Co–N₂P₂ single-atom catalysts enable efficient α-alkylation of aromatic ketones

Abstract

Carbonization of Co(PPh₃)₂Cl₂ on a covalent triazine polymer generates Co–N₂P₂ single-atom sites that enable efficient α-alkylation of aromatic ketones.

---

Single-atom catalysts (SACs) have emerged as a frontier material in heterogeneous catalysis owing to their atomically dispersed metal sites, well-defined coordination environments, and maximum atom utilization efficiency. The unique characteristics of homogeneous-like active sites and heterogeneous robustness have led to wide-ranging applicability in thermocatalysis, photocatalysis, and electrocatalysis.^1–5 3d-transition metals as single atoms on suitable supports have attracted increasing attention as promising alternatives to noble-metal SACs, due to their electronic structure, multiple accessible oxidation states, and strong affinity toward heteroatom coordination.^6,7

However, fabricating 3d-transition metals as isolated single atomic sites remains challenging, as these metals tend to aggregate into clusters or nanoparticles, rapidly.^8,9 Achieving stable single-atom sites requires rational support design with strong anchoring atoms accomplished through electronic and geometric stabilization.^10–12 Carbon-based materials have shown exceptional promise as supports due to their high surface areas, tunable porosity, chemical stability, and heteroatom functionalization ability.¹³ In particular, carbon supports derived from graphene^14,15 and covalent organic frameworks (COFs)¹⁶ have demonstrated excellent activity across organic, electrochemical, and photocatalytic transformations. Covalent triazine polymers (CTPs) have gained significant attention as precursors for nitrogen-doped porous carbons, owing to their outstanding thermal stability, structural tunability, and capability for incorporating heteroatoms such as N, B, F, P, and S.¹⁷ These, heteroatoms are not only abundant anchoring sites for metal atoms, but also modulate the electronic environment of the active metal sites, enhancing catalytic performance.¹⁸ To realize this, herein, we have successfully synthesized a Co single atom with P and N coordination (Co–N₂P₂) on porous carbon by introducing a Co(PPh₃)₂Cl₂ complex onto a covalent triazine polymer followed by subsequent pyrolysis. Atomic dispersion and local coordination were confirmed through HAADF-STEM and XAS analysis, respectively. Owing to the optimal configuration, the Co–N₂P₂ catalyst boosts α-alkylation of aromatic ketones with benzyl alcohol.

Initially, Co–N₂P₂, an isolated cobalt single atom catalyst with N and P coordination, was achieved through absorbing the Co(PPh₃)₂Cl₂ complex on a porous carbon and covalent triazene framework composite using a concise impregnation-pyrolysis process (Scheme 1 and Scheme S1). Initially, rice husk was activated by using H₃PO₄ and K₂CO₃ and carbonised to afford the porous carbon.¹⁹ Subsequently, the nitrogen-containing covalent triazine polymer (N-CTP) was synthesised using cyanuric chloride and melamine by a solvothermal method. The formation of N-CTP was confirmed by FT-IR spectroscopy (Fig. S1). The N–H stretching peak at ∼3400 cm⁻¹, along with the disappearance of –NH₂ bending (∼1630 cm⁻¹) and C–Cl vibrations, confirms the successful polycondensation. SEM images (Fig. S2) show a spherical morphology with a particle size of 0.8–1.2 µm. This N-CTP was combined with the rice husk-derived carbon and Co(PPh₃)₂Cl₂ by stirring and subsequent pyrolysis to afford Co–N₂P₂. During pyrolysis, N-CTP undergoes thermal condensation and partial decomposition, leading to the formation of an N-doped carbon matrix. Simultaneously, the rice husk-derived carbon restricts excessive graphitization and contributes high surface area, hierarchical porosity, and structural confinement that suppresses aggregation and promotes atomic dispersion of Co species during pyrolysis, while the phosphine ligand in the Co precursor provides the P for coordination.

Scheme 1 Schematic representation of the synthesis of atomically dispersed Co coordinated with N and P on hierarchically porous carbon derived from activated rice husk and a covalent triazine polymer (CTP) composite.

Powder X-ray diffraction (PXRD) studies were employed to examine the phase composition and crystallinity of the prepared Co–N₂P₂ (Fig. S3). The diffraction pattern of Co–N₂P₂ showed two broad peaks around 20–30° and 42–45° attributed to the disordered carbon (002) and (100) planes, respectively. Furthermore, the sharp diffraction peak observed at 22° (2θ) corresponds to the (002) reflection of turbostratic/hard carbon with an enlarged interlayer spacing, indicating partial graphitic ordering along with heteroatom (P/N) doping in the carbon matrix.²⁰ Meanwhile, no diffraction peaks for any crystalline metal species were observed, which may suggest the low content or high dispersion of the metal on the carbon matrix. The Co loading in the material was quantified using ICP-AES and found to be 0.2 wt% (Table S1). The morphology of Co–N₂P₂ was explored by scanning electron microscopy (SEM). In contrast to the N-CTP, the Co–N₂P₂ exhibits a highly porous, irregular, and rough surface structure, reflecting the intrinsic hierarchical porosity of biomass-derived carbon. The morphological differences highlight the complementary roles of CTP and biomass carbon supports in tailoring catalyst structure and performance (Fig. S2). Further analysis by high resolution transmission electron microscopy (HR-TEM) implies the absence of any crystalline Co NPs or clusters in the catalyst, Co–N₂P₂, suggesting that the Co species could be highly dispersed in the matrix (Fig. 1a). The SAED pattern, a diffuse ring pattern, of Co–N₂P₂ suggests that the material is highly amorphous in nature, which is in good agreement with the PXRD results (Fig. 1b). Furthermore, aberration corrected high-angle annular dark-field scanning transmission electron microscopy (AC-HAADF-STEM) was employed to illustrate the presence of Co species (Fig. 1c). The appearance of bright spots shows the presence of isolated Co atoms uniformly dispersed on the carbon matrix. The EDS mapping of Co–N₂P₂ also confirms the uniform distribution of metals on the carbon matrix (Fig. 1d). The porous nature and surface area of Co–N₂P₂ were examined by N₂ adsorption–desorption isotherm studies and the results revealed a type-IV curve suggesting mesoporous nature (Fig. S4). The BET surface area and pore size of the Co–N₂P₂ catalyst were 552 m² g⁻¹ and 2.1 nm, respectively.

Fig. 1 TEM image of Co–N₂P₂ (a), SAED (b), AC-HAADF-STEM image (c) and corresponding EDS elemental mapping images of Co–N₂P₂ (d).

The surface elemental composition of Co–N₂P₂ was investigated using X-ray photoelectron spectroscopy (XPS). The survey spectrum (Fig. S5a) reveals the presence of C, N, O, Si, P and Co in the Co–N₂P₂ catalyst. The C 1s spectrum (Fig. S5b) deconvoluted into four peaks at 284.52, 285.71, 288.83 and 288.75 eV, which were assigned to C–C/C=C, C–P, C–N and C–O, respectively.^21,22 Meanwhile, the N 1s spectra fitted into five peaks (Fig. S5c), suggesting the presence of pyridinic N (398.27 eV), M–N (399.66 eV), pyrrolic N (400.95 eV), graphitic N (402.12 eV) and the oxidized N (403.48 eV) species.^22,23 The pyridinic N acts as an anchoring site for metal atom deposition, and the graphitic N tailors the electron distribution of the active sites to enhance the adsorption during catalysis. The P 2p XPS spectrum (Fig. S5e) deconvoluted into three peaks corresponding to P–C (133.42 eV), P–N (135.02 eV), and P–Co (132.43 eV) species. The low-binding-energy P–Co peak confirms the formation of strong metal–phosphorus interactions.²⁴ The high-resolution Si 2p spectra (Fig. S5d) were deconvoluted into two characteristic peaks with binding energies at 103.78 and 102.48 eV attributed to Si–O and Si–O–C, respectively.²⁵ The Co 2p spectra (Fig. 2a) show the presence of two peaks at 796.41 and 781.98 eV attributed to Co 2p_1/2 and Co 2p_3/2, respectively. The binding energy difference between these two peaks was found to be 14.43 eV, which implies that the oxidation state of cobalt is +2. The peaks were further deconvoluted into six peaks. The peaks at binding energy 781.10 and 795.99 eV are assigned to Co–P species. Two additional peaks appeared at 784.24 and 797.96 eV ascribed to Co–N. Other peaks corresponded to satellite peaks with binding energies of 788.90 and 803.16 eV.^26,27

Fig. 2 The XPS Co 2p spectrum of Co–N₂P₂ (a), Co K-edge XANES data for CoPc, CoO, Co foil and Co–N₂P₂ (b), EXAFS spectra of CoPc, CoO, Co foil and Co–N₂P₂ (c), and EXAFS fitting for Co–N₂P₂ (d).

X-ray absorption fine spectroscopy (XAFS) analysis was employed to determine the coordination site of isolated Co atoms. The X-ray near edge spectrum of Co–N₂P₂ was compared with Co foil, CoO and cobalt phthalocyanine (Fig. 2b). The absorption edge of Co in Co–N₂P₂ in higher than the reference samples, suggesting that positive nature of Co.^28,29 The Fourier transform (FT) k³-weighted EXAFS spectra of Co k-edge (Fig. 2c) displayed the broad peak around 1.6 Å indicating the Co–N and Co–P bond.³⁰ Meanwhile, the absence of a peak at 2.2 Å (Co–Co bond) evidenced that atomic dispersion of Co present in Co–N₂P₂. Furthermore, EXAFS fitting was used to determine the local coordination environment of the Co atoms. The EXAFS fittings (Fig. 2d and Table S2) prove that Co atoms are well isolated and bonded to donor atoms (nitrogen and phosphorus) with the Co–N₂P₂ configuration on the porous carbon matrix. This mixed coordination environment and electronic metal–support interaction are ineluctable for augmenting the catalytic performance of single-atom catalysts.

The catalytic performance of Co–N₂P₂ was evaluated in the α-alkylation of aromatic ketones with alcohols via a borrowing-hydrogen pathway, a sustainable route for C–C bond formation. First, acetophenone and benzyl alcohol were chosen as model substrates to establish the optimal reaction conditions (Table S3).

Table 1 Substrate scope of Co–N₂P₂ catalyzed α-alkylation reaction for different benzyl alcohol derivatives

Reaction conditions: acetophenone (0.5 mmol), alcohol (0.6 mmol), KO^tBu (0.5 mmol), Co–N₂P₂ (10 mg, 4.047 × 10⁻⁷ mol), toluene (1 mL), 100 °C and 12 h.

---

Control experiments revealed that no product formation occurred in the absence of either the catalyst or base, confirming that both components are essential for the α-alkylation reactions. The negligible activity over the PC support matrix suggested that Co species acted as the catalytic centre. In addition, a control experiment using a homogeneous cobalt precursor, Co(PPh₃)₂Cl₂ (4.047 × 10⁻⁷ mol), produced a trace amount of product (Table S3, entry 4). Then, the effect of solvent was screened (Table S3 entries 6–9). Among the various solvents, toluene affords a good yield. Subsequently, several strong and weak bases were examined and demonstrated that KO^tBu afforded significantly higher conversions than KOH, K₂CO₃, Na₂CO₃ or NaOH (Table S3, entries 10–13). These results suggest that a strong non-nucleophilic base is required to promote alcohol dehydrogenation and subsequent condensation steps.

To evaluate the general applicability of the catalyst, a broad range of alcohols and ketones were utilized for α-alkylation under the optimized conditions (Table 1). Using alcohols bearing electron-donating groups such as –OMe and –Me as well as electron-withdrawing substituents including –Cl, –F, and –I were well tolerated, giving the desired products in excellent yields. Heterocyclic and polycyclic alcohols such as pyridine, naphthalene and anthracene methanol also reacted smoothly. Interestingly, cinnamyl alcohol underwent selective hydrogenation of the α,β unsaturated C=C bond, demonstrating the capacity of the Co single-atom site for chemoselective hydrogenation.

Furthermore, using a different substituted aromatic ketone was assessed (Table S4). Electron-rich and electron-poor acetophenones underwent α-alkylation in consistently high yields. Sterically demanding substrates such as o-methyl and o-chloro acetophenone also afforded the corresponding products, indicating that the porous structure and coordination environment of Co–N₂P₂ facilitate efficient substrate accessibility. These results collectively demonstrate that the synergistic N and P coordination environment in Co–N₂P₂ enables high activity and excellent functional group tolerance. In addition, the catalytic performance of the Co–N₂P₂ catalyst was compared with those of previously reported catalysts for the α-alkylation reaction (Table S5) involving high metal loading, the use of excess bases, and harsh conditions. But, our catalyst, Co–N₂P₂, showed excellent catalytic performance under mild conditions, with easy separability and reusability for the α-alkylation reaction.

Co–N₂P₂ also demonstrated excellent stability on a preparative scale (Table S6). The Gram-scale synthesis of 1,3-diphenylpropan-1-one proceeded smoothly with comparable yields, highlighting the practical applicability of the catalyst. Green-chemistry metrics further highlight its sustainability, giving an E-factor of 2.26, an atom economy of 92.1%, carbon efficiency 100%, atom efficiency 77.4% and high reaction mass efficiency 83.9%, clearly indicating the technical and sustainable advantages of this methodology.

To gain mechanistic insights, a series of control and intermediate-trapping experiments were conducted (Scheme S2). Benzyl alcohol oxidation under standard reaction conditions yielded benzaldehyde, establishing that Co–N₂P₂ is competent for the initial dehydrogenation step. The relatively low yield of benzaldehyde suggests that this transformation may be rate-determining. Furthermore, the aldol condensation between acetophenone and benzaldehyde produced chalcone in 62% yield in the presence of KO^tBu, indicating that base-mediated enolate formation and C–C coupling are integral components of the catalytic cycle. In the absence of alcohol, the α,β-unsaturated chalcone intermediate remained unreacted, confirming that hydrogen borrowed from alcohol oxidation is required for subsequent reduction. Indeed, chalcone was effectively hydrogenated in the presence of benzyl alcohol and Co–N₂P₂, consistent with a transfer hydrogenation. These combined observations support a mechanism (Scheme S3) wherein alcohol first coordinates to the Co active site and undergoes dehydrogenation via β-hydride elimination following base-assisted alkoxide formation. The resulting aldehyde condenses with the ketone through a base-promoted aldol pathway to generate an α,β-unsaturated ketone intermediate. Finally, hydrogen atoms adsorbed on Co sites are transferred to the enone, affording the α-alkylated product while regenerating the active catalyst.³¹

The stability and heterogeneous nature of the catalyst were examined by control experiments. Sheldon's hot-filtration test (Fig. S5) showed no further product formation after catalyst removal, confirming that there is no leaching of cobalt species in the reaction conditions. A KSCN poisoning test (Fig. S6) resulted in a significant reduction in activity, consistent with coordination of SCN⁻ to isolated Co sites and consequent inhibition of catalysis. The acid washed catalyst did not significantly affect activity, suggesting the robustness of Co–N₂P₂ sites and ruling out the involvement of Co nanoparticles. Together, these studies support that atomically dispersed Co sites are the true catalytic sites responsible for C–C bond formation. The catalyst was reused in up to five consecutive cycles without significant loss of activity (Fig. S7). PXRD and XPS analyses (Fig. S8 and S9) proved that the structural integrity of the catalyst was preserved in the recycled catalyst, and hence the activity remained unchanged. Overall, the combined catalytic, mechanistic, and sustainability assessments establish Co–N₂P₂ as a potent single-atom catalyst for α-alkylation reactions. The well-defined N,P-coordinated environment enhances electronic metal support interaction, stabilizes the Co sites during redox transitions, and enables efficient borrowing hydrogenation catalysis with excellent functional group tolerance and recyclability.

This work demonstrates a rational molecular-to-atomic strategy for constructing a cobalt single-atom catalyst with a precisely defined N and P coordination environment (Co–N₂P₂) on a sustainable, rice husk-derived carbon support. The incorporation of phosphorus through a molecular Co(PPh₃)₂Cl₂ precursor enables simultaneous control over atomic dispersion and electronic modulation of the metal centre. The synergistic N and P coordination enhances activity in α-alkylation reactions, delivering high functional-group tolerance, recyclability, and favourable green metrics. This study provides new insights into heteroatom-engineered SAC design for sustainable C–C bond formation using earth-abundant metals.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: experimental details, additional characterization, optimization table comparison studies and mechanism. See DOI: https://doi.org/10.1039/d6cc00398b.

Acknowledgements

We acknowledge the DST, New Delhi (SR/FIST/CS-1/2018/62; SRG/2023/001756; SR/PURSE/2023/165) for financial support and NMR facilities.

References

1. X. Cui, W. Li, P. Ryabchuk, K. Junge and M. Beller, Nat. Catal., 2018, 1, 385–397.
2. V. B. Saptal, V. Ruta, M. A. Bajada and G. Vilé, Angew. Chem., Int. Ed., 2023, 62, e202219306.
3. F. Zhang, Y. Zhu, Q. Lin, L. Zhang, X. Zhang and H. Wang, Energy Environ. Sci., 2021, 14, 2954–3009.
4. S. Mitchell, E. Vorobyeva and J. Pérez-Ramírez, Angew. Chem., Int. Ed., 2018, 57, 15316–15329.
5. K. Pang, Y. Tang, C. Qiu, M. Zhang, A. Tayal, S. Feng, C. Long, Y. Wang, J. Chang, B. Pang, A. Sikdar, S. S. Garakani, Y. Zhang, H. Wang, W. Zhang, G. Luo, Y. Wang and J. Yuan, Matter, 2024, 7, 655–667.
6. P. Wang, D. Zhao and L. Yin, Energy Environ. Sci., 2021, 14, 1794–1834.
7. H. Yan, C. Su, J. He and W. Chen, J. Mater. Chem. A, 2018, 6, 8793–8814.
8. Z. Li, S. Ji, Y. Liu, X. Cao, S. Tian, Y. Chen, Z. Niu and Y. Li, Chem. Rev., 2020, 120, 623–682.
9. Z. Zhang, C. Feng, C. Liu, M. Zuo, L. Qin, X. Yan, Y. Xing, H. Li, R. Si, S. Zhou and J. Zeng, Nat. Commun., 2020, 11, 1215.
10. S. Ji, Y. Chen, X. Wang, Z. Zhang, D. Wang and Y. Li, Chem. Rev., 2020, 120, 11900–11955.
11. J. Liu, ACS Catal., 2017, 7, 34–59.
12. X. Li, H. Rong, J. Zhang, D. Wang and Y. Li, Nano Res., 2020, 13, 1842–1855.
13. M. B. Gawande, P. Fornasiero and R. Zbořil, ACS Catal., 2020, 10, 2231–2259.
14. L. Zhao, X. Han, W. Kong, Y. Tong, Y. Ding, J. Wang, B. Li, Y. Liu, J. Xu and W. Xing, Catal. Sci. Technol., 2022, 12, 530–541.
15. H.-C. Zhou, J. R. Long and O. M. Yaghi, Chem. Rev., 2012, 112, 673–674.
16. Z. Xiang, D. Cao, L. Huang, J. Shui, M. Wang and L. Dai, Adv. Mater., 2014, 26, 3315–3320.
17. Y. Zheng, H. Song, S. Chen, X. Yu, J. Zhu, J. Xu, K. A. I. Zhang, C. Zhang and T. Liu, Small, 2020, 16, 2004342.
18. S. Guo, P. Yuan, J. Zhang, P. Jin, H. Sun, K. Lei, X. Pang, Q. Xu and F. Cheng, Chem. Commun., 2017, 53, 9862–9865.
19. N. Anbuselvan, S. Sivaprakash, K. Nithya, A. Anakkayath, M. K. Ravva, R. K. Rana, D. Suresh and A. J. Amali, ACS Appl. Nano Mater., 2025, 8, 22251–22261.
20. Y. Liu, Y. Chen, Y. Liu, Z. Chen, H. Yang, Z. Yue, Q. Fang, Y. Zhi and S. Shan, J. Catal., 2022, 407, 65–76.
21. K. Nithya, R. Anbarasan, N. Anbuselvan, V. S. Vasantha, D. Suresh and A. J. Amali, ACS Appl. Nano Mater., 2024, 7, 9554–9564.
22. V. Vyas, P. Maurya and A. Indra, Chem. Sci., 2023, 14, 12339–12344.
23. K. Nithya, R. Anbarasan, N. Anbuselvan, K. Lakshmanan, V. S. Vasantha, D. Suresh and A. J. Amali, ACS Appl. Nano Mater., 2024, 7, 25459–25469.
24. T. Su and C. Cai, ACS Appl. Mater. Interfaces, 2022, 14, 55568–55576.
25. H. Chu, Q. Wu and J. Huang, Colloids Surf., A, 2018, 558, 495–503.
26. B. Li, J. Kou, G. Zeng, J. Ma and Z. Dong, ACS Catal., 2023, 13, 16286–16299.
27. H. Luo, L. Wang, S. Shang, G. Li, Y. Lv, S. Gao and W. Dai, Angew. Chem., Int. Ed., 2020, 59, 19268–19274.
28. Q. Zhao, W. Ni, X. Tan, F. Cao, T. Liu, H. Huang, Z. Cheng, Y. Li, S. He, H. Ning and M. Wu, J. Mater. Chem. A, 2022, 10, 9435–9444.
29. C. Long, K. Wan, X. Qiu, X. Zhang, J. Han, P. An, Z. Yang, X. Li, J. Guo, X. Shi, H. Wang, Z. Tang and S. Liu, Nano Res., 2022, 15, 1817–1823.
30. F. Cao, W. Ni, Q. Zhao, L. Wang, S. Xue, Y. Li, D. Kong, M. Wu and L. Zhi, Appl. Catal., B, 2024, 346, 123762.
31. X. Zhang, G.-P. Lu, K. Wang, Y. Lin, P. Wang and W. Yi, Nano Res., 2022, 15, 1874–1881.

---