Hydration dominates dielectric spectroscopy-based high accuracy quantification of lactose content in cow's milk

Lactose is a critical factor in the quality of milk and dairy products. Achieving high accuracy and rapid detection of lactose content in cow's milk remains a challenge. Dielectric spectroscopy has emerged as a promising tool for detecting food components. We explored the effect of lactose content on the dielectric spectra of cow's milk and we propose a rapid analytical method for the quantitative determination of lactose content in cow's milk with high accuracy based on dielectric spectra.

We obtained the dielectric spectra of 316 cow's milk samples in the frequency range 20-4500 MHz and noticed a strong negative correlation between the lactose content and the value of the dielectric loss factor (ε″) below 1500 MHz. Lactose does not affect cow's milk dielectric properties by excluded volume effect, but dominates the effect on the dielectric properties of cow's milk by hydration. The support vector regression model based on the variable importance in projection has the best prediction performance for lactose content. Its root-mean-square error of prediction set and residual prediction deviation is 0.29 g kg^-1 and 6.968, respectively, and its prediction performance is better than that of the currently reported near-infrared (NIR) method and other methods.

Despite the weak polarity of lactose molecules, its hydration is a significant factor affecting the dielectric properties of milk. The present study provides a basis for high accuracy and rapid quantitative detection of lactose in cow's milk based on dielectric spectra. © 2023 Society of Chemical Industry.