Surface Evolution Of Nanostructured Crn And Si3n4 Films

junhua xu,lihua yu,isao kojima
DOI: https://doi.org/10.1063/1.1617358
IF: 2.877
2003-01-01
Journal of Applied Physics
Abstract:Comparative surface evolution of nanocrystalline (nc)-CrN and amorphous (a)-Si3N4 monolithic films and their alternative multilayer films has been studied in relation to film thickness in monolithic films, as well as to modulation periods in multilayers. Structural analysis was carried out using atomic force microscopy (AFM) and high-resolution transmission electron microscopy (HRTEM). The scaling behaviors of the AFM topographical profiles were analyzed using the one-dimensional power spectral density method (1DPSD). It is interesting to note that the growth exponents (beta) of both CrN and Si3N4 films can be divided into two regions of beta(1)=0.22+/-0.08 and beta(2)=0.87+/-0.1 for CrN films and beta(1)=0.09+/-0.01 and beta(2)=0.24+/-0.03 for Si3N4 films. The critical transition thickness, where the roughness jumps, is about 16.0 for CrN films and 168.0 nm for Si3N4 films. The 1DPSD analysis showed that the dependence at high frequency can be fitted to f(-4.3) and f(-4.2) power law decays, since the film thickness is less than 16.0 for CrN films and 168.0 nm for Si3N4 films. However, a region with an intermediate slope of -2 appeared as the film thickness exceeded 37.0 for CrN films and 222.0 nm for Si3N4 films. For CrN/Si3N4 multilayers, the roughness increases with increased modulation periods, and is much smaller than that of monolithic Si3N4 films at small modulation periods. Cross-sectional HRTEM analysis of the films revealed that the highly diffusive surface a-Si3N4 films are responsible for very smooth surfaces, while the oriented grain growth of nc-CrN films is a determining factor in the surface roughening of multilayers. The resulting parameters are discussed in terms of three competitive factors: surface diffusion, shadowing, and oriented growth, and are compared with the results of previous experimental and theoretical studies. (C) 2003 American Institute of Physics.
What problem does this paper attempt to address?