Spectroscopic Evidence of Intraband Gap States in α‐SnWO4 Photoanodes Introduced by Interface Oxidation

α‐SnWO4 is an emerging photoelectrode material for photoelectrochemical water splitting, with several promising properties such as the favorable bandgap of 1.9 eV and suitable positions of the valence and conduction band. However, a major challenge remains: unprotected α‐SnWO4 undergoes surface passivation that blocks further charge transfer, and α‐SnWO4 electrodes that are covered with a protection/catalytic overlayer (e.g., NiOx, CoOx) show limited photovoltage. Earlier studies reveal that interfacial oxidation occurs due to the deposition of the overlayer. This negatively impacts the photovoltage that can be extracted, which is attributed to Fermi‐level pinning at the interface. The exact origin of this Fermi‐level pinning mechanism, however, remains unclear. In the present study, a combination of surface photovoltage analysis and hard X‐ray photoelectron spectroscopy is used to elucidate the electronic structure of the α‐SnWO4/oxide interface. Both techniques offer compelling and consistent evidence for the presence of a defect state that is energetically located within the bandgap energy of α‐SnWO4 and is likely responsible for the Fermi‐level pinning.

DOI: 10.1002/solr.202201104 α-SnWO 4 is an emerging photoelectrode material for photoelectrochemical water splitting, with several promising properties such as the favorable bandgap of 1.9 eV and suitable positions of the valence and conduction band. However, a major challenge remains: unprotected α-SnWO 4 undergoes surface passivation that blocks further charge transfer, and α-SnWO 4 electrodes that are covered with a protection/catalytic overlayer (e.g., NiO x , CoO x ) show limited photovoltage. Earlier studies reveal that interfacial oxidation occurs due to the deposition of the overlayer. This negatively impacts the photovoltage that can be extracted, which is attributed to Fermi-level pinning at the interface. The exact origin of this Fermi-level pinning mechanism, however, remains unclear. In the present study, a combination of surface photovoltage analysis and hard X-ray photoelectron spectroscopy is used to elucidate the electronic structure of the α-SnWO 4 /oxide interface. Both techniques offer compelling and consistent evidence for the presence of a defect state that is energetically located within the bandgap energy of α-SnWO 4 and is likely responsible for the Fermi-level pinning.
properties. Grain boundaries have been reported to limit charge transport, [10] and the transport properties are reported to be highly anisotropic. [18,25] Epitaxial or highly oriented films are therefore expected to have much improved charge transport, and indeed recent reports on two-dimensional α-SnWO 4 crystalline nanosheets with preferred {001} orientation demonstrated superior performance. [18][19][20] Another limitation is the modest photovoltage that can be extracted from the material, despite the fact that the band positions straddle the water reduction and oxidation potentials (which would enable a maximum quasi-Fermi level splitting of at least 1.23 V under operating conditions). In our earlier study on NiO x -coated α-SnWO 4 , in which the NiO x serves as the protection layer, we showed that the limitation of the photovoltage can be correlated with the formation of an interfacial oxide layer at the interface of α-SnWO 4 and NiO x . [12] A similar observation of limited photovoltage has been reported for CoO x -coated α-SnWO 4 . [21] The interfacial layer contains Sn 4þ , most likely in the form of SnO 2 , which adversely affects the charge transfer at the interface. Fermi-level pinning due to the presence of intraband states was suggested as the possible underlying mechanism of the limitation in photovoltage.
In this study, we utilize spectroscopic techniques to further unravel the presence of such intraband states. Surface photovoltage (SPV) analysis reveals the wavelength-dependent photovoltage response of α-SnWO 4 and synchrotron-based hard X-ray photoelectron spectroscopy (HAXPES) measurements directly probe the density of states of energy levels around the valence band maximum of α-SnWO 4 . By analyzing the impact of depositing NiO x and SnO 2 on the surface of our pulsed laser deposited α-SnWO 4 films on the SPV and HAXPES data, we will show evidence that the previously proposed intraband states at the interface are indeed present and likely causing the photovoltage limitation.

Results and Discussion
First, SPV measurements (modulated illumination at 8 Hz) were performed on a pristine α-SnWO 4 film and on a film coated with a 20 nm-thick NiO x layer. The structural, optoelectronic, and photoelectrochemical characterization of these films has been discussed in our previous reports. [10][11][12][13] Figure 1a shows the overall photovoltage amplitude R as a function of the photon energy for both films. Three qualitatively different ranges can be distinguished in the SPV spectra. At higher photon energies, i.e., above the bandgap of α-SnWO 4 (E g , 1.9 eV [11] ), the SPV signals are related to fundamental absorption, i.e., band-to-band excitation. In the range between about 1.6 and 1.85 eV, the SPV amplitude increases exponentially due to local potential fluctuations, which can be caused by disorders or local variations of the bandgap. The corresponding energies of the exponential tails (E t ) are about 65 and 70 meV for pristine and α-SnWO 4 films coated with 20 nm NiO x , respectively. This means that local potential fluctuations increase near the surface of α-SnWO 4 after the deposition of NiO x . At photon energies below %1.45 eV, the SPV amplitude increases to a maximum at about 0.85 eV and decreases to a minimum at about 1.45 eV. This means that photogeneration from deep defect states in the bandgap of α-SnWO 4 could lead, depending on the process of charge separation, to an increase or decrease of modulated charge separation.
After the deposition of NiO x , the SPV amplitude increases at photon energies above %0.75 eV but tends to decrease at lower photon energies. This suggests that modulated charge separation increases for photogeneration from deep defect states above about 0.75 eV as well as for excitation from tail states and for fundamental absorption. In contrast, the role of transitions leading to a decrease of modulated charge separation decreases after the deposition of NiO x . The reduction of the SPV amplitude in a certain range below the bandgap has to be related to the reduction of the role of a certain kind of trap states at/near the surface. The separate analysis of the in-phase (X ) and phase-shifted by 90°(Y ) signals, from which the SPV amplitudes (R) are obtained (Equation (1)), allows us to distinguish more clearly the role of deep defect states for modulated charge separation In the low photon energy regime (below the bandgap of α-SnWO 4 ), the nonzero signal indicates the presence of surface or bulk defect states. In the high photon energy regime (above the bandgap), the amplitude increases after NiO x deposition. Representation of the in-phase (blue) and phase-shifted by 90°(red) signals for b) the pristine α-SnWO 4 sample and c) the sample coated with 20 nm NiO x . The deposition of 20 nm NiO x clearly modifies the in-phase and phase shifted by 90°signals at sub-bandgap energies.
www.advancedsciencenews.com www.solar-rrl.com Here, it is important to mention that the Xand Y-signals are related to the fast and slow response with respect to the modulation period and that the relation between the signs of both gives information about the preferential direction of charge separation with regard to trapping. [26] Figure 1b,c depicts the spectra of the Xand Y-signals of α-SnWO 4 before and after deposition of NiO x , respectively. In the two regions above about 1.5 eV, all in-phase signals are positive and all phase shifted by 90°signals are negative. This is typical for charge separation of holes toward the surface and of electrons toward the bulk. [26] In the region below about 1.5 eV, the X-signals for the pristine film (i.e., before deposition of NiO x ) are practically zero whereas the Y-signals are positive. This means that electrons are partially trapped at surface states and that the relaxation time of detrapping is long in relation to the modulation period. In addition, above about 0.8 eV, the Y-signals decrease and change toward negative signals (change of sign at about 1.4 eV). This behavior can be interpreted as a reduction of electrons trapped at surface states due to photogeneration from defect states leading to the separation of holes toward the surface where they can recombine with electrons.
After the deposition of NiO x , the behavior of the Xand Ysignals changes completely at photon energies below about 1.5 eV. First, positive X-signals appear and, second, the sign of the Y-signals changes to negative. This means that photogeneration from defect states leading to the separation of holes towards the surface dominates over photogeneration from defect states leading to the trapping of electrons at surface states and that the relaxation time of separated charge carriers decreased. Therefore, the deposition of NiO x leads to a reduction of electron trapping at surface states and therefore to an increase of the overall SPV signal. This is attributed to increased band bending and enhanced charge separation, which agrees with our previous HAXPES measurements that demonstrated additional %0.4 eV band bending upon the deposition of NiO x on α-SnWO 4 . [12] Incidentally, photogeneration from defect states leading to preferential modulated separation of holes toward the surface may be attributed to sub-bandgap absorption via surface or bulk defect states. [27] Overall, the SPV analyses suggest that a substantial modification of the charge generation, spatial redistribution, and relaxation in α-SnWO 4 films is introduced upon the deposition of the NiO x layer. This is indicative of the creation and/or modification of defect states caused by the NiO x deposition. In our previous studies, we showed that the deposition of NiO x introduces the formation of Sn 4þ species, most likely in the form of an SnO 2 layer, at the interface of α-SnWO 4 and NiO x . [12] We therefore hypothesize that the origin of the sub-bandgap defect states or their modification observed in the SPV data is caused by the presence of this interfacial SnO 2 layer.
To test the earlier hypothesis, we performed another HAXPES investigation on two different samples: pristine α-SnWO 4 and SnO 2 -coated α-SnWO 4 . The SnO 2 layer was intentionally grown by photoelectrochemical oxidation of α-SnWO 4 at an applied potential of 1.23 V versus RHE in a 0.5 M KP i buffer electrolyte (pH 7) with added 0.5 M Na 2 SO 3 . We have previously established that the integrated charge transfer from the initial current spike during this process can be assumed to fully contribute to the oxidation of SnWO 4 . [11,12] Based on this, an SnO 2 layer thickness of %1 nm was estimated. By comparing the core level and valence band spectra of the two samples measured using hard X-rays, the chemical nature of the α-SnWO 4 /SnO 2 interface could be studied. Figure 2 shows the W 4f and Sn 3d core-level spectra of the pristine α-SnWO 4 sample (black) and the SnO 2 -coated α-SnWO 4 sample (blue). Different photon energies (2210, 3880, and 6009 eV) were used, as indicated in the figure, to obtain depth-dependent information. The peak position of the nominal Figure 2. a) W 4f spectra of pristine α-SnWO 4 (black) and α-SnWO 4 /SnO 2 (blue) samples measured by HAXPES. The varying excitation energies are indicated. The SnO 2 layer was formed by photoelectrochemical oxidation in 0.5 M KP i þ 0.5 M Na 2 SO 3 (pH 7) using an applied potential of 1.23 V versus RHE, and the thickness is %1 nm. b) Sn 3d core level spectra measured by HAXPES. The oxidation of Sn 2þ to Sn 4þ can be observed by the peak asymmetry toward higher binding energy. Dashed vertical lines in (a) and (b) indicate the peak positions of W 5þ , W 6þ , Sn 2þ , and Sn 4þ .
www.advancedsciencenews.com www.solar-rrl.com oxidation state of W in α-SnWO 4 (W 6þ ) is indicated in Figure 2a by the dashed vertical black line. In addition, a small W 5þ contribution is observed and indicated by the dashed vertical magenta line. This observation of W 5þ agrees with our earlier work where it was discussed in more detail. [12] A comparison of the W 4f spectrum of the pristine sample and the SnO 2 -coated samples reveals a spectral broadening (see Figure S1, Supporting Information for the quantitative full-width at half-maximum, FWHM, of the W 4f 5/2 peak). This increase in FWHM can be attributed to an additional W 6þ contribution not related to the α-SnWO 4 phase. For example, the formation of WO 3 , which was suggested earlier as a possible degradation product, [11] may be the origin of this additional W 6þ species. It should be noted, however, that changes in band bending can also contribute to the broadening of the peak. Note that the earlier FWHM analysis was done for W 4f 5/2 peak, since the contribution from W 5þ overlaps with the lower binding energy peak of W 4f 7/2 , but not the W 4f 5/2 peak. The Sn 3d core-level spectra shown in Figure 2b show a clear peak asymmetry toward higher binding energies after the introduction of SnO 2 , due to spectral contribution from Sn 4þ . When varying the probing depth, we also observe a variation in the Sn 4þ contribution. For larger probing depths (i.e., higher photon energy), the Sn 4þ contribution is lower, which indicates that Sn remains in the 2þ oxidation state in the underlying α-SnWO 4 film. HAXPES measurements were also performed on an α-SnWO 4 sample coated with a thicker layer of SnO 2 (%12.5 nm), and the same trends are revealed (see Figure S1 and S2, Supporting Information).
A quantitative analysis of the relative contributions of the different species and their binding energies is reported in Figure 3. As described in the Experimental Section, the analysis was performed by a peak-fitting procedure, and an example of the fitted spectra for the pristine α-SnWO 4 film can be found in Figure S3, Supporting Information. The relative W 5þ contribution is shown in Figure 3a as a function of the photon energy for α-SnWO 4 films coated with SnO 2 layers of two different thicknesses: %1 and %12.5 nm. The contribution of W 5þ in both samples increases with increasing photon energy, and the overall W 5þ contribution is higher for the sample with the 1 nm SnO 2 layer. These observations suggest that the W 5þ contribution originates from the bulk of the α-SnWO 4 film, which agrees with our earlier discussion. [12] In contrast, a decrease of the Sn 4þ contribution with increasing photon energies can be seen in Figure 3b. This is expected since the intentionally grown SnO 2 layer, which contains the Sn 4þ species, is present on the surface. The overall higher contribution of Sn 4þ and the steeper decrease with increasing photon energy observed for the sample with a 12.5 nm-thick SnO 2 layer are also consistent with the explanation above. For comparison, the W 5þ and Sn 4þ contributions in the Figure 3. Analysis of the core-level spectra for the α-SnWO 4 þ 1 nm SnO 2 (blue data points) and α-SnWO 4 þ 12.5 nm SnO 2 (red data points) samples (the spectra are shown in Figure 2 and Figure S2, Supporting Information, respectively). a) Relative W 5þ contribution extracted from the W 4f 7/2 peak. b) Relative Sn 4þ contribution extracted from the Sn 3d 5/2 peak. c) Binding energies of the W 6þ species (W 4f 7/2 peak). d) Binding energies of the Sn 2þ species (Sn 3d 5/2 peak). The respective binding energies of the pristine sample are also indicated in (c) and (d) by the black data point.
www.advancedsciencenews.com www.solar-rrl.com pristine α-SnWO 4 sample are estimated to be 10 AE 3% and 19 AE 4%, respectively, from the spectra at the photon energy of 3880 eV. Figure 3c,d plots the binding energies of W 6þ and Sn 2þ (nominal oxidation states of W and Sn in α-SnWO 4 ) as a function of the photon energy. The binding energies remain rather constant with increasing photon energy, and a small increase of the binding energies can be observed for the SnO 2 -coated samples in comparison to the pristine film. While this may be attributed to changes in band bending related to the presence of the SnO 2 layer, the distribution of values falls within the experimental error range of AE0.1 eV. Thus, we conclude that no major changes are observed for the W 6þ and Sn 2þ binding energies upon the formation of SnO 2 .
Valence band measurements were then performed on the pristine α-SnWO 4 and α-SnWO 4 /SnO 2 samples. The valence band spectra recorded with a photon energy of 3933 eV are shown in Figure 4a. Note that the spectra are reported with the intensity axis on a logarithmic scale and are normalized to the peak maximum centered at %2.8 eV. Evaluation of the valence band maximum (VBM) shows that the VBM of α-SnWO 4 appears as expected at 1.5 AE 0.1 eV (see Figure S4, Supporting Information). This value agrees well with the earlier report of 1.6 AE 0.1 eV determined by ultraviolet photoelectron spectroscopy (UPS). [11] When comparing the pristine sample with the α-SnWO 4 /SnO 2 samples, two differences are noted. First, an increase in the intensity at higher binding energies (>4 eV) is observed for the α-SnWO 4 /SnO 2 samples. The intensity increases with the thickness of SnO 2 , which suggests that this change may originate from the valence band spectra of the SnO 2 layer. Indeed, SnO 2 is a large bandgap semiconductor with a reported VBM at >3 eV binding energy. [28][29][30] Second, an additional feature emerges at the binding energy of %0.5-1.5 eV (peak at %0.7 eV) for the α-SnWO 4 /SnO 2 samples. This energy range is within the bandgap of α-SnWO 4 and is therefore assigned to an intraband defect state. The intensity of this feature remains the same for the samples with 1 and 12.5 nm-thick layers of SnO 2 , thereby suggesting that its origin is not related to the SnO 2 layer itself, but it most likely arises from the α-SnWO 4 /SnO 2 interface. Based on the valence band spectra, a schematic band diagram of the α-SnWO 4 /SnO 2 interface region is constructed and shown in Figure 4b. The approximate locations of the energy levels and the defect state in the bandgap of α-SnWO 4 are indicated.
We can now combine the findings from the HAXPES investigation with the SPV data. The binding energy range of the intraband defect state observed in the HAXPES VB measurements agrees very well with the sub-bandgap signals observed in the SPV data. In addition, the relatively low intensity of the subbandgap signals in the SPV data is also consistent with the hypothesis that they originate from an intraband defect state at the interface. We note that a direct correlation cannot be made, since the SPV measurements were performed using α-SnWO 4 / NiO x samples, while α-SnWO 4 /SnO 2 samples were used for the HAXPES measurements. Unfortunately, HAXPES valence band measurements directly on the α-SnWO 4 /NiO x samples are unable to reveal the intraband defect state since its energy overlaps directly with the VBM of NiO x (see Figure S5, Supporting Information). Note that the density of states in the VBM of NiO x is much higher, and therefore masks that of the intraband defect state. However, since it was already established that an interfacial layer (SnO 2 ) forms between α-SnWO 4 and NiO x , [12] and SPV measurement on a α-SnWO 4 /SnO 2 sample also shows sub-bandgap signals ( Figure S6, Supporting Information), we postulate that the sub-bandgap signals observed in the SPV data of α-SnWO 4 /NiO x originate from the same intraband defect state that we observed in the HAXPES measurements of α-SnWO 4 /SnO 2 . Indeed, the defect structures of SnO 2 and SnO x have been investigated in several studies, [31][32][33][34][35][36] and the origin of the observed intraband states has been correlated with the presence of oxygen vacancies. Thus, oxygen vacancies in the SnO 2 layer are likely to be responsible for the intraband defect states observed in our study. Thus, we attribute the reduction in photovoltage observed in earlier studies, [11,12] which manifests itself in the form of Fermi-level pinning (see Figure S7, Supporting Information for band diagram illustration), to the defective Figure 4. a) HAXPES valence band spectra of the pristine α-SnWO 4 , α-SnWO 4 þ 1 nm SnO 2 , and α-SnWO 4 þ 12.5 nm SnO 2 samples. The spectra were normalized to the peak maximum at %2.8 eV. After formation of SnO 2 , an indication for an additional state above the valence band maximum (VBM) can be observed. b) A schematic band diagram is depicted based on the valence band spectra in (a) (rotated here). The approximate location of the state E D within the bandgap E g of α-SnWO 4 is shown, along with other energy levels: E F = Fermi level, E C = conduction band minimum, and E V = valence band maximum. The position of E V was determined from the nonlogarithmic plot, as shown in Figure S4, Supporting Information.
www.advancedsciencenews.com www.solar-rrl.com nature of the oxidized (SnO 2 ) layer at the interface of α-SnWO 4 and NiO x .

Conclusion
In summary, we have successfully detected the presence of an intraband state in α-SnWO 4 films covered with oxide overlayers. SPV spectroscopy measurements on NiO x -coated α-SnWO 4 show higher photovoltage amplitudes compared with the pristine film at photon energies above and below the bandgap of α-SnWO 4 . The former is attributed to increased band bending, while the sub-bandgap signals can be explained by the presence of an intraband defect state as a result of NiO x deposition. To confirm whether the intraband defect state is truly present at the interface, valence band measurements using HAXPES were performed. Since our earlier investigation has shown that the deposition of NiO x on α-SnWO 4 introduces an interfacial oxide layer likely to be SnO 2 , [12] and detecting the intraband state on NiO x -coated samples is hindered by the overlapping signals from the NiO x overlayer, we have intentionally grown a thin SnO 2 layer on the surface of our α-SnWO 4 films by performing a photoelectrochemical oxidation treatment. HAXPES analyses were then performed to further examine the interface between α-SnWO 4 and SnO 2 . The formation of the SnO 2 layer is confirmed by the increase of the Sn 4þ contribution in the Sn 3d core-level spectra. More importantly, valence band measurements of the SnO 2 -coated α-SnWO 4 films reveal the presence of an intraband state %0.7 eV above the VBM; this state is not detected in the pristine film. The signal intensity corresponding to this state does not change with increasing SnO 2 thickness, indicating that it originates from the interface between α-SnWO 4 and SnO 2 . Interestingly, the energetic position of the intraband state detected by HAXPES agrees very well with the sub-bandgap signals observed in the SPV analysis of NiO x -coated α-SnWO 4 , consistent with the presence of an SnO 2 layer at the α-SnWO 4 /NiO x interface. Consequently, the origin of Fermi-level pinning, which was previously shown to limit the photovoltage that can be extracted from NiO x -coated α-SnWO 4 photoanodes, [11,12] can be explained by the presence of an intraband state at the α-SnWO 4 /oxide interface.

Experimental Section
Thin-Film Preparation: The α-SnWO 4 films were prepared by pulsed laser deposition (PLD) as reported previously. [10][11][12][13] As the substrate material, a fluorine-doped tin oxide-coated glass (FTO TEC 7, Pilkington) was used, which was cut into pieces of approximately 24 Â 24 mm 2 . Before film deposition, the substrates were cleaned in five successive 15 min ultrasonic bath treatments using different solutions, starting with 1 vol% Triton solution (Triton X-100, laboratory grade, Sigma-Aldrich), followed by acetone, propanol, ethanol, and deionized water (Milli-Q, 18.2 MΩ cm). Subsequently, the substrates were dried under an N 2 gas stream. A custom-made α-SnWO 4 target was ablated in a custom-built PLD system (PREVAC, Poland). The preparation procedure of the α-SnWO 4 target has been reported previously. [11] A KrF-excimer pulsed laser (wavelength λ = 248 nm, LPXpro 210, Coherent) was focused on the target located in the preparation chamber. The base pressure in the chamber was 10 À7 mbar. During deposition, O 2 gas was used as the process gas such that the pressure in the chamber was increased to %10 À4 mbar. Target ablation was performed with a laser repetition rate of 10 Hz and a fluence of 2 J cm À2 . The laser beam spot size was 1.3 Â 2 mm 2 . Between the target and the substrate, a distance of 60 mm was set. In order to achieve a homogeneous film thickness, the substrate was rotated at 12°s À1 (2 rpm). After the films were deposited at room temperature, the samples were annealed at 520°C for 2 h in a tube furnace under argon flow to obtain the crystalline α-SnWO 4 phase.
The NiO x layers were deposited on the α-SnWO 4 films using the same PLD setup. A metallic Ni target (99.99%, Alfa Aesar) was used for film deposition at room temperature. An O 2 process gas (0.05 mbar chamber pressure) was used to deposit the NiO x layer in a reactive deposition process. The laser fluence was 2 J cm À2 , and the laser repetition rate was 10 Hz. The target rotation was set to 12°s À1 (2 rpm), and the target-tosubstrate distance was 60 mm.
The SnO 2 layers were formed by photoelectrochemical oxidation of the α-SnWO 4 samples. For that purpose, the samples were mounted in a custom-made Teflon cell and a potential of 1.23 V versus RHE was applied with an EG&G Princeton Applied Research (Model 273A) potentiostat. A three-electrode configuration was used with an Ag/AgCl reference electrode (XR300, KCl saturated, Radiometer Analytics) and a coiled Pt wire acting as the counter electrode. Sunlight was simulated with a WACOM super solar simulator (Model WXS-505-5 h, AM 1.5, Class AAA). The film oxidation was performed in a 0.5 M KP i buffer (pH 7) electrolyte, containing 0.5 M Na 2 SO 3 as hole scavenger.
Surface Photovoltage Measurements: Modulated SPV spectra were measured in a parallel plate capacitor configuration, composed of a quartz cylinder partially coated with the SnO 2 :F electrode, and a mica sheet as an insulator. [37] The measurements were performed at ambient conditions. The SPV signal is recorded as a change in the surface potential as a result of the illumination. Illumination from the front (i.e., light impinging the surface of the sample) was provided by a halogen lamp, coupled to a quartz prism monochromator (SPM2) and modulated at a frequency of 8 Hz using an optical chopper. The in-phase and phase-shifted by 90°S PV signals were detected with a high-impedance buffer and a dual phase lock-in amplifier (EG&G 5210). The photovoltage amplitude is defined as the square root of the sum of the squared in-phase and 90°phase-shifted SPV signals, [38] as defined in Equation (1). A more detailed description of the SPV setup, including the schematic and the specific roles of each components, has been reported in the literature. [26,37,38] Finally, additional discussion on the role of straylight and how it affects the sub-bandgap signals, as well as the consideration for choosing the modulation frequency, can be found in Supporting Information Note S1.
Hard X-ray Photoelectron Spectroscopy: The surface analysis of the α-SnWO 4 and α-SnWO 4 /SnO 2 samples was performed at the BESSY II synchrotron at Helmholtz-Zentrum Berlin, Germany. The measurements were performed at the KMC-1 beamline with the HIKE (high kinetic energy photoemission) end-station. [39,40] Photon energies of 2210, 3880, 3933, and 6009 eV were used for the measurements and adjusted with Si (111) and Si (311) double crystal monochromator planes. The photoelectrons emitted from the samples were detected with a Scienta R4000 electron spectrometer. Core level and valence band spectra were recorded with a step size of 0.1 eV and the pass energy was set to 200 eV. The base pressure of the analysis chamber was %10 À8 mbar. The calibration of the binding energy scale was done by acquiring the Au 4f 7/2 core level from a gold foil attached on the sample holder and in electrical contact with the sample. The Au 4f 7/2 core-level binding energy was set to 84.00 eV. During the measurements, a common ground was established between the sample and the electron spectrometer. Quantitative analysis of the peaks was performed with a peak fitting procedure. Prior to that, a Shirley baseline correction was performed. Pseudo-Voigt functions (e.g., linear combination of Gaussian and Lorentzian functions) were used for the procedure. Finally, chi-square minimization was performed to obtain the fitted spectra.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.