Activating oxygen deficient TiO₂ in the visible region by Bi₂MoO₆ for CO₂ photoreduction to methanol

Abstract

Fast photogenerated charge recombination and inappropriate bandgap for visible light driven charge generation hinders the performance of TiO₂. In this study, TiO₂ was activated for visible light driven CO₂ reduction in the presence of Bi₂MoO₆ as an electron donor. Furthermore, the introduction of oxygen vacancies resulted in enhanced CO₂ adsorption and conversion. The best catalyst gives 27.1 μmol g⁻¹ h⁻¹ methanol formation. DRIFTS was used to explain the methanol formation mechanism on oxygen deficient TiO₂.

---

Rapid escalations in the demand for energy globally have set researchers on a quest to design and develop new energy production and storage technologies that can meet this demand sustainably. One such route is the use of abundantly available solar energy to convert anthropogenic CO₂ to energy. As an added benefit, this can also mitigate the greenhouse effects caused by the rise in CO₂ levels in the atmosphere.¹ Transition metal oxide-based semiconductors are considered to be promising photocatalysts due to their suitable band energy and excellent stability against photodegradation. For example, TiO₂,² WO₃,³ Bi₂MoO₆,⁴ BiVO₄⁵etc. have been extensively used for the photocatalytic hydrogen evolution reaction (HER) and CO₂ reduction reaction (CO₂RR).⁶

Despite having promising photocatalytic activity, the practical application of TiO₂ is constrained by its poor visible light activity and its tendency to undergo extensive electron–hole recombination, reducing the extent of photogenerated electrons (e⁻) available for photoconversion. Similarly, Bi₂MoO₆, a promising visible light driven photocatalyst shows low quantum efficiency owing to its poor charge separation and slow carrier mobility.⁷ Establishing a heterojunction between two semiconductors is an excellent strategy to significantly increase the photoconversion efficiency by not only modulating the light sensitivity towards the visible region but also promoting the charge carrier separation.⁸ In addition, surface engineering leading to the generation of oxygen vacancies (O_v) can also increase the lifetime of photogenerated e⁻s by creating trap states between the conduction and valence band.⁹ Here, Bi₂MoO₆ is used as an e⁻ donor to activate TiO₂ under visible light. In addition, O_v was created deliberately to offer optimum CO₂ chemisorption opportunities at the TiO₂ surface. With the help of this dual strategy, we were able to achieve a dramatic increment of the CO₂RR performance by suppressing the drawbacks of the individual components. CH₃OH was obtained as the major product, which is regarded as a high energy dense fuel.¹⁰

TiO₂ and Bi₂MoO₆ and composites were synthesised by solvothermal processes. Pre-synthesized TiO₂ was added to an autoclave along with Bi(NO₃)₃ and Na₂MoO₄ to make a suitable interface between TiO₂ and Bi₂MoO₆ for facile e⁻ transfer within the heterojunction. The formation of the heterojunction is depicted via TEM and SEM images in Fig. 1a and b, which demonstrate the anchoring of Bi₂MoO₆ nano-seeds on TiO₂ nano ribbons. The HRTEM image in Fig. S1 (ESI†) depicts the (101) plane of TiO₂ and (131) plane of Bi₂MoO₆, which are in intimate contact with each other to form the heterostructure in the composite catalyst. The nano-seed like morphology of pristine Bi₂MoO₆ and the ribbon like morphology of pure TiO₂ were confirmed from EDX colour mapping which showed that the seed like morphology contains Bi, Mo and O (Fig. S2a–d, ESI†) and the ribbon like morphology mainly consists of Ti and O (Fig. S3a–d, ESI†). The ratio of Bi₂MoO₆ and TiO₂ was varied during the synthesis process to optimize the composite composition for best CO₂ photoreduction activity. The weight ratios of Bi₂MoO₆ and TiO₂ were varied from 3 : 1, to 2 : 1 and 1 : 1 and termed as Bi3@Ti1, Bi2@Ti1 and Bi1@Ti1. Henceforth this nomenclature will be used throughout the manuscript. The PXRD patterns of the TiO₂, Bi₂MoO₆, and Bi₂MoO₆/TiO₂ heterostructures are shown in Fig. S4a (ESI†). Six distinctive peaks observed for TiO₂ at 2θ = 25.28° (101), 37.80° (004), 48.05° (200), 53.89° (105), 55.06° (211) and 62.69° (204) corroborate with anatase TiO₂ (I4₁/amd).¹¹ In the case of Bi₂MoO₆, the diffraction peaks at 28.25°, 32.59°, 33.07°, 46.72°, 47.07°, 55.46°, 55.53° and 56.16° could be perfectly indexed to the (131), (002), (060), (202), (260), (331), (133) and (191) planes of orthorhombic Bi₂MoO₆ (Pca2₁).¹² For the Bi₂MoO₆/TiO₂ heterostructures, all of the peaks can be assigned to either TiO₂ or Bi₂MoO₆. The absence of extra peaks proves the formation of a pure heterojunction. The (101) plane of TiO₂ with very less intensity is observed in the XRD patterns of the composite, which is due to the lower diffraction ability of TiO₂ compared to Bi₂MoO₆. With an increase in the amount of TiO₂, the peak intensity also increases (Fig. S4a, ESI†). Instead of using commercially available TiO₂ (P25), here TiO₂ was treated with a base for increasing O_v in it. The presence of more O_v in the TiO₂ nano ribbons compared to P25 was understood from EPR spectroscopy (Fig. 1c). O_v induces the formation of Ti³⁺ along with Ti⁴⁺ (Ti⁴⁺O₂ → Ti³⁺Ti⁴⁺O_(2−X) + X/2O₂).

Fig. 1 Characterization and CO₂RR performance of Bi₂MoO₆/TiO₂ composites. (a) TEM and (b) SEM image of TiO₂ nanoribbon and Bi₂MoO₆ nanoparticle composite. (c) ESR spectra of as synthesised TiO₂ nanoribbon and commercially available P25 TiO₂ for qualitative understanding of O_v. (d) Comparison of the photocatalytic activity of all the catalysts after 6 h. Here Bi, 31, 21, 11 and Ti refers to Bi₂MoO₆, Bi3@Ti1, Bi2@Ti1, Bi1@Ti1 and TiO₂ respectively. (FA) means full arc contains UV-visible light and (V) means visible light illumination.

The presence of more Ti³⁺ compared to Ti⁴⁺ in the TiO₂ nanoribbons was depicted from X-ray photoelectron spectroscopy (XPS) in Fig. S5 (ESI†). The deconvoluted spectra of P25 also indicated the presence of Ti³⁺ although in much lower quantity compared to the TiO₂ nanoribbons. The 3d¹ e⁻ in Ti³⁺ generates an EPR signal.¹³ Hence, the more intense EPR signal of the TiO₂ nano ribbons compared to P25 clearly indicates the presence of more Ti³⁺ or more O_v in the TiO₂ nano ribbons. These O_v act as hole scavengers (Ti³⁺ + h⁺ → Ti⁴⁺) and oxygen deficient TiO₂ lowers the CO₂ adsorption energy on the TiO₂ surface, which in turn facilitates the CO₂ reduction process (Fig. S4b, ESI†).¹⁴ CO₂ Temperature Programmed Desorption (TPD) in Fig. S6 (ESI†) revealed that the TiO₂ with vacancies has better CO₂ uptake capacity (1.85 μmol g⁻¹) compared to pure TiO₂ (0.53 μmol g⁻¹) (Table S1, ESI†). The CO₂ photoreduction performance of all the catalysts was tested under solar simulated 450 W Xe light illumination in a sealed quartz tube. Methanol was obtained as the major product and CO as a minor product. The CO₂ reduction performance (rate of product formation) of all the catalysts is summarized in Fig. 1d and Table S2 (ESI†). Among all the composite catalysts, Bi2@Ti1 exhibited the best catalytic activity. It achieved a methanol formation rate of 27.1 μmol g⁻¹ h⁻¹ (Fig. S7, ESI†), which is better than most of the TiO₂ based catalysts reported for methanol (Table S3, ESI†). Interestingly, pristine TiO₂ produced more CO compared to methanol under UV-Visible light illumination. Upon switching the illuminated light from the UV-visible to visible region the activity reduced drastically as e⁻s and holes cannot be generated by visible light (>400 nm) illumination on TiO₂ due to its large band gap. On the other hand, pristine Bi₂MoO₆ also could not perform satisfactorily due to fast charge recombination. In the composite material photogenerated e⁻s of Bi₂MoO₆ were transferred to the neighbouring TiO₂ nano ribbons and activated TiO₂ for photocatalysis under visible light. The optimum supply of e⁻s from Bi₂MoO₆ and the presence of O_v in TiO₂ for CO₂ adsorption and activation rendered Bi2@Ti1 the best catalyst for methanol formation. The presence of more Bi₂MoO₆ facilitates more e⁻ for the kinetically demanding methanol formation process (6e⁻) in comparison to the production of CO (2e⁻).¹⁵ Therefore, better methanol formation rate was observed for Bi3@Ti1 compared to Bi1@Ti1 and pristine TiO₂. Furthermore, the role of each component of the CO₂ photoreduction reaction (CO₂RR) was confirmed from a series of control experiments. Finally, the CO₂RR was carried out with ¹³CO₂ and the obtained methanol and CO were examined by GCMS, which clearly shows the formation of ¹³CO (Fig. S8a and c, ESI†) and ¹³CH₃OH (Fig. S8b and d, ESI†) confirming that the generated products are solely coming from the CO₂RR. Cyclability of the spent catalyst showed an unaltered product formation rate for 6 consecutive cycles meaning that the catalyst is stable for repeated performance (Fig. S9, ESI†).

Photocatalytic CO₂RR in the absence of CO₂, light, catalyst and 0.1M NaOH solution showed neither gaseous nor any liquid product formation, indicating the crucial role played by these components. After 6 cycles the catalysts were examined via XRD (Fig. S10, ESI†), TEM (Fig. S11a, ESI†) and SEM (Fig. S11b, ESI†) analyses to understand the structural and morphological integrity. Post catalysis samples displayed similar XRD patterns, and morphology compared to the pristine catalyst. All the catalysts has produced some amount of H₂. However, upon composite formation with Bi₂MoO₆, the HER was suppressed dramatically compared to pristine TiO₂. Therefore, it can be inferred that Bi2@Ti1 not only performed better towards CO₂ reduction but also suppressed the e⁻ wasting competitive HER, which is considered as one of the biggest challenges in photocatalytic CO₂RR. Improved activity of composites and the role of Bi₂MoO₆ as an e⁻ donor can be understood by the analysis of the band structure and charge transfer process. As depicted in Fig. 2a, the absorbance spectra revealed that TiO₂ can only harvest the UV part of the solar spectrum. Therefore, TiO₂ could not perform under visible light illumination. On the other hand, Bi₂MoO₆ has an absorption edge in the visible region. Therefore, the composite materials have UV as well as visible light absorption capability. The agreement between the absorption onset of Bi2@Ti1 and overall apparent quantum yield (AQY%) was understood upon different wavelength chopped light illumination (Fig. 2b). The CO₂ reduction process terminated at more than 475 nm light irradiation and the AQY% trends followed the Bi2@Ti1 absorption pattern meaning that the photocatalysis proceeded via excited e⁻s of Bi₂MoO₆. Therefore, the highest AQY of 0.89% was obtained by 400 ± 10 nm light irradiation where Bi₂MoO₆ shows maximum absorption.

Fig. 2 Band structure and charge transfer process. (a) UV-DRS spectra of all the composites and pristine individuals. (b) Comparison of absorbance and AQY% to check the co-relation between absorbance and activity of Bi2@Ti1 catalyst. (c) Electron transfer route in Bi₂MoO₆/TiO₂ composite catalysts via type-II mechanism. (d) Time dependent photoluminescence spectra of pure Bi₂MoO₆ and Bi2@Ti1.

Further insight of e⁻ transfer was understood by band gap calculation through a Tauc plot (Fig. S12, ESI†) and band alignment calculation by Mott–Schottky measurements. It showed that TiO₂ and Bi₂MoO₆ have bandgaps of 3.19 eV and 2.92 eV, respectively. From the Mott–Schottky measurements it can be said that both the semiconductors are n-type and hence their conduction band maxima (CBM) stay near to the Fermi level. Therefore, the CBM position was obtained by adding −0.1 V with the flat band potential (V_fb) obtained from Mott–Schottky plots (Fig S13a and b, ESI†). As depicted in Fig. 2c the CBM positions of Bi₂MoO₆ and TiO₂ are situated at −0.71 V and −0.58 V. Therefore, the excited e⁻s of Bi₂MoO₆ can easily be transferred to TiO₂ upon visible light irradiation, which can further be used for the CO₂RR. Fig. 2c clearly shows that TiO₂ has enough potential for CO₂ to methanol or CO production by utilizing excited e⁻s of Bi₂MoO₆ because −0.36 V and −0.54 V is the CO₂ to methanol and CO formation potential, respectively. Therefore, the leftover photogenerated holes in the valence band of Bi₂MoO₆ participated in the water oxidation reaction. The evolved O₂ was quantified by gas chromatography (GC) which showed that the amount of evolved O₂ is approximately matching with the stoichiometric amount (Fig. S14, ESI†). The valence band positions were determined using band gap value and CBM position. Time dependent photoluminescence study further elucidated the e⁻ transfer pathway. Fig. 2d and Table S4 (ESI†) show that the pristine Bi₂MoO₆ had lower lifetime for the excited electrons (τ_avg = 2.2 ns) compared to Bi2@Ti1 (τ_avg = 2.7 ns). On the other hand, as shown in Fig. S15 (ESI†) static photo luminescence (PL) study showed that the composites have weaker PL intensities compared to pristine Bi₂MoO₆. This manifests that the excited e⁻s of pristine Bi₂MoO₆ quickly recombine with photogenerated holes, whereas in the composites the e⁻ goes to neighbouring TiO₂ instead of recombining. Therefore, the composites showed less PL intensity and more excited charge lifetime.¹⁶ Photoelectrochemical measurements including interfacial charge transfer resistance and light induce current generation capacity were understood from electrochemical impedance spectroscopy (Fig. S16a, ESI†) and transient photocurrent measurements (Fig. S16b, ESI†). Interestingly, Bi2@Ti1 has the least and pristine Bi₂MoO₆ has the highest charge transfer resistance. An interesting photocurrent generation feature was seen for TiO₂ and the composites. TiO₂ showed rapid photocurrent decay compared to the composites upon light off (Fig. S17, ESI†). Moreover, Bi2@Ti1 exhibited the highest photocurrent. Pristine TiO₂ has delivered no extra current under visible light illumination. The featured photocurrent for TiO₂ in Fig. S16b (ESI†) is due to UV-visible light irradiation. Pristine Bi₂MoO₆ has much less photocurrent production ability and high charge transfer resistance. Therefore, its performance as a photocatalyst is lacking. However, upon composite formation these drawbacks were overcome and the CO₂RR performance increased.

The methanol formation pathway was understood from the identification of intermediates formed during CO₂ hydrogenation via operando DRIFTS (Fig. 3a). The peaks observed at 1372 cm⁻¹ and 1583 cm⁻¹ represent the bidentate and monodentate carbonates generated via adsorption of CO₂ in the O_v–TiO₂ surface.¹⁷ Most importantly, the intensity of the band at 1655 cm⁻¹ corresponding to *COOH has increased with light irradiation time.¹⁸ The *COOH intermediate is regarded as the first and most common intermediate for CO₂ hydrogenation.¹⁸

Fig. 3 In situ study for understanding the methanol formation mechanism. (a) Operando DRIFTS with CO₂ and water vapour under 450 W Xe light illumination with the Bi2@Ti1 catalyst. (b) Plausible methanol formation mechanism on the O_v–TiO₂ surface. It also shows without Bi₂MoO₆, CH₃OH formation became difficult.

A tiny peak evolution at 2095 cm⁻¹ corresponding to the *CO intermediate was also found, which upon desorption from the surface can produce gaseous CO.¹⁸ The most important intermediates of methanol formation are *OCH₃ and *CHO. A monotonous evolution of the IR peaks at 1036 cm⁻¹ and 1114 cm⁻¹ implicitly indicates the presence of these 2 important intermediates.⁵ The *CO intermediate was rapidly converted to the *CHO intermediate; therefore, the intensity of the *CO intermediate remained negligible. On the other hand, the *OCH₃ intermediate showed a prominent peak. Therefore, it can be inferred that the *OCH₃ intermediate has chemisorbed on the O_v–TiO₂ surface for longer time and got protonated on the O site for methanol formation.¹⁹ The absence of methane eliminates the probability of proton adsorption on the C centre of the *OCH₃ intermediate (Fig. S18, ESI†). Based on the knowledge of these intermediates, the formate mechanism for methanol formation can be confirmed. A mechanistic scheme for methanol formation via a 6 proton coupled e⁻ transfer process is shown in Fig. 3b. CO and H₂O were obtained as the by-products through this mechanism. Upon *OCH₃ protonation, the catalyst bed can be regenerated without any chemical corrosion. Moreover, as mentioned previously, the formed heterojunction allows TiO₂ to utilize visible light energy more effectively via additional photo-generated e⁻ transfer from Bi₂MoO₆ for the reaction, whereby the yield of methanol formation is greatly enhanced.

In summary, introducing Bi₂MoO₆ nanoparticles into O_v TiO₂ allows e⁻ transfer through the intimate interface. Thereby, the active sites of TiO₂ were used for CO₂ adsorption and photogenerated charges of Bi₂MoO₆ were used for the CO₂RR. A series of photophysical and photoelectrochemical studies established the regions of best activity in the Bi2@Ti1 catalyst. Finally, the methanol formation mechanism was predicted based on the intermediate's information from operando DRIFTS. Further modification of the reaction conditions can lead to better increment of the performance and can be used for large scale application owing to the high durability of this composite.

Financial support from the Department of Science and Technology (DST) (DST/TMD/(EWO)/IC#5-2018/02) is gratefully acknowledged. SCP thanks DST for the SwarnaJayanti Fellowship (DST/SJF/CSA-02/2017-18). RD, KD and SRC thank CSIR and JNCASR for the research fellowships.

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. (a) W. Tu, Y. Zhou and Z. Zou, Adv. Mater., 2014, 26, 4607–4626; (b) S. Roy, A. Cherevotan and S. C. Peter, ACS Energy Lett., 2018, 3, 1938–1966; (c) S. C. Peter, ACS Energy Lett., 2018, 3, 1557–1561; (d) W. Zhou, Q. Xie, J. Li, C. Zhengwei and F. Peng, Bull. Mater. Sci., 2019, 42, 12; (e) D.-H. Lan, Y.-X. Gong, N.-Y. Tan, S.-S. Wu, J. Shen, K.-C. Yao, B. Yi, C.-T. Au and S.-F. Yin, Carbon, 2018, 127, 245–254.
2. R. Das, S. Chakraborty and S. C. Peter, ACS Energy Lett., 2021, 6, 3270–3274.
3. L. J. Zhang, S. Li, B. K. Liu, D. J. Wang and T. F. Xie, ACS Catal., 2014, 4, 3724–3729.
4. R. Das, K. Das, B. Ray, C. P. Vinod and S. C. Peter, Energy Environ. Sci., 2022 10.1039/D1EE02976B.
5. R. Das, S. Sarkar, R. Kumar, S. D. Ramarao, A. Cherevotan, M. Jasil, C. P. Vinod, A. K. Singh and S. C. Peter, ACS Catal., 2022, 12, 687–697.
6. J.-K. G. Wei Zhou, S. Shen, J. Pan, J. Tang, L. Chen, C.-T. Au and S.-F. Yin, Acta Phys. -Chim. Sin., 2020, 36, 1906048.
7. H. Yu, L. Jiang, H. Wang, B. Huang, X. Yuan, J. Huang, J. Zhang and G. Zeng, Small, 2019, 15, 1901008.
8. (a) L. Etgar, P. Gao, Z. Xue, Q. Peng, A. K. Chandiran, B. Liu, M. K. Nazeeruddin and M. Grätzel, J. Am. Chem. Soc., 2012, 134, 17396–17399; (b) R. Hao, G. Wang, H. Tang, L. Sun, C. Xu and D. Han, Appl. Catal., B, 2016, 187, 47–58; (c) E. Vesali-Kermani, A. Habibi-Yangjeh, H. Diarmand-Khalilabad and S. Ghosh, J. Colloid Interface Sci., 2020, 563, 81–91.
9. (a) Y. Zhang, Z. Xu, G. Li, X. Huang, W. Hao and Y. Bi, Angew. Chem., Int. Ed., 2019, 58, 14229–14233; (b) Y. Fang, Z. Liu, J. Han, Z. Jin, Y. Han, F. Wang, Y. Niu, Y. Wu and Y. Xu, Adv. Energy Mater., 2019, 9, 1803406.
10. (a) A. Cherevotan, J. Raj, L. Dheer, S. Roy, S. Sarkar, R. Das, C. P. Vinod, S. Xu, P. Wells, U. V. Waghmare and S. C. Peter, ACS Energy Lett., 2021, 6, 509–516; (b) D. Bagchi, J. Raj, A. K. Singh, A. Cherevotan, S. Roy, K. S. Manoj, C. P. Vinod and S. C. Peter, Adv. Mater., 2022, 2109426.
11. H. G. Yang, C. H. Sun, S. Z. Qiao, J. Zou, G. Liu, S. C. Smith, H. M. Cheng and G. Q. Lu, Nature, 2008, 453, 638–641.
12. Z. Jia, F. Lyu, L. C. Zhang, S. Zeng, S. X. Liang, Y. Y. Li and J. Lu, Sci. Rep., 2019, 9, 7636.
13. X. Pan, M.-Q. Yang, X. Fu, N. Zhang and Y.-J. Xu, Nanoscale, 2013, 5, 3601–3614.
14. (a) W. Pipornpong, R. Wanbayor and V. Ruangpornvisuti, Appl. Surf. Sci., 2011, 257, 10322–10328; (b) J.-B. Pan, B.-H. Wang, J.-B. Wang, H.-Z. Ding, W. Zhou, X. Liu, J.-R. Zhang, S. Shen, J.-K. Guo, L. Chen, C.-T. Au, L.-L. Jiang and S.-F. Yin, Angew. Chem., Int. Ed., 2021, 60, 1433–1440.
15. Y. Fang and X. Wang, Chem. Commun., 2018, 54, 5674–5687.
16. Y. P. Xie, Y. Yang, G. Wang and G. Liu, J. Colloid Interface Sci., 2017, 503, 198–204.
17. T. H. Tan, B. Xie, Y. H. Ng, S. F. B. Abdullah, H. Y. M. Tang, N. Bedford, R. A. Taylor, K.-F. Aguey-Zinsou, R. Amal and J. Scott, Nat. Catal., 2020, 3, 1034–1043.
18. X. Li, Y. Sun, J. Xu, Y. Shao, J. Wu, X. Xu, Y. Pan, H. Ju, J. Zhu and Y. Xie, Nat. Energy, 2019, 4, 690–699.
19. Z. Sun, T. Ma, H. Tao, Q. Fan and B. Han, Chem, 2017, 3, 560–587.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2cc00490a

---