(658d) Modeling and Simulation of Hydrothermal Synthesis of Nanoparticles in a Continuous Flow Reactor

Continuous hydrothermal synthesis of nanoparticles is gaining attention because of its simple and efficient operation, use of inexpensive solvent, short process time and inherent scalability. Typically, supercritical water is used as the reaction medium while an aqueous metal salt solution is used as the precursor. Streams of supercritical water and aqueous metal salt solution are continuously supplied to the reactor and the product is also continuously withdrawn. Upon mixing of reactants, product monomers form as a result of hydrolysis and condensation reactions, subsequently growing to nanoparticles via nucleation, diffusional growth, coagulation, coalescence, and Ostwald ripening. Now, physical, thermodynamic and transport properties of water under supercritical conditions are significantly different from those under ambient conditions. This results in enhanced rates of hydrothermal reactions and particle formation and growth. Besides the implementation of appropriate process conditions (temperature, pressure, pH, flow rates and concentrations of the reactants etc.), mixing of reactants also has a strong influence on the particle size distribution (PSD) and morphology of the product.

Various attempts, by both experimental and numerical investigations, have been made by several research groups to understand the mixing of reactants in different flow reactors. Particle size distribution (PSD) is the key product quality parameter but very little focus has been made to study the effect of mixing on the product PSD. The effect of mixing on temperature and velocity profiles inside the reactor has been studied using computational fluid dynamics (CFD) simulations. Similarly, population balance models (PBMs) have also been developed to predict the PSD. A few attempts have also been made to couple the CFD with PBM to predict the effect of mixing on PSD, but testing and validation of these coupled models with real experimental data has not been demonstrated so far. A few research groups have attempted to only correlate numerical results of flow and temperature fields inside the reactor with experimentally measured PSD. However, prediction of the PSD and its validation with experimental data is limited.

In the present work, we have developed a coupled computational fluid dynamics (CFD) and population balance model (PBM) to study the effect of reactor configuration and process parameters on the product PSD. The CFD model consists of mass, momentum, species and energy balance equations. The PBM is one-dimensional in nature and takes into account particle nucleation, diffusional growth and growth by particle coagulation and coalescence. The coupled CFD-PBM model was implemented in a commercial CFD software (FLUENT 14.5) and was tested with published experimental data for three different reactor configurations with a T-mixer, used for synthesis of ceria nanoparticles. The predicted thermal profiles are reasonably close to the published data obtained using neutron radiography. Similarly, the predicted velocity profiles are in qualitative agreement with the measured fluid density profiles. Experimental data were also published for the effects of different reactant flow rates and temperatures at a constant pressure on the PSD and the predictions of the coupled CFD-PBM for product PSD are in good numerical agreement with the data.

The validated model was subsequently used to study the effect of important process parameters. When the flow rates of supercritical water and aqueous metal salt solution are comparable, it was observed that vortex formation and circulation occur around the T-junction. This is primarily because of the buoyancy force which occurs due to the density difference between the streams. The mean size of the particles increases as residence time of particles increases due to recirculation. In other cases, where the difference between the flow rates is significant, the supercritical water slides over the metal salt solution and the mean particle size decreases. Since the proposed model can predict the effect of mixing on product PSD, it can be employed beneficially for reactor scale-up and for designing novel reactors including features targeted towards process intensification.