Miniaturized multispectral imaging for microfluidic pH sensing

Abstract

Multispectral lab-on-chip imaging based on complementary metal-oxide semiconductor (CMOS) is an innovative technology that supports portable microfluidic platforms and enables the extraction of spectral information from specimens under multiple wavelengths. Accurate pH sensing is essential across a diverse range of chemical, biological, medical, and environmental applications. In this paper, we presented a new approach for pH sensing using the CMOS-based multispectral lab-on-chip imaging. Using a six-band color filter array that spans the optical spectrum in the visible and near-infrared range and a U-shape PDMS microfluidic channel, images of solutions with specific pH values were captured by a monochrome CMOS image sensor. To predict pH values, we developed a Support Vector Regression (SVR) model, which was trained using captured images of the six bands. Predicted pH values were obtained by the model with an RMSE of 0.364. Our research demonstrates that multispectral lab-on-chip imaging represents a miniature and rapid method to achieve pH sensing with high sensitivity and accuracy.