Advancing process analytical technology for solution-phase reaction monitoring through data analysis and automated technologies

Process analytical technologies (PATs) provide real-time, automated measurements that enable deeper understanding of underlying chemistry, more efficient reaction conditions, and safer manufacturing processes. Recent advances in hardware and automation have allowed PATs to be integrated into increasingly complex chemical systems, while progress in computation and chemometric modelling enables researchers to extract useful chemical information from high dimensional PAT data. This thesis leverages these advances in both software and hardware to develop PAT methodologies for solution-phase reaction monitoring. First, a quantum mechanical spectral analysis (QMSA) workflow for benchtop nuclear magnetic resonance (NMR) spectroscopy is investigated. QMSA addresses the low signal dispersion commonly observed in benchtop ¹H NMR through first-principles modelling and reference data acquired on high-field NMR. Compared to the traditional indirect hard modelling approach, the proposed workflow demonstrates improved reaction monitoring performance and higher model transferability. Next, a Bayesian statistical workflow is proposed to enhance univariate calibration for infrared (IR) spectroscopy reaction monitoring. Three diagnostic probabilistic models combined with data preprocessing selection identify IR signals that improve calibration performance under non-linear conditions. Applied to a test reaction system, the workflow identified a fingerprint region signal and an unconventional preprocessing strategy that reduced prediction error by over 50% compared to conventional approaches. Finally, an online reaction monitoring platform integrating high-performance liquid chromatography (HPLC) and mass spectrometry (MS) is reported. By independently controlling dilution and solvent conditions for each detector, the system allows HPLC to monitor high-concentration species while MS captures low-level intermediates and catalyst species without chromatographic interference. Demonstrated on test reaction systems such as a Petasis multicomponent reaction and a Suzuki–Miyaura cross-coupling reaction, the platform delivers comprehensive kinetic profiles across both major and minor species in complex reaction networks.