Piezo-spectroscopic data analysis: a PC tool

Information from the many kinds of spectroscopy used by chemists and physicists is fundamental to our understanding of the structure of materials. Numerical techniques have an important role to play in the augmentation of the instrumentation and technology available in the laboratory, but are frequently viewed as separate from the laboratory procedures. We describe an integrated PC-based approach for obtaining directly the parameter estimates of transition types in piezo-spectroscopic measurements of crystalline materials. Typically, the analyses in question are required to handle complex secular matrices, to distinguish between components in the experimental results, and to identify the transition types as rapidly and as efficiently as possible. The method described, based on providing a discrete shell to the Powell algorithm, is shown to give both accurate identification of the transition type in the case of new data and improved fits (i.e. reduction in residual variation) when compared with results obtained via standard procedures. In addition it is flexible with respect to the language used and possesses a high degree of portability. We illustrate the success of the approach using (i) data previously reported on the solution of a trigonal defect which includes both mixing of states and spin orbit interactions and (ii) new data obtained for a defect related to beryllium impurities in silicon