Annealing dynamics of irradiation-induced defects in high-purity silicon in the presence of hydrogen

Abstract

A reaction model explaining (i) the hydrogen-mediated transformation of the vacancy-oxygen (V O) center into a vacancy-oxygen-hydrogen center (V OH * ), with an energy level at 0.37 eV below the conduction-band edge (E c ), and (ii) the passivation of the divacancy center is presented. V OH * dissociates with a rate of 2 × 10 −5 s −1 at 195 • C, causing V O to recover after long duration (>10 4 min), while a similar evolution occurs at 300 • C on a time scale of the order of 10 min. The diffusivity of the monatomic hydrogen used in the model agrees closely with the established values for the diffusivity of protons. After the recovery of V O, further annealing at higher temperatures and/or longer durations transforms V O into the "ordinary" vacancy-oxygen-hydrogen center with an energy level at E c − 0.32 eV (V OH). V OH is subsequently transformed into V OH 2 . For temperatures above 250 • C, two additional hydrogen-related levels occur (∼ 0.17 and ∼0.58 eV below E c ) with a one-to-one ratio and a possible association with different charge states of a V 2 OH center is discussed.

The effect of millisecond flash lamp annealing on electrical and structural properties of ZnO:Al/Si structures

“…However, in hydrogenated and diffusion-oxygenated float zone Si (DOFZ) Si wafers, the loss of the VO center at 400 C is accompanied by the growth of the so-called E7 peak. 53 E7 occurs on the high-temperature tail on the VO peak in the DLTS spectra after annealing, similar to that of T2 in Fig. 5, and it appears likely that T2 is identical to E7.…”

Interaction Between Hydrogen and Vacancy Defects in Crystalline Silicon

“…There is, however, no indication of the V 2 H(−/0) level at E c −0.44 eV, if H is introduced from hydrogen‐plasma or acid solution . In these studies both DLTS peaks corresponding to V 2 (−/0) and V 2 (=/−) have the same amplitude even after annealing, which implies the absence of the V 2 H(−/0) level overlapping with V 2 (−/0) .…”

“…There are, however, experimental indications on the presence of V 2 H(−/0) overlapping with V 2 (−/0). It has been consistently observed by DLTS that the amplitude of the V 2 (−/0) peak is greater than that of V 2 (=/−) with almost a two‐to‐one ratio for H‐implanted n‐type Si, in contrast to hydrogen‐free irradiated material where the V 2 (−/0) and V 2 (=/−) DLTS peaks show equal amplitudes . Therefore, an additional H‐related contribution in the DLTS peak of V 2 (−/0) was suggested.…”

“…This defect has an acceptor (−/0) level in the upper part of the band gap at E c −0.32 eV and a donor (+/0) level in the lower part of the band gap at E v + 0.27 eV ( E v being the valence band maximum) . These levels are observed together by DLTS and minority‐carrier transient spectroscopy (MCTS) after H‐implantation as well as after H‐diffusion from acid solution into electron‐irradiated n‐type silicon . Figure a shows the amplitudes of the signals from V OH−/0), V OH(0/+), and V O(−/0) deduced from DLTS and MCTS measurements as a function of annealing temperature.…”

“…Although there is a consensus on the most stable atomic configurations of V OH n complexes shown in Figure , the existence of other configurations cannot be ruled out. Bleka et al observed, for example, that the annealing of V O in chemically‐hydrogenated high‐purity n ‐type Fz‐Si gave rise to a DLTS peak, labeled E 1, with a position at E c −0.37 eV. The formation of E 1 occurred after annealing at 225 °C for 20 min and demonstrated one‐to‐one negative correlation with V O.…”