Variable strain energy in amorphous silicon

  • PDF / 816,853 Bytes
  • 7 Pages / 593.28 x 841.68 pts Page_size
  • 26 Downloads / 203 Views

DOWNLOAD

REPORT


Different Raman experiments on structural relaxation of a-Si and a-Ge are reviewed and discussed in relation to calorimetric measurements on a-Ge. On the basis of the correlation found between results from Raman spectroscopy and results from calorimetry in the case of a-Ge and of the strong similarity between a-Si and a-Ge in terms of their Raman spectra, it is suggested that the strain energy in a-Si may vary considerably with preparation conditions and subsequent treatments. Under this assumption the a-Si Gibbs free energy versus temperature has been constructed for material in different initial states of relaxation. It is shown that the melting temperature of amorphous silicon should increase when relaxation occurs during the heating phase prior to melting. Thus differences in apparent melting temperature, as observed under different laser heating conditions, may be explained.

I. INTRODUCTION The history of amorphous silicon (a-Si) dates back to the early days of ion implantation into crystalline silicon (c-Si)' and to the pioneering experiments on vacuum deposition,2 both roughly two decades ago. Studies of its properties and behavior upon various (heat) treatments have been carried out ever since, especially in relation to crystallization and solid-phase epitaxy. A boost in research activities occurred around 1975 when it became clear that a-Si in its hydrogenated form (a-Si:H) shows semiconducting properties and can be used as such in devices.3 a-Si:H differs in many respects from its "clean" counterpart a-Si, but fundamental questions concerning the atomic structure are similar for the two types of material. It was only after the first reports on pulsed-laser annealing in 1975 that the possibility of melting (ion-implanted ) a-Si became an issue.4 In the years following this discovery of damage repair by pulsed-laser irradiation, it was suggested that the process does not involve simple thermal melting5 but rather the formation of a high-density, electron-hole plasma. In the latter picture, the temperature of the Si lattice does not rise to the melting temperature.6 More recently, there is a general consensus that indeed a-Si can be melted by pulsed-laser irradiation.7 The temperature at which melting occurs, however, still poses major questions. In 1978 Bagley and Chen8a and Spaepen and Turnbull 8b predicted that the melting temperature of a-Si should be considerably ( > 200 K) lower than that ofc-Si (1685 K). Their prediction was based on calculations of the Gibbs free energy versus temperature of a-Si and /-Si, in comparison to that of c-Si. Experimental indications for a difference in melting temperature were found in 1980 by Baeri et al.,9 who studied pulsed electron beam heating of a-Si. Experiments on continuous J. Mater. Res. 3 (6), Nov/Dec 1988

http://journals.cambridge.org

wave (cw) laser heating10 and on line-source electron beam heating,'' in contrast, did not show any significant difference: solid-phase epitaxial regrowth of a-Si was found to occur even at temperatures relatively close to the m

Data Loading...