Photocatalytic C-N coupling from stable and transient intermediates for gram-scale acetamide synthesis

Reaction system design and efficiency evaluation

The commercial TiO₂ (P25) is applied as a representative photocatalyst to develop the photoredox route for C-N coupling, which only absorbs the photon and forms the active e⁻ and h⁺ in this system, excluding the specificity of other modified photocatalysts and establishes a general pattern for the superiority of the photocatalytic amides synthesis scheme. Other common photocatalysts are also investigated in the C-N coupling reaction, which processes similar properties, showing the universality of photocatalysts in C-N coupling (Supplementary Fig. 4 and Supplementary Note 3). At first, various photocatalytic routes, including co-oxidation, co-reduction, and combined redox reactions, are screened to examine the practicability of amide synthesis (Table 1). It is identified that the amide bonds can be constructed through the coupling of various C₁/C₂ and N₁ substances, in which significant promotion of the synthesis efficiency is observed in the reaction route of CH₃CH₂OH and NH₃ co-oxidation for CH₃CONH₂ production (109.24 μmol h⁻¹), remarkably exceeding the other photocatalysis routes. Therefore, CH₃CH₂OH and NH₃ reactants are assigned as the respective C- and N-source for the comprehensive efficiency evaluation and mechanism investigations.

After the optimization of the reaction parameters, including NH₃ concentration, CH₃CH₂OH proportion, O₂ proportion (in Ar), and catalyst dosage (Supplementary Figs. 5–9 and Supplementary Note 16), a high CH₃CONH₂ production rate is achieved at 105.61 ± 4.86 mmol g_cat⁻¹ h⁻¹ (Fig. 2a) from the co-oxidation of CH₃CH₂OH and NH₃, which is a significant progression as it reaches the hundred-mmol level for photocatalytic C-N coupling under ambient conditions. Then, as the complete redox reaction proceeds in a heterogeneous photocatalysis process, it is deduced that the co-oxidation reaction is contributed by the light-generated h⁺ and O₂-enabled active species, in which the detailed mechanism requires further investigation. Meanwhile, the hydrogen gas (H₂) is yielded by the cooperative e⁻-driven reduction reaction with the dehydrogenated hydrogen from CH₃CH₂OH and NH₃ (Supplementary Note 13 and Supplementary Figs. 10, 11), in which a minor amount of H₂ is detected due to competing O₂ activation and reduction for e⁻ consumption.

a catalyst dosage-dependent unit production rate; CH₃CONH₂ selectivity evaluation regarding the N-(b) and C-sources (c), respectively; d Efficiency comparison between different catalytic routes for photocatalytic C-N coupling, including the targets of production rate and selectivity, the corresponding research works are listed and cited in  Supplementary Information (Supplementary Table 3)^{54,55,56,57,58,59,60}; e Long-term stability test. f XRD pattern and the image (inset) of the collected CH₃CONH₂ sample with rotary evaporation after the long-term stability test. The respective standard curves for detecting the reaction species using ion chromatography (IC, NH₄⁺, NO₂⁻, NO₃⁻, CH₃COOH, and NH₂OH, Supplementary Figs. 31–35) are provided in  Supplementary Information. The error bars in (a–c) were drawn based on the calculated standard error of two parallel tests.

Subsequently, the CH₃CONH₂ selectivity is evaluated, corresponding to NH₃ (N) and CH₃CH₂OH (C), respectively (Supplementary Note 15). As depicted in Fig. 2b and Supplementary Figs. 12–20, the concentration of CH₃CONH₂ gradually increases along with the consumption of NH₃. It is noted that trace N₂ byproduct is generated at the first 30 min, which originates from the NH₃ peroxidation. However, as the reaction proceeds, the accumulation of N₂ is impeded due to the decrease of the NH₃ concentration, which illustrates that the controllable provision of the initial NH₃ concentration is vital to determine the oxidative pathways of the system. The concentrations of the other byproducts are exceedingly low, resulting in data points in Fig. 2b that are so closely aligned they appear to overlap, making individual distinctions hard to discern. Under the limited provision of NH₃ (100.00 mg L⁻¹), a near-complete conversion of NH₃ (97.67% ± 1.67%) for CH₃CONH₂ production is accomplished under the 60 mins’ light irradiation, presenting a superior CH₃CONH₂ selectivity (N) of 99.17 ± 0.39%. In addition, it is observed that the synthesis rate decreases after 45 min, which is reasonable due to the remaining NH₃ concentration being low after rapid consumption. Based on these results, the recovery of the reaction rate and maintaining of selectivity are expected by conducting cycled experiments through periodical input of limited NH₃.

Then, the CH₃CONH₂ selectivity regarding CH₃CH₂OH is also evaluated by comprehensively detecting the oxidative products in both the liquid and gaseous phases (Fig. 2c and Supplementary Figs. 21–25). The corresponding selectivity is ~100% at the first 30 min. Along with the consumption of NH₃, the generation of the trace byproduct (CH₃COOH) is observed due to the peroxidation of CH₃CH₂OH under low NH₃ concentration, which again verifies the importance of limited NH₃ provision to guarantee the superior selectivity. After the photocatalysis reaction for 1 h, the CH₃CONH₂ selectivity (C) is maintained at as high as 95.60 ± 0.22%. The reaction parameters of NH₃ concentration for the CH₃CONH₂ selectivity test regarding NH₃ and CH₃CH₂OH are set differentially. Specifically, a lower concentration of NH₃ (100.00 mg L⁻¹) is provided for its efficient conversion to evaluate the selectivity (N). While more NH₃ (800.00 mg L⁻¹) is included to impede the peroxidation of CH₃CH₂OH after the rapid NH₃ consumption, where objective evaluations can be established for the selective oxidation of NH₃ and CH₃CH₂OH for CH₃CONH₂ synthesis, respectively (Supplementary Note 17). These selectivity results imply that the key C- and N-intermediate must be clarified and optimized for the selective C-N coupling, which will, in turn, impede the individual peroxidation of CH₃CH₂OH and NH₃. Moreover, when compared with the conventional industrial synthesis route (Supplementary Table 2), the photocatalytic co-oxidation route for CH₃CONH₂ synthesis demonstrates significant progress in terms of synthesis conditions, raw material price, and synthesis efficiency.

After systematically optimizing the reactants and reaction parameters, the establishment of the highly effective and selective co-oxidation route of CH₃CH₂OH and NH₃ lays the theoretical foundation for developing the photocatalytic CH₃CONH₂ synthesis route. In comparison with the other reported routes for photocatalytic C-N coupling (Fig. 2d and Supplementary Table 3), it is significant that a high efficiency is accomplished by applying the photocatalytic co-oxidation routes, including the important indexes of synthesis rate of the targeted coupling products and the corresponding selectivity. Moreover, the as-constructed reaction system has been continuously operated with sufficient CH₃CH₂OH provision for 300 h (Supplementary Table 1), in which periodic NH₃ conversion for stable CH₃CONH₂ production is maintained, delivering a gram scale yield of CH₃CONH₂ (1.82 g, Fig. 2e). The collected solid sample, with rotary evaporation for its recovery, is verified as CH₃CONH₂ by the XRD and nuclear magnetic resonance (¹H NMR) technologies, consistent with the crystallography open database (COD #9007523, Fig. 2f) and chemical shift of the hydrogen atom (Supplementary Fig. 26) respectively. No deactivation of the P25 is observed based on the characterization results of the sample before and after the long-term stability test (Supplementary Figs. 27–30 and Supplementary Note 18).

Reaction coordinates for the selective oxidation of CH₃CH₂OH and NH₃

To identify the critical step of the co-oxidation of CH₃CH₂OH and NH₃. The variate-controlled in situ attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR, Supplementary Fig. 36) investigation is conducted by comparing the IR signals for individual ethanol oxidation (EOR), ammonia oxidation (AOR) and their co-oxidation reactions. As the individual EOR proceeds (Fig. 3a, the spectra for the adsorption equilibrium process in the dark are displayed in Supplementary Fig. 37 and Supplementary Table 4), the gradual consumption of CH₃CH₂OH (C-OH at 1603 and 1364 cm⁻¹) is observed³³, which generates the corresponding oxidative intermediates (C-O at 1657 and C = O at 1519 cm⁻¹)^34,35. It is found that the peroxidative product (COO⁻ at 1420 cm⁻¹)³⁶ is produced by the excess oxidative ability in the individual EOR. Similarly, the consumption of NH₃ (N-H at 3099 cm⁻¹)³⁷ yields the intermediated products of -NH₂ (1090 cm⁻¹)³⁸ and NH₂OH (1268 cm⁻¹)⁸ in the individual AOR (Fig. 3b, the spectra for the adsorption equilibrium process in the dark are displayed in Supplementary Fig. 38 and Supplementary Table 5), in which the peroxidative products are also observed (N-O at 1755 cm⁻¹, NO₂⁻ at 1530 cm⁻¹ and NO₃⁻ at 1626 cm⁻¹)^23,39. Since NH₂OH species are not detected during the reaction process (Fig. 2b, Supplementary Note 14, and Supplementary Fig. 20), it is clarified that NH₂OH is not desorbed from the catalyst surface but is instead converted to other intermediates and products.

Time-dependent IR signals for the individual ethanol oxidation reaction (EOR, a), ammonia oxidation reaction (AOR, b), and combined EOR and AOR (c), respectively; Normalized results of the IR signals of COO⁻ (d), NO₃⁻ (e), C=O (f), and -NH₂ (g), respectively. The full spectra of the adsorption and photocatalysis process are provided in Supplementary Information for individual EOR (Supplementary Fig. 37), individual AOR (Supplementary Fig. 38), and combined EOR and AOR (Supplementary Fig. 39), respectively.

Most importantly, as illustrated in the combined co-oxidation reactions of EOR and AOR (Fig. 3c, the spectra for the adsorption equilibrium process in the dark are displayed in Supplementary Fig. 39 and Supplementary Table 6), as the co-oxidation of CH₃CH₂OH (C-OH at 1364 cm⁻¹) and NH₃ (N-H at 3087 cm⁻¹) proceeds, the interaction of the C- (C=O at 1544 cm⁻¹) and N-intermediates (-NH₂ at 1159 and 1090 cm⁻¹) leads to the generation of the coupling products as indicated by the C-N bond formation at 1416 cm^{−1 40}. Therefore, it is identified that the amide bond (-CONH₂) can be effectively generated through the coupling of C=O and -NH₂-based intermediates. However, their exact structures require further investigation.

Then, the signals for the peroxidative byproducts (Fig. 3d, e) and critical coupling intermediates (Fig. 3f, g) are normalized for comparison between the individual and combined oxidation reactions. The detailed methods for normalization of the FTIR signals are presented in Supplementary Note 8. It is observed that more peroxidative byproducts, including COO⁻ from EOR (Fig. 3d) and NO₃⁻ from AOR (Fig. 3e), are accumulated in the respective individual oxidation reactions, which hinders the selective generation of key intermediates for C-N coupling. By the combination of EOR and AOR, the intensified signals for the intermediates of C=O (Fig. 3f) and -NH₂ (Fig. 3g) are noted, leading to the directed coupling rather than the individual peroxidation. Furthermore, since the CH₃CH₂OH is sufficiently provided, the continuous accumulation of C=O species is reasonable, which provides sufficient C-intermediate for coupling. On the contrary, the limited provision of NH₃ yields -NH₂ at the first 20 min. Then -NH₂ is consumed along with the decrease of NH₃ concentration, which follows the efficiency test result (Fig. 2b). Hence, it is deduced that the -NH₂-based species can be identified as the key N-intermediate which possesses higher reactive activity than that of the C=O species, due to its gradual accumulation and rapid consumption. It is again verified that continuous and effective C-N coupling can be accomplished by providing sufficient CH₃CH₂OH and periodically supplying limited NH₃. This approach not only generates C- and N-intermediates for coupling but also avoids their peroxidation.

Generation and coupling mechanism of the key intermediates

Based on the in situ ATR-FTIR investigation, it is acknowledged that the directed coupling of the C- and N-intermediates is vital to selectively producing CH₃CONH₂. Then, the generation pathways and coupling mechanism of these key intermediates should be revealed. The O₂-proportion-dependent CH₃CONH₂ synthesis efficiency is first evaluated to identify the oxidative driving force for the generation of these intermediates (Fig. 4a and Supplementary Figs. 40,41). It is observed that the CH₃CONH₂ yield increases along with the elevation of O₂ proportion (in Ar) from 0% to 75%, reaching a yield of 209.77 ± 7.68 μmol h⁻¹. However, a rapid decrease of CH₃CONH₂ yield (68.82 ± 1.56 μmol h⁻¹) is noted under excess O₂ provision (100%), which directly confirms that the O₂ molecules, as fundamental active species, require to be quantificationally provided for mild O₂ activation, which enables the precise generation of the coupling intermediates. In addition, the production of e⁻-driven H₂ is elevated at the O₂ proportion of 0% than that of 75%, which again confirms that the O₂ activation and reduction work as competing reactions with H₂ evolution (Supplementary Fig. 42).

a O₂ proportion-dependent CH₃CONH₂ yield; DMPO-trapping (with D-labeling, images above) and TEMP-trapping (images below) in situ EPR experiments under the O₂ proportion of 0% (b), 75% (c), and 100% (d) respectively. The error bars in (a) were drawn based on the calculated standard error of two parallel tests.

Then, a comprehensive iso-type (deuterium, D) labeled in situ electron paramagnetic resonance (EPR) experiment is conducted to dynamically track the selective EOR pathways and corresponding intermediates, by applying the 5,5-dimethyl-1-pyrroline N-oxide (DMPO) and 2,2,6,6-Tetramethylpiperidinooxy (TEMPO) as the respective trapping agents under different O₂ proportion (Fig. 4b–d)^41,42,43. The CD₃CD₂OD, rather than the other common reagents such as CH₃OH or H₂O, is applied as the only solvent for the D-labeled in situ EPR tests to precisely reproduce the selective EOR for acetamide synthesis. Under the 0% of O₂ proportion, it is found that the EOR proceeds under the oxidative driving force of h⁺, as evidenced by the successful detection of the oxidative intermediate of the DMPO trapped alkyl radical (DMPO-CD₃^●CDOD) and the absence of the O₂-enabled active species (Fig. 4b and Supplementary Figs. 43–45). Interestingly, DMPO trapped alkoxy radical (DMPO-CD₃CD₂O^●) and trace DMPO-CD₃^●CDOD are both detected under the condition of an O₂ proportion of 75%, which follows the optimum O₂ proportion parameter for acetamide synthesis (Fig. 4c, Supplementary Fig. 46, 47, and Supplementary Note 19), illustrating the direct relationship between CD₃CD₂O^● generation and acetamide yield promotion. There is a tautomerization transformation from CD₃^●CDOD to CD₃CD₂O^●, which is driven by the mild oxidative ability of singlet oxygen (¹O₂, Fig. 4c and Supplementary Fig. 48). The ¹O₂ is generated from either the activation of the triplet oxygen (³O₂) or the h⁺-triggered rapid oxidation of superoxide radical (^●O₂⁻)^44,45. Since ³O₂ and ^●O₂⁻ are difficult to distinguish in the current in situ EPR experiments, it is reasonable to infer that both of these two reaction pathways are involved for the generation of ¹O₂. Moreover, the generation of CD₃CD₂O^●, which is an O-centered active radical, implies that the C-OD bond in the CD₃CD₂OD molecule can be effectively activated by dehydrogenation, which yields the precursor of -C=O species for C-N coupling. As the concentration of O₂ increases, the signal intensity of ^●O₂⁻ radical gradually increases while the signal intensity of ¹O₂ gradually decreases. Finally, the excess provision of O₂ (100%) leads to the accumulation of the strong oxidative ^●O₂⁻ radicals (Fig. 4d, Supplementary Fig. 49, and Supplementary Note 20), in which no signals of the ¹O₂ are detected (Fig. 4d and Supplementary Figs. 50, 51). As the ^●O₂⁻ radicals bind to the reactants to generate peroxidative products, ¹O₂ species are no longer generated. Hence, the ^●O₂⁻-induced peroxidation is inevitable under such a high O₂ concentration, resulting in the decreased coupling efficiency.

By the quantified O₂ activation at the proportion of 75%, it is confirmed that the peroxidation of CD₃CD₂O^● is avoided. Therefore, it is reasonably deduced that stable C-intermediate of acetaldehyde (CH₃CHO), which possesses the critical -C=O group, can be generated through the hydrogen abstraction from CD₃CD₂O^●. The corresponding control experiment is thereby conducted by replacing CH₃CH₂OH with CH₃CHO as the initial reactant for C-N coupling (Fig. 5a, Supplementary Note 10, Supplementary Fig. 52, and Supplementary Table 7). It is noted that almost complete recovery of the CH₃CONH₂ yield is accomplished via the NH₃-CH₃CHO condensation (204.67 ± 6.58 μmol h⁻¹), consistent with that of the co-oxidation of NH₃ and CH₃CH₂OH (209.77 ± 7.68 μmol h⁻¹), which identifies the CH₃CHO as the stable C-intermediate, where the generation of CH₃CHO requires precise regulation to avoid the decrease of the acetamide. Furthermore, the key role of CH₃CHO is illustrated by the gas chromatography (GC, Fig. 5b) and corresponding mass spectra (MS, Supplementary Figs. 53, 54 and Supplementary Note 21) detection results. The provision of CH₃CH₂OH/CD₃CD₂OD feedstock results in the generation of H/D labeled acetaldehyde and acetamide, respectively, presenting a selective EOR along the route of CH₃CH₂OH → CH₃^●CHOH ↔ CH₃CH₂O^● → CH₃CHO → CH₃CONH₂. Also, it is clarified by the ¹H NMR (Fig. 5c, Supplementary Figs. 55 and 56 and Supplementary Note 22) results that the CD₃CONH₂ is detected with a deuterated ratio of 98.33%, which verifies the reliability and practicability of these designed D-labeled in situ experiments.

a Control experiments for CH₃CONH₂ synthesis by replacing CH₃CH₂OH with CH₃CHO as the C-source; GC (b) and ¹H NMR (c) results with D- (images right) and H-labeling (images left) respectively; d DMPO-trapping in situ EPR experiments with ¹⁴N- (image above) and ¹⁵N-labeling (image below) respectively; e in situ EPR spectra for e⁻-induced TEMPO consumption with (image below) and without (image above) NH₃ provision respectively; f HR-MS results for CH₃CONH₂ detection with ¹⁴N- (image above) and ¹⁵N-labeling (image below) respectively; g Proposed reaction pathways of CH₃CH₂OH and NH₃ co-oxidation for intermediates coupling and CH₃CONH₂ synthesis. The error bars in (e) were drawn based on the calculated standard error of two parallel tests.

After successfully identifying the EOR for accumulating the stable C-intermediate of CH₃CHO, the cooperative AOR mechanism is revealed by a series of ¹⁵N-/¹⁴N-labeled tracking experiments. The DMPO-^●14NH₂ and DMPO-^●15NH₂ are detected without O₂ provision, respectively (Fig. 5d, Supplementary Figs. 57, 58, and Supplementary Note 23)^46,47, manifesting that h⁺ is the dominant driving force for the mild oxidation of NH₃ for ^●NH₂ generation. By combining the in situ ATR FTIR and in situ EPR spectra (Fig. 3c, g, 4h), it is indicated that the absorbed -NH₂ is oxidized by h⁺ to produce ^●NH₂, which is involved in subsequent C-N coupling processes. Besides, further in situ EPR experiments are applied by using TEMPO as the indicator. Under the condition of individual TEMPO provision (Fig. 5e and Supplementary Fig. 59), a slight decrease of TEMPO signals is recorded due to the TEMPO reduction driven by e⁻ under light irradiation for 2 min On the contrary, near-complete TEMPO consumption is observed when adding NH₃ into the test solution (Fig. 5e and Supplementary Fig. 60). Since no other oxidative driving force is included due to the absence of O₂ (proportion at 0%), it is again confirmed that the selective AOR is dominantly driven by h⁺ for ^●NH₂ generation, which consumes more h⁺ and in turn provides extra e⁻ for accelerating TEMPO reduction.

By the combination of the efficiency test (Figs. 2b, 4a), the in situ ATR-FTIR results (Fig. 3c), as well as the ¹⁵N-labeled in situ EPR investigations for ^●NH₂ detection (Fig. 4h, i), it is acquired that no peroxidative N-containing intermediates or products are generated in the combined EOR and AOR under quantified O₂ activation, which leads to the ~100% selectivity (N) for CH₃CONH₂ synthesis. Hence, it is concluded that the ^●NH₂ is identified as the transient N-intermediate for C-N coupling. Since the reaction rate coefficient of NH₃-to-NH₂ oxidation is an order of magnitude faster than that of the NH₂-to-NH oxidation, the selective oxidation of NH₃ primarily yields NH₂-based intermediates or radicals than that of the NH^48,49. Moreover, the fast formation (~1 ps) and relatively long lifetime (>1 ns) endow the ^●NH₂ radicals with a high possibility to participate in the coupling reactions, in comparison with the other N-centered radicals possessing short lifetimes, such as ^●NH₄ (13 ps)^50,51, which again verifies that ^●NH₂ can act as the transient N-intermediate for C-N coupling. The illustration for ^●NH₂ detection parameters, especially the O₂ proportion, is listed in Supplementary Note 24, which explains that the absence of the O₂ provision is indispensable for the detection of ^●NH₂. Based on these results, it is clarified that the controllable generation and fast addition of transient ^●NH₂ radicals on stable CH₃CHO intermediate should contribute directly to the efficient and selective C-N coupling for CH₃CONH₂ synthesis. The AOR pathways, along with the route of NH₃ → ^●NH₂ → CH₃CONH₂, are then verified by the ¹⁴N/¹⁵N labeled high-resolution MS (HR-MS, Fig. 5f, Supplementary Figs. 61,62, and Supplementary Note 25) results, in which both CH₃CO¹⁴NH₂ and CH₃CO¹⁵NH₂ are detected, by introducing ¹⁴NH₃ and ¹⁵NH₃ (in CH₃CH₂OH) as the feedstocks respectively.

After these comprehensive mechanism investigations of the reaction pathways and coupling intermediates, the heterogeneous photocatalytic scheme for CH₃CONH₂ synthesis is proposed (Fig. 5g). Under the sufficient provision of CH₃CH₂OH and quantified O₂ activation (75% in Ar), the CH₃CHO, from selective oxidation of CH₃CH₂OH, is generated and accumulated as the stable C-intermediate for coupling. Meanwhile, the limited and periodic provision of NH₃ leads to the controllable generation and fast addition of transient N-intermediate (^●NH₂) on stable CH₃CHO intermediate, in which the rapid consumption of ^●NH₂ for C-N coupling avoids its peroxidation for side product generation. The coupling of long-lived C-intermediate and short-lived N-intermediate ensures a high selectivity and yield rate of CH₃CONH₂. The C-N coupling selectivity can be effectively maintained by the tailored oxidative ability and oxidative species, including h⁺ and e⁻-driven O₂-enabled active species (¹O₂ and ^●O₂⁻), in which the excess e⁻ contributes to the H₂ evolution to construct a complete photocatalytic redox reaction. For the overall photocatalytic synthesis pathway of CH₃CONH₂, optimizing the interaction between the initial reactive species and the photogenerated charge carriers is of crucial importance to achieve carrier-driven radical generation. By fully leveraging the radical characteristics of lifetime and migration distance, radical-mediated photocatalytic reactions can proceed effectively, which ensures the high efficiency of mass transfer and the sufficient utilization of light energy, thereby achieving a significant enhancement of the redox efficiency.

Here, we have demonstrated a compelling solar-driven photocatalytic synthesis system. By combining mild co-oxidation of CH₃CH₂OH and NH₃, efficient and selective access to CH₃CONH₂ synthesis is achieved with heterogeneous photocatalysis technology. Through the precise regulation of the O₂ activation and e⁻-h⁺ carriers, the CH₃CONH₂ yield reaches the gram scale with satisfying production rate, selectivity, and stability. Comprehensive mechanism investigation results indicate that the controllable generation and fast addition of the transient ^●NH₂ radical on stable CH₃CHO intermediate contributes essentially to this catalytic performance in acetamide synthesis. The C-N coupling from stable and transient intermediates could provide scientific and technical feasibility for the solar light-driven synthetic fields, which, with further promotion, might be industrially and economically practical, thereby illuminating valuable opportunities for the mild-photocatalysis-driven synthetic industry.