A study of S-doped TiO₂ for photoelectrochemical hydrogen generation from water

Abstract

Sulfur-doped titanium dioxide (TiO₂) was investigated as a potential catalyst for photoelectrochemical hydrogen generation. Three preparation techniques were used: first ballmilling sulfur powder with Degussa P25 powder (P25), second, ball milling thiourea with P25, and third a sol–gel technique involving titanium (IV) butoxide and thiourea. The resulting powders were heat-treated and thin-film electrodes were prepared. In all three cases, the heat-treated powders contained small amounts of S (1–3%). However, Rietveld analysis on X-ray diffraction (XRD) measurements revealed no significant changes in lattice parameters. For the samples prepared using thiourea, X-ray photoelectron spectroscopy (XPS) measurements indicated the presence of N and C in the heat-treated powders in addition to S. In all cases, visible-ultraviolet spectroscopy performed on bulk powders confirmed the extension of absorption into the visible region. However, the same spectroscopic technique performed on thin-film electrodes (∼0.5 μm) suggests that the absorption coefficients were very small in the visible region (≤10⁴ m⁻¹). The first and third methods yielded powders with substantially smaller photocatalytic activity relative to P25 powder in the UV region. The electrodes prepared from powders obtained using the second method yielded photocurrents comparable to those prepared from P25 powder.

Appendix

Assignment of XPS peaks

Sulfur peaks

The identification in the literature of S 2p peaks from XPS data has not always been consistent. While a peak at around 163 eV has been consistently identified as S²⁻ species, such as occur with Ti–S bonds [10]; different authors have identified a peak at around 168 eV as S⁶⁺/S⁴⁺ species, or alternatively as adsorbed SO₂.

Umebayashi et al. [31] referred to two articles by Onishi et al. and Sayago et al. [22, 28] to support their identification of the peak as absorbed SO₂. The former article in fact identifies this peak as ${\text{SO}}^{{2 - }}_{3}$ (i.e., S⁴⁺). The latter article, discussed below, provides a survey of different literature values. Zhang et al. [36] who also identified a peak at 169 eV as adsorbed SO₂, referred to one of the above articles [22] and to a paper [10] that is concerned only with a peak at around 163 eV, which is identified as an S²⁻ peak.

Sayago et al. [28] collected literature values of S 2p peaks for S-based species absorbed on different surfaces, mainly oxides. Their results are shown in Table 2. They noted that there is a large variation of reported values, because the bonding of the SO₂ molecule is strongly substrate-dependent for oxide surfaces.

Table 2 Range of S 2p _3/2 binding energies reported for sulfur species on oxide surfaces

Sayago et al. [28] referred to by articles Rodriguez et al. [25] and Polcik et al. [23] as a basis for identification of the physisorbed/multilayer SO₂ peak. Rodriguez et al. [24] observed a peak corresponding to physisorbed SO₂ between 168 and 171 eV in experiments in which a single-crystal ZnO surface was exposed to SO₂ at 110 K. However, these molecules desorbed upon heating to 150 K, and an ${\text{SO}}^{{2 - }}_{3}$ peak was formed. Moreover, no peaks corresponding to physisorbed SO₂ were observed when polycrystalline ZnO films or powders were exposed to SO₂. Polcik et al. [23] observed similar behavior when platinum was exposed to SO₂ at 90 K. An XPS peak corresponding to multilayer SO₂ was observed at 166.4 eV; however, the peak disappeared and was replaced by an ${\text{SO}}^{{2 - }}_{4}$ peak when the substrate was heated to room temperature.

The temperatures reached in ball-milling experiments are much greater than those at which physisorbed SO₂ becomes desorbed from single-crystal and metallic surfaces. Moreover, TiO₂ that is ball-milled with the sulfur compounds is in powder form. Hence, the possibility that the peak at around 168 eV observed in such experiments is associated with physisorbed SO₂ can be rejected. The peak should be identified as either ${\text{SO}}^{{2 - }}_{3}$ (S⁴⁺) or ${\text{SO}}^{{2 - }}_{4}$ (S⁶⁺); such peaks have been observed on oxides in single-crystal, polycrystalline, and powder form, and are present at temperatures up to at least 600 K. For example, Rodriguez et al. [24] identified ${\text{SO}}^{{2 - }}_{3}$ and ${\text{SO}}^{{2 - }}_{4}$ peaks resulting from exposure of polycrystalline ZnO to SO₂ at temperatures up to 500 K. They found that ${\text{SO}}^{{2 - }}_{3}$ and ${\text{SO}}^{{2 - }}_{4}$ resulted from the reaction of SO₂ with oxygen sites. Rodriguez et al. [24] observed an SO_x peak originating from exposure of single-crystal TiO₂ to sulfur. A peak associated with S atoms bonded to Ti rows on the surface was also observed; this disappeared upon heating to 550 K.

Nitrogen peaks

Two main N 1s peaks have been observed in XPS measurements of nitrogen-doped TiO₂. A peak between 396 and 397 eV has been consistently identified as being due to Ti–N bonds (e.g., [1, 7, 11, 30]). The identification of this peak is based on an XPS study of TiN oxidation chemistry [26].

The second peak, around 400 eV, is generally assigned to molecularly chemisorbed N₂, (e.g., [1, 5, 11]), again based on Saha and Tompkins’s study [26]. However, the peak at 400 eV has also been assigned to hyponitrite (${\text{N}}_{2} {\text{O}}^{{2 - }}_{2}$) species, [27] to the presence of NH_x species, [7] and to organic compounds. [8] The hyponitrite peak has been identified in studies of photo-oxidative nitrogen fixation on UV-illuminated TiO₂ surfaces [2, 15, 16]. These studies show that adsorbed nitrogen is oxidized to hyponitrite anions (${\text{N}}_{2} {\text{O}}^{{2 - }}_{2}$), which can be further oxidized to nitrite anions (${\text{NO}}^{ - }_{2}$) and nitrate anions (${\text{NO}}^{ - }_{3}$). The binding energies of these peaks are given in Table 3.

Table 3 N 1s binding energies reported for oxidized nitrogen species on TiO₂ surfaces after Navio et al. [15]

Oxygen peaks

The O 1s peak corresponding to Ti–O bonds occurs at 530 eV [26]. Navio et al. [16] have identified a further peak at 531.7 ± 0.1 eV corresponding to adsorbed hyponitrite species. Studies of SO₂ absorption on metal and oxide surfaces [4, 25] have identified a peak at around 531.5 eV that has been assigned to ${\text{SO}}^{{2 - }}_{3}$ species.