A MULTISPECTRUM NONLINEAR LEAST-SQUARES FITTING TECHNIQUE

An extension of the nonlinear least squares spectrum fitting technique has been developed for the simultaneous fitting of multiple spectra. This procedure's smaller number of fitted parameters as compared to fi tting one spectrum at a time improves the determination of spectroscop ic parameters. A more reliable evaluation of the errors associated wit h the solution is possible. Correlations among fitted parameters may p reclude their determination from a single spectrum fit. If the correla tions differ from spectrum to spectrum, however, separation of these p arameters is often possible when including the spectra in one solution . Overfitting is also avoided when combining spectra with sufficiently different experimental physical conditions.