Combined deconvolution and curve fitting for quantitative analysis of unresolved spectral bands

摘要: Curve-fitting programs usually only yield accurate parameters for unresolved spectral bands when the separation of neighboring exceeds their average half-width. By reduction width each feature in a heavily overlapped band multiplet by Fourier self-deconvolution (FSD), conditioning curve-fitting algorithms is improved to point that area can be estimated with much accuracy. The practical limitations this approach are discussed paper. Most FSD strictly applicable Lorentzian bands; effect increasing Gaussian fraction isolated on accuracy estimation has also been investigated. Finally, application combined deconvolution and curve fitting synthetic multiplets which relative peak absorbances bandwidths varied shown estimates areas until approaches instrument line-shape function, determined truncation, exponentiation, apodization domain array.