Positron spectroscopy study of nickel and cobalt silicide film growth.

Abstract

This work represents the first systematic Reemitted-Positron Spectroscopy (RPS) study of the growth of the various phases (M$\sb2$Si, MSi, and MSi$\sb2$) of Co and Ni silicide films. These films were grown in situ by the solid phase thermal reaction of metal overlayers deposited on Si substrates. The Si substrates were found to have a positive positron work function $\phi\sp+$, with the observed weak reemission attributed to epi-thermal positrons. In the case of the silicides, $\phi\sp+$ was found to be negative for all of the different phases. The parameter $\Sigma$ (defined by $\Sigma\equiv -(\phi\sp- +\ \phi\sp+)$, where $\phi\sp-$ is the electron work function) was found to have a surprisingly large variation with stoichiometry, thus allowing each phase to be clearly distinguished in the RPS spectrum. There is a correspondingly large variation in $\phi\sp+$ with relatively little variation in $\phi\sp+$. In all cases, there is a relatively large difference in the values of $\Sigma$ between the silicide film and Si substrate. In addition, in an attempt to observe and calibrate shifts in $\sigma$ due to strained pseudomorphic growth of thin CoSi$\sb2$ films, the positron deformation potential, $\rm E\sbsp{d}{+}\equiv V(\partial\Sigma/\partial V)$, was determined using the thermal expansion technique. All of the silicides were found to have a relatively low yield of reemitted positrons, due to a high density of misfit dislocations and/or vacancies which are trapping the positrons. Defect trapping in the CoSi$\sb2$ films was studied in detail, with the positron diffusion lengths determined to be relatively short, and not increased significantly by annealing at higher temperatures. However, the measured value of the positron deformation potential indicates that, in the absence of defects, the diffusion length should be relatively large. This indicates that a large fraction of the positrons are trapping in defects in the film, or at the interface/Schottky well. RPS data are consistent with a short diffusion length, and inconsistent with a long diffusion length with significant trapping at the interface. This gives support to the conclusion that the majority of the trapping is occurring in defects in the film, and not at the interface. In order to confirm the above conclusion, and to further investigate the fate of positrons at the interface, these films were studied with Doppler-Broadening Spectroscopy (DBS). DBS spectra clearly show the transitions from the different phases of the silicide growth, as well as the motion of the interface with increasing silicide thickness. The data do not give any indications of vacancy accumulation in the substrate near the interface, as one might expect to occur due to the diffusion of Si atoms during the silicide growth. In addition, the data yields positron diffusion lengths which are in reasonable agreement with the RPS results. This gives further support to the conclusion that there is a great deal of defect trapping, evidenced by the short diffusion length, and that interfacial trapping does not play a major role. Moreover, the effect of the built-in electric field in the depletion region of the substrate was investigated, and found to have no substantive effect on the positron motion.