Study of Particle Formation using Supercritical CO2 as an Antisolvent

Abstract

Particle design using supercritical CO2 has been of great interest in the pharmaceutical, microelectronic, catalytic, and related industries over the past 10 years. There have been numerous papers and patents published on the processes studied in this work. The solubility of most drug compounds in carbon dioxide is very low, making it a very attractive antisolvent for particle formation at suitable ranges of temperatures and pressures. This thesis explores the use of different CO2 antisolvent precipitation system designs for the formulation of small crystalline drug particles of a given size, morphology, and uniformity, using the precipitation of acetaminophen from ethanol as an example.

In order to understand the precipitation process, the equilibrium concentration of acetaminophen in CO2 and CO2 plus ethanol were measured at a range of temperatures and pressures in a high-pressure extraction system. This information is important in understanding the supersaturation of the drug at various precipitation conditions.

Several antisolvent processes were tested in order to determine their effectiveness in controlling the precipitation of acetaminophen from ethanol. The first system involved the use of Solvent Enhanced Dispersion by Supercritical Fluids (SEDS) patented by Hanna and York (WO9501221, 1994). This process uses a coaxial nozzle design where the solvent with the solute of interest is injected in the inner tube and the supercritical CO2 is injected in the outer tube. The two streams mix at nearly constant pressure and temperature in a small volume region of the nozzle before expanding through the nozzle tip into a chamber maintained at a fixed temperature and pressure. The fast mixing process rapidly expands the solvent with CO2 in order to induce phase split of the solid drug particles. The chamber pressure is maintained constant and nearly equal to the pressure in the nozzle.

This process was studied because it was claimed that SEDS gave the best control of system parameters. However, the thermodynamic, hydrodynamic and kinetic mechanisms resulting in particle formation are still not well understood. The effects of the nozzle dimensions and vessel dimensions on system performance had not been studies previously. In addition, little work has been published on the effects of variables such as liquid solvent and CO2 flow rates, solute concentration, temperature, and pressure on particle size and morphology.

A Design of Experiments (DOE) analysis was used to identify the more important process parameters that control particle size and morphology. DOE is a useful statistical tool to reduce the number of experiments necessary to find the most important variables at an early stage of experimentation. With DOE, a 512 full factorial run was reduced to 32 runs by confounding primary variables with higher order interactions (Example: concentration + temperature). The results of these experiments indicated that the most important factors in determining particle size and morphology are the concentration of acetaminophen in the solvent, the nozzle geometry (length of the mixing zone), pressure and temperature. These parameters were singled out for more detailed experiments aimed at determining the influence of these variables on particle size and morphology. A key feature of the experiments described in this thesis is the use of on-line monitoring of the acetaminophen concentration at the exit to the capture vessel in order to determine how the supersaturation of the solute varied with time during the process. In this way it was possible to determine the nozzle effectiveness in particle precipitation. In addition, the experiment performed in this thesis recognized that the SEDS process is in essence a batch process and it studied the effect of transients in co-solvent concentrations in the particle capture vessel on particle size and morphology.

In addition to SEDS, the precipitation of acetaminophen from ethanol was carried out using a Precipitation with Compressed Antisolvent (PCA) process, which is very similar to SEDS without the coaxial configuration. This system is simple to install and has been widely studied. The parameters that were important from the SEDS experiments were studied in the PCA to characterize their effects on particle size and morphology for this system. These results were compared to those obtained using the SEDS process. Both SEDS and PCA yielded equal particle size and morphology if designed properly. The major feature of this work was the emphasis on the design of effective nozzles for the PCA application. Similar to the SEDS results, a good mixing volume along with adequate residence time for micromixing are the best nozzle designs.