Amorphous Si, Crystallization and Melting
- PDF / 1,290,397 Bytes
- 9 Pages / 417.6 x 639 pts Page_size
- 88 Downloads / 263 Views
AMORPHOUS Si, CRYSTALLIZATION AND MELTING
J. M. POATE Bell Laboratories, Murray Hill, New Jersey 07974
ABSTRACT
Recent experiments dealing with the thermodynamics of crystallization and melting of amorphous Si are reviewed. Differential scanning calorimetry measurements give the heat of crystallization of implanted, amorphous Si to be 11.3±t.8 kJ/mole. Gibbs free energy calculation based on these measurements indicate that amorphous Si melts at a temperature of 1460'K compared to the crystalline value of 1685"K. Evidence for this reduced melting temperature also comes from rapid heating measurements using a) structural information after solidification and b) dynamic conductance measurements during the melt. Solid phase epitaxial regrowth experiments which apparently do not show such a depression in amorphous melting temperature will be discussed. INTRODUCTION
The advent of rapid heating techniques has created much interest in the basic thermodynamic properties of Si. The condensed phases of Si that take part in these rapid-heating transformations are the solid (amorphous and crystalline) phases and the liquid phase. The thermodynamic relationships of the liquid and crystal phases are quite well established but there is some uncertainty regarding the amorphous phase. The fact, however, that amorphous Ge and Si are in a higher free energy state than the crystalline state led Bagley and Chen1 ) and Spaepen and Turnbull2 ) to propose that the amorphous phase melts at a discrete temperature (Ta1) which is lower than the crystalline melting temperature (Tr). These initial predictions were based on calorimetric measurements of the heat of crystallization (AHaJ. of amorphous Ge as no measurements existed for Si. The Gibbs free energies of the amorphous and liquid phases were calculated with respect to the crystalline phase and the intersection of the liquid and amorphous free energy curves gave Tae.-0.8 Tce. The argument regarding this remarkable depression in melting temperature is principally based on the premise that the phase transition from the amorphous to liquid phase is discontinuous and first order. The reasoning behind this premise is quite plausible when it is realized that the transition involves a fundamental change in bonding from the covalent, amorphous phase with fourfold coordination to the metallic liquid with 11-12 fold coordination. Clearly this transition will only occur in very rapid heating measurements otherwise the amorphous phase will recrystallize directly in the solid phase as observed, for example, in furnace heating measurements. The transition should be reversible so that undercooling or supercooling the melt beneath T., should lead to growth of amorphous Si in the melt. This paper will review recent experiments addressing some of these issues. Firstly, measurements of the heat of crystallization of implanted, amorphous Si layers will be discussed along with calculations of the Gibbs free energies and amorphous melting temperatures. Experiments on the melting of amorphous Si will then be
Data Loading...
