Deconvolution of overlapping spectral polymer signals in size exclusion separation-diode array detection separations by implementing a multivariate curve resolution method optimized by alternating least square

Peaks eluting from a size exclusion separation (SEC) are often not completely baseline-separated due to the inherent dispersity of the polymer. Lowering the flow rate is sometimes a solution to obtain a better physical separation, but results in a longer retention time, which is often not desirable. The chemometrical deconvolution method discussed in this work provides the possibility of calculating the contribution of each peak separately in the total chromatogram of overlapping peaks. An in-house-developed MATLAB script differentiates between compounds based on their difference in UV-spectrum and retention time, using the entire 3D retention time UV-spectrum. Consequently, the output of the script offers the calculated chromatograms of the separate compounds as well as their respective UV-spectrum, of which the latter can be used for peak identification. This approach is of interest to quantitate contributions of different polymer types with overlapping UV-spectra and retention times, as is often the case in, for example, copolymer or polymer blend analysis. The applicability has been proven on mixtures of different polymer types: polystyrene, poly(methyl methacrylate) and poly(ethoxyethyl acrylate). This paper demonstrates that both qualitative and quantitative analyses are possible after deconvolution and that alternating concentrations of adjacent peaks do not significantly influence the obtained accuracy.