RESUMEN
A high-precision pH monitoring system over a wide pH range is introduced. The system comprises a cavity-enhancement module constructed by two high-reflectivity mirrors, a microfluidic pH sensing chip based on a binary-indicator membrane of Congo red and m-cresol purple, and a hyperspectral transmission module. This structure extends the effective absorption optical path of the sensing chip, significantly amplifying the spectral differences at various pH values. The spectrum of the transmitted light is recorded by a self-developed hyperspectral module and then converted to broadband cavity-enhanced absorption spectrum (BBCEAS) via the Beer-Lambert law. An artificial neural network (ANN) is employed to predict pH values of the solution. With such a design, this system exhibits a wide detecting range of 2 M [H+] - 2 M [OH-] (corresponding to pH -0.3-14.3) with a response time of about 120 s. The system can achieve a higher detection accuracy with root mean square error (RMSE) of 0.073, as compared to 0.137 without the cavity enhancement. The system also possesses good properties of repeatability, long-term stability, ion resistance, and organic corrosion resistance. These excellent properties make the proposed system a promising candidate technology for harsh environments, such as seawater acidification warning, chemical plant sewage monitoring, and biological sample detection.