Precise measurement of chiral spectrum in the ultraviolet band using weak measurement and a deep neural network model

Abstract

Circular dichroism (CD) spectrum and optical rotation (OR) spectrum, crucial for understanding molecular properties and configurations, present challenges due to limited testing methods and equipment accuracy in the ultraviolet (UV) region. This study proposes a weak measurement system for chiral signals in varying concentrations in the ultraviolet range, optimized using a deep neural network (DNN) model. Introducing different post-selections to detect the circular dichroism spectrum and optical rotation spectrum separately, with contrast as a probe, it achieves a detection resolution of up to 10−6 rad. Moreover, the fitted value of the training data can reach 0.9989, enhancing the prediction accuracy of chiral molecule concentrations. This method exhibits considerable promise for applications in chiral measurement and sensor technologies.