Abstract
An improved method of analysis of e.s.r. spectra is described, based on correlation methods. The most prominent line is identified by comparison of the experimental spectrum with a matching ‘test spectrum,’ which consists initially of a single Gaussian or Lorentzian derivative shape, but which includes a more and more complex pattern of lines as the analysis proceeds. A product function P(produced by correlation of the test spectrum with the experimental spectrum) is used as the criterion of improvement of fit. At each stage of the analysis, the product function is computed for a range of values for an additional coupling constant, a display of product function versus coupling constant constituting the a-spectrum. The coupling constant giving the biggest improvement in P is now incorporated in the test spectrum with the appropriate multiplicity, and the process is repeated until the a-spectrum shows a maximum at the origin, when the analysis is complete. The analysis allows coupling constants to be assigned with precision even in weak spectra; a unique analysis is often possible in a weak spectrum where few of the lines are recognizable by the naked eye. This feature should be especially valuable in identifying spectral satellites due to nuclei such as 13C and 29Si present in natural abundance. The benzyl and the 1-ethoxy-ethyl radical are analysed in detail, the latter being used in spectra with successively lower signal: noise ratios to demonstrate the power of the method, and finally the method is used to determined the position of the 29Si satellites of the triethylsilyl radical.

This publication has 0 references indexed in Scilit: