This document is the Accepted Manuscript version of a Published Work that appeared in final form in
Elsevier B.V Copyright © peer review and technical editing by the publisher. To access the final edited
and published work see
https://www.sciencedirect.com/science/article/pii/S0920586118307405?via%3Dihub

Cobalt oxide-based materials as non-PGM catalyst for HER in PEM electrolysis and in
situ XAS characterization of its functional state
Jordi Ampurdanés, Muralidhar Chourashiya, Atsushi Urakawa*
Institute of Chemical Research of Catalonia (ICIQ), The Barcelona Institute of Science and Technology,
Av. Països Catalans 16, 43007 Tarragona, Spain
* Email: aurakawa@iciq.es

Abstract:
The polymer electrolyte membrane (PEM)-based electrolysis technology is a promising mean to
split water and store renewable energy in the form of clean fuel, hydrogen. However, its high
price due to the use of platinum group metals (PGMs) as catalyst materials generally makes the
technology cost-restrictive. Herein we present the performance evaluation of cobalt oxide in
PEM electrolysis as cathode catalyst where hydrogen evolution reaction (HER) takes place.
Performance comparison of non-PGM catalysts (CoO, Co3O4 and MoS2) revealed that better
performance was attained with Co3O4 and it was further improved by mixing with an electrically
conducting carbon material (Vulcan). An optimum amount of the carbon additive was found, and
the best performance was recorded at 47wt% Co3O4 over the total amount. At 2.05 V, 47wt%
Co3O4-based catalyst (0.55 A/cm2) outperformed 47wt% MoS2-based one (0.51 A/cm2). More
importantly, the performance of the former at 2.3 V (1.12 A/cm2) surpassed even that of Pt-black
(1.11 A/cm2). In situ XAS study of the Co3O4-based material under PEM electrolysis conditions
revealed dynamic interchange of Co3+ and Co2+ fractions, which was attributed to ultimately
boost the HER performance.
Keywords: PEM water electrolysis; hydrogen evolution reaction (HER); Co3O4; carbon; X-ray
Absorption Spectroscopy (XAS)

1. Introduction
Storing energy from renewable sources, such as sunlight or wind, by converting it into
chemical fuel is an appealing way to match up with increasing energy demands and to decrease
the dependency on fossil fuels. However, such conversions with high efficiency, low cost and
environmental benignity remain a great technological challenge. One of the most attractive
chemical fuels for storing renewable energy is hydrogen, as it can be produced via water
electrolysis, where the required electric energy can be derived from intermittent renewable
energy resources [1-3]. Water electrolysis needs at least 1.23 V across cathode and anode, where
hydrogen evolution reaction (HER) and oxygen evolution reaction (OER) take place respectively,
to overcome its thermodynamic barrier [4]. However, in practice due to overpotentials
associated with various factors such as the reaction kinetics, water electrolysis based on proton
exchange membrane / polymer electrolyte membrane (PEM) typically operates at a cell voltage
of 1.8 to 2.0 V (82 to 74% voltage efficiency, respectively), even using precious noble-metal based
electro-catalysts (Pt, Ir and Ir-Ru; platinum group metals – PGMs). Although the PEM-based
electrolysis is a very promising technology to store renewable energy by producing hydrogen
with high purity at high current density, the high cost due to the catalysts presents barriers
toward commercialization of these electrolyzers [5].
For its practical application, there is a strong need to search for alternative catalysts based
on non-PGMs for both HER and OER [6-8]. As the result of this drive, innovative breakthroughs
based on cobalt compounds have been witnessed in the past decade [9]. Among these, Co3O4 is
identified as an attractive bi-functional catalyst for both HER and OER in water splitting. Co3O4
adopts the normal spinel structure, with Co2+ ions in tetrahedral interstices and Co3+ ions in the
octahedral interstices of the cubic close-packed lattice of oxide anions. The presence of mixed
valences of the same cation provides donor-acceptor chemisorption sites and thereby inducing
the possibility of the bi-functionality [10]. Moreover, these oxides exhibit relatively high electrical
conductivity by electron transfer taking place with relatively low activation energy barriers
between cations of different valences by hopping processes. Its high electrocatalytic activity,
thermodynamic stability, low electrical resistance and their abundance in Earth’s crust (low cost)
are the main profits of Co3O4 as electrode materials [10].
In the last two decades, the electrocatalytic activity of bare Co3O4 has been well researched
mostly as HER or OER catalyst in a typical 3-electrode cell configuration [11-14]. Therefore, there
is a need to evaluate its performance as a device in a functional PEM-based electrolyzer.
Reflecting this background, the aim of this work was to investigate the performance of cobalt
oxide-based catalysts as cathode material in PEM electrolysis. The Co3O4 catalyst was
investigated in unmixed and mixed forms with activated carbon material (Vulcan® XC72) and
their performance was compared with our earlier reported MoS2/Vulcan® XC72 [15] and Pt-black
based cathode catalysts. The stability of the optimum catalyst was also examined. Finally, we
clarified the unique electronic state of Co in the catalyst material under PEM electrolysis
conditions using in situ X-ray Absorption Spectroscopy (XAS).

2. Experimental
2.1. Electrolyte and electrode catalysts
Nafion® membrane (NAFION-117, 175 μm thickness, Ion-Power) was used as the polymer
electrolyte membrane (hereinafter mentioned as “Nafion”). The Nafion membrane was pretreated [16] by washing in 5% H2O2 for 1 h at 80°C to eliminate residual organic impurities [17],
then in 0.5 M sulfuric acid for 1 h at 80 °C to incorporate water molecules and activate the
membrane by hydration [18] and finally with boiling water for 1 h. The pre-treated membranes
were stored in water at ambient temperature.
Generally, an improvement in catalytic performance is achieved when particle size is
decreased, but the durability can be deteriorated as reported for oxygen evolution reaction (OER)
catalysts in PEM electrolysis [5, 19-22]. One can thus expect a similar type of correlation between
particle size and the catalytic properties for HER catalysts. In this investigation the commercial
materials are used “as received” and all the other protocols, like the material mixing procedure,
were kept identical.
As cathode catalyst, two types of commercial cobalt oxides, namely cobalt (II) oxide (CoO,
≥99.99%, <2 µm) and cobalt (II, III) oxide (Co3O4, >99 %, powder, <50 nm) from Sigma-Aldrich,
were used. To investigate the effect of incorporating an electrically-conducting secondary
element in the catalyst, an appropriate amount of Vulcan® XC72 (carbon black, average particle
size of 50 nm, Fuel Cell Store; hereinafter denoted as “Vulcan”) and cobalt oxide were physically
mixed [23] and thoroughly homogenized by grinding in a mortar to achieve the proportion of
cobalt oxide : Vulcan as 9:91, 33:67, 47:53, 70:30 and 100:0. For each proportion, Vulcan loading
was kept fixed (5 mg/cm2). The employed physical mixing did not compromise the materials itself
as evidenced by the unchanged XRD patterns.
As anode catalyst, Iridium (IV) oxide (IrO2, >99%, powder, Alfa Aesar) was used with a fixed
loading of 2 mg/cm2. Pt black (≤20 μm powder, surface area ≥25 m2/g, ≥99.97% trace metal basis,
Sigma-Aldrich) was used as reference cathode catalyst with 0.6 mg/cm2 loading. Table 1 shows
various anode and cathode catalyst formulations used in this study.

Table 1
Anode and cathode catalyst formulations used for the preparation of MEAs. *Fixed Vulcan
loading of 5 mg/cm2.
No.
Anode
Cathode-catalyst
ID
Ref.
MEAs with unsupported catalyst – screening
2)
a.
IrO2 (2 mg/cm
CoO (5 mg/cm2)
100% CoO
This work
2)
b.
IrO2 (2 mg/cm
Co3O4 (5 mg/cm2)
100% Co3O4
IrO2 (2 mg/cm2)
MoS2 (5 mg/cm2)
100% MoS2
[15]
c.
MEAs with supported catalyst – support% optimization
d.
IrO2 (2 mg/cm2)
9 wt% Co3O4 | 91 wt% Vulcan *
9% Co3O4
2)
e.
IrO2 (2 mg/cm
33 wt% Co3O4 | 67 wt% Vulcan *
33% Co3O4
This work
f.
IrO2 (2 mg/cm2)
47 wt% Co3O4 | 53 wt% Vulcan *
47% Co3O4
g.
IrO2 (2 mg/cm2)
70 wt% Co3O4 | 30 wt% Vulcan *
70% Co3O4

h.

Reference MEA with Pt black cathode catalyst
IrO2 (2 mg/cm2)
Pt-black (0.6 mg/cm2)
PtB

[15]

2.2.

Membrane Electrode Assembly (MEA)
The catalyst-coated membrane (CCM) preparation method of MEAs has important
advantages over other MEA preparation methods [24]. In CCM approach, the catalyst-electrode
are directly coated on either side of the membrane. In this work, the technique employed for
such coating is spray-deposition [24]. The catalyst-ink for spray coating consists in the calculated
amount of catalyst for desired catalyst loading, isopropanol (2-propanol, >99.7%, Panreac
Química), and Nafion ionomer (Nafion, D-521 dispersion, 5% w/w in water and 1-propanol, Alfa
Aesar). The amount of Nafion ionomer added to ink was adjusted to have dry Nafion amounting
15.4 wt% of total weight of catalyst and the carbon additive. The prepared catalyst-isopropanolNafion mixture was ultra-sonicated for 1 h, prior to the coating and spraying processes. Spraycoating of the catalyst-ink on a preheated (at 85°C) Nafion membrane was performed using a
hand-held spray-gun (0.25 mm nozzle diameter, Ventus Titan). The active area of these MEAs
was kept fixed to 4 (2 x 2) cm2. The homogeneity of the coated catalysts was assured by
reproducible electrolysis performance of the fabricated MEAs.
2.3. Electrolysis setup and configuration
To evaluate the performance of fabricated MEAs, they were mounted in an in-house
designed electrolyzer cell (Figure 1a). An MEA was sandwiched between two Ti porous gas
diffusion layers (GDL, sintered-Ti filter, 50 µm pore diameter, 1 mm thickness; Xinxiang Xinli Filter
Technology Co., Ltd, China). The GDL sandwiched MEAs was further sandwiched between two Ti
current collector plates (2 mm thickness) machined with a single-serpentine flow field with total
volume of 0.5 cm3. The resulting assembly and the metallic housing of the electrolyzer were
closed through PTFE gaskets. These components were kept under compression with the help of
aluminum-end plates and nut/bolt assembly (8 bolts, 5 Nm torque). End plates (with heating
accessories) were kept electrically insulated using PTFE plates. Water (milli-Q) was supplied at a
fixed rate of 0.5 mL/min to the anode side of the electrolyzer using a peristaltic pump (Ismatec
REGLO Digital). A temperature controller was used to set the desired working temperature (fixed
to 80 °C) of the electrolyzer. A LabVIEW based software was used to control using a power supply
(CPX400DP, 60V/20A, AIM & THURLBY THANDAR INSTRUMENTS, UK) to apply a DC potential
across the cell and record the resulting current. The schematic of the electrolyzer test setup is
shown in Figure 1b.
2.4. Performance and structural stability of cathode catalyst
Every polarization curves (V vs j) were recorded as a function of applied potential (1.2 to 2.7
V with a voltage step of 50 mV every 30 min). The resulting current was allowed to stabilize for
30 min and the observed current value on the last minute was used to plot the polarization
curves. The maximal cell voltage for PEM water electrolysis is 2.3V, however, according to
literature the maximal cell voltages for an industrial electrolyzer can be up to 2.6 V which is to
counter the high overpotential and large ohmic drop in electrolyzers [25]. The cell voltage at a
reference current density of 1 A/cm2 can be used as a formal criterion to quantify the efficiency

of electrocatalysts for hydrogen production in PEM water electrolyzers. Generally, with Pt-black
as anode (reference) catalyst the electrolyzer shows 1 A/cm2 at around 1.8 V but in our case, it
was obtained at ca. 2.2 V, indicating the presence of ohmic drop (ca. 0.4 V) in our electrolyzer
cell. As the data were used to compare with our own reference material in the indigenously built
electrolyzer cell, the data were not iR-corrected. Hence, to compensate the ohmic drop the
potential sweep was carried out up to 2.7 V (2.3 V + 0.4 V). Additionally, the activity of best
performing catalyst among the investigated catalyst was compared at around 2.2V.
To determine the catalyst stability, the chronoamperometry (current vs time at constant
potential) experiments were performed. X-ray diffraction patterns of catalyst were recorded
before and after catalyst stability experiments, to estimate phase-stability of the catalyst.
2.5. In-situ and ex-situ XAS of the cathode catalyst
The best performing catalyst among the investigated was further characterized by using insitu and ex-situ XAS with an in-house designed electrolyzer with X-ray transparent windows. The
challenges of designing the electrolyzer for XAS were to maintain its air-tight/compressed
assembly and acquire XAS data with adequate quality.
2.5.1. X-ray transparent Electrolyzer
The components and materials of the spectroscopic cell were same as those of the PEM
electrolyzer (section 2.3, Figure 1) with a few modifications so that the catalyst can be probed by
X-ray. The modifications are as follows: (i) 45° tapered slot-cut on aluminum endplates, (ii)
replacement of electrically insulating PTFE plate by Kapton® (polyimide, 75 µm thickness) film
which serves as X-ray transparent window, (iii) rectangular slot-cut on the Ti current collectors
and (iv) Ti-GDL drilled (ø = 9 mm) to allow X-ray beam to pass through.
A window in aluminum endplates, Ti current collector and Ti-GDL allowed a path for
entering/exiting X-ray through electrolyzer. The Kapton® film provided the electrical insulation
between the aluminum-endplate and the Ti-current-collector while providing an
assistance/support to direct the water from the inlet to outlet ports. Figure 2 shows the
schematic of the electrolyzer with X-ray transparent windows and also explains its features. This
design enables the evaluation of anode/cathode catalysts of PEM electrolyzer in transmission as
well as fluorescence mode of XAS. All the results reported here were collected in transmission
mode.
2.5.2. XAS measurements
The XAS measurements were performed at BL22 - CLÆSS (Core Level Absorption & Emission
Spectroscopies) beamline of ALBA Synchrotron Light facility (Cerdanyola del Vallès, Barcelona,
Spain). The catalysts were examined by X-ray absorption near edge structure (XANES) at Co Kedge for cathode catalyst and at Ir LIII-edge for anode catalyst. For this particular study, Iridium
black (99.8% metals basis, Alfa Aesar) as anode catalyst (2 mg/cm2) and 47% Co3O4 as cathode
catalyst were used and coated on an active area of 16 (4x4) cm2. The Ti-GDL with an average pore
diameter of 6 µm was used for this study to carry out low-intensity electrolysis and thus ensure
a homogeneous media (liquid + gas) inside the cell. For in-situ and ex-situ XANES characterization
two identical MEAs were prepared.

For ex-situ XANES characterization, one of the MEAs was operated in the normal electrolysis
setup (section 2.3) for five times (linear sweep voltammetry, LSV). The LSV was carried out from
1 to 3 V with a scan rate of 0.05 V/min. The cell was operated at 80 °C and the water (Milli-Q)
was supplied to the anode compartment at a flow rate of 0.5 ml/min. After completion of these
5 LSVs, the cell was disassembled and the MEA was then characterized for XANES at Ir LIII-edge
(ca. 11.2 keV) and Co K-edge (ca. 77.2 keV). Moreover, the powder scratched from the anode and
cathode side of same MEA were pelletized and again characterized by XANES. As references,
pellets of Iridium black and Co3O4 (with cellulose as diluent) were also characterized. For every
sample, at least 3 spectra were recorded to determine the quality of data.
For in-situ XANES characterization, the other MEA was mounted in the spectroscopic cell
(section 2.5.1). The MEA in the electrolyzer was hydrated for 5 h at room temperature before
measurements. After hydration at room temperature (25 °C), XANES spectra at Ir LIII-edge and Co
K-edge were recorded. Later, the temperature of the electrolyzer was raised to operating
temperature (80 °C) and after stabilization for 20 min, a series of XANES spectra were recorded.
To investigate the effect of applied potentials, a constant potential was applied to the
electrolyzer and after a fixed interval of time (10 min), XANES were recorded repeatedly at least
for 5 times. These steps were repeated at the applied potentials of 1.6, 1.8, 2.2 and 2.8 V (Table
2). Collected spectra were deglitched, aligned, normalized and exported using ATHENA (Demeter
0.9.21) XAS data processing software [26].
Table 2
Summary of conditions at which the XANES were recorded.
XANES obtained for condition(s)
Ex-situ - in transmission mode
After 5 LSV cycles – Anode of MEA
After 5 LSV cycles – Cathode of MEA
Powder scratched from anode layer of MEA – pelletized
Powder scratched from cathode layer of MEA – pelletized
Ir-black – pelletized
Co3O4 – pelletized
In-situ - through cathode side in transmission mode
0.0 V / 25 °C
0.0 V / 80 °C
1.6 V / 80 °C
1.8 V / 80 °C
2.2 V / 80 °C
2.8 V / 80 °C

at energy edge of
Ir - LIII
Co - K
Ir - LIII
Co - K
Ir - LIII
Co - K
Ir - LIII
Ir - LIII
Ir - LIII
Ir - LIII
Ir - LIII
Ir - LIII

Co - K
Co - K
Co - K
Co - K
Co - K
Co - K

3. Result and discussions
By optimizing electrolyzer parameters/design, such as coating strategy, the temperature of
the cell and supplied water, optimum contact/compression by tightening electrolyzer
components and supplying reactants with appropriate flow rate depending on applied potential,
it is possible to gain an enhancement in electrolyzer performance. Therefore, comparing a result
with others in literature might show a large deviation in performance in PEM electrolysis even

when a similar MEA is used. To compare the material aspects of the catalysts, one can prepare a
reference MEA with standard loading/catalysts and then compare its performance with their own
results using the same electrolyzer and protocol. Performance of MEA with such a reference
catalyst can act as a baseline for comparison. This approach avoids the ambiguity that might be
caused by one or more of the numerous experimental parameters. In this study, we compare the
performance of various prepared MEAs with the performance of our reference MEA (Table 1).
3.1. Co-based cathode catalyst for PEM electrolysis
Figure 3 shows the polarization curves of MEAs with CoO (100% CoO), Co3O4 (100% Co3O4)
and MoS2 (100% MoS2; from our earlier work) [15] as cathode catalyst in PEM electrolysis. The
polarization curves were also compared with that of Pt-black cathode catalyst MEA (reference
cathode catalyst). In general, the MEA with Co3O4-based cathode catalyst outperformed the
other two catalysts. With our electrolysis setup, a measurable current density of 1 mA/cm2 was
achieved at applied potentials of 1.55 V, 1.50 V, 1.50 V and 1.40 V for MoS2, CoO, Co3O4 and Ptblack (PtB) cathode catalysts, respectively. The cobalt oxide-based catalysts showed comparably
lower overpotential than the MoS2 catalyst. At 2.7 V, the current densities obtained for Co3O4,
CoO, MoS2 and Pt-black based MEAs were 1.25, 1.0, 0.40 and 1.65 A/cm2, respectively. This shows
that Co3O4 is a better performing catalyst among the two types of cobalt oxides. Therefore, we
focused on this material to optimize its performance by mixing it with an electrically conducting
carbon material (Vulcan), as this strategy was observed to improve the HER performance of MoS2
[15].
3.2. Effect of Vulcan amount on Co3O4-based cathode performance
The electrical conductivity is an important factor while designing electrodes for any
electrochemical devices, and the addition of electrically conducting material like Vulcan to the
active phase of the electrode catalyst can be beneficial for the charge distribution. On the other
hand, if the amount of the additive is too large it may deteriorate the catalytic activities
negatively. Hence, a larger than an optimum amount may lead to blockage of a fraction of
catalytic-reaction sites or their dilution, while less than the optimum amount could result in the
less conductive catalyst layer. This approach can be highly effective if the catalyst has poor
electrical conductivity as demonstrated previously [15].
Figure 4 shows the polarization curves of MEAs prepared with different proportion of Co3O4
mixed with Vulcan as the cathode catalyst. The polarization curves of the MEA with reference
(PtB) and MoS2 mixed with Vulcan (47%MoS2 [15]) based cathode catalyst are also included for
the sake of better comparison. Operating at the cell potentials of 1.2-2 V (Figure 4a), the MEA
with 100% Co3O4, 70% Co3O4 and 9% Co3O4 exhibited almost identical performance. This signifies
that either there is inadequate electrical connectivity (100% Co3O4 and 70% Co3O4) or insufficient
amount of catalytic reaction sites (9% Co3O4) in these cathode catalysts. With 33% and 47%
Co3O4, a clear improvement was observed, and the best performance was achieved with 47%
Co3O4. The results indicate the importance of the balance between the amount of the
catalytically active element and the electrically conducting medium in the electrode catalyst layer
and that 47% Co3O4 is at the optimum among the variations studied here.

At the cell potential of 2 V, the MEA with the Pt-black cathode catalyst achieved the current
density of 0.70 A/cm2, whereas the MEA with the Co3O4 and MoS2 based composite cathode
catalysts showed ca. 0.42 A/cm2 at the optimum amount of Vulcan. Intriguingly, at a high cell
voltage range (2 to 2.7 V, Figure 4b), the performance of the composite catalysts surpassed that
of the reference cathode catalyst. The polarization curves of 47% Co3O4 and 47% MoS2 crossed
that of Pt black at 2.3 and 2.55 V, respectively, and above these cell voltages, these composite
catalysts outperformed PtB. Particularly, the performance of the Co3O4-based composite catalyst
is promising for cost-effective implementation of PEM based hydrogen production or
hydrogenation reactions [27] operated at the high voltage range.

3.3. Stability of the optimum Co3O4-carbon composite catalyst
Stability tests were performed with the MEA coated with 47% Co3O4 at 2 and 2.5 V for 24 h
(Figure 5). At applied cell potential of 2 V, the current density showed a maximum of 0.42 A/cm2
(after 4 h) and remained almost stable. The initial phase of ca. 4 h to reach the maximum current
density at an applied potential of 2 V can be attributed to the period for complete hydration to
achieve optimum H+/H2O transport of Nafion membrane [28]. On the other hand, at the cell
potential of 2.5 V, the current density showed a maximum value of ca. 1.5 A/cm2 quickly and
remained stable without any indication of deactivation.
After the stability test of the MEA with the 47% Co3O4 catalyst, the MEA was studied by XRD.
Figure 6 shows the XRD patterns of the cathode catalyst layer, before and after the stability tests.
XRD patterns measured before and after the electrolysis showed reflections of Co3O4 (COD ID:
9005888) with a spinel structure [29] besides minor and broad reflections of Vulcan. Both XRD
patterns showed unchanged lattice parameters (a = b = c = 8.084 ± 0.001 Å; α = β = γ = 90°).
However, the calculated average crystallite size for catalyst before electrolysis (14.7 nm) and
after electrolysis (17.8 nm) showed an increase, signifying agglomeration of catalyst particles.
Nevertheless, the catalyst did not show any phase changes, indicating the robustness of 47%
Co3O4 catalyst as cathode material for PEM electrolysis.
3.4. In-situ/ex-situ XAS characterization of anode/cathode catalysts
The electronic structure of Co3O4 at the optimized composition (47 wt%) mixed with Vulcan
and deposited over MEA was further investigated by XAS. For this study, the MEA was coated
with Ir-black (pure Ir metal) as anode catalyst and with 47% Co3O4 as the cathode catalyst. The
nature of the anode compartment in PEM based (acidic) electrolyzer is highly electro-oxidative
and there is a strong possibility that the metallic element gets oxidized. The intention to use Irblack as an anode catalyst for this study was in anticipation of the possibility to observe, if any,
the transformation of Ir to IrOx phase.
Collected spectra showed random glitches caused by the interference of various
instrumentations used for water electrolysis with X-ray detectors as well as by fluctuations in the
fluid density during the measurements. Therefore, the spectra shown are initially de-glitched
using the ATHENA (Demeter 0.9.21) XAS data processing software and are used for qualitative
comparison. Normalized XANES spectra recorded under various conditions (Table 2) are
presented in Figures 7 and 8. In situ XANES spectra recorded at Ir LIII-edge for various operating

stages (Figure 7A(b-g)), namely at different temperatures and potentials, were all very similar
also in comparison to the spectrum of Ir-black (Figure 7A (a)). The principal peak maximum was
observed at about 11.22 keV, confirming the stability of Iridium as an anode catalyst for
electrolysis [30]. Ex situ XANES spectra also confirmed the same (Figure 7B).
In contrast, the electronic state of Co varies dynamically due to the environment as
evidenced by the XANES spectra at Co K-edge measured at different temperatures and applied
potentials (Figure 8). The spectrum obtained for the reference Co3O4 powder (Figure 8A (a)) show
mixed oxidation states of Co between +2 (relatively smaller shoulder peak at 7727 eV) and +3
(principle-maximum peak at 7733 eV) as expected for the material. A similar ratio of Co2+/Co3+
oxidation states was found for the MEA coated with 47% Co3O4 catalyst when it was hydrated
and kept at room temperature (Figure 8A (b)) in the PEM electrolyzer cell. To our surprise, as the
cell temperature was raised to 80 °C (at 0.0 V; Figure 8A (c)), peak intensity corresponding to Co2+
species exceeded the peak intensity corresponding to Co3+ species. In other words, a fraction of
Co2+ species becomes dominant over that of Co3+ species, clearly showing a reduction of Co3+ to
Co2+, which could be due to the acidic environment created by Nafion. Furthermore, this
reduction of Co was more pronounced when the cell potential was raised (Figure 8A (d-g)),
manifesting the highly dynamic nature of Co oxidation state in Co3O4.
The most prominent peaks for the MEA immediately after dismantling the electrolyzer
(Figure 8B(a)) and that for the MEA while operating at 2.8V (Figure 8B(b)) were observed at about
7727 eV, indicative of Co2+ abundance. This implies that the transition of Co oxidation state
occurred during the electrolysis has been retained even after dismantling the electrolyzer and
the transition was retained. However, when the catalyst was scratched off from the MEA and
measured ex-situ (Figure 8B(c)), their spectral features were very similar to those of Co3O4 (Figure
8B(d)). This clearly shows that the oxidation state of the material is dynamically and reversibly
changed and most likely the environment of the coated membrane with Nafion helps to keep the
oxidation state of Co more reduced (Co2+). The recovery of the electronic structure observed for
47% Co3O4 catalyst is analogous to the self-repairing mechanism reported for the so-called
cobalt-phosphate catalyst (also known as Co-Pi-Cat) [31]. As Co3O4 inherits the structural
properties of Co-Pi-Cat [32], the same structural and electronic changes may take place for Co3O4.
According to Bergmann et al [32], it is feasible that crystalline cobalt oxides form structurally
flexible, hence catalytically active, amorphous shells. This shell formation in the Co3O4-based
catalyst can be related to the interchanged proportion of Co3+ and Co2+ species, resulting in the
improved catalyst and therefore higher performance of the MEA.
4. Conclusions
Comparison of CoO, Co3O4 and MoS2 as cathode catalyst and as an alternative to platinumgroup-metals (PGMs) electrode materials in PEM water electrolysis revealed that Co3O4 is the
most promising one among the three. Its hydrogen evolution reaction (HER) activity was further
improved by addition of electrically conducting material (Vulcan) and by optimizing the amount
of Co3O4 with respect to Vulcan. These non-PGM materials showed high performance at high cell
potentials. The MEA based on 47wt% Co3O4 mixed with Vulcan outperformed the MEA with Ptblack as cathode catalyst at applied potentials of ≥ 2.3 V. The electronic structure of active Co
and Ir states at cathode and anode, respectively, was investigated by in situ XAS under water
electrolysis conditions. The electronic structure of Ir was found stable, whereas that of Co,

present as mixed states of +2 and +3 in Co3O4, underwent dynamic changes upon varying
temperature and cell potential. The acidic environment of Nafion, especially when it is wet at
high temperature, as well as at more positive potential facilitated the pronounced transition from
Co3+ to Co2+. Remarkably, the redox process was found reversible; enhanced amount of Co2+ is
transformed back to Co3+ once its environment changed to the initial state. The reversible redox
nature of Co3O4 can facilitate the charge transfer and was interpreted as the origin of its high HER
activity as well as the enabler for widely reported activity in oxygen evolution reaction (OER).
Acknowledgments
We thank the Generalitat de Catalunya for financial support through the CERCA
Programme and recognition (2017 SGR 1633), the ICIQ foundation and MCIU (CTQ2016-75499-R
(AEI/FEDER, UE)) for financial support. MGC also would like to acknowledge the fellowship
received under the ICIQ-IPM Programme partially funded by COFUND action of the European
Commission and MINECO for support through SEV-2013-0319. The XAS experiments were
performed at BL22 - CLÆSS beamline at ALBA Synchrotron with the collaboration of ALBA staff
and particularly the beamline scientists are greatly acknowledged.
References
[1] N.S. Lewis, D.G. Nocera, Powering the planet: chemical challenges in solar energy utilization, Proc Natl Acad Sci
U S A, 103 (2006) 15729-15735.
[2] N. Armaroli, V. Balzani, The hydrogen issue, ChemSusChem, 4 (2011) 21-36.
[3] G. Gahleitner, Hydrogen from renewable electricity: An international review of power-to-gas pilot plants for
stationary applications, International Journal of Hydrogen Energy, 38 (2013) 2039-2061.
[4] A. Ursua, L.M. Gandia, P. Sanchis, Hydrogen Production From Water Electrolysis: Current Status and Future
Trends, Proc. IEEE, 100 (2012) 410-426.
[5] M. Carmo, D.L. Fritz, J. Mergel, D. Stolten, A comprehensive review on PEM water electrolysis, International
Journal of Hydrogen Energy, 38 (2013) 4901-4934.
[6] S. Cobo, J. Heidkamp, P.A. Jacques, J. Fize, V. Fourmond, L. Guetaz, B. Jousselme, V. Ivanova, H. Dau, S. Palacin,
M. Fontecave, V. Artero, A Janus cobalt-based catalytic material for electro-splitting of water, Nat Mater, 11 (2012)
802-807.
[7] X. Zou, Y. Zhang, Noble metal-free hydrogen evolution catalysts for water splitting, Chemical Society Reviews,
44 (2015) 5148-5180.
[8] X. Li, X. Hao, A. Abudula, G. Guan, Nanostructured catalysts for electrochemical water splitting: current state
and prospects, Journal of Materials Chemistry A, 4 (2016) 11973-12000.
[9] V. Artero, M. Chavarot-Kerlidou, M. Fontecave, Splitting water with cobalt, Angew Chem Int Ed Engl, 50 (2011)
7238-7266.
[10] M. Hamdani, R.N. Singh, P. Chartier, Co3O4 and Co- Based Spinel Oxides Bifunctional Oxygen Electrodes,
International Journal of Electrochemical Science, 5 (2010) 556-577.
[11] J. Balej, Electrocatalysts for oxygen evolution in advanced water electrolysis, International Journal of Hydrogen
Energy, 10 (1985) 89-99.
[12] C. Iwakura, A. Honji, H. Tamura, The anodic evolution of oxygen on Co3O4 film electrodes in alkaline solutions,
Electrochimica Acta, 26 (1981) 1319-1326.
[13] Y.Y. Liang, Y.G. Li, H.L. Wang, J.G. Zhou, J. Wang, T. Regier, H.J. Dai, Co3O4 nanocrystals on graphene as a
synergistic catalyst for oxygen reduction reaction, Nature Materials, 10 (2011) 780-786.
[14] S. Du, Z. Ren, J. Zhang, J. Wu, W. Xi, J. Zhu, H. Fu, Co(3)O(4) nanocrystal ink printed on carbon fiber paper as a
large-area electrode for electrochemical water splitting, Chem Commun (Camb), 51 (2015) 8066-8069.

[15] T. Corrales-Sanchez, J. Ampurdanes, A. Urakawa, MoS2-based materials as alternative cathode catalyst for
PEM electrolysis, International Journal of Hydrogen Energy, 39 (2014) 20837-20843.
[16] S.J. Lee, S. Mukerjee, J. McBreen, Y.W. Rho, Y.T. Kho, T.H. Lee, Effects of Nafion impregnation on
performances of PEMFC electrodes, Electrochimica Acta, 43 (1998) 3693-3701.
[17] C. Yang, P. Costamagna, S. Srinivasan, J. Benziger, A.B. Bocarsly, Approaches and technical challenges to high
temperature operation of proton exchange membrane fuel cells, Journal of Power Sources, 103 (2001) 1-9.
[18] Z.J. Lu, G. Polizos, D.D. Macdonald, E. Manias, State of water in perfluorosulfonic ionomer (Nafion 117) proton
exchange membranes, Journal of the Electrochemical Society, 155 (2008) B163-B171.
[19] D.F. Abbott, D. Lebedev, K. Waltar, M. Povia, M. Nachtegaal, E. Fabbri, C. Copéret, T.J. Schmidt, Iridium Oxide
for the Oxygen Evolution Reaction: Correlation between Particle Size, Morphology, and the Surface Hydroxo Layer
from Operando XAS, Chemistry of Materials, 28 (2016) 6591-6604.
[20] J.C. Cruz, V. Baglio, S. Siracusano, R. Ornelas, L. Ortiz-Frade, L.G. Arriaga, V. Antonucci, A.S.J.J.o.N.R. Aricò,
Nanosized IrO2 electrocatalysts for oxygen evolution reaction in an SPE electrolyzer, 13 (2011) 1639-1646.
[21] E. Rasten, G. Hagen, R. Tunold, Electrocatalysis in water electrolysis with solid polymer electrolyte,
Electrochimica Acta, 48 (2003) 3945-3952.
[22] S. Siracusano, V. Baglio, A. Di Blasi, N. Briguglio, A. Stassi, R. Ornelas, E. Trifoni, V. Antonucci, A.S. Aricò,
Electrochemical characterization of single cell and short stack PEM electrolyzers based on a nanosized IrO2 anode
electrocatalyst, International Journal of Hydrogen Energy, 35 (2010) 5558-5568.
[23] J.C. Tokash, B.E. Logan, Electrochemical evaluation of molybdenum disulfide as a catalyst for hydrogen
evolution in microbial electrolysis cells, International Journal of Hydrogen Energy, 36 (2011) 9439-9445.
[24] B. Bladergroen, H. Su, S. Pasupathi, V. Linkov, Overview of Membrane Electrode Assembly Preparation
Methods for Solid Polymer Electrolyte Electrolyzer, in: D.J. Kleperis (Ed.) Electrolysis, InTech, 2012.
[25] M. Wang, Z. Wang, X. Gong, Z. Guo, The intensification technologies to water electrolysis for hydrogen
production – A review, Renewable and Sustainable Energy Reviews, 29 (2014) 573-588.
[26] B. Ravel, M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using
IFEFFIT, J Synchrotron Radiat, 12 (2005) 537-541.
[27] S.M.A. Kriescher, K. Kugler, S.S. Hosseiny, Y. Gendel, M. Wessling, A membrane electrode assembly for the
electrochemical synthesis of hydrocarbons from CO2(g) and H2O(g), Electrochemistry Communications, 50 (2015)
64-68.
[28] M. Cappadonia, J.W. Erning, S.M.S. Niaki, U. Stimming, Conductance of Nafion-117 Membranes as a Function
of Temperature and Water-Content, Solid State Ionics, 77 (1995) 65-69.
[29] X. Liu, C. Prewitt, High-temperature X-ray diffraction study of Co3O4: Transition from normal to disordered
spinel, Phys. Chem. Miner., 17 (1990) 168-172.
[30] J.E. Penner-Hahn, X-ray absorption spectroscopy in coordination chemistry, Coordination Chemistry Reviews,
190 (1999) 1101-1123.
[31] M.W. Kanan, D.G. Nocera, In situ formation of an oxygen-evolving catalyst in neutral water containing
phosphate and Co2+, Science, 321 (2008) 1072-1075.
[32] A. Bergmann, E. Martinez-Moreno, D. Teschner, P. Chernev, M. Gliech, J.F. de Araujo, T. Reier, H. Dau, P.
Strasser, Reversible amorphization and the catalytically active state of crystalline Co3O4 during oxygen evolution,
Nat Commun, 6 (2015) 8625.

Figure 1. (a) In-house designed electrolyzer components: (1) membrane electrode assembly
(MEA), (2) Ti gas diffusion layer (GDL), (3) PTFE gasket, (4) machined Ti current collector with
single serpentine flow, (5) PTFE electrical insulating plate and (6) Al end plate with holes for
heating cartilages. Inset: Assembled electrolyzer. (b) PEM electrolysis setup: (1) PEM electrolyzer
(a), (2) peristaltic pump and water reservoir, (3) temperature controller and (4) computercontrolled DC power supply.

Figure 2. Electrolyzer with Kapton windows to enable investigation of catalyst materials by
transmission/fluorescence mode of X-ray Absorption Spectroscopy (XAS).

Figure 3. Polarization curves for MEAs with pristine cathode catalysts. Current density was
measured as a function of potential difference across the electrolyzer. The cell potential was
varied from 1.2 V to 2.7 V with a step of 0.05 V, after stabilization for 30 min. Flow rate of water
was 0.5 mL/min and the electrolysis temperature was 80 °C.

Figure 4. Polarization curves measured for MEAs with Co3O4 (100% Co3O4), Pt-black (PtB), Co3O4Vulcan composite materials (70%, 47%, 33% and 9% Co3O4) and 47% MoS2 cathode catalyst.
Polarization curves are divided in (a) the common operating cell voltage range (1.2 to 2 V) and
(b) high cell voltage range (2 to 2.7 V). Flow rate of water was 0.5 mL/min and the electrolysis
temperature was 80 °C.

Figure 5. Stability tests of PEM water electrolysis with the MEA with 47% Co3O4 cathode catalyst
measured at 2 V and then at 2.5 V for 24 h. Flow rate of water was 0.5 mL/min and the
electrolyzer temperature was 80 °C.

47% Co3O4 - Before
47% Co3O4 - After

10

20

30

40

50

(135)

(044)

(333)
(224)

(004)

(222)

(022)

Vulcan

(111)

Intensity (a.u.)

(113)

Vulcan

60

70

2(degrees)
Figure 6. XRD patterns of Vulcan and 47% Co3O4 cathode catalyst coated MEA, recorded before
and after the stability tests.

Figure 7. Normalized XANES spectra at Ir LIII-edge recorded (A) in situ and (B) ex-situ. A: XANES
spectra of (a) Ir-black (reference) powder, (b) 47% Co3O4/Ir MEA after hydration at room
temperature (0.0 V), (c) at 80 °C (0.0 V), and then the cell potential was increased to (d) 1.6 V, (e)
1.8 V, (f) 2.2 V (f) and (g) 2.8 V. B: XANES spectra of (a) Ir-black (reference) powder, (b) the MEA
immediately after the in situ characterization and after dismantling the electrolyzer and (c) pellet
made using the anode catalyst powder scratched from the MEA (b).

Co2+
7700

7710

(d) Co3O4

A

(g) 2.8 V / 80°C
(f) 2.2 V / 80°C
(e) 1.8 V / 80°C
(d) 1.6 V / 80°C
(c) 0.0 V / 80°C
(b) 0.0 V / 25°C
(a) Co3O4

7720

Co2+

Co3+
7730

Energy (eV)

B

(c) Co-Scratched Powder
(b) 2.8 V / 80°C
(a) Co-On Membrane

7740

7750

7700

7710

7720

Co3+
7730

7740

7750

Energy (eV)

Figure 8. Normalized XANES spectra at Co K-edge recorded (A) in situ and (B) ex-situ. A: XANES
spectra of (a) Co3O4 (reference) powder, (b) 47% Co3O4/Ir MEA after hydration at room
temperature (0.0 V), (c) at 80°C (0.0 V), and then the cell potential was increased to (d) 1.6 V, (e)
1.8 V, (f) 2.2 V (f) and (g) 2.8 V. B: XANES spectra of (a) the MEA recorded immediately after the
in situ characterization and dismantling the electrolyzer, (b) the MEA at applied potential of 2.8
V across the electrolyzer while keeping the electrolyzer hydrated and at 80 °C (this is in situ
spectrum), (c) pellet of cathode catalyst powder scratched from the MEA and (d) Co3O4
(reference) powder.