Quantitative analysis of annealing-induced structure disordering in ion-implanted amorphous silicon

Abstract

Comment on "Quantitative analysis of annealing-induced structure disordering in ion-implanted amorphous silicon," by Ju-Yin Cheng et al.,We use fluctuation electron microscopy to characterize medium-range order in ion-implanted amorphous silicon. In fluctuation microscopy, intensity fluctuation in a dark-field image contains the information of high-order atomic correlations in the length scale of 1-3 nm. In this study, we heated as-implanted silicon at 500, 550, and 580°C for various times. Our results indicate that in the beginning amorphous silicon is a disordered phase with robust medium-range order. Thermal annealing leads to disordering of the structure. Furthermore, we find that the activation energy of the disordering is about 2.7 eV, close to the activation energy for thermal relaxation ͑about 2.2 eV͒. Our finding suggests a strong correlation between structure disordering and thermal relaxation.

Structural characterization of pressure-induced amorphous silicon

“…The lowering of the intensity variance upon annealing of the ion-implanted a-Si has been reported before, but the shift in the peak position has not. 27 Previously, this lower variance of annealed specimens has been attributed to the approach to a more ideal CRN for not only ion-implanted a-Si, 27 but also for amorphous germanium. 25 Thus, our observations not only suggest that relaxed PI a-Si is, within the sensitivity of our techniques, the same as relaxed ion-implanted a-Si, but also that this relaxed state of a-Si may be as close to the CRN model as is practically possible.…”

“…Since FEM data do not directly allow a unique model of the amorphous structure to be constructed, previous studies have inferred structural information from FEM data by comparing the measured variance with simulations from proposed models. For deposited a-Si, and also to some extent for ion-implanted a-Si, two distinct models have been proposed: a paracrystalline model 21,26,27,35,36 and a void model. 37,38 However, both the paracrystalline and the void model were found to fit the data equally well, demonstrating the essential nonuniqueness of structural models obtained by a particular solution to the FEM data.…”

“…To date, for the case of amorphous semiconductors the magnitude of the variance has been correlated with the degree of MRO by interpreting a higher variance as indicative of a higher degree of MRO. [25][26][27][28] This degree of MRO depends on a number of indistinguishable factors such as the number of regions of correlated structure, the size of these regions, and the relative order within such a region. Moreover, the positions of the peaks in the variance plot give some additional insight into the local bonding and the type of MRO within the regions of correlated structure.…”

Electron Irradiation on Amorphous Silicon

“…This observation may be due to the difference in sample geometry (although other causes are possible too). Before further discussion, it should be known that the bulk sample of ion-implanted amorphous silicon (prepared by P. M. Baldo [6]) was annealed at 500 °C for 135 minutes (by B. J. Kestel [6]). After that, the near-surface layer was chemically thinned to produce electron transparent regions.…”

“…Indeed, this prediction is confirmed by our experiments, which were done separately in the JEOL 4000 and Philips CM12 microscopes (we ever attempted to use one microscope to do both hollow cone and dark field, but did not succeed). The operating conditions for these microscopes can be found elsewhere [5,6].…”