RESEARCH ARTICLE www.advancedscience.com Selective Ethylene Glycol Oxidation to Formate on Nickel Selenide with Simultaneous Evolution of Hydrogen Junshan Li, Luming Li, Xingyu Ma, Xu Han, Congcong Xing, Xueqiang Qi, Ren He, Jordi Arbiol, Huiyan Pan, Jun Zhao, Jie Deng, Yu Zhang,* Yaoyue Yang,* and Andreu Cabot* 1. Introduction There is an urgent need for cost-effective strategies to produce hydrogen from renewable net-zero carbon sources using renewable energies. In this context, Molecular hydrogen is both a fundamen- tal precursor in the chemical industry and the electrochemical hydrogen evolution reaction can be boosted by replacing an extremely appealing, potentially carbon- the oxygen evolution reaction with the oxidation of small organic molecules, free, energy carrier.[1,2] However, H2 is cur- such as ethylene glycol (EG). EG is a particularly interesting organic liquid rently being produced mainly from fossil with two hydroxyl groups that can be transformed into a variety of C1 and C2 fuels, releasing large amounts of carbon chemicals, depending on the catalyst and reaction conditions. Here, a catalyst dioxide. The production of H2 from water electrolysis is not cost effective in front of is demonstrated for the selective EG oxidation reaction (EGOR) to formate on the steam reforming or gasification of fos- nickel selenide. The catalyst nanoparticle (NP) morphology and sil fuels due to its moderate energy effi- crystallographic phase are tuned to maximize its performance. The optimized ciency and the use of expensive noble metal NiS electrocatalyst requires just 1.395 V to drive a current density of 50 catalysts.[3] Thus alternative strategies for mA cm−2 in 1 m potassium hydroxide (KOH) and 1 m EG. A combination of in the production of hydrogen from net-zero situ electrochemical infrared absorption spectroscopy (IRAS) to monitor the carbon sources and using renewable ener- gies need to be developed. electrocatalytic process and ex situ analysis of the electrolyte composition In the electrochemical water splitting, shows the main EGOR product is formate, with a Faradaic efficiency above most energy is consumed at the anode 80%. Additionally, C2 chemicals such as glycolate and oxalate are detected due to the sluggish oxygen evolution re- and quantified as minor products. Density functional theory (DFT) action (OER).[3–6] To overcome this limi- calculations of the reaction process show the glycol-to-oxalate pathway to be tation, OER can be replaced with the ox- idation of small molecules having lower favored via the glycolate formation, where the C–C bond is broken and further oxidation potentials, which implies higher electro-oxidized to formate. energy efficiencies and the potential use of lower-cost catalysts based on abundant J. Li, L. Li X. Han, J. Arbiol Institute for Advanced Study Catalan Institute of Nanoscience and Nanotechnology (ICN2) Chengdu University CSIC and BIST Chengdu 610106, China Campus UAB, Bellaterra, Barcelona, Catalonia 08193, Spain L. Li, J. Deng C. Xing, X. Qi, R. He, Y. Zhang, A. Cabot College of Food and Biological Engineering Catalonia Institute for Energy Research—IREC Chengdu University Sant Adrià de Besòs, Barcelona, Catalonia 08930, Spain Chengdu 610106, China E-mail: yvz5897@psu.edu; acabot@irec.cat X. Ma, Y. Yang J. Arbiol, A. Cabot School of Chemistry and Environment ICREA Southwest Minzu University Pg. Lluís Companys 23, Barcelona, Catalonia 08910, Spain Chengdu 610041, China H. Pan E-mail: yaoyueyoung@swun.edu.cn School of Biological and Chemical Engineering Nanyang Institute of Science and Technology Nanyang 473004, China J. Zhao The ORCID identification number(s) for the author(s) of this article Hebei Key Laboratory of Photoelectric Control on Surface and Interface can be found under https://doi.org/10.1002/advs.202300841 College of Science Hebei University of Science and Technology © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH. Shijiazhuang 050018, China This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in Y. Zhang any medium, provided the original work is properly cited. Department of Materials Science and Engineering Pennsylvania State University DOI: 10.1002/advs.202300841 University Park, PA 16802, USA Adv. Sci. 2023, 10, 2300841 2300841 (1 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com elements.[7–9] Thus, potentially, hydrogen can be cost-effectively Ni metalloids (NiTx , T = P, S, and Se) on the coordination envi- evolved from aqueous solutions containing biomass-derived ronments of Ni sites to optimize the methanol electro-oxidation products or organic waste such as ethanol, glycerol, ethylene gly- performance.[42] Nevertheless, to the best of our knowledge, the col (EG), or sugars with a net-zero CO2 footprint.[10–12] An addi- electro-oxidation mechanism of glycols on nickel surfaces is yet tional advantage of the production of hydrogen from the reform- to be explored. ing of organic waste or biomass-derived organics is the poten- Herein, we prepared a series of nanostructured nickel dise- tial co-generation of valuable organic chemicals, which can im- lenides with tuned crystal phase and morphology from a simple prove the process economics and eventually reduce the use of single-precursor ink. The different nanostructures are tested to- fossil resources to produce these chemicals.[13–15] As an example ward the electrocatalytic EGOR. In situ infrared absorption spec- in this last direction, in previous studies we demonstrated that troscopy (IRAS), ex situ analyses of the electrolyte and materials, formate, a key precursor in the chemical and pharmaceutical in- and density functional theory (DFT) calculations are used to an- dustries, can be electrochemically produced from methanol in alyze the reaction product and gain insight into the process, the alkaline media using nickel catalysts, which can be economically active sites, and the influence of the different components. advantageous over current methods based on high temperature and pressure conditions.[16–18] Compared with methanol, EG is a nontoxic liquid with high boiling point and low volatility, which 2. Results and Discussion offers safe storage and high volumetric energy density. In addi- tion, EG can be renewably obtained through efficient biomass- 2.1. Nanomaterials’ Synthesis and Characterization to-ethylene glycol processes.[19] Besides, the EG market price is slightly below that of formic acid, thus the production of hydro- NiSe2 nanostructures were produced in two steps (Figure 1a). In gen from EG reforming to formate involves no loss of product the first step, a single Ni–Se precursor was prepared by dissolving value at the anode side.[20] Ni acetylacetonate and selenium (Se) powder in a solution con- Recently, the electrocatalytic oxidation of glycols has taining oleylamine (OAm, C18 H37 N) and 1-dodecanethiol (DDT, been extensively studied, especially on noble-metal-based CH3 (CH2 )11 SH). In the second step, this ink was thermally de- electrocatalysts.[21–27] Si et al. reported the electro-oxidation composed in the presence of a mixture of solvents/surfactants of glycolic acid with a very low cell potential on PdAg sup- (OAm/1-Octadecene (ODE)/oleic acid (OAc)) that determined ported on nickel foam.[25] In another recent study, Pd–PdSe the final particle shape and crystallographic phase (see the “Ex- heterostructures showed a high EG oxidation reaction (EGOR) perimental Section” for details). Figure 1b,c shows representa- selectivity, above 44%, to C1 products, with superior C–C bond tive transmission electron microscopy (TEM) images and the X- cleavage.[24] However, the use of noble metals hampers the ray diffraction (XRD) patterns of the materials produced using industrial application of these strategies. different OAm/OAc/ODE combinations. In the presence of just Nickel is an excellent alternative catalyst to noble metals for OAm, quasispherical nanoparticles (NPs) with a narrow size dis- the electro-oxidation of small molecules in alkaline media. In a tribution centered at ≈15 nm and pure cubic NiSe2 phase (JCPDS basic solution, the surface of nickel is oxidized to nickel oxyhy- 00-041-1495) were obtained. When adding ODE into the OAm droxide, NiOOH, which is the active surface for organic electro- solution (VOAm :VODE = 1:3), ≈200 nm branched nanostructures oxidation.[28–31] On a typical Ni-based catalyst, small organics with 50–100 nm arms and mixed cubic and orthorhombic phases containing only one hydroxyl can be electrochemically trans- were produced. Besides, when OAc was added into the solution formed into their corresponding acidic commodities, such as (VOAm :VODE :VOAc = 5:3:12), elongated bundles with a diameter methanol to formate and ethanol to acetate with high Faradaic of about 100 nm, a length of ≈0.5–0.8 μm and pure orthorhom- efficiency.[17,32,33] EG, functionalized with two hydroxyls, displays bic phase (JCPDS 00-018-0886) were obtained. As shown in Fig- different routes for the transformation to C1 and C2 products. ure S1 (Supporting Information), the Ni:Se atomic ratio for these Specifically, the EGOR primarily involves two pathways, either three structures was determined to be 1:2, in good agreement a 10-electron process to direct/complete oxidation that results with stoichiometric NiSe2 . in carbonate (in an alkaline environment), or a multielectron As displayed in Figure 2a and Figure S2a,b (Supporting In- process resulting in indirect/incomplete oxidation into C1 or formation), high-resolution TEM (HRTEM) and electron diffrac- C2 chemicals, including glyoxal, glycolaldehyde, glycolate, glyox- tion analyses confirmed the presence of the orthorhombic (space alate, oxalate, and formate.[34] group = Pnnm, a = 4.89 Å, b = 5.96 Å, and c = 3.67 Å) and Nickel compounds, such as nitrides, phosphides, and chalco- cubic lattices (space group PA3-, a = b = c = 5.9730 Å) in the genides, have frequently demonstrated promoted electrocatalytic NiSe2 bundles and spherical NPs, respectively. The crystal model performance over bare nickel.[35] We and others have demon- of NiSe2 and an atomic supercell model are presented in Fig- strated the surface of such compounds to be reconstructed to ure S2b,c (Supporting Information) for the spherical NPs and more active hydroxide and mainly oxohydroxide, which are recog- bundles, respectively. Electron energy loss spectroscopy (EELS) nized as the true catalytically active surfaces for OER in alkaline chemical composition maps show a homogeneous distribution media.[35–37] During this surface reconstruction, the surface an- of Ni and Se within each particle (Figure 2c; Figure S2d, Support- ions are simultaneously oxidized to nitrates (NO3 − ), phosphates ing Information). Figure 3a shows the high-angle annular dark- (PO4 3− ), sulfates (SO4 2− ), and selenates (SeO4 2− ).[38–40] Such an- field (HAADF)–scanning TEM (STEM) and HRTEM images of ionic species can modulate the surface electronic structure im- the branched structure. As can be seen in Figure 3b, the detail of proving the material catalytic performance.[41] As an example, Li the orange squared region and its corresponding power spectrum et al. studied the influence of different oxyanions obtained from reveals the presence of the NiSe2 cubic lattice in this sample. Adv. Sci. 2023, 10, 2300841 2300841 (2 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 1. a) Scheme of the two-step process used to produce nanostructured nickel selenide. b) TEM images. c) XRD patterns of the obtained materials. The reference patterns of the cubic (JCPDS 00-041-1495) and orthorhombic (JCPDS 00-018-0886) NiSe2 phases are also plotted. Adv. Sci. 2023, 10, 2300841 2300841 (3 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 2. a) HRTEM image, detail of the orange squared region and its corresponding power spectrum of a NiSe2 bundle. From the crystalline domain, the lattice fringe distances were measured to be 0.295, 0.267, and 0.299 nm, at 53.99° and 108.96°, which can be interpreted as the cubic NiSe2 phase, visualized along its [010] zone axis. b) 1*1*1 unit crystal model of NiSe2 and 3D atomic supercell model of the NiSe2 bundles. Red and green particles represent Ni and Se, respectively. c) EELS chemical composition maps obtained from the red squared area of the STEM image. Within our system, OAm acts as a strong ligand, preventing the branched sample displays three obvious peaks, including a the growth of the NiSe2 crystals and resulting in NPs with the shake-up satellite (marked “Sat.”) at 860.9 eV. The peak located thermodynamically stable cubic crystal structure and a quasi- at 853.5 eV is assigned to Ni2+ within the selenide lattice.[45] Be- spherical morphology associated with the high symmetry of the sides, the peak at 855.7 eV is assigned to Ni2+ within a more elec- cubic crystallographic phase. When adding ODE as a solvent, i.e., tronegative chemical environment probably generated by the sur- reducing the OAm concentration, the action of the OAm as a cap- face oxidation of the particles.[46] The high-resolution Se 3d XPS ping agent is reduced; a faster growth rate is obtained; and the for- spectrum displays two doublets. The first doublet is located at mation of some kinetically controlled orthorhombic NiSe2 takes 54.4 eV (Se 3d5/2 ) and it is assigned to Se2− within the metal dis- place. Because the orthorhombic phase is highly anisotropic, the elenide environment.[47] The second doublet at 58.9 eV (Se 3d3/2 ) growth rate along the different crystallographic directions is not is assigned to an oxidized Se environment (SeOx ) generated by homogeneous. In particular, we observe the growth rate along the exposure of the particles to the air atmosphere.[47] As can be the [001] direction to be much faster, which results in the growth seen in Figure S4 (Supporting Information), the XPS characteris- of some elongated structures from the surface of the crystals.[43] tics of the spherical NPs and bundles did not change with respect When adding significant amounts of OAc, there is a competi- to the branched NiSe2 morphology. tion between the OAm and OAc binding to the surface and also with the formation of OAm–OAc complexes.[44] With an excess amount of OAc, the growth rate is further accelerated and rela- 2.2. Electrocatalytic Performance tively large bundles with a pure orthorhombic crystal structure are formed. While the detailed growth mechanism is not com- NiSe2 nanostructures were supported on glassy carbon (GC) to pletely understood, it is worth noting that the phase-controlled characterize their electrocatalytic performance toward EGOR (see synthesis reported herein bears some similarity to the synthesis details in the “Experimental Section”). As shown in Figure 4a of other metal oxide and chalcogenide NPs in the presence of and Figure S5 (Supporting Information), cyclic voltammetry (CV) OAc and OAm.[44] in the potential range from 0.9 to 1.9 V versus reversible hy- X-ray photoelectron spectroscopy (XPS) was used to study drogen electrode (RHE) was initially conducted in 1 m potas- the surface chemistry of the produced nickel selenides (Fig- sium hydroxide (KOH) at a scan rate of 50 mV s−1 . In alkaline ure 3c; Figures S3 and S4, Supporting Information). As shown electrolyte, the anodic peak at 1.391 V measured in the forward in Figure 3c, the high-resolution Ni 2p3/2 XPS spectrum for sweep corresponds to the Ni(OH)2 /NiOOH oxidation, and the Adv. Sci. 2023, 10, 2300841 2300841 (4 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 3. a) TEM image of branched NiSe2 NPs. b) HRTEM image of an NP branch, detail of the orange squared region and its corresponding power spectrum. c) Ni 2p3/2 and Se 3d high-resolution XPS spectra. cathodic peak at 1.266 V in the backward sweep is attributed to the for all the samples, regardless of their morphology and crystal NiOOH/Ni(OH)2 reduction (black curve).[48] When adding 1 m phase.[49] EG into the electrolyte, a sharp rise in the current density was The surface coverage of active species (Γ*) and the apparent observed at ≈1.405 V (blue curve), pointing that NiOOH is rec- diffusion coefficient (D) were determined from CVs at different ognized as the main active species for EG electro-oxidation.[49] sweep rates, in the range from 10 to 100 mV s−1 (see Figure Without EG, 1.595 V was required to drive a current density of 10 S6 and calculation details in the Supporting Information).[50,51] mA cm−2 , while just 1.389 V was needed in the presence of 1 m From the linear fitting of the dependence of the anodic and ca- EG. However, Pd and Pt catalysts are certainly characterized by thodic peak currents with the sweep rate in the range from 10 much lower on-set potentials, down to just 0.5–0.6 V versus RHE, to 50 mV s−1 , a Γ*= 2.7 × 10−7 mol cm−2 was calculated for the which allows the manufacturing of direct EG fuel cells.[21,24,25] branched NiSe2 -based electrode (Figure 4d). This value decreased For the case of noble-metal-free electrocatalysts, the onset volt- to Γ*= 2.4 × 10−7 mol cm−2 for bundles and Γ*= 1.4 × 10−7 ages measured for NiSe2 in alkaline media in the presence of mol cm−2 for spherical nanostructures. The peak current den- EG are much lower than that required to drive the OER, which sities linearly change with the square root of the sweep rate in could concurrently promote a more efficient H2 generation at the the range of 60–100 mV s−1 , demonstrating the redox reaction to cathode. be diffusion limited. From the fitting, the apparent diffusion co- Among the different NiSe2 samples, the branched particles efficient was D = 2.23 × 10−9 cm2 s−1 for branched NiSe2 , which displayed the lowest EGOR overpotentials (Figure S5, Supporting was slightly above that of bundles (D = 1.83 × 10−9 cm2 s−1 ) and Information). Figure 3b compares the current densities obtained spherical (D = 1.49 × 10−9 cm2 s−1 ) nanostructured electrodes at 1.6 V from electrodes based on NiSe2 particles with different (Figure 4d). morphologies/phases. As can be seen, the branched nanostruc- CVs at different sweep rates in the non-Faradaic region allowed tures have the highest electrocatalytic performance, with 95.23 for estimating the electrochemical surface area (ECSA; Figure mA cm−2 , well above the 40.25 mA cm−2 obtained from spheri- S7, Supporting Information).[52,53] ECSA was found to increase cal NPs and the 83.95 mA cm−2 for bundles. from 5.3 cm−2 for spherical NPs and 6.3 cm−2 for bundles, to Figure 4c shows the CVs of the three electrodes in 1 m KOH 9.65 cm−2 for branched nanostructures. Besides, electrochemi- solution at a scan rate of 50 mV s−1 in the potential window from cal impedance spectroscopy (EIS) showed a higher electrical con- 0.9 to 1.6 V. The potential difference of redox peaks (ΔEp ), re- ductivity and lower charge-transfer resistance for the electrodes vealing the rate of the electron transfer kinetics, was 0.125 V based on branched structures, compared with those based on Adv. Sci. 2023, 10, 2300841 2300841 (5 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 4. a) CV curve of an electrode based on branched NiSe2 particles in 1 m KOH with and without 1 m EG in the potential range of 0.9–1.9 V versus RHE. b) Current density of NiSe2 -based electrodes at 1.6 V versus RHE in 1 m KOH with and without 1 m EG. c) CV curves of NiSe2 -based electrode in 1 m KOH in the potential range of 0.9–1.6 V versus RHE. d) Surface coverage of active species (Γ*) and apparent diffusion coefficient (D) of the nanostructured nickel selenide electrodes. spherical NPs or bundles (Figure S8, Supporting Information). To further examine the EGOR products, a long-term test was Overall, the higher EGOR performance of branched particles was carried out. Figure 5c displays the chronoamperometry (CA) re- consistent with their higher surface coverage of active species, sponse for these three electrodes at 1.6 V for a continuous 10 h diffusion coefficient, ECSA, and charge-transport properties pro- operation in 1 m KOH and 1 m EG. It should be noted that vided by their particular geometry. the fluctuation of the curve was associated with the release of oxygen bubbles during the CA operation (Figure S9, Support- ing Information). A sustained current density well above 50 2.3. In Situ and Ex Situ Electrocatalytic Study mA cm−2 was measured for both branched and bundled NiSe2 - based electrodes. On the other hand, spherical NPs showed Generally, IRAS was performed to study the electrocatalytic oxi- slightly lower current densities at ≈43 mA cm−2 after 10 h test- dation process. Figure 5a shows a schematic of the homemade ing. As it can be seen in Figure S10 (Supporting Information), spectro-electrochemical cell used for IRAS measurements (see at an external potential of 1.65 V, after continuous 2 days CA the “Experimental Section” for details). Figure 5b plots the IRAS testing, a current density of up to 30 mA cm−2 was still mea- spectra collected from a branched NiSe2 NP-based electrode in sured, higher than the initial value of most of the previously re- 1 m KOH containing 1 m EG. We observed the asymmetric ported catalysts (Table S1, Supporting Information). As shown stretching vibration of O–C–O (𝜈 as (OCO), 1568 cm−1 ), and –C–H in Figure S11 (Supporting Information), scanning electron mi- (𝛿(–CH), 1380 cm−1 ), and the symmetric stretching vibration of croscopy (SEM)–energy dispersive spectrometer (EDS) analysis O–C–O (𝜈 s (OCO), 1350 cm−1 ), that can be ascribed to the char- of the branched NiSe2 electrode after the long-term CA test acteristics of HCOO–, formate.[34,54] Based on the band at 1568 demonstrated that the Ni/Se atomic ratio increases from ≈3:7 cm−1 , oxalate is one of the possible products, since the feature (Figure S1b, Supporting Information) to 8:2. Besides, the XPS band of oxalate at 1307 cm−1 may be submerged by the bands at spectra show that the NiSe2 was oxidized to Ni(OH)2 or NiOOH, 1350 cm−1 . The simultaneous appearance of 1568, 1467, 1350, with a reduced amount of Se remaining (Figure S12, Support- and 1085 cm−1 bands also indicates that glycolate species could ing Information). The atomic ratio Ni/Se was quantified at ≈9, be generated, becoming one of the EGOR products.[34] Addition- which is consistent with SEM–EDS data considering just the ally, IRAS spectra showed signs of the presence of carbonate at reconstruction of the material surface. Besides, TEM charac- higher potential, which pointed out that formate was the main terization shows the particle morphology not to change during product of the EGOR at the NiSe2 surface and other carbonates CA (Figure S13, Supporting Information). The electronic struc- were minor products. ture and chemical composition change were associated with the Adv. Sci. 2023, 10, 2300841 2300841 (6 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 5. a) Schematic drawing of the IRAS set-up. b) In situ IRAS spectra of EGOR on branched nickel selenide in 1 m KOH and 1 m EG in the potential range of 0.9–1.7 V versus RHE, taken the spectrum collected at 0.9 V vs. RHE as the reference spetrum. c) CA response for the three electrodes at 1.6 V with continuous 10 h operation in 1 m KOH and 1 m EG. d,e) Ex situ 1 H NMR spectra and IC profiles for the electrolyte after CA measurement of the electrode based on branched nickel selenide. f) Faradaic efficiencies based on IC results. g) GC curves with three samplings at the cathode. h) Schematic drawing of the coupling hydrogen evolution reaction (HER) and EGOR. surface reconstruction under an external potential in alkaline based on the standard concentration curve, we determined that media.[55–57] around 12.23 mg L−1 glycolate, 83.21 mg L−1 formate, and 13.93 After 10 h CA, a small portion of the electrolyte was analyzed by mg L−1 oxalate were produced from the branched NiSe2 -based 1 H nuclear magnetic resonance (NMR) spectroscopy. As shown electrode. in Figure 5d, the NMR spectra confirmed that formate was the The Faradaic efficiency (FE) was calculated using the following main EGOR product. To further determine the reaction prod- equation ucts, the electrolyte after long-term CA was diluted a 20-fold, and analyzed by ion chromatography (IC). Figure 5e displays mol of product × n × F FE (%) = × 100% (1) the experimental IC curve (black) collected from the branched total charge passed NiSe2 -based electrode and the reference data obtained from 50 mg L−1 glycolate (red), formate (blue), and oxalate (pink). The IC where n is the electron transfer number (4 for glycol to glycolate, curve obtained from the electrolyte displayed peaks at 4.3, 5.2, 6 for glycol to formate, and 8 for glycol to oxalate) and F is the and 27.6 min, which matched well with the reference peak, re- Faradaic constant (96 485 C mol−1 ). As displayed in Figure 5f, the vealing the formation of glycolate, formate, and oxalate during Faradaic efficiency of the glycol-to-formate conversion was above EGOR.[58] Similar results were obtained from the three differ- 80% for the three electrodes. Besides, the FE values for glycol to ent NiSe2 -based electrodes (for spherical NP and bundle-based glycolate and glycol to oxalate were around 5% and 7%, respec- electrodes; Figure S14, Supporting Information). Quantitatively, tively. Adv. Sci. 2023, 10, 2300841 2300841 (7 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 6. a) Optimized structural model of NiOOH and NiOOH–SeO4 from the side and top views. b) PDOS of NiOOH and NiOOH–SeO4 regarding the Ni 3d orbitals. c) Gibbs free energy diagrams for the glycol-to-oxalate C2 conversion on a NiOOH surface. d) Gibbs free energy diagrams for the glycol-to-formate C1 conversion on a NiOOH surface considering the C–C bond cleavage at the glycolate step (red) and the glyoxylate step (yellow). The free energy of each configuration is given in eV. e) Proposed reaction pathway for the electrocatalytic glycol conversion. The concept of coupling HER and EGOR in alkaline media Ni in the SeO4 -modified NiOOH were found to significantly con- was further studied. As shown in Figure S15 (Supporting Infor- centrate around the Fermi level, leading to an enhancement of mation), an H-cell with two compartments separated by a proton the electroconductivity and electron transfer. Thus, the higher exchange membrane was used to produce hydrogen and quan- EGOR performance of the NiSe2 -based electrode could be associ- tify its evolution rate. Figure 5g displays the curve of the detected ated with the presence of SeOx species during the electrocatalytic product at the cathode side. As can be seen in Figure 5g, two process, in agreement with previous publications.[42] peaks were located at ≈1.1 and ≈1.5 min, which were associated To gain insight into the reaction mechanism and understand with molecular hydrogen and argon. By comparing with known the competition among the dehydrogenation, and C–C bond standard hydrogen concentrations, the calculated Faradaic effi- cleavage at the molecular level, the adsorption energies in the ciency was determined to be nearly 100%, indicating that hydro- reaction coordinates were determined using DFT calculations. gen was the only product on the cathode. As shown in Figure Considering NiOOH as the main active species for EGOR in al- S16 (Supporting Information), a sharp rise in the CA curve after kaline solution, we built a slab model of NiOOH (001). For the adding 1 m EG in the 1 m KOH solution was observed, indicat- first step, the glycol binds to the top site via one O atom that was ing that the presence of EG significantly promotes the HER at removed from the model, and the calculated binding energy for the cathode (Figure 5h). glycol is −0.67 eV, which is in good agreement with that obtained on the glycol on Pt (211) surface.[59] To figure out the reaction pathways for the EGOR in this study, 2.4. DFT Calculations first, we analyzed the pathway for glycol-to-oxalate conversion. Figure 6c shows the free energy of adsorbed intermediates for the To determine the role of Se in the electrocatalytic performance, glycol-to-oxalate conversion on the NiOOH surface. According DFT calculations were used to calculate the partial density of state to previous investigations, the adsorbed glycol (CH2 OHCH2 OH) (PDOS) of the NiOOH–SeO4 and the pristine NiOOH structural requires two dehydrogenations to form the intermediate of gly- models (Figure 6a). As shown in Figure 6b, the d electrons of colaldehyde (CH2 OHCHO). At this point, the next step involves Adv. Sci. 2023, 10, 2300841 2300841 (8 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com either further dehydrogenation on the same C and interacting C ∗H2 OHCOOHads +2OHn +NiOOH with OH− to form glycolate (CH2 OHCOOH) or another two de- hydrogenations on another C resulting in glycoxal (CHOCHO). → C ∗ HOCOOHads +Ni(OH)2 +2H2 O + 2e (7) As can be seen, both reactions are exothermal, and more energy is released to form glycolate (−6.24 eV) than glycoxal (−4.74 eV). C ∗ HOCOOHads + 2OH− + NiOOH This means that the former reaction pathway of glycolaldehyde to → (COOH)2 ads + Ni(OH)2 + H2 O + 2e (8) glycolate is favored. With another two dehydrogenations on an- other C, the glyoxylate (CHOCOOH) is formed. This can finally (COOH)2ads → (COOH)2sol (9) interact with OH− to produce the oxalate ((COO− )2 ). The glycol-to-formate reaction pathway displays more possibil- Based on the DFT calculations, the C–C bond cleavage ities due to the C–C bond cleavage that occurs on one of the in- preferentially occurs in the adsorbed CH2 OHCOOH, forming termediates of the glycol-to-oxalate conversion. Figure S15 (Sup- *CH2 OH and *COOH, and resulting in the presence of formate porting Information) displays the adsorbed intermediates from in the solution glycol-to-formate conversion on the NiOOH surface regarding the C–C bond cleavage in glycol (CH2 OHCH2 OH), glycolalde- CH2 OHCOOHads +OH− → ∗ CH2 OHads + ∗ COOHads (10) hyde (CH2 OHCHO), and glycoxal (CHOCHO). As can be seen in this graph, the reaction pathway for the two dehydrogenations to The intermediate *COOH interacts with NiOOH in the follow- CH2 OHCHO is more favored than for the direct C–C bond cleav- ing way age in the glycol (CH2 OHCH2 OH). Among these three pathways (Figure S17, Supporting Information), the C–C bond cleavage is ∗ COOHads +NiOOH+H2 O + 2e → HCOOHads +Ni(OH)2 +OH− more likely to occur in the glycoxal (CHOCHO), resulting in two pieces of formaldehyde (CHO). Figure 6d compares the reaction (11) pathway for glycol-to-formate conversion that involves the C–C The other intermediate *CH2 OH probably undergoes two de- bond cleavage in the glycolate (CH2 OHCOOH) or the glyoxylate hydrogenations (CHOCOOH). As it can be seen, the C–C bond cleavage is more favored in the former intermediates glycolate (CH2 OHCOOH). ∗ CH2 OHads +OH− +NiOOH → ∗ CHOads +Ni(OH)2 +H2 O + e By comparing the C–C bond cleavage in the glycolate (Figure 6d) and glycoxal (Figure S17, Supporting Information), less reaction (12) energy was needed for breaking the C–C bond in glycolate than in glycoxal. Thus, we hypothesize that formate was formed pref- ∗ CHOads +2OH− +NiOOH → ∗ HCOOHads +Ni(OH)2 +2e (13) erentially by the C–C bond cleavage at the glycolate and the pos- terior further oxidation to formate. Finally, the adsorbed formic acid is released from the electro- To sum up, the possible EGOR mechanism on a nickel sur- catalyst surface face in alkaline media is illustrated in Figure 6e. First, NiOOH was formed from the Ni surface reconstruction under the exter- ∗ HCOOHads → ∗ HCOOHsol (14) nal potential. Then, the glycol was adsorbed on the NiOOH sur- face, followed by two dehydrogenations on one carbon, and inter- 3. Conclusion acting with OH− to form glycolate To sum up, we reported on the production of nickel selenide OH− + Ni(OH)2 → NiOOH + H2 O + e (2) nanostructures with controlled morphology and crystallographic phases. These nanostructures were tested for the EGOR in al- CH2 OHCH2 OHsol → CH2 OHCH2 OHads (3) kaline electrolyte. The best performance was obtained from branched NiSe2 nanoparticles, displaying a current density of up CH2 OHC ∗H2 OHads + 3OHn + NiOOH to 95 mA cm−2 in an electrolyte containing 1 m KOH and 1 m EG at 1.6 V. Even after 10 h continuous CA operation, the cur- → CH2 OHC ∗ HOads + Ni(OH)2 + 2H2 O + 3e (4) rent density was sustained at 56 mA cm−2 . The improved activity was associated with improved surface coverage, diffusion of ac- CH2 OHC ∗ HOads + 3OHn + NiOOH tive species, and ECSA. Further in situ IRAS and ex situ NMR and IC investigations showed the main product of the EGOR to be for- → CH2 OHC ∗ OOHads + Ni(OH)2 + 2H2 O + 3e (5) mate with up to 80% FE, and with glycolate and oxalate as minor Some of the adsorbed CH2 OHCOOH was released from products. DFT calculations provided insights into the oxidation the surface and existed in solution in the form of glycolate mechanism, demonstrating that the presence of Se promotes the (CH2 OHCOO- ) catalyst activity and suggesting that the C–C bond cleavage takes place in the glycolate. CH2 OHCOOHads → CH2 OHCOOHsol (6) On the other hand, some of the adsorbed CH2 OHCOOH 4. Experimental Section undergo two dehydrogenations at another C to form oxalate Chemicals: All the chemicals, including selenium powder (200 mesh, ((COO− )2 ) 99.5%, Acros Organics), nickel(II) acetylacetonate (Ni(acac)2 , 96%, Adv. Sci. 2023, 10, 2300841 2300841 (9 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Sigma–Aldrich), OAm (80–90%, Acros Organics), OAc (99%, Sigma– measurement, a stable flow rate of 20 cm3 min−1 Ar (99.999%) was Aldrich), KOH (85%, Sigma–Aldrich), potassium carbonate (K2 CO3 , bubbled for 30 min at the Pt wire in an H-type two-compartment cell, 99.5%, Sigma–Aldrich), potassium bicarbonate (KHCO3 , 99.7%, Sigma– which was separated by a Nafion-115 proton membrane. The GC product Aldrich), ODE (90%, Sigma-Aldrich), DDT (98%, Sigma–Aldrich), carbon analysis was conducted every 12 min. The produced gas was identified black (CB, Vulcan XC72, Sigma–Aldrich), EG (99.8%, Sigma Aldrich), and and quantified by a calibration curve using known concentrations of Nafion (10 wt%, perfluorinated ion-exchange resin, dispersion in water, standard H2 gas. Sigma–Aldrich), were used as received, without further treatment. Analyt- In Situ Electrochemical IRAS: In situ IRAS measurements were per- ical grade chloroform and ethanol for the washing process were obtained formed using a homemade spectro-electrochemical cell. The incident an- from various sources. MilliQ water (18.2 MΩ cm) was used for the prepa- gle of the IR beam was set at 55°. Before spectra gathering, a ≈10 μm thick ration of electrolyte and catalytic ink, and IC measurements. An argon- layer containing 1 m KOH and 1 m EG electrolyte was generated by press- filled glove box was used to handle sensitive chemicals and reactive ink ing the WE onto the CaF2 prism surface, allowing the infrared radiation preparation during the synthesis. to travel through the CaF2 prism and reflect at the WE into the mercury– Synthesis of NiSe2 : Nanostructured NiSe2 was produced from a two- cadmium telluride (MCT) detector. The infrared spectrum was collected step process involving the preparation of a reactive ink and its subse- every 5 s, and potential scanning was 5 mV s−1 . Thus the spectrum was quent thermal decomposition. A series of nickel selenide samples, having generated by the average co-adding of ≈44 interferograms. All the spec- particles with different morphologies, were produced using the type and tra in this work were shown in the absorbance units defined as −log(I/I0 ), ratio of solvents and surfactants as the variable parameter.[60] First, the where I and I0 represent the absorption intensities at the sample and cor- precursor ink was prepared inside an argon-filled glove box by dissolving relative conditions, respectively. 0.4 mm Ni(acac)2 and 2.5 mm of Se powder inside a vial with 30 mL OAm Computational Method: The Vienna ab initio package (VASP) was and 6 mL DDT under vigorous stirring for 1 h at 60 °C. DDT was essen- used to perform all the DFT calculations within the generalized gradi- tial to dissolve the Se powder. The obtained solution was filtered using ent approximation (GGA) using the Perdew, Burke, and Ernzerhof (PBE) a 0.2 μm filter. Then, 10 mL of the prepared precursor together with the formulation.[63–65] The projected augmented wave (PAW) potential was appropriate amount of OAm, ODE, and/or OAc was placed in a 50 mL used to describe that the ionic cores and valence electrons were taken three-neck flask and vacuumed at 60 °C for 30 min followed by heating into account using a plane wave basis set with a kinetic energy cutoff of to 230 °C under Ar at a rate of 5 °C min−1 . After reacting for 20 min, the 400 eV.[66,67] Partial occupancies of the Kohn–Sham orbitals were allowed solution was rapidly cooled down to room temperature in a water bath. using the Gaussian smearing method and a width of 0.05 eV. The elec- Then, the crude solution was mixed with 10 mL of chloroform and cen- tronic energy was considered self-consistent when the energy change was trifuged at 8000 rpm for 5 min. This purification process, involving redis- smaller than 10−5 eV. Geometry optimization was considered convergent persion in chloroform and precipitation using ethanol and centrifugation, when the force change was smaller than 0.02 eV Å−1 . Grimme’s DFT-D3 was repeated several times. The final product was dissolved in chloroform methodology was used to describe the dispersion interactions.[68] and stored in the glove box. Different OAm/ODE/OAc ratios were used to Using a 7×11×9 Monkhorst–Pack k-point grid for Brillouin zone sam- produce NiSe2 with different geometries and crystal phases: NiSe2 quasi- pling, the equilibrium lattice constants of the monoclinic NiOOH unit cell spherical nanoparticles were obtained using 20 mL of OAm and no ODE; were optimized to be a = 5.168 Å, b = 2.847 Å, c = 4.516 Å, at 𝛼 = 90°, 𝛽 = NiSe2 branched nanostructures were produced using 5 mL of OAm and 107.1°, and 𝛾 = 90°. They were used to construct a NiOOH(001) surface 15 mL ODE; NiSe2 elongated bundles were obtained in the presence of model with p(2×3) periodicity in the x- and y-directions and one stoichio- 5 mL of OAm, 3 mL ODE, and 12 mL OAc. metric layer in the z-direction separated by a vacuum layer in the depth Material Characterization: Laboratory XRD was performed on a Bruker of 15 Å to separate the surface slab from its periodic duplicates. One O AXS D8 Advance (Cu K𝛼 radiation: 𝜆 = 1.5106 Å). SEM together with an atom was removed from the model to create an oxygen vacancy. During EDS was performed on a Gemini 300 field emission scanning electron mi- structural optimizations, the gamma point in the Brillouin zone was used croscope (ZEISS, Germany) equipped with an AZtecOne UltimMax40 en- for k-point sampling, and all atoms were allowed to relax. ergy spectrometer. HRTEM and STEM investigations were performed on The free energy of a gas phase molecule or an adsorbate on the surface a field emission gun FEI Tecnai F20 microscope. HAADF–STEM was com- was calculated by the equation G = E + ZPE − TS, where E is the total bined with EELS in the Tecnai microscope using a GATAN QUANTUM energy, ZPE is the zero-point energy, T is the temperature in kelvin (298.15 filter.[61,62] XPS analyses were conducted on a SPECS system. K), and S is the entropy. Electrochemical Characterization: Electrochemical characterization was performed at room temperature on a CorrTest potentiostat. A conventional three-electrode cell was used, containing a Pt wire as a counter electrode (CE), Hg/HgO as a reference electrode (RE), and GC Supporting Information (5 mm diameter) as the working electrode (WE). To prepare the catalytic Supporting Information is available from the Wiley Online Library or from ink, 5 mg of dried particles and 10 mg of CB were dispersed in a vial the author. containing 1 mL of MilliQ water and 1 mL of ethanol and 100 μL of 10 wt% Nafion, followed by half an hour vigorous sonication. Then, 5 μL of the prepared ink was drop-casted on the polished GC electrode and dried naturally in the open air. Conventional electrochemical techniques Acknowledgements including CV and CA were applied to study the activities and stabilities. Current densities were normalized to the geometrical area (0.196 cm−2 ) This work was supported by the start-up funding at Chengdu University of the GC electrode. The measured potentials (vs Hg/HgO) were con- and the Natural Science Foundation of Sichuan (NSFSC) project funded verted to the RHE according to the equation ERHE = EHg/HgO + 0.059 × by the Science and Technology Department of Sichuan Province (Project pH + Eϴ Hg/HgO (where EHg/HgO is the measured potential, Eϴ Hg/HgO is No. 2022NSFSC1229), and also the open project from Hebei Key Lab- the reference potential of 0.098 V according to the manufacturer guide, oratory of Photoelectric Control on Surface and Interface (Project No. and pH is the practical value of 13.6 for 1.0 m KOH). Electrochemical ZD2022003). It was also supported by the European Regional Devel- impedance spectra were obtained in 1 m KOH in the presence of 1 m EG opment Funds and by the Spanish Ministerio de Ciencia e Innovación from 10−1 to 105 Hz. To identify and quantify the reaction products, after through the project COMBENERGY (Project No. PID2019-105490RB- long-term CA, the electrolyte was characterized by 1 H NMR spectroscopy C32). Y.-Y.Y. acknowledges funding from the National Natural Science and IC. For IC measurement, a freshly prepared 4.5 mm K2 CO3 and Foundation of China (NSFC, Grant No. 22172121), the Natural Science 0.8 mm KHCO3 solution was used as leachate solution. At the cathode, Foundation of Sichuan Province (NSFSC, Grant No. 23NSFSC6266) and the generated hydrogen was determined by gas chromatography (GC, the Fundamental Research Funds for the Central Universities, Southwest 8890, Agilent Technologies) connected to the electrocatalytic cell. Before Minzu University (Grant No. xiao2021102). X.H. has received funding Adv. Sci. 2023, 10, 2300841 2300841 (10 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com from the CSC-UAB Ph.D. scholarship program. X.H. and J.A. acknowl- [18] J. Li, R. Wei, X. Wang, Y. Zuo, X. Han, J. Arbiol, J. Llorca, Y. Yang, A. edge funding from Generalitat de Catalunya 2021SGR00457. ICN2 ac- Cabot, C. Cui, Angew. Chem., Int. Ed. 2020, 59, 20826. knowledges support from the Severo Ochoa Programme from Spanish [19] J. Pang, M. Zheng, R. Sun, A. Wang, X. Wang, T. Zhang, Green Chem. MCIN/AEI (Grant No. CEX2021-001214-S). ICN2 authors thank the sup- 2016, 18, 342. port from the project NANOGEN (PID2020-116093RB-C43), funded by [20] ChemAnalyst, “Global Chemical and Petrochemicals, Specialty MCIN/AEI/10.13039/501100011033/ and by "EDRF a way of making Eu- Chemicals, Elastomer and Rubber, Fertilizer and Feedstock—Latest rope", by the "European Union". IREC and ICN2 were funded by the CERCA Chemical Prices, News and Market Analysis | ChemAnalyst,” https: Programme/Generalitat de Catalunya. Part of the present work was per- //www.chemanalyst.com/ (accessed: September 2022). formed in the framework of the Universitat Autònoma de Barcelona Ma- [21] K. Miyazaki, T. Matsumiya, T. Abe, H. Kurata, T. Fukutsuka, K. Kojima, terials Science Ph.D. program. This study was also supported by MCIN Z. Ogumi, Electrochim. Acta 2011, 56, 7610. with funding from European Union NextGenerationEU (Grant No. PRTR- [22] H. Xu, C. Liu, P. Song, J. Wang, F. Gao, Y. Zhang, Y. Shiraishi, J. Di, Y. C17.I1) and Generalitat de Catalunya. Du, Chem. Asian J. 2018, 13, 626. [23] B. Qiao, T. Yang, S. Shi, N. Jia, Y. Chen, X. Chen, Z. An, P. Chen, Small 2021, 17, 2006534. Conflict of Interest [24] Y. Qin, W. Zhang, F. Wang, J. Li, J. Ye, X. Sheng, C. Li, X. Liang, P. Liu, The authors declare no conflict of interest. X. Wang, X. Zheng, Y. Ren, C. Xu, Z. Zhang, Angew. Chem. 2022, 134, e202200899. [25] D. Si, B. Xiong, L. Chen, J. Shi, Chem. Catal. 2021, 1, 941. Data Availability Statement [26] S. Arulmani, P. V. Kumar, G. Landi, S. Anandan, ChemistrySelect 2019, 4, 6130. The data that support the findings of this study are available from the cor- [27] R. G. Da Silva, A. Rodrigues de Andrade, K. Servat, C. Morais, T. W. responding author upon reasonable request. Napporn, K. B. Kokoh, ChemElectroChem 2020, 7, 4326. [28] X. Wang, S. Xi, W. S. V. Lee, P. Huang, P. Cui, L. Zhao, W. Hao, X. Zhao, Z. Wang, H. Wu, H. Wang, C. Diao, A. Borgna, Y. Du, Z. G. Yu, Keywords S. Pennycook, J. Xue, Nat. Commun. 2020, 11, 4647. [29] M. T. Bender, Y. C. Lam, S. Hammes-Schiffer, K. S. Choi, J. Am. Chem. electrocatalysis, ethylene glycol electro-oxidation, formate, hydrogen, Soc. 2021, 142, 21538. nickel selenide, phase engineering [30] H. Y. Wang, Y. Y. Hsu, R. Chen, T. S. Chan, H. M. Chen, B. Liu, Adv. Energy Mater. 2015, 5, 1500091. Received: February 7, 2023 [31] W. Chen, C. Xie, Y. Wang, Y. Zou, C. L. Dong, Y. C. Huang, Z. Xiao, Z. Revised: February 21, 2023 Wei, S. Du, C. Chen, B. Zhou, J. Ma, S. Wang, Chem 2020, 6, 2974. Published online: March 22, 2023 [32] J. Li, X. Wang, C. Xing, L. Li, S. Mu, X. Han, R. He, Z. Liang, P. Mar- tinez, Y. Yi, Q. Wu, H. Pan, J. Arbiol, C. Cui, Y. Zhang, A. Cabot, Chem. Eng. J. 2022, 440, 135817. [33] Y. Yi, J. Li, C. Cui, Chin. Chem. Lett. 2022, 33, 1006. [1] I. Roger, M. A. Shipman, M. D. Symes, Nat. Rev. Chem. 2017, 1, 0003. [34] S. C. Chang, Y. Ho, M. J. Weaver, J. Am. Chem. Soc. 1991, 113, 9506. [2] T. Lu, T. Li, D. Shi, J. Sun, H. Pang, L. Xu, J. Yang, Y. Tang, SmartMat [35] L. Gao, X. Cui, C. D. Sewell, J. Li, Z. Lin, Chem. Soc. Rev. 2021, 50, 2021, 2, 591. 8428. [3] S. Wu, Y. Zhu, G. Yang, H. Zhou, R. Li, S. Chen, H. Li, L. Li, O. [36] S. Jin, ACS Energy Lett. 2017, 2, 1937. Fontaine, J. Deng, Chem. Eng. J. 2022, 446, 136833. [37] Y. Zuo, Y. Liu, J. Li, R. Du, X. Han, T. Zhang, J. Arbiol, N. J. Divins, J. [4] B. You, Y. Sun, Acc. Chem. Res. 2018, 51, 1571. Llorca, N. Guijarro, K. Sivula, A. Cabot, Chem. Mater. 2019, 31, 7732. [5] X. Wang, C. Xing, Z. Liang, P. Guardia, X. Han, Y. Zuo, J. Llorca, J. [38] X. Li, L. Xiao, L. Zhou, Q. Xu, J. Weng, J. Xu, B. Liu, Angew. Chem. Arbiol, J. Li, A. Cabot, J. Mater. Chem. A 2022, 10, 3659. 2020, 132, 21292. [6] X. Wang, X. Han, R. Du, C. Xing, X. Qi, Z. Liang, P. Guardia, J. Arbiol, [39] J. Y. Zhang, Y. Yan, B. Mei, R. Qi, T. He, Z. Wang, W. Fang, S. Zaman, A. Cabot, J. Li, ACS Appl. Mater. Interfaces 2022, 14, 41924. Y. Su, S. Ding, B. Y. Xia, Energy Environ. Sci. 2021, 14, 365. [7] W. J. Liu, Z. Xu, D. Zhao, X. Q. Pan, H. C. Li, X. Hu, Z. Y. Fan, W. K. [40] K. Fan, H. Zou, Y. Lu, H. Chen, F. Li, J. Liu, L. Sun, L. Tong, M. F. Toney, Wang, G. H. Zhao, S. Jin, G. W. Huber, H. Q. Yu, Nat. Commun. 2020, M. Sui, J. Yu, ACS Nano 2018, 12, 12369. 11, 265. [41] X. Wang, X. Han, R. Du, Z. Liang, Y. Zuo, P. Guardia, J. Li, J. Llorca, J. [8] S. Sheng, K. Ye, Y. Gao, K. Zhu, J. Yan, G. Wang, D. Cao, J. Colloid Arbiol, R. Zheng, A. Cabot, Appl. Catal. B 2023, 320, 121988. Interface Sci. 2021, 602, 325. [42] S. Li, R. Ma, J. Hu, Z. Li, L. Liu, X. Wang, Y. Lu, G. E. Sterbinsky, S. [9] G. Yang, Y. Jiao, H. Yan, Y. Xie, A. Wu, X. Dong, D. Guo, C. Tian, H. Liu, L. Zheng, J. Liu, D. Liu, J. Wang, Nat. Commun. 2022, 13, 2916. Fu, Adv. Mater. 2020, 32, 2000455. [43] B. Yuan, W. Luan, S. T. Tu, CrystEngComm 2012, 14, 2145. [10] X. Liu, Y. Han, Y. Guo, X. Zhao, D. Pan, K. Li, Z. Wen, Adv. Energy [44] S. Mourdikoudis, M. Menelaou, N. Fiuza-Maneiro, G. Zheng, S. Wei, Sustainability Res. 2022, 3, 2200005. J. Pérez-Juste, L. Polavarapu, Z. Sofer, Nanoscale Horiz. 2022, 7, 941. [11] S. Ghosh, B. Mondal, S. Roy, M. Shalom, M. B. Sadan, J. Mater. Chem. [45] J. F. Moulder, W. F. Stickle, P. E. Sobol, K. D. Bomben, in Hand- A 2022, 10, 8238. book of X-Ray Photoelectron Spectroscopy: A Reference Book of Stan- [12] H. Luo, J. Barrio, N. Sunny, A. Li, L. Steier, N. Shah, I. E. L. Stephens, dard Spectra for Identification and Interpretation of XPS Data, Perkin- M. M. Titirici, Adv. Energy Mater. 2021, 11, 2101180. Elmer Corporation, Minnesota, United States of America 1992, p. 84. [13] H. Zhou, Y. Ren, Z. Li, M. Xu, Y. Wang, R. Ge, X. Kong, L. Zheng, H. [46] R. Du, K. Xiao, B. Li, X. Han, C. Zhang, X. Wang, Y. Zuo, P. Guardia, Duan, Nat. Commun. 2021, 12, 4679. J. Li, J. Chen, J. Arbiol, A. Cabot, Chem. Eng. J. 2022, 441, 135999. [14] B. You, X. Liu, N. Jiang, Y. Sun, J. Am. Chem. Soc. 2016, 138, 13639. [47] J. F. Moulder, W. F. Stickle, P. E. Sobol, K. D. Bomben, in Hand- [15] L. T. Mika, E. Cséfalvay, Á. Németh, Chem. Rev. 2018, 118, 505. book of X-Ray Photoelectron Spectroscopy: A Reference Book of Stan- [16] D. A. Bulushev, J. R. H. Ross, ChemSusChem 2018, 11, 821. dard Spectra for Identification and Interpretation of XPS Data, Perkin- [17] J. Li, C. Xing, Y. Zhang, T. Zhang, M. C. Spadaro, Q. Wu, Y. Yi, S. He, Elmer Corporation, Minnesota, United States of America 1992, p. 96. J. Llorca, J. Arbiol, A. Cabot, C. Cui, Small 2021, 17, 2006623. [48] D. M. MacArthur, J. Electrochem. Soc. 1970, 117, 422. Adv. Sci. 2023, 10, 2300841 2300841 (11 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com [49] D. Chen, S. D. Minteer, J. Power Sources 2015, 284, 27. [58] L. Barron, E. Gilchrist, Anal. Chim. Acta 2014, 806, 27. [50] X. Cui, P. Xiao, J. Wang, M. Zhou, W. Guo, Y. Yang, Y. He, Z. Wang, Y. [59] X. K. Gu, B. Liu, J. Greeley, ACS Catal. 2015, 5, 2623. Yang, Y. Zhang, Z. Lin, Angew. Chem., Int. Ed. 2017, 56, 4488. [60] Q. Yang, S. Xu, Q. Yang, D. Zhang, Z. Li, H. Zhou, S. Zhu, Bioresour. [51] C. Zhang, S. Park, J. Electrochem. Soc. 1987, 134, 2966. Technol. 2020, 316, 123972. [52] C. Wei, S. Sun, D. Mandler, X. Wang, S. Z. Qiao, Z. J. Xu, Chem. Soc. [61] S. Bernal, F. J. Botana, J. J. Calvino, C. López-Cartes, J. A. Pérez-Omil, Rev. 2019, 48, 2518. J. M. Rodríguez-Izquierdo, Ultramicroscopy 1998, 72, 135. [53] C. C. L. McCrory, S. Jung, I. M. Ferrer, S. M. Chatman, J. C. Peters, T. [62] J. Arbiol, A. Cirera, F. Peiró, A. Cornet, J. R. Morante, J. J. Delgado, J. F. Jaramillo, J. Am. Chem. Soc. 2015, 137, 4347. J. Calvino, Appl. Phys. Lett. 2002, 80, 329. [54] A. Kowal, S. N. Port, R. J. Nichols, Catal. Today 1997, 38, 483. [63] G. Kresse, J. Furthmüller, Comput. Mater. Sci. 1996, 6, 15. [55] J. Li, Y. Zuo, J. Liu, X. Wang, X. Yu, R. Du, T. Zhang, M. F. Infante- [64] G. Kresse, J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys. Carrió, P. Tang, J. Arbiol, J. Llorca, Z. Luo, A. Cabot, J. Mater. Chem. A 1996, 54, 11169. 2019, 7, 22036. [65] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865. [56] J. Li, Z. Luo, F. He, Y. Zuo, C. Zhang, J. Liu, R. Du, X. Yu, T. Zhang, P. [66] D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, Tang, M. F. I. Carrió, J. Arbiol, J. Llorca, A. Cabot, J. Mater. Chem. A 1758. 2018, 6, 22915. [67] P. E. Blöchl, Phys. Rev. B 1994, 50, 17953. [57] X. Yu, Z. Luo, T. Zhang, P. Tang, J. Li, X. Wang, J. Llorca, J. Arbiol, J. [68] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132, Liu, A. Cabot, Chem. Mater. 2020, 32, 2044. 154104. Adv. Sci. 2023, 10, 2300841 2300841 (12 of 12) © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH