Spectroscopic Monitoring and Chemometric-Based Feedback Control of the Diazotization Reaction

Abstract

Inline Raman spectroscopy-based monitoring enables the development of control systems that allow the implementation of Process Analytical Technology (PAT) principles even in chemical reactions, thereby improving quality, efficiency, and safety. There are only a few articles in the literature that discuss such control systems. This may be due to the numerous challenges involved, including the need for robust measurement setups and the development of chemometric algorithms capable of delivering accurate and stable real-time evaluations. Thus, this study presents the Raman spectroscopic monitoring and chemometric-based feedback control of a highly exothermic and potentially hazardous diazotization reaction, namely, the synthesis of phenyldiazonium chloride from aniline, hydrochloric acid, and sodium nitrite. Raman spectra were analyzed offline using Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) and in real-time using Classical Least Squares (CLS) chemometric methods to monitor and control key reaction components. The developed feedback system successfully controlled concentration levels during both acid–base and diazotization phases. The reduction of the Raman signal-to-noise ratio, induced by the decomposition of the diazonium salt, was also observed. This factor significantly affected the monitoring of the diazotization experiment. Furthermore, a calibration was established to quantify the molar concentrations of key components based on Raman spectra. The calibration experiment successfully demonstrated well-fitted linear molar–spectral relationships for both aniline (R²_homogeneous = 0.997; R²_{heterogeneous} = 0.9947) and aniline hydrochloride (R² = 0.9987). For aniline, the molar–spectral relationship exhibited a distinct change in tendency that precisely correlated with its solubility limit in water, marking the transition from homogeneous solution to emulsion.