Quantitative analysis of formaldehyde solution based on terahertz attenuated total reflection and chemometric methods

In order to explore the possibility of using terahertz attenuated total reflection (THz-ATR) technology combined with chemometric methods to quantitatively analyze different concentrations of formaldehyde solutions, the THz-ATR was applied to measure the formaldehyde solution with a concentration of 0% to 10%, and the absorption coefficient and refractive index of the samples in the frequency range of 0.2 to 1.2 THz were calculated. Partial least squares (PLS) and back propagation (BP) neural network (BPNN) algorithms are used to establish the mathematical model to predict the relation between the absorption coefficient and the concentration as well as the refractive index and the concentration. Correlation coefficient R and the root-mean-square error RMSE obtained from the PLS prediction model based on the refractive index are 98.99% and 0.4604, respectively, and those based on the absorption coefficient are 99.05% and 0.4464, respectively. Meanwhile, the correlation coefficient R and root-mean-square error RMSE obtained from the BPNN prediction model based on the refractive index are 99.80% and 0.2407, while those based on the absorption coefficient are 99.93% and 0.1182, respectively.