Thank you for visiting nature.com. The browser version you are using has limited CSS support. For the best experience, we recommend using a newer browser (or disabling compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we will display the site without styles and JavaScript.
       Electrosynthesis of adipic acid (a precursor of nylon 66) from CA oil (a mixture of cyclohexanone and cyclohexanol) is a sustainable strategy that can replace traditional methods that require harsh conditions. However, low current density and competing oxygen evolution reactions significantly limit its industrial applications. In this work, we modify nickel double hydroxide with vanadium to enhance the current density and maintain high faradaic efficiency (>80%) over a wide potential range (1.5–1.9 V vs. reversible hydrogen electrode). Experimental and theoretical studies revealed two key roles of V modification, including accelerated catalyst reconstruction and improved cyclohexanone adsorption. As a proof of concept, we constructed a membrane-electrode assembly that produced adipic acid with high faradaic efficiency (82%) and productivity (1536 μmol cm-2 h-1) at an industrially relevant current density (300 mA cm-2), while achieving stability >50 h. This work demonstrates an efficient catalyst for the electrosynthesis of adipic acid with high productivity and industrial potential.
       Adipic acid (AA) is one of the most important aliphatic dicarboxylic acids and is mainly used in the production of nylon 66 and other polyamides or polymers1. Industrially, AA is synthesized by oxidizing a mixture of cyclohexanol and cyclohexanone (i.e., AA oil) using 50–60 vol% nitric acid as an oxidizing agent. This process has environmental concerns related to the emission of concentrated nitric acid and nitrogen oxides (N2O and NOx) as greenhouse gases2,3. Although H2O2 can be used as an alternative green oxidizing agent, its high cost and harsh synthesis conditions make it difficult to be practically applied, and a more cost-effective and sustainable method is needed4,5,6.
       Over the past decade, electrocatalytic chemical and fuel synthesis methods have attracted increasing attention from scientists due to their advantages of using renewable energy and operating under mild conditions (e.g., room temperature and ambient pressure)7,8,9,10. In this regard, the development of electrocatalytic conversion of KA oil to AA is very important to obtain the above advantages as well as to eliminate the use of nitric acid and nitrous oxide emissions encountered in conventional production (Figure 1a). Pioneering work was done by Petrosyan et al., who reported the electrocatalytic oxidation reaction of cyclohexanone (COR; cyclohexanone or cyclohexanol have been commonly studied as representing KA oil) on nickel oxyhydroxide (NiOOH), but low current density (6 mA cm-2) and moderate AA yield (52%) were obtained11,12. Since then, significant progress has been made in the development of nickel-based catalysts to enhance COR activity. For example, a copper-doped nickel hydroxide (Cu-Ni(OH)2) catalyst was synthesized to promote Cα–Cβ cleavage in cyclohexanol13. We recently reported a Ni(OH)2 catalyst modified with sodium dodecyl sulfonate (SDS) to create a hydrophobic microenvironment that enriches cyclohexanone14.
       a The challenges of AA production by electrooxidation of KA oil. b Comparison of electrocatalytic COR of previously reported Ni-based catalysts and our catalyst in three-electrode system and flow battery system11,13,14,16,26. Detailed information on the reaction parameters and reaction performance are provided in Supplementary Tables 1 and 2. c Catalytic performance of our NiV-LDH-NS catalysts for COR in H-cell reactor and MEA, which operate over a wide potential range.
       Although the above methods improved the COR activity, the described Ni-based catalysts showed high AA Faraday efficiency (FE) (>80%) only at relatively low potentials, typically below 1.6 V compared to the reversible hydrogen electrode (RHE, abbreviated VRHE). Thus, the reported partial current density (i.e., the total current density multiplied by FE) of AA is always below 60 mA cm−2 (Figure 1b and Supplementary Table 1). The low current density is far below the industrial requirements (>200 mA cm−2)15, which significantly hampers the electrocatalytic technology for high-throughput AA synthesis (Figure 1a; top). To increase the current density, a more positive potential (for the three-electrode system) or a higher cell voltage (for the two-electrode system) can be applied, which is a simple approach for many electrocatalytic transformations, especially the oxygen evolution reaction (OER). However, for COR at high anodic potentials, OER can become a major competitor in reducing the FE of AA, thereby reducing the energy efficiency (Figure 1a; bottom). For example, reviewing the previous progress (Figure 1b and Supplementary Table 1), we were disappointed to find that the FE of AA on SDS-modified Ni(OH)2 decreased from 93% to 76% with increasing the applied potential from 1.5 VRHE to 1.7 VRHE14, while the FE of AA on CuxNi1-x(OH)2/CF decreased from 93% to 69% with increasing the potential from 1.52 VRHE to 1.62 VRHE16. Thus, the reported partial current density of AA does not increase proportionally at higher potentials, which largely limits the improvement of AA performance, not to mention the high energy consumption due to the low FE of AA. In addition to nickel-based catalysts, cobalt-based catalysts have also shown catalytic activity in COR17,18,19. However, their efficiency decreases at higher potentials, and compared with Ni-based catalysts, they have more potential limitations in industrial applications, such as greater price fluctuations and smaller inventories. Therefore, it is desirable to develop Ni-based catalysts with high current density and FE in COR to make it practical to achieve high AA yields.
       In this work, we report vanadium(V)-modified nickel layered double hydroxide nanosheets (NiV-LDH-NS) as efficient electrocatalysts for AA production via COR, which operate over a wide potential range with significantly suppressed OER, achieving high FE and current density in both H-cells and membrane electrode assemblies (MEAs; Figure 1 b). We first show that the acetylene oxidation efficiency over a typical Ni(OH)2 nanosheet catalyst (Ni(OH)2-NS) decreases, as expected, at higher potentials, from 80% at 1.5 VRHE to 42% at 1.9 VRHE. In sharp contrast, after modifying Ni(OH)2 with V, NiV-LDH-NS exhibited higher current density at a given potential and, more importantly, maintained high FE over a wide potential range. For example, at 1.9 VRHE, it showed a current density of 170 mA cm−2 and FE of 83%, which is a more favorable catalyst for COR in the three-electrode system (Fig. 1c and Supplementary Table 1). Experimental and theoretical data indicate that the V modification promotes the reduction kinetics from Ni(OH)2 to high-valent Ni oxyhydroxides (Ni3+xOOH1-x), which serve as the active phase for COR. Moreover, the V modification enhanced the adsorption of cyclohexanone on the catalyst surface, which played a key role in suppressing OER at high anodic potentials. To demonstrate the potential of NiV-LDH-NS in a more realistic scenario, we designed a MEA flow reactor and showed an FE of AA (82%) at an industrially relevant current density (300 mA cm−2), which is significantly higher than our previous results in a membrane flow reactor (Fig. 1b and Supplementary Table 2). The corresponding yield of AA (1536 μmol cm−2 h−1) was even higher than that obtained using the thermal catalytic process (<30 mmol gcatalyst−1 h−1)4. Furthermore, the catalyst showed good stability when using MEA, maintaining FE >80% AA for 60 h at 200 mA cm−2 and FE >70% AA for 58 h at 300 mA cm−2. Finally, a preliminary feasibility study (FEA) demonstrated the cost-effectiveness of the electrocatalytic strategy for AA production.
       According to the previous literature, Ni(OH)2 is a typical catalyst that shows good activity for COR, so Ni(OH)2-NS13,14 was synthesized for the first time via coprecipitation method. The samples showed β-Ni(OH)2 structure, which was confirmed by X-ray diffraction (XRD; Fig. 2a), and the ultra-thin nanosheets (thickness: 2–3 nm, lateral size: 20–50 nm) were observed by high-resolution transmission electron microscopy (HRTEM; Supplementary Fig. 1) and atomic force microscopy (AFM) measurements (Supplementary Fig. 2). Aggregation of the nanosheets was also observed due to their ultra-thin nature.
       a X-ray diffraction patterns of Ni(OH)2-NS and NiV-LDH-NS. FE, throughput, and AA current density on b Ni(OH)2-NS and c NiV-LDH-NS at different potentials. Error bars represent the standard deviation of three independent measurements using the same catalyst. d HRTEM image of NV-LDH-NS. Scale bar: 20 nm. HAADF-STEM image of NiV-LDH-NS and the corresponding elemental map showing the distribution of Ni (green), V (yellow), and O (blue). Scale bar: 100 nm. f Ni 2p3/2, g O 1 s, and h V 2p3/2 XPS data of Ni(OH)2-NS (top) and NiV-LDH-NS (bottom). i FE and j are the AA performances on the two catalysts over 7 cycles. Error bars represent the standard deviation of three independent measurements using the same catalyst and are within 10%. Raw data for a–c and f–j are provided in the raw data files.
       We then evaluated the effect of Ni(OH)2-NS on COR. Using constant potential electrolysis, we obtained 80% FE of AA at low potential (1.5 VRHE) without OER (Figure 2b), indicating that COR is energetically more favorable than OER at low anodic potentials. The main by-product was found to be glutaric acid (GA) with an FE of 3%. The presence of trace amounts of succinic acid (SA), malonic acid (MA), and oxalic acid (OA) was also quantified by HPLC (see Supplementary Figure 3 for product distribution). No formic acid was detected in the product, suggesting that carbonate may be formed as a C1 by-product. To test this hypothesis, the electrolyte from complete electrolysis of 0.4 M cyclohexanone was acidified and the gaseous products were passed through a Ca(OH)2 solution. As a result, the solution became turbid, confirming the formation of carbonate after electrolysis. However, due to the low total electricity generated during the electrolysis process (Figure 2b, c), the concentration of carbonate is low and difficult to quantify. In addition, other C2-C5 products may also be formed, but their amounts cannot be quantified. Although the total amount of products is difficult to quantify, 90% of the total electrochemical equivalent indicates that most of the electrochemical processes have been identified, which provides a basis for our mechanistic understanding. Due to the low current density (20 mA cm−2), the yield of AA was 97 μmol cm−2 h−1 (Figure 2b), equivalent to 19 mmol h−1 g−1 based on the mass loading of the catalyst (5 mg cm−2), which is lower than the thermal catalytic productivity (~30 mmol h−1 g−1)1. When the applied potential increased from 1.5 to 1.9 VRHE, although the overall current density increased (from 20 to 114 mA cm−2), there was simultaneously a significant decrease in AA FE, from 80% to 42%. The decrease in FE at more positive potentials is mainly due to the competition for OER. Especially at 1.7 VRHE, the OER competition leads to a significant decrease in AA FE, thereby slightly reducing AA performance with increasing overall current density. Thus, although the partial current density of AA increased from 16 to 48 mA cm−2 and the AA productivity increased (from 97 to 298 μmol cm−2 h−1), a large amount of additional energy was consumed (2.5 W h gAA−1 more from 1.5 to 1.9 VRHE), resulting in an increase in carbon emissions of 2.7 g CO2 gAA−1 (calculation details are given in Supplementary Note 1). The previously noted OER as a competitor to the COR reaction at high anodic potentials is consistent with previous reports and represents a general challenge for improving AA productivity14,17.
       To develop a more efficient Ni(OH)2-NS-based COR catalyst, we first analyzed the active phase. We observed peaks at 473 cm-1 and 553 cm-1 in our in situ Raman spectroscopy results (Supplementary Fig. 4), corresponding to the bending and stretching of Ni3+-O bonds in NiOOH, respectively. It has been documented that NiOOH is the result of Ni(OH)2 reduction and Ni(OH)O accumulation at anodic potentials, and is essentially the active phase in electrocatalytic oxidation20,21. Therefore, we expect that accelerating the phase reconstruction process of Ni(OH)2 to NiOOH can enhance the catalytic activity of COR.
       We tried to modify Ni(OH)2 with different metals since it was observed that heteroatom modification promotes phase reconstruction in transition metal oxides/hydroxides22,23,24. The samples were synthesized by co-deposition of Ni and a second metal precursor. Among the different metal-modified samples, the V-modified sample (V:Ni atomic ratio 1:8) (called NiV-LDH-NS) showed higher current density in COR (Supplementary Fig. 5) and more importantly, high AA FE over a wide potential window. In particular, at low potential (1.5 VRHE), the current density of NiV-LDH-NS was 1.9 times higher than that of Ni(OH)2-NS (39 vs. 20 mA cm−2), and the AA FE was comparable on both catalysts (83% vs. 80%). Due to the higher current density and similar FE AA, the productivity of NiV-LDH-NS is 2.1 times higher than that of Ni(OH)2-NS (204 vs. 97 μmol cm−2 h−1), demonstrating the promoting effect of V modification on current density at low potentials (Figure 2c).
       With increasing applied potential (e.g., 1.9 VRHE), the current density on NiV-LDH-NS is 1.5 times higher than that on Ni(OH)2-NS (170 vs. 114 mA cm−2), and the increase is similar to that at lower potentials (1.9 times higher). Notably, NiV-LDH-NS retained the high AA FE (83%) and the OER was significantly suppressed (O2 FE 4%; Figure 2c), outperforming Ni(OH)2-NS and previously reported catalysts with much lower AA FE at high anodic potentials (Supplementary Table 1). Due to the high FE of AA in a wide potential window (1.5–1.9 VRHE), an AA generation rate of 867 μmol cm−2 h−1 (equivalent to 174.3 mmol g−1 h−1) was achieved at 1.9 VRHE, demonstrating favorable performance in electrocatalytic and even thermocatalytic systems when the activity was normalized by the total mass loading of the NiV-LDH-NS samples (Supplementary Fig. 6).
       To understand the high current density and high FE over a wide potential range after modifying Ni(OH)2 with V, we characterized the structure of NiV-LDH-NS. The XRD results showed that the modification with V caused a phase transition from β-Ni(OH)2 to α-Ni(OH)2, and no V-related crystalline species were detected (Fig. 2a). The HRTEM results show that NiV-LDH-NS inherits the morphology of the ultrathin Ni(OH)2-NS nanosheets and has similar lateral dimensions (Fig. 2d). AFM measurements revealed a strong aggregation tendency of the nanosheets, resulting in a measurable thickness of approximately 7 nm (Supplementary Fig. 7), which is larger than that of Ni(OH)2-NS (thickness: 2–3 nm). Energy-dispersive X-ray spectroscopy (EDS) mapping analysis (Figure 2e) showed that V and Ni elements were well distributed in the nanosheets. To elucidate the electronic structure of V and its effect on Ni, we used X-ray photoelectron spectroscopy (XPS) (Figure 2f–h). Ni(OH)2-NS exhibited the characteristic spin-orbit peaks of Ni2+ (female peak at 855.6 eV, satellite peak at 861.1 eV, Figure 2f)25. The O 1 s XPS spectrum of Ni(OH)2-NS can be divided into three peaks, among which the peaks at 529.9, 530.9 and 532.8 eV are attributed to the lattice oxygen (OL), hydroxyl group (Ni-OH) and oxygen adsorbed on surface defects (OAds), respectively (Figure 2g)26,27,28,29. After the modification with V, the V 2p3/2 peak appeared, which can be decomposed into three peaks located at 517.1 eV (V5+), 516.6 eV (V4+) and 515.8 eV (V3+), respectively, indicating that the V species in the structure mainly exist in high oxidation states (Figure 2h)25,30,31. In addition, the Ni 2p peak at 855.4 eV in NiV-LDH-NS was negatively shifted (by about 0.2 eV) compared to that in Ni(OH)2-NS, indicating that electrons were transferred from V to Ni. The relatively low valence state of Ni observed after V modification was consistent with the Ni K-edge X-ray absorption near-edge spectroscopy (XANES) results (see the “V Modification Promotes Catalyst Reduction” section below for more details). The NiV-LDH-NS after COR treatment for 1 h was designated as NiV-LDH-POST and was fully characterized using transmission electron microscopy, EDS mapping, X-ray diffraction, Raman spectroscopy, and XPS measurements (Supplementary Figs. 8 and 9). The catalysts remained as aggregates with ultrathin nanosheet morphology (Supplementary Fig. 8a–c). The crystallinity of the samples decreased and the V content decreased due to V leaching and catalyst reconstruction (Supplementary Fig. 8d–f). The XPS spectra showed a decrease in the V peak intensity (Supplementary Fig. 9), which was attributed to the V leaching. In addition, O 1s spectrum analysis (Supplementary Fig. 9d) and electron paramagnetic resonance (EPR) measurements (Supplementary Fig. 10) showed that the amount of oxygen vacancies on NiV-LDH-NS increased after 1 h of electrolysis, which may lead to a negative shift in the Ni 2p binding energy (see Supplementary Figs. 9 and 10 for more details)26,27,32,33. Thus, NiV-LDH-NS showed little structural change after 1 h of COR.
       To confirm the important role of V in promoting COR, we synthesized NiV-LDH catalysts with different V:Ni atomic ratios (1:32, 1:16, and 1:4, designated as NiV-32, NiV-16, and NiV-4, respectively) except 1:8 by the same coprecipitation method. The EDS mapping results show that the V:Ni atomic ratio in the catalyst is close to that of the precursor (Supplementary Fig. 11a–e). With the increase of V modification, the intensity of the V 2p spectrum increases, and the binding energy of the Ni 2p region continuously shifts to the negative side (Supplementary Fig. 12). At the same time, the proportion of OL gradually increased. The results of the catalytic test show that the OER can be effectively suppressed even after minimal V modification (V:Ni atomic ratio of 1:32), with the O2 FE decreasing from 27% to 11% at 1.8 VRHE after V modification (Supplementary Fig. 11f). With the increase of the V:Ni ratio from 1:32 to 1:8, the catalytic activity increased. However, with further increase of V modification (V:Ni ratio of 1:4), the current density decreases, which we speculate is due to the decrease in the density of Ni active sites (particularly the NiOOH active phase; Supplementary Fig. 11f). Due to the promoting effect of V modification and the preservation of Ni active sites, the catalyst with the V:Ni ratio of 1:8 showed the highest FE and AA performance in the V:Ni ratio screening test. In order to clarify whether the V:Ni ratio remains constant after electrolysis, the composition of the used catalysts was characterized. The results show that for the catalysts with initial V:Ni ratios from 1:16 to 1:4, the V:Ni ratio decreased to about 1:22 after the reaction, which may be due to the leaching of V due to catalyst reconstruction (Supplementary Fig. 13). Note that comparable AA FEs were observed when the initial V:Ni ratio was equal to or higher than 1:16 (Supplementary Fig. 11f), which may be explained by the catalyst reconstruction resulting in similar V:Ni ratios in the catalysts showing comparable catalytic performance.
       To further confirm the importance of V-modified Ni(OH)2 in enhancing the COR performance, we developed two other synthetic methods to introduce V into Ni(OH)2-NS materials. One is a mixing method, and the sample is referred to as NiV-MIX; the other is a sequential sputtering method, and the sample is referred to as NiV-SP. The details of the synthesis are provided in the Methods section. SEM-EDS mapping showed that V was successfully modified on the Ni(OH)2-NS surface of both samples (Supplementary Fig. 14). The electrolysis results show that at 1.8 VRHE, the AA efficiency on the NiV-MIX and NiV-SP electrodes is 78% and 79%, respectively, both showing higher efficiency than Ni(OH)2-NS (51%). Moreover, the OER on NiV-MIX and NiV-SP electrodes was suppressed (FE O2: 7% and 2%, respectively) compared to Ni(OH)2-NS (FE O2: 27%). These results confirm the positive effect of V modification in Ni(OH)2 on OER suppression (Supplementary Fig. 14). However, the stability of the catalysts was compromised, which was reflected by the decrease of FE AA on NiV-MIX to 45% and on NiV-SP to 35% after seven COR cycles, which implies the need to adopt appropriate methods to stabilize V species, such as V modification in the Ni(OH)2 lattice in NiV-LDH-NS, which is the key catalyst in this work.
       We also evaluated the stability of Ni(OH)2-NS and NiV-LDH-NS by subjecting COR to multiple cycles. The reaction was carried out for 1 h per cycle and the electrolyte was replaced after each cycle. After the 7th cycle, the FE and AA performance on Ni(OH)2-NS decreased by 50% and 60%, respectively, while an increase in OER was observed (Fig. 2i, j). After each cycle, we analyzed the cyclic voltammetry (CV) curves of the catalysts and observed that the oxidation peak of Ni2+ gradually decreased, indicating a decrease in the redox ability of Ni (Supplementary Fig. 15a–c). Along with the increase in the Ni cation concentration in the electrolyte during electrolysis (Supplementary Fig. 15d), we attribute the performance degradation (decreased FE and AA productivity) to the leaching of Ni from the catalyst, resulting in a greater exposure of the Ni foamed substrate that exhibits OER activity. In contrast, NiV-LDH-NS slowed down the decline in FE and AA productivity to 10% (Fig. 2i, j), indicating that the V modification effectively inhibited Ni leaching (Supplementary Fig. 15d). To understand the enhanced stability of the V modification, we performed theoretical calculations. According to previous literature34,35, the enthalpy change of the demetallization process of metal atoms on the active surface of the catalyst can be used as a reasonable descriptor to evaluate the catalyst stability. Therefore, the enthalpy changes of the demetallization process of Ni atoms on the (100) surface of the reconstructed Ni(OH)2-NS and NiV-LDH-NS (NiOOH and NiVOOH, respectively) were estimated (the details of the model construction are described in Supplementary Note 2 and Supplementary Fig. 16). The demetallization process of Ni from NiOOH and NiVOOH was illustrated (Supplementary Fig. 17). The energy cost of Ni demetallization on NiVOOH (0.0325 eV) is higher than that on NiOOH (0.0005 eV), indicating that the V modification enhances the stability of NiOOH.
       To confirm the OER inhibitory effect on NiV-LDH-NS, especially at high anodic potentials, differential electrochemical mass spectrometry (DEMS) was performed to investigate the potential-dependent O2 formation on different samples. The results showed that in the absence of cyclohexanone, O2 on NiV-LDH-NS appeared at an initial potential of 1.53 VRHE, which was slightly lower than that of O2 on Ni(OH)2-NS (1.62 VRHE) (Supplementary Fig. 18). This result suggests that the OER inhibition of NiV-LDH-NS during COR may not be due to its intrinsic low OER activity, which is consistent with the slightly higher current density in the linear sweep voltammetry (LSV) curves on NiV-LDH-NS than that on Ni(OH)2-NS without cyclohexanone (Supplementary Fig. 19). After the introduction of cyclohexanone, the delayed O2 evolution (possibly due to the thermodynamic advantage of COR) explains the high FE of AA in the low potential region. More importantly, the OER onset potential on NiV-LDH-NS (1.73 VRHE) is more delayed than that on Ni(OH)2-NS (1.65 VRHE), which is consistent with the high FE of AA and low FE of O2 on NiV-LDH-NS at more positive potentials (Figure 2c).
       To further understand the promoting effect of V modification, we analyzed the OER and COR reaction kinetics on Ni(OH)2-NS and NiV-LDH-NS by measuring their Tafel slopes. It is worth noting that the current density in the Tafel region is due to the oxidation of Ni2+ to Ni3+ during the LSV test from low potential to high potential. To reduce the effect of Ni2+ oxidation on the COR Tafel slope measurement, we first oxidized the catalyst at 1.8 VRHE for 10 min and then performed the LSV tests in the reverse scan mode, i.e., from high potential to low potential (Supplementary Fig. 20). The original LSV curve was corrected with 100% iR compensation to obtain the Tafel slope. In the absence of cyclohexanone, the Tafel slope of NiV-LDH-NS (41.6 mV dec−1) was lower than that of Ni(OH)2-NS (65.5 mV dec−1), indicating that the OER kinetics could be enhanced by the V modification (Supplementary Fig. 20c). After the introduction of cyclohexanone, the Tafel slope of NiV-LDH-NS (37.3 mV dec−1) was lower than that of Ni(OH)2-NS (127.4 mV dec−1), indicating that the V modification had a more obvious kinetic effect on COR compared with OER (Supplementary Fig. 20d). These results suggest that although the V modification promotes OER to some extent, it significantly accelerates the COR kinetics, resulting in an increase in FE of AA.
       In order to understand the promoting effect of the above V modification on the performance of FE and AA, we focused on the mechanism study. Some previous reports have shown that heteroatom modification can reduce the crystallinity of catalysts and increase the electrochemically active surface area (EAS), thereby increasing the number of active sites and thus improving the catalytic activity36,37. To investigate this possibility, we conducted ECSA measurements before and after electrochemical activation, and the results showed that the ECSA of Ni(OH)2-NS and NiV-LDH-NS were comparable (Supplementary Fig. 21), excluding the influence of the active site density after V modification on the catalytic enhancement.
       According to the generally accepted knowledge, in the Ni(OH)2-catalyzed electrooxidation of alcohols or other nucleophilic substrates, Ni(OH)2 first loses electrons and protons and then is reduced to NiOOH through electrochemical steps at a certain anodic potential38,39,40,41. The formed NiOOH then acts as a real active COR species to abstract hydrogen and electrons from the nucleophilic substrate through chemical steps to form the oxidized product20,41. However, it has recently been reported that although the reduction to NiOOH may serve as the rate-determining step (RDS) for the electrooxidation of alcohol on Ni(OH)2, as suggested in recent literature, the oxidation of Ni3+ alcohols may be a spontaneous process through non-redox electron transfer through unoccupied orbitals of Ni3+41,42. Inspired by the mechanistic study reported in the same literature, we used dimethylglyoxime disodium salt octahydrate (C4H6N2Na2O2 8H2O) as a probe molecule to in situ capture any Ni2+ formation resulting from Ni3+ reduction during COR (Supplementary Fig. 22 and Supplementary Note 3). The results showed the formation of Ni2+, confirming that the chemical reduction of NiOOH and the electrooxidation of Ni(OH)2 occurred simultaneously during the COR process. Therefore, the catalytic activity may significantly depend on the kinetics of Ni(OH)2 reduction to NiOOH. Based on this principle, we next investigated whether the modification of V would accelerate the reduction of Ni(OH)2 and thus improve COR.
       We first used in situ Raman techniques to demonstrate that NiOOH is the active phase for COR on Ni(OH)2-NS and NiV-LDH-NS by observing the formation of NiOOH at positive potentials and its subsequent consumption after the introduction of cyclohexanone, following the aforementioned “electrochemical-chemical” process (Figure 3a). Moreover, the reactivity of the reconstructed NiV-LDH-NS exceeded that of Ni(OH)2-NS, as evidenced by the accelerated disappearance of the Ni3+–O Raman signal. We then showed that NiV-LDH-NS exhibited a less positive potential for NiOOH formation compared to Ni(OH)2-NS in the presence or absence of cyclohexanone (Figure 3b, c and Supplementary Fig. 4c, d). Notably, the superior OER performance of NiV-LDH-NS results in more bubbles sticking on the front lens of the Raman measurement objective, which causes the Raman peak at 1.55 VRHE to disappear (Supplementary Fig. 4d). According to the DEMS results (Supplementary Fig. 18), the current density at low potentials (VRHE < 1.58 for Ni(OH)2-NS and VRHE < 1.53 for NiV-LDH-NS) is mainly due to the reconstruction of Ni2+ ions rather than the OER in the absence of cyclohexanone. Thus, the oxidation peak of Ni2+ in the LSV curve is stronger than that of NiV-LDH-NS, indicating that the V modification endows NiV-LDH-NS with enhanced remodeling ability (see Supplementary Fig. 19 for the detailed analysis).
       a In situ Raman spectra of Ni(OH)2-NS (left) and NiV-LDH-NS (right) under OCP conditions after preoxidation at 1.5 VRHE in 0.5 M KOH and 0.4 M cyclohexanone for 60 s. b In situ Raman spectra of Ni(OH)2-NS and c NiV-LDH-NS in 0.5 M KOH + 0.4 M cyclohexanone at different potentials. d In situ XANES spectra of Ni(OH)2-NS and NiV-LDH-NS at Ni K-edge in 0.5 M KOH and e 0.5 M KOH and 0.4 M cyclohexanone. The inset shows a magnified spectral region between 8342 and 8446 eV. f Valence states of Ni in Ni(OH)2-NS and NiV-LDH-NS at different potentials. g In situ Ni EXAFS spectra of NiV-LDH-NS before and after cyclohexanone insertion at different potentials. h Theoretical models of Ni(OH)2-NS and NiV-LDH-NS. Top: On Ni(OH)2-NS, the slow remodeling from Ni(OH)2-NS to NiOOH acts as the RDS, while cyclohexanone reduces the high-valent Ni species through chemical steps to maintain the low-valent Ni state to produce AA. Bottom: On NiV-LDH-NS, the remodeling step is facilitated by the V modification, resulting in the transfer of the RDS from the remodeling step to the chemical step. i The Gibbs free energy changes upon the reconstruction of Ni(OH)2-NS and NiV-LDH-NS. The raw data for aj and i are provided in the raw data file.
       To investigate the evolution of the atomic and electronic structures during catalyst reduction, we performed in situ X-ray absorption spectroscopy (XAS) experiments, which provided a powerful tool to probe the dynamics of Ni species in three successive steps: OER, cyclohexanone injection, and COR at open circuit potential (OCP). The figure shows the K-edge XANES spectra of Ni with increasing potential before and after cyclohexanone injection (Figure 3d, e). At the same potential, the absorption edge energy of NiV-LDH-NS is significantly more positive than that of Ni(OH)2-NS (Figure 3d, e, inset). The average valence of Ni under each condition was estimated by a linear combined fit of the XANES spectra and the regression of the Ni K-edge absorption energy shift (Figure 3f), with the reference spectrum taken from the published literature (Supplementary Fig. 23)43.
       In the first step (before the introduction of cyclohexanone, corresponding to the OER process; Figure 3f, left), at the potential of the unreconstructed catalyst (<1.3 VRHE), the valence state of Ni in NiV-LDH-NS (+1.83) is slightly lower than that of Ni(OH)2-NS (+1.97), which can be attributed to the electron transfer from V to Ni, consistent with the above-mentioned XPS results (Figure 2f). When the potential exceeds the reduction point (1.5 VRHE), the valence state of Ni in NiV-LDH-NS (+3.28) shows a more obvious increase compared with that of Ni(OH)2-NS (+2.49). At higher potential (1.8 VRHE), the valence state of Ni particles obtained on NiV-LDH-NS (+3.64) is higher than that of Ni(OH)2-NS (+3.47). According to recent reports, this process corresponds to the formation of high-valent Ni4+ species in the structure of Ni3+xOOH1-x (Ni3+x is a mixed species of Ni3+ and Ni4+), which has previously shown enhanced catalytic activity in alcohol dehydrogenation38,39,44. Therefore, the superior performance of NiV-LDH-NS in COR may be due to the enhanced reducibility to form catalytically active high-valent Ni species.
       In the second step (introduction of cyclohexanone after ring opening, Figure 3f), the valence state of Ni on both catalysts decreased significantly, which corresponds to the reduction process of Ni3+xOOH1-x by cyclohexanone, which is consistent with the results of in situ Raman spectroscopy (Figure 3a), and the valence state of Ni almost recovered to the initial state (first step at low potential), indicating the reversibility of the redox process of Ni to Ni3+xOOH1-x.
       In the third step (COR process) at COR potentials (1.5 and 1.8 VRHE; Figure 3f, right), the valence state of Ni in Ni(OH)2-NS increased only slightly (+2.16 and +2.40), which is significantly lower than at the same potential in the first step (+2.49 and +3.47). These results indicate that after cyclohexanone injection, COR is kinetically limited by the slow oxidation of Ni2+ to Ni3+x (i.e., Ni reconstruction) rather than by the chemical step between NiOOH and cyclohexanone on Ni(OH)2-NS, which leaves Ni in a low-valence state. Thus, we conclude that Ni reconstruction can serve as RDS in the COR process on Ni(OH)2-NS. In contrast, NiV-LDH-NS maintained a relatively high valence of Ni species (>3) during the COR process, and the valence decreased much less (less than 0.2) compared with the first step at the same potential (1.65 and 1.8 VRHE), indicating that the V modification kinetically promoted the oxidation of Ni2+ to Ni3+x, making the Ni reduction process faster than the chemical step of cyclohexanone reduction. The extended X-ray absorption fine structure (EXAFS) results also revealed a complete transformation of Ni–O (from 1.6 to 1.4 Å) and Ni–Ni(V) (from 2.8 to 2.4 Å) bonds in the presence of cyclohexanone. This is consistent with the reconstruction of the Ni(OH)2 phase to the NiOOH phase and the chemical reduction of the NiOOH phase by cyclohexanone (Fig. 3g). However, cyclohexanone significantly hindered the reduction kinetics of Ni(OH)2-NS (see Supplementary Note 4 and Supplementary Fig. 24 for more details).
       Overall, on Ni(OH)2-NS (Fig. 3h, top), the slow reduction step from Ni(OH)2 phase to NiOOH phase may serve as the RDS of the overall COR process rather than the chemical step of AA formation from cyclohexanone during the chemical reduction of NiOOH. On NiV-LDH-NS (Fig. 3h, bottom), the V modification enhances the oxidation kinetics of Ni2+ to Ni3+x, thereby accelerating the formation of NiVOOH (rather than consumption by chemical reduction), which shifts the RDS toward the chemical step. To understand the Ni reconstruction induced by the V modification, we performed further theoretical calculations. As shown in Fig. 3h, we simulated the reconstruction process of Ni(OH)2-NS and NiV-LDH-NS. The lattice hydroxyl groups on Ni(OH)2-NS and NiV-LDH-NS are deprotonated by extracting OH- in the electrolyte to form electron-deficient lattice oxygen. The corresponding chemical reactions are as follows:
       The Gibbs free energy change of the reconstruction was calculated (Figure 3i), and NiV-LDH-NS (0.81 eV) showed a much smaller Gibbs free energy change than Ni(OH)2-NS (1.66 eV), indicating that the V modification reduced the voltage required for the Ni reconstruction. We believe that promoting the reconstruction may lower the energy barrier of the entire COR (see the reaction mechanism study below for details), thereby accelerating the reaction at higher current densities.
       The above analysis shows that the V modification causes a rapid phase rearrangement of Ni(OH)2, thereby increasing the reaction rate and, in turn, the COR current density. However, the Ni3+x sites can also promote the OER activity. From the LSV curve without cyclohexanone, it is evident that the current density of NiV-LDH-NS is higher than that of Ni(OH)2-NS (Supplementary Fig. 19), which causes the COR and OER reactions to form competitive reactions. Therefore, the significantly higher FE of AA than that of NiV-LDH-NS cannot be fully explained by the V modification promoting the phase rearrangement.
       It is generally accepted that in alkaline media, the electrooxidation reactions of nucleophilic substrates typically follow the Langmuir–Hinshelwood (LH) model. Specifically, the substrate and OH− anions are competitively coadsorbed on the catalyst surface, and the adsorbed OH− is oxidized to active hydroxyl groups (OH*), which serve as electrophiles for the oxidation of nucleophiles, a mechanism that has been previously demonstrated by experimental data and/or theoretical calculations45,46,47. Thus, the concentration of reactants and their ratio (organic substrate and OH−) can control the reactant coverage of the catalyst surface, thereby affecting FE and the yield of the target product14,48,49,50. In our case, we hypothesize that high cyclohexanone surface coverage in NiV-LDH-NS favors the COR process, and conversely, low cyclohexanone surface coverage in Ni(OH)2-NS favors the OER process.
       To test the above hypothesis, we first conducted two series of experiments related to the concentration of reactants (C, cyclohexanone, and COH−). The first experiment was carried out with electrolysis at a constant potential (1.8 VRHE) on Ni(OH)2-NS and NiV-LDH-NS catalysts with different cyclohexanone C contents (0.05 ~ 0.45 M) and a fixed COH− content (0.5 M). Then, the FE and AA productivity were calculated. For the NiV-LDH-NS catalyst, the relationship between AA yield and cyclohexanone C showed a typical “volcanic type” curve in the LH mode (Fig. 4a), indicating that the high cyclohexanone coverage competes with the OH− adsorption. While for Ni(OH)2-NS, the AA yield increased monotonically with the increase of C of cyclohexanone from 0.05 to 0.45 M, indicating that although the bulk concentration of cyclohexanone was high (0.45 M), its surface coverage was still relatively low. In addition, with the increase of COH− to 1.5 M, a “volcanic type” curve was observed on Ni(OH)2-NS depending on C of cyclohexanone, and the inflection point of the performance was delayed compared with NiV-LDH-NS, further proving the weak adsorption of cyclohexanone on Ni(OH)2-NS (Supplementary Fig. 25a and Note 5). In addition, the FE of AA on NiV-LDH-NS was very sensitive to C-cyclohexanone and increased rapidly to more than 80% when C-cyclohexanone was increased from 0.05 M to 0.3 M, indicating that cyclohexanone was easily enriched on NiV-LDH-NS (Figure 4b). In contrast, increasing the concentration of C-cyclohexanone did not significantly inhibit the OER on Ni(OH)2-NS, which may be due to the insufficient adsorption of cyclohexanone. Conversely, further investigation of the dependence of COH− on the catalytic efficiency also confirmed that the adsorption of cyclohexanone was improved compared with NiV-LDH-NS, which could tolerate higher COH− during the COR process without decreasing the FE of AA (Supplementary Fig. 25b, c and Note 5).
       Productivity of AA and EF of b Ni(OH)2-NS and NiV-LDH-NS on cyclohexanone with different C in 0.5 M KOH. c Adsorption energies of cyclohexanone on NiOOH and NiVOOH. d FE of AA on Ni(OH)2-NS and NiV-LDH-NS in 0.5 M KOH and 0.4 M cyclohexanone at 1.80 VRHE using discontinuous and constant potential strategies. Error bars represent the standard deviation of three independent measurements using the same sample and are within 10%. e Top: On Ni(OH)2-NS, cyclohexanone with a low surface area C is weakly adsorbed by cyclohexanone, resulting in strong competition for OER. Bottom: On NiV-LDH-NS, a high surface area concentration of cyclohexanone C is observed with increased adsorption of cyclohexanone, resulting in suppression of OER. The raw data for a–d are provided in the raw data file.
       To test the enhanced adsorption of cyclohexanone on NiV-LDH-NS, we used an electrochemical coupled quartz crystal microbalance (E-QCM) to monitor the mass change of the adsorbed species in real time. The results showed that the initial adsorption capacity of cyclohexanone on NiV-LDH-NS was 1.6 times larger than that on Ni(OH)2-NS in the OCP state, and this difference in adsorption capacity further increased as the potential increased to 1.5 VRHE (Supplementary Fig. 26). Spin-polarized DFT calculations were performed to investigate the adsorption behavior of cyclohexanone on NiOOH and NiVOOH (Figure 4c). Cyclohexanone adsorbs onto the Ni-center on NiOOH with an adsorption energy (Eads) of -0.57 eV, while cyclohexanone can adsorb onto either the Ni-center or the V-center on NiVOOH, where the V-center provides a much lower Eads (-0.69 eV), consistent with the observed stronger adsorption of cyclohexanone on NiVOOH.
       To further verify that the enhanced adsorption of cyclohexanone can promote AA formation and inhibit OER, we used the discontinuous potential strategy to enrich cyclohexanone on the catalyst surface (for Ni(OH)2-NS and NiV-LDH-NS), which was inspired by previous reports. 51, 52 Specifically, we applied a potential of 1.8 VRHE to COR, then switched it to the OCP state, and then switched it back to 1.8 VRHE. In this case, cyclohexanone can accumulate on the catalyst surface in the OCP state between electrolyses (see the Methods section for detailed procedures). The results showed that for Ni(OH)2-NS and NiV-LDH-NS, using discontinuous potential electrolysis improved the catalytic performance compared with constant potential electrolysis (Figure 4d). Notably, Ni(OH)2-NS showed a more significant improvement in COR (AA FE: from 51% to 82%) and suppression of OER (O2 FE: from 27% to 4%) than NiV-LDH-NS, which was attributed to the fact that the cyclohexanone accumulation could be improved to a greater extent on the catalyst with weaker adsorption capacity (i.e., Ni(OH)2-NS) by intermittent potential electrolysis.
       Overall, the inhibition of OER on NiV-LDH-NS can be attributed to the enhanced adsorption of cyclohexanone (Figure 4e). On Ni(OH)2-NS (Figure 4e, top), the weak adsorption of cyclohexanone resulted in a relatively low cyclohexanone coverage and a relatively high OH* coverage on the catalyst surface. Therefore, the excess OH* species will lead to severe competition for OER and reduce the FE of AA. In contrast, on NiV-LDH-NS (Figure 4e, bottom), the V modification increased the adsorption capacity of cyclohexanone, thereby increasing the surface C of cyclohexanone and effectively utilizing the adsorbed OH* species for COR, promoting AA formation and inhibiting OER.
       In addition to investigating the effect of V modification on the reconstruction of Ni species and cyclohexanone adsorption, we also investigated whether V alters the AA formation pathway from COR. Several different COR pathways have been proposed in the literature, and we analyzed their possibilities in our reaction system (see Supplementary Fig. 27 and Supplementary Note 6 for more details)13,14,26. First, it has been reported that the first step of the COR pathway may involve the initial oxidation of cyclohexanone to form the key intermediate 2-hydroxycyclohexanone (2)13,14. To verify the process, we used 5,5-dimethyl-1-pyrrolidine N-oxide (DMPO) to trap the active intermediates adsorbed on the catalyst surface and studied the EPR. The EPR results revealed the presence of C-centered radicals (R ) and hydroxyl radicals (OH ) on both catalysts during the COR process, indicating that the Cα − H dehydrogenation of cyclohexanone forms an intermediate enolate radical (1), which is then further oxidized by OH* to form 2 (Fig. 5a and Supplementary Fig. 28). Although the same intermediates were identified on both catalysts, the area fraction of the R signal on NiV-LDH-NS was relatively higher than that of Ni(OH)2-NS, which may be due to the enhanced adsorption capacity of cyclohexanone (Supplementary Table 3 and Note 7). We further used 2 and 1,2-cyclohexanedione (3) as the starting reactants for electrolysis to test whether V would modify the subsequent oxidation step. The electrolysis results of the potential intermediates (2 and 3) on Ni(OH)2-NS and NiV-LDH-NS showed comparable product selectivities, indicating that the COR reaction on Ni(OH)2-NS or NiV-LDH-NS proceeded via similar pathways (Figure 5b). Moreover, AA was the major product only when 2 was used as the reactant, suggesting that AA was obtained through a direct oxidation process via the cleavage of the Cα − Cβ bond of 2 rather than subsequent oxidation to 3 on both catalysts, since it was mainly converted to GA when 3 was used as the starting reactant (Supplementary Figures 29, 30).
       EPR signal of NiV-LDH-NS in 0.5 M KOH + 0.4 M cyclohexanone. b Results of electrocatalytic analysis of 2-hydroxycyclohexanone (2) and 1,2-cyclohexanedione (3). Electrolysis was carried out in 0.5 M KOH and 0.1 M 2 or 3 at 1.8 VRE for one hour. Error bars represent the standard deviation of two independent measurements using the same catalyst. c Proposed reaction pathways of COR on the two catalysts. d Schematic illustration of the COR pathway on Ni(OH)2-NS (left) and d NiV-LDH-NS (right). Red arrows indicate the steps that the V modification promotes in the COR process. Raw data for a and b are provided in the raw data file.
       Overall, we demonstrated that Ni(OH)2-NS and NiV-LDH-NS catalyze COR via a similar pathway: cyclohexanone is adsorbed on the catalyst surface, dehydrogenated and loses electrons to form 1, which is then oxidized by OH* to form 2, followed by multistep transformations to produce AA (Figure 5c). However, when cyclohexanone was used as a reactant, OER competition was observed only on Ni(OH)2-NS, while the lowest amount of oxygen was collected when 2 and 3 were used as reactants. Thus, the observed differences in catalytic performance may be due to changes in the RDS energy barrier and cyclohexanone adsorption capacity caused by the V modification rather than changes in the reaction pathway. We therefore analyzed the RDS of the reaction pathways on both catalysts. The above-mentioned in situ X-ray acoustic spectroscopy results indicate that the V modification shifts the RDS in the COR reaction from the reconstruction stage to the chemical stage, keeping the NiOOH phase and high-valent Ni species intact on NiV-LDH-NS (Fig. 3f, Supplementary Fig. 24, and Note 4). We further analyzed the reaction processes represented by the current density in each part of the different potential regions during the CV measurement (see Supplementary Fig. 31 and Note 8 for details) and performed H/D kinetic isotope exchange experiments, which collectively demonstrated that the RDS of COR on NiV-LDH-NS involves the cleavage of the Cα − H bond in the chemical stage rather than the reduction stage (see Supplementary Fig. 32 and Note 8 for more details).
       Based on the above analysis, the overall effect of V modification is shown in Figure 5d. Ni(OH)2-NS and NiV-LDH-NS catalysts undergo surface reconstruction at high anodic potentials and catalyze COR via a similar pathway. On Ni(OH)2-NS (Figure 5d, left), the reconstruction step is RDS during the COR process; while on NiV-LDH-NS (Figure 5d, right), V modification significantly accelerated the reconstruction process and converted RDS into Cα−H dehydrogenation of cyclohexanone to form 1. In addition, cyclohexanone adsorption occurred at the V site and was enhanced on NiV-LDH-NS, which contributed to the suppression of OER.
       Considering the excellent electrocatalytic performance of NiV-LDH-NS with high FE over a wide potential range, we designed a MEA to achieve continuous production of AA. The MEA was assembled using NiV-LDH-NS as the anode, commercial PtRu/C as the cathode53 and an anion exchange membrane (type: FAA-3-50) (Figure 6a and Supplementary Fig. 33)54. Since the cell voltage decreased and the FE of AA was comparable with 0.5 M KOH in the above study, the anolyte concentration was optimized to 1 M KOH (Supplementary Fig. 25c). The recorded LSV curves are shown in Supplementary Fig. 34, indicating that the COR efficiency of NiV-LDH-NS is significantly higher than that of Ni(OH)2-NS. To demonstrate the superiority of NiV-LDH-NS, constant current electrolysis was carried out with a step current density ranging from 50 to 500 mA cm−2 and the corresponding cell voltage was recorded. The results showed that NiV-LDH-NS exhibited a cell voltage of 1.76 V at a current density of 300 mA cm−2, which was about 16% lower than that of Ni(OH)2-NS (2.09 V), indicating its higher energy efficiency in AA production (Fig. 6b).
       Schematic diagram of the flow battery. b Cell voltage without iR compensation on Ni(OH)2-NS and NiV-LDH-NS in 1 M KOH and 0.4 M cyclohexanone at different current densities. c AA and FE yield on Ni(OH)2-NS and NiV-LDH-NS at different current densities. Error bars represent the standard deviation of two independent measurements using the same catalyst. d Comparison of the catalytic performance of our work with other reported flow battery systems14,17,19. The reaction parameters and reaction characteristics are listed in detail in Supplementary Table 2. e Cell voltage and FE of AA on NiV-LDH-NS at 200 and 300 mA cm−2 in the long-term test, respectively. The raw data for be are provided as a raw data file.
       Meanwhile, as shown in Fig. 6c, NiV-LDH-NS basically maintained good FE (83% to 61%) at higher current density (200 to 500 mA cm-2), thereby improving the AA productivity (1031 to 1900 μmol cm-2 h-1). Meanwhile, only 0.8% of adipic acid anions were observed in the cathode compartment after electrolysis, indicating that the cyclohexanone transition was not significant in our case (Supplementary Fig. 35). In contrast, with the same increase rate of current density, the FE of AA on Ni(OH)2-NS decreased from 61% to 34%, which made it difficult to improve the AA productivity (762 to 1050 μmol cm-2 h-1). In particular, the performance of AA even slightly decreased due to the strong competition from the OER, and thus the FE of AA decreased sharply with the increase of the current density (from 200 to 250 mA cm−2, Supplementary Fig. 5). To the best of our knowledge, the catalytic results using MEA with NiV-LDH-NS catalysts significantly exceed the performance of previously reported flow reactors with Ni-based catalysts (Supplementary Table 2). Moreover, as shown in Fig. 6d, NiV-LDH-NS showed significant advantages in terms of current density, cell voltage, and FE of AA compared with the best performing Co-based catalyst, i.e., graphene-supported Co3O4 (Co3O4/GDY)17. In addition, we evaluated the energy consumption of AA production and showed that the AA consumption was very low, only 2.4 W h gAA-1 at a current density of 300 mA cm-2 and a cell voltage of 1.76 V (detailed calculations are provided in Supplementary Note 1). Compared with the best result of 4.1 W h gAA-1 for Co3O4/GDY reported previously, the energy consumption for AA production in our work was reduced by 42% and the productivity was increased by 4 times (1536 vs. 319 μmol cm-2 h-1)17.
       The stability of the NiV-LDH-NS catalyst for long-term AA production in MEA was evaluated at current densities of 200 and 300 mA cm-2, respectively (Figure 6e). Since OH− is consumed faster at higher current densities, the electrolyte renewal rate at 300 mA cm-2 is higher than that at 200 mA cm-2 (see the subsection “Electrochemical measurements” for details). At the current density of 200 mA cm-2, the average COR efficiency was 93% in the first 6 h, then slightly decreased to 81% after 60 h, while the cell voltage slightly increased by 7% (from 1.62 V to 1.73 V), indicating good stability. With the current density increasing to 300 mA cm−2, the AA efficiency remained almost unchanged (decreased from 85% to 72%), but the cell voltage increased significantly (from 1.71 to 2.09 V, corresponding to 22%) during the 46-h test (Figure 6e). We speculate that the main reason for the performance degradation is the corrosion of the anion exchange membrane (AEM) by cyclohexanone, which leads to an increase in the cell resistance and voltage of the electrolyzer cell (Supplementary Fig. 36), accompanied by a slight leakage of electrolyte from the anode to the cathode, resulting in a decrease in the anolyte volume and the need to stop the electrolysis. In addition, the decrease in FE of AA could also be due to the leaching of catalysts, which favors the opening of Ni foam for OER. To demonstrate the impact of the corroded AEM on the degradation of stability at 300 mA cm−2, we replaced it with a new AEM after 46 h of electrolysis. As expected, the catalytic efficiency was clearly restored, with the cell voltage significantly decreasing to the initial value (from 2.09 to 1.71 V) and then slightly increasing during the next 12 h of electrolysis (from 1.71 to 1.79 V, an increase of 5%; Figure 6e).
       Overall, we were able to achieve 60 h of continuous AA production stability at a current density of 200 mA cm−2, indicating that the FE and cell voltage of the AA are well maintained. We also tried a higher current density of 300 mA cm−2 and achieved an overall stability of 58 h, replacing the AEM with a new one after 46 h. The above studies demonstrate the stability of the catalyst and clearly indicate the need for future development of higher-power AEMs to improve the long-term stability of the MEA for continuous AA production at industrially ideal current densities.
       Based on the performance of our MEA, we proposed a complete AA production process including substrate feeding, electrolysis, neutralization, and separation units (Supplementary Fig. 37). A preliminary performance analysis was conducted to evaluate the economic feasibility of the system using an alkaline electrolyte electrocatalytic carboxylate production model55. In this case, costs include capital, operations, and materials (Fig. 7a and Supplementary Fig. 38), and revenues come from AA and H2 production. The TEA results show that under our operating conditions (current density 300 mA cm-2, cell voltage 1.76 V, FE 82%), the total costs and revenues are US$2429 and US$2564, respectively, translating into a net profit of US$135 per ton of AA produced (see Supplementary Note 9 for details).
       a Total cost of the AA electrochemical process under the base case scenario with FE of 82%, current density of 300 mA cm−2, and cell voltage of 1.76 V. Sensitivity analysis of the three costs to b FE and c current density. In the sensitivity analysis, only the studied parameters were varied and the other parameters were kept constant based on the TEA model. d Effects of different FE and current density on the profit of AA electrosynthesis and the profit using Ni(OH)2-NS and NiV-LDH-NS, assuming that the cell voltage is kept constant at 1.76 V. The input data for a–d are given in the raw data file.
       Based on this premise, we further investigated the impact of FE and current density on the profitability of AA electrosynthesis. We found that the profitability is very sensitive to the FE of AA, since a decrease in FE leads to a significant increase in the operating cost, thereby substantially increasing the overall cost (Figure 7b). Regarding the current density, a higher current density (>200 mA cm-2) helps to reduce the capital cost and plant construction cost, mainly by minimizing the electrolytic cell area, thereby contributing to an increase in profit (Figure 7c). Compared with the current density, FE has a more significant impact on profit. By characterizing the impact of FE and current density on profit, we clearly see the importance of achieving high FE (>60%) at industrially relevant current densities (>200 mA cm-2) to ensure profitability. Due to the high FE value of AA, the reaction system with NiV-LDH-NS as catalyst remains favorable in the range of 100–500 mA cm−2 (pentagram dots; Figure 7d). However, for Ni(OH)2-NS, decreasing the FE at high current density (>200 mA cm−2) led to unfavorable results (circles; Figure 7d), highlighting the importance of catalysts with high FE at high current density.
       In addition to the importance of catalysts in reducing capital and operating costs, our TEA assessment suggests that profitability could be further improved in two ways. The first is to co-sell potassium sulfate (K2SO4) on the market as a by-product of the neutralization unit, but with potential revenue of US$828/t AA-1 (Supplementary Note 9). The second is to optimize the processing technology, including material recycling or the development of more cost-effective AA separation technologies (alternatives to the neutralization and separation units). The currently used acid-base neutralization process can result in high material costs (which account for the largest share at 85.3%), of which 94% is due to cyclohexanone and KOH ($2069/t AA-1; Figure 7a), but as mentioned above, the process is still overall profitable. We suggest that material costs could be further reduced by more advanced methods for the recovery of KOH and unreacted cyclohexanone, such as electrodialysis for the complete recovery of KOH14 (estimated cost of US$1073/t AA-1 via electrodialysis; Supplementary Note 9).
       In summary, we achieved high efficiency of aluminum atom electrolysis at high current density by introducing V into Ni(OH)2 nanosheets. Under a wide potential range of 1.5–1.9 VRHE and high current density of 170 mA cm−2, the AA FE on NiV-LDH-NS reached 83–88%, while the OER was effectively suppressed to 3%. The V modification promoted the reduction of Ni2+ to Ni3+x and enhanced the adsorption of cyclohexanone. Experimental and theoretical data indicate that the stimulated reconstruction increases the current density for cyclohexanone oxidation and shifts the RDS of COR from reconstruction to dehydrogenation involving Cα − H scission, while the enhanced adsorption of cyclohexanone suppresses OER. The development of the MEA achieved continuous AA production at an industrial current density of 300 mA cm−2, a record AA efficiency of 82%, and a productivity of 1536 μmol cm−2 h−1. A 50-hour test showed that NiV-LDH-NS has good stability as it can maintain high AA FE in MEA (> 80% for 60 hours at 200 mA cm−2; > 70% for 58 hours at 300 mA cm−2). It should be noted that there is a need to develop more powerful AEMs to achieve long-term stability at industrially ideal current densities. In addition, the TEA highlights the economic advantages of reaction strategies for AA production and the importance of high-performance catalysts and advanced separation technologies to further reduce costs.