In‐Situ‐Grown Cu Dendrites Plasmonically Enhance Electrocatalytic Hydrogen Evolution on Facet‐Engineered Cu2O

Herein, facet‐engineered Cu2O nanostructures are synthesized by wet chemical methods for electrocatalytic HER, and it is found that the octahedral Cu2O nanostructures with exposed crystal planes of (111) (O‐Cu2O) has the best hydrogen evolution performance. Operando Raman spectroscopy and ex‐situ characterization techniques showed that Cu2O is reduced during HER, in which Cu dendrites are grown on the surface of the Cu2O nanostructures, resulting in the better HER performance of O‐Cu2O after HER (O‐Cu2O‐A) compared with that of the as‐prepared O‐Cu2O. Under illumination, the onset potential of O‐Cu2O‐A is ca. 52 mV positive than that of O‐Cu2O, which is induced by the plasmon‐activated electrochemical system consisting of Cu2O and the in‐situ generated Cu dendrites. Incident photon‐to‐current efficiency (IPCE) measurements and the simulated UV–Vis spectrum demonstrate the hot electron injection (HEI) from Cu dendrites to Cu2O. Ab initio nonadiabatic molecular dynamics (NAMD) simulations revealed the transfer of photogenerated electrons (27 fs) from Cu dendrites to Cu2O nanostructures is faster than electron relaxation (170 fs), enhancing its surface plasmons activity, and the HEI of Cu dendrites increases the charge density of Cu2O. These make the energy level of the catalyst be closer to that of H+/H2, evidenced by the plasmon‐enhanced HER electrocatalytic activity.


Introduction
[3] Cu 2 O also has sufficient redox capacity, abundant surface active sites, and easily tunable morphology, porosity, and particle size. [4,5][14][15][16] However, n-type Cu 2 O as an electrocatalyst has attracted less attention.To date the hydrogen evolution reaction (HER) performance of n-type Cu 2 Obased electrocatalysts is poor, and the active sites and catalytic mechanism of these catalysts are still unclear due to the limitation of characterization techniques. [17,18]he atomic arrangement and chemical status of the exposed surface have a great influence on the performance of electrocatalysts, [19][20][21][22][23][24] including atomic steps, corners, edges, coordination states, dangling bonds, and surface energy.[27] For example, Wang and co-workers proposed an in situ electrochemical reduction strategy to synthesize Pt nanosheets with high-index {311} planes and low-index {200} and {111} planes on carbon nanotubes, and it is found that the {311} planes of Pt nanosheets had a higher electrocatalytic hydrogen evolution activity compared with the other planes. [28]Kuo et al. prepared {100} facet-dominated silver nanocubes and {111} facetdominated silver nano octahedra via polyol reactions.The results showed that plasmonic Ag NOs with exposed {111} facets exhibited better catalytic performance for HER than Ag NCs with exposed {100} facets. [29]However, most catalysts consist of crystals with mixed facets, showing an averaged catalytic activity.To investigate the influence of crystal facets, it is necessary to control the anisotropic structure of catalysts with specific crystal planes.
Recently, it has been found that light can be used as the driving force of electrochemical reactions.For example, Ag nanocubes harvested visible light and directly drove catalytic oxidation reactions of CO and NH 3 . [30]Au nanoparticles served as both the light absorber and electrocatalytic active site to promote the occurrence of efficient electrochemical reaction of glucose electrooxidation by utilizing the energy of hot carriers. [31][42] Therefore, introducing the plasmonic effect into Cu 2 O is an effective method to promote the overall electrocatalytic process.
Herein  2− to the (111) crystal plane is lower than that of Cu 2+ , indicating that SDS is more conducive to occupying the binding site of the (111) crystal plane, [43,44] thus adsorbing on the (111) plane to stabilize the surface; on the contrary, the binding energy of Cu 2+ to the (100) crystal plane is lower compared with that of SO 4 2− , which indicates that Cu 2+ is more favorable to adsorb on the (100) crystal plane and is reduced by NH 2 OH•HCl, resulting to the growth of the (100) crystal plane.ii) The surface energies of Cu 2 O (111) and Cu 2 O (100) facets are 0.023 and 0.057 eV, respec-tively.The lower surface energy of the Cu 2 O (111) plane indicates that it is more stable, which is also responsible for the increase in the proportion of (111) planes during crystal growth. [45]The powder X-ray diffraction (XRD) pattern of Cu 2 O nanostructures is shown in Figure 3a.Five distinct peaks at 2 = 29.5°,36.4°,42.3°, 61.3°, and 73.5°correspond to (110), ( 111), ( 200), (220),   3b) were assigned to the Cu─O stretching mode, [46,47] and no impurity peaks were observed, confirming the high purity of the product.O-Cu 2 O nanostructures with different exposed facets, which can be divided into two peaks, corresponding to the Cu─O bond and the adsorbed oxygen, respectively. [48]For the Cu 2p signals (Figure 3i) of C-Cu 2 O and O-Cu 2 O, they are divided into two peaks, attributing to Cu + of Cu 2p 1/2 and Cu 2p 3/2 , respectively. [49,50]

Evaluation of the Electrochemical HER Performance
The electrocatalytic HER performance of the three Cu 2 O catalysts was investigated by using a standard three-electrode system in 1.0 m KOH.The FTO glass coated with the Cu 2 O samples was served as the working electrode, and a graphite rod and an Hg/HgO/OH − electrode were employed as the counter electrode and reference electrode, respectively.The linear sweep voltammetry (LSV) curves of the three Cu  S4, Supporting Information).For the convenience of comparison with the literature, Cu 2 O samples were prepared as slurries and dropcoated on the glassy carbon electrode (GCE) for testing and the results are shown in Figure S3 (Supporting Information).By comparing the hydrogen evolution activities of Cu 2 O samples on FTO and CCE, it is found that the overpotential of the catalyst on GCE is lower than that on FTO, which may be attributed to the better conductivity of GCE compared to FTO.The HER performance of bare FTO, Nafion-modified FTO, and bare GCE were performed to reveal the influence of substrates, and the polarization curves are shown in Figure S4 (Supporting Information).It is found that their overpotentials of bare FTO, Nafion-modified FTO, and bare GCE were 380, 382, and 347 mV, respectively, indicating that the addition of Nafion had little effect on HER activity, and bare GCE has better HER activity than bare FTO, which may be due to the better conductivity of GCE electrode.This is consistent with the HER activity of the electrodes loaded with Cu 2 O samples.
The intrinsic catalytic activity of the Cu 2 O samples was evaluated by the electrochemically active surface area (ECSA) and mass activity.ECSA is proportional to the double-layer capacitance (C dl ), which is calculated from cyclic voltammetry curves (Figure 4d; Figure S5, Supporting Information). [51]The C dl of O-Cu 2 O is 48.8 mF cm −2 , which is higher than that of C-Cu It is worth mentioning that the determined intrinsic activity values are inevitably underestimated as active sites are endowed only by surface metal species, rather than the entire transition metal component. [52]LSV curves of the Cu 2 O samples in Figure 4a 4f, inset), which may be because the (111) crystal plane has a higher atomic arrangement density than the (100) crystal plane, resulting in its relatively lower surface energy. [53]Chronoamperometry measurements were used to investigate the stability of O-Cu 2 O under an applied voltage of −0.16 V vs reversible hydrogen electrode (RHE) (Figure 4f), and there was no obvious degradation after the 50 h test, indicating good durability.The HER activity of the obtained O-Cu 2 O compares favorably to most previously reported Cu-based HER electrocatalysts (Figure 4g, Table S5, Supporting Information).
As shown in Figure 4h, the free energy differences of the first

Active Site Identification
To elucidate the catalytic active sites, the microstructure, chemical composition, and valence evolution of the O-Cu 2 O catalyst after HER testing (O-Cu 2 O-A) in alkaline electrolyte were investigated by post-mortem characterization.SEM images (Figure 5a) demonstrate that dendrites were grown on the surface of the three Cu 2 O samples after long-term chronoamperometry HER tests, and no obvious fragmentation was observed.TEM images (Figure 5b,c; Figure S5a,c 200) and (220) crystal planes of Cu (JCPDS card No. 04-0836), indicating that Cu 2 O reduction occurred during the test, which is consistent with the HRTEM results and further proves that the observed dendrites are Cu crystals.From the FT-IR spectra of O-Cu 2 O before and after HER (Figure 6b), it was found that the sample after the HER test showed a Cu─Cu stretching mode at 1137 cm −1 , [54,55] suggesting the formation of Cu.As shown in Figure 6c, the specific surface areas of O-Cu 2 O and O-Cu 2 O-A were 58.05 and 110.24 m 2 g −1 , respectively, and the pore volumes of the samples before and after cycling were 0.032 and 0.107 cm 3 g −1 , respectively, suggesting a significant increase in the pore structure of the sample after HER.The pore diameters were mainly distributed ≈0.2 nm, indicating the existence of a large number of micropores (Figure S9, Supporting Information). [56]perando Raman spectroscopy was performed to explore the chemical composition evolution of O-Cu 2 O in real-time HER (Figure 6d).The peaks at 111 and 218 cm −1 are attributed to the A 1g vibration modes of the Cu─O and Cu─Cu bonds, respectively. [57]When the voltage changed from 0 to −0.2 V vs RHE, the peak intensity of the Cu─O bond essentially remained unchanged, while that of the Cu─Cu bond increased significantly, and the E g vibration mode of the Cu─Cu bond appeared, [58,59]   appears in O-Cu 2 O-A, which may be due to the lack of lattice oxygen during the reduction of Cu 2 O, and this is consistent with the HAADF-STEM result. [60]The peak of Cu 0 appears in the Cu 2p orbital spectrum of O-Cu 2 O-A (Figure 6f), [61] indicating the formation of Cu.This is consistent with the results of HRTEM, XRD, and FT-IR, which proves that Cu  To further explore the plasmonic enhancement effect of Cu dendrites, IPCE was conducted to explore the absorption of light with specific wavelengths by catalytic materials (Figure 7d).The bump in the region of 500-700 nm is overlapped with the range of plasmonic response of Cu nanostructures (Figure S15, Supporting Information), and this range is higher than the bandgap of O-Cu 2 O-A (2.50 eV, 496 nm) (Figure S9, Supporting Informa-tion), directly demonstrating the hot-electron transfer from Cu dendrites to Cu 2 O.The shift of the Cu + peaks to lower binding energy in the Cu 2p XPS spectrum of O-Cu 2 O-A also indicated hot-electron transfer (Figure 6f).This hot-electron injection (HEI) process produces more efficient electron-hole separation states, which greatly contributes to the HER activity.In addition, the slight enhancement at 300-500 nm may be due to the light scattering caused by the increased light absorption by Cu dendrites.
The physical and chemical mechanisms behind the photogenerated electrons in the wavelength region greater than 500 nm were investigated to elucidate the reasons for the enhanced electrocatalytic HER performance.O-Cu The electron relaxation is accompanied by electron-phonon coupling, and this is due to time-dependent fluctuations of the vibrating atomic motions, which is obtained from the Fourier transform (FT) spectra of the plasma state and the electron-acceptor state of Cu 76 -(Cu 2 O) 24 compounds (Figure 7h).
In order to further investigate the photogeneration carrier dynamics process and determine the timescales of energy relaxation (Process II) and electron transfer (Process III), the time dependence of electron transfer from Cu to Cu 2 O was obtained using ab initio nonadiabatic molecular dynamics (NAMD) simulations, [62][63][64][65][66] as shown in Figure 7i,j.The results demonstrate the photoexcited electron transfer from Cu to Cu 2 O takes place with 27 fs fitting the population of the initial state.However, the electronic energy relaxation occurs within 170 fs, which is longer than the electron transfer timescales.Due to the lowfrequency phonons of heavy Cu atoms, the energy relaxation is slow.Especially, the energy level fluctuations caused by atomic thermal vibration of the system produce electron-phonon coupling at 300 K. Finally, the photogenerated hot-electron transfer from Cu particle to Cu 2 O nanomaterials enhances the conductivity of Cu 2 O semiconductors, implying that visible light improves its electrocatalytic performance.7k).In addition, the irradiation power is proportional to the overpotential (Figure S17, Supporting Information), indicating that the direct HEI from Cu dendrites to Cu 2 O was accompanied by a decrease in activation energy (Figure 7l), which further demonstrates that the enhancement of HER performance under illumination is mainly due to the electron transfer effect (Figure 8).The in-situ-grown Cu dendrites during HER introduce plasmonic effects into Cu 2 O with optimally exposed crystal facets, tune the carrier density of the catalyst to match the energy levels of target reactions, improve interfacial charge transfer, and thereby further promote the electrocatalytic HER efficiency.This method can also be extended to other inexpensive plasmonic metals, such as Al and Mg, to expand the spectral range of light absorption and reduce costs, enabling more applications in electrocatalytic and photocatalytic devices.

Conclusion
photon dissipation,[37][38][39] all resulting in accelerated electrochemical processes.[40][41][42]Therefore, introducing the plasmonic effect into Cu 2 O is an effective method to promote the overall electrocatalytic process.Herein, cubic, truncated octahedral and octahedral Cu 2 O particles (C-Cu 2 O, T-Cu 2 O, O-Cu 2 O) with different crystal effects were synthesized by a wet chemical method, in which NH 2 OH•HCl acts as both reducing agent and morphology control agent.The structures and chemical compositions of the Cu 2 O nanostructures were compared through detailed characterization.The hydrogen evolution performance of the three samples was tested in an alkaline environment.The adsorption-free energy of hydrogen intermediates on different crystal facets was calculated and RLS was determined by density functional theory (DFT).Post-mortem investigations including ex situ characterization and operando Raman spectroscopy studied the changes in structure, composition, and bandgap of the samples during HER, and identified the active site of the reaction.DFT calculations investigated the charge transfer at the Cu 2 O/Cu interface.The hydrogen evolution performance of O-Cu 2 O before and after HER was compared under light and dark conditions, respectively, and it was proved that the superior HER performance of O-Cu 2 O-A was due to the plasmon enhancement effect of Cu dendrites and the formation of Cu 2 O/Cu heterostructures.

Cu 2 O
nanostructures with different exposed crystal facets were synthesized by the wet chemical method, and the synthetic route is shown in Figure 1.CuCl 2 , SDS, NaOH, and NH 2 OH•HCl were sequentially added to deionized water, and Cu 2 O particles were obtained by chemical reduction.By gradually increasing the content of reducing agent NH 2 OH•HCl, three Cu 2 O samples with different morphologies were prepared, which were cubic, truncated octahedral, and octahedral Cu 2 O (C-Cu 2 O, T-Cu 2 O, O-Cu 2 O).Field-emission scanning electron microscopy (FESEM) images show that the Cu 2 O nanostructures were uniform and with narrow size distribution, with average sizes of 300, 370, and 400 nm, respectively (Figure 2a-c).Transmission electron microscopy (TEM) images (Figure 2a-c, inset) show that the Cu 2 O nanostructures are all solid.High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images and the corresponding atomic models highlighting the (111) and (100) facets of Cu 2 O are shown in Figure 2d,e, respectively, where each contrast site is contributed by only one atom.Each O atom is surrounded by six Cu atoms in the plan view of (111) facets, while each O atom is surrounded by four Cu atoms in the plan view of (100) facets.With increasing NH 2 OH•HCl concentration, the exposed (100) crystal plane disappeared and the (111) crystal plane appeared.This trend is attributed to two parts: i) The binding energy of SO 4 2− (the functional group of SDS) and Cu 2+ to (111) facets are 0.434 and 0.503 eV (Figure 2f,g), respectively; the binding energy of SO 4 2− and Cu 2+ to (100) facets are −0.392 and −0.994 eV, respectively (Figure 2h,i).The binding energy of SO 4

Figure 1 .
Figure 1.Schematic diagram of the synthesis of C-Cu 2 O, T-Cu 2 O, O-Cu 2 O and C-Cu 2 O-A, T-Cu 2 O-A, O-Cu 2 O-A.

Figure 3 .
Figure 3.Chemical composition analysis.a) XRD patterns and b) FT-IR spectra of C-Cu 2 O, T-Cu 2 O and O-Cu 2 O. c) Cu K-edge XANES spectra of C-Cu 2 O, T-Cu 2 O, O-Cu 2 O, Cu 2 O, and Cu foil (inset: an enlarged view of the dashed box).d-f) Wavelet transform-EXAFS spectra of K-edge of C-Cu 2 O (d), T-Cu 2 O (e), and O-Cu 2 O (f). g) R-space FT-EXAFS spectra of Cu K-edge of C-Cu 2 O, T-Cu 2 O and O-Cu 2 O. h,i) XPS high-resolution spectra at O 1s edge (h) and Cu 2p edge (i) of C-Cu 2 O and O-Cu 2 O.
Scanning transmission electron microscopy attached with energy dispersive spectroscopy (STEM-EDS) results demonstrate that the Cu 2 O samples were composed of the elements Cu and O without impurities (Figure S1, Supporting Information).X-ray absorption spectroscopy (XAS) was performed to further study the different exposed facets.First, by comparing the near-edge regions of C-Cu 2 O, T-Cu 2 O, and O-Cu 2 O with Cu foil and Cu 2 O (Figure 3c), it confirms that Cu 2 O of high purity was synthesized.Wavelet analysis of the XAS data (Figure 3d-f) indicates the similarity between the three facets, with peaks near (6 Å −1 , 1.5 Å) attributed to first shell Cu-O scattering, and peaks near (6 Å −1 , 2.8 Å) attributed to second shell scattering.Their similarity in peak intensities and positions indicates their resemblance in structures.However, for O-Cu 2 O, a reduction in intensity along with a broader peak width is observed.Fitting on the R-space of the X-ray absorption fine structure (XAFS) data is then performed to quantify the changes in scattering paths (Figure 3g), with the fitting parameters and k-space fitting available in Figure S2 and Tables S1-S3 (Supporting Information).For C-Cu 2 O, a coordination number of 1.93(63) at 1.83(5) Å from Cu is observed, which falls slightly shorter than the expected value of 2 for Cu 2 O.As the fraction of exposed (111) facets increases to T-Cu 2 O and O-Cu 2 O, an increase of coordination number of 2.07(68) and 2.23(63) at 1.86(4) and 1.87(4) Å from Cu was obtained.An exposed (100) facet would result in undercoordinated Cu near the boundaries of the nanostructures, whereas the exposed (111) facet would see fully coordinated Cu, aligning well with our fitted experimental observations.Figure 3h shows the O 2p Xray photoelectron spectroscopy (XPS) signals of the C-Cu 2 O and

Figure 4 .
Figure 4. Electrochemical performance for HER.a) Polarization curves and b) Tafel plots of C-Cu 2 O, T-Cu 2 O, and O-Cu 2 O in alkaline electrolyte.The FTO glass coated with the Cu 2 O samples, a graphite rod, and an Hg/HgO/OH − electrode were employed as the working electrode, counter electrode, and reference electrode, respectively in all the measurements.c) Comparison of Tafel slopes and the overpotentials needed to deliver cathodic current densities of 10, 50, and 100 mA cm −2 for C-Cu 2 O, T-Cu 2 O, and O-Cu 2 O. d) Capacitive current measured at 0.25 V vs RHE for C-Cu 2 O, T-Cu 2 O and O-Cu 2 O as a function of scan rate.e) Mass activity of C-Cu 2 O, T-Cu 2 O and O-Cu 2 O. f) Chronoamperometry measurement of O-Cu 2 O at an applied voltage of −0.16 V vs RHE (inset: electrolyte contact angle measurements of C-Cu 2 O, T-Cu 2 O, and O-Cu 2 O).g) Comparison of HER activity with some recently reported Cu-based catalysts in alkaline electrolyte.h) Gibbs free energy diagrams of C-Cu 2 O, T-Cu 2 O and O-Cu 2 O for HER in alkaline electrolyte.i,j) Bader charge analysis of (111) facets (i) and (100) facets (j) of Cu 2 O, and the Cu active sites are highlighted by the orange circles.
2 O samples are shown in Figure 4a.It is found that O-Cu 2 O with (111) exposed facets exhibits the best activity, and the overpotentials of O-Cu 2 O were 158, 250, and 307 mV for current densities of 10, 50, and 100 mA cm −2 , respectively, while the overpotentials for the same current densities for C-Cu 2 O and T-Cu 2 O were 231, 184; 382, 299; and 472, 364 mV, respectively.O-Cu 2 O also exhibited a relatively small Tafel slope (75.6 mV dec −1 ) compared with C-Cu 2 O (122.1 mV dec −1 ) and T-Cu 2 O (82.8 mV dec −1 ), which further supports its faster HER reaction rate (Figure 4b,c, Table 2 O (19.8 mF cm −2 ) and T-Cu 2 O (34.7 mF cm −2 ).The mass activity of O-Cu 2 O at an overpotential of 120 mV is 30.7 A g −1 , which is better than that of C-Cu 2 O (5.8 A g −1 ) and T-Cu 2 O (17.1 A g −1 ) (Figure 4e).The TOF of C-Cu 2 O, T-Cu 2 O, and O-Cu 2 O are 0.232, 0.505 and 0.851 s −1 per site at an overpotential of 200 mV, respectively, and the TOF of O-Cu 2 O is ≈4 times that of C-Cu 2 O.
have been normalized to the ECSA (Figure S6, Supporting Information), O-Cu 2 O still exhibited the highest current density under the same voltage condition, indicating its best activity among the Cu 2 O samples.The exchange current density was obtained from the Tafel plots after ECSA normalization (Figure S7, Supporting Information), and the values of C-Cu 2 O, T-Cu 2 O, and O-Cu 2 O are 0.574, 0.440 and 0.296 mA cm −2 , respectively.The higher value of O-Cu 2 O reveals its higher HER kinetics.Electrolyte contact angle tests show that O-Cu 2 O is more hydrophilic than C-Cu 2 O and T-Cu 2 O (Figure

H 2 O
molecular adsorption, the Volmer step, the second H 2 O molecular adsorption and the Heyrovsky step in the HER process on the Cu 2 O (111) facets are −0.965,−0.027, −0.412 and −0.566 eV, respectively.However, the free energy differences of the same four steps on the Cu 2 O (100) planes are −0.996,−1.177, −0.365, and 0.567 eV, respectively.The last step on the Cu 2 O (100) planes has a positive free energy difference, indicating that the Heyrovsky step (RLS) on Cu 2 O (100) facets has a much higher energy barrier than that on Cu 2 O (111) facets, leading to its relatively poor HER activity.The partial charges of active Cu sites on Cu 2 O (111) and Cu 2 O (100) facets were calculated using Bader charge analysis (Figure 4i,j).The partial charges of the Cu active sites on the (111) and (100) facets of Cu 2 O are +0.398 and +0.235 e, respectively.The less positive partial charge of the Cu site on Cu 2 O (100) indicates more electrons on it, thus forming a stronger covalent bond with H, resulting in a stable H * inter-mediate.However, the overly stable H * intermediate makes the desorption of H difficult, which gives rise to a high energy barrier for the Heyrovsky step on Cu 2 O (100).
, Supporting Information) indicate the length of the dendrites is ≈5-10 nm, while the solid shape of the Cu 2 O samples after HER remains unchanged.The highresolution TEM (HRTEM) image (Figure 5c, inset) reveals the heterostructure at the interface of the solid structure and the dendrites.The lattice spacings of 0.24 and 0.21 Å correspond to the (111) planes of Cu 2 O and Cu, respectively, indicating the generated dendrites are Cu, formed by the reduction of Cu 2 O. Selectedarea electron diffraction (SAED) pattern (Figure 5d) shows the single-crystal structure of O-Cu 2 O-A, and the STEM image and elemental mapping (Figure 5e; Figure S8b,d, Supporting Information) show that the chemical composition of the Cu 2 O samples remains unchanged after HER, still consisting of O and Cu.HAADF-STEM image of O-Cu 2 O-A and the corresponding atomic model are shown in Figure 5f, and oxygen vacancies appear after the HER test, which is marked with yellow hexagons.XRD was conducted to analyze the crystal structure of O-Cu 2 O after HER.As shown in Figure 6a, compared with the XRD pattern of O-Cu 2 O, the additional peaks of O-Cu 2 O-A at 43.3 o , 50.4 o and 74.0°correspond to the (111), ( further proving the formation of Cu.The fitted XPS spectra of O 1s orbital for O-Cu 2 O and O-Cu 2 O-A are compared in Figure 6e, an additional oxygen vacancy (O v ) peak

Figure 5 .
Figure 5. Crystal structure characterization after HER.a) FESEM and b) TEM images of O-Cu 2 O-A (scale bars: 200 nm).c) HRTEM image and d) SAED pattern of O-Cu 2 O-A (scale bar: 5 nm).e) STEM and elemental mapping images of O-Cu 2 O-A (scale bar: 200 nm).f) HAADF-STEM image of O-Cu 2 O-A and the corresponding simulated atomic model (scale bar: 0.5 nm), and the yellow hexagons represent oxygen vacancies.
2 O is reduced to Cu during the HER process.The peak of the UV-vis spectrum of O-Cu 2 O-A redshifted compared with that of O-Cu 2 O, and the bandgap decreased from 2.57 eV (O-Cu 2 O) to 2.50 eV (O-Cu 2 O-A), indicating that the formation of Cu dendrites enhanced the conductivity of the catalyst (Figure S10, Supporting Information).XAS analysis was performed on O-Cu 2 O-A for the characterization of O-Cu 2 O after HER.The near-edge region of O-Cu 2 O-A is significantly different from that of Cu 2 O (Figure 6g), suggesting the structure has deviated from Cu 2 O upon catalyst cycling.Comparing the wavelet transform of O-Cu 2 O and O-Cu 2 O-A (Figures 3f and 6h), a shift of the peak from (5.5 Å −1 , 2.78 Å) to (5.5 Å −1 , 2.53 Å) was detected, along with an enhanced intensity of the peak centered from (5.2 Å −1 , 1.34 Å) to (6.05 Å −1 , 1.47 Å), which can be explained by the shortened Cu─Cu distance from Cu 2 O (3.03 Å) compared to Cu metal (2.56 Å).Further quantification with XAFS fitting of O-Cu 2 O-A (Figure 6i, Table S6, Supporting Information) shows a much higher Cu─O first shell co-ordination number (3.13) compared to O-Cu 2 O (2.23) (Figure 3g, Figure S2, Supporting Information), which may be attributed to the formation of Cu(OH) 2 phases during catalysis cycling.Intriguingly, a good fitting (R-factor 2.38%) can be obtained with two Cu─Cu scattering paths from Cu foil and Cu 2 O between 2-3.5 Å for O-Cu 2 O-A (Figures S11, S12, Supporting Information).Cu─Cu scattering from Cu foil is seen at 2.67(4) Å with a coordination number of 3.0(4), while another Cu─Cu scattering from Cu 2 O is seen at 3.19(13) Å with a coordination number of 9.9(45).The observed Cu─Cu scattering from both Cu foil and Cu 2 O echoes with the observed Cu dendrites on the surface of Cu 2 O aforementioned characterized by other experimental techniques.Figure 6j shows the density of states of plain Cu 2 O and Cu 2 O (111) with one layer of Cu (111) deposition, as a model of Cu 2 O (111) before and after HER tests.The main difference between the two models lies in the Fermi surface.With Cu (111) deposition after cycles, more unoccupied states appear near the Fermi surface, leading to higher electron conductivity.This is beneficial for the electron-involved steps (Volmer and Heyrovsky) and contributes to the higher activity of Cu 2 O (111) after HER tests.The charge-transfer diagram shows a remarkable charge transfer between the Cu (111) layer and the Cu 2 O (111) facet, in which

Figure 6 .
Figure 6.Chemical composition analysis after HER.a) XRD patterns, b) FT-IR spectra, and c) N 2 adsorption/desorption isotherms of O-Cu 2 O and O-Cu 2 O-A.d) Operando Raman curves and the corresponding contour plot obtained from the voltage changing from 0 to −0.2 V vs RHE.e,f) XPS high-resolution spectra at O 1s edge (e) and Cu 2p edge (f) of O-Cu 2 O and O-Cu 2 O-A.g) Cu K-edge XANES spectra of O-Cu 2 O, O-Cu 2 O-A, Cu 2 O and Cu foil.h) Wavelet transform-EXAFS spectra of O-Cu 2 O-A.i) R-space FT-EXAFS spectra of Cu K-edge of O-Cu 2 O-A.j) DOS of Cu 2 O and Cu 2 O/Cu (inset: the corresponding optimized structural models).k) Charge-transfer diagram (yellow and blue isosurfaces indicate the gain and loss of electrons) of O-Cu 2 O-A at the junction of Cu 2 O and Cu.
2 O-A was simulated as Cu 76 -(Cu 2 O) 24 , and the photogenerated electron transfer and energy relaxation at the interface of Cu 76 -(Cu 2 O) 24 is discussed in detail.The projected density of states of Cu 76 -(Cu 2 O) 24 compounds with separated Cu and Cu 2 O are presented in Figure 7e.It is observed that the systems show a metallic property due to the use of a large copper particle.The PDOS peaks corresponding to the blue and pink dotted lines in Figure 7f represent the photoexcited plasmon (donor) state and the electron acceptor state, respectively, whose charge densities are shown in Figure 7f.The photogenerated electron is mainly localized on the copper particle, which exhibits continuous waves.The electron acceptor is completely localized on the Cu 2 O system. Figure 7g shows the mechanism for the enhanced HER activity of O-Cu 2 O-A under light irradiation, the charge transfer of the Cu 76 -(Cu 2 O) 24 compound involves five different steps: Process I: Surface plasmons proliferate near-field plasmonic waves at the interface between Cu 76 and (Cu 2 O) 24 , forming electron-hole pairs.The generated high-energy hot electrons are further away from the electronic equilibrium state on Cu 76 and may have three possible transport channels: Process II: recombination with holes in Cu 76 ; Process III: injection into the conduction band (CB) of (Cu 2 O) 24 semiconductor; Process IV: direct electrochemical reduction of water on Cu 76 .Process V: The injected hot electrons in the CB of (Cu 2 O) 24 may be relaxed and transferred to the electron-deficient Cu 76 , returning to its ground state.The contribution of Process IV for the enhancement of photoelectrochemical current is negligible, as Cu is not a good HER catalyst due to the low adsorbed hydrogen coverage on the surface.There is a low Schottky barrier between Cu and Cu 2 O, and Cu dendrites are in situ grown on Cu 2 O with tight bonding during the electrochemical HER process, and the electron injection from Cu to Cu 2 O is faster and more stable due to the strong interaction between Cu and Cu 2 O (Process III).The electron relaxation and charge recombination (Process II) are competitive to the electron injection (Process III).

Figure 7 .
Figure 7. Photoelectrochemical performance for HER.a) Polarization curves and b) Tafel plots of O-Cu 2 O and O-Cu 2 O-A in the absence or presence of light irradiation in alkaline electrolyte.The FTO glass coated with the Cu 2 O samples, a graphite rod, and a Hg/HgO/OH − electrode were employed as the working electrode, counter electrode, and reference electrode, respectively in all the measurements.c) Comparison of Tafel slopes and the overpotentials needed to deliver cathodic current densities of 10, 50, and 100 mA cm −2 for O-Cu 2 O and O-Cu 2 O-A in the absence or presence of light irradiation.d) IPCE plots measured at 0 V (vs RHE) of O-Cu 2 O and O-Cu 2 O-A in alkaline electrolyte.e) PDOS of Cu 76 -(Cu 2 O) 24 and Cu 76 .f) Photoexcited plasmon-like state and the electron acceptor state corresponding to the peaks of the blue and pink dotted lines in (e).g) Potential hot-electron transfer pathways during SPR.h) Fourier transform spectra of the plasmon-like state (blue) and the electron-acceptor state (red) of Cu 76 -(Cu 2 O) 24 .i) Charge transfer from Cu particle to Cu 2 O semiconductors, where black and red lines indicate donor and acceptor states, respectively, and the additional lines represent the states between the donor and the acceptor.j) The energy relaxation between the electron donor and acceptor.The colored strips denote the electron distribution on different energy states, and the dashed line indicates the averaged hole energy.k) Energy level diagram of Cu 2 O-Cu illuminating hot-electron injection from Cu to Cu 2 O. l) Schematic of the activation energy change of O-Cu 2 O-A by light irradiation.

Figure 8 .
Figure 8. Schematic diagram of hot electron of in-situ-grown Cu dendrites activates the electrocatalytic hydrogen evolution over Cu 2 O.

Cu 2 O
nanostructures with different exposed crystal planes were prepared by wet chemical method and applied to electrocatalytic HER.The effect of exposed crystal planes on their hydrogen evolution performance was explored.It was found that O-Cu 2 O with (111) exposed crystal planes had the smallest overpotential and Tafel slope.DFT calculations indicated that this is due to the lower free energy barrier of the Heyrovsky step (RLS) on Cu 2 O (111).Post-mortem characterization and Operando Raman spectroscopy were conducted to analyze the changes in crystal structure, chemical composition, and valence state of O-Cu 2 O during HER, and thus determined the active sites.It is found that Cu dendrites were reduced on the surface of Cu 2 O nanostructures to form a Cu 2 O/Cu composite structure.DFT calculations showed that the charge transfer of the Cu 2 O/Cu interface promoted its surface electron conductivity and tuned the adsorption strength of HER intermediates, thus greatly enhancing the electrocatalytic HER activity of O-Cu 2 O-A.Under light irradiation, ab initio NAMD simulations showed that the charge transfer from Cu dendrites to Cu 2 O nanostructures is faster than charge relaxation, improving its surface plasmon activity.O-Cu 2 O-A con-stitutes a plasmonic-activated electrochemical system: in-situformed Cu dendrites act as light absorbers to excite electronhole pairs through LSPR, and Cu 2 O nanostructures serve as active sites and hot-electron acceptors to facilitate the HER process.