SERI/TR-332-416

VOLUME 2 OF TWO VOLUMES UC CATEGORY: UC-61

REVIEW OF THERMALLY REGENERATIVE ELECTROCHEMICAL SYSTEMS

VOLUME 2

HELENA L. CHUM SOLAR ENERGY RESEARCH INSTITUTE

ROBERT A. OSTERYOUNG STATE UNIVERSITY OF NEW YORK AT BUFFALO BUFFA�O, NEW YORK

APRIL 1981

PREPARED UNDER TASK No. 3356.50

Solar Energy Research Institute A Division of Midwest Research Institute 1617 Cole Boulevard Golden, Colorado 80401

Prepared for the U.S. Department of Energy Contract No. EG-77-C-01-4042

I�I---------------------------------------------------------------------

Synopsis Summary,

��������--����--��----Clayton Smith, Manage t

!;::�I TR-416

PREFACE

This review was prepared by R. A. Osteryoung, formerly with Colorado State University and now with the State University of New York at Buffalo, and by H. L. Chum, formerly with Colorado State University and presently on the SERI staff. Work was performed largely under Contract No. AM-9-S07S-1 and SERI Task 3356. 10. Compiled in two volumes, the review covers the technical back­ground of thermally regenerative electrochemical systems and presents recom­mendations for further work. For <the reader interested in a general overview, V olume 1, and Executive is a condensed version ·of V olume 2. Volume 2, which discusses the thermally regenerative electrochem­ical systems in more detail, is intended for researchers in chemical and elec­trochemical areas and for engineers (although detailed coverage of the fields of engineering, corrosion, and materials problems is outside the scope of this report).

The authors wish to acknowledge T. A. Milne for the suggestion of the subject of this rep0rt and for helpful discussions. During the course of this review, discussion took place with a number of people involved in research and devel­opment of fuel cells and/or regenerative electrochemical systems. Among these were J. Appleby, B. Baker, M. Breiter, E. Cairns, T. Cole, G. Elliot, E. Findl, A. Fischer, F. Gibbard, L. Heredy, Huff, T. Hunt, C. Johnson,J .• R. Kerr, M. Klein, K. Kordesch, F. Ludwig, J. McBreen, L. Nanis, W. O'Grady, J. Plambeck, H. Shimotake, H. Silverman, R. Snow, S. Srinivasan, C. Tobias, R. Weaver, N. Weber, and E. Yeager. These discussions were very helpful. Special thanks are due to J. H. Christie for profitable discussions and care­ful editing of the manuscript. The technical support of the Colorado State University Library and the Solar Energy Information Center is gratefully appreciated.

Approved for SOLAR ENERGY RESEARCH INSTITUTE

Thomas A. Milne, Chief Biomass Thermal Conversion & Exploratory

Res earch Branch

Helena L. Chum Senior Electrochemist

(�S:� Chemical and Biological

iii

------------------------------------------------------------------------

TR-4 16 5i::�1 I�I

SYNOPSIS

OBJECTIVE

Thermally regenerative electrochemical systems (TRES) are closed systems that convert heat into electricity in an electrochemical heat engine that is Carnot cycle limited in efficiency.

In this report, past,and present work on TRES is reviewed and classified. Two broad classes of TRES can be identified according to the type of energy input required to regenerate the electrochemical cell reactants: thermal input alone (Section I and II) or the coupling of thermal and electrolytic energy inputs (Sections III-V). To facilitate the discussion, these two broad cate­gories are further divided into seven types of TRES (Types 1-3 for thermal re­generation; Types 4-7 for coupled thermal and electrolytic regeneration). The subdivision was made according to significant differences in either the elec­trochemical cells or in the regenerators.

D+SCUSSION

is formed from C and A in an electrochemical cellIn Type 1 TRES, compound CA at temperature T1 with concomitant production of electrical work in the exter­nal load. Compound CA is sent to a regenerator unit at. T2 through a heat ex­changer. In the regenerator, compound CA is decomposed into C and A, which

erated in one step has long been known. It involves the oxidation of tin and the reduction of chromitnn (III) species at a graphite electrode in the elec­trochemical cell generating power. The regeneration is accomplished by lower­ing the temperature when the spontaneous reverse reaction takes place. This is a periodic power source.

v

are separated and redirected to the electrochemical cell via a heat exchanger, thus closing the cycle. The thermodynamic requirements for the electrochemi­cal reaction are �G < 0, �S < 0, and �Cp as close to. zero as possible. This type of TRES is equivalent to a primary battery, in which electrodes A and C are consumed. By coupling the battery with a regenerator unit, the electrode materials are regenerated. The classes of compounds that were investigated or proposed for this type are metal hydrides, halides, oxides, ' and chalcoge­nides. One of the most thoroughly investigated systems is the lithitnn hydride system (T1�500°C, T2�900°C). The advantage of this system is that lithium hy­dride decomposes into liquid lithium and gaseous hydrogen, enabling relatively simple separation. The power delivered by this system was low. Problems were encountered in the gas electrode, in the rate of decomposition of lithium hy­dride, and in materials. The best conditions f or the electrochemical cell and for the regenerator were never realized in a practical system. A considerable fraction of the compounds investigated or proposed had T2�800o-900°C. Nuclear reactors were the heat source envisioned for that temperature range. Some materials decomposing at lower temperatures were tried. An example is antimo­ny pentachloride, which decomposes at �300°C into liquid antimony trichloride and gaseous chlorine; however, the performance of the electrochemical cell was very poor. All of the above systems had one electrochemical reaction product and the regeneration was accomplished in a one-step process. An interes ting example of a system in which two electrochemical reaction products are regen­

S=�II_I ____ --'�

_____________________

T_R_-4_1_

Type 2 is similar to Type 1, but the products C and E of the electrochemical cell reaction A + D + C + E are regenerated in a two-step process (C + A + B; E + B + D). It involves more complex plumbing and two regenerator units. The systems attempted include metal halides or oxides; for example, A tin (II) =

or tellurium (II) chloride and D antimony (V) or copper (II) chloride with=

B gaseous chlorine. If at the regeneration temperatures A is also in the=

gaseous state, the ,separation of A and B is the major obstacle to successful operation of this type of system. To date, no complete eOlectrochemical cell coupled with the regenerator has been demonstrated to 'be feasible.

Type 3 is also similar to Type 1 and involves a one-step regeneration. Liquid metal electrodes are composed of one electroactive metal C and one electroin­active metal B. C and B form alloys C(B) or bimetallic compounds C B • Thex yanode and the cathode have, respectively, high and low concentrations of the electroactive metal in the liquid electrode. The cells are concentration cells. The regeneration is accomplished by sending the electrode materials (combined or individually) to a distillation unit where the C+B mixture is separated into C-rich and C-poor components, which are returned to the anode and cathode compartments, respectively. Examples include C sodium or potas­=

sium and B mercury or lead. These are the systems for which the feasibility =

of the thermal regeneration coupled to the battery was demonstrated. The per­formance of the demonstrated systems was poor, due in part to constraints im­posed by the space applications envisioned. The . performance of this type of system can be improved.

In Type 4 systems, compound CA, formed in the electrochemical cell at tempera­ture T1, is sent to a regenerator, which is an electrolysis cell at tempera­ture T2• In the regenerator, reactions opposite to those occurring at Tl re­generate C and A by using two energy inputs--electric and thermal. The elec­trolysis cell uses a fraction of the voltage produced by the battery at Tl (the additional energy is supplied as heat) and the remaining v.oltage is used to perform work in the external load. These systems are analogous to, and have the same requirements as, secondary (rechargeable) ,batteries . A few systems have been investigated; for example, CA sodium chloride, lead io­=

dide, cadmium iodide, lithium iodide. If C and A are in the gaseous state (for instance, hydrogen and oxygen), the electrochemical cell is a fuel cell; the regeneration is performed by water electrolysis at high temperature. To date, no complete demonstration of the feasibility of these systems has been performed. In addition to the above-mentioned examples of high temperature electrolysis, very few systems operating at lower temperature were explored in this mode of regeneration.

Type 5 systems are 'a particular case of Type 4, in which the electrolysis is performed at low pressure. They include systems in which one of the electro­active species is in the gaseous stat.e. The battery and the electrolysis cell operate at the same temperature, and the pressure of the electroactive species is varied in these two cells by physical means, e. g. , by the coupling of the cells with cold fingers. The operation is periodic. Examples include gase.ous iodine as the working electroactive fluid in lithium l molten iodideliodine cells. Low voltages are expected from these devices as well as mass transfer problems. However, these systems have energy storage capability.

vi

S=�I I.I ________________________ T_R_-_4_16 -� �

syste-", ctr1: ferent

: -,

Type

means of a not

neces­lead

elec-· of the'

In avoided by

In Type 6 or thermogalvanic or nonisothermal cells, the two elec­trodes are at temperatures and the cell temperature is not uni­form. The electrodes can be metallic, liquid, or gaseous (with inert elec­trodes). The electrolyte can pe solid or liquid, homogeneous or heteroge­neous. During the passage of current through the thermogalvanic cell, matter is transferred from one electrode to the other as a result of the electro�hem­ical reactions at the electrode/electrolyte interface and ionic transport in the electrolyte. In some types of cells the transfer of matter is permanent, and therefore a mechanical means to reverse the temperature of the electrodes must be provided for continuous operation of the engine as a power source. In these cells the thermal and electrolytic paths are not separated. Examples include copper electrodes immersed in a variety of copper salt solutions and gaseous chlorine in solid electrolyte or molten salt media. Most data for these cells refer to scientific information (e. g. , irreversible thermodynam­ics) and not to power generation. The systems generate low power outputs but can be made much more cheaply than their solid-state analogs.

are based on pressure differences of the working electroactive an isothermal electrolyte (solid or liquid). The pressure dif­

7 engines fluid across

to the expansion of the working zone at through theT2

the working fluid is condensed in recyc.led to the high temperature, high pressure

concentration cells. changes, no regeneration

iodine vapor expanded through vapor expanded through

example, the major difficulty. is liquid electrolyte integrity when it is subjected

example (T1�200o-300°C, T2�800o-900°C),

ference is maintained by using the changes in the vapor pressure with the tem­perature of the working electroactive fluid. The work performed is equivalent

isothermal electroactive fluid from the high to the low pressure electrolyte and its interfaces. After expansion, a cold reservoir and can be

zone of the cell by pump. The cells are Because the working fluid does undergo chemical and separation steps are sary. Examples include isothermal liquid iodide and sodium isothermal solid beta-alumina trolyte. In the first maintainance

to a pressure gradient. the second this problem is using a solid superionic conductor electrolyte. The highest power outputs in TRES to date have been obtained with this type of engine. The present non­availability of other superionic conductors limits the extension of this con­cept to other practical energy converters.

CONCLUSIONS AND RECOMMENDATIONS

TRES cover temperature ranges from near room temperature to about 1200°C. To date, power outputs of 0. 1 mW/cm2 to about 1 W/cm2 have been achieved. The majority of the systems reported utilized molten salt electrolytes and high regeneration temperatures. In addition, several promising energy converters employed solid electrolytes, which are superionic conductors. Much less ex­plored are lower

expended on General

-temperature media--aqueous, nonaqueous, or molten salt. Lit­tle effort was the use of catalysts to improve the rates of ther­mal decomposition. problems included engineering and materials pro-b­lems. A considerable fraction of the research and development of these en­gines was performed around 15 to 20 years ago in connection with the produc­tion of secondary space power sources to utilize heat from nuclear reactors.

vii

--

----------------------------------------------------------------------TR-4 16 !; :: �I I�I -

In this report we recommend areas of research in either science or engineering that would have long-range benefit to a TRES program. These areas include molten salt chemistry and electrochemistry, solid-state chemistry, materials sciences, aqueous systems and electrochemistry under extreme conditions, elec­trochemical engineering, and systems analysis. It should be pointed out that because solar-derived heat covers a very wide range of temperatures (�800­lOOO°C), more TRES can be brought into consideration.

viii

TR-4 16 S-�I �.=� =� I��I--------------------------------------------------------

TABLE OF CONTENTS

Page

I ntroduct ion . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Obje'ct i ve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 Types of Thermal ly Regenerat i ve El ectrochemi cal Systems . . . . . .

2

I Therma l Regenerat ion : Metal Hydri des , Hal i des , Oxi des , and Cha1 cogen i des . . . . . . . . . . . � . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

1 . 1 Si ngl e or Mul t i p l e El ectrochemi cal React ion Products and S i ngl e-Step Regenerat ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

I . 1 . 1 r�eta 1 Hydri de Systems : L ith i urn Hydri de . . . . . . . 26 1 . 1 . 2 Hal i de-Conta i n i ng Systems . . . . . . . . . . . . . . . . . . . . . 47 1 . 1 . 3 Oxi de-Conta i n i ng Systems and Other Systems . . . . 53 1 . 1 . 4 Summary and Di scus s i on of TRES Type 1 . . . . . . . . . 56

1 . 2 Mul t i pl e El ectrochemi cal React ion Products and Mu1 t i p 1 e-Steep Regen rati on . . . . . . " . . . . . . . . . . . . . . . . . . . . . ' . . . . . ' . . . 57

1 . 2 . 1 Meta.1 Hal i des . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 1 . 2 .2. Metal Oxi des . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 1 . 2 . 3 D i scus s i on of TRES Type 2 . . . . . . . . . . . . . . . . . . . . . 66

I I Therma l Regenerat ion : A l l oys or B i metal l i c Systems . . . . . . . . . . 69

1 1 . 1 Amal gam and Thal l i um Cel l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 1 1 . 1 . 1 The Potass i urn-Mercury System. . . . . . . . . . . . . . . . . . 76 I 1 . 1 . 2 The Sod i um- �1e, rcu ry Sys tern . . . . . . . . . . . . . . . . . . . . . 85 1 1 . 1 . 3 The Potass i um-Thal l i um and Ana l ogous Systems . . 97

I I . 2 B i metal l i c Cel l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 1 1. 2. 1 Sodi um-Conta i n i ng Systems . . . . . . . . . . . . . . . . . . . . . 100 I I . 2 . 2 Lith i um-Contai n i ng Systems . . . . . . . . . . . . . . . . . . . . 1 09

1 1 . 3 Summary of the Performance and Di scus n of Thermal ly Regenerat i ve Al l oys or B imetal l i c Systems . . . . . . . . . . . . . i12

I I I Thermoga 1 van i c o r Non i sotherma1 Ce1 1 s . . . . � . . . . . . . . . . . . . . . . . . . 1 15

II!. 1 �·1 o l ten Sa lt Thermoga1 vani c Cel l s . . . . . . . . . . . . . . . . . . . . . . 12 1 1 1 1 . 1 . 1 Sol i d or L i q u i d El ectrodes . . . . . . . . . . . . . . . . . . . . 121 I I I . 1 . 2 Gaseous E1 ectrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 I I I. 1 . 3 The B i smuth-Bi smuth Iodi de System . . . . . . . . . . . . . 130

1 1 1 . 2 Thermoga1 van i c Cel l s with Sol i d El ectrolytes . . . . . . . . . . 132 I II . 3 The- rmoga1 vani c Cel l s in Aqueous and Nonaqueous Sol vents 140 Il I . 4 D i scus s i on of TRES Type 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148-

-e

. s i o

ix

TR-4 16 S=�II*I ----------------------------.----

TABLE OF CONTENTS (concl uded )

I V Coupl ed Therma l and El ectro lyti c Regeneration Based o n Pnessure Di fferences of the Working El ectroacti ve Fl u i d . . . . . . . . . . . . . . . . . 149

IV . l S i ngl e Ce, l l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 I V . l . l Conti nuous Gas Concentrati on Cells . . . . . . . . . . . . . . . . . . 150 I V . l . 2 The Sod i um Heat Eng i ne . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

IV . 2 Mu l ti p l e Cel l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

IV . 3 Summary and Di scuss ion of TRES Types 5 and 7 . . . . . . . . . . . . . . . . 167

V Coupl ed Thermal and El ectro lyti c Regeneration : Genera l . . . . . . . . . . 169

V . l Hi gh Temperature El ectro lys i s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 V . 1 . 1 �'1o l ten Sal t Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 V . l . 2 Aqueous r,1edi a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 V . l . 3 Hydrogen-Oxygen Fuel Cel l Coupl ed wi th H i g h

Temperature Water El ectrolysi s a nd Rel ated Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

V . 2 Fl uori des of Uran i um(V I ) or Ceri um ( I V ) and Arsen i um ( I I I ) : Spontaneous Charge Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

V . 3 Thermocel l Regenerators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

V . 4 D i scuss i on of TRES Type 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

V I Concl us i ons and Recommendati ons . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 189

V I I References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

x

S=�I I*I _________________________ T_R_- _4 _1_6

LIST OF F I GURES

Page

S- l Thermal Regenerati on : Type 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

S-2 Thermal Regenerati on : Type 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

S-3 Thermal Regenerati on : Type 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

S-4 Coup l ed Thermal and El ectro lyt i c Regeneration : Type 4 . . . . . . . . . . . . 8

S-5 Coupl ed Thermal and El ectrolyti c Regeneration : Type 5 . . . . . . . . . . . . 9

S-6 Coupl ed Thermal and El ectro lyti c Regenerati on : Type 6 . . . . . . . . . . . . 10

S-7 Coupl ed Thermal and El ectrolyt i c Regenerati on : Type 7 . . . . . . . . . . . . 1 1

1 - 1 General Scheme for a Thermal ly Regenerati ve El ectrochemical System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . '. . . . . . . . . . . 16

1-2 Theoreti cal Standard Cel l Potenti a l s As a Functi on of Temperature ' for Various �1etal Hydri des . . .. . . . . . . . . . . . : . . . . . . . . . . . . . . .. . . . . . . 30

1 - 3 Li th i um Hydri de Regenerative Cel l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

1-4 Lit h i um Hydride Regenerat i ve Fuel Cel l wi th a Col d-Sal t Sea l Fl ange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

1 - 5 Batch Li thi um-Hydrogen Cel l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

1-6 Permeati on I sobars of Hydrogen on Vanadi um and Armco I ron 40 Fo i l s 0 . 0254 cm Th i c k at 1 atm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 -7 Cal cul ated Equi 1 i br i urn Hydrogen Pressure for Li H-L i Cl M i xtures . . . . 43

1-8

1 - 9

Vo l tage-Current Curve for a Lith i um Hydri de Cel l wi th a 0 . 025-cm Vanadi um Di aphragm at 525°C . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45

Vo l tage-Current �u;rve; f'Glr a L tth ium �.lY�;i €l-� ·ael l w1i���� 1 45 0 . 005-cm Vanadlum Dlaphragm at 425 C . . . . . . . . . . . • . . . . . . . . . . . . . . . .

1 - 1 0 Sc heme of Therma l ly Regenerati ve El ectrochemi cal System Proposed 60 by McCu l ly et a l e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 - 1 1 Tafel Pl ots for the El ectrochemi cal Oxi dat ion of Te ( I I ) ( 1 6 Mo l e % i n A1 C1 3 ) at 200,oC . . . . . . . . . . .. . . . . .. . . . . . . . . . . . . . . . . . . . . . 62

xi

S=�I I.I ______ _

___________

______ --'T....LR..::-=4.L.U-16 -,� ��

L IST OF F IGURES ( conti nued )

1 - 1 2 Tafel Pl ots for the El ectrochemical Reducti on of Cu ( I I ) ( 4 �101 e % i n 3 : 1 Mol ar R�ti o A1 C1 3 : KCl ) at 200°C . . . . . . . -. . . . . . . . . . . . 62

1- 1 3 Po�enti a l s of the TeC1 2 (Anode ) / CuC1 2 ( Cathode ) Ga l van i c Cel l s 1 n Mo 1 ten A 1 Cl 3 ' . . . . . . . . . . . . . . . . . . . . . ' . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

1- 1 4 Di scharge Cu rves for TeC1 2 (Anode ) /CuC1 2 ( Cathode ) Cel l s under 50- and 1 00-Ohm Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

I I - l Schemati c Representation of a Therma l l y Regenerati ve Al l oy System ( a ) 'r-or Amal gam Cel l s C = Na , K; B = Hg ; and ( b ) for B i meta 1 1 i c Ce 1 1 s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . " . . . . . . . . . . . 70

1 1-2 ( a ) Constant-Pressure Phase Di agram for a 'Generi c B imetal l i c System C/B ; ( b ) Three-Dimens i onal Phase Di agram for a Two­Component System ; and ( c ) Phase Di agram Showi ng Equ i l i bri um between Vapor and Sol i d i n the V-CB Reg i on Resul ti ng from Overl ap of V-L and L-CBR,eg ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 72

I I- 3 Ternary Phase Di agram of KOH-KBr- K1 and Properti es of the Ternary Eutect i c . . . . . . . . ' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7S

1 1 -4 EMF of K/ K+g l ass/K(Hg ) Cel l s at 1 36°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

I I- 5 Vol tage-Time Pl ot of Cycl i ng D i fferent ia l Dens ity Potass i um-Mercury Ce· l l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

I I -6 Potass i um-r�ercury Liqu i d Metal Cell ( LMC ) . . .. . . . . .. . . . . . . . . . . . . . . . . S 1

1 1- 7 Phase Di agram of Hg- K at 1 atm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . �3 I I -S Open-Ci rcu i t Potenti a l s of Sodi um-Sodi um-Mercury Gal vani c Cel l s at

50Qoe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . 88

1 1- 9 Cross Section of Stati c El ectrode TRAC Cel l . . . . . . . . . . . . . . . . . . . . . . . . . SS

I I- 1 0 Di scharge Characteri st ics of a Stati c TRAC Cel l . . . . . . . . . . . . . . . . . . .. S9 1 1- 1 1 Cross Section of the F l owi ng El ectrode TRAC Cel l . . . . . . . . . . . . . . . . . . 3 1

1 1- 1 2 Vapor/Li qu i d Equi l i bri um Compos i tions of the Na/Hg System at 5.S atm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . " . " . . . . . 8 • • 93

I I- 1 3 TRAC Test· Loop Flow Di agram� . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95,

1 1- 1 4 Sol i d/L i qu i d Equ i l i br i um Compos i tions of the Potass i um-Tha l l i um System . . . . . . . . . . . . . . . . . . . . . . . . .. II • • • • • II • • . .. . . . . . II • • • • • • • 0 • • II II • • • co • •

99

xii

/.= TR-4 16 S=�I ".I --------------------------"-'-'-"---'­-� ��

L IST OF FIGURES ( conti nued )

Page

I I- 1 5 Therma l ly Regenerati ve Sod i um-Ti n System . . . . . . . . . . . . . . . . . . . . . . . . 103

1 1- 1 6 Thermal ly Regenerati ve Sodi um- Lead System Operated for 1 00 Hours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

1 1- 1 7 Vo l tage-Current Dens i ty Pl ot for the Sod i um-Lead Gal vani c Cel l Operati ng under Thermal Regenerati on . . . . . . . . . . . . . . . . . . . . . . . . . . 107

1 1 - 1 8 Improved Thermal Regenerator for the Sodi um-Lead System Operated Conti n uous ly for 1 000 Hours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

1 1- 1 9 EMF-Temperature-Compos i tion Characteri sti cs of the ' Cel l L i (fi, ) I Li Cl -Li F (fi, ) I L i xSn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 1

1 1 -20 Pressure-Compos i t ion D i agram for the L i -Sn System a t l 200°C . . . . . . 1 1 1

1 1-2 1 Vo l tage vs . Current Dens i ty of the L i th i um-Conta i n i ng B imetal l i c . Cel l s wi th the Li Cl - KCl El ectrol yte . . . . . . . . . . . . . . . . . . . . . . .. .. . . 1 13

1 1 1- 1 We i n i nger ' s I 2 I a-Ag I I 12 Thermocel l . . . . . . . . . . . . . . . . . . . . . . . : . . . . . . . 1 39

1 1 1 -2 D i agram of the Sul furi c Aci d Concentration Thermocel l and Current-Vo l tage Curves for the C�l l . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

I V- l :Laboratory Conti n uous Gas Concentration Cel l . . . . . . . . . . . . . . . . . . . . . 1 5 1·

I V- 2 Advanced Conti nuous Gas Concentration Cel l . . . . . . . : . . . . . . . . . . . . . . . 1 53

I V- 3 Schemati c D i agram o f the Sod i um Heat Engi ne . . . . . . . . . . . . . . . . . . . . . . 155

I V-4 Ca l cul ated Power-Efficiency Performance of the Sodi um Heat Engi ne for Various Va l ues of T2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 59

I V- 5 Experimental a nd Ca l cu l ated ( Eq . 1V- 5 ) Vol tage-Current Curves for the Sod i um Heat Engi ne as a Functi on of T 2" . . . . . . . . . . . . . . . . . . . 1 59

IV-6 Sc hemati c D i agram of the Sodi um Heat Eng i ne Cel l s wi th S "-Al umi na Tubes . . . . . . . . . . . . . . . . . . . . . � . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

1V-7 Low-Pressure El ectro lys i s Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

I V- 8 Laboratory Cel l L i / I 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

1V-9 Mu l ti p l e-Ce l l El ectrochemical Heat Engi ne . . . . . . . . . . . . . . . . . . . . . . . . 165

V- l Generi c Scheme for El ectrotherma l Regeneration . . . . . . . . . . . . . . . . . . . 172

xiii

S=�I I.I ______________ -------�TR=-...34�16 -� .

L IST OF F I GURES ( cont i nued )

V- 2 OCV of Cel l s Composed of As F3 and UF6 or CeF4 Separated by the So l i d El ectrolyte PbF4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

V-3 Regenerati ve, �·1oJ ten Sa l t , Thermoel ectri c Fuel Cel l Schemati c Di agram . . . . . . .. . . � . . . . . . . . . . . . . . . . . lit . " . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . 185

V-4 Laboratory Ce 1 1 Scheme for a Thermoce 1 1 Regenerator Coup l ed wi th the Ga l vani c Cel l . . . . . . . . . . . . . · . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

xiv /

- TR-4 16 S=�I I_I ----------------------------

L I ST OF TABLES

S-l Examp l es of the Thermal ly Regenerati ve El ectrochemi cal Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1 - 1 Thermodynami cal ly Su i tabl e Compounds for Thermal D i ssoc i ati on i n Thermal ly Regenerati ve Fuel Cel l s . . . . . . . . . . . . . . . . . . . . . . 20

1 -2 Compounds Sel ected as Poss i bl e Candi dates for Thermal Regeneration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

1 - 3 Summary of Gal van i c Cel l Performance . . . . . . . . . . . . . . . . . . . . . . . . 48

1-4 Cel l Vo l tages Obta i ned i n the Reacti ons of Metal Ha l i des . . . . 59

1 - 5 Summary of Experimental Regenerator Operati o n . . . . . . . . . . . . . . . 65

1-6 Thermal ly Revers i bl e Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

1-7 Characteri sti cs of Oxi de Cel l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

I I- l

I I 1 - 1

I I I -2

Operation of the NajPb Therma l l y Regenerati ve System . . . . . . . . 104

Summary of Thermoel ectri c Powers i n Ag i Mol ten Sal t i Ag Thermoce 1 1 s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

Summary of I n i ti a l Thermoel ectri c Powers in Metal i Mol ten Sal t i Metal Thermocel l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

I II-3 Summary of Transported Entrop i es ( SMn+ ) , Partial Mol al Entro-pies (SMn+), and Entropi es of Transfer (SMn+) for Molten Sa 1 t Thermoce 1 1 s . . . . . . . . . . . . . . . . -: . . . . . . . . . . . . . . . . . . . . . . . . . . 1�6

I I I-4 Summary of Thermoel ectri c Powers i n Thermocel l s X2 i �1o l ten Sa 1 t (JI,) i X2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

1 1 1 - 5 Summary of Thermoel ectri c Powers of Sol i d- El ectrolyte Thermoce 1 1 s . . . . . . . . . . . . . . . . . . . . . . . . . . '. . . . . . . . . . . . . . . . . . . . . 133

1 1 1 -6 Summary of Thermoel ectri c Powers of Some Ioni c Sa1ts . . . . . . . . 135

I I I- 7 Summary of Transported Entropi es ( SMn+ ) for Soli d El ectrolytes of Ordered Structure and Summary of Overal l Transported Entropi�s for Sol i ds of the a-Ag I Type and Ag I -Ag Oxysa l ts ( 75-80 Mol e % Ag I ) . . . . .. . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . � . . . 136

1 1 1-8 Performance of the Iod i ne Cel l . . . . . . . . . . . : . . . . . . . . . . . . . . . . . . 140

xv

- TR-4 16 S=!!!SI r�.-�r ----------------------� ���

L I ST OF TABLES ( conti nued )

1 1 1 - 9 Thermoel ectri c Powers o f the Cu l El ectrolyte l Cu Cel l s i n

I I I- l 0 ,

IV- l

I V- 2

V- l

V-2

V-3

Aqueous a nd Nonaqueous Sol vents . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Thermoel ectr ic Powers for Some Sel ected Thermogal van i c Cel l s i n Aqueous Medi a . . . . . . . . . . . . . . . . ; . . " . . . . . . . . . . . . . . . . . . . . . . . 144

Vol tage Characteri s t i cs of the I 2 ( T2 ) I Pb I 2 ( � ) I I2 ( Tl ) Cel l 1 52 wi th Ni El ectrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Characteri s t i cs of Conti nuous Gas Concentrati on Cel l s wi th Several Worki ng Fl u i d/ El ectrolyte Systems . . . . . . . . . . . . . . . . . 154

Cel l Potenti a l s and Res i stances as a Functi on of Temperature for the Fe ( CN�:- I Fe ( CN )�iand CU ( NH2 )�+ I Cu ( NH3 ); Systems . . . 179

c._" I " I deal Effic i ency for H2-02 Thermoel ectrochemi cal Power

18 1 eye 1 e . . 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .

I deal Effi c i ency for H2-02-HC1 ( aq ) Thermoel ectrochemi cal 183 Power Cycl e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xvi

TR-4 16 5i::�1 I�I --------------------�-------------------------------------------------­-� ��

INTRODUCTION

OBJECTIVE

This review of thermally regenerative electrochemical systems (TRES) was writ­ten upon request of T. A. Milne of the Chemical and Biological Division of the Solar Energy Research Institute . The bulk of the information contained in this report was. collected from February to October 1979. The information comes from literature searches and from visits to the laboratories of, and discussions with, technical personnel involved with this type of research. Work on TRES has been classified and analyzed, with emphasis on the operation of the electrochemical systems. It is important to emphasize, however, that TRES involve the merging of several disciplines in addition to electrochemis­try: thermal conversion, engineering, materials science, etc . The purpose of this review is to aid evaluations of the electrochemical technique of direct thermal-to-electrical conversion.

The majority of the systems published in the open literature or patented are reviewed. Because this area was developed from the late fifties to the late sixties with the utilization of heat from nuclear reactors as a major mission, most of the systems operated at high temperatures. Our review covers these systems and a variety of others developed or proposed, which operate under a wide variety of conditions. A cross-comparison, or ranking, of systems for different missions, at different stages of development, and at different oper­ating conditions is not feasible for this review.

The systems investigated in the past are reviewed in detail in Vol. 2 and more briefly in the Executive Summary. It is our purpose to suggest areas of research (in either science or engineering) that would benefit from a long­range TRES program, rather than to propose specific systems for further device exploration. In the past, most of the funding and expectations were device oriented in short-term programs. The systems for which there existed a better understanding of the chemical, electrochemical, engineering, and materials problems were pursued in relatively long-term and research-oriented projects.

The Executive Summary has the same structure as Vol. 2, so that references and further technical background can be located easily.

BACKGROUND

Regenerative electrochemical systems were one of a variety of complex methods of energy conversion investigated during the period from 19 58 to 1968. In these systems the working substance produced in an electrochemical cell (fuel cell, battery, galvanic system, emf cell) is regenerated by the appropriate input of energy (thermal, light, atomic, electrical, or chemical), thereby defining the thermal, photo-, nuclear, electrolytic, or redox regenerative electrochemical systems [1]. The major heat source envisioned during this period was nuclear reactor heat. Direct use of sunlight for photoregeneration also was attempted , as well as use of nuclear radiation. The electrolytically regenerative systems are essentially indistinguishable from secondary batter­ies and were explored chiefly for their possible utilization in load leveling,

1

TR-4 16 S::�I I�I -----------------------------------------------------------------------­-� �

for their storage systems appeared capability [2].

capability, and for space-flighto application. The redox particularly attractive because of their energy storage

Austin [3] critically reviewed the government-sponsored fuel cell research from 1950 to 1964, including regenerative types, for possible space power application or for silent and portable electric generators. Kerr [4] reviewed work up to 1967, comparing the different types of regeneration for space power application. Nuclear, photo-, <and redox systems were eliminat�d from consi­deration due to weight, complexity, and low efficiency. The proceedings. of a symposium on regenerative emf cells .[5], published in 1967, includes discus­sion of most types of regenerative systems.

Because of the low overall eff iciencies [3, 4, 5] of the regenera ti ve sys tems due to Carnot cycle limitations (thermal), problems of pumping, plumbing and separation (thermal and nuclear), and low quantum Yields (photo-), research after 1968 was concerned primarily with electrolytic regenerative [6] and· redox [2] systems. However, because thermal energy can be obtained by har­nessing the sun's rays, it is possible to envision TRES operating under con­ditions that differ markedly from those offered by nuclear reactors. It is therefore conceivable, as pointed out by Kerr [4], that the problems of TRES for some applications/are surmountable. In this report, we classify thermally regenerative electrochemical systems as systems regenerated by the input of thermal or coupled thermal and electrolytic energy. Because the seven types of TRES have unique features., a general introduction is not given at this point.

TYPES OF THERMALLY REGENERA�IVE ELECTROCHEMICAL SYSTEMS

Thermally regenerative electrochemical systems are closed systems that convert heat into electricity in an electrochemical heat engine that is Carnot-cycle limited in efficiency. In this repor't the TRES have been classified into two broad classes: systems regenerated thermally and systems regenerated by a coupling of thermal and electrolytic inputs. To facilitate discussion, the types of TRES within these two broad classes have been designated gccording to significant differences in either the electrochemical cells or regenerators.

Sections I and II concern thermal regenergtion, and the following three types of TRES are defined:

Type 1.

Figure S-1 illustrates this type of system. The electrochemical reaction product CA is formed from C and A in an electrochemical cell at T1, with concomitant production of electrical work in the external load. For such a production of electricity to be continuous, compound CA must be easily decomposed into C and A. Thus, compound CA is sent to a regenerator at T2 via a heat exchanger. In the regenerator, the thermal decomposition rea:c­tion takes place spontaneously. Compounds C and A formed in the regenera­tor at T2 are separated by physical (or chemical). means, and the isolated compounds C and A are returned to the electrochemical cell after being returned to temperature Tl through the heat exchanger, thus completing the

2

/. '", TR-4 16 S=�I I I.II ----------------------------------­-� ��

cycle. The most favorable thermodynamic properties of the electrochemical reaction for a thermally regenerative electrochemical system are: !J.G < 0, !J.S < 0, and !J.Cp as close to zero as possible.

Type 2.

Figure S-2 illustrates this type of system. In this case, a more complex set of galvanic cell reactions occurs at T1• Two or more products are formed in the electrochemical reaction; therefore, the regeneration of the anode and cathode materials (A and D) must be performed separately at T2 and T3, as indicated in Fig. S-2. It is a more complex scheme, requiring more plumbing, heat exchangers, and regenerator chambers than the simple system of Type 1.

Type 3.

Figure S-3 illustrates this type of system. In principle, this scheme is identical to that of Type 1. However, it applies to specific cases in which the electrochemical cell reaction involves only one electroactive couple C+/C in a concentration cell at T1• C(B) represents, for instance, an alloy or a bimetallic compound.

Sections III, IV, and V concern coupled thermal and electrolytic regeneration, and the following four types of TRES are defined:

Type 4.

As illustrated in Fig. S-4, compound CA formed in the galvanic cell at Tl is sent to a regenerator at T2 via a heat exchanger, where it is electro­lytically decomposed into C and A. The requirements for this type of regeneration are that the cell reactions C t C+ + e- and A + e- t A- are· reversible and of high coulombic efficiency and high exchange current. In addition, the voltage V(TI) must be larger than V(T2). The cells are con­nected in electrical opposition and the electrolysis takes place consuming V(T2). The remaining voltage can be used to perform useful work in the external load. The separation is inherent in this type of regeneration. Compounds C and A are returned to the galvanic cell via heat exchangers, and the loop is closed. If at a temperature Tx the reverse reactions of reactions in the galvanic cell take place spontaneously, then the regen­eration produces an additive voltage V(Tx) while regenerating C and A at Tx'

Type 5.

This type is illustrated in Fig. S-5. Two galvanic cells at the same temperature are arranged so that the activity of one of the electroactive species can be varied by some physical means. In the example shown, a cold finger reduces the pressure of the gaseous working electroactive fluid A. The galvanic cells are connected back to back. The galvanic cells are concentration cells in the A/A- species. As cell I discharges, cell 2 charges and work is performed in the external load proportional to the differences in activities of A in the two cells. The operation is interrupted as A-rich material is consumed. The switch reaction corresponds to heating the cold finger associated with cell 2 and cooling that associated with cellI. The roles of the two cells are now reversed

3

TR-4 16 !;::�I 1!4I1 ------------------------------------------�------------,---------------

and the system can again perform electrical work in the external load. This scheme is equivalent to an electrolysis performed at reduced pressure. Mass transfer could be the major limitation to this type of TRES.

Type 6.

This type of TRES is illustrated in Fig. S-6. The thermal and electroly­tic paths are not separated. Two or more electrodes are at different tem­peratures. These electrodes (not necessarily chemically identical or reversible) are in contact with the electrolyte (liquid or solid, not necessarily homogeneous in composition, and with or without permeable mem­branes interposed in the electrolyte), in which a temperature gradient exists. These TRES are called thermogalvanic or nonisothermal cells. During the passage of current in the cells, matter is transferred from one electrode to the other as a result of the electrochemical reactions at the electrolyte/electrode interfaces and ionic conduction.

If the transfer of matter is permanent, as occurs with electroactive metal electrodes, the electrodes must have their temperatures reversed periodi­cally for continuous operation of the engine as a power source. This tem­perature reversal operation can be avoided if gas electrodes, or redox soluble couples, ate used. These thermogalvanic cells are the electro­chemical analogs of thermoelectric devices. The efficiency in these devices is related to the Carnot efficiency. The upper limit is deter­mined by the use of expressions developed for solid-state, thermoelectric devices. These equations take into account the Carnot efficiency, the thermal and electrical conductivities, and the thermoelectric power (dE/dT) of the system, but they do not t�ke into account electrode polari­zation effects characteristic of the electrochemical reactions.

Type 7".

In this type, illustrated in Fig. S-7, the thermal and electrolytic paths are separated. An isothermal electrolyte (solid or liquid) separates the working electroactive fluid from two pressure regions. The work performed in these engines is equivalent to the isothermal expansion of the fluid from the high pressure, high temperature zone (PH,TH) to a low pressure zone (PL,TH) separated by the electrolyte and created by cooling one end of the engine. The element C undergoes oxidation, the electrons traverse the external circuit, and ions C+ cross the electrolyte as a result of the pressure differential across the electrolyte. The C+ ions are reduced at the electrode attached to the bottom of the electrolyte, at a lower pres­sure. At the cold trap, C is condensed. To produce electricity continu­ously with these engines, C( PL, TL) must be pumped to C( PH' TH) • These engines do not need a chemical separation step.

Table S-l presents a summary of typical examples of all seven types of TRES. Inspection of the table shows that these systems can cover various ranges of temperatures from near room temperature to 1000°C. In the systems studied to date, powers of 0.1 mW/cm2 to 1 W/cm2 have been achieved. Emphasis in past work was placed on systems regenerated at high temperatures.

4

Th�rma! Regenerator Electrochemical I � + e-

• C+ + e- Anode at T2 Cell �eactions • A-t

! Cathode

C+A T1• .CA C I I CA A

!"Ieat �xchange CA(T1) • CA(T2)

�1�ctrocheR'!icctl �y�tem: Ther!11ct! Regenerator CA -.:!i� C+ A 01 ----. Fu�1 Cell or· Gctlvanic

Cell (Battery) at T1 Heat �xchange C + A (T 2) --.. C + A (T 1 )

RL T1 7' T2

Figure S-1. Thermal Regen�ration: Type 1

In III N -".=� II II �= v

"""3 � I oj::. -0)

r-

� A

CJ)

L...

I

Anod� C�tllode Regenerator Regenerator .. at T2 B at T3 I....-. I....-.

C E

t t ��Ivanic Cel!� Two or �ore Reactioll Products at T 1

1\ I RL

T1 jt T2 � T3

�

I I

� o

. G�!vallic Cell R��ctions

Heat Excilange

A + B- � C + e- Anode o + e- � E + B- Cathode

T, A+D • C +E

C(T,) ----.. C(T2); E(T1) � E(T3) T2

Anode Regeneration C • A + B Cathode Regeneration E + � T 3.. 0 Heat Exchange. A(T2)----. A(T,); D(T�) ----. D(T,)

Figure 8-2. ThEnm,,1 Regeneration: Type 2

III III ,-u -/..ii� II II �-�

>-3 � "'" I-' CJ)

J� [C(B)1 !

-:II

Th�!,mal Regel:1erator at 1"2

C-Int [C(B)]

t Gillvanlc Cell: Concentratiol'! eel' at T1

T,"t T2

c:-popr [C(B)]

!;!e�tn��!le",ic!l! Cell Re!lctions

Heat Exchilnge

Thermal R�geije;'ator

Heat J:xchange

C+ + e- • C-poor [C(B)] Anode Cathode I C-rich [C(B)] • C+ + �-

C-rich [C(B)] T1 • C-poor [C(B)]

(C-rich + C-POor)[C(B)]T, � C-lnt [C(B)h2 T2

C-Int [C(B)] ---..-.� C-poor [C(B)] + C-rich [C(B)] (

C-poor [C(B)]T2 + C-rich [C(B)]T2 •

C-poor [C(B)h, + C-rich [C(B)h,

Figure 8-3. Tt,@rmal Regen�rCltion: Type 3

III III N -.,:;;:� II II ���

>-3 � I ,.)::>. ...... (j)

-

C l 00 I---

Electrolysis C�II at 12

G�lva!lic Cell �eactions

t � CA A ? c

> RL Heat E�cha!lge >

Electro!ysi� Galva!lic C�II at T1 Cell Reactjons

T1fT2 Heat Exch�nge

.-

C ---.. 6+ .+ e-

A + e- )I K

T, C+A )I

CA(T, )

A- � A + e-I C· •• - .C

CA T2 •

Anode

Cathode

CA

• CA(T2)

Anode

Cathode

C+A

[C + A] (T2) --_ ....... [C + A](T, )

Fig�rf) S-4. Couplect Th�rm�t an� Electrolytic Regeneration: Type 4

In III N -/..;;;� II II ���

>-3 1 ...... 0)

,......

to!

-

I Ga!vanic C�II 1 at T H A-rich

Trap or Cold Fin�,r at TH

I Galy�"ic Cell 1 AtTH A-poor

Trap Qr CQ!d Finger at TL

1 Galvanic Cell 2 Q) at TH Ol Q) .... em A-poor III ...

.r= <II 0 .r= (J) u Trap or Cold 6 I Finger at TL

�vA AL

C9r!!:en'r�Uon cell �Y�"1ll Opera,es: A:r!�h -+ A�poQr Swi,ch Colc:l flng!ilr Telllpera'Llre and Reverse Cell FUllctlons

I c:;alvanlc

Q) Cel, 2 Ol aJ TH Q) a; Ol A-rich .... .r= <II 0 .r= til u · is

'V' AALA ... v .... "

Trap or Cold Finger at TH

EI!ilctrolysis at Low Pr�ssu!"

h I

I

'-

Electrochemical c,�!i i ���ctions (Discharge)

J:'!tqtrQcl'!�mICil! pel! (2) �e!lctions (Ch!lrge)

Cold Fing",r

Swltcl'! R�ac'ion (�o P9wer to External Circuit)

�I�c�roch!tm!cal C�I! 1 Rj!actions (Charge)

Cold Finger

Electrocl'!ernlcal Cell 2 Reac,ions

JDisch,arge)

C .• C+ +e-

A-rich + e-

C + A-rich(g)

• A-TI-f

A- ---.. A(g) + e-

Anode

Cathode

• CA.

Anode

C.,. + e- --...; .. � C Cathode

CA TH • C + A(g)

"'" A(g)T A(I or S)TL H ' { Cell Cold Finger at T L.

Cell 2 Cold Finger at T H A- • A(g) + e- Anode

C+ + e- � C Cathode

TH CA • A(g) + C

A(g)TH :.. A(I or s)TL , I :-clCh Ig) + e-

li> C+ + e- Anode

!Jo A- Cathode

C + A-rich(g) TH • CA

f,'igi:Jre S�f;. Oo�pl�� T�ermitl@nd. Elecfrolytic Regenera�ion: Type 5·

III In N -• II II ��'P

� � � -en

...... o

I RL

'I'VVV\(

Electrochemical System: 6C1!val"!ic Cell.

'

EI�ctrodes at TL Clnd THo COl,fpl�d Electrolytic al"!d Thermal Paths

I C I ,�

Electrode at T,

Electrochemical Cell 'Reactions

C+A- I C

(Solid or Liql,fid) Electrode

{

at T2

C • C+ + e-

C+ + e- � C

Anode Tl

Cathode T2

IfC = Metal, the roles of electrodes at T, and T2 have to be reversed periodical ly.

Figure S-6. Coupled Therm�1 and ElectrQlytic Regeneration: Type 6

III III N -•-=t\'-II II �� v

>-3 l ,;.... (j)

·-l r � �'-;>

--

CPw TH I TH

C+ Conductor T!-I

C PL' TL --

tc - -

Electroc!1emical Cell Reactions

Cooling

P"mping �Ild !-Ieating

Cp T (9) H' H

+ CPH

C+ PL + e

CT P (9) H' L.

GT P (I) L' L

C+ + e PH

+ 711""""" CPL

Anode

Ionic Conductor

CT P (g) Cathode H' L

CT P (I) L' L

CT H,PH (9)

Figure 8-7. Coupled "Ttlermal �nd &Ieetrolytic Regeneration: Type 7

III III N -

".=01-II /I ���

� � "'" I-' Cf)

Table S- l .

Type of Regeneration System

Type I LiR

Type 3 Na l Rg T2>TI

- Type 3 Na l Pb to.:)

Type I or 4 Sn+Sn2++2e -T2<TI 2Cr(III)+2e -+2Cr(II)

Type 3 or 4 K I TI

Type 4 Na I NaCl l C12 T2>TI

Type 4 Li ILil 1 12

Type 4 WF6+2UF5+2e-+2F-T 2>T 1 ; lIC'T2 <0 AsF3+2e -+2F-+AsF5

EXAMPLES OF THE THERMALLY REGENERATIVE ELECTROCHEMI CAL SYSTEMS

Performance (cells at T I )

Voltage at Current

Projectedd Open Density

Carnot Circuit T2 TI P Efficiency Efficiency Voltage V I

lUec trolyte ( ·C) ( ·C). (atm) (%) (%) (V) (V) (mA/cm2) Comments

eutectic molten, LiR-LiCl-LiF 900 530 32 0 . 6 0 . 3 200 Static. cell ; 0 .025-cm

vanadium diaphragm as R2 electrode. Closed-loop sys�em not tested .

molten NaCN:NaI : NaF -685 -495 9 -20 -6 0 .32 0 . 2 2 5 High resistivity o f 58 :30 : 12 mol % alumina matrix impreg-

nated with electrolyte; total 1200-h operation of which 750 h were closed-loop operation.

molten NaF:NaCl :NaI 875 575 8/760 26 9-12 0 .39 0 .18 100 Complete system operated 1 5 . 2 : 3 1 . 6 : 53.2 mol % -100 h. Regenerator

only operated 1000 h.

aqueous , excess el- 20 80 -0. 1 0 .06 3a Periodical power source .

molten KCI 175 335 0 .6 Regeneration by cooling and separating the two phases (liquid and solid) •

molten NaC! 827 3 .24 I up to 4 . 3 Low coulombic efficien':'" A/cm2 with cies (40% Na utiliza-IR drop only tion ) . Low electrode

polarization on dis-charge at -lOOO·C.

molten L'LI 1 170 500 50 •• 6 -18 2 .50 1 .5 320 Closed-loop system not tested. Two mol % dis-solved in Lil .

solid electrolyte >900 25 -0 .5(25·C) No voluage-current datao PbF2(KF doped) +O .3(900· C)

In I'll jIU -•=� II II < - �

I >-3 � I � -0)

I-" �

Table S-l . EXAMPLES OF THE THERMALLY REGENERATIVE ELE CTROCHEMICAL SYSTEMS ( Concluded)

Type of Regeneration

Type 5

Type 6

Type 6

Type 6

Type 7

Type 7

aCurrent In mAe

System

12 1 mol ten alkali I 12 metal iodides

12(T1 ) I a-AgI 1 12(T2)

Cu(T1 ) I Cus04 1 Cu(T2)

4- 3- 1 Pt I Fe(CN)6 , Fe(CN)6 Pt

Na I S"-A1203 I Na

12 I PbI2( t) 1 12

bFor T2 = 500°C and T1 = 200°C.

Electrolyte

molten electrolyte

solid electrolyte a-AgI

aqueous acid

�queous

solid electrolyte

molten PbI2

T2 T1 P ( OC) ( O C) (atm)

350 350

340 184

100 20

80 30

800- 100-900 200

170- 20-400 100

cT2 enough to give 1 atm iodine; 24 . 5-ohm load; electrolyte temperature of 540°C.

Carnot Efficiency

(%)

25

21

14

-60

dSome of the numbers in this column were obtained from a minimum amount of experimental data.

Performance (cells at T1 )

Voltage at Current

Projectedd Open Density Circuit

Efficiency Voltage V I (%) (V) (V) (mA/cm2) Comments

-30 0 . 29 0.23 100 Cold finger at 25°C. Mass transfer problems .

_5b 0 . 2 0 . 1 1 .4a Internal resistance .... 70 ohms; expected practical efficiencey 1%-2%.

-0.09 0 . 03 8 Saturated solutions at each temperature.

0 . 08 Maximum power estimated <0 . 1 mW/cm2 •

-25 -1 .2 0 . 7 1000 Voltage losses due to interfacial polariza-tion and thickness of S" alumina electrolyte.

O . l 7c 6 . 2a , c Liquid electrolyte in-tegrity difficult to maintain due to the pressure gradient .

III III N -".;� II II �-�

>-3 � ,j::. I-" 0)

1 4

S=�I I.I ________________________ ....::T..::.R:.....,.-.,::.4 1-'-'-.-S

-� ��

SECTION I

THERMAL REGENERAT ION : �1ETAL HYDR I D ES , HAL I DES , OX IDES , AND CHALCOGEN I DES

The general scheme for a thermal ly regenerative el ectrochemi cal

system ( TRES ) of Type 1 i s shown i n Fi g . 1 - 1 . It cons i sts of an el ec-

trochemi cal cel l i n whi ch substance CA i s formed el ectrochemi cal ly from

C and A at temperature TL wi th producti on of el ectri cal energy . The

worki ng s ubstance CA i s then heated and fed to the regenerator at tem­

perature TH where- , t undergoes thermal decompos i ti on i nto A and C . A

coo l i ng s tep comp l etes the cycl e , thus regenerati ng C and A at the l ower

temperature TL .

El ectrochemi cal Ce l l reacti on at TL

Heati ng

Therma l Regenerati on at TH Cool i n g

C

A + e

C + A

CA

C + A ( TH )

-+ CA

C + A

I

I I

I I I

IV

Thi s cycl e can be consi dered a heat eng i ne whi ch converts part of the

energy absorbed at a h i gh temperature i nto useful work and rej ects the

rema i nder as heat at the l ower temperature . The study of TRES was s ug­

gested by Yeager [ 7 ] i n 1 958 .

L i ebhafs ky [8] has shown that though the effi c i ency of the fuel

cel l escapes the Carnot cycl e l imi tati on , the combi nation fuel cel l heat

source i n TRES does not, havi ng a maximum theoreti cal effici ency

1 5

TR-4 16 S=�I I;.��I -----------------------------

-� �� �

r ­I L._

___ TH --­

CA ----... C + A

-

A + e- --. A-

C + A �T CA . L

l I .J

Heat . Exchanger

Figure 1-1 . Gene,ral Scheme for a Thermally Regenerative Electrochemical System

16

S=�I I.I _______________________ --'T"-"'R"'--�4_=16"__ -� �

for revers i bl e steps I and I I I and 6Cp = 0 ( CCA = Cc + CA ) of

e l ectri cal work heat i nput

=

The open-c i rcu i t vol tage of the cel l i s

1 - 1

1 -2

where 6Hl i s the enthal py of di ssoci ati on �f CA i nto C and A at TL " One

of the chi ef sources of i rrevers i bi l i ty i n these cel l s i s the heat

exchanged duri ng heati ng ( I I ) and cool i ng ( IV ) . Thi s can be mi n i mi zed

by hav i ng 6Cp as c l ose to zero as poss i bl e . The des i rabl e thermodynami c

propert i es for a regenerati ve fuel cel l reacti on are 6G I < 0 ; 6S 1 < 0 , rand 6Cp_

�S c l ose �zero as--poss i bl e [9] ' 1 ! I n practice the cel l i rreversi bi l i ti es prevent the worki ng vol tages

. from atta i n i ng the revers i bl e potenti a l s ( Eq . 1-2 ) when the cel l i s

operatirig a t useful l oads [ 1 0 ] . From a practi cal po i nt of v i ew i t i s

important that CA decompose a t reasonabl e temperatures and that C and A

can be eas i ly separated , preferabl y by bei ng i n di fferent phys i cal

states . More e l aborate thermodynami c treatment of these systems were

gi ven by deBethune [9] and Fri auf [ 1 1 ] .

Henderson , Agruss , and Capl e [1 2 ] presented a practi cal approach to

the determi nation of effi c i enc ies of TRES and to the thermodynami c

sel ecti on of candi date systems . In the i r deri vati on of effi c i enci es ,

they assumed that the compound CA l eaves the heat exchanger to go to the

regenerator at a temperature TE , and that the compounds C and A, after

be i ng coo l ed through the radi ator , l eave th i s heat exchanger at a

temperature TR . The temperatures of the regenerator and of the fuel

17

- TR-4 16 S=�I I_J ----------------------------=---=--

cel l are respecti vely TH and TL · The effi ci ency of the heat exchangers

i s ass umed to be 95% . The cel l i rrevers i bi l i t i es ( po l ari zation and

ohmi c l osses ) are represented by El oss ' wh i ch i s the cel l operati ng vol ­

tage ( E ) l ess the revers i bl e cel l potenti a l ( ER ) .

The effi c i ency express i on for 6Cp = 0 wi th the i nc l us i on of cel l

i rrevers i bi l i ti es and heat exchanger l osses i s

n = �I

wh i ch reduces to the Carnot express i on for an i dea l and revers i bl e sys-

tern operati on .

In pr�cti ce , few systems present 6Cp = 0 , but CCA can be l arger or

smal l er than Cc + CA : The effi ci ency of the sys tem wi th CCA > Cc + CA '

for wh i ch the heat content of C and A i s not enough to ra i se the temper­

ature of CA from TL to TH , i s expressed by

n = [-TH6SH + TL6SL - CCA( TH- TE ) - 0 . 05 ( CC+CA ) ( TH-TL ) - nFEl oss J

[-TH6SH + CCA (TH-TE ) ]

I f CCA i s l es s than Cc + CA ' the heat conta i ned i n C and A i s

s uffi c i ent to ra i se the temperature of CA from TL to TH , and i f l arger ,

the excess wi l l have to be radi ated away . The effi c i ency express i on for

th i s case i s

n = [-TH6SH + TL6SL - 0 . 05CCA( TH- TL ) - ( CC+CA ) ( TR-TH ) - n FE1 os s ]

[-TH6SH + 0 . 05CCA ( TH-TL ) ]

1 8

I - 3

1 -4

1 -5

S=�I I_I __________________________ T

_R

_-4_

1_6

The s urvey of thermodynami cal l y fea s i bl e systems for TRES conducted

at the Al l i son Di vi s i on of General Motors by Henderson et a l . [ 1 2 ] re­

v i ewed over 900 compounds , of whi ch on ly 20 ( see Tabl e 1 - 1 ) were cons i d­

ered s u i tabl e for further eval uati on as thermal ly regenerati ve el ectro­

chemi cal systems . The cri teria to excl ude the rema i n i ng compounds and

some exampl es of excl uded systems are the fol l owi ng :

( 1 ) More than two products are formed after or duri ng decompos i ti on .

I f several compounds resu l t from the thermal decompos i ti on of CJS..� the separati on of the i ntermedi ates i s necessary before these compounds

can be recombi ned i n the e l ectrochemi cal cel l . The separation of vari ous

products i s a probl em for both terrestri al and space appl i cati ons . The

e l ectrode reacti ons for s uch an operation wou l d be h i ghl y comp l ex .

The fol l owi ng are exampl es o f compounds exh i biti ng mul ti product di ssoci ­

ati on . .

5Ba ( IO) 2 -+

Ba5 ( 1 06 ) 2 + 902 + 4 1 2 +

2CdS04 (JI,) -+ 2Cd ( g ) + 2S03 ( g ) + 02 ( g ) +

2Pb ( N03 ) 2 ( s ) -+

2PbO ( s ) + 4N02 ( g ) + 02 ( g ) +

2 Kf4n04 -+

K2Mn04 + Mn02 ( s ) + °2 ( g ) +

2NaHC03 -+

Na2C03 + H20 (g ) + CO2 ( g ) +

( 2 ) Two gaseous decomposi t ion products are formed .

Gas separation i s extreme ly d i ffi cul t . A n exception i s hydrogen ,

due to the poss i b i l i ty of separation by di ffu s i on through metal fo i l s .

The systems i n wh i ch one of the gases formed was hydrogen were not

19

- TR ... 4 16 S=�I I_I--

-----------------------

TABLE 1 - 1 . THER�,10DYNAM ICALL Y SUITABLE COMPOUNDS FOR THERMAL DI SSOC IATION IN THERMALLY REGENERATIVE FUEL CELLS [ 1 2 J a

HYDRIDES

2Li H ( �) -+ 2Li ( �) H2 ( g ) +- +

2KH ( �) -+ 2K( �) H2 ( g ) +- +

2CsH -+ 2Cs ( �) H2 ( g ) +- +

2RbH -+ 2Rb ( �) H2 ( g ) +- +

Ga2H6 -+ 2Ga ( �) 3H2 ( g ) +- +

CHLORIDES AND OXYCHLORI DES

2CuC1 2 ( �) -+ 2CuC1 ( �) C' 2 ( g ) +- +

2CuOC1 2 ( �) -+ 2CuC1 2 ( �) + °2 ( g ) +-

BROMI DES

2CuBr2 ( �) -+ Cu2Br2 ( �) + Br2 ( g ) +-

2HBr -+ H2 ( g ) Br2 ( g ) +- +

PbBr2 ( �) -+ Pb ( �) Br2 ( g ) +- + -+

2Ti Br3 +- Ti Br2 ( �) + Ti Br4 ( g )

IODI DES

2H I ( g ) -+ H2 ( g ) 1 2 ( 9 ) b +- +

PbI 2 ( � ) -+ Pb ( � ) 1 2 ( g )c +- +

SULF I DES

2Bi 2S3 ( � ) -+ 4Bi { � ) 3S2 ( g or � ) +- + -+

2Bi +- 2Bi{ � ) + S2 ( g or ,Q.) 4CeS (� )

-+ Ce ( g ) Ce 3S4 ( � ) +- +

-+ H2S +- H2 ( g ) + S ( � or g )

PEROXIDES

2H202 -+ 2H20 ( � or g ) + 02 ( g ) +-

CARBONATES

T1 2C03 (,Q. ) -+

T1 20 ( � ) CO2 ( g )d +- +

f a See text for cri teria empl oyed i n the sel ecti on of these compounds .

! b At 283°C , 1 8% of HI i s d i ssoci ated . l' c ' I ; At 750 °C , l og P I ( atm ) = - 3 . 22 .

\ d . 2 l� At 353° C , PCO 1 S 470 torr.

2

20

_ /. "' , TR-4 16 S-�I II.II ------------------------­-� �

exc l uded . If techni ques for separation of gases are i mproved , add i -

ti onal systems can be eval uated . Typ i cal exampl es are :

As2S2 -+ 2As ( g ) + S2 ( g ) +-

2 I Br ( g ) -+

I 2 ( g ) + Br2 ( g ) +-

2HgO ( s ) -+ 2Hg ( g ) + 02 ( g ) +-

CdI 2 ( g ) -+ Cd ( g ) + I 2 ( g ) +-

( 3 ) Decomposi ti on takes pl ace above 1 000°C .

The arbi trary l i mi t o f l OOO°C a s a maxi mum al l owabl e temperature

was chosen cons i deri ng the severi ty of materi a l s probl ems above thi s

temperature . Many compounds were found i n th i s c l ass , for i nstance :

Sb2S3 ( ,Q, ) -+ 2Sb ( ,Q, ) + 3S ( 9 ) +-

2Cu2O ( ,Q, ) -+

4Cu ( ,Q, ) + 02 ( g ) +-

2HC1 ( g ) -+

H2 ( g ) + C1 2 ( g ) +-

MgC1 2 -+ Mg ( g ) + C1 2 ( g ) +-

2NaCl -+ 2Na ( g ) + C1 2 ( g ) +-

( 4 ) Decompos i ti on i s exp l os i ve o r exothermi c .

On ly a few compounds were found i n thi s category , e . g . , NH4 I 04 ,

( 5 ) So l i d i s present among the decomposi t ion products .

Thi s restr icti on was imposed by the space appl i cati on envi s i oned at

the t ime . For terrestri a l appl i cation these systems cou l d be recon-

2 1

/.� TR-416 S=�I II.I ---------------------------­-� �

s i dered from a thermodynami c po i nt of v i ew . Some examp l es of th i s c l ass

are :

are :

CaC03 ( s ) -+ CaO ( s ) + CO2 ( g ) +-

CoBr2 -+ Co ( s ) + Br2 ( g ) +-

2CuS ( s ) -+

Cu2S ( s ) + S ( t ) +-

Cu2Br2 ( t ) -+

2Cu ( s ) + Br2 ( g ) +-

2AgCl ( t ) -+

2Ag ( s ) + C1 2 ( g ) +-

The cri teri a for sel ecti on of the compounds l i sted

( a ) apprec i abl e decompos i ti on at or bel ow 1 000°C ;

( b ) n o sol i ds present;

( c ) compound decomposes i nto only · two products ;

( d ) on ly one product gaseous ;

i n Tabl e I - l

( e ) compound and one product are l i qui d at the decompos i ti on temperature ; and

( f ) products rema i n l i qu i d at fuel cel l temperature .

Ki n g , Ludwi g , and Rowl ette [ 1 3 ] have analyzed twe l ve thermodynami c

model s , wh i ch treated the effects of revers i bl e and i rrevers i bl e cond i -

t i ons and fi n i te spec i fi c heat d i fferences to obta i n the thermodynami c

effi c i enc i es of the · systems . The twe l ve model s are based on :

( 1 ) D i ssoc i ati on occurs only at TH . Equ i l i br i um atta i ned at the re­

generator outl et but not at the i nl et . Four mode l s are poss i bl e :

( a ) revers i bl e cycl e wi th �Cp = 0 ;

22

( b ) Irrevers i bl e cycl e wi th �Cp = 0 ;

( c ) revers i bl e cycl e wi th �Cp � 0 ; and

( d ) i rrever�i bl e cycl e wi th �Cp � o .

( 2 ) Di ssoc i ati on and equi l i br i um are achi eved at a l l temperatures

between TL a nd TH . Two model s are po ss i bl e :

( e ) �C = 0 ; and p ( f ) �Cp � O.

( 3 ) I f the equi l i br i um i s not atta i ned a t the regenerator outl et , s i x

model s ( g through 1 ) are poss i b l e , equ i val ent to those i n ( 1 ) and ( 2 ) .

The effi c i enc ies for these model s are :

nd �- nf -� ng_ l � nb x n InC t - , c arno ... _----

1 -6

I-7

1 -8

1 -9

where KH and KL a re the equi l i bri um constants for the di ssoci at ion reac­

t i on at TH and TL and �H1 5 i s the extra heat to be rejected by the

materi a l as i t cool s s i nce i t cannot be exchanged tota l l y i n the heat

exchanger .

The cal cul ati ons performed by Ki ng et al . [ 1 3 ] are based on the

23

- TR-4 16 S=�I IWI -------.:..------------------------- � ��

ana lys i s of the term �GL/�HL i n Eq . 1 -6 . � H was assumed to be i nde-

pendent of temperature between TL and TH . If �GL/�HL was too l ow , the

effic i ency was l ow; i f i t was too h i gh ( cl ose to 1 ) , the rati o TH/TL was h i gh , mean i ng that the regenerator temperatures wou l d be too h i gh to

be feas i b l e . Tabl e I -2 shows the compounds cons i dered feas i bl e for

therma l regeneration ba sed on these cal cul ati ons . It i s important to

noti ce that a l l these val ues assume that no phase changes occur between

TL and TH . Phase changes do occur i n several cases , mak i ng the s i mpl i ­

f i ed cal cul ati ons i ncorrect . I n these cases , the systems may not be

therma l ly regenerabl e . One exampl e i s the Cd I2 system. The mel t i ng

po i nts of Cd a nd Cd I2 are 594 K and 660 K, respecti vel y . At 700 K , the

rati o �G/�H i s 0 . 63 . Thus , the regenerator temperature at wh i ch the I2 vapor press ure i s 1 atm , as cal cul ated from Eq . I-6, is TH = 700� �

1( l - Q��3·)l= 1890K. At thi s temperature , the Cd I 2 woul d be vapori zed . The

as sumpti on of constant �H i s no l onger appl i cabl e , and the new �H i s

much l ower . The system Cd I 2 i s , the.refore , unl i ke ly to be fea s i bl e for

therma l regenerati on ( see Section I . l . 2 . 1 and Refs . 3 and 1 0 ) .

I n thi s secti on the metal hydri des , ha l i des , oxi des , and cha l ­

cogen i des wh i ch have been studi ed are descri bed . Due to the cons i der­

abl e effort d i rected towards the i nvesti gati on of the system l i th i um

hydri de , a separate summary and concl us ion ar� i ncl uded . . The remai n i ng

systems i n th i s s ecti on have not been thoroughly i nvesti gated . Some

metal hal i des were proposed as thermodynami cal ly feas i bl e but the model s

were very rudimentary and the practi cal experi ence has shown no thermal

decompos i ti o n . The two major d i v i s i ons of th i s section are based o n the

24

t-:) c.n :

TABLE 1 - 2 . COMPOUNDS SELECTED AS POSS IBLE CANDI DATES FOR THERMAL REGENERATION [ 1 2 ]a

.\ Compound mp mp toHfus tiHfus toGo298

C CA C CA ( K) ( K) ( kca1 ) mol e ( kcal ) mol e ( kca1 ) mol e

Li H 459 957 1 . 6 7 . 0 - 1 6 . 7

A1 Br3 932 371 2 . 6 2 . 7 - 1 24

T i C1 2 2000 950 4 . 6 6 . 0 - 96

WC1 6 3650 548 8 . 4 5 . 7 - 74

Co I 2 1 766 790 3 . 7 6 . 0 - 26

ZnC1 2 693 566 1 . 6 5 . 5 -88 . 3

Cd I2 594 660 1 . 5 3 . 6 - 53 . 3

Ga I 3 303 485 1 . 3 3 . 9 - 58

SnC1 2 505 520 1 . 7 3 . 0 - 72 . 2

B i C1 3 544 505 2 . 6 2 . 6 - 76 . 3

As I 3 1 087 4 1 5 6 . 6 2 . 2 -2 1 . 9

TeC1 4 723 497 4 . 3 4 . 5 - 58 . 3

aFor sel ection cri teri a see text.

toW 298

( kcal ) ( mol e

-2 1 . 3

- 1 38

- 1 44

- 96 . 9

- 36

- 99 . 4

- 63 . 3

- 73

- 83 . 6 ·

- 90 . 6

- 35 . 9

- 77 . 4

toS0 298 -ca l )

deg mol e

- 1 5 . 5

-46

- 35

-93

-32

-37 . 4

- 35 . 1

-49

- 32

-47 . 8

-47

-64

L\Go toGo700 toHO at

298 K ( kca 1 ) mol e

0 . 78 - 1 0 . 0

0 . 86 - 1 08 . 3

0 . 84 - 90 . 5

0 . 76 - 33 . 3

0 . 72 - 1 3 . 6

0 . 83 - 74 . 8

0 . 84 - 38 . 6

0 . 80 - 38 . 3

0 . 86 - 59 . 2

0 . 84 - 57 . 4

0 . 61 - 4 . 5

0 . 75 - 64

toW 700

( kca1 ) mol e

- 1 5 . 9

- 1 37 . 9

- 1 1 3 . 6

- 99 . 6

- 33 . 7

- 95 . 7

- 6 1 . 2

- 70 . 5

- 79 . 8

- 90 . 6

- 40 . 3

- 77 . 2

In I I I N -11 II II � =�

toGo toW at

700 K

0 . 63

0 . 79

0 . 80

0 . 33

0 . 40

0 . 78 .-0 . 63

0 . 54

0 . 74

0 . 63

0 . 1 1

0 . 45 I >-3 . .

16 Cj)

S=�I I.I _____________________ T..;...R_-_4 1_6 -� .

number of products formed i n the e l ectrochemi cal reacti on and of re­

generati on steps . Secti on 1 . 1 deal s wi th systems wh i ch present one or

more el ectrochemi cal reacti on products wh i ch use one- step regenerati on

( TRES , Type 1 ) . In Secti on 1 . 2 , the systems i n wh i ch two or more com­

pounds are fo�med i n the gal van i c cel l reaction are exp l a i ned i n deta i l .

In these systems the regeneration of each gal van i c cel l component i s

performed i ndependent ly , and therefore these TRES requ i re a mu1 ti p1 e­

step generati on ( TRES , Type 2 ) . These systems have been ' se 1 ected based

on thermodynami c cal cul ations by Snow [1 4] . Thermochemi cal data for 58

fami l i es of hal i des and oxi des were comp i l ed . The metal s chosen exi sted

i n more than one val ence state . The heats of formati on and trans i ti on ,

trans i ti on temperatures , coeffi ci ents i n heat capac i ty equati ons , and

free energi es of formation were tabul ated . The temperature range over

wh i ch these cal cul at ions were performed was 25-1 200°C . The temperatures

at wh i ch these compounds coul d be generated were a l so gi ven [ 1 4 ] .

Sect ion I . 2 i nc l udes the systems i nvesti aged as a resu l t of these ca l cu­

l ati ons .

1 . 1 S I NGLE OR MULTI PLE ELECTROCHEMICAL REACT ION PRODUCTS AND S I NGLE­

STEP REGENERATION

. Such cel l s a re commonly referred to as thermal ly regenerati ve fuel

cel l s ( TRFC ) or therma l ly regenerati ve gal vani c cel l s ( TRGC ) .

1 . 1 . 1 Metal Hydride Systems : Li th i um Hydri de

1 . 1 . 1 . 1 Summmary

The metal hydri de systems were proposed as thermal ly regenerati ve

26

S=�I I.I ________________________ T_R_-_4_1 6_

-� ��

e l ectrochemi cal systems i � 1 958 as a resu l t of the research performed at

Mi ne Safety Appl i ance Research Corporation ( MSA ) [ 1 6-23J . Thi s research

conti nued through 1 96 1 at MSA and a l so at the TAPCO di vi s i on of Thompson-

RamO-Wool dri dge , Inc . ( TRW ) [24- 31 J . From 1 96 1 to 1 96 7 , the Chemi cal

Engi neeri ng Di v i s i on of Argonne Nati ona l Laboratory (ANL ) [33-46J per­

formed research on the l i th i um hydri de system at a more bas i c l evel .

Paral l el and earl i er per�i nent research efforts were made at the Wri ght

Patterson Ai r Force Base [32J and at Tufts Un i vers i ty [47-49 J .

The l i th i um hydri de system was the fi rst to be envi s i oned as a

practi cal TRES . The el ectrochemi cal cel l reacti ons are :

Li ( � ) + Li + + e

1 / 2 H2 ( g ) + e- + H

anode

cathode

cel l reacti on

Thi s system i s appea l� ng because pure Li H decomposes at 900 °C i nto

eas i ly separabl e l i q u i d l i thi um and gaseous hydrogen . At th i s tempera­

ture the pres sure of the hydrogen gas i s about 760 torr . The gas can be

eas i ly separated from the l i thi um at th i s temperature because i t d i f�

fuses very rap i dly through metal s , e . g . , i ron foi l .

For sc i enti fic and practi cal reasons , however , th i s system poses

probl ems . Fi rst, both l i th i um and the hydri de are suscepti b l e to

reaction wi th normal atmos pheri c components ( oxygen , n i trogen , water ) ,

and therefore an i nert atmosphere i s necessary . Second , a sol vent i s

requ i red that d i s sol ves the L i H but not the l i th i um metal , i s thermo­

dynami cal ly s tabl e to l i th i um meta l at the regenerati on temperature ,

27

- TR-4 16 S=�I I�.�I _____________________ --=...:;c:........=..� -� � � '7

and preferabl y has a l ow vapor pressure at th i s temperature . Mol ten

sa l t systems meet most of these requ i rements and some research was

devoted to the search for su i tabl e el ectrolytes . Th i rd , i t was nec-

essary to devel op l i th i um anodes and hydrogen gas cathodes operabl e i n

the mol ten sa l t medi um. Certa i n l y the most di ffi cu l t probl em i s the gas

e l ectrode .

I n the 1 0 years wh i c h fol l owed the i n i ti a l work on the metal

hydri de regenerati ve systems , emphas i s was fi rst pl aced on the porous

gas e l ectrodes . S i nce the el ectron transfer takes p l ace in the v i c i � i ty

of the el ectrode surface , three-phase contact s i tes must exi st to

ach i eve practi cal current dens i ti es . A frequent probl em wi th th i s type

of el ectrode i s the absorpt i on of l i qu i ds by capi l l ary acti on , wh i ch

causes the el ectrolyte-el ectrode i nterface to retreat wi thi n the el e�­

trode . Thus , fl ood ing of the el ectrode wi th e i ther the el ectrolyte or

gas can occur. Catalysts are often needed for fast heterogeneous re­

act i on rates at the porous el ectrode . The porous el ectrodes tested were

not stabl e i n the presence of the h i gh ly corros i ve mol ten sa l t medi a at

the worki ng temperatures ( a bove 400° C ) .

I t was then recogni zed that th i s system provi ded an unusual oppor­

tun i ty to avo i d the use of porous el ectrodes by us i ng sol i d , th i n fo i l s

of meta l s that are permeabl e to hydrogen gas at these e l evated temper­

atures . Thus , hydrogen can be transferred by i nteratomi c di ffus i on to

the meta l /mol ten sal t e l ectrolyte i nterface. However , the temperature

has to be h i gh for the hydrogen to di ffuse through the metal foi l at an

appreci abl e rate . Th i s d im i n i s hes the effi ci ency of the system .

28

TR-416 S=�I I_I --------------�--------------

The regeneration step i n the presence of the �ol ten sal t medi um

a l so proved l ess than strai ghtforward . Low hydrogen part ia l pressures

( 1 00-200 torr) for most of the mol ten sal ts tested forces the use of

pumps to bri ng the hydrogen pres sure to the 1 atm requ i red for better

performance of the gas el ectrode . Thi s further decreased the cycl e

effi ci ency .

The separation steps were comp l i cated by the presence of L i H i n the

hydrogen exi t 1 i ne . The Li H was formed i n the react ion of L i vapor and

hydrogen at the regenerator temperature . A further compl i cation was the

need to perform the separation under the zero- gravi ty condi ti ons i mposed

by the s-pace power app l i cati on env i s i oned at that ti me . The engi neeri ng

a nd materi al s probl ems due to corros i on were formi dabl e [3 , 4 , 35 J .

1 . 1 . 1 . 2 Deta i l ed Rev i ew

I n 1 958 Shearer and Werner [1 6 J reported the fi rst i nvesti gati on of

schemes for the conti nuous regeneration of a gal van i c system by thermal

energy based on the formation and decompos i ti on of a l kal i and a l kal i ne­

earth metal hydri des i n mol ten hal i de el ectrolytes . Fi gure 1 -2 shows

theoreti cal cel l output potenti a l as a functi on of temperature , ca l cu­

l ated assumi ng that a l l the chemi cal s are in the i r standard states at -

the vari ous temperatures [1 7 J . Thi s i s on l y a gu i de , s i nce the actual

cel l s woul d operate under di fferent condi ti ons . From Fi g . 1 - 2 , the

l ower the fuel cel l operating temperatures are , the h i gher the open­

c i rcu i t vol tages are . On the other hand , i ncreas i ng temperatures wi l l

decrease the po l ari zati on at the hydrogen el ectrode and i ncrease the

sol ubi l i ty of the i on i c hydri de , therefore fac i l i tati ng the conti nuous

29

TR-4 16 S=�I I;.;�I --------------------------

- � �� ,.

> -

° w

0.2 t----t---'"

. H2;Supply Pressure=1 atm

Cell Tefl1perature (0 C)

Figure 1-2. Theoretical Standard Cel l Potentials p.:s a Function of Temperatur� for Various Metal Hydrides [20]

30

S=�I IWI _________________________ -LT.....,R,.=:-.... 4...,...1 6 - � �

regeneration operation [ 1 6-21 ] .

Experi menta 1 work by �1SA [1 6- 23 ] i nc 1 uded Ce 1 1 stud i es of L i , Na ,

K, and Ca e l ectrodes wi th the fol l owi ng mol ten sal t sol uti ons as el ec-

tro lytes : ( a ) Li Cl -Li F ( 570°C ) ; ( b ) Li Cl - KCl ( 357°C ) ; and ( c ) Li Cl ­

NaCl -RbCl -CsCl ( 285°C ) [20 , 23 ] . Other el ectro lyte systems were a l so

i n vesti gated ( KBr-KF- KI ; Li Cl -L i F-L i I ; Li BH4- KBH4 ; L i I -L i Br- KI - KBr )

[2 3 ] . Cel l s were operated most successful ly on a batch bas i s wi th

external supply of hydrogen and l i qui d metal . Conti nuous regenerati on

was attempted us i ng cel l s of the type shown i n Fi g . 1 - 3 [1 7 ] , wh i ch

exh i b i ted severe materi a l s and l ea kage probl ems [3 ] . Better . seal s and

el ectri cal . i nsul ation were obta i ned by us i ng a fl ange system wi th a

col d-sal t seal ( F i g . I -4 ) [23 ] . For al l the cel l confi gurati ons and

el ectrolyte systems used , the current dens i ti es di d not exceed 65 mA/cm2

at very l ow worki ng vol tages (�O . l V ) [ 1 6-23 ] . One excepti on was a

batch operation i n wh i ch a current densi ty of 250 mA/cm2 at 0 . 3 V under

s teady-state cond i t i ons was obta i ned emp l oy ing a l i q u i d l i th i um anode ,

the L i F- L i Cl mol ten eutecti c el ectrolyte at 570°C , and a sta i n l es s steel

cathode ( screen or porous mi crometa 1 1 i c ) . No deta i 1 s about experiment

durat i on or other condi ti ons were gi ven [1 7 , 23 ] .

In general , due to the el evated temperatures emp l oyed , the porous

el ectrodes tested by MSA ( n i cke l , pl ati num , pal l adi um , and carbon )

exh i b i ted vari abl e cata lyti c act i v i ty at the surface of the fri ts ,

fl ood i n g of the fri ts wi th ei ther hydrogen gas or fused el ectro lyte ,

concentration pol ari zati on , and severe corros i on . The di ffus ion mem­

brane ( Pd-Ag ) corroded rapi dl y . For reasonabl e conti nuous regenerati on

3 1

S=�I I_I ______________________ ___ T_R_-_4_16

See Detail A -

Gas Pump

lh-in . Pipe

3-in. Pipe - 1/.!-in. Pipe

Li Electrode

H2 Electrode

Jj�::::=:"::: � 2-in . Tubing

HI.in . . ) � -- -:/ L , - - --

lh in� -� Detail A

2-in . P ipe

1/.!-in . Pipe

1 5 in:

\ \ 1 \ I.J

37lh in .

Figure 1-3. Lithium Hydride Regenerative Cell [20]

32

TR-4 16 S=�I I.I -----------------------------.------� �

Electric I nsulator

Hydrogen Line

Hydrogen Electrode Lith ium Electrode

Cel l Regenerator

Figure 1-4. Lithium Hydride Regenerative Fuel Cell with a Cold-Salt Seal Flange [23]

3 3

S· =!!!tl r.: ______________________ T=R:;:..--"'4=lS:<_ - � ��

operat i on , the hydrogen pres sure must be about 1 atm ; however , the

c l osed-l oop cel l s tested at the regenerati on temperatures of 800- 1 1 00 ° C ,

a n d at the l i th i um hydri de compos i ti ons o f 0 . 35- 0 . 42 w/w % , exh i b i ted

l ow ( 20-300 torr ) parti al hydrogen pressures [ 1 7 , 23J . The hydrogen was

ci rcu l ated by pump i ng argon , whi ch was reci rcul ated from the cel l to the

regenerator . The metal and the e1 e,ctro1yte were ci rcul ated by natural

convecti o n .

The pl aus i bl e coup l i ng of thi s system wi th nucl ear reactors was

env i s i o ned [20 J , a s wel l as the potent ia l appl i cation for thermal energy

convers i on and storage of sol ar energy [21 J .

Al l experimental work carri ed out at MSA was of an expl oratory

nature , search i ng for poss i bl e el ectrolytes and el ectrodes but at the

same time devi ce ori ented . The work carri ed out at TRW [24-31 J , pri nc i pa l l y

on the l i th i um hydri de systems , was a l so devi ce ori ented . Due to the

state of the art of TRES at the t ime , Austi n [3J questioned the val i d i ty

of the stated objecti ve to "deve l op a thermal ly regenerati ve l i th i um

hydri de fuel cel l su i tabl e for use under zero- gravi ty cond i t i ons wi th a

nucl ear heat source " [24J .

Fuscoe , Carl ton , and Laverty performed the i n i ti a l studi es at TRW

[24J wi th l i th i um and hydrogen reactants and the eutecti c el ectrolyte -

mi xture of Li Cl -Li F ( 7 9 : 2 1 w/w % ) at 51 0°C and 1 atm . An open-ci rcu i t

vol tage ( OCV ) of 0 . 5 V was observed , and 6 mA/cm2 coul d be obta i ned wi th

a 50% pol ari zation l oss . The cel l s were constructed of 31 6 sta i n l ess

s te.e l , wi th a sta i n l es s steel be l l -s haped l i thi um anode , and a metal

fo i l d i aphragm cathode ( i ron , 0 . 005 cm ) repl aci ng the porous

34

/."'� TR-4 16 S=�I I_J ----------------------------'==--=-><-

el ectrodes . Th i s repl acement el imi nated the need for a cata lyst and -

mi n imi zed the e l ectro lyte i nterface and separation probl ems , but the

fo i l eiectrodes behaved properly only at el evated temperatures . Maj or

probl ems rel ated to the puri ty of the mel t , el ectrode and cel l materi a l s ,

a nd gases ( a rgon and hydrogen ) were encountered l eadi ng to hi gh spurious

open-ci rcui t vol tages (wh i c h decayed ) and i rreproduci bl e resu l ts . The

maj or impuri ti es were water , oxygen , and ni trogen ; the sta i n l ess steel

contact wel ded to the cathode materi a l corroded . N iobi um fo i l was al so

tested as a cathode materi a l but the n i obi um oxi de , formed at the

s urface or present i n the metal , was not removed under the treatment

gi ven to the el ectrode ; i ron was more successful s i nce the i ron oxi des

formed were reduced by hydrogen gas [24 , 25 , 26J . The regenerator ,

separators , and rad i ati on effects were a l so studi ed [24J .

The equi l i bri um pressure of hydrogen on the Li C l -Li F eutecti c

conta i n i ng 5- 1 0 mol e % of Li H was i nvesti gated at 880 ° C . At th i s tem­

peratur�, appreci abl e vapor pressure of Li , i nsol ubl e i n the mel t , l ed

to the recombi nation of evol ved hydrogen and l i th i um , thereby formi ng

LiH in the exi t l i ne . Under these condi ti ons Li Cl is present in the

vapor phase as wel l . The di ssoci ation of Li H i n the presence of L i Cl

was a l so studi ed [24 J .

An i mproved batch cel l des i gn [25 J i s shown i n Fi g . I - 5 , where h i gh

puri ty i ron was used to conta i n the mel t , el imi nati ng some of the

spurious h i gh i n i ti a l vol tage . A steady OCV of 0 . 38- 0 . 58 V was ob-

ta i ned . The cathode was a O . Ol O-cm , h i gh puri ty , i ron fo i l . As expected ,

the cel l vol tage and current dens i ty decreased wi th decreased hydrogen

pressure . At l ow pressure « 200 torr ) the observed current dens i ti es

35

S=�I I_I ______________________ ...::.T.:::.::R_-4:.:1-=-6

o Tygon Ai r Lock

o Nylon Bushin�

8 Shell

Upper Portion 347 ss

Electrolyte Level

Upper I ron I nsert

Lith ium Pool

Lower I ron I nsert

Bottom Section 347 S5

Head

Water Cool ing

o Feed Tube

{;\ I ron Anode � Bell

60-Meshl_�����\ Screen

12-Mesh Screen

Upper I ron I nsert

Fusion Weld

Detai l of Joint

Thermocouple Wel l

Hydrogen Out Hydrogen I n

Figure- I-5. Batch Lithium-Hydrogen Cell [25]

36 .

5 /. = � TR-4 16 =�I I�I----------------------------------------------------------

agreed wi th cal cul ated val ues ( based on the rate of d i ffus i on of H2 through the i ron foi l ) a nd were found to be proporti onal to the square

root of the press ure of the hydrogen at the el ectrode (assumi ng the

parti a l press ure of hydrogen on the el ectrolyte s i de of the foi l to be

zero ) . The rate of d i ffus i on and the permeabi l i ty of hydrogen on the -4 2 -5 ( 2 ) i ron foi l were l ow : 3 . 75 x 1 0 cm / s and 2 . 6 x 1 0 ml / cm /s , respec-

t i vely , at 550 °C and 760 torr [25 ] .

Other mater ia l s studi ed i nc l uded rhen i um , n i c ke l , z i rcon i um , beryl -

l i um , tanta l um , pal l adi um , n i obi um , vanad i um , rhodi um , ti tan i um , and

thori um . Theoreti cal ca l cu l ations of the permeabi l i ty , ba sed on l i ter-

ature data , i nd i cated that corros i on-res i stant n i o bi um , tantal um , and

vanadi um had h i gher permeab i l i ty to hydrogen than d i d i ron , and that

n i obi um had a 1 000 t imes greater permeabi l i ty . Very prel i mi nary experi ­

ments i nd i cated a permeabi l i ty for n i"obi um only 20 t imes l arger than

that of i ron [25 ] .

A conti nuous thermal regenerati on uni t for a normal gravi ty en-

v i ronment was des i gned , fabri cated , and tested wi th di fferent pump i ng

systems . The degree of regeneration ach i eved was l ower than expected

[25 , 29 , 30 ] .

More extens i ve studies of the n i obi um fo i l cathode were carri ed out

by Carl ton [26 , 27 ] . Thorough puri fi cation schemes were devel oped for

the mel t , gases , and el ectrode materi a l s , as wel l as the manufacture of

an al l -n i obi um cel l of a confi gurati on i nverse to that shown in Fi g . 1 - 5

( anode on the bottom ; gases fed from the top ) . Ni obi um turn i ngs were

used to getter oxygen from the mel t and a 0 . 01 6% oxygen content was

achi eved . A new l i th i um anode on porous n i obi um was devel oped . As a

37

!;::�I 11f1 ____________________________________________________ T _R _- _4 _1 __ 6

resul t of the e l aborate puri fication procedure used duri ng th i s project ,

the s i ngl e cel l tested showed no hi gh spuri ous vo l tage . The cel l poten-\

t i a l rose to a steady 0 . 45-V OCV val ue . Due to the smal l amount of

l i th i um ava i l abl e i n the anode , the run l asted 45 mi n . El ectrode pol ari za-

t i on stud i es were carried out and i nd i cated l ow cathodi c and l ow anod i c

pol ari zat ion . In these stud i es current dens i ti tes as h i gh as 1 500

mA/cm2 at hal f of the OCV were ach i eved on a 0 . 01 2-cm fo i l cathode .

Under l oad ( 2 ohms ) current dens i ti es of 1 25 mA/cm2 were obta i ned at

0 . 1 2 V . The current den s i ty was l ater questi oned by Johnson and Hei nr ich

[33 J , who contended that the permeation woul d not support current den-

s i ti es of that magn i tude .

The separation of the hydrogen gas from the l i qui d meta l /e l ectro­

l yte mi xture was exami ned at TRW [28J us i ng a compact gas/l i qu i d separa­

tor . The l i qui d/gas mi xture was i ntroduced tangenti al ly at h i g h vel oc i ty .

Beneath the cyl i ndri cal section was a coni cal chamber . The gas was

removed from the base of the cone , the heav ier l i qu i d from the s i de , and

the l i ghter l i qui d from the apex .

Probl ems wi th the l i th i um hydri de systems i ncl uded the operation of

the separators in s pace ' [28 , 32 J , e l ectri cal i so l ation of the cel l s , l i fe

of the pumps , and materia l s probl ems [3 , 4 , 35 J .

From 1 961 to 1 967 , the Chemi cal Eng i neer i ng Di vi s i on of Argonne

Nati onal Laboratory [33-46J conti nued the research efforts on the

l i thi um hydri de TRES at a more bas i c l evel than that prev i ous ly descri ­

bed . No attempt to devel op a practi cal conti nuous regeneration cel l was

made , but gu ide l i nes for pract i cal cel l des i gn were suggested [33 J .

The techni ques emp l oyed were s imi lar to previ ous work . An i nert-

38

/.; � TR-4 16 S=�I II.II ----------------------��� -� .

atmosphere dry box was devel oped i n wh i ch puri fied hel i um was ci rcul ated

[34 ] . The performance of i ron cel l s s i mi l ar i n des i gn to that shown i n

Fi g . 1 - 5 was tested i n KC1 -Li Cl (41 : 59 mo l e % , 357 °C ) wi th · an i ron fo i l

d i ffu s i on cathode . Aga i n spurious h i gh vol tages were a l so found . The

pol ari zation of the l i th i um anode was found to be smal l us i ng a l i th i um

reference e l ectrode . In agreement wi th TRW resu l ts [25] , i ron foi l s

were found to support current dens i ti es l ess than 1 00 mA/cm2 [36] .

Cons i derabl e effort was expended by Hei nri ch , Johnson , and Crou­

thamel [36 , 37 , 38] i n measuri ng d i ffus ion rates through i ron , i ron-moly­

bdenum a l l oys , vanadi um , n i o b i um, and tanta l um [36] . Quanti tati ve

permeation studies of Armco i ron and i ron-molybdenum al l oys [37 ] con­

fi rmed that currents i n excess of 1 00 mA/cm2 cannot be obta i ned wi th '

these d i aphragms . However , permeation stud i es on pure vanadi um [38] , a

more reasonabl e materi a l than n i o bi um [4] from an economi ca l and eng in­

eeri ng " po i nt of v i ew, showed that practi cal current den s i t i es coul d be

susta i ned by these ' e l ectrodes . Fi gure 1 -6 compares the current dens i ti es

obtai nabl e wi th Armco i ron and vanadi um . The d i s advantage o f the vanadi um

(or n i obi um) d i aphragm i s that the foi l i s poi soned by oxi de formation ,

res ul ti ng i n a t ime-dependent performance of the el ectrodes .

The permeation stud ies suggest that i n i ti a l l y hydrogen gas i s

adsorbed a t the metal surface i n a moderately fast revers i bl e step .

Th i s i s fol l owed by d i ffus ion of the gas atoms i nto the meta l , the rate

determi n i ng step wh i ch control s the rate of saturati on of the meta l and

a l so the rate at wh i ch the equi l i bri um potent ia l s in the cel l are atta i n­

abl e ( s l ow for vanadi um and n i obi um ; fast for i ron ) . Fi nal l y , the

atomi c hydrogen at the surface accepts an el ectron and mi grates i nto the

sol uti o n .

39

TR-4 16 ' S=�I I;.;� I -------------------------- � ���

7,01;--------...---------, 1004 6,0 861 5,0 7 1 8 4,0 J 574 � I E 3,0 I:: _ 2,0 0 '" :;:; E CO u 0) ....... E '= 1.0 ... E 0,9 � ....... 0:8 c:- 0,7 '0 � 0,6 0) � 0,5 - U CO U 0,4 0:: _ 0,3

o Vanadium D Armco I ron

Temperatu re (0 C) 700 600 � 400 0.1 f---'--,-'----'-r--L-.,-'----'r---'-r--� 0,9 1.0 1 . 1 1,2 1.3 1.4 1.5

1 000 T

430 ,� « 287 §. :>. -'in I:: 144 � 1 1 5 -, I:: 86.1 �

::J 57.4 (J -I:: 0) � 28,7 '5 0' UJ

Figure 1-6. Permeation Isobars of Hydrogen on Vanadium and Armco Iron Foils 0.0254 cm Thick at 1 atm [33]

40

TR-4 16 S=�I I.I -------------:-------------------� �

In add i ti o n , Pl ambeck , El der , and La i ti nen [39J i n vesti gated the

ki neti cs of the el ectrode reaction on an i ron or tungsten fl ag el ectrode

i n the KC1 -Li Cl mel t at 375 °C by us i ng steady- state vol tammetri c and

chronopotenti ometric measurements . Whereas the anodi c oxi dation of

hydri de i ons to hydrogen was found to be a di ffus i on-control l ed , one-

e l ectron process , the reverse process , of i nterest in the l i th i um hy-

dri de cel l , was found to be monoel ectroni c but rather comp l ex and

s l ower , wi th a transfer coeffi c i ent a = 0 . 4 and an exchange current

dens i ty of about 1 mA/cm2 , based on chronopotenti ometri c data at l ow

cathodi c overvo l tages [39 J .

The l i th i um anodes emp l oyed were reta i ned o n s i ntered metal fi ber

s ponge ( SS- 430 ) [33 J . The i nterfaci a l tens i on of the l i qui d l i th i um i n

the sponge ensured that the metal was reta i ned i n pl ace bel ow the sur­

face of the el ectro lyte , as l ong as the dens ity of the el ectrolyte was

not very h i gh [33 , 36J .

The search for s u i tabl e mol ten el ectrolytes l ed to determi nation of

the phase d i agrams of the l i thi um-hydri de/al ka l i- metal-hal i de-based

bi nary [40-42J and ternary [43 , 44 J systems , compl ementi ng the exi sti ng

l i terature up to 1 96 1 [33 J . Two el ectrol ytes were sel ected for practi cal

cel l use because of the i r l ower l i q u i dus temperatures : Li H-Li Cl -L i F ,

predomi nantly a sol i d sol uti on wi th a mi n imum at 456 °C , a nd Li H-L i C l ­

L i I , a ternary eutecti c wi th a mel t i ng poi nt of 330°C [44 J .

From EMF data on the l i th i um hydri de/ l i th i um ha l i de bi nary mel ts i n

the 400-600°C range , the thermodynami c properti es o f l i th i um hydri de

were determi ned by Johnson , He i nri ch , and Crouthamel [45 J . Deri ved

val ues of the standard free energy , enthal py , and entropy of formati on

41

S=�I I_I _________________________ ...... T ..... R'--4......,1 ........ 6

° ° ° at 527°C ar� �Gf = -6 . 74 kcal /mo� e , �Hf = -20 . 9 kca l /mo l e , and �Sf =

- 1 7 . 7 cal l degree mol e [45J . The Nernst equati o n , E = EO - ( RT/nF ) ln­

[ALi H/ ( AL i / a�/ 2 ) J , for the reacti on Li ( l ) + 1 / 2 H2 ( g ) + L i H ( s ) reduces 2

to the EMF be i ng i denti cal to the standard EMF s i nce the acti vi ti es of

the sol i d , l i qu i d , and gas (1 atm ) are uni ty i n saturated sol utions .

The temperature dependence of the standard EMF i s EO = 0 . 908 - ( 7 . 70 x

1 0-4 ) T [45 J .

The extens ion of EMF studi es as a functi on of temperature gave

i nformati on on the regeneration characteri sti cs of the system [33 , 36J .

The H�F dependence on 1 i thi urn hydri de concentrati on i n unsaturated sol u-

ti ons and the val ue for EO l ed to the determi nati on of the l i th i um

hydri de acti v i ty . Wi th the acti v i ty of l i th i um hydri de and the Nernst

equation ( a bove ) i ndi rect data were obta i ned for the hydrogen equ i l i bri um

pressure at each temperature and compos i ti on ( see Fi g . 1 - 7 ) [ 33 , 36 J .

These data are i n a greement wi th di rect hydrogen part ia l pressure

measurements carr ied out by Fuscoe , Carl ton , and Laverty [24J at TRW .

Stud i es of the hydrogen di ffus i on through the d i aphragm showed that

the quanti ty of hydrogen di ffus i ng through the metal ' i s proporti onal to

the d i fference i n the square roots of the hydrogen pressures on each

s i de [33 , 36 J , i n agreement wi th earl i er studi es at TRW [24 J . The

l arger the pressure at the cel l , the h i gher i s the output vol tage [33 , 36J .

Therefore , i f the hydrogen pres sures at the regenerator are l ow ( see

Fi g . 1 - 7 ) , an i ncrease to 1 atm or more by means of a pump is needed to

improve the performance of the fuel cel l and the effi ci ency of the

sys tem . Johnson et a l . [33 , 36 J poi nted out that , in pri nc i p l e , i f the

ternary eutecti c ( Li H-L i Cl -L i I ) were used as the worki ng el ectrolyte , at

42

- TR-4 16 S=�I I;.-�I --------------------------� "' :�

200 1 00 80 60 40

- 20 ... ... 0 � � 1 0 ::J 8.0 VI 6.0 VI Q) ... a.. 4.0 c: Q) Cl 2.0 0 ... "U >-I 1 .0 0.8 0.6

0.4 0.2 0. 1 0.9 1 .0 1 . 1 1 000 (K-1 )_ ,.-

Figure 1-7. Calculated Equil ibrium Hydrogen Pressure for LiH-Li CI Mixtures [36]

43

S=�I I.I ______________________ -=.T=R--'-4=.=1-=-6 -� �

or near saturati on wi th respect to Li H at 330°C ( or above ) , the re-

generati on pressure coul d be s i gn i fi cantly i ncreased by sendi ng to the

regenerator a mi xture ri cher i n Li H ( 90- 95 mol e % L i H ) . The pump i ng

requ i rements woul d be s ubstanti a l l y smal l er , and , therefore , the ef­

fi ci ency of the system h i gher. Johnson et a l . [33 , 35 J a l so stud i ed the

vo l tage l os ses across the vanadi um di aphragms .

At Argonrie Nati onal Lab . a practical batch cel l wa s tested for 540

hours at 535° C , g i v i ng current dens i ti es i n excess of 200 mA/cm2 at

hal f the OCV ( 0 . 3 V ) [36J . Th i s cel l used a nontreated 0 . 025-cm-th i c k

vanadi um d i aphragm and Li H-Li Cl -Li F mol ten sal t . Fi gure I - 8 shows the

current-vo 1 tage characteri sti cs of th i s cel l [36J . The OCV fol l owed the

Nernst equati on . A p l ot of p�/2 as a function of the OCV i s l i near 2

wi th the s l ope correspond i ng to a revers ibl e H2/H- coup l e at the van-

adi um el ectrode .

Prel i mi nary practi cal cel l s were a l so tested wi th the eutecti c L i H­

L i Cl -Li I , wh i ch has a l ower melt i ng poi nt than Li H-Li Cl -L i F . Larger

Carnot cycl e effi ci enc i es , due to the l ower operati ng temperatures , and

pos s i bly more effecti ve regeneration can be expected ( see above ) .

Fi gure I - 9 shows two experiments wi th th i s mel t u s i ng 0 . 005- cm-th i c k

vanadi um d i aphragms . These experiments di ffer i n i nternal cel l resi s ­

tance by a factor of four . Th i s di fference was attri buted to the di ffer­

ent pretreatments of the d i aphragms , yi el d i ng materi a l s wi th di fferent

oxi de content . Th i s gave vari abl e vol tage l osses probably assoc i ated

wi th the overvo l tage at the hydrogen el ectrode .

Hesson and Sh imota ke [46J have di scussed i n deta i l the thermo­

dynami cs and therma l effi c i enc i es of the l i thi um hydri de systems .

44

S=�I I_I _----'-__________________________ T_R_-_4_1_6

Q) OJ co -'0 > 03 ()

0.8.------"-'--------��------____,

0.3

0.2

0. 1 o!-----=L"---:l,,...----,J..,...------L---L--L-.....L----L--JI S-O -2...L0-O -.-J220 Current Density (mA/cm2)

Figure 1-8. Voltage - Curr�nt Curve for a Lithium Hydride Cell with a O.025-cm Vanadium Diaphragm at 5250 C [36]

600

:;- 500 E -

Q) 400 OJ co - 300 '0 > 03 200 ()

1 00

Current Density (mA/cm2)

Figure 1-9. Voltage - Current Curve for a Lithium Hydride Cell with a O�005-cm Vanadium Diaphragm at 4250 C [36]

45

- TR-4 16 S=�I I_I -----------------�-----------

I . 1 . 1 . 3 Concl us i ons

In a l most 1 0 years ( 1 958- 1 96 7 ) of research on l i th i um hydri de

thermal ly regenerati ve systems , no practi cal cel l was devel oped under

defi ned cond i t i ons i n a conti nuous regenerati on mode . Cel l s for batch

operati on showed the poss i bi l i ty of obta i n i ng practi cal current dens i ti es

(�200 mA/cm2 ) at rel ati vel y l ow worki ng vol tages (� . 3 V ) . Mos t of the

i n i ti a l effort was l argely empi ri cal , and even the more bas i c work

l acked reproduci bi l i ty because of the d i ffi cu l ty i n obta i n i ng th i n metal

di aphragms of oxi de- free vanadi um .

A practi cal l i th i um hydri de cel l shoul d operate at the l owes t pos­

s i b l e temperature to ta ke advantage of the h i gher vol tage and h i gher

Carnot cycl e effi c i ency thus obta i ned . There are sti l l two key i ssues

unresol ved : the gas e l ectrode and engi.neeri ng probl ems .

Wi th respect to the gas el ectrode there are agai n two approaches to

be i nvesti gated . I f th i n fi l ms are used as cathode materi a l s , modern

surface s pectroscop ic techn iques can faci l i tate the defi n i t i on of oxi de

content and i mpuri ti es to ensure reproduc i bi l i ty of the foi l s . More

reproduci bl e surfaces can be obta i ned wi th more adequate pretreatment of

the fo i l s . Mel ts operati ng at l ower temperatures perm it renewed cons i der­

ati on of the abandoned porous el ectrodes as cathodes , parti cul arly i n

v i ew - of recent technol ogical advances .

The regeneration step i s not s impl e due to the l ow part ia l pressure

of hydrogen obta i ned i n most cases and the s l ow rate of hydrogen for­

mat ion . Perhaps the use of the ternary eutecti c Li H-Li Cl -L i I woul d

effecti vely ra i se the hydrogen equi l i bri um pressure i n practi ce .

However , even for the most promi s i ng mol ten sal t systems ( L i H-Li Cl ­

L i F and Li H-L i Cl -Li I ) the engi neeri ng probl ems conti nue - ci rcul ati on

46

- TR-416 S=�I I.I ------------------------------� �

and separati on of the l i q u i d/ so l i d mi xtures ( e . g ; , sol i d L i H i n Li H­

L i Cl -Li I ) ; the presence of L i H , Li Cl , and Li I i n the hydrogen exi t l i ne

( th i s probl em can be mi n i mi zed by taki ng advantage of the fast di ffus i on

of hydrogen through meta l s at these regenerator temperatures ) ; and

corros i o n , pri nci pal ly by L i H .

1 . 1 . 2 Hal i de-Conta i n i ng Systems

In 1 959 , when Werner and Shearer [ 1 9J proposed the i r patent on

metal hydri de thermal ly regenerati ve systems , they al so s uggested metal

hal i des as potenti a l l y i nterest i ng TRES . In aqueous sol ut ion , a gal ­

van i c cel l wi th a cuprous bromi de paste anode and a bromi ne gas el ec-

trode formed CuBr2 el ectrochemi cal l y , wi th an OCV of 0 . 66 V . They

proposed that the regenerati on cou l d be achi eved by heating and dri v i ng

off water and bromi ne and return i ng the cuprous bromi de to the cel l

anode . Mol ten sal t medi a were al so suggested but not tested , and these

cel l s were not pursued further.

Most of the hal i de-conta i n i ng systems i nvesti gated and descri bed i n

t h i s secti on d i d not operate successfu l l y . Some hal i de-conta i n i ng

systems wh i ch used two di fferent ha l i de compounds as anode and cathode

wi l l be descri bed i n Section 1 . 2 . 1 .

1 . 1 . 2 . 1 Metal Iod i des

Systems based on Sn I 2 , Pb I 2 , and Cd I 2 were proposed by Lockheed

Ai rcraft Corp . [50 , 51 J as potenti al TRES . The i n i ti a l work was devoted

to test i ng the performance of cel l s of metal I mo l ten i od i de l i od i ne gas .

The res ul ts are s ummari zed i n Tabl e 1 - 3 . When regeneration was attempted

at temperatures up to 1 000° C , no decompos i ti on was observed . Mi s l ead i n g

k i neti c arguments [3 J ( cf . Ref. 9 ,. 1 2 ) were used to exp l a i n the l ack of

47

S=�I I_I _________________________ ...... T.;;R=--...... 4"'-><...16

decompos i ti on , but l ater it was rea l i zed that these i od i des are thermo­

dynami cal ly stabl e and therefore unl i ke ly to be thermal ly decomposed at

th i s temperature [50 , 51 , 9J . The B i I 3 cycl e was exami ned in the thermal

regenerati on mode by Aeroj et Genera l Corp . [ 52 J , but l ater a thermo­

gal van i c type of operation ( see Section I I I ) was preferred . Later , the

i od i de systems were empl oyed more successfu l l y under a di fferent mode of

operati on , the coupl i ng of el ectri cal and thermal regenerati on [5 1 J ( see

Secti ons IV and V ) . - - -

[50Ja 'Tabl e 1 - 3 . SUMMARY O F GALVAN IC CELL PERFORMANCE

- ---=-- -:..-....::.... Cel l Performance

Temperature OCV { V ) I 2 ( V ) System ( O C ) cal cd . expt . (mA/cm ) Comments

Sn! I2 not g iven 0 . 8 0 . 4 el ectrolyte sol i d ifi ed

Pb/ 1 2 450 0 . 63 55 0 . 3 ohmi c pol ari zati on on ly

V70 0 . 87 55 0 . 35 ohmi c pol ari zati"on Cd/ 1 2 on ly

470 0 . 83 55 0 . 65

aThe cel l con s i sted of a porcel a i n beaker conta i n i ng mol ten sal t fl oati n g o n a mol ten metal anode and a porous carbon el ectrode as the i od i ne vapor cathode ( 500 torr above atmospheri c pressure ).

1 . 1 . 2 . 2 Mercury Hal i des and Systems Regenerated by Thermal Di sproporti onati on Reacti ons

Another system proposed by Lockheed [50J was the cel l : Hg I HgBr2 : KBr

( 50 : 50 mol e % ) I Br2 at 260°C . A ca l cu l ated OCV o f 0 . 61 V was expected ,

but under these condi ti ons sol i d Hg2Br2 was formed due to the react i on

Hg ( � ) + HgBr2 ( � ) + Hg2Br2 ( s ) . I ncreased temperatures s ubl i med the

Hg2Br2 ·

48

- TR-4 16 S=�I I_I ----------------------------

An anal ogous system , us i ng the d i sproporti onation reacti on of mer-

cury( I ) chl ori de as the thermal regenerati on step , was proposed by the

I I I i no i s Inst i tute of Technol ogy Research I.nsti tute ( I ITRI ) [53J as a

resul t of thermodynami c cal cu l at ions [ 1 4J :

Hg + HgC1 2 gal van i c cel l �-�����----�> ( . Hg2C1 2 thermal regenerat, on

Pre l i mi nary gal van i c cel l measurements wi th a cel l cathode contai n i ng

HgC1 2 : A1 C1 3 ( 0 . 33 : 0 . 67 mol e fracti on ) and an anode of pure mercury gave

an OCV of 0 . 73 V at 205 ° C , and 0 . 22 V wi th a 1 00-Q l oad [53 J . Prel im in-

a ry studi es of the regeneration encountered probl ems i n separati ng

mercury vapor from gaseous HgC1 2 [54J . D i ffus i on of mercury vapor

through a gol d foi l was attempted as wel l as the preferred , but not very

effecti ve , separation by preferent ia l absorption of HgC1 2 i n NaCl - KCl

mel ts [54 J .

A s i mi l ar system was recently proposed by I ITRI [55J : -+

.' � _ J<: . . VC1J + VC1 4 + 2VC1 3

The proj ected OCV i s 1 . 5 V at 1 75 ° C ; the regenerati on i s to be per­

formed at 700°C .

I . l . 2 . 3 Phosphorus Pentachl ori de

The fol l owi ng system was proposed [56 , 57 J as therma l ly regenerati ve

i n ' organ i c sol vents of h i gh d ie l ectri c constant :

. ' �- gal van i c cel l :> �. PC1 3 + C1 2 <th 1 . . . �<.: _ erma 'C,egenera t, on

The conducti vi ty of the sol utions was attri buted to the PC1 � and PC1 6 i ons [57 J . Several sol vents were tested : acetoni tri l e , d imethyl forma­

mide , di methyl su l foxi de ( v i ol ent reacti on i n contact wi th the phosphorous

49

'" TR-4 1G S=�I I.I --------------------------=:....=..:.......::...:..=... -� �

compounds ) , methyl thi ocyanate ( reacted wi th PC1 5 ) , and n i tromethane

( decompos i t ion reacti ons occurred wi th time ) . The support i ng el ec-

trolyte was L i N03 [57 , 58 J . The EO for the cel l was 0 . 28 V at 1 5 ° C . A

cel l was bui l t but no s i gn i fi cant resul ts were reported .

Some e l ectrode ki net i cs stud i es were carri ed out by chronopoten-

ti ometry of the chl ori des of P, S b , and W i n di methyl formami de at 25 °C

on pl ati num worki ng e l ectrodes [59 , 60 J .

1 . 1 . 2 . 4 Anti mony Pentachl ori.de

Th i s i s one of the ha l i de systems tested by I ITRI [61 , 62 J from

1 960- 1 96 7 , as a resu l t of theoreti cal thermodynami c cal cul ati ons [ 1 4 J i n

a devi ce-ori ented project . Th i s compound wa s s e l ected because of i ts

easy di ssoc i ati on i nto l i qu i d anti mony tri chl ori de and gaseous chl ori ne

at rel atively l ow temperatures . .

The drawback of th i s system ; s the expected poor i o n i c conducti v i ty

of the antimony ch l ori des , wh i ch can be improved s l i ghtly to about 1 0-3

ohm- l cm- l by addi ti on of arsen i um trichl ori de and mi nor amounts of

a l umi num ch l ori de . Cel l s constructed wi th SbC1 3 anodes and chl ori ne

cathodes conta i n i ng SbC1 5 gave about 0 . 3 V and th� mi xture SbC1 3 : AsC' 3 ( l : l ) gave about 0 . 4-0 . 5 V , but the cel l res i stance was sti l l very h i gh

[63 J . A sol i d el ectrol yte ( PbC1 2 doped wi th KC1 ) was used i n a smal l

cel l wi th some success . The conduct i v i ti es were stud i ed , . as a functi on

of temperature , of severa l other sol i d el ectrolytes based on PbC1 2 ,

PbF2 , B i C1 3 , and i on exchange polymers ( s i l i cates , sodi um phosphotungstates ,

and sod i um polyphosphates ) , but none exhi bi ted behav ior wh i c h woul d be

, s u i tabl e for th i s type of cel l [63 J .

The di ssoc i ati on reaction SbC1 5 ( � ) + SbC1 3 ( � ) + C1 2 ( g ) was studi ed

50

S=�I I.I ____ -'--_____________ ___ __ --=T:..,:R=---"..4 =.,:lSO--� ��

i n detai l . The regeneration was demonstrated sati sfactori ly (80-90%

d i ssoci ati �n of SbC1 5 ) over a wi de range of temperatures ( 250-350 °C ) and

pressures ( 1 -25 atm) , produci ng l i qui d SbC1 3 and gaseous chl ori ne [61 ,

62J .

The work wi th mi xtures of antimony , arsen i c , and a l umi num chl ori des

was l argely emp i r ical wi th respect to the spec i es present in sol ution . Pl ots

of EMF as a functi on of the rati o 9f SbC1 5 to SbC1 3 i n AsC1 3 sol vent

s how a vari ation of about 300 mV when the rati o i s changed by a factor

of 1 0 at l ow SbC1 5 content ( 70°C ) . From rati os of 1 : 1 to 1 00 : 1 , the EMF

i s essenti a l ly i ndependent of the compos i ton and equal to �0 . 65 V . It

was proposed that mi xtures of Sb and As ch l ori des can be oxi d i zed by

free ch l ori ne-formi ng spec i es conta i n i ng As ( V ) whi ch coul d part i c i pate

i n the el ectrode reacti ons . McCu l ly and coworkers [61 , 62J suggested the

fol l owi ng total cel l reacti on :

� �bAsCl l O + SbC1 3 + 2SbC1 5 + AsC1 3 �,�

A un i t of 500-W formal power was bui l t wi th 1 0 cel l s connected i n

series wi th �0 . 3-cm spaci ng between anode and cathode compartments . The

compartments 'were separa-ted - by the ' sol i d el ectro lyte (woven gl ass

c l oth impregnated wi th doped PbC1 2 ) . The anode and cathode compartment

compos i ti ons were SbC1 3 : AsC1 3 ( 2 : 1 mol e rati o ) and SbC1 5 : AsC1 3 ( 4 : 1

mo l e rati o ) , respecti vely , both conta i n i ng 4 w/w % A1 C1 3 . The cel l had

a heat exchanger between the batteri es and the regenerator un i t . The

regenerator worked successful ly but the cel l performance was very poor .

The system coul d be charged to an output potent ia l of 1 . 4 V , whi ch

decreased very rap i dly wi thout appreci abl e l oad . It was veri fi ed that ,

under pressure , cracks i n the el ectrolyte a l l owed the mi xi ng of the

5 1

S=�I I.I __________________________ TuR .... -::!l4 ...... 1..LL...6 - � ���

anode and cathode compartments [61 , 62 J .

The use of sol i d e l ectrolytes was di scouraged for th i s system . It

was proposed that research s houl d be conti nued i n the di recti on of

i mprovi ng the conducti v i ty of the antimony chl ori de sol uti ons [62 J ,

s i nce the regenerati on i n th i s system can be accomp l i shed eas i l y , wi th

mi nor corros i on probl ems [63J .

1 . 1 . 2 . 5 Hydrogen Hal i des and Other Ha l i des

Ri ghtmi re and Ca l l ahan [64 J descri bed a hydrogen i od i de therma l ly

regenerati ve system i n 1 963 . The fuel cel l operated at 1 20 °C wi th two

porous p l ati n i zed e l ectrodes sandwi ch i ng the el ectro lyte , an aqueous

sol ution of H I . The fuel cel l reactions were hydrogen gas oxi dat ion at ·

the anode and i od i ne reduction at the cathode . Regenerati on was per�

formed by fl owi ng the el ectro lyte through a heat exchanger for pre-

heati ng and then cata lyti cal ly decomposi ng it i n a reactor at 1 000°C

( us i ng a Pt cata lyst s upported on a l umi na or a natural cl ay base ) . The

tota l pressure of the c l osed system was 6- 7 atm . The equ i l i br i um

mi xture 2H I ( g ) �20 ( g » H2 ( g ) + I2 ( g ) was then coo l ed . Iodi ne di ssol ved

i n H I as 13 ( condensed) and the sol ution [H20 , H I , I 2 ( 13 ) J was fed to

the cathode . The hydrogen gas separated from the condensed l i q u i d and

was fed to the anode compartment of the fue l cel l . When 43% H20 and 57%

HI were used as el ectro lyte at 120°C , the OCV was 0 . 5 V . The authors

c l a im that a power output of 0 . 03-0 . 08 W/cm3 at 75% therma l effi c i ency

( the Carnot effi c i ency was 91 % ) can be obta i ned i n th i s system wi th

curr�nt dens i ti es a s h i gh as 1 00 mA/cm2 . The authors suggested that

hydrogen bromi de woul d a l so be su i tabl e for thermal regenerati o n .

Proposa l s by Rowl ette and others [3 , 1 3 J to use the iodi ne mono-

52

- TR-4 16 S=�I 1.1 ------------------------------� ��

bromi de system and other meta l hal i de systems are based on thermodynami c

and k i neti c data . The mel ti ng poi nt of IBr i s 42 °C and i ts bo i l i ng

po i nt i s 1 1 6° C . The conducti v i ty i s l ow but can be i ncreased by add i ti on

of KBr . At 300°C , IBr i s 20% decomposed , whereas at room temperature

ca . 8% i s decomposed . The system thus seems feas i bl e for thermal regener-

ation but no further experimental studi es were carri ed out .

1 . 1 . 3 Oxi de-Conta i n i ng Systems and Other Systems

1 . 1 . 3 . 1 Sul fur di oxi de- tri oxi de

In 1 96 1 - 1 962 , Kumm [65] i nvesti gated the system i n wh i ch the regen-

eration reacti on i s

;- --..

and the cel l S02 l el ectrol yte l 02 ' Thermodynami c studi es of the regenera­

t i on steps i nd i cated that at about 1 000°C the su l fur tri oxi de i s l arge ly

decomposed . Kumm [65 ] presented equi l i bri um data over a wi de range of

temperatures . To separate the two gases formed , two methods were

s uggested : absorption of S02 by water , or condensati on of S03 to act as

a scrubber for S02 ' The fi rst method i s not feas i b l e due to the h i gh

vapor pressures of water at temperatures at wh i ch S02 can be di sti l l ed -

from water .

and

The gal van i c cel l reactions

2- -S02 + 2S04 + H20 + 2HS04 + S03 + 2e

1 /2 02 + 2HS04 + 2e- + H20 + 2S0�-

were expected i n the two medi a tested : mol ten eutect ic Li HS04- KHS04 at

53

- TR-4 16 S=�I I_I ------,--------------------------

�1 50°C , and dimethyl su l fate saturated wi th L i H504-KH504 . For the

fi rst medi um at �1 80° C an OCV of �O . l V was found and for the second at

�95 ° C an OCV of �0 . 1 7 V was obta i ned . Current dens i t i es were very l ow

and i t was concl uded that a thermal ly regenerati ve el ectrochemi cal

system based on 5°2/5°3 was not feas i bl e .

�1ore recently , Wentworth [66 J has been i nvesti gati ng the 5°2/5°3 system i n an a l kal i ne medi um. An OCV of about 1 V i s expected . I n h i s

pri or stud i es the mol ten sal t NH4H504 has been shown to undergo decom­

pos i ti on i n two steps whi ch can regenerate 502 ' In the fi rst step at

about 400 ° C , NH4H504 is decomposed i nto NH3 ( g ) , H20 ( g ) , and 503 ( g ) i n a

reaction wh i ch stores sol ar thermal energy . After the three gases have

been separated the 503 can be decomposed i nto 502 ( g ) and 02 at a h i gher

temperature ( 950° C ) i n a cata lyzed reacti o n . 5tudi es of the system are

underway . 50 far , l ow current dens i ti es and vol tages have been obta i ned

[66 J .

I . l . 3 . 2 Metal Oxi des

From thermodynami c cons i derations but wi thout experimenta l data

Lyons [67 J proposed metal oxi de fuel cel l s i n wh i ch l ower-val ent metal

oxi des are oxi d i zed i n a l ka l i ne fuel cel l s to the h i gher-va l ent metal

oxi des . The l atter are reconverted to the l ower-val ent oxi des by heat

or chemi cal reducti on . The oxi des proposed i n th i s patent are of

copper , coba l t , manganese , l ead , and i ron .

Aga i n from thermodynami c cons i derations , McKenz i e and Howe [68J

proposed cel l s wi th oxi des or oxyan i ons of rare earths , and group V1 A or

54

- TR-4 16 S=�I I.I ---------------------------� ��

1 . 1 . 3 . 3 Other Compounds and the Sn I Sn ( I I ) , Cr ( I I I ) , Cr ( I I ) I c System

��erner and Shearer [1 9J tested the system of an i ron ( I I ) s.u l fi de

anode i n mol ten sodi um polysul fi de ( 1 1 0 ° C ) and a su l fur cathode . A

0 . 25-V OCV was obta� ned and i ron di su l fi de produced . S i nce the standard

free energy of formation of i ron di su l f ide from ferrous sul fi de and

su l fur is zero at 700 ° C , the authors suggested that at th i s temperature

the cathode and anode regenerati on coul d be accompl i shed . They al so

suggested l i th i um n i tr ide formati on as a poss i bl e TRES , though of very

l ow cel l potenti a l (�O . l V ) .

1 . 1 . 3 . 4 The Sn I Sn ( I I ) , Cr ( I I I ) , Cr ( I I ) I C System

I n 1 886 , Case [69J reported what appears to be the fi rst therma l ly

regenerati ve system descri bed i n the l i terature . It produced el ectri cal

energy peri od i cal ly i f a per iodi cal temperature change was appl i ed to

the sys tem . Thu s , at a temperature of 90- 1 00°C the system del i vered

e l ectri cal energy and at 1 5- 20°C i t was spontaneous ly chemi ca l ly regen­

erated . Case l s system i s based on the gal van i c cel l composed of a ti n

a node and an i nert cathode ( e . g . , porous graph i te ) revers i bl e to the

sol ubl e s pec i es Cr( I I ) and Cr( I I I ) . At 90- 1 00 °C the gal van i c cel l

operates and generates el ectri c i ty . When the reactants are exhausted ,

the regenerati on i s performed by d i sconnect i ng the el ectri cal ci rcu i t

a nd l etti ng Cr( I I ) i ons reduce chemi cal ly 4he Sn ( I I ) i ons to S n metal ,

wh i ch depos i ts on the anode . I n 1 895 , Ski nner improved the anode by

repl aci ng pure ti n wi th t i n amal gam [70 ; 1 28 , p . 348 J .

A conti nuous regeneration operati on mi ght be obta i ned by fl owi ng

the e l ectro lyte enri ched wi th Sn ( I I ) and Cr( I I ) ions to a l ower-tem­

perature compartment , and by return i ng ' the prec i pi tated ti n to the

55

S=�I· IWI _________________________ ---JTwR.L:-::!:I4u.l.u-6 -� �

anode and the Cr( I I ) -enri ched sol ution to' the cel l ( cf . wi th Section

I L 1 . 3 ) .

Al most a century l ater , Case l s system has been rei nvesti gated by

Vedel , Soubeyrand , and LeQuan [7 1 J . The temperature dependence of the

EMF of the ce.l 1 Sn I Sn ( I I ) I I Cr ( I I I ) , Cr ( I I ) I C was measured at vari o us

e l ectrolyte compos i ti ons ( HCl : CaC1 2 ) and changed s i gn between 25 °C and

95° C . These authors obtai ned a temperature coeffi c i ent for the cel l of

1 . 42 mV/ o C . By choos i n g the el ectrolyte composi ti on , the i nvers i on of

the s i gn can be made to occur at any temperature wi thi n the gi ven range .

Low vol tages , -60 to +60 mV , and l ow currents , �3 rnA , were obta i ned .

1 . 1 . 4 Summary and Di scuss i on of TRES Type 1

Tabl e S-2 represents a s ummary of the thermal l y regenerati ve el ectro­

chemi cal systems i nvesti gated or proposed i n the l i terature covered by

Sec . I . 1 .

The maj ori ty of the systems reported i n Sec . 1 . 1 uti l i zed mol ten

sal t e l ectrolyte systems and h i gh regenerati on temperatures ( 500 -

1 000° C ) . Several hal i de systems were s hown not to decompose apprec i ab ly

wi th i n the temperature range i nvesti gated . I n some cases s l ow ki net i cs

was res pons i bl e for the l ow decomposi ti on y iel ds . Cata lyt i c decompos i t ion

was attempted on ly i n the H I system. Wi th the use of s u i tabl e cata lysts ,

other systems may deserve renewed cons i derati o n . Very l i ttl e experimenta l

work has been done i n oxi de systems .

Aqueous and nonaqueous systems have recei ved far l ess attenti on

than mo l ten sa l t med i a- -pri nc i p a l ly because of the l ower operati ng

temperatures , wh i ch woul d not be su i tabl e for coupl i ng wi th nucl ear heat

sources but wh i ch certa i n ly wou l d be adequate for sol ar appl i cati ons . -

56

S=�I I.I ________________________ T ..... R""--....:I4�16"'_ -� �.

Systems operati ng at l ower temperatures [anal ogous to Sn I Sn ( I I ) ,

Cr( I I I ) , Cr ( I I ) I CJ , d i scovereci by Case i n the 1 9th century , have not been

thoroughly i nvesti gated . Th i s i s an area i n wh i ch exi st i ng thermodynami c

data or new experi mental data may i ndi cate s�stems- of better performance

than Case ' s system .

1 . 2 MULT I PLE ELECTROCHEM ICAL REACT ION PRODUCTS AND MULTI PLE-STEP REGENERATION

A more compl ex gal van i c cel l was devi sed i n wh i ch there are two

el ectrochemi cal products ( See Type 2 , i n the Introducti on ) . The anode

and cathode are composed of d i fferent compounds , e . g . , meta l hal i des or

oxi des . The anode and cathode are regenerated separatel y , general l y at

d i fferent temperatures [72 J . The general scheme of such a system for

metal hal i des i s :

Gal van i c Cel l Reactions

MX + mX­n

M I X + me n l +m

+ MXn+m + me

+ M ' X + mX n l

MX + M I X Ii> n n l +m

Regeneration Reactions

anode

cathode

gal van i c cel l

Anode regeneration : MX heat > MX + m/2 X ( g ) n+m T 2>T 1 n 2

After separation of MXn and X2 ' MXn i s returned to the anode and

X2 is a l l owed to react wi th M ' Xn " thus regenerati ng the cathode .

These systems are c l early much more compl ex than those descri bed i n

57

!;=�l rlfr ________________________________________________ T�R�-�4�1�6

Section 1 . 1 . A general probl em ari ses when the anode regeneration step

produces two products in the same phys i cal state . For i nstance , i f

SnC1 4 o r TeC1 4 are the resul t of the anodi c processes , at the thermal

decompos i ti on temperatures the two products SnC1 2 or TeC1 2 and C1 2 are

i n the gaseous state , and the di ffi cu l t separati on consti tutes a very

severe l i mi tati on to the practi cal appl i cati on of th i s type of system

[61 , 62 J . I f SbC1 5 or CuC1 2 are formed as a resu l t of the anodic reac­

t i on , the regeneration yi el ds SbCl l( £ ) or CuCl ( £ ) and C1 2 ( g ) , wh i ch can

be separated by a rel ati vely s imp l e process ( see Secti on I . l . 2 . 4 on

SbC1 5 ) . Sel f-di scharge processes pose add i ti ona l d i ffi cul ties to the

uti l i zati on of cel l s of thi s type .

Th i s approach to therma l ly regenerati ve gal van i c cel l s was proposed

and researched from 1 960 to 1 969 at I ITRI ( formerly Armour Research

Foundat ion ) by McCul l y , Rymarz , and Snow, among others . Thermochemi cal

and thermodynami c cal cul ati ons di scl osed several chem ica l compounds as

potenti a l ly s u i tabl e for thermal regeneration and gal van i c cel l opera­

t ion [1 4 J . Secti ons 1 . 1 . 2 . 2 and 1 . 1 . 2 . 4 descri be the SbC1 5 and HgC1 2 systems , wh i ch composed part of the research effort at I ITRI on s impl er

hal i de systems . Sections 1 . 2 . 1 and 1 . 2 . 2 deal wi th the more compl ex

cel l concept .

1 . 2 . 1 Metal Hal i des

The reactions of meta l hal i des studi ed at I ITRI duri ng the 1 960-

1 967 peri od are s ummari zed i n Tabl e 1-4 , i n wh i ch sel ected gal van i c cel l

resu l ts are assembl ed . The proposed d i agram for th i s type of system ,

i nc l udi ng the cel l , the regenerators , and the heat exchangers , i s shown

i n F i g . 1 - 1 0 .

58

S-�I /..=,. - II II - � TR-4 16

Tabl e 1 -4 . CELL VOLTAGES . OBTAINED IN THE REACTIONS OF METAL HAL I DES [52 , 73 J

Reactiona Composi t i on Temperature Anode Cathode ( OC )

SnCl 2+SbCl 5�SnCl 4+SbCl 3 SnCl 2 : A1 Cl 3 1 50 50 w/w .%

SnBr2+SbBr5�SnBr4+SbBr3 SnBr2 SbBr2+Br2 c not g i ven

Mo l e Fraction i n A1 Cl 3 SnCl 2+2CuCl 2�SnCl 4+2CuCl 0 . 23 0 . 25 205

0 . 32 0 . 25 205 0 . 41 0 . 25 205

CuCl+CuCl 2�CuCl 2+CuCl 0 . 31 0 . 25 253

TeCl 2+2CuCl 2�TeCl 4+2CuCl 1 0 . 25 205 0 . 40e 0 . 2� 205 0 . 06 0 . 1 1 95

HgCl+CuCl 2+HgCl 2+CuCl 0 . 40 0 . 25 200 0 . 40 0 . 25 200

0 . 40 0 . 25 230 0 . 40 0 . 25 21 0

al st reactant : anode; 2nd reactant : cathode .

Open-Ci r­cuit Vol t­age ( V )

0 . 47

0 . 1 5

0 . 70 0 . 85 0 . 97

0 . 25

0 . 92 0 . 58 0 . 46

0 . 75

0 . 71 0 . 66

Vol tage Under Comments l OO-ohm Load ( V )

0 . 29d 0 . 30 0 . 32

0 . 1 1

0 . 51 f 0 . 20

0 . 26

0 . 29

Stable performance for over 2 weeks

Expected OCV of 0 . 54 V ; l ow BrZ concentration re-, spon s i bl e for l ow OCV

Pl atinum el ectrodes

Pl ati num el ectrodes

Pl ati num el ectrodes

P l atinum el ectrode Graphite el ectrode­Cathode d i s i ntegrated Tungsten el ectrodes Tantal um el ectrotles

bEl ectrolyt�: A1 Cl 3 : KCl eutect i c ; l oad : 1 0 , 000 ohms; current den s i ty : 50 mA/ cm2 . cEl ectrol yte: A1 Br3 : KBr eutect i c . dCurrent den s i t i e s : 1 3 mA/cm2 for 9-mm el ectrolyte th i ckness ; 50 mA/cm2 for 3-mm thi ckness and stationary

operation; an·d 60 mA/cm2 for f l ow operation. e 1n A1 Cl 3 : KCl eutecti c . fCurrent dens i ty: 2 2 mA/ cm2 for 9-mm el ectrolyte thicknes s ; current den s i ty i ncreases wi th decreased

el ectrolyte th i c knes s .

59

en o

Cathode Regenerator

\ I

+

Recovered H&logen Gas j

Heat Exchanger

\

Spent Anode Liquid

Thermal Regenerator

FiglJre 1-1 0. Schem� of Thermally Regenerative Electrochemical System Proposed by McCully et at [72, 73] ·

In III N -.;;� II II � � �

;3 I "'" -en

S=�I I.I ____________________ --.-____ T:..,:R=-...... 4'-"1-=-6 -� ��

The research on systems conta i n i ng t i n hal i des was abandoned due to

h i gh decompos i ti on temperatures of SnC1 4 ( >1 700bC ) and the severe

corros i on probl ems associ ated wi th SnC1 2 ( g ) and C1 2 ( g ) at th i s tempera­

ture , i n add i ti on to the probl ems of separati ng the two gaseous products

[53 , 54 J . The anode regeneration i n CuCl systems d i d not appear very

d i fficul t [ ( CuC1 2 ( � ) � CuCl ( � ) + 1 / 2 C1 2 ( g ) J at temperatures bel ow the

mel ti ng poi nt of CuCl ( 1 365° C ) , but the gal van i c cel l performance [CuCl

( anode ) I CuC1 2 ( cathode ) J was poorer than that of other systems ( see

Tabl e 1 -4 ) . .

The TeC1 2 ( anode ) I CuC1 2 ( cathode ) system was cons i dered the most

promi s i ng based on the gal van i c cel l performance ( see Tabl e 1-4 ) .

Therefore , a more detai l ed analys i s of the el ectrode ki net i cs of both

e l ectrodes was performed [70 J . The el ectrochemi cal oxi dati on reacti on

Te ( I I ) + Te ( I V ) + 2e- was found to be faster on Pt el ectrodes than on Ta

e l ectrodes . On Pt el ectrodes an exchange current of about 1 00 mA/cm2

( 1 6 mol e % TeC1 2 i n A1 C1 3 ) was found . Tafel curves for th i s oxi dati on

on Ta and Pt are shown i n Fi g . 1 - 1 1 . The el ectrode reacti on Cu ( 1 1 ) + e­

� Cu ( 1 ) was found to be much s l ower than the correspondi ng Te ( 1 1 ) oxi ­

dat i on . Tantal um el ectrodes were found to catalyze th i s reduct ion

react i on . Tafel curves for the reduct i on on Pt and Ta are shown i n

Fi g . 1- 1 2 . The cupri c chl ori de cathode showed poorer performance than

the TeC1 2 anode . Current dens i ti es of about 35 mA/cm2 were obtai ned at

O . l -V pol ari zati on ( 1 R-free ) .

Cel l s four i nches i n d i ameter were constructed and tes ted wh i ch

produced current den s i ti es of about 20 mA/cm2 . Based on the el ectrode

ki neti cs studi es , u s i ng p l ati num anodes and tanta l um cathodes and 0 . 1-

6 1

TR-4 16 !;::�I 1111--------------------------------------------------------------

1000

� 0 __ 0-0--

E �

/0 Platinum Electrode /0 o « .s /

.i!' 'iii 100 c: " 0 c �. :J U

1 00.':---------:'0."-1 -----�0.2::---------:J0.3 Total Voltage (V)

Figure 1-1 1 . Tafel Plot,s for the Electrochemical Oxidation of Te( l I ) (1 6 M'ol:e;· "'% in Alels ) act 2000 C I . . .

_

Platinum Electrode 0-0-0-o:;:::::::.-ll-ll-ll-/tf. Tantalum Electrode

it1 ,{j t!

� //0 � / 0 i 10 hll/ " ! o / � . I :; u

I r

10�------+0.1::--------0�.2::------�0.3· Overvoltage (V)

Figure 1-1 2. Tafel Plots for the Electrochemical Reduction of Cu(l I) (4 Mole· % in 3:1 Molar Ratio AICI�:KCJ) at 2000 C . [74] . 6 2 "

- TR-4 16 S=�I I_I -----------------------------

cm i nterel ectrode spac i n g , performance improvements of a factor of 1 0

were proj ected [62 J . Th i s opti mi sti c [4J , esti mate i ncreased by another

factor of 2 . 5 by i ncreas i ng the TeC1 2 concentrati on i n the anode . A

un i t of formal 5-kW power based on thi s concept , was bui l t and operated

for about 30 mi nutes , after wh i ch l eakages resu l ted i n cel l shutdown

, [55 , 62 J .

The effect of the presence of A1 C1 3 i n the compartment , i n wh i ch

TeC1 2 and TeC1 4 were a l so present, i s shown i n F i g . 1 - 1 3 . The data re­

ferred to as " h i gh Te " were obta i ned from cel l s i n wh i ch the TelAl mol e

rati o was greater than uni ty , whereas the " l ow Te " data refer to rati o s

of 0 . 02-0 . 35 . The experimental l i nes obta i ned were cons i stent wi th the

theoreti cal Nernst s l opes [61 J . The l arge effect of the TelAl mol e

rati o suggested that there was an appreci abl e i nteracti on between the

two chl ori des [61 J . Spec i es characteri zation stud i es were not performed .

Data on tel l uri um species i n ch l broal umi nate mel ts are avai l abl e i n the

l i terature [75 J . More recent stud i es of el ectrode processes a� a func­

t i on of ch l oroal umi nate mel t aci d i ty are al so ava i l abl e [76J .

Exampl es of d i scharge c�rves for TeC1 2 1 CuC1 2 cel l s are shown i n

Fi g . 1- 1 4 . The vol tage under l oad was ma i nta i ned for a period of about

20 hours at a current effi c i ency l evel of 75% ( 50-ohm l oad ) [61 J .

Though the gal van i c cel l stud i es i nd i cated the feas i bi l i ty of a

TeC1 2 1 cuC1 2 cel l , the abi l i ty of thi s system to undergo thermal regen­

eration was not s uccessfu l ly demonstrated . At the regenerati on tem­

peratures ( >5 50° C ) , TeC1 4 decomposes i nto gaseous TeC1 2 and C1 2 , and

s everal years of research were s pent i n tryi ng to dev i se a s u i tabl e and

effi c i ent separation method [70 , 77 , 78 J .

6 3

TR-416 S=�I I.I --

----------------------------'-

o � a:: Q) "0 � :> Q)

1 0

� 1 Q) -I- 0 ::. o ...J

o

Q) 1- 0 .c Ol I

I CJ)

o

0.1 L--_-L-_----L-_----.l..-_----L_-I._----! 0.2 0.4 0.6 0:8 1 .0 1 .2 EMF (V)

Figure. 1-1 3. "PC!tential�of the TeCl2(Anode)lCuCl2 (Cathode) Galvanic Cells - l in Molten �ICli3 [�1]

- - e;- - Cel l A 0.6 - 0 - Cel l B .' 0.5 ,�> - - - - - - - - - - - - - - .ee - _ .

� OCV ....... � ... , __

Q) 0.4 ��8':'" Ol .s "0 >

0. 1

5 10 1 5 Ti me (h) 20 25 30

Figure 1-1 4. Discharge Curves for TeCI2(Anode)/CuCI2(Cathode) Cells under 5Q- and 1 QQ-·Ohm Loads [61 ]

64

S=�I I.I _______________________ ---=-TR=.:---.:4::..::1-=-6

-� � ,

A successful but energy- i neffi c i ent separati on was ach i eved by

rap i dly q uenchi ng the gaseous mi xture at the l i qui d n i trogen temperature ,

produci ng so l i d TeC1 2 . Measurements of the rate of recombi nati on of

TeC1 2 and C1 2 i nd i cated that at 500°C the recombi nation was compl ete

wi th i n 20 ms [61 J .

A second approach to separate TeC1 2 from C1 2 was to keep the TeC1 2 i n the l i qui d phase by formi ng a coordi nati on comp l ex wi th one of the

chl ori des of group I I IB of the period ic chart : (A1 , Ba , Te ) TeC1 4·

comp1 ex( � ) heat> TeC1 2 · comp1 ex( � ) + C1 2 ( g ) . Extens i ve phase di agram,

thermal analysi s , and conductometri c studi es were carri ed out [77 , 78J . . . .

Tabl e 1 - 5 s ummari zes the regenerat·i on study resu l ts . The best regen­

erati on yie l ds were on the order of 2% wi th TeC1 4 · 2A1 C1 3 · KC1 . I t i s not

very c l ear whether a l arge improvement i n the regeneration of thi s

system can be achi eved .

Tabl e 1 - 5 . SUMMARY OF EXPERIMENTAL REGENERATOR OPERATION [74 , 77 , 78J

% Regeneration Operati ng System ( s i ng l e pass ) Pressure ( atm abs ) Temperature ( O C )

TeC1 4 oT1 C1 O . l a 1 5 500 TeC1 4 o GaC1 3 0 . 6 1 8 650 TeC1 4 oA1 C1 3 0 . 7 1 0 572 TeC1 4 oA1 C1 3 0 . 9 1 1 608 TeC1 4 0 2A1 C1 3 0 . 4b 20 230 TeC1 4 o 2A 1 C1 3 ° KC1 1 . 8 1 8 548

a l n the presence of C1 2 , T1 Cl i s oxi d i zed to T1 C1 3 , wh i ch at th i s temperature and pressure s eems to decompose , y ie l di ng ca . 20% C1 2 [77 J .

bA1 C1 3 subl imes under these cond i ti ons .

6 5

S=�I I.I ________________________ T=-:R:.;:..--=4:..=..1 6=__ -� ��

1 . 2 . 2 Metal Oxi des

The thermochem'i cal and thermodynami c ca l cu l ati ons by Snow [ 1 4 J i n­

di cated that the oxi des l i sted i n Ta bl e 1-6 are thermal l y revers i bl e

compounds [62 J . Some of these oxi des were sel ected for pract i cal cel l .

tests i n 1 960 . Test cel l s were cyl i ndri cal pel l ets havi ng three l ayers -

anode l e l ectrolyte l cathode - obtai ned by press i ng dry powdered materia l s

[62 , 79J . The e l ectrolyte l ayer ( 70- 90% i nert Zr02 mi xed wi th an el ec­

tro lyte of Na2C03 : L i 2C03 eutecti c ) was pressed fi rst and then the anode

and cathode l ayers were pressed at oppos i te ends . The cel l s were p l aced

between p l ati num el ectrodes , and gal van i c cel l tests were conducted .

Typi cal ga l van i c cel l resu l ts wi th sel ected oxi des are shown i n

Tabl e 1 - 7 [79 J � The conducti v i ty of the cel l s improved upon add i ti on of

graphi te . However , th i s materi al was obvi ous ly i ncompati bl e wi th the

anode regeneration step due to the formation of carbon oxi des and re-

duction of the metal oxi des to meta l . The work on oxi des was abandoned

i n favor of the hal ide systems , wh i ch seemed more promi s i ng ( see Section

I . 2 . 1 ) [62 , 69 J .

1 . 2 . 3 Di scus s i on o f TRES Type 2

Th i s approach of mul t i p l e el ectrochem i cal reacti on products and

mul ti p l e-step regeneration ( see Fi g . S- 2 ) i s far more compl ex t han the -

rema i n i ng types of TRES . To date , none of the systems i nvest i gated

d i s pl ayed both good cel l performance and good regeneration performance .

It i s c l ear that most of the systems promi s i ng from an el ectrochemi cal

poi nt of v i ew had a very poor regenerator performance i f two gases were

the resu l t of the thermal regenerati on . It i s our fee l i ng that systems

of th i s type shou l d be i nvest i gated on ly i f the thermal decompos i ti on

66

S=�I rllr ________________________ T_R_-_4_16_

,

products are i n d i fferent physi cal states , or i f major brea kthroughs i n

the separati on of gases are made i n the near future .

Tabl e 1 - 6 . THERMALLY REVERSIBLE OXI DES [62]

Reacti on . +

3Mn02 + Mn304 + 02 + 2Mn304+ 6MnO + 02 + 2A920 + 4Ag + 02 + 4CuO + 2Cu20 + 02 + 2Sb205 +2Sb204 + 02 + 2Sb204 +2Sb203 + 02 + T1 203 + T1 20 + 02 . + 2Pb02 + 2PbO + 02 + 6U03 + 2U308 + 02 + U308 + 3U02 + 02 + 2Rb203 +4RbO + 02 + 4RbO + ?Rb20 + 02 + Rb203 + Rb20 + 02

Tabl e 1 - 7 .

Cel l

MnO l co�- I Sb205 Mno l co�- I Sb205

cU20 I CO�- I Sb205

Cal cu l ated Reversal Temperature ( OC )

652 1 427 1 27

1 1 27 527 927 677 362

1 327 1 827 527 777 652

CHARACTERIST ICS OF OX I DE CELLS [79]

Compos i ti on i n Mol ten Eutecti c Li 2C03 : Na2C03

a EMF ( V )

Anode Cathode Cal cd . Measured

80 wt % MnO 80 wt % Sb205 0 . 29 0 . 1 7

same as above wi th 30 0 . 29 0 . 27 vol % graph i te added

graph i te added 0 . 30 0 . 23

67

Res i stance

(ohm )

2800

29

2000

S-�I -""� - 1'.'1 - � "' �

68

S=�I 'I_I _________________________ T_R_-_4_1_6

SECTION I I

THERMAL REGENERATION : ALLOYS OR B I METALLI C SYSTEMS

A schemati c representati on of a thermal ly regenerati ve a l l oy system

i s shown i n F i g . 1 1 - 1 ( see Type 3 above i n the Introducti o n ) . L i q u i d

metal C i s oxi d i zed to the respecti ve C+ i ons a t the anode . These i ons

mi grate i nto the C+ c�nduct i ng el ectrolyte and undergo reduct i on and

sol ubi l i zation i n B or al l oy formation [ C ( B ) or CXBy ) ] at the cathode .

Th i s a l l oy , a l one or combi ned wi th anode materi al , i s pumped ( e . g . ,

el ectromagnet i cal ly ) or fl ows to a bo i l er where i t i s heated above the

boi l i ng po i nt of the meta l of l ower boi l i ng poi nt . I n a separator the

vapor phase , ri cher ( not necessari l y pure ) i n the more vol ati l e compo­

nent , i s separated from the l i qui d phase , ri cher i n the l es s vol ati l e

metal , and the two streams are i ndi v i dual l y returned to the gal vani c

cel l . Therefore , the el ectrochem ical reacti on product i s a l i qu i d metal

a l l oy [ C ( B ) ] or an i nt�rmetal l i c compound ( CXBy ) in a conc�ntration cel l

wi th respect to the e l ectroacti ve speci es C+/ C .

The el ectrode reactions are essenti a l l y i denti cal i n oppos i te

d i recti ons :

C ( or l / z C B ) anode> C+ + e­z y

C+ + e- cathode> C ( B) (or l /x C B ) B x Y Such cel l s can be represented schemati cal l y as

or

69 .

I I- l

I I -2

- TR-416 S=�I I�.-�I --------------------------� ���

a Separator

i Boi l i ng Mixture

Boi ler

B-Rich Vapor

" �;" " "'" C-Rich Liqu id

Interrupter

b

Alkal i Metal i n SS Sponge

" � I /�/ '" , -' Anodes { , - - -,'t. - -. - - ..... - -, Electrolyte �:-:- - � - -� - � �, : Cathodes ,', '"" 0

\

Anode Metal

r- Cathode Metal with 5 to 30 mole % Anode Metal

;- Cathode Metal with 1 5 to 40 mole'% Anode Metal

r - - - - - - -, i L _ _ _ _ _ --l

tHeat �xchanger

-- Condenser Section

.-

Evaporator -- Section ..

Figure 1 1-1 . i Sch'ematic Representation of a Thermally Regenerative Alloy Sy,stem (a) for Amalgam Cells c: = Na, K; B = Hg; and (b) for Bimetall ic Cells

70

h" TR-4 16 S=�I I_I -----------------------------

i n wh i ch C and B represent a vari ety of metal s ( e . g . , C = L i , Na , K, Rb ,

Cs , Mg , Ca ; and B = Hg , Cd , Zn , T1 P , I n , Ga , Pb , Sn , B i , P ) and a i represents the acti v i ty of metal C i n each el ectrode . The cel l potenti al

can be expressed as a function of the acti v i ti es of C as fol l ows :

or

E = - RT Ji.n � = - RT Ji.n al n F aO n F

RT a l E = - - Ji.n -nF a2

( i f aO = 1 , pure metal )

The potenti a l decreases wi th i ncreasi ng acti v i ty of C i n the a l l oy

formed at the cathode . The more negati ve the free energy of formation

of the i ntermeta] l i c compound or a 1 1 oy , the more thermodynami ca 1 1 y

I I- 3

I I-4

s tabl e it i s and . the l arger the cel l potenti a l i s that wi l l be obta i ned

due to the l ow acti vity of C i n the compound or a l l oy .

The el ectrolytes for most o f these cel l s are mol ten sal ts because

of the hi gh conducti vi ty of these med i a (�1 0- 1 - 1 0 ohm- l cm- l ) compared

to that of aqueous el ectrolyte sol uti ons (�1 0-4 � 1 0-2 ohm-l cm-l ) and

because of the h i gh exchange currents obta i ned at metal e l ectrodes due

to smal l acti vati on pol arization [80] . The major vol tage l osses i n th i s

type of cel l are ohmi c .

Many o f the i mportant parameters i n sel ecti ng a bi metal l i c system

for feas i bl e thermal regeneration are i mpl i c i t i n the phase diagram of

the system . An approximate ana lys i s of the system can be made by con­

s ider i ng j ust the CB bi nary d i agram ( e . g . , Fi g . II - 2 ) though in actual

o perat ion , the presence of the el ectrolyte components , sol ubl e to some

extent i n the anode and cathode materi al , comp l i cates the regenerati on

analys i s . C and B are chosen such that they have an appreci abl e d i f-

7 1

TR-4 16 S=�I I;.��I ----------------=------- � ���

a

T V

Composition

c

b

PI

C Composition B

Composition (a/o)

Figure 1 1-2. (a) Constant-Pressure Phase Diagram for a Generic Bimetallic System C/B; (b) Three-Dimensional Phase Diagram for a Two.-Compone.nt System; and (c) Phase Diagram Showing Equilibrium between Vapor and Solid in the V-CB Re.gion Resulting from Overlap of V-L and L-CB Regions [1 04]

72

S=�I IIlI ______________________ T.::...;R:;.,:..--=4.;:..lSO-

ference i n bo i l i ng po i nts . The equi l i bri um pressure of the cel l system

i s determi ned by the condensation temperature i n the radi ator of the

regenerator system [T, i n Fi g . I I -2 ( a ) ; T, can be equal to Tl , the

gal van i c cel l operati ng temperature ] . Th i s pres sure fi xes the appl i ca­

b l e T vs . compos i ti on phase di agram and wi l l determi ne whether the

l i qui d/vapor l oop i s separated from the sol i d/ l i qu i d equi l i bri um regions

[ Fi g . I I- 2 ( b ) i l l ustrates the effect of pres sure on the d i agram of Fi g .

I I- 2 ( a ) , and Fi g . I I- 2 ( c ) i l l ustrates a case i n wh i ch the two l oops

i nteract and present a sol i d/vapor equ i l i bri um reg i on ] . The compos i ­

t i ons of the streams return i ng to the anode and cathode are determi ned

by the temperature of the bo i l er of the regenerator [T2 i n Fi g . 1 1 -

l( a ) ] , and the anode and cathode streams have composi ti ons xl and x2 '

respecti vel.y , provi ded the boi l er i s fed wi th a' stream of compos i ti on

x3 . I f the stream return i ng to the anode i s not enri ched enough i n C ,

fracti onation can be cons i dered . S i nce the net compos i ti on of C and B

i n the cel l shoul d be kept constant , the rate of d i sti l l at i on of the C-

enri ched stream i s d i rectly rel ated to the current-generati ng capaci ty

of the cel l . I n other words , the rate of c i rcul ation of the streams

enri ched i n C and B i s di rectly proporti onal to the rate at whi ch

coul ombs of e l ectri c i ty are generated ( current den s i ty ) .

. Compound formation i n the cel l i s des i rabl e to l ower the acti v i ty

of C i n the cathode and thus i ncrease the obta i nabl e vol tage ; however ,

for thermal regenerati on , the mel ti ng poi nt of the compound shoul d not

be so h i gh that separati on of the vapor/ l i qu i d and l i�u i d/so l i d reg ions

cannot be achi eved at a practi cal operati ng pressure . The practi cal

probl em of a d i agram s uch as F i g . I I-2 ( c ) is that the sol i d phase CB

73

- TR-4 16 S=�I I_I -----------------------------

wi l l be present i n the condenser, a l l owi ng the bu i l dup of sol i d materi al s ,

eventua l ly a l teri ng the anode compos i ti o n , and consti tuti ng a serious

engi neering probl em . Azeotrope formation can a l so be · encountered in the

l i q u i d/vapor d i agram. It must be emphas i zed that the above cons i dera­

ti ons assume equi l i br i um pressures and wi l l not accurate ly represent a l l

condi ti ons i n an operati ng cel l , but they al l ow one to draw certa i n

concl us i ons wh i ch are genera l l y val i d and hel pful i n eval uati ng the

systems and pred i ct i ng best theoreti cal operati ng cond i t i ons . Fi sher

[ 1 04] presents a comprehens i ve revi ew of �hase d i agram cons i derati ons

app l i ed to b imetal l i c systems .

For el ectr ical regeneration , i . e . , operation as a rechargeabl e

secondary battery , i t i s necessary that the el ectrochemical reacti ons be

revers i b l e , wi th negl i g i bl e pol ari zat i on , l ow sel f-d ischarge , and hi gh

power dens i ti es . To mi n imi ze sel f-di scharge and coul ombi c i neffi c i ency ,

the mutua� sol ubi l i ti es of C and B i n the el ectrolyte must be as l ow as

poss i bl e under operati ng cond i t i ons .

Section 1 1 . 1 descri bes i n deta i l the l i qu i d meta l cel l s wi th B = Hg

and C = K, Na ; i . e . , amal gam cel l s wi th the boi l i ng po i nts of B < C .

Regenerati on i s performed by d i sti l l ation [see Fi g . I I - l ( a ) ] o f the

combi ned cathode and anode materi al s , whi ch y iel ds streams of sol vent ( H g ) - ­

ri ch d i sti l l ate and sol vent- poor res i due to be returned to the cathode

and anode , respecti vel y . The thal l i um- potass i um system , i n wh i ch the

regenerati on i s obta i ned by part i a l sol i d i fi cati o n , is a l so descri bed i n

th i s section . These systems operated reasonably we l l i n the thermal

decompos i ti on mode or as secondary batteri es and can be operated i n the

el ectrothermal regeneration mode .

74

TR-4 16 S=�I I*I -----------------------------

Secti on 1 1 . 2 d i scusses the rema i n i ng b imetal l i c systems , i n wh i ch

regenerati on i s performed by di sti l l ati on [ see Fi g . I I- l ( b ) J of the

a l l oy formed i n the cathode . The d i sti l l ate i s the el ectroacti ve

meta l -ri ch phase ( the anode meta l ) of a compos i ti on determi ned by the

phase d i agram characteri st ics and operati ng temperature and pressure

condi ti ons of the system; i t i s returned to the anode . The res i due i s

the el ectroacti ve metal - poor phase , wh i ch returns to the cathode . The

maj ori ty of the b imetal l i c systems i ntended for use in a thermal re­

generati ve mode empl oyed l i th i um or sod i um metal as anodes , but most of

these cel l s d i d not operate successful ly i n the thermal regenerati on

mode . However , the cel l s operated more successful ly i n the el ectri cal

regeneration mode , as secondary batteries , and operation in the el ec- '

trothermal regeneration mode shou l d be poss i bl e .

The groups i nvol ved i n the major research efforts on these systems

were : General Motors Corporation (Al l i son and Del co- Remy D i vi s i ons ) ,

wh i ch proposed the systems Na/Sn [81 - 84 J , K/ Hg [84-88J , and K/Tl [89 J ;

North Ameri can Av i ati on , Inc . ( Atomi cs Internationa l ) , wh i ch devel oped

the sodi um-mercury thermal l y regenerati ve al l oy cel l ( TRAC ) [93-1 01 J ;

and Argonne Nat i onal Laboratory ( Chemi cal Eng i neeri ng Di vi s i on ) , wh i ch

i nvesti gated several b imetal l i c systems ( Na/Pb ; Na/B i ; L i / Bi ; L i /Te ; '

Li /Sn ; L i /Cd ; L i / P b ; Li /Zn ) i n B research effort paral l el to that on the

l i th i um hydri de system ( see Section 1 . 1 ) [36 , 46 , 1 04 J . The phys i co-

chemi cal i nvesti gati on of the bi meta l l i c systems was very thorough and

careful , a iming at thermal or e l ectri cal regenerati on . Few of the sys­

tems were found s u i tabl e for thermal regeneration . The search for

coupl es exhi bi t ing good el ectri cal regeneration properti es ( secondary

75

TR-416 S=�I I.I ---------------------------=-=..:.-...::.-:....­-� ��

batteri es ) was fru i tful and i s be i ng continued [1 05J . S i nce th i s rev i ew

i nvol ves only the thermal ly and coupl ed thermal ly-e l ectrolyti cal ly

regenerabl e systems , the work at Argonne a imed at secondary storage

batteri es ( pure el ectrolyti c regenerati on ) i s not rev i ewed here .

To p l ace t hese act i v i t i es i n perspecti ve , i t must be rea l i zed that

at the time of these i nvesti gati ons ( 1 960- 1 966 ) the research was a imed

at provi di ng s pace power sources uti l i z i ng waste heat from a nucl ear

reactor . Therefore, the operati ng temperatures were di ctated by the .

characteri sti cs of the reactor and i ts cool i ng system . In add i t i on ,

zero gravi ty resul ted i n the expendi ture of only a smal l amount of

energy for pump ing the heavy fl u i d .

A n added advantage o f thermal ly regenerati ve systems emp l oy ing

a l l oy cell s i s the capabil i ty of storage of el ectri cal energy , s i nce , i n

pri nci p l e , these systems coul d be rather compact secondary batteri es .

1 1 . 1 AMALGAM AND THALLIUM CELLS

1 1 . 1 . 1 The Potass i um-Mercury System

The el ectrochem ica l and regenerati ve feas i bi l i ty of the K/Hg system

was i nvesti gated at the Al l i son D i v i s i on of Genera l Motors Corp . by

Agruss , Henderson , Kara s , Wri ght , and Mangus [84-88J . The l i q u i d metal -

cel l empl oyed was

,\ -�� .

--�--� . :(: . ..., K(a l ) ( Hg ) I KOH- KBr-KI I K( a2 ) ( Hg )

where the e l ectro lyte compos i t i on was 70 : 1 5 : 1 5 mol e % i n KOH : KBr : KI ,

wh i ch i s mol ten at 250 ° C . Some bi nary eutecti cs , such a s KOH- KBr and

KOH- KI , were a l so stud i ed but were d i scarded because they presented

76

S=�I I�I TR-416 -� .,

more serious corros ion probl ems . Some of the phys i cal properti es of the

ternary eutecti c , as wel l as the phase di agram for the KOH-KBr- KI sys­

tem , are shown i n Fi g . 1 I- 3 [84 ] . The el ectrolyte conducti vi ty i n ­

creases wi th temperature , and therefore temperatures h i gher than 300°C

were chosen for cel l operati on to mi n i mi ze ohmi c l osses . The mutual

so l ubi l i ty of potass i um i s greater than that of mercury i n the ternary

eutecti c ( s ee F i g . 1 I- 2 ) , g i v i ng ri se to sel f-di scharge and only 90- 95%

coul omb i c eff i c i ency in thi s system.

Potent ia l s ?f cel l s K I K�l ass I K(H9 ) obtai ned by Lantratov and

Tsarenko [90] are shown i n Fi g . 1 I -4 ( 1 36 °C ) . Add i t i onal measurements

at operati ng cond i t i ons were made by LaManti a and Bon i l l a [91 ] , who a l so

performed a more thorough thermodynami c i nvesti gati on of the system .

Curves cal cu l ated accord i ng to Eq . 1 I -3 agree wel l wi th experim�ntal

data ; �he s l i ght dev i ati on i s rel ated to compound formati on [84-86] .

Reference 92 presents some aspects of the thermodynami cs of th i s system .

Several cel l confi gurati ons were empl oyed by the General Motors re­

searchers . For batch operation a d i fferenti al dens i ty cel l [84-86 , 88]

hel d wi th i n a ceram ic cruci bl e was bu i l t wi th the K/Hg ama l gam on the

bottom , the ternary mel t (d = 2 . 4 g/cm3 ) fl oati ng on top of the ama l gam

l ayer , and f ina l l y , a l ayer of mol ten potass i um ( d = 0 . 78 g/cm3 )

fl oati ng on top of the el ectrolyte . Leads were i ntroduced i n to the K/Hg

and K l ayers by i ron wi res surrounded by a l umi na tubes . A fl owi ng

system of th i s type was a l so made by the addi ti on of su i tabl e i n l ets and

outl ets , u s i ng a sta i n l ess steel body , but was not successful . Batch

di fferenti a l dens i ty cel l s were operated in an el ectri cal regeneration

mode undergoi ng charge/di scharge cycl i ng ( 1 2 mi nutes of charge and

77

- TR-4 16 S=�I I_I ------------------------------

2.0 .-----,----.------,

'Conductivity . (j 1.0 1----+---=:!I""-":--i-'-----..1

(0-1 cm-1 ) lO.80 1----+-----=:!�---..1 10.60 1----+----+-----..1

0.40 L-.. __ -'-__ --l. __ _

1 .5 1 .6 1 .7 Reeiprocal of Absol ute

Temperature 1 IT 1 .8

70% KOH, 1 5% K I , 1 5% KBr p = 2 . 1 g/ml at 300°C m p = 225°C Mutual sol u b i l ity with Hg - nOAe M utual sol u b i l ity with K - 2 mol % K satu rates eutectic at 3000 C -· 1 mol % eutectic saturates K at 300°C

Figure 11-3. Ternary Phase Diagram of KOH-KBr-KI and Properties of the Ternary Eutectic [84]

78

S=�I I_I _________________________ T_R_-_4_16

Ol 1: -� c: E ::1 -·iii (/) ca -0 a.. -0 c: 0 � 0 ca .... u. �- Q) : "0 'f :E

1 .000 r-----------------.

0.1 000

0.0100

0.00 1 0 ....... _'"'--_ ........ __ ..... _ ........ __ '"'--....L---' o 200 400 600 800 1 000 1 200

Potential of K(Hg) (mV)

79

S=�I I_I _________________________ T_R_-_4_1_6

d i scharge at a current dens i ty of �90 mA/cm2 i nterrupted by a 3-mi nute

open-ci rcu i t rest per iod ) . Some resul ts of the cycl i ng tests are shown

i n F i g . 1 1- 5 [84J . The shapes of the curves for charge and di scharge

i nd i cate concentrati on pol arization [84 , 88 J . From the decrease i n OCV

at the successi ve rest peri ods , a cou l ombi c effi c i ency of 90- 95% was '

estimated [84 J . Ce l l s o f thi s type operated conti nuous ly for over 700

hours wi th no deterioration i n performance [85 J .

Wri ght [87J studi ed d i ffus i on propert i es of potass i um amal gams and

found a sma l l er d i ffus i on coeffi c i ent for K i n di l ute amal gams than i n

concentrated amal gams . Thi s expl a i ns why concentration pol ari zation

beg i ns at a l ower current dens i ty duri ng charge than duri ng d i scharge ,

at the same bul k e l ectrode compos i ti o n . At a gi ven current dens i ty and

bul k concentration , the di fferenti a l potass i um concentration ( surface to

bul k ) i n the ama l gam wi l l be l ower on di scharge than on charge . Thi s

corresponds to a l ower concentration po l ari zati on on di scharge .

A better ce'l l for f l ow operatio,n was bui l t by hol d i n g the el ectro­

l yte in pl ace by impregnati ng it in a porous , s i ntered MgO matri x

[84 , 85 J . However , the probl ems of matr ix strength and res i st i v i ty and

of seal s abl e to wi thstand pressure rema i ned unso l ved , pri nci pal l y for

h i gh current dens i ty and operation of l ong duration . Cracks i n the

matri x a l l owed mi xi ng of the anode and cathode materi al s , l eadi ng to an

i nternal short. F i gure 1 1-6 shows an exp l oded vi ew of the l i q u i d metal

cel l i n wh i ch a matri x sandwi ch confi gurati on was used , wh i c h a l l ows

e l ectrolyte feed after cel l assembly . Two ri bbed i nsu l ators "co i ned "

i nto the Kovar metal cel l ha l ves seal ed the three l i q u i d streams from

ea�h other and the outs i de of the cel l [84 , 88 J . Some progress i n

80

- TR-4 16 S=�I IIlI ---------------------------

1 .50

5 1 .00

Q) Cl til -(5 > 0.50

o

r--, . I 1 1 .5 cm2 Area . 1 Ampere

1 2 m ln� -j 1 .5-cm-Thick Electrolyte �- --- Irl ? - r 3 min

,..... 0 .84V

"-- �

r i n

� "-

I I I I

Time (min ) -

� � -

" /" 0.83V

"-...r-.. " r-

I I

Figure 1 1-5. Voltage-Time Plot of Cycling Differential Density Potassium-Mercury Cells [84] .

Mu.lt'iple LM _____ Cell Stack . ____

Exploded View of LMC

Figure 1 1-6. Potassium-Mercury Liquid Me.tal Cell (LMC) [84] .

8 1

- TR-4 16 S=�l lf.-�1 ---------------------------� ���

reduc i ng the res i s tance of the el ectrolyte- impregnated matr i x was

achi eved by reduc i ng the matri x thi ckness and improvi ng the preparation

of the paste el ectrolyte . However , the final res i stance ach i eved was

s ti l l rather l arge for h i gh current dens i ty operati on ( see Ref . 3 ) .

Data for the l i qui d/vapor equi l i bri um at 1 atm pressure ( the Hg

vapor pressure at �350 °C i s �l atm ) are shown i n F i g . 1 1 -7 [84 , 86 ] . As

i nd i cated by the phase d i agram, separation appears feas i bl e at the

operating temperatures , wi th the compos i ti on of the mi xture fed to the

bo i l er and the regenerator temperature and pressure d i ctati ng the

composi t i on of the streams to be returned to the anode and cathode . In

the tests performed at the Al l i son Di v i s i on of General Motors Corp . , the

mode of operation of the fl ow cel l s was not ful ly thermal ly regenera­

ti ve , s i nce the l i qu i d potass i um was fed from a tank . The potass i um­

enri ched mercury from the boi l er was stored and the mercury-enri ched

vapor from the boi l er was condensed and fed to the cathode . Fresh

mercury s upply Was furn i shed from t ime to time [88 , 3] .

The performance of the s i ngl e cel l shown i n F i g . 1 1-6 was tested

for about 430 hours wi th conti nuous di scharges up to 32 hours , produc ing

power output of 50- 1 00 mW/cm2 . Dur ing ca . 75 hours of the tes t , mercury

was fl owi ng through the cel l i n order to ma i nta i n vol tage and power . A

three-cel l stack , also shown i n F i g . II-6 , was operated a total of 1 04

hours with d i scharges total l i ng 64 hours , produci ng 50 mW/cm2 and wi th

mercury fl ow duri ng 60% of the operati ng ti me [84 , 88 ] .

A mathematical ana lys i s of the K/Hg system was performed , wi th the

cel l shown i n F i g . 1 1-6 d i v i ded i nto segments and wi th countercurrent

l i q u i d metal fl ow [84 , 88 ] . . Changes i n vo l tage , concentrati on , di f-

8 2

S=�l r.""'�r -� �� � --------------------------------------------�------------TR-4 16

\ 1 400

1 300

u. 1 200

e..... 1 1 00 Q)

� :::l 1 000 -co � Q) 900 a. E

800 Q) I-700

600 0 20 40 60 80 1 0

% Hg

Figure 11-7. Phase Diagram of Hg-K at 1 atm [84]

8 3

/.� TR-4 16 S=�I II.I ---..:..---------------==-=-===-­-� .

fus i on l ayer , and current dens i ty were cal cul ated for each segment and

then i ntegrated for the cel l . Resul ts were obta i ned after i ncorporati ng

data for the concentrations of potass i um l eavi ng and enteri ng the cel l ,

as fi xed by the phase d i agram. These compos i ti ons a l so fi x changes i n

heat capaci ty , vapori zati on , and therma l energy for a gi ven el ectri cal

output . From these cal cul ati ons [88J and experiments [84 , 88J , the K/Hg

thermal ly regenerati ve system was j udged competi t i ve wi th other regener­

ati ve systems ( even those i nvol v i ng Rank i ne cycl e mercury turbi nes ) i n

the 1 -50 kW range [84J .

These resu l ts s how that the K/ Hg system i s e l ectrochemi cal ly

s i mpl e and poss i bly capabl e of generati ng current densiti es of the order . 2 of 1 00 rnA/cm at ca . 0 . 5 V wi th rel ati vel y l ow sel f-di scharge rates .

The concentration pol ari zati on , wh i ch l eads to a bui l dup of a di ffus i on­

control l ed l ayer at the cathode , can be mi n imi zed by us i ng thi nner

cathode streams . Ohmic l osses can be decreased by improvi ng the con­

ducti vi ty of the el ectrolyte matr ix . Some mechani cal probl ems of the

cel l were : l eakage through the sea l s ; el ectrolyte l eakage out of the

matri x due to pressure d i fferences between anode and cathode compart­

ments ( e . g . , if the K i n the anode compartment i s starved of feed ) ; and

cracki ng of the matri x . One advantage of the l i q u i d metal cel l for h i gh

current dens i ty operati on , when the temperature i n s i de the cel l can ri se

apprec i ab ly , is the ease of heat management in the system due to the

fl ow operati ng condi ti ons and the i nherent excel l ent heat transfer

capabi l i t i es of these meta l s . Cl osed-l oop operati on ;- however , has not

been demonstrated .

84

/-=� TR-4 16 S=�I II-li ----------------------------­-� .

1 1 . 1 . 2 The Sod i um-Mercury System

The Thermal ly Regenerati ve Al l oy Cel l ( TRAC ) system was devel oped

at Atomi cs I nternati onal , a Di vi s i on of North Ameri can Avi ati on , Inc .

( presently Rockwe l l I ntl . ) by Recht , I verso n , Heredy , and Ol denkamp ,

among others [93- 1 01 J . The mi s s i on of the program was to i nvesti gate

the feas i bi l i ty of c l osed-cycl e , stati c devi ces for convert i ng heat i nto

e l ectri c i ty based on l i qu id metal amal gam cel l s wi th the sodi um-mercury

system. At the t ime , major emphasi s was pl aced on 'the use of waste heat

from nucl ear reactors for the production of el ectri c i ty i n compact ,

l i ghtwei ght space power pl ants wi th no mov i ng parts [Systems for Nucl ear

Auxi l l i ary Power ( SNAP ) J . As a resu l t , most of the research a imed at

devi ces wi th h i gh power output/we i ght rati o s and not at the h i ghest

poss i bl e effi ci ency of the systems . Moreover , a reactor source at

""700° C ( SNAP 8) and the constra i nt of mi n i mum radi ator area imposed

certa i n operati ng temperatures ( cel l : 'V460-5 1 0°C ; regenerator : 670-

700°C ) , thus a l l owi ng a maximum Carnot eff ic i ency of the order of 20% .

The effi c i ency was reduced to about ha l f of that val ue due to i rrevers i -

bi l i t i es and was reduced further by wei ght constra i nts to about 30% of

the Carnot effi c i ency [96 J . Performance ana lys i s of a 3- kW i sotope-

powered system, i n wh i ch the regenerati on was performed i n a two-stage -

d i sti l l ation process ( 900° C and 700 ° C ) , showed improved performance of

the cel l and h i gher Carnot effi ci ency ( 34% ) ( net effi ci ency estimated as

1 1 % ) , but th i s system had trade-offs in the need for two separators , two

pumps , and three heat transfer stages [ 93 J .

The TRAC program demonstrated the feas i bi l i ty of the cl osed- l oop

operation in the Na/Hg system ( cf . Section 1 1 . 1 . 1 ) . The program deve l -

8 5

TR-4 16 S=�I I.I ------------------� �

oped stati c cel l s [94J to furn i sh the proper background for the cl osed­

l oop operati on [95 J , as wel l as the l i q u i d/vapor equi l i br; u� d i agram

[94J under operati ng cond i tions . Section 1 1 . 1 . 2 . 1 descri bes the batch

cel l s and Secti on I I . l . 2 . 2 the c l o sed-l oop operat i o n .

1 1 . 1 . 2 . 1 Batch Cel l s [94J

The sodi um-mercury ama l gam cel l

1, ' 7 Na ( a l ) ( Hg ) / NaCN-Na l -Na F / Na ( a2 ) Hg

empl oyed by Atomi cs Internati onal conta i ned ama l gams of di fferent sodi um

act iv i ti es i n the anode and cathode compartments , wh i ch were separated

by a porous ( 40- 50% poros i ty ) beryl l i um oxi de matri x [98 J impregnated

wi th the ternary sa� t mi xture of eutecti c compos i t i on 58 : 30 : 1 2 mol e % of

NaCN : Na l : Na F , wh i ch i s mol ten at 477 ° C . Other el ectrolytes were a l so

i nvest i gated and found s u i tabl e : ternary or quaternary mi xtures of

anhydrous sodi um sa l ts of the ha l i de an ions ( I - , F- ) , cyan i de , and car­

bonate . The compos i tions and properti es of these el ectrolytes are

descri bed i n a patent [99J . These mol ten sal ts are therma l ly stabl e at

the proj ected cel l operati ng temperatures (�500 °C ) , have l ow res i st i v i ty

« 1 ohm cm) , have l ow el ectro lyte/sodi um amal gam mutual sol ubi l i ty

[�0 . 2 wt % sol ubi l i ty of sod i um i n the above ternary el ectrolyte was

found , and Hg was found to be i nsol ubl e ; the mutual so l ub i l i ty of Na i n

i ts sa l ts i s smal l er than that o f K i n i ts sa l ts J , and are chemi cal l y

i nert i n the presence o f cel l components , sea l s , matri ces , etc . The BeO

matri ces d i d not s how attack by sodi um ama l gam after about 1 200 hours at

EMF studi es of sodi um/sodi um ama l gams were carr ied out [94 , 1 01 J i n

86

S=�I I_I _________________________ ---'Tu,R.a.::-:.:I4 ...... 1 ....... 6 -� .

cel l s Na { � ) I Na+g1 ass I Na { Hg ) { � ) i n the 350-400°C range , extend i ng avai l ­

abl e l i terature data [ 1 02] . . At hi gher temperatures open-ci rcu i t vol -

tages of the stati c cel l s as a functi on of el ectrode compos i t i.on were

used . Fi gure I I-B shows an exampl e at 500 °C . The OCV val ues agreed

wi th i n 1 0- 1 5 mV wi th those cal cu l ated from Eq . 1 1-4 . The d i agram of the

stati c cel l used i n these measurements is shown i n F i g . 1 1 - 9 . Thi s cel l

was a l so used for stud i es of cel l res i stance , el ectrode pol ari zati on ,

and e l ectrolyte matr ix confi gurati ons as functi ons of temperature .

Si nce at the operati ng temperature (�500° C ) the vapor pressure of mer­

cury i s 5 . B-9 . 5 atm , the cel l was conta i ned i n a sta i n l ess steel pres sure

vessel under argon atmosphere at 9 . 5- 1 2 atm tota l pressure .

Batch cel l s of thi s type were operated for a per iod of �l BO hours

and the tests were termi nated vol untari l y . An examp l e o f the d i scharge

characteri sti cs of a stati c cel l wi th a BeO matri x ( 0 . 25-cm th i c k , 1 1 . 4-

cm2 effecti ve area ) , i mpregnated wi th the ternary el ectro lyte at 500 ° C ,

and conta i n i ng i n i ti a l ama l gam concentrations o f 37 . 6 and 2 . 9 atom % ,

respect i ve ly , i n the anode and cathode compartments i s shown i n Fi g . 1 1 - 10 ..

Cel l s were recharged el ectri cal ly after di scharge peri ods of 5- 15

hours . Fi gure 1 1- 1 0 s hows d i scharges at vari ous current dens i ti es

( upper numbers ) and the correspondi ng cel l res i sti v i t i es ( l ower numbers ) . -

Up to about 200 mA/cm2 , no apprec i abl e concentration pol ari zation was

observed , as i nd i cated by the approximate ly constant va l ues of the cal ­

cul ated cel l res i st i v i ti e s . The magn i tude of the concentrati on pol ar­

i zati on i s a functi on of the cel l vo l tage . For i nstance , for an OCV of

� . 4B V ( h i gher sod i um concentrati on at the anode ) , apprec i abl e concen­

tration pol ari zati on was found only at 300-360 mA/cm2 . Duri ng the

87

S=�I I_I ______________________ ._T_R_-4_1_6

0.9

0.7

0.6

� 0.5 UJ 0.4

0.3

0.2

0.1

o ���--����� o 0.2 0.4 0.6 0.8 1 .0

Concentration (Mole Fraction Na in Hg)

Figure 11-8 . Open-Circuit Potentials of Sodium-Sodium-Mercury Galvanic Cells at 5000 C [93]

Electrodes

Thermocouple '\

Condensation Coils

H\--+1-- Dense Ceramic Tube

���+�W-Electrolyte.-Matrix Disc

Figure� 11-9. Cross Section of Static Electrode TRAC Cell [94]

88

'" TR-4 16 !i::�I 'II'------------------------------------------------------------

0.5 50

50

3.2 85

:;- 3.2 44 - 1 25 (]) 0.3 3 .3 Cl

3.3 ell -(5 1 67 > Qi 0.2 3 .3

() 220

0.1 3.3

0

0 30 60 90 1 20 1 50 1 80 2 1 0

Time (min)

Figure 1 1-10. Discharge Charac�eristics of a Static TRAC Cell [94] I n itial anode and cathode sodium concentrations: 37.6 and 2.9 atom %.

Figure� above curve are current densities (mA/cm2) and below curve are calcu lated resistivities (ohm cm) .

89

TR-4 16 S=!!!II 1.1 -----------------------------

- � ��

recharg ing per iod a l arger concentrati on pol ari zati on was observed , i n

agreement wi th Wri ght ' s observati ons o n the K/Hg system [87 J .

The onset of contentration pol ari zati on was shown to be a function

of the qual i ty of the matr ix . The res i st i v i ty of the matri x el ectro lyte

i s greater by a factor of at l east s i x than that of the pure el ectrolyte

(� . 5 ohm cm) . Improvements i n the porous matri ces wi th res pect to

the i r s tructural characteri sti cs coul d l ead to better cel l performance .

I I . l . 2 . 2 Fl owi ng El ectrode Cel l and Cl osed-Loop Operation [95 J

. The cross-secti on sketch o f the fl owi ng el ectrode TRAC cel l i s

s hown i n Fi g . I I- l l . The cerami c matri x conta i n i ng the el ectrolyte i n

th i s cel l i s a porous tube o f h i gh puri ty al umi na , hel d between two

dense ceram ic end p i eces wi th i n the cyl i ndri cal cerami c cel l body . The

end p i eces , i n turn , are hel d between the cel l fl anges by i ron kni fe­

edge gas kets , wh i ch seal off the i ns i de of the cel l from the ai r . Th i s

seal was des i gned for easy test i ng but not for advanced cel l s ( see Ref .

1 00 for the practi cal cel l des i gn proposed ) . A thermocoup l e we l l passes

through the center of the al umi na tube . The two l i q u i d metal streams

are pumped through the cel l countercurrently , the mercury fl owi ng up­

wards i n the annul us ( cathode compartment ) between the cyl i ndri cal

ceramic cel l body and the outs i de of the matri x el ectrolyte , and the

amal gam fl owi ng downward i ns i de the porous a l umi na tube ( anode compart­

ment ) . A tubul ar sta i nl es s steel structure ho l ds the cerami c parts .

The enti re cel l was mounted i ns i de a constant-temperature oven for

testi ng . Th i s cel l was tested for a few days i n an el ectri cal regenera­

t i on mode , u s i ng reservo i rs for the reactants and products and us ing

90

TR-4 16 S=�I I;."'�I ------------------------------� � = �

Enriched Effl uent

Ceramic

�Io---++- Electrolyte Matrix

Fresh .-----+--+-- Mercury Feed

Insu lator-++f2tll����Bl Spent AmalgarT'!-al����4k�d1

Figure 1 1-1 1 . Cross Section of the Flowing Electrode TRAC Cell [93] I

9 1

TR-4 16 5=�1 1*1 -----------------------------

argon gas pressure to c i rcul ate the meta l streams . Current dens i ti es of

50- 1 00 mA/cm2 were mai nta i ned,· and a power of 35 mW/cm2 was achi eved wi th

an e l ectrode fl ow of 0 . 5- 2 . 5 cm/mi n . The res i st i v i ty of the el ectrolyte

matri x was two to four t imes h i gher ( 6- 1 2 ohm cm ) than that obta i ned for

the porous d i scs used i n the stati c cel l .

The coupl i ng of the fl owi ng e l ectrode cel l wi th the regenerati on

l oop is shown schemati cal ly for a l i q u i d metal cel l in Fi g . I I- l ( b ) .

The l iq u i d stream from the condens i ng radi ator i s nearly pure mercury ,

whi ch enters the cel l under i ts own vapor press ure at the rad i ator exi t

temperature . Thi s f ixes the system pressure at about 5 . 8 atm i f the

radi ator is at 485 ° C . The pressure i n the separator wi l l be very

nearly the same as the cel l i n l et pressure . Wi th the pressure and

temperature at the separator fi xed , the sodi um content i n the l i q u i d and

vapor phases i s determi ned by the equil ; bri urn val ues under these cond i ­

ti ons . The separator was des i gned as a centri fugal cycl one . Mercury

vapor and l i q u i d ama l gam enter the separator tangenti a l l y , where the i r

vel oci t i es cause them to s p i n around the i ns i de wal l , forc ing the vapor

to fl ow out the top . Provi ded that the boi l i ng materi al reaches the

separator wel l mi xed , equi l i bri um between l i qu i d and vapor i s approached

in the separator and a one-theoreti cal -pl ate separation shoul d occur

( th i s was veri fied under operati ng condi ti ons ) . Fi gure I I - 1 2 shows the

l i qui d/vapor equi l i br i um d i agram for sodi um amal gams at 5 . 8 atm . From

these data one can concl ude that at 685 ° C (�1 300 ° F ) the vapor phase i s

nearly pure mercury wi th 0 . 1 - 1 atom % sodi um ( cathode- stream ) whereas

the l i q u i d phase i s approximately 36 atom % sodi um ( anode stream ) .

If the regeneration can be performed at a h i gher temperature , e . g . ,

92

TR-416 S=�I I_I ----------------------------

21 00

2000

1 900 -

1 800 [ 1 700

1 200

1 1 00

1 000

900 �_'__....L...---'-�____L_�..L...__'___'____' 0 1 0 20 30 40 50 60 70 80 90 1 00

Composit ion (Atom % Na)

, Figute 1 1-1 2. Vapor/Liquid Equilibrium Compositions of the Na/Hg System at 5.8 atm [95]

93

TR-4 16 S=!!SI 1.1 -----------------------� .

81 5 °C (�1 500° F ) , a more s i gn i fi cant sodi um concentrati on wi l l appear i n

the vapor phase , wh ich coul d be parti a l l y condensed a t about 485 ° C , and

the condensed phase coul d be returned to the boi l er . Under these condi ­

ti ons the anode stream wi l l be essenti a l l y 60 atom % sodi um and the

cathode s tream wi l l be very ' l ow i n sodi um [93J .

A schemati c of the test l oop i s shown i n F i g . I I - 1 3 . A more com­

pl ete descri pti on of the cl osed- l oop assembly i s g i ven i n Ref . 95 . The

regeneration system test l oop was made of a 5- kW ( thermal ) el ectri cal l y

heated :bo i l er , a cycl one-type separator , a water-cool ed condenser , two

reservo i rs , an el ectromagneti c pump , and connecti ng tub ing as sembl ed i n

a l oop , coupl ed to the fl owi ng el ectrode ce l l ( F i g . 1 1 - 1 1 ) , a s shown i n

Fi g . 1 1- 1 3 . These parts were des i gned for operati on i n a free- fal l

envi ronment .

I n the fi rst test of the combi ned cel l /regeneration system , the

matri x was i mpregnated wi th el ectrolyte after the cel l was assembl ed and

connected to the system . After t he matri x was impregnated and the

excess el ectrolyte was drai ned out , the regenerati on l oop was l oaded

wi th 1 4 atom % amal gam and started up . The l oop was operated for a

s hort t ime to generate a supply of mercury and concentrated ama l gam i n

the reservo i rs . After th i s the cel l was fi l l ed , and the cel l and the

regeneration l oop were operated conti nuous ly for 1 1 8 hours and then shut

down for d i sassembly . Duri ng th i s test , it was found that s i nce the cel l

temperature was h i gher than expected ( to ma i ntai n the coo l est part of

the cel l oven at 485 ° C ) , the system pressure essenti aJ ly doubl ed and

reduced the separati on that cou l d be ach i eved . A maximum OCV of about

0 . 25 V was devel oped , but the system operated sati s factori l y , wi th the

94

TR-4 16 !i::�I [�[--

------------------------------------------------------------

Amalgam Reservoi r

Flowmeter

Sample Line

Galvanic Cel l r -,

Pump L --1

r.-""'T""--t----'-- Arg on Line

Flowmeter

Figure 1 1-13. TRAC Test Loop Flow Diagram [95]

95

TR-4 16 S=�I I.I ---------------------

-� �

cel l i nternal res i stance remai n i ng essenti a l ly constant . The l oop

performance was s teady but the pressure drop across the l oop i ncreased

at the end of the test . The matri x was i ntact ( dark col ored ) but excess

e l ectro lyte was found i n the cel l . The sea l i ng gas kets had been corrod­

ed by atmospheri c oxygen . The l oop conta i ned l oose bl ack materia l wh i ch

was found to be i ron . The bu i l dup of materi a l s i n the l oop and the

excess e l ectrolyte were bel i eved res pons i bl e for the pressure drop ,--"--

i ncrease duri ng the test .

The second test was conti nued unti l the system fai l ed . The matri x

e l ectro lyte i mpregnation was carr ied out i n a separate apparatus and the

i ron gas kets were prevented from contact wi th atmos pheri c oxygen by a

n i trogen fl ow system i nstal l ed i n the cel l oven . The test was performed

s i mi l arly to the prev i ous one . The cel l temperature was ma i ntai ned at

�495°C and the system pressure was ma i nta i ned at 9 . 2 atm . A better

separati on was obtai ned wh i ch a l l owed a maxi mum steady OCV of 0 . 32 V .

After a period o f about 625 hours the l oop port i on was shut down for

about 440 hours to repl ace l ea k i ng va l ves , and the cel l rema i ned i n

operati on i n a n el ectri cal ly rechargeabl e mode . The regeneration l oop

was then restarted and the who l e system operated for an add i ti onal 1 30

hours . The test was termi nated when the bottom i ron gasket of the cel l

began to l eak due to a fa i l ure i n the n i trogen purge system . The cel l

operated conti nuous ly for about 1 200 hours , duri ng whi ch the cel l i nter­

nal res i s tance rema i ned constant . The maximum power dens i ty generated

was 5 mW/cm2 from 25 mA/cm2 at 0 . 2 V . The power and current den s i t i es

were l ow as a resu l t of a h i gh res i sti v i ty of the el ectro lyte matri x ( 54

ohm cm) , caused by the l ow poros i ty ( 1 5% ) of the a l umi na tube empl oyed .

96

S=�I r_r ------------------------�T:..::.R:::...-..;::.4=-16

These tests demonstrated the compati bi l i ty of the a l umi na matri x

wi th the sodi um amal gam, though beryl l i um oxi de was found to be a better

mater ia l [ l OOJ , presenti ng h i gher res i stance to a l ka l i metal s and the i r

amal gams , and hav i ng h i gher thermal conducti v i ty than magnes i a or a l umi na .

No detectabl e amount of el ectrolyte l eached out of the matri x duri ng

about 500 hours . Ce l l materi al s do not seem to pose a probl em for l ong­

l i ved devi ces .

A performance analys i s of a 42- kW TRAC pl ant was carri ed out [96 J

coup l ed wi th a SNAP 8 reactor [600 kW (therma1 ) J , assumi ng a rather

smal l res i sti v i ty of the cerami c tubes (�3 ohm cm) , and 1 20 cel l s , us i ng

the regenerator at 693°C and 6 . 1 atm pressure and , therefore , anode and

cathode sodi um concentrati ons of 38 . 5 and 0 . 2 atom % . The pump i ng

requi rements were estimated as 2 kW for an overal l 42- kW net power

output , correspondi ng to a net effic i ency of 7% . Deta i l s of th i s analy­

s i s are gi ven i n Ref . 96 . A deta i l ed analys i s of a 3-W ( el ectri cal )

i sotope system wi th a two- step regenerati on , the two-stage TRAC cel l ,

was performed [93 J . Due to the h i gher regeneration temperature the

overal l effi c i ency was est imated as 1 1 % .

1 1 . 1 . 3 The Potass i um-Thal l i um and Anal ogous Systems

A d i fferent approach for regeneration i n the a l l oy systems has been

s uggested i n the l i terature ( cf . Ref . 86 , p . 808 , and Ref . 89 ) . A gal ­

van i c cel l of the type descri bed i n Secti on 1 1 . 1 ( A B I A+ I A B ) operates z y x y at a temperature above the mel t i ng po i nt of the a l l oys . The streams of

the a l l oys from the anode and cathode are combi ned , wel l mi xed , and

coo l ed down to a defi n i te temperature at wh i ch the mi xture parti a l ly

sol i di fi es , form ing a C-ri ch phase and a C-poor phase , one of whi ch wi l l

97

- TR-4 16 S=�I I_I -----------------------------

be i n the sol i d state and the other i n the l i qui d state . These two

phases can be mecha n i cal l y separated by conventi onal methods and the two

separated streams of regenerated anode and cathode materi al s i nd i v i dual ly

reheated to the cel l temperature and returned to the gal van i c cel l .

A system wh i ch seems su i tabl e for th i s type of regenerati on has

been proposed by Agruss , H i etbri nk , and Nagey [89J : the potass i um­

thal l i um system . Fi gure I 1- 1 4 shows the sol i d/ l i qu i d phase di agram for

thi s system obta i ned by Kurna kow and Puschen [ 1 03J . The gal vani c cel l

K (Tl ) al I K+ I K( Tl ) a2 cons i sts of a mol ten K-Tl sol ution r ich i n K as anode

and a mol ten K-Tl so l ution ri ch i n Tl as cathode , separated by a porous

matri x impregnated wi th mol ten KCl at an operati ng temperature hi gher

than 335 ° C . The combi ned streams of the anode and cathode materi al s are

taken to a contai ner i n wh i ch the mi xture , conta i n i ng 0 . 75 mo l e fraction

of tha l l i um , i s coo l ed down to 1 73 ° C v ia cool i ng coi l s . From Fi g . 1 1- 1 4

one can concl ude that a t th i s temperature a sol i d phase conta i n i ng 0 . 5

mol e fract ion o f Tl i s i n equi l i br i um wi th a l i qu i d phase conta i n i ng

0 . 84 mol e fracti on of Tl . The two phases are separated , i nd i v i dual ly

reheated to the cel l temperature , a nd returned to the anode and cathode .

Approximately 0 . 6 V has been obta i ned wi th cel l s of thi s type [89J .

A process for conti nuous transfer of the sol utions from the cel l tG

the separator and for cont i nuous ly reconvey i ng the separated , l ess dense

sol i d phase , fl oati ng on top of the l i q u i d phase , i s descri bed i n Ref .

89 . A mesh-type conveyor bel t transfers the sol i d to a mel ti ng pot

where i t i s heated by heati ng coi l s to the cel l temperature and returned

to the anode . The system i s made conti nuous s imp ly by feed i ng cel l

effl uent to the settl i ng conta i ner and removi ng an equal amount of

98

S=�I I_I ------------:-----------'------

I I I I I

- --- -- - - ----:--::------1

-1- - - - -

0. 1 0 0.20 0.30 0.40 0.50 0.60 0.70

Mole Fraction Tha l l i um

1 .00

TI

TR-416

Figure 1 1-1 4. Solid/Liquid' Equilibrium Compositions of the Potassium ­Thall ium System [103]

99

S=�I I*I _________________________ T_R_�....::c4__'_16

separated sol i d and l i qu i d .

Agruss , H i etbr ink , and Nagey [89 ] suggest that th i s same type of

regenerati on can be appl i ed to vari ous systems and menti on the fol l owi ng

speci fi cal ly in the i r patent : A1 /Se ; Al /Te ; B i /Ca ; Bi / K ; B i / Na ; Ca/Hg ;

Ga/Na ; Hg/Mg ; K/Sn ; l i / Pb ; Mg/Sn ; Te/Tl ; B i / l i ; B i /Te ; Ca/Pb ; Hg/ K ;

Hg/Na ; K/Se ; li /Sn ; Na/Sb ; Mg/Sb; B i /Mg ; B i /Tl ; Ca/Sb , l i /Tl ; and

Sn/Te .

From the thermodynam ic data for the K/Hg system [91 ] , Bon i l la ( i n

Ref . 86 ) concl udes that very l ow vol tages (�O . l V ) woul d be obta i ned i n

thi s mode of operation . The same concl u s i on appears to hol d for Na/B i

and l i /B i systems ( cf. Ref . 1 1 6 , 36 , and 46 for phase di agrams and EMF

data ) .

1 1 . 2 B IMETAllIC CEllS

1 1 . 2 . 1 Sod i um-Conta i n i ng Systems

1 1 . 2 . 1 . 1 The Sod i um-Ti n System

laboratory cel l s of the type Na l Na+gl aSs l Naxsn were operated i n

batch mode , i n the 500- 700°C range , by Agruss [81 ] . The cathode com­

pos i ti on was vari ed between 1 5- 30 mol e % of sod i um and the res ul t i ng OCV

were 0 . 42 -0 . 36 V ( 500 ° C ) and 0 . 43-0 . 33 V ( 700 ° C ) [81 , 84 ] . These resu l t�

agreed wi th previ ous EMF data obta i ned by Weaver et a l . [82 ] us i ng

Na l Na I -NaCl i NaxSn cel l s at 625° C , wi th the eutecti c el ectro lyte of

compos i ti on 62 . 5 : 37 . 5 mol e % Na I : NaC1 , mol ten at 562 ° C , in the stat i c

mode ( d i fferenti a l dens i ty cel l s ) and wi th fl owi ng el ectrodes ( us i ng a

fl ame-sprayed a l umi na H-cel l ) .

The stati c and fl owi ng cel l s were used to study charge/di scharge

100

/. = , TR-4 16 S=�I II-II ----------------------� -� �

behavior of these cel l s . One stati c cel l showed abi l i ty to undergo

cyc l i ng tests on a 20-mi nute charge/di scharge cycl e for about a month .

The average d i s charge current was �50 mA/cm2 . Due to an i mproper

sea l i ng of the cel l there was l os s of sodi um duri ng the test , thus

expl a i n i ng the l ow coul ombi c effi c i ency . The rea l coul ombi c effi ci ency

i n th i s system s hou l d approach 95% . No concentration pol ari zati on

effects were found i n these studi es . Fl ood i ng of the porous a l umi na

matri x wi th sodi um caused short c i rcu i ts i n the fl owi ng cel l s . Pro­

j ecti ons based on a l ess porous matri x impregnated wi th the mol ten

eutecti c i nd i cate a maximum oev of 0 . 5 V and �700 mA/cm2 at 0 . 25 V .

These cel l s shoul d operate sati sfactori l y i n an el ectrothermal regen­

eration mode [82 ] .

Agruss [81 ] i nvesti gated the abi l i ty of th i s system to undergo

thermal regeneration up to temperatures of �l OOO oe . S i nce the measured

acti v i t i es of Na i n the a l l oy ( from EMF data ) correl ated l i nearly wi th

the i nverse of the absol ute temperature i n the 500-100°8 range , the

acti v i ty data extrapo l ated to the 900- 1 0000e range were used to provi de

estimates of the sodi um vapor pressure over the a l l oy so l uti on ( 1 5-30

mol e % Na ) i n the des i red regeneration temperature range . The EMF can

be rel ated to the acti v i ty of sod i um in the al l oy , wh i ch i s equal to th�

rat i o of sodi um parti a l pressure over the a lloy sol uti on ( p ) and the

parti a l pressure of pure sodi um ( Po ) at that temperature , provi ded

Raoul t ' s l aw i s obeyed : E = - ( RT/n F ) tn ( p/ Po ) ' I t was found that only

above 1 1 000e coul d 200-400 torr sodi um vapor pressure be obta i ned , wh i ch

woul d fac i l i tate the thermal regenerati on . Separate di sti l l ati on

experiments i n the 900- 1 0000e range showed that the di sti l l ati on was

1 0 1

TR-4 16 S=�I 1.1 -----------------------------

-� �

very s l ow . One attempt was made to run i n a regenerati ve mode wi th the

system shown i n F i g . 1 1- 1 5 , u s i ng the cel l Na I Na I - NaCl impregnated

a l umi na l NaxSn at 625-650°C and the regenerator at 1 000°C . The cathode

was fi l l ed wi th 30 mol e % Na/Sn a l l oy , wi th no pure sodi um in the anode

cup·. The regenerator was operated for about 1 0 mi nutes to generate

enough sodi um to s tart the cel l operati on . Power was drawn from the

cel l for about 1 5 mi nutes at 0 . 3 V and 1 00 rnA unti l the meta l - to-

cerami c seal s were corroded by hot sodi um vapor and started to l eak

[81 , 84J . S i nce regenerati on und�r 1 000°C was not feas i bl e , the Al l i son

D i v i s i on of General Motors Corp . started i nvesti gat i ng the potass i um­

mercury sys tem descri bed i n Sect i on 1 1 . 1 . 1 .

1 1 . 2 . 1 . 2 The Sod i um-Lead System [36J

The gal van i c cell Na I NaF-NaCl -Na 1 I Na/b was chosen by Argonne

Nati onal Laboratory [36 J as a pos s i bl e therma l l y regenerabl e system ,

wi th the eutecti c el ectrolyte of compos i ti on 1 5 . 2 : 31 . 6 : 53 . 2 mol e % of

NaF : NaCl : Na I , mol ten at 530°C . Phase di agrams for th i s ternary system

[1 06J as wel l as for b i nary sodi um hal i de el ectrolytes were i nvesti gated

[36J . The ternary eutecti c was chosen for the i nvesti gati on of sodi um­

conta i n i ng cel l s because of i ts l ow mel ti ng po i nt . EMF data for Na l Na+ _

gl ass ( Na20 ) I NaxPb had been determi ned by Hauffe and Vi erk [ � 07J , by

Lantratov [ 1 08 J , and a l so by Porter and Fei n l e i b [1 09J u s i ng as el ectro­

l yte a l umi na impregnated wi th sodi um carbonate . These resul ts agreed

wi th va l ues cal cul ated from the vapor pressure of sodi um over sodi um

l ead a l l oy (40 atom % ) , whi ch are about 8 mV h i gher than the di rect EMF

measurements [36 , 1 1 0J . The vapor pressure data were successful l y

r02

- TR-4 16 S=�I I.I -------------------------------� �

To Vacuum

Metal to Ceramic Seal

Regenerator

Electrodes

Impregnated Porous Cup

Na Liquid

Figure 1 1-15. Thermally Regenerative Sodium-Tin System [81 ]

103

TR-4 16 S=�I I.I -------------�--------------- � ��

analyzed by a q uas i - i deal sol ution treatment (prev i ous ly uti l i zed i n the

analys i s of the sodi um- bi smuth system [ 1 1 5J ) , i n wh i ch the major i nter­

metal l i c spec ies were assumed to be NaPb, Na Pb3 , and Na3Pb [ 1 1 1 J .

The phase d i agram studi es [36J i nd i cate that th i s system does not

present probl ems of overl ap of the l i q u i d/vapor and sol i d/ l i qu i d equi l i ­

bri um reg i ons . In fact , the h i ghest mel ti ng po i nt of any compound i n the

system i s 400 ° C . At th i s temperature the vapor pressure of pure sod i um

i s 0 . 35 torr , and at thi s pres sure the bo i l i ng poi nt of l ead i s �900 °C .

H i gher operati ng pressure ( 6- 1 0 torr ) and h i gh regeneration temperature

(�900° C or hi gher ) shou l d a l l ow regeneration of th i s system .

The EMF of sod i um- l ead cel l s i s 0 . 3-0 . 5 V ( al l oy compos i ti on �1 0-40

atom % sodi um) . One compl ete cel l and regenerator system was operated

for approximate ly 1 00 hours . The apparatus empl oyed i s shown schemati -

cal l y i n Fi g . I I- 1 6 . The cel l was l oaded wi th the ternary el ectrolyte

and wi th 30 : 70 atom % Na : Pb al l oy i n the cathode . The regenerator was

operated to d i st i l l sodi um for the anode , to be consumed i n the cel l .

El even runs of 2- 7 hours , for a total operati ng t ime of 45 hours , wi th

the cel l at 545-600 °C and 5 . 7-9 torr pres sure , were performed . The

el ectrode area was 45 cm2 and the i nterel ectrode separation 1 . 9 cm .

Tabl e 1 1- 1 shows the conti nuous operati ng current obta i ned at three

temperature and pressure cond i t i ons of the regenerator .

Tabl e 1 1- 1 . OPERAT ION OF THE Na/ Pb THERMALLY REGENERATIVE SYSTEM [36 J

Regenerator Regenerator Cel l Temperature ( O C ) Pressure ( torr ) Current (A )

875 8 7 . 5

850 7 5 . 0

825 6 2 . 5

104

- TR-4 16 S=�I 'I_J -------------------------------

Ceramic Spacer Lead

vacuum-Pressure �=� .ft ..... Baffles

Li ne c.

Vacuum­Pressu re , Line

Sod ium

Electrical Heaters

Sti l l

Heat Exchanger

Figure 1 1-1 6. Thermally Re.generative -Sodium-Lead System Operated fQr 1 00 Hours [36]

.

105

TR-416 S=�I I�.��I --------------------------� � = '/'

Ce l l OCV a s h i gh as 0 . 41 V at 575°C were recorded , i nd i cat i ng that

the regenerator had reduced the cathode sodi um concentrati on from 30 to

1 8 atom % . One examp l e of the cel l performance at th i s temperature i s

s hown i n F i g . I I- 1 7 . An OCV o f 0 . 39 V and a current den s i ty of 1 00

mA/cm2 at 0 . 1 8 V were obta i ned wi th the regenerator at 875 °C and 8 torr .

The regenerator shown i n F i g . 1 1- 1 6 l ed to errati c sodi um- l ead c i r­

cul ati on . S i nce the pressure of the system was shown to be an important

parameter i n the l i q u i d c i rcul at ion , a better regeneration sys tem was

des i gned , provi d i n g a hydrostat i c head to improve ci rcul ati on ( see Fi g .

I I- 1 8 ) . No major probl ems of corrosi on were encountered i n approxi ­

mately 1 000 hours . The i n i t i a l l i qui d c i rcul ation ( fi rst 3 days ) was

erratic but improved wi th better degass i ng of the system .

The di fferent ia l den s i ty cel l ( cf. F i g . 1 1 - 1 5 ) was cons i dered as an

i n i ti a l des i gn but cel l s wi th immobi l i zed el ectrolytes ( cf . Secti ons

1 1 . 1 . 1 . , 1 1 . 1 . 2 . , and 1 1 . 2 . 1 . 1 ) were j udged superior . Several des i gn

concepts were proposed , i nc l ud i ng des i gns for mul t i ce l l operation . Due

to the l ow vol tages obta i ned per cel l , a practi ca l devi ce woul d have to

connect many cel l s i n series to ach i eve useful vol tages . The effi c i ency

of th i s system shou l d be 9- 1 2% . Therefore , the Na/Pb system was not

cons i dered attracti ve for a practi cal dev i ce by Argonne Nati onal Labora�ory

i nvesti gators [36J .

1 1 . 2 . 1 . 3 The Sod i um-B i smuth System

The cel l Na I NaF-NaCl -Na I I NaxB i was al so i nvesti gated at Argonne

Nati ona l Laboratory [36J with the ternary 'eutecti c el ectrolyte descri bed

106

- TR-4 16 S::�I I�.-�I --------------------�-------------------------------------------� � ��

0.3

� Q) '" � 0.2 o > �

0.1

O ������ __ �� o 50 100 giO 200 250 Current Density (rnA/em')

Figure 1 1-1 7. VOltage-Current Density Plot for the Sodium-Lead Galvanic Cell Operating under Thermal Regeneration [36]

Partition

Na-Pb Alloy

'----'I1--I!'- Sodium Liquid Return Line

Sti l l

Figure 1 1-1 8. Improved Thermal Regenerator for the Sodium-Lead System Operated Continuously for 1000 Hours [36]

107

TR-4 16 S=�I I*I -----------------------------

for the Na/Pb system ( Secti on I I . 2 . 1 . 2 ) . The sodi um-b i smuth cel l has an

EMF i n the 0 . 55-0 . 75 V range , ca . 0 . 2 V h i gher than the Na/Pb cel l s ,

thus promi s i ng better performance . The EMF decreases wi th i ncreased

sod i um content i n the cathode but between 20-40 atom % sodi um the

decrease i s s l ow ( approximate ly 0 . 6-0 . 5 V ) . At about 55 atom % sodi um,

a sodi um-saturated sol ution is found and the EMF is approximately 0 . 4 V

[36 , 46J . A c l osed cel l of the di fferent i a l den s i ty type , s i mi l ar to the

bottom hal f of that empl oyed i n the study of the Na/ Pb system us i ng the

sti l l ( see Fi g . 1 I - 16 ) as a s i mpl e reservoi r for chargi ng sod i um and

b i smuth , was operated i n charge and di scharge cycl es for a period of 1 7-

1 8 months at 550°C wi thout deteri orat ion of performance . The OCV of

th i s cel l was 0 . 7 V ( 20 atom % sodi um i n the cathode ) and current

dens i ti es on d i scharge of 90 and 1 1 0 mA/cm2 were obta i ned at 0 . 5 and

0 . 45 V, res pecti vel y . The el ectrode area was 45 cm2 and the ca l cul ated

i nterna l res i stance was 0 . 05 ohm . No decompos i ti on of the mol ten sa l t

was observed and mi n imum corros i on was detected after th i s peri od . An

improved cel l des i gn us i ng a sodi um reta i ner/current col l ector i n the

form of a sta i n l ess steel s p i ral , mounted i n the anode compartment ,

mai nta i ned a uni form current di stri bution throughout the anode area .

Th i s cel l coul d be charged and d i scharged at currents up to 50 A ( l . l

A/cm2 ) over 50 t imes and at var ious temperatures , wi th no deteriorati on .

The combi ned charge/di scharge effi c i ency was approxi mately 80% for a

current dens i ty of 665 mA/cm2 at 565°C [36 , 1 1 2 J .

One drawback of these cel l s i s the rel ati ve ly fast sel f-di scharge ,

associ ated partly wi th the l a rge so l u bi l i ty of the i ntermetal l i c com­

pounds ( e . g . , Na3B i ) i n the mol ten sa l t system, wh i ch i ncreases wi th

108

temperature [ 1 1 3J . Foster [1 1 4 J presents an i nteresti ng account of the

b imetal l i c cel l s wi th regard to the sol ubi l i ti es of the i ntermetal l i c

compounds and spec i es characterization i n the mol ten sa l t medi a .

Thermodynami c and phase di agram studi es o f the sodi um-b i smuth

system were performed [ 1 1 5 , 1 1 6J . In th i s system , the sol i d/ l i qu i d and

l i q u i d/ vapor equi l i br i um regions overl ap at l ow pressures . The total

pressure curves i nd i cate the appearance of a three- phase equi l i bri um

( Na3B i sol i d , of mel t i ng po i nt 842 ° C , i n equi l i bri um wi th l i qui d and

vapor) bel ow a pressure of approximately 240 torr . Therefore , the

operati onal pressure of a regenerati ve sodi um- bi smuth system must be

>240 torr . I n order to col l ect pure sodi um at th i s press ure , a conden-

ser temperature of approximately 770°C i s requ i red , wh i ch wou l d ra ise

the gal van i c cel l o perati ng temperature about 270°C . Another conse­

quence of operati ng at 240 torr pressure i s that i n order to obtai n a

reasonabl e cathode compos i ti on , the regeneration temperature shou l d be

1 200- 1 300°C . At th i s temperature , mater ia l s probl ems and dynami c cor­

ros i on by Na/Bi cou l d be very d i ffi cul t to overcome . The vapor obtai ned

woul d sti l l conta i n severa l atom % Bi [ 1 1 5 J , thus neces s i tati ng frac­

t i onati on ( refl uxi ng ) , wh i ch reduces the overa l l effi ci ency .

1 1 . 2 . 2 L i th i um-Conta i n i ng Systems

The systems Li /Sn [1 1 7 J , L i /B i [ 1 1 8J , L i /Te [1 1 9J , L i /Cd [36J ,

L i /Zn [36 J , and L i / Pb [36J were i nvesti gated at Argonne Nati onal Labora­

tory i n the 1 96 1 - 1 967 per iod , from the el ectr i ca l ly or therma l l y regen­

erati ve poi nt of v i ew [36J . EMF data for these systems [36 , 1 1 7- 1 1 9J

were obta i ned u s i ng as an el ectrolyte the bi nary eutect ic L i F-L i Cl or

109

TR-4 16 S=�I I.I ----------------------------� �

Li C l - KCl mol ten sa l ts at approximately 500°C . F i gure 1 1- 1 9 shows the

EMF-temperature-compos i ti on pl ot obtai ned for the Li/Sn system ; h i gher

EMFs were obta i ned for the Bi and Te systems . For therma l ly regenera­

ti ve operat i on , both the L i /B i and L i /Te systems were found i nadequate

due to the h i gh vapor pressure of B i and Te over the i r respecti ve l i th i um

al l oys . I n the L i /B i systems , the same probl ems encountered for Na/B i

that resul ted from the overl ap i n the phase di agrams can be foreseen

[36 , 1 1 6 J . The compound Li 3B i has a mel ti ng po i nt of approxi mately

1 1 50°C [1 1 6 J . The only attracti ve system for thermal l y regenerat ive

operati on i s the L i /Sn system , due to the very l ow ti n vapor pressures

over l it h i um- ti n a l l oys even at very h i gh temperatures ( e . g . , 1 200 °C )

[ 1 20J .

The phase d i agram for the Li /Sn system shows a reg ion of l i qu i d/

s ol i d coexi stence i nc l udi ng the compound L i 7Sn2 wi th a mel ti ng poi nt of

783°C [36 J . The l i qu i d/vapor d i agram at 1 200 °C is shown in Fi g . 1 1-20 .

At th i s temperature the sol i d/ l i q u i d and l i q u i d/vapor equi l i bri um

regi ons do not overl ap . The l i th i um vapor pressure over a reasonabl e

cathode compos i tion ( e . g . , 30 atom % l i th i um ) i s of the order of 2 torr .

The regeneration at th i s l ow pressure ( cf . 6- 1 0 torr for the Na/ Pb

system ; see Section 1 1 . 2 . 1 . 2 ) may pose probl ems of heat , mass , and

momentum transfer, i n add i ti on to the materi al s and corros i on probl ems

of operati ng the therma l regeneration at 1 200° C . However , s i nce the

cel l coul d operate wi th a ternary eutect i c L i Cl - L i F-L i I [ 1 06J of mel ti ng

po i nt <350° C , the expected Carnot effi ci ency for thi s system cou l d be

58% , and a much hi gher net effi c i ency coul d be expected (25-30%) [4J i f

the rna teri a 1 s probl ems can be. overcome .

n o

TR-416 !;::�I I�I------------------�-----------------------------------------------

o.8or----.-------.------.---.....:.----r-1

8 Alloy Saturated with LisSn2(S)

0.70 � Xu=0.1

� O.SO . . . " . .

:§ 0-C::::>-<:J-0:)--(:r.--,", "E � o a. � 0.50 t)

0.40

� XLi=0.3

�4

" . .... . .. . � . . " . ..... ' .

XLi=0�SU=0.5

. Xu=0.S5 0.30,'---'-____ -----L _____ -'--__ -'--'

800 900 1000 1 100 Temperature (K)

Figure 1 1-19. EMF-Temperature-Composition Characteristics of the Cell Li (f)/LiCI-UFCP)/Lij(Sn [36] XLi = atom fraction l i t h i u m

Atom % i n Liquid

500 0 10 20 30 40 50 SO 70 80 90 100

200

100

-;:: 50

g 20

� 10 UJ gJ 5 a: iii 2

;§ 1

0.5

0.2 0. 1 '--'----'-�'-:-_:':_---'::-____J�-'---'-----'�

10-8 10-7 10 -6 10-5 10-4 1 0-' 1 0-' 10-' 10. 10' 10' Atom % Tin in Vapor

Figure 1 1-20. Pressure-Composition Diagram for the U-Sn System at 1 2000 C [1 20]

i n

S=�I I.I ______________________ T_R_-_4_16_ -� �

The other b imeta l l i c systems wi th l i th i um anodes were i nvesti gated

as poss i bl e secondary batteri es . Fi gure 1 1 -2 1 shows exampl es of the

d i scharge characteri sti cs of these systems [36J . Most of these systems

produced h i gh current dens i ti es at reasonabl e vol tages ( espec i a l ly L i /Te

and L i /Se ) and exh i bi ted good cycl i ng capabi l i ti es . Due to these charac-

teri sti cs , the subsequent i nvesti gati ons at Argonne conti nued to expl ore

the secondary battery and el ectri cal energy storage appl i cati on of h i gh

temperature gal van i c cel l s wi th L i or L i /Al anodes us i ng immobi l i zed

e l ectrodes wi th mol ten sal t el ectrolytes .

1 1 . 3 SUMMARY OF THE PERFORMANCE AND D ISCUSS ION OF THERMALLY REGENERATIVE ALLOYS OR B IMETALL IC SYSTEMS

Tabl e S-5 assembl es resul ts for the a l l oy and b imetal l i c systems

reported i n our rev i ew . It i s fai r to state that the el ectrochemical

cel l performances l i sted in the tabl e are l ow l i mi ts . These systems ex-

h i bi ted better cel l performance in batch cel l s tested . They are a l so

good storage battery systems . The coup l i ng of the el ectrochemi cal cel l

a nd regenerator system were successful i n the Na l Hg and Na l Pb systems .

The operati ng temperatures were very h i gh because of the appl i cati on

envi s i o ned at the t ime . One can safely state that these systems were -

cl oser to s uccess i n the thermal regeneration mode than those descri bed

i n Secs . 1 . 1 and 1 . 2 , when compared for regeneration i n the >500 DC

temperature range . S i nce the cel l reacti ons C t C+ + e- are revers i bl e ,

the systems of h i gh coul omb i c effi c i ency ( e . g . , Na l Hg and Na l Pb ) shoul d

be s u i tabl e for operati on i n the coupl ed thermal and el ectro lyt i c re­

generati on mode . A more deta i l ed i ntegra l systems ana lys i s coul d be

1 12

...... ' 1-" �

lO r, --------------------, O.BO r, -----------------,

O.9P>-", ... 0.8

0.7 :;­; 0.6 Ol '" '6 O.S > � 004

0.3

0.2

... ... , .... u ........ "',

li(l)/liCI-KCI/li in Bi(l)

... '0... .................. ...... ...

...... ............ ... ... ,

.... 0 ........ ...... "" ,

0.72 li(J)/liCI-KCI/Li in Cd(.l)

0.6'P>-o.. ... "

� 0.S6 ',0.. Q) � 0048

g 0040

...... ... .... "00 ... ... .... ... ... ' ... Q; u 0.32

.... .... "-

"'" 0.24

0 . 16 '\ ,

0.72r' --------------,

0.64 1-'\ 0.S6

� OAB Q) Cl )!! 0040 o > _ 0.32 Q) u

0.24

0.16

, , , , , Li(J)lLiCI-KCI/Li in Pb(J)

, "-, , , ", , , ,

""- \,

.... ... "'

... ... ... .... " , ,

0.081 I , b 0.1 1 I I o 100 200 300 400 SOO 600 700 800

2.0 00oQ. "" -

1 .6

1 .2

O.B

004

"'-...

Anode Current Density (mA/cm')

(a) Lithium-Bismuth, T = 489 :±. 2° C

li (/)/liCI-KCI/Li in Te(l)

.... .... 1" .... .... .... ... ... 0, ... 0.. ... "' 0.. ... ... ...

'Q. '0... .... _0...

0 1 � o SO 100 1S0 200 2S0 300 3S0 400 4S0 SOO SSO 600 Anode Current Density (mNcm')

(<!) Lithium-Tellurium, T = 496 ± 5° C

Figure 1 1":21 .

0.08 'L--L_.L----L_..L.-..l_-'-_L---'-_..L.-..l_-'----' o SO 100 1 S0 200 2S0 300 3S0 400 4S0 SOO SSO Anode Current Density (mA/cm')

o SO 100 1S0 200 250 300 3S0 400 4S0 Anode Current Density (mA/cm')

(b) Lithium-Cadmium, T = 493 ± 20° C (c) Lithium-Lead, T = 483 ± 9° C

O.BOhr:------------------------, 0.72 1, ------------

0.64t 0.S6 \"

� 0048 Q) � 0040

g 0.32

� 0.24

0 . 16

0.08

, '",

li(J)/liCI-KCI/Li in Zn(/)

, , , , " , , , , , '0 , , \ , � I � O 'L-���_L-���_L-�_-'-�

b. ... o, 0.72 f- ' ... ... , 0.64

� 0.S6 f--Q) g> '6 OA8 f­> � 040

0.32 f-

0.24 f-

"" ... -0

......

Li(/)/LiCI-KCI/Li in Sn (i)

' ... " 'a. ' ... " .... ....... .... ... ....

.... ... ',,-

0.16 1 I I I I I I I I

"" " "" .... ........ b ...

o SO 100 1 S0 200 2S0 300 3S0 400 o SO 100 1S0 200 2S0 300 3S0 400 4S0 SOO Anode Current Density (mNcm') Anode Current Density (mA/cm')

(e) lithium-Zinc, T = 486 ± 6° C (I ) Lithium-Tin, T = 496 ± 2° C

Voltage vs. Current Density of the L ithium-Containing Bimetallic Cells with tile LiCI-KCI Electrolyte [36]

-

III III N -I.' •

� oj:>. I-' Q')

/. � , TR-4 16 S=�I I_I _________________________ ----'O.;.:.-"'-"=_

performed to s uggest s u i tabl e systems to rei nvesti gate once a sol ar­

deri ved , h i gh temperature source i s i dent i fi ed . Th i s ana lys i s i s par­

t icu l arly i mportant i n v i ew of the l arge quanti t i es of materi a l s pumped .

At th i s poi nt i t shou l d be emphas i zed that past i nves t i gati on of

concentration cel l s for power generati on purposes in med i a other than

mol ten sal ts i s rather l i mi ted . The reasons are the general l y s l ow

el ectrode ki net i cs i n other medi a and h i gh ohmi c l osses due to l ower

conducti v i ty . Mo l ten sal ts wi th l ower mel ti ng poi nts and other medi a

coul d be the bas i s for the i nvesti gati on o f other thermal or coupl ed

thermal and e l ectro lyti c regeneration systems .

1 14

- TR-4 16 S=�I I.I --------------------------� �

SECT ION I I I

THERMOGALVAN I C OR NON I SOTHERMAL CELLS

Thermogal van i c cel l s can be defi ned as ga l van i c cel l s i n wh ich the

temperature i s not un i form . They are the el ectrochemi cal equ i val ent of

thermoel ectri c dev i ces , whi ch convert heat i nto el ectri c i ty . I n these

cel l s two or more el ectrodes are at di fferent temperatures . These el ec-

trodes , not necessari l y chemi cal l y i denti cal or rever� i bl e , are i n

contact wi th a n e l ectrolyte , sol i d o r l i qui d , not necessari l y homogeneous

i n compos i ti o n , and wi th or wi thout permeabl e membranes i nterposed i n

the el ectrolyte . Duri ng the passage of current through the thermogal -

van i c cel l s , matter i s transferred from one el ectrode to the other as a

resul t of the el ectrochem ica l reactions at the el ectrode/el ectrolyte

i nterface and i on i c transport i n the el ectrolyte . In th i s respect , the

thermogal van i c cel l d i ffers from metal l i c thermocoupl es , or th�rmoel ectri c

devi ces i n general , i n wh i ch no net transfer of materi a l occurs and the

state of the conductor rema i ns unchanged wi th the passage of current .

In fact , thermoe 1 ectri c effects i n the meta 1 1 i c 1 eads from the el ectrodes

i n the thermogal van i c cel l s contri bute to the observed EMF of these

cel l s [ 1 21 - 1 25 J . The names " thermal cel l , " " thermocel l , " "e l ectro-

chemi cal thermocoupl e , " " noni sothermal cell , " " ga l van i c thermocoupl es , "

a nd " ga l van i c thermocel l s " have been empl oyed for thermogal vani c cel l s ,

and express i ons such as " thermogal van i c energy convers ion " and " i on i c

thermoel ectri c convers i on" are used . These cel l s are des i gnated as

Type 6 in the I ntroduct ion .

Thermocel l s have been i nvesti gated s i nce 1 879 ( see Ref . 1 2 1 ) and

1 15

TR-416 S=�I I*I -----------------------------

have been the subj ect of several rev i ews ; e . g . , deBethune , L i cht , and

Swendeman [ 1 21 J ; deBethune [ 1 22 J ; Agar [1 23J ; and Wagner [1 24J . Much of

the work has i nvol ved studi es of the temperature dependence of reference

el ectrodes i n aqueous sol ution [1 21 , 1 23 , 1 23 J and a l so i n mol ten sa l t

medi a and sol i d el ectrolytes [ 1 23 , 1 23 , 1 24 J . The research has focused on

thermodynami c aspects , pri nci pal l y on the appl i cation of i rrevers i bl e

thermodynami cs to these systems . Some papers have a l so deal t wi th more

practi cal appl i cations [ 1 25J , e . g . , thermoga1 van i c corros i on [1 21 J and

power generation [ 1 25 J . Sundhe im [ 1 26J cons i dered mol ten sal t thermo­

cel l s for spec i fi c power generation uses i n 1 960 and Chr'i sty [1 27]

exami ned the poss i bi l i ty of emp l oying i on ic so l i ds in thermoel ectr ic

devi ces .

The EMF of a thermoga1 vani c cel l i n i ts i n i ti a l state ari ses from

three factors : ( 1 ) the d i fferences i n el ectrode temperature , ( 2 ) the

therma l l i qui d j uncti on potent i al , and ( 3 ) the metall i c thermocoupl e

effect . I n general , the EMF ari s i ng from ( 1 ) and ( 2 ) i s about two

orders of magni tude l arger than the EMF from ( 3 ) , wh i ch ari ses at the

j uncti on i n the external c i rcu i t between two el ectrode metal s at di ffer­

ent temperatures ( the Seebeck effect ) . Wi th the passage of time , a

therma l cel l i s subj ect to thermal di ffus ion i n the el ectrolyte (Soret ­

effect ) , wh i ch tends to concentrate the el ectrolytes I n the col d reg ion .

The concentrati on gradi ent further changes the two el ectrode potenti a l s ,

and the cel l reaches a new stati onary state ( f i na l EMF ) . The formation

of the concentrati on gradient can be avo i ded by sti rr i ng or convecti on

( as l ong as the thermal gradi ent i s not destroyed ) and is not present i n

cel l s wi th sol i d el ectrol ytes i n wh i ch on ly one k i nd of i on i s mobi l e

[ 1 28 J .

1 16

S=�I I_I _________________________ --=T=R.:,...--=4-'-lS-'--

The thermogal van i c cel l can be wri tten as :

where T2 > Tl . The el ectrodes can be meta l s or gases wi th i nert el ec­

trodes [ 1 2l , 1 22 J . The el ectrolyte can be an aqueous sol ution , a fused

sa l t , or a sol i d . Temperature Tl i s fi xed and T2 i s vari ed . The most

wi dely used s i gn convent ion for thermogal van i c cel l s is that the EMF ( E )

i s pos i t i ve when the termi nal connected to the el ectrode a t T2 i s

pos i ti ve wi th respect to that connected to the col d el ectrode . There­

fore , the hot e l ectrode i s the cathode , and the ( dE/dT2 )T constant i s 1 .

pos i ti ve . Th i s coeffi c i ent , ( d E/dT ) thermal ' i s the thermoel ectri c

power , someti mes a l so des i gnated the I I Seebeck coeffi c i ent ll i n ana l ogy

wi th the nomencl ature used i n thermoel ectri c phenomena [1 29 J . The

thermoel ectri c power i s obtai ned from measurements at open ci rcu it . � · \

1 ( 1 = 0 ) .

The · thermoel ectr i c power can be descri bed as the sum of a hetero­

geneous term ( due to the el ectrode temperature effect ) and a homogeneous

part ( thermal l i qui d j unction potent ia l for so l utions or the thermo­

e l ectri c effect on sol i d or l i qu i d i o n i c conductors ) [ 1 25 , 1 27 , 1 30 J . The

dri v i ng force for the thermogal van i c cel l i s the transport of entropy -

from the h i gh temperature reservo i r ( at T2 ) to the l ow temperature s i n k

( a t Tl ) , a i i s the case for any heat engi ne . References 1 2 1 - 1 23 and 1 31

g i ve thermodynami c treatments of thermoga l van i c cel l s and c i te the

various important papers i n th i s fi el d . References 1 21 and 1 22 are

part icu l arly ori ented to aqueous sol uti ons and g i ve a comprehen s i ve

treatment of the subj ect [1 22J , as wel l as val ues for ( d E/dT) th 1 and erma

1 1 7

S=�I I.I _ _____________________ T.:...:R:..:..--=4=16=_ -� �

( dE/dT ) . th 1 [for cel l s SHE l l el ectrolyte l el ectrode , where SHE i s 1 S0 erma '

the standard hydrogen el ectrode and ( dE/dT ) i sothermal i s the deri vati ve

dE/dT of the EMF ( E ) of the i sotherma l cel l J for ca . 300 el ectrodes at

25°C [1 21 J . Reference 1 23 presents a comprehens i ve treatment of the

subject for a l l three med i a and i n parti cul ar i nc l udes a rev i ew of the

confusi ng thermodynami c nomencl ature encountered i n th i s subject . Re-

ference 1 24 revi ews the l i terature up to 1 972 on the thermoel ectr ic

power of i on i c sol i ds and mel ts . In the fol l owi ng bri ef descr ipti on of

the thermogal van i c cel l s , the nomenc l ature of Agar and Breck [1 32J i s

used .

For a thermocel l , for i nstance wi th pure metal el ectrodes and a

s imp l e el ectrolyte MXn , sol i d or fused , the EMF of the cel l i s the

e l ectri cal potent ia l of a wi re attached to the hot el ectrode mi nus the

potent ia l of a s imi l ar wi re attached to the col d el ectrode . The el ec­

tri cal work for n equi val ents of el ectri c i ty is determi ned by the

entropy absorbed from the heat reservo i r surroundi ng the hot el ectrode

when pos i t i ve e l ectri c i ty passes through the cel l from the col d to the

hot e l ectrode [ 1 31 J . Th i s entropy i s i denti cal to the sum of the

entropy absorbed i n the e l ectrode reaction [ i n th i s case , SM (mol a l

entropy of the metal ) - St.1n+( part ia l mol ar entropy of Mn+ ) - nSe- ( M ) ( parti al mol a l entropy of the el ectron i n M ) J and the entropy trans-

ported away from the hot el ectrode [ i n th i s case , -SMn+ ( entropy of

transfer of Mn+ ) - nS�- ( M ) ( entropy of transfer of e- ) ] [ 1 23 , 1 3 1 ] . The

entrop ies of transfer resul t from heat effects attributabl e to the

movement of el ectrons and ions through a thermal gradi ent under the

i nfl uence of � vol tage drop . S i nce S + S* = S ( tota l transported entropy ) ,

1 18

TR-4 16 S=�I I.I ----------------------------

·one can wri te

'nF .9I = S dT M I I I - l

Express i on I I I- l a l so hol ds for aqueous thermoce l l s after the Soret

equi l i bri um i s reached ( fi na l EMF ) [ 1 21 J . For the i n i ti a l EMF of such

aqueous thermoce l l s ( uni form e l ectrolyte di s tri buti on ) the term

t_ ( SMn+ + nSX- ) ' where t i s the transference number of the an i on , must

be added to Eq . I I I - l . Thi s sum of i on i c entropi es of transfer governs

the Soret equi l i bri um. I n a pure sal t ( sol i d or l i qui d ) the

transfer of both i ons i n the same di rection i s the gross l i near movement

of the sal t , wi th no net entropy of transfer : ( SMn+ + nSX- ) pure sal t = 0 [ 1 3 1 J . The entropi es of transfer for meta l ions i n fused sal ts are

s uffi ci ently smal l that the express i on

. nF ( dE ) = - -'- dT I=O

llSllT=O

I I I -2

can be deduced from Eq . I I I - l if the transfer terms are negl ected (as -

has been found experimental ly for several mol ten sa l t thermocel l s wi th

pure i o n i c fu'sed sal ts ) . The term ( dE/dT ) therma l i s then equal to

( dE/dT ) i sothermal '

The thermoel ectri c power , ( d E/dT) thermal ' measured at I = 0 or

cal cul ated from the appropri ate equati on has been used to ca l cul ate the

fi gure of meri t , Z, of the thermocel l [ 1 21 - 1 23 , 1 3 1 J , i n anal ogy to that

used for thermoel ectri c devi ces [ 1 33J :

1 19

- TR-4 16 S=�I I.I ----------------------------. - �

Z =

. 2 ( dE/dT ) I=O PK I I I -3

where P = spec i fi c res i st i v i ty in ohm cm , K i s the spec i fi c thermal

conducti v i ty i n W cm- l K- l , and dE/dT i s i n V K- l . Ion i c conductors

were found to have fi gures of meri t of approximately 1 0- 3 K- l , wh i ch are

of the order of magn i tude of the semi conducti ng thermoel ectri c dev i ces

[ 1 25 , 1 28 , 1 33J . The convers i on effi ci enc i es are Garnot cycl e l i mi ted for

both the thermoel ectri c and thermogal van i c cel l s . However , ant i c i pated

practi cal fi gures of meri t wou l d be cons i stently l ess than those obta i ned

from I = O .

The express i on o f Tel kes [1 34J for sol i d-state dev i ce effi ci enc i es

n has been appl i ed to thermogal van i c cel l s :

1 n = -----'------� x 1 00% I I I-4

where the fi rst term in the denomi nator is rel ated to the Garnot effi -

c i ency and the second i s rel ated to the fi gure of meri t . Wartanowi tz

[ 1 35J has expressed the effi ci ency of mol ten sal t thermogal van ic cel l s -

as

n = (M+m+l ) [A(M+m+l ) (B/m+l )+l ] - G (m+l ) I I I -5

where nc = Garnot effi c i ency ( T2 - T, /T2 ) ; A = 1 / ZT2 ; B = ( T� - T� ) ( L ) ( Z ) /

1 20

- TR-416 S=�I IWI ---------------------------=:...=..:........::..::..::...... -� ���

m = el ectrode res i s -

tance/ o mol ten sal t res i stance ; and M = l oad res i stance/mo l te� sa l t

res i stance . Wartanowi cz ' s [ 1 36J experi mental resul ts wi th a Ag 1 AgN03 ( t ) l Ag

cel l agree wi th Eq . 1 1 1- 5 .

The analyses of Z and n above do not cons i der el ectrode pol ari za­

t ion effects , wh i ch effecti vel y l imi t the power output of such devi ces

under current dra i n .

Zi to [ 1 25 J compares thermoel ectr ic dev i ces wi th the thermogal van ic

cel l s , based mostly on mol ten sal ts or sol i d el ectrolytes , s i nce the

wi der temperature range of these materia l s a l l ows hi gher Carnot effi c-

i enci es to be obtai ned . Th i s paper al so i ncl udes perti nent comments

about materi a l s problems associ ated wi th these hi gher-temperature cel l s ,

fabri cation techn i ques , and cel l construct ion .

S i nce the century-ol d l i terature on thermogal van i c cel l s i s very

extens ive , our revi ew i s not comprehens i ve but covers most of the

thermocel l s studi ed for power generati on i n mol ten sal t , sol i d el ec-

tro lyte , and aqueous med i a , as wel l as the more recent papers in the

fi el d . References 1 2 1 - 1 24 conta i n most l i terature c i tati ons pri or to

approxi mate ly 1 970 concern i ng pri nc i pal l y the thermodynami cs of i r-

revers i bl e processes , temperature effects on reference el ectrodes , and -

thermoel ectri c powers of sol i d el ectrolytes and mel ts .

I I I . 1 MOLTEN SALT THERMOGALVAN I C CELLS

1 1 1 . 1 . 1 So l i d or L iqu i d El ectrodes

The thermocel l Ag i AgN03 1 Ag has been known s i nce 1 890 [Po i ncare ;

see Ref . l 37 J . S i nce 1 950 th i s system has been the subj ect o f several

1 2 1

/. ; � TR-416 S=�I II.II ------------------:----------..::...=.::........:..::..=.... - � ��

i nvesti gati ons [ 1 31 J . Sundhe im and Rosenstre i n [1 37J measured i n i ti a l

thermoel ectri c powers for thi s system i n 1 95 3 . Other determi nati ons

have been made over a w ide range of temperatures wi th pure AgN03 [ 1 37 , 1 38J

or wi th mol ten n i trate mi xtures [ 1 39 , 1 40 , 1 41 , 1 42 J ; some resu l ts are

gi ven i n Tabl e I I I- l . The measurements of Sundhei m et a 1 . [ 1 39J i n

AgN03 : NaN03 were repeated by Haase et a 1 . [ 1 40J , who al so measured the

AgN03 : L i N03 system . Haase and co-workers extended the treatment of the

EMF of . the thermoce1 1 s [ 1 43 J to bi nary mel ts and deri ved new genera l

rel ations for the thermoel ectr i c power of a thermoce1 1 cons i st i ng of a

two-component i on i c mel t ( three i on consti tuents ) and two s i mi l ar el ec­

trodes revers i bl e to one of the i on consti tuents [1 40J . The rel ationsh i p

between the thermoel ectr i c power and the Soret coeffi c i ent was deri ved .

The transport q uanti t i es rel evant to therma l d iffus i on and rel ated

phenomena i n the mel t are the transported entrop ies of the two cati ons

and a l i near combi nation of the heats of transport of these cati ons .

The experimenta l work i nc l uded new measurements of thermoel ectri c powers ,

transport numbers , and acti vi ty coeffi c i ents . Transported entrop i es , the

d i fference between the two heats of transport , and the Soret coeffi ci ent

were ca l cul ated from these measurements .

Connan and Dupuy [ 1 44 J have measured the thermoel ectri c powers of

mi xtures of AgN03 : MN03 (M = Li , K, Rb , T1 ) ; i n i ti a l thermoe l ectri c

powers of 0 . 38 , 0 . 25 , 0 . 30 , 0 . 28 mVjdegree at the 0 . 5 : 0 . 5 mol e fracti on

compos i ti on were found for these mi xtures . The dependence of the

i n i ti a l thermoel ectri c power on the mi xture compos i tj on was i nvesti gated .

Abraham and Gauth i er [1 41 , 1 42 J extended stud i es on mi xtures of

bi nary mol ten sal ts ( AgN03 : T1 N03 ) , wh i ch gave l ow thermoel ectri c powers

122

TR-4 16 !;=�I �III ----------------------------------------------------� ��

Tabl e 1 I I- l . SUMMARY OF THERMOELECTRI C POWERS IN Ag l MOLTEN SALT I Ag THERMOCELLS

�40 l ten Sa I t Temperature ( O C )

I n i ti a l ( dE/dT) 1=0

(mV/degree )

Steady-State ( dE/ dT) 1=0 (mV/degree )

Ni trates

AgN03 305 -0 . 344 AgN03 250 - 0 . 32 to - 0 . 34 AgN03 31 0 -0 . 31 9 AgN03 : NaN03 ( 0 . 5 : 0 . 5 ) a 3 1 0 - 0 . 331 -0 . 248 AgN03 : NaN03 ( 0 . 05 : 0 . 95 ) a 3 1 0 -0 . 41 9 -0 . 1 49 AgN03 : NaN03 ( 0 . 5 : 0 . 5 ) a 31 0 - 0 . 328 -0 . 354 AgN03 : L i N03 ( 0 . 1 : 0 . 9 )a 31 0 -0 . 496 -0 . 549 AgN03 : L i N03 ( 0 . 5 : 0 . 5 ) a 31 0 -0 . 379 -0 . 405 AgN03 : T1 N03 ( 0 . 5 : 0 . 5 ) a 9.0 , 1 80 -0 . 221 , -0 . 254 A9N03 : ( Cd , Na , K ) N03

b 1 20 -0 . 820 ( 0 . 001 : 0 . 999 ) a

Hal i des AgC1 ' - 0 . 375 AgC1 -0 . 40 AgC1 -0 . 42 AgC1 627 -0 . 38 AgBr 477 -0 . 45 Ag 1 577 -0 . 43 Ag 1 577 -0 . 50

Sul fate A92S04 -0 . 31

a Mol e fracti o n compos i ti o n . b Eutecti c compos i ti on : 0 . 460 : 0 . 394 : 0 . 1 46 .

1 2 3

Temperature Range ( O C ) Referenc

240-31 0 1 37 225- 31 0 1 38

1 39 1 39 1 39

280-340 1 40 260- 340 1 40 240- 340 1 40

1 42

1 42

500- 900 1 46 450- 650 1 47 487-590 1 48

1 49 1 48 1 3 1

- 1 48

657- 750 1 50

S=�I I.I ______________ � _________ _"'T_"R"_-"'74 ........... 16 -� ��

( see Tabl e I I I- l ) , to ternary mol ten sal ts based on Cd ( N03 ) 2 and two

other n i trates of L i , Na , K, Cs , or T1 , at eutecti c and noneutecti c

compos i ti ons . The temperature of operation i n these me l ts was 84- 1 20°C .

Thermoel ectri c powers as hi gh as 0 . 82 mVjdegree at much l ower tempera­

tures than those l i sted i n Tabl e I 1 I - l were obta i ned . The mel ts i n-

vesti gated form g l asses a nd i n the supercool ed reg ion the temperature

dependence of the thermopotent ia l i s represented by a strai ght l i ne of

d i fferent s l ope than that of the normal l i qu i d range . The authors

proposed measurements on thermocel l s as a means of studying supercool i ng

and nucl eat ion [ 1 45 J . The presence of metal ions such as Cd2+ and T1 +

l eads to mi xed metal depos i tion at the cathodes ; these cel l s are not

su i tabl e for power generati on .

The mol ten sal t thermocel l s Ag I AgX ( t ) I Ag ( X- = Cl - , Br- , I - ) were

i nvesti gated by Markov [1 48J , Senderoff and Bretz [1 46J , Ho l tan [1 47J ,

and Anderson et a l . [ 1 49J . The thermoel ectr i c powers for the hal i de

thermocel l s are a l so shown i n Tabl e I I I- l , as i s the val ue for the

mol ten cel l Ag I A92S04 ( t ) l Ag .

Tabl e I I I-2 s ummari zes the thermoel ectric powers of other metal 1

mol ten sa l t l metal cel l s studi ed . Most of the l ead cel l s had thermo­

e l ectr i c powers of a few mi crovo l ts ( see Tabl e I I I- 2 ) . Deti g and

Archer [1 30J used tungsten - l ead contacts whereas Anderson et a l e [ 1 49 ,

see a l so Ref . 3 J ran the mol ten l ead d ischarged a t the col d el ectrode

back to the anode by grav ity fl ow . The cel l T1 I T1 20 l Tl was a l so i n­

vesti gated but was found to be very corros i ve [1 49J ._.

Tabl e I I I- 3 summari zes some of the data on transported entropi es i n

mol ten s a l ts ( s ee Eq . I I I- l ) for some of the reported cel l s . I n the

1 24

S=�I I.I ________ -r-____________ ......::To....:::.R=---4=1=-6 -� �

Tabl e I I I- 2 . 'Sm�MARY OF I N IT IAL THERMOELECTRIC POWERS IN METAL I MOLTEN

SALT I METAL THERMOCELLS

Cel l Temperature

( ° C)

Cu I CuCl I Cu 462- 588

W : Pb ( £ ) I PbC1 2 I Pb ( £ )W 500-700

W : Pb ( £ ) I PbBr2 I Pb ( £) 400- 700

Pb ( £ ) I PbC1 2 I Pb ( £ ) 627

Pb ( £ ) I PbI 2 I Pb ( £ ) 627

Zn l ZnC1 2 1 Zn 327

Sn l SnC1 2 1 Sn

1 25

( d E/dT) 1 =0 (mV/degree )

- 0 . 436

-0 . 006

-0 . 040

-0 . 008

-0 . 048

+0 . 1 3

-0 . 028

Reference

1 50

1 38

1 38

1. 49

1 49

1 3 1

1 49

S=�I I_I TR-4 16

Tabl e 1 1 1 -3 . SUMMARY OF TRANSPORTED ENTROP I ES ( SMn+ ) , PARTIAL MOLAL ENTROP I ES ( SMn+ ) , AND ENTROP I ES OF TRANSFER ( S�n+) FOR MOLTEN SALT THERMO­CELLSa

Cel l

Ag i AgN03 1 Ag Ag I AgCl l Ag Ag I AgC1 l Ag Ag l AgBr l Ag Ag l Ag I I Ag Cu I CuCl I Cu Zn l ZnC1 2 1 Zn Sn l SnC1 2 1 Sn

Ag I AgN03 , NaN03 I Ag

Ag I AgN03 , L i N03 I Ag

Temperature ( OC ) SMn+

2 27 21 . 0 527 26 727 26 . 7 477 27 577 27 525 24 . 2 327 8 327 22

A;N03b Temperature - ( OC )

0 . 3 31 0 0 . 5 31 0 1 . 0 31 0

0 . 3 31 0 0 . 5 31 0 1 . 0 31 0

1 9 . 0 22

22 24

1 1 4 1 6

S (Ag+ ) -

22 . 7 22 . 4 21 . 8

24 . 6 23 . 6 2 1 . 8

aA1 1 entrop i es are gi ven i n cal /degree 9 i o n . bMo 1 e fracti on of AgN03 .

1 26

2 4

5 3

-6 +6

= + + S ( L i or Na )

1 6 . 5 1 7 . 4

1 6 . 4 1 6 . 3

Reference

1 3 1 1 3 1 1 46 1 31 1 31 1 50 1 31 1 3 1

1 40 1 40 1 40

1 40 1 40 1 40

S=�I I.I __________________ �_---T-R--4'-1-6

-� �

case of pure mol ten sal ts the transfer entropi es ( S* ) were found to be

sma l l [ 1 3 1 J , of the order of 26 e . u . The parti a l mol al entropies i n the

s i mpl e bi nary sal ts were -ca l cu l ated by

SM�= t (SMX + i R tn :� I I I-6

where -

M . i s · the i on i c 'tJei ght [1 31 ] . 1 Tabl e 1 1 1 - 3 al so gi ves the

transported entropi es cal cu l ated by Haase et a 1 . [1 40J for the cati ons

i n the AgN03 : Li , NaN03 molten sal ts at 3 of the 1 0 compos i ti ons stud i ed .

1 1 1 . 1 . 2 Gaseous El ectrodes

The thermoel ectric powers of some X2 ( g ) l mo1 ten e1 ectro 1yte I X2 ( g )

cel l s wi th i nert el ectrodes ( e . g . , graphi te ) are shown i n Tabl e 1 11 -4 .

Ha l ogen revers i bl e el ectrodes general ly produce h i gher thermoel ectr i c

powers than those obtai ned wi th metal el ectrodes ( compare Tabl e 1 1 1 -4

wi th Tabl es 1 1 1- 1 and 1 1 1- 2 ) . The cel l C1 2 I AgC1 ( � ) I C1 2 has been i n­

vesti gated i n several l aboratori es [1 46 , 1 53 , 1 49J . Senderoff [1 53J

patented a thermoce1 1 of th i s type us i ng mo l ten AgC1 , KC1 : L i C1 , or

NaA1 C1 4 ( see Tabl e 1 1 1-4 ) . In the 500-900° C range , wi th mol ten KC1 : L i C1 ,

current den s i t i es as h i gh as 700 mA/cm2 , wi th very l i ttl e pol ari zati on ,

were reported . Thi s was achi eved us i ng a cel l des i gn cons i st i ng of two

i nert porous graphi te current col l ectors , provi ded wi th su i tabl e channel s

for transporti ng gas to the i r outer faces where the graphi te el ectrodes

were i n contact wi th the mol ten sal t el ectro lyte , wh i c h was sandwi ched

. between the current col l ectors . Ch l ori ne gas was fed at 500 °C and 900 °C

countercurrent 1y [1 53J . Lockheed Ai rcraft Corp . [ 1 49J re i nvesti gated

th i s system and obta i ned current-vol tage curves that i ndi cated on ly IR

1 27

S=�I I.I _____________________ ---'T"-"'R"'--........ 4 1=<-.B -� ��

Tabl e 1 1 1-4 . SUMMARY OF THERMOELECTRIC POWERS IN THERMOCELLS X2 1 MOLTEN SALT { t ) I X2

Ce l l Temperature { dE/dT ) I=O Reference ( O C ) (mV/degree )

C1 2 1AgCl I C1 2 a 727 -0 . 644 1 46

C1 2 I AgCl I C1 2 a 500- 900 - 0 . 655 1 52

C1 2 1AgC1 I C1 2 500- 900 -0 . 65 1 53

C1 2 I AgCl I C1 2 a 627 -0 . 675 1 49

C1 2 I Li Cl I C1 2 627 -0 . 534 1 54

C1 2 1 NaC1 I C1 2 860 -0 . 45 1 38

C1 2 I NaCl I C1 2 850 -0 . 475 1 46

C1 2 1 NaC1 I C1 2 827 -0 . 483 1 54

C1 2 1 KCl I C1 2 830- 950 - 0 . 40 1 38

C1 2 1 KCl I C1 2 850 - 0 . 475 1 46

C1 2 1 KCl I C 1 2 727 -0 . 504 1 54

C1 2 1 PbC1 2 1 C1 2 727 -0 . 544 1 54

C1 2 rCsC1 I C1 2 727 -0 . 533 1 54

C1 2 1 PbC1 2 1 C1 2 627 -0 . 579 1 49

C1 2 1 KCl : L; Cl I C1 2 500-900 -0 . 55 1 52 ( 5 4 . 5 wt % KC1 )

C1 2 1 NaA1 C1 4 1 C1 2 - 1 . 0- 1 . 4 1 53

I 2 1 Pb I2 1 I2 627 - 0 . 637 1 49

I 2 1 L i I I I 2 627 - 0 . 595 1 49

a One atmosphere C1 2 pres sure .

1 28

TR-4 16 S=�I I*I ----------------------------

pol ari zati on to at l east 300 mA/cm2 . S imi l ar resul ts were found wi th

the 12 ( g ) I Pb 12 ( � ) I 1 2 ( g ) and 12 ( g ) I Li 1 ( � ) I 12 ( g ) cel l s , a l though Pb12 had

too h i gh a vapor pressure to be of practi cal use [ 1 4-9 , see al so Ref .

3 J . Brodd [1 55 J des i gned a c l osed- system , mol ten sa l t thermoce l l wi th

gaseous el ectrodes .

Me i ssner et a l . [ 1 52J stud i ed the pressure dependence of thermo­

cel l s of the type C1 2 ( g ) I MCl I C1 2 ( g ) , where M = Ag , Na , K, and L i . The

effi ci enc i es of these cel l s were found to i ncrease wi th a decrease i n

total vapor pressure and wi th a n i ncrease i n the hot cel l temperature .

Hi gher temperatures a nd l ower total vapor pressures i ncrease the frac­

t i on of e l ectrolyte i n the vapor , whi ch woul d cause el ectrolyte carri ed

by chl ori ne gas to condense a l ong the way to the col d el ectrode . Es­

t imated effi c i enci es of these cel l s are 2-5% . . ( 1 -0 . 1 atm ; l ow tem­

perature at 500° C , h i gh temperature at 900- 1 300 ° C ) . A prel i mi nary

ana lys i s based on l i terature d�ta for the F2 1 L i F I F2 thermocel l gave an

estimated energy convers i on effi ci ency of 8- 1 0% between 900- 1 400°C , wi th

total pressure between 0 . 2- 1 . 0 atm [ 1 52 J .

Senderoff and Bretz [ 1 46J as wel l as Anderson et al . [ 1 49 J veri fi ed

experimenta l l y that the thermoel ectri c powers of the thermocel l s

M I MX ( � ) I M ( a ) I I 1 -7

and

graph i te , X2 ( g ) I MX ( � ) I X2 ( g ) ' graphi te ( b ) I 1 1 -8

and the temperature coeff ic i ent of the i sothermal cel l

M I MX I X2 ( graphi te ) ( c ) 1 1 1-9

1 29

TR-4 16 S=�I I.I -----------------------------� �

can be rel ated as fol l ows

( dE ) dT b (slI) = ( dE ) dT a dT c

_ ( dE ) dT graphi te/metal l I I- l O

Equati on I I I- l 0 can a l so be deri ved from

F (slI) dT a

= Sr� - SM+ - Se- ( M ) I I I- l l

and

F (dE ) = 1 S o + SC1 - - Se- ( G ) dT b - 2" Cl 2 I I I - 1 2

where the el ectron i c transported entropi es are negl i g i bl e for most

thermocel l s under cons i derati on ( note devi ations from Ref . 1 56 ) . Katel aar

[ 1 57 J has shown that the same type of corre l ation ( Eq . 1 1 1- 1 0 ) cannot be

used or deri ved for cel l s i n wh i ch mi xed mol ten sal ts are used .

1 I 1 . l . 3 The B i smuth-Bi smuth Iodi de System

The system B i /Bi 13 was itudi ed at Aerojet-General Corp . [52 , 1 58 J ,

a i mi ng i n i ti a l l y at therma l regenerati on whi ch evol ved i nto the thermo­

gal van i c mode . A worki ng sandwi ch-type cel l was constructed wi th a

porous cerami c materi al impregnated with B i 1 3 (wi th or without addi ­

ti onal el ectrolyte , e . g . , Zn 1 2 , AuCl , PtC1 2 , PdC1 2 , Cd 12 , Hg 1 , K1 ) and ­

seal ed between conducti ve materi a l s ( e . g . , Au ) . Heati ng one s i de and

coo l i ng the other produced current . Another cel l conta i ned a sandwi ch

of two porous carbon el ectrodes on ei ther s i de of the el ectrolyte .

S i nce B i 13 decomposes at 500 ° C , when the cel l s were operated wi th the

hot s i de at T > 500° C ( e . g . , 900°C ) , i od i ne was evo l ved at the hot

el ectrode and reduced at the col d el ectrode wi th mass transfer wi th i n

1 30

S=�I I_I __________________________ T_R_-_4_16_

the cel l . Prel imi nary work wi th unsea l ed cel l s gave about 0 . 5 V OCV and

1 00 mA/cm2 at 0 . 3 V. An attempt to sea l the cel l s gave 0 . 25 V at 1 4

mA/cm2 wi th the hot face a t 740 °C . The performance was of l i mi ted

durati on due to the poor seal of the cel l wh i ch l ed to oxi dation of the ,

graph i te anode �no of the bi smuth as wel l as l oss of both i od i ne and

el ectrolyte . Seal ed cel l s wi th a hot face temperature at 500 °C were

s hown to have a l i fe of �230 hours wi th a very poor performance .

Another group of researchers at North Ameri can Avi ati on i nvesti ­

gated the mol ten Bi /Bi I 3 system, studying the thermoel ectri c powers '

[ 1 59] , magneti c properti es [ 1 60] , el ectri cal conducti v i ti es [ 1 61 , and

thermal d i ffus ion [ 1 62] . Above 485 ° C , l i q u i d bi smuth i s mi s c i bl e i n a l l

proporti ons wi th B i I 3 . Thermoel ectr i c powers were determi ned for th i s

system between 400-500° C a s a functi on of compos i t ion ( from pure B i to

pure B i I 3 ) , and ranged from -0 . 053 to 0 . 1 25 mV/degree , wi th a maxi mum at

about 4'0 mo l e % Bi and mi n ima at ca . 1 0 and 85 mol e % Bi [ 1 59 ] . The

resu l ts were i nterpreted i n the framework of i rrevers i bl e thermodynami cs

and were cons i s tent wi th an i nterpretation i nvol v i ng el ectron i c conduc­

t ion i n metp1 -ri ch composi ti ons and a mi xture of el ectron i c and i on i c

conduction i n sal t-ri ch compos i t i ons . These concl us i ons were supported

by magneti c suscepti bi l i ty data i n wh i ch a dev i ati on from s i mp l e addi - ­

ti vi ty was observed at h i gh metal concentrations [ 1 60] . It was con­

cl uded that Bi di ssol ves i n Bi I 3 by reacti n'g to form Bi + , as was shown

i n anal ogous B i /Bi Br3 and B i /B i C1 3 systems [ 1 63 ] . Metal s and/or free

e l ectrons were conc l uded to be pres�nt at hi gh metal concentrati on .

The el ectri cal conducti v i ti es i ncreased conti nuous l y from 0 . 32 ohm-l cm-l

1 3 1

S=�I I.I ______ --'--_________________ ...::.T.=.:R'---4:..:1:...::.-6 -� ���

( pure Bi I 3 ) to 7 . 2 x 1 03 ohm- l cm- l ( pure Bi ) [ 1 61 ] . Fi nal thermoe l ectri c

powers were a l so determi ned i n th i s system wi th the cel l W ! B i - B i I3 !W over the

compos i t i on range of 0 . 01 - 0 . 90 mol e fraction of Bi . From 0 . 30 to 0 . 90

mo l e fracti on of B i the i n i ti a l and f ina l thermoel ectri c powers were

i dentical ( 0 . 097-0 . 01 5 mV/degree ) , whereas at l ower metal concentrations

the fi nal thermoel ectri c powers were l arger than the i n i t i a l ; e . g . , at

0 . 03 mol e fraction of B i , the i n i ti al and fi nal thermoe l ectri c powers

were 0 . 035 and - 3 . 1 mV/degree , respecti ve ly [ 1 62] . Di scuss i ons of

mechan i sms and transported entropy data are gi ven by Kel l ner [ 1 62] .

1 1 1 . 2 THERMOGALVAN I C CELLS WITH SOL I D EL ECTROLYTES

The poss i bi l i ty of u s i ng sol i d el ectrolytes composed of i on i c

materia l s for power generation has been env i s i oned i n batteri es and i n

thermoce l l s s i nce the 1 950s [ 1 25 , 1 27 , 1 64] . Exampl es of thermocel l and

gal van i c cel l devel opments us i ng sol i d el ectrolytes were gi ven by

Wei n i nger [ 1 65- 1 70] and Schi ral d i et a l . [ 1 72- 1 75 ] . The work usi ng S-

a l umi na sol i d el ectro lyte i n modi fi ed thermoga l van i c cel l s is descri bed

i n Sect i on I V . I n 1 972 , \!Jagner [24] rev i ewed the 1 i terature o n thermo­

e l ectri c powers i n sol i d el ectrolytes and mol ten sal ts . Most work

concern i n g the study of thermoel ectri c power in sol i d el ectrolytes a imed

at the i nvesti gat ion of the transport mechan i sms and the eval uati on of

heats of transport of the poi nt defects in the l atti ces of the systems

cons i dered [ 1 23- 1 25 , 1 31 , 1 47 , 1 76- 1 80 ] .

Tabl e 1 1 1- 5 l i sts thermoel ectri c powers for some thermogal van i c

cel l s wi th sol i d el ectrolytes ( these powers i n cl ude those for the i o n i c

sal ts pl us the e l ectrode temperature effect ) . Compari son of Tabl es 1 1 1 - 5 ,

1 32

h "' , TR-4 16 S=�I II.11 -------------------------::....::-'---=-� -� ��

Tabl e I I I- 5 . SUMMARY O F THERMOELECTRIC POWERS OF SOL I D- ELECTROLYTE THERMOCELLS [1 25J

Ce l T

Ag I Ag I ( 5 ) l Ag

Ag I a-Ag I I Aga

C1 2 I AgC1 ( s ) I AgC1 2 C1 2 1 a-Ag I I C1 2

a

Pb I PbC1 2 ( s ) I Pb

C1 2 I PbC1 2 ( s ) I C1 2

. Pb I PbBr2 ( s ) I Pb

Br2 I PbBr2 ( s ) I Br2

a From Ref . 1 68 .

( dE/dT ) 1=0 (mV/degree )

0 . 60

0 . 56- 0 . 60

1 . 29

1 . 2- 1 . 4

0 . 54

1 . 28

0 . 40

1 . 20

1 3 3

Temperature ( O C )

1 40- 500

1 50-400

300-41 0

1 60-500

200-470

260-400

350

320

- TR-4 16 S=�I I_I -------------------------"----

I I I- 4 , 1 I I- 2 , a n d I I I- l i ndi cate that for the same sal t system the

thermoel ectri c powers i ncrease in the order : M I MX ( �) I M < X2 I MX ( � ) I X2 � M I MX ( s ) 1 M < X2 I MX ( s ) I X2 ' thus refl ecti ng the l arger homogeneous thermo­

el ectr i c powers ( and obv i ous ly l arger transported entropi es ) i n the

sol i d sal ts as compared to � the mol ten sal ts .

Tabl e 1 1 1- 6 l i sts thermoel ectri c powers for some ion i c sa l ts wi th

ordered structure [ 1 64 ] ( l ow concentration of po i nt defects ) and wi th

di sordered cati on i c subl atti ces of the a-Ag 1 structural type [1 64] .

Recently determi ned thermoe1 ectri c powers of the i o n i c gl ass-

l i ke sol i ds Ag 1-A9nX04 (where X = Cr , Mo , W, S , Se , P , As ) and Ag 1 -

Ag2Cr207 ( 7 5-80 mol e % Ag 1 ) , wh i ch exhi bi t h igh i o n i c conducti vi ty , are

i n the range of 0 . 5-0 . 6 mV/degree i n the temperature range of 25-275 °C

[1 72] .

Tabl e 1 1 1- 7 l i sts val ues of the transported entrop ies ( SMn+ ) for

some sol i d el ectrolytes [ 1 23 , 1 31 ] and an overal l transported entropy for

a-Ag 1 and rel ated compounds (i n wh i ch chemi cal di sorder i s superimposed

on the therma l di sorder ) and for some of the gl ass-type sol i ds descri bed

above [1 72] ( cf . Tabl e 1 1 1- 3 ) .

A few cel l s were made for power generati on and the i r di scharge

characteri sti cs were studi ed . The work of Wei n i nger [1 65- 1 70] at

General El ectri c Co . i s descri bed i n more deta i l i n Secti on 1 1 1 . 2 . 1 .

Schi ral d i et a l . [ 1 72- 1 75 ] have reported only battery resul ts us i ng the

s i l ver i odi de- s i l ver oxysal ts sol i d el ectrolytes at l ow temperature ( 25-

60° C ) , but they estimate the fi gure of meri t of thei·r thermocel l system

as 1 0- 3_ 1 0-4 degree- l . . The 1 imi tati on imposed by mass transfer ; s

' emphas i zed by the authors [1 72 ] ; no current-vo l tage curves for these

thermocel l s have been measured .

134

TR-4 16 S=�I r.r" -------------------------

-� ��

Tabl e 1 1 1- 6 . SUMMARY O F THERMOELECTRIC POWERS O F SOME ION I C SALTS

Ion i c Sal t

NaCl

NaBr

KCl

KBr

AgCl

AgBr

y-Ag I

S-Ag I

a-Ag I

" RbA94 1 5 KAg4 1 5 NH4A94 15

- ( dE/dT ) 1=0 ( mV/qegree )a

944/T - 2 . 295

1 . 1

1 . 8- 1 . 9

1 . 8- 1 . 9

1 . 2

0 . 8

0 . 6

0 . 66

0 . 6

1 . 0

0 . 7

93/T + 0 . 36

78/T + 0 . 28

58/T + 0 . 31

a T i s absol ute temperature .

135

Temperature ( O C )

450- 750

600-700

550- 700

550- 700

300

300

350

350

25-1 46

1 46

1 48

25-200

25-200

25- 200

Reference

1 78

1 79

1 78

1 79

1 27

1 27

1 77

1 30

1 80

1 80

1 80

1 80

1 80

1 80

TR-4 16 S=�I I.I --------------------------� �

Tabl e I I I- 7 . SUM�1ARY OF TRANS�ORTED ENTROP I ES ( S�1n+) FOR SOLID ELECTROLYTES

El ectrolyte

AgCl

AgBr

S-Ag I

CuCl

PbC1 2 PbBr2

a-Ag I

RbA94 I5 NH4A94 I5 AgI ' A92Mo04 Ag I ' A92S04 AgI ' Ag2Te04

a Ref . 1 31 . b Ref . 1 72 .

O F ORDERED STRUCTURE [1 23 , 1 31 J AND SUMMARY OF OVERALL TRANSPORTED ENTROP I ES FOR SOL I DS OF THE a-Ag I TYPE [1 72 J and Ag I -Ag OXYSALTS ( 75-80 MOLE % Ag I ) [ 1 72J

Temperature ( O C )

1 27 427

1 27 327

1 27

1 27

227

227

227 227

227

227

227

227

227

136

S�1n+ (ca l / degree 9 i on )

44 31 . 8

46 32 . 8

40 . 6

33

46

40

26 . 3a

27 . 6b

23 . 9

23 . 2

27 . 0

29 . 0

32 . 8

S=�I I*I _________________________ T_R_-.....;4_1_6

The Ag !.a-Ag I ! Ag and I2 ! a-Ag I ! I2 Thermocel l s

I n an attempt to devel op h i gh temperature mi ni ature batteri es

acti vated by gaseous ha l ogens , We i n i nger [ 1 65 , 1 66 J stud i ed a sol i d

e l ectro lyte cel l cons i sti ng of s i l ver ha l i de wi th Ta and Ag wi res as

cathode and anode . When the cathode i n these cel l s was exposed to

hal ogen vapor ( Br2 or 12 ) , a h i gh temperature primary battery was form­

ed . W i th a-Ag I the cel l s coul d be recharged a few times . The cel l

( Ta ) I2 ! a-Ag I ! Ag , stud i ed i n the 1 50-550 °C temperature range , showed an

OCV of 0 . 67 V and produced short c i rcu i t currents of 1 8 rnA [1 66 J . The

current outputs i ncreased wi th temperature [1 66J , and improved sol i d

e l ectrolyte gaseous d i ffus i on cathodes were a l so proposed [1 67 J .

Para l l e l to thi s battery work , thermocel l s were devel oped [ 1 68- 1 70J i n-

vol v i ng the systems

Ag ! a-Ag I ! Ag ( the " s i l ver cel l " ) I I I - 1 3 .

and

I I I - 1 4

Several cel l confi gurati ons were tri ed [1 68J for the s i l ver thermo�

cel l . These i nc l uded ( 1 ) s i l ver el ectrodes sandwi ch i ng the a-Ag I el ec­

trolyte ( i on i c res i sti v i ty of 0 . 38 ohm cm at 500°C and 0 . 47 ohm cm at

350°C ) conta i ned in a g lass spacer to prevent deformation and ( 2 ) two

s i l ver e l ectrodes sandwi chi ng a porous S-al umi na . matri x i mpregnated wi th

a-Ag I o Duri ng operation , s i l ver i s di ssol ved from the hot anode ,

137

'" TR-4 16 S=�I I�I --------------------------------------------------���� - � �

s i l ver i ons mi grate from the hot to the col d el ectrode , and s i l ver metal

i s deposi ted at the col d el ectrode . Dendri te formati on i n these cel l s

l eads to l oss of contact at the el ectrode/el ectrol yte i nterface , l oss of

s i l ver , and s hort c i rcu i t of the cel l . The cel l needs a provi s i on for

revers i ng the hot and col d s i des for conti nuous operati on . I n the cel l

wi th a porous matri x the dendri te formation was s l owed down but at the

cost of i ncreased i nternal res i stance . Current-vol tage curves were

obta i ned for these cel l s . For i nstance , for �T = 226 ° C , a thermopoten-

t i a l of �0 . 1 3 V ( OCV ) was observed and at � . 05 V a curren t of 40 rnA was

drawn from the cel l ( R = 1 . 6 ohm ) .

These probl ems of the s i l ver cel l di d not appear i n the i odi ne cel l

due to the gaseous el ectrode . I n th i s case , the i odi de i ons are oxi -

d i zed to i od i ne at the co l d el ectrode , produc i ng an i ncreased i odi ne

pressure at the col d el ectrode . S i l ver i ons mi grate from the col d to

the hot el ectrode , where iodi ne i s reduced wi th the formation of s i l ver

i odi de . Fi gure I I I- l s hows a d i agram of the i od i ne thermocel l i n wh i ch

the porous i od i ne d i ffus i on el ectrodes are made of graph i te . Some of the

maj or probl ems of thi s cel l were : contract i on of the Ag I on heati n g ,

mai ntenance o f the three-phase el ectrode/el ectro lyte/ i od i ne i nterface ,

and encapsu l ation [1 68- 1 70J . For thermodynami c express i ons of the

thermoel ectri c power .ta k i ng i nto account the contracti on of the el ec­

tro lyte wi th temperature , see Ref . 1 7 1 . The d ischarge characteri st i cs

of l a boratory cel l s were studi ed as a function of temperature [ 1 68 J .

The data i n Tabl e II I-8 exemp l i fy the performance of the i od i ne

cel l . At �T = 280 °C , el ectrochemi cal pol ari zation l i mi ts the current

dra i n , but at �T < 1 70 °C , mostly I R pol ari zati on is observed . The

1 38

- TR-4 16 !;::�I [11[--------------------------------------------------------------

-----

350'C - . - �

-

\ A

-

-HeOI 550 'C -

Figure 1 1 1-1 . Weininger's 12 /a-AgU1 2 Thermocell [1 68] A: annular ceramic mounting; B: porous diffusion gas electrodes; C: iod ine vapor; 0: casing; E: solid a-Agi

139

TR-4 16 S=�I I.I -----------------------------� �

Tabl e 1 1 1-8 . PERFORMANCE OF THE IODINE CELL [ 1 68 J

t,E (mV ) t,E/t,T I at /:£/2 R . 1 T2 ( O C ) Tl ( O C ) (mV/degree ) (rnA ) ( ohms )

342 262 95 1 . 1 0 0 . 6 340 1 84 208 1 . 33 1 . 4 335 1 65 232 l . 36 1 . 8 498 2 1 8 347 l . 24 1 . 8

aCurrent < 0 . 6 rnA . bCurrent > 1 rnA .

coul ombi c capac i ty o f the cel l , l i mited by the geometry of the el ec­

trode , was determi ned : . one cel l had a capac i ty of 8 . 5 coul ombs on a

79 67a

61 9l b

1 00-ohm l oad . The effi ci ency of the iodi ne cel l i s improved by revers­

i ng th� temperature gradi ent to avo i d the advance of the sol i d el ectro­

lyte from the h i gh pressure zone . Th� thermoel ectric powers obta i ned ,

1 . 2-1 . 4 mV/degree ( see Tabl e 1 1 1- 8 ) , compare we l l wi th theoreti cal

val ues , 1 . 3- 1 . 5 mV/degree [see Refs . 3 and l 25 J . The effi ci ency of th i s

system, cal cu l ated from Eq . 1 1 1 -4 , was �5% .

1 1 1 . 3 THERMOGALVAN IC CELLS I N AQUEOUS AND NONAQUEOUS SOLVENTS

The work descri bed i n th i s secti on was a imed at power generation

[ 1 82 , 1 83 , 1 85 , 1 86J , practical uses of thermogal van i c cel l s [ 1 84J , or

e l uci dation of general pri nc i pl es ( e . g . , appl i ed to thermoga1 van i c cor­

ros i o n ) [ 1 2 l , 1 87- 1 90J . LudWi g and Rowl ette [ 1 91 J dev i sed a thermo-

gal vani c concentrati on cel l for power generation wh i c h is descri bed

l ater i n th i s secti on . Reference 1 21 l i sts ca . 300 temperature co-

140

S=�I I�.=� I ____________________ -=T=R'---4=1=-6 -� �� �

effi c i ents of e l ectrode potenti al s ( from ca l cul ations and some experi -

menta l data ) .

Tabl e 1 1 1-9 presents thermoel ectr ic powers for cel l s Cu l el ectro­

l yte l Cu i n aqueous and nonaqueous med i a , as wel l as L i ebhafsky ' s [ 1 84 J

resu l ts for cel l s wi th copper e l ectrodes sandwi chi ng cat i on exchange .

res i ns wh ich were treated wi th Cu2+ i ons i n aqueous sol utions or mi x-

tures of water and organi c sol vents . Due to the l i neari ty of the EMF

( OCV ) val ues wi th �T , these cel l s were suggested as thermogal van i c

thermocoupl es for restri cted operations [ 1 84J . Use of these cel l s as

power generators was not suggested , but cl early l arge IR po l ari zati on

shoul d be expected .

Cl amp i tt and German [1 82 J have suggested severa l confi gu�ati ons for

h 1 1 d h · 1 C 2+ . . d t ermoce s an ave g l ven some resu ts for u l ons l n aqueous an

nonaqueous sol vents , wh i ch are reproduced i n part in Tabl e 1 1 1 . 9 . These

a uthors have a l so measured the power output and vol tage of some of these

cel l s . For Cu i CuS04 i n H20 l Cu 'and Cu i CuS04 i n H20 + 20% H2S04 I Cu , 'the

vol tages atta i ned at maximum power were 31 and 28 mV , wh i ch correspond

to powers of 0 . 03 and 0 . 25 mW/cm2 , respecti vel y . For compari son , some

resu l ts comp i l ed by deBethune et a l . [ 1 21 J for aqueous sol utions are

a l so i nc l uded i n Tabl e 1 1 1 - 9 . Due to mass transfer occurr ing when the

cel l � s operati ng under current dra i n cond i ti ons , Cu i s d i ssol ved

( oxi d i zed ) at the col d el ectrode and deposi ted at the hot e l ectrode , and

therefore Cu2+ i ons are transferred from the col d to the hot el ectrode .

To draw power conti nuous ly from the cel l , i t i s necessary to reverse the

process and heat the col d el ectrode and coo l the hot el ectrode by some

mecha n i cal means ( see Ref . 1 82 ) . The growth of dendri tes and the l oss

141

TR-4 16 S=�I I.I -----------------------------� ��

Tabl e I I I- 9 . THERMOELECTRIC POWERS O F THE Cu i ELECTROLYTE i Cu CELLS IN AQUEOUS AND NONAQUEOUS SOLVENTS

El ectrolyte ( dE/dT ) 1=0 ( mV/degree )

Temperature Reference

CuS04 0 . 08 M 0 . 5 M 1 . 0 r1 saturated

0 . 01 M; pH = 4 . 65 0 . 01 r1 ; pH = 1 . 8-1 . 0

saturated i n H20 saturated i n 20% H2S04 + Na2S04 'V0 . 4 M

NaCl

1 5 w/w % 1 5 w/w %

Amberpl ex Cl - Cu2+ form

Pheno l sul foni c res i n : Cu2+ form i n H20

Same res i n above i n water/ ethyl ene g lyco l

CuS04 i n CH30Ha

CuS04 i n DMSOb , a

CuS04 i n D�1Fc , a

aSaturated sol uti ons . bDMSO = d imethyl su l foxi de . cDMF = di methyl formami de .

0 . 64 0 . 73

0 . 69- 0 . 79 0 . 9- 0 . 97

1 . 0 0 . 5-0 . 3

0 . 89

1 . 03

-0 . 35 -0 . 24

- 0 . 5-0 . 6

-0 . 6

-0 . 5-0 . 7

0 ,, 68 1 . 1 9 0 . 98

142

( OC )

0- 50°C

tco l d

25 25

20

20

1 9 1 9

o

thot

75 85

1 00

1 00

1 00 1 00

1 0- 50

not spec ifi ed

not spec i fi ed

20 68 2.0 1 52 20 1 25

1 2 1

1 81 1 8 1

1 82

1 82

1 83 1 83

1 84

1 84 _.

1 84

1 82 1 82 1 82

TR-4 16 S=�I I.I ------------------------------� �

of copper are the maj or probl ems ant i c i pated for l ong-term operati on .

Tabl e 1 1 1- 1 0 a ssembl es data for exampl es of other thermogal van i c

cel l s studi ed i n aqueous med i a . The l ast two exampl es i n Tabl e 1 1 1 - 1 0 ,

thermogal van i c cel l s i n wh i ch the two redox spec i es are sol ubl e , were

studi ed by Burrows [1 85 , 1 86 J . I n these cel l s the permanent mass tran.s­

fer probl ems associ ated wi th metal deposit i on are not present , though ,

i n both exampl es [1 85 , 1 86J the di scharge behavi or of these cel l s , wh i ch

was stud i ed i n deta i l , s howed that concentration pol ari zati on ( the rate

of mass transfer of the el ectroacti ve spec i es - Fe3+ , Fe2+ or Fe ( CN ) ci- ,

Fe ( CN )�- - to the e l ectrode ) 1 i mi ts ' the current and therefore the power

output of such devi ces . However , these cel l s d id produce conti nuous

steady-state pow.er output (current dra i n ) as l ong as the temperature

gradi ent rema i ned constant . The maxi mum power output for these systems

under the experi menta l condi ti ons was ....., 0 . 05 mW/cm2 for the Fe2+/ Fe3+

coup l e [1 85J and <0 . 1 mW/cm2 for the Fe ( CN ) 64-/ Fe ( CN ) 6

3- coupl e [1 86J ,

wh i ch has a h i gher thermoel ectri c power than Fe2+/ Fe3+ ( see Tabl e 1 1 1 -

1 0 ) . Improvement i n power output by a factor of four was obta i ned by

forced convecti on of the sol ut ion adj acent to e i ther el ectrode [ 1 86 J .

Obv i ous ly , improved mass transfer may destroy the temperature gradient

in a practi cal dev i ce . The effi ci ency of convers i on of 2- 5% was ca l cu­

l ated from the open-ci rcu i t data by us i ng the sol i d-state thermoel ectri c

express i on ( Eq . I I I-4 ) for the Fe2+/ Fe3+ coup l e .

References 1 2 1 and 1 23 gi ve transported entropi es for vari ous meta l

i ons i n aqueous sol uti ons .

Cl amp i tt and German [ 1 82 J and Deysher [ 1 92J descri be thermogal van i c

cel l s empl oyi ng el ectrodes o f the second ki nd , e . g . , Hg/H92S04 , i n wh i ch

143

S=�I I.I ______________________ -=-T.:;;:;R---'4=1""-S -� ��

Tabl e I I I- 1 0 . THERMOELECTRI C POWERS FOR SOME SELECTED THERMOGALVAN IC C ELLS I N AQUEOUS MED IA

System ( dE/dT ) 1=0 Temperature ( O C ) Reference ( mV/degree ) Col d Hot

Pb I Pb ( C2H302 ) 2 I Pba 0 . 2 20 1 00 1 82 Zn I Zn ( C2H302 ) 2 I Zna 1 . 1 20 1 00 1 82 Zn I pheno 1 i c i on exchange. 1 Zn

. . Z 2+ f reS l n l n n orm 0 . 5 L1T = 30 1 84 Pt 1 HC1 ( 20 w/w % ) I pt 0 . 1 8 1 0 80 1 83 Pt 1 HN03 ( 30 w/w % ) I pt 0 . 30 1 0 90 1 83 Pt I Fe2+ , Fe3+; 1 M HCl l pt

[ Fe2+]= [Fe3+] = 2 M 0 . 57 30 80 1 85 1 M 0 . 78 30 80 1 85

0 . 25 M 1. 0 30 80 1 85

Pt I Fe ( CN ):- , Fe ( CN )�- ; 0 . 5:

_M,_ �

S�4 �t [ Fe ( CN ):- ] = [ Fe ( CN )�- ]

= O . r M 1 . 4 30 80 1 86

aSaturated sol uti ons .

144

· TR-416 S=�I I.I -------------------------

-� �

two cel l s are connected back to back by a meta l wi re and the sol uti ons

conta i n saturated CUS04 and H92S04 ( hot s i de ) and H92S04 and unsaturated

CUS04 ( co l d s i de ) . Copper el ectrodes are immersed in these sol uti ons .

On the hot s i de , Cu2+ i ons are reduced and the SO�- i ons prec i p i tated as

H92S04 . At the col d s i de , copper is oxi d i zed and H92S04 ( s ) d i s so l ved .

For power generation , the rate of preci p i tate formation and di ssol ution

wi l l probably l imi t the output of such dev i ces . Cl amp i tt and German

[ 1 82J descri be several other cel l s us i ng el ectrodes of the second ki nd

wi th more mass transfer probl ems than th i s exampl e .

Thermoel ectri c powers and mi s l ead i ngly h i gh effi c i enc ies ( S-20%)

g i ven i n Ref. 1 83 shoul d be careful ly analyzed s i nce the authors worked

i n severa l cases wi th open cel l s , even at t�mperatures as h i gh as

1 20° C . Thermoel ectric powers of 4- 6 mY/degree gi ven for the system

Pb 1 H2S04-H20 1 Pb actual l y refl ect the concentrati on cel l formed duri ng

evaporation on the hot s i de . Thermoel ectric powers i n the temperature

range of 20-S0°C are of the order of <O . S mY/degree and compare favor­

ab ly wi th those ca l cul ated by deBethune et a l . [ 1 21 J .

Equati ons rel ati ng the thermogal van i c currents and the temperature

d i fference between e l ectrodes are g i ven by Kal uzh i na and co-workers

[ 1 87- 1 90 J as functi ons of the exchange currents of the di ssol ution and

depos i tion of the metal and the act i vation overpotenti a l s of these

processes .

Sul furi c Ac i d Concentrati on Thermocel l

Ludwi g and Rowl ette [ 1 91 J proposed the sul furi c ac id cel l shown i n

Fi g . I I I- 2 , i n wh i ch the two pl ati num el ectrodes sandwi ch a porous , non-

145

S=!!!!tl r;.� �r -------------------

- � ��'} TR-4 16

-> Q)' 0.3 Ol al

-"0 > (HOT FACE TEMPERATURE)

o L-�u--U� __ L-��_L�� __ ����� __ ����� o 5 10 15 20 Current Density (mA/cm2)

Figure 1 1 1-2. Diagram of the Sulfuric Acid Concentration Thermocell and Current - Voltage Curves for the Cell [1 91 ]

146

S=�I I_I __________________________ T_R_-_4_1_6

conduct i ng barri er. Heat i s appl i ed at the hot el ectrode . Water i s

evaporated from the heated portion and condenses at the col d el ectrode ;

therefore , the sol ution adj acent to the col d el ectrode wi l l be more

di l ute than that adjacent to the hot el ectrode . Cap i l l ary act ion i n the

porous barr ier compensates for the l oss of water and su l furi c ac i d as

the hot part dri es out , so that , at equi l i bri um , a bal ance between these

two processes i s achi eved . Si nce the acti v i t i es of ac i d and water i n

su l furi c aci d sol uti ons vary wi del y wi th compos i ti on and temperature

[1 93] , thi s cel l i s a concentrati on cel l on wh i ch the smal l er thermo-

gal van i c effect i s superimposed . Fi gure I I I -2 a l so shows exampl es of

vol tage-current dens i ty curves for these cel l s . When oxygen gas i s

passed through the cel l , the el ectrode reactions are :

2e- + 2H+ + 1 / 2 O2 ---7-) H20 (rv200° C ) cathode ( hot face )

externa l t -} c i rcu i t 2e- + 2H+ + 1 /2 O2 < H20 ( 6BOC ) anode ( col d face )

OCV as h i gh as 0 . 7- 0 . B V were achi eved and the maximum power output was

1 7 mA/cm2 at 0 . 5 V [ 1 3 , 1 92 ] . A cel l was tested conti nuous ly for l B

hours under l oad wi th no deteriorati on of performance . Oxygen acti va- -

t i on overvol tages on pl ati num are h i gher than hydrogen overvo l tages .

Cel l s us i ng hydrogen i nstead of oxygen were al so tested but the d i s­

advantage i s that hot , concentrated su l furi c aci d s l owly oxi di zes

hydrogen .

Other cel l des i gns were proposed wi th a fracti onator coupl ed to the

el ectrolyti c cel l i n wh i ch the system operates sol e ly in the therma l

147

- TR-4 16 S=�I I.I ------------------------------ � �

regenerati on mode . References 1 92 and 1 3 g i ve deta i l ed descri pti ons of

these proposed cel l s .

1 1 1 . 4 D I SCUSSION O F TRES TYPE 6

Tabl e S- l presents th� r�sul ts for some of the thermoga l van ;c cel l s

devel oped for power generation . Mo l ten sal t thermoce l l s can , i n pri nci pl e ,

g i ve more power due to the hi ghe� range of . l i qui dus i n wh i ch they exi s t .

However , some data exi st for aqueous medi a i ndi cati ng that modest

powers of �l OO �W/cm2 can be achi eved in these systems . In v i ew of the

ava i l abi l i ty of sol ar heat sources for temperatures l ess than 1 00°C

( e . g . , so l ar ponds ) , the i nvesti gation of these l ow- temperature , modest-

power , rel ati vely i nexpen s i ve engi nes shou l d be conti nued . Research i n

th i s area i s be i ng performed at S ERI .

148

- TR-416 S=�I I_I -----------------------------

SECTION IV

COUPLED THERMAL AND ELECTROLYTIC REGENERAT ION BASED ON PRESSURE D I FFERENCES OF THE WORKING ELECTROACTIVE FLU I D

I V . l S I NGLE CELLS

The el ectrochemi cal heat engi nes descri bed i n th i s secti on are

based on a pressure d i fference of the worki ng el ectroacti ve fl u i d across o

the i sothermal sol i d or l i q u i d el ectrolyte ( for non i sothermal el ectro­

l ytes , see Secti on I I I ) . The pres sure di fference i s ma i nta i ned by

v i rtue of the change i n the' vapor pres sure wi th the temperature of the

worki ng el ectroacti ve fl u i d . The work performed i s equi va l ent to an

i sothermal expans i on of the worki ng el ectroacti ve fl u i d from pressure P2 to Pl at T2 through the e l ectro lyte and i ts i nterfaces . After expans ion

the worki ng fl u i d i s condensed i n a col d reservo i r and can be recycl ed

to the h i gh temperature , h i gh pressure part of the cel l by a pump .

These cel l s are bas i cal ly concentration ce l l s . If the temperature

across the el ectrolyte rema i ns constant , one of the sources of i rrevers i -

bi l i ti e s i s mi n imi zed . One of the maj or advantages of thi s concept i s

that no chemi cal regeneration step i s necessary . Because the worki ng

fl u i d does not undergo chemical changes , regenerati on and separati on

steps are ,not necessary . These cel l s are di scus sed as Type 7 in the

Introduct ion .

At open c i rcui t the vol tage of these engi nes i s gi ven by the

Nernst equation as

E - RT2 - nr

149

IV-l

TR-4 16 S=�I I.I ----------------------------� ��

where f = fugac i ty of the work ing fl u i d at the hi gh pressure el ectrode ( f2 ) and at the l ow pressure el ectrode ( fl ) . I f the work i ng fl u i d behaves a s a perfect gas , the fugaci t i es are equal to the part ia l press ures , wh i ch can be estimated from the Cl aus i us-Cl apeyron equation . The rel at ionshi p between the open-ci rcu i t vol tage and pressure i s

, - --� -, �--dE RT 2 (T 2) /� . . -�:. : ': i�S .en P ) = - � ,,-

/ IV-2

where T1 ; s the condensation temperature and T2 i s the vapori zati on temperature .

Secti on I V . l . l descri bes the fol l owi ng conti nuo�s gas concentrati on cel l s proposed and stud i ed by Angus [1 94 , 1 95 J : 12 ( g ) I Pb I 2 ( .e ) I I 2 ( g ) ; Hg ( g ) I Hg2C1 2 ( .e ) I Hg ( g ) ; Na ( g ) I NaCl ( .e ) I Na ( g ) ; and K( g ) I KCl ( .e ) I K( g ) . Section IV . 1 . 2 descri bes Ford ' s i nteresti ng sodi um heat eng i ne

I V . l . l Conti nuous Gas Concentration Cel l s Angus [1 94 , 1 95 J dev i sed an el ectrochemi cal heat eng i ne based on

i od i ne vapor bei ng expanded through an i sothermal el ectrolyte , Pb I 2 ( .e ) , capab l e of d i ssoc i ati ng the worki ng fl u i d i nto ions . Fi gure IV- l shows the d i agram of the smal l - s�al e l aboratory vers ion of the iodi ne cel l [ 1 94J . N i c kel and pl ati num e l ectrodes were empl oyed . Tabl e IV- l shows the vol tage characteri st ics of the iod i ne cel l s .

1 50'

TR-4 16 S::�I I�I ---------------------------------------------------------------------­-� �

ELECTRODES --+------1

PYREX

LOW PRESSURE

RESERVOIR

T THS SECTION OF CELL IS INSERTED IN A TUBE FURNACE

1 AUXILIARY HEATER FOR HIGH PRESSURE RESERVOIR

Figure IV-1 . Laboratory Continuous Gas Concentration Cell [1 94]

1 5 1

TR-4 16 S=�I I_I ----------------------------

o

Tabl e I V- l . VOLTAGE CHARACTERIST ICS OF THE I2 ( T2 ) I Pb I 2 ( � ) I I 2 ( Tl ) q_L_V_!}H _ _ �� EL�CTRODES [1341

Vapori zati on temperature ( OC ) 1 77 1 93 Condensati on temperature ( OC ) 2A 24 El ectrolyte temperature ( OC ) 538 538 Vol tage ( l ow pres sure e l ectrodes ) 0 . 22 0 . 28 Predi cted vol tage from Eq . IV- l 0 . 207 0 . 22

Regenerati on was accomp l i shed by cool i n g the ori g i nal hot end and vapori z i n g the i od i ne from the ori g i na l co l d end ; the cel l vo l tage reversed accordi ng ly . The i nterna l res i stance of these cel l s was very h i g h . Advanced cel l des i gns were tested and the most successful i s shown i n Fi g . I V-2 . Porous n i ckel el ectrodes sandwiched the Pb I2 ( � ) el ectro lyte . The OCV of the cel l wi th enough 1 2 to g i ve 1 atm o f 1 2 vapor at T2 was 0 . 1 7 V ( Ri = 2 . 9 ohms ) -. W i th a 24 . 5-ohm l oad , 6 . 2 rnA were dra i ned from the cel l . However , as the cel l was operated under l oad , i ts i nterna l res i stance i ncreased conti nuous l y . I t was found that the e l ectrolyte had been forced out of the cavi ty i nto the n i ckel el ec-trode from the h i gh to l ow pres sure zones . Corros i on of the n i ckel e l ectrodes by I 2 vapor was a l so detected , a probl em avoi ded by the use _ of pl ati num e l ectrodes .

Laboratory cel l s were a l so operated wi th the system Hg ( g ) I Hg2C1 2 ( � ) 1 Hg ( g ) , wi th W or Pt e l ectrodes . The Na ( g ) I NaCl ( � ) I Na ( g ) system and the correspondi ng potass i um system were a l so cons i dered as poss i bl e candi dates for cel l s of th i s type . Tabl e I V-2 assembl es the characteri sti cs of the cel l s wi th these worki ng fl u i d s .

152

- TR-4 16 S=�I I�.-� I ------------------------------� �� �

HOT ZONE

I i

I i " I

EPOXY SEAL

THREADED BORON NITRIDE SLEEVE

GRAPHITE

QUARTZ INSULATOR

BORON NITRIDE SEPARATOR RING jI�3��t POROUS METAL ELECTRODES

ELECTROLYTE CAVITY

<+--.>,,-it-- CONDENSED WORKING

FLUID

Figure IV-2. Advanced Continuous Gas Concentration Cell -. ,

1 5 3

· '" TR-416 S=�I I.I -------------------------.:.::.::-=-==-=­- � �

Tabl e I V- 2 .

Worki ng Fl u i d/

CHARACTER I STICS OF CONT I NUOUS GAS CONCENTRATION CELLS W ITH S EVERAL WORKING FLU I D/ ELECTROLYTE SYSTEMS [ 1 94 J

p Vapori zati on Condensation OCV P 2 P I El ectro lyte ( ohm cm) Temp . ( OC ) Temp . ( O C ) ( V ) ( torr ) ( torr )

1 2/ Pb I2 2 . 1 410 1 1 9 0 . 3 31 , 900 93 . 4 Hg/H92C1 2 2 . 0 305 1 00 0 . 22 247 0 . 27 Na/NaCl 0 . 27 827 327 0 . 88 433 0 . 039 K/ KCl 0 . 44 827 327 0 . 71 1 408 0 . 64

The a l kal i -meta l - based systems were found to di spl ay a much more favorabl e OCV than 1 2 or Hg systems . The maj or di fficu l ty associ ated wi th th i s type of cel l i s the need to ma i nta i n the i ntegr ity of the l i q u i d el ectro lyte when it is subj ected to a pres sure grad i ent . These press ure di fferences are a l so smal l er for the a l kal i metal systems than those for the 12 cel l ( see Pl and P2 in Tabl e IV-2 ) .

I V . l . 2 The Sodi um Heat Engine The sod i um heat engi ne is s i mi l ar to the cel l s descri bed in Secti on

I V . l but does not have the probl ems associ ated wi th a l i q u i d el ectrolyte subjected to a pressure d i fference . The el egant sol uti on to these probl ems , advanced by Kummer and Weber [ 1 96J , i s to use the sod i um-ion­conducti n g , sol i d e l ectrolyte 13" -al umi na to separate the h i gh and l ow pressure zones i n a c l osed-cycl e dev i ce i n wh i c h fl u i d sodi um i s c i rcu -l ated [ 1 96-200 J . The schematic di agram of th i s el ectrochemical heat engi ne i s shown i n Fi g . I V- 3 . I n the hi gh temperature ( T2 ) area o f the dev i ce , fl u i d sodi um i s at a pressure P2 ( upper reg i on ) , h i gher than the

154

TR-4 16 S=�I r;.=�r -------------------------

-� � � 'f.-

r - - - ----- - -- - - - - � I Tem'perature T2 I I I I I I I I I I I I I I + I L_.__ __J

Pump

r -·-- - -� I I I I I I I I I : L_. _ _ _ ��e��� ____ .J

Figure IV-3. Schematic Diagram of the Sodium Heat Engine [1 9J] , _

1 5 5

TR-4 16 S=�I I.I ---------------;------------� ��

pressure Pl i n the l ower reg i o n . The l i qu i d sodi um i s heated to T2 , and sodi um i s oxi d i zed to sod i um i ons and el ectrons . The el ectrons l eave the h i gh pressure zone v ia the negati ve el ectrode ( current col l ector ) , The sodi um i ons mi grate through the sol i d el ectrolyte as a resul t of the press ure di fference ( P2 - Pl ) across the el ectrolyte . At the l ow pressure s i de of the sol i d el ectrolyte , a su i tabl e i nert porous e l ectrode ( the cathode ) coats the e l ectrolyte . Sodi um ions are reduced at the porous e l ect�ode by the e l ectrons arri v i n g through the externa l l oad , thus formi ng sod i um meta l . Neutra l sodi um evaporates from the porous el ectrode at pressure Pl and temperature T2 , pas s i ng i n the gas phase to a condenser at Tl ( Tl < T2 ) : Condensed l i qu id sod i um i s returned to the hi gh pressure ' s i de by an el ectromagneti c pump , thus compl eti ng the cycl e wi thout the movement of mechan i cal parts - on ly c i rcul ation of fl u i d sod i um . The work output i s purely e l ectri cal but equ i va l ent to the mechani cal work of the i sothermal expans ion of sodi um from P2 to Pl at T2 [ 1 96 , 1 97 J .

Weber [1 97J presented a theoretical ana lys i s of the effi c i enc ies of these devi ces under no l oad and under l oad , neg l ecti ng para s i ti c heat l os ses (wh i ch were i nc l uded i n l ater papers : Refs . 1 98 , 1 99 ) . Current­vo l tage rel ati onsh i ps were deri ved theoreti cal ly by us i ng Eq . IV-l and the rate of evaporation of the meta l , s i nce th i s rate i s proporti onal to the current dens i ty . An express i on of the type :

\ ,- i V = A - B Q,n I - I R '.. _ . 0 IV-3

was obtai ned , where Ro is the surface el ectri cal res i st i vi ty of the sol i d e l ectro lyte and A and B are coeffi ci ents cal cu l ated from phys i cal constants and the emp i rical rel at ion between vapor pressure and temperature

1 56

,,�, TR-416 S=�I II.II ----------------------------­_�r �

for sodi um ( A = 0 . 74 V and B = 0 . 086 V f�r Tl = 227°C , T2 = 727°C ) . Weber [ 1 97 J analyzed the several sources of pol ari zati on and gave " theoreti cal expres s i ons for some of them : charge transfer ( acti vation ) ; overvol tage at the l i qu i d sodi um/sol i d el ectrol yte and el ectrol yte/-porous el ectrode i nterfaces ; effect i ve ohmi c res i stance ; and mass transfer through the porous el ectrode . Weber [ 1 97 J a l so cons i dered el ectri cal l osses resul t i ng from the fi n i te temperature di fference needed to susta i n the heat fl ow across the el ectro lyte . Heat i s absorbed at the porous e l ectrode pri nci pal ly from the sodi um evaporati on duri ng current fl ow and by therma l radi ation to the condenser . Equati on I I I -l was emp l oyed to account for these effects .

The effi ci ency under no l oad i s the quoti ent of the net work output W1 ( approxi mated by the i sothermal expan s i on of sodi um from P2 to P1 at T2 under quas i -equi l i bri um) l ess the work W2 necessary to c i rcul ate the l i q u i d sodi um ( Wl » W2 ) d i v i ded by the total heat i nput . The total heat i nput i s W1 pl us the l atent heat L of the sodi um vapor and the enthal py d i fference of the l i q u i d between T2 and T1 [6H : Cp (T2 - T1 ) J . Paras i t i c heat l osses ( Q1 oss ) s houl d a l so be i nc l uded i n the tota l heat i nput . Thus the effi c i ency n i s

n = Wl + L + 6H + Q1 osses Under l oad , the express s i on

Wl + L + 6H + Q1 0sses I V-4

IV-5

represents the effi ci ency of the system . Ql osses i nc l udes rad i ati on 4 4 l osses of the form Qr = 0(T2 - Tl ) / z , where 0 = Stefan-Bol tzman constant ,

1.57

S=�I I*I __________________________ T_R_--;-4_16_

z = effecti ve radi ati on res i stance , and conducti on l osses Qc = KL x (T2 - T1 ) . Fi gure I V-4 s hows some cal cu l ated power-effi c i ency performance curves as a functi on of T2 , u s i ng an expres s i on for effi ci ency of the type of Eq . I V- 5 [ 1 99 ] .

Fi gure I V- 5 shows experimental vol tage-current curves for several val ues of T2 as wel l as curves cal cul ated from Eq . IV- 5 , wh i ch does not take el ectrode pol arization i nto account . Interfac i a l po l ari zation i s respons i bl e for most o f the d i fferences between cal cu l ated and experi -menta l curves .

Fi gure I V-6 i s a d i a gram of an operati ng sodi um heat eng i ne us i ng a S II - a1 umi na tube . Cel l s of th i s type as wel l as others wi th i nverted geometry were used i n the study of the eng i ne performance . The i n i ti a l performance ( 1 975 ) was 0 . 2 W/cm2 , wi th an overal l effi ci ency of �1 0% [ 1 96] and short l i fetimes (�ne week ) . Thi s performance was improved to �0 . 7 W/cm2 , wi th an overal l effi c i ency of �20%, the performance bei ng l i mi ted mostly by i nterfaci a l pol ari zation [1 97 ] . I t appears that power outputs of 1 W/cm2 are feas i bl e wi th thi s system [201 ] . For such effi c i encies , these devi ces are very l i ght (�30 kW/ 1 00 1 b ) compared to turbi nes ( 30 kW/ 750- 1 000 1 b ) [201 ] .

The stabi l i ty of the S I I - a1 umi na does not seem to be a l i mi t i n g factor on the future appl i cabi l i ty of the sodi um heat eng i ne a s a stati c and modul ar power source . Careful preparati on and water excl us i on i n handl i ng may achi eve stabi l i ty a t temperatures a s h i gh a s �1 000 ° C . Tubes wi th thi nner wa 1 1 s are necessary to reduce the __ spec i fi c surface res i sti v i ty so that l arger power outputs can be dra i ned from the system . References 202 and 203 are rel evant rev i ews on S-a1 umi na ( see a l so

158

!;::�I IIfI ___________________________________________________________ T_R_-_4_1_6

.Specific Power \ (W/cm2) .

Figure IV-4. Calcula�ed P�wer-Efficiency Perform�nce of the Sodium Heat Engine for Various Values of T2 [1 99]

Q) '" .'9 o >

Ro = 0. 1 8 ohm cm2;Tl = 200°C; z = 20 (see Eq. I V- 5')

1 .0 watt

Specific Current (A/cm2)

Figure IV-S . .. Experimental (e) and Calculated (-) (Eq. jV�.�) Voltage �Current Curves for the Sodium I Heat Engineas--ci Function-of T2-[1 99]

159

S=�I I;."'\ _______________________ TR_-_4_16

-� �=�

;Porous

Electrode

Electrical I nsu lator

Condenser

Ceramic Tu be ---t----Jl'Il

Liquid Sod ium

.... _ Sodium Vapor

+

E-M Pump

\ Figure IV-,5. Schematic Diagram of the, Sodium Heat E'ngine Cells with f3"-Alumina Tubes [199]

160

!;::�I r�4Ir __________________________________________________ �T�R�-�4�1�6

Ref . 1 64 ) . i nterface .

Another source of probl ems i s the el ectrolyte/porous el ectrode Mol ybdenum fi l ms (mi crons th i ck ) can be depo s i ted on 8 1 1 -

a l umi na by chemi cal vapor" depos i ti o n . These fi l ms have h i gh i n- pl ane e l ectri cal, conducti v i ty , good sodi um vapor permeabi l i ty , strong adhesi on to the sol i d e l ectrolyte , good h i gh- temperature stabi l i ty ( tested for �1 000 hours wi thout deteri orati on ) , and an expans i on coeffi ci ent that matches that of 8 1 1 -a l umi na . One of the probl ems of these el ectrodes i s that recrysta l l i zation at 900- 1 000°C l eads to a materi al l ess permeabl e to sodi um vapor . Co l e , Weber , and Kunt [201 J are i nvesti gat i ng other methods of preparation of these e l ectrodes ( coati ng by sputteri ng ) . Radi ati on and conducti on l osses shou l d be kept smal l i f effi c i enc i es approachi ng Carnot cycl e are to be achi eved . Neg l ecti ng heat l osses and pol ari zation at the e l ectrode s , effi c i enci es as hi gh as 46 . 7% are ca l cu -l ated when the Carnot effi c i ency i s 50% ( Tl = 227°C , T2 = 727°C ) .

I V . 2 MuLTI PLE C ELLS The el ectrochemi cal heat eng i nes descri bed in thi s secti on are

composed of at l east two cl osed cel l s operati ng at the same temperature ( for di fferent temperatures , see Secti on V) and connected i n el ectri cal oppos i ti on . These are the Type 5 cel l s descri bed i n the Introducti on . -The e l ectrochemi cal reacti ons i nvol ve one gaseous reactant . The engi ne produces external work on a l oad when , by some mechan i ca l means ( e . g . , by coo l i ng one trap or one col d fi nger associ ated wi th the cel l s ) , the equ i l i bri um part ia l pressure of th i s gaseous reactant is made to be d i fferent i n both cel l s . Th i s di fference in part i al pressure wi l l dri ve the react ion of the gaseous substance and wi l l have a mi n imal effect on

16 1

TR-416 S=�I I_I -----------------------------

the condensed phases of the el ectrochemi cal cel l s . The net dri v i n g force o f the cycl e i s i ndependent o f any property o f the e l ectro lyte at open c i rcu i t . If the el ectrode ki netics are fast , l i ttl e po l ari zation s houl d be observed under current drai n . These systems run in cycl es . After di scharge of one cel l and consequent charge of the other , and performance of el ectri cal work on the externa l c i rcui t , it i s necessary to coo l down the trap ( or fi nger ) ori g i nal l y hot and to heat the one ori g i na l l y col d , thus revers i ng the rol e of the el ectrochemi cal cel l s . I n pri nci p l e , one can concei ve of th i s type .of regenerati on be i ng appl i cabl e whenever the acti vi ty of a chemi cal substance i n one of the cel l s can be made di fferent from the acti v i ty of that substance i n the other cel l . I n fact , the type of regeneration performed i n these cel l s i s a parti cul ar case of the el ectrotherma l regeneration deta i l ed i n Sect i on V .

These cel l s were f irst reported by Anderson , Greenberg , and Adams [ 1 49J at Lockheed , who properly refer to th i s type of regeneration as e l ectro lys i s at l ow press ure . More recently , El l i ott [204-209 J at Los Al amos has extended the study of these engi nes wi th a more practi cal cel l des i gn ( cf . Ref . 3) wi thout mas s transfer from the battery stacks .

The cel l proposed by Anderson , Greenberg , and Adams [ 1 4 9 J i s shown­i n Fi g . I V-7 . The systems envi s i oned for th i s type of regenerati on were Pb l Pb I2 1 I 2 or Cd I Cd I 2 I I 2 . The part ia l pres sures of iodi ne vapor are kept di fferent by the di fferent temperatures in the traps , one at Tl and the other at T2 , the temperatures of the cel l s . Iodj ne i s at l ow equ i l l ­bri um parti a l press ure i n the col d trap a t Tl and , therefore , a t the

162

TR-4 16 S=�I '.' ------------------------------� �

T,

L

Figure IV-7. Low-Pressure Electrolysis Apparatus [149]

163

/.�, TR-416 S=�I I_I ------------------------------

l eft-hand s i de of the cel l , wh i ch di scharges i od i ne vapor . The condensed i od i ne i s transferred v i a a ' nonreturn val ve to the hot trap at T2 , where i t vapori zes and i s i on i zed at the ri ght el ectrode . Therefore , the system must be run i n cycl es , wi th the trap at the l eft be i ng fi rst col d , thus condens i ng 12 , and then heated so that the cel l reverses . El ectro lys i s of Cd I 2 at 550°C i nto a l i qu i d n i trogen trap for the i od i ne gave currents up to 2 A/cm2 with a coarse- pore carbon el ectrode . However , the cou1 0mb i c effi c i ency was l ow due to the h i gh sol ubi l i ty of cadmi um i n mol ten cadmi um i od i de ( 30-40% ) . Reference 1 49 deta i l s the ana lys i s o f the effi c i enc ies o f these engi nes . One of the maj or l imi tati ons to a h i gh thermodynami c effi ci ency i s the need for a hi gh T2 but not too h i gh an equ i l i bri um pressure ( to avo i d seal i ng probl ems ) . Ki neti cs probl ems at the gas e l ectrode are l i ke ly to be a very important source of l i mi tati on of th i s type of cel l s , as wel l as di fferences i n current effi ci enc i es between the two el ectrochemi cal cel l s . A seri ous shortcom-i ng of th i s type of cel l , u s i ng i od i ne or di atom ic gases , i s the l ow vol tage obtai ned per cel l , wh i ch requ ires stacki ng of several seri es-connected cel l s to obta i n practi cal vol tages ( see a l so Ref . 3 ) .

The cel l proposed by El l i ott [204-209J i s shown i n Fi gs . IV-8 and -

I V- 9 . The s i ngl e l aboratory cel l s are made on a Pyrex cup wi th tungsten el ectri cal contact ( dense graph i te a l so can be used ) hol d i n g three l ayers of materi al s . The top l ayer i s a porous graph i te di sc , the i od i ne e l ectrode . The mi ddl e l ayer i s z i rcon i a fel t impregnated wi th mol ten el ectrolyte . The mi ddl e l ayer supports the graphi te di sc and prevents contact between the gas el ectrode and the bottom l ayer of

164

TR-4 16 S=::S1 If."\ -------------------------� ���

Agl-KI-(Kb)

on Porous Graphite (Liquid Plus Solid)

�H- Agi on Glass Tape (Solids)

��,- Ag on Porous Graphite (Solids)

Dense Graphite Cup Tungsten ConneCtor

Figure IV-S. Laboratory Cell Li/1 2 [207]

Discharge T, = T2 Charge T, > T 2 (charge < (discharge (external '> 0

Evacuated Batteries in Pai rs

Cell Stack --ilii�t::::1il

An od e --1'''':;;'--1 Leads

Condensers Across Anode Leads

Cathode Leads

Figure lV-g. Multiple-Cell Electro.chemical Heat Engine [206]

165

- TR-4 16 S=�I I.I ----------------------------� ��

ni ckel fel t hol di ng the metal el ectrode . Prel i mi nary studi es were made for the systems Ag l a-Ag 1 I 12 ; Ag I Ag 1 - K 1 ( mo l ten eutecti c ) 1 1 2 ; and L i l mo l ten a l ka l i metal i odi des l I 2 ' A proposed mu l ti pl e cel l el ectrochemi ca l heat eng i ne ( l aboratory scal e ) i s' s hown i n Fi g . IV-9 . References 207-208 de­ta i l a proposed practical cel l des i gn for two stacks of 1 00 cel l s eac h .

The Ag I I 2 cel l wi th the sol i d e l ectro lyte a-Ag I showed pol ari zation even at smal l dra i ns [204 J . Therefore the sol i d el ectro lyte was re­p l aced by the eutecti c Ag I - K I ( 238°C ) . The el ectrode ki net i cs were faster i n the mol ten e l ectro lyte than i n the sol i d el ectrolyte , but the i nternal res i stances were sti l l h i gh ( 1 2 ohms on di scharge and 36 ohms on charge ) .

The L i 1 1 2 cel l was a l so tested i n the l aboratory scal e [206-208 J . El l i ott [206J reported that these cel l s have i nternal res i stances of 0 . 3 -0 . 8 ohm/cm2 , and he i nd i cated that these res i stances rema i ned constant from 0 . 25 A/ cm2 ( d i scharge ) to 0 . 80 A/cm2 ( charge ) (wi th Tl = 350 °C and T2 = 25°C ) . The net vol tage obtai ned at open ci rcu i t was 0 . 29 V . Ass umi ng that pol ari zati on effects are smal l , El l i ott i ndi cated that 0 . 23 V cou l d be obta i ned at 0 . 1 A/cm2 . Unfortunate ly , he presented few experimental detai l s or resul ts .

Appl i cati ons projected for these heat eng i nes i nvol ve coupl i ng energy convers i on and storage , whi ch i s one of the un i que qual i ti es of these systems (as compared wi th systems reported in Secti on I V . l ) . These i nteresti ng app l i cati ons are thoroughly descri bed i n Refs . 205-208 .

Other systems proposed i nc l ude add i ti onal al kal i metal systems , wh i ch are s upposed to g i ve h i gher effi c i enc i es but a l so operate at

166

S=.�I I:.I ________________________ T�R_-4�1_6 -� .

h i gher temperatures . The patent l i terature descri bes the proposed systems [209 J .

I V . 3 SUMMARY AND D I SCUSS ION OF TRES TYPES 5 AND 7

, Tabl e S- l shows that the sodi um heat eng i ne ( Type 7 engi ne ) i s the TRES of h i ghest power demonstrated to date . I n fact , th i s i s a very i nteresti ng approach to TRES wi th no need for chemi cal regenerati on steps but wi th a major l imi ti ng requ i.rement--a proper sol i d el ectro lyte that i s a s tabl e s uperi on i c conductor at the des i red temperature range . Mos t of the good sol i d e l ectrol ytes operate at el evated temperatures . For thi s reason , research of sol i d el ectrolytes for operati on at l ower temperatures shou l d be strongly encouraged .

The el ectro lys i s at reduced pres sures ( Type 5 ) has mass transfer probl ems . Very l ow power outputs have been obtai ned wi th the systems

. attempted to date .

16 7

�'('" 1 68

�'II_I _________________________ T�.R_-_4_1_6

SECTION V

COUPLED THERMAL AND ELECTROLYTIC REGENERATION: GENERAL

In the electrothermal regeneration mode (see Fig. V-l), the elec -

�rochemical reaction products of cell 1 at Tl

C ) C+ + e

":

A + e :;. A-

C + A = C+

A-V-l

are regenerated by electrolysis in cell 2 at a temperature T2 r Tl:

C+ -+ e :;. C

A- A + e -

:;.

C+A- = C + A V-2

The ce 11 s a re connected to oppose e'ach other e 1 ectri ca l1y (back to

back). The operation of the energy converter is made continuous by

sending C+

A- (the electrolyte), formed in the galvanic cell reaction at

T, (previously taken to T2), to the electrolyzer. The galvanic cell

reactions then drive the electrolysis of C+A- into C and A, which are

taken to Tl and returned to the galvanic cell l,.completing the closed­

cycle operation. This TRES Type 4 is discussed in the Introduction.

If Tl = T2, the operation is similar to that of a secondary re­

chargeable battery with only electrolytic regeneration taking place and

with a net OCV of zero. If Tl r T2, electrothermal regeneration is

operative. The basic requirement is that the cell voltage- at T2 be

less than that at T, (V2 < Vl). Thus, a fraction of the cell voltage at

169

/.� TR-4 16 S=�I I_I -----------------------------

Tl i s used to perform the regenerati on of the ori gi nal reactants at T2 , and the rema i ni ng vol tage i s used to perform useful el ectri cal work i n the external l oad . I f T2 > Tl , the el ectrothermal regeneration i s be i ng performed by el ectrolys i s at a hi�her temperature than that of the gal vani c cel l , and the free energy of formati on of C+A- i s l ess negati ve wi th i ncreased temperature . I f T2 < Tl , the regeneration i s performed by e l ectrolys i s at a l ower temperature . The s i gn of (dE/dT) p determi nes the type, of regenerati o n . Examp l es i n wh i ch T2 > Tl are more common i n the l i terature .

Another important requ i rement for thi s type of regeneration i s that -+ + - - -+ -the el ectrode reacti ons C + C + e and A + e + A be capabl e of sus-

ta i ni ng h i gh c�rrent dens i ti es ( h i gh exchange currents , l ow acti vat ion overpotenti a l s , l ow el ectrode pol ari zation effects i n general , and l ow i nternal ce l l res i stance) and that they present h i gh coul omb i c effi ci -enc i es ; i . e . , the same characteri st ics as a secondary battery .

There i s a temperature at wh i ch regeneration ( react i on V-2 above ) starts to take p l ace spontaneous l y [6G ( Tx ) = OJ , wi th no need for el ec­trolys i s ( at h i gher or l ower temperatures ) , An exampl e of th i s type of spontaneous regeneration is Case l s cel l [69J , descri bed in Secti on 1 . 1 . 3 . 4 , in wh i ch s pontaneous chemi cal regenerati on takes pl ace at a l ower temperature than does the ga l van i c cel l operati on . McCu l ly [21 0 J descri bes another examp l e i n the oppos i te di recti on based on the reac­ti ons of the fl uori des of U ( V I ) or Ce ( IV ) and AsF3 , I f T > Tx ' the resu l ti ng gal van i c cel l wi l l generate power to the external l oad and regenerate the reactants of the l ow temperature ga l van ic cel l ( see Section V-2 ) .

170

S=�I I*I __________________________ T_R_-4_1_6_

I n pri nc i p l e , another poss i bi l i ty for operati ng el ectrothermal ly regenerati ve systems wi thout transferri ng reactants and products from one cel l to the other , as i n Fi g . V- l , i s to reverse the operati ng temperatures of the two cel l s . After the el ectrolys i s at T2 i s com­pl eted , wi th the d i s�harge of the cel l at Tl ( and concomi tant el ectri cal work produced i n the external l oad ) one can envi s i on heati ng cel l 1 to T2 and cool i ng cel l 2 to Tl , thus revers i ng the rol es of the two cel l s

. and. o bta i n i ng work on th� e l ectri cal l oad peri odi cal l y . I n 1 958 , Yeager [7 ] suggested that the coupl i ng o f thermal and

el ectrolyti c modes of regenerati o n , as wel l as thermal regenerati on a l one , shoul d be studi ed . Si nce then , several of these coupl ed el ectro­l yt i c and thermal regenerati on engi nes have been studi ed ; they have been cal l ed " doubl e thermogal vani c cel l I I [59 , 51 ] , " el ectrotherma1 1y regenerati ve transducers I I [2 1 1 - 21 3 ] , and " el ectrochemi cal heat engi nes " [204-209J . Secti on I V . 2 di scusses a parti cul ar case cif the el ectro-thermal regeneration i n wh i ch Tl can be i denti cal to T2 and the work i s produced by varyi ng the part i a l equ i l i bri um pressure o f a gaseous work­i ng el ectroacti ve fl u i d by phys i cal means .

The effi c i ency of the el ectrotherma l ly regenerati ve system i s a l so Carnot l i mi ted , as i s the therma l regenerati on . For a system i n wh i ch ­T2 > Tl , the effi ci ency i s ( for �Cp = 0 )

n = = = = V-3

where E i i s the thermodynamic cel l potentia l at Ti ' and Q2 is the heat per mol e of reactant i ntroduced i nto the di ssoci ati on un i t at T2 . Hesson and Sh i mota ke [46] anal yzed the effi ci ency of th i s mode and gave ex-

1 7 1

TR-4 16 S=�I I.I -----------------------------� �

CA - C + A T2 Heat

EXChangerr CA , R1 I I -t- I L J

C - C+ + e-A + e- - A-

.. Figure V-1 . Generic Scheme for Electrothermal Re.generation

172

S=�I I_I __________________________ T_R_-_4_16_

pres s i ons for the overal l effi c i ency of the system ( the product of the gal vani c cel l el ectr ical effi ci ency and the regenerator el ectri cal effi c i ency ) as a function of El , E2 , 1 1 , 1 2 , and ( dE2/dT ) p ' Anderson , Greenberg , and Adams [50 , 51 J have al so anal yzed effi ci enc i es of these systems for cases i n wh i ch �Cp = 0 and �Cp f O . . Even when �Cp f 0 the domi nant term i n these effi c i ency expressi ons i s sti l l ( T2 - Tl )/T2 .

Compared to thermal regenerati on , el ectrothermal regeneration has the adva.ntage of an i nherently s impl e separatton of reactants . In the systems i n wh i ch there i s a fl ow of reactants and products , the com­p l ex i ty of the dev i ce des i gn i s s i mi l ar to that encountered i n thermal regenerati on . Di sadvantages of the el ectrothermal regeneration are pol ari zati on at the el ectrodes i n the el ectrol yzer and poss i bl e mater­i a l s prcibl ems at the h i gher temperatures . These di sadvantages have paral l e l s i n the thermal regeneration i n the s l ow rates of thermal decompos i t i on and materi a l s probl ems i n the regenerator .

V . l H I GH TEMPERATURE ELECTROLYS I S

V . l . l Mol ten Sal t Medi a

I n the l ate fi ft i es , two groups i ndependently pursued the el ectro­thermal regenerati on method appl i ed to metal l metal ha l i de l hal ogen sys­tems as i l l ustrated by Fi g . V- l ; i . e . , the col d ga l vani c cel l reacti on product fl ows to the h i gh temperature el ectro lyzer ( s ) , where the re­actants of the col d gal van i c cel l are regenerated el ectro lyti cal l y , cool ed down , and transferred back to the col d gal van-� c cel l . The system Na ( � ) l mo l ten NaCl I C1 2 was studi ed at the Del co- Remy D i v i s ion of General Motors Corp . by Landers , Smi th , and Weaver [21 1 J i n 1 959 and was descri bed

173

•

'" TR-4 16 S=�I I_I -----------------�------------'-

by these authors i n the patent l i terature [ 2 J for a satel l i te power sys­tem [2 1 2 J . Later , Weaver [21 3 J stud i ed the same system more thorough ly i n order to i nvesti gate the feas i bi l i ty of th i s system a s a power source for armored veh i c l es . The systems Li I L i Cl ( � ) I C1 2 and L i I L i I ( � ) I I2 were ' i n i ti a l l y i nvesti gated . Research on the Li I L i Cl I C1 2 system was a l so i ndependently performed at the Al l i son D i v i s ion of General Motors by Swi n kel s [21 4 , 21 5 J , but the maj or emphas i s was on the rechargeabi l i ty of the i;>attery . system.

At Lockheed Ai rcraft Corp . the research on el ectrothermal regener­ati on descri bed by Anderson , Greenberg , and Adams [50 , 51 ] was appl i ed to the systems Pb I Pb I2 I I 2 ' Cd I Cd I2 I I 2 ' and L i I L i I i I 2 ' as a conti nuation of the prev i ous efforts on regenerati ve systems . That work started wi th· the uns uccessful work on thermal regeneration ( see Secti on I I I . l . 2 . 1 ) and evo l ved i nto the thermogal van i c systems ( see Sect i on 1 1 1 . 1 . 1 ) and then i nto the doubl e thermogal van i c or el ectrothermal regenerati on between 1 959- 1 962 ( cf . Ref . 3 ) .

At the Del co- Remy Di vi s ion [2 1 2 , 21 3 J the cho i ce of systems was based on el ectrolytes that exhi b i ted a wi de l i qu i dus range and gal van i c cel l reacti ons y ie l di ng the el ectrolyte . Thi s ga l van i c cel l reacti on shoul d have a s u i tabl e decrease of OCV wi th i ncreased temperatures sucn that the e l ectro lyzer at a h i gher temperature woul d consume a fracti on of the cel l vol tage produced at the l ower temperature . Another requ i s i te was the abi l i ty of the gal vani c cel l reacti on to susta i n l arge current dens i ti es on charge and d i s charge over a wi de range pf temperatures , wi th good cou l omb i c eff ici enc i es i n both proces ses . The wi der the l i qu i dus range of the sol vent chosen , the l arger i s the Carnot effi c i -

174

- TR-4 16 S=�I I.I ----------------------------=-----­-� �

ency that can be expected from the system. The Na l NaC1 I C1 2 system ( NaC1 : m . p . 801 ° C ; b . p . �1 450 °C ) meets most of these cri teri a . " The cel l exhi bi ted a h i gh OCV , e . g . , 3 . 24 V at 827 ° C , wh i ch decreased wi th the i ncrease i n temperature , e . g . , to 3 . 1 4 V at the sodi um mel ti ng po i nt ( 880° C ) a nd to 2 . 55 V at 1 220°C ( a di fference i n OCV of 0 . 7 V for a 390°C temperature di fference ) ; one woul d anti c i pate a l arger vol tage di fference at h i gher temperature di fferenti a l s . These cel l s were studi ed on charg� and, on di scharge j n the .827-J 057°C r:ange and showed the abi 1 i ty to be charged and d i scharged at h i g h current dens i t i es ( e . g . , di scharge currents above 3 . 5 A/cm2 at 827 ° C , with IR pol ari zation on ly at current dens i ti es above 4 . 3 A/cm2 ) .

The best 1 aborat�ry des i gn tested had a commerci a l a l umi na U tube as the cel l body . The sodi um el ectrode was made from an i ron or n i c ke l tube fl ared at the end , wi th a porous metal di s k wel ded on the fl are . The ch l ori ne e l ectrode was hol l ow graphi t i zed carbon conta i ned i n one of the arms of the U tube . The cel l was not seal ed but kept i n an argon b l anket . The sodi um uti l i zation i n the fi ve cel l s tested averaged only 40% . D i ssol ution of sodi um i n the mel t was d imi n i shed by reduc i n g the pore s i ze of the i ron gri d el ectrode . The faradai c effi c i enci es were too l ow to be accepta bl e i n a practi cal cel l . The abi l i ty of these cel l s to undergo charge ( e1 ectro1 yze� reacti on ) and d i scharge ( ga l van i c cel l reacti o n ) over a wi de temperature range was ta ken a s a part ia l feas i bi l i ty demonstration of the concept of the el ectrotherma l transduc­er . No compl ete coupl i ng of the gal van i c cel l and e] ectro 1yzer , wi th the appropri ate fl ow of materi al s , was tested . Some system des i gns are g i ven i n Weaver ' s report [21 3 J . The systems l i l l i C1 I C1 2 and li I l i I I I 2

175

S=�I I.I ______________________ ----=T::.....:R..:---_4-'-16----.

-� �

were i nvesti gated . The former gave 3 . 52 V OCV at 650 ° C , good l oad vol t­ages , and current dens i ti es as h i gh as 1 . 9 A/cm2 , wi th on ly I R po l ari za­t i on . Other stud ies of these systems were. made by Swi nke l s [21 4 , 21 5 J . The i od i de system gave 2 . 42 V at 473 ° C . Both the l i th i um i od i de and l i th i um chl ori de systems cou l d be operated at l ower temperatures than the sodi um ch l ori de system.

The work at Lockheed [50 �5 1 ; see a l so Ref . 3J was performed wi th cel l s of th� type descri bed i n S�<:_tiOrl- 1 . 1 . 2 . 1L For the Li I Li 1 I 1 2 system at 500° C , the OCV was 2 . 5 V , c l ose to the theoreti cal val ue . The current-vo l tage curves showed that current dens i ti es of 320 mA/cm2 cou l d be obta i ned at 1 . 5 V . About 2 mol e % of l i th i um di ssol ved i n the l i thi um i od i de at 500 ° C , thus decrea s i n g the cou1 omb i c effi c i ency . The el ectrol ­yses of Cd I 2 ( 450° C ) and PbI 2 ( 81 5 ° C ) showed onl y I R pol arizati on . However , i n the case of Cd 1 2 , the sol ubi l i ty of cadmi um was very h i gh , l eadi ng to a hi gh contri buti on of e l ectroni c conducti vi ty , thus decreas­i ng the cou1 omb i c eff ic i ency . The mol ten Pbl2 d i d not di ssol ve Pb appreci ably and cou1 ombi c effi c i enc i es of 1 00% were achi eved .

A thorough theoret i cal ana lys i s of the performance of the best can­d i dates for el ectrotherma l regenerati o n , L i I and Pb I2 [50 , 51 J , was per­formed and i s g i ven i n deta i l in Ref . 5 1 . The Carnot effi ci enc i es and -T2 for L i I and Pb I2 were 50 . 6% and 40 . 1 % , and 1 1 70°C and 870°C , res pec­ti ve1y: The maxi mum operationa l effi c i enc ies wi th maxi mum power to l oad at these temperatures , but wi thout tak i n g i nto account heat exchanger l osses , were 1 8% and 1 5% for L i I and Pb I 2 , respecti v�ly . I n pract i ce , l ower val ues woul d obv i ous ly be obta i ned .

W i th these systems [50 , 51 , 2 1 2 , 21 3J the conti nuous operati on as

176

S=�I I*I _______________

__________ _=T:..::R.::...�_=4:...:.1::...6 _

energy converters wou l d be achi eved by sendi ng the el ectrolyte , heated to T2 , to the e l ectro lyzer at a proper rate , where the reactants of the col d ga l van i c cel l are regenerated at T2 , cool ed to Tl , and redi ­rected to the col d gal van i c cel l . Heat exchanger i neffi ci enc i es l ead to i rrevers i b l e l osses . However , easy separati on of reactants and products i s i nherent i n these systems . S i nce the gal van ic cel l i s a battery , energy can be stored i n these systems . The batte.ri es can be operated temporari ly . for. power generation-;-- ·in audi ti on to operati on as energy converters . The theoreti cal operati ona l effi � i enci es [51 ] are hi gher , in general , than the correspondi ng effi ci enci es for conventi onal thermo­el ectri c devi ces ( o r thermogal van i c cel l s ) , i n wh i ch the el ectri c and thermal conducti on paths cannot be separated . Most probl ems common to a l l el ectrochemical devi ces uti l i z i ng gas el ectrodes ( p l ugg i n g or fl ood­i ng of the e l ectrodes ) woul d obv i ous ly be encountered . F i na l l y , because the current effi c i enc ies of the el ectro lyzers are l ower than those of the gal van i c cel l s , one can envi s i on the need for more than one e l ectro lys i s cel l coupl ed wi th a battery to ach i eve practi cal regeneration rates .

I n pri nc i p l e , one can concei ve a di fferent approach to the el ectro-thermal regenerati on that does not requ i re materi al s transfer from one

-

cel l to the other but only heat transfer for peri odi cal power generati on as an energy converter and a l so for power generation as a secondary battery , wi th energy storage .

V . 1 . 2 Aqueous �1ed ia

Hammond and Ri sen [2 1 6 ] recentl y descri bed an examp l e of a poten­t ia l el ectrothermal ly regenerati ve system based on the reacti ons :

177

- TR-4 16 S=�I I.I --------------------------'�---

- � ��

( ) 2+ . - + V-4 Cu NH3 4 + e � Cu ( NH3 ) 2 + 2NH3 ( Fe ( CN )�- � ( Fe ( CN )�- + e- V-5

CU ( NH3 ) �+ + Fe ( CN )�- = CU ( NH3 )� + Fe ( CN ) �- + 2NH3 V-6

These - reacti ons were sel ected due to the i r l arge mol ar entropy changes ( 6S = -32 cal /degree mol e for V-4 and 6S = -28 cal /degree mol e for V-5 ) . The two hal f-reacti ons ( V-4 and V- 5 ) must be separated by semi permeable · . .---in.embranes . Open-circu i t - vb rtages-o:s- a function -of temperature are gi ven i n Tab l e V- l . These data were obta i ned wi th J aboratory cel l s cons i st i ng of a beaker wi th rubber stoppers accommodat i ng el ectrodes (work i n g and reference ) , and two concentri c tubes wi th l -cm2 wi ndows i n thei r $ i des . The wi ndows were covered wi th cel ) ophane membranes ( previ ous ly exposed to Ba2+ and SO�- ) , wh i ch are more eas i l y permeabl e to s i ngly charged i ons than to mul ti p ly c harged i ons . Pl ati num e l ectrodes ( 1 cm2 ) were set to face the membranes . The cel l was pl aced i n a constant temperature bath , and the i nner and o uter sol uti ons were agi tated magneti cal l y . The mi ddl e compartment contai ned supporti ng el ectrol yte ( NH4Cl + BaC1 2 ) . With t ime , a redd i s h-brown prec i pi tate formed on the membrane i n the s i de adjacent to the Fe ( CN )�- sol ut ion and the res i stance of the cel l i ncreased . Tabl e V- l a l so shows the res i stances of the forward (+R ' ) and reverse ( - R ' ) react i ons . The studi es performed encompas sed the var­i ation of OCV wi th temperature and a few pol ari zation studi es up to 25 mA/cm2 . No current-vol tage curves were g i ven and resu l ts were i nter­preted i n terms of I R pol ari zati on onl y ( cf . Secti on- I I I . 3 ) .

178

S-�I � ____________________ �T_R_-4_1_6 _ - 11.11 - �

Tabl e V- 1 . CELL POTENT IALS AND RES I STANCES AS A FUNCTION OF TEMPERATURE FOR THE Fe ( CN )�-/ Fe ( CN )�- AND CU ( NH2 )�+/CU ( NH3 ); SYSTEMS [216t "

Temperature ( O C ) E ( V )

2 5 -0 . 51 8 30 ... 0 . 505 40 -0 . 475 60 -0 . 41 7 90 -0 . 330 30 -0 . 502

+R ' ( ohm m2 x 1 04 )

1 1 . 6 ..g . 1 7 . 4 5 . 5 3 . 4

1 7 . 2

I - -

- R ' ( ohm m2 x 1 04 )

1 1 . 4 8 . 3 7 . 0 5 . 3 3 . 4

1 2 . 3

aThe compos i ti on of the sol ution i n the i nner vessel i s 1 M ( NH4 ) 4Fe ( CN ) 6 ' 1 . 75 M NH4C1 , and 0 . 05 M K2S04 ; i t i s el ectrolyzed to equ imo 1 ar i n Fe ( I I ) and Fe ( I I I ) . The compos i ti on of the sol ution between the outer vessel and the beaker i s 0 . 75 M CuC1 2 · 2H20 , 2 . 0 M NH4C1 , 4 M NH3 , 0 . 05 M K2S04 , and 0 . 25 M metal Cu ( added under N2 ) to g i ve equ imo1 ar Cu ( I ) and CU ( I I ) . The i ntersti t ia l sol ution i s 2 . 0 M NH4C1 and 0 . 5 M BaC1 2 · 2H20 .

179

'" TR-416 S5'l1 '1l' -----------------------------

Power output den s i ti es for these cel l s were cal cul ated ass umi ng I R po l ari zati on on ly , and the authors cl a imed that 6 . 4 W/m2 i s feas i bl e for operati on between 30°C and 90 ° C . No esti mates of the vari ati on of i n-ternal cel l res i stance wi th time nor of pump i ng and heat exchanger requ i rements were made . Effi c i e�ci �s of 8% were c l a i med ( hal f of Carnot ) . Other system effi ci enci es were anal yzed theoreti cal l y . The authors proposed coupl i ng th i s type of energy converter to fl at-pl ate sol ar col l ectors for home heati ng, powe�. - -

V . l . 3 Hydrogen-Oxygen Fuel Cel l Coupl ed wi th H i gh Temperature Water El ectrolys i s and Rel ated Systems

I n thi s sect ion , papers that ana lyze the thermodynami c feas i bi l i ty of the coup l i ng of fuel cel l s and el ectro lyzers are bri efly revi ewed . Li terature perti nent to the present state of the art i n high temperature el ectrolys i s i s c i ted . W i th current technol ogy th i s coupl i ng does not appear to be feas i bl e .

Hs u and coworkers [2 1 7-2 1 9] and, more recently , Ste i n berg [220]

have proposed the combi nati on of fuel cel l s operating wi th hydrogen and oxygen at a l ow temperature , produc i ng water wh i c h cou l d be el ectrolyzed at a , h i gh temperature ( >l OOO°C ) . Thermodynami c cal cu l ations i ndi cate that the operation of the el ectro lyzer at h i gh temperatures shoul d requ i re a l ower vol tage than that furn i shed by the operation ,of the fuel cel l at l ower temperatures . Tabl e V-2 reproduces some thermodynami c re­su l ts ' of Ste i n berg [220] . The i deal net vol tage for ,operation of the e l ectro lys i s cel l at �1 200°C i s �0 . 4 V . Pol ari zation l osses decrease thi s val ue apprec i ab ly , presenti ng di fficu l t i es i n the uti l i zati on of thi s system in th i s regeneration mode .

1 80

S=�I I.I ___ � _____________________ ..:::..T=R_-4-=-1--,-6 -� ��

-- -- -'-- .---�.-- - '_-. - � - .,; '"::!-- �. --- . .. . . ;:$.: .:. ·l;: . •

'- .- - -� . '--:�-.--. ,-

Tabl e V- 2 . I DEAL EFF I C I ENCY FOR H2 -02 THERMOELECTROCHEM I CAL POWER CYCLE [220 1

El ec. .. �ro l.Y.tjc C�l l Fuel Cel l I deal Cycl e Temp : Te El ec . Heat Temp . El ec . Net Eff . ( % )

Cel l ilGf+ilGe t { O C ) ( O C ) Energy TeilSe Tf{ OC ) Energy Vol t . ilGe ilGf ( V ) TeilSe

25 . 0 298 . 2 +54 . 64 +3 . 1 6 298 . 2 -54. 64 0 . 000 0 . 0 - - -

226 . 8 500 +52 . 36 +5 . 9'--298 . 2 -54 . 64 0 . 049 38 . 6 726 . 8 1 000 +46 . 03 +1 3 . 1 8 - 298 . 2 -54 . 64 0 . 1 87 65 . 3

1 226 . 8 1 500 +39 . 26 +20 . 58 298 . 2 -54 . 64 0 . 333 74 . 7 1 726 . 8 2000 +32 . 31 +27 . 95 298 . 2 -54 . 64 0 . 485 80 . 1

The l i terature on h i gh temperature water el ectro lys i s focuses on the production of hydrogen and on the uti l i zation of waste heat from fus i on reactors ( temperatures at �1 400°C ) . The proceedi ngs of a work­s hop on water el ectrolysi s [221 J presents the devel opments i n th i s area up to 1 975 . Reference 222 conta i ns research resul ts regard i ng water e l ectrolys i s up to 1 978 . Reference 223 descri bes the status of the rel evant fuel cel l research i n 1 979 . The h i gh temperature water el ectro­lys i s uti l i zes sol i d e l ectrolytes , e . g , Zr02- Y203 . The search for appropri ate i nterconnect ing materia l s and s u i tabl e el ectrocatalysts [224J i s bei ng pursued acti vel y at several l aboratori es , e . g . , Brookhaven Nati onal Laboratory [225J and Westi nghouse Research [226J . I f the research succeeds i n reduc i n g the overvol tage , IR l osses , and materi al s probl ems , the coupl i n g concept may become feas i bl e .

1 8 1

Carnot Eff. ( % ) Te-Tf

Te

0 . 0 40 . 0 70 . 2 80 . 1 85 . 1

S=�I I.I ______________________ T_R_--=4�1 _=__6 _ -� ��

Stei nberg [220J a l so extended the thermodynami c cal cu l ati ons to other thermoel ectrochemi cal cyc l e s . One exampl e i s a H2-02-C1 2 system wh i ch cons i sts of ( 1 ) h i gh temperature water el ectrolys i s ; ( 2 ) oxi da­t i on of aqueous HCl by the oxygen , yie l d i ng chl ori ne ; and ( 3 ) l ow temperature fuel cel l recombi nation of the C1 2 wi th H2 . Tabi e - V-3 presents some cal cu l ated resu l ts for th i s system [220 J .

V . 2 FLUORI DES OF URAN I UM ( V I ) O R CERIUM ( I V ) AND ARSEN I �M ( I I I ) : , SPONTANEOUS CHARGE REACTION

- --Th e patent l i terature [2 1 0J contai ns two exampl es of gal vani c cel l s that can be recharged by h i gh temperature el ectro lys i s or , i f the temperature i s h i gh enough , are transformed i nto the reverse ga l van i c cel l s , wh i ch a l so act as power generators and regenerate the ori g i na l reactants of the l ow temperature gal van i c cel l s . Therefore , the cel l operates at one pol ar ity at l ow temperature and at the oppos i te pol ari ty at a h i gher temperature ; by d i scharge at the hi gher temperature the cel l i s regenerated to i t s ori g i nal el ectrochemi cal state .

These cel l s are ba sed on the redox coupl e AsF3/AsF5! for wh i c h the rel ati ve stabi l i ty of the two fl uori des rap i d ly changes wi th i ncreased temperature . The AsF3 i s more stabl e at �1 200 °C . By comb i n i n g th i s redox coup l e wi th UF5/UF6 o r CeF3/CeF4 , whi ch do not change rel ati ve stabi l i ty wi th temperature , McCu l ly [2 1 0J was abl e to make gal van i c cel l s exhi bi ti ng spontaneous regenerati on . At room temperature , the fol l owi ng d i scharge reaction occurs :

The reactants are separated by the sol i d el ectrolyte l ead fl uori de

1 82

- : � I

Tabl e V- 3 . I DEAL EFFIC I ENCY FOR H2 -02-HC1 ( aq ) THERMOELECTROTHERMAL POWER CYCLE [220]a

E1 ectro 1ytic Cefi Fuel Cel l H20 ( g ) = H2 ( g ) + 1 /2 02 ( g )/ __

. . : 1 /2 H2 ( g ) + 1 / 2 C1 2 ( g ) = 2HC1 ( aq )

t ( OC ) T ( K) LlGe TLlSe Cel l t ( O C ) T ( K) LlG Net I Net Ideal Carnot

Vol tage LlG , I dea l Cycl e Eff. ( V ) Vol t . ( V ) Eff . ( % ) ( % )

-3 1 . 33b i 25 . 0 298 . 2 54 . 64 3 . 1 6 1 . 1 85 25 298 . 2 4 . 01', 0 . 1 74 32 . 0 0 . 4 226 . 0 500 52 . 36 5 . 91 1 . 1 35 25 298 . 2 -31 . 33 5 . 1 5 1 0 . 223 37 . 0 40 .. 0 726 . 8 1 000 46 . 03 1 3 . 1 8 0 . 998 25 298 . 2 -31 . 33 8 . 32 0 . 360 47 . 4 70 . 2

1 226 . 8 1 500 39 . 26 20 . 58 0 . 851 25 298 . 2 - 31 . 33 1 1 . 70: 0 . 507 55 . 1 80 . 1 . 1 726 . 8 2000 32 . 31 27 . 95 0 . 700 25 298 . 2 -31 . 33 1 5 . 1 8 0 . 658 60 . 9 85 . 1

aWater e l ectro1yzers at h i gh temperature and aqueous hydrochl ori c ac i d and fuel cel l operation at l ow temperature . -

b l ntermedi ate reaction : HC1 ( a q ) + 1 /4 02 ( g ) = 1 / 2 H20 ( g ) + 1 /2 C1 2 ( g ) ; LlH298 2 = +1 0 . 95 . Ce l l vol tage = 1 . 359 V . ' .

In III "" -• II II

'\': ��

>-3 l:d I oj:>. ....... 0)

/. � , TR-4 16 S=�I I/.II _______________________ =--=..c:.._'__ - � �

(contai ni ng KF to i mprove the fl uori de ion conducti v i ty ) i n a sandwi ch

type of cel l ; the fl uori de i ons are transferred across the so l i d el ec­

trolyte and el ectr ical work i s performed on the externa l l oad . After

the cel l i s d i scharged at room temperature , the battery temperature i s

rai sed to ,\;900° C , a t wh i c h temperature the reverse react i on proce.eds

s pontaneous ly , generates power to the external l oad , and regenerates the

reactants of the ori g i nal col d cel l . Fi gure V-2 shows the OCV of these

cel l s as a functi on of the �emp���ture. Currant-vol tage or i nternal

cel l res i stance data were not gi ven for these cel l s [220J .

V . 3 THERMOCELL REGENERATORS

Greenberg , Tha l l er , and Weber [227 , 228J descri bed a combi nati on of

a gal van i c cel l i n mol ten sal t medi a and a theY-mocel l ( see - Secti on I I I ) ,

i n wh i ch the i nert el ectrodes are short-ci rcu i ted for the regeneration

of the reactants of the gal van i c cel l , provi ded that the thermopoten­

ti a l s devel oped are h i gher than the decompos i ti on vol tage of the sal t .

To avo i d the accumu l ation o f el ectro lys i s products and , therefore , the

devel opment of a back EMF whi ch can stop the decompos i ti on , the decom­

pos i ti on products are conti nuous ly removed from the thermocel l and re­

turned to the gal van i c cel l . Greenberg et a l . [227 , 228J c�l l ed thi s

type of combi nati on a I I regenerati ve , mol ten sal t , thermoel ectri c fuel

cel l l l ; the i r schemati c cel l d i agram i s shown i n Fi g . V-3 .

Laboratory cel l s were bui l t as shown i n F i g . V-4 , one wi th ZnC1 2 and another wi th SnC1 2 . The cel l el ectrodes B , C , a�d 0 were at 400°C

and an auxi l i ary heater was empl oyed to ra i se the temperature of the

compartment of el ectrode A to 500-600°C . For the ZnC1 2 cel l , a thermo-

184

- TR-4 16 S::�I I�.- �I -------------------------------------------------------------------------

-� �

UF. / UFs

Temperature ( K)

Figure V-2. OCV of Cells Composed of AsF3 and UF6 or CeF4 Separated by the Solid Electrolyte PbF4 [21 0]

Decomposition Product

External Short Circuit Decomposition Product

'-----+1- Molten-Salt Thermoelectric Converter

L+====�===- Decomposition Products React to Form Salt

Load

Figure V-3. Regenerative, Molten Salt, Thermoelectric Fuel Cell Schematic Diagram [227]

1 85

- TR-4 16 S=�I I;.-�I -----------------------------

-� �� 7

��D---4 i n . -------:�

Zinc Chloride

4 i n .

Heat-Resistant G lass

Auxi l iary Heater C

Figur� V-4. Laboratory Cell Scheme for a Thermocell Regenerator (Elec.trodes A, 8) Coupled with the Galvanic Cell (Electrodes C, D) [227]

186

S=�I I.I _______________________ ..;;;:..TR�-4=_=1=6_ -� �

gal van i c potent ia l was generated between el ectrodes A and C whi ch was

used to decompose the sal t wi th currents of 1 0-3- 1 0-2 rnA . Meta l l i c

parti cl es were vi s i bl e around el ectrode B , and s i nce the metal was

i nsol ubl e i n the mel t , i t prec i p i tated out towards el ectrode C . Around

e l ectrode A a yel l ow col or was vi s i bl e . The decompos i ti on products ,

therefore , di ffused to or prec i p i tated at el ectrodes D and C , wh i ch

di s pl ayed an OCV of 0 . 2 V . The short-ci rcu i t current between C and D

was 1 rnA. The ha l f�reaGti o.ns prop_osed-are :

El ectrode

A ( hot) B ( co l d )

C ( col d ) D ( co l d )

2Cl -2+ -Zn + 2e

Zn ;;. C1 2 + 2e-

;;. C1 2 + 2e-

;;. In

2+ -Zn + 2e ;;. 2Cl

Thermocel l Reacti ons

Gal van i c Cel l Reactions

The thermocel l s empl oyed had very l ow effi c i enci es , wh i ch mi ght be

i mproved to some extent by i ncreas i ng the el ectrode area and by us i ng

mol ten sal ts wi th h i gher thermoel ectri c powers ( see Secti on I I I ) .

V . 4 D I SCUSS ION OF TRES TYPE 4

Al though the coupl i ng of thermal and el ectro lyt i c regenerati on wa�

suggested i n 1 958 , very few systems have been attempted . The very h i gh

temperature systems that were tri ed exh i bi ted materia l s probl ems and l ow

coul omb i c effi c i enc i es . The systems descri bed i n Sec . I I and some of

Sec . I can conce i vably operate i n thi s regeneration mode . Thi s type of

regeneration has been expl ored to a l esser extent than thermal or el ec ­

tro lyti c regenerati on a l one , but i t can perhaps broaden the nature of

the systems to be i nvesti gated . Medi a operati ng at l ower temperatures

shoul d be i nvesti gated . 1 87

- -

188

S=�I I_I ___________________________ T

_R_-_4

_1 6

SECT I ON V I

CONCLU S I ONS AND RECOMMENDATI ONS

Most of the techni cal acti v i ti es descri bed in th i s revi ew were ei ther per­

formed or ori gi nated many years ago a n d , most i mportantl y , wi th spec i fi c ap­

pl i ed obj ecti ves . Mos t of the work was ai med at the use of nucl ear heat

sources , and much of it had as i ts ul timate goal the devel opment of space

power systems , for wh i ch wei gh t and zero gravi tati onal operati on were vi tal

consi derati on s .

I t appears to us that as a resu l t of si g n i ficant ti me con s trai nts imposed by

the proposed space uti l i zati on ti metabl e , much of the work i nvol ved certai n

p rel i mi nary experi ments fol l owed by a perhaps prematu re sel ecti on of a parti ­

cul ar system for devel opment. I n certa i n i n stances , devi ces were fabri cated

and tested pri or to the avai l abi l i ty of bas i c chemi cal and/or el ectrochemi cal

system i nformati on . I n some cases ( e . g . , the thermoga l vani c cel l s descri bed

i n Sec . II I ) , pri mari l y sc i enti fi c i nformati on was sought . In some i nstances ,

al bei t far fewer ( we fel t ) , dev i ces were fabri c ated based upon a known and

determi ned sci enti fi c ba se .

I t i s worth poi nti ng out that the reactor heat source uti l i zati on , whi ch was

one of the dri v i ng forces beh i nd much of th i s work , was based on regene rati on

at a hi gh temperature--usual l y 800°C or hi gher . Desi re for hi gh current effi ­

c i ency ten ded to resul t i n a system capabl e of operati ng over a severa l ­

hundred-degree temperature gradi ent. Systems operati ng at much l ower tempera­

tu res have not been thorough l y i nvesti gated but may wel l be rel evant to match­

i ng wi th sol ar heat sources . Our recommendati ons do not consti tute an offi ­

c i al pos i ti on on the part of the Sol ar Energy Research I n sti tu te . These re­

commendati ons are i ntended to suggest areas for wh i ch research i n ei ther

sci ence or engi neeri ng woul d have l ong-range benefi t to a TRES program . They

a re not i ntended as i ndi cati ons of proposed or i ntended devel opment programs

of a spec i fi c natu re . I n mak i ng these recommendati on s , we have not engaged i n

engi neeri ng or system eval uati on of parti cul a r concepts and , i n fac t , have

189

- TR-4 16 S=�I I_I -----------------------------

purposely avoi ded doi ng so . Our recommendati ons attempt to assess whether

past research may be worthy of renewed i nvesti gati on because of the avai l ab i l ­

i ty of new i nformati on or tool s and tec h n i ques and because of the very s i gn i f­

i cant changes i n perspecti ve posed by a terrestri al , sol a r heat source . We

hope the report and sub sequent eval uati on of both i t and our recommendati ons

wi l l resul t in new consi derati on of the sci ence i nvol ved in TRES systems .

For terrestri al appl i cati ons u s i ng sol ar heat sources , some general recommen­

dati ons concern i ng therma l l y regenerati ve el ectrochemi cal systems can be made :

• Systems operati nq at l ower tel1!Peratu res shou l d be i nvesti gated more - .. - - -

thorough l y • .

• The sea rch for pos s i bl e new types of TRES shoul d be conti nued. Mos t of

the obv i ous candi date systems for TRES have been tri e d ; however, the

devel opme nt of new concepts is possi bl e .

Our recommendati ons of areas of research to pursue .that a re rel ated to TRES

are very broad general i zati ons ( as opposed to spec i fi c system deci si on s ) • Ou r

assessment of the work reported and eval uated i n th i s document convi nces u s

thi s i s the most reasonabl e approac h , gi ven the constrai nts di scussed above .

• Mol ten Sa l t Chemi stry and El ectrochemi stry

A broad vi ew of the TRE S area i ndi cates the si gni fi cant attempts to em­

pl oy a wi de vari ety of mol ten sal t systems . Research i n th i s a rea i s

rather l i mi ted at th i s ti me i n the Uni ted States . New cl asses of mol ­

ten sal ts h ave been devel oped in the past years , but deta i l ed under­

s tandi ng of these systems i s l acki n g . Perti nent questi ons have -not

been addres sed·, such as " Are e l ectrode processes i ntri n s i ca 1 1 y fast at

el evated temperatures , s i mp l y by vi rtue of the hi gher temperature? I I

Recent devel opments i nvol v i ng consi derati ons of tran sport and structu re

i n mol ten sal ts suggest that i mportant , germa ne sci enti f i c i nformati on

n ow exi sts ( or can be obtai ned) that may not have been k n own or recog­

n i zed when much of the work reported here was·

performed . Whi l e a good

deal Qf the work reported . here had a resea rch componen t , the maj or

thrust of most of i t was devel opmental . Hence , addi ti onal

190

S=�I I_I ______________________ --.:T:..:R�-__=_4.:...:..16

understa ndi ng of mol ten sal t sy stems --and we mi ght wel l i nc l u de even

concentrated el ectrol ytes and hydrate mel ts--i s i mportant to th i s

a rea . Structura l , tran sport, el ectrochemi cal , and general phy s i cal

p roperti es , i ncl udi ng val i' d thermodynami c data , are a reas where

research woul d be appl i cabi l e to TRES devel opment.

• Sol i d-State Chemi s try

I n a number of i n stances one can observe the need for ei ther new mate­

ri al s ( parti cul arly stabl e , superi on i c conductors ) or a better under­

standi ng of exi sti ng materi al s , such as stabi l i zed zi rcon i a or

beta-al umi na . Research efforts i n these a reas a re very 1 i m i te d , par­

ti cul arl y wi th regard to Jl.t gll temperatu�e materi a l s , except for oxi de

transport. In a n umber of cases , materi a l s empl oyed as i on i c conduc­

tors in work reported here were used because they exi sted and were

readi l y avai l abl e , not because they were opti mal . Ki neti cs and mecha­

n i sms of el ectrochemi cal reacti ons tak i ng pl ace at the sol i d

el ectrol yte/el ectroacti ve materi a l /current col l ector i nterface al so

consti tute an area of i ntere s t . The recent devel opment of a l arge

n umber of su rface spectroscopi C techni ques provi des tool s that woul d

resul t i n new understandi n g .

A s a subset o f th i s area , i t appears that gas permeati on th rough metal s

coul d be studi e d , agai n mak i ng u se of recently devel oped surface tech­

n i ques to obta i n ba si c i n formati on of benefi t to an embryon i c TRES

program.

• Materi al s Sci ence

A qui ck gl ance th rough the body of th i s report makes obvi ous the mate-- -

ri al s probl ems �-conta i nment, bondi n g , corros i on--that pl agued numerous

acti v i t i e s ; and , obvi ousl y , the hi gher the temperatu re , the greater the

probl em . A broad program i n supporti ng materi al s research i s

recommen ded.

• Aqueous Systems and El ectrochemi stry under Extreme Condi ti ons

Whi l e the properti es of aqueo u s systems at room temperature are wel l

studi e d , acti v i ti es rel ated to TRES suggest that one may seek to empl oy

aqueous el ectrol yte systems at unusual l y hi gh temperatures and

1 9 1

S=�I I�I _______________________ -::...TR_----'4"'-1_6 -� .

pres sure s . Chemi cal and el ectrochemi cal research i n aqueous sy stems at

much over 100°C is not a wel l - studi ed area , bu t it i s one wh i c h woul d

be of broad i nte rest to th i s and other sol ar programs .

• El ectrochemi cal Engi neeri ng

The devel opment of any l ow vol tage source--and al l TRES are i nherentl y

l ow vol tage--ul ti mately woul d req u i re el ectrochemi cal engi neeri ng sup­

port in such areas as tran s port and battery formati on . El ectrochemi cal

engi neeri ng does not appear to have pai d much attenti on to the h i gh

temperature area except for certa i n spec i fi c process work , and general

con s i derati on of th i s p robl em area woul d be most useful . I t i s our

opi n i on that suc h- inpu t i s-requ ired for -useful systems anal y s i s .

• System Anal ys i s

To ou r knowl edge , no i ntegral system anal ys i s has been performed to as ­

sess the p racti cal i ty of a compl ete TRES . I n fac t , such an anal y s ; s

may not be pos s i bl e at th i s ti me , based on avai l abl e l i terature data .

For exampl e , i t i s not c l ear whether suffi c i ent and s i gn i f i �ant pol a�­

i zati on data for cel l s can be fou n d . Neverthel ess , i t may be worth

consi deri ng th i s approach for one or more of the systems revi ewe d ,

keep i ng fi rml y i n mi nd that n o el ectrochemi cal system ( i . e . , seri es

arrangement of cel l s ) has been i nvesti gated and even cel l data i s

meager i n most cases . We are unabl e to address the ul ti mate use of a

TRES ( i . e . , central - or di spersed-power appl i cati on ) , but c l early th i s

con s i derati on i s i mportant to an i ntegral system ana1 ys i s .

192

'" TR-4 16 S=�I I_I -----------------------------

SECT ION V I I

REFERENCES

1 . H . A . L i ebhafsky , " Regenerati ve el ectrochemi cal systems : an i ntroduct ion , " i n Regenerati ve EMF cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , 1 - 1 0 , Ameri can Chemi cal Soc i ety , Wash i ngton , D . C . ( 1 96 7 ) .

-

2 . Exampl es o f redox regenerati ve systems : ( a ) " Redox flow cel l devel opment and demonstrati on proj ect - cal endar year 1 97 7 , I I �eport- No .- NASA---=r�79067 prep-erred under Interagency agreement E (49-28 ) - 1 002 ( 1 97 9 ) . ( b ) G . C i pri o s , W . Erski ne , and P . G . Grimes , " Redox bul k energy storage system study , " vo l s . 1 and 2 , F i nal Reports , N 77-33608 and N 77-33609 ( 1 967 ) . ( c ) S . As i mura and Y . r�iyaki , " Redox-type fuel cel l . V I I . Po l ari za­t i on' characteri sti cs of the redox-type fuel cel l anode at fl ow-through porous carbon el ectrodes , " Denk i Kagaku , 40 , 50-54 ( 1 972 ) .

3 . L . G . Austi n , " Fuel cel l s : a revi ew of government sponsored. researc h , 1 950- 1 964 , " NASA-SP- 1 20 , pp . 1 53-203 .

4 . R . L . Kerr , " Regenerati ve Fuel Cel l s , II Proc . Performance Forecast Sel ect-ed Stati c Energy Convers ion Dev i ces r·1eet . AGARD , Propul s i on and Ener-

.

geti cs Panel , 29th , L i ege , Bel g i um 1 967 G . W . Sherman and L . Devo l , eds . , pp . 658- 71 1 ( 1 968 ) .

5 . Regenerati ve EMF Cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , Ameri can Chemi cal Soci ety , Wash i ngton , D . C . ( 1 967 ) .

6 . Exampl es of el ectrolyti cal l y regenerati ve systems : ( a ) J . �1cBreen , R . S . Yeo , A . Beaufrere , and S . Sri n i vasan , " Hydrcrgen:.;. ch l ori ne energy storage system , " Report BNL;.. 23 670 ( 1 977 ) . ( b ) L . Swette and G . L . Ho l l eck , " Hydrogen-n i ckel regenerati ve fuel cel l s , " Fi nal Report No . 339 , contract F3361 5- 73-C-2057 ( 1 974 ) . ( c ) B . �1 . Wi l ner , H . A . Frank , E . Fi ndl , and M . Kl e i n , " El ectrolyti cal ly regenerati ve hydrogen-oxygen fuel cel l s , " U . S . Patent 3 , 507 , 704 ( 1 970) .

7 . E . Yeager , " Fuel cel l s : bas i c cons i derati ons , "- Proc . 1 2th Ann . Battery Research and Devel opment Conf . , ed . by Power Source Di v i s i on , U . S . Army S i gna 1 Research and Devel opment Laboratory , Fort �1onmouth , New Jersey , pp . 2-4 ( 1 958 ) . .

193

TR-416 S=�I I.I ----------------------------� ��

8 . H . A . L i ebhafsky , l i The fue l cel l and the Carnot cycl e , 1i J . El ectroch'em . Soc . , 1 06 , 1 068- 71 ( 1 959 ) .

9 . A . J . deBethune , " Fuel cel l thermodynami cs , " J . El ectrochem. Soc . , 1 07 , 937-9 ( 1 960) .

1 0 . L . G . Au sti n i n " Handbook of fuel cell techno l ogy , II C . Berger , ed . , Prenti ce Ha l l , I nc . , New Jersey , pp . 36-41 ( 1 968 ) .

1 1 . J . B . Fri auf , "Thermodynami cs of thermal l y regenerated fue l cel l s , " J . Appl . Phys . , 1£, 61 6-20 ( 1 96 1 ) .

1 2 . R . E . Henderson , B . Agruss , and W . G . Capl e , " Resume of thermal ly regenerati ve fuel cel l s , l i Progr. i n Astronaut . Rocketry , 2, 41 1 -23 ( 1 961 ) . _ _ _ _

1 3 . J . Ki n g , J r . , F . A . Ludwi g , and J . J . Rowl ette , "General eval uati on of chemi cal s for regenerati ve fuel - cel l s-, " Progr. in Astronaut Rocketry , 2, 387-40 ( 1 96 1 ) .

1 4 . R . H . Snow , " Thermochemi cal and thermodynami c data on sel ected com­pounds for the chemi cal convers i on of waste heat to el ectri cal en­ergy , " Report No . AD 265- 376L , I l l i no i s I nsti tute of Technol ogy Research I n sti tute , Chi cago , I l l i noi s ; under contract NOw-60-0760-c ( 1 96 1 ) .

1 5 . G . Wurtzbacher, "Thermal l y regenerabl e fuel cel l s for convers ion of heat to el ectrical energy , " Chern . I ngr . -Tech , 37 , 532-8 ( 1 965 ) .

1 6 . R . E . Shearer and R . C . Werner , "Thermal ly regenerati ve ion i c hydri de ga l van i c cel l , " J . El ectrochem. Soc . , 1 05 , 693 ( 1 958 ) .

1 7 . R . C . Werner, R . E . Shearer, and T . A . C i arl ari el l o , I I Fuel cel l batteries - regenerati ve type , " Proc . 1 3th Annual Power Sources Conf . , PSC Publ i cat i ons Commi ttee , Red Bank , New J ersey , p . 1 22 ( 1 959 ) . R . E . Shearer , J . W . Maustel l er , T . A . C i arl ari el l o , and R . C . Werner , I I Regenerati ve metal hydri de system, " Proc . 1 4th Annual Power Sources Conference , Red Bank , New Jersey , PSC Publ i cati ons Commi ttee pp . 76-7 ( 1 960 ) .

1 8 . T . A . C i arl ariel l o , I I Sel f-cont i nu i ng hydri de cel l s , 1I U . S . Patent 3 , 01 4 , 084 ( 1 96 1 ) .

1 9 . R . C . Werner and R . E . Shearer , I I Fuel cel l s , I I U . S . Patent 3 , 031 , 058 ( 1 962 ) .

20 . T . A . C i arl ariel l o and R . C . Werner , I I Fuel cel l based on nucl ear reactors , I I Chern . Eng . Progr. , 57, 42-5 ( 1 961 ) .

2 1 . R . C . Werner and T . A . C i arl ari el l o , "Metal hydri de fue l cel l s as energy storage dev i ces , 1 I Proc . U . N . Conf. New Sources Energy , Rome , 1 96 1 , �, 2 1 3-21 8 ( 1 963 ) .

194

S=�I I*I __________________________ T_R_-_4_16_'

22 . R . E . Shearer, " Study of energy convers i on devi ces , " Reports No . 1 ( AD 230 503 ; 1 95 9 ) , 2 ( AD 234482 , 1 959 ) , 3 (AD 238235 , 1 960 ) , and 4 ( AD 250695 , 1 960 ) .

2 3 . R . E . Shearer , " Study of Energy Convers i on Devi ces , " Fi nal Report No . 7 ( 7/ 59-5/61 ) , DA- 36-039-SC-78955 , USAERDL/NJ .

24 . J . M . Fuscoe , S . S . Carl ton , and D . P . Laverty , " Regenerati ve fuel cel l system i nvesti gat ion , " WADD Techn i cal Report 60-442 , Thompson­Ramo-Woo l dr i dge , I nc . , Cl evel and , Oh i o , contract AF33 ( 600 ) -39574 ( May 1 960) .

25 . H . J . Schwartz , S . S . Carl ton , and J . M . Fuscoe , " Regenerati ve fuel cel l system, . 1 Fi nal Report No . ASD-TDR- 62- 1 8 , Thompson-Ramo-Woo1 dri dge , I nc . , C1 eve1 a,nd , Qh i o ,' cont-ra-ct Ar'33 ( 600 r.;.-39573 (Apri l 1 962 ) .

26 . S . S . Carl ton , " El ectrode devel opment pr_ogram, " Fi nal Report ASD­TDR-62-24 1 , Thompson-Ramo-Woo1 dri dge , I nc . , Cl evel and , Oh i o , contract AF33 ( 600 ) -42449 ( June 1 962 ) .

27 . S . S . Carl ton , " Fuel cel l constructi on , " u . S . Patent 3 , 1 1 0 , 631 ( 1 966) .

.

28 . A . J . Stromqui st , " Zero gra.v i ty separator devel opment for regenera-ti ve fuel cel l , " Fi nal Report No . ASD-TDR-62-240 , Thorripson-Ramo­Woo l dri dge, Inc . , Cl evel and, Oh i o � contract AF33 ( 600 ) -42449 (J une 1 962 ) .

2 9 . D . , R . Snoke and J . M . Fuscoe , " L i th i um hydrogen fuel cel l seen feas i bl e , " SAE J . , .2.2.( 6 ) , 68-69 ( 1 961 ) .

30 . D . R . Snoke , J . M . Fuscoe , and S . S . Carl ton , " Extended-l i fe fuel ' cel l s for space , " Nati onal Aerospace El ectron ics Conference of Insti tute of Rad i o Engi neers , Dayton , Ohi o , May 8 , 1 961 .

3 1 . �1 . A . Del Duca , J . M . Fuscoe , and T . A . Johnson , " Fuel cel l s for s pace veh i cl es , " As tronauti cs , �, 36-44 ( 1 960 ) .

32 . R . L . Kerr, " Low-gravi ty separator i nvesti gati on , " Report No . ASD-TDR-62-776 ( 1 962 ) .

33 . C . E . Johnson and R . R . Hei nri c h , "Thermodynami cs of ·the l i th i um hydri de regenerati ve cel l , " i n Regenerati ve EMF Cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , 1 05- 1 20 , Ameri can Chemi cal Soci ety , Washi ngton , D . C . ( 1 967 ) .

34 . �1 . S . Foster, C . E . Johnson , and C . E . Crouthamel , " Puri fication uni t for h i gh puri ty i nert atmosphere boxes , " U-SAEC-ANL-6652 , Argonne Nati onal Laboratory ( 1 962 ) . C . E . Johnson , M . S . Foster, and M . L . Kyl e , " Puri fi cati on of i nert atmospheres , " Nucl ear Appl i cati ons , 1, 563-7 ( 1 967 ) .

195

- TR-4 16 S=�I I_I -----------------------------

35 . H . S h imotake and J . C . Hesson , " Corros i on by fused sal ts and heavy 1 i qu i d metal s - a survey , II i n Regenerati ve EMF Cel l s , C . E . Crouthamel and H . L . Recht, eds . , Advances i n Chemi stry Seri es , 64 , 1 49- 1 85 , Ameri can Chemi cal Soci ety , Wash i ngton , D . C . ( 1 967 ) . -

36 . E . J . Ca i rn s , C . E . Crouthame l , A . K. Fi sher , M . S . Foster , J . C . Hesson , C . E . Johnson , H . Sh imotake , and A . D . Tevebaugh , " Ga l vani c cel l s wi th fused sa l t el ectrolytes , " Chemi cal Engi neer i ng D i v i s i on , Report No . ANL-73 1 6 ( November 1 967 ) . Thi s fi na l report conta i n s summari es of al l parti a l reports "Thermal ly regenerati ve fuel cel l s " (ANL-6379 , 641 3 , 6477 , 6543 , 6569, 6596 , 6648 , 6687 , 6766 , 6725 , 6800 , 6875 ; coveri ng 1 96 1 - 1 964 ) and " Energy convers i on " (ANL-6900 , 6925 , 7055 , 71 25 , 7255 , 7325 ; coveri ng 1 964- 1 966 ) from Argonne Nati onal Laboratories germane to TRES , both L i H and bimeta l l i c systems .

3 7 . R . R . He i nri c-h , C-� E. Johnson , and C . E . Crouthamel , "Hydrogen permeation studi es . ! . Armco i ron and i ron-molybdenum al l oys , " J . El ectrochem . Soc . , ill, 1 067- 70 ( 1 965 ) .

_ .- .

38 . R . R . He i nrich , C . E . Johnson , and C . E . Crouthamel, "Hydrogen permeat ion s tudi es . I I . Vanadi um as a hydrogen el ectrode i n a l i th i um hydri de cel l , " J . El ectrochem. Soc . , ill, 1 07 1 - 3 ( 1 965 ) .

3 9 . J . A . Pl ambeck , J . P . El der , and H . A . La i ti nen , " El �ctrochemi stry of the l i th i um hydri de cel l , " J . El ectrochem. Soc . , 1 1 3 , 931 -7 ( 1 966 ) .

40 . C . E . Johnson , S . E . Wood , and C . E . Crouthamel , " Stud i es of l it h i um hydri de systems . I . So l i d- l i qu i d equ i l i bri um i n the l i th i um hydri de­l i th i um chl or ide system, " J . I norg . Chern. , 2, 1 487 ( 1 964 ) .

41 . C . E . Johnson , S . E . Wood , and C . E. Crouthame l , " Studi e s of 1 i th i um hydri de systems . I I . Sol i d- l i qu i d equi l i bri um i n the sodi um chl ori de- l i th i um hydri de system , " J . Chern. Phys . , 44 , 880-3 ( 1 966 ) .

42 . C . E . Johnson , S . E . Wood , and C . E . Crouthamel , " Stud i es of l i th i um hydride systems . I I I . So l i d- l i qu i d equi l i bri um i n the l i th i um bromi de- l i th i um hydri de and l i th i um i od i de- l i th i um hydri de systems , " J . Chern . Phys . 44 , 884-9 ( 1 966 ) .

-

43 . C . E . Johnson , E . Hathaway, and C . E . Crouthamel , " L i th i um hydri de systems . Sol i d- l i qu i d phase equi l i br ia for the ternary l i th i um hydri de- l i th i um

.

ch l ori de- l i th i um fl uoride system. II J . Chern . Eng . Data , ll, 372-4 ( 1 966 ) .

44 . C . E . Jo hnson and E. J . Hathaway , " L i th i um hydri de systems : Sol i d- l i qu i d equi l i bri a for the ternary l i thi um hydri de- l i th i um chl ori de- l i th i um i odi de systems . " J . Chern. Eng . Data , li, 1 74-5 ( 1 969 ) .

45 . C . E . Johnson , R . R . He i nr ich , and C . E . Crouthamel , "Thermodynami c pro­perti es of l i th i um hydri de by an el ectromoti ve force method , " J . Phys . Chern. , 70 , 242-6 ( 1 966 ) .

196

- TR-4 16 .S=�I I_I -----------------------------

46 . J . C . Hes son and H . Sh imotake , I IThermodynami cs and thermal effi c i enc ies of thermal ly regenerati ve b imeta l l i c and hydri de EMF cel l systems , 1 I i n Regenerati ve EMF cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , 82- 1 04 , Ameri can Chemi cal Soci ety , Wash i ngton , D . C . ( 1 967 ) . -

47 . C . E . Mes ser, I IA survey report on l i th i um hydride , 1 I USAEC-NYO-9470 ( October 1 960) .

48 . C . E . Mes ser and J . Mel l or , l iThe system l i th i um hydri de-l i th i um fl uori de , II J . Phys . Chern. , 64 , 503-4 ( 1 960 ) .

49 . C . E . Messer, E . B . Damon , D . C . Maybury , J . Mel l or , and R . A . Seal es , I I So l i d- l i q u i d equi l i bri um i n the l i th i um- l i th i um hydride system , 1I J . Phys . Chem. ,_ 62 s 2�0- 2 lli_SB) �

50 . L . B . Anderson , E . V . Ba l l ou , and S . A . Greenberg , I ISo l ar Regenerati ve Systems , 1I Fi nal Report AD 289294 , lockheed Ai rcraft Corp . , Mi s s i l e and Space Di v i s i on , Cal i forn i a , contract DA-36-039 SC-85245 ( 1 962 ) .

5 1 . L . B . Anderson , S . A . Greenberg , and G . B . Adams , I ITherma l l y and photo­chemi cal ly regenerati ve el ectrochemical systems , 1 I i n Regenerati ve EMF cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , 2 1 33-276 , Ameri can Chemi cal Soci ety , Wa shi ngton , D . C . ( 1 967) .

5 2 . R . F . Fogl e a nd H . E . Lawson , II I nvesti gati on of a n energy convers i on devi ce , II Fi na 1 Report AD285667 , Aeroj et -Genera 1 Corp . , Cal i forni a , contract DA36-039 SC87229 ( 1 962 ) .

5 3 . T . M . Rymarz et a l . , II Chemi cal convers i on of waste heat to el ectri cal energy , 1 I Quarterly Report No . 4 , ARF-3l 82-4 , Armour Research Foundation of I l l i no i s Insti tute of Technol ogy , contract NOw 60-0760-c ( 1 961 ) .

54 . T . M . Rymarz et a l . , II Chemi cal convers i on of waste heat to el ectri cal energy , 1I Quarterly Report No . 5 , ARF 31 82-6 ( 1 961) .

5 5 . Conversati on wi th M . J . Kl e i n and R . H . Snow ( H . L . Chum and R . A . Osteryoung ) , Feb . 22 , 1 979 , C h i cago , I l l i no i s .

56 . J . L . Reger and L . Sch i e l er , II Regenerabl e fuel cel l des i gn , 1I U . S . Patent 3 , 236 , 691 ( 1 966 ) .

5 7 . F . D . Hess and L . Sch i el er, I I Energy convers i on research program , I I Semi annual Tech . Report . AD 27621 , Aerospace Corp . , Cal i forni a , contract AFO 4 ( 647 ) - 930 ( 1 96 1 ) .

58 . · F . D . Hess and L . Schi e l er , II Energy convers ion research program , 1I Semi -. annual Tech . Report , Report No . TDR-69 ( 2220-30 )TR-2 ( 1 962 ) AD 285084 ; contract AFO 4 ( 695 ) -69 ( 1 962 ) .

59 . S . W . Mayer and W . E . Brown , Jr . , II Chronopotenti ometri c measurements of el ectrode ki net i cs for chl ori des of W , Sb , P and group IVA , II Report No . TD-R69 ( 2220-30J TM1 ( 1 962 ) AD 282902 , Aerospace Corp . , Cal i forni a , contract AF04 ( 695 ) - 69 ( 1 962 ) .

197

� TR-4 16 S=�I I_I ----------------------------------

60 . S . W . t�ayer and W . E . Brown , Jr . , " El ectrode ki neti cs for chl ori des of tungsten , ant imony , and phosphoru s , I I J . El ectrochem . Soc . , 1 1 0 , 306- 1 1 ( 1 963 ) .

6 1 . C . R . McCu l l y , T . M . Rymarz , and S . B . N icho l son , " Regenerati ve chl or ide systems for convers i on of heat to el ectri cal energy , " i n Regenerati ve EMF cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , 1 98- 21 2 , American Chemical Soc i ety , Wash i ngton , D . C . ( 1 967 ) .

62 . c . R . McCu l l y and T . M . Rymarz , " Chemi cal convers i on of waste heat to el ectri cal energy , " Fi nal Report No . I ITR-31 82- 1 0 , I l l i no i s Inst i tute of Technol ogy Research Insti tute , contract NOw 60- 0760c ( 1 962 ) .

63 . T . M . Rymarz et a l . , " Chemi cal convers i on of waste heat to el ectri cal energy , I I Quarterly_ Rep_ort No ._ J3 No . ARF-3.182-9 , I I I i no i s Insti tute of Technol ogy Researeh I n st itute , contract NOw 60-7060-c ( 1 962 ) . C . R . tkCul l y , I I El ectrochemi cal power supply regenerated by heat , I I U . S . Patent 3 , 52 3 , 829 ( 1 970) .

64 . R . A . Ri ghtmi re and J . L . Cal l ahan , " Energy convers i on system , " U . S . Patent 3 , 088 , 990 ( 1 968 ) .

6 5 . E. L . Kumm , " S02-S03 regenerati ve fue l cel l research , " NASA Doc . N62- 1 7 , 308 (1 962 ) .

66 . l� . E . Wentworth and E . Chen , " S imp l e therma l decompos i ti on reactions for storage of sol ar thermal energy , " So l ar Energy , 1 8 , 205- 1 1 ( 1 976 ) . Phone cal l wi th W . E . Wentworth ( H . L. Chum ) , Apri l 1 5-, 1 979 .

67 .. E . H . Lyons , J r . , "Meta l oxi de fue l cel l s , " U . S . Patent 3 , 1 00 , 1 63 ( 1 963 ) .

68 . D . E . McKenz i e and J . P . Howe , " El ectrochemi cal convers i on of heat to e l ectri c i ty , " U . S . Patent 3 , 368 , 92 1 ( 1 968 ) .

6 9 . W . E . Case , "Apparatus for converti ng heat energy i nto el ectri cal energy , " U . S . Patent 344 , 345 ( 1 886 ) . W . E . Case , J b . d . El ectrochemi e , £, 56 ( 1 896 ) .

70 . S . Sk i nner, l i The t i n-chromi c chl ori de cel l , " Proc . Phys i c . Soc . London , 1 3 , 477-81 ( 1 895 ) ; l iThe ti n-chromi c ch l or ide cel l , " Ph i l . t�a g . , 39 , 444-7( 1 895 ) . -

71 . J . Vedel , M . Soubeyrand , and H . Le Quan , l iThe convers i on of thermal energy i nto e l ectri cal energy usi ng cel l s havi ng a reversa l of potent ia l wi th temperature , " J . Appl . El ectrochemi stry , 2., 475-81 ( 1 979 ) .

72 . c . R . McCu l l y , l iThe chem ica l convers i on of sol ar energy to el ectri cal energy , " Proc . U . N . Conf . New Sources of Energy , Rome , 1 96 1 , 2 , 1 96-202 ( 1 963 ) .

.. -

73 . C . R . McCu l l y et a l . , " Chemi cal convers i on of waste heat to el ectri cal energy , " Quarterly Report No . 3 , ARF-31 82-3 , Armour Research Foundat ion of I l l i no i s Inst i tute of Technol ogy , contract NOw 60-0760-c ( 1 961 ) .

198

TR-4 16 S=�I I.I ------------------------------� �

74 . C . R . McCu l l y , " Chemi cal convers ion of heat to el ectri ca,l energy , " Report No . I ITRI-C-6006� 1 0 , Fi nal , I l l i no i s Insti tute of Technol ogy Research Insti tute , contract NOw 60-051 2-c ( 1 967 ) .

75 . N . Bj errum, " El ectrochemical and spectroscopi c stud ies of the cha l co­gens in chl oroal umi nate mel ts , " i n Characteri zati on of Sol utes in Non­Aqueous So l vents , " G . Mamantov , ed . , Pl enum Press , New York , pp . 25 1 -271 and references there i n ( 1 976 ) . T . vJ . Couch , D . A . Lo kken , and J . D . Corbett , l iThe crystal structure of tetratel l uri um ( 2+ ) tetrachl oroal umi nate and heptachl orod ia l umi nate , Te4 ( A1 Ch ) 2 and Te 4 ( AhCh ) 2 , " I norg . Chern . , 1 1 , 357-9 ( 1 972 ) . N . J . Bj errum and G . P . Smi t h , " Tel l uri um i n the formal el ectropo s i ­t i ve oxi dati on state one-'hal f i n ac i d i c ch l ori de med i a , " J . Am . Chern . Soc . , 90, 4472 ( 1 968 ) . D . J . Pri nce , J . 0:- COl'bett-,-and 130 Garbtsil, "B i atomic cati ons of tel l ui um and 'sel en i um l n chl oroal umi nate mel ts , " Inorg . Chern . , 9 , 2731 ( 1 970) . _

-

R . Fehrman , N . J . Bj errum , and M . A . Andreasen , " Lower oxi dati on states of tel l uri um 4 Te�+ , Te �+ , Te �+ i n chl oroa l umi nate mel ts , " Inorg . Chern . , 1 5 , 21 87 ( 1 976 ) . 'F'":- W . Paul sen , N . J . Bj errum, and D . R . N ie l sen , " Chl orocompl exes i n mol ten sal ts . I I I . Raman study of chl oro compl exes found i n the mol ten KC1 -A1 C1 3TeC1 4 system , " Inorg . Chern. , ll, 2693 ( 1 974 ) .

76 . J . Robi nson and R . A . Osteryoung , l iThe el ectrochemi cal behavior of Te ( IV ) i n sodi um tetrachl oroal umi nate melts , I I J . El ectrochem . Soc . , 1 25 , 1 784 ( 1 978) .

77 . T . ',M . . Rymarz , "Tel l uri um chl ori de thermoregenerati ve ga'l van i c cel l , " Fi nal Report No . I ITRI-c 6069-5 , contract NOw-65-0431 -c ( 1 967 ) .

78 . D . E . Anthes , " Tel l uri um ch l ori de thermoregenerati ve gal van i c cel l , " Fi nal Report No . I ITRI - c 61 42- 5 , contract NOw 1 9- 68-C-0361 ( 1 969 ) .

79 . C . R . McCu l l y et a l . , " Chemi cal convers i o.n of waste heat to el ectri cal energy , " Quarterly Report ARF-31 82-2 , Armour Research Foundati on of I l l i no i s Insti tute of Technol ogy , contract NOw-60-0760-c ( 1 961 ) .

80 . I u . K . De l i marski and B . F . Markov , El ectrochemi stry � Fused Sal ts , The S i gma Press , Washi ngton , D . C . ( 1 961 ) .

81 . B . Agrus s , l i The therma l ly regenerati ve l i qu i d metal cel l , " J . El ectrochem . Soc . , 1 1 0 , 1 097- 1 1 03 ( 1 963 ) .

82 . R . D . Weaver, S . W . Smi th , and N . L . Wi l l i am, l iThe sodi um/ti n l i qui d metal cel l , " J . El ectrochem . Soc . , 1 09 , 653-657 ( 1 962 ) . S . W . Smi th and R . D . Weaver ,� compari son of the system wei ght of the Na/Sn l i q u i d metal cel l for various appl i cati ons , 1 I Report No . 4337-E , Del co­Remy Di vi s i on , General Motors Corp . ( March 1 962 ) .

83 . B . Agruss , " Regenerati ve battery , " U . S . Patent 3 , 245 , 836 ( 1 966 ) .

199

TR-4 16 S=�I I:.I ------------------------------� �

84 . B . Agrus s and H . Ka ras , l iThe thermal l y regenerati ve l i qu i d metal con­centrati on cel l , " i n Regenerati ve EMF Cel l s , C . E. Crouthamel and H . L . Recht , eds . , Advances in Chemi stry Seri es , 64 , 62-81 , Ameri can Chemi cal Soci ety, Wa shi ngton , D . C . ( 1 967 ) .

85 . R . E . Henderson , il Li qui d metal cel l s , " Chern. Eng . Pr . , 59 , 56- 7 ( 1 963 ) . R . E . Henderson and E . H . Hi etbri n k , " Mercury space power systems , " i n Di rect Convers i on Proceed i ng s , Pac i fi c Energy Convers ion Conf . , San Franc i sco , Ca l i forn i a , pp . 1 6- 1 to 1 6- 1 2 ( 1 962 ) .

86 . R . E . Henderson , " Thermal ly regenerati ve fuel cel l s , I I Proc . 6th AGARD Combusti on and Propul s i on Col l oqui um on Energy Sources and Energy Con­Vers i o n , Cannes , France , March 1 6 , 1 964 ; AGARDograph , No . 81 , pp . 795-809 .

.

87 . R . B . Wri ght ,- " Diffus ion of potass i um i n a l i qu i d metal cel l , " Report No . N64- l 9798 , the Al l i son D i v i s ion of General Motors Corp . , Indi anapol i s , Indi ana , contract EDR- 38l 4 ( 1 964 ) . -

88 . J . D . Mangus , " Research and devel opment of an advanced l aboratory l i q u i d meta 1 regenerati ve cel l , I I Report No . AD-43851 9 ( 1 964 ) .

89 . B . Agrus s , E . H . Hi etbri n k , and T . F . Nagey, " Regenerati ng molten metal fuel cel l , " U . S . Patent 3 , 503 , 808 ( 1 970 ) .

_ _

90 . M . F . Lantratov and E . V . Tsarenko , " Investi gation of the thermodynami c properti es of l i qu i d metall i c sol uti ons i n potass i um-mercury system , " J . Appl . Chern . USSR , 33 , 7 ( 1 960 ) .

9 1 . C . R . LaMantia and C . F . Bon i l l a , I I Thermodynamics of the system potass i um­mercury , " Proc . Sympos i um on Thermophysi cal Properti es , 4th , J . R . Mozynski , ed . , Ameri can Soc i ety of Mechan i ca l Engi neers , New York , pp . 58-71 ( 1 968) .

92 . V . T . Vorogush i n , I IVari ati on of the free-energy wi th the heat of evapor­ation i n the thermodynamic cycl e of a regenerati ve fue l cel l , " Rus s . J . Phys . Chern . , 43 ( 3 ) , 435- 6 ( 1 969 ) .

�

93 . R . D . 01 denkamp , L . A . Heredy , and H . L . Recht , l iThe test program and performance analysi s of the Atomi cs Internati onal therma l ly regenerati ve a l l oy cel l ( TRAC ) system , 1 I Proc . I ntersoci ety Energy Convers i on Eng i ­neer ing Conf. , 1 st , Los Angel es , Cal i forn i a , pp . 324-331 ( 1 966 ) .

94 . L . A . Heredy , M . L . Iverson , G . D . Ul ri ch , and H . L . Recht , " Devel op­ment of a thermal ly regenerati ve sodi um-mercury gal van ic system . Part I . E l ectrochemi cal and chemi cal behavi or of sodi um-mercury gal van i c cell s , " i n Regenerati ve H1F Cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances in Chemi stry Seri es , 64 , 30-42 , Ameri can Chemi cal Soci ety , Wash i ngton , D . C . ( 1 967 ) . -

200

- TR-4 16 S=�I I.I ---------------------------� �

95 . I . J . Groce and R . D . Ol denkamp , " Deve1 opment of a therma l ly regener­ati ve sodi um-mercury gal van i c system. Part I I . Des i gn , constructi on and test i ng of a thermal ly regenerati ve sodi um-mercury gal van i c system , " i n Regenerati ve EMF Cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , 43- 52 , Ameri can Chemi cal Soci ety , Wash i ngton , D . C . ( 1 967 ) .

-

96 . R . D . Ol denkamp and H . L . Recht , " Deve1 opment of a thermal ly regener­ati ve sodi um-mercury gal vani c system. Part I I I . Performance ana lys i s for a nucl ear reactor- powered , thermal l y regenerati ve sodi um mercury gal van i c system , 11 i n Regenerati ve EMF Cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , 53-61 , Ameri can Chemi cal Soci ety , Wash i ngton , D . C . ( 1 967 ) . -,

97 . H . L . Recht and [)' - L -McKeflt-i-e-; I+t:nergy convers i on process and apparatus , " U . S . Patent 3 , 41 9� 435 { 1 968 ) .

98 . H . L . Recht and M . L . I verson , " POrOUS matri x for gal van i c cel l , " U . S . Patent 3 ,4 1 9 , 436 ( 1 968 ) .

99 . �1 . L . Iverson , " G� l van ic cel l el ectro 1yte , " U . S . Patent 3 , 441 ,41 1 ( 1 969 ) .

1 00 . L . A . Heredy , " Hi gh temperature gal va n i c cel l , II U . S . Patent 3 , 441 , 446 ( 1 969) .

1 01 . �1 . L . Iverson and H '. L . Recht , l iThe acti v i ty of sod i um i n sodi um amal ­gams from measurement , " J . Chem. Eng . Data , 1£, 262-5 ( 1 967 ) .

1 02 . K . Hauffe , " Determi n i ng the acti v i t i es of metal s i n bi nary systems whose behavi or i s widely d i vergent from the i deal , " Z . El ectrochem . , 46 , 348-50 ( 1 940 ) .

1 03 . N . S . Kurnakow and N . A . Pusch i n , Z . Anorg . Chem . , 30 , 87- 1 01 ( 1 902 ) .

1 04 . A . K . Fi scher, " Phase d iagram cons i derati ons for the regenerati ve bi :­metal l i c cel l , " i n " Regenerati ve EMF cel l s , " C . E . Crouthamel and H . L . Recht, eds . , Advances i n Chemi stry Seri es , 64 , 1 21 - 1 35 , Ameri can Chemi cal Soci ety , Wash i ngton , D . C . ( 1 967 ) . -

1 05 . For exampl es , see : S . M . Zi v i , I . Po l l ack , H . Kaci nskas , A . A . Ch i 1 ens kas , and D . L . Barney , " Battery engi neering probl ems i n des i gn i ng an el ectrical l oad l evel i ng p l ant for l i th i um/ i ron- su l fi de cel l s , " Proc . 1 4th Intersoc i ety Energy Convers i on Engi neeri ng Conference , Boston , Massachusetts , Ameri can Chemi cal Soci ety , �Jash i ngton , D . C . , pp . 722-729 ( 1 979 ) . F . J . Marti no , T . W . Ol szanski , L . G . Barthol me , E . C . Gay , and H . S h imotake , "Advances i n the deve l opment of L i -Al / FeS cel l s for el ectri c­automobi l e batteries , " Ameri can Chemi cal Soci ety , Wash i ngton , D . C . , p p . 660-664 ( 1 97 9 ) .

201

S=�I I.I _______________________ T_R_-_4_16

_

-� �

1 06 . C . E . Johnson and E . J . Hathaway, I I So l i d- l i qu i d phase equ i l i br ia for the ternary systems Li ( F , Cl , I ) a nd Na ( F , C l , Br ) , I I J . E 1 ectro­'chem . Soc . , 1 1 8 , 631 -4 ( 1 971 ) .

1 07 . K . Hauffe and A . L . Gri essback-V ierk , I IAct i v i ty measurements on l i qui d-thal l i um al l oys , 1I Z. E1 ektrochem . , 53 , 1 51 ( 1 949 ) .

1 08 . M . F . Lantratov , I I Thermodynami c properti es of l i qu i d-meta l sol uti ons i n the Na-Pb system , 1I Rus s i an J . I norg . Chern. ( transl . ) , i, 2043-5 ( 1 959 ) .

1 09 . B . Porter and M . Fei n 1 e i b , I I Determi nati on of acti v i ty of sod i um i n Na-Pb a l l oys a t h i gh temperatures , 1I J . E1 ectrochem . Soc . , 1 03 , 300- 303 ( 1 956 ) . .

1 1 0 . E . J . Ca i rn s , A • . D. Te.vebalJjJll�J � D . B i ngle , C . E . Johnson , �1 . S . Foster , E . J .' Hathaway , J . Peck , E . L . Gasner , T . F . Young , G . H . McCl oud , A . K . Fi scher , S . A . Johnso n , S . E . Wood , H . Sh imotake , G . Rogers , and J . Kargo 1 , I I Energy convers i o n , II Chemi cal Engi neeri ng Di v i s i on Semiannual Report, Ju ly- December 1 966 , ANL-7325 , Argonne , I l l i no i s : Argonne Nati onal Laboratory p . 1 79- 1 96 (Apri l 1 967 ) .

1 1 1. A . K . Fi scher and S . A . Johnson , i l L i qui d-vapor equi l i bri a i n the sodi um­l ead system , 1I J . Chern . Eng . Data , ]E., 492-495 ( 1 970 ) .

1 1 2 . H . Sh i motake and E . J . Ca i rn s , I IB imeta l l i c cel l s wi th ·fused-sal t e1 ectrolytes , 1I i n Advances i n Energy Convers i on Engi neeri n g , Ameri can Soci ety of Mechani cal Engi neers , pp . 951 - 962 ( 1 967 ) .

1 1 3 . J . C . Hes son , M . S . Foster , and H . S h imotake , I ISel f-di scharge i n a 1 kal i ­meta l -conta i n i ng b imetal l i c cel l s , 1 I J . El ectrochem. Soc . , 1 1 5 , 787-790 ( 1 968) .

-

1 1 4 . M . S . Foster, I I Laboratory stud i es of i ntermeta l l i c cel l s , 1I i n Regener­ati ve EMF ce1 1 s , 1I C . E . Crouthamel and H . L. Recht , eds . , Advances i n Chemi stry Seri es , 64 , 1 36- 1 48 , Ameri can Chemi cal Soci ety , Wash i ngton , D . C . ( 1 967 ) .

1 1 5 . A . K . Fi scher , S . A . Johnson , and S . E . Wood , i l L i qui d-vapor phase di agram and thermodynami cs of the sod i um- bi smuth system, II J . Phys . Chern . , 11, -1 465-1 472 ( 1 967 ) .

1 1 6 . C . E . Johnson and A . K . F i scher, I I New measurements for the sodi um­b i smuth phase d iagram , 1I J . Less-Common Metal s , 20 , 339-344 ( 1 970 ) .

1 1 7 . M . S . Fo ster, S . E. Wood , and C . E . Crouthamel , I I Thermodynami c of bi nary a l l oys . I I. The l i th i um-t i n system , 1I J . Phys . Chern. , 70 , 3042-45 ( 1 964) .

_ -

1 1 8 . M . S . Foster, S . E . Wood, and C . E . Crouthamel , I IThermodynami c of bi nary a l l oys . ! . The l i th i um- bi smuth system , 1I Inorg . Chern . , 1, 1 428- 1 431 ( 1 964 ) .

202

S=�I I.I _______________________ T.=...;R=.,...--=4:...=.1.=...6 _ -� ��

1 1 9 . M . S . Foster and C . C . L i u , I I Thermodynami c of bi nary al l oys . I I I . The l i th i um-tel l uri um system, 1I J . Phys . Chern. , 70 , 950 ( 1 966 ) .

1 20 . A . K . Fi scher and S . A . Johnso n , il L i qui d-vapor equi l i bri a and thermo­dynami cs of the l i th i um- ti n systems , 1I J . Chern . Eng . Data , 1l, 280-283 ( 1 972 ) .

1 21 . A . J . deBethune , T . S . L i cht , and N . Swendeman , li The temperature co­effi ci ents of el ectrode potenti a l s , 1 I J . El ectrochem . Soc . , 1 06 , 61 6-625 ( 1 959 ) .

-

A. J . deButhune and H . O . Da l ey, Jr . , l iThe thermal temperature coeffi ­ci ent of the cal omel el ectrode between 0° and 70° C , I I and I I I , I I J . El ectrochem . Soc . , 11 6 , 1 395- 1 401 , 1 401 - 1 406 ( 1 969 ) .

1 22 . A . J . deBethun� , . u lrre.vers i ble.-ther:modynaJlli cs i n el ectrochemi stry , 1I J . El ectrochem . Soc . , 1 07 , 829-842 ( 1 960 ) .

1 23 . J . N . Agar , I I Thermogal vani c ce1 1 s , U. i n Advances i n El ectrochemi stry and El ectrochem ical Engi neeri n g , P . Del ahay and C . W . Tobi a s , eds . , vol . 3 , Intersc ience , pp . 31 - 1 21 ( 1 963 ) .

1 24 . C . Wagner, I I Thermoel ectr ic power of cel l s wi th i on i c compounds i nvol v­i ng i oni c and el ectron i c conducti on , 1I Progr. So l i d State Chern . , 7 , 1 -37 ( 1 972 ) .

-

1 25 . R . Zi to , J r . , I IThermogal vani c energy convers i on , 1 I AIAA J . , 1, 21 33-21 38 ( 1 963 ) .

1 26 . B . . ' Sundheim , I I Mo l ten sal ts vs . thermoel ectri c materi al s , II chapter 2 i n Thermoel ectric Materi a l s and Devi ces , I . B . Cadoff and E . Mi l l er , eds . , Re i nho l d Publ . Corp . , New York ( 1 960 ) .

1 27 . R . W . Chri sty , II Ion i c Materi al s , 1I i n Thermoel ectric Materi a l s and Devi ces , I . B . Cadoff and E. Mi l l er , eds . , Rei nho l d Publ . Corp . , New York , pp . 1 73- 1 83 ( 1 960 ) .

1 28 . W . Vi el s t i c h , Fuel Cel l s , transl ated by O . J . G . I ves , Wi l ey­I ntersci ence, New York, p p . 345- 361 ( 1 970 ) .

1 29 . N . Fuschi l l o , I IThermoel ectri c phenomena , I I i n Thermoel ectri c Mater ia l s and Devi ces , I . B . Cadoff and E . Mi l l er , eds . pp . 1 - 1 6 ( 1 960 ) Rei nhol d Publ . Corp . , New York . S . W . Angri st , I I Di rect Energy Convers i o n , 1 I 3rd ed . , Al l yn and Bacon , I n c . , Boston , pp . 1 29- 1 90 ( 1 976 ) .

1 30 . Examp l es : L . Patri ck and A . W . Lawson , I IThermoe l ectri c power of pure and doped AgBr , I I J . Chern. Phys . , 22 , 1 492-5 ( 1 954 ) . R . W . Chri sty , E . Fukush i ma , and H . T . L i , I IThermoel ectri c power of s i l ver bromi de conta i n i ng cadmi um bromi de , I I J . Chern. Phys . , 30 , 1 36-8 ( 1 959 ) .

203

TR-416 S=�I I.I ------------------------------� �

1 3 1 . K . S . P i tzer , " Thermodynami cs of thermoce 1 1 s wi th fused or sol ; d e l ectrolytes , " J . Phys . Chem. , .§i, 1 47- 1 50 ( 1 961 ) .

1 32 . J . N . Agar and W . G . Breck , "Therma1 di ffus i on i n non i sothermal cel l s . I . Theoreti cal rel ation and experiments on so l uti ons of tha l lous sa l ts , " Trans . Faraday Soc . , 5 3 , 1 67-78 ( 1 957 ) .

1 33 . N . Fusch i l l o , "Thermoel ectric fi gure of meri t , " i n Thermoel ectri c Materi a l s and Dev i ces , I . B . Cadoff and E . Mi l l er , eds . , Rei nho l d Publ i s h i ng Corp . , New York, pp . 31 -46 ( 1 960 ) .

1 34 . M . Tel kes , " So l ar thermoel ectri c generators , " J . Appl . Phys . , 25 , 765-777 ( 1 954 ) .

1 35 . T . t�artanowi cz , "A_ the.oreti caJ_ana lys i s elL a mol ten sal t thermoce l l as a thermoe l-ectric generator , II Advan . Energy Convers i on , i, 1 49- 1 58 ( 1 964 ) .

1 36 . T . Wartanowi cz , "Analys i s performance and experimental i nvesti gati ons on a mo l ten sal t thermocel l as a thermoel ectrochemi cal energy convert­er , " Bu l l . Acad . Po l on . Sci. Ser. Sci . Tech . , Jl., 91 1 - 1 6 ( 1 965 ) .

1 37 . B . R . Sundhe im a nd J . Rosenstre i c h , "Mol ten sal t thermocel l s , " J . Phys . Chern. , 63 , 41 9-22 ( 1 959 ) .

1 38 . R . H . Det i g and D . H . Archer , "Thermoel ectric effects i n fused i oni c materi al s , " J . Chern. Phys . , 38 , 66 1 -666 ( 1 963 ) .

1 39 . B . R . Sundhe im and J . D . Kel l ner , " Thermoel ectri c properti es of mol ten s i l ver-s i l ver n i trate-sodi um n i trate system , " J . Phys . Chern. , 69 , 1 204-8 ( 1 965 ) .

1 40 . R . Haase , U . Pruser, and J . Ri chter, " Eval uati on of measurements on thermocell s conta i n i ng mol ten sal t mi xtures , " Ber . Bunsenges . Phys . Chern . , Qi, 577-84 ( 1 977 ) . See al so i bi dem , 508-51 4 .

1 41 . M . Abraham and M . Gauth i er , " Infl uence de l a temperature sur l e pouvo i r thermoel ectri q ue i n i ti a l de quel ques thermopi l es a mel anges fondus et s urfondus de AgN0 3 et Ti N0 3 avec el ectrodes de Ag , " El ectroch im . Acta� ..li, 1 399- 1 405 ( 1 970 ) .

1 42 . M . Abraham and M . Gauthi er , "Thermopi l es a ni trates fondus : evo l ution du pouvo i r thermoe l ectri que i n i ti al l ors de la trans i ti on l i qu i d-verre , " El ectroch i m. Acta , �, 953- 9 ( 1 971 ) .

1 43 . R . Haase , " Thermodynami cs of I rrevers i bl e Processes , " Readi ng , Massachusetts ( 1 969 ) . 1 44 . R . Connan and J . Dupuy , " Perturbations dans 1 es mel anges AgN0 3-MN0 3

fondus par mesures de thermopi l es - nouvel l es poss i bi l i tes d ' exp l o i ta­t i on , " El ectroch im. Acta , ..li, 977-85 ( 1 970 ) .

204

TR-4 16 S=�I I.I ------------------------------� �

1 45 . t4 . Abraham and M . Gauth i er , l i La thermopotenti ometri e appl i quee a 1 ' etude du comportement de certa i ns l i qu i des surfondus , 1I El ectroch im . Acta , 1I, 279-84 ( 1 972 ) .

1 46 . S . Senderoff and · R . I . Bretz , I I Ion i c transport entropy i n non - i so­therma l mol ten s i l ver chl ori de cel l s , I I J . El ectrochem. Soc . , 1 09 , 56-6 1 ( 1 972 ) .

1 47 . H . Hol tan , Jr . , I I Thermocel l s wi th sol i d and fused el ectrolytes , 1 I Kon i nkl . Ned . Akad . Wetenschap . , Proc . , Ser . B . , 56 , 498-509 ( 1 953 ) . H . Hol tan , Jr . I I Thermocel l s conta i n i ng el ectro lyti c sol uti ons , 1I J . El ectrochem . Soc . 56 , 5 1 0- 14 ( 1 953 ) .

1 48 . B . F . Markov , I I Thermal el ectromoti ve forces wi th fused sa l ts , II Dokl ad . Akad . Na u k . USSR,. 1 08 ,_ 1 1 5·:11- 1 1 956 ) .

1 49 . L . B . Anderson , S . ,A . Greenberg , and G .. B . Adams , "Thermal ly and photo­chemi cal ly regenerati ve el ectrochemi cal -systems , I I i n Regenerati ve EMF Cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , vol . 64 , American Chemi cal Soci ety , Was h i ngton , D . C . pp . 21 3-276 ( 1 967) . -

1 50 . A . R . N i cho l s and C . T . Langford , I I Entropy of the movi ng cuprous ion i n mol ten cuprous chl ori de from thermogal van i c potenti a l s , 1 I J . El ectro-chem . Soc . , 1 07 , 842-7 ( 1 960 ) . - - .

1 5 1 . A . Kvi st and A . Randsal i e , II Thermoel ectri c power of mol ten and sol i d Ag 2S04 , " Z . Naturforsch . , a 21 , 278-81 ( 1 966 ) .

1 52 . H . 'P . Mei s sner, D . C . Wh i te , and G . D . Uhl ri ch , "Thermocel l s : effect of pressure on vol tage , 1I Advan; Energy Convers i o n , i, 205- 1 6 ( 1 965 ) .

1 53 . S . Senderoff , I I Thermocel l , " U . S . Patent 3 , 294 , 585 ( 1 966 ) . S . Senderoff , "Thermocel l battery , II U . S . Patent 3 , 31 1 , 506 ( 1 967) .

1 54 . W . Fi scher , II Thermoel ectri c powers of cel l s of the type C1 2T l mol ten c h l ori de I T+flTC1 2 , 1I Z . Naturforsch . , a �, 281 - 6 ( 1 966 ) .

1 55 . R . J . Brood , I IThermoce1 1 , I I U . S . Patent 3 , 293 , 079 ( 1 966 ) .

1 56 . H . Hol tan , Jr . , I I Rel ati on between temperature coeffi c i ent of i sothermal cel l s and the thermopotenti a l s of the correspondi ng thermoce l l s , 1I Kon i nkl . Ned . Akad . Wetenscha p . , Proc . , 57B , 1 38-41 ( 1 954 ) ; i b i dem , 592-5 . H . Ho l tan , P . Mazur , and S . R . deGroot , Phys i ca , X I X , 1 1 09- 1 8 ( 1 953 ) .

1 57 . J . A . A . Ketel aar, l i The unatta i nabi l i ty of a un i fi ed EMF seri es for mo l ten sal ts , 1I J . El ectroanal . Chem. , 65 , 87- 93 - ( 1 975 ) .

1 58 . H . E . Lawson , I IApparatus and method for thermal regeneration of el ec­tri cal energy , 1I U . S . Patent 3 , 374 , 1 20 ( 1 968 ) .

205

- TR-416 S=�I '.' ------------------='-----� .

1 59 . D . O . Ra l e i gh and L . E . Topol , "Thermoel ectri c potent ial s i n mol ten B i -B i I g so l utions , " J . Chern. Phys . , 11, 31 79-84 ( 1 964 ) . .

1 60 . L . E . Topo l and L . O . Ranson , "Magneti c suscepti bi l i ti es of mol ten B i -B i I 3 so l uti ons , " J . Chern . Phys . , 38 , 1 633-70 ( 1 963 ) .

1 6 1 . L . F . Grantham and S . J . Yos i m , " El ectri cal conducti vi ti e s of mol ten B i -Bi I 3 sol uti ons , " J . Chern. Phys . , 38 , 1 671 - 76 ( 1 963 ) .

1 62 . J . D . Ke l l ner , " Therma l di ffus ion i n an oxi dation-reduction thermo­cel l - the bi smuth-bi smuth i od i de system , " J . Phys . Chern . , 70 , 2341 -47 ( 1 966 ) . -

1 63 . L . E . Topo l , S . J . Yos im , and R . A . Osteryoung , " EMF measurements i n mol ten bi smuth-bi.s'!1uth_ tri ch l ori � so l utto_n s , " J . Phys . Chern. , 65 , 1 5 1 1 - 1 9 ( 1 96 1 -) . - . . --L. E. Topol and R . A . Osteryoun g , " EMF pol arograph i c and chronopotenti ometri ( s tudi es i n mol ten bi smuth-bi smuth t-r i bremi de sol utions , " J . Phys . Chern . , 6 6 , 1 587- 91 ( 1 962 ) .

1 64 . " So l i d e l ectro lytes , general pri nci pl es , characteri zat ion , materi al s appl i cati ons , " P . Hagenmul l er and W . van Goo l , eds . , i n �1ateri a l s Sci ence Seri es , Academic Pres s , New York ( 1 978 ) .

1 65 . J . L . We i n i nger, " Ha l ogen acti vated sol i d el ectrolyte .ce l l , " J . El ectrochem . Soc . , 1 05 , 439-41 ( 1 958 ) . J . L . We i n i nger, " Iodi ne-acti vated sol i d el ectrolyte cell for use at h i gh temperature , " J . El ectrochem. Soc . , 1 06 , 475-81 ( 1 959 ) .

1 66 . J . L . We i n i nger , " So l i d el ectrode battery , " U . S . Patent 2 , 933 , 546 ( 1 960 ) .

1 67 . J . L . Wei n i nger and H . A . L i ebhafsky , " Sol i d el ectrolyte-gaseous cathode battery , " U . S . Patent 2 , 987 , 568 ( 1 96 1 ) .

1 68 . J . L . We i n i nger, "ThermogaTvan i c cel l s wi th s i l ver i od ide as a sol i d e l ectro lyte , " J . El ectrochem. Soc . , ill, 769-74 ( 1 964 ) .

1 69 . J . L . We i n i nger, " So l i d el ectro lyte thermoce1 1 , " U . S . Patent 2 , 890 , 259 ( 1 959 ) .

1 70 . J . L . We i n i nger , ." Non- i sothermal vol tai c cel l hav i ng i od i ne el ectrodes , " U . S . Patent 3 , 297 , 486 ( 1 967 ) .

1 71 . H . F . Hunger , " The s i l ver l s i l ver i od i de l si l ver thermocel l , " J . El ectrochem . Soc . , 1 20 , 1 1 57-61 ( 1 973 ) .

1 72 . A . Sch i ral d i , G . Ch i ode1 1 i , and A . Mag i stri s , "Thermoel ectri c power of · AgI-Ag oxysal t i o n i c sol i ds , " J . Power Sources , £, 257-64 ( 1 977/ 1 978 ) . ·

1 73 . G . Ch i odel l i , S . r�ag i stri s , and A . Schi ral di , " Sol i d el ectro lyte cel l s , " El ectroch im . Acta , 11, 655-6 ( 1 974 ) .

206

S=�I I*I _________________________ T_R_-4_1_6_

1 74 . A . Sch i ral di , G . Ch i odel l i , and A . �1agi stri s , " S i l ver i odi de-s i l ver oxysal t el ectrolytes for sol i d-state cel l s , " J . App1 . El ectrochem. ,. �, 251 - 5 ( 1 976 ) .

1 75 . A . Sch i ral d i and E . P i zzati , "Thermoel ectri c · powers of the systems s i l ver i odi de-s i l ver molybdate ( V I ) or tungstate ( V I ) , I I Z . Naturforsch . , 31 A , 1 077-80 ( 1 976 ) .

1 76 . R . E . Howard and A . B . Li di ard , "Thermoel ectri c power of i on i cal ly con­ducti ng crystal s , " Phy1 . �1a g . , £, 1 462- 1 467 ( 1 957 ) . R . E . Howard and A . B . Li ddard , "Thermoel ectri c power of i on i c crystal s , " Di s c . Faraday Soc i ety , 2 3 , 1 1 3-21 ( 1 957 ) .

1 77 . E . Haga , " Theory of thermoel ectri c power of i o n i c crysta l s . I - IV , " J . Phys . Soc . Japan , 1 3 , 1 090-5 ( 1 958 ) ( cf . Ref . 1 30 ) ; 1 4 , 992-6 ( 1 969 ) ; li, 1 1 76-81 ( J 95g1; - and l fr;-�"949=54 ( 1 96-o-} . -

1 78 . A . R . Al l natt and P . W . �1 . Jacobs , _"The ...thermoel ectri c power of ion i c crystal s . I . Theoreti cal , " and " I I . Resul ts for potass i um chl ori de , " Proc . Roy . Soc . ( London ) , A260 , 350- 69 ( 1 96 1 ) ; A267 , 31 -44 ( 1 962 ) . A . R . Al l natt and A . V . Chadwick , "Thermoel ectri c power of crystal -l i ne sod i um ch lor ide , " J . Chern. Phys . , 47 , 2372-8 ( 1 967 ) . A . R . Al l natt and t� . H . Cohen , " Stati sti cal mechani cs of defect­conta i n i ng sol i ds . I . General forma l i sm , " and " I I . Ion i c crystal s , " J . Chern . Phys . , 40 , 1 860- 70 and 1 87 1 - 90 ( 1 964 ) . P . W . M . Jacobs and J . M . Maycock , " Pol ari zati on effeds i n the i on i c conducti v i ty of a l kal i hal i des crysta l s . I . Al ternati ng current ca;paci ty , " J . Chern. Phys . , 39 , 757-62 ( 1 963 ) . A . R . Al l natt , P . W . �1 . Jacobs , and J . �1 . Maycock , " Po l ari zati on effects in the i on i c conducti vi ty of a l ka l i hal i de crystal s . I I . Current-t ime dependence , " J . Chern . Phys . , 43 , 2526-32 ( 1 965 ) .

1 79 . M . S h imoj i and H . Ho sh ino , "Thermoel ectri c power of i on i c crystal s . I . General theory , " J . Phys . Chern . Sol i ds , 28 , 1 1 55-67 ( 1 967 ) . H . Hosh i no and Iv1 . Sh imoj i , "Thermoel ectric power of i on i c crystal s . I I I . Thermoel ectr ic power and conducti vi ty of potass i um bromi de conta i n i ng barri um bromi de , " J . Phys . Chern. , 29 1 431 -41 ( 1 968 ) ; IL., 1 553-63 ( 1 970 ) . -

- -

1 80 . ( a ) A . Sch i ral di , " Thermoel ectri c power of S- and y- si l ver i od i de , I I Z . Phys . Chern. ( Frankfurt/Ma i n ) , 97 , 285- 93 ( 1 974 ) . ( b ) A . Magi stri s , E . P . Pezzat� and C . S i n i stri , "Thermoel ectric propert i es of h i gh-conducti v i ty sol i d el ectrolytes , " Z . Naturforsch , 27A, 1 379-81 ( 1 972 ) .

1 81 . P . D . Mi l l er , A . B . Tri pl er , J r . , and J . J . l�ard , l iThe appl i cation of i rrevers i b l e thermodynami cs to the thermogal van i c behav i or of copper­copper su l fate systems , " J . El ectrochem. Soc . , 1 1 3 , 746-9 ( 1 966 ) .

1 82 . B . H . Cl ampi tt and D . E . German , " El ectrochemi cal cel l for convers i on of heat energy , " U . S . Patent 3 , 253 , 955 ( 1 966 ) .

207

- . TR-416 S=�I I_I -----------------------------

1 83 . G . Hoffmann and A . Dav i d , " Investi gati ons of some el ectrolyte thermo­p i l es , " Acta Ch im . ( Budapest ) , 78 , 373-85 ( 1 973 ) .

1 84 . H . A . L i ebhafsky , "Thermogal van i c cel l , " U . S . Patent 2 , 882 , 329 ( 1 959 ) .

1 85 . B . W . Burrows , " Redox thermogal van i c cel l s for di rect energy conver­s i on , " Proc . l Oth Intersoc i ety Energy Convers i on Engi neeri ng Confer­ence , pp . 82 1 - 7 ( 1 97 5 ) .

1 86 . B . Burrows , " Di scharge behavior of redox thermogal van i c cel l s , " J . El ectrochem . Soc . , 1 23 , 1 54- 9 ( 1 976 ) .

1 87 . L . I . Bel l chi ns kaya , S . A . Ka l uzh i na , and A . Y . Shata l o v , "Temper­ature dependence of thermogal van i c current , " El ektrokh imiya , §" 228-30 ( 1 970 ) . . _ . _

1 88 . S . A . Ka l uzh i na , G . A . Mi troshki na , and A . Ya . Shata l o v , " El ectro­chemi cal aspects of thermogal van i c -cel i s. I I . Thermoga l vani c cel l s wi th i ron i n aci d i c sul fate el ectro lyte , " El ektrokhimiya , lQ, 924-7 ( 1 974) .

1 89 . S . A . Ka l uzh i na and G . . A . Mi troshki na , " El ectrochemi cal aspects of thermogal van i c cel l s . I I I . Temperature dependence of the current of thermogal vani c cel l s on ni cke l , in an ac id sulfate el ectrolyte , " El ektrokhi miya , l 2 � 1 01 3- 5 ( 1 976) .

1 90 . S . A . Ka l uzh i na and G . A . Mi troshk ina , " El ectrochemi cal aspects of thermoga l van ic e l ements . V . Temperature dependence of the current of thermQgal van i c e l ements in an aci di fi ed sol uti on of proper i ons , · · El ektrokhi miya , 11, 630-40 ( 1 978 ) .

1 91 . F . A . Ludwi g and J . J . Rowl ette , " Conti nuous concentration cel l , " U . S . Patent 3 , 23 1 , 426 ( 1 966 ) .

1 92 . R . H . Deys her , " Thermoel ectri c cel l , " U . S . Patent 2 , 31 0 , 354 ( 1 943 ) .

1 93 . H . S . Harned and B . B . Owen , Phys i ca l Chemi stry of El ectrolyti c Sol u­t i ons , Re i nhol d , New York p . 436 ( 1 950 ) .

1 94 . J . C . Angus , " Conti nuous gas concentration cel l s as thermal ly regenera­ti ve , ga l van i c cel l s , " i n Regenerati ve EMF Cel l s , C . E . Crouthamel and H . L . Recht , eds . , Advances i n Chemi stry Seri es , 64 , Ameri can Chemi cal Soc i ety , Wash i ngton , D . C . pp. 1 1 - 1 6 ( 1 967 ) .

--

1 95 . J . C . Angus , " Method and apparatus for di rect convers i on of thermal energy to el ectr ical energy, " U . S . Patent 3 , 51 1 , 71 5 (1 970 ) .

1 96 . J . T . Kummer and N . Weber , " Thermoel ectri c generators , " U . S . Patent 3 , 458 , 356 ( 1 968 ) .

1 97 . N . Weber , " A thermoel ectr ic dev i ce based on beta-al umi na sol i d el ec­trolyte , " Energy Convers ion , 11, 1 - 8 ( 1 974 ) .

208

S=�I 1:_1 __________________ � _______ T�R_-4_1_6 -� �

1 98 . T . K . Hu nt , N . We ber , a nd T . Col e , " Output power and effi ci ency for a sodi um thermoel ectr i c heat eng i ne , " Proc . 1 0th I n ters oci ety Energy Co nvers i on Engi neeri ng Conference , I , Newa r k , Del aware , pp . 231 -4 ( 1 97 5 ) .

1 99 . T . K. Hunt , N . Weber , and T . Col e , " Research on the sodi um hea t en g i ne , " Proc . 1 3th I n tersoc i ety Energy Convers i on Eng i neeri ng Conf . , San Di ego , Ca l i forn i a , pp . 201 1 - 7 ( 1 978 ) .

200 . " Thermoe 1 ectri c generator devi ces and methods , U . S . Patents 3 , 51 1 , 71 5 and 4 , 098 , 958 ( 1 978 ) .

201 . Co nversati on wi th N . Weber , T . Col e , and T . K . Hunt ( H . L . Chum) , Dea rborn , Mi c hi gan ( May 1 979 ) .

202 . G . J . May , l iThe devel opment of beta- al umi na for use i n el ectrochemi cal ce 1 1 s : a survey , I I J . Power Source�, 1, J - 22 ( 1 978 ) .

203 . R . Knod1 er and W . Bauka 1 , " Determi nation of the l i fe properti es of beta­a 1 umi na tubes , I I J . Power Sources , 1, 23-8 ( 1 978 ) .

204 . G . R . B . El l i ott , " E1 ectrochemi cal heat eng i ne " , LA-6632-MS ( 1 976 ) .

205 . G . R . B . El l i ott , W . J . Trel a , and G . E . D i a l s , " �l e�trochemi cal heat engi nes for d i rect el ectr i c power generat i on and energy storage , " Proc . A/�A/SAE 1 1 th Propui s i on Conference , Anahe i m , Cal i forn i a , AIAA Paper 75;- 1 237 ( 1 97 5 ) .

206 . G . R . B . El l i ott , " E1 ectri c i ty and storage for res i dences us i ng L i / I 2 e l ectrochem i ca l engi nes to augm�n t photovol ta i cs , ·· Proc . of the E1 ectrochem . Soc . Meet i n g , 78 , 428- 35 ( 1 978 ) .

207 . G . R . B . E l l i ott and N . E . Vanderborgh , " El ectrochemi cal heat engi nes fo r power generati o n , l oad- l evel i n g at s i te for underground coal conver­s i on , " Proceedi ngs of the 1 3th I ntersoci ety Energy Convers i on Engi neeri ng Co nf . , San D i ego , Cal i forni a , pp . 373·79 ( 1 978 ) .

208 . G . R . B . El l i ott and N . E . Vanderba rg h , " N i ght . storage and bac kup .generation wi th e lectrochemical engi nes , " LA-UR- 78- 606 . Los Al amos , New Mexi co : Los Al amos Sc i enti fi c La bo rato ry ( 1 978 ) .

209 . G . R . B . El l i ott , " El ectrochemi cal heat en g i ne , " U . S . Patent 4 , 090 , 01 2 ( 1 978 ) .

2 1 0 . C . R . McCu l l y , " Therma l l y regenerati ve ga l van i c cel l emp l oy i n g the fl uori des of a rsen i c , ceri um and uran i um , " U . S . Patent 3 , 31 8 , 734 ( 1 967 ) .

2 1 1 . Wri tten corres pondence between J . J . Lander ( D i rector of El ectrochemi cal Res earch , General Motors ) and R . D. Weaver i ndi cat i ng that the concept of the Na-C1 2 el ectrotherma l l y regnerati ve trans ducer was proposed i n September 1 959 and that actual l aborato ry work started a t that t i me .

209

TR-416 S=�I I.I ---.,---------------------------- � �

2 1 2 . J . J . Lander , S . W . -Smi th , and R . D . Weave r , " El ectrotherma1 transducer , " U . S . Patent 3 , 370, 983 ( 1 968 ) .

2 1 3 . R . D . Weaver , " Feasi bi l i ty study of the e1 ectrotherma l l y regenerati ve transduce r , I I AD607 2 93 ( 1 964 ) .

2 1 4 . D . A . J . Sw i n kel s , " Li th i um- ch 1 o ri ne battery , " J . E1 ectrochem . Soc . , 1 1 3 , 6 - 1 0 ( 1 96 6 ) . D . A . J . Swi n kel s and R . N . Seefurth , " Characteri zati on of a porous graph i te C1 2 e1 ectrode , " J . E1 ectrochem. Soc . , 1 1 5 , 994-9 ( 1 968 ) .

2 1 5 . D . A . J . Swi n kel s , " E1 ectro 1 ys i s of fused a l kal i meta l hal i des and al ka l i metal - a l ka l i metal ha l i de ha l ogen fue l ce1 1 s , " Bri t . Patent 1 , 1 44 , 388 ( 1 969) . _ __ _ _ _ _ __ _ _ _._ _ _

D . A . J . Swi nkel s, " El ectrochemi ca1 wel l wi th l ayered el ectrode of cerami c and ca rbon , " U . S . Patent 3 , 544 , 373 ( 1 970 ) .

2 1 6 . R . H . Hammond and W . M . Ri sen , J r . , "An el ectrochemi cal heat eng i ne for d i rect sol ar energy convers i on , " Sol a r Energy , fl, 443- 9 ( 1 980 ) .

2 1 7 . M . S . S . H su and T . B . Reed , " E1 ectrochemi ca 1 power and hydrogen genera,... t i on from h i gh temperatu re el ectrol yte ce 1 1 s , " Proceedi ngs of the 1 1 th I n tersoci ety Energy Convers i o n Engi neeri ng Conference , State L i ne , Nevada , p p . 443- 6 ( 1 976 ) .

2 1 8 . M . S . S . H s u , W . E . r�orrow, J r . , and J . G . Goodenough , " H i gh effi c i ency el ectrochem i ca l p1 ant , " Proceedi ngs of the 1 0th Intersoci ety Energy Convers i on Engi neeri ng Conference , Newark , De l aware , pp . 555- 63 ( 1 975 ) .

2 1 9 . W . E . Mo rrow, J r . , and �1 . S . S . Hs u , " E1 ectr i c power pl ant us i ng el ectro­l yti c cel l - fuel cel l comb i nati on , " U . S . Patent 4 , 087 , 976 ( 1 978 ) .

220 . Iv) . Stei nberg , " Thermoe1 ectrochemica1 power cycl es , " BNL-21 323 R , Brookhaven , New Yo rk : Brookhaven National Laboratory ( 1 978 ) .

2 2 1 . F . J . Sa l zano and S . Sr i n i vasan , Proceedi ngs of the Fi rst Internati ona l Energy vJater El ectro l ys i s Workshop , September 23- 25 , 1 976 , BNL 2 1 1 65 , B roo khaven , New York : Broo khaven Nati onal Laboratory .

222 . S . Sri n i vasan , F . J . Sal zano , and A . R . Landgrebe , eds . , Proceed i ngs o f the Sympos i um on I ndustr i a l Water El ectro l ys i s , 78- 4 , The El ectro­chemi cal Soci ety , I n c . , Seattl e , Wa s h i n gton ( 1 978 ) . --

223 . Na ti onal Fuel Ce l l Semi nar Abs tracts , coordi nated by Cou rtesy As soci ates , I n c . , Bethesda , Ma ryl and , June 26- 28 , 1 979 .

224 . W . E . O ' Grady , S . S ri n i vasan , and R . F . Dud1 ey , eds . , Proceedi ngs of the Workshop on the E1 ectrocata 1ysts of Fuel Cel l Reacti ons , Broo khaven Na ti onal Labo ratory , Brookhaven , New York , May 1 5 - 1 6 , 1 978 , Vo l . 79- 2 , The El ectrochemi cal Soc i ety , I nc . ( 1 979 ) .

2 10

S=�I I.I ______________________ T_R_-4.:.....1_6 _ -� �

225 . For examp l e : H . S . I s aacs and L . J . Ohner , l iThe overpotenti a l be-havi or of el ectrode mater i a l s at i nterfaces wi th Zr02-Y20 3 el ectrol ytes , " Proc . El ectroc hemi cal Socl ety Meet i n g , Los Angel es , Cal i forn i a , October 1 4- 1 9 , 1 97 9 , pp . 371 - 2 .

.

226 . Fo r exampl e : A . O . I s en berg , " Mass transport i n sol i d oxi des as re­l ates to the fabri cati on of hi gh temperature sol i d state fue l ce l l s , " Proc . El ectroc hemi cal Soc i ety Meeti ng , Los Angel es , Cal i forn i a , October 1 4- 1 9 , 1 97 9 , p p . 364 - 5 .

2 2 7 . J . Greenberg , L . H . Th a l l er., and D . E . ��eber , "A poss i bl e regenerat i ve mol ten sal t thermoel ectri c fue l ce 1 1 , " NASA Techn i ca l Note N64-27361 ( 1 964 ) .

. .

2 28 . J . Green berg a nd L-. H . -Thaller , " Combi ned el ectro l ys i s dev i ce and fue l cel l a n d method of o perati on , " U . S . Patent 3 , 357 , 862 ( 1 968 ) .

2 1 1

Document Control 1 1 . SER I Report No. 1 2. NTIS Accession No. Page _TR -332-�li.1[a.1 2.

4. Title and Subtitle Review of Thermally Regenerative Electrochemical Systems

Vol. 2 7. Author(s)

Helena L . Chum ; Robert A . Oste�y�unz 9. Performing Organization Name and Address

Solar Energy Research Institute 1617 Cole Boulevard Golden , Colorado 80401

1 2. Sponsoring Organization Name and Address

1 5 . Supplementary Notes

3. Recipient's Accession No.

5. Publication Date

A�ril 1981 . 6 .

8. Performing Organization Rept. No.

1 0. ProjecVTask/Work U nit No.

3356 . 50 1 1 . Contract (C) or Grant (G) No.

(C)

(G)

13. Type of Report & Period Covered

Technical Report 1 4.

L I:: I'" I-I 1::. I' ... t: r I::: t:

i [:: I: I: i ,.

I ,-\ [ I

1 6 , Abstract (L imit: 200 words) [ Thermally regenerative electrochemical systems (TRES) are closed systems that convert ' � heat into electricity in an electrochemical heat engine that �s Carnot cycle limited I in efficiency. Past and present work on such systems is revie,;ved . Two broad classes r:: of TRES are based on the types of energy inputs required for regeneration : thermal C alone and coupled thermal and electrolytic . The thermal regeneration alone encom- I passes electrochemical systems (galvanic or fuel cells) in which one or more products ' r are formed . The regeneration can b e performed in single or multiple steps . The i t compounds include metal hydrides, halides, oxides , chalcogenides , and alloys or bimetallic systems . The coupled thermal and electrolytic regeneration encompasses electrochemical systems ( galvanic or fuel cells) regenerated by electrolysis at a different temperature or different pressure . Examples include metal halides and water . Thermogalvanic or nonisothermal cells are included in this category . Also

i L L

( r: included are electrochemical engines in . which the working electroactive fluid is \ isothermally expanded through an electrolyte . TRES cover temperature ranges from i f about 20°C to 1000°C . Engines with power outputs of 0 . 1 mW/cm2 to 1 W/cm2 have been L demonstrated . Recommendations are made of areas of research in science and �ngineer- ( ing that would have long-range benefit to a TRES program . ( L

[ 1 7. Document Analysis '( 1

a, Descriptors Thermal Regeneration ; Electrochemical Cells ; Regenerative Fuel Cells ; Fue]: Cells ; Thermogalvanic Cells ; Thermal Reactors ; Distillation ; Electrochemical Engines ; .I Electrochemical Heat Engines ; Hydrides ; Halides; Chalcogenides ; Oxides ; Bimetallic I I Systems ; Alloy Systems ; Electrothermal Systems ; Coupled Electrolytic Thermal Regen- ( eration ; Double Thermogalvanic System I I

b . Identifiers/Open-Ended Terms Thermally Regenerative Electrochemical Systems ' l c. U C Categories (TRES) ( I { I

1 .... 1 8-.-A-v-a-i l-ab-i-l i-ty-S-t-a-te-m-e-n-t -----------------� --------------------------------i7.1 9�.�N�O�.-o�f�p�a�g�es�-----------,\ l

National Technical Information Service 227 (

61

u . S . Department of Commerce ( .. � 20. Price 5 285 Port Royal Road l Springf ield , V irginia 22161 $9 . 50 { L-����������....-.-......_ ________________ � ______________ � _

Form No. 8200-1 3 (6-79)