AD
DEVELOPMENT OF HIGH-ENERGY BATTERIES FOR ELECTRIC VEHICLES
Progress Report for the Period
July 1970-June 1971
By
E. J. Cairns
R. K. Steunenberg
J. P. Ackerman
B. A. Feay
D. M. Gruen
M. L. Kyle
T. W. Latimer
J. N. Mundy
R. Rubischko
H. Shimotake
D. E.Walker
A. J. Zielen
A. D. Tevebaugh
July 1971
Prepared for
The Division of Advanced Automotive Power Systems Development
Environmental Protection Agency
Ann Arbor, Michigan
Chemical Engineering Division
ARGONNE NATIONAL LABORATORY
9700 South Cass Avenue
Argonne, Illinois
60439
-------
Work performed under an agreement between the ENVIRONMENTAL
PROTECTION AGENCY, Office of Air Programs, Division of Ad-
vanced Automotive Power Systems Development, and the United
States Atomic Energy Commission.
-------
Argonne National Laboratory
Chemical Engineering Division
9700 South Cass Avenue
Argonne, Illinois
60439
DEVELOPMENT OF HIGH-ENERGY BATTERIES FOR ELECTRIC VEHICLES
Progress Report for the Period July 1970 - June 1971
by
E. J. Cairns
R. K. Steunenberg
J. P. Ackerman
B. A. Feay
D. M. Gruen
M. I. Kyle
T. W. Latimer
J. N. Mundy
R. Rubischko
H. Shimotake
D. E. Walker
A. J. Zielen
A. D. Tevebaugh
July 1971
Prepared for
the Division of Advanced Automotive Power Systems Development
Environmental Protection Agency
Ann Arbor, Michigan
-------
FOREWORD
This is the second annual technical progress report of a research and
development program conducted by the Chemical Engineering Division of Argonne
National Laboratory under an agreement between the United States Atomic
Energy Commission and the Environmental Protection Agency, Office of Air
Programs, Division of Advanced Automotive Power Systems Development. The
period covered by this report is July 1, 1970, through June 30, 1971.
The long-term goal of this program is to develop the technology for
the construction and testing of a lithium/sulfur battery suitable for power-
ing an electric family automobile. The immediate goals are the development
and scale-up of single lithium/sulfur cells, followed by the development and
testing of small (1-2 kW) batteries.
Overall program management is the responsibility of Dr. R, C. Vogel,
Division Director, and Dr. A. D. Tevebaugh, Associate Division Director.
Technical direction is provided by Dr. E. J. Cairns, Section Head, and
Dr. R. K. Steunenberg, Group Leader. The project officer for OAP is
Mr. Charles E. Pax.
-------
TABLE OF CONTENTS
Page
ABSTRACT 1
SUMMARY „ 2
I. INTRODUCTION 6
II. INVESTIGATIONS OF LITHIUM/SULFUR CELLS 9
A. Experimental Procedure 9
B. Results and Discussion 13
C. Conclusion 26
III. SUPPORTING LABORATORY INVESTIGATIONS 30
A. Phase Equilibrium Studies of Electrolyte-Containing Mixtures. 30
B. Studies of Sulfur-Bearing Species in Molten Alkali Halides. . 33
1. Electrodes 34
2. Direct Sulfide Analysis in Molten Salt 35
3. Electrochemical Generation of S2 36
4. Spectrochemical Studies 38
5. Analysis of Frozen Eutectic Samples 38
C. Solid-Electrolyte Studies 38
1. Li20-MgO-Al203 System 39
2. Li20-La203-Al203 System 40
3. Sodium 6- and g"-Alumina 41
4. Conductivity 43
D. Cathode Material Studies 44
E. Mass-Transport Studies. ... 47
IV. MATERIALS TESTING AND FABRICATION 52
A. Experimental Procedure 52
B. Metallic Components Studies 53
C. Seals and Insulating Component Studies 59
D. Development and Fabrication of Ceramic Insulators 61
1. Lithium Aluminate 62
2. Lithium-Aluminate Sintering Studies 62
3. Cell Insulator Production 63
E. Cathode Current Collector Development 66
iii
-------
TABLE OF CONTENTS
Page
V. ELECTRIC AUTOMOBILE PERFORMANCE CALCULATIONS 67
A. Cell Design 67
B. Method of Calculating Automobile Performance 69
C. Results and Discussion 76
D. Lithium Reserves 79
VI. STATUS AND FUTURE PLANS 80
REFERENCES 82
APPENDIX: SUMMARY OF PERFORMANCE OF CELLS OPERATED IN FISCAL YEAR
1971 84
iv
-------
LIST OF ILLUSTRATIONS
No. Title Page
1. Schematic Diagram of a Typical Unsealed Laboratory Cell 10
2. Schematic Diagrams of Some Cathode Current Collector Designs ... 13
3. Schematic Diagram of a Mixed Cathode 15
4. Typical Voltage-Current Density Curve for a Lithium/Sulfur Cell
With a Laminated Cathode Having All-Stainless-Steel Laminae. ... 16
5. Typical Voltage-Capacity Density Curve for a Lithium/Sulfur Cell
With a Laminated Cathode Having All-Stainless-Steel Laminae. ... 17
6. Voltage-Capacity Density Curves for a Lithium/Sulfur Cell With
a Laminated Cathode Having Porous Graphite and Stainless Steel
Feltmetal Laminae, , . , , 18
7. Voltage-Capacity Density Curves for a Lithium/Sulfur Cell With
a Comb-Type Cathode, .,.-,.-,......„.. 19
8. Voltage-Capacity Density Curves for a Lithium/Sulfur Cell With
an Enclosed Laminated Cathode , 20
9. Capacity Density as a Function of Cycle Number for a Lithium/
Sulfur Cell With an Enclosed Laminated Cathode 21
10. Voltage-Capacity Density Curves for a Lithium/Sulfur Cell With
a Reservoir Cathode. ,,,,,,,.. 21
11. Voltage-Capacity Density Curves for a Lithium/Sulfur Cell With
an Enclosed Reservoir Cathode.- , 22
12. Capacity Density as a Function of Cycle Number for a Lithium/
Sulfur Cell With a Mixed Cathode . . . , . . . 24
13. Voltage-Capacity Density Curves for a Lithium/Sulfur Cell With
a Mixed Cathode, ............ 24
14. Voltage-Capacity Density Curves for a Lithium/Sulfur Cell With
a Mixed Cathode Containing a Niobium Expanded-Mesh Spiral
Current Collector ,,.,..., 25
15. Schematic Drawing of a Sealed Lithium/Sulfur Cell 26
16, Phase Diagram of the Pseudo-ternary System Li2S-S(LiBr-RbBr) ... 30
17. Molten-Salt Cell Assembly for Sulfide-Ion Investigations 34
18. Precipitation Titration Curve for S2 With Generated Ni2+;
Experiment LK-4. ,...,, . » 35
-------
LIST OF ILLUSTRATIONS
No- Title Page
19. The Potential of the Ni/Ni2+ Electrode as a Function of S2~
Concentration 37
20. Ionic Conductivity of Aluminate Samples 44
21. Schematic Diagram of the Lithium Electrode Test Cell 48
22. Typical Voltage-Time "Behavior of a Li/LiCl-KCl/Li-Al Cell at
425°C and at a Current Density of 0.25 A/cm2 48
23. Lithium/Sulfur Test Cell 51
24. Corrosion Rates of Metals Exposed to 20 at. % Li-S Mixtures at
375°C 53
25. Schematic Diagram of a Chromium-Plated Cell 55
26. Voltage-Capacity Density Curves for a Chromium-Plated Li/S Cell
With LiBr-RbBr Electrolyte 56
27. Voltage-Capacity Density Curves for a Chromium-Plated Li/S Cell
With LiF-LiCl-KCl Electrolyte 58
28. Corrosion Rates of Electrical Insulators in Molten Lithium at
375°C 61
29. Photograph of a Rubber'Die and Sealing Plug 65
30. Photograph of Isostatically Pressed LiA102 Insulator Rings .... 65
31. Specific Energy and Schematic Design of a Li/LiCl-KCl/Li in S
Square Laminated Sealed Cell 68
32. Conceptual Li/S Cell Design Used in Automobile Performance
Calculations 70
33. Voltage-Capacity Density Data and Empirical Curves for Li/S Cells
. With Laminated Cathodes 74
34. Driving Profiles With Corresponding Power Profiles for the
Automobile of Table XXII 75
35. Electric Automobile Ranges for Selected Driving Profiles 78
36. Effect of Automobile Velocity on Range for Constant Velocity
Driving 78
vi
-------
LIST OF TABLES
No. Title Page
I. Lithium/Sulfur Battery Development Program Goals 8
II. Physical Properties of Molten-Salt Electrolytes 10
III. Physical Characteristics and Electrical Performance of Two
Mixed-Cathode Cells 23
IV. Construction and Performance Characteristics of a Sealed
Lithium/Sulfur Cell 27
V. Performance of Some Typical Lithium/Sulfur Cells 29
VI. Phases Present at Selected Compositions in the System
Li2S-S-(LiBr-RbBr) 31
VII. Solubility of L2 in L3 for Various Electrolytes at 360 and
400°C , 32
VIII. Titration of S2~ in LiCl-KCl Eutectic with a Nickel Anode;
Evaluation of Ni/Ni2+ Formal Potential and NiS Solubility
Product „ 36
IX. Starting Compositions (in wt %) of Li20-MgO-Al203 Specimens . . 40
X. Weight Loss and Density of Li20-MgO-Al203 Compositions Fired
at 1700 and 1800°C 40
XI. Starting Compositions (in wt %) of Li20-La203~Al203 Specimens . 41
XII. Weight Losses and Densities of Li20-La203^X263 Compositions
Fired at 1500, 1650, and 1700°C 41
XIII. Fired Densities of Monofrax H Specimens 42
XIV. Effect of Milling Time on Unfired and Fired Densities of Alcoa
3-alumina 42
XV. Electrical Conductivities of the Two Immiscible Liquid Phases
in the Sulfur-Rich Region of the Thallium-Sulfur Phase
Diagram . . , 46
XVI. Summary of Lithium Electrode Cell Tests 49
XVII. Corrosion of Aluminum and Ferritic Stainless Steel in LiCl-KCl
Electrolyte and 20 at. % Li-S Mixtures at 375°C 55
XVIII. Performance of a Chromium-Plated Li/S Cell 57
XIX. Corrosion Rates of Electrical Insulators in Molten Lithium. . . 60
vii
-------
LIST OF TABLES
No. Title Page
XX. Characteristics of LiA102 Powders 62
XXI. Characteristics of Sintered LiA102 Ceramic Bodies 64
XXII. Electric Automobile Characteristics 69
XXIII. Weight and Material Cost Breakdown for Conceptual Cell and
Battery Design 71
XXIV. Electric Automobile Performance Under Selected Driving
Profiles 77
Vlll
-------
DEVELOPMENT OF HIGH-ENERGY BATTERIES FOR ELECTRIC VEHICLES
by
E. J. Cairns, R. K. Steunenberg, J. P. Ackerman, B. A. Feay,
D, M. Gruen, M. L. Kyle, T, W, Latimer, J. N. Mundy,
R. Rubischko,* H. Shimotake, D. E. Walker, A. J. Zielen,
and A. D. Tevebaugh
ABSTRACT
The objective of the High-Energy Battery Development Program at Argonne
National Laboratory is to develop the technology required to construct sec-
ondary batteries having the performance capabilities required for pollution-
free electric automobiles. Batteries for this application should have an
energy-storage capability of 220 W-hr/kg and be able to deliver power at a
peak rate of 220 W/kg. Their cost should not exceed about $10/kW-hr of energy
storage capability. Lithium/sulfur cells using a molten lithium halide-contain-
ing electrolyte and operating at 360 to 390°C have achieved capacity den-
sities of up to 0.52 A-hr/cm2 (above 1 V) at a current density of 0.52 A/cm2.
These results are consistent with the specific energy and specific power
goals, but the cycle life (currently hundreds of cycles) and the sulfur elec-
trode performance require further improvement.
The cell development program is supported by laboratory studies in
various areas. The solubility of cathode materials in various electrolytes
and the identity of the soluble species are being studied. Preliminary results
indicate that electrolytes containing only fluoride and chloride anions have
the lowest solubility for sulfur-bearing species. Investigations of various
additives to sulfur indicate that thallium may be useful in reducing the
vapor pressure and increasing the conductivity of sulfur. A survey of candi-
date solid electrolytes has led to the investigation of the lithium form of
6-alumina. Mass transport studies have shown that an improved cycle life of
the lithium electrode (hundreds of cycles) can be obtained by using LiF-LiCl-
KC1 electrolyte and by minimizing the concentration of various impurities in
the cellc
The investigation of the corrosion rates of various materials at 375°C
has shown that chromium and molybdenum had low corrosion rates in 20 at. %
Li in S, and lithium aluminate, beryllia, aluminum nitride, yttria, thoria,
Y3Al50i2, and MgO • A^OS had low corrosion rates in lithium. Work has begun
on the preparation of various forms of some of these materials, with most
of the effort on LiA102 shapes.
Battery design and performance calculations carried out for a 1075-lb
Li/S battery in a 4300-lb (curb weight) electric automobile indicate that
a range of 88 to 225 miles can be expected under realistic driving con-
ditions, depending on the driving profile and accessory load.
*Resident Associate from Gould, Inc.
-------
- 2 -
SUMMARY
Investigations of Lithium/Sulfur Cells
Experimental studies of small cells of the types Li/LiX/S and
Li/LiX/P^S-j o, where LiX represents a low-melting salt mixture containing
lithium halides, operating at x375°C have indicated that these cells show
promise for meeting the requirements of a low-cost, high-specific-energy
and high-specific-power battery for electric automobile propulsion. The
immediate goals of the cell program are to produce laboratory cells with
a specific energy of 220 W-hr/kg and a specific power of 220 W/kg. These
goals, when related to single-cell performance, imply a capacity density
(at the 1-hr rate) of 0.4 A-hr/cm2, a sulfur utilization near 70%, a
capacity per unit volume of cathode of 1.0 A-hr/cm3, and for the battery
to be economical, a cycle life of over 1000 cycles.
During the last year, significant progress in cell design and per-
formance has been made. As a result of the investigation of a number of
new cathode current collector designs, laboratory cells have demonstrated
capacity densities (at the 1-hr rate) of up to 0.52 A-hr/cm2. Other cells,
operating at lower current densities and capacity densities, have demon-
strated a cycle life of more than 800 discharge-charge cycles (over 1100 hr
of operation). Although these results are encouraging, the capacity per
unit volume (currently --0.5 A-hr/cm3), the sulfur utilization (currently
30 to 50%), and the retention of good performance levels over long cell
lifetimes must be improved in order to satisfy the requirements for prac-
tical electric vehicle batteries.
Phase Equilibrium Studies of Electrolyte-Containing Mixtures
In order to minimize possible difficulties caused by transport of cathode
material through the electrolyte of lithium/sulfur cells, the mutual solu-
bilities of several electrolyte-cathode material combinations are being
determined. This information will be used to select the optimum electrolyte
and cell operating temperature range.
The present aim of the experimental program is to determine the pseudo-
ternary phase diagram of the Li2S-S-electrolyte system using various can-
didate cell electrolytes such as LiBr-RbBr, LiCl-Lil-KI, LiF-LiCl-LiI,
LiCl-KCl, and LiF-LiCl-KCl. Initial indications are that the system is
similar to the Li2Se-Se-electrolyte system, which has been studied in some
detailo There are apparently one solid-phase and three liquid-phase fields
in the diagram. The complete phase diagram has not yet been determined,
but it is clear from the results gathered so far that salt mixtures con-
taining only the small anions (chloride and fluoride) are the desirable cell
electrolytes if the choice is made solely on solubility considerations.
Studies of Sulfur-Bearing Species in Molten Alkali Halides
The goals of this research are to study the molten salt chemistry,
electrochemistry, and spectroscopy of the sulfur-sulfide-polysulfide system.
Sulfide solutions free of elementary sulfur have been prepared at 400 and
450CC in LiCl-KCl eutectic salts by electrolysis with a nickel sulfide
cathode: NiS + 2 e~ = Ni + S2~. (The reverse reaction, or precipitation
-------
- 3 -
titration with a nickel anode, has been used for coulometric in situ sulfide
analysis.) The results have yielded solubility data on Li2S and NiS and
the formal potential of the Ni/Ni2+ couple. It has been demonstrated that
the Ni/Ni2+ couple can serve as a sulfide ion monitor and that the cal-
culated concentrations so obtained are in agreement with the coulometric
generation values. A "standard" sulfide solution was prepared electro-
chemically, and it is being used in a spectrophotometric study of sulfur,
sulfide, and polysulfide species. Analytical methods for sulfide and total
sulfur in solid salts are under evaluation. Excellent results have been
obtained for sulfide ion using an Orion silver/sulfide specific ion electrode
for potentiometric end-point detection in argentometric titrations in
alkaline ammoniacal solutions.
Solid Electrolyte Studies
Solid electrolytes that are capable of adequate lithium-ion transport
in lithium/sulfur cells are being investigated as alternate electrolytes
to the molten salt in current use. The Li20-MgO-Al203, Li20-La203~Al203,
and sodium B- and 6"-alumina systems have been studied. The activation
energies for the lithium-containing materials were between 17 and 21
kcal/mol. Conductivity was found to be linear when plotted as log con-
ductivity versus 1/T and in the range of 10~8 to 10"^ ohm"1 cm"1 in the
temperature range of 350 to 900°C. The conductivities of the sodium aluminates
show a curved relationship and an activation energy of ^5 kcal/mol when plotted
on similar coordinates. Conductivities in the range of 10~^ to 10"1 ohm"1 cm"1
were measured in the temperature range from 350 to 900°C. This conductivity
of the sodium aluminates is lower than anticipated and it is expected that with
improved fabrication techniques a conductivity near 10"1 ohm"1 cm"1 at 400°C
can be obtained. The molten-salt electrolytes in current use have conductivities
near 1 ohm"- cm"1. None of the solid-electrolyte data obtained to date appears
sufficiently promising to warrant cell operation with these electrolytes. Con-
sequently, this investigation will concentrate on determining the materials
systems, and the methods of producing them, that show promise as solid lithium-
ion conductors.
Cathode Materials Studies
Although the small-scale tests have shown that elemental sulfur alone
can be used as the active cathode material in lithium/sulfur cells, the use
of an additive to increase its electronic conductivity or decrease its vapor
pressure may prove beneficial. Iron does not appear to be a useful additive
because of the slow dissolution rate of iron sulfides in sulfur. The addi-
tion of thallium to liquid sulfur results in the formation of two liquid
phases. The lower-density, sulfur-rich phase has a conductivity of ^10~9
ohm"1 cm"1; the higher-density, thallium-rich phase has a much higher con-
ductivity of "'4 x- 10"] ohm"1 cm-1 at 367°C. Thallium may be a useful additive
if it is effective at low concentrations. Sulfur containing a proprietary
additive, which was obtained from a commercial source, has a relatively
high conductivity of 1.4 * 10~3 ohm"1 cm"1 at 158°C, but it does not appear
to be suitable for use in the cathodes of lithium/sulfur cells because of
its tendency to release gases at cell operating temperatures. It also appears
that the additive and the sulfur tend to segregate.
-------
- 4 -
.'•lass-Transport Studies
In an effort to improve the cycle life of the lithium anode in lithium/
sulfur cells, a study has been undertaken on the transport characteristics
of lithium between a lithium electrode of the type used in cells and a Li-Ai
counter electrode. Four cells were operated. The first three cells using
LiCl-KCl, LiCl-Lil-KI and LiCl-Lil-KI electrolytes and A1203 electrical
insulators, failed in less than 10 cycles (100 hr) of operation. The
principal mode of failure was due to lithium not soaking into the Feltmetal
during charging, This characteristic appeared to be strongly affected by
the presence of contaminants such as traces of moisture or corrosion products
resulting from lithium attack on the A12C>3 in the electrolyte.
In the operation of the fourth cell, an effort was made to eliminate
contaminants. The electrolyte, LiF-LiCl-KCl, contained no iodide salt
(normally among the salts most difficult to obtain in an anhydrous form), and
the cell insulator was changed to BeO, which is less susceptible to lithium
attack than is A1203. This cell operated well for 480 hr and 138 cycles,
after which 10 g of LiCl-Lil-KI was added. The resulting bubbling indicated
the presence of moisture and the performance of the cell rapidly deteriorated.
Under proper operating conditions, the present lithium anode, comprised of
lithium in Felcmetal (Huyck Metals Co., Type 302 stainless steel, 90% porosity,
67-pm pore size), appears capable of good cycle life.
In order to increase the capacity density of the sulfur cathode, a
study of various cathode designs and materials has been undertaken. An attempt
to develop a cell with a quartz housing failed because the quartz was too
susceptible to lithium attack. Preliminary tests with a new cell design have
been encouraging.
Materials Testing and Fabrication
The corrosion resistance of various materials to simulated cell envi-
ronments is being studied to identify materials w'.th potential usefulness
for cell applications. Two different classes of materials resistant to cell
conditions are being studied. Corrosion-resistant materials possessing good
electrical conductivity are required to serve as current collectors and
housings. Corrosion-resistant, electrically insulating materials are required
to prevent short-circuit contacts within the cell. Among the conductive
materials that were tested at 375°C in 20 at. % Li-S mixtures, molybdenum and
chromium had low corrosion rates in both short-term (100-300 hr) and long-term
(600 hr) tests. Molybdenum, in particular, has been used in many of the exper-
imental cells for periods up to several hundred hours and has shown good corrosion
resistance under these conditions.
The austenitic stainless steels, such as Types 2RK65, 347, and 205
(15 wt % Mn), had low corrosion rates in short-term tests, but much higher
rates in the longer tests. The nickel-base alloys, such as Inconels 702
and 600 and Hastelloy X, as well as Zircaloy-2, showed similar behavior.
The corrosion rate of aluminum by 20 at. % lithium-sulfur mixtures has proved
to be highly variable, with rates ranging from 0.01 to 2.4 mm/yr in 600-hr
tests. It is suspected that these variations are associated with the nature
of a surface film that is formed during the exposure. Tantalum, iron,
-------
- 5 -
titanium, nickel, and beryllium showed poor corrosion resistance to lithium-
sulfur mixtures and have been excluded from consideration for use in the
cathodes of lithium/sulfur cells.
Beryllia, the most corrosion-resistant insulator tested, showed variable
rates that were dependent upon the purity and method of fabrication. High-
purity (>99.9%), .hot-pressed beryllia showed the best corrosion resistance.
Thoria, aluminum nitride, yttria, Y3Al5Oi2, MgO ' A12C>3, LiA102, and boron
nitride also had corrosion rates less than 0.5 mm/yr. Boron nitride had a
relatively low corrosion rate, and it maintained its physical integrity,
but it was attacked by a mechanism that caused it to become electrically
conductive. This phenomenon has also been observed when boron nitride was
used in test cells.
A program to produce electrical insulators in the sizes and shapes
required for testing in operating cells has been initiated. Lithium aluminate
(LiA102), 90 wt % Y203-10 wt % Eu203, and Y203 bodies have been produced
successfully and will be tested under cell conditions.
electric Automobile Performance Calculations
Calculations of the performance characteristics that might be obtained
from an electric automobile powered by a lithium/sulfur battery have been
made. The calculations were based on a vehicle of 2086-kg (4600-lb) test
weight powered by a 488-kg (1075-lb) battery. The electrical performance
of the battery was assumed to be identical (per unit of electrode area) to
that obtained from experimental laboratory cells. In the battery design
allowances were made for the casings, thermal insulation and other items
that are peculiar to an operational battery which were not included in the
cell tests. The method of calculation differed from others of this kind in
that specific cognizance was taken of the fact that the deliverable energy
from the battery depends upon the power level of its operation. The results
indicate that the present performance of laboratory cells is sufficient to
permit a range of about 142 to 362 km (88 to 225 miles) for an electric
family automobile, depending upon the accessory load and the driving con-
ditions. These ranges are adequate to suggest that lithium/sulfur batteries
possess the potential for this application. The power levels of the cells
are satisfactory, but an effort should be made to increase the available
specific energy of the batteries in order to extend the range of the vehicle-
-------
- 6 -
I. INTRODUCTION
Research on lithium/chalcogen cells began at Argonne National Laboratory
in 1967 as a result of an energy conversion program directed toward the
development of thermally regenerative bimetallic galvanic cells for the
direct conversion of heat to electricity. Power densities over 3 W/cm2, cur-
rent densities over 1 A/cm2, and plateau voltages over 1.5 V were obtained
trom lithium/tellurium and lithium/selenium cells using a low-melting lithium
halide electrolyte. This performance indicated the potential utility of lith-
ium/chalcogen cells as secondary (electrically rechargeable) energy storage
devices capable of delivering high power and energy per unit weight of cell.
Lithium is well-suited for use as an anode because of its low weight per unit
of electricity delivered, low affinity for electrons, and high electrochemical
reactivity. The chalcogens (tellurium, selenium, and sulfur) are attractive
as cathode materials because they have high affinities for electrons, high
electrochemical reactivity, and reasonably low weight per unit of electricity
delivered.
The OAP program on lithium/sulfur cells began at ANL in February of
1969 with a modest experimental program on cells of the types Li/LiX/Li in
S and Li/LiX/Li in PxSy to determine if these cells could meet the require-
ments of a low-cost, high-specific-energy and high-specific-power battery
for electric automobile propulsion. The lithium/sulfur and lithium/PxSy
cells were selected for investigation because the cost and availability of
these materials, relative to the other chalcogens, made these systems the
most attractive for electric automobiles and similar mass-market applications.
Between February 1969 and June 1970 about 30 lithium/sulfur and
lithium/phosphorus sulfide (PXSV) laboratory cells were tested. These small-
scale (^1-cm electrode area) cells were operated at current densities from
0,3 to 1 A/cm2 to determine if capacity densities of about 0.6 A-hr/cm2 were
feasible. These current-density and capacity-density goals were estimates or
the requirements of a battery in an electric automobile under urban driving
conditions. The cells exhibited capacity densities from 0.1 to 0.2 A-hr/cm2
at a current density of 0.5 A/cm2. Sulfur utilization was usually less than
35%. Sulfur utilization is defined on the basis of the reaction
2 Li + S - Li2S
where 100% utilization indicates that all the sulfur has been converted to
Li2S. While 100% utilization is unobtainable in an operating cell, the
sulfur utilization figure is useful in comparing various cell designs and
cathode geometries. A utilization of 50-70% (above a 1-V cut off) is desir-
able for a high-specific-energy cell in which a high percentage of the cell
weight is devoted to reactants and a lesser percentage to housing, current
collector, insulators, and other electrochemically inactive components.
Cell performance during the first year of the program was limited
by two main factors, (1) the low diffusion rate of the cell reaction pro-
ducts away from the sulfur-electrolyte interface and (2) the low electronic
conductivity of sulfur. The most significant findings of the initial phase
of the program were (1) the lithium sulfur system appeared more promising
-------
- 7 -
for further development than either the Li/P^Sjo or Li/P^Ss systems, (2)
cell geometry and current-collector design greatly influenced cell per-
formance, and (3) the cells operated at acceptable current densities,
but capacity density, cycle life, and sulfur utilization required further
improvement.
Corrosion tests (^600 hr, 375°C) of potential materials of construction
for lithium/sulfur cells, including some representative austenitic and
ferritic stainless steels, manganese-chromium stainless steels, nickel-
base alloys and pure elements, indicated that only molybdenum, chromium,
and niobium showed corrosion rates sufficiently low for use in cells of
anticipated lifetimes greater than 1000 hr.
Experimental work has been carried out at a significantly higher
level since July, 1970, when the funding level was increased and several
tasks were added to the program. The major emphasis continues to be cell
and battery development, but the increased funding permits the experimental
program to undertake the following supporting studies necessary for further
cell development:
(1) The solubility of sulfur-bearing species in various elec-
trolytes and that of electrolytes in the cathode material
are being measured in order to select the best electrolyte,
cathode compositions,-and operating temperature of the cells.
(2) Studies of sulfur-bearing species in molten alkali halides
are being conducted to aid in understanding the chemical
processes occurring in an operating cell.
(3) Exploratory work was initiated on solid electrolytes which
may prove more suitable for some lithium/sulfur cell appli-
cation than the alternative liquid electrolyte.
(4) Although it appears that elemental sulfur alone can be used
satisfactorily as the active cathode material in lithium/
sulfur cells, the use of an additive to increase its elec-
tronic conductivity and/or decrease its vapor pressure is
being studied since it may increase the electrical performance
(specific power and specific energy) of the cell.
(5) Interfacial phenomena, as they relate to the wetting of
current collectors and the containment of lithium and
sulfur, and mass-transfer effects occurring within a cell
are being investigated. For example, in an effort to im-
prove the cycle life of the lithium anode in lithium/sulfur
cells, a study was undertaken on the transport characteristics
of lithium between a lithium electrode of the type used as
the anode in the laboratory cells and an aluminum-lithium
counter electrode. Various electrode designs, electrode
materials, and electrolytes were evaluated. Successful
operation of an experimental cell for over 200 cycles in-
dicated that under the proper conditions a long cycle life
can be achieved.
-------
- 8 -
(6) Materials evaluation has been performed by means of simple
screening tests for initial identification of promising
materials. Emphasis is now being directed toward the
successful utilization of aluminum, titanium, or chromium-
plated structures, which will probably be necessary for low-
cost, light-weight batteries,
(7) Electrical insulator tests have demonstrated the usefulness
of BeO, LiA102, and ^2°3 f°r cell application, but additional
work on alternative insulators and better methods of fab-
rication into cell components is necessary.
The goals for the cell program have been restated to provide additional
insight into the relationship between the performance achieved in labora-
tory cells and what might be expected of an automobile battery constructed
from similar cells. Table I presents both the battery and single-cell per-
formance goals. These single-cell goals were developed from an analysis of
the energy and power requirements of an automobile operated under urban
conditions (see Section V).
The present laboratory cell performance fails to meet the prescribed
goals, particularly in the areas of capacity per unit volume of cathode,
sulfur utilization, cycle life, and lifetime. Consequently, the principal
objective of the cell program during the past year has been to increase
the capacity density, sulfur utilization, and lifetime of the laboratory
cells. The principal variables investigated during the year are the
design of, and materials used for, the cell cathode structure.
TABLE I. Lithium/Sulfur Battery Development Program Goals
Assembled Battery
Specific Energy (at the 4-hr rate) 200 W-hr/kg
Specific Power (at the 1-hr rate) 200 W/kg
Cost3 $10/kW-hr
Lifetime 3-year minimum
Cycle Life 300-1000 cycles
Single Cell
Specific Energy 220 W-hr/kg
Specific Power 220 W/kg
Cost3 $9/kW-hr
Capacity Density (at 1-hr rate) 0.4 A-hr/cm2
Sulfur Utilization (above 1-V cut off) 70%
Capacity per Unit Volume of Cathode 1.0 A-hr/cm3
Cycle Life 1000 minimum
3Capital equipment cost of installed capacity; not direct energy cost.
-------
- 9 -
II. INVESTIGATIONS OF LITHIUM/SULFUR CELLS
(H. Shimotake, M. L. Kyle, F. J. Martino, R. Rubischko)
Liquid-electrolyte cells with lithium anodes, molten-halide salt
electrolytes, and sulfur-containing cathodes are being operated at
temperatures from 370 to 400°C to determine the cell configuration,
materials, and operating conditions necessary to achieve capacity densities
of about 0.4 A-hr/cm2 at current densities near 0.4 A/cm2. Other near-
term objectives are to increase the cycle life of the cells to about 1000
cycles at these performance levels and to increase the sulfur utilization
to the range of 50-70%. These goals for cell performance were chosen as
necessary to meet the performance requirements of a battery for an elec-
tric automobile having a range of 200 miles under practical driving con-
ditions (see Section V).
The attempts to increase the capacity density of the cells have center-
ed around the design of the cathode. Since diffusion of the product in the
cell cathode appeared to be a limiting feature of the initial cell design,
several designs which increased the interfacial cathode area were tested. These
cell designs are intended to increase the area over which product diffusion
can occur. A higher interfacial cathode area should increase cell performance
since more of the sulfur is available to the reaction sites at the sulfur-elec-
trolyte interface. This should result in an increase of the sulfur utiliza-
tion and a decrease in the diffusional overvoltage caused by the product
layer at the sulfur-electrolyte interface.
The other main cathode design problem investigated was how to decrease
sulfur loss from the cathode, which would increase the cycle life of the cells.
Earlier cells exhibited lifetimes of less than 48 hr because sulfur was lost
from the cathode during operation. Recent tests identified the mechanism of
sulfur loss, and the cell design was modified to substantially reduce these
losses and increase cell lifetime.
A. Experimental Procedure
Approximately 50 experimental cells of various designs were tested
during the last year (July 1970 - June 1971). A generalized version of the
cells is illustrated schematically in Fig. 1.
The main features common to all of the cells were an anode consisting
of a stainless steel Feltmetal* current collector soaked with liquid lithium,
a porous cathode current collector soaked with liquid sulfur or PitSjo* and
a molten-salt electrolyte containing lithium halide. The geometric areas of
the electrodes used in these cells were 0.7 to 2.6 cm2, and the interelec-
tfode distance was usually 1.0 cm. The operating temperatures were in the
range 370 to 400°C.
*A product of Huyck Metals Co.
-------
- 10 -
ELECTRICAL LEADS
INSULATING SPACER
CATHODE
CURRENT COLLECTOR
CONTAINING SULFUR
OR P4S,o
ANODE
Fig. 1.
STAINLESS STEEL
FELTMETAL CONTAINING
LITHIUM
MOLTEN SALT ELECTROLYTE
Schematic Diagram of a Typical
Unsealed Laboratory Cell
A typical cell assembly procedure was as follows. The anode current
collector was soaked in molten lithium (Foote Mineral Co., 99.98% purity)
at 550°C, the excess lithium was removed, and the electrode was cooled to
room temperature and placed in the anode holder (a stainless steel cup with
current and voltage leads attached). The cathode was prepared similarly,
soaking the cathode current collector in sulfur at a temperature near
160°C where sulfur has a low viscosity. The sulfur was N. F. sublimed grade
from the Mallinckrodt Chemical Works. The PI+SJQ (obtained from the Hooker
Chemical Co.) was purified by recrystallization from C$2. The electrolyte
was melted in an alumina or niobium crucible and the anode was immersed,
followed by the cathode. The cell was then ready for electrical performance
measurements. All of the above procedures, including operation of the cell,
were performed in a glovebox containing a high-purity (< 2 ppm each of 02
and N£, < 1 ppm H20) helium atmosphere. The compositions, melting points,
and densities of the electrolytes used in the cell tests are listed in
Table II; all of the electrolytes are eutectic compositions. The elec-
trolytes used in these tests were normally prepared from anhydrous materials
of purity greater than 99.5%. The eutectic mixture was premelted and mixed
at a temperature slightly above the melting point of the lowest-melting
constituent and contacted with molten lithium at about 50°C above the eutectic
melting point. The lithium contacting was performed to remove any residual
TABLE II. Physical Properties of Molten-Salt Electrolytes
Composition,
mol %
58.8 LiCl-41.2 KC1
59.0 LiBr-41.0 RbBr
11.7 LiF-29.1 LiCl-59.2 Lil
8.5 LiCl-59.0 LiI-32.5 KI
3.5 LiF-56.0 LiCl-40.5 KC1
Melting
Point,
°C
352
278
340.9
260
346
Density
at 380°C,
g/cm3
1.68
2.85a
2.77
2.92a
1.68
Reference
2
3
4
4
5
aEstimated.
-------
- 11 -
water and heavy metal ions. The lithium-contacted salt was filtered through
porous quartz frits to remove the L±20 and other solids in the electrolytes.
The cell reaction during discharge involves the electrochemical oxidation
of lithium at the anode
Li -> Li+ + e~
transport of lithium ion through the electrolyte, and its reaction with
sulfur at the cathode to form Li2S
2Li+ + S + 2e~ -»• Li2S
The maximum amount of electrical energy that can be delivered by the cell is
determined by the quantities of lithium and sulfur in the electrodes. An
excess of lithium was used in all of the cell tests, and the amount of sulfur
in the cathode was usually 0.3 to 0.6 g/cm2, which corresponds to a capacity
density of 0,5 to 1.0 A-hr/cm2, assuming Li2S to be the final reaction product.
The assembled cells were placed in a furnace well which was attached to
the floor of the helium-atmosphere glovebox. The furnace temperature was
held constant by a controller at about 380°C for most of the experiments.
Electrical connections to the cell were made by four wires, two for cell
voltage and the other two for cell current readings. The wires were fed
through the glovebox wall by means of a hermetically sealed connector to ex-
ternal metering equipment.
Usually, two types of electrical performance measurements were made-
Voltage-current density data were obtained from short-time measurements, in
which the cell voltage (at various current densities) was observed within a
few seconds after the circuit was closed. Voltage-time measurements, used
in determining the capacity densities (A-hr/cm2) of the cells, were performed
at constant current. The constant discharge and charge currents were obtained
with the aid of a regulated dc power supply.
In most cases discharge data were taken from the fully charged condition;
state of charge was determined by the open-circuit voltage of the cell. Charge
data were obtained from the partially or totally discharged cells. The cells
were usually discharged to a cut-off voltage of 1 volt, and left on open cir-
cuit for a prescribed period of time (about 10 minutes) before charge was begun.
After the cell was charged to a prescribed cut-off voltage, it was left on
open circuit for a prescribed period of time (about 10 minutes) which was then
followed by the next discharge-charge cycle.
In the cycle-life tests, the cell was usually connected to a meter re-
lay that automatically opened the circuit and reversed the polarity of the
power supply when the cell reached a pre-set cut-off voltage, thereby causing
the cell to be charged and discharged at prescribed current densities. Fre-
quent measurements were made of IR-free voltages by interrupting the current
for a short time.
The cell capacity was calculated from the discharge current and the
duration of the cell discharge. The charge accepted by the cell was
-------
- 12 -
similarly determined. The charge-storage capability of the cell is reported
in terms of the capacity density by dividing the cell capacity (A-hr) by the
projected cathode area. The cell capacity density can be expressed math-
ematically as follows:
q = / i dt
where
q = capacity density, A-hr/cm2
i = current density, A/cm2
t = duration of cell operation, hr
The average voltage of the cell (volts) during the specified discharge or
charge was determined from the following:
y = I V dt
where
V = cell voltage, volts.
The voltage-current density data were gathered by taking voltage
readings (with a digital voltmeter having a 50-msec response time) at a
predetermined current within a few seconds after closing the circuit.
Because of the rapidity of the measurement there was a negligible over-
voltage due to diffusion of reactants or products. The internal resistance
of the cell was calculated from the following equation:
A di
where
R = cell internal resistance, ohm
i = current density, A/cm2
A = projected cathode area, cm2
The cells were disassembled after testing to relate the condition of
the cell to its electrical performance characteristics, to establish causes
of failure, and to identify corrosion problems. Frequently samples of
cathode products, electrolyte, structural materials, anode materials, and
insulators were taken and examined by means of X-ray diffraction, micro-
scopy and wet-chemical analysis. A mass balance was also made for some
cells for the materials contained in the cells. Photographic techniques
were used to record the appearances of cell cross sections and cell parts
as well as the cell assembly.
-------
- 13 -
B. Results and Discussion
The principal variables investigated in the cell tests were the cathode
current collector design and the materials used for its construction. An
earlier series of experiments6 had been conducted on porous metal structures
for use as the cathode current collector. The materials that were tested
in those investigations included stainless steel Feltmetal and Fibermetal*
and chromium foam.'*' Although considerably better short-term electrical
performance was obtained with the Feltmetal and Fibermetal than with the
chromium foam, stainless steel is not a suitable cathode current collector
material because of its poor corrosion resistance to lithium-sulfur mixtures,
Therefore two additional materials, molybdenum foam§ and porous graphite or
carbon, were investigated for this application. Graphite in particular has
the desirable characteristics of corrosion resistance, electrical conduc-
tivity, light weight, low cost, and durability.
Several different cathode designs that were used in the experimental
cells are illustrated in Fig. 2. The objective of the laminated design
(Fig. 2a) was to provide an extended interfacial area between the sulfur
(o) LAMINATED CATHODE
(d) RESERVOIR CATHODE
( CROSS SECTION )
(b) COMB CATHODE
(e) ENCLOSED RESERVOIR
CATHODE
( CROSS SECTION )
(e) ENCLOSED LAMINATED CATHODE
( EXPLODED VIEW )
Fig. 2. Schematic Diagrams of Some Cathode
Current Collector Designs
*A product of Brunswick Corp.
TA product of Astro Met Associates, Inc.
^A product of Spectra-Mat, Inc.
-------
- 14 -
and the electrolyte. In the first design, this cathode consisted of
alternating layers of stainless steel Feltmetal filled with sulfur and
electrolyte, respectively. The later designs employed alternating layers
of molybdenum foam, which is wet preferentially by the electrolyte, and
porous graphite or carbon, which is wet preferentially by the sulfur. The
comb design (Fig. 2b) embodies the same concept, but in this case the
electrolyte-containing current-collector layers were eliminated in an attempt
to decrease the cathode weight and to allow better access of the electrolyte
to the sulfur at the surface of the graphite current collector.
The laminated and comb designs both resulted in reasonably good initial
electrical performance, but they lost their capacity rapidly. The loss in
capacity was attributed to the escape of sulfur from the cathode structure,
which was confirmed in a special series of experiments. These experiments
showed that sulfur extrudes from porous carbon and disperses into the molten-
salt electrolyte, which becomes a yellow color. Upon freezing, the salt was
an orange to yellow color, and when it was heated to 600°C the color gradually
disappeared, presumably because of the vaporization of sulfur. The loss in
capacity of the laminated and comb cathodes was therefore believed to result
from the escape of sulfur from the current collector into the electrolyte
and subsequent vaporization of the sulfur.
In order to minimize the migration of sulfur into the electrolyte,
the enclosed laminated cathode in Fig. 2c was tested. If the sulfur entered
the electrolyte as a dispersion rather than as a true solution, the process
might be inhibited by surrounding the cathode by a porous layer that is wet
preferentially by the electrolyte. This approach was shown to be success-
ful first in a series of specific experiments for this purpose and later
in enclosed laminated cathode cells which exhibited greatly extended life-
times and showed little evidence of sulfur loss.
The reservoir cathode, shown in Fig. 2d, consisted of a thin, porous
current collector that confined liquid sulfur within a reservoir. As sulfur
was consumed by the cell reaction at the current co.1.lector-electrolyte inter-
face, additional sulfur was wicked through the current collector to the
site of the reaction. The purpose of this design was to minimize the
volume and weight of the current collector. However, significant sulfur
losses occurred, probably by the same mechanism as that in the laminated
and comb designs (Figs. 2a and 2b). This problem was overcome in the
same way, by interposing a porous molybdenum layer, which is wet pre-
ferentially by the electrolyte (Fig. 2e).
An attempt to increase the capacity density of the lithium/sulfur cell
by intimately mixing the sulfur in the cathode with conductive materials
such as carbon powder or metal fibers and electrolyte has been initiated.
The major potential advantage of this design, called the "mixed-cathode"
design, is that it is intended to increase the electrochemical reaction
area, thereby increasing the cell capacity. A schematic drawing of the
mixed cathode is shown in Fig. 3. The experimental cell housing was
borosilicate glass tubing for ease of construction, but metal housings
could also be
-------
- 15 -
CURRENT TERMINAL
BOROSILICATE HOUSING
MIXED SULFUR,
ELECTROLYTE,
CARBON POWDER
AND NIOBIUM
MOLYBDENUM FOAM
ELECTROLYTE
Fig. 3. Schematic Diagram
of a Mixed Cathode
Detailed construction and performance data for all the small-scale
cells operated during the last year are listed in the appendix. The results
of these tests will be discussed by groups of cells which have common
characteristics.
Lithium/sulfur cells were tested which utilized the laminated cathode
configuration shown in Fig. 2a in which the laminae were 80 to 85% porous
Type 302 or 304 stainless steel, with mean pore sizes in the range 25 to
40 um. Alternate layers were filled with electrolyte and with sulfur.
The laminae were 11 mm x 11 mm of various thicknesses in the range 0.07 to
0.16 cm and were mounted in a stainless steel housing. The cathode area
used in current-density and capacity-density calculations is defined as the
area of the cathode projected upon the anode, including both sulfur and
electrolyte elements. The thickness of the elements was varied in these
tests since it was expected that thin sulfur elements would produce improved
sulfur utilization as the distance over which diffusion must proceed is
diminished. The electrolyte elements were considered to serve only as a
means of providing a continuous phase for lithium-ion transport to the
sulfur-electrolyte interface. Voltage-current density and voltage-capacity
density data that are considered to be typical for these cells are shown
in Figs. 4 and 5. The best performance achieved with a laminated cathode
configuration was 0.52 A-hr/cm2 at 0.52 A/cm2 with a sulfur utilization of
25%, assuming that 100% utilization corresponds to conversion of all of the
sulfur to Li2S. Several difficulties with this type of cathode were en-
countered, the most severe being that the capacity density decreased markedly
-------
- 16 -
after 2 or 3 cycles. Examination of the cathode components after testing
indicated that considerable corrosion of the stainless steel elements had
occurred. The significant decrease in capacity density that took place
with cycling was attributed to the formation of corrosion products and
the loss of sulfur from the cathode structure by diffusion or dispersion
into the electrolyte. It was noted that thinner sulfur-bearing elements
permitted greater utilization of the available sulfur, but the corrosion
and sulfur vaporization problems still limited the cell life to a few cycles.
Since corrosion of the cathode current collector appeared to be a
severe problem, an alternate current-collector material was desired. It
was found that porous graphite and carbon were readily wet by sulfur and
were corrosion resistant to the lithium/sulfur cell environment. These
structures were not wet by the majority of molten-salt electrolytes used
in the cell tests. These characteristics made it possible to create a
laminated cathode in which the containment of the sulfur could be greatly
improved and the formation of corrosion products reduced. A series of cells
were fabricated using porous graphite (Poco Graphite, Inc., Grade AX Fuel
Cell Carbon, 64% porosity, 1.4-pm pore size) as sulfur-bearing elements and
porous stainless steel (Huyck Metals Corp., Type 302 stainless steel, 80%
porosity, 30-pm pore size) as electrolyte-bearing elements. These elements
4.0
3.5
3.0
Li/LiBr-RbBr/Li in S
ANODE AREA 2 7 cm*
CATHODE AREA I 58 cm'
INTERELECTRODE DISTANCE I cm
TEMPERATURE 395 "C
CATHODE CURRENT COLLECTOR
9 ELEMENTS
TYPE JO2-SS 80 % POROSITY 50 fj. PORE SIZE
THEORETICAL CAPACITY DENSITY i
I 25 A-hr/cmJ j
SHORT-TIME DATA J
CURRENT DENSITY, A/cmz
Fig. 4. Typical Voltage-Current Density
Curve for a Lithium/Sulfur Cell
With a Laminated Cathode Having
All-Stainless-Steel Laminae
-------
- 17 -
3.5
3.0
> 25
o
>
1.5
1.0
0.5
Li/LiBr-RbBf/Li in S
ANODE AREA 27 cm»
CATHODE AREA 158 cm'
INTERELECTROOE DISTANCE I cm
TEMPERATURE 395'C
CATHODE CURRENT COLLECTOR
9 ELEMENTS
TYPE 302-SS 80% POROSITY JO M PORE SIZE
THEORETICAL CAPACITY DENSITY
I 25 A-hi/cm*
-0.27 A/cm1 CHARGE 2
0.25 A/cm' CHARGE I
032 A/cm' DISCHARGE I
-0.63 A/cm' DISCHARGE 2
-0.33 A/cm' DISCHARGE 3
0.10 0.20 030
CAPACITY DENSITY, A-hr/cm1
0 4 6 12 16 20 24
PERCENT OF THEORETICAL CAPACITY DENSITY
Fig. 5. Typical Voltage-Capacity Density
Curve for a Lithium/Sulfur Cell
With a Laminated Cathode Having
All-Stainless-Steel Laminae
were contained in a stainless steel housing. The capacity density and cycle
life of test cells containing such cathode structures were better than those
for cells with all-stainless-steel cathode current collectors. Voltage-
capacity density data of the best cell of this series are shown in Fig. 6.
On the sixth discharge at 0.25 A/cm2, a capacity density of 0.34 A-hr/cm2
with 45% sulfur utilization was achieved above a 1.0-V cut off. Operation
of this cell was voluntarily discontinued after the sixth cycle and the cell
was disassembled. As in other cells of this type, excessive corrosion of
the stainless steel electrolyte elements was observed.
The next cathode design did not contain any stainless steel Feltmetal.
A porous-graphite comb structure as shown in Fig. 2b was tested. The first
discharge yielded 0.40 A-hr/cm2 at a current density of 0.53 A/cm2 with 28%
sulfur utilization above a 1.0-V cut off. However, subsequent discharges
yielded much lower capacity densities as shown in Fig. 7. It is noteworthy
that the cell voltage during discharge was significantly higher than in
previous cells due to the lower internal resistance of this cell. It appeared
that the absence of the electrolyte-containing elements permitted more rapid
transport of sulfur from the porous carbon into the bulk electrolyte. A
second cell of identical configuration was constructed and P^SIO was
-------
- 18 -
substituted for sulfur. Some improvement in performance was noticed, and
the rate of deterioration with cycling was reduced, but the conclusions re-
mained unchanged.
The next series of laminated-cathode cells tested was designed to
eliminate the corrosion problems associated with the use of stainless steel,
The cathode design was the laminated configuration except that in this
series of cells the electrolyte-bearing elements were molybdenum foam
(Spectra-Mat, Inc., ,5% porosity, 25-um pore size). The graphite elements
were filled with P^SKJ and all the elements were contained in a niobium
housing rather than the stainless steel housing used previously. Al-
though the capacity density of the&a cells was somewhat lower than that of
the earlier cells, the cycle life was substantially increased. The first
cell operated for 17 cycles (three days) and was frozen and restarted twice
without difficulty. Another cell operated continuously for 35 cycles. The
capacity density, however, decreased during operation.
2.4
2.2
2.0
1.8
^ 1.6
ui
j» ,4
O
> 1.2
_i
ui 1.0
o
0.8
0.6
0.4
0.2
T
T
Li/LiBr-RbBr/Li in S
ANODE AREA 2.92 cm2
CATHODE AREA 1.33 cm*
INTERELECTRODE DISTANCE I cm
TEMPERATURE 390-400°C
CATHODE CURRENT COLLECTOR
4 SULFUR ELEMENTS
GRAPHITE 64 % POROSITY 1.4
5 ELECTROLYTE ELEMENTS
TYPE 302-SS 80 % POROSITY 30 ^ PORE SIZE
THEORETICAL CAPACITY DENSITY 0.76 A-hr/cm*
0.25 A/cm2 DISCHARGE 6
PORE SIZE
0.5 A/cm2 DISCHARGE 4
-0.5 A/cm2 DISCHARGE 3
A/cm2 DISCHARGE I
1.0 A/cm2 DISCHARGE 5
I
I
I
0.05
0.10
0.15 0.20 0.25 0.30
CAPACITY DENSITY, A-hr/cm2
0.35
0.40
0.45
I
I
I
10 15 20 25 30 35 40 45
PERCENT OF THEORETICAL CAPACITY DENSITY
50
55
60
Fig. 6. Voltage-Capacity Density Curves for a Lithium/Sulfur Cell
With a Laminated Cathode Having Porous Graphite and Stain-
less Steel Feltmetal Laminae
-------
- 19 -
UJ
o
o
UJ
o
0.53 A/cm* DISCHARGE 3
0.53 A/cm2 DISCHARGE 2
0.53 A/cm2 DISCHARGE I
1.0
0.5
^0.53 A/cm2 DISCHARGE 4
I I
Li/LiBr-RbBr/Li in S
ANODE AREA 2.92 cm*
CATHODE AREA 1.89 cm2
INTERELECTROOE DISTANCE I cm
TEMPERATURE 395°C
CATHODE CURRENT COLLECTOR
GRAPHITE 63 % POROSITY 1.4 /i PORE SIZE
THEORETICAL CAPACITY DENSITY 1.45 A-hr/cm2
O.I
0.2 0.3
CAPACITY DENSITY, A-hr/cm2
j I I I I
0.4
I
I
10 20
PERCENT OF THEORETICAL CAPACITY DENSITY
30
Voltage-Capacity Density Curves for a
Lithium/ Sulfur Cell With a Comb-Type Cathode
Fig. 7.
In order to prevent the escape of sulfur (or P^Sjo)* a cathode current
collector of an enclosed laminated configuration was designed and fabricated
(see Fig. 2c). In this configuration the sulfur -bearing, porous-graphite
elements were completely surrounded by porous-molybdenum elements filled with
electrolyte. This design reduced the sulfur loss into the electrolyte, with
the result that the cell operated for over 800 cycles. The best capacity
density for this cell was 0.16 A-hr/cm2 (above a cut-off voltage of 1.0 V),
at a current density of 0.20 A/ cm2 with 49% sulfur utilization. Selected
voltage-capacity density data for this cell are shown in Fig. 8. The
trend of sulfur utilization was downwards, averaging 30% over the first 10
cycles, 25% in the next 60 cycles, and about 15% over the following 100 cycles.
The capacity density-cycle number performance of the cell is presented in
Fig. 9. The decrease in capacity density was attributed to the fact that
lithium was being lost from the anode current collector. The cell capacity
was recovered by replacing the lithium anode.
In another series of experiments, the concept of using a sulfur res-
ervoir in the cathode was evaluated. In the reservoir configuration (see
Fig. 2d) the cathode is a pool of sulfur which is contained by a thin porous-
graphite housing. The purpose of this design is to increase the capacity per
unit volume of cathode (A-hr/cm3). During discharge, sulfur should wick
-------
- 20 -
through the graphite current collector to the graphite-sulfur-electrolyte
interface at a rate sufficient to sustain the cell reaction. If the reaction
product, Li2S, diffuses back into the sulfur pool at an adequate rate, this
type of cell design would be very attractive and should have a high specific
energy. The voltage-capacity density curves for the first reservoir cell
are shown in Fig. 10. It is noteworthy that the capacity density, 0.28
A-hr/cm2, achieved in the first cycle required that sulfur from the res-
ervoir be utilized. However, the capacity decreased to 0.05 A-hr/cm2 by
the 5th cycle and wh^n the cathode was disassembled the reservoir was found
to be filled with electrolyte; no sulfur was present.
Based upon the experience gaiaed with the cell just discussed, two
design modifications were made: (1) Grafoil seals were used to contain
the sulfur and to prevent leakage of electrolyte into the reservoir, and
(2) a porous-metal element filled with electrolyte was placed in front of
the porous-graphite current collector. (See Fig. 2e.) One such cell was
operated for 803 cycles (1100 hr). The 193rd discharge yielded 0.55
3.5
3.0
2.5
2.0
bJ
O
5 '-5
O
=: 1.0
0.5
I i I
I I 1 T
i i i
i r
0.18 A/cmz CHARGE - 34
-0.2 A/cm2CHARGE -17
Li/LiCI-Lil-KI /Li in S
ANODE AREA 2.5 cm*
CATHODE AREA 2.53 cm2
INTERELECTRODE DISTANCE Icm
TEMPERATURE 380'C
CATHODE CURRENT COLLECTOR
3 GRAPHITE ELEMENTS
63 % POROSITY lAp. PORE SIZE
MOLYBDENUM FOAM CASING
THEORETICAL CAPACITY DENSITY 0335 A-hr/cm2
0.2 A/cm DISCHARGE-34
0.2 A/cm* DISCHARGE- 17
I I I I
I I I I
I I I I
0 0.05 O.I 0.15
CAPACITY DENSITY, A-hr/cm2
I l l i l I I l I l I l l l l I I l I I I I I l l
j i
10 20 30 40
PERCENT OF THEORETICAL CAPACITY DENSITY
50
Fig. 8. Voltage-Capacity Density Curves for a
Lithium/Sulfur Cell With an Enclosed
Laminated Cathode
Product of the Union Carbide Corp.
-------
0.15
t 0.10
z
u
a
o
s
<0.05
- 21 -
Li/LiCI-Lil-KI/Li inS
ANODE AREA 2.5cm2
CATHODE AREA 2.53 cm2
INTERELECTRODE DISTANCE I cm
TEMPERATURE 380" C
CATHODE CURRENT COLLECTOR
3 GRAPHITE ELEMENTS
63% POROSITY, l.4um PORE SIZE
MOLYBDENUM FOAM CASING
THEORETICAL CAR4CITY DENSITY 0.335 A-hr/cm2
I
CYCLE NUMBER
20
40
60
80
100
120
140
160 200
600
75
170
240
340
530
1100
TIME, hr
Fig. 9. Capacity Density as a Function of Cycle Number for a
Lithium/Sulfur Cell With an Enclosed Laminated Cathode
o
> 1.4
u
o
1.2
1.0
0.8
0.6
0.4
"0.15 A/cm* CHARGE 1
Li/LiCL-KCL/Li in S
ANODE AREA 2.95 cm'
CATHODE AREA I cm1
INTERELECTROOE DISTANCE I em
TEMPERATURE 390 - 400*C
CATHODE CURRENT COLLECTOR
GRAPHITE DISK
63% POROSITY 1.4 p. PORE SIZE
-0.3 A/cm2 DISCHARGE 1_
I
I
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
CAPACITY DENSITY, A-hr/cm'
| | | | |
6.77 13.5 21.7 27.0
PERCENT OF THEORETICAL CAPACITY DENSITY
Fig. 10. Voltage-Capacity Density Curves for a
Lithium/Sulfur Cell With a Reservoir Cathode
-------
- 22 -
A-hr/cm2 at 0.165 A/cm2 with 27% sulfur utilization; the 207th cycle yielded
0.52 A-hr/cm2 at 0.27 A/cm2 with 26% sulfur utilization; and the 209th cycle
yielded 0.15 A-hr/cm2 at 0.48 A/cm2 with 8% sulfur utilization. The voltage-
capacity density data for this cell are shown in Fig. 11.
o
o
_)
_1
UJ
CJ
CHARGE
Li/LiCI-LiI- KI/Li in S
CATHODE AREA I cm*
INTERELECTROOE DISTANCE ~ 2 cm
TEMPERATURE 380° C
CATHODE CURRENT COLLECTOR
CARBON DISK
49% POROSITY 3.5 ji PORE SIZE
MOLYBDENUM FOAM DISK
78 % POROSITY 25 fj. PORE SIZE
0.165 A/cm' CYCLE NO. 193
THEORETICAL CAPACITY DENSITY
2.04 A-hr/cm*
0.165 A/cm* CYCLE NO. 193
0.267 A/cm' CYCLE NO. 20
0.48 A/cm* CYCLE NO. 209
DISCHARGE
O.I
02 0.3 0.4
CAPACITY DENSITY, A-hr/cm2
0.5
5 10 15 20
PERCENT OF THEORETICAL CAPACITY
25
Fig. 11. Voltage-Capacity Density Curves for a
Lithium/Sulfur Cell With an Enclosed
Reservoir Cathode
The present reservoir-cathode design appears to increase the capacity
density of the cell but does so at a reduced capacity per unit volume and
a reduced sulfur utilization. Alternate design variations, especially with
a larger-pore-size-graphite reservoir to promote product diffusion, will be
attempted. This design would be even more attractive if cathode additives
which increase the rate of product diffusion can be identified (see Section
III. D).
A series of experiments to test the "mixed-cathode" design was con-
ducted. The first cell contained a cathode composed of a mixture of 1-10
urn dia niobium fibers, carbon powder (Vulcan XC-72R Carbon Black),
LiF-LiCl-KCl, and sulfur. The niobium fibers and LiF-LiCl-KCl were premixed
by adding niobium fibers, which were obtained from a sheet of niobium felt
(Brunswick Co., 98% porosity, 12-35 pm pore size), and LiF-LiCl-KCl to a
Petri dish at AOO°C. The resulting mixture was ground into a fine powder
-------
- 23 -
(estimated at about 75 um dia). This mixture was then added to a mixture of
sulfur and carbon powder which had been prepared separately. The weight and
volume percentages of each component are listed in Table III. The cathode
housing was loaded with the mixture (0.4 vol % Nb - 5,1 vol % C - 48.:> vol 7.
electrolyte - 46.0 vol % S) and sealed by a disk made of molybdenum foam
(Spectra-Mat, Inc., 78.2% porosity, 67-um pore size, 0.15 cm thick).
TABLE III. Physical Characteristics and Electrical
Performance of Two Mixed-Cathode Cells
Cathode
Metal
Weight, g
Vol %
Carbon
Weight, g
Vol %
Electrolyte
Weight, g
Vol %
Sulfur
Weight, g
Vol %
Anode
Lithium, g
Cathode Area, cm2
Theoretical Capacity Density, A-hr/cm2
Capacity Density,3 A-hr/cm2 (> 1-V cut off)
Capacity per Unit Vol. Cathode, A-hr/cm3
Current Density, A/cm2
Energy Density, W-hr/cm2 (> 1-V cut off)
Percent of Theoretical Capacity
Average Voltage, V
Number of Cycles
Short-Term Cell Resistivity, ohm-cm
Cell
67
Nb fibers
0.226
0.40
0.51
5.1
3.95
48.5
4.59
46.0
2.11
2.55
3.0
0.218
0.114
0.1
0.407
8.1
1.78
82
0.5
No.
68
Nb Mesh
2.02
4.4
0.8
10.0
2.69
43.5
3.2
42.1
1.85
2.55
2.1
0.257
0.172
0.1
0.41
12.2
1.58
3
0.5
aBased on volume of sulfur charge, electrolyte, and current-collect-
ing material (excluding housing and molybdenum sheath).
The electrical performance of the cell during 70 cycles and 230 hr of
operation is shown in Fig. 12. During most of the cycles the capacity density
was about 0.2 A-hr/cm2 with a cut-off voltage of 1 V at current densities
ranging from 0.1 to 0.3 A/cm2. The low capacity densities during the first
24 cycles were due to a high resistance in the current terminal. When a new
current terminal was installed, the performance improved.
The cut-off voltage for charge was gradually raised during operation.
The initial cut-off voltage, 2.67 V, was raised to 2.75 V at the 47th charge
to 2.95 V at the 53rd, and to 3.10 V at the 58th charge. The effect on the
-------
- 24 -
cell performance is shown in Fig. 13. A significant energy gain at voltages
above 2 V in the 59th discharge is noticeable. It is suspected that this
energy gain was due to the possible formation of S2C12 at the sulfur electrode
above 2.5 V.
0.5
0.4
i
<
0.3
i
UJ
Q
i °2
I
o
O.I
CATHODE MIXTURE (VOL%)
04 Nb
5.1 C
48.5 ELECTROLYTE
46.0 SULFUR
CELL NO. LIS-67
LI/LIF-LICI-KCI/LI IN S
ANODE AREA 6.4cm2
CATHODE AREA 2.55cm2
INTERELECTROOE DISTANCE ~2 cm
TEMPERATURE 380'C
THEORETICAL CAPACITY DENSITY 3.0 A-hr/cm
ANODE CURRENT COLLECTOR
TYPE 302 STAINLESS STEEL FELTMETAL
90% POROSITY, BT-fLin AV PORE SIZE
10
20
30
40 50
CYCLE NUMBER
60
70
80
90
Fig. 12. Capacity Density as a Function of Cycle Number for
a Lithium/Sulfur Cell With a Mixed Cathode
01 A/cm
DISCHARGE 25. 32.59
0.3-
CELL NO. LIS-67
LI/LIF-LICI-KCI/LI IN
ANODE AREA 64cm2
CATHODE AREA 2.55cm'
INTERELECTROOC DISTANCE 2cm
TEMPERATURE 380*C /
THEORETICAL CAPACITY DENSITY 30 A-hr/cmz
ANODE CURRENT COLLECTOR
TYPE 302 SS FELTMETAL
90% POROSITY, 67>im «/. PORE SIZE
CATHODE MIXTURE (VOL %)
04 Nb
3.1 C
48 5 ELECTROLYTE
46 0 SULFUR
I L
0050 0100 0150 Q 200 0.250 0.300
CAPACITY DENSITY. A-hr/cm2
0350
0.400
0.450
Fig. 13. Voltage-Capacity Density Curves for a
Lithium/Sulfur Cell With a Mixed Cathode
-------
- 25 -
A modification to the above cell was made by adding more carbon powder
and using a niobium expanded-mesh coil in place of the niobium fibers. A
sheet of niobium expanded mesh (Die Mesh Corp., 0.23-cm die size) was wound
in a coil form, 1.8 cm dia. x 2.2 cm high. Alumina cloth strips (Zircar,
Type ALK-15, 0.038 cm thick) presoaked with LiF-LiCl-KCl were inserted in
the gaps of the niobium coil. The purpose of the alumina cloth was to in-
sert a desired amount of the salt at desired locations in the cathode
assembly. The cathode housing was loaded with the niobium coil and a pre-
mixed powder of sulfur and carbon and sealed with a molybdenum foam disk
(Spectra-Mat, Inc., 78.2% porosity, ^67-um pore size, 0.15 cm thick). The
composition of the mixture was 4.4 vol % Nb - 10.0 vol % C - 43.5 vol %
electrolyte - 42.1 vol % sulfur. The electrical performance of the cell is
shown in Fig. 14. Three cycles having capacity densities ranging from
0.44 to 0.26 A-hr/cm2 with a cut-off voltage of 1 V at current densities from
0.1 to 0.4 A/cm2 were obtained before indications of a malfunction were noted.
An examination showed that the heater had burned out. Operation of the cell
was terminted at this point.
3.0
3.5
2.0
ui
(9
1.5
ui
o
1.0
CELL NO. LIS-68
LI/LIF-LICI-KCI/LI IN S
ANODE AREA 6.4cm2
0.5
DISCHARGE 3. O.IA/c
I
CATHODE MIXTURE(VOL%)
4.4 Nb
10.0 C
43.5 ELECTROLYTE
42.1 SULFUR
I
CATHODE AREA 2.55cm'
INTERELECTRODE DISTANCE 2cm
TEMPERATURE 380*C
THEORETICAL CAPACITY DENSITY 2.1 A-hr/cm
ANODE CURRENT COLLECTOR
TYPE 302 SS FELTMETAL
90% POROSITY, 67>im AV. PORE SIZE
J_
I
0.100
0.200
0300
0.400
0500
CAPACITY DENSITY, A-hr/cm'
Fig. 14. Voltage-Capacity Density Curves for a Lithium/Sulfur
Cell With a Mixed Cathode Containing a Niobium Expanded-
Mesh Spiral Current Collector
From the experiments that were made with the mixed-cathode design, it
appears that (1) the cell yields a consistent capacity density without loss;
however, (2) more study is needed to achieve higher utilization of the sulfur
in the cathode.
A sealed lithium/sulfur cell with an active electrode area of 5 cm2 has
been designed, fabricated, and operated three times. Figure 15 shows the
construction of the cell. This cell was operated to determine if a lithium/
sulfur cell could be operated in a "cathode-down" position, since most cells
-------
- 26 -
have been operated "cathode-up" and reversal of the electrode positions might
simplify cell design. Another objective was to identify problems associated
with the design and operation of cells utilizing Grafoil as a sealing material,
The performance characteristics of these cells are listed in Table IV. It
appears that a lithium/sulfur cell may be operated with the anode uppermost.
It also appears that a gasket seal of Grafoil that is effective in containing
the molten-salt electrolyte can be designed. Post-test examination of the
first two cells revealed that the anode had not been reabsorbing the lithium
during cell recharge; in both cases beads of lithium were found in the salt.
This phenomenon is under investigation (see Section III. E). In the third
cell a layer of frozen electrolyte was formed over the molybdenum foam be-
fore the sulfur was added to the porous graphite. This technique should help
to contain the sulfur during cell start up (i. e., before the salt becomes
molten). Free sulfur could create a Li2S film on the anode surface, and
this film could inhibit lithium retention by the anode. A significant
decrease in lithium beading in the third cell was observed.
C. Conclusion
In summary, lithium/sulfur cells that have been operated at ANL have
shown performance characteristics commensurate with electric automobile
requirements (capacity densities of about 0.4 A-hr/cm2 at a current density
of 0.4 A/cm2) for several cycles and have also demonstrated lifetimes of
over 800 cycles and 1100 hr at reduced capacity densities. No cell, how-
ever, has been capable of meeting both requirements simultaneously. Capacity
per unit volume, sulfur utilization, and the retention of high performance
MOLYBDENUM PLATEX
0.88 In ID, 1.500,
0.020 In. THICK
THREADED MOLYBDENUM
ROD AND NUTS 5-44
MOLYBDENUM CELL ENDS'
GRAFOIL GASKET
0.88 in ID, 1.5 In. 00,,
0.010 In. THICK
INCONEL SPRINGS
In. 00, I In. LONG
ALUMINUM PLATE
3 In DIAMETER
MOLYBDENUM FOAM
85% POROSITY,
0.991 In. DIAMETER, 0.082 In. THICK,
WT. 2 271 g
STEEL WASHERS
SS SPACER
LAVITE INSULATOR
THREADED STEEL ROD
10-32, 41 In. LONG
GRAFOIL GASKET
1.75 In. ID, 1.9 In. OD, 0.015 In. THICK
QUARTZ
45mmOD. 1.5mm WALL
HUYCK TYPE 302 SS
90% POROSITY, 0.990 In. DIAMETER.
0.255ln. THICK WITH O.I25ln. HOLE,
WT. 2.594g; 4.885g WITH
ROD AND NUTS
STEEL WASHER
STEEL HEX NUTS 10-32
POROUS GRAPHITE, GRADE FPA-20.
91% POROSITY, 0.991 In. DIAMETER.
0.122 In. THICK, WT. 0.440 g
Fig. 15. Schematic Drawing of a Sealed Lithium/Sulfur Cell
-------
TABLE IV. Construction and Performance Characteristics
of a Sealed Lithium/Sulfur Cell
Cell Design
Anode Area, 5.0 cm2; Cathode Area, 5.0 cm2
Interelectrode Distance ^1 cm
Temperature 380°C
Anode Current Collector:
Type 302 SS Feltmetal, 2.51 cm dia x 0.65 cm thick, 90% porosity, 67-pro pore size
Cathode Current Collector:
Porous graphite, 2.51 cm dia x 0.31 cm thick, 91% porosity, 50-ym pore size
Molybdenum foam sheath, 2.51 cm dia x 0.21 cm thick, 75% porosity, 25-pm pore size
Theoretical
Capacity
v y Open-
Cell A-hr/ A-hr/ Circuit
No. cm2 cm3 Voltage
1 0.478 0.956 2.35
-
-
2.35
2.38
2.35
2 0.501 1.00 2.30
2.29
2.34
2.27
2.18
2.13
3 0.541 1.08 2.40
2.31
2.35
2.35
2.37
_
_
Cell Performance
Cycle
No.
1
2
4
6
8
9
1
2
3
5
9
11
1
2
10
15
20
40
75
Discharge
Current
Density,
A/ cm2
0.25
0.25
0.20
0.20
0.20
0.20
0.20
0.20
0.10
0.20
0.20
0.20
0.20
0.20
0.20
0.20
0.20
0.20
0.20
Capacity
Density,
A-hr/cm2
0.136
0.079
0.095
0.101
0.090
0.080
0.109
0.074
0.102
0.068
0.066
0.032
0.136
0.071
0.089
0.105
0.061
0.013
0.0046
% of
Theo-
retical
Capacity
29.8
16.6
19.8
21.2
18.8
16.7
21.8
14.8
20.3
13.6
13.2
6.4
25.1
13.1
16.5
19.4
11.2
2.4
0.9
Average
Discharge
Voltage
1.6
1.75
1.8
1.75
1.75
1.7
_
-
1.85
1.9
1.90
1.77
1.87
1.85
1.80
Total
No. of
Cycles Comments
11 Control cir-
cuit malfunc-
tion caused
cell to over-
charge to
^4. 8 V
11 Lithium
beading
caused
short
circuit
108 Electrolyte
and sulfur
leakage
-------
- 28 -
levels for extended cycle life require further improvement. The performance
of the various cell designs is summarized in Table V, and the Appendix con-
tains complete information on the cells operated during the year. The
operation of long-life, high-performance cells remains the primary objective
of the laboratory cell program. The cells operated to this time have been
relatively small (<30 cm2) and unsealed. Initial scale-up efforts are
necessary to provide information on problem areas peculiar to the sealing,
design, construction, and operation of larger cells. Finally, the assembly
of cells into batteries will present new problem areas of cell matching,
safety, reliability, long-term hermetic sealing, and temperature control.
The program until now has been a laboratory effort to demonstrate the
technical feasibility of lithium/sulfur cells. A great deal of research and
development remains to be done to reach the stage of a reliable battery.
-------
TABLE V. Performance of Some Typical Lithium/Sulfur Cells
Cell
No.
30
37
44
45
50
55
56
fQ
Do
Electrolyte '
LiBr-RbBr
LiBr-RbBr
LiBr-RbBr
LiBr-RbBr
LiCl-KCl
LiCl-Lil-KI
LiCl-Lil-KI
T 4 f* 1 Vf*1 T 4 T?
L 1 U i. — tx.L» 1 •" L 1 r
Temp.,
°C
395
390
395
395
390
380
380
Structure
Type
Laminated
Laminated
Laminated
Comb
Reservoir
Enclosed
Laminated
Enclosed
Reservoir
xe
Depth,
cm
1.1
1.1
1.1
1.3
0.56
2.1
1.3
1 A
J. * *T
Materials3
& Thickness,
B,0.71;A,1.6
A,1.6;C,0.45
E,0.91;A,1.6
E , 1 . 3 ; none
E.0.63
E,1.5;G,3
F,3.0;G,4.0
G i 5
— — — ,800
>600
3
Life,
hr
6
6
26
24
8
>1100 to
I
>700
42
aThe materials are given first for the sulfur-bearing, then for the electrolyte-bearing laminae. The meanings of the
code letters are given below.
A = 302 SS, 80% porosity, 30um pore size, Huyck Metals Co.
B = 304 SS, 85%, 40wm, Brunswick Corp.
C = 304 SS, 83%, 25ira, Brunswick Corp.
D = 347 SS, 90%,. 67um, Huyck Metals Co.
E = Grade AX graphite, 63%, 1.4um, Poco Graphite Co.
F = FC-14 graphite, 49%, 3.Sum, Pure Carbon Co.
G = Mo foam, 78%, 25ym, Spectra-Mat, Inc.
H = Mo foam, 75%, 20um, Spectra-Mat, Inc.
bThe area given is the area facing the anode (the projected area).
average power density at the one-hour rate is the product of the average voltage and current density.
-------
A.
- 30 -
III. SUPPORTING LABORATORY INVESTIGATIONS
Phase Equilibrium Studies of Electrolyte-Containing Mixtures
(J. P. Ackerman, E~. S. Tani)
The electrolyte of an electrochemical cell must be in intimate contact
with both the anode and the cathode. If material from either electrode should
dissolve in the elec-rolyte, diffusion to the other electrode could produce
several adverse effects such as chemical self-discharge, the formation of an
insulating film on one of the electrodes, or dewetting of the electrode struc-
ture by the active material. Lithium does not dissolve in' its salts to any
significant extent;7 but the solubility in the electrolytes of sulfur, Li?S,
LixSy and other materials present in the cell cathode is essentially unknown.
The purpose of these investigations is to evaluate a number of electrolytes
with respect to their solvent power for cathode materials and to measure the
variation of solubility with temperature. The information obtained will
facilitate the choice of the electrolyte and of the operating temperature
range for lithium/sulfur cells. The same type of experiment permits the in-
vestigation of the solubility of cathode materials containing various addi-
tives.
Mixtures of various amounts of electrolyte, sulfur, and lithium sulfide
were heated in a specially constructed furnace and observed while at tem-
perature through a viewing port. Visual determination of the number of phases
indicated a general similarity of the Li2S-S-electrolyte phase diagram to that
of the system Li2Se-Se-(LiBr-RbBr), which has been studied in some detail.8
A typical Li2S-S-electrolyte phase diagram at 360°C is shown in Fig. 16.
There are apparently three liquid phases and one solid phase. These are
A THREE PHASES
D TWO PHASES
LiBr-RbBr
Fig. 16. Phase Diagram of the Pseudo-ternary System
Li2S-S-(LiBr-RbBr)
-------
- 31 -
(1) The electrolyte-rich liquid phase, L3. It is deep red at tempera-
tures near 360°C when in equilibrium with the other components of the
system, and yellow when quenched. This phase is more dense than any
other phase in the system when the electrolyte contains bromide or
iodide ions, and is less dense than the L2 phase when the electrolyte
contains fluoride and/or chloride ions only. The density of this
phase is probably near the density of the pure electrolyte.
(2) A liquid phase, L2, composed largely of Li2S and sulfur with lesser
amounLs of electrolyte. It is black and opaque both at temperature and
red in quenched samples, and has a density intermediate between those
salts containing iodide or bromide and those which contain only
chloride and fluoride, i.e., 2.28-2.95 g/cm3 at 360°C. The L2 field
meets the Li2S-S side of the pseudo-ternary phase diagram at the com-
position of the monotectic in the Li2S-S binary system. This has been
determined by P. T. Cunningham to be at 37 at. % lithium in sulfur.9
(3) A liquid phase, LI, whose composition never differs greatly from pure
sulfur. It is essentially insoluble in electrolyte and is the least
dense of the phases present. It has a deep clear ruby color at tem-
peratures near 360°C and is virtually indistinguishable from pure
sulfur when quenched.
(4) The solid, Li2S. It is essentially insoluble in electrolyte10 and
has been detected only in quenched samples by X-ray diffraction.
Some typical compositions which were observed in the system Li2S-S-
(LiBr-RbBr) are shown in Table VI. For other electrolytes, obser-
vations were confined to samples having compositions in the fields
of L3 and L2 + L3.
TABLE VI. Phases Present at Selected Compositions
in the System Li2S-S-(LiBr-RbBr)
Sample
S-l
S-6
C-l
C-3
mol %
Li2S
1.0
19.0
24.8
13.0
mol %
S
85.2
69.1
24.7
39.5
mol %
LiBr-RbBr
13.8
11.9
50.5
37.5
Phases Observed
LI H
L2 H
L2 H
h L2 -
h L2 -
1- L3 -
I- L3
«• L3
1- L3
h Li2S
The solubilities of Li2S and of L\ in L3 are both very low. Thus, the
remaining area of concern related to the solubility of sulfur-bearing species
is the solubility of L2 in L3. Determination of this solubility should yield
information of importance in the selection of the electrolyte which has the
lowest solubility for sulfur-bearing species.
A straight line, drawn from the intersection of the L2 field with the
Li2S-S side of the phase diagram to the pure electrolyte corner, passes through
the L2, L3, and L2 + L3 fields. Intersection of this line with the boundary
between the L3 and L2 + L3 fields is a measure of the solubility of L2 in L3.
This intersection is being determined at 360°C for LiBr-RbBr, LiCl-Lil-KI, and
LiCl-KCl-CsCl and is being determined at 400°C for those electrolytes and the
-------
- 32 -
higher-melting-point electrolytes LiF-LiCl-LiI, LiCl-KCl, and LiF-LiCl-KCl.
The experimental procedure consists of encapsulation of mixtures of the
desired compositions in quartz capsules, equilibration at temperature,
centrifugation at 100-120 g, and quenching in liquid nitrogen. The quenched
samples are microscopically examined to determine the number of phases present
At first it was thought that the solubility of L2 in 1,3 could be deter-
mined by analyzing the 1% phase for sulfur and Li2S. Sulfide analysis was
attempted by dissolving portions of the 1,3 phase in water, and titrating with
a AgNC>3 solution using a sulfide-specific electrode as an indicator. Samples
submitted for total sulfur analysis by oxidation and Ba + precipitation,
followed by turbidimetric or gravimetric determination of BaSO^, showed that
the sulfide ion determination did not provide the required accuracy. The
experimental procedure was changed to one in which the limits of the L2 field
are determined by microscopic examination of quenched samples. Samples in
which the sulfur-to-sulfide-ion ratio is maintained constant were mixed with
varying amounts of electrolyte and equilibrated to determine the percentage of
electrolyte present when the L2 phase disappears. The samples were held in
the centrifuge for 3-4 hr at a temperature 10-15°C above the desired tempera-
ture, cooled to the desired temperature, and equilibrated for another 3-4 hr.
After liquid nitrogen quenching, the quartz capsules were broken just above
the level of the contents, cast in Polylite,* sectioned by grinding parallel
to the force field of the centrifuge, and examined microscopically. It is
possible to estimate within approximately 1% the composition at which the L2
phase disappears. The results to date are given in Table VII for 360 and
400°C. Where inequalities or ranges are given, the experimental data are not
yet complete.
TABLE VII. Solubility of L2 in L3 for
Various Electrolytes at 360 and 400°C
Solubility, mol %a
Electrolyte
LiBr-RbBr
LiCl-Lil-KI
LiCl-KCl-CsCl
LiCl-KCl
360°C
92 + 2
94 + 2
97 + 1
-
400°C
<70
60-70
>90
»90
aMol percent electrolyte at which all L2 is dissolved.
From these results it can be seen that generally those salts having small,
difficultly-polarizable anions have a far less solvent power for L2 than do the
salts containing the large "soft" anions. At present LiCl-KCl ana LiCl-KCl-CsCl
electrolytes are clearly better for cell use from the point of view of solubility
considerations. A similar electrolyte, LiF-LiCl-KCl, has not yet been investi-
gated, but it is felt that, because of the presence of the small, hard fluoride
ion, it may be somewhat better than those which have been examined already.
*A product of Reichold Chemicals.
-------
- 33 -
B. Studies of Sulfur-Bearing Species in Molten Alkali Halides
(D. M. Gruen, A. J. Zielen)
The objective of this study is the characterization of sulfur-bearing
species in fused alkali halide melts. Complex equilibria between species such
as Sn, S2~, and SnS2~ undoubtedly occur and must be understood in order to
optimize cell performance.
A variety of techniques, including electrochemical and spectrophotometric
measurements, ic being employed. The first major goal was the preparation of
sulfur-free sulfide solutions in molten-salt electrolytes. It is essential to
establish the properties of sulfide ion (and elemental sulfur) before the added
complications of polysulfide species can be accommodated. Electrolytic reduc-
tion of a metal sulfide was selected as the preparative method. The "standard"
sulfide solution so obtained could then be used in spectrophotometric studies
and in evaluating the performance of sulfide-ion indicating electrodes.
A second major goal was the establishment of reliable analytical tech-
niques for sulfide and sulfur, particularly direct determinations in the melt
by electroanalytical methods. Precipitation titration with a coulometrically
generated metal ion combined with a potentiometric end point was employed for
in situ sulfide ion analysis. More conventional analytical procedures for
sulfide and total sulfur in frozen eutectic samples are also under evaluation.
All experiments to date have been performed at 400 or 450°C in LiCl-KCl
eutectic (59.5 mol % LiCl), which had been treated with magnesium metal and
filtered, treated with chlorine gas, purged with argon, and sealed in ampoules
of 50 g under an argon atmosphere. A simplified sketch of the molten salt cell
assembly is presented in Fig. 17. The cell top actually contained six standard-
taper entry ports instead of the three shown. The extra ports were used for a
thermocouple well, a tube with Rotaflo stopcock and ball joint for argon entry
or vacuum, and a fourth (if desired) electrode compartment. The electrode and
thermocouple-well positions were readily adjusted after easing the 0-ring com-
pression. Pyrex instead of quartz fritted ware was used in some of the earlier
runs, but the results were poor because of severe attack by sulfide on Pyrex
glass.
In a typical experiment, the apparatus was charged with frozen eutectic
(50 g) and completely assembled in a nitrogen glove box. The sealed cell was
then moved to the furnace, flushed with argon, and maintained under a positive
argon pressure throughout the run. The eutectic was slowly melted and seeped
into the fritted compartments with a period of 16-20 hr allowed for the cells
to fill. The furnace was mounted with counter weights on vertical tracks so
that it could be readily moved for visual inspection of the cell. Temperature
control of the melt is estimated at +2°C.
Electrode and thermocouple potentials were measured to 0.1 and 0.01 mV,
respectively,, with Rubicon Portable Potentiometers. The constant-current
source, built by the Electronics Division of Argonne National Laboratory,
supplied dc current continuously variable from 1 to 100 mA and constant within
*0btained from Anderson Physics Laboratories.
-------
- 34 -
0.1%. Timing to 0.001 sec was provided by a Beckman Universal Eput and Timer
with the start and stop triggers supplied by the voltage drop across a resistor
in series with the constant-current source.
BRASS ROD
BRASS HOLDER
—NEOPRENE 0-RING
• VENT FOR ARGON
— WATER JACKET
— QUARTZ TEST TUBE
QUARTZ BEAKER
FURNACE WALL
SET SCREW
HOLE FOR REMOVAL
QUARTZ COMPARTMENT
NIS CATHODE (PT LEAD)
NICKEL ANODE (PT LEAD)
PLATINUM REFERENCE
MOLTEN SALT
FINE QUARTZ FRIT
Fig. 17,
Molten-Salt Cell Assembly for
Sulfide-Ion Investigations
1. Electrodes
The reference electrode was a spiral of 30-mil platinum metal anodized
at constant current for a measured length of time,11 and the concentration of
Pt(II) determined by the weight of salt in the compartment. The density values
of the LiCl-KCl eutectic are 1.673 and 1.647 g/ml at 400 and 450°C, as inter-
polated from literature data.2
Silver sulfide cathodes were prepared by passing sulfur vapor in an
argon stream over strips of heated silver foil. The result was a heavy coat
of coarse crystalline needles, identified by X-ray and chemical analyses as
pure monoclinic Ag2S.
Nickel sulfide was prepared by sealing a 1:1 atomic weight ratio of
powdered nickel and sulfur in an evacuated quartz tube and gradually heating in
a vertical furnace for several hours until the temperature exceeded 1000°C (NiS
melts at ^980°C). The product was a golden ingot with a striking metallic ap-
pearance. The material, which was found to be an excellent conductor, was
confirmed as NiS by chemical and X-ray analyses. Other nickel sulfide ingots
were prepared by remelting portions of the original product, which is very
brittle and can be crushed with a mortar and pestle. It was found that the
addition of extra nickel metal produces an ingot, probably a mixture of NiS
and Ni3S2, of improved strength and machinability.
-------
- 35 -
Metal anodes, used as isolated counter electrodes and for in situ
titrations, were prepared by spot-welding 5-mm strips of metal foil (Ni, Ag,
or Pd) to 30-mil platinum lead wires.
2. Direct Sulfide Analysis in Molten Salt
Our first notable success was the in situ precipitation titration
of S2~ with coulometrically generated Ni2+, viz.
Ni + S2- = NiS + 2e~ (1)
A typical titration curve is presented in Fig. 18; the equilibrium potential of
the nickel electrode was determined after each anodization increment. Similar
results were also obtained using a silver anode, but the end-point break was
smaller and less sharp.
1.6
1.4
UJ 1.2
1.0
EXPERIMENT LK-4
TEMPERATURE, 450* C
LiCI-KCI EUTECTIC
LI2S ADDED, 4.3mg
NI GENERATING ANODE
Pt/Pt(H) (I M) REFERENCE ELECTRODE
8 12
COULOMBS
16
20
Fig. 18. Precipitation Titration Curve for S2~ With
Generated Ni2+; Experiment LK-4
If the potential of the nickel electrode is determined by the
solubility equilibrium for NiS
NiS = Ni2+ + S2~,
ENi/Ni2+ ' ENi/Ni2+ - (RT/2F)
K = (Ni2+)(S2~)
sp
(2)
(3)
Letting C equal the number of coulombs passed, A the initial concentration of
S2~ in mol/liter, V the solution volume in liters, and dropping the subscripts
of Eq. 3
E = E° + (RT/2F) In [-A + C/(2FV) + /(A - C/2FV)2 + 4K ]/2
sp
(4)
-------
- 36 -
The smooth curve of Fig. 18 was calculated from Eq. 4 after evaluating the ad-
justable parameters E°, K , and A by a least-squares computer fit to the data.
Since the emf measurements in the end-point region are always less accurate
than the other data, these values ("X's" in Fig. 18) were used obtain A, and
the remaining data established the E° and K values.
Table VIII summarizes the Ni/Ni2+ formal potential and NiS solubility
product values obtained by this technique. A value of -0.795 V at 450°C had
been reported for E1'. ^ L The found (S2~) values should correspond to the
solubility limit of Li?S in the eutectic except for experiment LK-4, where losses
becauses of sulfide attack on the Pyrex glass cell may have resulted in an un-
saturated solution. The percentage S2~ recovery entries will be discussed
below.
TABLE VIII. Titration of S2~ in LiCl-KCl Eutectic with
a Nickel Anode; Evaluation of Ni/Ni2+ Formal Potential
and NiS Solubility Product
Experiment
No.
LK-4
LK-11
LK-12
°C
450
400
400
E°,
voltsa
-0.809
-0.801
-0.790
Ksp,
(mol/liter)2
1.8 x icr13
3.7 x i(T15
2.1 x IQ-15
Found S2~,
mol/liter
0.0122
0.0147
0.0118
%c2-
o
Recovery
53. Ob
7.8C
94.3^
aVersus Pt/Pt2+ (1 M) reference electrode.
bAdded as Li2S (Foote Mineral Co.), Pyrex cell.
cGenerated from Ag2S cathode; no correction for S2~ removed with the
generating electrode.
^Generated from NiS cathode; corrected for Li2S precipitate found on
generating electrode.
The Table VIII results present a consistent picture that confirms the
successful in situ analysis for S2~ ion. A similar least-squares calculation
was not made for the precipitation titration with a silver anode, but an
estimate of 2 x lO"11 (mol/liter)3 was made from the data for K of Ag2S at
450°C. v
3. Electrochemical Generation of S2~
The first attempts to generate sulfide ion employed Ag2S cathodes:
Ag2S + 2e~ = 2Ag + S2~ (5)
The procedure was to pass a constant current for a measured length of time and
then to determine the electrode potential vs a Pt/Pt(II) reference. Usually
15-20 min were allowed for the potential to reach a (reasonably) steady value.
Plots of electrode potential vs log (S2~) or log (coulombs passed) were reason-
ably straight and reproducible but gave a Nernst slope for a one-electron
reaction, possibly indicating the reaction
Ag2S + e~ = Ag + AgS~ (6)
-------
- 37 -
Check determinations made by similar semilogarithmic plots for the metal anode
(Ag or Ni) used as the isolated counter electrode invariably gave excellent
straight lines with standard deviations usually less than 1 mV and Nernst
slopes within 1-3% of the theoretical values.
Another disturbing fact was that sulfide analyses made at the end of
the generating run, either in situ as previously described or on the frozen
eutectic, indicated very low yields of sulfide (cf. LK-11 in Table VIII). It
was finally discovered that the missing sulfide was trapped or precipitated
on the relatively large, irregular surface of the crystalline Ag2S, and it
was "lost" by the simple act of removing the generating electrode prior to
the sulfide analysis.
Nickel sulfide was then tried as the generating cathode, i.e., the
reverse of Eq. 1. The first attempt (Experiment LK-12) was very gratifying.
Over the useful range of Li2S solubility, the NiS electrode gave a Nernst
response within 10% of the theoretical two-electron value, and the total
sulfide found and generated values were in agreement (cf. Table VIII).
A mixed NiS-Ni3S2 cathode was used in the next experiment, LK-13.
This run also featured a small nickel electrode in the cathode compartment
with the intent of monitoring the sulfide ion concentration by means of the
Ni/Ni2+ couple and Eq. 2. It was hoped that such a nonworking electrode would
be better behaved than the generating cathode, and that is what occurred. The
mixed-sulfide cathode gave very erratic potentials, probably indicative of
mixed potential behavior. The nickel electrode data are presented in Fig. 19.
The straight line is drawn with the theoretical slope; the horizontal break
occurs due to Li2S precipitation. Furthermore, if the observed Ni/Ni2+
potentials are combined with the LK-4 results of Table VIII to calculate log
(S2~) via Eq. 3, the results agree with the "coulombs-passed" values within
a mean of 0.10 log unit. This is well within the error limits of Table VIII.
1.68
1.64
t.60
1.56
1.52
EXPERIMENT LK-13
TEMPERATURE.450°C
LICI-KCI EUTECTIC
NI/NI++ MONITORING ELECTRODE
Pt/Pt++(IM)REFERENCE ELECTRODE
NIS-NI3S2GENERATING CATHODE
-2.4
-2.0 -1.6
LOG(S-). CALCULATED FROM COULOMBS PASSED
-1.2
Fig. 19. The Potential of the Ni/Ni2+ Electrode as
a Function of S2~ Concentration
-------
Thus i ln> olr;el rochiMnlr.al nom-ration of S" ~ in molten salt lias boon
• IfiiiiMi::! i :ii i-.l ::.-i I I s I ac t or (I y . l-'uluiv work wlU tf.st tho response of NiS as a
II.MIU-.M k iny. . i ml leal Iny. i-U'ctrodr. A fully ox.tdizt-d electrode of tills type
will !•«•. ii'.|iilrod In po I ysul I lilc Collie Ions .
•i. Spec I roi-hem lea I Studio::
A "si ;nulard" sul Mill- solution was y.ono.rated with a nickel suJt'ido
I'alhodo. jiiul i lio e 1 eel rol yt. e was t ro/.en . Kerne l.tod portions of this material
with v.-irvliiK ainouiil s ol added sullur are bo i ny, examined sped vopholomel r ica 1 1 y
'Vlu- results :nv loo preliminary lor il i souss I on except to state lliat a dolinite
polvsull lili- ;ihsorpf Ion has been observed I hat I:: reproduc J b Lc and responds
irverslhlv ( i' I i-mpor.-i( tiro .'uul HIM:::: act Ion.
'» . An.-i I v:i I s i'l !•' ro/.en Kut cell i- SampJ_es
An.ilyt ir.-il iiift.hoils l\oinl dottv.tor . Analysts m.-uU* v<-ilh
.-i pmv- N.-i.'S nioik solul Ion In I M Na(>ll havt- K'VOII vi-ry sat l.sl 'ui'lory rosnlts
^O-l''- arrr.r.-irv') il. a:; prov i OXIM 1 y snj-'.y.i-sl.i'il , tlu1 Lltratlon is made l\\
I .',' NII..OII » II. I .V NaOII. Opt I mum. pi-rl ormaiu'i.1 Is obl.ti lucd if tho Orion oli-t.--
i ro.li'. I:: nol i nsrr t.nl until noar I he I'.iul piilnl.. 'I'll is miniml.zcs i'oa.«nlal Ion
fl AH _-S pov i |i i I M( >• on t.ln- sonsini! mrml>rane • Thi1 cust. omai'y 1 .i t oraLuro pro-
i-i-iliiiv ol I itr.-it iny. In I -I.' N.-iOIl slioulil bo aliaiiiloni'il bi'fausii c>f errors ctiusoil
l :.! u 1 f ate . ' '' Thus for a ]io 1 ysul f f vlo solut ion
10- ,~ » 3.S..S-'- i tjnOir - (-'i < :in)l~ » 3(.l. K n)SO;- «• :»nll .-0
An IniU'pi'iuli'iit i:\illiilo ion iloliM'ini 11:11. ion nui:>l lu: comb.l.no.il with I ho l.oilato rosult
! »' il i t I i-n-iif l.-it i- holwrrn S ami S'". l'robU.'.niy arise In tin; loilaU1. tltralion
hor.'iiuu1 .-aillnr i ;: I orinr.il an an i nturim.'il i at.o. prtnlnct find furt.hrr oxiilation i ;••
sliM.'. Thus low ivsults can he oht.;« luc«l hircause ol. l.ncomp l.c.t c i'xiilalii?n an. I/or
lo:;r: ol siillur hy slcHin iH.yt i. 1 1 a I ion . Using our standard Nw/s? sjolntlon. valiu-:;
rany.iny. from dS t,o T/.H!? of l.lie c.-ilirulatoil su-H ido hrtv« but?n Dhsorvo.d. dopoiul
iny. upon tin- c.xfi'sg of (.culatt: used and tho bol.IJ.ng time- However, it Is an-
( li: ipali'd that, with i ho proper tond.l.t l.ons. accurate: result.;? wi.U be; obtainoii.
i : . So I i .1 - -V. 1 ri- 1. 10 1 y t^e. _-?Jlu
(T. W. l.al..lniM.M. electrolytes -In r he s .11111-
t I'mporatni-i.' ranu.i: hsivc condvict.lv tttes of 1 to 10 ohm"1 cm"1. Wcc.au.ir soli, I
olivt rolytv.s »M'l«-r som« doslgn advantages, nin'l offer the ponw.iljj.l.l ty ol. avoid
Iny. «.-alhodi- mat.or.lal solubility In tho. e..l«i:l.ro.lyti?. f»n oxp. lovsil.ovy lt\ver:My..-n ton
is boiny. vnulerl ake.n in soared ol: ooUd c l.cc ;t ro 1 y I. RM with hly.h lithium- i. MI ,-on
due l Ivl I v.
-------
- 39 -
In soiid materials, ions move through the interlaced cation and anicn net-
work by means ot detects in -.he network. These detects, such as ions on inter-
stitial sites 01 vacant ion sites in the network, can be generated thermally
or be present because of charge-balance impurities in the material. In normal
crystals the number of detects present is relatively small, and consequently
their ionic conductivity is small. There are, however, a number of solid
compounds whi'Ch have ionic conductivities greater than 0,1 ohm" cm~ , These
compounds show a much lower temperature dependence of conductivity than that
found for normal crystals-
There appear to be three broad classes ot compounds with high ionic con-
ductivity. These are (1) the silver haiides and a number c>f ternary compounds
of silver haiide with an alka.1 1 metal, (2) Li2S04 , Li^WOi, , and the mixtures
(Li, Na)2SCL and (Li, AgJ^SOj, and (3) the ceramics based on 6-alumina
(MU Al;;0-. 7, where Mi, can be alkali metal or silver): The first class of ma-
terials has essentially a rigid crystalline arrangement, of anions with a large
number of interstitial sites on which the cations ai e distributed in a dis-
ordered fashion- There are always far mere sites than caticns. The situation
is similar in Li2SOu , where in each cell the two ^.ithium ions have three cation
sites. The 6-alumina lattice has a hexagonal layer structure. Every fifth
layer of oxygen is no longer a close-packed layer but contains only one quarter
of the usual number of oxygen ions. In this layer alkali metal ions can be
present on both oxygen and interstitial ion sites.
In all these classes of material there are always more cation sites
available than the number or cations present, and this is the basic reason
for the high conductivity. The 6-alumina is slightly different from the other
materials in that the cation disorder is only present in well-defined layers
and not throughout the entire structure. While this limits the overall ionic
conductivity, measurements have shown that this should not prove a limiting
factor in the use ot this material as a solid electrolyte in high-temperature
batteries.
Both Li2SOa and Li^WOu have cation-disordered structures above ^50 to
500CC. This temperature is too high for the lithium/sulfur battery, and in
any case there would be compatibility problems using Li2SO.j . On the basis of
our survey ot cation-disordered structures, we decided that the greater pos-
sibility of finding a high-lithium-ion conducting material lay in materials
possessing the 6-alumina layered structure. We believe that there should be
lithium compounds analogous to the sodium compounds possessing the 6- and
6"-type phases r The following sections describe the methods we have employed
to synthesize these structures and to determine whether the materials possess
a high conductivity .
1, Li20-MgO-Al203 System
This system was investigated because of the possibility of obtaining
lithium 6- or 6"-aluminas by conventional ceramic fabrication techniques. Eight
compositions in the Al20j corner of the phase diagram were chosen for investi-
gation and are given in Table IX,
-------
- 40 -
TABLE IX. Starting Compositions (in wt %) of
Li20-MgO-Al203 Specimens
No. Li20 MgO A1203
1
2
3
4
5
6
7
8
5.54
5.15
2.50
2.50
3.75
5.00
5.00
6.25
_
6.95
2,50
5,00
3.75
2.50
5,00
3.75
94.46
87.90
95.00
92.50
92,50
92.50
90.00
90.00
The raw materials were Li2C03, MgO, and A1203 (Alcoa Type A-15).
After calcining in air at 685°C for 1 hr, the specimens were fired at 1700 and
1800°C for 1 hr in helium. Weight losses at the two firing temperatures and
densities at 1800°C are given in Table X. The weight losses at the firing
temperatures are attributed to the vaporization of Li20.
TABLE X. Weight Loss and Density of Li20-MgO-Al203
Compositions Fired at 1700 and 1800°C
No.
1
2
3
4
5
6
7
8
1700°C
Wt Loss, %
0.72
1.02
0.38
0.40
0.44
0.67
1.02
1.90
1800°C
Wt Loss, %
1.42
2.39
0.34
0.46
0.75
1.31
1.59
-
Density, g/cm3
2.92
3.11
2.86
2.85
2.86
3.03
3.18
-
These weight losses were considerably less than those found with
similar Na20-MgO-Al203 compositions. The conductivities of all the specimens
were of the order of 2 * 10~8 ohm"1 cm"1 at 400°C. The low conductivities
and low weight losses of these compositions indicate no tendency to form the
hexagonal $"-alumina structure. The compound Li20 • 5 A1203 (Composition 1)
belongs to the cubic crystal system.11+
The use of Alcoa Type A-16 alumina (crystallite size
-------
- 41 -
in the area of 1,3203 • 11 A1203 was begun in order to determine the extent of
the phase field containing the 6-alumina structure. The starting compositions
are given in Table XI,
TABLE XI.
Starting Compositions (in wt %) of
Specimens
No.
Li20
La203
A1203
L-l
L-2
L-3
L-4
L-5
L-6
_
2.00
4.00
2.36
4.65
4.88
22.51
22.06
21.61
17.15
16.00
12.00
77.49
75.94
74.39
80.49
79.35
83.12
and
The raw materials for these compositions were Li2C03,
A1203 (Alcoa Type A-16). The compositions containing 4 wt % or more
were overfired at 1650°C. The weight losses and densities of the samples
fired at 1500, 1650, and 1700°C are given in Table XII.
TABLE XII. Weight Losses and Densities of
Compositions Fired at 1500, 1650, and 1700°C
Weight Loss, %
Density, g/cm3
No.
1500°C
1650°C
1700°C
1500°C
1650°C
1700°C
L-l
L-2
L-3
L-4
L-5
L-6
1.46
3.12
3.14
2.48
2.49
1.99
14.2
3.72
OFa
3.02
OF
OF
1.63
3.84
-
3.10
-
-
3.10
3.15
3.49
3.10
3.25
3.27
3.44
3.85
OF
3.89
OF
OF
3.94
3.96
-
3.99
-
-
aOverfired.
Conductivity measurements of these samples indicated that no high-
conductivity phases were present. However, the weight losses may have in-
fluenced the crystal structure, and the firings will be repeated in an alkaline
atmosphere, similar to the method used in firing sodium 6-alumina.
3. Sodium B- and g"-Alumina
Since the direct synthesis of lithium 6- or 6"-alumina structures
does not appear to be very promising, the investigation of a method for their
production from sodium 6-alumina by ion exchange was begun. The first source
of sodium 6-alumina was Monofrax H,* a commercial refractory product, con-
taining 5.7 wt % Na20. Approximate weight losses on firing Monofrax H were
measured (in vacuum) on a single l-in.-dia specimen by means of a thermo-
gravimetric balance. The weight-loss rates were 4.8 mg/hr at 1400°C, 11 mg/hr
at 1500°C, 58 mg/hr at 1600°C, and 120 mg/hr at 1700°C.
Product of the Carborundum Co., Falconer, N.Y.
-------
- 42 -
Three compositions (S-4, S-5, and S-6) were prepared after milling
-12 mesh Monofrax H for 14 hr. Composition S-4 was Monofrax H; Compositions
S-5 and S-6 were Monofrax H with additions of 1.78 wt % H3B03 and 4.56 wt %
Na2C03, respectively. Based on the thermogravimetric data, a fast firing in
helium consisting of a 30-min soak at 1400°C, an increase to 1675°C in 15 min,
and a 5-min soak at 1675°C was tried. Other specimens were packed in -20 mesh
Monofrax H and fired in helium for 1 hr at 1700°C. Table XIII gives the fired
densities obtained hv these methods; neither yielded specimens of sufficiently
high density,
TABLE XIII. Fired Densities of Monofrax H Specimens
Fast Fire to 1700JC Packed in
No. 1675°C, g/cm3 Monofrax H, g/cm3
S-4 2.66 2,81
S-5 2.57 2.68
S-6 2.75 2.69
The use of Monofrax H as a starting raw material was discontinued
upon receiving 100-325 mesh 6-alumina (sodium content = 7.3 wt %) from Alcoa.
This material was ground for 16 hr with sampling at 4-hr intervals to deter-
mine the effect of grinding time on density. The specimens were fired at
300°C/hr to 1700eC in helium (1-hr soak). Table XIV shows the unfired and
fired densities as a function of milling time.
TABLE XIV. Effect of Milling Time on Unfired and
Fired Densities of Alcoa B-alumina
Milling Time, Unfired Density, Fired Density,
hr g/cm3 g/cm3
4 2,15 2.78
8 2 .25 3.00
12 2.27 3.04
16 2.28 3.05
An attempt to fire the specimens at 550°C/hr resulted in specimens
having densities of under 3.0 g/cm-. A slightly higher firing temperature is
expected to increase the density to 3.15-3.20 g/cm3 for specimens to be used
for lithium ion-exchange studies.
Two compositions in the Na20-MgO-Al203 systems were chosen for
preliminary study of sodium 6"-alumina- These were S-2 (8 wt % Na20, 3 wt %
MgO, 89 wt % A1203) and S-3 (10 wt % Na20, 4 wt % MgO, 86 wt % A1203). Firing
these compositions in helium at 1800°C after prefiring at 700°C resulted in
high weight losses of 4.8-7.2 wt % for S-2 and 7.8-10.6 wt % for S-3. The
conductivity of S-3 was 3 * 10~3 ohm-1 cm"1 at 400°C despite the large Na20
loss. Firing in an alkaline atmosphere (packed in Monofrax H) resulted in
lower Na20 losses, but expansion caused by the density change involved in the
formation of the ^"-alumina structure cracked the specimens. The effects of
prereaction temperatures and times (before forming) on the fired products are
being investigated.
-------
- 43 -
<*. Conductivity
Measurements were made of the ionic conductivity of each sample.
The results were used to determine whether any high-ccnductivity phase was
present in the sample-. This method of "screening" samples proved to be a
relatively quick and reliable procedure.
The measurements were made with an ac bridge (Electro Scientific
Industries, Model 291-A), operating at a frequency of 1000 Hz, and an oscil-
loscope (Tektronix Model 515A) to improve the sensitivity™ A two-probe cell
was used because of the simplicity of sample changing. The problems associated
with two-prcbe cells did not present any difficulties with the measurements.
The aluminate disks were coated with graphite on each flat surface and mounted
between platinum electrodes. The conductivity of each sample was measured in
a flowing stream of argon (purity, 99,999%)- To confirm that the ac bridge
was operating correctly, the conductivity of a sample of NaCl was measured.
The results agreed with previous data on NaClr
Hsueh and Bennion-" found that the hygroscopic nature of 6-alumlna
made their conductivity measurements strongly dependent on prior heat treat-
ment. None of the lithium-containing samples investigated appeared to be
hygroscopic. It is possibly not surprising that when the "layered" 6-alumina
structure is formed, the material becomes hygroscopic. To eliminate the pos-
sibility of this effect en our conductivity measurements, we have followed
Hsueh and Bennion's procedure of heating the samples tc 9003C in an argon
atmosphere and holding at that temperature for 1 hr. Conductivity measurements
were then made as the sample was slowly cooled from 900 to 350°C. From the
data obtained, plots were made of log conductivity versus 1/T. The results of
all the measurements are shown in Fig. 20. It was found that the plots were
linear for all the lithium-containing samples. The data for the individual
samples are not shown in Fig- 20; however, each plot lay within the shaded
area appropriate to the composition series. The activation energies for all
the lithium-containing samples were between 17 and 21 kcal/mol.
It can be seen in Fig, 20 that the conductivities of all the samples
in the two lithium series have similar activation energies to those of LiCl
and NaCl in their extrinsic conductivity range (lower slopes). It appears
likely that the conductivity of the lithium-containing samples occurs by the
same processes that are responsible for the conductivity of the alkali halides.
In the extrinsic range of conductivity, the defects responsible for the con-
ductivity occur because of impurities which are present in the material- The
greater the amount of impurity, the higher the conductivity. Unfortunately,
impurity solubilities limit the amount of conductivity gain that can be achieved
The range of values represented in each shaded area of Fig. 20 gives an indica-
tion of the changes that were obtained by varying the compositions of the
samples in any one series.
The samples of sodium aluminates (S-3 and Monofrax H) show a very
different variation of conductivity with temperature. The plots are curved
and at about 400°C have an activation energy of approximately 5 kcal/mol.
This value of the activation energy is similar to that found by Imai and
Harada'7 in their investigation of the effects of divalent impurities on the
conductivity of 6-alumina. A low value for the activation energy is to be
-------
expected in a "layer" structure like B-alumina. In normal ionic conductors,
the activation energy is higher because it includes the energy to form the
defect responsible for the conduction and also the energy to move an ion by
means of that defect. In a cation-disordered structure such as B-alumina,
there is a large fraction of vacant cation sites always present in the net-
work, and the energy required to move the cations is small.
id
TEMPERATURE. *C
65Q 450 550
10
12 14
I01/ T °K
Fig. 20. Ionic Conductivity of Aluminate Samples
While the activation energies for the ionic conductivities of our
samples of sodium B-aluminates are the same as that found by Imai and Harada,
the absolute values of the conductivity are considerably lower. This is almost
certainly due to the small amount of high-conducting phase present in our
samples. Only a small amount of the B-alumina phase was formed due to large
Na20 losses during the firing of the samples. The improved procedure for
sample firing should minimize these losses and result in compacts having ionic
conductivities of the order of 10"1 ohm"1 cm"1 at 400°C.
D. Cathode Material Studies
(R. K. Steunenberg, R. M. Yonco)
Small-scale cell tests have shown that elemental sulfur alone can be used
as the active cathode material in lithium/sulfur cells. However, the use of
additives to increase the electronic conductivity and decrease the vapor pres-
sure of the sulfur may result in improved electrical performance of the cells,
and possibly permiL a simpler, lighter-weight cathode current-collector design.
Iron, thallium, and other materials have been under consideration as potential
sulfur additives.
-------
- 45 -
The solubility of iron in liquid sulfur at bQQ°C is reported to be about
5 at, %,-ri Several unsuccessful attempts were made to prepare a 4 at. % solu-
tion of iron in sulfur by heating FeS in liquid sulfur at temperatures up to
about 45CPC, Thermal-analysis results and visual observation indicated that..
the FeS had not dissolved completely, even after several days at cemperatuies
around 400:)C, One sample of FeS in sulfur was held for several days at '-i50'JC
and filtered through a fine Pyrex frit* The electrical conductivity of the
filtrate was <10~7 ohm"1 cm"1 at 127 and 377°C. This sample was then ignited
in air= The weisht of the residue was negligible ('0,01 wt %), which, together
with the conductivity data, indicates that the sulfur contained little or no
iron, (The conductivity of high-purity sulfur at 377:'C is "lO"9 ohm~: cm-5.)
These results may be attributed either to a slow dissolution rate or iron
sulfides in sulfur, which is consistent with observations reported in the
literature, ^ or to an error in the published phase diagram,"" In either case,
iron does not appear to be a premising additive to sulfur cathodes in lithium/
sulfur cells and no further studies are planned on the iron-sulfur system.
Sulfur-thallium mixtures are reported to have a high electrical conduc-
tivity ('"1 ohm"' cm"-) at 400°C.2C These data, however, were obtained at
relatively high thallium concentrations (-37.5 at, %), which would be un-
desirable in a lithium/sulfur cell because of the high atomic weight of
thallium (204,37 g/g-atom)„ Therefore, the electrical conductivities of
sulfur containing much lower concentrations of thallium are of interest as
a possible cathode material. Because the thallium-sulfur system is reported
to have a liquid miscibility gap in the sulfur-rich region of the phase dia-
gram and the phase boundaries of the gap appeared to be poorly known,"' a
brief investigation was undertaken to define the compositions of the two liquid
phases. In the first experiment, a sulfur-20 at, % thallium mixture was pre-
pared by heating the elements together in a sealed quartz tube- The existence
of the miscibility gap was confirmed by the formation of two liquid layers-
In subsequent experiments, the boundaries of the miscibility gap were
determined over the temperature range 350 to 444'JC- Quartz ampoules con-
taining approximately 20 at - % thallium in sulfur were equilibrated for periods
of 15 to 25 days at temperatures of 350, 410, and 444°C and then quenched.
Samples of the upper and lower phases were analyzed for thallium by wet chem-
ical methods. The results were as follows:
Equilibration temperature, °C 350 410 444
Thallium concentration, at. %
Upper phase -0.02 <0,08 0.09
Lower phase 26*5 26,0 26,9
These results indicate that the solubility of thallium in sulfur is consid-
erably lower than that reported in the literature and that the thallium con-
tent of the higher-density phase is slightly less than the literature value.
The electrical conductivities of the two liquid phases were measured,
using a Pyrex "U-tube" conductivity cell with molybdenum electrodes- The
conductivity results are presented in Table XV, The fact that the conduc-
tivity of the sulfur-rich phase is very low suggests that small concentra-
tions of thallium do not have a significant doping effect on the sulfur. The
relatively high conductivity of the thallium-rich phase may prove beneficial
-------
- 46 -
to a lithium/sulfur cell containing thallium as a cathode additive if a
sufficiently small amount of thallium can be used.
TABLE XV. Electrical Conductivities of the Two
Immiscible Liquid Phases in the Sulfur-Rich
Region of the Thallium-Sulfur Phase Diagram
Conductivity Cell
Material:
Shape:
ElectroJes:
Cell Constant:
Pyrex
U-tube
molybdenum
27.4 cm'1
Upper (Sulfur-Rich) Phase
Specific Conductivity,
ohm-1 cm
Frequency,
kHz
0 (dc)
0.1
1.0
10.0
150°
9.5 x
2.2 x
1.6 x
5.5 x
C
io-9
io-9
10~8
io-7
345
1.4 x
2.0 x
2.6 x
5.6 x
°C
io-9
io-y
io-7
io-7
Lower (Thallium-Rich) Phase
Specific Conductivity,
ohm
-1
cm
-1
Frequency,
kHz
0
0
1
10
(dc)
.1
.0
.0
1.
1.
1.
1.
142
57
59
59
60
°C
X
X
X
X
367
10
10
10
10
-2
-2
-2
-2
3
4
4
4
.8
.4
.4
.4
x
X
X
X
°C
10-
10"
10-
10-
1
1
1
1
cm
_1
In addition to the above studies, samples of sulfur containing a propri-
etary additive that is claimed to raise its conductivity to about 10"^ ohm"1
at 130°C were obtained from the Freeport Sulphur Company, Belle Chasse, La.
(The conductivity of high-purity sulfur at 130°C is 10"12 ohm"1 cm'1 or less.22)
Preliminary tests with a volt-ohmmeter showed this material to be conductive
both when molten and at room temperature after it had solidified. Subsequent
conductivity measurements were made, using a Pyrex cell with molybdenum elec-
trodes. The following results were obtained:
-------
- 47 -
Specific Conductivity,
ohm ; cm~~
frequency,
kHz
0-
1,
10.
dc
1
0
0
1
1
1
1
136
,23
.23
.23
.16
X
X
X
X
°C
10"
10"
10"
10"
158
3
3
3
3
1
1
1
,43
.43
,42
X
X
*
°C
10"
10-
10-
3
3
•3
^
These results show no dependence of the conductivity upon frequency in the ac
measurements, and the ac and dc data are in reasonable agreement. In the dc
measurements, the system exhibited Ohm's law behavior up to the maximum potential
that was applied (5 V). The conductivity appears to increase somewhat with in-
creasing temperature. It was not possible to obtain consistent conductivity
data at higher temperatures because of a tendency of the material to form gas
bubbles in the cell.
The results of the above studies indicate that iron is unlikely to be a
useful additive to the cathodes of lithium/sulfur cells. Thallium may prove
to be somewhat beneficial, but its toxicity is a potential disadvantage. The
proprietary "high-conductivity" sulfur supplied by the Freeport Sulphur Company
appears to be unsuitable for use in lithium/sulfur cells because of its ten-
dency to segregate and to release gases. There are other potential additives,
however, which are believed to merit investigation, and further work is planned
in this area.
E. Mass-Transport Studies
(B, A. Feay)
Work in this area has been concentrated on two problems: (1) increasing
the cycle life of the lithium anode and (2) increasing the capacity density of
the sulfur cathode. Considerable progress has been made in the work on the
lithium electrode. It appears that the problem is understood and is amenable
to solution. The work on the sulfur electrode has been hindered by the dif-
ficulty in developing a suitable test cell. However, recent results are
encouraging.
To study the lithium electrode, the cell shown in Fig. 21 was designed to
allow the simple operation of lithium electrodes of various designs in molten-
salt electrolytes of different compositions. The lithium electrode shown in
the figure is typical of the anodes presently used in lithium/sulfur cells *
It consists of lithium soaked into Type 302 stainless steel Feltmetal (90%
porosity, 67-um pore size). The lithium was added to the Feltmetal in a
furnace well at 650°C. The counter electrode is a disk of aluminum. When
the cell is discharged, lithium is transferred from the Feltmetal to the
aluminum, creating a Li-Al alloy. When the cell is charged, lithium is trans-
ferred from the aluminum back to the Feltmetal,. The aluminum counter electrode
was chosen because of the stable voltage and good cycling capability of the
Li-Al alloy, which allows an evaluation of the lithium electrode alone,
-------
- 48 -
ELECTRODE LEAD.
1/8 In. MOLYBDENUM""
ROD
TANTALUM CRUCIBLE-*
STAINLESS STEEL NUTS-
STEEL FELTMETAL.
0.318cm THICK,
3.18cm DIAMETER
ELECTROLYTEX
^^
™~~illlt
~5^~r=™
«*%rfw^_^^^^^_^^_
~~jf rS^^^
\
\
\
s
I^^fl. |
^~~~ !
^JP1
-
sfe;
y&z
^=s:
**^w**n
^-ELECTRODE LEAD.
1/4 In. STAINLESS
STEEL ROD
^CERAMIC SLEEVE
>STAINLESS STEEL
es-Dcuic
— 4.75cm-dlo
ALUMINUM DISK
STAINLESS STEEL PLAT
STAINLESS STEEL SCREWS
Fig. 21. Schematic Diagram of Che Lithium Electrode Test Cell
The cell was operated in the furnace well of a helium-atmosphere box and
was heated to temperatures between 380 and 425°C. The cell was cycled con-
tinuously by charging and discharging between maximum and minimum cutoff
voltages. A voltage-time trace for a typical cycle at a current density of
0.25 A/cm2 (2.0 A) is shown in Fig. 22. As the cell was discharged, the voltage
decreased gradually, and then dropped rapidly to the cutoff voltage of 0.05 V
when the lithium became depleted in the Feltmetal. As the cell was charged,
the voltage increased gradually, and then rose rapidly to the cutoff voltage
of 0.55 V when the lithium became depleted in the aluminum. During this cycle,
the capacity density was 0.8 A-hr/cnr (based on the weltmetal electrode).
CELL VOLTAGE. V
/>7K
Q50
0.25
000
~f\9*.
— DISCHARGE AT 0.25 A/cmZ —
i i i
V
i i i
\t
f
HV.UI 1 1
•• — CHARGE AT 0.25 A/cm *|
— "^
i i i
•9—
.
TIME.hr
Fig. .22. Typical Voltage-Time Behavior of a Li/LiCl-KCl/Li-Al Cell at
425°C and at a Current Density of 0.25 A/cm2
The performance of the four cells that have been run is summarized in
Table XVI. The maximum capacity density is a measure of the amount of lithium
-------
- 49 -
the Feltmetal could hold if the lithium were available. The observed capacity
density is based on the amount of available lithium transferred during a cycle.
The capacity density based on lithium added compared with the observed capacity
density is a measure of the unavailable lithium. Most of this unavailable
lithium is probably tied up in remote regions of the lithium-aluminum electrode.
TABLE XVI. Summary of Lithium Electrode Cell Tests
Lithium Electrode: Lithium in Huyck Type 302 stainless steel
Feltmetal (90% porosity, 67-um pore size)
Counter Electrode: Lithium in Aluminum
Interelectrode Distance: 1 cm
Cell Number
Lithium electrode
dimensions, cm
Projected area, cm2
Counter electrode
dimensions, cm
Projected area, cm2
Electrolyte, mol %
Position of lithium
electrode
Insulator
Temperature, °C
Hours of operation
Number of cycles3
Two disks, One disk, One disk,
3.18 dia 2.54 dia 2.54 dia
x 0.318 * 0.318 x 0.318
8.0 5.0 5.0
Aluminum disk 4-75 x 0.635 thick,
with four concentric grooves
(0.318 x 0.318) cut in surface
17.8 17.8 17.8
Lid 58.5 LiCl 8-5 LiCl 8,5
KC1 41.5 Lil 59.0 Lil 59.0
KI 32.5 KI 32.5
At surface of electrolyte
A1203 A1203 A1203
425 390 390
100 4 2
10 2 1
One disk,
2.86 dia
< 0.318
6,4
Aluminum disk
5.1 dia x
0.318 thick
20.3
LiF 3.5
LiCl 56.0
KC1 40.5
Top surface
0.32 cm below
electrolyte
surface
BeO
395
675
202
Capacity density,
A-hr/cm2
Maximum^
Based on total
1.18
0.59
0.59
0.59
lithium
Observed
Remarks
1.11
0.8 at
0.25 A/cm2
Failure by
lithium
beading on
Feltmetal
during
10th cycle
recharge
1.11
0.25 at
0.25 A/cm2
Lithium
beading
on first
recharge
1.11
_
Lithium
beading
on in-
serting
lithium
electrode
1.11
0,51 at
0.125 A/cm2,
0.375 at
0.25 A/cm2
10 g of LiCl-Lil-
K.I added at cycle
No. 139 (480 hr);
instant bubbling,
lithium beading,
deterioration in
cell performance
aAbove a current density of 0.125 A/cm2.
on quantity of lithium required to fill void volume of the Feltmetal,
-------
- 50 -
During the operation of the first three cells, a problem occurred which
appears to be a major factor that limits the cycle life of the present lithium
electrodes. While the ceil was charging, the returning lithium failed to soak
into the Feltmetal. Instead, it formed beads on the surface of the Feitmetal
which escaped from the electrode by floating to the surface of the electrolyte.
Cell No. 4, which had a LiF-LiCl-KCl eutectic electrolyte, performed
very well for 480 hr and 138 cycles. The electrolyte remained clear and color-
less and there was no evidence of lithium beading on the stainless steel Felt-
metal. During the first few days of operation, the capacity density decreased
gradually, probably because lithium migrated to remote regions of the aluminum
disk and became inaccessible. When more lithium was added to the Feltmetal,
the lithium soaked in immediately, and the capacity density increased. After
the first few days, the capacity density stabilized. At a current density of
0.125 A/cm2 the capacity density was 0-51 A-hr/cm2, and at 0.25 A/cm2 the
capacity density was 0-375 A-hr/cm2
On Cycle No. 139, after 480 hr of operation with no noticeable deteri-
oration in performance , 10 g of LiCl-Lil-KI eutectic was added to the elec-
trolyte. Immediate bubbling occurred around the electrodes, the electrolyte
became clouded, there was some lithium beading, and the capacity density
decreased. The electrolyte gradually cleared, and the capacity density
stabilized. At 0.125 A/cm2 the capacity density was 0,44 A-hr/cm2, and at
0.25 A/cm2 the capacity density was 0.125 A-hr/cm2. The lithium beading
continued and became more and more pronounced- After 460 hr of operation,
the lithium electrode was removed and inspected. Most of the lithium appeared
to be on the surface of the Feltmetal. The electrode was replaced and a small
piece of lithium was added. The lithium did not soak into the Feltmetal, and
the capacity density did not increase. On Cycle No, 202, after 675 hr of
operation, the cell was shut down and disassembled.
The significant differences between the fourth cell and the three previous
cells were
(1) LiF-LiCl-KCl eutectic was used as the electrolyte;
(2) beryllium oxide was used instead of A^Oi for ceramic sleeve;
(3) the lithium-soaked Feltmetal was completely immersed in the electrolyte
instead of being partially exposed to the box atmosphere.
Under these operating conditions, the present lithium anode design appears capable
of good cycle life, The addition of the LiCl-Lil-KI had an obvious deleterious
effect, A probable cause was moisture in the iodide salt. If iodide-containing
electrolytes are to be used, it appears that they must be of much higher purity.
It may be that iodide has such a strong tendency to pick up moisture that it
cannot be used. It is also possible that iodide ion has a significant effect
on the ability of the electrolyte to displace lithium from the Feltmetal.
The second problem under study is increasing the capacity density of the
sulfur cathode. An attempt was made to develop a cell which would run in a
quartz housing. This housing would have allowed visual observation of the cell
during operation- However, the failure of several cells indicated that quartz
was too susceptible tc lithium attack, even when L1-A1 alloy was used in place
of pure lithium for the anode.
-------
- 51 -
From the experience gained with the quartz cell housing and from cells
developed for other tasks, the cell shown in Fig. 23 was designed and con-
structed. Preliminary tests with this cell indicate it will be an excellent
cell for studying sulfur cathodes. It is sealed and is easy to assemble and
disassemble. It is easy to use various cathode materials, and the cathode
design can be changed without changing other parts of the cell. The first
series of tests to be run with this cell will be to study the effect of dif-
ferent carbon and graphite materials, used as cathode current collectors, on
cell performance.
INCONEL SPRING
INSULATOR
Mo ANODE HOLDER
ELECTROLYTE
INSULATOR. RING
Mo FOAM
CATHODE' S IN GRAPHITE
GRAFOIL
Al PLATE
Mo HOUSING -
GRAFOIL
ANODE' Li IN FELTMETAL
Mo RING
Mo HOUSING
Al PLATE
INSULATOR
Fig. 23. Lithium/Sulfur Test Cell
-------
- 52 -
IV. MATERIALS TESTING AND FABRICATION
The corrosion resistance of various materials- to simulated cell environ-
ments is being studied to identify materials with potential usefulness for
cell applications. Two distinctly different classes of materials resistant
to cell conditions are required. First, corrosion-resistant materials possess-
ing good electrical conductivity are required to serve as current collectors
and cell housings. Secondly, a corrosion-resistant, electrically insulating
material is required to prevent short-circuit contacts within the cell. The
materials chosen for the cell cathode must be resistant to attack by sulfur,
lithium-sulfur mixtures, and electrolyte at 375°C for thousands of hours.
The anode housing and current collector must withstand exposure to molten
lithium and the electrolyte under the same time and temperature conditions.
Electrical insulators may require resistance to all of the cell constituents.
Still other materials may be needed for hermetic sealing of cells, electrical
feedthroughs, thermal insulation, and exterior housings. The performance of
these various materials can be evaluated fully only through their use in
operating cells and batteries. This type of materials testing, however, is
expensive and time-consuming, and many candidate materials can be eliminated
from consideration by relatively simple corrosion tests.
In addition to corrosion testing, scaling-up the lithium/sulfur cells
from the successful laboratory sizes to larger sizes has indicated a need
for materials and materials-processing development. Two types of materials
that are undergoing such development are (1) corrosion-resistant, high-
surface-area, open-structure, cathode current collectors and (2) corrosion-
resistant ceramic insulators and electrical feedthroughs. The bulk of the
work was concentrated on the development of corrosion-resistant ceramic in-
sulators because of a more immediate need,
A. Experimental Procedure
Corrosion tests in a 20 at. % lithium-80 at^ % sulfur mixture were normally
conducted as dynamic isothermal tests- In these experiments, the samples (about
0-32 * 0.32 * 2,5 cm) were loaded into quartz capsules (about 2.5 cm dia, 6.25
cm long) with the appropriate amount of lithium-sulfur mixture. The capsules
were sealed and held at temperature (usually 375°C) for 200 to 600 hr. The
samples were held in a rocking furnace which rotated 180° every few minutes.
This type of furnace was used to insure relatively uniform contact between the
surface of the specimen and the lithium-sulfur mixture.
After the sample was removed from the furnace the adhering lithium-
sulfur mixture was removed with water or CS2- The weight loss or gain of the
specimen was measured. The weight loss was converted to an annual corrosion
rate which is reported in millimeters per year penetration; extrapolation is
based on an assumed linear rate. For samples gaining weight because of film
formation, extrapolation is based on an assumed parabolic rate. The exposed
samples were examined metallographically to determine corrosion effects not
apparent from simple weight-change considerations. All metal samples were also
tested for electrical conductivity after exposure, since some sulfide films
tend to be electrically insulating.
-------
- 53 -
The samples were tested in mixtures of lithium and sulfur rather than in
pure sulfur because preliminary investigations indicated that materials re-
sistant to sulfur alone are not necessarily resistant to a lithium-sulfur mix-
ture. For example, stainless steel is commonly protected from corrosion be-
cause of the presence of a surface oxide film. Oxidized Type 304 stainless
steel is not severely attacked by sulfur at 375°C because of this oxide film.
Lithium-sulfur mixtures, however,.appear to attack the oxide film and thereby
permit reaction of the stainless steel with sulfur. Therefore, corrosion
tests were performed with mixtures of 20 at. % lithium-80 at. % sulfur. This
composition was chosen as. representative of the conditions in the cathode com-
partment of a partially discharged cell.
Corrosion tests in molten lithium were normally performed as static
isothermal tests in which the specimen was immersed in molten lithium
(helium atmosphere) for 100-1000 hr at 375°C. Upon termination of the ex-
periment, the molten lithium was removed with methyl alcohol and the specimens
were analyzed as indicated above.
B. Metallic Components Studies
(M. L. Kyle, J. R. Pavlik)
Corrosion rates of the more corrosion-resistant electrically conducting
materials that were obtained in short-term and long-term screening tests in
20 at. % lithium-80 at. % sulfur at 375°C are presented in Fig. 24. Among
the materials that were tested, molybdenum and chromium had low corrosion
rates in both short-term tests of 100-300 hr and long-term tests of about
600 hr. Molybdenum, in particular, has been used in many of the experimental
cells for periods up to several hundred hours and has shown good corrosion
resistance under these conditions.
MOLYBDENUM
CHROMIUM
INCONEL<|02)
2RK65 SS
ZIRCALOY-2
347 SS
HASTELLOY-X
ALUMINUM
205 SS
I
I
AVE. RATE, I00-300hr
MAX RAT., 620hr
AVE. RATE. 620hr
ssgcgssjsssjgso
0123456
CORROSION RATE, mm/yr
Fig. 24. Corrosion Rates of Metals Exposed to
20 at. % Li-S Mixtures at 375°C
-------
- 54 -
The austenitic stainless steels, such as Types 2RK65, 347, and 205
(15 wt % Mn), had low corrosion rates in short-term tests, but much higher
rates in the longer tests. The nickel-base alloys, such as Inconels 702
and 600 and Hastelloy X, as well as Zircaloy-2, showed similar behavior.
The reason for the higher corrosion rates in the long-term tests is not
fully understood, but it may be related to intergranular attack or to the
temporary existence of a protective film.
The corrosion rate of aluminum by 20 at, % lithium-80 at. % sulfur
mixtures has proved to be highly variable, with rates ranging from 0.01 to
2.4 mm/yr in 600-hr tests. It is suspected that these variations are asso-
ciated with the nature of a surface film that is formed during the exposure.
Not shown in Fig. 24 are the corrosion rates for tantalum, iron, titanium,
nickel, and beryllium. All of these metals showed poor corrosion resistance
to lithium-sulfur mixtures and have been excluded from consideration for use
in the cathode of lithium/sulfur cells.
The metallic conductors used in lithium/sulfur cells should be both
lightweight and low cost if batteries of these cells are to have a specific
energy of at least 220 W-hr/kg. Two materials, aluminum and chromium, are
of particular interest in the current materials program since they meet the
weight and cost requirements. Molybdenum, on the other hand, is a high-density,
relatively high-cost product. It is usable in laboratory and scaled-up cells,
but its eventual replacement by a lighter, lower-cost material is desirable.
Much of the current materials program is concentrated upon methods of using
aluminum and chromium as cell components.
Three aluminum alloys (A-288, 5050, and 3004) and a ferritic chromium
steel (E-Brite 26-1) were tested for corrosion resistance to the molten
LiCl-KCl eutectic at 375°C for 113 hr and to 20 at. % lithium-80 at. % sulfur
at 375'C for 480 hr. The aluminum alloys selected were those which are re-
ported to be more resistant to chloride-ion attack.23 The ferritic stainless
steel is high in chromium content, and since chromium performed well in the
corrosion tests, it was hoped that an alloy high in chromium would exhibit
similar characteristics. The results of these tests are presented in Table XVII.
The electrolyte tests showed aluminum to be attacked at a rate of several
tenths of a millimeter per year (six-sample average, 0.67 mm/yr). The attack
was not uniform and most samples were noticeably pitted. This corrosion rate
is too high for practical cell application. The ferritic stainless steel
with the high chromium content showed excellent corrosion resistance. The
sample surface was shiny before and after test and no sign of attack was evident.
When these materials were retested in the cathode material, different
results were obtained. The three aluminum alloys, which had been attacked by
the electrolyte, had a slight gray-to-yellow coloration and were electrically
conductive™ Except for some surface roughening, the samples were almost un-
affected. The ferritic stainless steel, which was unaffected by the electrolyte,
was severely attacked in the lithium-sulfur mixture by a mechanism which ap-
peared to be dissolution.
-------
- 55 -
TABLE XVII. Corrosion of Aluminum and Ferritic Stainless Steel
in LiCl-KCl Electrolyte and 20 at. % Li-S Mixtures at 375°C
Observed Corrosion Rate, mm/yr
Material
113-hr Exposure to LiCl-KCl
480-hr Exposure
to 20 at. % Li-S
A-288 Ala
5050 Alb
3004 Alc
E-Brite 26-ld
0.65, 0.78
0.20, 1.3
0.47, 0.66
3.C x IQ-1* , 2.4 x i(T3
0.000,
0.048
0.031
0.46
0.013
Composition (wt %): 1 Ni, 0.5 Fe, 0.1 Ti, balance Al.
bComposition (wt %): 1.4 Mg, balance Al.
Composition (wt %): 1.2 Mn, 1.0 Mg, balance Al.
Composition (wt %): 0.005 C, 0.25 Si, 26 Cr, 1.0 Mo, balance Fe;
product of Airco Vacuum Metals, Berkeley, California.
The above results are taken to indicate that the corrosion resistance of
aluminum will depend to a large extent upon the presence of a surface film to
render the material inert to the electrolyte. The emphasis now will shift to
the identity and method of formation of an inert surface film. The initial
attempts will be with cathode, nitride, and sulfide films. High-chromium-
content steels will be included in tests. Corrosion tests have been screening
tests to determine the ability of various materials to withstand chemical and
thermal effects in cells for times up to 1000 hr. These tests are not defi-
nitive, however, since electrical and mass-transfer effects in an operating
cell cannot be readily simulated. The definitive materials tests must be
performed in an operating cell; this phase of the program has been undertaken.
The design of a chromium-plated cell is shown in Fig. 25. The cell
ANODE LEAD
ATHODE LEAD
CHROMIUM-PLATED
COMPRESSION PLATE
DEPTH OF
IMMERSION
IN
ELECTROLY
ELECTROLYTE-FILLED
Mo FOAM
SULFUR-FILLED
POROUS GRAPHITE
Mo SCREW
MOLYBDENUM FOAM SHEATH
LITHIUM-FILLED ANODE
Fig. 25. Schematic Diagram of a Chromium-Plated Cell
-------
- 56 -
is of a laminated-cathode design using a molybdenum foam (Spectra-Mat, Inc.,
78.2% porosity, 25-um av. pore size, 0.19 cm thick x 1.1 Cm x 1.2 cm) sheath
filled with electrolyte and porous graphite (Poco Graphite, Inc., 63% porosity,
1.4-ym av. pore size, 0.15 cm thick x 1.1 Cm x 1.2 cm) filled with sulfur.
The cell was fabricated of Type 304 stainless steel which was chromium-plated
by conventional aqueous electroplating techniques.
The initial operation of the cell, using the LiBr-RbBr electrolyte and
a stainless steel anode current collector, consisted of about 95 hr at 375°C
and 20 charge-discharge cycles at current densities from 0.06 to 0.15 A/cm2.
Cell operation was terminated because of a short circuit caused by molten
lithium floating on the electrolyte surface. Inspection of the cell showed
that nearly all the chromium plating exposed to the electrolyte had been
darkened or removed.
The cell performed smoothly during the experiment before the short circuit
caused termination. The first discharge was made at a current density of 0.15
A/cm2, and the cell delivered a capacity density of 0.14 A-hr/cm2 (27% sulfur
utilization to a final product of Li2S above a 1-V cutoff). By the third
discharge (at a current density of 0.13 A/cm2), the capacity density had de-
creased to 0.05 A-hr/cm2 (8.5% sulfur utilization) and remained at or near
this value through the twentieth discharge. The discharge curves are shown
in Fig. 26. Table XVIII presents the discharge data for the 20 cycles.
DISCHARGE 20. 0.11 A/cin2
0.5
CELL JRP-I
Li/LIBr-RbBr/LI IN S
CATHODE AREA 1.89cm'
INTERELECTROOE DISTANCE I cm
TEMPERATURE 375 *C .
THEORETICAL CAPACITY DENSITY 0.54 A-hr/cir/
CATHODE CURRENT COLLECTOR
FIVE POROUS GRAPHITE PLATES
63 X POROSITY
1.4-ftm AV. PORE SIZE
SIX MOLYBDENUM FOAM PLATES
78% POROSITY
2.3-^mAV. PORE SIZE
I
I
0.05
0.10
CAPACITY DENSITY, A-hr/cm2
Q15
10 IS 2O 25
PERCENT OF THEORETICAL SULFUR UTILIZATION
30
Fig. 26. Voltage-Capacity Density Curves for a
Chromium-Plated Li/S Cell With
LiBr-RbBr Electrolyte
-------
- 57 -
TABLE XVIII. Performance of a Chromium-Plated Li/S Cell
Temperature: 375°C
Electrolyte: LiBr-RbBr
Anode: Lithium contained in Huyck Feltmetal, Type 302 SS,
90% porosity, 67-um av. pore size
Cathode: Electrolyte elements (6)
Molybdenum foam, Spectra-Mat, Inc.,
78% porosity, 25-pin av. pore size, 0.19 cm thick
Sulfur elements (5)
Porous graphite, Poco Graphite, Inc., 63% porosity,
1.4-ym av. pore size, 0.15 cm thick
Discharge
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Current
Density,
A/ cm2
0.15
0.13
0.13
0.11
0.11
0.11
0.11
0.11
0.12
0.12
0.06
0.11
0.11
0.11
0.11
0.11
0,11
0.11
0.11
0,11
Capacity Density,
> 1 Volt,
A-hr/cm2
0.141
0.100
0.047
0.037
0.075
0.070
0.053
0.053
0.068
0.047
0.081
0.062
0.060
0.062
0.056
0.051
0.053
0.055
0.051
0.048
Sulfur
Utilization,
%
27
18
8.5
6.8
14
13
9.7
9.7
12
8.5
15
11
11
11
10
9.4
9.7
10
9.4
8.7
The cell was operated a second time with two modifications: (1) the
electrolyte was changed from LiBr-RbBr to LiF-LiCl-KCl in order to obtain
better wetting of the molybdenum foam, and (2) the anode current collector
was changed from Type 302 stainless steel (Huyck Metals Co., 90% porosity,
67-um pore size) to nickel (Huyck Metals Co., 85% porosity, 85-um pore size)
to test nickel as an anode current collector and determine its wetting pro-
perties in cell applications.
The cell operated for only a single discharge cycle, which is displayed
in Fig. 27. The discharge at 0.14 A/cm2 produced a capacity density of 0.17
A-hr/cm2 with 41% sulfur utilization (based on a 2 Li + S •> LiaS overall
cell reaction). Most of this capacity was obtained at a terminal voltage
>1.5 V. These results were considered very encouraging, but when the cell
was recharged, the lithium did not return to the Feltmetal, but floated to the
electrolyte surface. After repeated attempts to recharge the cell, its
operation was terminated.
-------
- 58 -
2.0
>
UJ
o
1.0
UJ
o
CELL JRP-2
LI/LIF-LICI-KCI/LI IN S
ANODE AREA 2.7 cm*
CATHODE AREA 2.14cm2
INTERELECTRODE DISTANCE-I cm
TEMPERATURE 375*C
THEORETICAL CAPACITY DENSITY 0.42A-hr/cmz
ANODE CURRENT COLLECTOR
NICKEL FELTMETAL. 1.5 x 1.8 x 0.5cm,
86% POROSITY. 85-/wn AV. PORE SIZE
CATHODE CURRENT COLLECTOR
5 PIECES POROUS GRAPHITE. 1.2 x 1.1 x 0.13cm,
63% POROSITY, 1.4-/Am AV. PORE SIZE
6 PIECES MOLYBDENUM FOAM, 1.2 x I.I x 0.19cm,
78% POROSITY, 25-/im AV. PORE SIZE
005 0.10 2
CAPACITY DENSITY. A-hr/cm
0.15
10 20 30
PERCENT OF THEORETICAL CAPACITY DENSITY
4O
Fig. 27. Voltage-Capacity Density Curves for a Chromium-Plated
Li/S Cell With LiF-LiCl-KCl Electrolyte
The anode Felttnetal was nickel, rather than Type 302 stainless steel used
in most cells, and had a characteristic pore size of 85 rather than 67 um.
Whether the nickel, the larger pore size, or neither of these caused the cell
failure is not known. The conclusion that nickel is an unacceptable anode
current collector should not be drawn on the basis of this single test. Other
cells using nickel Feltmetal will be operated.
The chromium-plated cell was destructively examined after the second cell
test. Metallographic examination indicated that the chromium coating on the
cathode housing was almost completely destroyed, while that on the anode
housing was present but badly cracked. Analysis of parts remote from the active
electrode surfaces showed the presence of an integral, well-bonded coating.
It is concluded that the chromium-plating operation, performed by aqueous
techniques, was done well. The coating apparently cracked from thermal stresses
during operation, and in those areas of the cell (i.e., the cathode) where the
base metal was not resistant to the corrosive environment, base-metal attack
caused coating failure. In other areas (i.e., the anode) where the base
metal was resistant, the coating remained in place. These results indicate
that chromium plating should be explored further with added emphasis on cell
design, plating temperature, and plating process, all of which are capable of
reducing the thermal stresses that probably caused the plating to fail.
-------
- 59 -
C. Seals and Insulating Component Studies
(M. L. Kyle, J. R. Pavlik)
Lithium/sulfur cells will probably require an electrical insulator re-
sistant to lithium and electrolytes and, perhaps, to cathode mixtures.
Currently BeO and BN are the only insulators which have demonstrated any
appreciable lifetime in cell service, and BN is not thought to be usable in
long-lived cells because of a corrosion mechanism which produces an elec-
trically conducting film on the insulator surface. Although BeO has operated
very well in cell service, alternative insulators are still being sought
because of the cost and health hazard associated with BeO.
The results of 300-hr tests to determine the corrosion resistance of
several potential electrical insulators to molten lithium at 375°C are shown
in Table XIX. The tests indicated that y-LiA102, high-purity A1203,
Li20 • 5A1203, Li20 • MgO spinel structures, and B1202 are not resistant to
molten lithium. These materials have been eliminated from further consider-
ation as lithium/sulfur cell insulators.
Three insulators, Y203, Y3Al5Oi2, and the MgO • A1203 spinel, were trans-
lucent before testing and darkened during testing but otherwise resisted attack
very well. The observed corrosion rates of <0.06 mm/yr compare favorably with
those of BeO and other insulators tested to date. It is interesting to note
that these insulators were fabricated by hot-pressing. Hot-pressed BeO also
appears to be superior in corrosion resistance to BeO fabricated by other
techniques.
The corrosion resistance of lithium aluminate (LiA102) also appears to
be strongly affected by its method of preparation. Specimens of a-LiA102
prepared by Foote Mineral and material compacted by McGraw-Edison from ANL-
produced a-LiA102 were both destroyed in testing. Other compacts prepared
at ANL by pressing and sintering high-purity a-LiA102 at 1600°C (y-LiA102
after sintering) were more corrosion resistant with observed corrosion rates of
0.12, 0.29, and 0.50 mm/yr in molten lithium at 275°C. The latter two samples
were retested in LiF-LiCl-KCl mixtures at 375°C to simulate the presence of
cell electrolyte and showed corrosion rates of 0.98 and 0.29 mm/yr. The
corrosion resistance of the various insulators tested in lithium is shown in
Fig. 28.
It is concluded that the purity of the material and the method of
fabrication are important variables affecting the corrosion resistance of
electrical insulators to molten lithium. Work at ANL will concentrate on
procedures for preparing LiA102- and Y203~based ceramic insulating materials
and on preliminary testing of other potential insulators, particularly
nitrides, while external sources of high-purity A1N, Y203, and BeO will be
sought to begin evaluating some of these insulators in actual cell operation.
Graf oil,'* a flexible graphite material containing no resins or fillers,
is being evaluated as a sealing material. This material has been used in
several lithium/sulfur cells and appears to provide a good seal. Grafoil
*A product of Union Carbide Corporation.
-------
- 60 -
will be evaluated more extensively. It is expected that it is corrosion resis-
tant to both sulfur and the molten-salt electrolyte but not to molten lithium.
TABLE XIX. Corrosion Rates of Electrical Insulators in Molten Lithium
Temperature: 375°C
Exposure Time: 300 hr
Sample Starting
Number Material
Sourct
Observed
Corrosion
Rate,
mm/yr
Remarks
Li-281 Y-LiA102
Li-282 Y-LiA102
Li-283 A1203 (Coram)
Li-284 A1203 (Lucalox)
Li-285 a-LiA102
Li-286 Li20 • 5A1203
Li-287 Li20 • MgO
Li-288 a-LiA102
Li-292 a-LiA102
Li-289 a-LiA102
Li-293 a-LiA102
Li-290 a-LiA102
Li-291 Y203
Li-303 B1202
Li-304 Y3A15012
Li-306 Y203
Li-307 Li20 • 5A1203
Li-305 MgO • A1203
Foote Mineral 1.9
Foote Mineral >3
Corning Glass >3
General Electric >3
McGraw-Edison >24
ANL >26
ANL 6.0
ANL 0.29a
ANL 0.98b
ANL 0.50a
ANL 0.29b
ANL
ANL
U. of 111.0
AMMRCd
AMMRC
AMMRC
AMMRC
0.12
>20
High
1.6 x 1CT2
1.6 x ID"2
0.97
5.9 x 10
-2
Severe attack
Sample destroyed
Sample destroyed
Sample destroyed
Sample destroyed
Sample destroyed
Severe attack
Sample intact,
little dimensional
change
Retest of Sample
Li-288 with salt
present;sample
friable after test
Sample intact;
little dimensional
change; some areas
green-brown
Retest of Sample
Li-289 with salt
present; sample
friable after test
Sample intact
Sample destroyed
Destroyed
Mottled gray surface
Dull black surface
film
Glaze destroyed;
sample friable
No longer translucent,
gray surface
aCorrosion rate based on observed weight change. Samples did not change
dimensions.
bExposed to Li-KCl-LiCl mixture.
cPrepared by Professor Nelson of the University of Illinois.
•^Obtained from Mr. G. E. Gazza of the Army Materials and Mechanics Research
Center.
-------
- 61 -
a. HOT-PRESSED. HIGH-PURITY
D RECRYSTALLIZED GRADE
c. COMMERCIAL GRADE
0.5 1.0
CORROSION RATE, flim/yr
I I I
120
20 40
CORROSION RATE, milt/yf
480
Fig. 28. Corrosion Rates of Electrical Insulators
in Molten Lithium at 375°C
D. Development and Fabrication of Ceramic Insulators
(D. E. Walker)
Boron nitride has been used as the cell insulator because it had shown
good short-term corrosion resistance to lithium, sulfur, selenium, and the
LiF-LiCl-KCl electrolyte and because it was readily available and machine-
able into the required shapes. In service, these insulators had lifetimes
of 200 hr or less because of the formation of a conducting film on the in-
sulator surface when dissolved lithium was present in the electrolyte.
The corrosion tests of various other ceramics discussed above indicated
acceptable corrosion resistance for materials such as BeO, Y203, and LiA102«
However, these materials are not readily available in the required shapes,
and the commercial materials were not uniform, showing varying degrees of
corrosion resistance from one lot to another.
It was believed that impurities in the grain boundaries as well as low-
density bodies with interconnected porosity were major reasons for poor cor-
rosion resistance. Therefore, the approach selected was one that would pro-
duce a high-density body without the use of sintering additives. High
density was to be obtained by sintering at a temperature as close to the melting
point as possible.
-------
- 62 -
The ceramic materials studied were LiA102, Y203, and 90 wt % Y203-10 wt %
Eu203. The bulk of the work was concentrated on LiAlC>2.
1. Lithium Aluminate
The LiA102 powders used in this work derived from two basic batches.
One batch was made by calcining a mixture of A1203 and Li2C03 at about 700°C
to form LiA102 by the reaction:
A1203 + Li2C03 + 2LiA102 + C02t
This powder contained excess Li2C03, which was not considered troublesome for
high-temperature sintering processes because it was expected to decompose and
volatilize during sintering. The bulk of this material (JK-2) was washed to
remove the excess Li2C03 and vacuum dried. The second material was a commercial
product from Foote Mineral Company. This material was dry ball-milled in
alumina equipment before use to reduce the particle size for better handling
characteristics. The powders were characterized as indicated in Table XX.
TABLE XX. Characteristics of LiA102 Powders
Surface Average
Area, Density,3 Particle
LiA102 Source m2/g g/cm^ Size, ym Structure13
Remarks
CEN-1
Used as
received
CEN
JK-2
Foote
Foote
Mineral
Mineral
17
3
7
.15
.18
.33
2
2
2
.73
.79
.82
25%
77
66
<44
.4
.1
Y-LiA102b
a-LiA102
Y-LiA102b
a-LiA102
Washed
vacuum
and
dried
As-received
Ball-milled
34 hr
aPowder density determined by an air pycnometer and is higher than the
handbook value of 2.55 g/cm3 at 25°C.
^Indicates the major constituent.
2. Lithium-Aluminate Sintering Studies
Wafers (approximately 2.5 cm dia, 0.2 cm thick) were used to develop
the best sintering conditions for the LiA102 ceramic bodies. These wafers
were cold-pressed in a steel die at pressures from 6.3 x 106 to 9.1 * 106 kg/m2,
The melting point for the 1:1 Li20-Al203 compound has been reported24 as
1700 + 15°C. A eutectic exists at 45 mol % Li20 with a melting point of
1670 + 15°C. These values are in agreement with our work. The as-received
LiA102 containing excess Li2C03 could be sintered at 1675°C without melting,
while the washed batch, JK-2, melted at this temperature. It was apparent
that the higher-melting LiA102 was stable at the 50 mol % composition during
sintering because of the presence of Li20 from decomposition of the excess
Li2C03. This Li20 compensated for the usual Li20 volatilization loss from the
compound at elevated temperatures. The washed LiA102 did not have an excess
of Li2C03 present, and the normal loss of Li20 during sintering moved the
composition towards the eutectic mixture and the lower melting temperature.
-------
- 63 -
Small amounts of an organic binder and lubricant had to be added to
the powders in order to produce a pressed body with sufficient strength to with-
stand removal from the die and necessary handling prior to sintering. Benzene
solutions of stearic acid + polyvinyl alcohol, stearic acid + carbowax, stearic
acid + acroloyd, and a commercial binder, Mobilcer-R (produced by Socony Mobile
Oil Company), were evaluated. The Mobilcer-R binder is a water emulsion of
microcrystalline wax. These binders were incorporated into the LiAlC>2 powder
by mixing a slu::ry of powder and binder emulsion. The benzene solutions were
added at the rate of 2 cm solution per gram of powder. The water emulsion
was added to the powder in the amount of 1 cm3 per gram of powder. After the
blended mix dried, the powder cake was crushed and sieved through a No. 20
sieve (U. S. sieve series). The best results were obtained with Mobilcer-R
added in the amount of 3-5 wt % wax.
Sintering was accomplished in two steps„ The first, in air, con-
sisted of heating the pressed samples slowly to 300°C, holding at this tem-
perature until the organic addition had been driven off, and then continuing
to heat to 700°C. The higher temperature completed the organic removal step
by oxidizing the residual carbon to C02, which diffused easily out of the
low-density green compacts. The sintering step was accomplished in a helium
atmosphere in a high-temperature, resistance-heated, tungsten, cold-wall
furnace. The sintering temperature was originally set at 1675°C but was
lowered to 1600°C to avoid melting the samples.
The disks produced during this study ranged in density from 1.84
to 2.50 g/cmd or from 72 to 98% of the theoretical density of 2.55 g/cm3.
Table XXI lists the characteristics of the LiA102 ceramic bodies produced.
Very little work was done on producing Y203 ceramic bodies for in-
sulators because the LiA102 ceramics appeared to be satisfactory. However,
the same criterion of high density for corrosion resistance was applied and
several samples were made. Yttrium oxide powder of 99.99% purity from Con-
solidated Astronautics, Inc., was pressed in a 2.5-cm steel die. These wafers
were sintered in helium at 2175°C. The samples attained a moderate density
of 89% of the theoretical 5.03 g/cm3. A second group of samples was made
from a mixture of 90 wt % Y203~10 wt % Eu203- This mixture was prepared with
the intention of utilizing the lower melting temperature of the Eu203 (2050°C)
to provide liquid-phase sintering, which generally promotes the formation of
high-density bodies. It was expected that the resulting ceramic would be a
single-phase solid solution and as corrosion resistant in the cell environment
as the Y203. Six 2.5-cm dia disks were pressed and sintered at 2200°C in
helium. The results were inconclusive because trouble with the furnace con-
trols caused rapid cooling from 1400°C. The wafers were badly cracked. Eval-
uation of one sample indicated a density of 4.99 g/cm3, which is 94.7% of the
calculated theoretical density for this mixture. The sintered material appear-
ed grey-brown in color and was translucent,
3. Cell Insulator Production
One type of insulator being considered for the lithium/sulfur cells
is a ceramic ring 73.6 mm OD * 64.6 mm ID * 0.79 mm thick. Procedures were
developed to press and sinter this ring to the required diameter. The thick-
ness was obtained by grinding the sintered ring to size.
-------
- 64 -
TABLE XXI. Characteristics of Sintered LiA102 Ceramic Bodies
Sample Material
No. Source
Sintering Sintered Percent Sintering
Weight Density, Theoretical Temp.
Loss, % g/cm3 Density, % °C
Remarks
Al- 1
- 2
- 3
- 4
Al- 9
-10
CEN-1
CEN-1
CEN-1
CEN-1
CEN-1
CEN-1
3.8
3.6
3.6
3.7
1-5
1.8
2,43
2.48
2,43
2.43
2,49
2.50
95
97
95
95
97.6
98.0
1675
1675
1675
1675
1675
1675
No organic
additions; no
pre-f iring
in air ,
Organic bind-
er, air-fired
to 700°C
Al-11
-12
-13
-14
AL-28
-29
-30
AL-34
-35
-36
CEN JK-2
CEN JK-2
CEN JK-2
CEN JK-2
Foote
Mineral,
as-received
Foote
Mineral,
ball-milled
10 hr
0.76
1.33
1,98
0.78
1.77
1.57
1,50
1.60
1.18
1.21
2.21
2.00
2 ,,20
2.21
1.84
Io85
1.85
2.33
2,34
2.36
86.6
78.4
86.3
86.6
72.16
72.55
72.55
91.37
91.76
92.55
1630
1630
1630
1630
1600
1600
1600
1600
1600
1600
Organic bind-
er, air-fired
to 700°C
Organic bind-
er, air-fired
to 700°C
Organic bind-
er, air-fired
to 700°C
The rings were pressed from powder in an isostatic press using a
rubber die, such as that shown in Fig. 29. A pressed ring is shown in place
in the die. The metal part on the right is the plug that seals the rubber die
cavity from the working liquid in the isostatic press. In use, dry powder
containing binder is poured into the ring-shaped portion of the die cavity and
leveled wich the surface. The metal plug is pressed into the die cavity until
it bottoms against the powder-filled ring. This assembly is then placed into
a steel support and lowered into the fluid in the isostatic press chamber.
The press chamber is closed and the fluid is pressurized to the desired level.
The pressure release rate is automatically controlled by means of a micrometer
adjustment on the pressure-relief valve. After isostatic pressing, the working
fluid is washed off the rubber die assembly and the metal plug is removed. The
pressed ring is then removed. Figure 30 shows a group of five pressed rings
after the 700°C binder removal step. The sintering procedure was the same as
that used for the development samples. After burning out the binder, the rings
were placed in a tungsten, resistance-heated, cold-wall furnace and heated to
1600°C in a helium atmosphere. The heating and cooling schedule was automati-
cally controlled by means of a Data-Trak control unit to provide an 8 1/2-hr
heat-up to 1600°C, 1/2 hr at temperature, and a 9-hr cool down.
A total of 46 rings have been pressed and sintered during this devel-
opment. Twelve of the rings have been ground to produce flat, parallel faces
and the desired thickness. One of the 12 rings broke during grinding, and an-
other had a visible crack.
-------
- 65 -
» i - « r s g io 11 ja ia 14 13 i« ir 11 i• i* 11 aa
Fig. 29. Photograph of a Rubber Die and Sealing Plug
Fig. 30. Photograph of Isostatically Pressed
LiA102 Insulator Rings
-------
- 66 -
In summary, the development of ceramic insulators is proceeding
satisfactorily. Lithium aluminate insulator rings have been produced and will
be tested in operating cells. Development work will now concentrate on the
methods for producing Y203~, Bed-, and AIN-based ceramics and testing chese
pieces in operating cells. In addition, this task will be broadened to in-
corporate the development of ceramic insulators and feedthroughs which may be
required in a battery,
E, Cathode Current Collector Development
(R, Rubischko)"
A modification of the cathode design of Fig. 15 is being investigated-
The purpose of this modification is to eliminate the molybdenum foam sheath-
ing that prevents sulfur from dispersing into the electrolyte. An effort is
currently being pursued to achieve a cathode structure of porous graphite into
which chromium has been vapor-deposited so that the surfaces of the large,
interconnected pores are coated with chromium. This structure will be par-
tially filled with sulfur, which is expected to be located in small uncoated
pores and the remaining volume will be filled with electrolyte, which is ex-
pected to wet the chromium-coated surfaces * The potential advantages are
twofold. First, the electrolyte-filled pores are expected to provide a means
of containing the sulfur without resorting to the porous metal elements in
the cathode. Secondly, the number of sulfur-electrolyte-current collector
reaction sites is expected to be greatly increased, which might result in
greater sulfur utilization.
Fabrication of such a cathode structure has been attempted by thermally
decomposing dicumene-chromium (DCC) vapor as it passed through porous graphite
with a heated carrier gas. The apparatus consisted of an argon flow control
unit, an argon heater, a DCC vapor generator, a quartz reaction tube, and an
after condenser. A discussion of this technique of vapor deposition of chromium
i-\ f-
can be found in the literature, 3 It appears that dicumene-chromium,
Cr(CgH5CH(CH-,)2)2> is one of a class of organo-chromium compounds which can be
utilized. Other materials which may be applicable are dibenzene chromium,
Cr(C^HG)2, and chromocene, Cr(C5H5>2> The technique may also provide a method
for producing chromium-plated aluminum or aluminum alloy cell housings.
A sample of Union Carbide "Pyrofoam" Graphite, Grade FPA-20, was par-
tially plated with chromium utilizing the dicumene-chromium vapor decomposition
technique. The sample was not chromium-plated as desired because of non-uniform
heating or the sample and leakage of the vapor past the porous graphite sample.
The test did, however, demonstrate the feasibility of the technique. Micro-
scopic examination of the sample and a mounted section of the sample revealed
chromium deposits on the forward surface of the graphite and on surfaces of
approximately 20% of the pores throughout the porous structure.
Changes have been made in the equipment design to remedy the above defi-
ciencies. Samples are now being coated, and process control is of immediate
concern. Suitably coated samples will be included in future cell tests to
establish the performance characteristics of these current collectors.
-------
V. ELECTRIC AUTOMOBILE PERFORMANCE CALCULATIONS
Calculations of the performance characteristics that might be obtained
from an electric automobile powered by a lithium/sulfur battery have been
made. These calculations, entailing extrapolation of laboratory cell data to
full-sized batteries, are useful in that they place the cell performance now
achievable in perspective relative to the intended application, and help to
point out areas in which increased cell performance is necessary to achieve
the ultimate aims of the program.
A. Cell Design
(R. Rubischko)
Early in the development of an electrochemical system, it is often useful
to consider those problems associated with transferring the technology from
small laboratory test cells to practical large-scale cells. An understanding
of the ways in which the specific energy and specific power of large-scale
cells are affected by variations of parameters in the laboratory cells permits
investigators to design more meaningful experiments. A first approximation to
large-scale cell performance is obtained from linear extrapolation of cell per-
formance data (on the basis of identical performance per unit area of electrode)
from labora.ory cells to cells of a practical size for automobile propulsion.
A scaled-up lithium/sulfur cell was analyzed to determine the effects of
cell size (area and thickness), materials, electrolytes, and configuration upon
the specific energy of a battery. The configuration that was evaluated is not
intended to be considered as a final design but rather to serve only as a
means of evaluating the manner in which design can affect performance. The
purpose of these studies was to estimate the values of the design parameters
necessary to achieve a sealed cell with a specific energy in excess of 220
W-hr/kg and to provide quantitative data relating cell parameter variation to
specific energy.
The assumptions used in the specific energy calculations were a sulfur
utilization of 50% and an average cell voltage of 1.8 V. The electrolyte was
assumed to be LiCl-KCl eutectic with a density of 1,67 g/cm3 at 400°C. The
overall cell reaction was assumed to be 2Li + S ->• Li2S. The densities of
lithium, sulfur, and Li2S at a cell temperature of 400°C were taken as 0.49,
1.64, and 1.6 g/cm3, respectively. Thus, since 0.433 g of lithium reacts
with each gram of sulfur, 1.449 cm3 of lithium reacts for each cubic centimeter
of sulfur utilized. The battery weight on which the specific energy was based
did not include outer battery cases, battery connectors, or outer thermal
insulation. It did take into consideration the cell heaters and thermal in-
sulation around the individual cells.
The design was that of a square cell with a laminated cathode, as shown
in Fig. 31. The square shape was chosen since it is expected that fabrication
of a cell with a laminated cathode would be simplest with this shape. The
cathode was located below the anode so that, as the lithium is consumed, the
electrolyte would remain in position. The anode housing material was assumed
to be 0.25-r™-thick stainless steel, and the cathode housing material was
assumed to be an aluminum alloy of 0.5-mm thickness. The anode current col-
lector was taken to be 95%-porous stainless steel, and the laminated cathode
-------
- 68 -
was assumed to be composed of 1-mm-thick elements of 91%-porous graphite filled
with sulfur and 90%-porous aluminum alloy filled with LiCl-KCl. The specific
energy of the cell was calculated as a function of cathode thickness and elec-
trode area. The effect of electrode area upon specific energy is not large for
cells of electrode area greater than 400 cm . Beyond this size, the increase in
specific energy would not be expected to be balanced by the increased diffi-
culties encountered in the use of the larger components. Increasing the cathode
thickness increases the specific energy at a decreasing rate. If we consider
a cell of 350-cm2 electrode area and 0.5-cm cathode thickness, a specific
energy of 204 W-hr/kg is expected. A cell having the same electrode area and
a cathode thickness of 1.2 cm would have a specific energy of 234 W-hr/kg
(an increase of
€
INSULATION,
HEATER SUPPORT,
347 SS, 25 mm
SL ASS AND
GRAFOIL SEAL,
0.3cm, 2.5g/cc
NODE
95% POROSITY SS
w/LITHIUM
LICI-KCI
T 0.05 cm
90% POROSITY
Al ALLOY, I mm
91% POROSITY
GRAPHITE, I mm
•/SULFUR
ALLOY, 0.50 mm
0.4 0.8 1.2
CATHODE THICKNESS, cm
1.6
0.25 0.50 Q75 ,1.00
CAPACITY DENSITY, A-hr/cni
Fig. 31. Specific Energy and Schematic Design of
a Li/LiCl-KCl/Li in S Square Laminated
Sealed Cell
It is concluded from the design calculations that the cell housing mate-
rials must be on the order of 0.25-mm-thick stainless steel or 0.5--mm-thick
aluminum since the use of heavier case materials would reduce the specific
energy. It is desirable to investigate chromium plating of aluminum and mag-
nesium to combine their low densities with the excellent corrosion resistance
of chromium. It is also apparent that electrolyte elements in the cathode
must have a density not much greater than that of aluminum. The use of an
aluminum alloy which is adequately corrosion resistant to the cathode material
and products would be desirable. High-porosity graphites such as Union
Carbide Corp., Grade FPA-20, which is 91% porous with a 26-80 urn pore size,
should be evaluated.
-------
B.
- 69 -
Method of Calculating Automobile Performance
(M. L. Kyle, E. J. Cairns)
Various estimates of the performance of electric automobiles are avail-
able, -6 but they commonly assume a certain deliverable battery energy, in-
dependent of the manner in which the automobile is driven. This assumption
is incorrect for battery-powered automobiles because the deliverable energy
of any battery xs a function of the operating power level of the battery. In
practice, the power level is related to the mode of driving (commonly termed
the driving profile), and consequently must be incorporated properly in the
vehicle range calculations. The effect of the details of the driving profile
on the vehicle range is significant. For example, our calculations estimate
a range of 530 km (329 miles) when a specified vehicle is operated at a con-
stant velocity of 45 km/hr (28 mph) . The range of this same vehicle drops to
about 158 km (98 miles) on a driving profile typical of urban driving condi-
tions with frequent stops, starts, low-speed driving, and accelerations. Re-
generative braking, if used to charge the battery, could increase the vehicle
range appreciably.
The calculations of expected vehicle performance reported here have
eliminated the assumption of constant delivered specific energy. The vehicle
parameters (weight, accessory requirements, frontal area, etc.) were based on
OAP recommendations. The vehicle characteristics used for this study are pre-
sented in Table XXII. The test weight of the vehicle is 2086 kg (4600 Ib).
TABLE XXII. Electric Automobile Characteristics3
Curb weight of vehicle
Vehicle test weight (W)
Weight of power plant (included in curb weight)
Motors
Controls
Transmission and drive train
Air conditioner and heater
Miscellaneous
Battery
Total
Ee)
Air drag coefficient
Frontal area of vehicle (Af)
Efficiency (battery output to wheels, EJJ,
Total accel./linear accel.
Air density (pA)
Accessory power
None
Lights, windshield wipers, blower, etc.
Above plus air-conditioning
1950 kg
2086 kg
4300 Ib
4600 Ib
68-0 kg
90.7 kg
31.8 kg
36.3 kg
11.3 kg
487.6 kg
725.7 kg
150 Ib
200 Ib
70 Ib
80 Ib
25 Ib
1075 Ib
1600 Ib
0.5
2.23 m2 24 ft:
0.82b
1.1
9.6 x 1CT9 kg-hr:/m4
0 kW
0.250 kW
2.983 kW
0.33 hp
4 ho
aThe assumptions and estimates used in the electric vehicle calculations
do not necessarily represent the opinion of the Office of Air Programs.
^The transmission efficiency of 0.82 is probably optimistic for urban and
suburban driving conditions; data for efficiencies under various driving
cycles are not now available but should be in the near future.
-------
- 70 -
The complete power plant weighs 726 kg (1600 Ib) and includes a 488-kg (1075-lb)
battery. A breakdown of the components of the power plant was made on the basis
of information in the literature.27'28 Performance estimates were made at
three different accessory power levels: 0 kW, 0.250 kW (sufficient to operate
service lights, windshield wipers, radio, and other normal automobile acces-
sories exclusive of air-conditioning), and 2.983 kW (which includes air-condi-
tioning) .
The conceptual Lattery design used in these calculations is composed of
498 cells of the type depicted in Fig. 32. The square cell is of the enclosed
laminated cathode design which hat, performed well in laboratory experiments.
The cathode compartment is 0.5 cm deep, and both the sulfur and electrolyte
elements are 1 mm thick. The sulfur elements consist of 91%-porous graphite
filled with sulfur and the electrolyte elements are 90%-porous low-density
metal (e.g., an aluminum alloy) filled with the LiCl-KCl eutectic salt mix-
ture. The lithium anode of the cell is impregnated in a 95%-porous stainless
steel current collector. Each cell is approximately 20 cm (7.9 in.) square,
has a projected active area of 350 cm2 (54 in.2), weighs about 0.91 kg (2 Ib),
and is 1.33 cm (0.5 in.) thick. In the battery, the cells are connected in
series-parallel and contained in an insulated 46 * 67 * 117 cm case (18 * 26 *
46 in.) having a volume of 0.36 m3 (12.7 ft3).
1.344cm
).7cm
0.025 cm SS
GRAFOIL GASKETS
INSULATOR
i-
r
95% POROUS SS FILLED
LI AND LiCI-KCL
LiCI -KCI ELECTROLYTE
0.694 cmV
i I J\
0.050 HOUSING
91% POROUS GRAPHITE
FILLED WITH SULFUR
0 1 cm
90% POROUS METAL
FILLED WITH LiCl-KCl
Fig. 32. Conceptual Li/S Cell Design Used in
Automobile Performance Calculations
A cost analysis of the conceptual cell and battery design was performed
to estimate both the current materials cost for one of these batteries and
the anticipated materials cost if mass-production techniques were employed in
production of the battery components. This analysis is also useful for iden-
tifying the battery components that need significant cost reductions to pro-
duce a battery with an overall cost near $2.20/kg ($1.00/lb). (The existing
cost goal of $10/kW-hr of electrical storage capacity is equivalent to $2.20/kg
if the battery specific energy is 220 W-hr/kg.) The results of these materials
cost estimates are shown in Table XXIII.
-------
- 71 -
TABLE XXIII. Weight and Material Cost Breakdown
for Conceptual Cell and Battery Design
Unit Cost, $/kg
Cost per CclJ,
Material
Weight, kg Present Projected Present ProjecliV.
Cell components
Electrical insulator
Anode current collector
Electrolyte
Anode material
Cathode current collector
Housing
Seal
Cathode material
Total cell cost
BN
SS or Ni
LiCl-KCl
Li
Porous Graphite
Treated Al & SS
Treated Al & SS
S
0
0
0
0
0
0
0
0
.062
.090
.335
.053
.Oi9
.145
.054
.122
77.
30.
4.
19.
11.
3.
3.
0.40
a
3.
2.
10.
3.
2.50
b
0.40
S 4
2
1
1
0
0
0
0
$11
.77
.70
.34
.01
.54
.44
.16
.05
.01
$0,
0.
0.
0,
0.
0,
b
0,
S2.
.30
,27
,67
.53
,15
,36
.05
,33
Material
Unit Cost, S/kg Cost per Battery, $
Weight, kg Present Projected Present Projected
Battery components
498 cells
Inner can
Outer can
Thermal insulator
Supports, connectors, etc.
Feedthroughs
Total battery cost
-
Al
Al
Various
Various
Y20, & Al
453.2
6.1
6.8
12.0
7.5
2.0
11.01C
0.14
0.14
50.
2.00
50.00
2.33C
0.14
0.14
25.00
2.00
6.00
$5482.
1.
1.
600.
15.
100.
$6199.
$1160.
1.
1.
300.
15.
12.
$1489.
aReplaced by feedthrough.
°Not required if feedthrough is used.
cCost per cell.
The goal for the cost of the conceptual cell used in the vehicle cal-
culations is about $2.00. The present cost for materials only is about $11.00,
and in mass production it would approach $2.33 if the cell electrical insulator
used in this design were replaced by an electrical feedthrough. In order to
reach the cost goal, further cost reductions will be required, associated pri-
marily with cell design, in such areas as reduction in the quantity of electro-
lyte in each cell and replacement of the anode current collector with a less
expensive porous current-collector material. These changes appear reasonable
since the conceptual cell described above was designed from the current state
of cell development, and advances in both design and performance are antici-
pated.
The calculation of vehicle performance was based upon the requirements
for the battery to provide power levels sufficient to overcome tire-rolling
resistance (friction between the tires and the road surfaces) and aerodynamic
drag, to provide power for hill climbing and acceleration and to operate ac-
cessories.
The retarding effect of the rolling resistance was calculated by the fol-
lowing expression:
-------
- 72 -
65
where
= r [1 + (1-3 x iQ-3 v) + (1Q-5 V2)]*
Rr = rolling resistance, kgf
W = weight of fully loaded vehicle,
V = vehicle velocity, km/hr
Aerodynamic drag resistance was approximated by the following expression:
R =
w
where
^ = wind resistance, kgf
PA = air density, kg^-hr^/m4
Op = air-drag coefficient, dimensionless
Af = frontal area of vehicle, m2
V = velocity of vehicle, km/hr
The grade resistance, the downhill component of vehicle weight, was cal-
culated from
Rg = W sin 6 (3)
where 6 is the angle of the grade.
The acceleration resistance is represented by the following relation:
*a = f£
where
g = gravitation constant, km/hr2
t = time, hr
The acceleration required was obtained from the various driving profiles used
in these calculations. The above relation applies only to linear acceleration.
Some parts of the vehicle (such as the wheels, driveshaft, and motor) require
rotational acceleration that is not accounted for in the above relation.
Therefore it was assumed that an additional 10% of the value for linear ac-
celeration is necessary.
The power that must be provided to overcome the various forces can be
obtained by multiplying the sum of these forces by the vehicle velocity:
*Derived from an CAP suggested equation by conversion from English to metric
units.
-------
- 73 -
Pr = V(Rr + Rw + R + 1.1 R
a'
This power requirement represents the power that must be delivered to the
wheels and does not include losses resulting from electrical and mechanical
inefficiencies nor the power required to operate the accessory load. Conse-
quently, the power that must be delivered by the battery is
?r
where
Pb = required battery power, kW
Pa = accessory power requirement, kW
EJJJ = mechanical efficiency
Ee = electrical efficiency of the drive system
Eae = electrical efficiency of accessories
The range of an automobile depends upon the design, the driving profile
that is assumed, and the energy-storage capability and operating power level of
the battery. To estimate the battery performance, data on the voltage-capacity
density relationships (taken from constant-current discharges) from several
laboratory cells were combined and fitted by the empirical equation:
V = 2.18621603 - 5.97476987 i + 13.48490411 i2 +
0.04421195 q - 595.18189456 iq + 867.25499666 i2q +
10116.62994535 iq2 + 17990.37652883 i2q2 -
39707.09967978 iq3 - 1231165.26761080 i2q3
This equation, which is quadratic in i (current density) and cubic in q
(capacity density), relates V (voltage), q, and 1 for the laminated-cathode
cell design. This relationship and some of the data used in its derivation
are plotted in Fig. 33 where V is a function of q at various values of i. It
should be noted that the electrode area used in calculating current densities
and capacity densities in this case is the actual interfacial area between the
electrolyte and sulfur-bearing elements of the laminated cathode and not the
projected area of the cathode that faces the anode. The interfacial area was
used because a better correlation of the data from several cells was obtained,
as might be expected.
The equation represented by the curves in Fig. 33 was used in the cal-
culation of the range of the vehicle. Each driving profile was divided into
1- to 15-sec segments, and the power and energy requirement for each segment
was calculated. The battery current necessary to provide the desired power
level was determined from the empirical equation. The energy requirement was
the power level multiplied by the time interval. The capacity density (q) of
the battery was incremented by this amount before the next interval was cal-
culated. The range of the vehicle was then computed as the distance traveled
before the battery was unable to deliver sufficient power for the next profile
-------
- 74 -
interval. For the purpose of these calculations, individual cell voltages in
the range of 2.2 to 1.0 V were considered acceptable.
Li/ LiBr-RbBr/Li in S
375*C
EMPIRICAL FIT
CURRENT DENSITY
A/cm2
A 0.095
A 0.094
V 0.091
• 0.049
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
CAPACITY DENSITY, A-hr/cm*
Fig. 33. Voltage-Capacity Density Data and Empirical Curves
for Li/S Cells With Laminated Cathodes
In most cases, a significant amount of energy remained in the battery at the
range reported. This energy was available at the required power levels; however,
only at cell voltages below 1 V. This lower level cf acceptable cell voltage
was chosen because it is thought that a practical control system would not oper-
ate over a range of input voltages greater than about a factor of two. Some
of this variation could be eliminated by parallel-series switching circuits in
the control system, but this degree of complexity is beyond the intent of the
current calculations.
The driving profile is an important determinant of vehicle range. Eight
different profiles were used in the calculations to estimate the vehicle range
under a variety of driving conditions. Four of the profiles (Urban I; Federal
Register, July 15, 1970, and Nov. 10, 1970; and Public Health Service) represent
urban driving conditions. Two are suburban driving profiles, and two are high,
relatively constant-velocity profiles. The velocity and power requirements* of
the profiles are shown in Fig. 34. The driving profiles differ in the number of
starts and stops in a given time period, in the length of time at selected ve-
hicle velocities, and in the acceleration required to complete the profile. These
various driving conditions are reflected in different battery requirements of
average power and peak power.
*The power requirements were based on a vehicle test weight of 4600 Ib and an
accessory power requirement of 2983 W.
-------
- 75 -
VELOCITY PROFILES
300 600 900 1200
URBAN DRIVING PROFILES
URBAN I
FEDERAL
REGISTER
7/15/70
FEDERAL
REGISTER
11/10/70
PUBLIC
HEALTH
SERVICE
SUBURBAN DRIVING PROFILES
POWER PROFILES
300 600 900 1200
80
SUBURBAN I
SUBURBAN H
ww
300 600 900 1200
300 600 900 1200
80
60
40
20
0
60
40
20
0
CROSS-COUNTRY DRIVING PROFILES
J
r
'/
ZOH
n| i i , i , i i i i
120 [
IOO
60
40
.
n
"1
i i i i i i i t i
CROSS-COUNTRY I
n
/ :
i
i
CROSS-COUNTRY JL
1
-
-
i 1 1 1 1 >
60
40
20
120
100
fiO
40
20
0
300 600 900 1200
TIME, sec
300 600 900 1200
TIME, StC
Fig. 34. Driving Profiles With Corresponding Power
Profiles for the Automobile of Table XXII
-------
- 76 -
C. Results and Discussion
(M. L. Kyle, E. J. Cairns)
The results of the calculations are shown in Table XXIV. The effect of
accessory power requirements and driving conditions on the range of the vehicle
for three of the driving profiles depicted in Fig. 34 are shown in Fig. 35.
Under urban driving conditions averaging around 32 km/hr (20 mph) and for an
accessory load of 2.983 kW (4 hp), the vehicle range is 158 to 248 km (98 to
154 miles) depending upon the exact driving conditions. At the lower ac-
cessory load of 0.250 kW, the range increases to around 322 km (200 miles).
Urban driving requires peak power levels of 99 to 150 W/kg (45 to 68 W/lb);
the battery delivers 120 to 195 W-hr/kg (54 to 89 W-hr/lb) with the greater
energy capability and longer ranges obtained at the lower power levels.
Suburban driving averages around 64 km/hr (40 mph) and requires higher
peak power levels of up to 162 W/kg (73 W/lb). The higher power levels re-
duce the delivered specific energy of the battery (at the 2.983-kW accessory
power level) to as low as 97 W-hr/kg (44 W-hr/lb) and the vehicle range drops
to 142 to 206 km (88 to 128 miles).
Constant-speed driving, even at the higher velocities of 96 to 113 km/hr
(60 to 70 mph), id not as demanding on the battery as some of the other profiles
because the average power level is lower. Once the initial power required to
accelerate the vehicle to its velocity has been satisfied, the power require-
ment drops off so that, at a constant velocity of 113 km/hr (70 mph) and an
accessory load of 2.983 kW, the range of the vehicle increases to 277 km
(172 miles). At lower constant velocities of 32 to 40 km/hr (20 to 25 mph),
the vehicle's range is over 483 km (300 miles). The effect of (constant)
velocity on vehicle range is shown in Fig. 36 for constant velocities of 16
to 113 km/hr (10 to 70 mph) and a 2.983-kW (4-hp) accessory load. The range
tails off rapidly at velocities above 88 km/hr (55 mph) primarily because of
the increased retarding effect of wind.
The possible benefits obtainable from regenerative braking were investi-
gated briefly. The automobile has a calculated range of about 187 km (116
miles) when driving according to the Urban I driving profile. This same ve-
hicle would have a range of about 232 km (144 miles), a 24% increase, if 50%
of the energy dissipated during braking could be recovered. A similar auto-
mobile, driven according to the Federal Register, Nov. 10, 1970, urban driving
profile, increases about 12% in range from 193 km (120 miles) to 217 km (135
miles) if regenerative braking is employed. The calculations suggest that re-
generative braking should be considered for any electric automobile powered
by a lithium/sulfur battery. Regenerative braking is attractive because of
the lithium/sulfur battery's ability to accept recharge at a rapid rate. The
increased range obtained is significant even if some weight must be added to
the vehicle to provide the regenerative-braking capability.
The conclusions that can be drawn from these calculations are that
lithium/sulfur cells possess the potential for powering an all-electric vehicle.
Laboratory cells have been shown to have a sufficient power density to ac-
complish all driving profiles tested. Energy storage, however, should be im-
proved to increase the range of the vehicle. Two different methods of increasing
the energy storage are available, namely, (1) increasing the capacity density
-------
TABLE XXIV. Electric Automobile Performance Under Selected Driving Profiles3
Driving Profile
(Average Velocity)
Federal Register
Nov. 10, 1970
(19.5 mph)
Federal Register
July 15, 1970
(19.6 mph)
Urban I
(15.7 mph)
Public Health
Service
(23.9 mph)
Suburban I
(41.3 mph)
Suburban II
(40.3 mph)
Cross-Country I
(70 mph)
Cross-Country II
(67.2 mph)
Average
Energy/
Ton-Mile,
kW-hr
0.18
0.18
0.26
0.18
0.18
0.26
0.21
0.22
0.31
0.20
0.21
0.27
0.20
0.20
0.23
0.21
0.21
0.25
0.25
0.26
0.28
0.24
0.24
0.26
Accessory
Power ,
W
0
250
2983
0
250
2983
0
250
2983
0
250
2983
0
250
2983
0
250
2983
0
250
2983
0
250
2983
Delivered
Specific
Energy,
W-hr/kg
157
156
146
132
131
120
185
185
170
215
213
195
105
102
97
164
162
148
256
254
225
219
221
240
Peak
Power
Interval,
sec
1
1
1
1
1
1
4
4
4
7
7
7
1
1
1
14
14
14
10
10
10
5
5
5
Peak
Power ,
W/kg
128
129
135
142
142
149
112
113
120
99
100
107
154
155
162
121
122
129
145
146
152
262
262
269
Peak
Power
Density,
W/cm5
0.36
0.36
0.38
0.39
0.40
0.41
0.31
0.31
0.33
0.28
0.28
0.30
0.43
0.43
0.45
0.34
0.34
0.36
0.40
0.40
0.42
0.73
0.73
0.75
Capacity
Density,
A-hr/cm*
0.22
0.22
0.20
0.18 .
0.18
0.16
0.27
0.27
0.24
0.32
0.32
0.29
0.15
0.15
0.14
0.29
0.29
0.22
0.40
0.40
0.36
0.32
0.33
0.37
Range
miles
187
179
120
158
150
98
186
179
116
225
217
154
114
109
88
169
164
128
213
209
172
193
193
193
km
301
288
193
254
241
158
299
288
187
362
349
248
183
175
142
272
264
206
343
336
277
310
310
310
aThe transmission efficiency of 0.82 used in these calculations is probably optimistic for the urban
and suburban driving profiles; data for efficiencies under various driving cycles are not now avail-
able but should be in the near future. An approximate correction for the listed ranges for small
variations of the transmission efficiency can be estimated from Rg = (3.75E - 2.1)R82 where RE = range
of vehicle at transmission efficiency E, km; E = transmission efficiency, fraction; Rs2 = range of ve-
hicle at transmission efficiency of 0.82, km.
-------
- 78 -
I
O
UJ
300
200
100
0
ACCESSORY LOAD -^
Y////A 2983 W
F!v!Tl 250W
—
™
1
F^-ai OW
^
— -r
'
••
1
i
XSjfil —
m
•'.-'.
:
-
200
ISO
w
i
3 Oi O
0 0
VEHICLE RANGE,
URBAN SUBURBAN CONSTANT
DRIVING DRIVING VELOCITY
31.4 km/hr 65.0 km/hr 113 km/hr
(19.5 mph) (40.3 mph) (70 mph)
Fig. 35. Electric Automobile Ranges for
Selected Driving Profiles
EFFECT OF VEHICLE VELOCITY ON RANGE
z
500-
400
300
200
100
ASSUMPTIONS
CONSTANT VELOCITY
VEHICLE TEST WEIGHT 4600 Ib
BATTERY WEIGHT 1075 Ib
AIR DRAG COEFF. 0.5
FRONTAL AREA 24ft*
EFFICIENCY 0.82
ACCESSORY POWER 2983 W
I
I
I
I
20 40 60 80
VELOCITY, km/hr
1,1,
100
I
350
300
250
200 I
UJ
150 2
100
50
120
10 20
30 40 50
VELOCITY, mph
60 70
80
Fig. 36. Effect of Automobile Velocity on Range
for Constant Velocity Driving
-------
- 79 -
and capacity per unit volume of the cells and (2) increasing the average
voltage of the cell during discharge by reducing the diffusional and resis-
tive overvoltages. Both of these areas are under active investigation in
the cell program, and it is hoped that these improvements in cell operation
will result in projected vehicle ranges in excess of 325 km (200 miles).
D. Lithium Reserves
(M. L. Kyle)
The availability of lithium ore reserves capable of supporting an electric
automobile industry utilizing lithium/sulfur batteries has been investigated.*
The resources were subdivided into three categories: (1) measured and in-
dicated resources, (2) resources inferred from current operations, and (3)
potential resources, which include low-grade ores of currently uneconomic
recovery potential. Estimates of lithium available from these sources are
shown below:
Source Kg of Li as Metal
Measured and indicated 4.3 * 109
Inferred 7.3 * 108
Potential resources 1.7 * 1010
Total 2.2 x 1010
There has been little economic incentive for locating additional lithium-
bearing deposits because the known reserves are adequate for many years at
the present rate of usage. Production statistics for lithium are not re-
leased to the public, but it is our estimation that the measured reserves re-
present a substantial supply at the current rate of lithium usage. (The 1965
edition of the U. S. Bureau of Mines' Minerals Yearbook estimates the world's
production of lithium ores to be only about 6 * 10y kg/year.) The chances
that a persistent exploration effort would yield further discoveries of lithium
ores must be recognized since this has proved to be true for other minerals.
The automobile described previously is powered by a lithium/sulfur
battery containing about 46.7 kg of lithium metal or equivalent. This value
assumes that LiCl-KCl is the cell electrolyte and LiA102 is the primary cell
insulator. The measured reserves are sufficient to produce about 108 bat-
teries of this size. Total reserves are sufficient to produce about 5 x 108
batteries of this size.
Lithium availability does not appear to be a hindrance to the utiliza-
tion of lithium/sulfur batteries as electric vehicle propulsion devices. The
present supply seems ample for any foreseeable application and, given an
economic incentive to develop new sources, additional reserves could probably
be located.
*The figures quoted were supplied by Dr. H. R. Grady of the Foote Mineral
Company.
-------
- 80 -
VI- STATUS AND FUTURE PLANS
The lithium/sulfur cell program has been, until now, essentially a lab-
oratory program to determine it these cells show promise for use in a high-
specific-energy, high-specific-power battery. At this stage in its development,
the Li/S cell still appears promising, and many of the problems that must be
solved to make the Li/S battery both technically and economically attractive
have been identified. Laboratory cells of less than 10-cm2 electrode area
have demonstrated capacity densities (at the 1-hr rate) of 0.4 A-hr/cm2 and
lifetimes of over 800 cycles and 1100 hr (at reduced capacity densities).
No cell operated to date, however has achieved the required high performance
levels over an extended lifetime. The sulfur electrode limits cell perfor-
mance — the capacity per unit volume and sulfur utilization must be improved
to provide a viable battery. Sulfur electrode design modifications, addi-
tives to the sulfur, and the proper choice of electrolyte are all capable of
increasing the cell performance.
The cells operated to this time have been relatively small ('10-cm2
electrode area) and unsealed. The full-scale cells will be much larger ("*•> 350
cm ) and must be sealed. Initial scale-up efforts are necessary to provide
informacion on problem areas peculiar to the sealing, design, construction,
and operation ot larger cells. Materials are available for laboratory and
first-generation, scaled-up cells, but the applicability of low-cost, light-
weight construction materials that are necessary for a high-energy-density,
high-power-density, low-cost battery is yet to be demonstrated. Finally, the
assembly of cells into batteries will present new problem areas of cell match-
ing, safety, reliability, long-term hermetic sealing, and temperature control.
The program until now has been a laboratory effort to demonstrate the tech-
nical feasibility of lithium/sulfur cells. A great deal of research and
development remains to be done to reach the stage of a reliable battery.
The FY 1972 OAP program at ANL for the development of lithium/sulfur
batteries will be concentrated in ^hree main areas of activity: a laboratory
program for the support of the other portions of the program, a materials
program for the development and testing of materials and components of cells
and batteries, and a cell-development program centering around the improve-
ment of cell performance and the development of sealed, scaled-up cells, lead-
ing into battery development in FY 1973.
The laboratory program will provide the necessary physicochemical in-
formation to the materials and cell-development programs. This information
will include the measurement of the solubility of cathode materials in molten-
salt electrolyte of various compositions and the physicochemical properties
of electrolytes, the preparation of electrolytes, the investigation and syn-
thesis of solid electrolytes having high lithium-ion conductivity, the study
ot the electrochemistry and chemistry of the lithium/sulfur cell, the evalua-
tion of new current collector and electrode structures, and safety studies.
The materials program will have as its objective the identification and
synthesis, as necessary, of corrosion-resistant metals and ceramics and the
construction of certain cell components such as current collectors and elec-
trical feedthroughs. Both short- and long-term corrosion tests will be carried
-------
- 81 -
out. The promising materials will be obtained in appropriate forms for cur-
rent collectors, housings, and other components, or these forms will be pre-
pared at ANL. Insulating seals and feedthroughs will be prepared for cor-
rosion testing and for use in cells. Suppliers for the seals and feedthroughs
will be identified.
The cell development program for FY 1972 will center around the improve-
ment of cell life at high capacity density and the scale-up of cells to larger
sizes. Small-scale cell studies will be performed to test new electrode
structure ideas, to improve cell lifetime, and to develop cell sealing tech-
niques. The cell scale-up work will involve the preparation and testing of
larger-area electrodes (^30-60 cm2), and the construction and testing of
larger-size sealed cells, preparatory to work on a 1-2-kW battery.
-------
- 82 -
REFERENCES
1. C. E. Johnson, M. S. Foster, and M. L. Kyle, Purification of Inert At-
mospheres, Nucl. Appl. J3» 563 (1967),
2. E. R. Van Artsdalen and I. S. Yaffe, Electrical Conductance and Density
of Molten Salt Systems: KCl-LiCl, KCl-NaCl, and KC1-KI, J. Phys. Chem.
_5_2, 118 (1955).
3. G. H. Kucera, P. T. Cunningham, Argonne National Laboratory, private
communication y!970).
4. C. E. Johnson, Argonne National Laboratory, private communication (1970).
5. S. I. Berezina, A. G. Bergman, and E. L. Bakumskaya, Phase Diagram of
the_System KCl-LiCl-KF-LiF. Zh. Neorgan. Khim. 8^ (9), 2140 (1963).
6. H. Shimotake, M. L. Kyle, V. A, Maroni, and E. J. Cairns in Proc. 1st
Intern. Electric Vehicle Symp., Nov. 5-7, 1969, Pheonix, Ariz.,
Electric Research Council.
7. A. S. Dworkin, H. R. Bronstein, and M. A. Bredig, Miscibility of Metals
with Salts. VI. Lithium-Lithium Halide Systems, J. Phys, Chem. 66, 572
(1962).
8. E. J. Cairns, G. H. Kucera, and P. I, Cunningham, Thermodynamic Studies
of the Lithium-Selenium System by an EMF Method, presented at the CITCE
Meeting, Prague, Sept. 28-Oct. 2, 1970; see also Extended Abstracts.
9. P. T. Cunningham, Argonne National Laboratory, private communication (1971)
10. V. A. Maroni, Argonne National Laboratory, private communication (1970).
11. H. A. Laitinen and C. H. Liu, An Electromotive Force Series in Molten
LiCl-KCl Eutectic. J. Amer. Chem. Soc. J50_, 1015 (1958).
12. C. H. Liu and S. Shen, Argentometric Titration of Sulfide in Alkaline
Solution, Anal. Chem. 36, 1652 (1964).
13. P. 0. Bethge, On the Volumetric Determination of Hydrogen Sulfide and
Soluble Sulfides. Anal. Chim. Acta 1£, 310 (1954).
14. R. Collongues, Les Phenomenes d'Ordre-desordre en Chimie Minerale, Ann.
Chim. (France) 8_, 395 (1963).
15. R. S. Roth and S. Hasko, Beta-Alumina-Type Structure in the System
Lanthana-Alumina, J. Amer. Ceram. Soc. ^(4), 146 (1958).
16. L. Hsueh and D. N. Bennion, Ionic Conduction in Beta-Alumina, Interim
Technical Report No. 6, Report No. 69-68, Dept. of Army, Mobility
Equipment Research and Development Center, Ft. Belvoir, Va.
17. A. Imai and M. Harada, Ionic Conduction in Impurity Doped B-Alumina,
Abstract No. 277, Extended Abstracts, The Electrochemical Society,
137th National Meeting, May 1970.
18. F. A. Shunk, Constitution of Binary Alloys, Second Supplement, p.344,
McGraw-Hill Book Co., N. Y. (1969).
19. G. Kullerud and H. S. Yoder, Pyrite Stability Relations in the Fe-S
System, Economic Geology _54_(4), 533 (1959).
20. A. A. Velikanov and V. A. Tertykh, Electrical Conductivity, Polarization,
and Electrolysis of Thallium-Sulfur Melts, Zh. Fiz. Khim. ^3_, 2580 (1969).
21. M. Hansen and K. Anderko, Constitution of Binary Alloys, p.1168,
McGraw-Hill Book Co., N. Y. (1958).
22. R. K. Steunenberg, C. Trapp, R. M. Yonco, and E. J. Cairns, Electrical
Conductivity of Liquid Sulfur and Sulfur-Phosphorus Mixtures, presented
at Third Annual Mardi Gras Symposium on Sulfur Chemistry and Theoretical
Chemistry, New Orleans, February 18-19, 1971.
23. R. W. Flourncy, Reynolds Aluminum Corp., private communication
(Dec. 28, 1970).
-------
- 83 -
24. D. W. Strickler and R. Roy, Studies in the System
J. Amer. Ceram. Soc. ^(5), 225 (1961).
25. J. E. Knap, B. Pesetsky, and F. N. Hill, Vapor Plating With Dicumene-
chromium Preparation and Properties of the Chromium Plate, Plating 53,
772 (1966).
26. J. H. B. George, L. J. Stratton, and R. G. Acton, Prospects for Electric
Vehicles, A Study of Low-Pollution-Potential Vehicles — Electric, Arthur
D. Little, Inc. report to the U. S. Department of Health, Education and
Welfare, National Center for Air Pollution Control (May 1968).
27. P. D. Agarwal, I. M. Levy, A High-Performance A. C. Electric Drive
System, SAE preprint 670178, Annual Meeting, January 1967.
28. E. A. Rishaun, W. D. Bono, and T. A. Zechin, Electrovair - A Battery
Electric Car, SAE preprint 670175, Annual Meeting, January 1967.
-------
APPENDIX: SUMMARY OF PERFORMANCE OF CELLS
OPERATED IN FISCAL YEAR 1971
-------
CELLS OPERATED IN FISCAL YEAR 1971
Cathode
Sulfur-Containing Element
Cell
No.
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Celia Hate-
Type rlalb
L
R
L
L
R
L
L
L
L
L
L
L
L
L
L
L
L
C
C
L
L
L
R
SSH
SSH
SSB
SSB
SSH
SSH
SSH
SSH
SSH
SSH
SSB
SSH
SSH
SSH
POCO AX
POCO AX
POCO AX
POCO AX
POCO AX
POCO AX
FC-14
POCO AX
POCO AX
Ity, X
80
80
85
85
80
80
80
80
80
80
83
80
80
80
63
63
63
63
63
63
49
63
63
Pore
30
30
40
40
30
30
30
30
30
30
25
30
30
30
1.4
1.4
1.4
1.4
1.4
1.4
3.5
1.4
1.4
Thick-
cm
0.16
0.16
0.071
0.071
0.16
0.16
0.16
0.16
0.16
0.16
0.05
0.09
0.09
0.09
0.09
0.09
0.09
1.3
1.3
0.094
0.17
0.09
0.06
Elect ro lyte-Cont .
rlalb
SSH
SSH
SSB
SSB
SSB
SSH
SSB
SSB
SSB
SSH
SSH
SSH
SSH
SSH
SSH
...
Mo
Mo
Mo
Ity,
80
80
85
85
85
80
83
83
83
80
80
80
80
80
80
75
75
75
Element
Pore Thlck-
I .m
30
30
40
40
40
30
25
25
25
30
30
30
30
30
30
None -
20
20
20
' °nc
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
cm
.16
.16
.071
.071
.071
.16
.045
.045
.05
.09
.09
.09
.16
.16
.16
--
.17
.17
.09
Mo Foam Sheathc
Pore Thick-
1 1 y , 'f. -in era
None
None
None
None
None
None
None
None
None
None
. None
None
None
None
None
None
None
None
None
None
None
None
None
Active
LI, K
1.0
0.5
0.5
0.5
0.2
0.5
0.5
0.5
0.5
1.0
1.0
0.3
0.3
0.5
0.5
0.6
0.6
0.7
0.7
0.7
1.7"
0.7
0.6
s. g
1.23
0.30
0.51
1 .06
0.30
1.24
1.27
1.18
1.16
1.18
0.68
0.74
0.77
0.77
0.62
0.59
0.60
1.64
1.S5J
0.6&J
0.71
1.27J
0.89
lyted
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
B
B
B
B
Cath-
ode
cm-
1 .59
i .:
1 .70
1.64
1 .2
1.09
1 .09
1.58
0.96
0.9t>
0.97
0.92
0.92
0.92
1.33
1.33
1.31
1.89
1 .89
1.75
1 .71
2.43
1.00
Theo-
ret leal
Capac 1 tv
A-hr/ccr
1.29
0.42
0.50
1. 08
0.40
1.90
1.95
1.25
2.01
2.06
1.17
1.34
1.39
1.39
0.79
0.74
0.76
1.45
1.08
0.50
0.56
0.69
1.48
Capacity
Density,
A-hr/cm1'
0.33
0.16
0.23
0.03
0.23
0.18
0.29
0.02
0.52
0.27
0.13
0.19
0.02
0.21
0.33
0.40
0.30
0.09
0.04
0.07
0.28
Capac Ity
per Unit Percent
Volume Current of Theo-
A-hr/cm2
0.30
0.14
0.21
0.16
0.21
0.16
0.26
0.018
0.47
0.25
0.12
0.18
0.019
0.20
0.30
0.23
0.18
0.05
0.03
0.06
0.20
A/cm' Capacity
0
0
0
0,
0,
0,
0,
0.
0,
.30
.29
.30
,33
,37
.38
.32
.52
,52
0.54
0.
0.
0
0.
0.
0.
0.
0.
0.
,54
.54
.26
38
25
53
32
32
25
0.25
0.30
26
31
21
8
12
9
24
1
25
20
9
14
1
28
44
27
28
17
8
9
19
Cell Life
V Cycles hr
1.4
1.3
1.4
1.4
1.4
1.4
1.5
1.2
1.5
1.3
1.5
1.5
1.5
1.4
1.3
1 .5
1.5
-
2.0
6
0
3
4
1
1
3
3
1/2
3
0
4
2
1/2
1/2
2
6
3
17
1
35
6
100
2
5
0.3
1
1
3
0.1
5
0.1
3
0.9
0.5
0.1
2
18
7
5
15
2
152
5
1
00
1
-------
CELLS OPERATED IN FTS:AL YEAR 1971 (cont.)
Cathode
Sulfur-Containing El<
Cell
No.
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
Cello
Type
L
ER
ER
ER
EL
ER
ML
ML
ER
ER
ER
ER
ER
LY
LY
ER
M
M
rial1" Ity, Z
POCO AX
FC-14
FC-14
FC-14
POCO AX
FC-14
PC-60
POCO AX
FC-14
PC-25
PC-25
FC-14
PC-25
Nb
CF
PC-25
Metal
Ut, g
0.226
2.0
63
49
49
49
63
49
48
63
49
48
48
49
48
90
97.3
48
Vol Z
0.40
4.4
um
1 .4
3.5
3.5
3.5
1.4
3.5
33
1.4
3.5
Thlck-
0.09
0.16
0.2
0.3
0.15
0.3
0.4
0.3
120 0.3
120 1.0
3.5 0.25
120 0.32
12/35 0.16
0.6
120 1.3
Carbon
Ut. g Vol Z
0.51 5.1
0.8 10.0
Elect rolyte-Cont . Element
Mo 75 20 0.09
Mo 78.2 25 0.19
Mo 78.2 25 0.2
Mo 78.2 25 0.2
Nb 90 12/35 0.16
Electrolyte Sulfur
Ut, g Vol Z Ut , g Vol Z
3.95 48.5 4.59 46.0
2.69 43.5 3.2 42.1
Mo Fo,
Ity, Z
am Sheathc
..m
cm
78.2
90
90
78.2
78.2
78.2
78.2
78.2
78.2
78.2
78.2
78.2
78.2
78.2
78.2
78.2
78.2
25 0.3
67 0.4
67 0.3
25 0.3
25 0.4
25 0.6
50 0.48
25 0.16
25 0.16
25 0.16
25 0.33
25 0.64
25 0.16
25 0.39
25 0.64
25 0.15
25 0.15
Act
i ve
Material
LI. g S, g
0.7
1.0
0.4
0.3
0.7
0.7
14.4
13.9
0.7
0.7
0.7
0.7
O.B
0.8
0.8
0.8
2.11
1.85
0
3
0
0
0
1
19
16
1
1.
0.
1.
11
0.
8.
9,
4.
3.
.28
.82
.98
.98
.51
.22
.4
.1
.18
.52
.74
.24
.8
,71
.34
,43
.59
.2
Cat li-
tre- Area.e
lyted cm-
B 1.47
B 1.24
C 0.28
B 1.00
D 2.53
D 1.00
D 20.0
D 27.6
B+ 1.0
B+ 1.0
B+ 1.0
D 1.0
D 2.6
D 1.0
B+ 1.8
D 2.6
B+ 2.55
Theo-
ret leal
Dens Ity.'
A-hr /cm''
0
5
6
1
0
2
1
0.
1 .
2.
1.
2
7
1
7.
6
3.
2.
.32
.14
.01
.63
.33
.04
.60
.97
.97
.54
.24
.0'
.58
,19
.61
. 15
0
1
Capacity
F , i v
A-hr /cm'
0.10
0.45
0.095
0.15
0.17
0.55
0.60
0.18
0.15
0.13
0.21
0.57
0.51
0.048
0.17
0.23
0.13
0.14
0.58
0.70
0.75
0.218
0.257
Capac Ity
per Unit
Cathode, S Density,
A-hr/ctrr A/en"
0
0
0
0
0
0
0
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0.
0
0.
0.
.19
.046
.13
. 14
.33
.30
. 17
.14
.18
28
51
52
05
.09
.05
.29
.30
.20
.24
.26
,114
,172
0.20
0.24
0.28
0.25
0.17
0.20
0.18
0.25
0.20
0.10
0.10
0.10
0.20
0.20
0.32
0.20
0.10
0.10
0.20
0.10
0.1
0.1
r ercent
ret leal
Capacity
31
9
2
9
50
27
20
19
7
22
41
8
11
11
11
8
11
12
8
12
Voltage,
V
1.7
1 .7
1.6
1.8
1.6
1.9
1.9
1.8
1.7
1.9
1 .7
1.9
1 .6
1.7
0.6
1.9
1.8
1.6
Cell
Life
h No. of Time,
Cycles hr
7
1
1
9
803
600
23
7
141
29
253
131
1/2
36
7
11
82
3
3
3
1
26
1100
750
30
25
184
89
90
175
5
117
27
70
230
42
00
ON
-------
- 87 -
FOOTNOTES FOR APPENDIX TABLE
aCell type classification:
L = laminated, C = comb, R = reservoir, ER = enclosed reservoir,
EL = enclosed laminated, ML = multilayered, LY = layered, M = mixed.
bSSH = Type 302 stainless steel Feltmetal, Huyck; SSB = Type 304 stainless
steel Brunspore Fibermetal, Brunswick; POCO AX = porous graphite, Poco
Graphite; FC-14 = porous carbon, Pure Carbon; PG-25, PG-60 = porous
graphites, Union Carbide Corp.; NB = niobium; GF = graphite felt.
cMolybdenum foam = a product of Spectra-Mat, Inc. In Cells No. 53 and 54,
Type 347 stainless steel Feltmetal was used.
dElectrolytes (in mol %):
A = 59 LiBr-41 RbBr (mp = 278°C), B = 58.5 LiCl-41.5 KC1 (mp = 352°C),
C = 11.7 LiF-29 LiCl-59.3 Lil (mp = 342°C), D = 8.5 LiCl-59.0 L1I-32.5 KI
(mp = 265°C), B+ = 3.5 LiF-56 LiCl-40.5 KC1 (mp ^345°C).
eCathode area: projected electrode area facing the anode.
^Theoretical capacity density: calculated based on Li£S as the final dis-
charged product.
gCapacity per unit volume of cathode: based on the volume of reactants and
current collectors excluding housing and sheath.
^Average voltage = ——— where V = cell terminal voltage,
t = discharge time (cut-off voltage = 1 V).
jIn these cells Pt+SjQ was used instead of S.
this cell a Li-Al alloy was used instead of Li.
------- |