(343g) Prediction of Density From the Uniaxial and Roller Compaction of Viscoelastic Pharmaceutical Powder Mixtures

Roller compaction is a pharmaceutical unit operation typically scaled-up and controlled through trial and error to produce compacted powders with specific qualities. The goal of this research is to develop a methodology to predict final roller compact properties based on raw material properties and process conditions for use in scale-up and control. The objective of this work was to develop a methodology to predict the density of roller compacts of Microcrystalline Cellulose (MCC) with Acetaminophen (APAP), from the rheological properties of the raw material and the processing parameters. Moisture sorption isotherms were determined for pure powders and their mixtures at 25°C, 35°C,45°C and 55°C from 0-90% RH. Uniaxial compacts were created with an Instron Universal Testing Machine to simulate roller compaction ribbons. Compression curves at rates of 0.5,2,5 and 10 mm/min and their Scanning Electron Microscopy images were obtained to further describe particle deformability. The linear viscoelastic region was determined using a 3 point bending tool in the DMA. Each compact was tested via dynamic strain sweep at a frequency of 0.5 Hz and temperatures of -60°C, 25°C, and 80°C. The viscoelastic nature of MCC and MCC-APAP mixtures as viscoelastic solid were determined through stress relaxation experiments and Maxwell models. Both compaction and stress relaxation experiments were conducted at moisture contents corresponding to relative humidities of 40% and 70%, temperatures of 25˚C and 55˚C, APAP composition from 0-60% and compression pressures of 16-55 MPa. Roller compaction studies were performed for samples at the same moisture contents and compositions as above, roll speeds of 8-12 rpm and hydraulic pressures from 7.5-12 MPa. Viscoelastic parameters from compaction and Maxwell models, and average powder particle size were used to model the viscoelastic contact area between two deforming particles in the system. Squeezing flow studies with pre-compacted powders were carried out to measure the shear properties. Moisture sorption isotherms showed very good fit with the GAB model (R²>0.99) with the amount of moisture adsorbed decreasing with an increase in APAP. Compaction and stress relaxation data showed a good fit with the respective models (R²>0.99). Statistical regression models were developed for the dependence of compression curve and Maxwell model parameters on the process variables. A strain of 10^-3% was determined to be in the linear viscoelastic region of all samples. The calculated contact area increased with increases in applied pressure and moisture content while a decrease was observed with an increase in APAP content and temperature. The viscoelastic contact area was coupled with process parameters from uniaxial and roller compaction in statistical least squares regression models to predict experimentally determined final compact density, with a correlation coefficient of R²> 0.90 at 90% confidence. Lower contact areas were reflected by lower uniaxial and roller compact densities. These models showed great improvements over a model without contact area and another one with elastic contact area. Statistical regression models for compact density based on viscoelastic contact area and composition of APAP were developed. It was shown that the density of roller compacts could be predicted fairly accurately based on uniaxial compaction experiments. The density of roller compacts was higher compared to uniaxial, presumably due to pressure and shear effects, which are absent in unaxial compaction. This study also showed that viscoelastic properties (influenced by raw material properties) in conjunction with process parameters have the potential to predict density of powders compressed using a uniaxial/roller compaction process.