Measurement of Antioxidant Concentration Using Terahertz Time-Domain Spectroscopy (THz-TDS) in Ethylene-Vinyl Acctate-Copolymer (EVA)

Ethylene-Vinyl Acctate-copolymers (EVAs) are widely used in the manufacture of wires and cables as electrical insulating media in electric power fields. Oxidation is one of the primary mechanisms leading to the aging of polymer materials, and the addition of antioxidants can effectively slow the oxidation reaction. Current nondestructive testing methodologies face challenges in the measurement of antioxidant concentrations, which constitute a critical metric for aging assessments. In this article, we propose a method for measuring antioxidant concentration using terahertz time-domain spectroscopy (THz-TDS). In comparison to other techniques such as liquid chromatography, no elaborate sample preparation is necessary, and calibration measurements can be conducted rapidly. The spectral changes caused by the addition of different concentrations of antioxidants were investigated, and the results showed that the refractive index and dielectric constant increased significantly with a decrease in the antioxidant concentration. When the concentration of antioxidant 1010 in the sample was increased from 0% to 10%, the refractive index decreased from 1.55 to 1.42, and the dielectric constant decreased from 2.38 to 2.03 average over the band of 0.5–2.5 THz. Moreover, the numerical calculation using density functional theory (DFT) showed the introduction of oxygen-containing functional groups of antioxidant 1010 molecules into EVA molecules slightly reduced the polarization, resulting in a consistently decrease in the dielectric constant of the constructed materials compared with original EVA materials. The classification of antioxidant concentration in EVA was predicted using machine learning, including partial least squares regression (PLSR), random forest (RF), support vector regression (SVR), and extreme gradient boosting (XGboost). XGboost was found to have the highest prediction accuracy with an root mean square error (RMSE) of 0.0412. In order to ascertain the characteristic spectrum of the 1010 antioxidant, we selected the dielectric constant at the 0.9 THz position for a linear fit. The RMSE of cross-validation (RMSECV) value for leave-one-out cross validation (LOOCV) was found to be 0.38.