This thesis is focused on investigating the scale-up of the roll compaction process by firstly considering the effect of plastic/brittle mixtures and process parameters, secondly, the impact of the change in scale on ribbon properties, and finally, the application of models to try to successfully scale the process up. Seven binary mixtures of MCC (plastic material) and mannitol (brittle behaviour) were investigated. The blends were roll compacted at different conditions of specific compaction force, gap width and roll speed determined by a design of experiments. A total of six compactors organized in three couples of different scale depending on the supplier were investigated, what means a total of three individual scale-up studies. 

In the first part of the thesis, the effect of varying the fraction of plastic/brittle material (variation of the proportion of MCC) was evaluated together with the process parameters, by characterizing ribbon density and microhardnes, as well as the granule size distribution. The results showed that the ribbon relative density decreases linearly when increasing the MCC fraction. However, when plotting the microhardness or the granule properties (expressed as the percentiles D10, D50 and D90 as well as the amount of fines) against the proportion of MCC, a non-linear relationship is observed. The percolation theory was applied in order to investigate this fact deeper. As conclusion, it was proven the importance of the MCC/mannitol fraction on the product properties. In the second and main part of the thesis, 3 individual scale-up studies were performed: Gerteis, L.B. Bohle and Freund-Vectors line, classified according to the compactors’ provider. Ribbon relative density was characterized for all cases. For the Gerteis and L.B. Bohle studies, it was generally concluded that that the higher the specific compaction force, the lower the gap and content of MCC, the denser the ribbons. However, the compactors scale or its interaction with another factors was also affecting the results proportionally or inversely. MCC compacted at the L.B. Bohle machines was the exception. For the Freund-Vector line, a general tendency of the large scale to have denser ribbons was observed, although unexpected as for the same conditions, the small scale applies a higher roll force and densifies the product more. Therefore, in this second chapter, the importance of the scale and the difficulties to transfer the process has been confirmed. Finally, in the third chapter, the modelling of the roll compaction scale-up was investigated in order to successfully transfer the process between scales. After testing several approaches, it was concluded that the best one is the mechanistic model proposed by Reynolds et al. 2010, which gives good fitting and predictions. 

The scalability of the roll compaction process for some couples of compactors available on the market has been investigated, by identifying the most important parameters affecting the results and adapting them through the application of a mechanistic model proposed by Reynolds et al.