Ac/dc conductivity and ML-based evaluation of electric characteristics of methylene blue solution

Abstract

Utilizing a precision LCR meter, the complex permittivity ε*(f) of methylene blue (MB) solutions in distilled water (DW) was determined at various temperatures and concentrations in the lower frequency range from 20 Hz to 2 MHz. The complex ac conductivity σ*(f) of the liquid samples was calculated using ε*(f). DC conductivity, σdc of the samples was also calculated. The effect of changes in temperature and concentration on the σ*(f) and ε*(f) of solutions are examined. A solution of MB in DW, ranging from diluted to concentrated, greatly enhances the electrical conductivity σdc. The empirical Casteel-Amis (CA) formula was employed to fit the experimental σdc values at different concentrations. Accurate measurement of complex permittivity of liquid solutions over wide range of concentration at broad range of frequencies and temperature is laborious and uneconomical. Therefore, we performed the complex permittivity measurement of the aqueous solution of MB over a limited concentration and temperature range over 201 frequency point. Then we used this experimental data in CatBoost regression model to predict the dielectric properties at intermediate frequencies and various concentrations. This model was employed to assess the performance of material characteristics using R-squared score. The study demonstrated that the CatBoost regression model can predict material performance at intermediate frequencies and concentrations while decreasing measurement resource needs by 60 %.