Journal of The Electrochemical

Society

OPEN ACCESS

Inexpensive and Efficient Alkaline Water Electrolyzer with Robust Steel-

Based Electrodes
To cite this article: Billal Zayat et al 2020 J. Electrochem. Soc. 167 114513

View the article online for updates and enhancements.

This content was downloaded from IP address 95.168.121.31 on 15/12/2021 at 09:41

 Journal of The Electrochemical Society, 2020 167 114513

Inexpensive and Efficient Alkaline Water Electrolyzer with Robust

Steel-Based Electrodes
Billal Zayat,* Debanjan Mitra,* and S. R. Narayananz
Loker Hydrocarbon Research Institute, University of Southern California, Los Angeles, CA 90089, United States of America

Electrolysis of aqueous solutions of alkali is a promising approach for the production of pure hydrogen. For this approach to be
economical on a large scale, the overpotentials for the electrode reactions and the high-cost of nickel-based electrode substrates
must be reduced. We report here on the performance of an “all-iron” electrolyzer cell that uses inexpensive steel-based electrodes.
This alkaline water electrolyzer uses a steel mesh coated with a thin catalytic coating of alpha-nickel hydroxide for the oxygen
evolution electrode, and another steel mesh sputter-coated with nickel and molybdenum for the hydrogen electrode. An alkaline
electrolyzer with these steel-based electrodes, a commercial Zirfon® separator, and a solution of 30% potassium hydroxide
exhibited an electrolysis cell voltage of 1.83 V and 1.71 V at 100 mA cm−2 when operating at 23 °C and 70 °C, respectively. We
show that the performance of the steel-based electrodes is comparable to commercial electrodes based on nickel substrates. When
the cell was operated continuously for 100 h at 1 A cm−2 at 23 °C, there was no measurable loss in performance, providing a
preliminary confirmation of the robustness of these iron-based electrodes and electrocatalysts. We conclude that cost-effective iron-
based electrolyzers could be a promising route to low-cost hydrogen production.
© 2020 The Author(s). Published on behalf of The Electrochemical Society by IOP Publishing Limited. This is an open access
article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/
by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/
1945-7111/aba792]

Manuscript submitted June 15, 2020; revised manuscript received July 16, 2020. Published July 29, 2020.
Supplementary material for this article is available online

Almost 95% of the hydrogen produced in the U.S. today is by 100 mA cm−2 and be durable for several years. To this end, we
steam-reforming of methane in which methane gas reacts with steam have focused on replacing nickel substrates with iron-based sub-
at high pressure and temperature to produce hydrogen and carbon strates, specifically those made with low-carbon steel.
dioxide, contributing to 3% of global carbon dioxide emissions.1 Iron is 600 times more abundant than nickel in the Earth’s crust.
Water electrolysis, on the other hand, is a promising alternative for Iron is also 50 times cheaper than nickel. The cost savings from the
clean hydrogen production where water is converted to hydrogen use of iron is further amplified when the cost of industrial steel mesh
and oxygen in a single step using electricity from renewable sources. is compared with specialized nickel foam. Unmodified iron, how-
Yet, water electrolysis only accounts for less than 5% of hydrogen ever, suffers from chemical instability under anodic potentials in
production.2 Water electrolysis may be carried out under either alkali solutions. The insulating passive film formed on the surface of
acidic or alkaline conditions. Alkaline water electrolysis allows the iron in alkali breaks down under anodic conditions resulting in the
use of less expensive catalysts for the hydrogen evolution reaction dissolution of iron. In previously published work from our
(HER) and the oxygen evolution reaction(OER).3,4 In addition, group,22,23 iron was rendered stable and active towards oxygen
alkaline systems have been in operation for over 100 years and their evolution by surface modification with a thin layer of high-surface
performance can rival that of the acidic electrolyzers.5 For large- area α-nickel hydroxide. Such an electrode was stable for over
scale adoption, the water electrolyzers must use low-cost and 1000 h of operation with performance comparable to nickel-based
abundantly-available materials, so that the approach is economically catalysts. Further, we have found that when the steel substrate is pre-
competitive and sustainable. Thus, our studies have focused on treated by anodic polarization in a solution of sodium sulfide, the
enabling the economic competitiveness of alkaline water electrolysis interfacial area can be increased substantially prior to surface
through designing electrodes with low-cost materials. modification. Under cathodic polarization, iron is quite stable and
Nickel is widely used as an electrode material in alkaline media evolves hydrogen. Thus, a steel mesh that is surface modified by co-
due to its high stability under cathodic and anodic conditions. Nickel sputtering of nickel and molybdenum as electrocatalysts yielded an
also exhibits high electrocatalytic activity compared to many other efficient and durable hydrogen evolving electrode. With such an
transition metals.6–8 Furthermore, nickel alloys such as nickel- inexpensive combination of electrodes, we show here that we can
molybdenum exhibit higher activity for hydrogen evolution than conduct continuous and efficient water electrolysis at high current
nickel itself, with overpotentials around 150 mV at 10 mA cm−2.9–17 densities and relatively high temperature. Such an approach to
Transition metal oxides that combine nickel, iron and cobalt are alkaline water electrolysis has the advantage of scalability and
commonly used for OER with overpotentials of approximately reduction in cost without compromising on the electrical perfor-
300 mV at 10 mA cm−2.18–21 However, such electrodes are usually mance.
prepared by applying a thin coating of the electrocatalyst on a high-
surface area nickel foam electrode substrate. The nickel substrate Experimental
constitutes more than 95% of the mass of the electrode and adds a
Electrode fabrication.—Electrodes were prepared from commer-
significant cost to the electrolysis system. Although the core of the
cial low-carbon steel mesh (Mcmaster-Carr, wire diameter of
Earth is believed to be made of molten nickel, the abundance of
0.2 mm, open area of 31%, opening size of 0.23 mm, and a mesh
nickel in the Earth’s crust is about 0.01%. Consequently, the large-
size of 60 × 60). Each of these electrodes were cut to be 5 cm ×
scale adoption of nickel as an electrode substrate material for water
5 cm. The cut mesh was rinsed with acetone to remove any grease
electrolysis would be a challenge from the viewpoint of cost and
and then dipped in 0.1 M sulfuric acid for one minute at room
sustainability. Alternate materials must not only be inexpensive and
temperature to remove surface oxides. The mesh was sonicated in
abundant but also support current densities in excess of
acetone for 15 min and then dried under vacuum at 55 °C for 15 min.
To prepare the hydrogen evolution electrode, nickel and molyb-
denum were co-sputtered on the steel mesh substrate using direct
*Electrochemical Society Student Member. current (DC) magnetron sputter deposition (Denton Vacuum
z
E-mail: srnaraya@usc.edu Explorer 20) with a pre-sputtering time of 5 min to clean the target
 Journal of The Electrochemical Society, 2020 167 114513

surface and a sputtering time of 40 min DC co-sputtering was carried material consisted of an open mesh polyphenylene sulfide fabric
out under 6.5 mTorr of argon with a nickel cathode power of 80 W that is coated with a mixture of a polysulfone and zirconium oxide. A
and molybdenum cathode power of 15 W resulting in a total metal solution of 30% potassium hydroxide was used as the electrolyte. A
loading of approximately 0.6 mg cm−2. nickel plate with a columnar flow field was used as the current
The steel substrate for the oxygen electrode was subjected to the collector on the oxygen electrode, and a graphite plate with a
cleaning procedure used for the hydrogen electrode. To increase the columnar flow field was used on the side of the hydrogen electrode.
surface area of the substrate, the steel mesh was polarized in a Centrifugal pumps (24 W BC-2CP-MD, March Pumps) were used to
solution of 30% potassium hydroxide in the presence of 0.5 g circulate the electrolyte through the electrolyzer at a rate of 1 l min−1.
sodium sulfide per liter of solution per gram of steel. At first, the The cell was heated using two 40-watt heating pads (2 in × 2 in)
steel mesh was polarized for 2 h at a constant reduction current of attached on to the current collector plates. The temperature of the cell
10 mA g−1 of steel. Under these conditions, the potential of the and electrolyte was controlled using a temperature controller and a
electrode was −1.2 V vs the Mercury/mercuric oxide reference K-type thermocouple (92000-00, Cole-Parmer Instrument Co.) placed
electrode (MMO), while hydrogen evolved at the electrode. A at the cell.
mercury-mercuric oxide (MMO, Eo = +0.098 V) electrode in 30%
potassium hydroxide was used as the reference electrode in all the Physical characterization of electrodes.—Scanning electron
experiments. RHE potential vs MMO for this electrolyte was microscopy (SEM, Nova NanoSEM 450) was used to examine the
calculated according to Eq. S1 (available online at stacks.iop.org/ morphology of the substrates and the “as-prepared” OER and HER
JES/167/114513/mmedia) to be −0.932 V. Following the cathodic electrodes. The morphology was also examined after electroche-
polarization, the mesh was polarized anodically at a constant current mical testing. Energy dispersive X-ray Spectroscopy (EDS, JSM
of 1 mA g−1 of steel during which the potential of the electrode was 7001F) was used to determine the atomic ratio of nickel to
about −0.75 V vs MMO until a cutoff voltage of −0.7 V vs MMO molybdenum on the surface of the HER electrode.
was reached. Under these conditions, the conversion of iron to iron The crystallinity and phase composition of the substrates and
(II) hydroxide occurred. To form the electrocatalyst layer of the catalysts were investigated by X-ray diffraction (Rigaku Ultima IV
oxygen electrode, the pre-treated steel substrate was placed in a diffractometer) using a Cu K-alpha source, while the surface oxidation
0.8 M solution of nickel nitrate for 15 min accompanied by sonica- states were examined using X-ray photoelectron spectroscopy (Kratos
tion, following which the substrate was heated to 150 °C to form an Axis Ultra DLD) using an aluminum K-alpha source (1486.6 eV).
evaporated layer of nickel nitrate. The electrode was dried and then Further, the thickness of the co-sputtered nickel-molybdenum layer for
annealed at 200 °C in air for 30 min to decompose the coating of the the HER electrode was determined using X-ray reflectivity measure-
nickel nitrate precursor. The resulting coating was a 100 nm layer of ment (Rigaku Ultima IV diffractometer) using a glass slide with the
α-nickel hydroxide. thin layer deposited under the same sputtering conditions.

Electrolyzer assembly.—An electrolyzer was assembled using Electrochemical characterization of electrodes.—Electrochemical

the 5 cm × 5 cm steel mesh-based hydrogen and oxygen electrodes characterization of individual electrodes was performed in three-
(Fig. 1). A commercially available special porous polymeric material electrode cells. The working electrode was a 5 cm × 5 cm catalyzed
stable in alkali with low gas permeability (ZIRFON PERL, Agfa steel electrode with nickel mesh counter electrodes. A mercury-
Specialty Products) was used as the separator. This separator mercuric oxide (MMO, Eo = +0.098 V) electrode in 30% potassium

Figure 1. Schematic of an all-iron alkaline water electrolyzer: (1) Nickel-molybdenum coated iron electrode, (2) Zirfon PERL separator, (3) graphite plate with
columnar flow field, (4) heating pad, (5) nickel-coated plate with columnar flow field, (6) gold-coated current collector, and (7) K-type thermocouple. Reactions
at individual electrodes are as indicated. Overall cell reaction: H2O → H2 + 1/2 O2.
 Journal of The Electrochemical Society, 2020 167 114513

hydroxide was used as the reference electrode in all the experiments. electrolyzer was measured under the passage of direct current. The
The area used in the current density calculations was the geometric direct current was held for 15 min over the range of 15 to 40 mA cm−2
area of both sides of the working electrode after subtracting the open with the current stepped up in 5 mA cm−2 increments until a steady
area of the mesh. The electrocatalytic activity of the HER and OER state was reached at each current density. Then, the electrochemical
electrodes was determined by steady-state potentiostatic polarization impedance was measured at each steady-state cell voltage under the
experiments at room temperature with a 15-min hold at each potential passage of direct current. Finally, the electrochemical stability of the
using a multi-channel potentiostat (AMETEK Scientific Instruments, cell was examined under galvanostatic polarization at 1 A cm−2 for
VersaSTAT 4). Potentiostatic data for the hydrogen electrode was 100 h at room temperature.
collected between −0.8 V and −1.4 V vs MMO in 50 mV increments
while potentiostatic data for the oxygen electrode was collected
Results and Discussion
between 0.3 V and 1 V vs MMO in 50 mV increments. The measured
potential was corrected for the drop across the uncompensated solution Physical characterization of the electrodes.—The SEM images
resistance using the solution resistance value obtained from the of the as-prepared HER catalyst (Fig. 2) showed that co-sputtering of
electrochemical impedance measured at 10 kHz. The Tafel slope was nickel and molybdenum for 40 min fully covered the surface of the
obtained by plotting the potentiostatic data from −1.1 V to −1.4 V low-carbon steel mesh and protected it from being exposed to the
(vs MMO) for the hydrogen electrode and from 0.5 V to 0.8 V (vs alkaline electrolyte. The cross-sectional image of the HER electrode
MMO) for the oxygen electrode. These potential windows corre- (Fig. 2b) showed the typical columnar growth observed in d.c.
sponded to the current density regime where the formation of gas sputtering and an approximate thickness of around 400 nm. This
bubbles did not impact the measurement. Furthermore, electrochemical value of thickness was confirmed by the X-ray reflectivity measure-
impedance at various values of electrode potential was measured in the ment which also showed a coating thickness of 400 nm and a
frequency range of 10 kHz to 100 mHz with a sinusoidal voltage roughness of 6.6 nm (Fig. S1). EDS measurements (JSM 7001F)
excitation of 2 mV peak-to-peak. To determine the durability of the showed that nickel and molybdenum are homogenously distributed
electrode, three sets of potentiostatic experiments were performed with an atomic ratio of nickel to molybdenum of 9:1 (Fig. S2). The
(each run conducted over 8 h) followed by 24 h of open circuit in 30% observation of iron in the spectrum is due to the penetration of
potassium hydroxide between each potentiostatic run. The electrode the electron beam in the EDS measurement past the coating into the
was then placed in 30% potassium hydroxide for one week. Finally, carbon steel mesh. The cross-sectional image of the HER electrode
the electrochemical stability of the electrode was tested by passing a after electrochemical tests (Fig. 2c) showed that there was no
constant current at 10 mA cm−2 for 100 h and observing the changes in reduction in the thickness of the sputtered catalytic layer.
electrode potential. Furthermore, there was no observed change in the surface mor-
phology or composition after the electrochemical tests (Fig. 2d).
Electrochemical characterization of electrolyzer.—All electro- The SEM images showed that upon pre-treating the low carbon
chemical tests on the assembled electrolyzer were performed steel substrate by cyclic polarization in sodium sulfide, the relatively
using a VersaSTAT single-channel potentiostat (AMETEK Scientific smooth surface of the substrate (Fig. 3a) was transformed into a
Instruments) with a single channel power booster (Princeton Applied rough surface with a relatively high-surface area (Fig. 3b). The
Research, 20 Amp max). The electrochemical performance of the increase in the substrate surface area is manifested in a eleven fold
electrolyzer was determined by galvanostatic polarization studies increase in double layer capacitance from 0.550 F to 6.41 F. Details
at room temperature between 100 mA cm−2 and 1 A cm−2 in of the electrochemical impedance spectroscopy (EIS) measurement
100 mA cm−2 increments with a 15-min hold at each current density to determine the interfacial capacitance at open circuit potential are
value. These galvanostatic measurements were repeated at 45 °C, 60 ° included in the supplementary information. We observed a “coral-
C, and 70 °C. In addition, the electrochemical impedance of the like” structure on the surface of the mesh electrode. After the

Figure 2. SEM images of HER electrode. (a) and (b) Cross-section and top-view of the as-prepared Ni-Mo co-sputtered surface. (c) and (d) Cross-section and
top-view of the Ni-Mo co-sputtered surface after electrochemical testing.
 Journal of The Electrochemical Society, 2020 167 114513

Figure 3. SEM images of OER electrode. (a) Bare steel mesh, (b) electrochemically modified substrate, (c) the as-prepared Ni-treated electrode, and (d) the Ni-
treated electrode after electrochemical tests.

thermal treatment we formed a catalytically active nickel hydroxide ambient atmosphere. The XPS investigation of the surface of the
layer (Fig. 3c). Thus, we concluded that pre-treating the steel mesh as-prepared OER electrode at the Ni 2p binding energy (Fig. 5c)
by cyclic polarization in sodium sulfide resulted in a high-surface revealed a peak at 855.6 eV corresponding to Ni2+ consistent with
area substrate upon which the electrocatalytically active layer was
deposited. This pre-treatment resulted in enhanced activity towards
OER higher than previously reported by our group.22 The coral-like
structure was maintained after 100-h stability tests as seen from the
SEM image of the OER electrode after electrochemical tests
(Fig. 3d). This stability is further supported by the absence of
significant change in the double layer capacitance values at 538 mV
of 6.06 F and 6.11 F, before the test and after the test, respectively.
The EDS spectrum of the as-prepared OER electrode (Fig. S3)
indicated the presence of nickel and oxygen on the surface of the
electrode along with sulfur that most likely was incorporated during
the sulfide pre-treatment process. Iron was also observed in the EDS
spectrum due to the electron beam penetration below the thin
catalytic layer.
The XRD pattern of the as prepared HER electrode (Fig. 4a)
showed 2 main peaks corresponding to Fe(110) and Fe(200)
indicating that the bulk of the electrode was elemental iron. The
XRD pattern of the HER electrode after electrochemical testing
indicated that there was no discernible change to the electrode
substrate as a result of the tests. The XRD pattern of the as-prepared
oxygen electrode (Fig. 4b) showed two main peaks corresponding to
Fe(110) and Fe(200) and two small peaks at 2-theta values of 23°
and 34° corresponding to Fe2O3(012) and Fe2O3(104), respectively.
The XRD pattern of the OER electrode after electrochemical tests
showed the two main elemental iron peaks indicating that the bulk of
the substrate had remained unchanged after the tests.
X-ray photoelectron spectroscopy (XPS, Kratos Axis Ultra
DLD) was used to characterize the electrode surfaces, as-prepared.
For the HER electrode, the Ni 2p binding energy values at
852.7 eV and 856.0 eV (Fig. 5a) indicated the presence of Ni0 and
Ni2+, respectively. The Mo 3d binding energy peaks at 228.1 eV
and 232.3 eV (Fig. 5b) were attributed to Mo0 and Mo6+,
respectively. Using the areas under the curves, it was estimated
that the molar ratio of Ni2+ to Ni0 was 3.4 to 1 while the molar
ratio of Mo6+ to Mo0 was 6.1:1. Although elemental Ni and Mo
were sputtered on the steel electrode, Ni2+ and Mo6+ were Figure 4. XRD spectra of (a) HER electrode before and after electroche-
spontaneously formed due to the exposure of the electrode to mical tests and (b) OER electrode before and after electrochemical tests.
 Journal of The Electrochemical Society, 2020 167 114513

Figure 5. Binding energy of (a) Ni (2p) and (b) Mo (3d) for the as-prepared
HER electrode and (c) Ni (2p) for the as-prepared OER electrode.

earlier studies from our laboratory that showed nickel(II) hydro- Electrochemical characterization of electrodes.—Steady-State
xide as the catalytic layer on the surface of the oxygen electrode.22 potentiostatic polarization measurements were performed by holding
Our attempt to include NiO to fit the Ni 2p peaks in Fig. 5c did not the potential of the working electrode for 300 s and recording of
yield a satisfactory fit. the steady state value of current density in the last 25 s of the
 Journal of The Electrochemical Society, 2020 167 114513

Table I. HER mechanism in alkaline conditions.

Step Reaction Tafel slope (mV decade−1)

Volmer Surface + H2O + e-  Hads + OH- 120

Heyrovsky Hads + H2O + e-  H2 (g) + OH- 40
Tafel 2Hads  H2 (g) 30

300-second period. After the 300-second potentiostatic measure- corresponded to the charge-transfer resistance of the HER electrode,
ment, the electrochemical impedance was measured over the we used the relationship between charge-transfer resistance and
frequency range of 10 kHz to 100 mHz. After completion of three overpotential (Eqs. 1 and 2). RCT corresponds to the charge-transfer
repeat sets of steady-state polarization experiments and one week in resistance while ηCT and I represent the charge-transfer overpotential
30% potassium hydroxide at open circuit, the electrochemical and the current density, respectively. B and io are the Tafel slope and
stability of the electrodes was determined by monitoring the exchange current density, respectively.
potential for 100 h at a constant current density of 10 mA cm−2.
The steady-state polarization data of the HER electrode (Fig. 6a) dhCT
R CT = [1]
showed that the catalytic layer was highly active towards HER with di
an overpotential of 166.8 mV at 10 mA cm−2, while the Tafel slope
(Fig. 6b) was calculated to be 115.4 mV decade−1. The performance h CT
of this electrode is comparable to hydrogen electrodes that are nickel
based.7,11 It is well known in the literature that the first step in the
log ( ) = log ( ) +
1
R CT
io
B B
[2]

hydrogen evolution reaction is the surface dissociation of water in Thus, according to Eq. 2, a plot of log(1/RCT) vs overpotential would
the Volmer step (Table I) resulting in adsorbed hydrogen atoms on be a straight line and the inverse of the slope would correspond to
the catalyst surface. The adsorbed hydrogen then combines to form the Tafel Slope. Plotting log(1/R2) vs overpotential (Fig. S4) yielded
hydrogen gas by the Heyrovsky pathway or the Tafel pathway.24 The a Tafel slope of 116.3 mV decade−1, which is close to the value of
Tafel slope values can be used to determine the rate limiting step. Tafel slope value of 115.4 mV decade−1 obtained from the potentio-
The Tafel slope of 115.4 mV decade−1 indicated that the Volmer static polarization measurements. Thus, R2 was attributed to the
step was the rate determining step. This observation is consistent charge-transfer resistance of the HER electrode.
with reports in the literature on Ni-Mo-based catalysts.14,25 The stability test performed on the HER electrode (Fig. 6d)
The Nyquist plots of the HER electrode (Fig. 6c) showed two showed that the electrode was stable under continuous polarization
semicircles with the diameter of the second semicircle (appearing at at 10 mA cm−2 for at least 100 h whereupon the overpotential
lower frequency) decreasing with increasing overpotential. The EIS increased at 3.4 μV h−1 during 100 h of continuous operation. This
data fitting (Fig. 6c inset) showed a series resistance appearing at miniscule increase of overpotential is a preliminary indication of
high frequency, which corresponded to the ohmic resistance of the good stability for extended operation.
potassium hydroxide solution. The resistance, R1, derived from the The OER electrode also showed high electrocatalytic activity
first semicircle at high frequency corresponded to a thin conductive with a low overpotential of 235 mV at 10 mA cm−2 (Fig. 7a). The
oxide film on the substrate that is present under ambient conditions. Tafel slope (Fig. 7b) was 46.9 mV decade−1 over the range of
The assignment of R1 to such a resistive layer is consistent with the 1–100 mA cm−2. This Tafel slope under high surface coverage of
observation that this resistance value does not change with over- hydroxyl ions was consistent with the Kobussen Pathway for OER
potential distinguishing it from a charge-transfer resistance. The (Table II) with the second step being rate determining.26 These
diameter of the second semicircle at low frequency corresponded to results were consistent with previously published work from our
R2 the charge-transfer resistance for the faradaic process of oxygen group on surface-modified iron electrodes for oxygen evolution.22,23
evolution. This assignment is consistent with the decrease of the The electrochemical impedance measurements of the OER
value of R2 with increasing overpotential. To confirm that R2 electrode yielded a Nyquist plot with two semicircles (Fig. 7c).
The first semicircle appearing at high frequency remained relatively
unchanged with changes in electrode potential while the diameter of
the second semicircle appearing at lower frequency decreased with
increasing overpotential. Just as in the case of the HER electrode, the
impedance data was fitted to an equivalent circuit model (Fig. 7c
inset) with Rs corresponding to the resistance of the potassium
hydroxide solution, R1 corresponding to the resistance due to the
compact oxide layer formed while fabricating the electrode, and R2
corresponding to the charge-transfer resistance of the catalytic nickel
hydroxide layer. To confirm that R2 corresponded to the charge-
transfer resistance in the OER electrode, log(1/R2) was plotted vs
overpotential (Fig. S5) yielding a Tafel slope of 52.9 mV decade−1
which is close to the value of 46.9 mV decade−1 obtained from

Table II. The Kobussen pathway.

Rate determining step Tafel slope

Step 1 M + OH-
aq + e
-  MOH + e- —
Step 2 MOH + OH-  MO + H2O + e- 2RT/3F
Figure 6. (a) Polarization data of the HER electrode, (b) Tafel slope of the Step 3 MO + OH-  MO2 H- ∞
HER electrode, (c) EIS Nyquist plots as a function of overpotential and EIS
data fitting (inset), and (d) chronopotentiometry of the HER electrode at Step 4 MO2 H + OH-  MO-2 + H2O + e-
- 2RT/F
10 mA cm−2. Step 5 MO-2  M + O2 + e- ∞
 Journal of The Electrochemical Society, 2020 167 114513

Table III. Comparison with the performance of various industrial electrolyzers.27–30

Company Temperature (°C) Pressure Cell voltage at 300 mA cm−2

De Nora 80 Ambient 1.85 V

Electrolyzer Corporation 70 Ambient 1.9 V
Teledyne Energy Systems 82 2 bar 1.9 V
General Electric 80 4 bar 1.7 V
This work 70 ambient 1.84 V

Figure 7. (a) Polarization data of the OER electrode, (b) Tafel slope of the Figure 8. (a) Galvanostatic data of the electrolyzer at different temperatures
OER electrode, (c) EIS data as a function of overpotential and the with the corresponding slopes shown in the inset, (b) EIS data as a function
corresponding EIS data fitting (inset) of the OER electrode, and (d) of overpotential, (c) EIS data fitting at 35 mA cm−2, and (d) constant current
chronopotentiometry of the OER electrode at 10 mA cm−2. operation of electrolyzer electrode at 1 A cm−2 at room temperature.

potentiostatic polarization experiments. This type of cross verifica- plots at all cell voltage values showed two semicircles with their
tion established that R2 was indeed the charge-transfer resistance of diameters decreasing with increasing current density suggesting that
the OER electrode. these semicircles corresponded to charge-transfer processes
Finally, the electrochemical stability of the OER electrode was (Fig. 8b). The two semicircles in the Nyquist plot at 35 mA cm−2
examined over 100 continuous h of polarization at 10 mA cm−2. The were fitted with two charge-transfer resistances, R1 and R2 (Fig. 8c).
electrode potential showed a marginal increase of overpotential The values of R1 and R2 for various overpotentials are shown in the
corresponding to 1.2 μV h−1 (Fig. 7d) indicating excellent stability Supplemental Material Table SI. Using Eqs. 1 and 2 to determine the
characteristics for the OER electrode. resulting Tafel slopes from each charge-transfer resistance, it was
concluded that R1 corresponded to the hydrogen evolution reaction
Electrolyzer electrochemical characterization.—An “all-iron” while R2 corresponded to the oxygen evolution reaction. To test for
alkaline electrolyzer was assembled using the HER and OER durability, the cell was operated at 1 A cm−2 for 100 h at room
electrodes fabricated with the modified steel substrates, as described temperature. The cell voltage of 2.13 V was maintained for the entire
above. The electrodes with a geometric are of 25 cm2 were period with no measurable change in this value (Fig. 8d). The all-
assembled with a commercially available Zirfon separator (Agfa), iron electrolyzer demonstrated here was found to have performance
and 30% potassium hydroxide as the electrolyte. The electroche- comparable to industrial electrolyzers, based on the literature
mical performance of the electrolyzer was measured using steady- data27–30 in Table III. Most importantly, we could achieve this
state galvanostatic polarization up to 1 A cm−2 and up to 70 °C. The performance with steel mesh based electrodes that could potentially
electrolyzer, at 100 mA cm−2, operated at voltages of 1.83 V and reduce the cost of the electrolyzer significantly as 50% of the cost of
1.71 V, at room temperature and 70 °C, respectively (Fig. 8a). The electrolyzer stacks is attributed to the nickel-based electrodes.3,31,32
Tafel slopes obtained from the IR-corrected voltage-current data
(Fig. 8a) are not constant, showing increasing values of slope with Conclusions
increasing overpotential. An example of the variation of the Tafel
slopes with potential is included in the Supplementary Material We have demonstrated the feasibility of building an “all-iron”
(Fig. S6). We ascribe the increasing slopes at high current densities alkaline electrolyzer with both hydrogen and oxygen electrodes
in Fig. 8a inset to the excessive generation of gas that leads to made of inexpensive iron-based substrates. We have shown that an
shielding of the electrode surface with bubbles. This issue is central inexpensive steel mesh may be catalyzed with a thin layer of nickel-
to electrolyzers operating at high current densities and often special molybdenum by co-sputtering to yield a hydrogen electrode that
expedients are devised in industrial electrolyzers to remove the operated at an overpotential of 167 mV at 10 mA cm−2. The
bubbles from blocking the electroactive surface. Furthermore, the preliminary test of durability of this electrode for 100 h at 10 mA
Tafel slope does not vary as a function of temperature indicating that cm−2 did not show any signs of degradation. Similarly, a steel mesh
the reaction mechanism of water splitting is not changing with first pre-treated anodically in sulfide-containing alkaline solutions
temperature. Electrochemical impedance measurements were col- and then coated with a thin layer of nickel hydroxide yielded an
lected between 15 and 35 mA cm−2 (1.62 V to 1.69 V). The Nyquist electrocatalytic oxygen evolution electrode. This electrode operated
 Journal of The Electrochemical Society, 2020 167 114513

at an overpotential of 235 mV at 10 mA cm−2. The performance of 5. M. Schalenbach, G. Tjarks, M. Carmo, W. Lueke, M. Mueller, and D. Stolten,
this type of oxygen electrode exceeds previous reports from our J. Electrochem. Soc., 163, F3197 (2016).
6. N. Mahmood, Y. Yao, J. W. Zhang, L. Pan, X. Zhang, and J. J. Zou, Adv. Sci., 5,
laboratory. Preliminary tests on the durability of this oxygen 1700464 (2018).
electrode indicated no signs of degradation even after 100 h of 7. C. C. L. McCrory, S. Jung, I. M. Ferrer, S. M. Chatman, J. C. Peters, and
operation at 10 mA cm−2. When these hydrogen and oxygen T. F. Jaramillo, J. Am. Chem. Soc., 137, 4347 (2015).
electrodes were combined to form an “all-iron” alkaline electrolyzer, 8. A. Garat and J. M. Gras, Int. J. Hydrogen Energy, 8, 681 (1983).
this cell could be operated continuously for 100 h at 1 A cm−2 with 9. L. Zhang, K. Xiong, Y. Nie, X. Wang, J. Liao, and Z. Wei, J. Power Sources, 297,
413 (2015).
no noticeable degradation. At 100 mA cm−2, the cell voltages were 10. J. R. McKone, B. F. Sadtler, C. A. Werlang, N. S. Lewis, and H. B. Gray, ACS
1.83 V and 1.71 V, at room temperature and 70 °C, respectively. The Catal., 3, 166 (2013).
current-voltage performance of this new “all-iron” alkaline electro- 11. M. Gong, D. Y. Wang, C. C. Chen, B. J. Hwang, and H. Dai, Nano Res., 9, 28 (2016).
12. L. Birry and A. Lasia, J. Appl. Electrochem., 34, 735 (2004).
lyzer was comparable to industrial electrolyzer systems. These 13. J. Kubisztal, A. Budniok, and A. Lasia, Int. J. Hydrogen Energy, 32, 1211 (2007).
results show that relatively inexpensive electrodes made with low- 14. Q. Han, S. Cui, N. Pu, J. Chen, K. Liu, and X. Wei, Int. J. Hydrogen Energy, 35,
carbon steel mesh could be a cost-effective alternative to nickel- 5194 (2010).
based electrodes for water electrolysis. Furthermore, these iron- 15. M. Xia, T. Lei, N. Lv, and N. Li, Int. J. Hydrogen Energy, 39, 4797 (2014) .
16. R. Solmaz, A. Döner, and G. Kardaş, Electrochem. Commun., 10, 1909 (2008).
based electrodes can be fabricated by processes that are readily 17. R. Solmaz, A. Döner, and G. Kardaş, Int. J. Hydrogen Energy, 34, 2089 (2009).
scalable for use in industrial systems. 18. D. E. Hall, J. Electrochem. Soc., 132, 41C (1985).
19. L. Trotochaud, S. L. Young, J. K. Ranney, and S. W. Boettcher, J. Am. Chem. Soc.,
Acknowledgments 136, 6744 (2014).
20. M. S. Burke, L. J. Enman, A. S. Batchellor, S. Zou, and S. W. Boettcher, Chem.
The Loker Hydrocarbon Research Institute and the University of Mater., 27, 7549 (2015).
Southern California supported the research. SEM images and the 21. E. Fabbri, A. Habereder, K. Waltar, R. Kötz, and T. J. Schmidt, Catal. Sci.
Technol., 4, 3800 (2014).
XPS data were acquired at USC’s Core Center of Excellence for 22. D. Mitra, P. Trinh, S. Malkhandi, M. Mecklenburg, S. M. Heald,
Nano Imaging. M. Balasubramanian, and S. R. Narayanan, J. Electrochem. Soc., 165, F392 (2018).
23. D. Mitra and S. R. Narayanan, Top. Catal., 61, 591 (2018).
ORCID 24. N. Krstajić, M. Popović, B. Grgur, M. Vojnović, and D. Šepa, J. Electroanal.
Chem., 512, 16 (2001).
Billal Zayat https://orcid.org/0000-0002-3794-3462 25. L. Mihailov, T. Spassov, I. Kanazirski, and I. Tsvetanov, J. Mater. Sci., 46, 7068
S. R. Narayanan https://orcid.org/0000-0002-7259-3728 (2011).
26. J. O. Bockris and T. Otagawa, J. Phys. Chem., 87, 2960 (1983).
References 27. K. Kinoshita, Electrochemical Oxygen Technology (John Wiley & Sons, Inc.,
Hoboken, NJ) (1992).
1. U.S. Department of Energy, Fuel Cell Technologies Office Multi-Year Research, 28. A. Buttler and H. Spliethoff, Renew. Sustain. Energy Rev., 82, 2440 (2018).
Development, and Demonstration Plan (2012), https://energy.gov/eere/fuelcells/ 29. I. Dincer and C. Acar, Int. J. Hydrogen Energy, 40, 11094 (2014).
downloads/fuel-cell-technologies-office-multi-year-research-development-and-22. 30. M. David, C. Ocampo-Martínez, and R. Sánchez-Peña, J. Energy Storage, 23, 392
2. S. Dunn, Int. J. Hydrogen Energy, 27, 235 (2002). (2019).
3. S. Marini, P. Salvi, P. Nelli, R. Pesenti, M. Villa, M. Berrettoni, G. Zangari, and 31. Z. Liu, S. D. Sajjad, Y. Gao, J. J. Kaczur, and R. I. Masel, ECS Trans., 77, 71
Y. Kiros, Electrochim. Acta, 82, 384 (2012). (2017).
4. K. Zeng and D. Zhang, Prog. Energy Combust. Sci., 36, 307 (2010). 32. Y. Nakajima, N. Fujimoto, S. Hasegawa, and T. Usui, ECS Trans., 80, 835 (2017).