Spectrally resolved optical rotation dispersion (ORD) is vital for analyzing chiral substances in biomedical and pharmaceutical research, yet practical applications are hampered by weak chirality signals, limited spectral range, and complex multicomponent drug formulations. To address these challenges, we have developed a broadband, high-sensitivity ORD spectroscopy technique that integrates weak value amplification with convolutional neural network (CNN) modeling. In this innovative approach, the ORD is directly correlated with the real component of the weak value, achieving a measurement accuracy on the order of 10-6 rad across a spectral range of 750 to 1600 nm while enabling the detection of concentrations as low as 3.2 × 10-3 g/mL. Our technology not only facilitates high-sensitivity measurements of sugars and amino acids but also allows for monitoring temporal changes in specific rotation. In practical applications, our sensor can effectively identify the type of pure honey based on its specific rotation and can distinguish between adulterated honey that contains glucose and fructose. Additionally, by utilizing CNNs, we can accurately assess the concentrations and mixing ratios of various samples, achieving a fitted value of R2 ≥ 0.982. This technology offers significant prospects for chiral measurement and sensor applications, providing valuable insights for drug analysis and food authenticity verification.
Xu et al. (Tue,) studied this question.