On the electron mobility and the donor centres in reduced and lithium-doped rutile (TiO2)

Abstract

Hall-and Seebeck-data of reduced and lithium-doped rutile both with and without alumina added are reported. The data are interpreted by a one band model. Anisotropy of the relaxation time must be taken explicitly into account. The donors in reduced rutile are probably titanium interstitials, which are multiple donors. The first and second of their ionization potentials are derived from the experimental data. Lithium is a single shallow donor. It is demonstrated that aluminium introduces more than one acceptor level. One of these levels is situated close to the conduction band.

Influence of annealing atmosphere on formation of electrically-active defects in rutile TiO2

“…Si substitutes for Ti in rutile TiO 2 71,72 and is not expected to introduce localized electronic states in the bandgap but rather to give rise to bandgap narrowing. 60,72 Li in an interstitial configuration is potentially a shallow donor, 26,60 but the predicted E A value is only 30 meV and does not agree with those of our three common levels. Also, the anticipated electron trap positions associated with Cr in rutile TiO 2 deviate substantially from those of our three common levels, exhibiting values of 190 meV and 410 meV below the conduction band edge.…”

“…There is quite a spread in values reported for m à e . 11,34,[59][60][61][62][63][64] In this study, a value of ten-times the free electron mass m e is assumed. Figure 5 displays exemplary DLTS spectra measured on Pd/TiO 2 /InGa junctions.…”

“…Al itself substitutes for Ti and gives rise to an acceptor level in the lower part of the bandgap, i.e., Al on the Ti site acts as a p-type dopant. 34,73 However, on the basis of generation noise recombination and temperature-dependent Hall effect measurements, it is argued by different authors 60,74 that Al-related shallow acceptor states also occur close to the conduction band. In particular, Acket and Volger 60 have ascribed a level at $50 meV below the conduction band edge to a 2Al Ti -V O complex which may be associated with our most shallow "general" defect level (E A % 70 meV) or alternatively with the prominent one at %60 meV detected after forming gas and N 2 gas anneals only (D 2,FG , D 2,N ).…”

Continuum and atomistic description of excess electrons in TiO2

“…There is a large amount of uncertainty concerning the value of electron effective mass in TiO 2 obtained experimentally (typically, they are obtained from the analysis of Hall coefficients, conductivity and thermoelectric power): values for rutile have been reported in the range between 5 m e and 13 m e [79] [80] [81] [82] and as large as 12-32 m e [83]; however, it has been argued that the measured value is the mass of a polaron, while the mass of a bare electron is smaller, of the order of 3 m e [80] or 6-7 m e [84]. For anatase nanoparticles, the measured values are much smaller, ~1 m e [85] or 0.71-1.26 m e [86] depending on particle size.…”

“…We find, by numerical differentiation of the lowest CB level near the point, the values of 2.1 m e and 1.1 m e for the X and Z directions in bulk rutile and 0.9 m e and 0.8 m e , respectively, in bulk anatase. However, for the model we adopt the values based on experiment: 1.22 m e consistent with anatase [86] and 10 m e consistent with rutile [79] [81] [82] and we will discuss the effect of changing the effective mass on the electron localisation.…”