Datensatz

Zusammenfassung

Several techniques for the synthesis of nanoparticles on the laboratory-scale have been developed in recent years [1]. With respect to industrial-scale processes it is of major importance to identify a suitable production route for the formation of narrowly distributed nano-scale particles. The precipitation of crystals induced by chemical reaction in conventional stirred tank reactors is a widely used industrial process. Due to the ionic character of the participating reactants a very fast and irreversible chemical reaction takes place. The production of sparingly soluble product is followed by subsequent nucleation and growth of particles. Additionally, agglomeration of single crystals in highly concentrated reactors influences the mean particle size. Furthermore, differences in time constants between chemical reaction and micro mixing cause inhomogenieties in crystallization kinetics. The combination of both effects results in broad size distributions of micrometer particl
es. In order to produce narrowly distributed particles on the nanometer scale emulsion-assisted precipitation techniques overcome the drawbacks of bulk precipitation processes. In particular the application of precipitation reactions in water pools of water-in-oil(w/o)-emulsions with a mean droplet size on the submicrometer scale is of interest for industrial applications [2].
This contribution provides a population balance model-based analysis of an emulsion-assisted precipitation process. We consider a case where the precipitation reaction is induced by mass transfer of one of the reactants from the continuous phase of the emulsion across the liquid-liquid interface into the droplets, while the other reactant is located in the aqueous dispersed droplet phase. The actual precipitation reaction takes place in the droplets. This technique can be considered as a semi-batch process. In a first process step a well-prepared emulsion is filled into the reactor, whereas one of the reactants is solved in the aqueous core of the droplets. In a second step the other reactant is added as a pure substance into the reactor by a dosing pump.
A mathematical model of this process consisting of mass balances in the continuous and dispersed phase and a population balance in the droplet phase has been derived. Assuming a spherical shape and symmetric boundary condition at the droplet interface the description of spatial effects of mass and population dynamics on the radius coordinate in a spherical coordinate system is sufficient. Consequently, a 1+1 dimensional population balance equation for the particle formation was deployed, whereas aggregation and breakage of particles are disregarded. As shown experimentally by Sathaye et al. [3] nucleation at the liquid-liquid interface should not be neglected in emulsion assisted precipitation processes. Therefore, two nucleation mechanisms, namely homogeneous nucleation in the volume of the droplet and heterogeneous nucleation at the droplet interface are both incorporated in the population balance. Hence, the interplay of different nucleation mechanisms and particle growth
on the one hand, and the influence of process parameters such as the droplet size, concentration levels and dosing rate on the particle properties on the other hand are investigated by computer simulations.
It is shown that the predominant nucleation mechanism determines the width and the modality of the resulting particle size distribution. At fixed nucleation and growth kinetics, the particle size and the polydispersity of the distribution can be varied by adjusting the process parameters. We show for example that for the production of small primary particles with a narrow size distribution a reduction in size of the emulsion droplets is favorable.
A number of model parameters need more experimental input because a detailed modeling of interfacial phenomena is necessary to assess nucleation and mass transfer at the liquid-liquid interface. In order to prove interfacial transport and nucleation mechanisms, applicable experiments have to be developed. The derived model will be validated by experiments using a suitable reaction system.