Multiple comonomer concentrations prediction from FTIR spectra with quantum chemistry-based interpretation

In this study, we enhance our previously proposed modeling approach, which combines wavelet transform and linear regression and utilizes Fourier transform infrared spectroscopy spectra, to simultaneously estimate the concentrations of two comonomers in binary copolymerization reactions. The advanced multi-output models were evaluated using reaction systems involving methyl methacrylate against two styrene-based and three methacrylate-based monomers with the help of quantum chemical calculations for peak interpretation. The multi-output model surpassed individual concentration prediction models in prediction accuracy for the shared comonomer of methyl methacrylate across the data sets, with identified peaks uniquely associated with the comonomer. Graphical abstract: (Figure presented.)