FLOW CHEMISTRY

200 20 99.4 97.2

200 5 74.8 73.9

210 10 97.2 94.4

210 5 83.4 85.6

Figure 3 – Case study in development : Predictions v . results

‣ batch . “ Basically , we took all of our batch data and some of our process safety data and used DynoChem to create a predictable model ,” said Oliveira . “ We also accounted for the pH equilibria in the reaction , which justified the use of excess reagent in the process .” Due to safety concerns , the batch reactions that were used to build the model required the mixture to be cooled before sampling . This is not an ideal procedure to generate results for a kinetic study , “ but as long as the modelling software accounts for these temperature changes , it should be fine – and that was the result . The model actually picked up the experimental points very well .” The lab set-up used to screen the reactions from microlitre- to litrescale comprised two feeds , a pressure sensor , heating and cooling zones and a pressure regulator . At 195 ° C , the top temperature at which the set-up could operate , a yield of 92 % was achieved in 20 minutes of residence time . “ This is not the best but it ’ s not too bad either for a starting point ,” Oliveira said . Before scaling the process up , the team conducted a series of experiments , involving a step-wise increase in temperature , adjusting the pressure , increasing the residence time and mixing , in order to add to its knowledge . The results were close to what had been predicted and the model picked up the results of the actual experiments quite well .

Based on this , it was determined that the process was feasible in flow . The process was then transferred from a 20- µ l chip reactor to a 5-ml coil reactor , which meant a 250-fold increase in throughput . The results ( Figure 3 ) were close to what the model and the experimental results had predicted . “ At this stage , we were quite happy because one of the conditions was already at the same performance level as the batch process . Without spending too much time in the lab and doing a lot of experiments , we were able to get a full process that was equivalent to the batch process .” With the proof-of-concept to hand , Hovione moved on to a development plan . The aims were to : increase productivity by reducing residence time ; complete the integration of process analytical technology to improve its knowledge ; tune the temperature and pressure ; and reduce the excess of reagent , which was still similar to that in the batch process . Using this approach , the company ran another case study , which Oliveira discussed briefly . A 60-hour batch reaction that ran at 30 ° C using sodium methoxide as a catalyst was successfully converted to flow at 120 ° C , with no catalyst required and in ten minutes of residence time . The yield is already good and the excess reagent was reduced by 18 %. In a third study , a cryogenic reaction that took place over two hours at -50 ° C in batch was converted to -35 ° C in one minute in continuous flow . Continuous conditions were also safer , required less energy – because there was no longer any need to cool a large reactor - and saw a 30 % improvement in yield . In a final example , a batch reaction generated a large volume of overreaction impurities and thus required an extra chemical step to remove them . In flow , Hovione was able to reduce the impurity level from the first reaction and avoid the extra step , thus telescoping the whole process .

Conclusion

“ Continuous flow is clearly an alternative to batch , thought it is not necessarily the answer for everything ” Oliveira said . “ Its main features come from the good control of process parameters and the resulting safety benefits . Yield and efficiency gains are possible , though these are not generally the main driver .” “ In these examples , our development workflow was designed for fast implementation to get meaningful data and a data-driven decision to go or not to go with continuous flow , based on the process understanding we gained during development .” •

Isabel Pina

DIRECTOR EXTERNAL COMMUNICATIONS

HOVIONE k + 351 21 9829000 J ipina @ hovione . com j www . hovione . com

16 SPECIALITY CHEMICALS MAGAZINE ESTABLISHED 1981