Investigating the interfacial properties of halide perovskite/TiO_x heterostructures for versatile photocatalytic reactions under sunlight

Abstract

Heterostructures of metal halide perovskites and TiO_x are efficient photocatalytic materials owing to the combination of the advantages of each compound, specifically the high absorption coefficients and long charge-carrier lifetimes of perovskites, and efficient photocatalytic activity of TiO_x. However, chemical reduction of CO₂ using PNC/TiO_x heterostructures without organic solvents has not been reported yet. Here, we report the first solvent-free reduction of CO₂ using amorphous TiO_x with embedded colloidal perovskite nanocrystals (PNCs). The combination was obtained by carrying out hydrolysis of titanium butoxide (TBOT) on the PNC surface without high-temperature calcination. We proposed a mechanism involving photoexcited electrons being transferred from PNCs to TBOT, enabling photocatalytic reactions using TiO_x under visible-light excitation. We demonstrated efficient visible-light-driven photocatalytic reactions at PNC/TiO_x interfaces, specifically with a CO production rate of 30.43 μmol g⁻¹ h⁻¹ and accelerated degradation of organic pollutants under natural sunlight. Our work has provided a simple path toward both efficient CO₂ reduction and photocatalytic degradation of organic dyes.

---

Introduction

As human civilization went through the industrial revolution and the population has increased, our world faces serious global crises, especially global warming¹ and water pollution, both of which threaten the environment and human civilization. Various strategies for decreasing CO₂ emissions and water pollutants have been attempted: biological,² electrochemical,³ photochemical⁴ and photoelectrochemical⁵ conversion techniques have been tested for converting CO₂ to valuable chemicals and fuels; and biodegradation,⁶ electrochemical oxidation⁷ and photocatalytic reduction⁸ have been applied to eliminate dye contaminations. Of these strategies, use of photocatalysts is considered the most promising one for decreasing both atmospheric CO₂ gas and organic pollutants because photocatalysts have wide applicability and can degrade organic materials to harmless products such as H₂O and H₂. This ability can be exploited to reduce CO₂ emission and degrade water pollutants. The process can be also powered by sunlight, which is free and essentially infinite. Photochemical reactions to decrease the amounts of CO₂ or water pollutants have been achieved using various materials such as metal oxides (e.g., TiO₂, Cu₂O, WO₃, Bi₂WO₆, Zn₂GeO₄, ZnGa₂O₄, oxide perovskites),^9–11 metal–organic frameworks,¹² nitrides (e.g., C₃N₄, GaN, (Zn_1+xGe)(N₂O_x)),¹³ sulfides (e.g., CdS, ZnS and Cu₂ZnSnS₄),¹⁴ and phosphides (e.g., InP, GaP).¹⁵ However, these materials have wide bandgaps (for instance, >3.2 eV for TiO₂,¹⁶ and >3.25 eV for WO₃¹⁷), so they initiate catalytic reactions only under ultraviolet (UV) light (wavelength λ < 400 nm). However, only less than 5% of sunlight has photons in these wavelengths, so photocatalytic reactions in conventional photocatalytic materials require an additional UV source. Also, in these conventional materials, recombination of charge carriers is rapid, so the photocatalytic reaction efficiency is relatively poor. Therefore, efficient chemical reduction of CO₂ and degradation of organic pollutants under sunlight require new materials that use wavelengths >400 nm, have high absorption coefficients, and show slow recombination of charge carriers and efficient photocatalytic reaction capabilities.

To achieve these goals, heterostructures in which an “absorber” traps light and transfers excited electrons to a “photocatalyst” have been tested. Halide perovskites have been regarded as promising light absorbers due to their high absorption coefficients, long carrier diffusion paths, long charge carrier lifetimes and tunable bandgaps.^18,19 Amorphous TiO₂ (TiO_x) or crystalline TiO₂ are efficient photocatalysts due to their electrical conductivity, decent photocatalytic activity and excellent chemical stability levels.²⁰

Halide perovskite nanocrystals (PNCs) embedded in TiO_x matrices²¹ or in TiO₂ nanofibers²² have been developed to facilitate chemical reduction of CO₂ by virtue of the photoinduced excitation of charges in the PNCs and transfer and extraction of these excited charges to TiO_x. These reactions are generally carried out with an external power supply (under Xe-lamp irradiation, which covers the ultraviolet-visible light range (100 nm < λ ≤ 800 nm)), with the CO₂ reduction rates measured in the presence of organic solvents (ethyl acetate,²¹ acetonitrile²²). Natural sunlight, which is an essentially infinite resource, can be utilized to realize the photocatalytic reactions, which would decrease both CO₂ emissions and generation of water pollutants without use of an additional external power consumption. To the best of our knowledge, the chemical reduction of CO₂ using PNC/TiO_x (or TiO₂) heterostructures under sunlight illumination has not been reported. Moreover, perovskite-photocatalyzed reduction of CO₂ is often overestimated in CO₂ reduction experiments that are conducted in the presence of organic solvents because the organic solvents can be transformed (at a rate of, for example, ∼1 mmol g⁻¹ h⁻¹) into reduced products such as CO and CH₄.²³ So far, CO₂ reduction using PNC/TiO_x heterostructures without organic solvents has not been reported.

Here we developed a reduction of CO₂ not involving the use of an organic solvent by using PNC/TiO_x heterostructures in which the PNCs absorbed visible sunlight and boosted the photocatalytic reactions at the TiO_x component. TiO_x formed on the surfaces of CsPbBr₃ PNCs as a result of hydrolysis of titanium butoxide (TBOT) without high-temperature calcination. We proposed a mechanism involving transfer of photoexcited electrons to TBOT, allowing for the first solvent-free CO₂ reduction based on PNC/TiO_x heterostructures. PNC/TiO_x heterostructures decreased CO₂ concentration at a rate of 30.43 μmol g⁻¹ h⁻¹ under outdoor sunlight without use of any other external energy. Moreover, we demonstrated more efficient degradation of organic dye under visible white light (λ = 400 nm–780 nm), green light (λ = 500–520 nm) and outdoor sunlight by PNC/TiO_x heterostructures than by pure PNCs and TiO_x.

Results and discussion

Synthesis of PNC/TiO_x heterostructures

PNC/TiO_x heterostructures were synthesized by introducing and then hydrolyzing TBOT (O–Ti(OC₄H₉)₄) on the surface of pre-synthesized CsPbBr₃ PNCs (Fig. 1a). Before synthesis of the heterostructures, the PNCs were cubic with an average size of 9.4 nm, and showed a peak at 511.1 nm in the PL spectrum, with an FWHM of 20 nm (Fig. S1a and b†). After injection of TBOT, the PNC solutions showed a significant decrease in PL intensity with a sudden redshift of the peak in the PL spectrum from 511.1 nm to 514.3 nm, attributed to TBOT having adhered to the surfaces of the PNCs and inducing severe exciton dissociation in them (Fig. 1b and c). As the TBOT was hydrolyzed in air with 40% humidity, composite solutions showed a gradual change of color from bright green to turbid yellow (Fig. 1b, inset), and a gradual decrease of PL intensity with a redshift of the PL signal (Fig. 1c, and S2†).

Fig. 1 (a) Schematic illustrations of the introduction of TBOT into CsPbBr₃ PNCs (left), subsequent hydrolysis of TBOT (middle), and photocatalytic reactions of CsPbBr_3/TiO_x heterostructures under sunlight (right). (b) PL spectrum (inset: optical images of solutions), (c) PL intensity and PL peak wavelength as a function of reaction time, (d) time-resolved PL spectra, (e) PL lifetime and PLQE and (f) schematic illustrations of charge carrier behaviors for formation of CsPbBr_3/TiO_x heterostructures under continuous hydrolysis.

Photophysical properties

To quantify the charge carrier dynamics in the composites, we measured PL lifetimes, and did so by conducting time-correlated single-photon counting (TCSPC) with excitation laser irradiation (wavelength = 405 nm) of the composite solutions. After injection of TBOT into CsPbBr₃ PNCs, the heterostructures showed dramatically decreased τ_ave (from τ_ave of 26.51 ns for CsPbBr₃ PNCs to τ_ave of 1.8 ns for CsPbBr₃/TBOT composites) due to exciton quenching and electron transfer by TBOT on the PNC surfaces (Fig. 1d, e, and Table S1†). This result corresponded well to the observed sudden decrease in PL intensity (Fig. 1b and c). After the initial rapid decrease in PL lifetime upon addition of TBOT, the composite solutions showed gradually increasing τ_ave (τ_ave = 10.33 ns, heterostructures with 180 minutes of reaction). These increases in the PL lifetime were opposite the decrease in the PL intensity (Fig. 1c, d, and S2†). We attributed the increase in PL lifetime upon hydrolysis of TBOT to the aggregation of nanoparticles (Fig. S3–6†).

Using measured average PL lifetimes, we investigated the rate of electron transfer k_ET from PNCs to TBOT. The k_ET parameter was derived using the equation²⁴

where τ_CsPbBr3/TBOT and τ_CsPbBr3 are the PL lifetimes of the CsPbBr₃/TBOT composites and CsPbBr₃ PNC, respectively. And k_ET of CsPbBr₃/TBOT was calculated to be 5.18 × 10⁸, a value 5 times that of the previously reported CsPbBr₃/TiO₂ (k_ET = 9.7 × 10⁷ s⁻¹),²⁴ indicative of efficient charge transfer at the CsPbBr₃/TBOT interface.

Chemical and structural properties

To study the chemical properties of the heterostructure surfaces upon hydrolysis of TBOT, we performed FTIR spectroscopy experiments (Fig. S7a and b†). When TBOT was introduced to the PNC solutions, increases were observed in the intensities of the C–H signals of O–Ti(OC₄H₉)₄ at 3430 cm⁻¹, the COO⁻ and CH₃ signals at 1468–1380 cm⁻¹, the C–H stretching signals at 2923–2854 cm⁻¹ (Fig. S7a†) and the Ti–O–C bond signals at 1125–1035 cm⁻¹ (Fig. S7b†). As the hydrolysis proceeded, the intensities of these signals decreased linearly; this trend confirmed that TBOT became converted to TiO_x as suggested in the PL and time-resolved PL spectra analysis. When the hydrolysis reaction had finished (180 minutes after initiation of hydrolysis when the PL lifetime, intensity and spectrum stopped changing), heterostructures were collected by performing sequential centrifugation and dried by performing evaporation of residual solvent in vacuum.

Then we determined the crystal structure of CsPbBr₃/TiO_x heterostructures by conducting X-ray diffraction (XRD) spectroscopy (Fig. 2a). The heterostructures yielded clear XRD peaks at 15.1°, 21.1°, 29.8°, 29.9°, 34.7°, 36.9° and 44.6°, which corresponded to (100), (110), (004), (200), (210), (211) and (220) crystal planes, respectively; these observations were consistent with peaks of CsPbBr₃ PNCs. Amorphous TiO_x without PNCs did not show any clear XRD peaks. High-resolution TEM showed PNCs embedded in the TiO_x matrix with clear lattice constants of 3.0 and 4.1 Å, which corresponded to, respectively, (002) and (110) planes in the CsPbBr₃ PNCs (Fig. S8†). To study the chemical formation and electrical interaction at the interfaces in the heterostructures, we collected X-ray photoelectron spectroscopy (XPS) data from control TiO_x, PNCs and the heterostructures (Fig. 2b–d, and S9†). The heterostructures showed XPS peaks assigned to Br (∼68.0 eV), Pb (∼137.9 and 143.0 eV) and Ti (∼464.2 eV, 458.2 eV), which indicated the presence of both CsPbBr₃ and TiO_x. Signals of Pb and Br from CsPbBr₃ in the heterostructures were blue-shifted toward higher binding energy compared to those in PNCs (from ∼142.5 eV for 4f_5/2, ∼137.8 eV for 4f_7/2, ∼67.7 eV for 3d_3/2, ∼66.6 eV for 3d_5/2 in PNCs to ∼143.0 eV for 4f_5/2, ∼137.9 eV for 4f_7/2, ∼68.0 eV for 3d_3/2, ∼66.8 eV for 3d_5/2 in heterostructures). These XPS peak shifts indicated that electron transfers from CsPbBr₃ to TiO_x occurred at the interface, in turn attributed to a shallower conduction band minimum (CBM) of CsPbBr₃ (∼−3.9 eV)²⁵ than of TiO_x (∼−4.00 eV).²⁶ This transfer focused the electrons towards TiO_x and boosted the photocatalytic activity of TiO_x.

Fig. 2 (a) XRD patterns and (b) Pb 4f, (c) Br 3d and (d) O 1s regions of the XPS spectra of the CsPbBr₃/TiO_x heterostructures, PNCs and control TiO_x.

Photocatalytic reactions

To evaluate the photocatalytic activities of the heterostructures, we conducted photochemical CO₂ conversion analysis of control TiO_x, PNCs and the heterostructures (results summarized in Table S2†). To prevent side reactions and enable correct measurement of the CO₂ conversion efficiencies of our samples, the light intensity was measured during the reaction (Fig. S10 and Table S3†) and the experiments were performed in a gas–solid reaction system without any sacrificial reagents or organic solvents.²³ First, an external light source was tested for practical analysis. Under visible white light irradiation (400 ≤ λ ≤ 780 nm, Fig. S10†), CsPbBr₃/TiO_x heterostructures showed a CO production rate of 12.77 μmol g⁻¹ h⁻¹ and photoelectron consumption rate of 25.54 μmol g⁻¹ h⁻¹ (Fig. 3a and b). The CO₂ was reduced to CO by consuming two electrons, and no other gasses such as H₂ were detected. PNCs showed a CO production rate of 7.1 μmol g⁻¹ h⁻¹, a value lower than that of the heterostructures, and control TiO_x did not show any CO production, attributed to the inability of TiO_x with its wide bandgap (∼4 eV) to absorb visible light (400 ≤ λ ≤ 780 nm). Furthermore, control TiO_x, PNCs and CsPbBr₃/TiO_x heterostructures showed CO production rates of 3.87, 7.44 and 30.43 μmol g⁻¹ h⁻¹, respectively, under natural sunlight without any additional light source such as an Xe lamp or LED light (Fig. 3c and d). To the best of our knowledge, this was the first reported example of a perovskite/TiO_x-heterostructure-based reduction of CO₂ carried out without using organic solvent.

Fig. 3 (a) CO-production rates and (b) GC chromatograms of a sample without any photocatalytic materials (“none”), control TiO_x, PNCs and heterostructures in white-light illumination, (c) CO-production rates and (d) photograph of CO₂ reduction experiment of heterostructures under outdoor natural sunlight.

To further investigate the photocatalytic activities of the heterostructures, we conducted a photocatalytic degradation of orange-light-emitting PPV copolymer (Merck Orange) under white light (400 ≤ λ ≤ 780 nm), green light (450 ≤ λ ≤ 600 nm) and outdoor natural sunlight. The degradations were monitored by tracking maximum absorbance (Fig. S12 and S13†), observing the solution color with the naked eye, and FTIR spectroscopy (Fig. S14†). The dye degraded much more rapidly in the solution with heterostructures than with control TiO_x and without any catalytic materials: degradation rates of 77.1% with heterostructures, 62% with control TiO_x and 18% without any photocatalytic reactors for 9 hours (h) under white light (Fig. 4a, S12a–c†); 57% for heterostructures, 47.7% for control TiO_x and 8.6% for samples without any photocatalytic reactors for 16 h under green light (Fig. S12d–f, and S15†); and 86.8% with heterostructures, 65.4% with control TiO_x and 47.1% with samples without any photocatalytic reactors for 4 h under outdoor sunlight (Fig. 4d, S12g–i†). We attributed the degradation of organic dyes with pure TiO_x, which cannot absorb visible light, to the charge carriers excited in orange PPV copolymer having been transferred to TiO_x and then inducing the photocatalytic reaction.²⁷ All of the degradations of organic dye under light irradiation followed a pseudo-first-order decay according to the equation ln(C/C₀) = kt, where C₀ is the initial concentration of organic dye, C is the concentration of organic dye at reaction time t, and k is the absorption coefficient.

Fig. 4 (a) Photocatalytic degradation activities, (b) kinetic curves, (c) photographs of PPV copolymer degraded without any photocatalytic materials (top), with control TiO_x (middle), and with PNC/TiO_x heterostructures (bottom) under visible white light. (d) Photocatalytic degradation activities, (e) kinetic curves (inset: photograph of experiment) and (f) photographs of PPV copolymer without any photocatalytic materials (top), with control TiO_x (middle), and with PNC/TiO_x heterostructures (bottom) under outdoor natural sunlight.

Conclusions

We have demonstrated an efficient reduction of CO₂ without organic solvent in PNC systems by forming TiO_x on CsPbBr₃ PNC surfaces as a result of hydrolyzing TBOT without high-temperature calcination. Photoexcited electrons in the resulting heterostructures were concluded to be transferred to TBOT and then TiO_x upon conversion of the TBOT to TiO_x, with this transfer slowing the recombination of charge carriers and having achieved in our experiments a solvent-free CO₂ reduction rate of 30.43 μmol g⁻¹ h⁻¹ under sunlight. To the best of our knowledge, this was the first reported example of PNC/TiO_x-heterostructure-based CO₂ reduction without any organic solvent. We also demonstrated that PNC/TiO_x heterostructures can be used to boost photocatalytic degradation of organic dye under visible white light, green light and outdoor sunlight. This result has suggested the use of halide perovskites as a step towards the development of environmentally benign CO₂ reduction and pollutant reduction.

Note added after first publication

This article replaces the version published on 31^st January 2023. Since the first publication, the authors were made aware of a late reviewer's concerns regarding the time-resolved photoluminescence (PL) lifetime data, which the authors and Editorial Office felt it would be best to address. The outcome of this consideration was that the authors re-conducted the experiment and have now attributed the increase in PL lifetime with hydrolysis of TBOT to an artefactual aggregation of nanoparticles during hydrolysis. Minor changes to the text of the main article, including a revision of eqn (1) and removal of eqn (2) in the Results and discussion section, have been made with the approval of the handling Associate Editor. Changes have also been made to ESI Fig. S3–S6,† most notably the addition of further discussion of the PL lifetimes following Fig. S4.† These changes have been reviewed by the handling Associate Editor and the authors confirm that the main findings of the article have not been significantly altered.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the research fund of National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (2022R1A5A1032539, 2022R1C1C1008282) and Industrial Strategic Technology Development Program-Alchemist Project (1415180859, Chiral perovskite LED smart contact lens based hyper vision metaverse) funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea) and Korea Evaluation Institute of Industrial Technology (KEIT, Korea).

References

1. M. Mendez-Galvan, B. Alcantar-Vazquez, G. Diaz, I. A. Ibarra and H. A. Lara-Garcia, React. Chem. Eng., 2021, 6, 828–838.
2. F. Tassi, S. Venturi, J. Cabassi, O. Vaselli, I. Gelli, D. Cinti and F. Capecchiacci, Org. Geochem., 2015, 86, 81–93.
3. S. Nitopi, E. Bertheussen, S. B. Scott, X. Liu, A. K. Engstfeld, S. Horch, B. Seger, I. E. L. Stephens, K. Chan, C. Hahn, J. K. Nørskov, T. F. Jaramillo and I. Chorkendorff, Chem. Rev., 2019, 119, 7610–7672.
4. U. Ulmer, T. Dingle, P. N. Duchesne, R. H. Morris, A. Tavasoli, T. Wood and G. A. Ozin, Nat. Commun., 2019, 10, 3169–3180.
5. V. Kumaravel, J. Bartlett and S. C. Pillai, ACS Energy Lett., 2020, 5, 486–519.
6. I. Jambon, S. Thijs, G. Torres-Farradá, F. Rineau, N. Weyens, R. Carleer, P. Samyn and J. Vangronsveld, Front. Microbiol., 2019, 10, 1892–1904.
7. C. A. Martínez-Huitle and S. Ferro, Chem. Soc. Rev., 2006, 35, 1324–1340.
8. C. Lu, P. Zhang, S. Jiang, X. Wu, S. Song, M. Zhu, Z. Lou, Z. Li, F. Liu, Y. Liu, Y. Wang and Z. Le, Appl. Catal., B, 2017, 200, 378–385.
9. C. S. Ribeiro, J. Z. Y. Tan, M. M. Maroto-Valer and M. A. Lansarin, J. Environ. Chem. Eng., 2021, 9, 105097–105106.
10. H. Wang, L. Zhang, K. Wang, X. Sun and W. Wang, Appl. Catal., B, 2019, 243, 771–779.
11. L. Liu, H. Zhao, J. M. Andino and Y. Li, ACS Catal., 2012, 2, 1817–1828.
12. H. Zhong, M. Ghorbani-Asl, K. H. Ly, J. Zhang, J. Ge, M. Wang, Z. Liao, D. Makarov, E. Zschech, E. Brunner, I. M. Weidinger, J. Zhang, A. v. Krasheninnikov, S. Kaskel, R. Dong and X. Feng, Nat. Commun., 2020, 11, 1409–1418.
13. P. Huang, J. Huang, S. A. Pantovich, A. D. Carl, T. G. Fenton, C. A. Caputo, R. L. Grimm, A. I. Frenkel and G. Li, J. Am. Chem. Soc., 2018, 140, 16042–16047.
14. R. Zhou and M. I. Guzman, J. Phys. Chem. C, 2014, 118, 11649–11656.
15. B. Chon, S. Choi, Y. Seo, H. S. Lee, C. H. Kim, H. J. Son and S. O. Kang, ACS Sustainable Chem. Eng., 2022, 10, 6033–6044.
16. C. Dette, M. A. Pérez-Osorio, C. S. Kley, P. Punke, C. E. Patrick, P. Jacobson, F. Giustino, S. J. Jung and K. Kern, Nano Lett., 2014, 14, 6533–6538.
17. R. S. Vemuri, M. H. Engelhard and C. v. Ramana, ACS Appl. Mater. Interfaces, 2012, 4, 1371–1377.
18. M. Pawar, S. T. Sendoǧdular and P. Gouma, J. Nanomater., 2018, 2018, 5953609–5953621.
19. G. Rainò, M. A. Becker, M. I. Bodnarchuk, R. F. Mahrt, M. v. Kovalenko and T. Stöferle, Nature, 2018, 563, 671–675.
20. A. Inbal, J. P. Ampuero and R. W. Clayton, Science, 2016, 354, 88–92.
21. Y. F. Xu, X. D. Wang, J. F. Liao, B. X. Chen, H. Y. Chen and D. B. Kuang, Adv. Mater. Interfaces, 2018, 5, 1801015–1801022.
22. F. Xu, K. Meng, B. Cheng, S. Wang, J. Xu and J. Yu, Nat. Commun., 2020, 11, 4613–4621.
23. J. S. Martin, N. Dang, E. Raulerson, M. C. Beard, J. Hartenberger and Y. Yan, Angew. Chem., Int. Ed., 2022, 61, e202205572–e202205579.
24. X. Liu, H. Zhao, L. Wei, X. Ren, X. Zhang, F. Li, P. Zeng and M. Liu, Nanophotonics, 2021, 10, 1967–1975.
25. H. Algadi, C. Mahata, J. Woo, M. Lee, M. Kim and T. Lee, Electronics, 2019, 8, 678–690.
26. J. Fujisawa, T. Eda and M. Hanaya, Chem. Phys. Lett., 2017, 685, 23–26.
27. S. A. Kumar, J. S. Shankar and B. K. Periyasamy, J. Mater. Sci., 2022, 57, 12449–12462.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2nr06840k

---