Prediction of antibacterial activity in ionic liquids through FTIR spectroscopy with selection of wavenumber by PLS

Ionic liquids (ILs) are novel organic salts with a wide melting temperature range that have enormous potential for industrial use. The study aims to predict the antibacterial activities of ILs against most commonly occurring bacteria in the environment i.e. Escherichia coli (E. coli), Enterobacter aerogenes (E. aerogenes), Klebsiella pneumoniae (K. pneumoniae), Proteus vulgaris (P. vulgaris), Pseudomonas aeruginosa (P.aeruginosa) and Steptococcus pyogenes (S. pyogenes) thorough Fourier transform infrared (FTIR) spectroscopy. The cationic parts of prepared ionic liquids include Imidazolium, Pyridinium, Quaternary ammonium and Tetraalkylphosphonium, while the anionic part includes Bromide, Methane sulphonate, Bis(trifluoromethanesulfonyl)imide, Dichloroacetate, Tetrafluoroborate and Hydrogen sulphate. For ILs prediction Partial Least Squares (PLS) based models were implemented by FTIR spectral data. Spectral wavelength selection and number of latent variables were optimized by minimizing the cross-validation root mean square error (RMSE). PLS based models showed good prediction of antibacterial activity against E. coli (RMSE = 2.26), E. aerogenes (RMSE = 2.37), K. pneumoniae (RMSE = 1.25), P. vulgaris (RMSE = 1.95), P. aeruginosa (RMSE 3.40) and S. pyogenes (RMSE = 2.83). Moreover, the data properties are also studied and are used to explain the prediction behavior of validated models. The results indicated that FTIR could be used as a rapid method to estimate the antibacterial activities of ionic liquids.