Process Analytical Technology (PAT) for Smart Chemical Manufacture
: AI-Integrated Online Fluorescent Multiarray Sensing System
The monitoring of critical quality attributes (CQA) or process parameters such as concentration of products, raw materials, media and metabolites, cell viability, product aggregations or viral contaminations of the manufacturing line is a longstanding challenge of process engineering. It fundamentally requires a sensitive, selective, and multiplexing analytical line for an on-site rapid monitoring. Traditional analytical biochemistry including spectroscopy, mass spectrometry, chromatography, or electrophoresis have been utilized for PAT in current industry. However, it significantly delays in analysis time and spacing based on off-line monitoring.
We are interested in new analytics for smart PAT based on online/inline sensing system of the reactors and production line. We develop a multi-array microfluidics or fiber optics based on label-free fluorescent transducer design technique. We design and fabricate a compact, integrated fiber optic nanosensor element, measuring a wide range of chemical process parameter libraries including drugs, decontaminants, and product qualifiers. We introduce a lab-on-fiber design with the potential for at-line monitoring with the integration of 3D-printed miniaturized sensor tips having high mechanical flexibility. These results in total constitute an effective form factor for nanosensor-based transducers for applications in industrial process monitoring. Based on this, we aim to maximize the production efficiency.
Not only for the advanced PAT hardware, we are also interested in the new software for the smart PAT: AI-based multivate nanosensor signal analysis. We develop a high-performance deep learning-based real-time concentration analytic programs and utilize them for high-throughput single-cell image analysis, hidden signal detection, and process parameter tracing in the baseline noise of the sensor.
