Initial Nucleation and the Effects on Epitaxial Silicide Formation
- PDF / 1,094,349 Bytes
- 6 Pages / 417.6 x 639 pts Page_size
- 6 Downloads / 240 Views
256 coherent epitaxy is likely. It has been well established that Auger spectroscopy is quite useful for determining when interface or surface atomic intermixing has occured, but the technique is not sensitive to long range order. High resolution TEM is used to characterize the epitaxial nature of the interface, but the preparation methods preclude in situ measurements. Additionally, the preparation steps subject the sample to temperatures ~-150°C, thus as-deposited TEM specimens can not be routinely prepared. Raman spectroscopy has, however, proved to be applicable to examining the initial nucleation properties [6,8]. The measurement displays the optic phonons which have a wavelength similar to that of the laser light in the sample. This is --1200,A and corresponds to nearly Brillouin zone center excitations. Crystalline Si thus exhibits one strong feature due to the triply degenerate zone center phonon and a continuum of weaker features due to two phonon excitations. In contrast, amorphous materials exhibit a continuum spectrum similar to the vibrational density of states. Thus if sharp silicide lines are observed then crystalline phase formation can be inferred [9]. FREE ENERGY MODEL In general the largest driving force for the reactions at a metal/Si interface is the chemical energy gained by creating the silicide. This is traditionally modelled by heats of formation of the compounds involved. A difficulty of this formalism is that microscopic configurations which deviate from the well ordered phase are not accurately described. We suggest that the chemical energy can be more accurately described by an atomistic free energy model where the chemical energy is accounted for by the energy gained by creating metal-Si bonds at the expense of Si-Si and M-M bnds. At the interface the significant diffusion leads to environments that do not have the long range order of the crystalline silicide compound, and in fact the short range order of all the atomic sites will not exhibit the preferred arrangement of the crystalline compound. In this atomistic free energy model, the major components are expressed in terms of the bond energy and the atomic scale strain. This is more or less a tight binding approach to approximating the energetics of the system. The free energy can thus be represented schematically in a configuration type diagram with two axes; one representing the degree of intermixing and the other representing the strain. The evolution of a metal/Si reaction could then be described as shown in Fig. 1. The point labeled I represents the free energy of the unreacted system. If the reaction occurs in a layer-by-layer way directly forming the silicide, then the dashed-dot line to A would represent the evolution of the system. Here A represents the stable silicide. A more likely evolution is that the interdiffusion leads to atomic environments which are not exactly that of the ordered crystal. Thus while the chemical energies are similar to the layer-by-layer situation, a significant contribution due to strain occurs.
Data Loading...
