A Comparative Study of NiCo2O4, NiO, and Co3O4 Electrocatalysts Synthesized by a Facile Spray Pyrolysis For Electrochemical Water Oxidation

Exploiting low‐cost, highly active, and robust oxygen evolution reaction (OER) electrocatalysts based on earth‐abundant elements by a simple synthesis approach holds paramount importance for green hydrogen production through water electrolysis. In this work, the NiO, Co3O4 and NiCo2O4 nanoparticle layers with identical surface morphologies are prepared under same deposition conditions by a simple spray pyrolysis method and their OER activities are comparatively investigated. Among all these three electrocatalysts, NiCo2O4 shows the lowest overpotential of 420 mV to drive benchmark current density of 10 mA cm−2 and the smallest Tafel slope (84.1 mV dec−1), which are comparable to the OER performance of the benchmark commercial RuO2 electrocatalyst. The high OER activity of NiCo2O4 is attributed to the synergy effect and the modulation of electronic properties between Co and Ni atoms, which drastically reduces the overpotential required to drive OER activities. Therefore, it is believed that the NiCo2O4 synthesized by this simple method would be a competitive candidate as an industrial electrocatalyst with high‐efficiency and low cost for large‐scale green hydrogen production via water electrolysis.


Introduction
[6] The electrochemical water splitting process consists of the hydrogen evolution reaction (HER) at the cathode and the oxygen evolution reaction (OER) at the anode. [7][12][13][14] Therefore, the OER is regarded as the bottleneck of the overall water splitting.Currently, the commercial water electrolysis cells mainly rely on noble metals (Pt, Ru, Ir, etc) and their oxides that exhibit the benchmark HER and OER catalytic activities. [15]However, their high cost and scarcity have significantly hindered large-scale production of low-cost hydrogen. [16][23][24][25] Particularly, nickel oxide, cobalt oxides, and their modified compounds as OER electrocatalysts have been investigated.Acedera et al. reported that porous Co 3 O 4 nanoparticles exhibited a low onset potential (353 mV) and a small Tafel slope (74.93 mV dec −1 ) by employing fuel-lean. [26]Tran-Phu et al. have synthesized Co 3 O 4 spinels with high Co 2+ /Co 3+ ratio that boosted the OER activity (an overpotential of 346 mV @ 10 mA cm −2 and a Tafel slope of 40 mV dec −1 ). [27]Meanwhile, nickel oxides are more attractive for practical application due to its abundancy, low cost, and less toxicity.For instance, Silva et al. investigated the dependency of surface morphology on the OER activity of NiO and found that 1D hollow nanofiber NiO showed a low overpotential (322 mV @ 10 mA cm −2 ) and Tafel slope (78.8 mV dec −1 ). [28]Liang et al. reported a novel leaven dough method to prepare NiO/Ni foam, which exhibited an overpotential of 345 mV at 10 mA cm −2 and a Tafel slope of 53 mV dec −1 . [29][32][33][34] Fang et al. showed that both the (111) and (110) facets in NiCo 2 O 4 nanosheets exhibited high OER activity. [30]Bao et al. reported that mesoporous Ni x Co 3-x O 4 nanoneedle arrays exhibited a large electrochemically active area and showed a promising OER activity (320 mV at 10 mA cm −2 , Tafel slope of 38 mV dec −1 ). [31]However, it is not straightforward to compare the OER performance of NiO, Co 3 O 4 , and NiCo 2 O 4 reported in literatures since these materials were synthesized by a large variety of methods and exhibited quite different surface morphologies and structural features.[34] To the best of our knowledge, a direct comparison study of OER performance of NiO, Co 3 O 4, and NiCo 2 O 4 with identical surface morphologies prepared by a method under same conditions has not been done yet.
In this work, in order to compare the OER performance, we deposited the NiO, Co 3 O 4 and NiCo 2 O 4 layers on fluorine tin oxide (FTO) substrates under same deposition conditions by a simple spray pyrolysis method.The prepared NiO, Co 3 O 4 and NiCo 2 O 4 layers exhibited identical surface morphologies, which enabled us to comparatively study their OER performance.Among all these three electrocatalysts with same surface morphology, NiCo 2 O 4 shows the lowest overpotential and the smallest Tafel slope, which are close to the activity performance of commercial OER electrocatalyst, RuO 2 .The high OER activity of NiCo 2 O 4 is attributed to the synergy effect and the modulation of electronic properties between Co and Ni atoms, which drastically reduce the overpotential required to drive OER activities.This facile spray pyrolysis approach paves the way for developing the low-cost and efficient electrocatalysts toward large-scale production of renewable hydrogen via water electrolysis.

Results and Discussion
The NiCo 2 O 4 , Co 3 O 4, and NiO layers were deposited on 500 nmthick FTO under the same conditions by a spray pyrolysis approach followed by a post calcination process (Figure 1).All three oxide layers exhibit an identical surface morphology, consisting of dense nanoparticles covered completely on FTO without exposed FTO grains (Figure 2).The cross-sectional SEM images show a quite similar thickness for all three layers (330 nm for NiCo 2 O 4 , 280 nm for Co 3 O 4 , and 430 nm for NiO) (Figure S1, Supporting Information).The XRD pattern shows that the NiCo 2 O 4 is indexed to JCPDS # 20-0781 (Figure 3) and three obvious diffraction peaks are indexed to (111), (311), and (222) at 2 = 18.9°, 36.7°, and 38.4°, respectively.XRD pattern of NiO layer is ascribed to PDF # 47-1049.Notably, <111> is the dominant orientation for all NiCo 2 O 4 , Co 3 O 4 , and NiO layers.Facet (111) of NiCo 2 O 4 was reported to show a higher OER activity than facet (110). [30]It is worth noting that all three NiCo 2 O 4 peaks are leftshift 0.1 degree compared with Co 3 O 4 (JCPDS # 42-1467) due to the lattice expansion by substitution of Co atom with Ni atom. Figure 4a shows the Raman spectrum of the NiCo 2 O 4 .The peaks at 184, 460, 505, and 650 cm −1 are related to the Ramanactive modes of F 2g , E g F 2g , and A 1g , respectively. [35,36]The Raman peaks of Co 3 O 4 at 477, 518, 615, and 683 cm −1 are ascribed to the active modes of E g , 2F 2g and A 1g (Figure 4b). [37,38]The Raman spectrum of NiO shows five peaks (Figure 4c).The stronger peaks at 555 and 1091 cm −1 represent the first-order longitudinal (LO) mode and second-order longitudinal (2LO) of NiO, respectively, which are related to the vibration of Ni═O bonds.The peaks at 461 and 780 cm −1 represent the first-order transverse mode (TO) and second-order transverse mode (2TO), respectively.Besides, the broad peak from 946 to 951 cm −1 was reported to be attributed to the combination of TO + LO phonon excitation mode of NiO. [39,40]The Raman results further confirm the phase formation of NiCo 2 O 4 , Co 3 O 4 , and NiO, consistent with the XRD results.It should be noted that the identical surface morphology, similar thickness, and same dominant orientation of all three NiCo 2 O 4 , Co 3 O 4 , and NiO layers prepared by the same method enable us to directly compare their OER performance with minimized effects from morphologies, facets, and synthesis methods.
X-ray photoelectron spectroscopy (XPS) measurements were carried out to investigate the surface composition and chemical states of NiO, Co 3 O 4, and NiCo 2 O 4 .XPS survey spectra confirm existence of Co, Ni, and O with their characteristic peaks (Figure S2, Supporting Information).In addition, core level signals from C, Na, and Ca were also detected.The former is due to the adventitious carbon that accumulates on all surfaces exposed to air, [41] while the latter might be caused by minor impurities from precursors with inert catalytic ability in OER.As shown in Figure 5a, the Ni 2p spectrum is characteristic of NiO and features two Ni 2p 3/2 peaks (853.4 and 855.1 eV), two Ni 2p 1/2 peaks (870.9 and 872.8 eV), and the satellites (860.2 & 862.8 eV and 877.8 & 881.2 eV). [42][45] The O 1s spectrum of NiO exhibits a peak at 529.0 eV, which is ascribed to the lattice oxygen (Figure 6a).Another broad peak is due to C─O (531.5 eV) and C═O (530.4 eV) species from the adventitious carbon contamination. [46]The Co 2p spectrum from Co 3 O 4 (Figure 5b), reveals that two doublets are present with the 2p 3/2 peaks at 778.9 and 780.1 eV corresponding to Co 3+ and Co 2+ . [47]The corresponding O 1s spectrum shows three peaks at 529.1, 530.8, and 531.8 eV (Figure 6b).The peak at 529.1 eV is ascribed to the lattice oxygen, and the latter peaks (530.8 and 531.8 eV) are related  6c.These XPS results are consistent with the previously reported data. [48,49]he OER activities of NiCo     The electrochemically active surface area (ECSA) can be assessed from determination of the double-layer capacitance (C dl ) using cyclic voltammetry (Figure S3, Supporting Information) because C dl is proportional to the ECSA of the electrocatalyst.As shown in Figure 8a, the C dl of the NiCo 2 O 4 is determined to be 4.90 mF cm −2 , which is about 10 times higher than that of Co 3 O 4 (0.46 mF cm −2 ), and 245 times higher than that of NiO (0.02 mF cm −2 ).This indicates that the ECSA of NiCo Catalytic kinetics is further evaluated by electrochemical impedance spectroscopy (EIS).The Nyquist plots can be fitted by an equivalent circuit shown in the inset of Figure 8b and the fitted results are given in Table S2 (Supporting Information).NiCo 2 O 4 exhibits the smallest charge transfer resistance (R ct ) of 3.46 Ω cm 2 compared to Co 3 O 4 (6.43 Ω cm 2 ) and NiO (1272 Ω cm 2 ).The smallest R ct can be explained by the largest ECSA of NiCo 2 O 4 among all three oxides, which enhances the charge transfer from the anode to electrolyte.This result confirms that NiCo 2 O 4 has higher intrinsic OER activity than its counterparts (Figure 8).
Another crucial factor is the stability of the catalysts under high current densities.Chronopotentiometry technique is adopted to assess the durability of NiCo 2 O 4 .At a high current density of 100 mA cm −2 , NiCo 2 O 4 shows remarkable stability without any change of the applied voltage for 10 h (Figure 8c).After the stability measurement, the LSV curve was measured again, and it showed negligible difference compared to the LSV curve before stability test (Figure 8d).A comparison of XRD patterns and SEM images measured before and after the stability tests shows no changes of crystalline structure and surface morphology, (Figures S4 and S5, Supporting Information), which further confirms the robust stability of NiCo 2 O 4 .
To further understand the mechanism of the superior OER performance of NiCo 2 O 4 over Co 3 O 4 and NiO, we compare our experimental results with the density functional theory (DFT) calculations.The OER reaction pathway and steps are generally suggested as follows: where * represents an OER catalyst.For OER to start, the ability to bond hydroxyl group (OH − ) is crucial.Transition metal atom sites are generally more active for bonding the hydroxyl group than the oxygen atom sites.Additionally, the vicinities around metal atom are generally active in bonding OH − . [50,51]DFT calculations indicate the bonding of OH step (Equation 2) on NiCo 2 O 4 and Co 3 O 4 is relatively easier than that on NiO, suggesting fast kinetics on NiCo 2 O 4 and Co 3 O 4 over NiO at the beginning of OER. [50]Our experimental result shows that NiO exhibits the largest overpotential and the Tafel slope among the three catalysts (Figure 7b,c), suggesting more sluggish OER on NiO.This result agrees with the DFT calculations. [51]In NiCo 2 O 4 , with replacing one Co atom with Ni atom, the electronic properties in the vicinities of Co and Ni have been modified notably compared with that of Co 3 O 4 .We believe that the enhanced OER activity of NiCo 2 O 4 is attributed to the synergy effect and the modulation of electronic properties between Co and Ni atoms, which drastically reduces the overpotential required to drive OER activities.This is also confirmed by the DFT calculations. [52,53]Specifically, Gong et al. reported that the Gibbs free energy barrier in the step of OH * to O * (Equation 3) is 1.68 eV for the NiCo 2 O 4 , [52] which is much smaller than that for Co 3 O 4 under the same step (1.98 eV) reported by Dong et al.. [53] This synergy effect around Co atom sites in NiCo 2 O 4 drastically reduces the energy barrier in the step of OH * to O * (Equation 3), which is reported to be the rate-determining step (RDS) for both NiCo 2 O 4 and Co 3 O 4 in OER. [52,53]Therefore, the reduction of energy barrier in the RDS explains the better OER performance of NiCo 2 O 4 than Co 3 O 4 , thus, elucidates the mechanism behind the superior OER performance of NiCo 2 O 4 with the lowest overpotential and R ct among these three catalysts (Figures 7b and 8b).

Conclusion
In summary, the NiO, Co

Experimental Section
Growth of Materials: The NiO, Co 3 O 4, and NiCo 2 O 4 layers were deposited on FTO glasses under identical conditions using a facile chemical spray pyrolysis method.The distance between the nozzle and the substrate was kept at 30 cm while the substrate temperature is kept at 450 °C for deposition of all NiO, Co 3 O 4, and NiCo 2 O 4 layers (Figure 1).To obtain homogeneous thin films, the spray time, and the pressure of the carrier gas were kept at 20 min and 1.5 bar, respectively for all prepared samples.The as-prepared layers were further heat treated at 500 °C in air for 4 h.Nickel chloride and cobalt chloride precursor solutions were used for preparing NiO, Co 3 O 4, and NiCo 2 O 4 thin films.0.1 m of Nickel chloride, and cobalt chloride solutions were dissolved in distilled water, in order to obtain clear solutions suitable for the spraying process for NiO and Co 3 O 4 thin films; respectively.These solutions with ratio 1:2 were mixed with continuous stirring, to deposit NiCo 2 O 4 thin film.The used FTO glass substrates were cleaned with acetone, distilled water, and alcohol; successively.
Characterizations of Materials: The structure of the films was analyzed by a Panalytical X′Pert diffractometer operating CuK1 radiation ( = 0.154 nm) at 45 kV and 40 mA within the values of 2 between 10 and 90°.Raman analysis was performed by a confocal Raman microscope (Witec Alpha 300 RA) using laser excitation 532 nm.The surface morphology of the obtained samples was studied by scanning electron microscope (SEM) model Leo 1550 Gemini.XPS measurements were performed using a Kratos Ultra DLD photoelectron spectrometer (Kratos Analytical, Manchester, and UK) equipped with a monochromatized Al K (1486.6 eV) excitation source.The base pressure during spectra acquisition was lower than 1.1 × 10 −9 Torr (1.5 × 10 −7 Pa), achieved by a combination of turbomolecular and ion pumps.The anode power was set to 150 W. Ni 2p, Co 2p, and O 1s core level spectra were recorded from the 0.3 × 0.7 mm 2 area at the normal emission angle.The analyzer pass energy was set to 20 eV, resulting in the full width at half maximum of 0.55 eV for the Ag 3d 5/2 peak from reference Ag sample.Fermi edge of the FTO substrate was used for charge referencing to avoid unreliable C 1s method. [54]Samples were analyzed in the as received state, i.e., with no Ar + etching.Flood gun was used during measurements.
Electrochemical Measurements: All electrochemical measurements were carried out using an electrochemical workstation (Princeton Applied Research, VersaSTAT 3).A typical three-electrode cell was adopted with 1.0 m KOH as the electrolyte, where a graphite rod as the counter electrode and Hg/HgO as the reference electrode.The RuO 2 electrode was prepared as the following method: 5 mg of RuO 2 powder was mixed with 80 μl of nafion (5 wt.%), 250 μl isopropanol and 750 μl deionized water to form homogenous solution under ultra-sonication for 30 min.Afterward, 50 μl of the solution was cast-deposited on the FTO glass with size of 1 × 1 cm 2 .All the potentials were referenced to reversible hydrogen electrode (RHE) according to the following equation: E (RHE) = E (Hg∕HgO) +0.098 + 0.0592 × pH (6)   LSV results were recorded at a scan rate of 5 mV s −1 , and the chronoamperometry measurement was done at a high constant current density of 100 mA cm −2 to evaluate the stability.Nyquist measurements were performed at 1.7 V versus RHE in the frequency range from 0.1 Hz to 100 kHz with an amplitude potential of 10 mV.All the results were recorded without IR-correction.The overpotentials () were calculated using the equation ( = E(RHE) − 1.23 V).

Figure 1 .
Figure 1.The illustration of a facile spray pyrolysis followed by post calcination synthesis method.

2 O 4 ,
Co 3 O 4 , and NiO were measured by the linear sweep voltammetry (LSV) at a scan rate of 5 mV s −1 .NiCo 2 O 4 shows the lowest onset potential (356 mV at 3 mA cm −2 ) to initiate the OER process compared with Co 3 O 4 (452 mV at 3 mA cm −2 ) and NiO (423 mV at 3 mA cm −2 ), indicating a lower energy required for NiCo 2 O 4 to drive OER process than Co 3 O 4 and NiO (Figure 7a).To drive a benchmark current density of 10 mA cm −2 , NiCo 2 O 4 requires an overpotential of 425 mV, which is lower than that for Co 3 O 4 (551 mV) and NiO (608 mV).As the overpotential goes as high as 990 mV for all three catalysts, NiCo 2 O 4 drives the highest current density of 100 mA cm −2 , whereas Co 3 O 4 drives just ≈60 mA cm −2 and NiO only gives about a quarter of current density (24 mA cm −2 ) compared with that of NiCo 2 O 4 .Figure 7b compares the overpotentials of NiCo 2 O 4 , Co 3 O 4, and NiO at ten and 50 mA cm −2 , respectively.It clearly shows that NiCo 2 O 4 gives the smallest overpotential among these three oxides at both ten and 50 mA cm −2 .The reaction kinetics of three catalysts was assessed by the Tafel slope, which gives us more information about how fast the OER process progresses at the vicinity of onset potential, namely, at the beginning of the OER process.NiCo 2 O 4 gives the lowest Tafel slope of 84.1 mV dec −1 (Figure7c), compared

Figure 5 .
Figure 5. High resolution XPS spectra of a) Ni 2p of NiO, b) Co 2p of Co 3 O 4 , c) Ni 2p, and d) Co 2p of NiCo 2 O 4 .
2 O 4 is tremendously enhanced compared to the individual Co 3 O 4 and NiO catalysts.The larger ECSA of NiCo 2 O 4 suggests more active OER sites involved in water splitting, which is favorable for enhancing the OER performance.Table S1 (Supporting Information) lists the OER performance and C dl values of some recently reported NiCo 2 O 4 nanostructures prepared by different methods.We find that the OER activity of our NiCo 2 O 4 (C dl = 4.90 mF cm −2 ) is very close to that of the reported NiCo 2 O 4 nanostructures with similar C dl values (for instance, C dl = 4.5 mF cm −2 ).
3 O 4, and NiCo 2 O 4 nanostructure layers with the <111> dominant orientation and identical surface morphologies were deposited on FTO under same conditions by a simple spray pyrolysis method and their OER activities were comparatively investigated.Among these three oxides, NiCo 2 O 4 exhibits the smallest Tafel slope (84.1 mV dec −1 ) and the lowest overpotential (420 mV) to drive a benchmark current density of 10 mA cm −2 , which are comparable to those of the benchmark commercial RuO 2 electrocatalyst.In contrast, Co 3 O 4 and NiO show relatively high overpotentials (551 mV for Co 3 O 4 and 608 mV for NiO) and Tafel slopes (106.4 mV dec −1 for Co 3 O 4 and 128.6 mV dec −1 for NiO).These experimental results agree with DFT calculations that show the energy barrier of the ratedetermining step from OH * to O * is significantly reduced in NiCo 2 O 4 .This work paves the way for developing low-cost and efficient electrocatalysts toward large-scale production of renewable hydrogen via water electrolysis.