Spontaneous Chemical Amine Oxidation in Organic Solvents Paired With Electrochemical Hydrogen Production

ABSTRACT

Benzonitrile (BN) is a versatile chemical building block extensively applied across multiple sectors. Electrochemical benzylamine oxidation reaction (BOR) provides a sustainable method for BN synthesis, simultaneously enabling H₂ fuel generation via hydrogen evolution reaction (HER). Alkaline electrolytes are commonly used to reduce the BOR overpotential but promote the hydrolytic decomposition of benzylimine intermediates and BN products, contributing to significant carbon loss. These side reactions become more pronounced under industrially relevant conditions (>50 mM). In this study, redox reservoirs (RRs) were employed to decouple BOR from HER, allowing their independent operation. Utilizing a highly positive formal potential, RRs facilitated spontaneous BOR in hexane, effectively circumventing hydrolysis issues and simplifying the product purification. Furthermore, NiOOH species within RRs functioned as reactive sites with strong affinity toward benzylamine (BA) and negligible self‐reduction, thereby achieving high BN yield and Faradaic efficiency even at BA concentrations exceeding 100 mM. To further demonstrate the versatility, BA derivatives bearing various functional groups were also tested, all of which were successfully converted into their corresponding nitriles. These results highlight the flexibility of RR system to accommodate diverse substrates and meet specific market demands.

Electrochemical benzylamine oxidation reaction (BOR) offers sustainable benzonitrile (BN) synthesis paired with hydrogen evolution reaction (HER), but alkaline electrolytes trigger hydrolysis of benzylimine intermediates and BN. Here, redox reservoirs (RRs) decouple BOR from HER and drive spontaneous BOR in hexane, suppressing hydrolytic carbon loss. NiOOH sites within RRs deliver high BN yield at benzylamine concentrations above 100 mM.

1. Introduction

Nitriles are essential precursors for manufacturing commercial products, including dyes, herbicides, agrochemicals, pharmaceuticals, electronic materials, and polyesters [1, 2, 3, 4]. Their widespread utility has generated significant interest across numerous market sectors. To meet this demand, various organic synthesis pathways for nitriles have been developed. Prominent among these are Kolbe nitrile synthesis, Sandmeyer reaction, Rosenmund‐von Braun reaction, and amide dehydration (Figure S1) [5, 6]. These traditional approaches often involve the use of hazardous reagents, harsh conditions, and result in the production of significant contaminants. Thus, there is a critical requirement for a more environmentally benign method to synthesize nitriles under less severe conditions.

On‐site production of nitriles through electrochemical amine oxidation reaction (AOR), powered by renewable energy, offers a sustainable avenue for nitrile synthesis (Figure 1a). This method offers numerous benefits, including elimination of toxic chemicals and mild operational conditions. Besides, green H₂ fuel can be generated via hydrogen evolution reaction (HER) at the cathode. The thermodynamic potential for AOR typically ranges between 0.5 and 0.8 V versus the reversible hydrogen electrode (vs. RHE) [7]. For instance, converting benzylamine (BA) to benzonitrile (BN) theoretically requires a potential of 0.76 V (vs. RHE). This reaction occurs via an initial two‐electron transfer step forming benzylimine (BI) intermediates, which subsequently undergo a second two‐electron oxidation to yield BN, as depicted in Figure 1b (pathway I). Consequently, the involvement of consecutive multi‐electron processes tends to be kinetically sluggish, with the onset potential typically observed at ≥ 1.3 V (vs. RHE) [1, 2, 7, 8, 9, 10]. To reduce the overpotential and enhance current density, highly alkaline electrolytes are often utilized for electrochemical benzylamine oxidation reaction (BOR), enabling high selectivity and yield from dilute BA solutions (≤ 50 mM) [1, 2, 7, 8, 9, 10]. However, under such conditions, both BI intermediates and BN products are inherently unstable, undergoing hydrolysis in the presence of H₂O and OH⁻ ions (pathways II and III in Figure 1b) [11, 12]. In industrial scenarios that involve higher concentrations (> 50 mM), this intrinsic instability becomes more pronounced, contributing to the formation of substantial by‐products.

FIGURE 1.

(a) Diagram of a conventional electrochemical cell used for benzylamine oxidation reaction (BOR) alongside hydrogen evolution reaction (HER). (b) Proposed catalytic pathways for BOR, including mechanisms of side reactions in an alkaline environment. (c) Schematic illustration of a device designed to decouple chemical BOR and electrochemical HER processes using a redox reservoir (RR), where the oxidative (RR_oxi) and reductive (RR_red) states of the RR electrodes correspond to NiOOH and Ni(OH)₂, respectively.

In conventional electrochemical BOR systems (Figure 1a), the concurrent generation of BN and H₂ with a fixed stoichiometric ratio (1:2) poses economic challenges due to the mismatch in market demands for these two products [13, 14, 15]. The global markets for BN and H₂ are projected to reach USD 360.7 million and USD 47.83 billion, respectively, by 2032, reflecting compound annual growth rates (CAGR) of 3.6% and 10.4% [16]. Consequently, coupling BOR with HER to fulfill H₂ demand would lead to excessive BN production, surpassing market requirements and resulting in undesirable chemical waste. Besides, to overcome crossover issues in electrochemical devices, ion‐exchange membranes are commonly employed to separate anodic and cathodic compartments during BOR [7, 17, 18]. However, their practical application is constrained by high costs and limited durability, typically lasting only ∼1000 h due to performance degradation from ion contamination [19, 20]. As an alternative, membrane‐free undivided cells have been adopted for BOR at relatively low BA concentrations (≤ 50 mM) [21, 22, 23, 24]. Nevertheless, under high concentration conditions, the cathode exerts a detrimental influence on BN products, promoting the generation of nontarget molecules—a critical issue largely overlooked in prior studies and further elaborated upon in detail later. Additionally, supporting electrolytes (i.e., electrochemically inert salts) are generally indispensable for maintaining ionic conductivity in electrochemical setups. Their presence complicates downstream purification and often necessitates additional energy‐intensive separation steps, thereby increasing operational costs [25, 26]. Nitrile synthesis has also been achieved via the electrochemical reduction of nitrogen oxide species (NO_x) or nitrate ions (NO₃ ⁻) into hydroxylamine (NH₂OH) [27, 28, 29], followed by condensation with carbonyl substrates to form oxime intermediates and their subsequent conversion to nitriles. However, this approach remains constrained by competing HER, limited NH₂OH accumulation, and inefficient oxime conversion, all of which diminish the overall yield.

In this work, a nickel oxide hydroxide/nickel hydroxide [NiOOH/Ni(OH)₂] electrode was utilized as a redox reservoir (RR), functioning analogously to water reservoirs in pumped storage hydropower systems, to decouple the chemical BOR and electrochemical HER steps (Figure 1c). Thanks to the highly positive formal potential (E ₀), NiOOH species within the RR electrode spontaneously initiated BOR without requiring an applied bias, operating effectively under open‐circuit conditions. During the reaction, NiOOH was reduced to Ni(OH)₂, facilitating the conversion of BA into BN molecules. To circumvent the hydrolysis side reactions encountered in alkaline environments (Figure 1b), various organic solvents were employed for the RR‐driven BOR, as these media are devoid of H₂O or OH⁻ ions. Besides, this methodology eliminates the need for supporting electrolyte salts, thereby simplifying the product purification through a straightforward distillation step. Among the solvents tested, BOR conducted in hexane delivered the fastest reaction rate, attributed to the strong interaction between BA and the NiOOH surface, along with minimal self‐reduction of the RR electrode. This enabled high BN yield and selectivity in hexane, even at BA concentrations exceeding 100 mM. Following the BOR step, the RR electrode was repurposed as the anode to perform HER for regenerating Ni(OH)₂ into NiOOH at designated timings (Figure 1c). This strategy effectively eliminated cathodic degradation of BN molecules. Remarkably, high HER Faradaic efficiency was maintained even in the absence of membranes. To validate the flexibility of RR system, a series of BA derivatives containing distinct functional groups were subjected to the BOR process. In all cases, efficient transformation into the target nitriles was achieved. These findings demonstrate that the spontaneous oxidation method is highly adaptable to structurally diverse substrates, supporting customized synthesis aligned with varying industrial demands.

2. Results and Discussion

2.1. Chemical Properties of BA and BN Molecules

Electrochemical oxidation of BA proceeds via a stepwise four‐electron transfer mechanism (Figure 1a,b): initially, BA substrate undergoes a 2‐electron oxidation to form BI intermediates, followed by a second 2‐electron step converting BI into BN, as illustrated in Figure 1b (pathway I). Alkaline electrolytes are commonly employed to facilitate this process by lowering the overpotential and enhancing the current response. Although several studies have reported high BN selectivity or yield, these findings were primarily based on low BA concentrations (≤ 50 mM) [1, 2, 7, 8, 9, 10]. This raises concerns that reactants, intermediates, or products may undergo undesired side reactions under alkaline conditions—an issue that has not been thoroughly examined. To the best of our knowledge, BA oxidation reaction (BOR) has yet to be explored under industrially relevant conditions, particularly at concentrations exceeding 100 mM.

To assess the stability of substrates in alkaline environments, we monitored the time‐dependent concentration changes using high‐performance liquid chromatography (HPLC), as shown in Figure S2. Stirring BA molecules in 1 M KOH revealed no significant concentration variations (Figure 2a,b). Intriguingly, when tracking the BN concentration under the same condition, we noted a substantial drop to about 40.5% of its original value within 12 h (Figure 2c,d). During this interval, new peaks emerged in the HPLC profiles, identified as benzamide and benzoic acid, both of which showed a gradual increase. It is noted that under alkaline conditions, benzoic acid predominantly exists in its deprotonated form as benzoate, the corresponding carboxylate anion. Nevertheless, as the reactions in this study were conducted under varying pH conditions and solvent systems (as discussed in later sections), the term “benzoic acid” is used throughout for consistency and ease of compound identification. Further experiments demonstrated that when benzamide alone was stirred in 1 M KOH (Figure 2e,f), its concentration declined over time, accompanied by a corresponding increase in benzoic acid. These results suggest that in alkaline solutions, BN molecules are susceptible to hydrolytic decomposition, as depicted in Figure 1b (pathway II), initially forming benzamides as intermediates (Figure S3a). This process is followed by further transformation into carboxylic acids, facilitated through the nucleophilic addition of OH⁻ ions and the elimination of NH₂ ⁻ groups (i.e., amide hydrolysis), as shown in Figure S3b. Although a less alkaline medium was employed to decelerate such undesired side reactions, it failed to suppress them entirely (see Note S1 and Figures S4–S6 for details) [30]. These degradation pathways likely account for the limited exploration of BOR at elevated BA concentrations (>100 mM), as the resulting instability of the products ultimately constrains BN yield.

FIGURE 2.

High‐performance liquid chromatography (HPLC) patterns of 5 mM (a) benzylamine, (c) benzonitrile, and (e) benzamide dissolved in 1 M KOH, along with (b, d, f) corresponding time‐dependent changes in quantitative percentage over 12 h.

To further elucidate the effect of BN quantities on the degradation behaviors under alkaline conditions, a series of experiments was performed by introducing different amounts of BN into 1 M KOH (Figure S7). Due to the hydrophobic nature of BN [1, 30], its limited aqueous solubility may influence its susceptibility to hydrolytic decomposition. After stirring for 12 h, about 35–40% of the initial BN remained in both the 75 and 750 µmol cases, with the majority converting into benzamide and benzoic acid. Although the degradation percentages were similar, the absolute amount of side products was substantially higher in the 750 µmol condition. This finding emphasizes that operating BOR at elevated BA concentrations, although essential for industrial applications, results in greater overall product loss and presents a critical challenge for scaling up in alkaline electrolytes.

Ion‐exchange membranes, though widely employed in electrochemical BOR [7, 17, 18], have a limited operational lifespan of approximately 1000 h due to impurity‐induced decay [19, 20]. Using undivided cells (i.e., batch mode) may offer a more cost‐effective approach, yet there has been no research into whether the cathode negatively impacts BA or BN molecules during HER. To examine this, 750 µmol of BA in 1 M KOH was isolated as the catholyte using a membrane to prevent it from anodic influences (Figure S8a). Following the chronopotentiometry test at the current of −15 mA, no significant changes were detected (Figure S8b). In a separate experiment, an emulsion containing BN in 1 M KOH was used as the catholyte. The degradation of BN became more pronounced during HER (Figure S8c), compared to the relatively milder decomposition observed in the absence of applied bias—i.e., the non‐Faradaic pathway discussed earlier (Figure S3). This bias‐induced deterioration is herein termed the “cathodic effect”. While BA reactants remained stable at the cathode during HER, the resulting BN products exhibited limited resistance, highlighting their vulnerability under cathodic conditions in undivided cells. This degradation pathway may have gone unrecognized in earlier studies, likely due to the use of lower substrate concentrations, where insufficient BN accumulation near the cathode failed to provoke observable side reactions.

2.2. Characterizations and Electrochemical Behaviors of RRs

In this study, RRs were utilized to overcome the challenges of conventional BOR in alkaline media, as outlined in previous sections. The oxidative species of RR, NiOOH (denoted as RR_oxi), enabled the spontaneous BOR under open‐circuit conditions (Figure 1c). This process concurrently generated Ni(OH)₂ as the reductive components of RR (denoted as RR_red). At designated timings, RR_red electrode acted as the anode in an electrochemical cell, driving HER catalysis and regenerating Ni(OH)₂ back into NiOOH. Thus, this approach effectively decoupled the BOR and HER half‐reactions, allowing for their independent operation. To meet these requirements, modified‐Ni(OH)₂ electrodes sourced from commercial nickel‐metal hydride (Ni‐MH) batteries were employed owing to their high capacity. X‐ray diffraction (XRD) profile of the RR_red electrode revealed two sharp peaks at 2θ values of 43.8° and 50.7°, attributed to the Ni foam substrate (Figure S9a). Additional peaks at 18.9°, 32.7°, 38.0°, 57.1°, and 60.3° were identified as the β‐Ni(OH)₂ phase (JCPDS No. 14–0117), corresponding to the crystal planes (001), (100), (101), (110), and (111), respectively [31]. Upon full oxidation to RR_oxi, Ni(OH)₂ converted into the β‐NiOOH phase (JCPDS No. 06–0141). This transformation was evidenced by peaks at 18.9° and 37.9° in the XRD pattern (Figure S9a), correlating with the (001) and (002) Miller indices, respectively [31]. Scanning electron microscopy (SEM) images of the RR electrode displayed spherical particles with diameters ranging from 5 to 20 µm (Figure S9b).

Herein, Ni K‐edge x‐ray absorption near‐edge structure (XANES) was used to investigate the RR oxidation state (Figure S10a). The K‐edge absorption process involves the transition of core‐level electrons from the 1s orbital to the unoccupied np orbitals, with higher‐valent cations exhibiting a strong affinity for these electrons and requiring more excitation energy. A linear correlation between the absorption energy and oxidation state was established by reference standards (Figure S10b). The RR_red and RR_oxi electrodes closely matched those of the NiO and LiNiO₂ standards, which feature Ni²⁺ and Ni³⁺ cations, respectively. Extended x‐ray absorption fine structure (EXAFS) analysis of the RR_red electrode revealed two signals (Figure S11), identified as Ni(OH)₂ with Ni–O (1.60 Å) and Ni–Ni (2.76 Å) coordination. For the RR_oxi electrode, both the Ni–O and Ni–Ni bond lengths displayed contraction, consistent with the formation of the NiOOH phase [32, 33].

To analyze the redox behaviors of RR, cyclic voltammetry (CV) was performed in a 1 M KOH electrolyte (Figure 3a). It revealed an anodic peak at approximately 0.64 V versus the standard hydrogen electrode (vs. SHE), corresponding to the oxidation of Ni(OH)₂ to NiOOH, and a cathodic peak at 0.46 V (vs. SHE), indicating the reduction of NiOOH back to Ni(OH)₂. The highly positive formal potential (E ₀) of 0.55 V (vs. SHE), contributes to a strong electromotive force (EMF) that drives the spontaneous BOR (to be elaborated later). To demonstrate the capability of decoupling reactions using RR, water splitting served as the model reaction (Figure S12). The RR_red electrode was oxidized to form NiOOH during the charging progression, paired with HER catalysis at the cathode. In the subsequent discharging step, the RR_oxi electrode was reduced, converting NiOOH back to Ni(OH)₂, while the anode simultaneously performed the oxygen evolution reaction (OER). Decoupling performance was assessed using galvanostatic charge–discharge (GCD) tests at a current density of 5 mA cm⁻² (Figure S13), tracking the potential variations of RR. Throughout 20 cycles, the Coulombic efficiency (CE) averaged 98.3%, presenting excellent electrochemical stability.

FIGURE 3.

(a) Cyclic voltammetry (CV) of RR powders deposited on carbon paper in 1 M KOH at a scan rate of 0.05 mV s⁻¹. (b) Chronoamperometry (CA) of a NiOOH electrode at 0.65 V versus the standard hydrogen electrode (vs. SHE) in 1 M KOH, with 50 mM benzylamine introduced at 25 min. (c) Galvanostatic charging of RRs at a current density of 5 mA cm⁻² in 1 M KOH following 90 min immersion of a RR_oxi electrode (1 cm²) in specific media. HPLC profiles for BOR using fully oxidized RR (total 12 cm²) with 50 mM benzylamine dissolved in 15 mL of (d) KOH and (f) hexane, accompanied by (e, g) time‐dependent changes in product yield during the reaction. (h) Benzylamine conversion, benzonitrile yield, benzonitrile selectivity, and rate constant of RR‐driven BOR in various organic solvents.

2.3. Performance Evaluation of RR‐Driven Spontaneous BOR

Herein, chronoamperometry (CA) was conducted on the RR_oxi electrode at 0.65 V (vs. SHE) in 1 M KOH, where only a minimal current was initially observed (Figure 3b). At 25 min, BA molecules were introduced into the electrolyte, immediately triggering a significant increase in anodic current. This surge was caused by the NiOOH species within RR acting as chemical oxidants with a strong EMF for driving BOR [11, 34, 35], owing to the highly positive E ₀ value (Figure 3a). This process resulted in the conversion of NiOOH to Ni(OH)₂, which was then re‐oxidized to NiOOH under the applied anodic potential. To further assess whether the intrinsic EMF triggers spontaneous BOR in the absence of external bias, a RR_oxi electrode was immersed in a 50 mM BA/1 M KOH solution for 90 min under an open‐circuit condition, followed by electrochemical charging in 1 M KOH (Figure 3c). Interestingly, a charging capacity of 17 mAh was observed, indicating that the stored charge (Q _S) in RR had been depleted. A control experiment was also carried out in which the RR_oxi electrode was immersed in neat 1 M KOH for 90 min and then fully recharged. Only a minimal charge uptake of ∼1 mAh was detected, possibly due to the self‐discharge of NiOOH [4 NiOOH + 2 H₂O → 4 Ni(OH)₂ + O₂]. These findings validate that NiOOH species in the RR electrode consumed Q _S to automatically oxidize BA molecules.

To quantify the products, HPLC analysis was performed for spontaneous BOR using RR_oxi electrodes (four pieces of each 3 cm²) immersed in a 50 mM BA/1 M KOH solution. The BN yield reached a maximum of approximately 39% after 30–45 min of reaction (Figure 3d,e). Thereafter, it gradually declined to about 31% at 90 min, which was attributed to the instability of BN in the alkaline medium, leading to its conversion into benzamide and benzoic acid (Figure S3). Notably, an unexpected peak, identified as N‐benzylidenebenzylamine (N‐BB), emerged at the retention time of ∼23.2 min in the HPLC profile (Figure 3d). We suggest that in the BOR process, the hydrolysis of BI intermediates led to the generation of benzaldehydes, which were then attacked by the nucleophilic BA substrates, forming N‐BB (azomethine) by‐products (pathways III and IV in Figure 1b; see proposed mechanisms in Figure S14) [12]. Furthermore, benzaldehyde is susceptible to the Cannizzaro reaction in alkaline solutions, leading to disproportionation into benzyl alcohol and benzoic acid (see details in Note S2 and Figure S15) [36, 37]. These side reactions significantly hindered BN yield and exacerbated “carbon loss”—defined as the formation of by‐products not detectable by the HPLC system in this work—during BOR (Figure 3e). To alleviate this issue, reaction environments with reduced alkalinity, including 0.1 M KOH/0.5 M Na₂SO₄ and pure water, were also tested [30, 38]. However, only marginal improvements were observed (see Notes S3 and S4 and Figures S16–S20 for details). These findings further indicate that lowering the OH⁻ ion concentration merely slows the progression of undesired reactions rather than eliminating them completely.

Based on the above results, both BI intermediates and BN products generated during BOR exhibit intrinsic instability, readily undergoing hydrolytic decomposition in the presence of H₂O or OH⁻ ions (Figure 1b). To overcome these challenges, the versatility of RR electrodes was utilized to explore alternative reaction environments [15, 38, 39]. Specifically, a series of organic solvents—hexane, toluene, chloroform, diethyl ether, ethyl acetate, tetrahydrofuran, acetonitrile, and dimethyl sulfoxide—were examined, spanning a wide polarity range based on their XLog P values (Table S1) [40]. These solvents were selected to suppress hydrolysis pathways that occur in aqueous solutions and to leverage the chemical durability of BN in these media (Figure S21). Among the tested solvents, hexane (XLog P = 3.9) delivered the strongest hydrophobicity, effectively preventing the formation of benzoic acid in the BOR process (Figure 3f,g). Upon 90 min of reaction with RR_oxi electrodes in 50 mM BA/hexane, high BN yield and selectivity were achieved, reaching approximately 96.6% and 97.8%, respectively (Figure 3h). The formation of BN and the stability of hexane under oxidative conditions were further verified by nuclear magnetic resonance (NMR) spectroscopy (Figure S22), supporting the suitability of this solvent for RR‐driven BOR under open‐circuit conditions enabled by the highly positive E ₀ property.

Previous studies have shown that aldehyde oxidation to carboxylic acid readily proceeds via NiOOH species, corresponding to pathway V in Figure 1b [38, 41, 42]. Notably, when BOR was conducted in hexane, trace amounts of N‐BB by‐products emerged without any detectable benzoic acid (Figure 3f). To probe the mechanistic origin of N‐BB, a control experiment was performed by immersing RR_oxi electrodes in a 50 mM benzaldehyde/hexane medium (see Note S4 for details), which led to benzoic acid formation and carbon loss (Figure S20). These results indicate that benzaldehyde is unlikely to serve as a key intermediate in the hexane‐based BOR process (Figure 3f,g). Based on prior reports [43, 44, 45, 46], we propose that in organic solvents, BI intermediates may undergo direct nucleophilic attack by BA molecules via a concerted single‐step mechanism, as outlined in Figure 1b (pathway VI). This contrasts with the stepwise pathway typically investigated in aqueous solutions, wherein BI is first prone to hydrolyze into benzaldehyde, which subsequently reacts with BA to generate N‐BB (Figure S14). These findings underscore the critical influence of the reaction environment in shaping intermediate speciation and determining the nature of side products during spontaneous oxidation.

As shown in Figure 1c, oxidizing BA to BN requires four electrons, whereas converting NiOOH to Ni(OH)₂ involves only a single‐electron transfer. Consequently, four NiOOH species are depleted for each BN molecule produced during BOR. This consumption of Q _S within the RR electrodes correlates with the BA conversion rate, which can be formulated as:

$$\mathrm{Rate}=k\times {\left[\mathrm{BA}\right]}^x\times {Q_{\mathrm{S}}}^y=k_{\mathrm{obs}}\times {\left[\mathrm{BA}\right]}^x$$

Here, k is the rate constant, while x and y denote the reaction orders concerning BA concentration ([BA]) and Q _S, respectively. Given that the Q _S value (∼480 mAh for RRs with a total area of 12 cm²) far exceeds the theoretical charge required for full conversion of 50 mM BA to BN in 15 mL (80.4 mAh, denoted as Q _BOR), Q _S was regarded as quasi‐constant and embedded into the observed rate constant (k _obs). Under this assumption, the BA conversion rate exhibited a linear dependence on [BA], indicating first‐order kinetics (x = 1) in hexane, with a k _obs of 2.42 h⁻¹ (Figure S23).

Among the organic solvents evaluated for BOR performance (Figures S24–S30), chloroform was the only medium that showed negligible BA conversion. By contrast, all other solvents enabled spontaneous BOR that followed first‐order kinetics with respect to [BA], with corresponding k _obs values summarized in Figure 3h. Notably, acetonitrile delivered a BN yield of approximately 82.9% after 90 min, together with a high k _obs of 2.78 h⁻¹ (Figure S29). Nevertheless, several unknown peaks appeared in the HPLC profiles (Figure S31a). Under oxidative conditions, previous studies have suggested that BA can generate reactive intermediates (e.g., radical cations), which attack the electrophilic carbon of acetonitrile to form amidine‐type adducts (Figure S31b). Further transformation of these species is a possible origin of the carbon loss observed in acetonitrile (Figure S29) [47, 48]. Consequently, only moderate BN selectivity of 85.8% was obtained (Figure 3h). By contrast, the other organic solvents employed in this study lack comparably activated electrophilic centers and are therefore less likely to participate in such side reactions.

To further elucidate the influence of solvent properties on the BOR process, hexane (which afforded both high reaction rate and BN selectivity) and chloroform (which showed minimal reactivity) were chosen for detailed comparative analysis in the following sections.

2.4. Mechanistic Insights into BOR Facilitated by RRs in Organic Solvents

In light of the stark contrast in BOR efficiency between using hexane and chloroform as reaction solvents, we sought to elucidate the underlying causes by systematically examining four critical factors potentially governing the overall reactivity: (i) the mass transport of BA from the bulk medium to the RR surface, (ii) the adsorption affinity of BA toward NiOOH species, (iii) the EMF magnitude for driving the reaction, and (iv) the oxidative stability of RR_oxi electrode in each solvent environment.

Diffusion‐ordered spectroscopy (DOSY) NMR analysis was employed to determine the diffusion coefficients (D) of BA in organic solvents (see Note S5 and Figure S32 for details). This technique relies on tracking the attenuation of NMR signal intensity under a series of pulsed field gradients, which encode molecular displacement during the diffusion period. Based on the Stejskal–Tanner equation, the D values can be extracted from the observed signal decay. As shown in Figure 4a, BA molecules exhibited similar D levels in both hexane and chloroform. Under conditions of well‐controlled convection, where the diffusion boundary layer is effectively minimized, the D magnitude can be regarded as approximately proportional to the mass transport rate of substrates. Accordingly, these findings indicate that molecular diffusion of BA is unlikely to be the rate‐limiting factor accounting for the observed differences in reactivity during BOR.

FIGURE 4.

(a) Diffusion coefficients (D) of benzylamine in hexane and chloroform solvents determined by diffusion‐ordered spectroscopy (DOSY) nuclear magnetic resonance (NMR) analysis. (b) Structural model of the β‐NiOOH slab with O⁻/OH⁻ terminations used in density functional theory (DFT) calculations. (c) Adsorption energies (E _ads) of benzylamine co‐adsorbed with either hexane or chloroform on the NiOOH surface. (d) CV curves of RR powders deposited on carbon paper in 1 M KOH containing 10% organic solvents under vigorous stirring at a scan rate of 0.05 mV s⁻¹. (e) Galvanostatic charging of RRs at a current density of 5 mA cm⁻² in 1 M KOH following 90 min immersion of RR_oxi electrodes in specific media.

To elucidate the interfacial interactions between BA and β‐NiOOH, later identified by in situ characterizations as the reactive species for BOR, density functional theory (DFT) calculations were performed using a structural model of β‐NiOOH (Figure 4b) to determine the most energetically favorable adsorption configurations. After first exploring the adsorption energies (E _ads) of individual solvent and BA molecules on the NiOOH surface (see details in Note S6 and Figures S33–S35), we further examined the cooperative effects on adsorption characteristics of BA in conjunction with either hexane or chloroform (Figures S36–S39). Across all tested adsorption configurations, the BA molecule consistently acted as the primary adsorbate, anchoring to NiOOH via strong polar interactions. The presence of a nearby solvent species further modulated the overall E _ads levels, as summarized in Figure 4c. Notably, in both solvents, adjacent (denoted as A‐type) geometries consistently exhibited more negative E _ads values than their stacked (denoted as S‐type) counterparts, suggesting that direct interaction between solvent molecules and NiOOH enhances the interfacial stability of neighboring BA. In the case of hexane, the A6 arrangement—featuring a side‐by‐side orientation—was identified as the most stable, with an E _ads of –1.25 eV. This pronounced stabilization likely stems from a synergistic effect, where BA forms robust surface interactions, while hexane reinforces the adsorption through lateral van der Waals forces. A similar trend was observed in the chloroform system, with the A6 arrangement once again emerged as the most favorable, although its E _ads (–1.18 eV) was slightly less negative, indicating a weaker cooperative effect. This contrast underscores the superior compatibility between BA and the β‐NiOOH surface in a nonpolar medium, where hexane not only avoids disrupting interfacial interactions but also enhances adsorption stability.

Inspired by the work of Stahl and coworkers, who demonstrated that open‐circuit potential (OCP) tests provide useful thermodynamic benchmarks for correlating reaction driving force with product yields in organic solvents [49], we initially sought to apply a similar strategy to evaluate the redox properties of RR electrodes. However, direct OCP measurements were impractical in our case because no universal supporting electrolyte was compatible with all of the solvents examined. To address this limitation, a vigorously stirred emulsion containing 10% organic solvent in 1 M KOH was used as the reaction medium. Its feasibility was first assessed by chronoamperometric measurement of the RR_oxi electrode at 0.65 V (vs. SHE) in 1 M KOH, followed by injection of organic solvents containing BA at 25 min (Figure S40). Relative to chloroform, hexane provided a higher current response, indicating faster BOR driven by the oxidative equivalents stored in RRs, consistent with its large k _obs value (Figure 3h). This emulsion system was therefore adopted for CV tests to estimate the apparent formal potentials (E ₀′), as shown in Figure 4d. Notably, the measured E ₀′ values deviate from standard E ₀ obtained under quiescent conditions, because convection was required to prevent phase separation of the emulsion during the CV scans. Even so, comparable E ₀′ values were obtained in hexane and chloroform (Figure 4d), suggesting that the EMF level required to trigger BOR is largely independent of solvent identity.

Following NiOOH‐driven BOR in organic solvents, the resulting Ni(OH)₂ species within RRs were re‐oxidized in 1 M KOH via a galvanostatic charging step, concurrently enabling HER at the cathode (Figure 1c). Notably, the charge consumed during BOR matched the amount retrieved in the subsequent charging process. After reaction in hexane containing 50 mM BA (15 mL), a recovery of ∼82 mAh was observed (Figure 4e), closely aligning with theoretical Q _BOR value (80.4 mAh) and indicating that nearly all the charge was effectively directed toward substrate oxidation. In contrast, the same procedure in chloroform required an input of 139 mAh but yielded only ∼4.5% BN. To further investigate this discrepancy, the RR_oxi electrode was incubated in pure chloroform for 90 min. Upon recharging in 1 M KOH, a capacity of ∼75 mAh was measured (Figure 4e), indicating substantial self‐reduction of NiOOH in chloroform. Thus, the diminished oxidative capacity of RR electrode markedly hindered its ability to sustain BOR in chloroform.

Taken together, the above findings indicate that the superior BOR efficiency observed in hexane using RRs arises primarily from favorable substrate adsorption and retained oxidative stability of NiOOH. On this basis, several general principles for medium selection in spontaneous oxidation can be outlined. The reaction environment should avoid severe transport limitations, especially those associated with high viscosity. It must also maintain suitable substrate‐electrode interactions consistent with the Sabatier principle, preserve sufficient thermodynamic driving force, and enable the stored oxidative equivalents to be only delivered to BOR. In addition, the solvent is essential to be chemically inert toward the substrate, intermediates, and products, since undesired solvent‐participated reactions may lead to by‐product formation and diminished selectivity.

2.5. Identification of Active Sites in RRs for Spontaneous BOR in Hexane

Herein, dissolved O₂ was unable to trigger BOR, either in the absence or presence of RR_red electrodes with the Ni(OH)₂ phase (Figure S41). In electrochemical oxidation, hydroxyl radical (•OH) is among the most commonly proposed reactive oxygen species (ROS) [50, 51]. However, addition of tert‐butyl alcohol (TBA), a well‐established •OH scavenger, caused a slight decrease in BN production (Figure S42). This minor change is more reasonably attributed to alteration of the reaction medium, which may weaken BA adsorption on the RR surface and thereby reduce the BOR rate, rather than to quenching of ROS. These observations suggest that neither dissolved O₂ nor •OH acts as an effective oxidant under the present conditions. Instead, initiation of BOR requires the RR_oxi electrode.

Fleischmann et al. suggested that NiOOH (Ni³⁺) serves as the active site for BOR in Ni‐based materials, proceeding through a hydrogen atom transfer (HAT) step to form an α‐aminobenzyl radical intermediate (Figure S43a) [34, 35]. Conversely, Choi's group proposed that NiO₂ (Ni⁴⁺) acts as the reactive species, facilitating a hydride transfer from the α‐position of adsorbed amine to NiO₂ [11]. Both mechanistic pictures were established in aqueous environments, whereas the reactive sites in an organic medium have remained unclear. For RR‐driven BOR in hexane, no electron paramagnetic resonance (EPR) signal was observed when 5,5‐dimethyl‐1‐pyrroline N‐oxide (DMPO) was used as the spin‐trapping agent (Figure S43b) [52]. Nevertheless, the HAT pathway is not necessarily precluded, as the rate‐limiting step is proposed to generate highly transient α‐aminobenzyl radicals that would be difficult to capture [11]. This absence of EPR signals is again taken as evidence against the involvement of •OH in the reaction, given the widespread use of DMPO for •OH detection.

To further identify the reactive species, in situ XANES and XRD characterizations were performed on the RR_oxi electrode in the NiOOH phase during BOR. Relative to the spectrum measured in air, a subtle shift of the absorption edge to higher energy was observed in hexane (Figure S44), suggesting a modified interfacial electronic environment with slightly more electron‐deficient Ni centers. Such modulation is expected to strengthen BA adsorption through its electron‐donating amine group, in line with the DFT results (Figure 4c). Following BOR initiation by BA addition in hexane, diffraction peaks assigned to β‐Ni(OH)₂ gradually emerged in the XRD patterns (Figure S45) [31], while the Ni K‐edge absorption edge shifted progressively to lower energy with time (Figure 5a). Together, these observations consistently indicate conversion of β‐NiOOH into Ni(OH)₂ during BOR, thereby supporting NiOOH as the reactive site, most likely through a HAT‐type mechanism, with no evidence for NiO₂ formation in this system.

FIGURE 5.

(a) In situ Ni K‐edge x‐ray absorption near‐edge structure (XANES) spectra of the fully oxidized RR (3 cm²) in the BOR process, along with (b) variations in Ni oxidation states. (c) Evolution of state of charge (SOC), charge stored (Q _S), and charge depletion rate (− ΔQ _S/Δt) in RR during the reaction. (d) In situ Ni K‐edge extended x‐ray absorption fine structure (EXAFS) profiles of the RR_oxi electrode throughout the reaction.

Using the linear correlation derived from the standards (Figure S10b), in situ XANES was applied to quantify the evolution of RR oxidation states during BOR (Figure 5b). The resulting data were used to estimate the NiOOH fraction, from which the state of charge (SOC) and Q _S value of the RR electrode were further determined (Figure 5c). The BA conversion efficiency is proportional to the charge depletion rate (− ΔQ _S/Δt), which decreased as the reaction proceeded. This trend suggests that RRs with higher SOC levels, corresponding to larger Q _S, provide a greater amount of NiOOH species available for facilitating BOR. Strategic manipulation of Q _S, crucial for achieving full BA conversion within an acceptable timeframe, will be demonstrated in detail later. Besides, upon BA introduction into hexane to trigger BOR, the EXAFS features assigned to Ni–O (1.44 Å) and Ni–Ni (2.36 Å) became elongated with reaction time (Figure 5d) [32, 33]. These results further support NiOOH in the RR_oxi electrode as the reactive site responsible for spontaneous oxidation in organic solvents.

2.6. Efficiency Assessment of Decoupled Reactions Using RR Electrodes

To elucidate the strategy for controlling the BOR rate by manipulating the Q _S values within RRs, as discussed in the previous section, we adjusted the ratio between Q _S (amount of NiOOH species) and Q _BOR (the theoretical charge required for complete BA‐to‐BN conversion) [38]. As shown in Figure 6a, increasing this ratio contributed to faster BA conversion, confirming that a high stoichiometry of NiOOH accelerated BOR efficiency. By systematically tuning the Q _S/Q _BOR ratios (Figure 6b), full BA conversion was achieved within reasonable timeframes. At concentrations up to 150 mM, high BN yield was obtained after 4.5 h, together with excellent Faradaic efficiency of BOR (FE_BOR). The corresponding BN production rate reached approximately 0.5 mmol h⁻¹, surpassing that of several prior studies about direct BOR electrocatalysis [8, 21, 53]. Notably, such performance under industrially relevant concentrations (≥ 100 mM) has not been reported to date (Table S2). Besides, BOR was successfully carried out at an even higher BA concentration of 200 mM (Figure 6b), affording a slightly reduced BN yield of 91.8% and a FE_BOR of 88.3%. This decline in selectivity was possibly associated with the elevated local concentration of BA around adsorbed BI intermediates on the RR surface, which may promote formation of the N‐BB by‐product, as shown in Figure 1b (pathway VI).

FIGURE 6.

(a) Time‐dependent residual benzylamine percentage during BOR at different ratios between stored charge in RRs (Q _S) and the theoretical charge required for complete benzylamine‐to‐benzonitrile conversion (Q _BOR). (b) Spontaneous oxidation of highly concentrated benzylamine in hexane (15 mL), carried out by the RR electrode. (c) Operating potentials of charging the RR anode (E _a) and performing hydrogen evolution reaction (HER) at the Pt foil cathode (E _c). (d) Benzylamine conversion, benzonitrile yield, and Faradaic efficiencies (FEs) for BOR using RRs paired with HER on Pt foil over 15 consecutive cycles. The BOR reaction times are 1.5 and 4.5 h for 50 and 150 mM benzylamine conditions (15 mL), respectively, corresponding to Q _S/Q _BOR values of 6 and 2. (e) ¹H nuclear magnetic resonance (NMR) profile of distilled benzonitrile product obtained from BOR in hexane. (f) Substrate scope of RR‐driven BOR for benzylamine derivatives bearing different functional groups. The substrate concentration is 50 mM in hexane (15 mL), with Q _S/Q _BOR set to 6.

As illustrated in Figure 1c, when employing the RR electrode to decouple BOR and HER, energy input was only required during the charging step, in which Ni(OH)₂ was electrochemically regenerated to NiOOH. This process operated at a cell potential of approximately 1.80 V (Figure 6c). By contrast, conventional direct electrocatalysis coupling BOR with HER in 1 M KOH necessitated a substantially higher voltage (∼3.96 V), primarily due to the large ohmic resistance introduced by the high‐impedance membrane used to suppress the “cathodic effect” (see Note S7 and Figures S46 and S47). Nevertheless, BA crossover into the catholyte remained evident (3.90 mM), and the BN yield was limited to 42.1%, largely attributed to hydrolytic degradation under strong alkaline conditions. To assess the stability and reusability of RR electrode, long‐term decoupled cycling was performed by alternating BOR and HER steps—10 cycles with 50 mM BA followed by 5 cycles with 150 mM BA, totaling 15 operations (Figure 6d). By temporally separating the two half‐reactions, the adverse cathodic effect on BN molecules during HER was completely avoided. Full BA conversion was consistently achieved, maintaining high BN yield and FE_BOR over all 15 cycles. By eliminating the need for supporting electrolytes, this approach streamlined post‐reaction purification, allowing efficient solvent removal through simple distillation, as confirmed by NMR analysis (Figures 6e and S48). Additionally, the Faradaic efficiency of HER (FE_HER) during the charging steps was quantified by gas chromatography, confirming persistently high performance (Figure 6d). Characterizations of the spent RR electrode by XRD and SEM revealed no significant structural or morphological variations (Figure S49).

Despite the widespread use of NiOOH as a catalytic motif for BOR (Table S2) and in decoupled systems [14, 54, 55], the key advance of our work arises not from the introduction of a new material. Rather, by integrating established NiOOH redox chemistry with a non‐aqueous environment that suppresses side reactions, the present strategy addresses a fundamental limitation that has long restricted the practical applicability of previous BOR studies, while improving product yield and lowering the energy demand. Notably, although RR‐driven BOR achieved a high BN production rate of 0.5 mmol h⁻¹ (Table S2), this efficiency corresponds to only ∼4.5 mA cm⁻², assuming a four‐electron process. By contrast, electrochemical regeneration of Ni(OH)₂ to NiOOH can be operated at substantially higher current densities [38]. These results suggest that chemical oxidation is the rate‐determining step in the overall decoupled system.

Herein, to illustrate the feasibility of continuous operation in the decoupled platform, we propose a dual‐cell flow reactor (Figure S50) [55]. While one cell proceeds spontaneous BOR with NiOOH serving as the chemical oxidant and being reduced to Ni(OH)₂, the other cell concurrently regenerates NiOOH electrochemically through anodic oxidation of Ni(OH)₂, paired with HER at the cathode. By alternating the roles of the two cells, BOR and HER can be sustained in a complementary manner, enabling uninterrupted operation. Inspired by recent advances in mechanically assisted reaction systems [56], we are also exploring a roll‐to‐roll reactor analogous to a cable car (Figure S51), in which the RR electrode continuously shuttles between the BOR and regeneration compartments, thereby spatially separating the two half‐reactions while maintaining continuous throughput.

To showcase the flexibility of RRs in driving spontaneous oxidation, a series of BA derivatives bearing either electron‐withdrawing (fluoro) or electron‐donating (methoxy/methyl) substituents were investigated (Figure 6f). In most cases, full conversion was achieved within 1.5 h in hexane, providing the corresponding nitriles with high efficiency. These results highlight the versatility of RR electrodes in oxidizing BA substrates with different functional groups, enabling tailored applications to meet various industrial demands. Notably, methyl‐substituted BA molecules required a longer reaction time of 3 h to reach complete conversion. This observation suggests that the reaction kinetics are not solely governed by the electrophilicity of substituents. Ongoing efforts in our group focus on establishing a reaction library to systematically explore the influence of molecular structure on BOR efficiency. These insights will guide the rational design of more active materials for performing BOR process, offering a sustainable alternative to traditional oxidation protocols in pharmaceutical and fine‐chemical manufacturing. The implications of this RR strategy, however, are not limited to BOR. RR‐driven conversions have also been demonstrated for alcohols and aldehydes [31, 38, 57]. Collectively, these studies highlight the critical influence of medium selection in controlling reaction pathways and broadening the synthetic scope of this platform.

3. Conclusion

In summary, this study demonstrates the effective application of RR electrodes for decoupling BOR and HER, enabling independent control over the two half‐reactions within a desired timeframe. This strategy avoids the detrimental cathodic effect on BN molecules typically encountered during HER in undivided electrochemical cells. Owing to the highly positive E ₀ of RRs, spontaneous BOR was successfully conducted in organic solvents, thereby circumventing the hydrolysis of BI intermediates and BN products that commonly occurs in alkaline solutions. Among the solvents tested, hexane delivered the highest BN yield and selectivity, attributed to the stronger adsorption of BA on RR surface and the suppressed self‐reduction behavior—factors that collectively contributed to the high overall efficiency. In situ characterizations confirmed that NiOOH species (Ni³⁺ cations) served as the reactive sites responsible for driving BOR. Additionally, this electrolyte‐free approach simplified post‐reaction purification by allowing straightforward solvent removal through simple distillation. The spontaneous conversion of BA derivatives bearing diverse functional groups into the corresponding nitriles with high yields highlights the flexibility of RR system for meeting specific market requirements. Collectively, these findings provide a sustainable and modular platform for green nitrile synthesis and fuel production, opening new avenues for renewable energy‐driven chemical manufacturing.

Author Contributions

Tzu‐Ting Weng: writing – review and editing, data curation, formal analysis, validation, and investigation. Song‐Chi Chen: writing – review and editing, data curation, formal analysis, validation, and investigation. Jing‐Mei Huang: formal analysis, investigation, data curation, and software. Jyh‐Pin Chou: methodology, writing – original draft, formal analysis, data curation, investigation, and software. Ying‐Rui Lu: methodology, resources, formal analysis, data curation, validation, and investigation. Chih‐Jung Chen: conceptualization, methodology, funding acquisition, writing – original draft, writing – review and editing, supervision, project administration, resources, formal analysis, investigation, and validation.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Detailed experimental procedures, characterization methods, and additional figures and tables.

Supporting File 1: anie72791‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.