Mössbauer Spectroscopy on Fe Impurities in Si Materials

Abstract

Based on a series of Mössbauer spectroscopic investigations on Fe impurities in p-type and n-type Si materials, we propose a new model for Fe impurities in Si matrix, consisting not only of interstitial Fe, but also substitutional Fe atoms with different charge states. Mössbauer spectroscopy enables us to observe directly these components which transform each other by changing external conditions such as under light illumination, under external voltage, and also under external stress. This means that not only interstitial Fe impurities, but also substitutional Fe impurities appear to be a source for producing “electrically active Fe impurities” in Si materials.

Mössbauer parameters of Fe-related defects in group-IV semiconductors: First principles calculations

“…118 They are also at variance with a recent proposal that the Fe i B s pair gives rise to a single resonant line (presumably without or with unresolvable quadrupole splitting) at δ = 0.42 mm/s. 43 On the other hand, they are in much better agreement with earlier work by Bergholz on 57 Co-contaminated ptype Si, where at least three doublets were observed with intensities that depended on the temperature of the measurements, and were all assigned to 57 Co i B s pairs. All these doublets had δ values of about 0.7 mm/s, and were split by either 0.21, 0.54 and 1.04 mm/s, the one with larger ∆ being assigned to the 111 -oriented structure studied here.…”

“…42 Extensive measurements in n-and p-type samples, where 57 Fe was introduced by several methods, were also carried out by Yoshida and his group. 13,26,43 Although their data for p-type Si seems to be in line with other reported data (δ = 0.8 mm/s), for n-type material a resonance at δ = 0.40 mm/s has been assigned to Fe 0 i . Since our calculations conflict with this assignment, we investigated several possible sources of error, including (i) increasing the BZ sampling grid to MP-4 3 , (ii) bringing the 3s state of Fe into the valence to be treated within the Kohn-Sham scheme, (iii) including an on-site correction to the exchange-correlation energy by means of the GGA+U +J approach as proposed by Liechtenstein et al 113 (with Hubbard and Hund parameters in the range U = 2-5 eV and J = 0-1 eV), and (iv) using the local density approximation to the electronic exchange-correlation interactions.…”

“…The ISOLDE consortium reported a δ(Fe s ) = −0.04 mm/s peak which is effectively independent of the dopant type and concentration, and was therefore assigned to the neutral charge state of substitutional iron (Fe 0 s ). From the same experiments, neutral and positively charged Fe i defects were assigned to resonances at δ(Fe 0 i ) = 0.72 mm/s and δ(Fe + i ) = 0.78 mm/s in n-type and p- 26,43 Other defects were tentatively assigned using MS, namely a quadrupole-split doublet labeled Fe N with ∆(Fe N ) = 0.51 mm/s and δ(Fe N ) =0.43 mm/s, assigned to an iron-vacancy (Fe i V) pair, 12 and another doublet labeled Fe D which has been associated to Fe in regions damaged by the ion-implantation process, with ∆(Fe D ) = 1.02 mm/s and δ(Fe D ) =0.33 mm/s. 33 Ab-initio calculations of Mössbauer parameters have played an important role in the interpretation and validation of experimental data.…”

“…12,42 More recently, and for the same defects, Yoshida and his group reported IS values of δ(Fe 0 s ) −0.17 mm/s, δ(Fe 0 i ) = 0.40 mm/s and δ(Fe + i ) = 0.80 mm/s. 26,43 Other defects were tentatively assigned using MS, namely a quadrupole-split doublet labeled Fe N with ∆(Fe N ) = 0.51 mm/s and δ(Fe N ) =0.43 mm/s, assigned to an iron-vacancy (Fe i V) pair, 12 and another doublet labeled Fe D which has been associated to Fe in regions damaged by the ion-implantation process, with ∆(Fe D ) = 1.02 mm/s and δ(Fe D ) =0.33 mm/s. 33 Ab-initio calculations of Mössbauer parameters have played an important role in the interpretation and validation of experimental data.…”