Abstract
Oxygen vacancy (VO) has been recognized to possess an effect to promote the charge separation and transfer (CST) in various n‐type semiconductor based photoelectrodes. But how external stimulus will change this VO effect has not been investigated. In this work, external polarization is applied to investigate the effect of VO on the CST process of a typical ferroelectric BiFeO3 photoelectrode. It is found that negative poling treatment can significantly boost VO effect, while positive poling treatment will deteriorate the CST capability in BiFeO3 photoelectrodes. This poling history determined VO effect is rooted in the VO induced defect dipoles, wherein their alignment produces a depolarization electric field to modulate the CST driving force. This finding highlights the significance of poling history in functionalizing the VO in a photoelectrode.
Keywords: oxygen vacancy, defect dipole, dipole moment, charge separation and transfer
The effect of oxygen vacancy in influencing charge separation and transfer capability depends on the poling history the photoelectrode has experienced, including the poling direction, poling voltage, relaxation duration and so on.

Introduction
Oxygen vacancy (VO) widely exists in different metal oxides due to the intrinsic oxygen deficiency during synthesis.[ 1 , 2 ] Due to the charge compensation of VO, [3] it endows the n‐type conductivity for many metal oxide semiconductors, which lays the foundation for many significant applications such as blue light emission diode,[ 4 , 5 ] field effect transistor[ 6 , 7 ] and solar energy conversion.[ 2 , 5 ] Especially in the community of solar energy conversion, it is commonly accepted that moderate concentration of VO is beneficial to charge separation and transfer (CST) process of n‐type semiconductors applications, such as photoanodes for photoelectrocatalysis.[ 12 , 13 , 14 , 15 ] But this knowledge is acquired under classical environment with light irradiation only, how the VO effect will change under external stimulus is still yet to be understood.
As a kind of typical lattice defect, VO can be positively charged and theoretically create defect dipoles in metal oxides, [16] which can change the polarization of a crystal.[ 17 , 18 ] In photocatalysts design, polarization has been applied to generate additional electric field for efficient photogenerated charge separation and transfer (CST).[ 19 , 20 , 21 , 22 ] In our previous research based on ferroelectric BiFeO3 photoanodes, we also have revealed that proper alignment of the polarization direction can largely tune the photoresponse in photoelectrodes. [10] Considering that the VO induced defect dipoles may interfere with the intrinsic dipoles in ferroelectric materials,[ 24 , 25 ] it raises an interesting question: what is the influence of VO on the performance of ferroelectric photoanodes? Clarifying this question will provide a more comprehensive understanding about the nature of VO, providing guide to the application of VO in various fields.
Herein we use BiFeO3, a ferroelectric with narrow bandgap and high Curie temperature, [12] as a prototype to gain insight into the mystery of VO’s function. The VO concentration is first tailored via conventional heating treatment. Further photoelectrochemical (PEC) investigation shows the influence of VO is dramatically dependent on the poling history. Higher VO concentration results in more profound poling voltage dependent VO effect. In addition, relaxation of aligned dipoles leads to photocurrent decay, which in turn can be fully recovered via another poling process. This poling history dependent VO effect on the CST process in a semiconductor photoelectrodes can provide new insights into the application of VO in tuning the optoelectrical feature of semiconductors.
Results
The BiFeO3 photoelectrode was prepared with previously reported spin coating method, achieving a condensed film with a thickness around 100 nm (Figure S1). [10] The light absorption (Figure S2) and X‐ray diffraction patterns (Figure S3) have demonstrated the formation of well crystallized BiFeO3. To induce VO, the BiFeO3 photoelectrode (BFO) undergoes another annealing treatment at 600 °C in N2 atmosphere for 4 hours with the samples noted as BFO‐VO (See Supporting Information for detail). This heat treatment does not show significant change of the light absorption and crystal phase. The UV‐Visible light absorption spectra of the photoanodes display similar band gap of around 2.2 eV before and after N2 treatment (Figure S2). Moreover, this treatment has been reported to create oxygen vacancies in other semiconductors and tuning heating duration can tailor the VO concentration. [29] To further check whether this treatment has successfully produced extra VO, we first tested the Mott‐Schottky (M‐S) curves. With the generation of VO, extra‐charge will be compensated to achieve higher carrier concentration with lower M‐S slope. As shown in Figure 1a, the M‐S slope of BFO‐VO (1.48×1010 F−2 cm4 V−1) has been reduced by one order of magnitude compared to BFO (2.58×1011 F−2 cm4 V−1), suggesting a much higher VO concentration in BFO‐VO. From the intersection of M‐S curves of BFO and BFO‐VO on E‐axial (Figure S4), the flat band potential of them is 0.20 V and 0.22 V vs RHE, respectively.
Figure 1.
The comparison between BFO (black) and BFO‐VO (red) photoelectrodes in terms of (a) the Mott‐Schottky curves, (b) X‐ray absorption of O K‐edge spectra, and (c) X‐ray absorption of Fe L‐edge spectra.
Furthermore, the X‐ray photoelectron spectroscopy (XPS) has been adopted to show the O 1s chemical states.[ 8 , 9 , 10 , 11 , 29 ] With the N2 treatment varied from 0 to 4.0 h, the XPS peak area attributes to VO gradually increased, aligned with the change of the carrier concentration in Figure S5. These results further confirm the presence of VO during N2 treatment.
To future check the electronic structure change by VO, we have investigated the chemical states change of oxygen and iron via near edge X‐ray absorption fine spectroscopy (NEXAFS).[ 9 , 10 , 11 ] According to previous reports,[ 30 , 31 ] the formed VO will possess positive charges with the balanced electrons been accepted by the neighbor cations (Fe in BiFeO3), wherein the NEXAFS can detect the electron cloud change around the atoms.
From the O K‐edge spectra (Figure 1b), it observes a shift towards the high energy for the electrons in t2g and eg orbitals, indicating that O atoms are electron deficient.[ 32 , 33 ] As an obvious comparison, the Fe L‐edge spectra show a negative shift with the VO existence (Figure 1c), suggesting Fe is in an electron rich status.[ 34 , 35 , 36 ] As a conclusion, the change in the electron cloud around O and Fe atoms clearly evidences the existence of VO in BFO‐VO.
With the above characterizations about the structure and electronic status changes, it solidly confirms the generation of VO in the BFO‐VO sample. The presence of VO will interfere with the intrinsic ferroelectric property of BiFeO3.[ 37 , 38 , 39 ] The phase images of piezoelectric force microscope (PFM) of BFO and BFO‐VO show that both samples have a 180° phase switch due to the random orientation of the ferroelectric domains (Figure 2 a and b). [40] But the PFM amplitude loop shows significant difference between pristine BFO and BFO‐VO. As shown in Figure 2c, a symmetric butterfly loop in the amplitude response is observed in BFO. After introducing VO, the PFM amplitude loop is significantly distorted, indicating the presence of defect dipole originated from VO (Figure 2d). Actually, due to the charge compensation, VO will be positively charged and the surrounding cations been negatively charged as revealed in NEXAFS. As a result, defect dipole moments will be established between these positive and negative charges. These defect dipoles can interfere with the intrinsic ferroelectric dipoles in BiFeO3, resulting in the distorted butterfly loop in Figure 2d.[ 41 , 42 ]
Figure 2.

The phase images of (a) BFO and (b) BFO‐VO photoelectrodes. The hysteresis loops of the PFM amplitude and phase for (c) BFO and (d) BFO‐VO photoelectrodes.
With the BiFeO3 photoelectrodes, we investigated the photoresponse under external poling treatment (Figure S6). The CST photocurrent density (jCST) is measured to reflect the CST driving force change in the samples. [43] Using the jCST at 1.23 V versus reversible hydrogen electrode (VRHE) as a benchmark, the jCST‐poling voltage (j‐P) curves are acquired as shown in Figure 3a. Compared to BFO, the jCST of BFO‐VO has improved from 0.20 mA cm−2 to 0.31 mA cm−2 due to the VO induced CST driving force (that is the VO effect). When the poling voltage continually changes from −120 V to 120 V, it is found that the BFO‐VO show a gradually decreased jCST. Especially, when BFO‐VO is poled at −120 V, the jCST will increase to 1.31 mA cm−2, five times larger than the original jCST of BFO. In contrast, when poled at +120 V, the jCST of BFO‐VO will reduce to 0.06 mA cm−2, which is only one third of the original jCST of BFO. Actually, due to the ferroelectric feature, the jCST of BFO will also be tuned by external poling as shown in Figure 3a. When comparing the BFO and BFO‐VO at same poling voltage, we can conclude that the VO effect is gradually decreased with the poling voltage varied from −120 V to 120 V. With the presence of VO, we can observe a more profound change upon external poling treatment. Taking the slope of j‐P as example, under negative poling, the BFO‐VO shows a slope of 8.5 μA cm−2 V−1, much steeper than that of BFO photoelectrode (4.04 μA cm−2 V−1). Using the carrier density (Nd) to reflect the VO concentration change, the relationship between the slope and charge carrier density in Figure 3b further shows the significant role of VO in poling treatment. Under positive poling, both samples experience a gradual photocurrent decay, but the VO makes the jCST decay faster.
Figure 3.

(a) The change of CST capability with the change of external poling voltage. (b) The change of j‐P slope for the negative poling with the change of carrier concentration. (c) The linear sweep voltammetry curves for water oxidation process on BFO‐VO photoelectrodes under −120 V poling, no poling and +120 V poling treatment.
The dependence of VO effect on poling treatment can be explained with the defect dipole as discovered in Figure 2. Due to the interference of defect dipole with the intrinsic dipole in BiFeO3, it will strengthen the polarization of the semiconductors, leading to enhanced depolarization electric field produced when the dipoles are aligned under external poling.[ 19 , 20 , 21 , 22 , 44 ] With negative poling, the depolarization electric field has the same direction as the built‐in electric field for the CST process. Therefore, the adding of defect dipole into the depolarization process will strengthen the CST process, which is reflected as the larger slope of BFO‐VO in j‐P curves. With the generation of more VO in BiFeO3, the defect dipole moment will be stronger, resulting in more significant VO effect and steeper j‐P slope at high VO concentration in Figure 3b. Under positive poling, since the depolarization electric field is against the built‐in electric field, the additional defect dipole will further weaken the CST driving force produced by the built‐in electric field. As a result, the presence of VO will reduce the photocurrent more significantly and eventually eliminate VO effect under positive poling.
Moreover, we have investigated the PEC water oxidation to show the influence of poling treatment on VO modified samples. As shown in Figure 3c, we can observe over 100 times improvement of water oxidation photocurrent density (at 1.23 VRHE) from 21 μA cm−2 to 261 μA cm−2 after −120 V poling, which represents as the top photoresponse based on BiFeO3 photoelectrodes (Table S1). While poling at +120 V has reduced the photocurrent density to nearly null response, together with significantly improved onset potentials.
Actually, all above PEC tests are carried out right after the poling treatment. But the polarization in the BiFeO3 is at a metastable status, and relaxation of the aligned dipoles will slowly occur when the poling electric field is perished. [45] To investigate this process, we monitored the jCST of the BiFeO3 photoelectrodes over a long period. For both BFO and BFO‐VO, they experienced a steady jCST photocurrent change. The negatively poled samples will gradually lose the photocurrent improvement, while positively poled ones slowly recover in a daily based monitoring process, until both samples achieve their original performance after 7 days (Figure 4). More interesting, the effect of polarization can be regained with another same poling stimulation and no significant photocurrent loss is observed with the three rounds poling‐relaxation processes. Furthermore, we can see the BFO‐VO has experienced a very similar recovery rate as the BFO (Figure S7), indicating that the relaxation of polarization process is similar in these two samples. That is reasonable considering that the polarization relaxation is a temperature dependent process, wherein BFO and BFO‐VO are both stored at room temperature. To further check the above assumption, the storage temperature of BFO‐VO has been reduced to −20 °C in a freezer. From Figure 4, we can observe that the photocurrent recovery process has been slowed down, confirming the temperature dependent relaxation process.
Figure 4.

The CST photocurrent density recovery process after negative poling (−120 V, square) and positive poling (+ 120 V, circle). The BFO (black), BFO‐VO (red), and −20 °C freezed BFO‐VO (blue) samples are monitored for comparison.
The above findings reveal that VO can strengthen the polarization in BiFeO3 photoelectrodes, leading to tunable band bending and CST process during PEC research. At negative poling, the VO induced defect dipoles can further strengthen the depolarized electric field established by the ferroelectric dipoles, thereby providing stronger band bending as schemed in Figure S8. Consequently, the VO will promote the CST much more significantly than the non‐polarized BFO samples. While at positive poling, the VO will deteriorate the CST process, leading to even weaker band bending and worse CST performance than the pristine BFO. In addition, the polarization will experience a relaxation process, leading to a dynamic role of the VO induced defect dipole.
Conclusions
In conclusion, we have investigated the influence of VO on the polarization of typical BiFeO3 photoelectrodes. The VO induced defect dipole was found to strengthen CST in ferroelectric BiFeO3 under negative poling, while positive poling led to reduced PEC performance. In addition, polarization experienced a slow relaxation process, leading to the decay of the polarization influence. Therefore, the role of VO in a polarized system is drastically determined by the poling history of the materials, including the poling direction, intensity and relaxation time and temperature. These findings enhance our understanding on the nature of VO and shed light on the proper application of VO in tuning optoelectrical behavior of a semiconductor photoelectrode.
Supporting Information
The experiment details and additional data and references are available within the Supporting Information.
Conflict of Interests
The authors declare no conflict of interest.
1.
Supporting information
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
Supporting Information
Acknowledgments
The authors would like to acknowledge the support by Australian Research Council through its DECRA (DE210100930, DE230100357), Discovery (DP200101900, DP230100462), Future Fellowship (FT230100251) and Laureate Fellowship (FL190100139) schemes. This work was performed in part at the Queensland node of the Australian National Fabrication Facility. A company established under the National Collaborative Research Infrastructure Strategy to provide nano and microfabrication facilities for Australia′s researchers. This research was undertaken on the Soft X‐ray Absorption Spectroscopy beamlines at the Australian Synchrotron, part of ANSTO. The authors acknowledge the facilities, and the scientific and technical assistance, of the Australian Microscopy and Microanalysis Research Facility at the Centre for Microscopy and Microanalysis, The University of Queensland. X. L., Y. B. and J. Y. acknowledges scholarship support from UQ Graduate School. Open Access publishing facilitated by The University of Queensland, as part of the Wiley ‐ The University of Queensland agreement via the Council of Australian University Librarians. Open Access publishing facilitated by The University of Queensland, as part of the Wiley ‐ The University of Queensland agreement via the Council of Australian University Librarians.
Li X., Wang Z., Bao Y., Lu H., You J., Wang L., ChemSusChem 2024, 17, e202400946. 10.1002/cssc.202400946
Contributor Information
Dr. Zhiliang Wang, Email: zhiliang.wang@uq.edu.au.
Prof. Lianzhou Wang, Email: l.wang@uq.edu.au.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Peng W., Mun J., Xie Q., Chen J., Wang L., Kim M., Noh T. W., npj Quantum Mater. 2021, 6, 48. [Google Scholar]
- 2. Li S., Zhao Z., Zhao J., Zhang Z., Li X., Zhang J., ACS Appl. Nano Mater. 2020, 3, 1063. [Google Scholar]
- 3. Wang Z., Zhang L., Schulli T. U., Bai Y., Monny S. A., Du A., Wang L., Angew. Chem. Int. Ed. 2019, 58, 17604. [DOI] [PubMed] [Google Scholar]
- 4. Li J., Guo Z., Li F., Lin H., Li C., Xiang S., Zhou T., Wan N., Liu Y., Superlattices Microstruct. 2015, 85, 454. [Google Scholar]
- 5. Maruska H. P., Rhines W. C., Solid-State Electron. 2015, 111, 32. [Google Scholar]
- 6. Quinn J. T. E., Zhu J., Li X., Wang J., Li Y., J. Mater. Chem. C 2017, 5, 8654. [Google Scholar]
- 7. Some S., Kim J., Lee K., Kulkarni A., Yoon Y., Lee S., Kim T., Lee H., Adv. Mater. 2012, 24, 5481. [DOI] [PubMed] [Google Scholar]
- 8. Li W., Shi J., Zhang K. H. L., MacManus-Driscoll J. L., Mater. Horiz. 2020, 7, 2832. [Google Scholar]
- 9. Wang Z., Wang L., EcoMat 2021, 3, 12075. [Google Scholar]
- 10. Wang S., Chen P., Bai Y., Yun J. H., Liu G., Wang L., Adv. Mater. 2018, 30, e1800486. [DOI] [PubMed] [Google Scholar]
- 11. Nishioka S., Hyodo J., Vequizo J. J. M., Yamashita S., Kumagai H., Kimoto K., Yamakata A., Yamazaki Y., Maeda K., ACS Catal. 2018, 8, 7190–7200. [Google Scholar]
- 12. Zhang Y., Bu Y., Jiang F., Li H., Chen X., Ao J.-P., Electrochimica. Acta. 2021, 366, 137288. [Google Scholar]
- 13. Wang S., Chen P., Yun J. H., Hu Y., Wang L., Angew. Chem. Int. Ed. 2017, 56, 8500. [DOI] [PubMed] [Google Scholar]
- 14. Wang J., Waters J. L., Kung P., Kim S. M., Kelly J. T., McNamara L. E., Hammer N. I., Pemberton B. C., Schmehl R. H., Gupta A., Pan S., ACS Appl. Mater. Interfaces 2017, 9, 381. [DOI] [PubMed] [Google Scholar]
- 15. Zhou T., Li L., Li J., Wang J., Bai J., Xia L., Xu Q., Zhou B., Chem. Eng. J. 2021, 425, 131770. [Google Scholar]
- 16. Li X., Liu Q., Deng F., Huang J., Han L., He C., Chen Z., Luo Y., Zhu Y., Appl. Catal. B 2022, 314, 121502. [Google Scholar]
- 17. Li Z., Thong H. C., Zhang Y. F., Xu Z., Zhou Z., Liu Y. X., Cheng Y. Y. S., Wang S. H., Zhao C., Chen F., Bi K., Han B., Wang K., Adv. Funct. Mater. 2020, 31, 2005012. [Google Scholar]
- 18. Gao P., Nelson C. T., Jokisaari J. R., Baek S. H., Bark C. W., Zhang Y., Wang E., Schlom D. G., Eom C. B., Pan X., Nat. Commun. 2011, 2, 591. [DOI] [PubMed] [Google Scholar]
- 19. Chen X., Xu Y., Ma X., Zhu Y., Natl. Sci. Rev. 2020, 7, 652. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Hu Y., Pan Y., Wang Z., Lin T., Gao Y., Luo B., Hu H., Fan F., Liu G., Wang L., Nat. Commun. 2020, 11, 2129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Zhen C., Ren Z., Kang Y., Wang L., Liu G., Acc. Mater. Res. 2023, 4 (7), 591. [Google Scholar]
- 22. Huang J., Kang Y., Liu J., Yao T., Qiu J., Du P., Huang B., Hu W., Liang Y., Xie T., Chen C., Yin L., Wang L., Cheng H. M., Liu G., Nat. Commun. 2023, 14, 7948. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Li X., Wang Z., Ji W., Lu T., You J., Wang K., Liu G., Liu Y., Wang L., ACS Nano 2023, 17, 22944. [DOI] [PubMed] [Google Scholar]
- 24. Islamov D. R., Zalyalov T. M., Orlov O. M., Gritsenko V. A., Krasnikov G. Y., Appl. Phys. Letter. 2020, 117, 162901. [Google Scholar]
- 25. Song B., Wu S., Yan H., Zhu K., Xu L., Shen B., Zhai J., J. Mater. Sci.Technol. 2021, 77, 178. [Google Scholar]
- 26. Kubel F., Schmid H., J. Cryst. Growth 1993, 129, 515. [Google Scholar]
- 27. Dhole S., Chen A., Nie W., Park B., Jia Q., Nanomater. 2022, 12, 835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Lee D., Jeon B. C., Baek S. H., Yang S. M., Shin Y. J., Kim T. H., Kim Y. S., Yoon J. G., Eom C. B., Noh T. W., Adv. Mater. 2012, 24, 6490. [DOI] [PubMed] [Google Scholar]
- 29. Wang Z., Mao X., Chen P., Xiao M., Monny S. A., Wang S., Konarova M., Du A., Wang L., Angew. Chem. Int. Ed. 2019, 58, 1030. [DOI] [PubMed] [Google Scholar]
- 30. Moos R., Menesklou W., Härdtl K. H., Appl. Phys. A: Mater. Sci. Process. 1995, 61, 389–395. [Google Scholar]
- 31. Ralf M., Heinz H. Karl, J. Am. Ceram. Soc. 1997, 80, 25492562. [Google Scholar]
- 32. Chi X., Mandal L., Liu C., Fauzi A. D., Chaudhuri A., Whitcher T. J., Jani H. K., Chen Z., Xi S., Diao C., Naradipa M. A., Yu X., Yang P., Castro-Neto A. H., Breese M. B. H., Loh K. P., Venkatesan T. V., Rusydi A., NPG Asia Mater. 2022, 14, 19. [Google Scholar]
- 33. Huang Z. F., Xi S., Song J., Dou S., Li X., Du Y., Diao C., Xu Z. J., Wang X., Nat. Commun. 2021, 12, 3992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Gimenez-Marques M., Bellido E., Berthelot T., Simon-Yarza T., Hidalgo T., Simon-Vazquez R., Gonzalez-Fernandez A., Avila J., Asensio M. C., Gref R., Couvreur P., Serre C., Horcajada P., Small 2018, 14, e1801900. [DOI] [PubMed] [Google Scholar]
- 35. Kumari P., Zzaman M., Jena S., Kumar M., Bharadwaj R. R., Verma V. K., Shahid R., Amemiya K., Singh V. R., J. Supercond. Nov. Magn. 2021, 34, 1119. [Google Scholar]
- 36. Nomura K., Catal. 2019, 9, 1016. [Google Scholar]
- 37. Weymann C., Lichtensteiger C., Fernandez-Peña S., Naden A. B., Dedon L. R., Martin L. W., Triscone J. M., Paruch P., Adv. Electron. Mater. 2020, 6, 2000852. [Google Scholar]
- 38. Yang S. M., Yoon J.-G., Noh T. W., Curr. Appl. Phys. 2011, 11, 1111. [Google Scholar]
- 39. Gottschalk S., Hahn H., Flege S., Balogh A. G., J. Appl. Phys. 2008, 104, 114106. [Google Scholar]
- 40. Liu Q., Zhou Y., You L., Wang J., Shen M., Fang L., Appl. Phys. Letter. 2016, 108, 022902. [Google Scholar]
- 41. Liu L., Xu K., Li Q., Daniels J., Zhou H., Li J., Zhu J., Seidel J., Li J. F., Adv. Funct. Mater. 2020, 31, 2005876. [Google Scholar]
- 42. Kitanaka Y., Noguchi Y., Miyayama M., Phys. Rev. B 2010, 81, 094114. [Google Scholar]
- 43. Wang Z., Gu Y., Zheng L., Hou J., Zheng H., Sun S., Wang L., Adv. Mater. 2022, 34, 2106776. [DOI] [PubMed] [Google Scholar]
- 44. Liu Y., Ye S., Xie H., Zhu J., Shi Q., Ta N., Chen R., Gao Y., An H., Nie W., Jing H., Fan F., Li C., Adv. Mater. 2020, 32, 1906513. [DOI] [PubMed] [Google Scholar]
- 45. Assavachin S., Osterloh F. E., J. Am. Chem. Soc. 2023, 145, 18825. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
Supporting Information
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.

