Muonium donor in rutileTiO2and comparison with hydrogen

Abstract

Muonium, a positive muon and an electron, is often used as an experimentally accessible substitute for hydrogen in materials research. In semiconductors and insulators, a large amount of information on the hydrogen behavior is deduced from this analogy; however, it is seldom demonstrated that this procedure is justified. We show here, via a comparison of the hyperfine interactions, that in TiO 2 muonium and hydrogen form the same configuration with the same basic electronic structure. A detailed description of the bonding characteristics of the muon to the Ti 3+ polaron is presented. The special role of muon motion within the so-called oxygen channel in the rutile structure, which occurs at a lower temperature than for hydrogen, is emphasized. Muonium (Mu) is a pseudoisotope of hydrogen in which the proton is replaced by a positive muon (μ + ), with a factor of 9 lighter mass. Muon spin spectroscopy (μSR) uses muons implanted with 100% spin polarization and offers a very sensitive method to study the properties of this isolated pseudohydrogen in solids [1,2]. It is usually assumed that information obtained from μSR can be transferred with appropriate modifications to H. However, overlapping experiments to support this assumption are scarce. A particularly relevant case is the doping character of H in semiconductors and oxides [3][4][5][6], where practically all calculations refer to the electronic structure of H whereas most experimental information comes from μSR [7][8][9][10][11][12][13]. Overlapping data exist only for ZnO where proton-ENDOR (electron-nuclear double resonance) data [14] can be compared directly with μSR results [15][16][17][18]. A number of properties (e.g., ionization energy) are indeed similar for the two species. However, the measured hyperfine interaction (hfi), scaled with the magnetic moments, differs by almost a factor of 10. This raised the question of whether the same configuration is measured, or if the H center may involve an additional defect [19,20].Here, we report a case where the same configuration can be established for H and Mu. We compare the hfi of the μSR experiment with the proton-ENDOR result, both for rutile TiO 2 . The H center in TiO 2 was extensively studied by Brant et al.[21] using electron paramagnetic resonance (EPR) and ENDOR, who found that the electron is located at the Ti ion reducing it from Ti 4+ to Ti 3+ . H is bound to one of the six O atoms surrounding Ti and the magnetic interaction between the proton and the electron is mainly dipolar. This specific hfi permits a sensitive comparison of the two experiments. We have observed a dramatic change of the μSR spectra with increasing temperature and a complete disappearance of the hyperfine splitting at 10 K. We show that this is due to rapid jumps of the muon between neighboring bonding positions to O atoms around Ti 3+ . The very strong angle dependence of the dipolar interaction and the averaging over values in different * ruivilao@fis.uc.pt positions lead to the reduction and final disappearance of the hfi...

Isolated hydrogen configurations in zirconia as seen by muon spin spectroscopy andab initiocalculations

“…Muonium spectroscopy has the particular advantage of corresponding to the high-dilution limit for the muonium impurity, which can thus quite generally be considered isolated and is affected only indirectly by other defects and impurities (including hydrogen impurities) through the overall Fermi energy. Extensive studies have thus been carried out to characterize the muonium centers formed in different semiconductors and oxides [12,13,[19][20][21] and the respective results for the muonium configurations compare well with those obtained with protons, for the very few cases allowing comparison [9,[22][23][24][25].…”

“…These polaronic centers can also be created when samples are irradiated with UV light or x rays or subjected to treatment in hydrogen atmosphere [61]. A polaronic state has also been observed in similar systems, notably TiO 2 [24,25,62]. The difference between the experimental value of the hyperfine interaction and the predicted computational value for the lowest-energy neutral oxygen-bound configuration is probably an indication of spin dynamics [25] or that the lattice does not reach its full relaxed configuration around the muon during the time of the experiment (about 15 μs; see Fig.…”

“…6 we present a Fourier transform of a high-statistics spectrum obtained for YSZ at 8.5 K. This Fourier spectrum was obtained using the Lomb method [58]. From the similar case of TiO 2 we know that the hyperfine interaction is mainly dipolar [24,25]. We may immediately estimate its approximate value from the full width at half maximum of the line, yielding D ≈ λ/π = 0.09 MHz.…”

“…[64] as an alternative model. Another possibility, in analogy to the H/Mu behavior in TiO 2 [25], is the onset of muon motion between nearly equivalent hydrogen/muon sites. However, the lattice geometry is possibly less favorable in ZrO 2 .…”

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

“…18 The identification of defects in rutile TiO 2 is particularly challenging as the energy scheme of the defect levels may be influenced by polaronic effects. [19][20][21][22][23][24][25][26][27][28][29] First, this makes the theoretical description of defects in rutile TiO 2 exceptionally challenging. 19,24 Second, polaronic effects can also lead to different defects having similar experimental signatures as electrons are trapped at similar Ti sites no matter which defect they actually originate from.…”