Introduction

EA is a fundamental chemical widely utilized in industry and organic synthesis. The annual demand of EA exceeded half a million tons in 2022, with a value exceeding 600 million dollars1,2,3,4,5. Beyond synthesized through the amination of ethanol6,7, acetaldehyde or acetone8,9, the most prevalent method of industrial EA production is the direct hydrogenation of acetonitrile, due to the availability and cost-effectiveness of acetonitrile as the raw material. In industrial process, the hydrogenation of acetonitrile is catalyzed under elevated temperature and pressure, and employed explosive hydrogen gas as the proton source1,10,11. Moreover, the intrinsic nucleophilicity of amines leads to the reaction of EA with imine intermediates to form secondary and tertiary amines. As a result, the product is usually a mixture of EA, dimethylamine and triethylamine, which requires further separation, leading to increased costs10,12.

Electrocatalytic hydrogenation of acetonitrile, which can be powered by renewable electricity, has gained great interest in the past several years3,4,5,13,14. The process is operated under room temperature by using water as the hydrogen source. ARR to EA takes place at the cathode as a four-proton and electron transfer process. In comparison, the HER undergoes a two-proton and electron transfer mechanism, significantly competing with ARR. To promote ARR and suppress HER, an ideal catalyst should exhibit a moderate affinity to the nitrile but not form too strong interaction with the amine group to hinder desorption of EA product4,11,14. Metals such as Pt, Pd, and Ni have demonstrated both intense adsorption properties towards nitrile and amine, inhibiting desorption of EA. In addition, these metals also serve as excellent HER catalysts, leading to poor FE toward EA4,10,15. Metallic Cu nanoparticle has been explored as a promising catalyst for ARR, which exhibits remarkable EA selectivity and current density. Nonetheless, the FE at high current densities remains suboptimal, with the formation of diethylamine as a byproduct5. Zhang et al. introduced CO2 to suppress the formation of diethylanime or triethylamine on Cu catalysts. CO2 reacts with the amine group to prevent EA from condensing into dimers and trimers13. The excess free electrons on Cu surface lead to strong amine adsorption, thus oxygen atoms were introduced into the lattice to reduce the free electrons4. Despite these efforts, the stabilities of most studied Cu-based catalysts only maintain for less than 20 h. Notably, in the previous studies, the Cu-based catalysts are prepared by reducing CuO or Cu2O through electrochemical treatment or thermal reduction, thus the majority of CuO is reduced to metallic Cu that may prompt the HER. Moreover, the hydrogenation of acetonitrile process was not extensively investigated, especially the source of hydrogen atoms has not been identified.

Herein, we demonstrate that the HER can be suppressed by coupling the Volmer step and ARR process on a self-supported CuO@Cu heterostructure. The CuO@Cu heterostructure was prepared through electrochemical treatment of a Cu2O film in the presence of acetonitrile. A disproportionation reaction occurs, which converts Cu2O into Cu and CuO, rather than the deeper reduction of Cu2O to Cu. The CuO@Cu heterostructure demonstrates high FEEA approaching 100% across the entire tested potential range of −0.1 to −0.4 V vs. RHE. The catalyst remains stable over 1000 h of testing at a constant current density of 0.2 A cm−2 with FEEA consistently exceeding 95% and no notable increase in overpotential. The ARR process is systematically investigated by electrochemical impedance spectroscopy (EIS), in situ spectra analysis, and density functional theory (DFT) calculations. We reveal that although HER and ARR were competing reactions on the cathode, they share the same elementary step, namely the Volmer step. The electronic structure of CuO@Cu facilitates the migration of hydrogen atoms generated in the Volmer step to the adjacent unsaturated bond of AN. Thus, a synergistic hydrogenation process is established, which involves both surface hydrogen migration and proton addition from solution.

Results

Synthesis and structural characterizations

The preparation of the CuO@Cu heterostructure is depicted in Fig. 1a. Initially, the Cu foam is subjected to high-temperature treatment (450–550 oC) in air to form a surface oxide layer. In terms of ARR activity, the optimized annealing temperature is 450 oC (Supplementary Fig. 1). Thus, we focused on the optimized CuO@Cu in the following investigations. As shown in the X-ray diffraction (XRD) analysis in Fig. 1b, after annealing for 2 h, the Cu foam displays characteristic peaks of both Cu2O and Cu. Herein, the surface of the Cu foam converts to Cu2O while the inner metallic Cu is preserved. Subsequently, the oxidized Cu foam undergoes 100 cycles of electrochemical cyclic voltammetry (CV) at negative potentials in 1 M NaOH with acetonitrile (12 wt%). The resulting material exhibits a typical CuO structure on the surface, corresponding to the black line in Fig. 1b. Normally, Cu2O is unable to be oxidized under negative potential. We posit that a disproportionation reaction occurs, in which Cu2O transforms into metallic Cu and CuO:

$${{\mbox{Cu}}}_{2}{\mbox{O}}\to {\mbox{Cu}}+{\mbox{CuO}}$$
(1)
Fig. 1: Preparation illustration and structural characterizations.
figure 1

a Schematic diagrams of synthesis CuO@Cu catalyst; b XRD patterns of Cu2O/Cu and CuO@Cu; c L3M45M45 Auger spectra of Cu2O/Cu and CuO@Cu; d Raman spectra of Cu2O/Cu and CuO@Cu; e TEM image of Cu2O/Cu; f TEM image of CuO@Cu; g Dependence of Cu concentration in electrolyte on CV cycles. Source data for (b–d) and (g) are provided as a Source Data file.

Full size image

Cu L3M45M45 Auger spectroscopy was utilized for a detailed investigation of the surface Cu chemical states. As illustrated in Fig. 1c, the peak at 916.8 eV corresponds to Cu+, while 917.7 eV and 918.6 eV represent Cu2+ and Cu0, respectively16,17,18. In the annealed Cu foam, Cu+ predominates, related to the formation of Cu2O. After CV activation in the presence of acetonitrile, Cu2+ peak becomes predominant, and Cu0 peak appears as a shoulder of Cu2+ peak, indicating a CuO@Cu heterostructure. Raman analysis was carried out to investigate the surface species. The peaks at 147 cm−1, 217 cm−1, and 632 cm−1 in the spectrum of the annealed Cu foam correspond to Cu2O (Fig. 1d)19,20. In comparison, the CuO@Cu heterostructure exhibits distinct characteristic peaks of CuO (98 cm−1, 344 cm−1, and 630 cm−1). The absence of specific Cu2O peaks in the spectrum of CuO@Cu heterostructure confirms the transformation of Cu2O. Since metallic Cu do not show Raman signals, Cu substrate cannot be detected in the Raman spectra. Moreover, we performed transmission electron microscope (TEM) test to direct observe the alteration in the morphology and structure before and after CV activation. As shown in Fig. 1e, the interplanar distance of 0.242 nm corresponds to the Cu2O (111) crystallographic plane. Following CV activation, the interplanar distance increases to 0.246 nm, aligning with the (002) crystal plane of CuO (Fig. 1f). These results are consistent with each other, and clearly demonstrate that Cu2O undergoes a disproportionation reaction to metallic Cu and CuO in the CV cycling under negative potentials in the presence of acetonitrile.

We should emphasize that the presence of AN is crucial in suppressing the reduction of Cu oxides. As shown in Supplementary Fig. 2, in the absence of AN, the Cu oxides are reduced to metallic Cu under the same electrochemical operation. However, no reduction peak is observed when acetonitrile is present in the electrolyte. Throughout the activation procedure, the concentration of Cu2+ in the electrolyte initially increases and subsequently decreases (Fig. 1g), implying that Cu2+ is captured by the surface.

Electrochemical ARR to EA

A typical three-electrode H-type cell is employed to assess the electrochemical performance, with CuO@Cu heterostructure as the working electrode, Ag/AgCl as the reference electrode (the calibration curve of Ag/AgCl was shown in Supplementary Fig. 3), and Pt mesh as the counter electrode. As shown in Supplementary Fig. 4, the reduction of AN to EA occurs much more readily than hydrogen evolution. For comparison, the Cu foam and Cu2O foam after activation in 1 M NaOH solution without AN (oxide-derived Cu, OD-Cu) were also tested. Figure 2a shows the linear sweep voltammetry (LSV) curves of the CuO@Cu heterostructure, Cu foam, and OD-Cu in the Ar saturated 1 M NaOH electrolyte containing 12 wt% acetonitrile (LSV measurements without iR compensation are provided in Supplementary Fig. 5, and the values of the compensated solution resistance are detailed in Supplementary Table 1). CuO@Cu heterostructure exhibits significantly higher current density than Cu foam and OD-Cu, demonstrating a greatly faster electron transfer on the catalyst surface. During the electrochemical measurements, abundant small bubbles generated on the surface of Cu foam, indicating the evolution of hydrogen and a resulted low selectivity for EA (Supplementary Fig. 6). For OD-Cu, the apparent evolution of hydrogen was also observed at high current density. On the contrast, no visible bubble was observed on CuO@Cu during the test. The CuO nanoparticles was also tested to compare with the CuO@Cu heterostructure for acetonitrile reduction. As shown in Supplementary Fig. 7, CuO@Cu heterostructure exhibits much higher ARR activity than that of CuO nanoparticles. Tafel slopes are analyzed to investigate the electrochemical kinetics. As shown in Fig. 2b, the Tafel slope of CuO@Cu is 127.1 mV dec−1, smaller than those of Cu foam (158.4 mV dec−1) and OD-Cu (163.5 mV dec−1), indicating a more rapid catalytic kinetics. In addition, we evaluate the electrochemical active surface area (ECSA) by double-layer capacitances (Cdl) based on CV cycling with different scan rates in the non-Faradic region. As depicted in Fig. 2c, the Cdl value of CuO@Cu is 43.6 µF cm−2, larger than those of Cu foam (36.1 µF cm−2) and OD-Cu (35.3 µF cm−2), suggesting a higher ECSA related to more accessible sites on CuO@Cu. To quantitatively evaluate the FE for EA, 1H nuclear magnetic resonance (1H-NMR) and gas chromatography (GC) were employed that detect the liquid products and gaseous products, respectively. According to the 1H-NMR and GC results recorded at different potentials, the FE for EA on CuO@Cu is consistently maintained over 95% in the whole potential windows from 0.237 V to −0.462 V vs. RHE (Fig. 2d), whereas OD-Cu and Cu foam show apparently lower FE values, especially at high potential region due to the competition of HER (Supplementary Fig. 8 and Supplementary Fig. 9). These results demonstrate that CuO@Cu heterostructure can selectively catalyze ARR to EA. To clarify the impact of pH on the electroreduction of acetonitrile, we investigated several aqueous electrolyte solutions with varying pH values. The results are displayed in Supplementary Fig. 10. The ARR activity of the catalyst increases with the concentration of NaOH in the electrolyte, indicating that a higher pH is favorable for the ARR process. To examine the influence of acetonitrile concentration on ARR, we investigated the catalytic performance of CuO@Cu in 1 M NaOH solutions with acetonitrile contents of 8 wt%, 12 wt%, and 16 wt%. As illustrated in Supplementary Fig. 11, the activity and selectivity for EA are nearly independent of the acetonitrile concentration in the studied range.

Fig. 2: Electrochemical measurements.
figure 2

a LSV plots of CuO@Cu, OD-Cu, and Cu foam in 1 M NaOH containing 12 wt% acetonitrile; the potentials in (a, b, d and e) are provided after 100% iR correction, the compensated solution resistance are detailed in Supplementary Table 1; b Tafel slops of CuO@Cu, OD-Cu and Cu foam; c Cdl values of CuO@Cu, OD-Cu and Cu foam; d FEEA at different potentials in CuO@Cu; e Chronoamperometry test of CuO@Cu at current density of 200 mA cm−2 in H-type cell. Source data for Fig. 2 are provided as a Source Data file.

Full size image

The stability of CuO@Cu during the acetonitrile electroreduction was investigated using an H-type cell at a constant current density of 0.2 A cm−2 in 1 M NaOH electrolyte containing 12 wt% acetonitrile. During the 1000-h stability test, the ratio of peak area of EA to that of internal standard tetramethyl silane (TMS) remains basically unchanged (Supplementary Fig. 12), suggesting that the high EA selectivity is maintained. After 1000-h stability test, FEEA determined by 1H NMR experiment is still above 95% (Fig. 2e) and the potential remains stable as well. There is no notable variation in the LSV curves of CuO@Cu before and after stability test (Fig. 3a). Furthermore, the color of the CuO@Cu electrode before and after the 1000-h test at 0.2 A cm−2 remains unchanged (black), which corresponds to the color of Cu oxide (Supplementary Fig. 13). In contrast, when the CuO@Cu electrode undergoes HER operations (without acetonitrile in the electrolyte) at the same current density for 12 h, its surface turns reddish-orange, indicative of metallic Cu formation (Supplementary Fig. 14). In addition, structural characterizations have been conducted after the 1000-h stability test. The morphology of CuO@Cu remains virtually unchanged (Supplementary Fig. 13), and the XRD patterns before and after stability test are almost identical as well (Fig. 3b). The TEM images confirm the coexistence of CuO and the amorphous Cu domain (Supplementary Fig. 15). In Fig. 3c, we can obverse that after 1000-h chronoamperometry measurement, the Cu+ is negligible and Cu0 valence peak became more obvious, suggesting a more complete transformation of Cu2O to CuO@Cu. The peak intensities of Cu0 and Cu2+ after the 1000-h test are much closer to each other, consistent with the theoretical mole ratio of 1:1. We also performed in situ and operando X-ray absorption spectroscopy (XAS) under electrochemical operating conditions to further investigate the variation in Cu valence of the CuO@Cu catalysts. The Cu K-edge energy position was determined from the first peak in the derivative of the XAS spectrum. As shown in Supplementary Fig. 16, the energy position of the Cu K-edge is not shifted during electrochemical operation, indicating that the Cu valence remained unchanged. Therefore, CuO@Cu heterostructure is stable over the entire stability test. These results suggest the high stability of CuO@Cu for ARR to EA, which outperforming other reported catalysts (Supplementary Table 2). Encouraged by the activity and stability, we fabricated an electrolyzer. Figure 3d illustrates the schematic diagram of the overall electrochemical reactor, whereby the NiFe layered double hydroxide (NiFe-LDH) supported on foam nickel serves as the anode for oxygen evolution reaction, and the CuO@Cu catalyst functions as the cathode for ARR. The anode is flowed by 1 M NaOH solution, whereas the cathode is flowed by 1 M NaOH solution containing 12 wt% acetonitrile. The anode and cathode are separated by Nafion 117 proton exchange membrane. Nafion 117 was used for its relatively high stability in the presence of AN. On the anode side, water dissociates to generate oxygen. On the cathode side, acetonitrile receives protons to generate EA. Simultaneously, a portion of the protons coupled with electrons to produce H2. As shown in Fig. 3e, FE of EA can remain above 90% even after a test up to 100 h at a high current density of 0.5 A cm−2. Moreover, the CuO@Cu catalyst in the flow cell maintains its original morphology in contact with acetonitrile, whereas the catalyst in contact with the electrode plate shows the brownish-red color of metallic copper, further proving that the surface CuO is not reduced during the ARR process (Supplementary Fig. 17).

Fig. 3: Stability characterizations.
figure 3

a LSV plots of CuO@Cu and CuO@Cu after 1000-h stability test; the potentials in (a) are provided after 100% iR correction; b XRD patterns of Cu2O/Cu, CuO@Cu and CuO@Cu-after 1000 h; c L3M45M45 Auger spectra of Cu2O/Cu, CuO@Cu and CuO@Cu-after 1000 h; d Schematic diagram of the liquid flow cell; e Chronoamperometry test of CuO@Cu at 0.5 A cm−2 in liquid flow cell; the potentials in (e) are provided without iR correction. Source data for (a–c) and e are provided as a Source Data file.

Full size image

Insight into the catalytic mechanism

As depicted in Fig. 2a, Supplementary Fig. 6, and Supplementary Fig. 9, HER is the primary reaction on metallic Cu. Hence, the catalytic species for ARR could be the CuO component. To elucidate the reaction steps involved in the ARR process, the EIS spectra of CuO@Cu in 1 M NaOH with and without acetonitrile are both collected21. The corresponding Nyquist plots at different potentials are compared in Fig. 4a, d. The depressed semicircles of the Nyquist plots suggest a charge-transfer process, whose radius reflects the charge-transfer resistance. With the applied potential increasing, the radius of the semicircle decreases, demonstrating an accelerated kinetics of electrochemical reaction. Corresponding Bode plots are depicted in Fig. 4b, e. In the case without AN, only HER needs to be considered. At 0.154 V and 0.004 V vs. RHE, HER cannot occur. The reduced phase angle derived from the comparison of Bode plots is due to underpotentially deposited hydrogen (Hupd) related to the Volmer step (H2O + M + e− → M-H + OH−)22. The rapid consumption of the reductive protons can significantly contribute to the high stability of CuO@Cu under high current density. The Tafel slope for HER is calculated as 129 mV dec−1 (Supplementary Fig. 18), indicating the Volmer step as the rate-determining step23. There is only one phase angle peak in the Bode plots, which is associated with the Volmer step. Thus the electron-transferring Heyrovsky step (M-H + H2O + e− → M + H2 + OH−) should be absent. On the other hand, the Tafel step (M-H + M-H → M + H2) does not involve electron transfer, thus it is not discernible in the Bode plots. Therefore, HER on CuO@Cu follows Volmer-Tafel mechanism. In comparison, when acetonitrile is added to the electrolyte, a new peak at lower frequency emerges (Fig. 4e), which should denote to ARR process. By comparing the frequencies and evolutions of phase-angle peaks with and without acetonitrile, the electrochemical reaction corresponding to the almost unchanged phase angle at ~1 Hz is related to the Volmer step. Since the following generation of H2 is inhibited, the surface hydrogen generated in the Volmer step may transfer to the acetonitrile molecule adsorbed at the adjacent site to form CH3CHxN species. The phase-angle peak at lower frequency is supposed to correspond to hydrogenation of CH3CHxN species. Distribution of relaxation times (DRT) analysis is further employed to directly examine the frequency and resistance shift of electrochemical process15. Each peak in the DRT spectra stands for a sub-reaction in ARR or HER. The peak frequency represents the characteristic frequency of the corresponding sub-reaction, and the integral area of each peak is related to the resistance. The relaxation time τ is calculated as a function of the frequency, i.e., τ = 1/f. As shown in Fig. 4c, for HER process of CuO@Cu, only one DRT peak is observed, corresponding to the Volmer step. For ARR process, two peaks are observed (Fig. 4f), suggesting the existence of two different steps. The peak at τ = 100 s is close to the one observed in HER process, which we can ascribe to the Volmer step. So, the peak at τ = 1 s is attributed to the hydrogenation of CH3CHxN species. These DRT results further support the inferences obtained from Bode plots. To analyze adsorbed intermediates and products, in situ attenuated total internal reflectance Fourier transform infrared spectroscopy (ATR-FTIR) was utilized (Supplementary Fig. 19). We established the system spectrum under open circuit conditions as a baseline, monitoring spectral changes over time. In Fig. 4g, the peaks at 1133, 1241, 1297 cm−1 suggest the presence of C-NHx, according to previous study15 and our theoretical calculation (Supplementary Table 3). As the hydrogenation proceeds, the C-N bond is weakened and the absorption peak is generally red-shifted. We attribute the 1241 cm−1 and 1297 cm−1 to C-N bond stretching of the initial hydrogenated specie. The band at 1133 cm−1 took longer to reach their maximum intensity is attributed to C-NH2 in EA, also confirmed by the simulation. The emergence of a negative peak at 2273 cm−1 corresponds to the triple bond of acetonitrile, suggesting its consumption. A very broad peak at 3000–3400 cm−1 can be clearly observed. This peak cannot be solely attributed to the ν (N-H) which should be much sharper24. We attribute this broad to the overlap of ν (N-H) and ν (O-H). This peak intensifies over time, indicating the progression of the Volmer step11,15. In addition, we have extended the electrocatalytic hydrogenation process to other nitriles, including butyronitrile, pentanenitrile, and cyclopropanecarbonitrile. For all these substrates, the selectivity towards the primary amine remains at 100%, with no detectable formation of secondary or tertiary amines (Supplementary Fig. 20 and Supplementary Table 4).

Fig. 4: EIS measurements and in situ spectra analysis.
figure 4

a Nyquist plots of CuO@Cu at different potentials in 1 M NaOH; b Bode plots of CuO@Cu at different potentials in 1 M NaOH; c DRT plots of CuO@Cu at different potentials in 1 M NaOH; d Nyquist plots of CuO@Cu at different potentials in 1 M NaOH with 12 wt% acetonitrile; e Bode plots of CuO@Cu at different potentials in 1 M NaOH with 12 wt% acetonitrile; f DRT plots of CuO@Cu at different potentials in 1 M NaOH with 12 wt% acetonitrile; g in situ ATR-FTIR characterization of the ARR at −0.146 V. Source data for Fig. 4 are provided as a Source Data file.

Full size image

To dive deep into the enhanced performance on electrocatalytic ARR, DFT calculations were carried out. A Cu0/Cu2+ interfacial model, i.e., CuOx, was constructed based on CuO matrix (Supplementary Fig. 21). According to the charge analysis, one can see that the Cu sites in the partially reduced surface are in various states, including Cu0, Cuδ+ (0 < δ < 1) and Cu2+ (Supplementary Fig. 22). In comparison, the reaction process on Cu (111) surface was simulated. As shown in Fig. 5a, unsaturated bond in acetonitrile undergoes hydrogenation reactions, with the hydrogen transferring from aqueous electrolyte or hydrogen atoms adsorbed on the catalyst surface. Since the electrochemical reaction takes place under the alkaline condition, H2O molecule serves as the proton donor in electrolyte, instead of hydrated proton. All possible intermediates and reaction pathways were first simulated according to computational hydrogen electrode (CHE) model, as shown in Supplementary Fig. 23. The thermodynamically favorable pathways on CuOx and metallic Cu surfaces were screened out. Considering an external potential of −0.4 V (vs. RHE), the kinetic barriers on these two thermodynamically favorable paths were calculated to examine whether the hydrogen atom in each hydrogenation step is from *H or proton in H2O, as shown in Supplementary Figs. 24, 25. The thermodynamics energy changes and favorable kinetics barriers of ARR towards EA are summarized in Fig. 5b, and the structures of transition states are shown in Fig. 5c, d. Acetonitrile is adsorbed on the surface (*CH3CN) by nitrogen, thus the nitrogen site is passivated at first. For Cu (111), the activation of *CH3CN to form *CH3CHN should overcome a barrier of 0.91 eV, in which the hydrogen migrates from water to the carbon in the cyano group. The following reduction of *CH3CHN undergoes a much lower barrier. The subsequent hydrogenation of *CH3CHNH towards *CH3CH2NH is the rate-determining step (RDS) of the whole reduction process on Cu (111) with a barrier of 1.03 eV. Reduction of *CH3CH2NH has a lower barrier, and the *CH3CH2NH2 desorption is facile with an energy change of 0.25 eV. On CuOx surface, ARR takes place on Cu0 and Cuδ+ sites (0 < δ < 1) (Fig. 5c, Supplementary Fig. 22, Supplementary Fig. 27). Though the CH3CN adsorption is slightly weakened, the activation towards *CH3CHN is significantly promoted with a barrier of 0.53 eV, in which surface hydrogen addition is preferred. Following this step, *CH3CHN is reduced via *CH3CHNH and *CH3CH2NH intermediates to *CH3CH2NH. The hydrogenation of *CH3CHNH to *CH3CH2NH is the RDS with a barrier of 0.77 eV. The desorption energy of *CH3CH2NH2 is 0.29 eV, remaining a facile process at room temperature. Overall, CuOx shows preferable affinity for intermediates than Cu (111). Furthermore, CuOx shows a lower kinetic barrier of RDS than Cu (111) (0.77 eV vs. 1.03 eV). For all the hydrogen steps on Cu and CuOx surfaces during the simulation, hydrogen atoms transfer from the surface-adsorbed state to carbon and from water to nitrogen, except the first step on Cu surface. It is understandable that the positively charged C atoms may prefer the surface *H with partially negative charge, while the negatively charged N atom is attractive to the positively charged proton in H2O (Supplementary Table 5). Similar situation has been reported in electrochemical reduction of CO2, where the hydrogenation of C atoms favors the surface hydrogen, while the hydrogenation of O atoms prefers the proton from aqueous H2O25,26. The transfer of hydrogen from a polar water molecule to carbon may lead to suppressed activation of acetonitrile, whereas the hydrogen migration on CuOx surface could expedite this step substantially. It suggested that the synergistic hydrogenation process involving both surface hydrogen migration and proton addition from aqueous water plays an important role in the electroreduction of acetonitrile. Moreover, due to the stronger *H adsorption compared with Cu (111), the competitive HER should be less favored on CuOx, (Supplementary Fig. 26), potentially enhancing the selectivity of acetonitrile reduction.

Fig. 5: Mechanism illustration and theoretical calculations.
figure 5

a A simplified, atomic-scale illustration of the hydrogenation process for acetonitrile on the CuOx surface; b Free energy diagrams for acetonitrile electroreduction along the favorable pathway over Cu (111) and CuOx at an external potential U = − 0.4 V (vs. RHE, pH = 14). The hydrogenation process via aqueous H2O pathway and surface adsorbed hydrogen pathways are depicted in orange and blue, respectively; Optimized structures of transition states (TS) for c CuOx and d Cu (111). The transferred hydrogen atoms in the hydrogenation steps were highlighted. The atoms in orange, white, red, gray and cyan colors represent copper, hydrogen, oxygen, carbon, and nitrogen atoms, respectively. The atomic coordinates for the adsorption structures are listed in Supplementary Data 1. Source data for (b) are provided as a Source Data file.

Full size image

Discussion

In summary, we demonstrate a self-supported CuO@Cu heterostructure for the stable electrochemical reduction of acetonitrile to EA over a 1000-h period, achieving an EA faradaic efficiency of nearly 100% at a high current density of 0.2 A cm−2. CuO serves as the catalytic site, with metallic Cu species significantly enhancing electron transfer. The migration of hydrogen atoms generated during the Volmer step to the adjacent site where acetonitrile is adsorbed is identified as the root cause of the remarkably high faradaic efficiency for EA and the enhanced acetonitrile reduction activity. Additionally, DFT calculations reveal the effective transfer of hydrogen atoms from surface-adsorbed states to the carbon atoms of acetonitrile, consistent with experimental observations, and suggest a hydrogen transfer process from water to the nitrogen atoms of acetonitrile. This study presents a strategy for the efficient electrochemical reduction of organic molecules with high selectivity and activity. Future investigations could focus on fine-tuning the electronic properties of catalytic sites and designing catalytic interfaces to facilitate the Volmer process in the electrochemical reduction of organic molecules.

Methods

Materials

All chemicals were purchased and used without purification. The copper foam (CF, 99.8%) and Ni foam (NF, 99.8%) were bought from Sinerno. The sodium hydroxide (NaOH, 98%) were bought from Aladdin. The Nafion 117 membrane was supplied by Fuel Cell Store. Ethanol, acetone, and acetonitrile were procured from Sinopharm. Chemical Reagent Co., Ltd.

Synthesis of CuO@Cu heterostructure

The copper foam was cleaned with acetone and ethanol, each repeated three times, before usage. The cleaned copper foam was heated in a muffle furnace at a temperature of 350 °C for a duration of 2 h, resulting in the formation of Cu2O/Cu. Subsequently, the Cu2O/Cu was immersed in a 1 M NaOH solution containing 12 wt% acetonitrile and underwent CV-activation within a potential range of 0.237 V vs. RHE to −0.462 V vs. RHE. The surface of Cu2O underwent reconstruction and ultimately yielded CuO. The synthesis of CuO@Cu-450 and CuO@Cu-550 is the same as the above synthesis steps, except for the change in annealing temperature to 450 °C and 550 °C.

Synthesis of OD-Cu

In situ electrochemical reduction of annealed Cu2O/Cu in 1 M NaOH solution for 6 h at −200 mA cm−2 to obtain OD-Cu.

Characterization

The morphologies and microstructural information were examined using a Hitachi S-4800 field-emission scanning electron microscopy (SEM), and Transmission Electron Microscopy (TEM, Tecnai F20, JEM-ARM200F, FEI, America). X-ray powder diffraction (XRD) analysis was conducted using a D8-Advance Davinci diffractometer with Cukα (λ = 1.5418 Å) radiation at room temperature. The surface chemical states of the samples were analyzed using X-ray photoelectron spectroscopy (XPS) with Axis SUPRA Kratos.

In situ surface-enhanced attenuated total internal reflectance Fourier transform infrared spectroscopy (in situ SEATR-FTIR)

In situ SEATR-FTIR measurements were performed using a Linglu Instruments ECIR-II cell mounted on a Nicolet-6700 Fourier transform infrared spectrometer. The cell featured a single-bounce silicon crystal coated with an Au membrane (Supplementary Fig. 19), operated in internal reflection mode. Chronoamperometry (CA) experiments were monitored using a CHI 760E electrochemical workstation. The in situ electrochemical FT-IR cell was configured as a three-electrode system, with an Ag/AgCl electrode as the reference, a Pt wire as the counter electrode, and the CuO@Cu catalyst as the working electrode. The CuO@Cu catalyst was carefully sonicated from the electrode and loaded on the Au membrane as working electrode. The electrolyte consisted of 1 M NaOH with 12 wt% acetonitrile. The spectra were first recorded at open circuit potential (OCP), and this point was set as 0 min. Then the potential was held −0.146 V vs. RHE and maintained 3000 s by CA method. The in situ FTIR was recorded at 10, 100, 200, 300, 500, 800, 1000, 1500, 2000, 3000 s during the CA method.

In situ and operando XAS measurement

The X-ray absorption spectra (XAS) measurements for the Cu K-edge were conducted in transmission mode (fluorescence mode) on the Laboratory-Based XAFS and XES Spectrometer (easyXAFS300+, USA). The energy was calibrated using corresponding metal foil as references. XANES and EXAFS data reduction and analysis were analyzed by Athena 3 and Artemis software. For the measurement, the CuO@Cu catalyst was carefully sonicated from the electrode and then collected. Cu XAS was first recorded at OCP, followed by XAS during CV cycling in the range of −0.8 V vs. RHE to −1.2 V vs. RHE. The sample was then subjected to chronoamperometric testing at −1.0 V vs. RHE for 1 h, and XAS was collected again under these conditions.

Electrochemical measurements

The electrochemical performance of the prepared catalysts was evaluated using a CHI 760E conventional three-electrode system from CH Instruments, Inc. Shanghai, in an H-type cell in a 1 M NaOH containing 12 wt% acetonitrile electrolytes at room temperature. The electrolyte was prepared just before use. The Ag/AgCl electrode served as the reference electrode, and a platinum net was used as the counter electrode. The CuO@Cu and Cu/CF were employed as the working electrodes. The measured potentials were subsequently converted to the RHE scale as described below:

$${E}_{{{\rm{RHE}}}}={E}_{{{\rm{Ag}}}/{{{{\rm{Ag}}}{{{\rm{Cl}}}}}}}+0.059\times {{\rm{pH}}}+0.210\,$$
(2)

Prior to testing, all working electrodes were conditioned by cyclically sweeping within the potential range from 0.237 V versus the reversible hydrogen electrode (vs. RHE) to −0.462 V vs. RHE at a scan rate of 50 mV s−1 until reproducible cyclic voltammograms (CV) were achieved. The electrocatalysts underwent 50 cycles of CV prior to performing the linear sweep voltammetry (LSV) curve at a scan rate of 10 mV s−1. Electrochemical impedance spectroscopy (EIS) measurements were carried out within the frequency range of 100 kHz to 0.01 Hz. The distribution of relaxation times of EIS was conducted by DRT tools to separate the anode and cathode reactions with high resolution27,28,29.

Faradaic efficiency measurements

The resulting liquid product was analyzed using 1H NMR spectroscopy (Bruker 600 MHz spectrometer) with water suppression. Typically, 400 μL of electrolyte after ARR was mixed with 200 μL of MeOD containing 0.03 vol% TMS as an internal standard. The FE of the ethylamine was calculated using the following formula.

$${{FE}}\left(\%\right)=\frac{nzF}{Q}\times 100\%$$
(3)

where n is the amount of product in moles, z is the number of electrons transferred (the value of z is 4 for ARR to ethylamine), F is the Faraday constant, Q is the charge.

The FE of H2 is calculated using the following formula:

$${{FE}}\left(\%\right)=\frac{nfCFP}{IRT}\times 100\%$$
(4)

where n is the number of electrons transferred to H2 formation (the value is 2), f is the gas flow rate 0.33 mL s−1), C is the measured concentration of the product by GC (in ppm), F is the Faraday constant (96485 C mol−1), P is the pressure (1.01 × 105 Pa), I is the imposed current (in A), T is the reaction temperature (298 K); R is ideal gas constant (8.314 J mol−1 K−1).

Chronopotentiometry test

Chronopotentiometry experiments were performed in an H-type cell to evaluate the performance of acetonitrile electroreduction using a three-electrode system. The Ag/AgCl electrode was employed as the reference electrode, and a platinum net was used as the counter electrode. The electrolyte was refreshed every 12 h. For each data point of Faraday efficiencies, 400 uL of liquid product was collected at intervals and the corresponding time was recorded. The measured potentials were corrected using iR compensation to account for the voltage drop caused by solution resistance.

Flow cell measurement

A two-compartment microfluidic flow cell electrolyzer was fabricated for ARR electrolysis. NiFe LDH (1 cm × 1 cm) and CuO@Cu (1 cm × 1 cm) were employed as the anode and cathode, respectively. The NiFe-LDH used as anode in the stability test was prepared according to a method reported in previous literature30. Typically, Ni foams of 1 cm × 1 cm were treated with ultrasound for 5 min in acetone, ethanol and 0.1 M hydrochloric acid, respectively. Then, the Ni foams were rinsed with a large amount of deionized water. Subsequently, the Ni foams were immersed in a 0.2 M Fe(NO3)3 solution for 60 s. After the submersion, the Ni foams were dried in air. A Nafion 117 membrane was used to separate the anode and cathode. The catholyte consisted of 1 M NaOH with 12 wt% acetonitrile, while the anolyte was 1 M NaOH. The flow rate for both catholyte and anolyte was 10 mL min−1.

Computational details

The Density Functional Theory (DFT) calculations presented were executed with the Vienna Ab-initio Simulation Package (VASP)31,32, employing the projector-augmented wave (PAW) method33,34. The generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof exchange-correlation functional35 was used to describe electronic exchange and correlation. The water environment was treated implicitly using the VASPsol code with a dielectric constant of 78.436,37. To identify the transition states, the VTST package38 with climbing image nudged elastic band (CI-NEB)39 combined with the improved dimer method (IDM)40 was employed including solvation. The energy cutoff was set as 600 eV. The convergence criteria were set as 0.05 eV/Å for force on each relaxed atom and 10−4 eV and 10−7 eV for energy in adsorption structure optimization and CI-NEB simulations, respectively.

A five-layer 4 × 4 Cu (111) slab was employed with the top two Cu layers relaxed. Based on the experimental XRD, TEM, and Auger spectroscopy characterizations, a partially reduced CuOx slab model was constructed to represent the CuO@Cu catalyst. A 4 × 4 CuO (001) supercell was used, in which all the surface oxygen atoms and half of the subsurface oxygen atoms were removed. The optimized CuOx slab (Supplementary Fig. 21) maintains CuO (001) matrix, while there exhibits Cu0 and Cu2+ with a ratio close to 1:1 on the surface, validated by the Bader charge analysis shown in Supplementary Fig. 22. Bader charge analysis in Supplementary Fig. 22 indicates that this model comprises Cu0, Cu2+, and Cuδ+ whose valence is between 0 and +1. The top two layers were relaxed during structure optimization. The vacuum height was set around 12 Å to prevent the inter-layer interactions in the periodic model. A 3 × 3 × 1 and 2 × 2 × 1 k-point grid generated with the Monkhorst–Pack scheme41 was used for Cu (111) and CuOx, respectively. For the CuOx, the spin polarization was considered due to its nonnegligible spin effects, and a Hubbard correction of U = 5.0 eV was applied to the d-orbitals of Cu for a better description of the properties of Cu oxides42,43.

Electrochemical reaction thermodynamics were benchmarked using the CHE due to its thermodynamic rigor and wide utilization in electrocatalysis44. The thermal correction to Gibbs free energy was calculated by VASPKIT45. For the barrier calculation, H2O was chosen as the proton donor according to the alkaline conditions. Considering that OH− ions need at least 3–4 H2O molecules for solvation, an explicit hydrogen-bonded water cluster comprising 6 H2O molecules was considered during searching for the transition state for the aqueous H2O pathway. All the water molecules were relaxed during the transition state search. While for the surface-adsorbed hydrogen pathway, the explicit water cluster was not included as the H2O does not participate in the reaction process and should have little effect on the barrier. Barriers from various water cluster structures were investigated, and the lowest of which should dominate the activity. A sodium atom was added to help simulating the reductive potential. The charge extrapolation method46,47 developed by Karen Chan et al. was applied to deduce the potential-dependent activation barriers. The combination of CHE model and the charge extrapolation method has garnered validation as an effective tool for calculating free energy diagrams of electrocatalytic processes, like HER and CO2RR48,49. When it involves hydrogenation via surface pathway, we assume that the barrier is independent of the electrode potential, premised on that there is no excess electron transfer during this process from the perspective of the charge extrapolation method.