One-pot/self-template synthesis of mesostructured vanadium oxide embedded carbon nanofiber as a visible-light photocatalyst

Abstract

Based on a one-pot/self-template synthesis strategy, mesostructured 1D vanadium oxide-embedded carbon nanofiber was prepared by combining electrospinning and carbonization methods. Cost-saving ammonium metavanadate was used as a semiconductor precursor and porogen. The novel membrane exhibited highly efficient and stable visible light-driven photocatalytic activity.

---

Much effort has been exerted to solve the widespread biorecalcitrant and organic water pollutant problems produced by some industries.¹ Among decomposition methods of organic pollutants, photocatalysis is a “green” method that offers considerable perspective for complete decomposition of organic pollutants in the environment.^2–4 Thus, numerous kinds of semiconductor metal oxide photocatalysts, such as Bi₂O₃, Fe₂O₃, TiO₂, and ZnO, have been developed to create a better environment for humans.^5–7 Vanadium oxides are extensively used as catalysts with a band gap of approximately 2.6 eV.⁸ This narrow band gap could make it active in a wide region of visible light frequencies (<600 nm) covering most of the solar light region. Thus, vanadium oxides are a type of prospective solar light active catalyst. However, the rapid electron–hole pair recombination makes the vanadium oxides containing catalysts to exhibit low activity in both ultraviolet (UV) and visible (vis) lights.⁹ To enhance the photocatalytic performance of catalysts, two approaches are commonly implemented. The first approach centers on developing nanoscaled or porous photocatalysts with greatly enhanced surface areas. These large surface areas provide more active sites where the photocatalytic reaction can occur. In addition, the nanoscaled materials shorten the distance in which electrons and holes migrate from bulk to reaction active sites.^10–12 The second approach focuses on improving the separation efficiency of photogenerated carriers (photogenerated electrons and holes) by supporting appropriate co-catalysts (such as noble metals and metal oxides)^13,14 or constructing catalysts with various compositions.^15,16 Therefore, we believe that high-performance photocatalysts can be obtained when these two strategies are combined.

Carbon–semiconductor hybrid materials are a new class of photocatalysts that have elicited considerable attention.^17,1 Highly conductive, long carbon nanofibers (CNFs) can efficiently capture and transport photogenerated electrons.⁷ However, low surface area is the critical disadvantage for these 1D carbon materials, as it inhibits their practical application.^18,19 Given this reason, numerous studies have focused on exploiting state-of-the-art porous CNFs to build new devices with improved functions.^20,17 Conventional processes of producing porous carbon materials require a post-treatment called activation. Activation method involves complex chemical and physical processes that take place at multiple times and at different temperature scales. And, the template methods still have numerous unsolved technical difficulties.²¹ Electrospinning is a simple and low-cost approach to engineer nanofibrous membranes with large surface area and controlled pore structure.^22,17 Only a few methods have been reported for the direct synthesis of porous CNFs by carbonization of electrospun 1D polymers. Li et al.²¹ reported on the self-template method using ethylene glycol as the carbon precursor and Zn(CH₃COO)₂ as the structural constructor and porogen to prepare mesoporous CNFs with 3D interconnected mesoporous structure and large surface area. Poly(acrylonitrile) (PAN) a well-known polymer with good stability and mechanical properties, has been widely used in producing carbon nanofibers (CNFs). However, PAN has not been used as a carbon precursor for porous CNF production by the self-template/direct synthesis method. Therefore, developing a template-free and economical method to conveniently synthesize porous CNFs with high surface area and large pore size for different electrochemical applications is a great challenge. Mesoporous CNF can be a good matrix for the separation of photon electrons and holes. Thus, a highly efficient photodegradation catalyst can be prepared by incorporating vanadium oxide (VO_x) into CNF matrix.

This study is the first to report a new photocatalyst of 1D mesostructured vanadium oxide VO_x membrane (V-MCNF) obtained through simple combination of electrospinning and carbonization based on the one-pot/self-template system. The novel photocatalyst membrane has several important features. First, the 1D structure not only provides shortened pathways for electron transport, but also facilitates electrolyte penetration from the direction perpendicular to the longitudinal axis along the fiber. Second, the generated pores in CNFs can connect the vanadium oxide to the CNF surface and create multi-channeled reaction paces available for mass transfer, which offer remarkably higher accessible surface areas for contaminant photodegradation. Third, long fibril morphology of the membrane is useful for recovery and reuse. The combination of all of these important features in one material results in a distinct photocatalysis with considerable efficiency and stability.

The reagents, preparation process of V-MCNF membrane, and characterization of V-MCNF membrane are shown in the ESI.†

Inorganic ammonium metavanadate (NH₄VO₃) was used as a precursor, as it is cost effective and more suitable for industrial production. NH₄VO₃ is not only a precursor of vanadium oxide but also provides a critical contribution by generating a mesoporous structure. NH₄VO₃ decomposes at low temperature of 200 °C to 280 °C and produces intermediate vanadyl compound species [e.g., (NH₄)₂V₆O₁₆ and NH₄V₄O₁₀] before V₂O₅ formation at 450 °C.²³ V₂O₅ has a melting point of 700 °C (2B1258; Junsei), which can be sufficiently dissociated by heat treatment as metal and oxygen at a higher temperature than the m.p. of metal oxide. Dissociated oxygen in carbon fibers formats carbon monoxide and carbon dioxide during carbonization and generate the path between metal catalyst and pores over the structure. Polyvinylpyrrolidone (PVP) was selected as a co-polymer with PAN to improve the spinnability of polymer mixture solution by reducing the viscosity, thereby decreasing the electrospun nanofiber diameter.

Fig. 1(A) shows an SEM image of pure CNFs. These randomly oriented CNFs have a smooth and uniform surface. The diameter of the pure CNFs ranged from 250 nm to 300 nm. The SEM image of V-MCNF membrane shown in Fig. 1(B) indicates that the nanofiber became curvy, and the diameter shrank to 150 nm to 200 nm. The latter images suggest that the nanofiber morphology could be kept intact after carbonization. Application of higher concentration of NH₄VO₃ (0.2 g) resulted in aggregation of vanadium oxide on the CNF matrix (ESI Fig. S1†).

Fig. 1 (A) SEM image of pure CNFs and (B) SEM image of V-MCNF.

Fig. 2A and B present the TEM and HRTEM images of pure CNF, respectively. The CNF exhibited a smooth and uniform surface without any other nanostructure, which is consistent with the SEM observation. Fig. 2C and D show the TEM and HRTEM images of V-MCNF, respectively; a homogenous incorporation of vanadium oxide (VO_x) in the CNF matrix was observed. VO_x incorporation was indicated as connected white spots in a spherical morphology, which previously grew in the black CNF. The white spots caused by NH₄VO₃ decomposition and oxygen dissociation in CNF during carbonization finally resulted in pore generation.

Fig. 2 (A) TEM image of pure CNF membrane, (B) HRTEM image of pure CNF membrane, (C) TEM image of V-MCNF membrane, and (D) HRTEM image of V-MCNF membrane.

Fig. 3(a) shows the XRD patterns of CNF and V-MCNF. The XRD patterns of the CNF and V-MCNF show a graphite diffraction peaks at 25° and 43° corresponding to (002) and (101) planes, indicate the presence of graphite-like carbon in both formulations. However, the broad peaks and a complete absence of higher order peaks in the XRD patterns suggest low graphitization and disordered amorphous structure for the prepared materials.²⁴ The XRD pattern of V-MCNF showed that no peaks attributed to vanadium oxides were detected. This indicating that vanadium oxides were successfully incorporated into the crystal lattice of CNFs or it were too little to be detected by XRD. However, the vanadium ion concentration was 0.48%, as detected by the ICP. In addition, the intensity of graphitic CNF peaks became weaker because of the destruction of graphitic layers while the decomposition of NH₄VO₃ by carbonization.

Fig. 3 (a) XRD patterns of the samples of pure CNF and V-MCNF membranes and (b) UV/vis diffused reflectance spectra of the samples CNF and V-MCNF membranes.

Fig. 3(b) shows the UV-vis spectra of CNF and V-MCNF membranes. The adsorption spectrum of CNF membrane displayed no absorption peaks from 400 nm to 800 nm. However, the adsorption spectrum of V-MCNF membrane showed a broad absorption from 400 nm to 800 nm, which is in accordance with previous reports. These reports indicated that V⁵⁺ absorption was lower than 570 nm and that V⁴⁺ absorption band centered at 770 nm.^25,26 Therefore, the electron spectrum of V-MCNF suggested the co-existence of V⁵⁺ and V⁴⁺ species in the membrane. Consequently, the composite materials can make use of the whole visible light region. The band gap energy (E_bg) of the V-MCNF photocatalyst was estimated using the following equation:

E_bg = 1240/λ[eV],

where λ is the wavelength in nanometers.²⁷ and E_bg was approximately 3.8 eV, which is larger than the value of 2.6 eV reported for the V₂O₅.⁸ These results also confirm that the obtained vanadium oxide phase is different from the well-known monoclinic and tetragonal phases (∼0.7 eV)²⁸ and may be a new phase in the vanadium oxide family. To further study the chemical and bonding environment of the components of the prepared V-MCNF membrane including carbon phase and vanadium oxide, X-ray photoelectron spectroscopy (XPS) was applied. ESI, Fig. S2(a)† shows the high resolution XPS spectra of the C1s region at approximately 285 eV. The binding energy with a peak at 284.4 eV was attributed to C=C.²⁹ The peaks at 286 and 289 eV were characteristic of the oxygen-bound species C–O and C=O,^30,31 respectively. ESI, Fig. S2(b)† show the high resolution XPS spectrum of V2p. The binding energy at 517.2 eV corresponded to the vanadium species in the oxidation state +5. A small component at lower binding energy (516.0 eV) could be attributed to vanadium species in a lower oxidation state, i.e., V⁴⁺.^32,33 The XPS analysis of O s1 [ESI, Fig. S2(c)†] detected a binding energy at 533.4 and 535 eV. These binding energies indicate that the different oxygen species present in the near surface region, such as OH⁻.³⁴ The hydroxyl groups are responsible for the improved surface hydrophilicity during visible-light irradiation. Hydrophilicity is one of the reasons for enhanced visible-light photocatalytic performance of V-MCNF.^34,35 Another reason is the synergetic effect between CNFs and semiconductive metal oxides.⁷

Fig. 4(a) shows the nitrogen adsorption–desorption isotherms for pure CNF and V-MCNF. The adsorption–desorption isotherms of CNF showed a typical type I behavior representing the microporous adsorption. Nitrogen adsorption was nearly complete at a low relative pressure (P/P₀ < 0.1). However, the adsorption isotherms of V-MCNF exhibited a combination of type I and II characteristics. Hysteresis was observed at a relative pressure higher than P/P₀ = 0.5, which was a typical type II behavior of mesoporous adsorption. Micropore filling was observed at a low relative pressure of P/P₀ < 0.1, which was a typical type I behavior.³⁶ The micropores were generated during carbonization.²¹ The mesopores were created by the decompositon of the embedded NH₄VO₃ in the carbon matrix and the formation of carbon monoxide and carbon dioxide at a higher temperature than the m.p. of V₂O₅ during carbonization. NH₄VO₃ decomposition was confirmed by measuring the concentrations of vanadium ions using the ICP, which decreased from 1.23% before the carbonization to 0.48% (25.3 mg kg⁻¹) after the carbonization. The results suggest the porogen and the structural constructor function of NH₄VO₃. The generated mesopores contributed to the BET V-MCNF surface area of 600 m² g⁻¹. This value is immensely larger than the BET surface area of the pure CNFs (254 m² g⁻¹). The higher surface area of V-MCNF sample clearly indicated that most of the surface area was produced upon the removal of embedded NH₄VO₃ in the carbon matrix. This high specific surface area could supply sufficient space for the MB molecule to be adsorbed onto the surface and into the pore. MB adsorption greatly enhanced the photocatalytic performance. The pore size distribution curves of V-MCNF and pure CNF were calculated using the Barret–Joyner–Halenda model and presented in Fig. 4(b). The V-MCNF curve clearly confirmed the existence of trimodal pores with detectable sizes of 1.7, 4, and 6 nm to 20 nm. Pure CNF exhibited micropores at a mean value of 1.7 nm, which were generated during the carbonization process.

Fig. 4 (a) N₂ adsorption–desorption isotherms and (b) the corresponding pore size distribution of the sample of pure CNF and V-MCNF membranes.

The photocatalytic activity and stability of the prepared V-MCNF membrane were investigated by examining the organic pollutant MB under visible-light irradiation conditions. The absorption spectra of MB aqueous solution showed the change of its concentration. The MB concentration (%) after various intervals of time was estimated using the following equation:

%MB concentration = (A/A₀) × 100,

where A₀ is the initial MB absorbance at 566 nm and A is the absorbance obtained after various intervals of time. The order of photocatalytic reaction was calculated by ln(C/C₀) = k(min⁻¹)t, where k is the apparent reaction rate constant, C₀ is the initial concentration of aqueous dye solution, and C is the concentration of aqueous dye solution at the reaction time t.

Fig. 5(A) shows that the adsorption–desorption equilibrium of MB in the dark was established within 30 min. V-MCNF recorded the highest adsorption capacity (23%) because of its higher surface area with mesoporous structure. The increase of irradiation time resulted in a steady progress of MB decomposition. The decomposition over the V-MCNF catalyst was completed in 60 min of visible-light irradiation. Fig. 5(A) also shows that no significant removal of MB occurred with the addition of CNF or VO_x within 90 min under visible-light irradiation. Notably, in the presence of V-MCNF membrane, rapid degradation was achieved, and the degradation efficiency of MB was approximately 50% after 15 min and reached approximately 98% after 60 min. These results indicated that the bleaching reaction followed a pseudo-first-order reaction, and the MB decomposition rate over V-MCNF was estimated to be 0.0033 min⁻¹, which is much faster than that of commercial P25 (Degussa) (2.66 × 10⁻⁴ min⁻¹). The photocatalyst stability is very important for application in environmental technology. Therefore, four-cycle experiments of MB degradation over the V-MCNF membrane were carried out under the same conditions. Fig. 5(B) shows that the photocatalytic degradation efficiency of the novel membrane under visible-light irradiation remained at high efficiency even after four cycles. More importantly, this heteroarchitecture membrane with high photocatalyst activity could be easily separated and recovered by sedimentation, which can greatly promote its practical application to eliminate the organic pollutants from wastewater.

Fig. 5 (A) Photodecomposition curve of MB (30 ml of aqueous MB, 20 mg L⁻¹, and 10 mg catalyst). (B) Four cycling experimental results of the V-MCNF membrane photocatalyst. Estimated error: 0.15.

Based on the above-mentioned experimental facts and analysis, we could propose an explanation for the high photocatalytic performance observed from our V-MCNF membrane. The CNF-supported catalysts were believed to exhibit cooperative or synergetic effects between CNFs and semiconductive metal oxides. In this study, a high-energy photon excited an electron from the valence band to the conduction band of VO_x. Photogenerated electrons in VO_x may have moved freely toward the CNF surface, and excess valence band holes were left in the VO_x to migrate to the surface and react with H₂O or OH⁻ to produce active species, such as OH˙.¹² This finding suggests that the photogenerated electrons and holes were efficiently separated. These actions of electrons not only prevent the recombination of the hole and electron efficiently, but also avoid being captured by VO_x. Thus, the activity was maintained at a high level. More importantly, the stability was ensured simultaneously. In addition, the remarkably fast photodegradation performance was due to the unique mesostructured material and the high surface area (600 m² g⁻¹) that increased the surface adsorption capacity of the reactants and led to the enhancement of photocatalytic reactions.

In summary, a mesostructured vanadium oxide-embedded CNF (V-MCNF) membrane was demonstrated by combining electrospinning technique with carbonization process based on one-pot/self-template synthesis approach. We used PAN as carbon precursor, NH₄VO₃ as semiconductor precursor and porogen, and PVP as a copolymer to improve mixture polymer spinnability. The prepared membrane displayed high photocatalytic degradation capacity of MB under visible-light irradiation. Furthermore, the V-MCNF membrane could be easily recycled without decrease of the photocatalytic activity because of its 1D nanostructure. The novel membrane proved to be a potential candidate to eliminate organic pollutants from wastewater.

This work was supported by Ministry of Science and Technology of China (Grant nos 2010DFA92820, 2010DFA92800 and 2012DFG91870) and China-American Cooperation for 10+10 (Grant no. 2010DFA91130).

Notes and references

1. Q. Xiang, J. Yu and M. Jaroniec, Chem. Soc. Rev., 2012, 41, 782–796.
2. T. Arai, M. Horiguchi, M. Yanagida, T. Gunji, H. Sugihara and K. Sayama, Chem. Commun., 2008, 5565–5567.
3. H. Zeng, W. Cai, P. Liu, X. Xu, H. Zhou, C. Klingshirn and H. Kalt, ACS Nano, 2008, 2, 1661–1670.
4. X. Zhou, C. Hu, X. Hu, T. Peng and J. Qu, J. Phys. Chem. C, 2010, 2746–2750.
5. L. Ji, O. Toprakci, M. Alcoutlabi, Y. Yao, Y. Li, S. Zhang, B. Guo, Z. Lin and X. Zhang, ACS Appl. Mater. Interfaces, 2012, 4, 2672–2679.
6. J. Yu, J. Fan and B. Cheng, J. Power Sources, 2011, 196, 7891–7898.
7. J. Mu, C. Shao, Z. Guo, Z. Zhang, M. Zhang, P. Zhang, B. Chen and Y. Liu, ACS App. Mater. Interfaces, 2011, 3, 590–596.
8. T. F.-R. Shen, M.-H. Lai, T. C.-K. Yang, I.-P. Fu, N.-Y. Liang and W.-T. Chen, J. Taiwan Inst. Chem. Eng., 2012, 43, 95–101.
9. F. Chen, T. Wu and X. Zhou, Catal. Commun., 2008, 9, 1698–1703.
10. O. K. Varghese, M. Paulose, T. J. Latempa and C. a. Grimes, Nano Lett., 2009, 9, 731–737.
11. H. Li, Z. Bian, J. Zhu, Y. Huo, H. Li and Y. Lu, J. Am. Chem. Soc., 2007, 4538–4539.
12. K. Woan, G. Pyrgiotakis and W. Sigmund, Adv. Mater., 2009, 21, 2233–2239.
13. H. Zeng, W. Cai, P. Liu, X. Xu, H. Zhou, C. Klingshirn and H. Kalt, ACS Nano, 2008, 2, 1661–1670.
14. H. Tada, T. Kiyonaga and S. Naya, Chem. Soc. Rev., 2009, 38, 1849–1858.
15. H. Zhang, G. Chen and D. W. Bahnemann, J. Mater. Chem., 2009, 19, 5089.
16. H. Zhang, X. Lv, Y. Li, Y. Wang and J. Li, ACS Nano, 2010, 4, 380–386.
17. L. Liao, M. Zheng, Z. Zhang, B. Yan, X. Chang, G. Ji, Z. Shen, T. Wu, J. Cao and J. Zhang, Carbon, 2009, 47, 1841–1845.
18. P.-L. Taberna, G. Chevallier, P. Simon, D. Plée and T. Aubert, Mater. Res. Bull., 2006, 41, 478–484.
19. C. Portet, P. L. Taberna, P. Simon and E. Flahaut, J. Power Sources, 2005, 139, 371–378.
20. H. Wang, X. Bo, J. Ju and L. Guo, Catal. Sci. Technol., 2012, 2, 2327.
21. W. Li, F. Zhang, Y. Dou, Z. Wu, H. Liu, X. Qian, D. Gu, Y. Xia, B. Tu and D. Zhao, Adv. Energy Mater., 2011, 1, 382–386.
22. Y. Wu, W. Jia, Q. An, Y. Liu, J. Chen and G. Li, Nanotechnology, 2009, 20, 245101.
23. W. M. Shaheen and I. H. A. El Maksod, J. Alloys Compd., 2009, 476, 366–372.
24. A. G. El-Deen, N. a.M. Barakat, K. A. Khalil and H. Y. Kim, New J. Chem., 2014 10.1039/c3nj00576c.
25. F. Amano, T. Yamaguchi and T. Tanaka, J. Phys. Chem. B, 2006, 110, 281–288.
26. X. Gao and I. E. Wachs, J. Phys. Chem. B, 2000, 104, 1261–1268.
27. V. Štengl, S. Bakardjieva, N. Murafa, V. Houšková and K. Lang, Microporous Mesoporous Mater., 2008, 110, 370–378.
28. S. D. Perera, B. Patel, N. Nijem, K. Roodenko, O. Seitz, J. P. Ferraris, Y. J. Chabal and K. J. Balkus, Adv. Energy Mater., 2011, 1, 936–945.
29. X. Mao, F. Simeon, G. C. Rutledge and T. A. Hatton, Adv. Mater., 2013, 25, 1309–1314.
30. P. Zhang, C. Shao, Z. Zhang, M. Zhang, J. Mu, Z. Guo, Y. Sun and Y. Liu, J. Mater. Chem., 2011, 21, 17746.
31. M. Zhang, C. Shao, X. Li, P. Zhang, Y. Sun, C. Su, X. Zhang, J. Ren and Y. Liu, Nanoscale, 2012, 4, 7501–7508.
32. Y. Wang, Y. R. Su, L. Qiao, L. X. Liu, Q. Su, C. Q. Zhu and X. Q. Liu, Nanotechnology, 2011, 22, 225702.
33. A. Mekki, G. D. Khattak and L. E. Wenger, J. Electron Spectrosc. Relat. Phenom., 2009, 175, 21–26.
34. M. Long, W. Cai, J. Cai, B. Zhou, X. Chai and Y. Wu, J. Phys. Chem. B, 2006, 20211–20216.
35. H. U. Lee, S. C. Lee, S. Choi, B. Son, S. M. Lee, H. J. Kim and J. Lee, Chem. Eng. J., 2013, 228, 756–764.
36. D.-W. Wang, F. Li, M. Liu, G. Q. Lu and H.-M. Cheng, Angew. Chem., Int. Ed., 2008, 47, 373–376.

---

Footnote

† Electronic supplementary information (ESI) available: The reagents, preparation process of V-MCNF membrane, characterization of V-MCNF composite nanofibers, description of the photocatalytic degradation tests, SEM image of CNF with higher concentration of NH4VO3 (0.2 g), XPS spectra of C1s, V2p and O1s for V-MCNF were shown in ESI. DOI: 10.1039/c3ra42695e

---