N-doped carbon nanofiber-supported amorphous Nb₂O₅ with synergistic Brønsted–Lewis sites for converting sorbitol to isosorbide

Abstract

Isosorbide, a versatile platform chemical in pharmaceuticals, biodegradable polymers, and food additives, is usually produced through sulfuric acid catalysis, which suffers from corrosive waste generation and tedious product purification. In this work, we have prepared a class of nitrogen-doped carbon nanofiber network-supported amorphous niobium oxide (Nb₂O₅/N-CNF) through coupling electrospinning and thermal annealing. The optimized catalyst achieves 100% sorbitol dehydration conversion with an isosorbide yield of 84% within 1 h. The mass-normalized activity surpasses reported values by 8–10 fold, establishing a new benchmark for niobium-based catalysts. The experimental characterization studies and theoretical simulations demonstrate that the Lewis/Brønsted acid ratio has been modulated by creating electron-deficient Nb sites through Nb–N coordination. The electronic engineering facilitates charge transfer between Nb₂O₅ and sorbitol, reinforces the adsorption of polyol radical intermediates, and decreases the energy barrier for hydroxyl group elimination. The active centers of Nb–N preferentially interact with hydroxyl lone-pair electrons, effectively polarizing C–O bonds to initiate stepwise cascade dehydration. This work provides fundamental insights into the design of nanocatalysts for biomass conversion, establishing a new paradigm for functional catalytic architectures towards various applications.

---

Introduction

With growing attention on the sustainable biomass utilization, the catalytic conversion of biomass-derived sorbitol to isosorbide has emerged as a pivotal green chemical process.^1–3 As a bio-based platform compound, isosorbide has substantial commercial potential as a renewable alternative to petroleum-based chemicals.^4,5 It is essential in medical applications, covering effective treatments for glaucoma and cardiovascular diseases^6–8 and also acts as a superior polymer monomer for enhancing thermal stability, mechanical strength, and optical properties in the fields of spanning cosmetics, plasticizers, and combustion promoters.^9–14 These multifaceted roles of isosorbide stimulate the necessity for developing efficient and environmentally benign methodologies.

Conventionally, the conversion of sorbitol to isosorbide predominantly employs acid catalysts, where homogeneous systems like concentrated sulfuric acid demonstrate high activity,^15,16 leading to environmental hazards, equipment corrosion, and complex product separation.¹⁷ In contrast, heterogeneous solid-acid catalysts, such as metal oxides, inorganic salts, sulfonic acid resins, phosphotungstic acid, molecular sieves, and heteropoly acids, offer distinct advantages.¹⁸ Metal oxides, such as zirconia (ZrO₂), titania (TiO₂), and niobium oxide (Nb₂O₅) with good Lewis acidity,^21,22 strong acidity and excellent thermal stability,^19,20 have been applied for catalytic dehydration of sorbitol. For example, Ford et al. synthesized sulfonated zirconia using a solgel method, achieving a 76% yield of isosorbide at 150 °C within 2 h.²³ Similarly, Zhang investigated the impact of different zirconia crystal structures and found that the monoclinic phase demonstrated superior activity and high isosorbide yield via sorbitol dehydration.²⁴ Mesoporous TiO₂ achieved complete conversion of sorbitol, with an isosorbide selectivity of 70%.²⁵ The acidified niobium oxide realized an isosorbide yield of 84%.²⁶ These studies highlight the potential of metal oxides to improve the efficiency and selectivity of sorbitol-to-isosorbide conversion. However, the development of cost-effective and efficient green catalytic systems still remains a challenge.

Herein, a class of nitrogen-doped carbon nanofiber-supported Nb₂O₅ (Nb₂O₅/N-CNF) has been prepared through coupling processes of electrospinning and thermal annealing. The Lewis acid distribution and electronic properties of Nb₂O₅ can be regulated by tuning N types in the substrate. The sorbitol has been totally converted on the optimized catalyst, yielding 84% isosorbide, corresponding to a mass-specific activity of 168 g_isosorbide g_Nb2O5⁻¹, and surpassing state-of-the-art benchmarks. The incorporation of N atoms is responsible for the enhanced accessibility of acid sites and charge transfer dynamics, collectively promoting dehydration kinetics. This work provides a pathway for building a catalytic system towards green biomass conversion.

Experimental

Catalyst preparation

Nb₂O₅/N-CNF. The Nb(C₂O₄H)₅/TEOS/PAN composite nanofibers were first electrospun on Al foil from a mixed solution of Nb(C₂O₄H)₅ (0.8 g), TEOS (1.5 g), PAN (0.6 g), and DMF (8 mL). A voltage of 15 kV was set with a distance of 20 cm and the flow rate of the spinning solution was set at 1 mL h⁻¹. A white film of Nb(C₂O₄H)₅/TEOS/PAN was obtained. The as-prepared precursor was kept for 2 h in air at 250 °C, which was further annealed at 650 °C for 2 h in a furnace under Ar. After etching SiO₂ in NaOH, the sample was dried at 40 °C overnight. Subsequently, acidification was performed by immersing the dried sample in 2 M H₂SO₄ solution with a solid–liquid ratio of 1 : 16 (catalyst : H₂SO₄ solution, w/v) and stirring at 60 °C for 2 h. The sample was then filtered, washed with deionized water until the filtrate was neutral (pH = 7), and dried at 80 °C for 6 h before use. The obtained Nb₂O₅/N-CNF was calcined under NH₃ atmosphere at different temperatures to engineer the nitrogen content.

C-Nb₂O₅/N-CNF. The Nb(C₂O₄H)₅/TEOS/PAN composite nanofibers were first electrospun on Al foil from a mixed solution of Nb(C₂O₄H)₅ (0.8 g), TEOS (1.5 g), PAN (0.6 g), and DMF (8 mL). A voltage of 15 kV was set with a distance of 20 cm and the flow rate of the spinning solution was set at 1 mL h⁻¹. A white film of Nb(C₂O₄H)₅/TEOS/PAN was obtained. The as-prepared precursor was kept for 2 h in air at 250 °C, which was further annealed at 850 °C for 2 h in a furnace under NH₃.

Catalytic performance evaluation

0.5 g sorbitol and 1 g of the catalyst were introduced into the reactor. The temperature was increased to 140 °C, and the vacuum pump was started to decrease the vacuum degree to less than 6 kPa.

Product analyses

The catalyst (Nb₂O₅/N-CNF) and sorbitol were added to the reaction flask at 140 °C in an oil bath under magnetic stirring. After reacting for a certain time, the reaction mixture was quenched with water. The filter and catalyst were dried under reduced pressure. The reaction mixture was concentrated under reduced pressure for 48 h. Quantification of reaction species was realized using ¹H NMR in 10% D₂O with dimethyl sulfoxide (DMSO) as an internal standard as described above.

Catalyst characterization

The morphologies of the samples were observed by scanning electron microscopy (SEM) (Hitachi S-4800) and transmission electron microscopy (TEM) (TECNAI G2 Spirit TWIN). High resolution TEM (HRTEM) was performed with a JEM-2100F equipped with an energy-dispersive X-ray spectrometer (EDS). Aberration corrected high angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images of atomic dispersion of Nb were examined using a FEI-Titan Cubed Themis G2 300. X-ray diffraction (XRD) was recorded with a Rigaku Corporation UItimaIV diffractometer using a Cu Kα source (λ = 0.15406 nm). The adsorption isotherms of nitrogen were measured by using automatic volumetric adsorption equipment (BELSORP mini II, BEL), and the surface areas were calculated using N₂ sorption data at 77 K. X-ray photoelectron spectroscopy (XPS) was performed on a scanning X-ray microprobe (PHI 5000 Verasa, ULAC-PHI, Inc.). Thermogravimetric analysis (TGA) was performed on a Netzsch (TG 209 F3, Germany) with a heating rate of 10 °C min⁻¹ under a N₂ atmosphere. The content of Nb was determined by ICP-MS (ICAP RQ, Germany). Infrared spectra were obtained with a Frontier Mid-IR FTIR Fourier spectrometer.

Results and discussion

Fig. 1a shows the schematic process for the preparation of an Nb₂O₅/N-CNF film. Firstly, polyacrylonitrile (PAN), tetraethyl orthosilicate (TEOS), and niobium oxalate (Nb(C₂O₄H)₅) were dissolved in N,N-dimethylformamide (DMF), producing a light-yellow mixture.²⁷ The obtained nanofiber film of Nb(C₂O₄H)₅/TEOS/PAN was fabricated via electrospinning. After the film was calcined under Ar (or NH₃), the black Nb₂O₅/SiO₂/N-CNF film was formed. SiO₂ was selectively etched by NaOH, creating a porous Nb₂O₅/N-CNF. Finally, the produced porous Nb₂O₅/N-CNF film was immersed in a dilute sulfuric acid solution to introduce Brønsted acid sites. The SEM image shown in Fig. 1b reveals that the catalyst exhibits interwoven nanofibers with an average diameter of 400 nm and several hundred micrometers length, respectively. High-resolution TEM (HRTEM, Fig. 1c, S1) and SEM images (Fig. S2) further demonstrate that the nanofibers possess a porous and rough surface. Barrett–Joyner–Halenda (BJH) pore-size distribution (Fig. S3) confirms the presence of abundant mesopores in the nanofibers with a pore diameter of 2–20 nm. Selective area electron diffraction (SAED) and X-ray diffraction (XRD) patterns (Fig. S4 and S5) indicate that Nb₂O₅ and the carbon substrate are amorphous in nature. The energy dispersive X-ray spectroscopy (EDS) mapping (Fig. 1d–h) gives a uniform distribution of elemental C, N, O, and Nb. The results indicate that amorphous Nb₂O₅ was dispersed within porous N-doped carbon (NC) nanofiber framework.

Fig. 1 (a) Schematic diagram of Nb₂O₅/N-CNF film preparation and catalytic conversion of sorbitol to isosorbide. (b) SEM image, (c) TEM image, (d) HAADF-STEM image, and (e–h) elemental EDS-mapping images of the sample.

The preparation conditions were systematically investigated to produce an efficient catalyst for converting sorbitol to isosorbide. The conversion rate and isosorbide selectivity were quantitatively analyzed with ¹H NMR spectra (Fig. S6). Here, the influence of different polymer sources as electrospinning additives was examined for sorbitol dehydration. As shown in Fig. 2a and Table S1, three types of carbon-supported Nb₂O₅ were separately synthesized using polystyrene (PS), polyvinyl alcohol (PVA), and polyacrylonitrile (PAN) as spinning sources. The catalysts were assigned as Nb₂O₅/CNF, Nb₂O₅/O-CNF, and Nb₂O₅/N-CNF, respectively. The induced coupled plasma (ICP) analysis revealed that Nb₂O₅ accounts for ∼2%. Fig. 2a and Table S1 show that bare CNF exhibits negligible activity. Single Nb₂O₅ exhibits a low isosorbide yield (3%), which is much lower than the Nb₂O₅-loading catalysts, e.g., 51% (Nb₂O₅/O-CNF), 57% (Nb₂O₅/CNF), and 63% (Nb₂O₅/N-CNF). Apparently, Nb₂O₅/N-CNF has the highest conversion efficiency. Thus the conditions for the optimized catalyst were utilizing PAN as the typical carbon source (Table S1), including calcination temperature, atmosphere, and catalytic reaction temperature (Table S2). Following the calcination at 650 °C under a NH₃ atmosphere (Table S2, entry 10), the obtained Nb₂O₅/N-CNF catalyst achieved complete sorbitol conversion within 40 min, with an isosorbide yield of 84%, corresponding to a mass-specific activity of 168 g_isosorbide g_Nb2O5⁻¹. This performance surpasses most reported values (Fig. 2c, Table S3).^{9,19,26,28–31} For direct comparison with the literature, the mass-normalized activity of our optimized catalyst (∼2 wt% Nb₂O₅ loading) is 5 g_isosorbide g_catalyst⁻¹ (Table S3, entry 11), a value that remains competitive with reported systems. As a reference to the amorphous phase, crystalline Nb₂O₅/CNF (C-Nb₂O₅/N-CNF) (Fig. S8) shows significantly lower performance in isosorbide dehydration (yield: 23.14% vs. 84.3% for amorphous, Table S2, entry 13). This gap stems from structural differences: the amorphous phase features a disordered structure with abundant defects and polarized Nb–O bonds, enabling rich Brønsted (surface –OH/SO₄²⁻) and Lewis (exposed Nb⁵⁺) acidity. In contrast, the ordered crystal lattice of crystalline Nb₂O₅ restricts acid site exposure, reducing both total acidity and catalytic activity. Additionally, catalyst recyclability was assessed over five cycles under optimized conditions. After each run, the catalyst was filtered, ethanol-washed, and dried at 80 °C before reuse. The isosorbide yield remained above 80% throughout, declining only modestly from 84% to 80%, confirming excellent stability (Fig. S7a). XRD and XPS analysis of the spent catalyst (after cycle 5) revealed retained amorphous Nb₂O₅ (Fig. S7b) and nearly constant N-content (9.06 at% fresh vs. 9.2 at% spent), indicating minimal leaching and structural integrity.

Fig. 2 (a) Comparative experiment on the catalytic performance of different catalysts (CNF, Nb₂O₅, Nb₂O₅/O-CNF, Nb₂O₅/CNF, and Nb₂O₅/N-CNF) for the dehydration of sorbitol to prepare isosorbide. (b) The catalytic activity of Nb₂O₅/N-CNF under different atmospheres and temperatures. (Reaction conditions: 140 °C, 2 h, 50 mg of the catalyst, 250 mg sorbitol.) (c) Comparison of catalytic performance between Nb₂O₅/N-CNF and reported catalysts; the number in the upper right corner of the catalyst denotes references. (d) Pyridine-adsorbed FTIR spectra of Nb₂O₅/N-CNF with different treatment methods, UA indicates no acidification.

In the conversion of sorbitol to isosorbide, both Lewis acid and Brønsted acid sites in the catalyst play critical roles in facilitating the dehydration reaction.^23,26 The Lewis acid acts as an electron-pair acceptor, interacting with the lone-pair electrons of the hydroxyl oxygen atoms of sorbitol. The interaction reduces the electron density around the hydroxyl groups, promoting departure and initiating the dehydration process. On the other hand, the Brønsted acid provides protons during the reaction, which directly bind to the hydroxyl groups of sorbitol, forming hydronium ions, further accelerating the removal of hydroxyl groups and driving the dehydration reaction. The pyridine-IR was done to analyse the types and distribution of acidic sites over Nb₂O₅/N-CNF. Fig. 2d and Table S5 show that the annealed catalyst in NH₃ exhibits a significant increase in Lewis acid sites vs. the sample treated under Ar. Furthermore, as the annealing temperature increased, the proportion of Lewis acid sites initially increased and then declined. The sample treated at 650 °C has the highest proportion of Lewis acid sites, aligning with the variation of the N content (Table S4). The results suggest that the N introduced during annealing enhances the Lewis acid content in the catalyst. The increasing Lewis acid sites are beneficial for promoting the dehydration of sorbitol to isosorbide. To confirm the role of Brønsted acid sites, the performance of Nb₂O₅/N-CNF was also compared before and after acidification (Table S2, entry 12 vs. entry 10). The non-acidified catalyst (UA-Nb₂O₅/N-CNF, dominated by Lewis acid sites) primarily produced 1,4-anhydro sorbitol (yield: ∼12.2%) with negligible isosorbide (∼3.5%), indicating that Lewis acid sites alone can only drive the first dehydration step (sorbitol → 1,4-anhydro sorbitol) but fails to complete the second dehydration (1,4-anhydro sorbitol → isosorbide). In contrast, acidified Nb₂O₅/N-CNF (with both Brønsted and Lewis acid sites) achieved 84% isosorbide yield, confirming that Brønsted acid sites are essential for the second dehydration. To further investigate the effect of H₂SO₄ acidification time on Brønsted acidity, tests were conducted under fixed conditions (2 M H₂SO₄, solid–liquid ratio 1 : 16, 60 °C) for durations ranging from 0.5 to 4 h. As summarized in Table S7, the isosorbide yield reached a maximum (84.31%) after 2 h of acidification, indicating optimal Brønsted acid site formation. Shorter treatment (0.5 h, 49.21% yield) resulted in insufficient acid sites, while prolonged exposure (3–4 h, ∼77% yield) led to a slight decline, likely due to active site blockage by excess sulfate species. These results clarify the role of acidification time in tuning Brønsted acidity and offer practical guidance for process optimization. Therefore, a synergistic effect between Brønsted and Lewis acid sites is responsible for the enhanced performance. Additionally, Nb XPS and FTIR analyses (Fig. S9) confirm the intact structure of Nb₂O₅/N-CNF after acid treatment. The appearance of a sulfate ion peak (Fig. S9b, 1030 cm⁻¹) in the FTIR spectrum indicates the successful introduction of Brønsted acid sites through adsorption, with no evidence of structural degradation.

X-ray photoelectron spectroscopy (XPS) was used to examine the chemical composition and bonding characteristics of the catalyst. Fig. 3a shows that the Nb 3d in pristine Nb₂O₅/N-CNF appeared at 210.96 eV and 208.16 eV, corresponding to Nb 3d_3/2 and Nb 3d_5/2 for Nb⁵⁺.^32,33 As the N content increased, the binding energy shifted to a high energy direction, indicating a decrease in the electron-cloud density around Nb. The result implies that the incorporation of N alters surface electron distribution of the catalysts. Further analysis of N in the sample (Fig. 3b) gives four distinct nitrogen species: N–Nb (398–399 eV),^34,35 pyridinic N (399–401 eV), pyrrolic N (401–402 eV), and graphitic N (above 402 eV).^36–39 Their proportions varied with the change of the N content (Table S6). The sample treated at 550 °C under NH₃ leads to the domination of pyrrolic N (59.51%). As the temperature increased, it was converted into other forms. The pyridinic N reached the peak proportion at 650 °C (44.98%). When the NH₃ annealing temperature reaches 750 °C, pyrrolic N and pyridinic N partially converts to graphitic N (15.84%), which has weaker ability to form Nb–N bonds (Table S6). This reduces the density of Lewis acid sites and thus their activity (Fig. 2b). The superior performance at 650 °C is rationalized by the distinct acid-modulating effects of N types: pyridinic N donates a lone pair of electrons to create stronger Lewis acid sites, in contrast to the weaker acidity induced by pyrrolic N's delocalized electrons. The conversion of N species aligns with the instability trend of pyrrolic N at high temperatures.^40–42 Among these species, the pyridinic N, typically situated at the edges or voids of the graphite carbon layer, possesses a lone pair of electrons as the Brønsted acid.^43,44 Additionally, as the N content increased, the rising proportion of N–Nb bonds boosts the density of Lewis acid sites.

Fig. 3 XPS spectra of (a) Nb 3d and (b) N 1s for Nb₂O₅/N-CNF annealed under an Ar/NH₃ atmosphere at 550 °C, 650 °C and 750 °C, respectively.

Theoretical calculations were also conducted to evaluate the average local ionization energy (ALIE) at various positions of the sorbitol molecules.^45,46 The ALIE quantifies the ease of electron ionization at a specific location in a three-dimensional space. The low values indicate weak electron-binding capability, making electrophilic reactions more favourable. The terminal hydroxyl groups of sorbitol exhibit lower ALIE compared to the intermediate ones (Fig. 4a). This suggests that the terminal hydroxyl groups are more susceptible to electrophilic attack and more likely to interact with the Lewis acid sites of the catalyst. The density functional theory (DFT) calculation quantifies the impact of N in the substrate on the surface charge of Nb₂O₅. Fig. 4b shows that the presence of N increases the positive charge of Nb atoms in comparison with the bare sample, thus enhancing the catalyst's Lewis acidity.⁴⁷ The modification enables the catalyst to interact with the terminal hydroxyl groups of sorbitol, promoting the dehydration to isosorbide. As revealed in Fig. 4c (step 1), the process begins with the Lewis acid site for the terminal hydroxyl group of sorbitol, and to be protonated. Then the hydroxyl group on the active sites undergoes an S_N2 reaction, leading to the elimination of water molecules and the formation of 1,4-dehydrated sorbitan intermediate. The Lewis acid and protons activate the terminal hydroxyl groups in the intermediate in a similar way, facilitating the second dehydration.

Fig. 4 (a) Average local ionization energy of sorbitol, (b) charge change and structural model of niobium oxide on catalyst surface and (c) the proposed reaction mechanism for the dehydration of sorbitol to isosorbide.

Conclusions

In summary, a class of Nb₂O₅/N-CNF catalysts has been developed by coupling electrospinning and annealing under various atmospheres, exhibiting exceptional conversion of sorbitol-to-isosorbide. The optimized catalyst achieved complete sorbitol conversion with 84% isosorbide yield. Experimental and theoretical investigations revealed that the N in the carbon substrate effectively modulates the balance of Lewis and Brønsted acid sites, significantly enhancing catalytic performance. This work provides a new strategy for designing nanocatalysts for carbohydrate dehydration.

Author contributions

Jiaqi He: data curation, formal analysis, methodology, resources, software, writing – original draft, and writing – review and editing; Jun Fu: data curation, resources, and writing – review and editing; Yongli Shen: data curation and software; Chengcai Pang: methodology; Wen Zhang: conceptualization, data curation, software, and writing – review and editing; Yixin Song, Lijuan Zhang, Xintai Su, and Bekchanov Davronbek: analysis and manuscript preparation; and Changhua An: conceptualization, funding acquisition, project administration, supervision, and writing – review and editing. Jiaqi He and Jun Fu contributed equally to this work.

Conflicts of interest

The authors declare no conflict of interest.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: additional characterization data for materials, XRD, SEM, HADDF, and additional electrochemical characterization data. See DOI: https://doi.org/10.1039/d5nr03255e.

Acknowledgements

This work was supported by the Natural Science Foundation of China (22275139) and the Key Project of Natural Science Foundation of Tianjin City (Contract No. 22JCZDJC00510).

References

1. C. Dussenne, T. Delaunay, V. Wiatz, H. Wyart, I. Suisse and M. Sauthier, Green Chem., 2017, 19, 5332–5344.
2. Y. Wang, K. Van Der Maas, D. H. Weinland, D. Trijnes, R.-J. Van Putten, A. Tietema, J. R. Parsons, E. De Rijke and G.-J. M. Gruter, Environ. Sci. Technol., 2024, 58, 2293–2302.
3. T. Yan, M. Zhang, R. Yuan and W. Dai, Sci. China: Chem., 2024, 67, 3588–3613.
4. F. Aricò, Curr. Opin. Green Sustainable Chem., 2020, 21, 82–88.
5. A. Ismagilova, L. Matt, P. Jannasch, V. Kisand and L. Vares, Green Chem., 2023, 25, 1626–1634.
6. W. Ni, D. Li, X. Zhao, W. Ma, K. Kong, Q. Gu, M. Chen and Z. Hou, Catal. Today, 2019, 319, 66–75.
7. G. Li, M. Ling, K. Yu, W. Yang, Q. Liu, L. He, X. Cai, M. Zhong, Z. Mai, R. Sun, Y. Xiao, Z. Yu and X. Wang, J. Nanobiotechnol., 2022, 20, 471.
8. D. Döpping, J. Kern, N. Rotter, A. Llevot and H. Mutlu, ACS Sustainable Chem. Eng., 2024, 12, 13401–13408.
9. D. Yuan, F. Xiao, N. Zhao, C. Deng, X. Huang, H. Zhang, Q. Yang and Y. Qiao, Chem. Eng. J., 2023, 460, 141780.
10. H.-S. Yang, S. Cho, M. Lee, Y. Eom, H. G. Chae, S.-A. Park, M. Jang, D. X. Oh, S. Y. Hwang and J. Park, Mater. Des., 2021, 198, 109284.
11. Z. Yang, X. Li, F. Xu, W. Wang, Y. Shi, Z. Zhang, W. Fang, L. Liu and S. Zhang, Green Chem., 2021, 23, 447–456.
12. Y. G. Daniel and B. A. Howell, Polym. Degrad. Stab., 2018, 156, 14–21.
13. Y. Yang, Z. Xiong, L. Zhang, Z. Tang, R. Zhang and J. Zhu, Mater. Des., 2016, 91, 262–268.
14. M. Ben Abdelkader, N. Azizi, A. Baffoun, Y. Chevalier and M. Majdoub, Ind. Crops Prod., 2018, 123, 591–599.
15. F. Brandi and M. Al-Naji, ChemSusChem, 2022, 15, e202102525.
16. J. Saska, S. Dutta, A. Kindler, S. J. Zuend and M. Mascal, ACS Sustainable Chem. Eng., 2021, 9, 11565–11570.
17. D. Yuan, N. Zhao, Y. Wang, K. Xuan, F. Li, Y. Pu, F. Wang, L. Li and F. Xiao, Appl. Catal., B, 2019, 240, 182–192.
18. O. A. Rusu, W. F. Hoelderich, H. Wyart and M. Ibert, Appl. Catal., B, 2015, 176–177, 139–149.
19. N. Lertna, P. Srisamut, N. Supapan, P. Athikaphan, S. C. Rood, A. Neramittagapong and S. Neramittagapong, Mater. Today: Proc., 2023, 75, 91–98.
20. Y. Zhang, T. Chen, G. Zhang, G. Wang and H. Zhang, Appl. Catal., A, 2018, 562, 258–266.
21. R. Otomo, T. Yokoi and T. Tatsumi, Appl. Catal., A, 2015, 505, 28–35.
22. G. Morales, J. Iglesias, J. A. Melero, J. Moreno, R. Sánchez-Vázquez, Á. Peral and A. Cubo, Top. Catal., 2017, 60, 1027–1039.
23. J. T. Hopper, R. Ma, J. B. Rawlings, P. C. Ford and M. M. Abu-Omar, ACS Catal., 2023, 13, 10137–10152.
24. X. Zhang, A. I. M. Rabee, M. Isaacs, A. F. Lee and K. Wilson, ACS Sustainable Chem. Eng., 2018, 6, 14704–14712.
25. A. A. Dabbawala, S. M. Alhassan, D. K. Mishra, J. Jegal and J.-S. Hwang, Mol. Catal., 2018, 454, 77–86.
26. J. Guo, Y. Song, S. Liu, L. Huang, X. Wang, S. Liu and C. Li, Catal. Sci. Technol., 2021, 11, 4226–4234.
27. Y. Li, H. Wang, L. Wang, Z. Mao, R. Wang, B. He, Y. Gong and X. Hu, Small, 2019, 15, 1804539.
28. Y.-R. Du, B.-H. Xu, S.-P. Xia, G.-R. Ding and S.-J. Zhang, ACS Appl. Mater. Interfaces, 2021, 13, 552–562.
29. M. R. Kamaruzaman, X. X. Jiang, X. D. Hu and S. Y. Chin, Chem. Eng. J., 2020, 388, 124186.
30. Y. Zhang, T. Chen, G. Zhang, G. Wang and H. Zhang, ChemistrySelect, 2020, 5, 1751–1759.
31. Y. Zhang, T. Chen, G. Zhang, G. Wang and H. Zhang, Appl. Catal., A, 2019, 575, 38–47.
32. X. Zhang, D. Yu, J. Zhao, W. Zhang, Y. Dong and H. Huang, Catal. Commun., 2014, 43, 29–33.
33. A. C. Kozen, J. M. Woodward, L. B. Ruppalt, H. Cho, C. A. Ventrice, A. H. Rowley, N. Zhe, A. Mesiti, E. Sargent, J. H. Michels and Z. R. Robinson, ACS Appl. Electron. Mater., 2022, 4, 4280–4287.
34. A. J. Howarth, T. C. Wang, S. S. Al-Juaid, S. G. Aziz, J. T. Hupp and O. K. Farha, Dalton Trans., 2016, 45, 93–97.
35. G. Jouve, C. Séverac and S. Cantacuzène, Thin Solid Films, 1996, 287, 146–153.
36. Y.-F. Wang, X. Yu, Z. Li, M. Pan, M. Jiang and L. Xu, Sens. Actuators, B, 2023, 376, 132978.
37. S. Wang, C. Han, J. Wang, J. Deng, M. Zhu, J. Yao, H. Li and Y. Wang, Chem. Mater., 2014, 26, 6872–6877.
38. P. Zhu, W.-X. Shi, Y. Wang, Z.-M. Zhang, L. Li and C. An, Nano Res., 2023, 16, 6593–6600.
39. H. Feng, J. Ma and Z. Hu, J. Mater. Chem. A, 2010, 20, 1702.
40. A. Wang, R. Fan, X. Pi, Y. Zhou, G. Chen, W. Chen and Y. Yang, ACS Appl. Mater. Interfaces, 2018, 10, 37407–37416.
41. O. L. Li, K. Prabakar, A. Kaneko, H. Park and T. Ishizaki, Catal. Today, 2019, 337, 102–109.
42. Q. Yang, C. Yang, C. Lin and H. Jiang, Angew. Chem., Int. Ed., 2019, 58, 3511–3515.
43. K. Ham, D. Shin and J. Lee, ChemSusChem, 2020, 13, 1751–1758.
44. L. Lai, J. R. Potts, D. Zhan, L. Wang, C. K. Poh, C. Tang, H. Gong, Z. Shen, J. Lin and R. S. Ruoff, Energy Environ. Sci., 2012, 5, 7936.
45. F. A. Bulat, J. S. Murray and P. Politzer, Comput. Theor. Chem., 2021, 1199, 113192.
46. Y. Xie, X. Wang, C. Li, J. Li and J. Zhang, ACS Sustainable Chem. Eng., 2025, 13, 10851–10863.
47. Md. A. Ali, S. K. Moromi, A. S. Touchy and K. Shimizu, ChemCatChem, 2016, 8, 891–894.

---