Paired electrolysis enables para-C–H amination of phenols with nitroarenes and mechanistic visualization via multifunctional electrochemical mass spectrometry

Abstract

Direct C–N bond formation from nitroarenes offers an appealing alternative to conventional routes that rely on aniline intermediates, yet existing metal-catalyzed approaches often suffer from narrow substrate scope and the need for stoichiometric additives. Herein, we report a paired electrolysis strategy that enables highly selective para-C–H amination of phenols with nitroarenes, delivering unprotected p-hydroxy diphenylamines in a single step without transition-metal catalysts or external reductants. Central to this advance is a multifunctional automated injection electrochemical mass spectrometry (AIEC-MS) platform, which accelerates reaction screening, identifies productive electrochemical conditions, and reveals broad compatibility with diverse phenolic and nitroarene substrates. Real-time EC-MS analysis captures key transient intermediates—including a putative aryl-nitrene species—thus providing mechanistic visualization that clarifies how anodic and cathodic events cooperatively generate the reactive partners. Together, these insights demonstrate both the synthetic utility and mechanistic distinctiveness of paired electrolysis for direct C–N bond construction.

Direct C–N bond formation from nitroarenes offers an appealing alternative to conventional routes that rely on aniline intermediates. Here, the authors report a paired electrolysis strategy that enables highly selective para-C–H amination of phenols with nitroarenes, delivering unprotected p-hydroxy diphenylamines in a single step without transition-metal catalysts or external reductants.

Introduction

C–N bond formation has long been a cornerstone of synthetic chemistry^1,2, traditionally achieved through methods like Buchwald-Hartwig^3–9, Ullmann-type^10–13, and Chan-Lam^14–18 couplings. These approaches typically rely on amines derived from the reduction of nitro compounds. Direct utilization of nitro compounds as starting materials for C–N bond formation offers economic benefits and broad substrate compatibility^19–28. Recent progress has shown the effectiveness of selective reductive amination of nitroarenes using iron²⁸ and phosphorus catalysts^29–32, demonstrating their utility in C–N bond synthesis. This strategy was further refined by integrating electrochemistry with reductive amination, enabling the coupling of nitrobenzene with phenylboronic acid without the need for external reducing agents³³. Despite these advancements, these reactions still depend on functionalized aryl reagents as coupling partners for nitrobenzene. An emerging strategy is C–H bond amination, which provides a compelling route to construct C–N bonds from readily available substrates^34–37.

p-Hydroxy diphenylamine serves as a key building block for dyes, pesticides, and ethanol fuel antioxidants, and this scaffold is widely found in both pharmaceutical agents and natural products (Fig. 1a)^38–42. Typically, its synthesis from p-nitrophenol requires acetyl protection of the phenol hydroxyl group, adding complexity to the process (Fig. 1b)⁴³. Although methods for synthesizing hydroxyl-compatible compounds exist, they often involve expensive metal catalysts and Grignard reagents^44,45. The direct arylation of aminophenol from nitrobenzene offers significant advantages, while Weinmayr’s direct arylation from nitrobenzene demands high-pressure hydrogen and ultrapure hydrofluoric acid⁴⁶. Developing a mild and selective C–H bond amination of phenols using nitrobenzene therefore remains an attractive yet challenging goal.

Fig. 1. Research background and overview of the research work.

a Important molecules containing diphenylamine structures. b Generic synthesis of p-hydroxy diphenylamine and palladium-catalyzed reaction. c Our paired electrolysis strategy for nitroarene-phenol coupling.

Electrochemistry operates under mild conditions, eliminating the need for external oxidizing or reducing agents, and facilitates both oxidation and reduction within a single system^47–50. Paired electrolysis enables simultaneous oxidative and reductive transformations at the same electrode, effectively overcoming mass transfer limitations that typically hinder such reactions^51–54. While the electrochemical reduction of nitroarenes to generate reactive intermediates⁵⁵ and the C–H activation of phenol under electrochemical conditions^56–59 have been individually studied, their combined application remains underexplored.

Herein, a paired electrolysis method is developed for the para-C(sp²)–H amination of phenols with nitroarenes, enabling the efficient synthesis of p-hydroxy diphenylamines (Fig. 1c). This approach is guided by a multifunctional EC-MS platform, which streamlines reaction optimization and demonstrates broad substrate compatibility. Moreover, Real-time EC-MS monitoring facilitates the real-time detection of key intermediates, including aryl nitrene, providing valuable mechanistic insights into electrochemically driven C–N bond formation. This integrated system not only enhances the understanding of the underlying reaction mechanisms but also expands the scope of electrochemical transformations in organic synthesis, demonstrating the versatility and potential of this methodology for constructing C–N bonds from readily available substrates.

Results and discussion

Electrochemical mass spectrometry (EC-MS) has emerged as a valuable technique^60–66 for efficient reaction screening and reactive intermediates characterization in synthetic chemistry^67–76. In this work, we developed an onsite EC-MS strategy for rapid reaction screening and mechanistic investigation. This method combines an automated sample introduction system with in situ EC-MS analysis⁷⁷, allowing for batch detection of reaction samples with minimal material consumption. The rapid and cost-effective approach makes it ideal for identifying key products and potential reactants. Moreover, by utilizing high-resolution mass spectrometry (HR-MS) and tandem mass spectrometry (MS/MS), the onsite EC-MS strategy enables the identification and characterization of active intermediates during electrochemical reactions, providing compelling evidence for elucidating reaction mechanisms.

As illustrated in Fig. 2a, we developed an in situ EC-MS system by integrating a coaxial flow electrolytic cell, composed of stainless-steel capillaries and platinum wires, with atmospheric pressure chemical ionization (APCI). During EC-MS detection, the reaction solution is introduced to flow through the electrolytic region, and any intermediates or products generated are immediately directed to the mass spectrometer by sonic spray for APCI-MS analysis. The APCI based in situ EC-MS system is also coupled with an autosampler capable of automated batch injection and mass spectrometric analysis (Fig. 2b further device details are provided by Fig. S2).

Fig. 2. Introduction to mass spectrometry instruments and experimental procedures.

a Schematic of the APCI-EC-MS setup; b Photograph of the AIEC-APCI-MS device; c Electrochemical Mass Spectrometry Screening Research Strategy; and d Electrochemical Mass Spectrometry Screening Process.

By utilizing the automated injection in situ EC-MS system (AIEC-MS) for reaction development, we established an electrochemical reaction and mechanistic investigation strategy (Fig. 2c). Initially, reaction solutions containing nmol-level candidate substrates were prepared and subjected to batch mass spectrometry screening using AIEC-MS. By monitoring specific m/z channels, we rapidly identified potentially reactive substrates. Furthermore, the integration of high-resolution and tandem mass spectrometry allowed for the in situ identification and structural elucidation of reaction intermediates, providing robust mechanistic insights. Specifically, to evaluate the reactivity of nitrobenzene, we tested a range of substrates with 4-nitrotoluene using a simple electrochemical system guided by our analysis strategy. Screening reactive candidates was performed by establishing m/z channels based on potential products from candidate substrates and nitroarenes, and matching these channels with sample ion flow diagrams (Fig. 2d). Among ten potential reactants (S-1 to S-10, as shown in Fig. S3), phenol was identified as capable of undergoing electrochemical reductive coupling with nitrobenzene to yield the para-coupled product 4i, as confirmed by NMR analysis. We extended this approach to various substituted phenols (S-10-1 to S-10-4) and consistently observed mass spectrometric signals corresponding to these phenols and their coupling products. These results demonstrate that the reaction is compatible with a broad range of substituted phenols, showcasing the versatility and wide applicability of our electrochemical synthesis method.

Based on the AIEC-MS results, we optimized the reaction conditions to achieve selective C(sp²)–H bond amination of phenol using nitrobenzene as the amino source via paired electrolysis in an undivided cell (Table S1). We first examined various electrode combinations (Table S1, Entries 2-5) and found that electrode choice significantly affected the reaction outcome. Changing the anode reduced the product yield to 24%, while using nickel or platinum as cathodes resulted in yields of 53% and 44%, respectively. Modifying both anode and cathode materials further decreased the yield to 22%. Factors such as current intensity and electrolyte selection had a minimal impact on the yield of 4i (Table S1, Entries 6-8). Hexafluoroisopropanol (HFIP) proved essential for target product formation, whereas trifluoroethanol, a similar solvent, produced the product with a lower yield (Table S1, Entries 9-11). A control experiment confirmed that electrical input was essential for the reaction to proceed (Table S1, Entry 12).

We then investigated the performance of various nitroarenes to evaluate their suitability as substrates (Fig. 3). Halogen-containing functional groups (Br, Cl, F) were well-tolerated, delivering the target products with good yields (4a-4c). Substrates with diverse substitution patterns and multiple halogen atoms (4e-4g) consistently gave positive results. Notably, even fluorinated nitrobenzenes, which are less commonly reported, achieved a 70% yield (4c). Substrates with multiple fluorine atoms and CF₃ groups (4d, 4h) also readily produced the desired compounds. Electronic effect analysis revealed that substrates with electron-withdrawing groups (4i-4l) outperformed those with neutral or electron-donating groups. The position of the electron-withdrawing group (4p) had minimal impact on the reaction outcome. Substrates with electron-neutral groups (4m-4o) consistently yielded satisfactory results. Representative functional groups frequently employed in C–N bond formation were successfully accommodated (4q-4u). Substrates bearing long-chain ketones (4v) and sulfonylamine moieties (4w), commonly encountered in organic synthesis, delivered 41% and 74% yields, respectively. Benzoheterocyclic substrates (4x) and those bearing additional functional groups (4y) afforded the corresponding products in good yields. Even amides and aliphatic halides (4z, 4aa) were effectively converted into the target compounds in moderate yields. The inherent selectivity of electrochemistry enabled the reduction of a single nitro group in dinitro substrates (4ab-4ad), affording the desired products with high specificity.

Fig. 3. Scope of nitrobenzene substrates.

Demonstration of the reaction effects of nitroarenes with different substituent groups. a Reaction conditions: carbon cloth anode (15 mm × 15 mm × 0.1 mm), platinum plate cathode (15 mm × 15 mm × 0.1 mm), constant current = 10 mA, 1a (0.10 mmol), 2a (3.0 equiv), ⁿBu₄NBF₄ (0.20 mmol, 0.033 M), DCE (3 mL), HFIP (3 mL), room temperature, N₂, 8 h. undivided cell. Isolated yields. b platinum plate cathode (15 mm × 15 mm × 0.1 mm), carbon cloth anode (15 mm × 15 mm × 0.1 mm), 1a (0.20 mmol), 2a (3.0 equiv), 5 h. c 6 h. d platinum plate cathode (15 mm × 15 mm × 0.1 mm), carbon cloth anode (15 mm × 15 mm × 0.1 mm), 2a (1.0 mmol). e 11 h. f 12 h.

Substrates substituted with deuterium also showed favorable reactivity. The reaction efficiently tolerated substrates with multiple deuterated sites (5a, 5b) without significant changes in yield. Deuteration studies on substrates with specific functional groups (5c-5e) further demonstrated that the reaction accommodates both the functional groups and deuteration with minimal impact on yields.

We further explored the compatibility of phenol substrates by examining halogenated derivatives (Fig. 4a). Fluoro-, chloro-, and bromo-substituted phenols (6a-6c) consistently yielded the desired products with efficiencies ranging from 43% to 54%. Additionally, phenols bearing electron-donating groups (6d, 6e) produced favorable yields. Substrates with substituents at various positions on the phenolic ring (6f, 6g) demonstrated comparable efficiencies, achieving yields between 50% and 54%. Furthermore, unsaturated phenols (6h, 6i) produced moderate yields, suggesting their potential for further modification and functionalization. Additionally, phenols with amide substituents were also compatible with the reaction, and the desired product was afforded(6j).

Fig. 4. Scope of phenol substrates, flow scale-up experiments, and control experiments.

a Nitrobenzene coupling substrate expansion. a Reaction conditions: carbon cloth anode (15 mm × 15 mm × 0.1 mm), platinum plate cathode (15 mm × 15 mm × 0.1 mm), constant current = 10 mA, 1a (0.10 mmol), 2a (3.0 equiv), ⁿBu₄NBF₄ (0.20 mmol, 0.033 M), DCE (3 mL), HFIP (3 mL), room temperature, N₂, 8 h. undivided cell. Isolated yields. b platinum plate cathode (15 mm × 15 mm × 0.1 mm), carbon cloth anode (15 mm × 15 mm × 0.1 mm), 1a (0.20 mmol), 2a (3.0 equiv), 5 h. c 1a (3.0 equiv), 2a (0.20 mmol), 10 mA, 2 Hz, D = 80%, 5 h. b Scale-up experiment in an electrochemical continuous flow cell and application in the synthesis of p-hydroxy diphenylamines. For details of the reaction, see Fig. S7. c Control experiments.

Our study extended beyond phenolic substrates. Using a high-throughput screening system, we broadened the substrate scope to include ethylbenzene derivatives (6k) and phenyl methyl ether substrates (6l, 6m). This expansion facilitated the synthesis of identical products via two distinct synthetic pathways (4j), thereby demonstrating the versatility and broad applicability of our electrochemical amination methodology.

Subsequently, we focused on scaling up the reaction. After optimizing parameters such as flow rate, we performed the reaction in a flow cell at the gram scale, demonstrating compatibility with aldehydes (Fig. 4b). Compared to conventional multi-step protocols, our approach eliminates the need for metal catalysts, reduces overall reaction time, and achieves higher yields. In certain instances, we observed aniline as a reduction byproduct and investigated its potential involvement in the reaction mechanism. Control experiments involving nitrosobenzene, phenylhydroxylamine, and aniline revealed that while nitrosobenzene slightly decreased the product yield, aniline did not facilitate the formation of the desired product (Fig. 4c). Real-time EC-MS experiments corroborated these results, indicating that aniline does not participate in the primary reaction pathway.

Cyclic voltammetry (CV) experiments were performed to elucidate the electrochemical behavior of the reaction, revealing notable shifts in the reduction potentials of the substrates (Fig. 5a). In the absence of hexafluoroisopropanol (HFIP), the reduction peak for substrate 1i appeared at a lower potential. Upon the initial addition of reagent 2a, no significant reduction peaks were observed. However, the subsequent introduction of HFIP resulted in the emergence of clear reduction peaks. A comparative analysis of the onset potential for nitrobenzene reduction revealed a value of −1.4 V under standard conditions, while a notably more positive onset potential of −0.7 V was observed in the presence of HFIP., indicating enhanced electron affinity. This shift suggests that HFIP facilitates the stabilization of the reduced species, thereby promoting the electron transfer necessary for the reaction. These findings highlight the critical role of HFIP in modulating the electrochemical environment to favor the desired reduction processes. Linear sweep voltammetry (LSV) under single-electrode conditions showed that compound 2a exerts only a negligible influence on the cathodic reduction of 1i (Fig. 5b). By contrast, square-wave voltammetry (SWV)—which records both cathodic and anodic events—revealed that the presence of 2a markedly accelerates the formation of the reduced product of 1i (Fig. 5c).

Fig. 5. Electrochemical measurements and KIE experiments.

a CV test conditions: Unspecified parts are consistent with the general reaction conditions. Blank: DCE, HFIP and ⁿBu₄NBF₄. b Effect of phenol on the cathode (LSV). c Effect of phenol on the reaction (SWV). d Parallel reactions with 2a and D₃-2a of kinetic isotope effect experiments. e Kinetic profiles under different concentrations of 1i. f Kinetic profiles under different concentrations of 2a.

To assess whether the activation of the C–H bond in phenol controls the reaction rate, competitive intermolecular experiments were conducted between phenol and D₃-phenol under standard conditions (Fig. 5d). The analysis of the kinetic isotope effect (KIE) curve yielded a value of 1.18, indicating that C–H bond cleavage in phenol is not the rate-determining step. Further kinetic studies demonstrated that the reaction follows zero-order kinetics to nitrobenzene (1i) and first-order kinetics to phenol (2a) (Fig. 5e, f).

To deepen our understanding of the coupling process between nitroaromatics and phenol, we utilized in situ EC-MS to study the reaction mechanism (Fig. 6a). By comparing the state of 1i before and after electrolysis, we observed that 1i ([M + H]⁺, m/z 138.0550) was converted into several compounds, including nitrosobenzene (1i-1, [M + H]⁺, m/z 122.0600), N-(p-tolyl)hydroxylamine (1i-2, [M + H]⁺, m/z 124.0757), p-methylaniline (1i-3, [M + H]⁺, m/z 108.0808), the less common N-oxide (1i-4, [M + H]⁺, m/z 106.0651), an intermediate (1i-5, [M + H]⁺, m/z 198.0913), and the product (4i, [M + H]⁺, m/z 200.1070). For clarity, the HR-MS spectra of each substance before and after electrolysis are presented separately in Figure S4.

Fig. 6. Identification and structural analysis of intermediates by online mass spectrometry.

a HR-MS before and after electrolysis. b Extracted ion chronogram for the electrolysis experiment. c HR-MS/MS of 1i-1, 1i-2, 1i-3 and 1i-4.

During a one-minute electrolysis experiment (Fig. 6b), we observed temporal variations in the mass spectrometric signals of reaction-related intermediates and products. Approximately 0.15 min before initiating electrolysis, signals for the reactant 1i and trace amounts of the product 4i were detectable. After electrolysis commenced at 0.15 min, a slight decrease in the mass spectrometric signal for 1i was noted, along with a marked increase in signals for 4i and intermediates 1i-4 and 1i-5. These changes in mass spectrometric signals confirm their close association with the electrochemical reaction and suggest that nitrobenzene can be reduced to N-oxides during the process. To further investigate the role of intermediates 1i-1 and 1i-3, we introduced them directly into the reaction. Interestingly, while reactions involving 1i-1 resulted in a significant decrease in the yield of 4i, no 4i was detected in the reaction involving 1i-3, suggesting that 1i-3 has a limited role in the reaction. HR-MS/MS analysis was also conducted to identify intermediates 1i-1 to 1i-4 (Fig. 6c), confirming that the structure of 1i-4 matches the nitrene structure identified in previous work⁷⁸.

The quasimolecular ion of 198.0918 caught our attention, and we identified it as intermediate 1i-5. To determine the structure of intermediate 1i-5, we employed HR-MS/MS (Fig. 7a), which revealed several neutral losses: 15.0235, 27.9949, 43.0184, and 55.0058. These losses matched exactly with the masses of CH₃ (15.0235), CO (27.9949), C₂H₃O (43.0184), and C₂HON (55.0053), confirming the presence of unsaturated alkanes containing N and O heteroatoms. The characteristic fragment ion peaks at 105.0533, 106.0649, and 107.0727 suggest that intermediate 1i-5 could have a p-methylaniline structure (Exact Mass = 107.0735), providing further evidence to validate its structure. Similar structures were also detected in isotope-labeling experiments; the analogous fragment peaks observed for D₃-5 confirm the accuracy of the preceding data. The unique structure of intermediate 1i-5 has not been previously reported, indicating the presence of a novel mechanism in this reaction.

Fig. 7. Online mass spectrometry control experiments.

a HR-MS/MS of intermediate 1i-5 and D₃-5. b Differences between online and offline tandem mass spectrometry of product 4i. c Deuteration experiment of phenol.

Moreover, product 4i ([M + H]⁺, m/z 200.1070) was detected by in situ EC-MS, revealing that the HR-MS/MS spectra of 4i obtained online and offline differed (Fig. 7b). The fragment peaks at 106.0652, 107.0730, and 108.0445 were observed in the online mass spectrum, with 107.0730 being the base peak. In contrast, the offline spectrum showed a much weaker intensity for the fragment ion 107.0730. Using D₃-1i instead of 1i in the reaction yielded similar results, confirming this phenomenon (Fig. S6). This led us to hypothesize that intermediate 1i-6, an isomer of 4i, was detected by in situ EC-MS. To explore the relationship between these fragment ions and the parent ion, we employed Quantum Chemical Mass Spectrometry (QC-MS), an open-source tool for fully automated calculations of positive ion collision-induced dissociation mass spectra of singly charged molecular ions^79,80. QC-MS simulations showed a strong intensity for the fragment ion 107.0735 of intermediate 1i-6, highlighting its contributions to the online tandem mass spectrum (Fig. S23).

Using tritiated phenol D₃-2a as the starting substrate, we identified a target product with an exact mass of 202.1195, incorporating two deuterium atoms. Tandem mass spectrometry confirmed the presence of structure D₂-6, analogous to intermediate 1i-6 (Fig. 7c). These findings indicate that product formation requires both reduction and hydrogenation, whereas direct coupling involving nitrosobenzene would necessitate reduction and dehydration. The differences suggest that nitrosobenzene does not directly participate in forming intermediate 1i-6 (Fig. S24). Instead, intermediate 1i-6 is generated through the reduction and hydrogenation of intermediate 1i-5. Additionally, nearly all detected intermediates corresponding to 1i-5 lack three deuterium atoms, indicating that this step is the sole pathway to 1i-6. Among nitrobenzene reduction intermediates, only aryl nitrene 1i-4 can eliminate the ortho hydrogen of phenoxide via C–H bond insertion during coupling. Therefore, we propose that aryl nitrene 1i-4 directly contributes to the formation of intermediate 1i-5.

In subsequent experiments, we replaced nitrobenzene with nitrosobenzene and performed online MS analyses. As shown in Fig. 8a, during the electrochemical reduction of nitrosobenzene (1I-1), the mass spectrum before electrolysis, obtained via EC-APCI-MS, revealed a characteristic peak for 1I-1 at [M + H]⁺, m/z 108.0442. Upon initiation of the electrolysis, mass spectral signals corresponding to a series of hydrogenation-dehydration products were detected, including phenylhydroxylamine ([M + H]⁺, m/z 110.0600), aniline ([M + H]⁺, m/z 108.0442), as well as the coupling intermediate 1I-5 ([M + H]⁺, m/z 184.0757) and the product 4I ([M + H]⁺, m/z 186.0913). These results confirm that nitrosobenzene undergoes reduction, analogous to nitrobenzene, without altering the intermediate profile of the reaction. Importantly, the post-electrolysis mass spectrum also revealed the generation of the nitrene intermediate 1l-4 ([M + H]⁺, m/z 92.0495), suggesting that nitrosobenzene acts as an intermediate in this system, undergoing further reduction to form the nitrene species.

Fig. 8. Validation experiments and proposed mechanisms.

a Intermediates for electrochemical nitrosobenzene coupling reactions by APCI-EC-MS; b Synthesis of aziridines by olefin capture of nitrogen; c Identification of Cadogan reaction by APCI-EC-MS, d Proposed reaction mechanism based on the in situ EC-MS results.

As a prerequisite for this reaction pathway, we sought to determine whether nitrobenzene undergoes further reduction within the reaction system. According to the LSV experiment shown in Fig. S16, the reduction potential of nitrotoluene (1i) is approximately −0.52 V, corresponding to the initial reduction potential required for converting 1i into nitroso-toluene. We maintained cathodic potentials of −0.40 V, −0.60 V, and −1.0 V for 4 h, respectively (Fig. S25). After the reaction, only the system with a cathodic potential of −1.0 V detected product formation. This result indicates that nitrobenzene must undergo further reduction to proceed with the subsequent reactions.

To confirm that the active intermediate has a sufficiently long lifetime in paired electrolysis, we employed a combination of a separation cell and alternating current (AC) to conduct the coupling reaction (Fig. S26). Due to the alternating nature of AC, where the polarity switches between positive and negative, only one type of electrode reaction can occur at a time in each chamber of the separation cell. Consequently, the reduction of nitrobenzene and the oxidation of phenol cannot occur simultaneously. By increasing the AC frequency to 5 Hz, we observed significant product formation in the system. Considering that the electrode polarity switches every 0.01 s at this frequency, we conclude that in the original integrated system, nitrene possesses a sufficient lifespan to facilitate the coupling reaction.

To establish a direct correlation between nitrene formation and electrochemical reactions, we designed an intermediate capture experiment (Fig. 8b). When 1,2-diphenylvinyl (7a) was used to replace phenol in the reaction, the formation of aziridine (7b, [M + H]⁺, m/z 286.1590) was detected through online mass spectrometry, and HR-MS/MS detected its neutral losses of 78.0481, 92.0638, 107.0749, and 178.0804 correspond to the exact masses of C₆H₆ (78.0470), C₇H₈ (92.0626), C₇H₉N (107.0735) and C₁₄H₁₀ (178.0783), suggesting that 7b has a structure of benzene ring and N atom. In the electricity on-off experiment, the extracted ion chronograms (EIC) demonstrated that the generation of both the nitrene intermediate and 7b was dependent on the applied current, confirming that the nitrene detected in our system arises from the electrochemical reaction and is not an artifact of the mass spectrometry analysis. Additionally, when benzaldehyde was used directly in the reaction instead of phenol, we successfully isolated the oxazolidine tricyclic product. This reaction was also monitored by online mass spectrometry, which further confirmed that the generation of the nitrene intermediate and the product was electricity-dependent.

The Cardogan reaction, an established method for carbazole synthesis, provides a useful model for demonstrating the involvement of nitrene intermediates^81–83. We adapted this reaction to a system comprising nitrobenzene and phenol (Fig. 8c). Due to its intramolecular nature, we utilized a magnesium electrode as a sacrificial anode. Upon completion of the reaction, we isolated product 7f with a yield of 41%. In situ EC-MS analysis revealed that intermediates, including hydrogenation products 7e-1 ([M + H]⁺, m/z 184.0757) derived from nitrobenzene, were associated with the electrochemical process. This behavior, resembling that of the template reaction, supports the conclusion that our system effectively reduces nitrobenzene to generate nitrene 7e-2 ([M + H]⁺, m/z 168.0808), which then participates in the subsequent cyclization reaction to form product 7f ([M + H]⁺, m/z 168.0808). EIC showed that the generation of 7e-2 & 7f and other intermediates were dependent on electrolysis and confirmed the intermediates’ mass spectrum signals caused by electrochemistical process.

Based on the intermediates identified by in situ EC-MS and our experimental results, the proposed reaction mechanism is outlined in Fig. 8d. Initially, substrates 1i and 2a undergo paired electrolysis at the electrodes. Substrate 1i is reduced and deoxygenated in two steps at the cathode via nitroso intermediate 1i-1, leading to the formation of aryl nitrene 1i-4 in a monoclinic state. Concurrently, phenol undergoes oxidation at the anode, forming a radical intermediate 2a-1, which is further oxidized to generate the cationic intermediate 2a-2. The monoclinic aryl nitrene 1i-4 then inserts into the C–H bond of cationic intermediate 2a-2, resulting in their coupling and the release of a hydrogen ion, forming intermediate 1i-5. Intermediate 1i-5 is subsequently reduced at the cathode to form intermediate 1i-6, which then re-aromatization to produce the target product 4i.

In conclusion, we have established a paired electrolysis strategy that enables efficient C(sp²)–H amination of phenols with nitroarenes, providing a straightforward route to p-hydroxy diphenylamines under mild and practical conditions. The integration of multifunctional EC-MS played a central role by accelerating substrate evaluation through rapid and reliable screening and by offering direct evidence for transient intermediates that clarify key steps of the coupling mechanism. This combination of electrochemical synthesis with real-time analytical capabilities enhances mechanistic understanding and supports the continued advancement of electrochemical reaction design. Looking ahead, further refinement of EC-MS–based screening and mechanistic visualization is expected to broaden the scope, efficiency, and applicability of future electrochemical transformations.

Methods

General information

All glass wares were oven dried at 110 °C for hours and cooled down under vacuum. Unless otherwise noted, materials were obtained from commercial suppliers and used without further purification. The carbon cloth, platinum wire and plugs were commercially purchased. The instrument for electrolysis was dual display potentiostat (DJS-292B) (made in China). Thin layer chromatography (TLC) employed glass 0.25 mm silica gel plates. Flash chromatography columns were packed with 200-300 mesh silica gel. Gradient flash chromatography was conducted eluting with a continuous gradient from petroleum to the indicated solvent, and they were listed as volume/volume ratios. MS data were acquired by LTQ Velos Orbitrap Elite mass spectrometer (Thermo Fisher Scientific Co., Ltd., USA). The temperature of the ion transmission tube was set lo 275 °C; the energy of the S-lens was set to 60 %; the number of micro scan was set to 1. MS/MS data were acquired by setting isolation window at 1.0 Da and normalized collision energies from 20 to 35 (units were defined by the instrument manufacturer; the collision gas was helium). The resolution of MS was set to 60000. All MS data were analyzed through the Qual browser (Xcalibur workstation, version 4.2.47). The high voltage power supply was from the mass spectrometer. Both the mass spectrometer and the external voltage power supply were well grounded. Unless otherwise stated, normalization is performed by dividing by the maximum value.

General procedures for the synthesis of p-hydroxy diphenylamines

In an oven-dried undivided two-necked cell (10 mL) equipped with a stir bar. Arylhydrazine 4-Nitrotoluene (27.4 mg, 0.2 mmol), Phenol (56.5 mg, 0.6 mmol), ⁿBu₄NBF₄ (65.8 mg, 0.2 mmol) were combined and added. The cell was equipped with a piece of platinum (15 mm × 15 mm × 1 mm) as the anode and a piece of carbon cloth (15 mm × 15 mm × 1 mm) as the cathode, and then charged with argon to replace air atmosphere. DCE (3.0 mL) and HFIP (3.0 mL) were added to the tube through a syringe. The reaction mixture was stirred and electrolyzed at a constant current of 10 mA under room temperature (25 ^oC) for 5 h (2.9 F/mol). When the reaction finished, the desired products were obtained in the corresponding yields after purification by flash chromatography on 200-300 mesh silica gel (petroleum: Ethyl acetate = 4: 1). 3m, 3s, 3t, 3u, 3v: Otherwise above general procedure, carbon cloth anode (15 mm × 15 mm × 0.1 mm), platinum plate cathode (15 mm × 15 mm × 0.1 mm), constant current = 10 mA, 8 h, 1a (0.10 mmol), 2a (0.60 mmol).

Mass spectrometric screening process of reaction substrates

Device Setup. The schematic diagram of the EC-APCI-MS apparatus is depicted in Figure S2a. The flow EC cell, featuring a coaxial structure, comprises a platinum wire inserted into a stainless steel tube. The platinum wire is exposed for ~2 mm, with the rest insulated to prevent shorting. At the left outlet of the stainless steel tube, a T-connector fixes the platinum electrode and introduces the reaction solution to the flow EC cell. On the right side of the stainless steel tube, pneumatic nebulization of the reaction solution is realized by using a T-connector and sheath gas. A stainless steel needle with voltage of 3.5 kV is placed near the pneumatic nebulization nozzle for atmospheric chemical ionization. The spray nozzle, APCI electrode tip, and mass spectrometer inlet are aligned linearly with a 3 mm gap between each of them. All connections are sealed with PEEK tubing of matching sizes to ensure proper integration. By connecting the inlet of the Electrochemical Mass Spectrometry to the outlet of an autosampler (Ultimate 3000, Thermo Scientific) via a fitting, we established an Auto Injection EC-MS system capable of rapid electrochemical reaction screening. Utilizing the batch processing function of the interaction software (Thermo Xcalibur 3.0.63) between the autosampler and the mass spectrometer, Auto Injection EC-MS facilitates automated sample injection through flow injection. Unless otherwise specified, acetonitrile (HPLC grade, degassed via ultrasonication prior to use) was used as the mobile phase, with a sample volume set to 5 μL, an injection interval of 2 min, a reaction voltage of 15 V, the voltage of the corona electrode of +3.5 kV, the flow rate of 10 μL/min, and the sheath gas pressure of 0.4 MPa.Preparation of Reaction Solutions. To screen potential substrates for coupling with p-nitrotoluene, a solution containing 1 mM p-nitrotoluene, 1 mM candidate substrate, and 10 mM lithium triflate in a 9: 1 acetonitrile/water mixture was prepared and placed in the vials. Additionally, a 1 mL solution of 10 mM lithium triflate in a 9: 1 acetonitrile/water mixture was prepared as a blank sample and also placed in the vials.

Online electrochemical reactions monitoring

Online electrochemical reactions were monitored by EC-MS, and the electrochemical reaction conditions have been slightly modified for compatibility with mass spectrometry systems. The solution of acetonitrile: water = 9: 1 (v: v) with 10 mM lithium triflate as supporting electrolyte containing 0.6 mM phenol and 0.2 mM p-nitrotoluene was loaded into EC-MS. The voltage was applied to that the current of electrolytic cell reaches 10 mA.

Deuterated nitrobenzene, Deuterated phenol, nitrosobenzene or aniline were also used as a reaction substrate for online monitoring, and the experimental conditions were consistent with the above.

Author contributions

H.Y. supervised the project. T.J.P., Z.Y.C., X.C., H.N.J., L.Y.L., Z.W.W., A.W.L., and H.Y. conceived the idea and designed the experiments. T.J.P., Z.Y.C., X.C., H.N.J., and L.Y.L. carried out all the experimental work. P.T.J., Z.Y.C., and X.C. contributed to data analysis and manuscript editing. Z.W.W., A.W.L., and H.Y. cowrote the manuscript. All authors discussed the results and assisted during the manuscript preparation.

Peer review

Peer review information

Nature Communications thanks Hamid Salehzadeh, Yunfei Zhang, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

All data are available in the main text or the Supplementary Information. All data are available from the corresponding author upon request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-70691-x.