Search

• Experimentally-validated model to predict size and distribution of nanoparticles. • Model couples time-resolved fluid dynamic simulations with population balance. • Increasing mixing rate, increases overall nanoparticle synthesis but not local rate. • Combination of fast mixing and high rate leads to a burst of nuclei formation. Achieving size control during nanomaterial synthesis is critical for their deployment in numerous applications due to their strong size-properties relationships. However, it remains a challenge in the field due to multi-step mechanisms in nanoparticle synthesis. In this paper, we present an experimentally-validated model able to predict the particle size and distribution of silver nanoparticles synthesised in flow reactors by coupling time-resolved fluid dynamic simulations with population balance. The model reveals fundamental correlations between reactor design and the corresponding size of the nanoparticles. It shows that increasing mixing rate, increases the overall rate of nanoparticle formation, where nanoparticle synthesis takes place homogeneously in the whole of the reaction volume. In contrast, when there is a distinguishable interface between reactant streams (i.e. low mixing rate), nucleation and growth take place in a small fraction of the reaction volume in the interface between the inlet streams. The combination of fast mixing and high average nucleation rate can lead to a burst of nuclei formation translated in small size and narrow size distribution. This work presents a new approach for the use of computational-guided design of reactors for the synthesis of nanomaterials to move the field away from current trial-and-error strategies. [ABSTRACT FROM AUTHOR]

Abstract: A three‐dimensional numerical model was solved for a turbulent incompressible steady flow and the discrepancies and similarities between the software OpenFOAM® and ANSYS® Fluent on the flow modeling were analyzed. The evaluation was performed for the meshing, setup, and simulation step of the modeling process. Outcomes were contrasted with experimental data to validate the fluid flow obtained with both softwares. With this aim, an experimental pulse‐response technique was applied to obtain the residence time distribution (RTD) curve to be compared with the RTD curves achieved from the modeling. Additionally, images of the tracer transport in the experimental reactor were taken, showing good agreement with the computational fluid dynamics (CFD) predictions. [ABSTRACT FROM AUTHOR]