Continuous is the industry’s buzzword right now, but flow chemistry has long been used in the chemicals sector. Is this the unsung hero of continuous processing and why has it taken pharma so long to realise its potential? There are many reasons why pharma is behind the chemical and other industries in the adoption of continuous processing, notes Sam, including the following:
The bulk chemical sectors (such as sugar and fertiliser) have been using flow chemistry for decades. These industries deal with high volume, low price products, and rely heavily on productivity to provide the required volumes — which is exactly what flow technology does. By contrast, the pharmaceutical industry requires low volumes of high quality products, meaning that productivity is important … but not such a high priority. As a result, flow has not been at the top of their agenda.
The majority of facilities making bulk chemicals have been manufacturing the same products for decades. For these industries, flow systems can be implemented once and used for many years. But for pharma, with its frequently changing products, many flow processes would need to be custom developed. On the surface, that makes the process a less attractive proposition.
In addition, Big Pharma traditionally relies on organic chemists who are trained in the development of batch processes. The inclusion of chemical engineers in product development is a relatively recent concept (going back 20 years or so). Medicinal and early stage chemists are hardwired to establish proof of principle in the shortest possible time. By the time people need to think about commercial-stage productivity, it’s almost too late to make major changes ... such as implementing flow.
The pharma industry as a whole, including the regulators, takes time to adjust to new approaches; it’s conservative and risk averse. The transition from batch to flow, even when supported by justifiable drivers, involves resolving many quality assurance and control issues, as well as taking a novel route to process development. Leading companies in flow chemistry view these issues as a challenge to be overcome … and are solving them one by one. Yes, some companies do share their experiences, but the majority keep their cards very close to their chests. This culture of not sharing knowledge is a hindrance to the widespread implementation of flow.
When asked about the best time to implement flow chemistry, Sam explains that it should be part of the commercial process. The latest time to initiate flow for a GMP step, he adds, is before registration. Flow chemistry is associated with a number of processing advantages, such as increased safety, improved product quality, cost efficiency and flexibility, but what about data acquisition (reaction rates, endpoints, etc.) opportunities and Process Analytical Technology (PAT) compatibility? During the early phases of development, a deep understanding of the process and the relationship between inputs/outputs is useful … but not critical, says Sam. By contrast, a tight timeline and/or lack of starting materials are potential issues that could exacerbate the process. From this aspect, it’s very similar to batch.
However, as a project advances to the later stages, information such as kinetic data becomes increasingly important. PAT is the right tool to capture kinetics, notes Sam; it facilitates analysis and helps to optimise the operational parameters. “In one of our flow projects, we measured the purity of the product using online technology, which was designed and set up in such a way that if the online HPLC system indicated that the product quality was out-of-specification, it was redirected to waste. This, of course, helped to maintain the quality (and purity) of the product. So, although PAT can be used to capture kinetic information in the later stages of a project, it’s also a useful tool to monitor early phase projects,” he says.
The pharma industry is just beginning to use PAT in flow production. Although PAT is now used more as an in-process control in batch processes, it is more amenable to flow than batch. The reason for this is that flow systems are steady state (variables don’t change with time) and the Critical Quality Attributes (CQAs) need to be maintained within a narrow range to guarantee quality. In batch lines, quality is measured at the end of the process and, for that, offline measurement may be sufficient. However, the application of PAT in flow processes requires certain issues to be resolved — such as online equipment qualification. As a caveat, it should be noted that even when PAT is applied to a flow process, it is used as an In-Process Control (IPC) tool and the product should still be sampled at end-of-production using standard offline analytical methods.
For companies already running batch-based reactions, could converting to flow chemistry result in commercial or other benefits? Sam explains: “Of the six technical drivers for flow —process safety, short residence time requirements, unstable intermediates, scale-up issues, high temperature/high pressure needs and productivity — the first five apply to both early and advanced stage projects. However, from a productivity point of view, continuous flow processing is more beneficial to late- and commercial-stage projects than early ones.”
The chemical route to a final API may comprise 10 distinct steps. Combining these with unit operations such as distillation, crystallisation, filtration, washing, drying, quenching, etc. adds to an already complex process. In batch mode, product derived from each step needs to be dried and stored until the plant/facility for the next stage is available. Continuous flow chemistry uses the intermediate in the subsequent “step” as soon as it’s produced. This has two immediate impacts: first, production times are significantly reduced because everything works in parallel; second, inventory management is almost eliminated because there’s no need to store the intermediate.
Sam agrees that switching to flow for a GMP step at the commercial stage would be difficult, time consuming and expensive … but may still be justifiable if quality and safety are potential concerns. Here, flow chemistry can be used as an enabling technology to resolve probable issues as opposed to simply cutting costs. It might perhaps be necessary to start over and repeat the commercialisation steps, he concedes, and there may be problems dealing with changes in the purity profile. “It’s a difficult question to answer, but with time and investment, anything’s possible. We recommend including flow as part of the registration batch. For non-GMP steps, flow chemistry and the associated changes don’t require regulatory approval if the quality of the starting material in the first GMP step is not affected.”
Discussing the downside
In a rather more quickfire part of our discussion, I postulated that the literature cites a number of disadvantages related to flow chemistry, to which Sam responded, such as the following:
Q: Dedicated equipment is needed for precise continuous dosing, such as pumps, connections, etc.
A: The measurement and control of the feed rate are fundamental requirements in flow chemistry. This can be easily done using a mass flowmeter and a PID control loop. We have multiple feeding platforms at WuXi STA and we have never considered this to be a problem. The chemical industry has been using these monitoring and control systems for a few decades, so I don't understand why it’s considered to be a disadvantage in pharma. Regarding the precision of the pumps, I think that if a reaction can be run in batch mode, it shouldn’t be sensitive to back-mixing and can therefore be run in flow mode with a reasonable control system.
Q: Start-up and shutdown procedures have to be established (for flow but not batch).
A: It’s correct that continuous processes involve these steps. For example, to do a high temperature reaction in a perfusion flow reactor (PFR), the unit needs to be heated with pure solvent until the set point is reached before starting; otherwise, valuable material would be lost during the warm-up period. Once the required temperature is reached, then the actual feed stream is pushed into the reactor. So, yes, there’s a delay as the process ramps up. Similarly, at the end of the run, reaction solvent is pushed through the reactor to clean it; at the same time, however, this does increase the recovery yield by collecting material left inside the reactor. These are additional requirements compared with batch … but certainly not disadvantages. It’s not a particularly complex activity and not very time consuming. Often, flowing the solvent for 3–4 residence times is sufficient for these procedures.
Q: The scale-up of micro effects such as high area to volume ratios is not possible and economy of scale may not apply.
A: For an early stage project and because of the low volume requirement, the same or slightly larger microreactors can be used for both lab and production applications. For commercial projects, microreactors with larger channels are required to accommodate higher flow rates, but the increase in channel size is not significant. Regarding economy, the cost of the reactors at the commercial stage is not a significant factor as utilisation rates are high. Plus, the level of investment is very much related to the product or project; some flow reactors are relatively inexpensive.
Q: Safety issues for the storage of reactive material still apply.
A: Because the intermediate is immediately consumed in the next step as soon as it’s produced, there’s no need for storage in continuous flow manufacturing. Therefore, any “energetic” material can be produced in flow and immediately used in flow. This is a significant contribution to reducing process safety risk when dealing with energetic material and, it could even be said that continuous flow mitigates the safety risk of processes and intermediates.
Offering advice for companies developing an API or looking to use flow chemistry to expedite their time to market, Sam recommends that, for early phase projects, interested parties should carefully evaluate the synthetic route and consider flow whenever a driver exists. Key examples are
- the batch process is not safe
- the intermediate is not stable
- long reaction times that generate impurities
- engineering problems such as two immiscible phases leading to scalability issues
- high temperature reactions (beyond boiling point).
In conclusion and pondering — once industry fully appreciates the benefits of the technique — the advantages it can bring to drug manufacturers in the short- and long-term, Sam says: “We understand that implementing flow requires a number of challenges to be resolved. Once dealt with, however, there are benefits to be gained. In the short-term, the process offers better quality and product safety. In the long-term, however, there are environmental and economic advantages.”
“Companies need to recognise the advantages of flow mode operation. There are some exciting possibilities out there, such as high temperature–high pressure reactions involving azide chemistry that will enable chemists to design shorter synthetic routes that will deliver significant cost savings when the project reaches the commercial stage. There’s a lot to be gained by going with the flow.”