Four scaling criteria were used to scale-up the precipitation with a cornpressed-fluid antisolvent (PCA) process from the laboratory to the pilot plant scale using an injector with a confined mixing chamber. The scale-up criteria consisted of maintaining either: (1) constant Reynolds numbers, (2) constant axial velocity, (3) constant residence time and suspension density, or (4) a constant energy dissipation rate, as process scale was changed. Experiments were conducted in a semi-continuous PCA flow apparatus and the biodegradable polymer poly (L-lactic acid) (PLLA) was used as the model solute. Successful scale-up was based on maintaining measured particle size distributions for PLLA microparticles as process scale was increased. The scale-up criteria of maintaining constant Reynolds numbers or constant axial velocities were not sufficient to ensure equivalent process performance during scale-up. Scale-up at constant Reynolds number resulted in a 40% increase in the average PLLA diameter with a feed concentration of 0.125 wt% PLLA, and a 36% increase in the average diameter was calculated when the feed concentration was increased to 0.25 wt% PLLA. Scale-up maintaining a constant axial velocity resulted in a 32% increase in the average PLLA diameter with a feed concentration of 0.125 wt% PLLA. The shift in the PLLA particle size distributions towards larger size fractions with this criteria was attributed to a low energy dissipation rate in the confined mixing chamber, which directly affected mixing quality and secondary nucleation mechanism(s) in the confined mixing chamber. Scaling the injector with a constant residence time and Suspension density, or a constant energy dissipation rate resulted in similar PLLA particle size distributions at each scale of operation, and was attributed to maintaining comparable mixing quality and similar precipitation kinetics with a change in process scale. (c) 2005 Elsevier B.V. All rights reserved.