Chemical Engineering Journal, Vol.371, 43-54, 2019
Mechanistic aspects in the formation of nano- and submicron particles in a batch and a continuous microfluidic reactor: Experiment, modeling and simulation
To investigate mechanistic contrasts in particle formation between a batch and a continuous microreactor, a coupled computational fluid dynamics (CFD)-population balance equation (PBE) based model is proposed; incorporating mixing, reaction, nucleation, diffusion-growth, Brownian-coagulation, shear-induced coagulation and Ostwald ripening (OR), occurring simultaneously. This enables prediction of nanoparticle size over a wide range of temperatures, flow-fields and solvents. This has been possible for the first time, since we included: (i) effect of evaporative-loss of solvent at a high synthesis temperature in our batch reactor experiments, along with (ii) turbulent-shear coagulation of nanoparticles in this reactor, due to the imposed strong forced convection; while accounting for, in contrast, (iii) laminar-shear coagulation in the microreactor [Gutierrez et al. (Chem. Eng. J., 2011, 674-683)], due to prevailing low Reynolds number flow-synthesis in their experiments. We find that, evaporation significantly reduces particle size by reducing critical radius for OR. Consequently, without it, a strikingly high over-prediction error (about 40% with respect to our experimental particle size) is seen for synthesis of SiO2 particles in ethanol itself, with a more pronounced effect in methanol, a more volatile solvent. Furthermore, in contrasting flow fields of the two reactors, inclusion of the correct function, either turbulentshear or laminar-shear based coagulation, for batch or microreactor, respectively, enables us to predict temporal particle size evolution accurately. For batch reactor, prediction is completely a priori, while for microreactor, coagulation efficiency as a single adjustable parameter achieves that; without these, existing models suffer about 20% and 15% under-prediction in particles size, respectively, compared to experimental data.