76 77 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 1 / 2 2 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 1 / 2 2 Structural insight into the photocatalytic water-splitting activity of N-TiO 2 Related publication: Foo C., Li Y., Lebedev K., Chen T., Day S.J., Tang C., Tsang, S. C. E. Characterization of oxygen defects and nitrogen impurities in TiO 2 photocatalysts using variable-temperature X-ray powder diffraction, Nat. Commun. 12 , 661 (2021). DOI: 10.1038/s41467-021- 20977-z Publication keywords: Water splitting; Photocatalysis; Oxygen vacancies; X-ray diffraction T iO 2 -based powder materials have been widely studied as efficient photocatalysts for water splitting due to their low cost, photo- responsivity, abundance, and chemical and thermal stability. Nitrogen-doping TiO 2 enhances the presence of structural defects and dopant impurities at elevated temperatures and results in amaterial with impressive visible-light absorption for photocatalytic activity. Although the electronic and optical properties of these materials have been extensively studied, the structure-activity relationship and photocatalytic mechanism remain ambiguous. Using the High-Resolution Powder Diffraction beamline (I11), researchers at the University of Oxford have detailed the structure of nitrogen- doped TiO 2 . The high collimation and angle-rejection on I11 can lead to very high real-space resolution. The ease of variable-temperature environments and the reasonable exposure time on I11were also paramount for this investigation. They found that an unusual anisotropic thermal expansion of the anatase phase can reveal the intimate relationship between sub-surface oxygen vacancies, nitrogen-doping level and photocatalytic activity. They also identified a new cubic titanium oxynitride phase for highly dopedanatase,whichprovides important informationon the fundamental shift inabsorptionwavelength, leading toexcellent photocatalysis using visible light. These results show that visible light can drive efficient photocatalytic water-splitting over nitrogen-doped TiO 2 , yielding plentiful hydrogen gas at elevated temperatures. Crucially, the absorptionprofile is shifted to longerwavelengths relative to undopedTiO 2 , whichusually absorbs primarily in the UV region. Strong absorption of visible light enables more complete utilisation of the solar spectrum. TiO 2 -based powder materials have been widely studied as efficient photocatalysts for water splitting due to their low cost, photo-responsivity, earthly abundance, chemical and thermal stability, etc. In particular, the recent breakthrough of nitrogen-doped TiO 2 , which enhances the presence of structural defects and dopant impurities at elevated temperatures, exhibits an impressive visible-light absorption for photocatalytic activity. Although their electronic and optical properties have been extensively studied, the structure-activity relationship and photocatalytic mechanism still remain ambiguous. Previous research within the Tsang group at the University of Oxford has shown that nitrogen-doped TiO 2 can be an exemplary photocatalyst for water- splitting, and hence the production of hydrogen. This report details the structural characterisation, primarily by X-ray powder diffraction. Data was collected at the High-Resolution Powder Diffraction beamline (I11) at Diamond Light Source, as well as supplementary collections at BL02B2 at SPring-8, Japan. The authors report an in-depth structural study of rutile, anatase and mixed phases (commercial P25) with and without nitrogen-doping by variable-temperature synchrotron X-ray powder diffraction (SXPD). An unusual anisotropic thermal expansion of the anatase phase can reveal the intimate relationship between sub-surface oxygen vacancies, nitrogen-doping level and photocatalytic activity. For highly doped anatase, a new cubic titaniumoxynitride phase is also identified which provides important information on the fundamental shift in absorption wavelength, leading to excellent photocatalysis using visible light. Efficient photocatalysts must utilise the highly abundant visible part of the solar spectral-irradiance. In the publication, the authors have shown that visible light can drive efficient photocatalytic water-splitting over nitrogen-doped TiO 2 , yielding plentiful hydrogen gas at elevated temperature. Crucially, the absorption profile is shifted to longer wavelengths relative to pristine TiO 2 , which usually absorbs primarily in the UV region. Strong absorption of visible light enables more complete utilisation of the solar spectrum. In additional work, It has also been shown that the temperature requirements of the reaction can also be provided by solar energy through the concentration of light in a light furnacewith no electrical heating 1 . The photoactivity of TiO 2 is well-known to be linked with the oxygen vacancy concentration at a given temperature. The formation of an oxygen vacancy as ‘hole’ is accompanied by the release of two electrons. Usually these are thought to reside at Ti 4+ sites forming Ti 3+ (electron). In N-doped TiO 2 , these high energy electrons can also be localised at nitrogen impurity sites, which significantly decreases the energy of formation for oxygen vacancies. The high concentration leads to linearly increasing rate of photocatalytic hydrogen evolution, as shown by EPR in Fig. 1. However, structurally there is a lack of information about the location of the nitrogen impurity, and the coordination effects of the absence of oxygen and the presence of undercoordinated titanium. The insight into the system lies in the unusual thermal expansion of TiO 2 and N-TiO 2 . The high resolution of the I11 instrument, leads to very high real- space resolution. This, coupled with the ease of use of variable-temperature environments and fast exposure times was also paramount to the success execution of this investigation. Though known to be non-linear, this degree of correlated anisotropic thermal expansion is a novel observation. There is a significant change in the c-axis lattice expansion in anatase under elevated temperature, resulting in a unique non-linear expansion due to the Jahn-Teller effect (Fig. 2). However, as surface oxygen vacancies are formed at temperatures above 200°C, a concomitant decrease in the c-axis expansion in lattice is observed, which have been shown to be correlated to photocatalytic activity. The authors demonstrate oxygen-vacancy mobility from the disordered surface to formordered sub-surface vacancies, as resolved by SXPD. The technique indicates that the effects of the large quantities of oxygen vacancies and Ti 3+ , which facilitate photocatalysis in anatase, extends beyond the surface trilayer and into the bulk material, particularly in comparison to the inactive rutile polymorph which experiences no anisotropy. The inclusion of nitrogen can clearly increase the quantity of sub-surface oxygen vacancies and stabilise the anatase phase as reflected by the different degrees of observed unit-cell distortion. By doping high levels of nitrogen into anatase, a new cubic titanium oxynitride phase is identified. This phase was markedly different to standard titanium nitride in its (1) lattice parameter, (2) thermal stability, and (3) excess electron density. Firstly, while lattice parameters for the anatase component of the powder agreed with the known values, the value refined for TiNO x is much lower than expected (exp. 4.1843Å vs. lit. 4.235 Å). Secondly, the phase is only stable to 240°C, after which the reflections are no longer observed at 520°C (Fig. 3). This is a remarkably lowthermal stability compared to non-oxygen-doped titanium nitride which only reaches full degradation at 1,000°C 2 .Thirdly, the real-space Fourier difference maps for the Rietveld phase showed excess electron density at the tetrahedral hole of the framework. This phase identification provides an additional component in the effort to achieve a more complete fundamental understanding of the bandgap modification in N– TiO 2 materials and will informwider hypotheses and perspectives on this system. In summary, the authors present important links for the first time between structure, sub-surface oxygen vacancies, nitrogen-doping, and photocatalytic activity of anatase catalysts at various temperatures. Ultimately, the authors report the influence of well-characterised structural modifications on electronic phenomena in this photocatalytic hydrogen-evolution powder catalyst. References: 1. Li, Y. et al. Photocatalytic water splitting by N-TiO2 on MgO (111) with exceptional quantum efficiencies at elevated temperatures. Nature Communications , 10 , 4421 (2019). DOI: 10.1038/s41467-019-12385-1 2. Hou, X.-M. et al. Kinetics of thermal oxidation of titaniumnitride powder at different oxidizing atmospheres. Journal of the American Ceramic Society 94 , 570–575 (2011). DOI: 10.1111/j.1551-2916.2010.04084.x Funding acknowledgement: This work is supported by the EPSRC (1947428) and Diamond Light Source, both of whom are thanked for the joint DPhil Studentship undertaken by C. Foo. Corresponding author: Professor Shik Chi Edman Tsang, University of Oxford,
[email protected] Crystallography Group Beamline I11 R0 R1 R2 R3 A0 A1 A2 A3 a b c Intensity / arb. units H 2 evolution rate / μmol g -1 h -1 g factor g factor V o concentration / (x10 14 ) g -1 Intensity / arb. units Figure 1. EPR spectra for (a) rutile and (b) anatase materials; (c) Hydrogen evolution of anatase catalysts under visible light against EPR-derived oxygen vacancy concentration. Error bars indicate the standard deviation. 3.784 3.786 3.788 3.790 3.792 3.794 3.796 9.50 9.51 9.52 9.53 9.54 9.55 9.56 0 100 200 300 400 500 4.590 4.595 4.600 4.605 4.610 4.615 0 100 200 300 400 500 2.955 2.960 2.965 2.970 2.975 0 50 100 150 200 250 300 350 400 450 2 4 6 8 10 12 14 16 18 Rutile, c axis Rutile, a axis Anatase, c axis Lattice parameter / Å Anatase, a axis Lattice parameter / Å Temperature / °C Temperature / °C Linear thermal expansion ( x ) / (10 -14 ) °C -1 Temperature / °C P0 a P0 c P3 a P3 c a b Figure 2. (a) Lattice parameters of anatase and rutile phases for P0 and P3 from room temperature to 500 °C; (b) Linear thermal expansion coefficient α for the a-axis and c-axis of the anatase phase in P0 and P3. Error bars indicate the standard deviation. 11.0 11.1 11.2 11.3 11.4 16.0 16.2 16.4 16.6 16.8 17.0 18.4 18.6 18.8 19.0 19.2 19.4 0 100 200 300 400 500 600 100 200 300 400 500 10 12 14 16 18 20 22 24 26 28 30 Temperature / °C 2 θ / ° e d c Ana (101) Intensity / arb. units 2 θ / ° 25°C 40°C 60°C 80°C 100°C 120°C 140°C 160°C 180°C 200°C 220°C 240°C 260°C 280°C 300°C 320°C 340°C 360°C 380°C 400°C 420°C 440°C 460°C 480°C 500°C 520°C 540°C 560°C 580°C b a Ana (103) Ana (112) Ana (004) TiNO x (111) Intensity / arb. units 2 θ / ° Intensity / arb. units 2 θ / ° TiNO x (200) Intensity / arb. units Temperature / °C Ana (101) TiNO (200) x Figure 3. (a) In-situ VT-SXPD patterns; (b–d) enlarged regions showing the decomposition of TiNO x and increase in anatase intensity; (e) intensity of anatase (101) and TiN (200) plotted against temperature (negligible standard deviation not plotted).