Carbon cartridges and their use as a purification step in pharmaceutical API processes
Chemical Product Design and Engineering (CPD&E)
Chemical Product Design & Development - IV (CPD&E - 4)
Keywords: Carbon cartridges, carbon treatment, impurity removal, colour removal, scale-up
In the pharmaceutical industry powdered carbon is frequently used to remove impurities, both those present in unacceptable percentage amounts and those present in almost undetectable amounts, but still affecting the colour of the API. The handling of powdered carbon is problematic due to its dusty nature and cleaning of carbon from the equipment can be difficult. As a result, carbon cartridges have been developed. The carbon is immobilized in a cellulose matrix and made into filter discs, which are contained in a jacketed filter housing. After a flush to remove “extractables” (material leached from the cellulose matrix), the process fluid flows through the filter discs and the impurities are adsorbed onto the carbon. The cartridges are disposed of after use. This effectively removes the charging of dusty, powdered carbon, the stirring operation, the filtration of carbon particles and the cleaning of carbon particles from the process. Carbon cartridges are also more effective than powdered carbon since the process fluid is transferred through immobilized carbon, which essentially acts like a carbon bed with gradual build-up of impurities as the process fluid passes through. This means that as the process fluid passes through the bed it gradually becomes cleaner and the bed it comes into contact with will also contain fewer impurities, allowing a higher removal of impurities than powdered carbon.
Carbon cartridges from suppliers CUNO and PALL have been tested for use in two pharmaceutical API processes. In process 1, the goal was to reduce impurity A from ~3% to below 2%. In process 2 the goal was to reduce a low-level, unidentified, red coloured impurity B to such levels that the API was judged as white to off-white.
For process 1, a suitable carbon type had been selected. The process was run using two CUNO lab scale filters (~80ml and ~8L Process Volume) and impurity A and product concentrations were measured at the CUNO outlet at intervals during the transfer. We were able to see a breakthrough curve and a mass balance was completed for the CUNO transfer step to quantify impurity A adsorption. Using these results, the process was scaled up to a ~300L manufacture. We proved, experimentally, that the flush was successful and “extractables” did not end up as product impurities. We found that we were able to scale up the process based on constant Process Volume / CUNO Area (effective disc surface area) and that lowering the product concentration whilst maintaining a constant impurity A concentration led to increased impurity A adsorption.
For process 2, only one type of powdered carbon had been tested and because of that, discs with different carbon types from both CUNO and PALL were tested. The colour reduction varied between different discs, but the best disc from each supplier had equal colour reduction.
Large-scale filter housings have been purchased for both processes and will be used during manufacturing campaigns. For process 1, the filter housing is a CUNO (one or two 16”, 14 disc cartridges, 3.2 or 6.4 m2 CUNO Area) and the large-scale process volume is ~0.3 m3. For process 2, a four-module 16” jacketed PALL filter housing was chosen and the large-scale process volume will be 4 m3. The results from these campaigns will be included in the full paper.
See the full pdf manuscript of the abstract.
Presented Thursday 20, 12:12 to 12:30, in session Chemical Product Design & Development - IV (CPD&E - 4).
