Continuous Flow or Batch for API Manufacturing: When to Convert

Continuous flow is attractive when controlled mixing, heat removal or residence time solves a specific limitation in an active pharmaceutical ingredient (API) route, and the downstream process can use the resulting output. Batch may remain preferable for intermittent demand, difficult solids or established flexible capacity. The decision requires a molecule-specific quality, safety and economic comparison across the whole manufacturing process.

Reaction suitability and comparable API output

The starting question is which operation prevents reliable production at the required quality and volume. A fast reaction limited by mixing or heat removal may benefit from a compact flow reactor. A slow transformation needing a long hold can demand substantial reactor volume even in continuous operation. Flow chemistry is a manufacturing choice around reaction and transport behaviour; reshoring an active pharmaceutical ingredient does not by itself make that choice technically or economically superior.

A useful comparison keeps the product specification and manufacturing boundary constant. Compare isolated, acceptable API output rather than conversion measured at a reactor outlet. Improved selectivity may reduce purification work, but the result matters only if impurity control, recovery and final material attributes remain suitable. Establish a batch baseline that includes charging, processing, isolation, cleaning and quality activities. A flow experiment lasting a few hours should not be compared with an entire batch campaign while omitting the flow train’s start-up, shutdown and maintenance requirements.

Solids handling and downstream capacity

Solids often determine the equipment architecture. Precipitation, catalyst particles or fouling can disrupt narrow passages, alter pressure drop and shorten a campaign. The COSMIC project’s final reporting describes research addressing continuous processing of solids-containing fluids in intensified equipment, among other challenges. It is evidence of a substantial engineering problem and work to solve it, rather than proof that any slurry can use a microreactor. The process trial needs representative concentrations, impurity levels and run duration to reveal accumulation that a short clean-feed test misses.

Downstream integration can overturn an attractive reaction result. Solvent exchange, extraction, crystallisation, filtration and drying have their own capacities and operating rhythms. If the isolation equipment can handle less material than the reactor produces, a faster reaction alone will not increase saleable output. Intermediate storage can decouple operations but introduces hold-time and quality questions. A hybrid route, with the demanding reaction in flow and an appropriate isolation step in batch, may capture most of the benefit without requiring every operation to become continuous.

ICH Q13 controls and process safety

The control strategy must connect what enters the process with the material collected. ICH Q13, issued by FDA in March 2023, addresses process dynamics, material traceability and diversion, including residence-time distribution: the range of times material spends passing through the system. These relationships matter when a feed changes or a disturbance occurs. Instrumentation needs a demonstrated connection to quality and a strategy for data gaps. A stable temperature reading, or the presence of process analytical technology, does not automatically establish acceptable product or qualify real-time release.

Small reacting inventories can reduce some consequences of a process upset, but the safety assessment must include feed storage, pressure, blockage, cooling loss and downstream accumulation. It should address the actual chemistry and installed equipment through the site’s process-safety procedures. Likewise, extending run time can reveal fouling or catalyst changes that do not appear during development. Numbering up parallel units requires evidence for flow distribution and common utilities; replicating the nominal channel geometry is not sufficient to assume identical performance throughout the train.

Campaign economics and approval of a process change

Campaign economics depend on utilisation and demand. In an illustrative comparison, €300,000 of annual fixed cost contributes €15 per kilogram at 20,000 kg of accepted output, but €60 per kilogram at 5,000 kg. Those figures exclude variable costs and do not predict the price of a particular API. They show why a dedicated continuous system can lose its advantage when demand is small or uncertain. Add process development, controls, analytical methods, validation, cleaning, spare parts and supply continuity during transfer before comparing an investment with available batch capacity.

A conversion decision should therefore follow a representative campaign demonstrating acceptable product, controlled disturbances and credible operating duration, with an agreed route through the pharmaceutical quality system. ICH Q13 explicitly addresses conversion from batch to continuous and the need to seek regulatory approval before implementing conversion of an approved process. The architecture does not automatically establish compliance with good manufacturing practice (GMP). Proceed where the specific process improvement, qualified capacity and total cost justify the development and transition work; retain batch or a hybrid arrangement where those advantages have not been demonstrated.