ENVIRONMENT**
PROTECTION
AGENCY
DALLAS, TEXAS
EPA-600/2-76-007
January 1976 Environmental Protection Technolo!
MIXED OXIDES FOR FUEL CELL ELECTRODES
i
viV- iv, '
Industrial Environmental Research Laboratory
Office of Research and Development
:;;::;;^^
R^irch Triangle Part, Mtrtft Carolina 27711
-------
EPA-600/2-76-007
MIXED OXIDES
FOR
FUEL CELL ELECTRODES
by
U. Bertocci, M. Cohen, W. S. Horton,
T. Negas, andA.R. Siedle
National Bureau of Standards
Department of Commerce
Washington, DC 20234
EPA Interagency Agreement No. D4-0528
ROAP No. 21BKR-007
Program Element No. 1AB013
EPA Project Officer: Stanley J. Bunas
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
January 1976
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research-and Development,
U.S. Environmental Protection'Agency, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed
to develop and demonstrate instrumentation, equipment and
methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the
new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U. S. Environmental Protection
Agency, and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
This document is available to the public through the National
Technical Information Service, Springfield, Virginia 22161.
-------
ABSTRACT
This report presents studies of the preparation, chemical stability,
and electrochemical behavior of mixed oxides in order to determine if
these are potential candidates as oxygen-reducing electrocatalysts in an
acid fuel cell. Materials studied included strontium and barium cobaltates
and manganates with and without added titanium; lanthanum titanates, with
and without calcium or strontium; calcium, strontium, and barium ruthen-
ates; and mixed oxides of the systems Ti-Ta-O, V-Nb-0, Ce-Ta-0, Pr-Ta-0,
Ce-Nb-0, and Ce-Pr-Ta-O. The choices were based upon producing variable
valence for a given transition metal and upon conferring stability at
elevated temperatures U 150 °C) in phosphoric acid using oxides of known
stability. Barium ruthenate and the systems Ti-Ta-O, V-Nb-0, V-Ta-0,
Ce-Ta-0 were hot-acid stable. The high temperature reactions of CeTaOtf+x with
0 $ x £ 0.5 were studied in air up to about 1960 °C.
Potentiodynamic and galvanostatic studies are reported for materials
from the Ti-Ta-O system, Ti02 as grown, Ti02 reduced with hydrogen, Ti02
with 0.1% Nb, lanthanum titanates with and without calcium or strontium,
a tungsten bronze, barium ruthenate, and strontium titanate with 0.03%
and with 0.15% Nb.
Because the reduction of oxygen is easier than that of nitrogen,
and because biochemical reduction of nitrogen takes place rather easily,
preparation of inorganic compounds with ternary metal-sulfur arrays
similar to the arrays -in nitrogen reductase was attempted. These
compounds were also to be tested in a manner similar to the mixed oxides.
The following were made: (Ph3?) i+Cu2W2S6 , (PhsP)sAg5W2S602, and
t(Ph3?)3Ag]2W2Sg, where Phs? refers to the triphenylphosphine moiety.
Also prepared were (PhsPAu) 2W2Stt, (PhsAsAu) 2WStf, (diphos)NiW02S2 ,
(diphos)2Pd3W2Se02, and (Phs?)3PtW2Sg02, where "diphos" refers to
1,2-bis(diphenylphosphino)ethane.
111
-------
TABLE OF CONTENTS
Page
Abstract < . . ii
List of Figures v
Acknowledgements vi
I. CONCLUSIONS 1
II. RECOMMENDATIONS 2
III. INTRODUCTION 3
General 3
Objectives 4
Sections 4
IV. ACID STABILITY 6
Method of Test 6
Results 7
V. PHASE STUDIES AND MATERIALS PREPARATION 9
VI. ELECTROCHEMICAL MEASUREMENTS 16
Introduction 16
Experimental Arrangement 16
Description of Experimental Set-up 16
Description of Cells Used 16
Electrodes 18
Experimental Methods 23
Potentiodynamic Measurements 23
Titanates 23
Tungsten Bronze 26
Barium Ruthenate 26
Redox Measurements 33
Titanates 33
Tungsten Bronze 34
Barium Ruthenate 35
Discussion 39
-------
Page
VII. MATERIALS OTHER THAN OXIDES 42
Introduction 42
Results 42
Aqueous Systems 42
Nonaqueous Systems 43
VIII. REFERENCES 46
IX. LIST OF PUBLICATIONS 48
VI
-------
LIST OF FIGURES
1. Block diagram of electrochemical instrumentation. Page 17.
2. Electrochemical cell used at room temperature with Type I electrode.
Page 19.
3. Type II electrode shown in front view and cross-section. Page 21.
4. Current-potential curve for 95Ti02:5Ta£05 in 5 mol/1 J^PO^ at roorr,
temperature. Page 24.
5. Current-potential curves for the lanthanum titanates at room
temperature. Page 25.
6. Current-potential curves for a Type II BaRuOs electrode in concen-
trated phosphoric acid at 60 °C. Page.27.
7. Effect of oxygen and the current-potential curve for a Type II
BaRu03 electrode. Page 29.
8. Current-potential curve for a Type II BaRuO3 electrode in neutral
solution. Page 30.
9. Comparison of two BaRuOa electrodes (Type II) in 6 mol/1 NaOH at 60 °C.
Page 31.
10. The effect of 0.007M H2C>2 on a Type II BaRuOs and on smooth Pt.
Page 32.
11. Steady-state performance of Type II BaRuOs electrode with
ferro-ferricyanide reaction? semilog scale. Page 36.
12. Steady-state performance of Type II BaRuOs electrode with
ferro-ferricyanide reaction; linear scale. Page 37.
13. Steady-state performance of Type II BaRuC>3 electrode with
iodine-iodide reaction. Page 38.
711
-------
-ACKNOWLEDGEMENTS
The work reported herein was performed by Dr. U. Bertocci,
Mr. M. I. Cohen, Dr. T. Negas, and Dr. A. R. Siedle, with the assis-
tance of Mr. N. K. Adams, Mr. J. L. Mullen, Mr. C. D. Olson, and
Mr. J. Broussalian.
Dr. A. D. Franklin provided overall scientific guidance. Dr.
H. P. R. Frederikse participated constructively in a number of the
discussions particularly those dealing with catalytic effects of a
variety of perovskites. Dr. R. S. Roth provided valuable crystallo-
graphic data and advice on the cerium tantalates and the metal
thiotungstate derivatives. Mr. H. S. Parker gave valuable assistance
with preparation of single crystals. Dr. G. A. Candela performed and
interpreted magnetic measurements. All of the above collaborators
are at NBS.
Mr. N. L. Loeffler, University of Tennessee, engendered the
interest in and collaborated in the preparation of the lanthanum
titanates.
vni
-------
SECTION I
CONCLUSIONS
On the basis of consideration of stoichiometry and the expected
effect on electron transport properties, the mixed oxides containing
rare earths (cesium and/or praeseodymium) and certain transition metals
(titanium, vanadium, niobium, and tantalum) may be quite interesting as
candidates for electrocatalytic applications in hot-acid fuel cells.
It has been shown that some are stable in hot, concentrated phosphoric
acid. The high temperature phase relations important to their prepara-
tion have been studied.
Barium ruthenate, although exhibiting stability in hot, phosphoric
acid, appeared to change chemically during the passage of electric
current. This material may not be useful as a catalyst for the oxygen
electrode of an acid fuel cell.
Alkaline earth titanates, manganates, and cobaltates are soluble
in hot, phosphoric acid. They are, therefore, unsuitable as catalysts
for fuel cells of the type considered here.
Although some materials appear promising, the work reported here was
too preliminary to decide definitely that oxide materials will or will
not serve adequately as oxygen electrodes in a hot phosphoric acid
fuel cell.
-------
SECTION II
RECOMMENDATIONS
There are several fruitful areas of research in which appropriate
work may well lead to a substitute for platinum and increased under-
standing of the electro-catalytic process involved. Of the materials
studied here, barium ruthenate appeared to have promise as a model
material, but much remains to be done to provide an understanding
of its chemistry as an electrode.
The phase relations for the mixed oxides containing rare earths
(e.g., cesium and/or praeseodymium) and certain transition metals
(e.g., titanium, vanadium, niobium, and/or tantalum) should be eluci-
dated in order to prepare suitable materials for testing. Then the
acid stability tests should be made where still needed, and the
electrochemical studies performed.
Appropriate electrodes of triphenylphosphine copper thiotung-
state should be prepared and studied electrochemically.
-------
SECTION III
INTRODUCTION
GENERAL
The fuel cell is related to the well-known "storage battery"
through the fact that both convert chemical energy into electrical
energy. The principal difference is pointed up by the words fuel and
storage. The storage battery, a regenerative device, stores electrical
energy by converting it to chemical energy for later release again as
electricity. The fuel cell, on the other hand, can convert fuel into
electrical energy continuously.
In ordinary use, fuel is burned and the heat thereby generated
is used directly, or among other possibilities, it may be converted
to steam which is used to power an electric generator. Other means
to burn fuel and produce electric energy include magnetohydrody-
namic generation. In a fuel cell, on the other hand, the chemi-
cal reaction between the fuel and oxidant produces, in the appro-
priate electrochemical device, electrical energy directly. For the
simplest devices hydrogen and oxygen serve directly as fuel and
oxidant. More complex systems are required in order to use fossil
fuels or derivatives therefrom. Air is a practical and economical
oxidant.
The fuel cell offers advantages of high efficiency at partial
and at full load, silence, low-pollution, unattended operation, and
construction in modular units to allow for optimum distribution of
sites. There are, however, a number of areas where further
research and development are needed to make fuel cells economically
competitive so that they will make a significant impact on the
energy conversion technology of the United States. A large factor
in the economics of these devices is the cost of the catalyst used
on the electrodes. Currently, platinum is the most widely used catalyst
-------
in the acid fuel cell, the leading candidate for widespread use in
electric power generation. For example, it is the catalyst for the
United Aircraft device for power generation which has had the most
development, testing, and demonstration to date. The great cost of
platinum affects the fuel cell economics by leading to high initial
cost and to high replacement cost because platinum has a relatively
short life-time. Another factor is that its catalytic action is not
perfect and permits formation of intermediate peroxides cutting
efficiency by about 40%. For these reasons a fruitful program of
research would be to look for alternate catalysts which may not suffer
from some of the disadvantages of platinum. An important auxiliary
benefit could be to gain further understanding of the electrochemical
processes involved and of the nature of pertinent catalytic materials.
The work reported here was based upon the assumption that the
pertinent fuel cell for power generation would have hot phosphoric
acid as an electrolyte. Again, this is used in the cell which is
the forerunner . Furthermore, attention was focussed upon the
oxygen electrode, where the electrochemical efficiency loss appears
to be greatest. Consequently, two required properties are stability
in hot phosphoric acid and catalytic activity for the electroreduction
of oxygen.
OBJECTIVES
The objectives of the project were to assess the potential of
some selected transition metal oxides as electrocatalysts for fuel
cells and to develop an understanding of the important material
parameters for this application. It is a long-term goal, beyond
the term of the contract, to establish an interrelationship between
the structure, defect content, and concentration of dopants of
metal oxides and their behavior as electrodes in low temperature
fuel cells, with emphasis upon the cathode.
-------
ORGANIZATION OF THE REPORT
In the remainder of the report, separate sections will be
devoted to phase studies related to the discovery and preparation
of new mixed oxide materials. Brief studies of the acid stability
of some of these materials are reported. Electrochemical studies
of those materials which appeared at the time to be potential
catalysts for oxygen reduction at the cathode of a fuel cell are
included.
Early in the course of the project the idea was conceived that
the molecular structure of nitrogen-reducing enzymes might offer a
clue to oxygen reduction catalysts. Consequently, some effort was
aimed at producing pertinent compounds. A number of such compounds
were made, and some of these are being subjected to tests involving
solid electrodes. The concept and the chemistry involved in the
preparation are sufficiently interesting to be reported here.
-------
SECTION IV
ACID STABILITY
INTRODUCTION
A systematic program that integrated efforts in the acid
stability and materials preparation (see Section V) studies was
followed. Initially, acid stability testing was conducted using
available and newly-synthesized oxide materials which showed poten-
tial for suitable electronic properties. These materials primarily
were oxides containing first-row transition metals. This avenue of
approach proved unsuccessful and emphasis, therefore, was shifted
toward materials having potential for acid stability. Materials
in this group were oxides containing tantalum or niobium. These
oxides proved to be acid stable but are poor electronic conductors.
Research, therefore, was focussed on new Ta- and Nb-containing
oxides which could be modified chemically to enhance their electronic
properties.
METHOD OF TEST
Stability in 85% phosphoric acid at a mininum of 100 °C was
chosen as an initial screening test for candidate oxide materials.
This test was conducted in two stages. A qualitative determination
first was made by simply immersing powdered oxides in a bath of hot
acid. Stability of a material was noted by observing apparent
degradational processes (i.e., dissolution/decomposition, solution
color changes). Residues from these tests were examined by powder
x-ray diffraction to determine any changes in phase constitution
from the original, single phase, untreated material. A material
which did not show apparent degradation during this stage was
fashioned into a rod- or disc-shaped body and sintered at an
appropriate temperature. This body was placed in a Pt-wire basket
which then was immersed and rotated in a bath of hot acid. Possible
degradation of the material was monitored gravimetrically. In most
-------
cases this test was performed first at 110 °C and then at 150 °C
for durations of up to one month.
RESULTS
The initial stage of this investigation concentrated on the
acid stability characteristics of oxides containing first-row
transition metals previously prepared in this laboratory. For this
stage, however, only those materials were chosen which, from stoichiometric
considerations, had at least the potential to be metallic or semiconducting.
Materials from the following systems were tested:
a. SrMnC>3_x, 0 £ x $ 0.35 (see ref. 2 for complete details of
the system)
b. BaMn03_x, 0 $ x $ 0.28 (see ref. 3 for complete details)
c. Bai_ySryMnC>3_x, (see ref. 4 for complete details)
d. BaCo03~x and BaNiC>3_x (see ref. 5 and 6 for complete details)
e. SrMnC>3 - "manganese oxide" (see ref. 7 for complete details)
These systems feature materials having numerous crystal structures
related to and including the perovskite type. Moreover, the transi-
tion metal (Mn, Ni, or Co) can exist in multiple oxidation states,
a phenomenon (nonstoichiometry) recently shown to be important in
some types of catalytic activity (i.e., CO oxidation and NOX
reduction). None of the materials in these systems withstood attack
by hot phosphoric acid. It was noted, however, that the low tempera-
ture form of SrMnC>3_x (O £ x $ 0.10) , with a so-called four-layer
2
hexagonal structure was not attacked at room temperature. The
compound began to dissolve at 60 °C. It was thought that this
material might be rendered more acid stable by suitable partial
substitution (solid solution) of acid stable, oxide end-members
having similar, perovskite-related structures. BaTiC>3 and SrTiC>3
were chosen as candidate materials for this purpose in view of
their resistance to some acids. However, this effort was curtailed
when tests showed that even these materials undergo dissolution in
hot phosphoric acid.
-------
Another series of nonstoichiometric transition metal oxide
materials of the perovskite type, La(2/3)+xTi3?xT:i-^-3x03 (x S 1/3)
appeared promising. This series contains solid solutions with
properties ranging from metallic (LaTi3+C>3) through semiconducting
U4- 8
to insulator (La£/3TiH 03). It had been suggested that these
materials were stable in cold HC1, HNC>3, perchloric acid and in
hot, dilute HN03, and in boiling perchloric acid. They were
slowly attached by aqua regia. Phases in this series, as well as
Gd and Sm analogues, were prepared either in this laboratory or
obtained from an external source (Mr. Neil Loeffler, University
of Tennessee Space Institute). Similar materials,
LaO .7 5A0 .25^1^75Tio .25^3 ^ = ^r or Ca) were also prepared in
cooperation with Mr. Loeffler. All materials were tested and
found to dissolve in hot phosphoric acid and 3N sulfuric acid.
In the second stage of this study, oxide materials containing
Ru, Ta, and Nb were investigated. BaRuOs, SrRuOs, and CaRuC>3,
9 10
characterized as having "good" conductivity ' and metallic
conductivity were tested. Only BaRuC>3 was found to be stable
(110 °C, 5 days) in phosphoric acid. All of the niobium- and
tantalum-containing phases that were prepared (see Section V)
showed remarkable resistance to attack by hot phosphoric acid.
Many of these phases, however, are poor conductors and it would
be necessary to enhance their electronic properties, perhaps, by
suitable doping.
-------
SECTION V
PHASE STUDIES AND MATERIALS PREPARATION
Perovskite materials in the series La(2/3)+xTi§xTi]f-3x03
(x $ 1/3) and in the series Lag .7sAg .25Ti$t75Ti(Jt2503
(A = Sr or Ca) were synthesized. The latter materials were chosen
hoping to enhance the acid stability of the first
series without a sacrifice in electronic conductivity. Materials
were prepared from La2C>3 [from La (OH) 3], Ti02/ Ti metal, and
SrTiOs or CaTi03 starting materials, proportioned to yield the
desired stoichiometries. The starting materials were mixed then
melted and cooled in high purity Mo crucibles in an induction furnace
containing a "gettered" argon atmosphere. The resulting products,
examined by powder x-ray diffraction, were found to be single-phase
perovskites (non-cubic). These materials were utilized subsequently
for electrochemical and acid stability experiments.
BaRuOs, SrRuOs, and CaRuOs were prepared from the appropriate
alkaline-earth carbonate and RuO2 or Ru metal. The blended
starting materials were reacted in gold envelopes at 1000 °C in
air for one day. SrRuOs and CaRuOs have the perovskite structure
while BaRuOs has a so-called nine-layer rhombohedral-hexagonal
structure in which face-sharing of Ru-containing oxygen octahedra
9
occurs . As indicated in Section IV only BaRuO3 remains stable in
phosphoric acid and was used in subsequent electrochemical
experiments. The acid stability of BaRuO3 was somewhat unexpected
as the phase Bag.gSro.iMnOs (< 1265 °C in air) has the same
4
crystal structure yet is attacked by phosphoric acid. Phase
transitions in or reduction of BaRuOs were not detected up to
1600 °C in air by differential thermal and thermogravimetric
analysis. Significant vaporization, presumably of Ru through
oxidation to gaseous RuO£ or RuO3, was observed above 1500 °C.
This vaporization precludes the preparation of dense sintered discs
at these elevated temperatures. For example, a pressed disc of
-------
initially single phase BaRuOs was fired 22 hours at 1500 °C. The
product consisted of two phases, BaRuOa plus a ruthenium-containing
phase richer in barium. If dense discs are necessary for electro-
chemical testing, BaRuOs could be hot-pressed or isostatically
pressed at lower temperatures (< 1500 °C) to prevent loss of
compositional integrity.
Past experience in this laboratory has shown that oxide
materials containing Nb20s or Ta2Os resist attack by many acids.
For this reason, several of these oxides were synthesized and
tested with phosphoric acid prior to initiating an investigation
12
of new materials. In the system Ta2O5-TiC>2, Roth and Waring at
NBS showed that Ti02 (rutile) can accommodate as much as 10 mole
percent Ta20s, depending on temperature, without gross structural
changes. A composition midway in this range, 13305'19Ti02, was
chosen and prepared using high purity Ti02 and Ta20s starting
materials. After a preliminary calcination at 1000 °C for
three days, this mixture was fired at 1600 °C for one day and
quenched in water. The final blue-black product consisted of
single-phase rutile solid solution as revealed by powder x-ray
diffraction analysis. This material was utilized for acid and
electrochemical testing. The compounds V2C>5'9Nb205 and
V205«9Ta205 were also prepared according to methods in reference
13 and tested in hot phosphoric acid.
In view of the apparent resistance of 13205- and Nb2O5~containing
oxides to attack by hot I^PO^, an investigation of new materials based
on these oxides was initiated. Cerium oxide was chosen as an
additional component because the cerium cation can exist in
crystalline phases in mixed 3+, 4+ oxidation states. This could
enhance the probability for reasonable electronic conductivity of
potential electrode materials. Furthermore, cerium oxide was
14
shown to be a catalyst for related gas phase reactions . When
reacts with Ta20s (or Nb20s) at elevated temperatures, compounds
are formed that contain the reduced oxidation state, Ce3 , even
10
-------
under oxidizing conditions as in air. In the Ta£C>5-cerium oxide
system, three compounds, CeTayOjg, CeTasOg, and CeTaO^ are stable.
The crystallographic and magnetic properties of these, as well
as additional, unusual phases, were detailed fully in this
laboratory. The phases CeTasOg and CeTaOi^ resist attack by hot
H3POI+, but, apparently, are not good conductors. Attempts,
therefore, were made to enhance the electrical conductivity by
altering the Ce3+ content toward mixed Ce3+/Ce't+ variants for
each phase.
CeTa30g can be reformulated as Cei/3TaC>3. Single crystal
and powder x-ray diffraction studies revealed that this material
has the perovskite structure which is distorted to orthorhombic
symmetry. The formulation, Cei/^TaO^, indicates that the A-sites
of this perovskite (ideally ABC>3, A/B = 1.0) are not fully
occupied, A /B = 1/3. Attempts, therefore, were made to
C.6 X Si
establish Ce cations at the A-sites by appropriate substitutions
of A2+ cations. This would lead to mixed Ce3+/Celt+ oxidation
states and possibly to enhanced conductivity. The substitution
mechanism can be formulated as,
[Ceft/3)-xCe^24>2]Ta03, x < 1/3 (1)
The feasibility of this type of substitution was tested by
additions of Sr2 according to,
[Ce^30Ce]$20Srl/2oJTa02» x = 1/10 in (1) (2)
Single phase perovskite products could not be synthesized and this
effort was discontinued. It is noted, however, that Cei/sTaOs,
an acid stable material, might be used as a component to enhance
the acid stability of perovskite and perovskite-like phases of
the first-row transition metal oxides previously tested (see
Section IV).
Attempts to vary the Ce3+/Ce't+ content of CeTaO^ met with
success. Phase equilibrium studies (reaction and
quenching of materials from elevated temperatures), single crystal
and powder x-ray diffraction analysis, thermogravimetric analysis,
11
-------
and magnetic susceptibility measurements were utilized to
completely define phase relations and oxidation-reduction
reactions in the system CeTaO^-oxygen (in air) . CeTaOit is stable
in air from 1265 °C to a melting point near 1960 °C. At 1265 °C,
the material oxidizes (reversibly) according to,
SCeTaO^ + 1/2 02 « 1265—^-*- CeTa3Og + 2Ce02 (3)
Notice that the bulk composition of the decomposition products is
expressed by CesTasOis or CeTaOi+^33. The reaction proceeds to
the right rapidly within the range 1265°-1100 °C but sluggishly
between 1100°-1000 °C. If CeTaOi* is quenched to room temperature
and reheated below 1000 °C, or, if the material is rapidly cooled
from above 1265 °C to below 1000 °C, the oxidation-decomposition
reaction (3) is completely by-passed. Instead, CeTaO^ absorbs
oxygen (oxidizes) according to,
CeTaOtt + 1/2 x 02 •- CeTaO^+x (4)
The x parameter in CeTaOit+x is variable but temperature dependent.
Four distinct reaction series were defined for the system
CeTaO^-oxygen (CeTaOi^+x). A summary of the observed reactions and
reaction paths is given in the flow diagram provided. The
following symbols are used:
(* »), rapid, reversible reaction
(^ .».) f sluggish, time dependent, but reversible reaction
( ».) f sluggish, non-reversible reaction
(^ ».) f reversible reaction but rapid or sluggish in the
direction indicated.
It is emphasized that every phase shown in the flow diagram can be
prepared by the pertinent reaction path and can be quenched (and
retained) to room temperature. Within series 2, 3, and 4, x varies
continuously with temperature. For this reason, only typical x
parameters in CeTaO^+x are given for selected temperatures.
Ce3+/Cel+"1" contents can be computed for given x parameters using
Ce?-2xCe£xTa0tf+x' This formulation was confirmed by magnetic
12
-------
The System CeTaO^+y (0 $ x $ 0.5) in Air
1265°.
+ Ce02 (decomposition products) Series 1
350c
CeTaOij.50
500e
575
CeTaO
4.480
(quenched from > 1265° to room temp.)
600C
CeTa04.165
800C
CeTaOit ,UB
I
950°
950<
800C
700C
CeTa04.383
700° to room
temperature
Series 2
Series 3
Series 4
13
-------
susceptibility measurements.
Starting with preformed CeTaOi^ (> 1265 °C, quenched to room
temperature) any reaction series can be initiated by heating the
materials within the pertinent temperature range indicated. Once
series 4 phases (not related crystallographically to CeTaOit) are
formed, however, series 3 and 2 materials subsequently will not
form. Series 4 materials will transform to series 1 phases at
elevated temperatures. Within each series, oxidation-reduction
reactions are almost instantaneous and reversible, with x being a
function of temperature. Reactions between any two series are
generally sluggish and not necessarily reversible. For example,
reactions in series 3 can be followed (thermogravimetrically)
continuously to at least 950 °C (x = 0.058). Phases in this series
are crystallographically related to CeTaOtt. If the sequence is
stopped within the 900-950 °C interval, and the series 3 material
is maintained at constant temperature, series 4 materials will
slowly develop over several hours. Thus the transition from
series 3 to series 4 materials, although indicated at 950 °C, can
proceed sluggishly between at least 900-950 °C. Another example
is the oxidation reaction which results in the formation of com-
pletely oxidized cerium phase CeTaO^.so (350 °C) from CeTaOt,. or from
any series 3 phase. This reaction proceeds over periods of at
least several days. Indeed, one sample of CeTaOi^ was maintained
at 350° for 61 days to insure formation of a completely oxidized
material which could be used for further experimentation. Phases
in series 2, 3, and 4 will transform to series 1 phases when reacted
above 1000 °C. Once CeTaO^ is decomposed (series 1, < 1265 °C),
further reactions at lower temperatures, of course, are precluded.
The complex reactions observed for the CeTaO^+x system provided
the impetus to investigate similar Ta- and Nb-containing materials.
The phases Pr3+TaOit and Ce3+NbOi+ were reported by Bodiot . PrTaO^
was found to have an x-ray diffraction powder pattern (unindexed)
similar to that of CeTaOit. CeNbOit, however, has the monoclinic
(M)-fergusonite structure (see, for example, ref. 16) at room
14
-------
temperature. These materials were also prepared attempting to vary
their Ln^V^n3"1" contents by appropriate annealing at low tempera-
tures. PrTaOit is isostructural with CeTaOit, and the magnetic and
crystallographic properties of the material have been characterized.
In air, PrTaOi+ does not undergo the reaction series shown by CeTaO^.
This can be attributed to the difficulty of oxidizing Pr3+ to Pr4"1".
Indeed, P^OH (= PrO^sss) i-s ^e highest oxide that is stable in an
air environment. PrTaOit+x might be prepared at much higher oxygen
pressures. As PrTaOi+ is isostructural with CeTaO^ and because it does
not oxidize, it is conceivable that this component can be utilized to
stabilize CeTaO^+x phases at specific x values and at lower temperatures,
For example, series 3 CeTaOm-x Phases are stable only between 600 °C
to about 900 °C, but may be quenched to room temperature. If these
materials are subjected to high oxygen activities at lower tempera-
tures, complete oxidation to CeTaOi+.so ultimately results. Suppose
PrTaOif is introduced in solid solution with CeTaO^, to give, for
example, the composition (Ce^sPr^sJTaOit. If it were possible
to completely oxidize this phase the maximum oxygen content would be
(Ce&tsPrtJtsJTaO^s, * = 0.25, but not (CefttsMts) TaO^ .5 . The
stoichiometry, when x = 0.25, however, is never encompassed by the
x parameter variation within any given CeTaO^+x series. If phases
with x > 0.25 cannot be obtained, then x parameters typical only
of series 3 phases remain possible. Single-phase (Ce^sPrjj^sJTaO^
was prepared and annealed at temperatures < 900 °C in air. Only
series 3 type phases were obtained.
The Ce^/Ce3* content in CeNbO^ (not isostructural with
CeTaOtJ also can be varied according to preliminary data. Further-
more, x-ray diffraction data suggest that more than one reaction
series is operative. One series appears to be characterized by
oxidation together with exsolution of Ce02 from CeNbO^. Such a
process would yield materials having variable Ce^Vce3"1" contents
and vacancies on the Ce sublattice.
15
-------
SECTION VI
ELECTROCHEMICAL MEASUREMENTS
INTRODUCTION
Potential materials for the oxygen electrode in fuel cells
must be screened, under conditions similar to those encountered in
a real cell, for both stability and catalytic activity. Although,
ultimately, any electrode found must be tested by construction
of an actual cell, there are simpler methods of screening avail-
able that can be used to select the most likely candidates. Such
techniques were initiated during the course of this study.
Two basic approaches to the electrochemical characterization
of oxide electrodes were utilized. The primary method was to
subject the samples to a potentiodynamic scan ranging from the
hydrogen evolution potential to anodic potentials greater than the
expected potential for oxygen evolution. This was performed with
the sample in electrolytes similar to those used in real fuel cells.
If a cathodic current were to be observed in the potential range of
oxygen reduction, the possibility that the current was due to chemical
instability would have to be checked before it could be ruled out.
A secondary, galvanostatic, method was to observe the potential at
various fixed currents on both the anodic and cathodic side of
the equilibrium potential for several oxidation-reduction couples
dissolved in varying electrolytes.
EXPERIMENTAL ARRANGEMENTS
Description of Experimental Set-up
As shown in Figure 1, the experimental arrangement is straight-
forward. Minor details such as a compensator for the recorder and
switching circuits for AC work have been omitted. A detailed
discussion of the cell and of the electrode format used will follow
later in this section.
The potentiostat is a commercial unit capable of providing
slightly more than 100 ma. Its control voltage is provided by the
output of the control voltage amplifier which is driven by the
16
-------
Potentiostat
CE Ref.
Cell
Control Voltage
Amplifier
WE
Differential
Amplifier
I Ramp
Generator
2-Pen
Recorder
X - Y
Recorder
C R 0
Switch positions
!• Potentiostatic
2»Galvanostatic
Figure 1. Block diagram of electrochemical instrumentation.
-------
ramp generator. The ramp generator consists of a motorized multi-
turn potentiometer wired as a voltage divider. Kamp rates range
from 1.6 x 10~3 V/sec to 8 x 10~2 V/sec. Adjustment of the gain
of the control voltage amplifier gives the capability of rates
up to 1.6 V/sec.
Description of Cells Used
Depending on the choice of sample electrode, two types of
cells were used. The first, used with electrodes of Type I
(described later) is shown in cross section in Figure 2. This
cell consists of four interconnected compartments. The sample-to-
counter electrode connection is through a fritted glass disc. The
two remaining compartments are for a Pt and a calomel reference
electrode, respectively. Provisions were made in the sample
container for stirring and for bubbling gasses over the sample.
The sample is electrically connected to the system by a threaded
stainless steel rod which is protected from the electrolyte by
glass tubing sealed to the sample by a Teflon gasket, in the manner
of Stern and Makrides
The second type of cell which was used, is intended for Type
II electrodes (described later). Essentially, it was a beaker
contained in a water bath. The entire assembly was enclosed in
an insulated box with feed-throughs for electrical contact and
gas input. The counter electrode was platinum black and of
Type II, and the reference electrode was either a smooth Pt strip
or a calomel electrode. For most of the experiments, the reference
was placed in a compartment connected to the electrolyte near
the sample by a capillary.
Electrodes
For convenience, in this report the two kinds of electrodes
used are labelled as Type I and Type II. Type I electrodes (shown
in cross section as part of Fig. 2) were made from poly- or mono-
18
-------
Counter Electrode
(Pt)
Stainless Steel
Teflon
Reference Electrode
(Pt)
Fritted
disc
Figure 2. Electrochemical cell used at room temperature with Type I
electrodes. This is a cross section of three of the four
compartments. Compartment 4, which contains the Calomel
reference, is situated behind the compartment for the Pt
reference. A type I electrode is also shown in cross
section. The heavy black line, connecting sample and Cu
block, denotes a Pt wire for electrical contact.
19
-------
crystalline samples or from sintered powders. A Ft wire lead was
attached to the rear surface of the sample by appropriate means
(indium soldering, peening, silver paste, etc.), and the other
end of the lead was soldered to a Cu block about 1 cm on each edge.
This assembly was placed in a mold made from Al foil and cold-
setting epoxy mounting compound was poured in. The mold was then
placed in an evacuated chamber in order to remove air from any
pores and allow them to be filled by the epoxy. After a few
minutes, the mold was returned to atmosphere and allowed to set
for 24 hours. When set, the Al foil was stripped off and the
mounted sample was ground to a convenient shape. A hole was drilled
through one end into the Cu block, and this was then threaded to
receive the stainless steel rod previously mentioned. The sample
surface was then ground with 600 grit.
Type II electrodes (Fig. 3) were made from powders and are
porous. These electrodes were manufactured by a modification of
18
the method of Neidrach and Alford . The substrates for Type II
electrodes were strips 1.5 cm by 15 cm of either electrodeposited
Ni or stainless steel screen, 0.017 cm thick with approximately
20% open area. This was coated with a 0.005 cm thick layer of
Teflon film except for two regions. At one end, about 5 cm were
left uncoated for purposes of electrical connection to the
measuring instruments. Close to the other end, an area of about
0.5 cm by 1.5 cm was left clear on one side of the strip. A paste
composed of the sample powder mixed with #30 Teflon emulsion was
placed on this area and covered with Al foil. The electrode was
hot pressed at 320 °C under an applied load of 2 x 103 kg for
2 minutes. When cool, the Al foil was removed by immersion in
6 mol/1 NaOH.
Table 1 shows the electrochemical systems that were examined.
20
-------
Ni
•Teflon film
Sample
Teflon 30
Figure 3. Type II electrode shown in front view and cross section.
-------
Table 1. ELECTROCHEMICAL EXPERIMENTS
Material
Ti02:5% Ta
TiO2:5% Ta
TiO2:H2 reduced
Electrode Type
I
II
I
Study Method
potentiodynamic
potentiodynamic
potentiodynamic
Ti02 : 0 . 1% Nb
La
(nom.)
g
La0.75Ti°3
La0.75Ca0.25Ti03
SrTiO3:0.03% Nb
SrTiO3:0.15% Nb
Ti02 as grown
Ni
Pt black
I
I
I
I
II
II
II
galvanostatic
potentiodynamic
galvanostatic
potentiodynamic
galvanostatic
potentiodynamic
galvanostatic
potentiodynamic
potentiodynamic
potentiodynamic
galvanostatic
potentiodynamic
potentiodynamic
galvanos tati c
potentiodynamic
galvanostatic
potentiodynamic
galvanostatic
potentiodynamic
galvanostatic
potentiodynamic
potentiodynamic
22
-------
EXPERIMENTAL METHODS
Potentiodynamic Measurements
Titanates -
Voltammetric scans of the titanates were done in both 5M and
concentrated phosphoric acid electrolytes with voltage scan range
of the order of 10 volts. Normal procedure was to sweep the potential
at a linear rate from values more negative than required for
hydrogen evolution to those more positive than needed for oxygen
evolution. This was done at room temperature and at 60 °C nominal.
On the basis of behavior, the samples studied may be separated
into three groups. The first of these, (SrTiOs :0 .03% Nb and
SrTi03:0.15% Nb) is soluble in phosphoric acid and will not be
discussed in this section. The second group (TiO£ as grown, Ti02:
5% Ta, Ti02:0.1% Nb, Ti02:H2 reduced) gave I-V curves typified by
Figure 4. Starting at a potential yielding good evolution of
H£ and scanning in an anodic direction, a peak due to oxidation
of H2 was first observed followed by a region of zero current. At
about 2.5 volts positive vs. S.C.E., a sizeable current began to flow.
On the reversed sweep direction there was no observable negative
current until the potential for H2 evolution was reached. Bubbling
of oxygen gas over the electrode yielded no detectable response.
The third group of samples (LaQ.ysTiOa, Lao.g2Ti03»
Lag .ysSrg .25T^-°3 ' an<^ LaO.75CaO.25Ti°3 ~ aH values nominal) gave
a family of I-V curves (Figure 5) that appear to depend on the
value of x in the generalized formula LaxTiO3. (Addition of Sr 2
or Ca+2 yields an effective new x value.) Starting with Lag 751103
fairly large anodic currents were drawn at potentials on the order
of +2 volts. As indicated in Fig. 5, however, these currents fell
off with continued cycling of the potentiodynamic sweep. As x was
increased, there was a shift toward higher anodic potentials for
23
-------
+ 0.5
x
o.
Q.
O
t CM
5 E
Z o
UJ X
Q <
E
H
z
UJ
en
oe
I>
o
ITa
I
I
I
-2
0 +2
POTENTIAL (vs SCE)
VOLTS
Figure 4. Current-potential curve for 95
at room temperature. Type
-: 5 Ta Q in 5 mol/1 H PO
I electrode configuration.
-------
TU.O
CM
O
X
<
E
K
0.
0.
o
^^ o
t
'
® La_KTiO, Run 1
.75 3
© La__TiO, Run 4
.75 3
® La.75Sr25Tl°3
^ L T 0
1 l i i i
-2024
POTENTIAL (vs Pt) —VOLTS
Figure 5. Current-potential curves for the Lanthanum Titanates at
room temperature. Type I electrodes in 5 mol/1 H PO .
-------
current flow, until x = 0.92 where the observed current approached
zero. There was no visible evolution of oxygen, and bubbling of
oxygen over the electrode gave no response.
Tungsten Bronze -
A Nag 7WC>3 single crystal was subjected to the same conditions
as above. Starting at the hydrogen evolution potential, a large
negative current was observed. As the scan went toward anodic
potentials the positive current peak due to oxidation of hydrogen
appeared (much larger than in the case of the titanates). The
current then fell to zero and as the potential went positive, the
surface of the electrode changed color and after a short positive
transient the current returned to a low value and remained low.
There was no indication of response to oxygen.
Barium Ruthenate -
Barium ruthenate, BaRu03, was the subject for the major portion
of the experiments. Potentiodynamic scans were performed with
samples in a number of different electrolytes at both room temperature
and 60 °C. The electrolytes used were concentrated phosphoric acid,
6 mol/1 NaOH, 1:1 H3B03-Na2B407, and 4 mol/1 NaOH with H2C>2 additions.
Although there was a great deal of scatter in the magnitudes
of the currents observed at particular potentials, the overall shape
of the curve for each BaRuOs electrode was essentially the same.
Figure 6 shows a typical result for concentrated phosphoric acid
at 60 °C. Visual observation of the electrode showed both
hydrogen and oxygen evolution. Oxygen evolution began at
potentials less anodic than the peak at + 0.9 V. In addition,
visual observation indicated dissolution of the BaRuOs as current
flowed. This is also evident in the I-V curves, since the observed
currents decrease linearly with number of scans (ranging from
hydrogen evolution to oxygen evolution). The scatter in the
current for a fixed potential scan was about one order of magnitude
when all electrodes were compared.
26
-------
nom. ar«a«5.l cm*
I
+40
+ 20
o
E
o •
u
-20
-2 0 +2
Potential (Volts)
Figure 6.
Current-potential curves for a Type II BaRuO electrode in
concentrated phosphoric acid at 60 °C. The decrease in
current with time is probably related to dissolution of
the sample.
27
-------
Observed responses to oxygen bubbling vary with individual
electrodes. Some electrodes exhibit a change (Figure 7) while
others appear inert.
Voltaimnetry in 1:1 saturated 113603 :Na2Bk°7 electrolyte at
room temperature differs from both acidic and alkaline (discussed
later) media. Figure 8 shows the results over a limited sweep
range. A wider potential sweep shows that the results of the
limited sweep illustrated are extended to greater potentials with
only small changes in shape. Interestingly/ the sharp changes in
slope associated with hydrogen and/or oxygen evolution, are missing.
Also missing are the peaks observed in acid or alkaline media.
Figure 9 shows the results of voltammetry for two electrodes in
6 mol/1 NaOH at 60 °C. Note that the presence of 02 gives rise to
fairly significant changes. The observed shape of the curves is
different from that observed in concentrated phosphoric acid. Other
electrodes yielded similar responses although in one case the
reaction to oxygen addition was small. It should be noted that the
overall shape of these curves is similar to that observed for RuO2
on Ti in 4 mol/1 NaOH, by O'Grady, et al19. Because of this similarity,
the effects of additions of 0.007 mol/1 H202 were studied for a 4 mol/1 NaOH
electrolyte on BaRuOs electrodes. Figure 10 shows the effect as
compared to a smooth Pt electrode. Note that there is only a small
increase in cathodic current at potentials less than about -0.2 V
for BaRuOs in the presence of ^02, and no changes above that
potential. Compare with the large changes on Pt electrodes.
Finally, it should be noted that a large hysteresis in current
was observed with all media when the scan was halted. On stopping
the scan, the current rapidly fell toward zero. On restarting, the
current returned rapidly to its initial value.
28
-------
H3P04 60°C
Pt Ref.
BaRu0
p*>m. area-3.8
o
£
0
3
u
- 5
0
0.15
Potential (Volts)
0.30
Figure 7. Effect of oxygen and the current-potential curve for a
Type II BaRuO electrode.
-------
H3B°3:Na2B4°7
22°C Pt. Ref.
nom. area -4.5cm
I
-f-50
o
E
0
O
-50
-1.60
1.85
Potential (Volts)
Figure 8. Current-potential curve for a Type II BaRuO electrode in
neutral solution. Scan range limited. Curve 1 is for an
air saturated electrolyte.
electrode surface.
Curve 2 is
for 0 bubbling over
-------
6M NaOH 60°C
Pt Ret.
Argon
Oxygen
+ 5
o
E
c
0)
•-5
Potential (Volts)
Figure 9. Comparison of two BaRuO3 electrodes (Type II) in 6 xool/1 (=M) NaOH
at 60 °C.
-------
4M NaOH
Calomel Ref.
Argon
+ .007 M H202
/ Pt (smooth)
-1.18
+20
+10
o
E
3
O
-10
-20
Potential (Volts)
Figure 10. The effect of 0.007 M HO on a Type II BaRuO and on smooth Pt.
M = mol/1.
-------
Redox Measurements
In general, measurements of the performance of electrode
materials with respect to single redox reactions were made in
neutral solutions of 0.1 mol/1 K3[Fe(CN)6] +0.1 mol/1 K4[Fe(CN)6] ,
0.1 mol/1 K3[Fe(CN)6] + 0.02 mol/1 Ki+[Fe(CN) 6] , and 0.1 mol/1 KI3 + 0.5 mol/1 KE
Most of the measurements were carried out by potentiodynamic
sweeping except in the cases of Type II BaRuO3 electrodes, where,
because of their porous nature, long times at constant current or
voltage were required to attain steady state conditions.
In K3[Fe(CN)6]/K)+[Fe(CN)6] solutions for all electrodes
tested, the open circuit potential corresponded to the equilibrium
potential for the redox couple, so that the potential difference
between the working electrode and the Pt reference electrode was
at most a few millivolts. In KI/KI3 solutions, however, the
SrTi03 and TiC>2 electrodes tended to drift to -400 mV (vs. Pt)
indicating that the potential was not determined by the I3/I~
equilibrium.
With the exception of tungsten bronze in KI3/KI, all materials
were stable in these redox solutions. Some electrodes, however,
showed increasing resistance with repeated anodic oxidation,
indicating that the properties of the surface were changing,
probably because of the formation of an insulating oxide layer.
Larger currents could at times be restored by cathodic polariza-
tion to sufficiently negative potentials so that hydrogen evolution
occurred. This behavior was seen with H2 reduced TiC>2, 0.1% Nb
doped Ti02, La0.75CaO.25Ti°3' and La0.75Sr0.25TiO3 .
Titanates -
For some of these materials a pronounced dissymmetry was
observed, i.e., the anodic currents were much smaller than the
cathodic ones. Indeed, in some cases the rate of oxidation of
]1*" was negligible, as with Lao%7sTiO3 and
33
-------
Rutile (as grown) also exhibited similar dissymmetry, but the
anodic current was larger. In contrast, a fairly symmetric
current-potential curve was observed on Ta-doped TiC>2, as well as
on two samples of SrTiO3 with different amounts of Nb doping (0.03
and 0.15%).
At equal polarization, current densities varied considerably
from material to material. To give an idea of their magnitude,
exchange current densities very roughly estimated from An/Ai
(where n is the over-voltage) are reported in Table 2. In some
cases, however, the dissymmetry is such that it is impossible to
derive reliable values for the exchange current density.
Table 2. EXCHANGE CURRENT DENSITIES
Electrode
TiO3:5% Ta
SrTiO3:0.03% Nb
SrTi03:0.15% Nb
La0.92Ti°3
La0.75Ti03
La0.75Ca0.25Tio3
i (A/ cm2)
o
ID'7 •
6 x 10~6
2 x 10~6
2 x 10~9
7 x 10~8
3 x 10~7
Behavior
Symmetry
Symmetry
Symmetry
Dissymmetry
Dissymmetry
Symmetry
None of the Ti02 electrodes tested in KI3/KI solution showed
any detectable current (< IpA) around the equilibrium potential,
at least, from -300 to +700 mV. Nb doped SrTiO3 showed no anodic
current, and cathodic current only below -600 mV vs. Pt, and as
already mentioned, anomalous behavior at open circuit.
Tungsten Bronze -
Tungsten bronze, Nag 7W03 showed peculiar behavior. Around the
equilibrium potential currents were large, but on the anodic side a
limiting current of ^ 8 mA at an over-voltage (n) of about 300 mV was
reached. For polarizations higher than 0.6 V the current rapidly
decreased to 500 MA or less (for n of the order of 1.2V) and further
34
-------
increases in potential had no effect. It is unlikely that this
behavior was due to chemical changes on the surface layers , but
could have been due to the formation of a depletion layer. The
exchange current density close to the equilibrium potential was fairly
large, and can be estimated at about 4 mA/cm2 . This corresponds to a
rate constant of 4 x 10""1* cm/sec, in good agreement with the values
of Amjad and Fletcher in acidified solution (0.4 mol/1 I^SQ) . As
already mentioned, in KI3/KI, the tungsten bronze dissolved; (the
dissolution reaction seemed to occur in a regular fashion, leaving a
flat and smooth surface) . Because of this , the results are difficult
to interpret, but they are consistent with the existence of a cathodic
limiting current due to charge-carrier depletion on the electrode .
Barium Ruthenate -
Barium Ruthenate -
Steady-state data for a BaRu03-Type II electrode in two con-
centrations, 0.1M/0.1M and 0.1M/0.02M K3 [Fe (CN) &] /Kt+ [Fe (CN) 6] ,
are reported in Figures 11 and 12 , where M = mol/1 . It is evident
that a limiting current was reached on the cathodic side which was
not due to transport in solution; (as a further indication, stirring
had no effect on the limiting current) . The effect was probably due
to limited supply of charge-carriers in the electrode. The exchange
current for the redox reaction was of the order of 500 yA for the
first solution and 200 yA for the second. The current density
cannot be established because the surface area is not known.
The behavior of BaRuC>3 with respect to the [Fe(CN)6]^~/
[Fe(CN)g]1+~ redox reaction is very interesting when compared with
the behavior in a O.lM KI3/0.5M KI solution. The results, for the
latter system, shown in Figure 13, indicate that on the cathodic
side the limiting current was of the order of 3 to 4 mA, a value
consistent with diffusion control. The current, therefore, was not
limited by charge carrier supply on the electrode side .
35
-------
-1-300
E
I
UJ
o
<
l-
o
>
o:
u
-300
BoRuO, Electrode Pt Ref.
K,[Fe(CNL] /K.tFe(CN).]
3646
- • O.IM/O.IM
--• O.IM/0.02M
0.05 O.I 0.2 0.5 I 2
CURRENT — mA
Figure 11. Ferro-ferricyanide reaction. Steady-state performance of
Type II BaRuO, electrode. Plot of overvoltage versus
logarithm of current. M = mol/1.
-------
+ 200 —
1
UJ
o
o
oc
UJ
o
-200
Electrode
Reference
K_[Fe(CNV]/K tFe(CN) J
o 4 o
O.I M/O.I M
• O.IM/0.02M
CURRENT
-I
mA
Figure 12. Ferro-ferricyanide reaction. Steady-state performance of
Type II.
M
mol/1.
BaRuO electrode.
Plot of overvoltage versus current.
37
-------
o
>
£•
Id
O
<
cc
UJ
-I
-2.0
BoRuO, Electrode
Pt
Reference
O.IM K I3 /0.5M K I
0.2
0.5 1.0 2
CURRENT— mA
10
Figure 13. Iodine-iodide reaction. Steady-state performance of Type II
BaRuO- electrode. Plot of overvoltage versus logarithm
of current. M = mol/1.
-------
DISCUSSION
The electrochemical measurements carried out to screen the
materials for electrocatalytic properties have been perhaps more
extensive than anticipated at the beginning of the contract, since
less time than expected was used in readying the necessary instru-
mentation. Nevertheless, for the purpose of both forming a
systematic picture of electrochemical behavior as related to
chemical composition, and of investigating the reaction mechanisms
occurring on non-metallic electrodes, the research done so far
must be considered very preliminary, and few conclusions can be
drawn at this point.
The following discussion, therefore, will be aimed at out-
lining a few general conclusions that emerge from the results, but
even more at describing the research problems that have been
uncovered, so as to help in directing future efforts in the field.
A first conclusion that can be drawn concerning experimental
techniques is that electrodes produced from powders hot-pressed
with binders tend to be porous with considerable disadvantages for
the achievement of reliable data in a reasonable time: a large
amount of the current observed on these electrodes during
21 22
potentiodynamic scanning can be attributed to capacitative effects '
Even more troublesome than their long relaxation times is the
difficulty of reproducible fabrication. For these reasons it is
suggested that great effort be made to produce massive specimens as
electrodes, and alternatively — or better, concurrently — to find
methods to make electrodes from powders more reproducible and
suitable for laboratory research.
As far as stability to hot, concentrated HaPOit is concerned,
a number of materials have shown increasing electrical resistance,
that is, lower currents for the same scanning voltages particularly
at more positive potentials. This indicates slow deterioration
39
-------
of the surface properties, probably with formation of insulating
oxides. These results point to the need to test for chemical
stability under realistic polarization conditions.
Some of the materials tested, such as the various La titanates
and BaRuC>3, exhibit metallic conductivity. However, in the elec-
trochemical studies semiconducting properties have been observed.
This is particularly evident in solutions containing redox couples,
where, in many cases, the current tended to flow dissymmetrically
in anodic and cathodic directions for reasons that appeared to be
due to saturation of the charge-carrier current within the electrode.
Asymmetric behavior was also observed for some titanium oxide
electrodes in phosphoric acid: on rutile (as grown) as well as
0.1% Nb doped Ti(>2 no significant anodic current was detected.
Whether this occurs because of surface depletion of electrons
(these materials are n-semiconductors) or because of formation of
an insulating phase is not clear. The situation is further confused
since anodic current and oxygen evolution were obtained on 5% Ta
doped TiC>2 and on H2 reduced TiC>2, which are supposed to possess
similar semiconducting characteristics.
As shown in Figure 5, the anodic behavior of the lanthanum
titanates in HsPO^ seems to show an increase in anodic current with
decreasing concentration of lanthanum, which corresponds to an
increase in the valency of Ti from 3 (for LaTiC^) to 4 (for
Lao.67Ti°3- This could be attributed to a more pronounced
n-semiconducting behavior for the more electron-rich titanate.
However, if the effect is related to the formation of a depletion
layer, the n-character appears to be a complicated function of
composition, since the nominal valency of Ti is the same in
La0.92Ti°3 and in Lao.75Cao.25Ti°3/ but the anodic current is
larger on the latter. An alternative explanation might be based
on a greater ease in forming an insulating layer (possibly TiC>2)
by anodic oxidation in the case of lower valency titanate.
40
-------
Most of the materials under investigation have also been
tested as electrodes for two single redox reactions K3[Fe(CN)g]/
KijFeCCNjg] and KI3/KI. The results obtained with BaRuO3 and
sodium tungsten bronze have been described in more detail and are
quite interesting. Saturation currents are of opposite sign
(cathodic for BaRuOs and anodic for Nao.75^103) . Since BaRuC>3
exhibits metallic conductivity, the semiconducting behavior might
be explained by assuming that BaRuOs is a p-metal, and cathodic
bias is sufficient to deplete holes at the surface. No saturation
currents could be detected for BaRuOs in KI3/KI. The reason might
be linked to the equilibrium potential being more positive in
KIs/KI (+286 mV vs. SCE) compared with the ferri-ferrocyanide
couple (+285 mV for the 0.1M/0.1M solution and +235 mV for the
0.1M/0.02M), but the differences in equilibrium potential seem
rather small compared with the substantial (several tenths of a
volt) polarizations applied, although it is not known how the
potential difference is partitioned inside and outside the electrode.
The W bronze, which should tend to exhibit n-semiconductor
properties, appeared to develop an exhaustion layer by anodic
polarization in ferri-ferrocyanide, so that the current dropped
by increasing bias. The other redox system used was not suitable
since the bronze underwent dissolution, probably caused by hole
injection on the part of the reducing iodine.
More redox systems have to be tested before a meaningful
correlation with the oxygen reaction can be attempted, but it
appears that materials on which little if any current can be drawn
in H3PO4 exhibit similar behavior in ferri-ferrocyanide. The
significance of the behavior in KI3/KI is much more confused,
since for many of the materials tested even the equilibrium
potential is not determined by the I3~/I~ equilibrium and no
current is obtained.
41
-------
SECTION VII
MATERIALS OTHER THAN OXIDES
INTRODUCTION
In agreement with molecular orbital theory, the reduction of
oxygen is easy and that of nitrogen difficult. Biochemical reduc-
tion of nitrogen, however, takes place under mild conditions, the
final product being ammonia. The ordering of molecular orbitals in
nitrogen and oxygen is very similar, leading to the expectation that
similarities between the reduction of oxygen and nitrogen may exist,
and that the nitrogen reductase enzyme (nitrogenase) might serve as
a useful model for compounds capable of catalyzing the reduction of
oxygen at electrode surfaces. The active site of nitrogenase
consists of a molybdenum-nonheme iron-sulfur array in which molyb-
denum is thought to be involved in the final electron transfer
steps. In order to pursue this analogy further, the synthesis
of ternary metal sulfide arrays, in which one of the metals is
molybdenum or tungsten, was undertaken with a view to providing an
alternative approach to the central thrust of this program.
RESULTS
1. Aqueous Systems. Metathetical reactions in water, e.g.,
Fe + Mos£~ »- FeMoSi+
failed and the products were ferrous sulfide and molybdenum disul-
fide. Analogous reactions were attempted with dithiotungstate(2-),
W02si~» i° the hope that the transition metal oxythiotungstates
would be more stable. However, unusual solvolysis reactions
supervened and sulfur-free products were isolated, e.g.,
+++ o_
Cr + WO2S5 *• Cr2 (WOi+J 3-3H2O
A publication describing these results is in press. In order to
avoid complications introduced by reactions with water, nonaqueous
syntheses were studied.
42
-------
2. Nonaqueous Sytems. In order to solubilize transition
metal ions in organic solvents, it is usually necessary to employ
large, hydrophobic ligands. Triphenylphosphine was chosen for use
in initial experiments, primarily because of its ready availability.
Use of tertiary phosphines confers another potential advantage,
for these ligands are often labile. Thus dissociation in solution
of the phosphine may uncover a coordination site on a metal atom
to produce a species, in equilibrium concentration, too reactive
to be isolated.
A simple metathetical reaction between (PhsP)sCuCl and
(Ph3PCH3)2WO2S2 in dichloromethane was attempted. The product,
(Ph3P)i+Cu2W2S6, isolated by chromatography on silica gel, was
oxygen-free. This suggested that triphenylphosphine could serve
as a reducing agent to remove oxygen and sulfur from W02s£~
to produce the desired ternary metal sulfides; the by-products are
Ph3PO and Ph3PS.
This copper-tungsten-sulfur complex has been studied by a
variety of physical techniques and some of its physical properties
are tabulated below:
yellow-orange crystalline solid, soluble in chlorinated
hydrocarbons
mp. 227-230° (dec.) (under vacuum)
{1H}31P NMR: broad singlet at -8.1 ppm from K^POi,
^•H NMR: complex multiplet centered at -7.5 ppm
IR: 448 cm"1(v )
ws
Raman: 467(s), 254(w)
Electronic spectrum: X at 440, 267 nm
max
E (oxidation) - 1.14 V; E (reduction) - 0.94, - 1.5 V (vs.
P P
SCE, irreversible; by cyclic voltametry).
The general synthetic route (metal ion, triphenylphosphine,
and W02s£~) has been extended to other coinage metals. The
Ag-W-S system has proved to be particularly instructive. The
43
-------
initial product is a yellow solid whose elemental analysis
corresponds to (PhsP)eAgsV^SgC^. Attempts to further purify this
product led to its decomposition into a mixture of an orange and
a yellow material. The elemental analysis of the yellow material,
separated by fractional crystallization, corresponds to
[(Ph3?)3Ag]2W2Sg. It is converted on boiling in dichloromethane-
acetone to the orange material, whose analysis corresponds to
(PhaP)aAgaWSs. It is possible that the orange compound is not
a single phase and this question is being pursued. These observa-
tions suggest that the initial product in the Ag-PhsP-WC^S^"
system is a complex oxythiotungstate which undergoes further
reaction on the acidic sites on silica gel to form a V^Sg complex.
Using PhsPAuCl as the starting material, the only product
isolated was (PhsPAu)2^284. Like the related copper complex, this
material had a molecular weight half that of the expected value,
suggesting that the phosphine ligands are indeed labile. Only in
the case of gold was a new product formed when triphenylarsine was
used as a reducing agent. It appears to be a simple tetrathia-
tungstate(2-) derivative, (PhsAsAu) 2WSit.
The physical techniques employed to date do not provide an
adequate structural characterization of these ternary metal
sulfides, which are new and quite novel materials, and x-ray
diffraction studies are needed. The triphenylphosphine complexes
generally have low molar solubilities, making it difficult to
grow crystals and those that are obtained are usually thin flakes.
Preparation of tri-p-tolylphosphine and diphenylmethylphosphine
analogues is currently underway. Good single crystals of the
diphenylmethylphosphine-gold-tungsten sulfide have been obtained.
o o
They are monoclinic with a = 15.46 A, b = 13.51 A, and 3 = 96°.
An x-ray study of this material has just begun. Some single
crystals of orange (PhaP)3AgsWS3 have also been grown and cell
parameters are now being measured. These more soluble derivatives
44
-------
will be useful for studies of 31P NMR spectra, since the triphenyl-
phosphine complexes are so poorly soluble that exhorbitant amounts
of spectrometer time are required to obtain even marginal data.
Materials characterization has now proceeded sufficiently far
that applications in real systems can be considered. Arrangements
have been made for fabrication of composite electrodes containing
(Ph3P)i+CU2W2S2 and graphite. These will be examined in a working
cell to test for catalytic activity in oxygen reduction.
A limited effort has been made to prepare ternary arrays
containing nickel, palladium, or platinum. Metathesis between
WC^S^" and (diphos)NiCl2 is straightforward and (diphos)NiWC>2S2
is produced. This material appears to be a mixture of geometric
isomers. Similar reactions with (diphos)PdCl2 and cis-
(Ph3P)2PtCl2 led to (diphos)2Pd3W2S6O2 and (Ph3P)3PtW2S602.
These yellow microcrystalline materials are virtually insoluble
in nonreactive solvents and their characterization is extremely
difficult [diphos = 1,2-bis(diphenylphosphino)ethane].
45
-------
SECTION VIII
REFERENCES
1. Gillis, A. E. Advanced Development. In: Seventh Status
Report on Fuel Cells, Huff, J. R. (ed.). U.S. Army Mobility
Equipment Research and Development Center, Fort Belvoir, Va.
Report 2039. NTIS No. AD-755 106. October 1972. 174 p.
2. Negas, T., and R. S. Roth. J. Solid State Chem. 1_: 409-418,
1970.
3. Negas, T., and R. S. Roth. J. Solid State Chem. 3_:323-339,
1971.
4. Negas, T. J. Solid State Chem. 6_: 136-150, 1973.
5. Negas, T., and R. S. Roth. Proceedings of 5th Materials
Research Symposium, National Bureau of Standards, Washington,
D.C. NBS Special Publication No. 364. July 1972. 31 p.
6. Candela, G. A., A. H. Kahn, and T. Negas. J. Solid State
Chem. 7^:360-369, 1973-
7. Negas, T. J. Solid State Chem. 7_:85-88, 1973.
8. Kestigian, M., and R. Ward. J. Am. Chem. Soc. 77_:6199-6200,
1955.
9. Donohue, P. C., L. Katz, and R. Ward. Inorg. Chem. 4_: 306-310,
1965.
10. Callaghan, A., C. W. Moeller, and R. Ward. Inorg. Chem.
5^:1573-1576, 1966.
11. Bouchard, R. J., and J. L. Gillson. Mat. Res. Bull.
7_: 873-878, 1972.
12. Waring, J. L., and R. S. Roth. J. Res. NBS. 27A;175-186,
1968.
13. Waring, J. L., and R. S. Roth. J. Res. NBS. 69A:119-129,
1965.
14. Nicks, L. J., and D. J. MacDonald. Report of Investigations
7841, U. S. Department of Interior, Bureau of Mines,
Washington, D.C. 1973. 9 p.
46
-------
15. Bodiot, D. Revue de Chemie Minerale. 5_: 569-607, 1968.
16. Graham, J. Am. Mineral. 59_: 1045-1046, 1974.
17. Stern, M., and A. C. Makrides. J. Electrochem. Soc.
10_7:782, 1960.
18. Neidrach, L. W., and H. R. Alford. J. Electrochem. Soc.
112_: 117-124, 1965.
19. O'Grady, W., C. Iwakura, J. Huang, and E. Yenger. Ruthenium
Oxide Catalysts for the Oxygen Electrode. ONR Technical
Report 37, Case Western Reserve, Cleveland, Ohio. 1974. 22 p.
20. Amjad, M., and D. Fletcher. Electro-analytical Chem. and
Interfacial Electrochem. 59_:61-67, 1975.
21. Austin, L. G., and E. G. Gagnon. J. Electrochem. Soc.
120_:251-254, 1973.
22. Gagnon, E. G. J. Electrochem. Soc. 120;1052-1056, 1973;
Ibid. 121,512-515, 1974; Ibid., 122:521-525, 1975.
47
-------
SECTION IX
LIST OF PUBLICATIONS
1. Siedle, A. R., T. Negas, and J. Broussalian. "Reaction of
Transition Metal Ions with the Dithiotungstate(2-) Ion."
J. Inorg. Nucl. Chem. In press.
48
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-007
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Mixed Oxides for Fuel Cell Electrodes
5. REPORT DATE
January 1976
6. PERFORMING ORGANIZATION CODE
. Bertocci, M. Cohen, W.S. Horton,
T. Negas, and A. R. Siedle
8. PERFORMING ORGANIZATION REPORT NO
NBSTJR 75-742
3. PERFORMING OR9ANIZATION NAME AND ADDRESS
National Bureau of Standards
Department of Commerce
Washington, DC 20234
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21BKR-007
11. CONTRACT/GRANT NO.
IAG D4-0528
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 5/74-5/75
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES Project Officer Bunas is no longer with EPA: direct questions to
G.L. Johnson, Mail Drop 63, 919/549-8411, Ext 2815.
16. ABSTRACT
The report gives results of studies to determine if mixed oxides can act as
oxygen-reducing electrocatalysts in an acid fuel cell. Choices were based upon pro-
ducing variable valence and upon conferring stability at elevated temperatures (less
than or equal to 150 C) in phosphoric acid. Barium ruthenate and the systems Ti-Ta-
O, V-Nb-O, V-Ta-O, Ce-Ta-O were hot-acid stable. Potentiodynamic and galvano-
static studies are reported on several materials. Preparation of inorganic com-
pounds with ternary metal-sulfur arrays similar to the arrays in nitrogen reductase
was attempted: several were prepared.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATl Field/Group
Fuel Cells
Electrodes
Transition Metals
Oxides
Catalysis
Electrolysis
Oxygen
Reduction (Chemistry)
Phosphoric Acids
Tungsten Containing
Alloys
Electrocatalysis
Mixed Oxides
Ternary Metal-Sulfur
Arrays
Triphenylphosphine
10B
09A
07B
07D 11F
3. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
57
Unlimited
20 SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |