Two FeH pairs inn-type Si and their implications: A theoretical study

Abstract

Experimental evidence for interstitial ͕FeH͖ pairs in n-type Si stems from thermally stimulated capacitance ͑TSCAP͒. Electron-paramagnetic resonance ͑EPR͒ data have also been interpreted in terms of ͕FeH͖ pairs. We present theoretical studies of two ͕FeH͖ pairs. The properties of the first match those of the TSCAP center but are incompatible with the EPR center. The second is a possible candidate for the EPR center. If true, this suggests that high-temperature anneals can introduce substitutional Fe in concentrations higher than commonly believed.

First-principles calculations of iron-hydrogen reactions in silicon

“…No further levels were found for Fe i H n defects. Although our results for Fe i H are not far from previous theoretical reports, 20 the Fe i H(0/+) level seems too deep to be connected to the 'E v + 0.32 eV' trap of Ref. 16.…”

“…In line with the calculations reported by Szwaki et al, 20 we predict that after trapping one hydrogen atom, the Fe i atom becomes more stable near the hexagonal site while connecting to H along the trigonal axis. The resulting Fe i H structure is depicted in Fig.…”

“…The modest bind-ing energies of the H atoms to Fe i seem consistent with the annealing temperature in the range of 125-175 • C reported for a hole trap at E v + 0.32 eV and assigned to an iron-hydrogen complex. 16,17,20 However, an assignment to Fe i H (with a single H atom) conflicts with its predicted migration barrier, which is close to that of Fe i . Accordingly, both defects are expected to anneal out at close temperatures (just above room temperature).…”

“…In the case of Fe s H, we found it to be bistable, with H connecting to Fe s along the 100 direction in the neutral charge state, or along 111 towards the tetrahedral interstitial site in the negative charge state. These structures are represented in the configuration coordinate diagram of 20 (0.08 eV) and we can only suggest that in that work the Brillouin-zone sampling that was employed (Γ-only) was not sufficiently dense considering the size of the supercells (64 atoms).…”

“…While the early literature indicates a relatively low thermal stability for Fe-H complexes (the measured binding energies and annealing experiments on one hand, 16,17,19 and the first-principles calculations by Szwacki et al 20 on the other hand, suggest that these complexes can only survive to temperatures of at most 125-200 o C), more recently the solar-Si community has turned the attention to the effect of hydrogenation on Fe diffusivity and gettering at higher temperatures. For instance, Ref.…”