Microwave chemistry promises to improve the sustainability and carbon footprint of pharm production. Dr Jayne E Muir, Cambrex, looks at the recent development and benefits of production scale reactors
Fig. 1: The CaMWave KiloLAB continuous-flow microwave reactor
Microwave chemistry promises to improve the sustainability and carbon footprint of pharm production. Dr Jayne E Muir, Cambrex, looks at the recent development and benefits of production scale reactors.
The biggest questions surrounding microwave chemistry or microwave-assisted organic synthesis (MAOS) were when and if the technology would be scaled to commercial manufacture.
Increasingly popular in the laboratory and routinely used for drug discovery, the short reaction times make it ideal for route scouting and optimisation. But when would pharmaceutical manufacturing be able to enjoy the benefits of increased yields and higher purity products - in other words, to use MAOS to improve productivity, quality and ultimately lower costs?
There are several large laboratory microwave systems available, which use a variety of approaches to achieve kilogram scale; parallel vessels, stopped-flow and continuous-flow, for example, although some of these systems have inherent limitations.1 The commercial scale-up of microwave heating has been achieved in many other industries, including food, ceramics and mineral processing, but it has not proved easy in organic synthesis since the first academic publications in 1986.2
There are a number of reasons for this, not least that scaling up pharmaceutical reactions demands very versatile equipment and systems capable of bringing many different results and methods from the laboratory to the plant. Microwave energy also has a limited penetration depth defined by the material's dielectric properties; for most pharmaceutical reactions of interest this lies between 1cm and 10cm. Therefore, retrofitting micro-wave heating to large stirred vessels is not an option since only the outside layer would be heated as with conventional jacketed heating today. An in-pipe configuration is needed to scale up MAOS, which strongly suggests continuous processing.
Cambrex has more than 50 years of expertise in operating multi-tonne continuous processes, which are rapidly gaining momentum and acceptance within the pharma industry to address increasing cost, quality and control pressures. Thus this apparent "disadvantage" of microwave heating was really a "hidden" advantage. Expertise in microwave heating was then added by forming a partnership with technology company, C-Tech Innovation, whose 30-year expertise lies in the application of dielectric heating as a manufacturing solution.
Joint design/brainstorming discussions started at the end of 2006 to address the problem of scaling up MAOS as a continuous-flow reactor. The solution came from an interdisciplinary approach with chemists, physicists and engineers pooling ideas and experiences. The result was the idea for the CaMWave KiloLAB continuous-flow microwave reactor (Figure 1).
This is a tubular reactor of 6mm inner diameter. Fluid is pumped from feed vessels through a vertical section surrounded by a waveguide where it is heated by microwave energy before being collected in receiver vessels. The current materials of construction are acid-proof stainless steel and quartz glass in the microwave-heated section, where the tube needs to be transparent to microwave energy. A computer controls the reactor's temperature, pressure and microwave power and this data is logged. The current reactor specification is shown in Table 1.
Safety features include cut-offs on high temperature and pressure with bursting discs before and after the microwave-heated section. The waveguide is flooded with nitrogen during operation and will contain the reactor contents in the case of failure. The waveguide is further surrounded by a safety cabinet, which must be closed for operation and the reactor assembly surrounded by an extracted fume hood.
The first such reactor was built as a homogeneous reactor to prove the concept that heating a continuous chemical reaction by microwave energy was possible. A Hantsch dihydropyridine synthesis was chosen as the test reaction because these reactions are known to work well under microwave heating but also since suitable quantities of the starting materials were available on-site (Figure 2).
Starting reaction parameters were determined in our small-scale laboratory microwave reactor with a one minute reaction time at 150°C giving an acceptable 75% conversion. Optimisation was not deemed necessary for commissioning, so this reaction time was translated into a residence time within the microwave-heated section giving a flow rate of 20ml/min. In total, 7.8L was processed, from which 1.8kg product (68% yield) crystallised with good purity. The flow conversion was 76%, indicating that lab protocols can be easily scaled to kg quantities.
Following this success, the configuration of the CaMWave KiloLAB reactor was changed with the addition of a custom pressure control system that would allow the handling of slurries and suspensions. A heterogeneously-catalysed Suzuki test reaction was chosen as again these reactions are known to proceed well under microwave heating (Figure 3). This time, 2 mins residence time gave full conversion and 2L were processed yielding 128g product (98% yield) with excellent purity. It was subsequently discovered that only 30s was needed to give the same yield and purity, and potentially the reaction could go even faster.
With commissioning complete, final changes were implemented to give the CaMWave KiloLAB reactor production versatility. A further feed vessel was added to enable two streams to be simultaneously fed rather than having to pre-mix all the reaction components. An additional receiver vessel was also added, such that the two receivers can operate independently, allowing one to be emptied while the other continues to collect product. The reactor can, therefore, operate 24/7.
With this versatility in place, the reactor's potential as a viable manufacturing option was tested. The reactor was operated continuously for 32hrs giving 22.3kg of a Suzuki-coupled product after work-up (88% yield, >99% crude purity by HPLC) (Figure 4). This is the longest microwave-enabled reaction as a single run, published to date, which corresponds to an impressive 5m t.p.a. capacity for a 20ml laboratory reactor.
The reactor can undertake most reported MAOS and Cambrex is using the reactor to synthesise early kilograms of material for development and scale-up projects. A wide range of chemical reactions have been carried out, most being heterogeneous in nature up to 10 weight% solids. Palladium-catalysed reactions have been studied in depth along with nucleophilic substitutions and alkylations with long batch reaction times.
The high temperatures achieved with the reactor ensure very fast reactions. The shortest residence time to date is 15s at 150°C and a conventional 24hr reflux was bought down to 60s at 170°C. Reactions have also been carried out at the top temperature of 200°C. No new impurities have been observed to date, by either the high temperature or the switch to continuous-flow. In fact, impurities are usually greatly reduced such that downstream purification can be minimised or avoided.
Metal catalyst loadings can easily and often be dramatically reduced under microwave heating. Usually at least one order of magnitude reduction can be achieved. In one example, the conventional batch loading of 1,300ppm palladium was reduced to 5ppm with the same yield and purity of product.3
It is often beneficial to use water as the reaction solvent or to form part of the solvent mixture and it has been found that we can often substitute aqueous ammonia for methanolic ammonia. In some cases, reactions have been found to work best in the absence of solvent and even reactions that produce gases have been handled safely and performed successfully.
Cambrex is actively looking to switch internal commercial manufacturing routes to using the CaMWave technology. The CaMWave KiloLAB can manufacture around 25kg/day with larger reactors planned for tonnage production. A 2cm inner diameter CaMWave Pilot reactor, capable up to 200kg/day and handling higher weight% slurries, is already in testing.
The footprint of the reactors is relatively small as is the case with continuous-flow equipment in general. The current KiloLAB reactor fits into a large walk-in fume hood (3m x 2m x 1.5m) as will the Pilot reactor, and even the larger Plant reactor intended to manufacture up to 1,000kg/day will not be significantly larger. As such, these reactors make good-sized mobile units that can easily be added to current batch or continuous plants. Indeed large batch vessels can be used as convenient feed and receiver vessels for the larger reaction volumes. This plug-and-play concept could even be used to "recycle" currently under-utilised plants into becoming continuous-flow and microwave-heated facilities with very little investment.
The future will take the CaMWave technology and its reactors into cGMP and add automation and in-line analysis for process control. Quantification of the CaMWave technology against conventional batch and flow reactors has already started in terms of productivity, costs and energy efficiency. Part of this work is being supported by a grant from The Carbon Trust at C-Tech Innovation.4
Finally, to consider why the CaMWave technology won an award. First, it answered a long-running debate of whether microwave chemistry was scaleable. Second, the technology went even further to make continuous-flow applicable to handling slurries and suspensions. This opens up continuous-flow to virtually all reactions of pharmaceutical interest, well beyond the estimated 30% that are or can be made homogeneous.
The technology is a universal methodology, not product or process specific, allowing pharmaceutical manufacturing to catch up with other large chemical industries, well versed in the benefits of continuous-flow and microwave-heating to deliver productivity and quality benefits.
1 J. D. Moseley et al, Organic Process Research & Development, 2008, 12, 30-40
2 R. Gedye, et al Tetrahedron Letters, 1986, 27, 279-282; and 4945-4948.
3 The Suzuki reaction is not metal-free as control experiments without addition of catalyst have shown. Only 1% coupled product is formed. There is also no observed "conditioning" of the reactor walls with palladium metal.