Unambiguous Characterization of Advanced Semiconductor Alloy Films & Multi-Layer Superstructers
Heteroepitaxial layers on single-crystal substrate materials are of increasing importance in various fields of material engineering. An enhanced semiconductor alloy multi-layer structure allows one to achieve remarkable improvement of the electronic properties (such as high transit frequencies, low noise, etc) of the devices made on this basis. In particular, group III-V & II-VI compound structures and group IV elements grown on silicon & III-V alloy substrates are widely used in opto- and micro-electronics in designs for new generation semiconductor technology.
High-resolution x-ray diffraction methods are well-established non-destructive techniques for efficient diagnostics of such systems. Information about crystal-lattice strain conditions can be directly related to impurity concentration quantities in the case of alloyed layers, and the layer thickness can be determined with high precision. The widely used method of least-square fitting (LSF) of a calculated reflectivity to the experimental diffraction profile relies on an a priori known model of crystal-lattice deformation distribution. The method works well for a large class of crystal structures and allows one to obtain information about the lattice strain profiles as well as the geometrical macro-parameters of the crystalline layers grown on the top of a semiconductor wafer. However, in some cases a mistake or uncertain conditions in the growing process could lead the resulting fitted parameters of the layer to quite different values from the expected ones. In such a case use of the expected parameters as the initial seed can cause the fitting procedure to fall rapidly into an unreasonable situation. In addition, the fitting procedure is not, ultimately, physically correct since there is no direct relation between the features of the experimental set-up and the analysed strain profile dimension parameters. The spatial resolution which can be actually achieved in the resulting strain distribution depends directly on the angular aperture of the measured intensity profile. This fact, however, is not taken into account in the fitting procedure.
A physically substantiated self-consistent method for a model-independent determination of the crystal-lattice strains in a single-crystal with a surface layer having mismatched lattice spacing from the underlying substrate has recently been developed (Nikulin, et.al., Phys. Rev. B, 1996). The technique, however, does not give a unique solution, and often there are two resulting strain profiles which are very similar but not identical.
This page presents a new approach to the unambiguous solution of the inverse problem which uses high-resolution x-ray diffraction data collected for two different radiation wavelengths. The difference in absorption coefficients allows us to localise complex zeros, i.e. to distinguish between the true and artificial ones, and, thus, to solve the problem uniquely. The method does not need special radiation energies and can be performed using laboratory equipment, i.e. at characteristic radiation wavelengths (Nikulin, et.al., J. Appl. Phys., 1996).
|
