated w i t h i n the TA and the dewar tested as a stand-alone subsystem. The dewar w i l l be f i l l e d w i t h normal helium and allowed t o s t ab i l i ze ; then the helium w i l l be converted t o super f lu id a t 'abou t 1.6K by pumping through the vent 1 ine. A long. term performance t e s t w i l l be made w i t h the dewar up r i gh t (porous p lug n o t wet by the l i q u i d helium), t i l t e d 90' i n t o the launch a t t i t u d e (p lug may OP. may no t be wet), and t i l t e d beyond 90' t o insure t h a t the p lug i s wet and operat ing as i t w i l l i n space where f l u i d f lows e a s i l y t o it. Actual performance wi 11 be compared t o pred ic ted performance .

I n i t i a l estimates o f the thermal balance i n t h i s complex dewar l e d t o the nominal temperatures ind icated i n Figures 2 and 3: 20K f o r the inner VCS, 60K f o r the middle VCS, and 120K f o r the outer VCS. The f i n a l con f igu ra t ion o f the system, p a r t i c u l a r l y the neck tube s t ruc tu re and the support straps, has l e d t o a r ev i s i on of the ana l y t i c p r e d i c t l ~ n s f o r steady s t a t e operat ion [2]. I n the case i n which a l l b o i l o f f gas i s used t o cool the dewar VCS (none t o the c ryos ta t ) , the sh ie lds are pred ic ted t o operate a t 30K, 126K, and 240K respect ive ly . The corresponding pred ic ted heat load on the l i q u i d vessel i s 152 mW and the b o i l o f f r a t e i s 6.7 mg/s, o r 4.0 l i t e r / d a y of super f lu id hel'ium.

When the stand-alone t es t s are completed, a c r yos ta t subs t i t u t e w i l l be i n s t a l l e d on the c ryos ta t gas l i n e o f the TA. This t e s t equipment w i l l a l l ow us t o simulate the d i v i s i o n o f gas f l ow between the dewar and c r yos ta t sh ie lds and t o make p re l im inary t es t s o f the f l i g h t s i t ua t i on .

2.2 Cryostat

The cryostat , shown schematical ly i n Figure 5, w i l l be a spec ia l mod i f i ca t ion o f an open neck laboratory dewar. The essent ia l special features inc lude mounting f langes f o r the two sect ions o f the in f ra red telescope, an access p o r t f o r i n f r a r e d detector i n s t a l l a - . t i on , a s ide extension on the c ryos ta t r o t a t i o n ax i s through which the co l d gas from the TA enters, and a gas heat exchanger and VCS system. A comandable vacuum cover insures t h a t a h igh vacuum can be maintained w i t h i n the c ryos ta t and telescope before the experiment i s i n space. The co ld gas from the TA i s f i r s t de l i ve red a t a temperature o f about 2K t o a co l d f i nger , designed t o mainta in a temperature o f about 2.5K, t o which the I R detector b lock w i l l be clamped, and theh t o the heat sh i e l d system. The lower and upper telescope sect ions are designed t o operate a t maximum temperatures o f 8K and 60K, respect ive ly . They w i l l be cooled by conduction t o t h e i r mounting f langes. The two telescope mount f l ange r ings , and two add i t i ona l r i ngs which are a t temperatures o f approximately 120K and 200K are suspended from the top of the c ryos ta t on la rge diameter f iberglass-epoxy tubes. Each r i n g ca r r i es a VCS and each de l i ve r s i t s co l lec ted heat t o the co l d vent ing helium gas. The vent tube, a f t e r leav ing the co l d f i nger , i s fastened t o the f o u r r i ngs i n turn, e x i t s the c ryos ta t a t ambient temperatures, then connects t o external f l ow con t ro l valves. Addi t ional design in format ion was given p rev ious ly [I]. The c ryos ta t was a lso b u i l t by Cryogenic Associates, Inc.

123

VACUUM

LOW THERMAL INERTIA SHIELD

VACUUM COVER SHROUD AND LATCM MECHANISMHOT SHOWN

ELECTRICAL FEEOT~~ROUQM

INTERMEDIATE THERMAL VACUUM PUMP-OUT PORT RtNG l120KI WITH INTEGRAL RELIEF VALVE LQW CO~WCTIVITY

W P W R T TUBES

UPPER TELESCOPE FLANGE AND THERMAL RING G O K I

UPPER TELESCOPE TUBE

MOUNTING RING

MULTILAVER INSULATION

LOWER TELESCOR~ FLANGE AN0 THERMAL RING l8K) FROM DEWAR SUBSYSTEM

COLD FINGER l2.5K) ACCESS PORT

MWNTING RING

VAPOR COOLED SHIELDS

OUTER SHELL

LOWER TELESCOPE TUBE

@ THERMOMETER

Figure 5. Schematic diagram o f the IRT cryos ta t subsystem.

The c r yos ta t was tested cryogenica l ly , using a t r ans fe r assembly subs t i t u t e which permit ted the con t r o l l ed f l o w o f co l d hel ium gas i n t o the c r yos ta t l i n e fromean ord inary helium dewar. The r e s u l t s o f these tests , discussed elsewhere [2], i nd i ca te t h a t a f l o w r a t e of about 17 mgls w i l l be required, a t an i n l e t temperature and pressure o f 2K and 6 t o r r (.008 a h ) , respect ive ly , i n order t o mainta in the i n f r a r e d o p t i c a l components a t t h e i r design temperatures.

2.3 Cryogenic System Performance

A f t e r t he c r yos ta t and dewar subsystems a re in tegra ted together, a cryogenic systemi t e s t w i l l be run t o v e r i f y the combined operat ion o f the system. We w i l l be p a r t i c u l a r l y in te res ted i n conf i rming the pred ic ted steady s t a t e and t r ans i en t behavior as we vary the d i v i s i o n o f vent gas f l ow between the two heat exchangers. We ,must develop an optimum s t ra tegy f o r the experiment operat ions prelaunch and on o r b i t i n order t o conserve helium and s t i l l permi t successful i n f ra red observations whenever required. The on ly source o f co l d gas i s vapor i za t ion o f l i q u i d helium from the dewar/porous p lug by e i t h e r p a r a s i t i c heat, o r e l e c t r i c a l heat from the heater on the porous plug, The p a r a s i t i c heat can, o f course, be con t r o l l ed over a wide range by vary ing the amount o f gas i n the dewar vent l i ne . For a g iven r a t e o f heat input , t o t a l pumping ra te , and d i v i s i o n o f flow,the l i q u i d w i l l come t o some equ i l i b r ium state. Regardless o f t he d i v i s i o n o f f low, t o t a l pumping r a t e must be adequate t o keep the l i q u i d i n the supe r f l u i d phase.

When the 17 mgls r e u i r e d f o r in f ra red observat ions i s d ive r ted t o the cryostat , the dewar f l ow i s pred ic ted q21 t o be on ly 3 mgls. I n t h i s s i t u a t i o n the dewar sh ie lds w i l l warm t o temperatures of 90K, 205K, and 260K respect ive ly . The r e s u l t i n g t o t a l system flow r a t e o f 20 mgls o r 12 li ter/day could be maintained on ly f o r about 21 days, if the dewar were i n i t i a l l y f u l l o f superf lu id. With t he Spacelab 2 on-o rb i t mission durat ion o f 9 days and an an t i c ipa ted prelaunch per iod a f t e r dewar topof f of as much as 15 days, i t w i l l n o t be possible, nor, i n fact , necessary t o f u l l y cool the c r yos ta t u n t i l a f t e r launch. As shown e a r l i e r , dedicat ing a l l f l ow t o the dewar r e s u l t s i n lowest dewar sh i e l d temperatures, a warm cryostat , and a t o t a l supe r l f u i d b o i l o f f r a t e of 4 l i t e r s l d a y o r a f u l l dewar l i f e t i m e of 62 days. C lear l y an in termediate s i t u a t i o n w i l l be chosen w i t h the i n f r a r e d op t i c s he ld a t a f a i r l y low temperature u n t i l t he experiment i s i n space.

Adjustment o f dewar and c ryos ta t vent f lows w i l I be poss ib le dur ing the mission by operat ing a s e t o f three commandable warm valves i n each of the vent l i n e s as shown i n F igure 2. The three p a r a l l e l o n l o f f valves i n each vent are s ized w i t h o r l f i c e s a l low ing . 14%, 29%,and 57% of f u l l f l ow i n the vent l i n e . Thus, par t i , cu la r combinations of 0, 14, 29, 43, 57, 71, 8 6 , ~ 100% of f u l l f low can be se lected f o r each vent l t n e w i t h the se t t i ng o f three valves. The system, and p a r t i c u l a r l y the telescope and 1 i'quid he1 ium temperatures w i l l respond ra the r s lowly t o changes i n these va lve se t t ings . We an t i c i pa te the occa- s iona l need t o r a p i d l y de l i ve r add i t i ona l coo l ing t o the c r yos ta t as, f o r example, when c losure of t he warm cover dumps a bu r s t o f heat onto the co ld op t i cs . This r ap i d de l i ve r y o f ex t r a coo l ing can be achieved by a c t i v a t i n g a small heater on the vapor s ide o f the porous plug. The well-known super f l u i d he1 ium " founta in-ef fect " [ 3 ] then draws 1 i q u i d through the plug.

3. Prelaunch Cryogenic Operations

Cryogenic se rv ic ing o f the IRT experiment before launch presents a number o f unusual and chal lenging requirements. The Spacelab 2 payload w i l l be i n s t a l l e d i n t o the Space Shu t t le O rb i t e r several weeks before launch. Perhaps 3 weeks before launch the IRT w i l l be purged, f i l l e d w i t h normal helium, and allowed t o s t a b i l i z e . Since the supply dewars must

' b e located on workstands o f l i m i t e d weight ca r ry ing capaci ty a t the sides o f the Orb i te r bay, several small supply dewars w i l l be requi red and the vacuum jacketed f l e x i b l e f i l l l i n e w i l l be 4.5 m long. Venting o f the co l d gas must be through long insu la ted hoses t o prevent condensation and r e s u l t i n g d r ipp ing o f water, onto other experiments o r o r b i t e r hardware. The he1 ium w i l l then be pumped down t o and maintained i n the super f lu id phase by means of large, remote vacuum pumps- As l a t e as poss ib le i n the ground processing f low, nominal ly 15 days before launch, the dewar w i l l be topped o f f w i t h super f lu id . Recent super f lu id t opo f f s performed a t B a l l Aerospace D i v i s i o n on the super f lu id helium dewar f o r the I n f r a red Astronomical Sa te l l i t e (IRAS) have demonstrated t h a t essent ia l l y 100% super- f l u i d f i l l can be achieved.

125

The requirement t h a t the l i q u i d helium be he ld a t a temperature and va o r pressure below 2.17K and 38 t o r r (.O5 a h ) , respect ive ly , ( the super f lu id t rans i t i on ! means t h a t essent ia l l y continuous pumping must be maintained u n t i l launch. Pumping through long umb i l i ca l vacuum l i n e s t o overboard pumps i s no t feas ib le f o r the Spacelab 2 mission, so we must prov ide our own on-board vacuum maintenance system, powered almost cont inuously through an umb i l i ca l e l e c t r i c a l l i n e . This pump must operate s a t i s f a c t o r i l y when the Orb i te r i s hor izonta l and a lso when i t i s i n the v e r t i c a l launch a t t i t ude . Thvough t h i s per iod the equipment i s inaccessible. We are using a commercial, d i r e c t d r i v e labora to ry pump and valves f o r t h i s purpose. The pump was found t o operate s a t i s f a c t o r i l y , over an angular range o f about -50" t o 45O, measured about the shaft ax i s from normal up r i gh t labora to ry a t t i t ude . The pump i s therefore mounted t i l t e d -50" on a bracket; when the Orb i te r i s ra ised t o the v e r t i c a l , the pump i s a t 40" and s t i l l operates.

There w i l l be a few unavoidable power outages f o r the vacuum maintenance assembly as the O rb i t e r i s moved, ra ised t o the v e r t i c a l , mated t o the External Tank and So l i d Rocket Boosters,and r o l l e d t o the launch pad. When the pumping i s thus diasbled, the helium w i l l warm; i t must no t warm above the super f lu id t r ans i t i on . We have estimated t h a t one four- hour power outage and several shor ter ones can be to lerated, i f adequately separated t o g ive the system t ime t o recool. Pumping w i l l be continued u n t i 1 a few minutes before launch. During ascent some heat ing o f the l i q u i d i s expected, mainly due t o the absence of vent f low, Recent v i b r a t i o n t es t s on the co l d IRAS dewar i nd i ca te t h a t very l i t t l e f r i c t i o n a l heat ing o f the super f lu id he1 ium occurs. When the O rb i t e r reaches an a1 t i t u d e a t which t h e atmospheric pressure i s below f i v e t o r r , a ba ros ta t i c swi tch w i l l open a va lve between the experiment vent tube and the overboard vent. Th is w i l l permi t pumping on the super f lu id t o resume several hours before the Spacelab i s act ivated.

We w i l l monitor and con t ro l the cryogenics dur ing the miss ion from the ground i n the Spacelab Payload Operations Control Center (POCC) . Emergency moni t o r i n g and c ~ n t r o l w i l l a l so be poss ib le by the on-board Payload Spec ia l i s ts ( s c i e n t i s t s ) . Sho r t l y before landing the experiment w i l l be stowed, the vacuum cover sealed, and the vent valves closed. The r e l i e f valves on the vent l i n e s w i l l permi t the experiment t o warm up safe ly and untended.

4. Conclusion

The cryogenic system fo r the Spacelab 2 I n f r a r e d Telescope Experiment has been designed as a low-cost approach t o the storage o f l i q u i d helium dur ing the preparat ion of a Spacelab payload f o r launch and. the supply o f adequate coo l ing t o a separate rn f rared Telescope f o r the dura t ion of a t yp i ca l Spacelab mission. The r e s u l t i s a system which w i l l sa t i s f y the s c i e n t i f i c requirements of the IRT and a lso be capable of f u rn i sh i ng coo l ing t o a v a r i e t y o f f o l low-on Spacelab experiments. I t i s expected t h a t as the Spacelab and Space Transpor- t a t i o n System f l i g h t program matures, f 1 i g h t durat ions wi 11 increase and prelaunch preparat ion times wi 11 decrease. The supe r f l u i d he1 ium system descr ibed here in should be capable o f s a t i s f y i n g experiment helium cooled r e f r i g e r a t i o n requirements f o r the next several years.

5. References

[l] Urban, E. W., Katz, L., ,Karr, G. R., and Hendricks, J. B., Spacelab 2 i n f r a r e d telescope cryogenic systems, Society o f Photo Opt ica l Inst rumentat ion Engineers, Volume 183 - Space Optics, 1979.

[2] Karr, G. R. e t a1 . , Cryogenic subsystem performance o f t he i n f r a r e d telescope f o r Spacelab 2, Proceedings Eighth I n t e rna t i ona l Cryogenic Engineering Conference, ICEC-8, Genoa, 1980.

[3] Karr, G. R. and Urban, E. W . , Super f lu id p lug as a con t ro l device f o r helium coolant, Cryogenics , 266-270 (1 989).

Requirements f o r and Sta tus o f a 4 . 2 ' ~ Adsorpt ion R e f r i g e r a t o r Using Z e o l i t e s

W i l l i a m H. Har tw ig Department o f E l e c t r i c a l Engineer ing

The U n i v e r s i t y o f Texas a t A u s t i n Aust in , Texas 78712

Previous work byothe author and h i s assoc ia tes has demonstrated t h e f e a s i b i l i t y o f 77 K adso rp t i on r e f r i g e r a t o g s us ing one N 0 stage which p rov ides t h e adsorp t ion temperature (185 K) f o r N . ~ 8 e con- cep t o f adding an H stage and He stage has been examinid t h e o r e t i c a l l y and exper imen ta l l y 6 s i ng zeol i t e s and/or o t h e r molecu lar adsorbers. The r e s u l t s a r e descr ibed which shows t h e ~ e q u i r e m e n t s f o r each stage and suggests t h a t a re1 i a b l e low-power 4.2 K non-mechanical r e f r i g e r a t o r can be b u i l t w i t h a cons iderab le development e f f o r t and some a d d i t i o n a l research.

Key words : Adsorpt ion; compressor; cryogenics ; molecu lar adsorpt ion; r e f r i g e r a t i o n ; z e o l i t e s .

1. Molecular Adsorption/Desorption Compressors for Cryogenic Gases

It is well known El] the clay minerals called Zeolites adsorb large volumes of such gases as N20, C02 and NH3 at room temperature and pressure. For example, the synthetic sodium aluminum silicate zeolite Nay will adsorb 21% by weight of N20 at room temperature and pressure. The family of adsorption isotherms for N20 and the zeolite Nay is s h m in Figure 1 as an example. The adsorption process is exothermic, generating a heat of adsorption, Hay which is less than the heat of vaporization of the normal liquid. The adsorbed gas density is, however, approximately that of the liquified gas at its normal boiling point. For that reason, addition of heat will release a large mass of adsorbed gas in a small volume, creating a high pressure.

Cryogenic gases with lower boiling points can also be adsorbed with a useful mass fraction if the zeolite temperature is suppressed to about 2-3 times their normal boiling points in absolute degrees. The exact pressure ratios and desorption temperatures for nitrogen,hydrogen,and helium depend upon the choice of zeolite. The practical problem is to find adsorbing media for helium at ZOOK, for H at 77OK, and for N2 at 1850K. A successful N20 refrigerator has been built f 2,3,4] Ghich delivers 150 watts of cooling.

a. Advantages of Adsorption/Desorption Gas Compressors

The Molecular Adsorption Refrigeration System (MARS) is being developed to take advantage of several desirable features; no shaft seals or lubricant contamination, low-tolerance fabrication techniques, no mechanical vibration or acoustic noise background, long mean time to repair or failure, low cost materials, and the ability to use waste heat.

Zeolite NaY

+ T = 2 0 2 O K o T = 240° K x T = 270' K a T = 3004 K * T = 33s0 K

0.0 3.30 5.10 6.90

P R E S S U R E ( A T M )

Fig. 1 Weight ratio of gas t o zeolite, x/m fur N20 ~dsorbed in zeolite

The disadvantages include a lower C.O.P. than mechanical compressors, but this is partially offset by the large compression ratio. To operate with lower boiling point gases than N20 the adsorption temperature must be suppressed below room temperature, but some of the heat can be cycled reversibly with proper design.

b. Theory of Gas Adsorption on Zeolites

The crystal structure of zeolite is that of an open lattice of Al, Si,and 0 atoms with dominant cations that may be Na, K, Cs,or other metals. The open structure is-a three dimensional periodic array of voids, typically 13A in diameter, with 5A openings between adjacent voids. Normally each of these voids fill with 10-20 Hz0 molecules. When heated in a vacuum the water is desorbed, and may then be cooled and filled with polar molecules of other gases. The adsorption of gas molecules on the interior surfaces of the voids is an ionic interaction with a characteristic potential energy for each different gas molecule at each different ion site. The functional relationship of amount of adsorption versus temperature and pressure for various. zeolite-gas pairs is extremely complex, and as such emperical methods have been commonly used to generate adsorption data.

Recent work by Woltman and Hartwig [5,61 has led to a theory of gas adsorption derived from first principles. A lattice model of solutions is the basis of the theory, and it treats adsorbed gas molecules and the vacant sites where there are no adsorbed molecules as a two-component solution. The theory is capable of accounting for the detailed physical behavior over a wide range of pressures, and thereby establishes a reliable formalism to predict behavior and design apparatus.

c. Closed Cycle MARS Compressor Configurations

While many configurations are possible, a typical one consists of four cylindrical pressure vessels filled with zeolite. Each compressor is able to heat and cool the zeolite with a finned heat exchanger, using electric heat and circulating liquid coolants. Each compressor successively adsorbs gas, pres- surizes with addition of heat, desorbs gas, and then cools to adsorption pressure. Figure 2 is a cycle diagram, where the vertical scale is heat power, positive if added and negative if removed.

In cryogenic refrigeration service the desorbed high pressure gas is delivered to a Joule-Thomson orifice or expansion engine, and the low pressure retyn gas is adsorbed. The adsorption cycle is illustrated in Figure 3. Gas is adsorbed at 1 Atm. pressure up to point a, where x/m is 0.15 grams of gas per gram of zeolite. Upon addition of heatFgas is desorbed and pressure rises to 21 Am. at b. Gas is delivered at 21 Atm until x/m reaches c. The cooling interval removes - Keat and the pressure drops to 1 Atm. at d. -

2. Compressor Design Theory

For a given mass flow, , the compressor delivers a fraction, Ax/m,of the adsorbed gas from MZ grams zeolite in the desorption time, td. Hence

To pressurize and desorb the gas, a fixed energy must be added to the compressor per gram of zeolite, or

Fig. 2 Heat Cycle diagram for NARS compressors. Each one, A-D, is always in one of four states; (1) adsorbing gas at low pressure, (2) pressurizing by addition of heat with no flow of gas, (31 desorbing gas under pressure, and (4) cosiin~ the zeolite bed to reach adsorption pre: \ su re .

where K1 is a constant, Cp is the specific heat and AT rise. The average power Input during the pressurizing

K M C A T qdMz = 1 z p Q = - t +t P d

t +t P d

is the zeolite temperature and desorbing intervals is

(3)

Eliminating M, between (1) and (3), the mass flow of gas is proportional to the input power, or

where K2 includes K1 and td/(tp+td).

Pressure ( Atm )

Fig ( 3 ) Adsorption cycle of N20 in 13X with delivery pressure of 21 Atm. Temperature i s

peak value for zeolite at the end of the desorption phase, point "c".

Desorption pressure, Pa, is the dominant gas parameter that links gas flow to cooling power. By referring to the T-s or H-P diagrams of the cryogenic gases in their superheated regions, an emperical relation can be derived.

wherg A is a constant over a wide range of pressure. Equation (5) can be solved for Qo,

The Coefficient of Performance is the ratio of cooling power developed by the stage to the heat removal rate for adsorption, which means we divide Equation (6) by Equation (4).

Good design practice will increase the constant, K, by reducing the relative mass of non-adsorbing materials, such as the compressor shell, using isentropic expansion instead of J-T orifices, and taking advantage of low Cp materials at low temperatures. AS with conventional refrigerators, the COP of a multiple- stage MARS is the product of the separate COP'S.

'MARS compressors would be optimized in part, by selecting the lowest possible adsorption temperature to maximize the total adsorbed mass of gas; which.in turn, insures the highest desorption pressure for a given zeolite. This choice is limited in practice to the boiling points of natural gases, such as N20, N2,and Hz, but more favorable combinations may include Ne; argon, CH4,and COz.

The optimum zeolite for each gas is the one that maximizes the combined COP, not necessarily the individual COP'S. At the time of this report no definitive choices have been made between various synthetic and natural zeolites, and also charcoal.

Table 1 summarizes typical values found so far, for the parameters in the equations above, all using a J-T expansion orifice. The values selected for the various constants are realistic approximations in the present state of the art. The optimum energy budget for any design maximizes pd(~x/m)/~T, and it is noteworthy to point out neither zeolite mass nor the cycle time appear in Equation (7). A critical parameter, concealed in K, is the effect of free volume in limiting the maximum value of desorption pressure.

Desorption pressure can be thought of as a function of gas temperature, free volume (exclusive of the zeolite and adsorbed gas volumes), and the mass of desorbed gas. The adsorption isotherms show the adsorbed mass is almost independent of pressure, so if more gas is "vaporized" by changing the zeolite temperature, the pressure will increase. To the extent this is true, MARS desorption pressure will be enhanced by limiting the volume of the external tubing and valves, and reducing the interstitial volume between zeolite grains.

TABLE 1

Adsorb. Desorb . P aX AT ax/m Stage Gas Temp. OK Temp. OK (atm) K % (OK) COP AT

1. 70° is the temp. reduction load on the first stage. Reduction from 475OK to 2700K is by thermoelectric-cooled circulating fluid.

2. AT for Hz compressor may be less for some zeolites. 3. AT for He compressor can be less since Ax is probably optimum at about 600K

and the zeolite need not be taken to 773. 4. COP of a working N20 stage is experimental data. Energy for desorption was

added by joule heat and dissipated to the ambient.

3. Conclusions

The data in Table 1 is not indicative of optimum performance because it was processed with equations for isenthalpic expansions. Some errors are possible, although the data is consistent with published values[7] for common zeolites. It is reasonable, we believe at this writing, that optimum COP'S with the zeolites tested will be about twice those reported for N20, N2 and He. The very low values for H2 are not consistent with the others, and other zeolites will increase then. As it stands now, a 4-stage M S could have an overall COP of about

This would produce 10 m of cooling at 4.2K with about 10 KW input power.

a. Directions for Future Research

The Woltman-Hartwig model for adsorption and an experimental program for verification is now appropriate to match all the cryogenic gases with particular adsorbers. This will produce a better understanding of the physics of adsorption/ desorption systems, and permit the MARS compressors to achieve the maximum values of a(x/m)/aT by improved design and temperature control.

b. Directions for Future Development

Many avenues for innovation and design remain for increasing the COP of multistage MARS. The Hz stage has the lowest individual COP, and it can be increased by adsorbing at higher pressure, finding more effective zeolites, or substituting other technology. A LaNi5 compressor which utilizes the reversible hydride formation is one possibility now being studied elsewhere 181. It would operate with a room temperature heat sink but needs a very small 77OK stage to get the H2 gas below its inversion point. This could be supplied with a two-stage MARS. If metal hydride compressors prove to be feasible, the marriage of the two new technologies will be straightforward.

Many areas for compressor design improvements must be pursued to optimize the cycle and give proper attention to reliability and maintainability. Work by the author and his associates [9] has led to optimizing the heat exchanger geometry for MARS compressors. Other problems involve the logic of valve networks for routing gases throughout the system, and design of the valves themselves. The best design approach appears to be rotary ganged valves that synchronize the cycle in each stage.

He MARS Compressors 1 watt

LaNi5

Compressors

300 watts

/( I He MARS Compressors

Fig. 4 Schematic diagrams of two configurations of cryocoolers which utilize non-mechanical gas compression. The system (a) has J-T expansion in four stages, and in (b) turbine expansion permits the system to be reduced to a single 1,aNi5 compressor operating against ambient temperature followed be a He MARS adsorption/desorption stage using zeolite or charcoal. Power levels shown are approximate.

The use of simple expansion engines to improve the COP of each stage is an obvious choice. Miniature cylindrical or spherical rotors with gas bearings are possible configurations, but the development effort will be significant. Figure 4 is a schematic diagram of two He cry.ocoolers which compares JT and rotating expansion.

A most important problem,for a multistage MARS design that is otherwise feasible,is heat management. Advantage can be taken of heat leaks and waste heat in hotter stages to provide desorption heat in colder stages. If the optimum gas/zeolite combination is found for a particular stage the final desorption temperature can be controlled to achieve the optimum Pd(AX/m)/bT and COP.

4. Acknowledgements

The work was supported by NASA-LBJ Space Center, Houston, and by Rockwell International,Anaheim,by a contract with the Navy. The author acknowledges their contributions to the work.

5. References

D.W. Breck, Zeolite Molecular Sieves, John Wiley, New York, (1974)

Hartwig, W.H., Steinfink, H., Masson, J.P. and Woltman, A.W., Proc. 1976 Re- gion V IEEE Conf., IEEE Cat. 76CH1068-6Reg5, 1976, p. 80. Hartwig, W.H., "Adv. in Cryogenic Engineering," Ed. by K.D. Timerhouse, Plenum Publishing Co., N.Y., 1978, 23, 435-447. Hartwig, W.H., Woltman, A.W., andMasson, J.P., Proc. 1978 Int'l. Cryogenic

IPC Science and Technology Press Guildford, Surrey GU2 SAW,

Woltman, A.W., Ph.D. Dissertation, The University of Texas at Austin, Austin, Texas, March 1978. A.W. Woltman and W.H. Hartwig, "The Solution Theory Modeling of Gas Adsorption on Zeolites ," ACS Symposium Series 135, pp. 3-25, W.H. Flank, Ed., (1980) A.J. Kidnay and M.J. Hiza, "High Pressure Adsomtion Isotherms of Neon, Hydrogen &d Helium at 760~~" xdv. in cryogenic* ~ngineerin~, Paper K-5; pp. 730-740, (1978), K.D. Timerhaus, Ed., Plenm, New York. William Steyert, Los Alamos Scientific Labs., (personal communication of a paper submitted for publication) . R.M. Flores and W.H. Hartwig, "Optimization of Heat Exchangers. in a ,Molecular Adsorption Refrigeration System," Rpt. No. TSR-3, Electrical Engineering Dept., The University of Texas at Austin, Austin, Texas 78712, @lay 1980).

COOLING OF SQUID DEVICES BY MEANS OF LIQUID TRANSFER TECHNIQUES

FOR REDUCED HELIUM CONSUMPTION AND ENHANCED TEMPERATURE STABILITY

El don A. Byrd

Naval Surface Weapons Center White Oak, S i l v e r Spring, Maryland 10901

R. G. Hansen

R. G. Hansen & Associates Santa Barbara, Ca l i f o rn i a 93101

I n an e f f o r t t o es tab l i sh a s ta te -o f - the -a r t SQUID device f o r both labora to ry and f i e l d appl icat ions, a developmental e f f o r t has been undertaken by Naval Surface Weapons Center t o u t i l i z e a l i q u i d helium t r ans fe r system t o perform coo l ing o f SQUID devices i n the temperature region o f 4.5 t o 16 Kelvin. Various techniques have been employed t o enhance the temperature s t a b i l i t y o f the inherent system as we l l as t o provide increased system por tab i 1 i t y and f l e x i b i 1 i t y o f appl i ca t i on .

The s p e c i f i c r e f r i g e r a t i o n system used i n t h i s program has been f u r t h e r modi f ied f o r expanded app l i ca t ions i n the low temperature t e s t - i n g o f i n f r a r e d sensors, low noise amp1 i f i ers, and o ther devices.

This p a r t i c u l a r developmental e f f o r t has r esu l t ed i n a system which al lows the user t o func t ion i n e i t h e r a hor i zon ta l o r v e r t i c a l o r ien ta t ion . Operation had p rev ious ly been l i m i t e d t o s i ng l e ax i s o r i en ta t i on as a r e s u l t o f l i q u i d cryogenic systems.

This paper presents a concept o f u t i l i z i n g s ta te -o f - the -a r t cryogenic systems t o i n t e r f ace w i t h d iverse technica l needs i n the area o f e lect ro-opt ics , SQUID devices, and support ing e l ec t r on i c accessories.

Key words: Cryogenics, helium, in f ra red , SQUID, temperature sensors

This presentat ion, a1 though o r i g i n a l l y intended t o be a r epo r t on actua l performance r e s u l t s o f a funct ional system, i s r espec t f u l l y submitted as an i n t e r i m repo r t on " r esu l t s t o date1' o f a r ecen t l y de l ivered system designed f o r both labora to ry and f i e l d operat ion of SQUID devices.

I n the i n i t i a l d e f i n i t i o n o f the cryogenic support requirements, considerat ions such as package size, system e f f i c i ency , cos t o f operat ion, and p o r t a b i l i t y were def ined as primary ob ject ives. The spec i f i c func t ion f o r the system was t o perform the necessary r e f r i g e r a t i o n t o operate SQUID magnetometers i n both labora to ry and f i e l d experiments.

Secondary app l i ca t i on such as i n f r a r e d detector cool ing, sample coo l ing f o r cryospec- troscopy, and the t e s t i n g o f low noise amp l i f i ca t i on systems were a lso envisioned f o r f u t u re use w i t h the basic cryogenic system.

A r e f r i g e r a t i o n capaci ty o f 0.250 watts a t 4.2 Kelv in was the requi red r e f r i g e r a t i o n capaci ty w i t h a goal o f 2 wat ts a t 4.2 Kelvin. Temperature i nd i ca t i on and con t ro l over the, e n t i r e range o f 4.2 Kelv in t o 300 Kelv in was necessary w i t h des i red s t a b i l i t y o f f 0.001 Kelv in over the spec i f i c reg ion o f 4.2 t o 16 Kelvin.

Re f r ige ra t ion techniques such as closed cyc le r e f r i g e r a t i o n systems, t r a d i t i o n a l dewars, and l i q u i d t r ans fe r techniques were reviewed.

With a l l des i red parameters being considered i n view o f ava i lab le systems, the A i r Products and Chemicals' Model LT-3-110 Hel i -Tran l i q u i d t r ans fe r r e f r i g e r a t i o n system was selected. O f p a r t i c u l a r importance i n t h i s se lec t ion were the fo l low ing ra ther unique features o f t he Hel i -Tran as i t r e l a t e d t o our needs.

1. The Hel i -Tran 's compact s i ze and a b i l i t y t o operate i n e i t h e r a hor i zon ta l o r v e r t i c a l o r i en ta t i on was bene f i c i a l .

2. The "o f f - the -she l f " system was e a s i l y adaptable t o our spec i f i c needs w i thou t cost1 y a1 te ra t ions .

3. Rapid cooldown of samples could be e f f k c t ed w i t h i n 45 t o 60 minutes w i t h cryogenic consumption o f cu 4 one l i q u i d l i t e r o f helium per hour o f operat ion.

4. The "0" r i n g i n t e r f ace o f the Hel i -Tran co l d head was r e a d i l y adaptable t o var ious vacuum shroud conf igurat ions, thus a l lowing v e r s a t i l i t y o f app l i ca t ions w i t h i n vary ing technologies.

5. Temperature con t ro l could be e f fec ted w i t h r a p i d response t ime due t o the " f i n e tune" c o n t r o l a b i l i t y o f the l i q u i d helium t r ans fe r and the co l d head t i p heater c i r c u i t .

The commercially ava i lab le Hel i -Tran LT-3-110 l i q u i d hel ium t r ans fe r system i s shown i n Figure 1 schematical ly. The func t iona l system i s shown i n Figure 2.

I n b r i e f l y summarizing the operat ion o f the Heli-Tran, we provide the fo l lowing:

A spec i a l l y designed f l e x i b l e t r ans fe r l i n e i s inser ted i n t o any standard hel ium storage container. An i n s e r t i o n tube on the dewar end o f the l i n e contains a f i t t i n g f o r pressuring the dewar from 3 t o 5 p s i and incorporates a safety r e l i e f valve. The l i n e i t s e l f cons is ts o f two very small f l e x i b l e tubes, one i ns i de t he other, and a separate r e t u r n tube which acts as a heat sh ie ld . Operation o f the l i n e i s as fo l lows: l i q u i d i s t rans fe r red t o the specimen through the innermost passage o f the two concentr ic tubes a t the pressure of the dewar. To ensure 100% l i q u i d and extremely steady f low i n the center tube, a second stream o f helium i s t rans fe r red a t a s l i g h t l y lower pressure through the annular space between the concentr ic tubes. This lower pressure (and therefore lower temperature) l i q u i d i s a r e s u l t of f l ash ing the l i q u i d t o a s l i g h t l y lower pressure by means o f a c a p i l l a r y tube through which the l i q u i d enters the l i n e from the dewar. B o i l i n g and f l o w ' o s c i l l a t i o n s thus occur i n the annular flow passage wh i le the l i q u i d i n the center tube remains sub-cooled. The l i q u i d t o the specimen i s expanded a t the co ld end through an ad justab le needle o r i f i c e . This regulates the f low r a t e and therefore r e f r i g e r a t i o n t o accommodate varying heat loads. The exhaust vapor from the specimen cools the r ad i a t i on sh i e l d surrounding the specimen and i s then heated as i t vents from the co l d end t o avoid f r o s t i n g . The pressure o f the exhaust helium from the specimen may be lowered by connecting a vacuum pump permi t t i ng temperatures as low as 2 Ke lv in t o be achieved.

The sh i e l d l i q u i d i n the annular space i s vaporized by the heat leak i n t o the l i n e . This cons is ts p r i m a r i l y o f conduction through the support system and rad ia t ion . Radiat ion i s minimized by use o f super insu la t ion around the con- c e n t r i c tubes. The co l d gas from the annular f l ow passage i s returned

through a tube p a r a l l e l t o the insu la ted concentr ic tubes. This f u r t h e r decreases r ad i a t i on and support conduction .losses by using the sens ib le heat o f the sh i e l d f low. This gas i s vented a t the dewar end o f the t r ans fe r l i n e .

Temperature i s monitored and regulated through temperature sensors and a heater mounted on the co l d t i p . The top o f the conduct ive ly ,cooled u n i t has.a 1A" -28 UNF thread i n a h igh conduc t i v i t y copper t i p a t the co l d end which permits a v a r i e t y o f specimen holders t o be e a s i l y interchanged. [I]

1. HELIUM DEWAR 5. VACUUM SHROUD

2. HELIUM CYLINDER 6. VACUUM SHROUD PUMP

3. PRESSURE REGULATOR 7. VACUUM PUMP FOR OPERATION

4. MERCURY MANOMETER, 0 t o BELOW 4.2OK

20" Hg OR PRESSURE GAUGE 8. ACCESSORY FLOW CONTROL PANEL ON HELIUM DEWAR 9. ACCESSORY TEMPERATURE CONTROLLER,

MANUAL OR AUTOMATIC

FIGURE 1

Typical Test Set Up [2]

I n the i n t e r f ace ' o f our p a r t i c u l a r SQUID device t o t he standard Heli-Tran, i t was necessary t o design a vacuum shroud approximately 46" i n leng th so as t o accommodate the requi red e l e c t r i c a l feedthrough. This was accomplished by means o f a two-piece telescoping vacuum shroud.

The heat l o ss o r load f o r the e l e c t r i c a l feedthrough was ca lcu la ted t o be 0.230 wat ts . This heat load was received on the incoming cable by means o f a thermal attachment o f the cable t o the exhaust gas r ad i a t i on s t a t i o n thus reducing the base load on the r e f r i g e r a t i o n system and e f f e c t i v e l y u t i l i z i n g the r e f r i g e r a t i o n provided by the exhausting hel ium gas.

To provide f o r the temperature i nd i ca t i on and con t ro l o f the SQUID, a ga l l i um arsenide diode thermometer was i n i t i a l l y selected. Although no t o f f e r i n g the most favorable

1 38

CRYOGENIC REFRIGERATION SYSTEM

CONTROLLER 7 I V A C U U M PORT

COPPER BLOCK

HEATER T 0 CONTROLLER

F I G U R E 2 .

s e n s i t i v i t y , t h i s type o f thermometer provided the most favorable non-magnetic in te r fe rence f o r the SQUID.

As a secondary means o f temperature monitor ing, an iron-doped gold vs. Chrome1 thermocouple was a lso inser ted i n the co ld head. A t the present t ime we are i n s t a l l i n g a ca l i b ra ted s i l i c o n diode thermometer i n t o the SQUID i n t e r f ace i n a n - e f f o r t t o ascer ta in the u l t ima te s t a b i l i t y o f the system over the temperature reg ion o f 4.2 t o 16 Kelvin.

To perform t he requi red temperature d isp lay and con t ro l , the S c i e n t i f i c Instruments Incorporated Model 3800 Ind ica to r /Con t ro l l e r was used. Th is p a r t i c u l a r instrument i s com- p a t i b l e w i t h e i t h e r s i l i c o n diodes o r ga l l i um arsenide diodes and provides propor t ional heater output spec i f i ed by the programmed temperature se t po i n t .

The s a l i e n t parameters o f the c o n t r o l l e r as i t appl ies t o our primary app l i ca t ion (4.2 K t o 16 K) are as fo l lows:

1. Readout Resolut ion and Repeat ib i l i t y (Kelv in) * .002 K from 1K t o 10K * .010 K from 10K t o 20K

2. With Si-400 S i l i c o n Diode the Readout Accuracy.in Ke lv in i s * .005 K from 1K t o 10K * .010 K from 10K t o 20K

3 . Control S t a b i l i t y a t Fixed Set Po in t w i t h Constant Load and Ambient ( i n Kelv in u n i t s ) * .001 K * .04% o f temperature f o r one (1 ) minute * ,010 K * .04% o f temperature f o r one (1 ) hour

4. Power Output One (1 ) ampere a t 30 v o l t s (maximum f o r 46 ohm 20 wat t heater in tegra ted i n t o Hel i -Tran co ld head. ) [3]

Having had bu t one oppor tun i ty t o f unc t i ona l l y t e s t ' t h e above system, we accomplished a cooldown from 300 Kelv in t o 11 Ke lv in w i t h i n n ine ty (90) minutes a f t e r commencement o f cryogenic t rans fe r . Unfor tunate ly the des i red temperature could no t be achieved due t o a l a t e r discovered vacuum leak i n a braze j o i n t on the vacuum shroud. Temperature s t a b i l i t y , however, over the temperature range o f 11-17 Ke lv in was recorded w i t h a s t a b i l i t y o f greater than 0.01 Kelvin; t h i s represented the GaAs readout accuracy.

At the present t ime the vacuum leak i s being corrected and the add i t i ona l s i l i c o n diode i s being i n s t a l l e d on the i n t e r f ace between the r e f r i g e r a t i o n assembly and the SQUID device .

References

[l] Meier, R. N. and KOURY, A. G., "Cryogenic Systems f o r Research a t Helium Temperatures" A i r Products & Chemicals, Inc., Advanced Products Department, A1 lentown, PA 18105

[2] Operation Manual f o r Model LT-3-110 L i qu i d Transfer He1 i -Tran Refr igerators , A i r Products & Chemicals, Inc., Advanced Products Department, Allentown, PA 18105

[3] S c i e n t i f i c Instruments Incorporated Data Sheet Number 301 277 e n t i t l e d "Model 3800 Temperature Ind ica to r /Con t ro l l e r " .

A CONTAMINATION FREE COMPRESSOR FOR SMALL SCALE

STIRLING REFRIGERATORS

J.G. Daunt*, C. Heiden

Institut fur Angewandte Physik der Justus-Liebtg-Universitat Giessen Heinrich-Buff-Ring 16, D-6300 Giessen, West Germany

Contamination of the working gas appears to be'a major problem encountered in the design of reliable small scale ~txrling cycle refrigerators for cooling cryoelectronic devices over long uninterrupted periods. Seals for the reciprocating shafts of such a refrigerator may be a source for contamination due to lubrication and possible small leaks. The use of bellows provides the potentiality for a hermetically sealed refrigerator. As a first step towards such a system, a bellows compressor was designed. For increased lifetime, mechanical load on the bellows is reduced significantly by keeping the differential pressure acting on the membranes at a low level. This is achieved by placing the bellows in the volume of a conventional piston compressor. Such a compressor system is tried in connection with a three stage plastic displacer unit. First results on the performance of the scheme are reported and discussed together with possible design variations that may lead to a.further reduction of mechanical vibrations and magnetic interference of the cooling unit.

Key words: Bellows; compressor; contamination; continuous operation; cryocooler; low power refrigerator; Stirling refrigerator.

1. Introduction

Low Power Cryocoolers such as developed by Zimmerman et al. [ I ] , [2] , [3] are very attractive for use with cryoelectronic components such as SQUIDS or other devices based on the Josephson effect. Long uninterrupted working periods of such a cooler are desirable not only when needed in continuous measurements but also from the point of view of the rather long cool down times of the order of several hours of the refrigerator. It may for example be more convenient to leave thz cooling unit for a magnetocardio- graphic system using a SQUID running continuously than always turning it on many hours before actual use. Since one of the major problems encountered with the design of a continuously running small scale cryocooler is contamination of the working gas, efforts towards a hermetically sealed system were undertaken using bellows as a means for sealing. This paper reports first experiences made during this work.

2. Basic Low Pow ,er Cryocooler Design

t 0 Vacuum- Pump

Fig. 1. Basic Stirling cryocooler design.

Our Stirling refrigerator is based on a design similar to that of Zim- merman [ I ] . The three stage displacer and surrounding cylinder are made of glass fibre enforced plastic (Cadillac G I 0) with close fit (cf. fig. 1). The lowest stage of the displacer has a bore which is filled with lead shot in order to increase the low temperature heat capacity. The displacer.unit is connected to a small compressor which consists of a graphite piston moving with close fit in a copper cylinder. The compressed gas is cooled by a package of air cooled copper tubes located at the exit of the compressor. The piston shafts of displacer and compressor are sealed with double O-rings. The buffer volume between the O-rings is filled with helium gas of the same average pressure as in the compressor. The whole system is connected via needle valves to a helium supply cylinder.

Running this machine with 5 bar average helium pressure and a speed of Il5rpm with a compression ratio of about 3 : l resulted in an almost linear initial cool down behavior . with a temperature drop $% 120K/h, measured at the bottom of the lowest displacer stage. The final temperature was 24K. A reduction in average pressure by a factor of two then resulted in a further decrease of the lowest temperature towards a-value near 18K. This is somewhat higher than the final temperature that was reached by Zimmerman and coworkers[l] We expect to obtain a lower temperature after optimization of the present displacer unit.

After an uninterrupted running period of three days, a noticeable reduction in performance however took place resulting in a slow temperature increase of the coldest stage. The origin of this was found in the formation of deposits at the lowest displacer stage causing increased friction. A possible source may be small leaks due to the O-ring seals and also due to some remaining poros- ity of the plastic material in the displacer unit that lead to losses of helium gas. They are compensated by connecting the refrigerator at appro-

/

priate locations with high flow impedance (needle valves) to the pressure cylinder of the helium supply. This however requires helium gas of high purity. Otherwise the refrigerator will act as a cryopump on the contaminations inthe gas eventually leading to the observed deposits. The transport of air molecules through the O-ring seals along the greased shafts also has to be considered in this context.

3. Cryocooler Using a Bellows Compressor

Since contamination appears to be a fundamental problem with regard to long uninterrupwd working periods a straight forward solution to it should be the construction of a hermetically sealed system or at least one that comes close to it. As a first approach we started to build a contamination free compressor. A simple well known means to separate different media for instance in pumps is the use of bellows. Using this principle the present set was modified by inserting a bellows into the piston compressor as shown in fig. 2. In addition to being rather simple, this design has the advantage

of keeping the mechanical load on the bellows at a low level which

t to Displacer unit

Fig.2: Sealing Bellows inside Piston Compressor

should be of importance for a long lifetime of the bellows. Except for a difference due to the mechanical stiffness of the membranes, the pressure inside and outside the bellows is identical as long as it can breathe.

The experience with a stainless steel bellows used in the first experiment however was disappointing. Driven with a compression ratio of about 2 its lifetime was about lo5 strokes, terminated by fatigue frac- ture along one of the welded seems.

A lifetime of 1 o8 strokes on the other hand would be quite appropriate. A lifetime exceeding this is quoted for commercial bellows made of teflon. Using such a bellows the compressor has been operated with an average helium pressure of 5 barand without apparent degradation in performance now for a period totaling more than 4 million strokes. Although the compression ratio due to increased dead volume in the compressor was reduced to a value of 2.3 resulting in a somewhat slower cool down rate of about 80K/h, the final temperature again was close to 18K. A problem encountered during the cool down period is the change of average helium gas pressure inside the bellows, which has to be accounted for

by a corresponding reduction of the outside average gas pressure. So far this was done manually, but should be effected automatically in a more elabo- rated design. Much longer testing periods of course are needed before anything definite can be stated about the long term performance of the compressor system, and there may be still other materials superior to teflon for the bellows.

However,.before any long period testing is reasonable, the other weak spot in the system has to be eliminated which is the seal at the displacer shaft. This can be done using the same principle (cf. fig.3). To reduce again mechanical load acting on the membranes the sealing bellows is surrounded by a pressure compensation cylinder connected to the outside volume of the compressor bellows. The time constants for fil- ling the volumes of the displacer unit and of the compensation cylinder should of course have about the same value in order to avoid larger dynamic pressure differences at the second.bellows.

Fig.3. Hermetically sealed Stirling cryocooler.

4. Refrigerator with Pneumatic Bellows Compressor

Instead of using a piston compressor to provide the driving pressure of the bellows, one could em- ploy a pneumatic system as shown in fig.4. The compressor bellows is activated by periodically pressurizing and depressurizing the surrounding cylinder via two valves that are controlled by the crankshaft of the displacer unit. This can be done either mechanically as shown or by magnet valves. W e tried a

Fig.4. Stirling refrigerator with pneumatic bellows compressor

setup with the latter valve type and found that such a system can be operated successfully with practically the same cooling performance as with bellows in the piston compressor. The inherent advantages of such a system are increased flexibility of design which allows for instance to place sources of electromagnetic interference and of mechanical vibration at locations remote from the cold end of the displacer unit.

Acknowledgements: Part of this work was made possible by a Senior US Scientist award granted one of the authors (J.G.D.) by the Alexander von Humboldt Foundation.

5. References

[ I ] Zimmerman, J.E., Radebaugh, R. and Siegwarth, J.D., Possible Cryocoolers for SQUID Magnetometers, Superconducting Quantum Interference Devices and Their Applications. Hahlbohm, H.D. and Lubbig, H., eds., Walter de Gruyter, Berlin (1977) p.287.

[2] Zimmerman, J.E. and Radebaugh, R., Operation of a SQUID in a Very Low- Power Cryocooler, Applications of Closed-Cycle Cryocoolers to Small Superconducting Devices, Zimmerman, J.E. and Flynn, Th.M., eds., US Dept. of Commerce, NBS Special Publication 508, Washington DC 1978, p. 59

[3] Zimmerman, J.E., Cryogenic Techniques for SQUIDS, Proc. of the second intern. Conf. on Superconducting Quantum Devices, Berlin 1980, in print

* Present address: Physics Departmenf Queen's University,Kingston, Ontario, Canada K7L 3N6.

A BI-DIRECTIONAL LINEAR MOTOR/GENERATOR WITH INTEGRAL MAGNETIC BEARINGS FOR LONG LIFETIME STIRLING CYCLE REFRIGERATORS

Philip A. Studer and Max G. Gasser

Space Technology Division Goddard Space Flight Center

Greenbelt, Maryland

Linear motors are well suited to drive Stirling Cycle machines since they reduce the machine dynamics to straight line motion. The first device is being constructed as a dynamic balancer for the Stirling Cycle cooler, which is described in a companion paper at this conference. The motor concepts also applies to Stirling Cycle cryogenic coolers, and its advantages in this application are presented. The permanent magnets, pole pieces and drive coils, which form the stator, are stationary and external to the working fluid. The only moving part is the iron piston, which is not in physical contact with anything except the working fluid. It is magnetically suspended and positioned to within 10 microinches by differential capacitive sensors. This allows the clearance between the piston and cylinder to be small enough to form a clearance seal. The piston is also driven electromagnetically and its axial position sensed capacitively. Heat generated in the motor coils does not enter the working fluid. Gas springs allow a controlled resonant drive to minimize power. The same magnetic structure is used for both the drive function and the magnetic suspension. The electromagnetic design along with the sensing and control implementation are described in the paper. Preliminary test data is presented. Magnetic bearing performance characteristics are discussed.

Key words: Clearance sed; gas compressor; linear motorlgenerator; magnetic bearings; Stirling Cycle.

1. Introduction

The linear motor/generator is the heart of a dynamic balancer which is one element of a. long-life mechanical refrigerator system being developed by the Goddard Space Flight Center. The need for cooling capability in space is growing with the use of larger sensor arrays and larger optics as well as future plans for cooled instruments and superconducting devices. Long maintenance-free life is the prime goal since satisfactory short term missions can be accomplished with stored cryogens and satisfactory refrigeration systems are available where maintenance can be easily provided. The goals of this development are directed toward technological advancement of the state-of-the- art as well as providing an effective balancer. To illustrate, the motor concept was the result of an analysis of a basic limitation of linear moving magnet motors and an effort to minimize the mass of the moving element. It establishes a basis for significant system improvements in future refrigeration systems. The major thrust of this overall development &to eliminate all life limiting mechanical wearout phenomena by employing non-contacting bearings and seals and reducing the number of moving elements to a minimum. The first step in this direction is the line& drive motor which does away with the rotary to linear conversion mechanism. There is no crankshaft, connecting rod, or wrist pin, each of which requires a set of bearings. Also to this end,magnetic bearings, a development pioneered by the GSFC in the early 70's['l for rotary devices, plays a major role. The close clearances imposed by the gas sealing requirements represent a new challenge in bearing design. Good thermodynamic performance requires additional analysis and tradeoffs to operate efficiently with clearance seals. The initial motor analysis rejected some motor types based on highly destabilizing side forces and focused on moving coil and moving magnet d.c. motors. Moving coil motors, which offer lower moviq mass (at the expense of power) were dis- counted because of the need for flexing leads, a reliability hazard. A generalized expression for motor weight and power was programmed (Figure 1) based on fundamental motor equations.

n * lee w

EFFICIENCY

Figure 1. Typical Plot for Linear Motor Sizing

This plot is representative of the way power dissipation and the weight of the motor can be expected to vary for a &en set of parameters meeting a particular refrigerator requirement. This procedure established a minimum weight for a given thrust and efficiency with currently available permanent magnet materials with the clear choice of rare-earth magnets. The flux density of these magnets is one fourth the saturation value of magnetically soft steels. These considerations led to the concept about to be described in which the piston is magnetically soft steel and both the permanent magnet and the drive coils are stationary.

2. Linear Motor Description

A permanent magnet biased electromagnet approach was selected. This allows the use of the highest magnetic flux densities, on which the force generation is dependent. The piston, illustrated in Figure (2) is essentially a solid iron bar with the maximum magnetic flux limited only by saturation of the central cross sectional area. This determines the peak axial force which can be generated. In the final design the two end portions are composed of thinring laminations to reduce magnetic losses.

Figure (3) shows the bare stator structure which is symmetrical both axially and radially. This photograph was taken prior to coil winding to show the structural design. A pair of poles sandwich the permanent magnets at each end and produce a high airgap flux in the major diameter at each end of the piston when it is inserted in the bore. There is some magnetic axial centering force. The permanent magnet and pole assemblies are joined by four bridging elements whose functions will be explained later. A rather thick wall non-magnetic cylinder is interposed between the pole surfaces and the piston.

The magnetic circuit and typical flux paths are seen more clearly in Figure (4). On the right is a cross section of the stator assembly with the piston in place and the motor coil identified.

PERMANENT MAGNETS (8)

CAPACITOR

MOTOR COIL

MAGNETIC SUSPENSION LINEAR MOTOR (AT MID STROKE)

---- PERMANENT MAGNET FLUX PATHS -.-.-. CONTROL FLUX PATHS AND RESULTANT FORCE

Figure 4. Cryogenic Cooler Balancer at Mid Stroke

Note that none of the permanent magnet flux passes axially through the length of the piston when it is at mid-stroke. The flux from the permanent magnets is radially outward in both of the central gaps which are also part of the motor coil circuit. When the motor coil is energized the flux level on one side increases with a decrease on the other resulting in an axial force. A good motor should also be a good generator and this equivalent recipro- cal action is evident in Figure (5).

Here the piston is shown displaced to the left which causes permanent magnet flux to pass down the length of the piston following a path which encircles the motor drive coil. This change of flux occurring in the bore of the coil generates a voltage proportional to the rate of change. The magnitude of this flux change from one end of stroke to the other can be equal to twice the saturation flux level of the piston mid-section. And the mid- section of the piston can nearly fill the bore of the motor coil. Therefore, a highly effective generator and motor action occur. The efficiency of static force generation is the ratio of mechanical power generated to the input power and the differences in the electrical input and mechanical output is dissipation which results in heating. The drive coil resistance can be kept small because of the large volume available for the coil winding and, in contrast to rotary motors, there is no wasted energy in "end turns." All of the copper is effective at all times so that the electromagnetic efficiency can be very high. All of the stator structure in the final design is laminated to reduce magnetic losses. The permanent magnets are Samarium Cobalt, which has a coercive force ten times that of Alnicos, which makes it possible to drive large airgaps.

On the left of Figure ( 5 ) you see that each of the poles is actually four elements forming a cruciform shape around the centered piston. The large magnetic gap of 0 .25 cm is shown to scale. The piston, when radially centered, is at a point of unstable equilibrium. Huge side forces would be developed if the piston were free to move close to the poles in any direction.

The requirements of the long life refrigerator however, demand clearance seals and the piston is physically limited to move less than *0.003cm within the cylinder wall. This means that a relatively thick wall pressure cylinder can be interposed between the moving piston and the magnetic poles of the stator assembly removing all of the windings and magnets outside the pressurized working fluid. This is advantageous from several standpoints; contamination, heat rejection, minimization of dead volume and the structural weight of the pressurized envelope.

149

MAGNETIC SUSPENSION LINEAR MOTOR (END OF STROKE)

---- PERMANENT MAGNET FLUX PATHS

-.-.-. CONTROL FLUX PATHS Figure 5. Cryogenic Cooler Balancer at End of Stroke

3. Magnetic Bearing Description

Referring back to Figure (4) left, note that the permanent magnet flux is always similarly directed radially from each of the four surfaces making up the cruciform shape (in this case inward). The suspension control coils are mounted on the ring which connects the outer ends of the cruciform. When a pair of these coils are energized, the normally symmetric flux distribution is altered with an increase in two sectors and a decrease in the other two. A force is generated in the diagonal vector shown which is proportional to the product of the permanent magnet flux and the control flux[2]. This force is directly proportional to the control ampere turns and, just as in this motor, provides linear control at high efficiency. The efficiency is due to the permanent magnet bias flux level for which no power is required.

Active (servo controlled) magnetic suspension is required to obtain the precision necessary to center the close fitting piston within the cylinder bore. Stiffnesses of the order of 1 X 106 Newton per meter are required to overcome static and gas dynamic loads anticipated while maintaining zero contact (displacements less than 20 micrometers). Also required are means to sense the displacement of the piston relative to the cylinder wall.

4. Sensing and Control

We have selected differential capacitive sensing to measure and control the wall clearances. Capacitive sensing is compatible with the large area short stroke requirements. The piston surface itself acts as the passive plate. Figure (6) illustrates the etched capacitor plates which are formed on the inner bore of the cylinder.

There are eight of the smaller plates encircling each end of the piston. Four of these are electrically common and are excited by an A.C. source in the mega Hertz range. The other four form two p,airs of sensors whose coupling varies as a function of piston clearance. The inverse action in opposing gaps linearizes the output. A simple demodulation circuitF3] developed by H. Machlanski of Mac Bar Mechanisms, Inc., has been found to be a very effective signal source. Figure (7) shows schematically the four identical radial control channels and a diagram- atically similar axial drive servo.

150

DESIGN CONSIDERATIONS FOR MICROMINIATURE REFRIGERATORS USING LAMINAR FLOW HEAT EXCHANGERS

W. A. L i t t l e

Physics Department, S tan fo rd U n i v e r s i t y Stanford, CA 94305

Conventional counter f l o w hea t exchangers a re designed t o operate

w i t h t h e f l u i d streams i n t u r b u l e n t f low. The r e s u l t a n t m ix ing o f t h e

f l u i d a l lows heat t o be t r a n s f e r r e d f rom the w a l l s o f t h e exchanger t o

the body o f t he f l u i d . I f, however, t h e dimensions o f t h e exchanger a re

s u f f i c i e n t l y smallyenough heat may be t r a n s f e r r e d by conduct ion alone

t h a t ope ra t i on i n t h e laminar f l o w regime i s poss ib le . We discuss t h e

design parameters o f a micromin ia ture Joule-Thomson cryogen ic r e f r i g e r -

a t o r ope ra t i ng i n t h i s regime. Such opera t i on o f f e r s t h e advantages of

s i m p l i c i t y o f design and low noise. The l a t t e r i s o f p a r t i c u l a r ad-

vantage f o r t h e c o o l i n g o f cryogenic sensors.

Key words : Micromin ia ture r e f r i g e r a t o r s ; 1 ami nar f l o w heat exchanger;

Jou l e-Thomson r e f r i g e r a t o r s .

We have p rev ious l y discussed t h e problem o f s c a l i n g m i n i a t u r e c ryocoo lers t o micro-

m in ia tu re s i z e ill. As discussed by Daunt 121, t h e pressure g rad ien t i n a hea t exchanger i s

r e l a t e d t o the mass f l o w A o f t h e f l u i d , and t h e diameter d o f t h e t u b i n g through which i t

f lows, by the expression

4; where t h e Re i s Reynolds number (Re = -) and p i s t h e dens i t y and 5 t h e v i s c o s i t y o f t h e

f l u i d . We had shown p r e v i o u s l y t h a t f o r a Joule-Thomson (J-T) r e f r i g e r a t o r w i t h t h e hea t

exchangers ope ra t i ng over t h e same temperature and pressure regime the diameter o f t h e ex- * 0.5

changer tub ing shou ld be sca led so t h a t d = m . Where A i s p r o p o r t i o n a l , o f course, t o the

r e f r i g e r a t i o n capac i t y o f t h e device. Reynolds number f o r t h e f l u i d f l o w i n each p a r t o f . 0.5 t h e exchanger i s then p r o p o r t i o n a l t o m too. Thus, as one sca les t h e device t o sma l le r and

sma l le r s ize,eventual ly one comes t o t h e p o i n t where Reynolds number i s t o o low f o r tu rbu-

l e n t f l o w o f t he f l u i d t o occur. For t h i s reason we have considered t h e p o s s i b l i t y o f de-

s ign ing, a heat exchanger f o r ope ra t i on under laminar f l o w cond i t i ons .

154

It should be noted t h a t tu rbu len t f l ow i s normal ly considered essent ia l f o r the oper-

a t i on o f a heat exchanger f o r the r esu l t an t mixing o f the f l u i d allows heat t o be t rans fe r red

from the wa l l s o f the exchanger t o the body o f the f l u i d . I f , however, the dimensions o f the

exchanger are made s u f f i c i e n t l y small, t h i s mixing no longer i s so important and s u f f i c i e n t

heat may be t rans fe r red t o the f l u i d by conduction alone. Laminar f l ow operat ion i s then

poss ib le w i thou t loss o f e f f i c i ency . I n t h i s paper we consider the condi t ions under which

such operat ion can be achieved.

I n the t r u l y laminar f l ow regime,one can ca lcu la te the v e l o c i t y f l ow p r o f i l e o f the

f l u i d d i r e c t l y from the d e f i n i t i o n o f v i s cos i t y . For a counter flow exchanger, such as t h a t

i l l u s t r a t e d i n Figure 1,the f low follows the contour i l l u s t r a t e d i n Figure 2.

Figure 1. I l l u s t r a t i o n o f counter f low heat exchanger

0

VELOCITY

Figure 2. Ve loc i t y prof i , le i,n counter f low exchanger

155

As t h e aspect r a t i o o /a f o r t h e gas channel becomes larger, t h e edge e f f e c t s f rom t h e

f i n i t e w i d t h o f t h e channel becomes unimportant and the v e l o c i t y contour becomes almost

independent o f p o s i t i o n . I n t e g r a t i n g t h e volume flow through each channel then g ives a

mass f l o w

Again f rom f i r s t p r i n c i p l e s one can c a l c u l a t e t h e temperature p r o f i l e knowing the

thermal c o n d u c t i v i t y o f t h e two f l u i d s and t h a t of t he w a l l separat ing t h e two streams. For

t h e f l o w p r o f i l e of F igu re 2 one ob ta ins a temperature p r o f i l e shown i n F igure 3.

0 AT, ~AT~LAT,-- VELOC l TY AT

F igu re 3. Temperature p r o f i 1 e AT across a counter-cUrrent hea t exchanger f o r t h e f l o w p r o f i l e on t h e l e f t

Assuming t h a t no heat en te rs t h e f l u i d s except through t h i s w a l l , and t h a t t h e s p e c i f i c heat

pe r u n i t mass o f each o f t he two streams i s t he same; as i s t he therinal c o n d u c t i v i t y ; and,

dl = d,,then one can c a l c u l a t e t h e t o t a l temperature drop AT f rom one s i d e o f t h e exchanger

t o t h e o ther . One obta ins

Where KI i s t h e thermal c o n d u c t i v i t y o f t h e f l ~ i d , ~ ~ t h a t o f t h e w a l l , c t h e s p e c i f i c heat P

o f t h e f l u i d , and TI and T2 t h e temperatures o f t h e h o t and c o l d ends. Th is expression has

a very s imple i n t e r p r e t a t i o n . L e t us r e w r i t e i t as

156

The l e f t hand s ide represents the heat ca r r i ed down the exchanger by the movement o f the

f l u i d wh i le the r i g h t hand s ide represents the heat conducted across the exchanger perpen-

d i cu l a r t o t h i s f low. For exchangers o f d i f f e r e n t geometry, the only d i f fe rence i n t h i s ex-

pression w i l l be some numerical fac to rs i n the denominator of (4).

It i s customary t o def ine the e f f i c i e n c y q o f a heat exchanger through the expression

Where the h igh and low temperatures and AT are as def ined as i n Figure 4 and expression (3) .

Figure 4. Temperatures i n the i npu t (T) and r e t u r n (TI) channels o f a counter f l ow exchanger as a func t ion o f distance z along the exchanger

Using (5 ) i n (3) one ob ta ins t h e r e s u l t

I f we assume t h a t t he dens i t y of t he gas i s p r o p o r t i o n a l t o t h e pressure so p = po(p/p0)

then (2 ) can be i n t e g r a t e d over t h e l e n g t h o f t h e exchanger t o g i v e

SCALING CONSIDERATIONS

We cons ider now the e f f e c t s o f s c a l i n g such a laminar f l o w r e f r i g e r a t o r . We assume t h e

r a t i o d2/d l , and t h e thermal c o n d u c t i v i t i e s K I and K Z are kep t constant . If we e l i m i n a t e

rh/o f rom (6) and (7) we f i n d t h a t

So f o r r e f r i g e r a t o r s ope ra t i ng over t h e same pressure regime and having t h e same e f -

f i c i e n c y R shouZd be made proportionaZ t o d2. Note t h a t t h i s express ion i s independent o f

w t h e w i d t h o f t h e channels so t h a t i f we f i x pl , p2, R and r~ then we o b t a i n dl from (8 ) .

Using t h i s va lue i n (7) we can o b t a i n w f o r a g iven mass f l ow . D i f f e r e n t we f r i ge ra t i on ca-

p a c i t i e s can be obta ined by va ry ing t h e w id th o f the, channel w i t h no change o f t h e e f f i c i e n -

cy.

Consider t h e f o l l o w i n g example o f a 25 mW J-T r e f r i g e r a t o r ope ra t i ng w i t h n i t r o g e n gas

a t a b o i l e r pressure o f 3 atmospheres (87K) and an i n l e t p ressure o f 120 atmospheres a t 300K.

I f one consu l t s a T-S c h a r t f o r nitrogen,one f i n d s t h a t f o r a heat exchanger e f f i c i e n c y o f

0.95 a mass f l o w o f about 2.5 mg/sec would be requ i red. L e t R = 5 cm f o r both t h e h i g h and

low pressure channels and assume d2K1/d1~2 << 1 ( a reasonable assumption i n c e r t a i n conf igu-

r a t i o n s ) then we f i n d f rom (8) t h a t dl = 35 microns and f rom (7 ) t h e w i d t h of t he channel

o = 1.1 mm. The pressure drop i n t h e h igh pressure channel w i l l a l s o be g iven by ( 7 ) and i n

t h i s case would be about 0.03 atmospheres. A c a p i l l a r y would then be r e q u i r e d t o drop t h e

pressure from 120 atm t o 3 atm. Again us ing (7) we f i n d t h a t t h i s cou ld be accomplished w i t h

a c a p i l l a r y 5 cm long, 110 microns wide, and 6 microns deep. For t h e f l o w through t h e c a p i l -

l a r y where a l a r g e pressure drop i s r e q u i r e d t u r b u l e n t f l o w w i l l p e r s i s t f o r r e f r i g e r a t o r

designs even o f very small dimensions. For such cases, t h e dimensions should be c a l c u l a t e d

us ing t h e t u r b u l e n t f l o w express ion (1 ) r a t h e r than (2 ) .

This s imple ana lys i s shows t h a t micromin ia ture heat exchangers ope ra t i ng i n the h m i n a r

f l o w regime can have e f f i c i e n c i e s comparable t o t h e i r 1 a rge r macroscopic cousins ope ra t i ng

under t u r b u l e n t cond i t i ons . One should a l s o note t h a t i t i s o n l y the area wR which determi-

nes t h e e f f i c i e n c y o f t h e exchanger i n (6) n o t R alone. Th is i s i n c o n t r a s t t o the case f o r

t u r b u l e n t f l o w where the requirement o f a l a r g e Reynolds number and hence a l a r g e f l o w ve-

l o c i ty makes i t necessary f o r t h e exchanger t o be l ong and narrow. For laminar flow, slow

f l o w o f t he f l u i d ove r a broad su r face can be used w i t h o u t l o s s o f e f f i c i e n c y . Th is s i m p l i -

f i e s the design o f -min ia ture 1 ami na r f l o w exchangers and reduces t h e dimensional cons t ra in t s

i n t h e i r f a b r i c a t i o n . I n addi ti on,the absence o f tu rbu lence i n t h e exchanger should reduce

mechanical n d i se i n t h e device. The mi c r o f a b r i c a t i on techniques descr ibed i n t h e accompa-

ny ing paper by Hollman and L i t t l e [31 t oge the r w i t h a laminar f l o w design should make i t

poss ib le t o b u i l d e f f i c i e n t J-T r e f r i g e r a t o r s an o rde r o f magnitude sma l le r than those de-

s c r i b e d h i t h e r t o .

Support f o r t h i s work was prov ided by t h e O f f i c e o f Naval Research Cont rac t N00014-78-

C-0514.

REFERENCES

[l I L i t t l e , W. A., "Sca l i ng o f M i n i a t u r e Cryocoolers t o Micromin ia ture Size" Proc. NBS Cryocool er Conf.

Ed.J. E. Zimmerman and T. M. Flynn, Spec ia l P u b l i c a t i o n 508 ( A p r i l 1978)

[21 Daunt, J. G., Encyclopedia o f Physics, Ed.S. FlUgge, Vol X I V , Low Temperature Physics I , Spr inger Verlag, B e r l i n (1956).

[3] Hollman, R., and L i t t l e W. A., "Progress i n t h e Development o f Micromin ia ture R e f r i g e r a t o r s us ing Photo1 i thographic, F a b r i c a t i o n Techniques", t h i s issue.

LOW TEMPERATURE REGENERATORS

Lawrence A. Made

Lockheed Palo Alto Research Laboratory* 3251 Hanover Street , Palo A1 to, Cal ifornia 94304

A study was conducted to examine ways of enhancing the heat capacity of regenerators used in refrigerators for achieving refrigera tion below 30'~. I t was found that , due to an increase i n the energy of their surface vibrational modes, t h i n films, t h i n wires, and ultrafine particles will exhibi t a higher specific heat than that which would be normally expected for a bul k sample of the same material.

Based on an idealized refrigerator performance model, i t i s predicted that a Stir l ing cycle refrigerator having the thi . rd stage of i t s . regenerat- o r replaced with an ul trafine particle (22 8 dia. lead particles) matrix whose void volume i s partial ly f i l l ed w i t h mercury, will deliver a t a cooling temperature of ~ O ~ K , 2.75 times as much refrigeration as one with a typical third stage regenerator made u p of 0.005'' dia. lead spheres.

A, table showing the expected improvements i n refrigerator performance from the use of several ultrafine particle configurations i s presented.

Key words : Heat capacity; regenerators ; specific heat; ref rigeration; cryogenics.

1 . Specific heat enhancement

The efficiency of refrigerators for cooling below 20 '~ has been, to date, severely limited by poor regenerator performance. The primary cause of poor regenerator performance a t these temperatures i s the fall-off of the regenerator matrix's heat capacity.

This paper deals w i t h how the heat capacity of the regenerator might be improved.

The model typically used to describe a material ' s specific heat a t constant volume i s : * Author's present aff i l ia t ion. This paper presents work partial ly done while with Sci tech. Corp. I t was supported by AFWAL Flight Dynamics Laboratory, Wright-Patterson AF Base, Ohio, under Contract F33615-78-C-3425. Also presented i s work done a t Lockheed Palo Alto Research Laboratory under IR & D funding.

This model i s based on t h e assumption t h a t t h e mate r ia l i s an e l a s t i c continuum. I n t he case o f extremely l a r g e surface t o volume r a t i o s t h i s assumption i s not accurate. The

number o f l a t t i c e s i t e s and t h e number o f poss ib le v i b r a t i o n modes were s i g n i f i c a n t l y under-

counted. E. W. Mon t ro l l discerned t h i s and der ived a se t o f frequency d i s t r i b u t i o n func-

t i o n s tak ing t h i s i n t o account. Using the same approach and assumptions as Debye, w i t h t he

except ion already noted, he der ived a new s p e c i f i c heat model.

where

and

where S ' i s t he s p e c i f i c sur face area, p i s densi ty, M i s atomic weight, OD i s the Debye

temperature, No i s Avogradros number,and n* i s the average number o f conduction electrons per

l a t t i c e s i t e . The surface con t r i bu t i on would be expected t o become s i g n i f i c a n t on ly a t very

low temperature and f o r l a r g e surface t o volume r a t i o s . The d e t a i l s o f t h i s de r i va t i on a re

beyond t he scope o f t h i s paper, bu t they may be found i n Reference 11J.

A ser ies o f measurements o f t he s p e c i f i c heat o f u l t r a f i n e p a r t i c l e s have been made

over t he years. They have been performed on metals such as pal ladium and vanadium, as we l l

as d i e l e c t r i c s , such as magnesium oxide. They i nd i ca te t h a t Montro l l underestimated t he

surface con t r i bu t i on by a f a c t o r o f between 4 and 6 (Ref. [ I ] ) .

Modifying t he correct ions made by Montro l l t o take t he experimental r e s u l t s i n t o

account, we f i n d

Applying equat ion (3) t o lead and mercury, i t i s evident t h a t t he surface e f f ec t i s

indeed q u i t e s i gn i f i can t (see f i gu res 1 and 2).

I t was assumed i n t h i s study t h a t t h e assumption from which equations (1) - (3) a re

der ived &hat T << eD) was accurate up u n t i l T = eD/8. Thereafter i t was assumed t h a t t he

surface enhancement e f f e c t would decrease t o 0 a t 8d3. No measurement o f u l t r a f i n e

p a r t i c l e s p e c i f i c heat has been attempted y e t i n t he range where t h e asslumption T << OD i s

no t reasonably va l i d . 165

160 I VOLUMETRIC SPECIFIC HEAT I

Figure I . Lead volumetric specific heat. Figure 2 . Mercury volumetric specific heat

There are a number of different 1 a t t i ce dynamics models which have been developed since

the work by Montroll. However, to date, none of them appear to be superior to . the simple model of eq ( 2 ) in predictive value. For th is reason, the Montroll model was modified to

f i t the existing experimental data (eq ( 3 ) ) .

An experimental program examining the specific heat of ultrafine particles up t o 100 '~ could give some insight into the physics of specific heat which would be valuable to future

workers in this f ie ld . A program of this type would allow an empirical modification to be

made on eq (2) which would improve on the accuracy of the predictions made by eq ( 3 ) . In addition, the assumption made as to the degradation of the surface contribution, as a function of temperature, could be eliminated and instead, an empirically based model used to

predict refrigerator performance.

2. Utilization of ultrafine particles

There are two candidate methods for the uti l ization of ultrafine particles, both of which are derived from the powdered metallurgy industry. The f i r s t method i s derived from

techniques used during the formation of sintered parts.

2.1 Green compacts

The f i r s t stage of preparation of sintered parts involves pressing metal powder into "green compact". A t th is stage the metal will hold together b u t the surface welding between particles i s minimal. The extent to which the surface welding occups i s related to the pres-. sure exerted during compaction. The second stage involves heating the green compact to just under the materials melting point. A t th is temperature heavy surface welding takes place, "sintering" the powder together, and thereby increasing i t s mechanical strength. For th is appl ication, heavy surface welding would be undesirable as i t would destroy the discreteness of the particles. An optimum pressure must be determined which will provide maximum strength and density of the element without causing an excessive amount of surface welding.

Once formed into a green compact, gas flow passages will need to be fabricated in order to minimize the pressure drop through the regenerator. The end result would approximate a conventional screen. If the gas flow passages are aligned, pressure drop losses could be kept to a minimum.

I t i s also expected that the he1 ium gas used as the working f luid will contribute t o the regenerators heat capacity. In the model used when predicting the performance of a refrigerator with an ultrafine particle regenerator, the helium.gas was a t a base cycle pressure of 40 atm. I t would seem reasonable that some of the he1 i u m gas would be adsorbed and absorbed, and would thereby contribute to the regenerator's heat capacity. In the case of 228 dia. particles th is effect i s expected to be maximized because the particles

2 are only about 5 l a t t i ce s i t e s across and have a large specific surface area (200 m /g).

Absorption refers to the diffusion of the gas through the host metal's l a t t i c e struct-

ure, thereby forming. a solution. Gas which has been absorbed into the l a t t i c e can be considered to represent, from the standpoint of the gas specific heat, a sol id. Unfortunately,

it i s found that helium will not diffuse through a metal l a t t i c e structure (Ref. [2J). There- fore, there will be no contribution to the regenerator matrfx's heat capacity from the absorption of he1 iurn gas.

Adsorption i s caused by an attractive force which i s exerted in a direction normal t o the surface by the atoms a t the surface of a solid. The attractive force i s a result of Van der b h l S forces which ar ise from dipole interactions between the solid and the gas.

167

This i s re fer red t o as physisorpt ion, and i s t h e on l y k ind o f adsorpt ion which can occur

between an i n e r t gas and a metal. It i s the re fo re found t h a t a t a gas /so l id i n t e r f ace

t he concentrat ion of gas exceeds t h a t o f t he gas no t i n t h e s o l i d ' s immediate v i c i n i t y . The

excess gas a t t h e surface i s re fe r red t o as having been adsorbed. A t t he warm end of t h e regenerator ~ 3 5 ' ~ ) ~ i t i s reasonable t o expect t h a t on l y a few percent o f a monolayer of gas

w i l l be adsorbed. At t h e c o l d end (lo°K), about one monolayer w i l l be adsorbed (Ref. 133,

141). I n terms o f t h e s p e c i f i c heat con t r i bu t i on from t h e adsorbed gas, i t i s best t o

consider it t o be a l i q u i d .

A f u r t h e r con t r i bu t i on t o t he regenerators heat capaci ty comes from t h e hel ium trapped

i n t he pore s t r uc tu re o f t h e green compact which i s no t i n t he gas f l ow stream.

Green compacts do have t he disadvantage o f low mechanical s t rength. It i s uncer ta in

what t he long term effect of gas cyc l i ng would be on a green compact. As i t i s no t present-

l y poss ib le t o conta in p a r t i c l e s smal ler than 200 a, i f any p a r t i c l e should break o f f t he

compact some mig ra t ion would r e s u l t . Th is i s undesirable, as i t i s conceivable t h a t seal

and bearing de te r i o ra t i on could r e s u l t .

2.2 Mercury impregnated green compacts

The second method o f u t i l i z a t i o n a l so invo lves t h e f a b r i c a t i o n o f a green compact;

a f t e r which mercury i s pressed i n t o t h e pore s t ruc tu re . A s i g n i f i c a n t p o r t i o n o f t h e mer-

cury (greater than 50% i f t h e process i s repeated) w i l l be re ta ined i n t h e green compact.

Th is r esu l t s i n t h e ma t r i x having a s i g n i f i c a n t l y h igher heat capacity. Mercury has a

h igher low temperature s p e c i f i c heat than lead, a higher densi ty, and a lower Debye tem-

perature. This r e s u l t s i n a l a rge sur face con t r i bu t i on t o t h e s p e c i f i c heat. This process

i s r e f e r r ed t o as mercury poros imetry.

This approach has several advantages. Pressed i n t o t h e pore s t ruc tu re , t he mercury

w i l l have a h igh surface t o volume r a t i o , and i t i s hoped w i l l thereby evidence a surface

enhancement o f i t s s p e c i f i c heat. The thermal conduction o f t h e compact i s a lso expected

t o be enhanced. This w i l l r e s u l t i n a more e f f i c i e n t heat t r a n s f e r between t h e regenerator

and t he r e f r i g e r a n t gas. The con t r i bu t i on t o t h e regenerator 's s p e c i f i c heat by t he mercury

i s t he most important advantage. Because o f t h e p roper t ies already stated, mercury would be

the p re fe r red regenerator mater ia l i f i t weren' t a l i q u i d a t room temperature; but, by

conta in ing t h e mercury w i t h i n t h e u l t r a f i n e p a r t i c l e pore s t ruc tu re , t h i s d i f f i c u l t y i s

overcome.

It i s expected t h a t some amalgamation w i l l occur between t h e mercury and t h e host

metal, which w i l l r e s u l t i n a much s t ronger s t ruc tu re . This r e s u l t s i n t h e i nd i v i dua l par-

t i c l e s being more t i g h t l y bound t o t he compact s t r uc tu re and w i l l probably so lve t h e

po ten t i a l problem o f p a r t i c l e migrat ion. It i s unknown whether a s u f f i c i e n t boundary w i l l

e x i s t a f t e r amalgamation t o form a wave re f l ec t i on , o r scrambling boundary, f o r t h e low

frequency acoust ic v ib ra t ions , and therefore, whether t h e f u l l Montro l l e f f e c t w i l l be

rea l ized.

3. Anomaly mate r ia l s

An a l t e r n a t i v e t o using u l t r a f i n e p a r t i c l e s would be t o f i n d a mate r ia l w i t h a very

h igh bu lk s p e c i f i c heat. The bes t mate r ia l s which has so f a r been found i s GdRh. It

evidences .a very h igh heat capaci ty compared w i t h lead below 20K. There would appear t o be

a t t h i s t ime no reason why GdRh could no t be formed i n t o .005" d ia . spheres o r i n t o 400 mesh

screen.

Because GdRh has a r e l a t i v e l y h igh Debye temperature; i n ca l cu l a t i ng t he s p e c i f i c heat

of GdRh from equation (3) i t i s found t h a t t he surface con t r i bu t i on i s on ly about 3% o f

i t s t o t a l s p e c i f i c heat. Therefore, i n t h i s repor t , i t i s examined as having been fabr i ca ted

i n t o ,005" d ia . spheres.

GdRh i s except ional i n t h a t i t evidences a broad band s p e c i f i c heat anomaly extending

from 20°K t o below 2 ' ~ (Ref. [5 ] ) . Most mate r ia l s which evidence an enhanced s p e c i f i c

heat do so over a narrow range of temperatures; o r t y p i c a l l y on ly a few degrees. Once t he

temperature p r o f i l e s o f a regenerator a re b e t t e r understood, another approach might be t o

u t i l i z e a number o f mate r ia l s which evidence t h i s s o r t o f enhancement (ca l led a lamda

t rans i t i on -norma l l y caused by a phase t r a n s i t i o n ) i n a segmented approach t o regenerator

design. Examples o f . t h i s t ype of mate r ia l a re DyRh B and (ErHo)Rh B x Y x Y '

4. Regenerator analys is

The r e f r i g e r a t o r performance model i s based on an idea l i zed 3-stage S t i r l i n g cyc le

design. The r e f r i g e r a t o r i s assumed t o have a base cyc l e pressure o f 40 atm and t o operate

a t 600 rpm. The t h i r d stage i s designed t o prov ide .3 watts o f ne t r e f r i g e r a t i o n a t lo0K

w i t h a gross output o f 2.785 wat ts u t i l i z i n g .005" dia. l ead spheres as t he regenerator

matr ix . This w i l l r equ i r e a 167 design NTU value (A) f o r t he regenerator, and a r e f r i g e r -

an t gas mass f l o w r a t e o f ,8077 g/sec. The NTU value i s def ined as

where AT i s the heat t rans fe r area, UAV i s t h e ove ra l l c o e f f i c i e n t of heat t rans fe r and

Cg = hc i s t h e gas f l o w stream capaci ty ra te . The r e f r i g e r a t o r performance model , is des- P

c r ibed i n d e t a i l i n Ref. [I I . The regenerator e f f i c i e n c y was determined using a computer

program based on work by ~amber tson (Ref. [6 ] ) . The program uses t he f i n i t e di f ference

method t o ca l cu l a t e t he regenerator 's temperature d i s t r i b u t i o n and e f f i c iency . It i s

assumed t h a t the re i s no outgassing of the hel ium contained i n t he regenerator 's pores

due t o pressure f l u c t ua t i ons . 169

The results described are functions of regenerator heat capacity only. A1 1 el se , including pressure drop ( 7 psi assumed), i s held constant for a l l the cases.

The cases described u t i l i ze the .00511 dia. lead spheres regenerator matrix as the base1 ine. Case 2 ut i l izes .005" dia. GdRh spheres in the l a s t third of the regenerator, the f i r s t two-thirds (warm end) being made up of lead spheres. Case 3 deals w i t h 22 8 lead particles i n a green compact without gas contributions taken into account. Case 4 is the same as Case 3 except that gas contributions to the resenerator's heat capacity are

taken i n t o account as described in section 2.1. Case 5 i.s a mercury Impregnated 22 # lead particle regenerator without amalgamation. Case 6 i s the same as Case 5 except that i t i s assumed that 75% of the surface specific heat contribution i s los t to amalgamation.

The data shown i s the percent of required refrigeration delivered where the required refrigeration is 0.3 watts a t l o O ~ , the design NTU value (A), the effective NTU value (Ao), where

- e l o - l-E

and the regenerator efficiency (E ) . Note that A. and E are the values for the third stage of the regenerator, not for the entire regenerator.

From the results presented in Table 1 i t can be concluded that:

GdRh substituted for lead i n the cold end of a regenerator gives a significant improvement in refrigeration over a lead sphere regenerator. *

A1 though a significant portion of the 22 8 green compact regenerators heat capacity i s contributed by gas adsorption and gas i n the in te r s t i t i a l pores; lead ultrafines, even without the gas contributions, have a high enough specific heat to give significant improvements i n regenerator efficiency over conventional regenerators. I t can a1 so be seen that i f 1 ead proved to be unsuitable for green compact fabrication, a material w i t h a lower specific heat might be utilized i n the ul trafine particle configurations, and by taking advantage of the gas contributions, s t i l l significantly improve regenerator efficiency. That lead ultrafine particle green compacts which are impregnated with mercury offer the greatest potential gains i n refrigeration.

Although amalgamation might reduce the surface contribution to the regenera-

to r ' s heat capacity, the mercury bulk specific heat i s large enough to pro-

vide significant improvements i n refrigerator performance. I t i s apparent that maximizing the quantity of mercury contained within the small part icle pore structure will result i n optimum regenerator efficiency.

TABLE 1 . A comperison of d i f f e r en t regenerator configurations

% of Required Case No. Description Refrigeration

Del ivered X l o E

1 .005" dia . lead spheres 100% 167 54.6 .9820

2 GdRh spheres + lead spheres 193% 167 74.2 .9867

3 22A lead green compact 24 5% 167 92.5 .9893

4 Case 3 + gas e f f ec t s 253% 167 96.1 .9897

5 Mercury impregnated 228, lead 275% green compact

6 Case 5 + amalgamation base 271 % 167 105.4 ,9906

5. Conclusions

An enhancement of materials spec i f ic heat can be accomplished by the use of u l t r a f ine par t ic les of t he same material. The u t i l i z a t i on of u l t r a f ine par t ic les presents a number of

d i f f i c u l t problems; however, t he improvement in re f r igera tor performance appears t o be s igni -

f i can t enough t o merit experimental research t o determine the mechanical properties of

green compacts in d i f fe ren t configurations and t o make an accurate determination of t h e i r

heat capacity over a wide range of temperatures. This will allow more accurate regenerator

eff iciency predictions t o be made and the f e a s i b i l i t y of u t i l i z a t i on analyzed. Fabrication

of GdRh into .005" dia . spheres would a l so seem merited a s i t presents a re la t ive ly simple solut ion t o immediate re f r igera tor needs.

Par t of t h i s work was sponsored by Fl ight Dynamics Laboratory,

A.F. Wright Aeronautical Laboratories, Wright-Patterson Air Force Base,

Ohio, under contract F33615-78-C-3425, Project 2402, Task 240204, "Aerospace Vehicle Environmental Control." Mr. Ronald mite was the

Project Engineer f o r AFWAL/FIEE. This work was conducted a t Scitech

Corporation, Santa Ana, Ca. The project manager was Mr. Henry Yoshi- mot0 and the principal invest igator was Mr. P. J . Walsh.

The extensions on t h i s work were carr ied out a t Lockheed Palo

A1 t o Research Laboratory, Palo Alto, Ca. and sponsored through Internal

Research funds.

I would especial ly 1 ike t o thank Mr. Jer ry Walsh f o r the help and

guidance he gave me throughout t h i s en t i r e project.

6. References

[I] P. J. Walsh, Low Temperature Regenerator Study, I n t e r im Report, A i r Force F l i g h t Dynamics Laboratory, Wright-Patterson A i r Force Base, Ohio, 45433, AFFDL-TR-79-3099, Aug. 1979,

A0 80892.

[2 ] R. M. Barrer, D i f f us i on i n and Through Sol ids, Cambridge Un i ve r s i t y Press, 1973, Chapters

3-4.

[3 ] R. C. Tompkins, Chemisorption o f Gases on Metals, Academic Press, 1978, Chapters 1 and 3.

[4 ] J. H. de Boer, The Dynamical Character of Adsorption, Oxford, 1953, Chapters 1-4.

[51 K. H. J. Buschow, J. F. O l i j oek , and A, R. Miedema, "Extremely Large Heat Capacit ies

between 4 and IOK," Cryogenics 15, 261 (1975).

[61 T. J. Lambertson, "Performance Factors o f a Periodic-Flow Heat Exchanger," Trans. ASME 80, 586 (1 958). -

J. E. Zimmerman, D. 6. Su l l ivan, R. L. Kautz, and R. D. Hobbs tt

National Bureau o f Standards Electromagnetic Technology D i v i s i on

Boulder, CO 80303

Mater ia ls used i n several low-power cryocoolers are charac- t e r i z e d by very low thermal conduc t i v i t i es and moderate t o l a rge heat capaci t ies a t low temperature. Consequently, thermal equ i l i b r ium times are i no rd i na te l y long f o r measurements o f thermal proper t ies t o be made by the usual quasi-steady-state methods. We have developed a method o f measurement i n which t he thermal conduc t i v i t y and the s p e c i f i c heat are der ived from observations o f the time-dependent temperatures a t two o r more po in ts on a c y l i n d r i c a l sample i n response t o a step f unc t i on of heat appl ied t o one end o f the sample. Some pre l im inary r e s u l t s have been obtained on G-10 epoxy laminated tubing. These t r ans i en t measurements g ive r e s u l t s i n much less t ime than t h e steady-state method, and have the.addi t i o n a l advantage o f being adaptable, w i t h f a i r accuracy, t o t he cryocooler s t r uc tu re i t s e l f , o r t o a separate s t r uc tu re geometr ica l ly s i m i l a r t o the cryocooler.

Key words : Composites ; cryocoolers ; p l a s t i c s ; re f r i ge ra t i on ; regenerators; s p e c i f i c heat; thermal conduc t i v i t y .

I. Int roduc t ion

Measurement o f thermal p roper t ies of p l a s t i c and composite mate r ia l s used i n t he con- s t r u c t i o n o f several low-power-cryocoolers requi res long times i f done by conventional methods. For example, a sample o f nylon whose longest dimension i s 5 cm would r equ i r e something l i k e two hours t o reach thermal equ i l i b r ium a f t e r an i npu t of heat a t one end. TR?s e s t p t e i s based on a thermal conduc t i v i t y K Q, 0.5 mW/cm K and a spec i f i c heat c s 25 tr,J/cm K a t 10 K. Thermal equ i l i b r ium times w i l l o f course be p ropor t iona te ly grea!er f o r special composite mate r ia l s w i t h 1 arge heat capaci t ies . We have there fo re developed a method o f measurement i n which we apply a step func t ion i n heat f l ow t o one end o f a c y l i n d r i c a l sample and observe the propagation o f the t r ans i en t i n temperature as the heat d i f f u ses along t he sample. Thus, thermal proper t ies of the sample mate r ia l a re der ived from measurements made dur ing the i n i t i a l development o f the t rans ien t , r a t he r than a f t e r the t r ans i en t has d ied ou t as i n conventional equ i l i b r ium o r steady-state ac methodsE11. This gives an order o f magnitude reduct ion i n t ime requi red f o r the measurements, no t t o mention the f a c t t h a t both the thermal conduc t i v i t y and the s p e c i f i c heat are der ived from the same se t o f measurements. I n addi t ion, the method can be used f o r rough measurements on the cryocooler s t r uc tu re i t s e l f , the geometry being appropriate.

?Contr ibut ion o f t h e U. S. Government, no t subject t o copyr ight . Work supported by the O f f i ce o f Naval Research. ??Present address: Physics Department, Indiana State Un ive rs i t y , Terre Haute, Indiana.

Trans ien t measurements o f thermal p r o p e r t i e s have been made p r e v i o u s l y on c y l i n d r i c a l samples by a heat pu lse method. Our reasons f o r us ing a s tep f u n c t i o n r a t h e r than a pu lse was t h a t an i d e a l heat pu l se momentar i ly r a i s e s t h e end o f t h e sample t o a h i g h temperature, thereby enhancing t h e p o s s i b i l i t y o f heat l oss by extraneous mechanisms such as r a d i a t i o n , and a l s o comp l i ca t i ng t h e ana lys i s through t h e temperature dependence o f t h e thermal con- d u c t i v i t y and s p e c i f i c heat. The magnitude o f t h e s tep f u n c t i o n can be ad jus ted t o g i v e a slow, steady, and r e l a t i v e l y smal l , increase i n temperature o f t h e end o f t h e sample.

11. Ana lys is o f t h e Method

The i d e a l geometry i s a un i fo rm s e m i - i n f i n i t e bar w i t h a heater on t h e f r e e end, x = 0, and thermometers t o measure temperatures T and T2 a t two p o i n t s x and x (Fig. 1 ) . The s imp les t i n i t i a l c o n d i t i o n i s t o l e t t h e bdr come t o thermal equilibrium i t a temperature T , and then a t t ime t = 0, t u r n on t h e heater a t a power l e v e l P. The one-dimensional d i f f u s i 8 n equat ion f o r t h i s s i t u a t i o n i s

where k and C a r e r e s p e c t i v e l y t h e thermal c o n d u c t i v i t y and s p e c i f i c heat per u n i t volume of t h e m a t e r i a l . The s o l u t i o n f o r t h e s t a t e d i n i t i a l c o n d i t i o n i s :

where e r f c means t h e complementary e r r o r f unc t i on , and A i s t h e cross-sect iona l area. Th i s

s o l u t i o n i s p l o t t e d as a f u n c t i o n o f x and o f t i n F ig . 2.

HEATER I Y'

Fig. 1. Experimental arrangement f o r thermal p r o p e r t i e s measurement. The heater i s t u rned on a b r u p t l y a t t = 0 ( s t e p f u n c t i o n ) and t h e r e s u l t a n t temperature v a r i a t i o n (sensed by t h e two thermometers) prov ides t h e means f o r determin ing t h e thermal c o n d u c t i v i t y and s p e c i f i c heat. The r i g h t hand end i s connected t o a heat s ink , t h e temperature o f which i s c o n t r o l l e d t o s e t t h e e q u i l i b r i u m va lue o f t h e temperature a t To. The sample i s suspended i n a vacuum and surrounded by a l i q u i d he l ium coo led shroud t o min imize heat leaks t o it. The leads f rom t h e carbon r e s i s t a n c e thermometers a r e heat sunk t o t h i s shroud f o r s i m i l a r reasons.

DISTANCE. xl(2&75

T IME, t ~ 4 k / C k 2 )

F ig . 2. Spat ia l ( a ) and temporal ( b ) so lu t ions of d i f fus ion equation.

175

Our method of measurement i s t o record the temperatures, TI and T ,as a func t ion of time. his gives a p a i r of curves o f the shape of Fig. 2b, bu t w i t h d f f f e r e n t t ime scales s ince T1 changes more r a p i d l y a f t e r the power i s turned on than T2, and o f course T2 i s always l ess than TI. From these curves we can read off a p a i r o f numbers TI and T2 f o r a p a r t i c u l a r t ime ti. For t h a t p a r t i c u l a r t ime we then have both coordinates (T (TZ,xZ) . f o r two po in ts on the curve o f Fig. Za, which estab l ishes the ho r i zon ta i :ad and v e r t i c a l scales o f t h a t curve. This process invo lves use of t h e r a t i o s x /x and T /T2 and an i n t e r a c t i v e procedure (using a computer) which locates t he two unique Jo i8 ts having the cor rec t values o f these r a t i o s . The r a t i o , k/C, i s obtained d i r e c t l y from t he hor i zon ta l sca le f ac to r and the product kc can be der ived a t x = 0 from t he v e r t i c a l sca le factor and equation (2). The combination of t h i s r a t i o and product y i e l d t he values o f k and C.

It i s on l y essent ia l , o f course, t o measure T and T a t some convenient t ime t. How- ever, measuring T and T2 as a func t ion o f t ime pr!wides h igh degree o f redundancy which i s usefu l i n checking f o r e r ro rs and inconsistencies i n the method.

Since k and C are usua l l y temperature dependent a t low temperatures, i t i s important t o choose a power l e v e l and t ime scale such t h a t T and T do no t r i s e very much above the i n i t i a l value To. Otherwise, the temperature dApenden& o f k and C must be considered i n equation (2), making the analys is d i f f i c u l t i f no t impossible. The use o f small tempera- t u r e i n t e r v a l s i s standard p rac t i ce i n s p e c i f i c heat and thermal conduc t i v i t y measurements.

S'ince the bar i s no t semi - in f in i te , t he measurement t ime ti should be chosen small enough t h a t no appreciable amount o f heat reaches the f a r end o f the bar, otherwise the spa t i a l and temporal p r o f i l e s given i n Fig. 2 w i l l no longer be accurate. Since the spa t i a l p r o f i l e , Fig. Za, approaches zero very qu i ck l y a t l a rge x , i t seems sa t i s f ac to r y t o choose x ( the pos i t i on a t which T2 i s measured) a t about the middle o f the r ea l bar, and then chgose ti such t ha t the r i s e i n T i s no more than about 25% o f the r i s e i n T . It i s assumed t h a t the pos i t i on x should Be near the f r e e end o f the bar. Any deviation of the temporal p r o f i l e from the h r v e of Fig. 2b might i nd i ca te t h a t one o r more o f these condi t ions has been v io la ted.

6.0 r

TIME (SECONDS) F i g . 3. Temperature as a func t ion o f t ime f o r two thermometers on a G-10 epoxy f ibe rg lass

rod. The thermometers were 1.2 cm and 4 cm from the heater. The rod had a cross- sect ional area o f 0.895 cm .

176

111. Results

F igure 3 shows the temperature a t 2 thermometers on a G-10 epoxy-f iberglass r od as a func t ion o f t ime a f t e r a step func t ion was appl ied t o the heater. The rod area was .895 cm and the thermometers were loca ted 1.2 cm and 4.0 cm from the heater.

We found t h a t t h e consistency o f the r e s u l t s was poor i f we se lected a t ime t f o r which t he temperature change a t the second thermometer was small. This i s presumably r e l a t e d t o problems o f r eso l u t i on i n the temperature a t t h i s po in t . We a lso found t h a t t he r e s u l t s va r ied considerably i f t he heat wave was al lowed t o reach the heat s ink ( t he temperature con t r o l l ed end o f t he bar) before the temperature po in ts were taken. This c l e a r l y r e l a t es t o a breakdown i n t h e assumption t h a t t he rod was semi - in f in i te .

For the data o f Fig. 3, reasonable consistency i n r e s u l t s was achieved over the t ime i n t e r v a l o f 40 t o 75 seconds. While the temperature d i f ferences are a b i t l a rge r than desired, we used the method described i n the previous sect ion t o determine the thermal conduc t i v i t y and spec i f i c heat o f t h i s mate r ia l a t 5 K.3 The thermal conduc t i v i t y was found t o be 1.1 mW/cm-K and the spec i f i c heat was 4.7 mJ/cm -K.

These numbers f a l l w i t h i n the range o f values given by o ther inves t iga to rs . Addi t ional work must be done t o a r r i v e a t a quan t i t a t i ve c r i t e r i o n f o r se lec t ion o f v a l i d data resions. The measurement scheme i s t o be automated so as t o provide f o r r ap i d acqu is i t i on o f informa- t i o n over a modest temperature range fo r a va r i e t y of mater ia ls . One should note t h a t t h i s technique has been devised t o deal w i t h r a p i d eva luat ion o f a wide range o f mater ia ls . An accuracy of 10-20% should su f f i ce f o r such a study and t h i s method seems w e l l su i t ed t o the task.

IV. References

[I I. G. C. Danielson and P. H. Sidles, Thermal D i f f u s i v i t y and Other Non-Steady-State Methods, i n Thermal Conduct iv i ty, ed i ted by R. P. Type, 149-200 (Academic Press, New York,NY, 1969).

GAS HEAT SWITCHES

Emanuel Tward

J e t Propulsion Laboratory 4800 Oak Grove Drive

Pasadena, Cal i fo rn ia 91 103

Thermally operated gas heat switches for use i n re f r igera t ion

systems operating in t he temperature range 4K to 300K have been de-

signed and constructed. The switches a r e operated by a miniature ad-

sorpt ion pump which when heated supplies gas t o the switch, turning

i t on, and when cooled removes gas from the switch, turning i t o f f .

The switches have been operated a t 4K, 77K, and 300K with He o r H p

gas adsorbed on e i t he r charcoal o r z eo l i t e 13X. Switching times as

low as a second and switching r a t i o s of 80 have been achieved in

miniature devices whose to ta l mass i s 8 grams.

Key words: Cryogenics; gas adsorption; heat switch; r e f r i ge r a to r s .

1 . Introduction

In order t o cool sensors in spacecraf t , re f r igera t ion systems

covering a broad range of temperatures from l e s s than 1K to ambient

and having service l i fe t imes of up to 10 years a r e being designed and

bu i l t .

Some of these systems require thermal switches both fo r switching

the load to the re f r igera tor and as an essent ial component of the re -

f r i ge r a to r s themselves. Since the primary goal being addressed i s

one of 1 ifetime and re1 iab i l i t y with the potential f o r miniatur izat ion,

i t was decided to develop a gas heat switch, Since the switching

function i s accomplished by changing the gas pressure between two

surfaces, a gas supply and a pump a r e required. Both of these functions

a r e provided by an adsorption pump connected t o the switch. This

switch, together with i t s adsorption pump operator , form a permanently sealed system with no moving parts . This permanent seal ing, to-

gether with the absence of mechanical motion, should provide the re-

l i a b i l i t y and long 1 ifetime which a r e required.

2. P r i n c i p l e o f Operat ion

The thermal conductance between two p a r a l l e l surfaces o f area A

a t temperatures TI and Tp separated by a d i s tance d depends on t h e gas

pressure between t h e surfaces (1,2). A t "h igh" pressures where t h e

mean f r e e path L <<d ( c l a s s i c a l conduct ion r e g i o n ) t h e thermal con- P

d u c t i v i t y o f t h e gas i s e s s e n t i a l l y independent o f p ressure and v a r i e s

a lmost l i n e a r l y w i t h temperature (3) as shown i n F igure 1. A t

" low" pressures where L >>d t h e thermal c o n d u c t i v i t y o f t h e gas i s P

g i ven by (1)

where a i s an accommodation c o e f f i c i e n t f o r t h e s u r f a c e s , ~ = Cp/Cv

i s t h e r a t i o o f s p e c i f i c heats f o r t h e gas, R i s t h e gas constant , P

i s t h e pressure, M i s t h e molecu lar we ight o f t h e gas, and T i s t h e

temperature. I t i s t h e r e f o r e apparent t h a t a t some f i x e d ope ra t i ng

temperature T t h e thermal conductance between two sur faces may

be v a r i e d from some h igh f i x e d va lue i n t h e c l a s s i c a l regime t o some

lower va lue which i s pressure dependant. I n a r e a l swi tch , however,

t h e minimum conductance i s determined by o t h e r thermal conductances

which a r e ope ra t i ng i n para1 1 e l t o t h e conductance through t h e gas.

These conductances a r i s e due t o t h e heat t r a n s f e r by r a d i a t i o n be-

tween t h e two sur faces ( u s u a l l y n e g l i g i b l e i n our a p p l i c a t i o n s ) and

due t o t h e heat t r a n s f e r by conductance through t h e suppor t i ng

s t r u c t u r e ho ld ing the two sur faces apa r t . A schematic o f a des ign

f o r a r e a l heat sw i t ch i s shown i n F igu re 2 . Two h igh c o n d u c t i v i t y

p l a t e s (e.g. copper) o f area A a r e separated by a small gap d and

he ld a p a r t by a low c o n d u c t i v i t y suppor t (e.g. t h i n w a l l s t a i n l e s s

s t e e l tube). The gas gap i s connected v i a a s t a i n l e s s s t e e l c a p i l -

l a r y t o a small adso rp t i on pump. The thermal conductance between

t h e two p l a t e s a t t h e d i f f e r e n t temperatures TI and T2 can be modeled

as i n F igu re 3 (a) , where t h e p l a t e s 1 and 2 have thermal res i s tances R1

and R2, r espec t i ve l y , t h e i s o l a t i o n suppor t has a thermal r e s i s t a n c e Rs 3

t h e r a d i a t i o n r e s i s t a n c e i s Rr,and t h e thermal r e s i s t a n c e due t o

changing t h e gas pressure p i s R(p). I n p r a c t i c e f o r t h e devices which we have made R1, R2<<RS and RrxR(p) , Rs and t h e r e f o r e t h e equ iva len t

c i r c u i t i s adequately descr ibed i n F igu re 3 (b) . Therefore, i f t h e

gas pressure p i s changed from some va lue phigh t o some va lue plow

where R ( ~ ~ ~ ~ ) ~ ~ R ~ , t h e sw i t ch r a t i o i s g i ven as

TEMPERATURE (K) Figure 1. Thermal conductivity of gases as a function of

( 3 ) temperature i n the c lass ica l conduction regime .

CONDUCTIVITY SUPPORT

7 HIGH CONDUCTIVITY PLATE

#:; . ., 1.. , . 6 . *: . . .-:. - - * - 8 , ;@K - . . '

AREA A . - .,.- - ; - , .," ; .;. - -. : .- . ' Z ' . ' . - < . .

*

THERMAL RESISTANCE

R, - PLATE 1 RESISTANCE

- ISOLATION SUPPORT Rs RESISTANCE

Rr - RADIATION RESISTANCE

- VARIABLE PRESSURE R(p) RESISTANCE

Rq - PLATE 2 RESISTANCE

Figure 3. (a) Model of Heat Switch resistances.

(b ) Simp1 ified Heat Swi tcti model.

3 . Adsorption Pump

The adsorption pump ac t s as a source and sink fo r the gas. I t

i s operated by heating and cooling the adsorber which causes gas to

be e i t he r released o r readsorbed respect ively (4,5) . A number of

materials can be used as adsorbers. In a heat switch where t he thermal

conductivity of the gas i s important, the most appropriate gases f o r

use a r e He and H 2 (Figure 1 ). These gases may be adsorbed onto char-

coal o r z eo l i t e f o r example, o r i n t he case of H 2 on t o metal hydrides.

An important fea ture of these adsorbers i s t h a t the adsorbers have ex- 2 tremely l a rge e f fec t ive surface areas per un i t volume ( -500 m [gm)

and hence can s t o r e l iqu id dens i t ies of the gas a t .low pressures. 3 Storage capaci t ies of 200 cm STPIgm adsorber a r e possible. I f t he

temperature i s ra ised, gas i s re1 eased and the pressure builds up.

Since a l a rge amount of gas i s stored on the adsorber and the pres-

sure range of operation i s usually in the mill i t o r r to few t o r r range, therefore orders of magnitude pressure changes a r e possi b1 e with small

temperature changes and with small amounts of adsorber. The principl e

of operation of the switch operator (adsorption chamber) i s shown i n

Figure 4 .

TO SWITCH

Figure 4 . Heat Switch Operator. The adsorption pump i s heated to t u r n the switch on and cooled to t u r n the switch o f f .

182

When t h e heater i s o f f , t h e adsorber i s h e l d a t t h e temperature Tf.

A t t h a t temperature, t h e b u l k o f t h e gas i s adsorbed and t h e pressure

has been ad jus ted t o be plow, s u f f i c i e n t l y smal l t h a t t h e conductance

between t h e p l a t e s i s determined by t h e i s o l a t i o n suppor t i n t h e sw i t ch .

When t h e adsorber i s heated by t h e heater t h e pressure r i s e s t o a

va lue phigh where t h e mean f r e e path L <<d ( t h e gap). For example,

us ing He i n a s w i t c h ope ra t i ng a t 4K, p 1 L~ [ a p ] i s 10-~mm a t 1 t o r r .

Therefore, f o r a gap l a r g e r than, say 10-~mm,a pressure of 1 t o r r i s

l a r g e enough f o r phigh.

The q u a n t i t y o f adsorber and t h e power r e q u i r e d t o opera te t h e

s w i t c h a r e a f unc t i on o f t h e temperature Tf a t which t h e s w i t c h i s

operated. Th is r e s u l t s f rom t h e s t r o n g dependance o f t h e adso rp t i on

c a p a c i t y of t h e m a t e r i a l s on temperature. I f Tf i s near t h e b o i l i n g

p o i n t o f t h e ope ra t i ng gas, then v e r y smal l q u a n t i t i e s o f t h e adsorbers

such as charcoa l and z e o l i t e a r e r e q u i r e d because o f t h e l a r g e s t o -

rage c a p a c i t i e s o f t h e adsorbers. For example, one o f t h e swi tches -1 3 which we have made has a gas volume o f approx imate ly 1 0 cm . A t

3 4.2K zeol i t e 13X adsorbs -1 50 cm STP/gm a t 1 0 -3 t o r r (6) . To pro-

-4 3 duce 1 t o r r i n t h i s volume r e q u i r e s o n l y 1 0 cm STP o f gas o r l o q 4 o f t h e gas s t o r e d on grams o f z e o l i t e .

The power r e q u i r e d t o ope ra te t h e s w i t c h i s a l s o a f u n c t i o n o f

t h e s w i t c h i n g t ime r e q u i r e d . For r a p i d t u r n o f f t ime t h e thermal

r e s i s t a n c e o f t h e thermal 1 i n k must be reduced, thus causing an i n -

creased hea t l o a d a t t h e temperature Tf and increased power r e -

quirement t o t u r n t h e sw i t ch on.

4. Measurements

Measurements were made i n a m i n i a t u r e hea t s w i t c h s i m i l a r t o

t h a t shown i n F igu re 2. The sw i t ch i s 4" diameter x .200" l o n g and con-

t a i n s a gap o f .001". The i s o l a t i o n suppo r t i s a s t a i n l e s s s t e e l

tube 4" diameter x .001" w a l l . The p l a t e s a r e made o f copper. To ta l

mass o f t h e s w i t c h i s 8 grams.

F igu re 5 shows t h e response o f t h e s w i t c h t o t h e a1 t e r n a t e

hea t i ng and c o o l i n g o f t h e charcoal pump c o n t a i n i n g a few mgms o f

charcoa l . The data was taken w i t h one s i d e o f t h e s w i t c h (TI) f i x e d

a t 4.2K. 6.5 mW i s a p p l i e d con t i nuous l y t o t h e l ower p l a t e o f t h e

sw i t ch c a r r y i n g the increased temperature Tp recorded i n F igu re 5.

I n t h i s example, RL i s r e l a t i v e l y smal l so t h a t t h e s w i t c h o f f t ime

can be made f a s t . The He working gas i s adsorbed on t h e a c t i v a t e d

301 I I I 1 I I I I ' 20 - I :26 - 24 - SWITCH NON-CONDUCTI NG

22 - 20-

TIME (sect) ___t

0 20 40 60 80 100 120 140 160 180

F igu re 5. Response o f Heat Switch t o ope ra t i on o f adso rp t i on pump.

charcoal a t Tf=4.2K. A t t=O seconds t h e 6.5 mW i s be ing app l i ed t o

t h e bottom o f t h e heat sw i t ch w h i l e t h e sw i t ch i s non-conducting.

A t t=110 seconds t h e charcoal pump i s heated. W i t h i n 1 second t h e

sw i t ch i s conduct ing. A t t=15O seconds t h e charcoal pump heater i s

tu rned o f f . W i t h i n 1 second t h e bottom o f t h e heat sw i t ch begins

warming i n d i c a t i n g t h a t t h e sw i t ch has turned o f f . The sw i t ch r a t i o s

f o r t h i s sw i t ch were a l s o measured a t 77K. Wi th He as t h e working gas,

t h i s s w i t c h has a r e s i s t a n c e w h i l e on Ron=l 6 K/W. Rof f= l l 60K/W

g i v i n g a sw i t ch r a t i o of 72.5. Wi th H2 as t h e working gas, Ron=15K/W

g i v i n g a s w i t c h r a t i o o f 77.

The research descr ibed i n t h i s paper was c a r r i e d o u t a t t h e

J e t Propu ls ion Laboratory, C a l i f o r n i a I n s t i t u t e o f Technology, under

NASA Cont rac t NAS7-100.

5. References C 1.

[I] R.R. Conte, Elements de cryog;nie, pp. 17-24> Masson e t Cie, Pa r i s (1 970).

(21 R.J. Cor ruc inn i , Vacuum, Vo1. V I I & V I I I , 1 9 (1959).

(31 G.K. White, Experimental Techniques i n Low Temperature Physics, p. 82, Oxford U. Press, Oxford (1968).

, . [4] R.R. Conte, Elements de cryog;nie, p. 68, Masson e t Cie,

Pa r i s (1 970).

151 A.C. Rose-Innes and J.J. Rowland, Journal o f Physics E: S c i e n t i f i c Instruments, 5, 939 (1 972).

J.G. Daunt and C.F. Rosen, Journal Low Temp. Phys., 3, 89 (1 970).

OPERATION OF A PRACTICAL SQUID GRADIOMETER IN A LOW-POWER STIRLING CRYOCOOLER'

D. B. S u l l i v a n , J. E. Zimmerman, and J. T. I ves t i

Nat iona l Bureau o f Standards Electromagnet ic Technology D i v i s i o n

Boulder, CO 80303

A commercial Nb p o i n t - c o n t a c t SQUID has been used i n t h e c o n s t r u c t i o n o f a s imple magnetic gradiometer system supported by a s p l i t S t i r l i n g r e f r i g e r a t o r . The magnetic i n te r fe rence generated by t h e c ryocoo le r was s u f f i c i e n t l y smal l t o pe rm i t t h e a c q u i s i t i o n ( f o r demon- s t r a t i o n purposes) o f a reasonable magnetocardiogram. P r e l i m i n a r y i n v e s t i g a t i o n o f t he source o f t h e r e s i d u a l r e f r i g e r a t o r i n t e r f e r e n c e i n d i c a t e s t h a t c y c l i c temper- a t u r e v a r i a t i o n i s t he l a r g e s t s i n g l e f a c t o r i n f l u e n c i n g t h e gradiometer output . A simple e l e c t r o n i c temperature c o n t r o l l e r has been added t o a d i f f e r e n t c ryocoo le r t o s tudy temperature s t a b i l i z a t i o n . B e t t e r than 1 mK s t a b i l i t y has been achieved a t 8.5 K.

Key words : Cryocoolers; 70w temperature; magnetic gradiometer, r e f r i g e r a t i o n ; SQUID; S t i r l i n g cyc le ; superconduc t i v i t y .


The economic v i a b i l i t y o f superconduct ing i ns t rumen ta t i on would be g r e a t l y enhanced i f p r o p e r l y matched cryocoo lers were r e a d i l y a v a i l a b l e t o support t h e i r opera t ion . By p r o p e r l y matched we mean: 1. c o s t comparable t o o r l e s s than t h e supported ins t rument , 2. c o o l i n g capac i t y n o t s i g n i f i c a n t l y g rea te r than t h a t r e q u i r e d ( o f t e n o n l y a few m i l 1 i w a t t s ) , 3. package s i z e comparable t o o r n o t much l a r g e r than t h a t o f t h e device i n a l i q u i d - h e l i u m c ryos ta t , and 4. v i b r a t i o n and magnetic i n t e r f e r e n c e l e v e l s which do n o t app rec iab l y a f f e c t t h e ins t rument opera t ion . Th i s l a s t requirement i s e s p e c i a l l y c r i t i c a l f o r SQUID magnetometers and magnetic gradiometers which a r e designed t o sense environmental magnet ic f i e l d s o r g rad ien ts (e.g., geophysical16ields o r b i o l o g i c a l f i e 1 ds). The SQUID magnetometer has a s e n s i t i v i t y o f t h e o rde r o f 10 t imes the e a r t h ' s magnetic f i e l d and thus extremely smal l angu lar v i b r a t i o n s t r a n s l a t e d i r e c t l y i n t o i n t e r f e r e n c e i f t h e dev ice i s operated i n such a f i e l d . I n t h i s paper we discuss t h i s and o t h e r p o t e n t i a l i n te r fe rence sources and descr ibe a SQUI'D gradiometer supported by a low-power c ryocoo le r [1,21. As a demonstrat ion o f t h e l e v e l o f i n t e r f e r e n c e f o r t h i s system, a magnetocardiogram was recorded.

2. Cryocool e r

The c ryocoo le r used i n t h i s work has been r e c e n t l y descr ibed by Zimmerman [ ? I . F igu re 1, taken f rom t h i s re ference, shows t h e d e t a i l s o f c o n s t r u c t i o n of t h e c o l d end o f t h i s s p l i t S t i r l i n g r e f r i g e r a t o r . The r e f r i g e r a t o r ope ra t i ng a t 1 Hz has achieved a temperature o f 7K w i t h no heat l o a d (an aluminum r o d occupied t h e ' c o l d space f o r dev ice ' shown i n F igu re 1). The compressor displacement i s 90 cm3 and optimum opera t i ng he l ium pressure i s j u s t over 2 atmospheres (gauge pressure). The d i s p l a c e r s t r o k e i s 0.7 cm.

t C o n t r i b u t i o n o f t h e U. S. Government, n o t s u b j e c t t o copy r igh t . Work supported by t h e Navy Coastal Systems Laboratory. tt Present address: Engineer ing Department, U n i v e r s i t y o f Utah, S a l t Lake G i t y , Utah.

Displacer Aluminum A Flange

-Pinned Displacer

Joint

/ Annular ) Expansion

Space

Solid Nylon Stepped Displacer (in lowest position

\~luminum Spacer (where required)

\ Glass-Epoxy Composite Cylinder Walls

Thin Aluminum Radiation Shields

Aluminized Plastic ?(10-20 layers in

vacuum space)

7 Aluminum End Cap

Cold Space For Device

I; Vacuum Wall

Fig. 1 . Details of 5-stage, split Stir1 ing cryocooler. Dimensions are in millimeters. Figure taken from reference 121.

187

The s e l e c t i o n o f many o f t he design fea tu res f o r t h e c ryocoo le r was i n f l uenced by i n t e r f e r e n c e cons idera t ions. These design fea tu res a re descr ibed i n s e c t i o n 4 below under t h e app rop r ia te p o t e n t i a l i n t e r f e r e n c e sources.

3. SQUID Gradiometer

The gradiometer and i t s p r i n c i p a l dimensions a re shown schemat i ca l l y i n F igu re 2. The diameter o f t h e gradiometer was se lec ted such t h a t none o f t h e s h i e l d s o f i t s suppor t r e - f r i g e r a t o r (F igu re 1 ) would need m o d i f i c a t i o n . Th is was done t o keep t h e system as s imple as poss ib le , b u t t h e r e s u l t i n g gradiometer was n o t p a r t i c u l a r l y sensi t i v e w i t h t h e 1 i m i t e d diameter and base l ine . The commercial t o r o i d a l SQUID i s centered w i t h i n a niobium tube between t h e end c o i l s . Th is tube i s moved v e r t i c a l l y t o achieve a x i a l balance. Small superconduct ing tabs were a t tached t o t h e tube ( i n p a i r s , one a t each end) t o achieve t ransverse balance.

1.59 cm 4

ALUMINUM PLUG (TAPPED) FOR CONNECTION TO COOLER

COIL (10 TURNS)

/ ACCESS TO SQUID

( , EPOXY-GLASS TUBE

Fig . 2, SQUID gradiometer. The SQUID i s i n s i d e a niobium c y l i n d e r .

188

The system was balanced t o approximately 1 x lom4 both ax ia l ly and in the trgnsverse direct ion while operated in l iqu id h e l i u ~ , The s e n s i t i v i t y was found t o be Q 10 T/m per f lux quantum w i t h a noise level of Q 10- T/m in a 40 Hz bandwidth. For thermal anchoring, four s t r i p s of aluminum f o i l (4 mm wide by 0 .3 mm th ick) were t i ed along the e n t i r e length of the gradiometer with the ends a l so ti'ed t o the a1 uminum cold end of the re f r igera tor . The SQUID was a l so independently anchored (thermally) t o the cold end. The gradiometer co i l s ( l a turns each) were connected i n paral le l t o the SQUID ra ther than the usual s e r i e s connection. Thl's was done so t h a t more turns could be used on each coi l and s t i l l have the combined inductance mqtch t R q t of the SQUID co i l .

For the i 'nterference t e s t 3 the gradiometer was operated par t ly within an aluminum shieided room [3 ] , as shown in Pfgure 3. The shielded room provides some i so la t ion between compressor and displacer s ince i t s cutoff frequency i s 113 Hz. The cyl inder with displacer was mounted r i g id ly t o the roof of the room t o ayoid any rocking motion induced by the rotqti'on of the displacer dr ive shaf t (see Fig. 3 ) . This r i g id mounting led t o additional microphonTc noise in t he 3 t o 10 Hz range. Vibrations in t h i s frequency range were generated by some heavy refr igerat l 'on equipment i n the building.

COMPRESSOR A N D MOTOR DISPLACER DRIVE SHAFT

HELIUM GAS LINE CRANKSHAFT

DISPLACER A N D CYLINDER IN VACUUM JACKET

I

I A L U M I N U M ROOM

Fig. 3. Mounting of the cryocooler i n an aluminum shielded room.

4. Potential Sources of Interference

In prevfous papers we emphasized the re f r igera t ion aspects of the cryocooler while a l - l udi'ng t o cer ta in ch'aracteri'stics which would minimize the magnetic s ignature of the device. In t h i s sect ion we expand upon the design features re la t ing t o magnetic interference.

4.1. 13otional Interference

The use of low speed ( i n the neighborhood of one Hz) and the low power requirement re- s u l t in low vibrational noise, because of the low s t a t i c and dynamic forces involved. With the p l a s t i c regenerator mater ials , t h i s low-speed i s , in f a c t , 'essent ial s ince much higher speed operation reduces the thermal penetration depth and thus r e su l t s in regenerator inef- f ic iency, par t icu la r ly in the bottom stage_

The djsplacer requi'res very l iY t l e power f o r operation, so the mechanical components which produce t he d i rp lacer motion can be very smooth in operation. The compressor and motor a r e coupled t o the cold portion of the cooler through a drSve sha f t and a tube through which the helium gas i s transported between compressor and displacer sections. I t i s es- senti'al t h a t v f t ra t ion transmission throu h these two connections be minimized. To t h i s end we use ra tner l i g h t weight and flexibye mater ials (thin-wall brass fo r the dr ive shaf t and small diameter copper tube fo r the helium gas) f o r these components. By separating the compressor and displacer one meter from one another we f ind l i t t l e vibration transmission from compressor t o displacer . Additional f l e x i b i l i t y and v e r s a t i l i t y might be achieved through use of a simple hydraulic system t o transmit the mechanical motion from the motor- compressor t o the displ acer.

The one component of motion which s t i l l remains as a potent ial problem i s t h a t of pressure f lexing of the displacer cylinder. I t i s unlikely t h a t t h i s cylinder i s perfect ly uniform around i t s circumference and the cyc l ic pressure variat ions could therefore r e su l t in some cyc l ic transverse motion of the cold end. This problem would be reduced by operation a t the lowest possible pressure.

An e a r l i e r paper [ I ] described the r e su l t s of operation of a simple SQUID magnetometer in a four-stage cooler. The peak-to-peak s ignature a t the cooler frequency of 15Hz was 0.5 nT (SQUID operated in the e a r t h ' s f i e l d ) . This corresponds t o a motion of 10 radian i f t h i s i s the only interference source, emphasizing the extreme s t a b i l i t y required by t h i s type of instrument. This f i e l d o sc i l l a t i on was dependent upon the or ientat ion of the SQUID in the horizontal plane, so we concluded t h a t a t l e a s t par t of t h a t s ignature was due t o flexing of the displacer cylinder presumably induced by the pressure o sc i l l a t i ons . There was no d i r ec t evidence f o r other interference mechanisms, but they could ea s i l y have been present as well.

For magnetic gradiometers, motional interference i s of ten not as severe. I f the e a r t h ' s f i e l d i s uniform in the v i c in i t y of the device ( the usual case), the motional s ignature i s reduced by the degree of balance of the gradiometer. That i s , i t i s the remnant magnetometer response which picks up the f i e l d motion. In t h i s par t icu la r case the gradiometer i s balanced t o a par t in 10 so motional inteference i s reduced by the same fac tor .

4.2 Compressor Interference

I t i s necessary t o m i n i ~ i z e d i r ec t magnetic interference by the compressor a t the cold end. Two precautions a r e taken. F i r s t , the compressor ;s designed w i t h a minimum of ferromagnetfc components. Materials u t i l i z ed Sncluded brass , aluminum, s t a in l e s s s t e e l , and nylon. We have used bronze and nylon bushings ra ther than ball and r o l l e r bearings which a r e commonly ferromagnetic. The piston i s nylon in a s t a in l e s s tube. The motor and gearbox selected f o r the machine has ca s t aluminum housing components and appears t o have only a modest magnetic s ignature ( t h i s motor was once operated within 15 cm of the gradiometer with no severe in te r fe rence) . The second precaution involves modest separation (1/2 t o 1 meter) of the motor-compressor and the displacer sect ions of the re f r igera tor . I t i s conceivable t ha t some s o r t of shielding could be added t o the motor and compressor t o fur ther reduce t h e i r magnetic s ignature, but t h i s does not ye t appear t o be a major i n t e r - ference source.

4.3. Magnetism of Cold End Materials

P las t ics were chosen f o r the c r i t i c a l cold parts of the re f r igera tor because they a re nominally nonmagnetic. However, a t t he s ens i t i v i t y level of the SQUID, the magnetic proper t ies of v i r t ua l l y a l l mater ials a r e not negligible. Small remanent moments and magnetic su scep t ib i l i t i e s often r e su l t in readi ly measured e f f ec t s . The mater ials concerns f a l l in to two categories , magnetic interference from moving components and slow d r i f t s which a r i s e from temperature dependent magnetic e f f ec t s i n mater ials which a r e fixed r e l a t i ve t o t he SQUID.

In the f i r s t instance two phenomena a r e important. The presence of a remanent magnetic moment in the displacer leads t o an obvious magnetic signature. In general, remanent magnetism in such mater ials i s associated with t races of ferromagnetic substances e i t he r introduced in to the bulk of the material during manufacture, o r qu i te commonly, by machine work during shaping of the component. The l a t t e r contamination can be reduced sharply by careful cleaning procedures. Remaining permanent moments can be reduced by demagnetizing procedures. Even without remanent magnetism, moving components can generate a magnetic s ignature s ince the ambient f i e l d (often the ea r th ' s f i e l d ) induces a moment in anything with a non-zero magnetic suscept ib i l i ty .

I t i s d i f f i c u l t t o be quant i ta t ive in the case of remanence, but we can estimate the s i z e of the induced moment a t the bottom end of the nylon displacer . Assuming fo r the moment t h a t the lowest par t of the displacer produces the s t rongest e f f ec t , we suggest t h a t the displaced volume in the bottom stage i s the relevant volume of nylon t o consider. That i s , we assume t h a t the e f f ec t of cyc l ic axial motion of a semi-inf ini te rod in a magnetic f i e l d i s the same as the cyc l ic appearance of a volume of the rod equal t o the cross sectional area time t h s t roke. Thus f o r the 5-stage cooler of igure 3, we have a volume of 0.7 x 2 5 6 x (2.4) cm . For nylon with a s u j i e p t i b i l i t y of 10- in a f i e l d of 5 x T t h i s yields a cyc l ic induced moment of; % 5 x 7'0- ~ m 2 . This t r ans l a t e s t o an interference gradient f i e l d of approximately 10-l4 Tlm. (The spacing between the end of the displacer and the top coi l of the gradiometer i s % 3 cm). This i s well below any interference level which might be achieved in the near future.

The magnetic suscept ib i l i ty of many mater ials can be a ra ther strong function of temperature leading t o long term d r i f t in the induced magnetic moment. The re f r igera tor components most li.kely t o experience such temperature d r i f t a r e the radiat ion sh ie lds and myl.ar superinsulation. Dr i f t s of many degrees C in temperature a r e 1 i kely s ince radiat ion heat t r ans f e r i s very small and the time constant f o r temperature equilibrium might then be extremely long. Lacking spec i f ic knowledge of e i t he r the magnitude of such temperature variat ion o r the temperature dependence of the relevant material su scep t ib i l i t i e s , i t i s not pract ical t o dwell on t h i s question. Needless t o say, the appearance of d r i f t might eventually lead us t o attemp t o c l a r i f y t h i s s i tua t ion .

4.4. Temperature Variation

Besides t h i s l a s t e f f e c t (induced magnetism which d r i f t s with temperature), there i s a more d i r e c t influence which temperature change can have on a SQUID. I f the c r i t i c a l cur- ren t of the SQUID i s a function of temperature, then temperature change will have an e f f ec t on the operating point of the SQUID. In pr inciple t h i s should not change the output of a SQUID operated in a phase-locked loop, but measurements indicate tha t t h i s i s not the case. This re la t ion between SQUID output and operating point Crf level o r c r i t i c a l current) i s probably a second order e f f e c t re la ted t o some asymmetry in the interference pat tern. I t ' s presence means t h a t temperature changes ( e i t he r d r i f t o r osci 11 atory changes) can a f f ec t the SQUID output d i rec t ly . The cures fo r t h i s problem are twofold. F i r s t eliminate the temperature variat ions and second, eliminate (or reduce) the temperature dependence of the SQUID'S c r i t i c a l current. Probably both solut ions should be considered.

5. Interference Measurement

The gradiometer sensed an interference signal a t 1 Hz (compressor frequency) with an amplitude of % 6 x 10-I' T/m. We conducted several experiments t o i so l a t e the source of the noise. F i r s t , we placed a high permeability shield around the SQUID and found no

observable change in the amplitude of the signature. This immediately eliminates compressor- generated magnetic s ignals as the predominant interference source. Furthermore, t h i s sh ie ld attenuated the e a r t h ' s f i e l d by a t l e a s t a fac tor of 100. The lack of any change in in te r - ference with t h i s shi'elding suggests t h a t both gradiometer motional noise and induced moments (non-zero su scep t ib i l i t y ) a r e a l so uni'mportant a t t h i s level . Final ly , we noted rapid d r i f t when the cooler was turned off momentarily. During such shutdown the S Q U I D remained f u l l y in opeyat-?on. This suggests a strong dependence of SQUID output on temperature as discussed above. From d i r ec t measurements we f ind t h a t the temperature of the bottom end varies by a t l e a s t several mK a t the re f r igera tor frequency and previous measurements on the S Q U I D indicated a strong dependence of c r i t i c a l current on temperature. Thus we conclude t h a t the primary source of t he apparent magnetic s ignature a t t h i s level i s temperature f luc tua t ions a t the SQUID. This i s a l so consis tent with the observation t h a t the signal amplitude i s somewhat dependent on temperature s ince the slope of t he Ic vs. T curve i s a function of T and some dependence i s thus expected.

We a l so noted t h a t the very low frequency noise ( l e s s than 1 Hz) was small. After achieving temperature equilibrium we noted a ra ther s t ab l e signal [less than 2 x 10-11 T/m variat ion over a period of three minutes). This could probably be improved through tempera- tu re regulation which wil l be discussed in the next sect ion.

6. Interference and Dr i f t Reduction

We used a very simple expedient t o cancel most (90%) of the 1 Hz signature. This was t o a t tach a small magnet t o the shaf t of the compressor so as t o generate a f i e l d which j u s t canceled the observed signature. The proper way t o deal with t h i s s o r t of interference reduction i s t o use a commutative f i l t e r with a reference derived from the compressor shaf t . The advantage of such a f i l t e r i s t h a t i t can a l so handle harmonics of the fundamental interference which a r e c l ea r ly present. Nonetheless, the simple cancel lat ion of most of the interference allowed the following measurement.

As a demonstration of the system we recorded two magnetocardiograms (see Figure 4) . The f i r s t was taken with the gradiometer in a simple c ryos ta t and the second with the gradiometer supported by the 5-stage cooler. Bandwidth f o r the measurements was 2 t o 40 Hz ( t o fur ther reduce the one Hz s igna ture) . The added noise in the 3 t o probably a r e s u l t of the r i g id mounting t o the shielded room (see Section

10 Hz range i s 3 ) -

Fig. 4. Magnetocardiograms 5-stage cryocooler

IN CRYOSTAT

IN CRYOCOOLER

taken with gradiometer in a simple cryostat ( a ) and in the ( b ) .

This r e s u l t provides a f a i r l y c l ea r indication t ha t mechanical re f r igera tors can be compatible w l t h SQUID instruments, although ~ i ~ e r e i s room f o r much i'mprovment. Obviously, the problem of temperature f luctuat ion and d r i f t must be addressed f i r s t . (The reduction of SQUID temperature s e n s i t i v i t y will a lso help, but we assume tha t complete elimination of such temperature s e n s i t i v i t y i s very d i f f i c u l t ) .

We have experimented with two means f o r s t ab i l i z ing the temperature of the cryocooler. These experiments have been carr ied out with a 4-stage S t i r l i n g cryocooler, the bottom end of which i s shown in Figure 5. The carbon-resistance thermometer and the heater were used as sensor and control elements in a simp1 e e lec t ron ic temperature regulator (servo system). This system was qu i te e f fec t ive in eliminating d r i f t . Over a period of two hours, the temperature change was no more than a f rac t ion of a mK with the temperature a t 8 .5 K . Because of t he ra ther slow thermal response, the simple regu1atio.n c i r c u i t did not s ign i f ican t ly reduce the 2 m K o sc i l l a t i on ( a t 1 Hz) which was observed without the regulator. Further improvement of t h i s regulation i s c lear ly needed. One should note t h a t a l te rna t ing current a t a frequency above the SQUID sense band can be used fo r both the heater and thermometer t o eliminate any possible interact ion between the SQUID and a d i r ec t current in these elements .

NYLON DISPLACER

EPOXY-GLASS CYLINDER

EXPANSION SPACE

ALUMINUM END PLUG

GAS THERMOMETER BULB (BRASS)

HEATER (10 kR)

CARBON RESISTANCE THERMOMETER

(220 R NOMINAL) - BRASS BLOCK

Fig. 5. Details of the f o r e lec t ron ic thermal mass o

bottom end of the 4-stage cryocooler with thermometer temperature control . The helium gas bulb permits con

f t h f s bottom stage.

and heater t r o l of the

The hel ium gas bu lb shown i n Figure 5 provides a simple means f o r changing the heat capaci ty o f the co ld end o f the r e f r i ge ra to r . Even though the volume o f t h i s bu lb i s on ly 1 cm3, the t o t a l heat capacity o f the bottom stage could be increased by a f a c t o r o f 10 by pressur iz ing i t w i t h hel ium gas t o on ly 4 o r 5 standard atmospheres. This r e s u l t s i n a ten- f o l d reduct ion i n the o s c i l l a to ry temperature v a r i a t i o n and s i g n i f i c a n t l y reduces d r i f t even wi thout the e l ec t r on i c temperature regulator .

7. Summary

I n summary then, a SQUID gradiometer supported by a s p l i t S t i r l i n g cryocooler has been shown t o operate a t an in te r fe rence l eve l s u f f i c i e n t t o ob ta in a magnetocardiogram o f modest c l a r i t y . Temperature f l u c t ua t i ons produce the primary in te r fe rence s ignature and methods f o r temperature s t a b i l i z a t i o n have been studied. We expect t h a t w i t h each improvement which reduces magnetic in ter ference, we w i l l i d e n t i f y o ther in te r fe rence sources which a f f e c t the system a t a lower l e ve l . The most l o g i c a l approach i s t o proceed t o deal w i t h these as they are uncovered.

8. References

[I] J. E. Zimmerman and R. Radebaugh i n "Appl icat ions of Closed-Cycle Cryocoolers t o Small Superconducting Devices," NBS Special Pub l i ca t ion SP-508, J. E. Zimmerman and T. M. Flynn, eds. ) , Superintendent o f Documents, U. S. Government P r i n t i n g Of f ice, Washington, D. C. 20234 (1978). (Also ava i l ab l e from the authors), p. 59.

121 J. E. Zimmerman i n "IC-SQUID (Conference)" Ber l i n , West Germany, May 1980 ( t o be pub1 ished) .

131 J. E. Zimmerman, J. Appl. Phys. 48, 702 (1977).

CURRENT STATUS OF HIGH TEMPERATURE JOSEPHSON DEVICE TECHNOLOGY

M. Nisenoff

Code 6854, Naval Research Laboratory, Washington, D.C. 20375

The fabrication of Josephson devices from refractory thin films with high superconducting transition temperatures (Tc) should yield devices which are rugged and reliable, and capable of withstanding repeated thermal cyclings between room and cryogenic operating temperatures. If the onset of Josephson response could be raised to above 15K, the devices could be operated near 10K at a reduced temperature of 0.6 where the temperature dependences of the relevant superconducting parameters are relatively weak. This temperature region is readily accessible using compact, energy-efficient closed cycle refrigeration systems.

A review of recent progress in the fabrication of various types of refractory high temperature Josephson devices is presentecl. Certain configurations require dimensional resolutions approaching 0.1 micrometer or less for successful operations. In other structures, there are problem areas in the growth of suitable thin films of the superconducting materials of interest and of suitable barrier layers and controlling the properties of the interfacial regions between these layers and the superconducting electrodes. Current progress and accomplishments are summarized and critical areas for future R&D are identified. Based on recent progress, the prospects for preparing all-refractory Josephson devices with onset temperatures exceeding 15K appears very promising.

Key words: Superconducting devices; Josephson effects; proximity effect devices; tunnel devices; weak links; closed cycle refrigeration systems.

1. Introduction

The use of superconductive sensors and circuits as well as cooled circuits employing conventional technologies is slowly gaining acceptance. The recent development of small, reliable and energy efficient cryogenic refrigeration systems operating in the 10-30K region has satisfied the requirements for such electronic circuits as IR detectors and for cryogenic vacuum pumps ("cryopumps") . Such cryocoolers are not particularly suited for superconductive applications as the current generation of superconductive electronic devices are fabricated using niobium (T -9.2K) or lead and lead alloys (T -7.2K) and an environment in the viciniey of 4.2K must be provided. ~gese low temperatures require the use of an additional low tempera- tpre stage for the cryocooler such as a Joule-Thomson stage or a wet expansion engine. These approaches introduce additional complexity, weight, cost

and an associated reduction in reliability and convenience of use, Never- theless, a number of industrial and government laboratories have undertaken programs to develop small, energy-efficient cryocoolers whose charactexis- tics are designed to optimally match the requirements of present generation superconductive sensors and circuits.

An alternative approach to the problem of integrating superconductive electronic systems and presently available cryocoolers is the development of high operating temperature superconductive device technology. If techniques can be perfected for the fabrication of superconductive devices and circuits using materials with superconductive transition temperatures in excess of 16K, operation in the vicinity of 10K would be practical without imposing severe restrictions on temperature stability and present day cryogenic technology might be adequate. In this paper, the current status of high temperature superconductive device, technology will be reviewed, existing problem areas identified,and recommendations made for possible research and development .

2. Background

A partial list of practical superconductors is given in Table 1 (Devices using superconductors with T below 4.2K are not deemed "practical"). The present generation of ~~~erconductive devices use niobium or lead alloy technologies and are usually operated in an environment provided by a liquid helium bath. Thus, these devices are operating at reduced temperatures between 0.5 and 0.6 (see Table 2 ) . At such reduced temperatures, the temperature dependence of material parameters (e.g., superconducting energy gap, critical current, penetration depth, etc.) are relatively weak compared to the behavior close to Tc. If operation at reduced temperatures of 0.6 or lower is acceptable, operation in the vicinity of 10K would be acceptable for devices and circuits fabricated using materials with T in excess of 16K (see Table 11). Thus, we will adopt the definition that hYgh temperature superconductive devices are those that can be operated at temperatures in the vicinity of 10K, at a reduced temperature of 0.6 or less.

The advantages that can be gained by the use of cryocoolers operating near 10K relative to those that operate in the vicinity of 4.2K is illustrated in Fig. 1. For convenience of comparison, each set of data has been nornalized at 10K. The solid line is the "inverse" Carnot efficiency, that is,the number of watts of power at room temperature required to produce one watt of refrigeration at the indicated operating temperature. The data points in Fig. 1 are for characteristics (thermal efficiency, weight, volume and cost) of crycoolers providing one to several watts of cooling at the indicated temperature [I], [ 2 ] . It is not too meaningful to draw lines through a given series of data points as it would be too difficult to correct the data for the wide variation in all the other parameters of the systems. Thus, a shaded region is shown to indicate the trend in these parameters as a function of operating temperatures. It can be seen that although 4K operation is theoretically only 2% times more difficult to achieve, for practical system, a factor of about 10 increase is observed in the parameters such as efficiency, size, weight. and cost. Hence the cryogenic support portions of an integrated superconductive electronic system can be simplified significantly if 10K operation is feasible compared to 4K operation.

Table 1. Properties of Superconducting Materials of Possible Technological Importance

6 8 0 ) Mechanical Substrate Material ( 1 Properties Temperature

Lead/Lead Alloys 7.2 800 Soft -RT Niobium 9.2 40 0 Hard RT+40O0C Mo-Re 15 50 Hard LN +RT NbN 16 5 0 Hard 'u <380°C NbCN 18 40 Hard -5OO0C V Si 17.3 3 0 Hard -8OO0C d , s n 18.3 3 0 Hard -800°C Nb3Ge 23 30 Hard -9OO0C

Table 2. Operating Temperature for Superconducting Devices

Lead and Lead Alloys Niobium NbN Nb3Ge

Transition Temperature 7.2 K 9.2K 16K 23K

Usual Operating Temperature 4.2K 4.2K - - Reduced Temperature 0.58 0.48 I, (0.6 I, (0.6 Projected Operating Temperature - - I, <9.6K $13.8

30

10 THERMAL EFFICIENCY

2 f z i ; a 3.0 z P

z 0.3

0.1

TEMPERATURE IKI

Fig. 1. Variation of selected parameters of closed cycle cryogenic refrigeration systems as a function of operating temperature. The solid line is the inverse Carnot efficiency as function of operating temperature. Each set of data has been normalized to unity at 10K (see text).

In the remaining portion of this paper we will describe recent progress in high temperature Josephson device technology and speculate on the prospects for its eventual realization. Although there are a large number of Josephson configurations, we will limit this review to thin film materials (as contrasted with bulk materials) and to tunnel devices, proximity effect devices, and "weak link" constrictions.

3. High Tc Materials

There is a large volume of literature on the preparation of thin superconducting films with transition temperatures in excess of 15K [3], [4]. However, for the preparation of a given type of device there may be special requirements on certain fabrication parameters such as substrate temperature, film thickness, crystalline grain size, nature and degree of granularity or "inhomogeneityrlI etc., which may warrant additional activity in material research. Some of these aspects of thin film preparation will be identified in the following sections dealing with specific Josephson device configurations.

4. Proximity Effect Devices

A proximity effect Josephson device can be fabricated by locally sup- pressing the T of the host film. The initial demonstration of this concept was by ~ o t a r ~ g and Mercereau [5] who used a very narrow strip of normal material across a film either as an overlay or an underlay to reduce the Tc by the proximity effect between a normal metal and a superconductor. Prox- imity devices have also been fabricated by reducing the thickness of the best film over a local region by anodization 153, radiation damage 161, ion implantation [7], [8], etc, Josephson proximity effect behavior can be observed whenever the length of the region of suppressed T (that is, the dimension along the direction of current flow) is comparable %o a characteristic length which is a function of the normal state parameters. When the temperature is maintained below the T of the contacts (or "banks") but above the TI of the localized region, &ere will be a critical current with a temperatuge dependence that will depend exponentially on the ratio of the characteristic length relative to the physical length of the bridge and on the difference in temperature between the operating temperature and some characteristic temperature which will be close to the suppressed T' of the region. Josephson effects, both dc and ac, can be observed in thiz regime which typically spans a temperature interval of 0.2 to 0.5K.

Proximity effect devices have been fabricated from high T superconducting films using many of the techniques described above. !?he highest temperature of operation is about 17K reported by Palmer, Notarys and Mercereau [9] for Nb3 Sn films in which the Tc of a 1-2 pm long region had been suppressed by about 0.5K from the orlginal T -17.5K by selective anodization. DC SQUIDS made in this fashion showed hear modulation pat- terns with good signal-to-noise ratios although no minimum detectable field change was reported. The devices operated over a temperature range of a few tenths of a degree andwere reliable, did not degrade under repeated thermal cycling, etc., and exhibited the other desirable characteristics expected for devices fabricated from refractory materials.

The proximity effect behavior can be easily observed provided that the characteristic length associated with the proximity region can be obtained with the resolution available from existing fabrication techniques. The primary disadvantage is that the critical current varies exponentially with temperature and that the temperature range of operation in the proximity mode is quite narrow requiring precise control and regulation of the cryogenic environment.

5. Tunnel Device

The "classical" Josephson device configuration is the oxide tunnel device. This configuration has been extensively studied theoretically and extremely good agreement has been obtained between theory and observed behavior of carefully prepared devices. However, there are serious problems associated with the fabrication of refractory high Tc tunnel devices. First, it is difficult to grow native oxides with controllable and reproducible characteristics on refractory metals, and especially on alloy and compounds of these metals [4]. Secondly, almost all of the high Tc alloys and compounds require elevated substrate temperatures (300°C and above) for proper film growth. During the deposition of the top ,,electrode, such elevated temperatures may de rade the properties of the base electrode and may destroy the very thin (-20 % ) oxide barrier region.

Techniques have been developed for reliably growing oxide barriers on the soft superconductors which have very reproducible thicknesses and exhibit very small values of excess current below the energy gap voltage in the I-V tunneling curve [lo]. When these techniques have been applied to niobium and other high T materials, noticeably poorer quality tunnel junctions have been obtained Yll], [12]. This is partially due to the fact that niobium and vanadium, the other refractory material commonly used for high T alloy preparation, have a great affinity for oxygen and hence the surface of the base electrode must be thoroughly cleaned by some technique such as sputter etching or glow discharge before the oxide growth is initiated [13]. These cleaning procedures may mechanically damage a shallow region just below the surface, thus degrading the superconductivity at the surface of the base electrod . Since the coherence length of these high T materials is less than 100% (see Table I), even a very thin surface layef of damage will be evident in the properties of the resultant tunnel device (In the case of the soft, low Tc materials, such as lead and lead alloys, a thin region of damage at the surface of the base electrode does not noticeably affect device operation as the coherence length in these materials is of the order of 1,000~ or greater). Furthermore, in the case of niobium, there are a number of natural oxides, sub-oxides etc., some of which are insula- tors and some conductors and even superconductors (bulk NbO has T -1.5K) [13 ] . Hence depending on the details of the growth process, the oxidg layer grown on niobium and niobium alloys will have varying electrical characteristics.

In the case of the elemental superconductor, niobium, the group at IBM has developed techniques for sputter cleaning the surface of the niobium base electrode, growing an oxide barrier by the Greiner technique and then sputter treating the surface of the oxide layer to minimize the possibility of interaction between surface atoms of the oxide layer and the impinging niobium atoms that form the counter electrode [13]. With these techniques IBM has succeeded in producing fair1.y good 'Nb-Nb 0 -Nb tunnel junctions as gauged by the low value of the excess current ju& %elow the voltage corresponding to the energy gap. However, the characteristics of these devices are still not as good as those obtained for lead alloys devices or for Nb-Nb205-Pb alloy devices.

In the case of high T alloy systems, standard oxidation techniques do not work. When they have geen used, the quality of the device characteristics are quite poor (see upper trade in Fig. 2) [14]. Noticeably improved tunneling characteristics have been obta'ned if the high T base electrode is first coated with a thin layer (-20 3( ) of some materih which readily oxidizes such as aluminum or silicon and this artifical layer is then allowed to oxidize [14], [ E l , [16], [17], [18], [19]. The improvement in the tunneling characteristics of devices with artificial oxide barriers as compared to devices with native oxide barriers is obvious by comparison of the several traces in Fig. 2. The concept of using artificial oxide bar-

riers for refractory Josephson devices was apparently first used by Laibowitz and Mayadas [I51 for niobium devices and has subsequently been adapted by other groups interested in tunneling into high T materials. Using these artificial oxide barriers and soft low T lead of; lead alloy counter electrodes, tunneling characteristics of reasonable quality have been obtained.The detailed mechanism of formation of these artificial barriers is not fully understood [20] as a 202 thick film is obviously not continuous and therefore the nature of the resulting oxide barrier probably is not continuous and "pinhole free." However, once the optimum thickness for the oxidized aluminum or silicon layer has been determined, good quality tunnel junctions with reproducible and controllable thicknesses can be obtained.

0 0 1 2 3 4 5 6 7

VOLTAGE (mV)

Fig. 2. Tunneling characteristics of V Si-Pb junctions for several types of tunneling barriers, natural as well as axidized silicon layers (see Ref. 14)

Good quality tunneling characteristics have been obtained using artificial oxidebarrierson high T baseelectrodes wherethe counter electrode has been a soft material (such &i lead or lead alloys), which can be deposited at or below room temperature [14], [17], [18], [191. The tunneling characteristics are noticeably poorer if high Tc materials are used for the counter electrode. Good high T behavior can only be obtained for the potentially high T materials, it? they are deposited at elevated temperatures - at least 3 5 0 ~ and, in certain cases, at temperatures as high as 1000C. These elevated temperatures can have disastrous effects on the properties of the base electrode and the oxide barrier region thathas been grown on the base electrode. It is well known that oxygen can diffuse readily through niobium and niobium alloys at relatively low temperatures. Diffusion of oxygen can either suppress or destroy superconductivity in the layer into which the oxygen has diffused. If superconductivity is suppressed at the surface of the electrode, the desired abrupt transition between uniform superconducting and insulating regions will be smeared out and the quality of the tunneling characteristic will be degraded, and excess current will be observed below the energy gap voltage. Hence, the deposition of the high T counter electrode for an all-refractory superconducting tunnel device prese%ts a formidable problem.

An attempt to fabricate an all high Tc Josephson device has been reported recently by Tarutani et al. [21], who used Mo-Re as the counter electrode material. The system of Mo-Re, which in the bulk can have a T as high as 12.2K1 can be deposited at temperatures near or below room temp&a- ture and, if the deposition conditions are suitable, the compound Mo Re will be formed with a T as high as 11.2K. They report on the prepara$io6 of tunnel devices with 6 3 ~ i as the base electrode (T -16.4K when deposited at 900°C), an artificial oxidized silicon barrier (nat%ve oxide barriers did not exhibit tunneling) and Mo-Re counterelectrodes. Films of Mo-Re deposited at ambient temperatures did not exhibit superconductivity as interdif- fusion of atoms between the Mo-Re, as it was being deposited, and the remainder of the structure probably occurred as the temperature of the device was raised to about 130C during the Mo-Re deposition. When the deposition was ,made at LN temperatures, the device probably remained below lOOK and Mo Re2 was forme6 which, even after being warmed to room temperature. exhibi2ed superconductivity with a broad transition kxtending from 10.5K down to 8.3K. The article did not contain details about the tunneling characteristics of the devices nor did it indicate the highest temperature at which Josephson effects could be detected.

Another likely candidate for the counter electrode in an all refractory tunnel device is NbN. This material is normally deposited at temperatures in the range of 300-500C and transition temperatures of the order of 1% are obtained for films of thickness of 1000 a or greater. If the device is reactively sputtered at ambient temperature, T is reduced to the vicinity of 10-12K. An attempt to fabricate an all &N tunnel junction has been reported by Shinoki et al. [22], who used an artificial layer of oxidized silicon to form the tunneling barrier. The base electrode was deposited at 500C and exhibited T of about 15K. In order to minimize degradation of the base electrode and %he oxidized silicon barrier region, the NbN counter electrode was deposited at near ambient temperatures. The tunneling characteristics at 4.2K show a distinct energy gap at 4mv which suggests that the energy gap for the upper film is below 10K. Although no specific value is quoted, the excess current below the gap appears to be large and the device as reported was not suitable for switching application.

6. Semiconductor Barrier Devices

Another Josephson tunneling configuration employs a semiconductor barrier several hundred angstroms thick instead of the oxide barrier which is normally tens of angstroms thick. The Naval Research Laboratory has a joint effort with Sperry Research Center to fabricate all refractory superconducting tunnel devices with CVD deposited polysilicon barrier [23], [24]. The base NbN electrode is reactively sputter-deposited at a substrate temperature of about 500C which results in a T of about 15K. A several hundred angstrom thick silicon layer is grown byCcVD on the base electrode which is held at a temperature of 300-500C and a counter electrode of NbN is then reactively sputtered at (nominally) ambient temperatures. After fabrication, the base electrode has a Tc-13.5K while the Tc of the counter electrode is found to be about 12K. Josephson currents were observed at temperatures slightly below 10K. Unfortunately, the preliminary results were such that it could not be unequTvocally established whether the supercurrent was due to tunneling through the silicon barrier or along shorts across the barrier region. At the time of the preparation of this communication, Sperry had improved their techniques for depositing silicon layers on NbN and new all refractory, semiconductor barrier devices will be fabricated in the near future.

7. Weak Link/Constriction Devices

Another class of device configuration which can exhibit Josephson effects is a structure which contains a very severe reduction in width over a very short distance - the so-called "weak link". The theoretical con- straint for the observation of Josephson effect ' is that length (and width) of the constriction must be comparable to or smaller than three times the superconducting coherence length of the material [25]. In the case of the mechanically soft, low T materials, such as tin and aluminum, coherence lengths are of the order 0% thousands of angstroms and ideal weak links have been fabricated by conventional optical lithographic techniques or scribing with sharp diamond points, However, the coherence length of materialstends to decrease with increasing transition temperature and materials such as niobium and the high T materials have coherence lengths so small that conventional techniques gannot be used to fabricate constrictions that are small enough to satisfy the conditions for obtaining ideal Josephson behavior.

In the case of niobium, which has a coherence length of the order of 4008, recently developed "contamination resistn techniques have been used to fabr'cate constrictions with widths of the order of 5008 and lengths about lZOOi which have shown ideal Josephson behavior both in their response to incident microwave radiation [26] and when used to fabricate SQUID magnetometers[27]. The techniques for fabricating these so-called "nan~bridges~~ (SO named as they are smaller than lrmicrobridgesll) cannot be employed to fabricate Josephson weak levels in the refractroy alloy materials, especially those that have transition temperatures in excess of 10K. These materials have coherence lengths of the order of 502 or less (see Table I), a dimension not reachable with any current device fabrication technique.

There are a number of references where the authors have reported phenomena at temperatures approaching 17K which they have attributed to the Josephson effect. Golavashkin et al. [28], [29 ] , has reported on a Nb Sn device in which they mechanically scribed two constrictions 1-5 pm wide and 1-2 pm long to produce what is supposedly a DC SQUID. The device was sen- sitive to changes in the applied magnetic field in a manner resembling a DC SQUID although the response behavior was poor compared to that observed for DC SQUIDS fabricated from low Tc materials. The poor quality of the response might be due to operation by some type of synchronized "flux flow" which one would expect to observe for the case of apparent bridge dimensions at least 100 times larger than the coherence length of Nb Sn. Another possible explanation for this type of response (f. rom a device aith constrictions with these dimensions) is that there are flaws (such as cracks or grain boundaries) within the constricted region such that the dimensions at some point are very small compared to the apparent dimensions. Another possibility is that the region of the film where that constriction was produced was not uniform and that the operation of the device is attributable to granular or inhomogeneous superconductivity (see below). Until the true nature of the constriction can be established, reservations must be exer- cised before attributing the observed behavior to true Josephson effect.

The same reservations should be made in considering the recent report by Lee and Falco [30] on the observation at temperatures near 17K of RF induced microwave steps in the I-V characteristics of constrictions made in Nb Sn films. A constriction with dimensions of the order of micrometers was sckptured into the film and then the transition temperature of the constriction was adjusted downward by applying across the constriction voltage pulses with microsecond duration. By a mechanism as yet not understood, this technique permitted a systematic and controlled depression of the T of the constriction. Using a device in which the T had been depresse8 to about 17K from an original value of about 18K, t%ey observed RF induced steps in the device I-V curve at voltages corresponding to the integer har-

monics of the Josephson frequency: no subharmonic or non-integer harmonic steps were obtained. This type of response is expected for ideal Josephson behavior while synchronized flux flow, which would be expected for a constriction in Nb3Sn with these dimensions, would also have exhibited subharmonic or non-integer harmonic steps. The modification of the film that occurred during the 'Ipulse" treatment is not understood: small fractures may be created or, possibly, local heating may expedite the diffusion of oxygen along grain boundaries thus creating a region within the constriction which exhibits inhomogeneous superconductivity. (In the latter case, as will be discussed below, the coherence length does not impose a severe con- straint on the dimensions of an inhomogeneous weak link.) Notwithstanding the lack of knowledge about the details of the physical and electrical nature of the constriction after the pulse treatment, the preliminary results appear quite convincmg that true Josephson effects have been observed in refractory high Tc materials at temperatures as high as 17K. Furthermore, this fabrication technique should be suitable for producing Josephson devices in other refractory materials. This technique, which was first employed in producing superconducting weak link constrictions in niobium films by Duret et al. [31], might be considered to be the thin film equivalent of the bulk mechanical point contact configuration. In both cases, the device is adjusted while observing the superconducting or electrical properties, the thin film version by repeated pulsing while the bulk contact by very careful adjustment of the tension on the point. (Unfortu- nately the thin film version is - not reversible while the point contact is.) The pulsing technique may permit extensive investigations of Josephson effects in high Tc films, much the same way that the point contact was an extremely valuable tool in studying bulk materials. Although this technique may prove to be very useful in making single Josephson devices, it is doubt- ful that it can be employed to fabricate complex circuits and/or arrays of Josephson devices.

8. Granular Weak Links

As pointed out in the previous section, theory predicts that ideal Josephson behavior should be observed in weak link constrictions only if their dimensions are of the order of or less than three times the coherence length of the material [25]. For refractory alloy materials, coherence lengths are of the order of 100 A or less and thus the observation of Josephson behavior requires constrictions with dimensions of the order of 300 A or less, a value that cannot be achieved by conventional electron beam lithography and is at the limit of what may be achieved with I1contamination resistt1 lithography developed by IBM.

However, the group at NRL has shown that weak linkconstrictions in refractory alloy films can exhibit Josephson effects with micrometer sized dimensions if the weak link region is sufficiently thin to exhibit granular or inhomogeneous superconducting behavior [32]. When a film of, for example, niobium or niobium nitride of thickness of lOOOA is thinned down by anodization or ion milling, the transition temperature begins to decrease (see Fig. 3). The variation of the transition temperature with film thickness obeys the following relation

Fo both niobium and niobium nitride, the constant d was found to be about 1 The significance of this parameter is that Tc 'of the film appear to be reduced by the proximity effect with non-superconducting layers of 5g on each surface of the film.

As the film thickness is reduced, a critical value is reached, about 1008 for niobium nitride and about 508 for niobium films, below which there

is a noticeable decrease in the critical current density of severalorders of magnitude (see Fig. 3). The same reduced value of critical current density values is observed for thinner films. However, the Tc varies smoothly through this critical thickness value and obeys the above relation for all values of film thickness.

FILM THICKNESS IAngSroms)

Fig. 3. Variation cal current density

of the superconducting transition tempeature and criti- of niobium nitride films as a function of film thickness

The decrease in critical current density is believed to be due to a transition from bulk conduction for the thicker films to a granular or hopping conduction for films with thicknesses less than the critical value. As the film is deposited, it initially grows as isolated islands of materials which develop to sizes of the order of 100A. As the second layer of grains begin to grow, they touch and "bulk-likeff conduction occurs. Con- versely, when a thick film is thinned until the thickness of the region becomes comparable to the grain size, the conduction changes from bulk-like to tunneling between the isolated grains. The proposed nature of the conduction mechanism for thin films is supported by the observation that many of these films exhibit a negathe coefficient of resistance for temperatures above T which can be attributed to hopping conduction between isolated grains. 'similar behavior has been observed in granular aluminum and tin films which were deposited in an oxygen partial pressure which aided in the formation of isolated grains coated with oxide layers.

SQUIDs have been fabricated using granular regions with dimensions of the order of one micrometer in length which exhibited good Josephson behavior [33], [34]. (Granular regions with lengths greater than several micrometers exhibited rather poor or smeared out Josephson responses.) Depending on the degree of granularity of the weak link region, the width of the localized region can be varied from submicrometer for films which were marginally granular to widths greater than 40 micrometers for weak links which were clearly granular in nature. For niobium nitride SQUIDs with weak links that were clearly in the granular regime, Josephson behavior could be followed from T , which for some devices was as high as 10K, down to about 1.2K, the lowe& temperature that could be reached (see Fig. 4). Ideal Josephson behavior was obtained for critical currents greater than I$~/L (the condition for hysteretic behavior required by the Kurki jarv i theory) but less than about 6 I$o/L [34], 1351 .. The degradation of response for large values of critical current is not understood but is being investigated. Several of these devices have been examined at microwave frequencies (about 10 GHz) and exhibited Josephson behavior with extremely good signal-to-noise ratios, over a corresponding wide temperature range from just below Tc down to about 1.2K [34].

I,+ (ARBITRARY)

Fig. 4. I-V characteristics of NbN RF SQUID containing a granular weak link measured at 20 MHz as a function of temperature. The onset temperature for observing Josephson behavior was 10K while 1.27K was the lowest temperature that could be achieved in the particular apparatus (see Refs. 34 and 35)

If the weak link region has been thinned down so that it is just marginally granular in nature, Josephson behavior can be observed at much higher temperatures but over a restricted temperature region. For example, the data in Fig. 5 is the response of a NbN SQUID containing a marginally granular weak link which exhibits an onset of SQUID behavior at near 14K. The response can be followed down to about 13K before the critical current becomes too large. As described in the previous paragraph, ideal magnetic flux noise,characteristic of the temperature, is observed for critical currents greater than 1 I$ /L but less than about 6 $ /L. corresponding to a one degree temperature ranBe of operation. The qual?ty of the response and the temperature range over which it is observed is superior to the previously reported response for any RF-biased SQUID fabricated from refractory alloy materials.

T=13.0eK

T=13.06'K

-

C_ m

Q ... >-

T = 13.35'K

11 l,f (ARBITRARY)

Fig. 5. I-C characteristics for a NbN RF SQUID containing a weak link constriction that is marginally granular in nature (see Refs. 34 and 35)

The physics behind the observed operation of these granular weak link constriction devic s is not understood [32], 1351. The grain size in these films is about 10$ and thus in a l micrometer by 1 micrometer region. there are more than 10 grains. However, a granular region of this dimension built into an l$F biased SQUID behaves as if it were a single ideal Josephson device--the 10 grains are strongly coupled together and appear to be characterized by a single quantum mechanical phase difference common to the entire array of grains.

Thus far, granular behavior has been observed in two refractory thin film systems - niobium and niobium nitride. If similar type of behavior can be observed in higher Tc material systems, it may be feasible to fabricate weak link structures with micrometer dimensions which may show onsets of SQUID behavior at temperature near 1% and thus, could be operated near 10K (that is, at a reduced temperature of 0.6) where the temperature coefficients of the various superconducting parameters are weak. At these operating temperatures, the cryogenic environment could be provided by small, compact, energy-efficient cryogenic refrigeration systems.

9. Conclusions

Materials. There are a fairly large number of superconducting materials that are potentially suitable for the fabrication of high temperature devices and circuits. Many of these materials have been prepared in thick film form (thicknesses greater than 1 micrometer) with the purpose of exploring their use in high power superconducting applications. However, there has been only limited work on preparing thin refractory films (thicknesses of the order of or less than 1,000A) in a form that would be suitable for fabricating Josephson devices and superconducting electronic circuits. Additional effort will be required to develop procedures for preparing fine grain structures, that adhere well to substrate materials and that can be prepared with desired properties under fabrication conditions that are compatible with multi-layered device processing.

Device Processinq. In the fabrication of tunnel junctions using all refractory high temperature superconducting electrodes, the major problem areas are the barrier region, either the natural oxide or artificial barriers, suitable maskin-9, and the deposition of the refractory counter electrode without degrading the base electrode or the barrier region (see Table 3). The fabrication of weak link constrictions in refractory materials that can exhibi ideal Josephson behavior requires device geometry resolutions in the 300k regime, which is beyond the capabilities of presently available device processing technology. Granular weak links have been fabricated with micrometer dimensions with operating temperatures as high as 14K. Additional materials research would be required for the fabrication of granular systems which have transition temperatures at values in excess of 15K so that devices made from these materials can be operated at 10K at a reduced temperature of 0.6.

Table 3. Problem Areas for Fabricating High Temperature Josephson Devices (Top 2 10K)

Highest Reported Type of Structure Operation Problem Area

Tunnel Junction

Proximity Effect Device

Barrier layer growth Defining masks Counter electrode

Limited temperature range Exponential Ic dependence

Weak Link/Constriction ? Small dimensions (51002)

Granular/Inhomogeneous -13K Preparation of granular Weak Link specimens of high Tc

materials

206

Prospects for F_uture Progress. Despite the limited amount of research activity directed at t h e f a b r i c a t i o n of all refractory high operating temperature Josephson devices and circuits, the prospects are quite good that these materials will be used in the near future in Josephson technology. The primary obstacles to their use are materials problems such as producing controlled barrier layers and suitable counter electrodes in the case of tunnel junctions, and producing high T materials which exhibit granular behavior in the case of weak links. 9hese are quite serious obstacles, especially given the refractory nature of the materials and the affinity of these materials for such undesirable impurities as oxygen. However, the modest efforts to date at IBM, NRL and in Japan have yielded quite encourag- ing results.

All refractory devices would be extremely rugged and reliable and would withstand repeated thermal cyclings to cryogenic operating temperatures. These properties are not characteristic of the presently used lead alloy Josephson device technology but are essential if JosepHson technology is to become a viable one, either on a commercial basis or for military applications. In addition, these refractory materials often have superconducting transition temperatures in excess of 15K which makes possible their operation near 10K(at a reduced temperature of 0.6). This would permit the use of fairly compact, highly reliable closed cycle refrigeration systems for deployment with Josephson sensor systems and, possibly small Josephson circuits and systems. It would appear that large scale digital processing systems would continue to be operated at liquid helium temperatures as the liquid is probably essential for adequate cooling of the very complex packaging associated with such systems.

The ruggedness and reliability of Josephson devices and circuits fabricated from refractory materials provide the impetus that will motivate the developing of this technology. Current activities are limited but should expand in the near future so that one may confidently expect to see alL- refractory Josephson devices within the next five years and fairly complex integrated circuit structures fabricated with refractory metallization within the next ten years. The advent of such circuit technology is crucial to the development of a viable superconducting electronic technology and their development will be pursued despite the serious materials problems that must be overcome.

10. References

[l] Strobridge, T.R., llCryogenic Refrigerators - An Updated Survey," NBS Technical Note 655, June 1973 (U.S. Government Printing Office, Washington, D.C. 20402, catalogue number C13.46:655), unpublished.

[21 Daunt, J. G., and Goree, W. S., "Miniature Cryogenic Refrigerators," Stanford Research Institute Report (ONR Contract N00014-67-C-0193), unpublished.

[3] Beasley, M. R., "Improved Materials for Superconducting Electronics,lr Future Trends in Superconductive Electronics, ed. B.S. Deaver, Jr., C.M. Falco, J.H. Harris and S.A. Wolf (American Institute of Physics, New York, 1978), pp. 389,

[4] Beasley, M. R. , "Progress Report on High T Superconducting Devices, " Applications of Closed-Cycle ~r~ocoolerg to Small Superconducting Devices; NBS Special Publication 508 (U.S. Government Printing Office, Washington, D.C. 20434, Stock No. 003-003-0910-1, 1978), p. 167.

[5] Notarys, H. A. and Mercereau, J. E., "Proximity Effect Bridges and Superconducting Micro~ircuitry,~~ J. Appl. Phys. 44, 1021 (1973).

[6] Wu, C. T. and Falco, C. M., "High Temperature Nb3Sn Thin-film SQUID1sI1l Appl. Phys. Lett., 30, 609 (1977).

[7] Gu, J., Cha, W., Gamo, K. and Namba, S., "Properties of Niobium Super- conducting Bridges Prepared by Electron Beam Lithography and Ion Implantation," J. Appl. Phys., - 50, 6437 (1979).

[8] Kraichman, R. K., Hutchby, J. A., Burgess, J. W., McNamara, R. P. and Notarys, H. A., "Comparative Studies of Ion-Implant Josephson-Effect Structures," IEEE Trans. Magnetics, MAG-13, 731 (1977)-

[9] Palmer, D. W., Notarys, H. A. and Mercereau, J. E.,"Quantum Interference in High Transition Temperature Thin Film Materials," Appl. Phys. Lett. 25, 527 (1974).

[lo] Greiner, J. H.,ltJosephson Tunneling Barriers by R-F Sputter Etching in an Oxygen Plasma,If J. Appl. Phys. 42, 5151 (1971).

[lll Laibowitz, R. B. and Cuomo, J. J., I1Tunneling Sandwich Structures Using Single Crystal Niobium FilmsItt J. Appl. Phys. 5, 2748 (1970).

[12] Broom, R. F., Jaggi, R., Laibowitz, R. B., Mohr, Th. 0. and Walter, W., nThin Film Josephson Tunnel Junctions with Niobium ElectrodesIw Proceedings of 14th International Conference on Low Tempeature Physics; Otaniemi, Finland; August 1975 (North Holland Publishing Co., Amsterdam); Vol 4, p. 172.

1131 Broom, R. F., Laibowitz, R. B., Mohr, Th. 0. and Walter, W., "Fabrica- tion and Properties of Niobium Josephson Tunnel J~nction,'~ IBM J. Res. Develop. 24, 212 (1980).

[I41 Moore, D. F., Zubeck, R. B., Rowell, J. M. and Beasley, M. R., "Energy Gap of A-15 Superconductors Nb Sn, V3Si and Nb3Ge measured by Tunnel- ing, " Phys. Rev. =, 2721 71979).

[15] Laibowitz, R. B. and Mayadas, A. I?., I1Josephson Junctions with Nb/A1 Com- posite Electrodes," Appl. Phys. Lett. a, 254 (1972).

[16] Laibowitz, R. B., Tsuei, C. C., Chaudhari, P., Raider, S.I., Drake, R. and Viggiano, J. M., "Proximity Effect Tunnel Junctions with Bar- riers Formed from Amorphous Alloys, It Jour. de Phys. =, C6-1340 (1978 ) .

[17] Yeh, J. T. C. and Tsuei, C. C., "Tunneling Studies on Thin Film Nb-A1 Alloysl" IEEE Trans. on Magnetics, MAG-15, 591 (1979).

[181 Buitrago, R. H., Goldman, A. M., Toth, L. E. and Cantor, R. "Prepara- tion of Nb Ge Superconducting Tunneling Junctions," IEEE Trans. on Magnetics, h ~ - l 5 , 589 (1979).

[I91 Howard, R. E., Rudman, D. A. and Beasley, M. R., "Josephson Properties of Nb3Sn/Pb tunnel junction^,^^ Appl. Phys. Lett. 33, 671 (1978).

[20] Rudman, D. A. and Beasley, M. R., Ifoxidized Amorphous-Silicon Super- conducting Tunnel Junction Barriers," Appl. Phys. Lett. =, 1010 (1980).

[211 Tarutani, Y., Yamada, K. and Kawabe, U., "Superconducting Tunneling Junctions with V3Si-Si02-Mo3Re2,I1 Appl. Phys. Lett. 37, 239 (1980).

208

[221 Shinoki, F., Takada, S., Kosada, S. and Hayakawa, H., "Fabrication of High Quality NbN/Pb Josephson J~nctions,~~ Jap. Jour. Appl. Phys. 19, Suppl. 1, 591 (1980). -

[23] Kroger, H., Potter, C. N. and Jillie, D. W. llNiobium Josephson Junc- tions with Doped Amorphous Silicon Barriers," LEEE Trans. on Magnetics, MAG-15, 488 (1979).

[24] Kroger, H. "Josephson Devices Coupled by Semiconductor Links,I1 IEEE Trans. Electron Devices ED-27, 2016 (1980).

1251 See for example, Likharev, K. K., "Superconducting Weak Links, Rev. Mod. Phys. 51, 101 (1979).

[261 Laibowitz, R. B., Broers, A. N., Yeh, J. T. C. and Viggiano, J. M., "Josephson Effect in Nb Nanobridges," Appl. Phys. Lett. 35, 891 (1979).

1271 Voss, R. F., Laibowitz, R.B. and Broers, A.N., 'INiobium Nanobridge dc SQUID," Appl. Phys. Lett. 37, 656 (1980).

1281 Golovashkin, A.I., Levchenko, I.S., Lykov, A.N. and Makhashvili, L.I., "Quantum Interference on Nb Sn Thin Film Contacts at Hydrogen Tem- peratures," JETP Lett. 24, 531 (1976).

[29] Golovashkin, A. I., Levchenko, I. S. and Lykov, A. N., "Quantum Inter- ference at Nb and Nb Sn Superconducting Microbridges," Soviet Physics Solid State 8, 2121 (1977).

[30] Lee, T. W. and Falco, C. M., "Josephson Effects in Nb Sn Microbridges," Applied Superconductivity Conference, Santa Fe , N.. , October 1980; IEEE Trans. on Magnetics, Spring 1981, to be published.

C311 Duret, D., Bernard, P. and Zenatti, D., "A UHF Superconducting Magnetom- eter Utilizing a New Thin Film Sensor," Rev. Sck. Instr. El 475 (1975).

[32] Claassen, J. H., Cukauskas, E. J. and Nisenoff, M., llGranular Weak Link Josephson DevicesIn Inhomogeneous Superconductors-1979, ed. Gubser, D. U., Francavilla, T. L., Leibowitz, J. R. and Wolf, S. A. (American Institute of Physics, New York, 1980), p.. 169.

[33] Rachford, F. J. and Cukauskas, E. J., "High T Refractory Thin-Film Microwave SQUIDS, Appl. Phys. Lett. 35, 891 b979).

[34] Cukauskas, E. J. and Nisenoff, M., "Intrinsic Noise Characteristics of NbN SQUIDS," J. Appl. Phys. 52 (Feb 1981), accepted for publication.

[35] Cukauskas, E. J. and Nisenoff, M., I1Noise Properties of Thin Film Granular Weak Link SQUIDsI1' Applied Superconductivity Conference, Santa Fe, N.M., October 1980; IEEE Trans. on Magnetics MAG-17, Spring 1981, to be published.

[361 Claassen, J. H., "Josephson Behavior in Granular NbN Weak Links,I1 Appl. Phys. Lett. 36, 771 (1980).

ATTENDANCE LIST Conference on Refrigeration for Cryogenic Sensors and Electronic Systems

October 6 & 7, 1980

Robert A. Ackermann Fernand 0. Bedard Project Manager, S t i r l i n g Engine Sys. Div. Lab f o r Physical Sciences Mechanical Technology Incorporated 4928 College Avenue 968 Albany-Shaker Road Col lege Park, MD 20740 Latham, NY 02110

W. G. Baechler Leybol d-Heraeus Bonner S t r . 504 5600 Cologne West Germany

Abe Bakst 27764 Edgeton Road Los Altos H i l l s , CA 94022

M i ke Bal i ster National Radio Astronomy Observatory P. 0. Box 2 Green Bank, WV 24944

C. Barbe L' A i r Liquide Centre de Recherche C1 aude-Del orme 75321 Paris Cedex 07 France

John Barclay Group P-10 M. S. 764 Los Alamos Sc ien t i f i c Lab. Los Alamos, NM 87545

Carl i s l e Barnes Lockheed Aerospace P. 0. Box 4346 Foster City, CA 94404

Me1 Bel lo The Aerospace Corporation Bldg. 05 - P. 0. Box 92957 Los Angeles, CA 90009

G. M. Benson ERG, Inc. Lowel 1 and 57th Street Oakland, CA 94608

B. W. Birmingham National Bureau o f Standards 325 S. Broadway Boulder, CO 80303

R. D. Blaugher Manager, Cryogenic Technology Westinghouse E lec t r i c Corp. 1310 Beulah Road Pittsburgh, PA 15235

Richard G. Brandt Of f ice o f Naval Research 1030 E. Green Street Pasadena, CA 91106

R. W. Breckenridge A. 0. L i t t l e , Inc. Acorn Park Cambridge, MA 02140

ATTENDANCE LIST, continued

Howard Brown P. 0. Box 2 National Radio Astronomy Observatory Green Bank, WV 24944

Eldon Byrd Naval Surface Weapons Center Whlte Oak Laboratory Mail Code R-34 S i lver Spring, MD 20910

Pasquale Care l l i Consiglio Nazionale Del le Ricerche Laboratorio D7 Elet t ronics Del lo Stato Solido Romano 42 Rome I t a l y

Charles Class Mart in Marietta Aerospace Denver Div is ion P. 0. Box 179 M. S. 30484 Denver, CO 80201

Will iam Colyer Energy Systems Corp. 169 Los Cordovak R t . Taos, NM 87571

Larry Daddari o P. 0. Box 2 National Radio Astronomy Observatory Green Bank, WV 24944

David E. Daney NBS - Thermophysical

Properties Div is ion Boulder, CO 80303

A. Daniels 345 Scarborough Rd. Phi?ips Lab. B r l arc1 i f f Manor, NY 1051 0

V. J. deWaal Laboratori um voor Technische Naturhunde University o f D e l f t D e l f t The Netherlands

Denis Duret t. E. T. I. Commissariat a 1'Energie Atomique 85X 38041 Grenoble, Cedex France

Nlchol as Eber Dept 6/65 Sul zer Brothers Limited CH-8401 Winterhur Swi t z e r l and

Edgar A. Edelsack Dept. o f the Navy Code 427 Off ice o f Naval Research 800 N. Qunicy St. Room 332 A r l Sngton, VA 22217

Alain Faure LIAir Liqufde-Etablissement de Sassenage Dept. de Construction e t de Vente Mecanique B. P. 15 38360 Sassenage France

A. A. F i f e CTF Systems #?5 1750 McLean Ave. Port Coquitlam B. C. Canada V3ClMG


Ephraim 8. F l i n t IBM - Thomas J. Watson Research Center P. 0. Box 218 Yorktown Heights, NY 10598

M. X. Francois Laboratoire de Thermodynamique des Fluides Campus Universitaire Bat 502 Ter 91405 Orsay Cedex France

Max G. Gasser NASA Goddard SFC Code 713 Greenbelt, MD 20771

Peter E. G i f fo rd President, Cryomectt, Inc. 314 Alnsley Drive Syracuse, NY 13210

Dave Giguere Hughes A i r c ra f t Company EED Bldg. 240 3100 Lamitom Blvd. Torrance, CA 90509

M i ke Go1 dows ky Phi l ips Lab. 345 Scarborough Road Br ia rc l i f f Manor, NY 10510

Geoffrey Green OTNSROC (Annapol i s Lab. ) Code 2712 Annapolis, MD 21402

Robert W. Guernsey, Jr. IBM - Watson Res. Center P. 0. Box 218 Yorktown Heights, NY 10598

J. Bindslev Hansen Physics Lab. - Univ. o f Copenhagen Universitet3parken 5 DK-2100 Copenhagen Denmark

Robert G. Hansen R. G. Hansen 81 Associates 1324 State Street Santa Barbara, CA 93101

W. H. Hartwig Dept. o f E lec t r i ca l Engineering P. 0. Box 7728 University o f Texas a t Austin Austin, TX 78712

W i l l l a m Haskin Adv. Cryogenic Systems Group Environmental Control Branch Dept. o f the A i r Force Wright-Patterson A i r Force Base, OH 45433

C. Heiden I n s t i t u t f u r Angewandte University o f Giessen Heinrich-Buff-Ring 16 0-6300 Gfe5en W. Germany

Thomas E. Hoffman Arthur D. L i t t l e , Inc. Acorn Park Cambridge, MA 02140

ATTENDANCE LIST, cont i nued

R. Hollman Physics Department Stanford Un ive rs i t y Stanford, CA 94305

S tuar t Horn U. S. Army Night V is ion Lab. NV & E.O. Lab. Ft . Belvo i r , VA 22060

Yoshf h i r o I sh i zak i Un ive rs i t y o f Tokyo Tokyo, Japan

Ly le M. I sho l Code 792 Naval Coastal Systems Center Panama C i ty , FL 32407

A. L. Johnson Aerospace Corp. D-5 P. 0. Box 92957 Los Angeles, CA 90009

Dean L. Johnson J e t Propul s ion Laboratory M. S. 238-737 4800 Oak Grove Dr ive Pasadena, CA 91106

Yasuharu Kami aka Un ivers i t y o f Ca l i f o rn i a 7671 Boel ter Ha l l Los Angeles, CA 90024

Ken j i r o Kasai Kasado Research Dept. H i tach i Ltd. 794 Kudamatsu-Ci t y Yamaguchi-Ken 744 Japan

Dale Kennedy AFWAL/FIEE W r i ght-Patterson AFB Ohio 45433

Peter J. Kerney CTI Cryogenics Kelv in Park Waltham, MA 02254

T. Kondo Engineering Research I n s t i t u t e Faculty o f Engineering Un ive rs i t y o f Tokyo Bunkyo-Ku , Tokyo Japan

Tadashi Kondoh Manager, 2nd Technical Lab. A i s i n Sei k i Co., Ltd. 1-Asahi-Machi, 2-Chome Kariya City, A ich i Pref. 448 Japan

Herbert Korf AGE-Tel e f unken Theresienstr. 2 D-7100 Hellbronn W. Germany

Tomoo Kurumada Osaka Oxygen Indust r ies , Ltd. Utajima, Nishiyodogawa-Ku Osaka Japan

ATTENDANCE L I S T , continued

Kurt Kwasni tza ETH - Zurich I n s t i t u t e o f Sol id State Physics Zurich Switzerland

Rudy Latasa National Radio Astronomy Observatory P. 0. Box 2 Green Bank, WV 24944

Will iam Lawless Lake Shore Crybtronics 64 E. Walnut Street Westervil le, OH 43081

Otto Ledford Aerospace Corp. Box 92957 Los Angeles, CA 90009

Daniel Lehrfeld Magnavox 34 Greenwood Drive New City, NY 10956

Robert Lepobs ky CT I, Cryogenics 266 Second Ave. Waltham, MA 02154

3ames Lester Ba l l Aerospace Box 1062 Boulder, CO 80306

Del Linenberger NBS-Thennophysical Properties Div is ion Boulder, CO 80303

Will iam A. L i t t l e Physics Department Stanford University Stanford, CA 94305

Ralph C. Longsworth A i r Products & Chemicals, Inc. Advanced Products Department Box 2802 Allentown, PA 18105

J. C. Lo t t i n Commissariat Al'Energie Atomique C. E. N. Saclay-Orme des Meri siers, Stipe Postale No. 2, 91-Gif-Sur-Yvette France

K. Lueders Frele Univ. Be r l i n Fach Physlk (FB 20) I n s t i t u t f u r Atom & Feskorperphysik Koenigi n- Lui se-Str. 28/30 1000 Ber l i n 33 W. Germany

John T. Lynch MIT - Lincoln Labs. Room 329-Box 73 Lexington, MA 02173

Thomas J. Marusak Mechanical Technology Incorporated 968 Albany-Shaker Road Latham, NY 02110


Yoichi Matsubara Atomic Energy Research I n s t i t u t e Nihon Univers i ty Kanda-Surugadal Chiyoda Ku Tokyo Japan 101

Ken Myr t le Dept. o f Physics Simon Fraser Univers i ty Burnaby El. C. V5A156 Canada

Lawrence G. Nelsen Rockwell In te rna t iona l 10202 Fl intndge V i l l a Park, CA 92667

M. Nisenof? Naval Research Laboratory Code 6854 Washington, D. C. 20375

3. J. OrPh Hughes A i r c r a f t Company Culver C i t y , CA 90230

Isao Oshima Osaka Oxygen Indust r ies , Ltd. Utajima, Nishiyodogawa-Ku Osaka Japan

J.*A. Pals Phi 1 i p s Research Lab. E i ndhoven Neder 1 and

George Patton DTNSRDC (Annapol i s Lab. ) Code 2712 Annapolis, MD 21402

Dusan Petrac J e t Propulsion Lab, CIT 4800 Oak Grove Drive, 183/401 Pasadena, CA 91103

Warren Pierce Lake Shore Cryotronics 64 E. Walnut St reet Westervi l le, OH 43081

H. Quack Sulzer Brothers L imi ted CH-8401 Winterthur Switzerland

Ray Radebaugh NBS - Thermophysical

Propert ies D iv i s ion Boulder, CO 80303

Richard M. Ra l l Hughes A i r c r a f t Company EED-Bldg. 24O/ll3 3100 W. Lomita Blvd. Torrance, CA 90509

P i t e r Roos Magnavox 46 Industr ia7 Ave. Mahwah, NJ 07430


P ie r re M. Roubeau A1 1 an Sherman Commissariat A 1'Energie Atomique NASA Goddard Space F l i g h t Center C. E. N. Saclay -Orme des Meris iers, Bo i te Code 713 Postale No. 2, 91-Gif-Sur-Yvette Greenbelt, MD 20771 France

S.C. Russo B l dg. l2/Vl37 Hughes A i r c r a f t Company Culver C i t y , CA 90230

Ronald E. Sager 5. H.E. Corporation 4174 Sorrento Val ley Blvd. San Diego, CA 92121

Raymond Sarwi ns k i S. H. E. Cosporation 41 74 Sorrento V A l 1 ey 61 vd. San Diego, CA 92121

E. Saur I n s t i t u t f u r Angewandte Physik der Un ive rs i ta t Giessen Heinr ich Buff-Ring 16 6300 GieBen West Germany

Aron Sereny Ph i l i p s Lab. 345 Scarborough Rd. B r i a r c l i f f Manor, NY 10510

G, W. Shepherd Magnavox 46 I ndus t r i a l Ave. Mahwah, NJ 07430

Arnold H. S i l v e r Aerospace Corp. Box 92957 Los Angeles, CA 90009

J. F. Skinner Hughes A i r c r a f t Company Bldg. 5 M.S. B lS l Culver Cf ty , CA 90230

Hans A. Steinherz CTI Cryogenics Ke lv in Park Waltham, MA 02254

Wi l l iam Steyer t Los Alamos S c i e n t i f i c Lab. Mai l Stop 764 - P. 0. Box 1663 Los Alamos, NM 87545

Phi 1 I p Studer NASA Goddard SFC Code 716 Greenbelt, MD 20771

Donald 6. Sul l i van D i v i s i on 724.03 - Room 2137 Nat ional Bureau o f Standards 325 Broadway Boulder, CO 80303


Jean-Jacques Thibault Lawrence A. Wade L' A i r Liqui de , Centre d' Etudes Cryogeni ques Lockheed Research Lab. Claude-Delorme, Bol te Postale No. 15 3251 Hanover Street 38360 Sassenge Bldg. 205 - Dept. 52-32 FRANCE Palo A1 to, CA 94304

K. D. Timmerhaus Engineering Center AD1 -25 Univers i ty o f Colorado M. S. 423 Boulder, CO 80309

Bryan C. Troutman IBM 2L5O 18100 Frederick Pike Gai thersburg, MD 20760

E. Tward Je t Propulsfan Laboratory M. S. 183-901 4800 Oak Grove Drive Pasadena, CA 91 103

A. R. Urbach Ba l l Aerospace Corp. Box 1062 Boulder, CO 80306

Eugene Urban Es. 63, NASA Marshall Space F l i g h t Center Huntsvf 1 le, AL 35812

R. J. Vincent ERG, Inc. Lowell and 57th Street Oakland, CA 94608

G. Walker Dept. o f Mechanical Engineering 2920 24 Ave. N.W. Univers i ty o f Calgary Calgary, Alberta CANADA

Sander Wei nreb National Radio Astronomy Observatory P. 0. Box 2 Green Bank, WV 24944

Nancy K. We1 ker Lab f o r Physical Sciences 4928 College Avenue College Park, MD 20740

R. White Adv. Cryogenic Systems Group Environmental Control Branch Dept. o f the A i r Force Wright-Patterson AFB Ohio 45433

Calvin Winter Simon Fraser Univ. Dept. o f Physics Burnaby B. C. V5A156 Canada

James E. Zimmerman D iv is ion 724.03 - Room 2137 National Bureau o f Stds. 325 Broadway Boulder, CO 80303

* U.S. GOVERNMENT PRINTING OFFICE : 1981 0 - 344-816 : QL 3

4. T I T L E AND SUBTITLE

NBS-114A (REV. 2 -88 ] -

Refr igerat ion fo r Cryogenic Sensors and Electronic Systems Proceedings of a Conference held a t the National Bureau of Standards, Boulder, C O Y October 6-7, 1980

5. AUTHOR(S)

J . E. Zimmerman, D. B . Su l l ivan , and S. E . McCarthy, Edi tors 6. PERFORMING ORGANIZATION ( I f jo int or other than N B S , see instructions) 1 7. C o n t r a d G r a n t No.

u.s. DEPT. O F COMM.

BIBLIOGRAPHIC DATA SHEET (see instructions)

N A T I O N A L BUREAU O F STANDARDS D E P A R T M E N T O F C O M M E R C E WASHINGTON, D.C. 20234

'1. PUBLICATION OR REPORT NO.

NBS SP 607

2. Performing Organ. Report No.

Final

3. Publ icat ion Date

May 1981

- I 9. SPONSORING ORGANIZATION NAME AND COMPLETE ADDRESS (Street. City, State, Z IP)

Internat ional I n s t i t u t e of Refrigeration - Commission A 112 Office of Naval Research - Naval Research Laboratory Cryogenic Engineering Conference National Bureau of Standards

LO. SUPPLEMENTARY NOTES

Library of Congress Catalog Card Number: 81 -600038

CJ Document describes a computer program; SF-185, F lPS Software Summary, i s attached.

11. ABSTRACT ( A 200-word or less factual summary o f most s igni f icant information. If document includes a signif icant bibliography or l i terature survey. mention i t here)

This document contains the proceedings of a meeting of r e f r i ge r a t i on s p e c i a l i s t s held a t the National Bureau of Standards, Boulder, C O Y on October 6 and 7 , 1980. Par t i c ipa t ion included represen ta t ives of industry , government, and academia. The purpose of the meeting was t o discuss progress in the development of r e f r i ge r a t i on systems which have been speci a1 ized f o r use w i t h cryogenic sensors and e l ec t ron i c systems. The meeting focused primarily on the temperature range below 20 K and cooling capaci ty below 10 W The meeting was j o i n t l y sponsored by the Internat ional I n s t i t u t e of Refrigeration-Commission A 112, the Off ice of Naval Research, the Naval Research Laboratory, the Cryogenic Engineering Conference, and the National Bureau of Standards.

12. KEY WORDS (Six to twelve entries; alphabetical order; capi ta l ize only proper names; and separate key words by semicolons)

Cryocoolers ; cryogenic sensors; he1 i u m ; r e f r i ge r a t i on ; superconducting devices.

L3. AVA ILAB IL ITY 114. NO. O F

Unl imi ted

n For Of f i c ia l Distr ibution. Do Not Release to NTIS

P R~NTED PAGES I 223

Order From Superintendent o f Documents. U.S. Government Pr int ing Office, Washington. D.C. 20402.

Order From National Technical Information Service (NTIS), Springfield, VA. 22161 1 $5.50

USCOMM-DC 6043-P80

U S . DEPARTMENT OF COMMERCE National Bureau of Standards Washmgton. D.C. 2 0 2 3 4 - OFFICIAL BUSINESS

Penalty for Private Use. $ 3 0 0

POSTAGE AND FEES PAID U S . DEPARTMENT OF COMMERCE

COM-2 15

SPECIAL FOURTH-CLASS RATE BOOK