Reactor modelling reveals that pharma developers could use CSTRs to perform commercial-scale operations using lab-scale reactors.
BATCH manufacturing has long been the staple of the pharma industry due to the flexibility of operation and the simplicity of batch testing and release procedures. However, recent trends have seen a move towards continuous processing, which has benefits in terms of reduced investment costs, significantly reduced working capital, and shorter production lead times. While pharma research for continuous processing has typically focussed on plug flow reactors (PFRs), it is missing a trick by ignoring more common reactors from other industry sectors, such as the continuous stirred-tank reactor (CSTR).
We created an advanced process simulation within Aspen Plus to model and compare the outputs from PFRs and CSTRs, highlighting the impact of the reactor design upon reaction time, heat transfer, cost and practicality, demonstrating the significant advantages of the CSTR.
PFR vs CSTR
A plug flow reactor (PFR) is a tube or tube bundle where the reactants are continuously introduced at one end and the products continuously removed from the other end. The reaction conditions (concentration, temperature, reaction rate) vary along the length of the tube(s) as the reactants are consumed.
A continuous stirred tank reactor (CSTR) is an agitated vessel with a continuous feed of reactants and a continuous discharge of the reaction mixture (product). The feed and discharge rates are controlled to maintain constant reaction conditions (concentration, temperature, reaction rate) ensuring a consistent product stream is produced.
Our simulations model the Mannich reaction, an exothermic reaction producing an intermediate product, which is subsequently used in the synthesis of active pharmaceutical ingredients (APIs). The reaction is typically performed at low temperatures (10°C) to prevent the formation of side products, which makes the control of the exotherm more challenging due to the close approach temperature with the cooling medium (50% glycol coolant). As the reaction is an aqueous process we limited the inlet coolant to a minimum of 2°C to prevent localised freezing at the heat transfer surface.
The reaction also has a high conversion (95%), resulting in relatively long residence times and large reactor volumes to achieve the desired conversion, which present practical challenges to the reactor design, particularly for the PFR option. The general consensus among the pharma majors is that a capacity between 3–10 t/y will be sufficient for the majority of APIs coming through the development pipeline, so for this work we modelled a continuous rate of 1 kg/h, equivalent to 8 t/y on an 8,000 h operation.
The model investigates the relative performance of the reactors in terms of the control of the process temperature, and the residence times (reactor size) required to achieve the target conversion or capacity. For the CSTR we modelled both single reactors and reactors in series, whilst for the PFR we have modelled both single-tube reactors and multi-tube reactors.
Catch up on the latest news, views and jobs from The Chemical Engineer. Below are the four latest issues. View a wider selection of the archive from within the Magazine section of this site.