Efficient Scale-Up of a Solvent-Free Photooxidation Using a High-Power LED-Based Photomicroreactor

Recent advancements in light-emitting diode (LED) technology and photomicroreactors have provided a prudent roadmap for scaling-up the photooxidation processes. In this work, we presented a highly efficient, chromoselective, and wavelength-selective high-power LED-based photomicroreactor (HLPMR) for multikilogram ascaridole production through solvent-free α-terpinene photooxidation. First, we segregated kinetically controlled and mass transfer limited zones by implementing various operating schemes. Numerical investigations indicated that a simultaneous drop in the molar diffusion coefficient and reaction rate led to the decrease of α-terpinene conversion in the mass transfer limited zone. Kinetic investigations showed that apparent rate constants were improved by increasing the LED input power. Based on the rise in apparent constants and the control over the selectivity, we varied perfluoroalkoxy alkane (PFA) microchannel design configurations and its volume to increase the irradiated area for efficiently scaling-up the ascaridole productivity, space-time yield, and photochemical space-time yield to 3.5 kg/day, 9303.5 g L^-1 h^-1, and 34.7 mol kW^-1 day^-1, respectively. Predictive modeling using various supervised machine learning classifiers was executed to select the best possible photoreactor for this benchmark photooxidation. Random forest classifier showed better prediction performance using the train test split, cross-validation, and folding approaches among the implemented classifiers. Overall, HLPMR ensured a green and sustainable production of ascaridole with a multikilogram per day capacity.