Excess Activity Tuned by Distorted Tetrahedron in CoMoO4 for Oxygen Evolution

Oxygen vacancies enable modulating surface reconstruction of transition metal oxides containing metal–oxygen polyhedrons into metallic oxyhydroxide for oxygen evolution reaction (OER), while revealing reconstructing mechanism is stuck by the requirement to precisely control exact sites of these vacancies. Herein, oxygen vacancies are localized only within MoO4 tetrahedrons rather than CoO6 octahedrons in CoMoO4 catalyst, guaranteeing coherent reconstruction of CoO6 octahedrons into pure CoOOH with tunable activities for OER. Meanwhile, distorted tetrahedron accelerates the dissolution of Mo atoms into alkaline electrolyte, triggering spontaneous transition of partial CoMoO4 into Co(OH)2. CoO6 octahedrons in both CoMoO4 and Co(OH)2 can transform pure CoOOH completely at lower potential, resulting in excess intrinsic activity whose summit is identified by overpotential at 10 mA cm−2 with 22.9% reduction and Tafel slope with 65.3% reduction. Well‐defined manipulation over the distorted polyhedrons offers one versatile knob to precisely modulate electronic structure of oxide catalysts with outstanding OER performance.


Introduction
Transition metal oxides (TMOs) are considered as low-cost alternative candidates for oxygen evolution reaction (OER), expectantly strengthening one arm to boost electrocatalytic water splitting with industrial production. [1,2]Outstanding performance fundamentally depends on the number of active sites [3,4] and intrinsic activity [5] of individual site onto the catalysts.Refined nanoparticles and hierarchical structures potentially promote the sluggish dynamics, [6] in favor of sufficient exposure of intermediate species onto active surface [7] and prompt ion transport [8] within hollow channels.Intrinsic activity strongly relies on electronic structure of upper layer that is created via surface reconstruction during potential supply. [9]In previous investigations, substitution of foreign atoms, [10,11] vacancies, [12,13] strain, [14,15] interfaces, [16,17] enables regulating electronic structure to optimize bonding energy of some key species.Among them, the vacancies can weigh up atomic coordination [18,19] and charge distribution onto catalyst surface that provides some opportunities to enhance the OER further.
Oxygen vacancies (V O ), the deficit of oxygen atoms within the lattices, usually distort metal-oxygen (M-O) polyhedrons (tetrahedrons and octahedrons) in some perovskite [20] and spinel [21] structures.Metallic oxyhydroxides (NiOOH, CoOOH, FeOOH) [22][23][24] or their derivatives are finally created from these polyhedrons via surface reconstruction.In principle, lattice distortion seriously relies on the concentration of V O , [25,26] so that different reconstructing behaviors occurring at various types of polyhedrons make distinct contributions to ultimate OER performance. [27]So far, convincing correlation between intrinsic performance and metallic oxyhydroxide is not built up because both the types of initially distorted polyhedrons containing Ni, Co, and Fe elements and coherent reconstruction from the same kind of polyhedron cannot be guaranteed throughout the experiments.Especially, M-O tetrahedrons and octahedrons [28,29] extensively exist in most TMOs, and manipulating exact locations of V O cannot be achieved in most synthesizing methods and subsequent treatments.Thus, some precise characterization techniques are necessary to confirm the location and content of oxygen vacancy in TMOs and further investigate the mechanism of oxygen vacancy enhancing catalytic activity, such as in situ electrochemical Raman spectra. [30]erein, CoMoO 4 nanosheet precursors synthesized by hydrothermal method are randomly anchored on nickel foam (Figure S1, Supporting Information).The CoMoO 4 precursor was annealed in Ar atmosphere Oxygen vacancies enable modulating surface reconstruction of transition metal oxides containing metal-oxygen polyhedrons into metallic oxyhydroxide for oxygen evolution reaction (OER), while revealing reconstructing mechanism is stuck by the requirement to precisely control exact sites of these vacancies.pure CoOOH completely at lower potential, resulting in excess intrinsic activity whose summit is identified by overpotential at 10 mA cm −2 with 22.9% reduction and Tafel slope with 65.3% reduction.Well-defined manipulation over the distorted polyhedrons offers one versatile knob to precisely modulate electronic structure of oxide catalysts with outstanding OER performance.
to form nearly V O -free CoMoO 4 (AP-CoMoO 4 ). [31]Subsequently, the AP-CoMoO 4 was annealed in a reducing mixture atmosphere of H 2 and Ar for different durations of n (n = 20, 30, 60, 90, and 120) minutes, producing n-CoMoO 4 with distinct levels of V O .Most V O are localized only within MoO 4 tetrahedrons that are identified by X-ray photoelectron spectra (XPS) and Raman spectroscopy, so that coherent surface reconstruction only occurs at CoO 6 octahedrons in favor of tuning intrinsic activity of pure CoOOH (Figure 1).

Results and Discussion
The AP-CoMoO 4 completely seals entire Ni skeletons, and typical surface contains many standing nanosheets distributed randomly with a thickness of ~80 nm (Figure S2, Supporting Information).Sketchy geometries of these nanosheets are well inherited after reduced annealing.Appearances of some crevices imply existence of internal stress in these oxides annealed in the reducing environment.Meanwhile, it is visible that the widths of crevice are distinctive and vary with the annealing time in Figure S2, Supporting Information.Crevice hardly appears when the annealing time is <60 min.However, apparent crevices appear when the annealing time is over 90 min.Interestingly, the annealed oxides are still monoclinic rutile-type β-CoMoO 4 (JCPDS card no.00-021-0868), representing identical phase to AP-CoMoO 4 based on X-ray diffraction (Figure S3, Supporting Information). [32]Transmission electron microscopic image shows that the CoMoO 4 is highly crystallized, and lattice fringes are clear.The d-spacing is 0.245 and 0.381 nm, being assigned to (4 0 0) and (0 2 1) planes of CoMoO 4 , respectively.Meanwhile, the V O -rich CoMoO 4 with d-spacing of 0.199 and 0.186 nm can be indexed to (5 1 1) and (1 3 3) planes of CoMoO 4 (Figure 2a-c and Figure S4, Supporting Information).Unlike AP-CoMoO 4 , the crystallinity of some regions decreases in n-CoMoO 4 (yellow dotted circle part) and results in partial structural distortion during the introduction of oxygen vacancies.Notably, the low crystallizing region is more obvious following the increase of reducing time, which is consistent with the shrinking crevices in SEM images.
In order to characterize the details of atomic bonding, XPS and Raman spectroscopy are used to confirm ionic states and bonding vibrations.XPS spectra of AP-CoMoO 4 show the signals of Co 2+ and Mo 6+ cations.Two peaks at 797.5 and 781.6 eV are Co 2p core-level spectra, corresponding to Co 2+ 2p 1/2 and Co 2+ 2p 3/2 , respectively (Figure S5a, Supporting Information). [33,34]Compared with pristine counterpart, all n-CoMoO 4 exhibit the identical Co 2p spectra, indicating that bonding state of Co 2+ cations does not change after reduced annealing.Distinct from Co 2p, obvious changes are detected in Mo 3d spectra (Figure 2d).Besides two peaks at 235.3 and 232.2 eV corresponding to 3d 3/2 and 3d 5/2 of Mo 6+ cations, [35] new peaks at 233.4 and 230.3 eV appear in n-CoMoO 4 , which are assigned to 3d 5/2 and 3d 3/2 of Mo 4+ cations, [36] respectively.Additionally, the proportion of Mo 4+ cations increases with annealing time, implying partial transition of Mo 6+ into Mo 4+ cations (Figure 2e and Table S1, Supporting Information).O1s XPS spectra contain three peaks at 530.5, 531.4,and 532.8 eV, confirming the coexistences of M-O, [37] O vacancies, [38,39] and absorbed water [40] (Figure S5b, Supporting Information).Furthermore, the oxygen vacancy peaks from XPS spectra and their deconvolutions of AP-CoMoO 4 and n-CoMoO 4 corresponding to O1s peaks were conducted with normalization.Similar conclusion can be derived, indicating that 30-CoMoO 4 and 60-CoMoO 4 possess the highest level of oxygen vacancy (Figure S6 and Table S2, Supporting Information).In addition, the existence of oxygen vacancies is investigated by electron paramagnetic resonance (EPR).Obvious EPR signal peak at g = 2.003 indicates the electrons have been captured by the defects, confirming the existence of free electrons trapped in oxygen vacancies (Figure 2f).However, the 90-CoMoO 4 and 120-CoMoO 4 show obvious paramagnetic effect due to excessive structural distortion of MoO 4 .The signal of oxygen vacancy is immerged by the background (Figure S7, Supporting Information), in which case the detection of oxygen vacancy cannot be reliably detected by EPR because of paramagnetic feature of CoMoO 4 .
Raman spectra of all CoMoO 4 also show significantly different characteristics after reducing thermal annealing (Figure 2g and Figure S8, Supporting Information).Raman peak of AP-CoMoO 4 at 345 cm −1 corresponds to the vibration of MoO 4 tetrahedrons, [41,42] while dual peaks (C and A in Figure 2g) at 800 and 927 cm −1 are assigned to Co-O-Mo stretching modes. [41]In n-CoMoO 4 , the individual peak at 345 cm −1 splits into double peaks (E and D in Figure 2g) at 334 and 365 cm −1 .This splitting stems from the distortion of MoO 4 tetrahedrons due to the existence of V O .This notion is also supported by newborn band (B in Figure 2g) at ~870 cm −1 , corresponding to distinct vibrations of Co-O-Mo bonding mediated by V O .After quantitative analysis signify relative proportion of integrated intensity from peak D located at 365 cm −1 (denoted as P D ) rises monotonously in n-CoMoO 4 (Figure 2h and Table S3, Supporting Information).Meanwhile, the P E decreases with a subsequent saturation annealing time, indicating accordance with the concomitances of Mo 6+ and Mo 4+ cations in XPS analysis.In addition, highest P B is obtained in 30-CoMoO 4 and 60-CoMoO 4 , and P A exhibits a minimum in the same case, which is consistent with the O1s XPS spectra and EPR results.Based on Raman spectroscopy and XPS characterization, the level of oxygen vacancy significantly depends on the annealing time.Its proportion is almost saturated gradually with the elongated duration of up to 60 min, implying that lattice distortion reaches the summit after 60 min annealing.These results indicate that short time reduction in the Ar/H 2 atmosphere can introduce oxygen vacancies into AP-CoMoO 4 effectively, while excessive V O will lead to severe distortion of MoO 4 tetrahedrons.
Oxygen evolution reaction performances of all CoMoO 4 are evaluated in 1.0 M KOH electrolyte based on three-electrode configuration (Figure 3).Cyclic voltammetry (CV) measurements reveal dual quasireversible redox peaks (Figure S9a, Supporting Information) which are attributed to the oxidation of Ni and the Co 2+/3+ transition.Intrinsic activity is quantitatively derived from linear sweep voltammetry polarization (Figure 3a) and Tafel plots (Figure 3b).AP-CoMoO 4 indicates mediocre catalytic activity evidenced by an overpotential of 328 mV at 10 mA cm −2 and Tafel slope of 167 mV dec −1 .While n-CoMoO 4 holds better activity, improving magnitude sensitively depends on exact levels of V O .Both overpotential and Tafel slope reduce with reduction treatment time toward 30 min, followed by subsequent rising toward 120 min.30-CoMoO 4 exhibits the best performance in terms of an overpotential of 253 mV and Tafel slope of 58 mV dec −1 , corresponding to 22.9% and 65.3% declines, respectively (Figure 3c).S9c, Supporting Information).It implies that the number of catalytically active sites is increased through introducing oxygen vacancy into CoMoO 4 .The C dl is derived from the CV curves at different scan rates of 20-140 mV s −1 by the linear fitting of the slopes (Figure S10, Supporting Information).Although high levels of V O are involved, multi-step chronopotentiometry tests of AP-CoMoO 4 (gray), 30-CoMoO 4 (red), and 120-CoMoO 4 (blue) without iR-corrected indicate negligible decay at various potentials after 100 h operation (Figure S11, Supporting Information), suggesting excellent stability of the n-CoMoO 4 in OER.Besides, the cyclic voltammetry cycling test is also carried out for the durability testing.As shown in Figure S12, Supporting Information, even after 5000 CV cycles the polarization curve of 30-CoMoO 4 indicated excellent stability and durability in alkaline solution.This result provides more evidence for 30-CoMoO 4 to be a promising catalyst in the field of water oxidation.
To investigate the underlying mechanism for enhanced activity and stability, systematic structural characterization of AP-CoMoO 4 , 30-CoMoO 4 , and 120-CoMoO 4 is carried out after OER test.SEM images confirm that the nanosheet arrays structure are still well preserved, but the gap between nanosheets became narrower, which illustrates the structure of catalysts has undergone extraordinary exquisite evolution during OER process (Figure S13, Supporting Information).In addition, the Co 2p spectrum of all samples after OER is fitted with two different peaks, where the peaks at 780.3 and 795.3 eV are ascribed to Co 3+ , revealing that Co 2+ in CoMoO 4 has been partly oxidized into oxyhydroxides phase [31,43] (Figure S14a, Supporting Information).While the Mo 3d and O1s spectra after the OER tests were approximately the same as before (Figure S14b,c, Supporting Information).This phenomenon reveals that the Co oxyhydroxide maybe is the real active sites.Notably, the intensity of Mo 3d spectra is extremely weak.The signal of Mo 3d and Mo 2p almost disappears from the full XPS spectrum after going through the reconstruction evolution, suggesting that the reconstructing to form Co oxyhydroxide may be accompanied by the dissolution of Mo atoms (Figure S15, Supporting Information).
In situ Raman spectroscopy enables elucidating the detailed composition of catalysts subjected to surface reconstruction (Figure 4a,b and Figure S16, Supporting Information).As shown in Figure 4a, AP-CoMoO 4 exhibits three typical bands (A, C, and E) of CoMoO 4 , and their intensities decrease gradually but do not disappear completely with increasing potential from 1.0 to 1.25 V versus RHE.With further positive potential to 1.3 V, two bands that are respectively assigned to E g at 456 cm −1 and A 1g and 535 cm −1 of CoOOH [44,45] S6, Supporting Information).V O plays a role for both leaching out of Mo atoms and the interaction between Co 2+ and OH − ions in favor of prompt formation of CoOOH at lower potential (Figure 4d). [46]Inductively coupled plasma optical emission spectrometer was carried out on part catalysts before and after OER test.Before OER test, the Co/Mo molar ratios of AP-CoMoO 4 , 30-CoMoO 4 , and 120-CoMoO 4 are 1.07, 0.89, and 0.90, respectively.However, the ratio separately increases to 5.78, 8.21, and 8.43 of each sample after performance testing (Figure S17a, Supporting Information).Meanwhile, the Mo/K molar ratio in electrolyte after OER test is much higher than that of pure KOH solution, which shows Mo atoms dissolved into electrolyte (Figure S17b, Supporting Information).The defective CoMoO 4 after OER test holds higher Co/Mo and Mo/K molar ratio, which indicates the presence of V O accelerates the dissolution of Mo atoms and further promotes the reconstruction process to form the active phase CoOOH.Besides, only CoOOH character peak appears, while no other signal was detected in Raman spectrum of three samples after OER.This result further confirms that the surface CoMoO 4 is completely converted into CoOOH (Figure S18, Supporting Information).Secondly, additional peak at 682 cm −1 that is assigned to Co (II) -O vibration mode in Co(OH) 2 [47]   appears when all n-CoMoO 4 are soaked into 1 M KOH solution without potential supply.It manifests that CoMoO 4 with V O tends to generate new species of Co(OH) 2 spontaneously (Figure 4e and Table S7, Supporting Information).The longer reduced annealing time is, the more Co(OH) 2 is born (Figure 4e).Only Co(OH) 2 phase is detected with the absences of Co-O-Mo stretching and MoO 4 tetrahedron in 120-CoMoO 4 .It demonstrates that the level of V O directly affects the distorted magnitude of MoO 4 tetrahedron and affects Mo dissolution in different catalysts.X-ray diffraction pattern confirms this spontaneous process, and only Co(OH) 2 is observed with the complete absence of CoMoO 4 (Figure 4f).

Conclusions
In summary, oxygen vacancies localized within MoO 4 tetrahedrons of CoMoO 4 catalysts can be manipulated by thermal annealing.The distorted MoO 4 tetrahedrons facilitate the formation of the Co(OH) 2 via spontaneous dissolution of Mo atoms without potential supply.Therefore, pure and active phase CoOOH is finally created from CoO 6 octahedrons in both CoMoO 4 and Co(OH) 2 at lower potential in comparison with AP-CoMoO 4 .Low overpotential of 253 mV at 10 mA cm −2 with excess 22.9% reduction is observed in 30-CoMoO 4 , manifesting a fascinating route to improve intrinsic activity of TMO catalysts for OER by controlling oxygen vacancies precisely.in deionized (DI) water with magnetic stirring for 30 min.Then, the obtained uniform purple solution was transformed into Teflon-lined stainless steel autoclave liner.A piece of nickel foam (4 cm × 3 cm) was sonicated in 3 M HCl for 10 min to remove the possible surface oxide layer and then washed with deionized (DI) water and ethanol for several times.The prepared nickel foam was subsequently immersed into solution in hydrothermal reaction vessel, which were sealed and maintained at 180 °C for 6 h and then cooled to room temperature.After cleaning with distilled water and ethanol to remove the accumulation of surface particles, the as-prepared samples were then dried in air and annealed at 400 °C in an Ar atmosphere for 2 h, the pristine CoMoO 4 was obtained, donated as AP-CoMoO 4 .

Experimental Procedures
Preparation of of Defective CoMoO 4 Electrode: The as-prepared AP-CoMoO 4 sample treated in H 2 /Ar (H2: 20 sccm, Ar: 100 sccm) atmosphere at 400 °C for several minutes, donated as x-CoMoO 4 (where x is the reduced treated time, x = 10, 20, 30, 60, 90, 120 min).All the anneal treatment heating rate was completed in CVD system with a heating rate of 5 °C min −1 .In order to investigate the structural evolution of CoMoO 4 during hydrogenation by XRD test, similar processes were also applied to the powders formed in the hydrothermal process.
Material Characterization: The morphology and structure of products were characterized with a field-emission scanning electron microscope (FESEM, FEI, Nova NanoSEM 450).X-ray photoelectron spectroscopy (XPS) was performed on a Kratos-Axis spectrometer with monochromatic Al Kα (1486.71eV) X-ray radiation (15 kV and 10 mA) and a hemispherical electron energy analyzer.X-ray diffraction (XRD) patterns were carried out on a diffractometer (Lab XRD-6100, Shimadzu, Japan) equipped with a Cu Kα radiation (λ = 0.15406 nm) at a voltage of 40 kV and a current of 30 mA.
Electrochemical Measurements: All of the electrochemical measurements were carried out on a CHI 760E electrochemical analyzer (CH Instruments, Inc., Shanghai, China) at room temperature in a standard three-electrode system with a Hg/HgO electrode as reference electrode and a graphite rod as counter electrode in 1.0 M KOH solution.The resultant catalysts were used as working electrodes.Linear sweep voltammetry (LSV) and cyclic voltammograms (CV) measurements were performed at a scan rate of 5 mV s −1 .All potentials were iR-corrected to 85% with the built-in program.The EIS was obtained by AC impedance spectroscopy within the frequency range from 0.01 Hz to 10 kHz in 1.0 M KOH.The amplitude of the alternating voltage was 5 mV around a potential of 1.5 V versus RHE for OER.The long time durability test was conducted by using controlled-potential electrolysis method.All measured potentials were calibrated to reversible hydrogen electrode (RHE) scale.
Raman Spectrum Measurements: All Raman experiments were conducted with a Raman spectrometer (Ramos S120, Ostec-ArtTool, LLC, Moscow) with a 50× long working distance (LWD) visible objective.The light source excitation wavelength is 532 nm at a range of 200-1000 cm −1 .To ensure the reliability and reproducibility of the acquired spectra, the Raman shift was calibrated using a silicon standard sample (520.7 cm −1 ).The individual Raman spectrum was obtained with an acquisition time of 100 s.The in situ Raman experiments were performed using a three-electrode system and custom-made in situ electrolytic cell at room temperature (25 °C).For the in situ measurements, CoMoO 4 nanosheets loading on the nickel foam were used as the working electrode.The counter electrode was a graphite rod, and the reference electrode was Hg/ HgO.A laser power of around 15 mW was used in all in situ Raman measurements.The Raman spectrum was obtained in situ under each applied potential using a potentiostat (CHI 760E CH Instruments, Inc.), and al spectrum were collected after reaching steady current density.In all in situ Raman experiments, 1 M KOH electrolyte was used.
Furthermore, the electrochemical impedance spectroscopy (EIS) analysis and double-layer capacitance (C dl ) measurement were performed on AP-CoMoO 4 and n-CoMoO 4 to elucidate the oxygen vacancy effect.The Nyquist plots reveal that 30-CoMoO 4 exhibits a lower charge transfer resistance (1.41 Ω) than that of AP-CoMoO 4 (3.79Ω) and other defective ones, implying a better charge-transfer capacity for 30-CoMoO 4 (Figure S9b and are visible.Dual bands become stronger with potential, in accompany with degenerated characteristic of CoMoO 4 could be contributed to surface reconstruction and Mo leaching at high overpotential.Finally, Co-O-Mo and Mo-O

Figure 2 .
Figure 2. TEM images of a) AP-CoMoO 4 , b) 30-CoMoO 4 , c) 120-CoMoO 4 .d) XPS spectra of Mo 6+ and Mo 4+ cations.e) Relative proportion of the integrated area corresponding to Mo 6+ and Mo 4+ cations in (a).f) Electron paramagnetic resonance (EPR) of AP-CoMoO 4 and n-CoMoO 4 .g) Raman spectra for AP-CoMoO 4 and n-CoMoO 4 subjected to reduced annealing.h) Relative proportion of main Raman bands denoted by A, B, D, E varies with reduced annealing time.The integrating area of all bands is assigned as 1 for each spectrum, and relative proportion of each band is normalized by overall integration.
On one hand, high level of V O promotes Mo dissolution in alkaline electrolyte, resulting in the decrease or even disappearances of Co-O-Mo stretching and MoO 4 tetrahedron.On the other hand, more active Co 2+ sites are exposed into alkaline electrolyte, promoting the formation of intermediate Co(OH) 2 .In our investigation, V O localized in MoO 4 tetrahedrons of n-CoMoO 4 can trigger spontaneous dissolution of Mo atoms significantly in alkaline electrolyte.The CoO 6 octahedrons existing in newborn Co(OH) 2 phase also transform into CoOOH completely, just like the octahedrons in CoMoO 4 .Although the same phase of CoOOH is produced from Co(OH) 2 and CoMoO 4 , different supports may modulate the detailed characters of reconstructing layer.Relative ratio between Co(OH) 2 and CoMoO4 makes a dominant contribution to the refined structure of final CoOOH phase (Figure 4e), outputting different activities for OER.Middle level of V O can enhance the transition of CoMoO 4 into CoOOH at lower potential, resulting in effective overpotential shift as shown in Figure 4c.In situ Raman spectroscopy also indicates the complete reconstruction of pure Co(OH) 2 into CoOOH at high oxidation potential (Figure S19, Supporting Information).The overpotential and Tafel slope of Co(OH) 2 for OER locates within those of AP-CoMoO 4 and 120-CoMoO 4 (Figure S20, Supporting Information).Therefore, CoOOH in different states attained from AP-CoMoO 4 and n-CoMoO 4 actually demonstrates the difference of intrinsic activities.

Figure 3 .
Figure 3. Electrochemical characterizations of OER performance of different CoMoO4 catalysts in 1 M KOH electrolyte.a) LSV polarization at a scan rate of 5 mV s −1 ; b) Tafel slopes derived from LSV polarization curves.c) The overpotential at current density of 10, 100 mA cm −2 and the corresponding Tafel slopes for various catalysts.
Herein, oxygen vacancies are localized only within MoO 4 tetrahedrons rather than CoO 6 octahedrons in CoMoO 4 catalyst, guaranteeing coherent reconstruction of CoO 6 octahedrons into pure CoOOH with tunable activities for OER.Meanwhile, distorted tetrahedron accelerates the dissolution of Mo atoms into alkaline electrolyte, triggering spontaneous transition of partial CoMoO 4 into Co(OH) 2 .CoO 6 octahedrons in both CoMoO 4 and Co(OH) 2 can transform