(586h) Operational Aspects Of Continuous Pharmaceutical Production | AIChE

(586h) Operational Aspects Of Continuous Pharmaceutical Production


Mitic, A. - Presenter, Technical University of Denmark (DTU)
Gernaey, K. V., Technical University of Denmark
Johansen, K. D., Technical University of Denmark
Skovby, T., Lundbeck A/S


Organic synthesis is essential for the production of an important class of pharmaceuticals. The implementation of organic chemistry on an industrial scale is a great challenge. To date, pharmaceutical manufacturing is mainly based on batch and semi-batch processes. However, besides flexibility and versatility of batch vessels there are many disadvantages. The main problems are linked to long reaction sequences, the occurrence of non-uniform conditions inside the vessels, the limited potential to apply process control and as a consequence difficulties in automating the production, and complicated cleaning procedures. Additionally, Process Analytical Technologies (PAT) cannot reach the full benefits in this type of production. Therefore, a solution to those problems might be moving to continuous production mode and establishing in-line process monitoring and control.


In theory, continuous production conserves energy and natural resources leading to more eco-friendly and economical processes. Furthermore, due to the high degree of automation and process controllability of continuous processes, PAT may be implemented in a very efficient way. However, when switching an organic synthesis-based process from batch to continuous operation, it is important to realize that not all reactions are suited to be operated in a continuous mode. Slow reactions, for example, result in operational problems when operated in continuous mode.

Therefore, acceleration of slow chemical reactions is the main focus here in order to incorporate also this type of reactions in a continuous process efficiently. As an example process, the synthesis of zuclopenthixol – a product of H. Lundbeck A/S – is studied, and more specifically one production step – the dehydration of “Allylcarbinol” (9‐Allyl‐2‐Chlorothioxanthen‐9‐Ol) to cis/trans - “Butadiene” (cis/trans - 9H‐Thioxanthene, 2‐chloro‐9‐(2‐propenylidene)‐(9CI)).

Despite acceleration and continuous manufacturing, increased stereo-selectivity is needed because just the cis-isomer will lead to the medically active product. Hence, the cis isomer of “Butadiene” is a precursor to cis-clopenthixol (zuclopenthixol) via the consequent hydroamination reaction with 1‐(2‐Hydroxyethyl)piperazine. Trans isomers are separated afterwards in a sophisticated and expensive way and will undergo the isomerization procedure.

Results and discussions

Zuclopenthixol synthesis could be divided in two parts. The first part involves fast and exothermic chemical reactions, such as Grignard reaction and hydrolysis whereas the second part includes slow and endothermic chemical reactions – dehydration and hydroamination. The main focus here is on the dehydration reaction which is traditionally performed in batch mode in the Lundbeck A/S company. Acetic acid anhydride and acetyl chloride are used as chemical catalyst, as well as toluene as a solvent.

However, long reaction times and low selectivity influenced the development of new production processes. Furthermore, the previous reaction steps (Grignard alkylation and hydrolysis) are performed in THF and thereby avoidance of the solvent swap procedure was a desired choice. Hence, the simplified synthetic route to cis/trans “Butadiens” is proposed involving usage of Brønsted acids, such as hydrochloric or sulfuric acids. The first tests were performed in batch mode leading to the total conversion of “Allylcarbinol” and producing a mixture of equal amounts of cis/trans “Butadienes”. As a result, a cheaper and a very efficient synthesis step was developed, but the reaction time was still relatively high in view of performing the chemical reaction in continuous mode.

Hence, following the Arrhenius equation led to the conclusion that doubling of the reaction temperature could lead to a significant acceleration of this chemical reaction. As a consequence, a high-pressure process was developed by increasing the boiling point temperature of THF to 120⁰C. For this purpose, a back-pressure regulator was used. The reaction time was consequently decreased from 2h to just 3 min. Furthermore, in-/at-line process monitoring was established by using a NIR spectrophotometer, as well as off-line analysis by applying HPLC. 

Despite development a fast manufacturing procedure, yield of the desired product is still low. Therefore, an increase of the fraction of the cis-isomer is necessary. For this purpose, different Lewis acids and bases were tested, as well as special dehydration agents. As result, an increased amount of the cis-isomer was achieved by using thionyl chloride and triethyl amine as Lewis acid and base, respectively. The yield was increased from 50% to 59% of the desired “Butadiene”,

However, by switching dehydration agent the reaction conditions were drastically changed due to very low reaction temperatures (-68⁰C) and longer reaction times – up to 2h. Temperature increase in this case leads to an increase in the reaction rates, but has the disadvantage of causing a significant decrease of the yield of the produced cis-isomer.

Conclusions and future perspectives

Acceleration of the dehydration reaction in the zuclopenthixol production was performed in a successful way. Hence, the reaction time was decreased from 2h to just 3 min, while obtaining the same composition of the products. However, a complicated downstream processing of this process led to the investigation of new synthetic routes which would increase the amount of desired cis-isomer. Therefore, instead of using Brønsted acids, suitable combinations of Lewis acids and bases were tested. The results show increased yields of the desired product, but further investigations in this direction are necessary. As a final step, establishment of a control strategy for the most suitable chemical process has been done as well.