Precipitation of Mandelic Acid with a Supercritical Antisolvent Process: Experimental and Theoretical Analysis, Optimization, and Scaleup AÄ ngel Martı ´n,* Laura Gutie´ rrez, Facundo Mattea, and Marı ´a Jose´ Cocero Departamento de Ingenierı ´a Quı ´mica y Tecnologı ´a del Medio Ambiente, Facultad de Ciencias, UniVersidad de Valladolid, 47011 Valladolid, Spain An experimental and theoretical analysis of the precipitation of mandelic acid with a semicontinuous supercritical antisolvent (SAS) process is presented. The experimental section comprises a study of the effect of different operating parameters on particle size, including pressure, temperature, solution concentration, and flow rates. Prismatic or needlelike particles with lengths of 30-200 μm have been obtained, with precipitation yields between 20% and 80%. The parameter with a stronger influence on particle size is temperature, while an increase in the initial concentration allows a large increase in the precipitation yield with small variations in particle size. Variations in the injection velocities in the nozzle had only a minor influence on particle size. In the theoretical section the application of a detailed mathematical model of the SAS process to this system is discussed. The model is used for the interpretation of the different experimental trends and to propose the optimum process parameters. The model is also used to study the scale-up of the process and particularly the design of the nozzle for higher flow rates. As experimentally it has been found that the parameters of the nozzle do not affect particle characteristics, this discussion is focused on the determination of the required precipitator volume for different nozzle designs. Finally, a scaled-up process with a 5-20 times increase in flow rates and product amount has been tested experimentally, obtaining results similar to those in the lower scale experiments. Introduction Since size, size distribution, and morphology of particles are of the utmost importance in pharmaceuticals, cosmetics, and other applications, 1 particle design technologies have an increas- ing relevance. The supercritical antisolvent (SAS) processes for the production of fine powders have several advantages over other precipitation methods due to the peculiar properties of supercritical fluids: the mixing between the supercritical antisolvent and the liquid is much faster than in conventional liquid antisolvent processes, thus leading to higher supersatu- rations and smaller particle diameters. Moreover, it is possible to control the particle size distribution (PSD) with changes in the operating conditions. The supercritical antisolvent can be easily removed from the final product by reducing pressure, in contrast to the complex purification processes often required when organic antisolvents are used. With a proper selection of the antisolvent, it is possible to carry out the process at near- ambient temperatures, avoiding the thermal degradation of the product. For these reasons supercritical antisolvent processes have been extensively studied in recent years for applications which include pharmaceuticals, natural compounds, explosives, polymers, and pigments. 2-4 One of the main challenges in the development and optimiza- tion of an SAS process is the large number of process parameters that affect the performance of the process. Further, the interac- tions between these parameters due to their simultaneous influence on different process steps such as the fluid mechanics, mass transfer, and particle formation and growth kinetics make it difficult to relate changes in the process conditions to changes in product characteristics, or to predict the effect of a certain variation in the operating parameters. For this reason the development of an SAS process often requires an extensive experimental study with different combinations of process parameters. 5-7 Likewise, when an SAS process is scaled up, it is difficult to predict the effects of the variations in the process design on its performance, and particularly the variations in the fluid mechanics and mass transfer behavior of the process. There are very few references in the literature which deal with the scale-up of SAS processes. Some of the most representative are the papers of Reverchon et al. 8 and Jarmer et al. 9 In the first, the authors applied their results on the precipitation of amoxicillin in a laboratory scale 5 to develop a precipitation process on a pilot scale. They found that the parameters that determine the product characteristics were the same in the laboratory scale and pilot scale. In particular, the initial concentration of the solution had a strong effect on the particle size and the PSD of the product. However, the design of the nozzle had a limited influence on the process, and although different nozzles were used in the laboratory- and pilot-scale experiments, product characteristics remained almost the same. In contrast, Jarmer et al. 9 studied a case (the precipitation of poly(lactic acid)) in which the design of the nozzle had a strong influence on the performance of the process. These authors studied different criteria for scaling up the nozzle. They found that the best scale-up criterion was a constant energy dissipation rate in the nozzle, while maintaining a constant Reynolds number or axial velocity was not enough to ensure the reproducibility of the results. Those two papers illustrate which are the critical parameters of the process depending on the operating conditions and in particular on the position of the operating point with respect to the mixture critical point: the thermodynamic parameters above the mixture critical point, and the fluid mechanic parameters under conditions of partial miscibility. In a previous work, 10 a mathematical model of the SAS process was developed. This model considers the main steps of the SAS process, including the fluid mechanics, the mass transfer, and the kinetics of particle nucleation and growth. One * To whom correspondence should be addressed. Tel.: (+34) 983- 423-174. Fax: (+34) 983-423-013. E-mail: mamaan@iq.uva.es. 1552 Ind. Eng. Chem. Res. 2007, 46, 1552-1562 10.1021/ie0608051 CCC: $37.00 © 2007 American Chemical Society Published on Web 02/08/2007