Highly Active and Stable NiCuMo Electrocatalyst Supported on 304 Stainless Steel Porous Transport Layer for Hydrogen Evolution in Alkaline Water Electrolyzer

Several functionalized porous transport layers with Pt‐free electrocatalysts for hydrogen evolution reaction in alkaline conditions, based on Ni, Cu, and Mo, are prepared through electrodeposition onto a 304 stainless steel mesh. Morphological characterization confirms the fabrication of electrodes with high electrochemical surface active area due to the formation of hierarchical nanostructures. Mo presence into the electrocatalysts increases the activity toward the hydrogen evolution reaction. The optimization of electrodeposition process leads to the preparation of highly active NiCuMo electrocatalyst that exhibits near zero onset overpotential and overpotentials of 15 and 113 mV at 10 and 100 mA cm−2, respectively, in 1 m KOH electrolyte. Moreover, this electrocatalyst shows superior stability with respect to other Pt‐free electrocatalysts, reaching 100 h of durability with low overpotentials value demonstrating the successful preparation of very promising functionalized porous transport layers for future‐generation alkaline electrolyzers.


Introduction
Renewable energy storage became more and more critical in the path for a decarbonized society. [1][2][3] A way to store renewable electrical energy is producing fuels and chemicals that do not suffer of self-discharge phenomena. In this context, producing (and storing) green hydrogen is the key of the so called "hydrogen economy" and can be achieved through the electrochemical water-splitting process that produces hydrogen and oxygen at cathode and anode electrodes, respectively. [4][5][6] In particular, Cu-doped nanotubular Ni electrocatalyst was found to be highly active due, also, to the synergistic effect of Cu that, substituting to Ni sites in crystal structure, leads to an optimal ΔG H* for hydrogen adsorption, therefore favoring H 2 desorption. [20] Based on these findings, we developed high surface area NiCuMo electrocatalysts for HER with excellent activity and long-term durability. NiCuMo was deposited on a cheap 304 stainless steel (SS) mesh through an optimized one-step electrodeposition process in order to produce cost-effective and active CCS for zero-gap alkaline water electrolyzer. Electrodes were studied through morphological and chemical analysis carried out with SEM and EDX. Electrochemical characterization was carried out to have information about catalysts surface area and activity in 1 m KOH aqueous solution. Then, galvanostatic long-term durability test was performed in a flow-through cell.

Results and Discussion
NiCuMo electrocatalysts were fabricated by an electrodeposition process on a mesh made in AISI 304 SS, as described in the Experimental section.
The electrodeposition was conducted in an aqueous solution containing 0.5 m NiSO 4 , 12.5 m M CuSO 4 , 0.5 m H 3 BO 3 , and a variable concentration of (NH 4 ) 6 Mo 7 O 24 , from 7 mM to a concentration of 28 mM. Samples were then identified on the basis of the concentration of Mo-containing salt in the electrodeposition electrolyte, i.e., 0Mo is the sample obtained with no Mo in solution during the electrodeposition, 1Mo is the sample obtained with a 7 mM concentration of Mo salt arriving to 4Mo for the sample obtained with a 28 mM concentration. The electrolyte pH was corrected to 2.5 by the addition of H 2 SO 4 . The electrodeposition potential was set to -2 V (Ag/AgCl), so that Ni, Cu, and Mo are thermodynamically stable in their metallic state according to the following reactions and corresponding equilibrium potentials: [21] During the electrodeposition process, a cathodic current slightly increasing with time was recorded for all the investigated bath composition (see Figure S1, Supporting Information).
After the electrodeposition process, it was important to estimate the roughness factor of the NiCuMo electrodes to assess the electrochemically active surface area (ECSA). At this aim, we recorded EIS spectra in 1 m KOH in O 2 -free environment at 0.15 V RHE to hinder the possible O 2 reduction process, thus having only capacitive current relating to the double layer capacitance (C DL ). Indeed, at this electrode potential, water reduction reaction is not thermodynamically possible. EIS spectra related to all the electrodes produced by electrodeposition process are shown in Figure 1a) in Nyquist representation. For comparison, EIS spectra of not functionalized 304 SS mesh and 0Mo electrode are reported in Figure S2 (Supporting Information).
EIS spectra were fitted with the equivalent electrical circuit reported in the inset of Figure 1b), i.e., a series between R s , the electrolyte resistance, and a parallel between R CT , a resistance to model the charge transfer resistance, and Q DL,el , a constantphase element (CPE) modeling the nonideal electrode double layer capacitance. The electrode double layer capacitance, C DL,el , is derived from Q DL,el by the following equation: [22] Fitting parameters (see Table S1, Supporting Information) clearly show a significant reduction of charge transfer resistance and a relevant increase of the double layer capacitance after the electrodeposition process that can be explained by the enhanced roughness (and thus active area) of the coated electrodes. Assuming a double layer capacitance of 50 µF cm −2 (as that measured using a flat mirror-finished AISI 304 coupon), the roughness factor r was estimated according to the following equation: r values are reported in Table 1 for all the electrodes.

www.advsustainsys.com
Bare 304 SS mesh has a r of ≈4, while coated electrodes have r values increasing with an increased concentration of Mo salt in the solution, ranging between 71 for 1Mo sample and 164 for 4Mo sample. Therefore, tailoring the electrodeposition process, it is possible to obtain electrodes with an increased ECSA, necessary to have high activity for HER reaction. However, the roughness factors reported in Table 1 can only in part explain the huge differences in the R CT estimated by the fitting procedure. Indeed, the charge transfer resistance for the 304 SS mesh is very high (3 × 10 4 Ω cm 2 ) as expected for an ideally polarizable interface, and two order of magnitude higher than that estimated for coated electrodes. The very low charge transfer resistance can be explained considering that at this potential a Faradaic process can occur (namely Mo oxidation in agreement with the thermodynamic predictions of Molybdenum Pourbaix diagram).
Based on catalyst mass loadings estimated from mass differences before and after the electrodeposition process (see Figure S3, Supporting Information), it is possible to obtain high specific surface area values ranging between 3 m 2 g −1 for 1Mo sample to 11 m 2 g −1 for 4Mo sample.
SEM images, related to 0Mo and 3Mo electrodes, are shown in Figure 2a-d), while for 1Mo, 2Mo, and 4Mo electrodes, SEM micrographs are reported in Figure S4 (Supporting Information). In the case of electrodeposition without Mo ions in solution (0Mo sample), NiCu coating is not uniform throughout the 304 SS mesh substrate and it presents agglomerates with needle-shaped tips. This leads to a not high roughness factor (see Table 1), only ≈four times higher than that estimated for 304 SS mesh sample. The presence of MoO 4 2− ions in the electrodeposition bath (see Figure 2c,d and Figure S4, Supporting Information) leads to higher roughness factors due to more  www.advsustainsys.com ordered nanostructure. In fact, in the case of 3Mo electrode, after the electrodeposition process, the 304 SS mesh surface is totally covered with particles with average diameter of ≈600 nm, organized in hierarchical structures (see Figure 2d). Same hierarchical structures can be observed for 1Mo, 2Mo, and 4Mo electrodes (see Figure S4, Supporting Information). The crystal structures of the electrodeposited samples were characterized by XRD. As shown in Figure 3, the XRD patterns for Ni and 0Mo are crystalline with three main reflections at 2θ = 44.44°, 51.26°, and 76.35°, corresponding to (111), (200), and (220) planes of face-centered cubic Ni. For 0Mo, the shoulder peaks appear at the base of Ni (111) and (200), indicating the partial formation of NiCu alloy. [23] This result suggests that partial replacement of Ni with Cu leads to an expansion of pure Ni lattice, as Cu has larger lattice parameter (i.e., 0.362 nm) than that related to Ni structure (i.e., 0.350 nm), therefore in agreement with the alloy formation. Several papers report such experimental findings during co-electrodeposition of Ni and Cu. [24][25][26] When the Cu content is higher than almost 50%, there can be segregation of the individual atoms rather than the formation of a homogeneous alloy. [25] It is worth noting the effect of Mo presence on the crystallinity degree of Ni. The characteristic Ni peaks disappeared in the Mo-containing Ni sample, suggesting a highly disordered structure. Moreover, no characteristic peaks corresponding to Mo or NiMo are present, indicating that Mo atoms are dissolved into the lattice of Ni. For NiCuMo samples, a small and broad peak at ≈43.5° was observed, which suggests the poor crystallization of metallic Ni coating, as well as that a portion of Ni crystalline structures were kept in spite Mo deposition. [27] The XRD diffraction peak of the NiMo alloy should have a small shift relative to the pure Ni coating [27] but, in the case of NiCuMo alloys, it overlaps to NiCu diffraction peak. All these experimental findings suggest that solid solution of Cu and Mo in Ni is formed in the exploited composition range. Figure 4 shows the XPS Cu 2p, Mo 3d, and Ni 2p core level spectra. In Figure 4a, the Cu 2p core level spectra exhibit peaks at 932.5 eV related to Cu 0 , 933.8 and 934.5 eV related to Cu 2+ . Another peak corresponding to Cu 2+ were observed at higher binding energy close to 954 eV. [28][29][30] For Ni 2p core level spectra (Figure 4b), three peaks are detected: peak at 852.9 eV related to Ni 0 while peaks at 856 and 861 eV are related to Ni 2+ . [31][32][33] Mo 3d core level spectra ( Figure 4c) show two peaks at 232.6 and 235.7 eV, both related to Mo 6+ species. [31] In Table 2, the atomic compositions obtained by XPS fitting are reported.
XPS characterization reveals the presence of oxidized species for Cu, Ni, and Mo on the samples, while XRD characterization detects patterns diffraction related to only metallic species. This could be explained by considering that XRD is a bulk technique while XPS is a surface technique and thus it is able to detect thin-air-formed passive films on metallic substrates. Moreover, when the electrode works as cathode for the hydrogen evolution process, it is likely that these species are also reduced due to the very negative working potential, thus cancelling the effect of air exposure.
Electrocatalytic performance of NiCuMo electrodes was evaluated by electrochemical measurements, more specifically galvanodynamic and electrochemical impedance spectroscopy measurements.
Galvanodynamic measurements (with 95% iR drop correction) carried out in 1 m KOH aqueous solution are reported in   It is immediately noticeable that the presence of Mo in the electrodes plays a crucial role, greatly improving the electrocatalytic performance. In fact, 304 mesh and NiCu sample (i.e., 0Mo electrode) display an onset overpotential for HER of 300 and 70 mV, respectively. The presence of Mo, regardless of Mo salt concentration during the electrodeposition process, leads to an onset overpotential for HER close to 0 mV. This result highlights that a high surface area is not sufficient to obtain highly active electrocatalysts but it even more important the chemical composition of the catalyst surface. By looking at η 10 and η 100 values (reported in Figure 5b), it can be seen that 3Mo sample is the best electrode with respect to electrocatalytic performance. η 10 and η 100 values of 443 and 577 mV, respectively, were estimated for bare 304 SS mesh, while η 10 and η 100 values of 302 and 486 mV, respectively, were estimated for the electrode not containing Mo. Optimizing Mo ions concentration during the electrodeposition process led to obtain an electrode with highly performing characteristics, in fact, very small η 10 and η 100 values of 15 and 113 mV, respectively, were estimated for 3Mo electrode. η 10 value is currently one of most reported benchmarks for the activity of an electrocatalyst and it is related to the potential at -10 mA cm −2 corresponding, for a photoassisted water-splitting process, to a solar-to-hydrogen efficiency of 12.3% (i.e., the required efficiency for cost-competitive solar water splitting). [34] The value obtained for 3Mo electrode is close to the lowest ever-reported η 10 value for heterostructures electrocatalysts for HER in alkaline solutions. [20,[35][36] This result is outstanding, considering also that 3Mo electrode is obtained in a one-step electrodeposition process using a very cheap 304 SS mesh as PTL. Figure 5c shows linear portions of the Tafel plot (overpotential η vs log(i) related to the curves reported in Figure 5a) for all the synthesized electrocatalysts. Linear part of Tafel plot can be fitted according to the following equation: Adv. Sustainable Syst. 2023, 7, 2200486

www.advsustainsys.com
where a is directly proportional to the logarithm of the exchange current density that is a measure of the inherent electrocatalytic properties of the material, and b is called Tafel slope that is another important kinetic parameter, and it is related to the reaction mechanism. The exchange current density, i 0 , is significantly affected by the composition of the catalytic layer and, to cancel the contribution due to the different roughness factors on i 0 , the latter were normalized for r (see Table 3). The results show that the coating deposited from 3Mo has the best i 0 and k 0 , i.e. the standard rate constant. Estimated Tafel slope values are reported in Figure 5c. The lower is b, the better is the electrocatalytic performance of the electrode. The lowest Tafel slope (114 mV dec −1 ) was assessed for 3Mo electrode, confirming the best electrocatalytic performance for this electrode. This Tafel slope value is compatible with a reaction mechanism that foresees the Volmer reaction (i.e., water molecule reacts with an electron on the electrode surface to generate an adsorbed H* atom) as rate-determining step of HER process. [13,34] EIS spectra, recorded at -0.15 V RHE, are shown Figure 5d in Nyquist representation. EIS spectra were fitted according to the equivalent electrical circuit reported in Figure 1b. Fitting parameters are reported in Table S2 (Supporting Information) for all the electrodes. As it is possible to note, the lowest R CT value (0.21 Ω cm 2 ) was found for 3Mo electrode, in agreement with the lowest overpotential recorded for this electrode with respect to the other electrodes synthesized for this work. The presence of Mo into the electrocatalyst can facilitate the kinetics of water adsorption, [37] i.e., of the first step of HER in alkaline medium, therefore having a synergistic action with Ni and Cu, which facilitate hydrogen desorption for the surface, leading to very high catalytic activity.
In order to understand the effect of single elements, therefore of the layer composition, on the performance toward HER, we recorded current density versus potential curves also for electrodeposited Ni, Cu, Mo, NiMo, CuMo samples to have a complete view of the catalysts effect (see Figure S5, Supporting Information). Ni, Cu, Mo, and CuMo samples have worse performance toward hydrogen evolution reaction, in terms of η 10 and η 100 values with respect to the best NiCuMo alloy. This is due to lower exchange current density values corrected for the roughness factor estimated by the procedure described in the manuscript. In Tables 1 and 3 are also reported roughness factor values and exchange current density values, respectively, for the Ni, Cu, Mo, Ni-Mo, and Cu-Mo samples.
It is possible to note the effect of Cu addition in increasing the real-active surface area, as also confirmed by SEM images reported in the Supporting Information file for Cu and CuMo samples (see Figures S7 and S9, Supporting Information). It is worth noting that, in the case of NiMo sample, the roughness factor remains low, due to the absence of Cu, but the electrocatalytic activity (i.e., real i 0 , see Table 3) is higher than any other sample. Anyway, the performance is lower than best NiCuMo alloy, highlighting that the composition of the layer plays a key role. Indeed, H-adsorption on Cu-doped Ni(111) is weaker than in the case of Ni with consequent enhancement of HER activity as reported in the literature [20,23] and as confirmed by the experimental results of this work.
It is very well known that the interaction between H atom and catalyst surface should be neither too weak, to correctly adsorb reactants, nor too strong, to efficiently release produced hydrogen. The presence of two empty d-orbitals in Ni, useful for H atom adsorption/desorption, can efficiently promote electron transfer and H-O bond breakage. According to Brewer Engel valence bond theory, the electron density of transition metals with more-filled d orbitals, as Ni, can be modified when alloyed to metals as Mo with empty or half-filled d-orbitals, that in turn facilitate M-H ads formation. [38] The presence of both Cu (i.e., a M-H bond weakener) and Mo (i.e., M-H bond strengthener) allows a good compromise with consequent enhancement of the catalytic activity.
Another crucial criterion for the practical application of these electrocatalysts is the electrochemical stability, i.e., a constant overpotential for HER with operating time, sign of a stable electrochemical activity of the electrode. Chronopotentiometric curve of 3Mo electrode recorded at -100 mA cm −2 in 1 m KOH is shown in Figure 6a.
After a transitory of ≈1 h, the overpotential for HER remains quite stable for (at least) 100 h, with a value of 165 ± 45 mV, demonstrating the high stability of this electrode. This is the lowest overpotential value if compared with other values reported in the literature for stability tests carried out at 100 mA cm −2 for 100 h. [23,39] To further demonstrate the outstanding stability of 3Mo electrocatalyst, the electrode was studied after the stability test with morphological characterization and evaluating Ni, Cu, and/or Mo loss during 100 h polarization by ICP-OES.
By ICP-OES analysis, Ni and Cu ions were not detected in the electrolyte samples collected during the stability test. This result can be explained by considering that, during shutdown times, Ni and Cu are thermodynamically stable in 1 m KOH aqueous solution while, during operating time, they are cathodically protected operating in immunity zone of corresponding Pourbaix diagrams. [21] Regarding Mo, a certain concentration of Mo species in the electrolyte during the stability test was detected. In fact, both during shutdown and operating times, Mo can dissolve according to the following reaction: [21]   Anyway, concentration of Mo ions in the electrolyte keeps really low values during stability tests, from 0.469 mg L −1 at 0 h to 0.664 mg L −1 at the end of stability test (see Figure 6b), corresponding to a dissolution rate equal to 40 nmol cm −2 h −1 , further demonstrating the high stability of the electrode during operation time.
SEM images and catalyst composition reported in Figure 7 were obtained after 100 h stability test.
Average particles diameter is ≈1.3 µm, i.e., average diameter value doubled after the stability test indicating the loss of ECSA. At the same time, chemical composition indicates Mo loss during 100 h polarization test. However, HER overpotential and R CT values remained almost unchanged during the stability test, indicating a preserved electrochemical activity of the catalyst.

Conclusions
We optimized an electrodeposition process to produce Ptfree and CRM-free NiCuMo electrocatalyst, supported by a 304 SS mesh, for HER reaction showing high catalytic activity and high stability. Electrocatalysts' high electrochemical surface active area was assessed by electrochemical measurements and by SEM and it is due to hierarchical nanostructures. Best η 10 and η 100 values of 15 and 113 mV, respectively, were estimated that are lower of most of the Ptfree and CRM-free state-of-art electrocatalysts for HER. Furthermore, outstanding stability was demonstrated, i.e., 100 h at -100 mA cm −2 with an overpotential value of 165 ± 45 mV, with neglecting Ni and Cu dissolution and low dissolution rate for Mo during operating conditions. The combination of Ni and Cu with Mo led to very active electrocatalyst, demonstrating the successful preparation of very promising functionalized porous transport layers for future-generation alkaline electrolyzers.

Experimental Section
Preparation of Stainless Steel Mesh Electrodes: All electrodeposition processes were carried out in a standard three-electrode electrochemical cell with a 304 SS mesh or fluorine tin oxide (FTO)-coated glass substrates as the working electrode, a platinum mesh as the counter electrode, and Ag/AgCl/3.5 m KCl (0 V vs Ag/AgCl = 0.2 V vs SHE) as the reference electrode.
Prior to every electrodeposition, SS mesh was first ultrasonicated in 0.5 m H 2 SO 4 solution for 10 min, rinsed subsequently with water, and then dried in air. The electrodeposition was conducted in 50 mL    4 2ions. The electrodeposition was carried out with a VERSASTAT 3 workstation at -2.0 V vs Ag/AgCl for 600 s at room temperature.
Electrochemical Measurements: All electrochemical measurements were carried out with a PARSTAT 2263 workstation in a flow-through cell. The as-prepared electrode on 304 SS mesh was directly used as the working electrode without any further treatments. The exposed surface area of the steel mesh electrode was 0.5 cm 2 . A dimensionally stable anode (DSA) and Hg/HgO/1 m NaOH were used as the counter electrode and reference electrode, respectively. All potentials measured were then referred to the reversible hydrogen electrode (RHE) using the following equation: EIS spectra were carried out at 0.15 V RHE in the case of estimating double layer capacitance (in a N 2 atmosphere) and at -0.15 V RHE in the frequency range of 0.1-10 5 Hz with an a.c. signal amplitude of 10 mV in 1 m KOH electrolyte by using PARSTAT 2263 workstation. EIS spectra were then fitted with ZSimpWin software with a suitable equivalent electrical circuit, described in the Results and Discussion section.
Galvanodynamic tests were carried out with a scan rate 1 mA cm −2 s −1 between -1 and -100 mA cm −2 . Stability tests were carried out as chronopotentiometric measurement at -100 mA cm −2 for 100 h.
Morphological and Structural Characterization: The specimens were examined by scanning electron microscopy (SEM) using a FEI Quanta 200 FEG SEM instrument at several magnifications, combined with a X-ray energy dispersive system (EDX).
X-ray diffraction (XRD) patterns for FTO after electrodeposition were recorded at room temperature on a PANalytical Empyrean diffractometer equipped with a PIXcel1D (tm) detector using the CuKa radiation. A Ni filter was mounted at the exit of the X-ray source. In order to achieve, we used the highest signal-to-noise ratio, the operating conditions were 40 V and 40 mA.
XPS Characterization: X-ray photoelectron spectroscopy (XPS) analysis was carried out using a PHI5000 VersaProbe II scanning microprobe (ULVAC-PHI), operating with a monochromatic Al Ka source and X-Ray beam of 100 µm diameter. A take-off angle of the emitted photoelectrons of 45° relative to the surface was used. Peaks fitting was performed with MultiPak 9.9.2 (ULVAC-PHI), using Gauss-Loretz model and a Shirleytype background.
ICP-OES Test: For ICP-OES tests, all chemicals used were of analytical grade (Sigma Aldrich). To estimate metal release in solution, a calibration procedure was carried out with 0 (ultrapure water), 2, 5, 10, 20, 50, 100, 200, and 500 ppb for Ni, Cu, and Mo standard solutions. PerkinElmer Inc.-Optima 2100 DV was used for ICP-OES analysis. Samplings were carried out at the beginning of stability test, then after 25, 50, 75 h and at the end of test.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.