The aim of this work was to develop a quantification method based on image analysis, able to characterize sticking during pharmaceutical tableting. Relationship between image analysis features and relevant mechanical parameters recorded on an instrumented tablet press simulator were investigated. Image analysis, based on gray levels co-occurrence matrices (GLCM), generated textural features of the tablet surface. The tableting simulator (StylcamA (R) 200R, Medelpharm), instrumented with force and displacement transducers, provided accurate records. The tablet defects and compaction process parameters were studied using three pharmaceutical powders (Fast-FloA (R) lactose, anhydrous EmcompressA (R) and AvicelA (R) PH200 microcrystalline cellulose), five compression pressures (60 to 250 MPa), five lubricating levels, and three types of punches (standard steel, amorphous hard carbon and anti-sticking punches). Texture parameters made it possible to quantify with precision tablets' aspect. The selected parameter IC2 (Information on Correlation 2) plotted versus the ratio between the ejection shear stress (Esh) and the compression pressure (Cp) let appear a relevant knowledge space where it was possible to identify a normal and a degraded tableting mode. A positive link between those two parameters was shown. Since the Esh/Cp ratio was related to image analysis results, it proved to be an interesting defect tag.
