The photochemical identification of two chemisorption states for molecular oxygen on TiO2(110)

Abstract

We report the first experimental observation of two chemisorption states for molecular oxygen on a TiO2(110) surface containing anion vacancy sites. The first molecular species can be photoactivated to oxidize coadsorbed CO to CO2 (α channel) and undergoes slow photodesorption. The second molecular oxygen species only undergoes fast photodesorption (β channel). Conversion from α-O2, to β-O2 occurs upon heating the surface to above 200 K.

TiO2-based Photocatalysis: Surface Defects, Oxygen and Charge Transfer

“…The UV induced photodesorption of O 2 from TiO 2 has been shown to be first-order in O 2 coverage, and careful analysis of the photodesorption signals show clear evidence for at least 2 photodesorption channels for O 2 from TiO 2 [17]. The two channels are deconvolved based on their respective cross sections, where the a-channel undergoes slower photodesorption (Q ¼~10 )17 cm 2 ) and the b-channel undergoes faster photodesorption (Q ¼~10 )15 cm 2 ).…”

“…They find multiple states of adsorbed O 2 which exhibit differing cross sections, but one state observed (which desorbs at 400 K) cannot be related to the a-O 2 species or the b-O 2 species shown in figure 7, because both either convert or desorb by 400 K [17,18]. Figure taken from reference [17].…”

“…Lu et al [16] were able to clearly show that by adsorbing a test molecule on the defective surface as well as on a fully stoichiometric surface and comparing the results of temperature programming, one can observe the reactivity of thermally-created defect sites and can measure their relative population. Specifically in that work, thermally-created defect sites were shown to be active for the reduction of 13 [17,18] where it was shown that the presence of defects is necessary for the UV-induced photodesorption of molecular oxygen. Later work done by Rusu et al [19] showed that the intensity of photodesorbing oxygen has a direct correlation to the defect density on the thermally annealed TiO 2 (110) surface.…”

“…This observation suggests that a-O 2 species can convert to b-O 2 species either by a chemical change at the O 2 adsorption site, or by migration to another TiO 2 site, where the species changes chemically. Lu et al [17] have defined the 2 species in the following way: the first channel (a) has the ability to photooxidize CO into CO 2 and has a lower photodesorption cross section than the second species (b) which does not play a role in CO oxidation and is observed to only photodesorb. The authors of this work propose that the 2 photodesorption channels for adsorbed oxygen may be explained by the presence of both a superoxide type O 2 (O 2 ) ), which has Photon Energy (eV) Figure 6.…”

“…For O 2 adsorbed at 105 K, a photodesorption curve exhibiting a long tail (and hence a low photodesorption cross section (Q a » 10 )17 cm 2 ) is measured. As the adsorbed oxygen is annealed to higher temperatures in sequential experiments, the desorption tail becomes sharper, and measurements of its shape yield a higher photodesorption cross section (Q b » 10 )15 cm 2 ) [17]. When the adsorbed O 2 species are used to photooxidize coadsorbed CO, it is seen that the a-O 2 species are singularly active for CO 2 formation, as shown by the sequence of CO 2 photoproduction plots in the right-hand side of figure 7.…”