Solid acid proton conductors: from laboratory curiosities to fuel cell electrolytes

TLDR: The result indicates that enhancing power output beyond the present levels will require improving cathode properties rather than further lowering the electrolyte thickness, and a discussion of the entropy of the superprotonic transition and the implications for proton transport is presented.

Abstract: The compound CsH2PO4 has emerged as a viable electrolyte for intermediate temperature (200–300 °C) fuel cells. In order to settle the question of the high temperature behavior of this material, conductivity measurements were performed by two-point AC impedance spectroscopy under humidified conditions (p[H2O] = 0.4 atm). A transition to a stable, high conductivity phase was observed at 230 °C, with the conductivity rising to a value of 2.2 × 10−2 S cm−1 at 240 °C and the activation energy of proton transport dropping to 0.42 eV. In the absence of active humidification, dehydration of CsH2PO4 does indeed occur, but, in contradiction to some suggestions in the literature, the dehydration process is not responsible for the high conductivity at this temperature. Electrochemical characterization by galvanostatic current interrupt (GCI) methods and three-point AC impedance spectroscopy (under uniform, humidified gases) of CsH2PO4 based fuel cells, in which a composite mixture of the electrolyte, Pt supported on carbon, Pt black and carbon black served as the electrodes, showed that the overpotential for hydrogen electrooxidation was virtually immeasurable. The overpotential for oxygen electroreduction, however, was found to be on the order of 100 mV at 100 mA cm−2. Thus, for fuel cells in which the supported electrolyte membrane was only 25 μm in thickness and in which a peak power density of 415 mW cm−2 was achieved, the majority of the overpotential was found to be due to the slow rate of oxygen electrocatalysis. While the much faster kinetics at the anode over those at the cathode are not surprising, the result indicates that enhancing power output beyond the present levels will require improving cathode properties rather than further lowering the electrolyte thickness. In addition to the characterization of the transport and electrochemical properties of CsH2PO4, a discussion of the entropy of the superprotonic transition and the implications for proton transport is presented.

Figures

Fig. 1 Schematic impedance spectra resulting from electrode | electrolyte | electrode systems (a) at low temperatures and (b) at high temperatures.

Fig. 6 Comparison between as-measured and membrane resistance (IR) corrected polarization curves for the fuel cell shown in Fig. 5.

Fig. 7 Configuration for three-point electrical measurement (a) schematic and (b) photo of cell showing working and reference electrodes.

Fig. 12 Crystal structure of CsH2PO4 in its room temperature, paraelectric form, taken in space group setting B21/m. 44 The disordered hydrogen bonds formed by the H(2) atoms create a 1-dimensional chain of phosphate groups along [010], in turn, linked by the ordered hydrogen bonds formed by H(1) to generate a layered structure.

Fig. 2 Equilibrium phase diagram for the decomposition of CsH2PO4. Reproduced with permission from ref. 28.

Fig. 5 (a) Schematic of thin-film fuel cell and (b) resulting polarization and power density curves for operation under humidified (p[H2O] = 0.3 atm) H2/O2 at 240 1C. 36 A peak power density of 415 mW cm 2 was obtained. Reproduced with permission from ref. 36.

Full text

Solid acid proton conductors: from laboratory
curiosities to fuel cell electrolytes
Sossina M. Haile,* Calum R. I. Chisholm,w Kenji Sasaki,
Dane A. Boysenw and Tetsuya Udaz
Received 24th March 2006, Accepted 4th May 2006
First published as an Advance Article on the web 7th August 2006
DOI: 10.1039/b604311a
The compound CsH
2
PO
4
has emerged as a viable electrolyte for
intermediate temperature (200–300 1C) fuel cells. In order to settle the
question of the high temperature behavior of this material, conductivity
measurements were performed by two-point AC impedance spectroscopy
under humidified conditions (p[H
2
O] = 0.4 atm). A transition to a stable,
high conductivity phase was observed at 230 1C, with the conductivity rising
to a value of 2.2  10
2
Scm
1
at 240 1C and the activation energy of
proton transport dropping to 0.42 eV. In the absence of active
humidification, dehydration of CsH
2
PO
4
does indeed occur, but, in
contradiction to some suggestions in the literature, the dehydration process
is not responsible for the high conductivity at this temperature.
Electrochemical characterization by galvanostatic current interrupt (GCI)
methods and three-point AC impedance spectroscopy (under uniform,
humidified gases) of CsH
2
PO
4
based fuel cells, in which a composite mixture
of the electrolyte, Pt supported on carbon, Pt black and carbon black served
as the electrodes, showed that the overpotential for hydrogen
electrooxidation was virtually immeasurable. The overpotential for oxygen
electroreduction, however, was found to be on the order of 100 mV at
100 mA cm
2
. Thus, for fuel cells in which the supported electrolyte
membrane was only 25 mminthicknessandinwhichapeakpowerdensityof
415 mW cm
2
was achieved, the majority of the overpotential was found to be
due to the slow rate of oxygen electrocatalysis. While the much faster kinetics
at the anode over those at the cathode are not surprising, the result indicates
that enhancing power output beyond the present levels will require improving
cathode properties rather than further lowering the electrolyte thickness. In
addition to the characterization of the transport and electrochemical properties
of CsH
2
PO
4
, a discussion of the entropy of the superprotonic transition and
the implications for proton transport is presented.
Introduction
Solid acid proton conductors, based on tetrahedral oxyanion groups, have received
attention as electrolytes in next generation fuel cells. Compounds within this class,
Materials Science, California Institute of Technology, Pasadena CA 91125, USA. E-mail:
smhaile@caltech.edu
{ Present Address: SuperProtonic Inc. Pasadena, CA 91101, USA.
{ Present Address: Materials Science and Engineering, Kyoto University, Sakyo, Kyoto 606-
8501, Japan.
PAPER www.rsc.org/faraday_d | Faraday Discussions
Faraday Discuss., 2007, 134, 17–39 | 17This journal is
c
The Royal Society of Chemistry 2006

such as CsHSO
4
,
1
Rb
3
H(SeO
4
)
2
,
2
and (NH
4
)
3
H(SO
4
)
2
,
3
exhibit anhydrous proton
transport with conductivities of the order of 10
3
to 10
2
Scm
1
at moderate
temperatures (120–300 1C). Unlike the polymers in more conventional proton
exchange membrane fuel cells (PEMFCs), proton conduction in oxyanion solid
acids does not rely on the migration of hydronium ions. Consequently, the require-
ment for humidification of the electrolyte is, in principle, eliminated as is the need for
delicate water management.
4
The temperatures of operation accessible to fuel cells
based on solid acids furthermore imply that catalysis rates will be enhanced relative
to PEMFCs, opening up possibilities for reduction in precious metal loadings or even
the elimination of precious metals entirely. These temperatures additionally imply a
high tolerance of the catalysts to poisons, particularly CO, in the fuel stream. While
these many features render solid acids very attractive as fuel cell electrolytes several
challenges must be addressed in technologically relevant fuel cell systems. Prominent
amongst these is the water solubility of all known solid acids with high conductivity,
which requires the implementation of engineering designs to prevent condensed
water from contacting the electrolyte, particularly during fuel cell shutdown.
In contrast to the application potential of solid acid proton conductors, which has
been explored over only the past five years,
4
the fundamental physical and chemical
characteristics of these materials have been studied for well over twenty years. This is
a consequence of the fascinating sequence of phase transitions that occur in these
compounds in response to heating, cooling or application of pressure. In general,
these transitions involve changes to the network of hydrogen bonds which link the
oxyanion groups to form dimers, chains, layers, or three-dimensional structures. In
the particular case of proton transport, the dynamic disordering of the hydrogen
bond network above the so-called superprotonic transition leads to a dramatic
increase in proton conductivity by several orders of magnitude.
1–3,5
Subtle changes
in the local hydrogen bond geometry similarly give rise to the well-known ferro-
electric transition in compounds such as KH
2
PO
4
6
and Cs
3
H(SeO
4
)
2
.
7
This rich
phase behavior has spawned the production of at least 500 papers that broadly
address solid acids with stoichiometry MHXO
4
,M
3
H(XO
4
)
2
,M
2
H(X
0
O
4
), or some
variation thereof, where M = alkali metal or NH
4
; X = S, Se; and X
0
= P, As.
In this work we review the scientific and technological status of selected solid
acids, with emphasis on recent developments in the authors’ laboratory in the study
of CsH
2
PO
4
. After addressing the ongoing literature controversy regarding the high
temperature properties of this material, in particular, the nature of the transforma-
tion occurring at approximately 230 1 C, we present new data supporting the position
that a true polymorphic transition occurs in this material. We then evaluate the
behavior of CsH
2
PO
4
as a fuel cell electrolyte, examining the relative rates of
hydrogen electrooxidation and oxygen electroreduction. We then close with a
speculative discussion of the configurational entropy of the high temperature phase
of CsH
2
PO
4
, in which we propose that the disorder associated with the hydrogen
bond network should be considered independently of the oxyanion group disorder.
Thus, despite its rather innocent chemical formula, CsH
2
PO
4
provides a rich variety
of scientific challenges and technological opportunities.
Phase transition behavior
The literature debate
The controversy surrounding the high temperature properties of CsH
2
PO
4
stem
from the decomposition behavior of the material. Specifically, it has been argued by
some that the dehydration of the compound
CsH
2
PO
4
(s) - CsH
22x
PO
4x
(s) + xH
2
O(g), (0 r x r 1) - CsPO
3
+H
2
O(x =1)
induces a transient rise in conductivity as water leaves the structure, but that there is
no true polymorphic transition to a high conductivity phase. Others, however, have
18 | Fa rada y Discuss ., 2007, 134, 17–39 This journal is
c
The Royal Society of Chemistry 2006

argued that, while decomposition can interfere with the observation of the poly-
morphic transition, it nevertheless occurs. A selection of the relevant papers
documenting this controversy is provided in Table 1, along with a notation
indicating whether the paper supports, refutes or remains neutral on the matter of
a superprotonic phase transition.
The early thermal analyses of CsH
2
PO
4
were very much in contradiction with one
another. The first papers on the topic of the thermal behavior of CsH
2
PO
4
(beyond
simple weight loss measurements) appear to be two reports from Rashkovich et al.,
and even these are in disagreement. The earlier paper concludes that two transitions
occur prior to decomposition,
8
whereas the latter concludes that the thermal events
are entirely due to decomposition.
9
In two papers co-authored by Clark, two
polymorphic transitions are reported for CsH
2
PO
4
.
10,11
The latter, occurring at
230 1C, was found to be fully reversible. However, it was associated with a slight
weight loss (B1.5%) for powder samples examined under ambient conditions. This
feature would become the point of significant controversy in later years. Wada
subsequently confirmed the 230 1C transition by dilatometry measurements of single
crystal samples, observing a sharp increase in lattice constants at this temperature.
12
Almost simultaneously, Gupta reported, again on the basis of calorimetry and
thermal gravimetric analysis, a polymorphic transition in CsH
2
PO
4
at 235 1C just
prior to the maximum in the decomposition process.
13
Again, however, initiation of
the weight loss coincided with the reported polymorphic transition. In contradiction
to these results, Nirsha et al. published a study two years later concluding that
thermal events at 233 1C and higher in CsH
2
PO
4
are entirely due to decomposition.
14
The matter of a polymorphic phase transition in CsH
2
PO
4
may have remained an
obscure point in the field of solid state chemistry were it not for the results of
Baranov et al. showing a so-called superprotonic transition to occur at 230 1C,
5
precisely the temperature of the reversible, higher temperature transformation first
reported by Clark.
10,11
The conductivity was shown to increase by five orders of
magnitude at the transition, and apparently reliable data were obtained to tempera-
tures of B250 1C. In hindsight, it is clear that Baranov was able to observe the
transition because single crystals, in which dehydration is slow compared to
powdered materials, were utilized for the experiments. Shortly after Baranov’s
study, Bronowska and Pietraszko reported the structure of superprotonic CsH
2
PO
4
and provided the first clear demonstration of the significance of water partial
pressure in suppressing dehydration.
15
All of the peaks in the high temperature
X-ray powder diffraction pattern, along with their relative intensities, could be
explained on the basis of the proposed high temperature structure. The subsequent
Raman study of Romain and Novak
16
supported Bronowska’s conclusions regard-
ing the structural features of the superprotonic state, while Vargas and Torijano in
1993 also agreed (initially) with the existence of a reversible, but hysteretic, phase
transition at 227 1C on the basis of differential scanning calorimetry.
17
The controversy surrounding the properties of CsH
2
PO
4
began in earnest in 1996
with a publication by Lee suggesting that the observed conductivity effects were
artifacts of thermal decomposition and partial polymerization at the surfaces of the
CsH
2
PO
4
particles.
19
The hypothesis was based on a review of literature data,
without the benefit of new experimental results. A later paper from this same author
repeated these conclusions, but in this case experimental support was provided in the
form of a limited set of optical micrographs showing the degradation of single crystal
surfaces.
24
Inspired by Lee’s work, Ortiz, Vargas and Mellander published a series of
papers also taking the view that only decomposition occurs at the supposed super-
protonic transition.
21,22,33
In this case, the conclusions were based on thermal
analysis and high temperature X-ray diffraction experiments performed on pow-
dered samples and on conductivity measurements performed on single crystal
samples. Thermal events were found to coincide with weight loss events, increases
in conductivity at 230 1C were found to diminish in significance with repeated
thermal cycling, and the high temperature diffraction data showed a rather messy
Faraday Discuss., 2007, 134, 17–39 | 19This journal is
c
The Royal Society of Chemistry 2006

Table 1 Selected publications describing the high-temperature structural and/or transport
properties of CsH
2
PO
4
. S = supporting the conclusion of a superprotonic transition;
R = refuting the conclusion; and ? = uncommitted
Authors Article Title Year Ref. S/R/?*
L. N. Rashkovich, K. B.
Meteva, Ya. E
´
. Shevchik,
V. G. Hoffman, and
A. V. Mishchenko
Growing Single Crystals of
Cesium Dihydrogen
Phosphate and Some of Their
Properties.
1977 8 S?
L. N. Rashkovich and
K. B. Meteva
Properties of Cesium
Dihydrophosphate.
1978 9 R
E. Rapoport, J. B. Clark
and P. W. Richter
High-Pressure Phase
Relations of RbH
2
PO
4
,
CsH
2
PO
4
, and KD
2
PO
4
.
1978 10 S
B. Metcalfe and J. B. Clark Differential Scanning
Calorimetry of RbH
2
PO
4
and
CsH
2
PO
4
.
1978 11 S
M. Wada, A. Sawada and
Y. Ishibashi
Some High-Temperature
Properties and the Raman-
Scattering Spectra of
CsH
2
PO
4
.
1979 12 S
L. C. Gupta, U. R. K. Rao,
K. S. Venkateswarlu and
B. R. Wani
Thermal-Stability of
CsH
2
PO
4
.
1980 13 S
B. M. Nirsha, E. N.
Gudinitsa, A. A. Fakeev, V.
A. Efremov, B. V. Zhadanov
and V. A. Olikova
Thermal Dehydration Process
of CsH
2
PO
4
.
1982 14 R
A. I. Baranov, V. P.
Khiznichenko, V. A. Sandler
and L. A. Shuvalov
Frequency Dielectric-
Dispersion in the
Ferroelectric and Superionic
Phases of CsH
2
PO
4
.
1988 5 S
W. Bronowska and
A. Pietraszko
X-Ray Study of the High-
Temperature Phase-
Transition of CsH
2
PO
4
Crystals.
1990 15 S
F. Romain and A. Novak Raman Study of the High-
Temperature Phase-
Transition in CsH
2
PO
4
.
1991 16 S
R. A. Vargas and E. Torijano Phase-Behavior of RbH
2
PO
4
and CsH
2
PO
4
in the Fast-Ion
Regime.
1993 17 S?
A. Preisinger, K. Mereiter,
and W. Bronowska
The Phase Transition of
CsH
2
PO
4
(CDP) at 505 K.
1994 18 S
K. S. Lee Hidden Nature of the High-
Temperature Phase
Transitions in Crystals of
KH
2
PO
4
-Type: Is It a
Physical Change?
1996 19 R
Y. Luspin, Y, Vaills, and
G. Hauret
Discontinuities in the Elastic
Properties of CsH
2
PO
4
at the
Superionic Transition.
1997 20 S
E. Ortiz, R. A. Vargas and
B. E. Mellander
On the High-Temperature
Phase Transitions of
CsH
2
PO
4
: a Polymorphic
Transition? A Transition to a
Superprotonic Conducting
Phase?
1999 21 R
20 | Fa rada y Discuss ., 2007, 134, 17–39 This journal is
c
The Royal Society of Chemistry 2006

evolution of peaks.
21
From the diffraction data the authors identified what they
believed to correspond to the most intense peak of the first dehydration product,
however, this peak (at 2Y E 25.31) coincides almost precisely with the most intense
peak for superprotonic CsH
2
PO
4
, as reported previously by Bronowska et al .
18
Additional support for the view of dehydration rather than polymorphic phase
transitions comes from the work of Park who has published two papers agreeing
with the position that only decomposition occurs in CsH
2
PO
4
upon heating.
27,29
Here, the conclusion is based primarily on the results of AC impedance
E. Ortiz, R. A. Vargas and
B. E. Mellander
On the High-Temperature
Phase Transitions of Some
KDP-Family Compounds: a
Structural Phase Transition?
A Transition to a Bulk-High
Proton Conducting Phase?
1999 22 R
W. Bronowska Does the Structural
Superionic Phase Transition
at 231 1C in CsH
2
PO
4
Really
Not Exist?
2001 23 S
K. S. Lee Surface Transformation of
Hydrogen-Bonded Crystals at
High-Temperatures and
Topochemical Nature.
2002 24 R
J. Otomo, N. Minagawa,
C. J. Wen, K. Eguchi and
H. Takahashi
Protonic Conduction of
CsH
2
PO
4
and Its Composite
With Silica in Dry and Humid
Atmospheres.
2003 25 S
D. A. Boysen, S. M. Haile,
H. J. Liu and R. A. Secco
High-Temperature Behavior
of CsH
2
PO
4
Under Both
Ambient and High Pressure
Conditions.
2003 26 S
J. H. Park, C. S. Kim,
B. C. Choi, B. K. Moon and
H. J. Seo
Physical Properties of
CsH
2
PO
4
Crystal at High
Temperatures.
2003 27 R
D. A. Boysen, T. Uda, C. R.
I. Chisholm and S. M. Haile
High-Performance Solid Acid
Fuel Cells Through Humidity
Stabilization.
2004 28 S
J. H. Park Possible Origin of the Proton
Conduction Mechanism of
CsH
2
PO
4
Crystals at High
Temperatures.
2004 29 R
K. Yamada, T. Sagara,
Y. Yamane, H. Ohki and
T. Okuda
Superprotonic Conductor
CsH
2
PO
4
Studied by H-1, P-
31 NMR and X-Ray
Diffraction.
2004 30 S
J. Otomo, T. Tamaki, S.
Nishida, S. Q. Wang, M.
Ogura, T. Kobayashi, C. J.
Wen, H. Nagamoto and
H. Takahashi
Effect of Water Vapor on
Proton Conduction of
Cesium Dihydrogen
Phosphate and Application to
Intermediate Temperature
Fuel Cells.
2005 31 S
A. I. Baranov, V. V.
Grebenev, A. N. Khodan,
V. V. Dolbinina and E. P.
Efremova
Optimization of
Superprotonic Acid Salts for
Fuel Cell Applications.
2005 32 S
Table 1 (continued)
Authors Article Title Year Ref. S/R/?*
Faraday Discuss., 2007, 134, 17–39 | 21This journal is
c
The Royal Society of Chemistry 2006

Frequently asked questions

Q1. What is the effect of gas flow rates on dehydration kinetics?

even gas flow rates, in addition to particle/crystallite size, can be expected to have a strong impact on dehydration kinetics. 

Q2. How many hydrogen bond configurations are possible in a hypothetical compound?

By analogy to ice, the number of hydrogen bond configurations possible in this hypothetical compound results from the placement of two protons over six hydrogen positions, modified by the probability (4/6) that any proton site is open or available. 

Q3. What is the expected effect of the solid nature of the electrolyte on the fuel cell?

It can further be expected that the solid nature of the electrolyte will ensure zero fuel crossover and thereby enable the use of high concentrations of alcohol in the fuel stream. 

Q4. How many configurations of hydrogen bonds can be considered as a transition entropy?

If these sources of disorder were to be considered entirely correlated such that the orientation of the phosphate group entirely fixed the location of hydrogen bonds, then the number of configurations per formula unit would be only 6, implying a transition entropy of only 9.13 J mol 1 K 1, far less than the experimentally measured value of 23 J mol 1 K 1. 

Q5. What is the molar configurational entropy of CsH2PO4?

The molar configurational entropy that one would expect to result from the disordered hydrogen bond in CsH2PO4 is Rln(2) = 5.76 J mol 1 K 1. 

Q6. How many tetrahedral group orientations are possible for a hydrogen bond?

if only oxygen atoms directly adjacent to a cube face can serve as donors to a bond extending out of that face, then the number of tetrahedral group orientations compatible with ‘L’ type hydrogen bond configurations is only two (Otetr = 2). 

Q7. What are some rather profound implications for the proton transport behavior of CsHSO4?

Buried within the modified ‘extended’ ice rules are some rather profound implications for the proton transport behavior of CsHSO4. 

Q8. How many tetrahedral group orientations are possible for hydrogen bonds?

for hydrogen bond configurations in which the bonds extend out of opposing faces of the cube, there are four tetrahedral group orientations possible, if one only considers short hydrogen bonds (Otetr = 4). 

Q9. What is the sulfate group orientation in the Jirak model?

Other models, however, such as those of Belushkin50 and Merinov51,52 in particular, distinguish between donor and acceptor oxygen atoms, and thus the sulfate group orientation in these structural models is fixed by the location of the hydrogen bonds. 

Q10. What is the length of the oxygen neighbor in Fig. 13a?

Each of the oxygen atoms marked with a number sign (#) in Fig. 13a has an oxygen neighbor in the phosphate group that resides directly to the right of the one shown with which a bond of this length could be formed.