Determining the crystalline degree of silicon nanoclusters/SiO2 multilayers by Raman scattering

We use Raman scattering to investigate the size distribution, built-in strains and the crystalline degree of Si-nanoclusters (Si-nc) in high-quality Si-rich oxynitride/SiO2 multilayered samples obtained by plasma enhanced chemical vapor deposition and subsequent annealing at 1150 °C. An initial structural characterization of the samples was performed by means of energy-filtered transmission electron microscopy (EFTEM) and X-ray diffraction (XRD) to obtain information about the cluster size and the presence of significant amounts of crystalline phase. The contributions to the Raman spectra from crystalline and amorphous Si were analyzed by using a phonon confinement model that includes the Si-nc size distribution, the influence of the matrix compressive stress on the clusters, and the presence of amorphous Si domains. Our lineshape analysis confirms the existence of silicon precipitates in crystalline state, in good agreement with XRD results, and provides also information about the presence of a large compressive stress over the Si-nc induced by the SiO2 matrix. By using the Raman spectra from low temperature annealed samples (i.e., before the crystallization of the Si-nc), the relative scattering cross-section between crystalline and amorphous Si was evaluated as a function of the crystalline Si size. Taking into account this parameter and the integrated intensities for each phase as extracted from the Raman spectra, we were able to evaluate the degree of crystallization of the precipitated Si-nc. Our data suggest that all samples exhibit high crystalline fractions, with values up to 89% for the biggest Si-nc. The Raman study, supported by the EFTEM characterization, indicates that this system undergoes a practically abrupt phase separation, in which the precipitated Si-nanoclusters are formed by a crystalline inner part surrounded by a thin amorphous shell of approximately 1-2 atomic layers. © 2014 AIP Publishing LLC.
The research leading to these results has received funding from the European Community's Seventh Framework Programme (FP7/2007–2013) under Grant Agreement No: 245977, under the project title NASCEnT. The present work was supported by the Spanish national project LEOMIS (TEC2012-38540-C02-01) and MAT2010-16116.