When Less is Much, Much More

Aug. 8, 2006
Microreactors are challenging the concept of "economies of scale."

Are microreactors heading to a pharmaceutical facility near you? Smaller alternatives to traditional batch reactors are already being used commercially by specialty chemical companies, while the world’s leading drug companies are all using microreactors for R&D.

Compared to traditional batch reactors, the smaller devices can tolerate extremely high temperatures and pressures, above 650ºC and 25 bars. Microreactors can also use reagents that are between five and 100 times more concentrated — 5 M is common — than the solutions used for solution batch reactors.

Without altering fundamental reaction kinetics, microreactors permit extremely exothermic reactions to proceed smoothly, allowing chemical reactions that usually take hours to be completed within seconds. Volker Hessel, Ph.D., vice R&D director for the Institut für Mikrotechnik Mainz (IMM) in Mainz, Germany, knows of a 1-mL reactor that can produce 50 kg of nitroglycerine per hour.

New microreactors, incorporating innovations from microchannel analytical devices, can run chemical reactions in reaction volumes ranging from microliters to about one liters. Micro bioreactors are even being used to run biopharma fermentations (see "Invitrogen’s Push in Bioprocess Intensification," July/August 2006).

Scaling up microreactors is usually simple. Manufacturers can either use more of them or adjust the temperature, flow or reagent concentrations.

And they are already being applied to scaleup and manufacturing. Bayer Technology Services (BTS), the contract services subsidiary of Bayer AG (Leverkusen, Germany) has rapidly incorporated microreactors into its businesses, and offers the technology to clients as well. “There is a lot of interest in microreactors in pharmaceuticals,” explains Thomas Daszkowski, PhD, director of process technology, because of the easy scaleup from lab through development and manufacturing. Bayer runs one production-scale process with microreactors, and has “many more” in development.

Benefits of microreactors include:

  • Fewer side products;

  • Rapid, straightforward scaleup;

  • Shorter development time and time to market;

  • Less down-time; greater capacity utilization;

  • Minimal hold-up volumes and reagent/product degradation;

  • Safer handling of unstable, even explosive reagents and intermediates;

  • Shorter reaction times and improved quality, yield and selectivity through more efficient mixing and heat transfer (due to microreactors’ high surface-to-volume ratio);

  • Rapid analytics (often built into the reactor);

  • Lower reagent use, operator costs, material and solvent requirements, waste disposal and capital costs;

  • More manageable investment-related risks for construction or acquisition of new plants, facilities, equipment and personnel;

  • Easier process scaleup by adding more microreactors or by adjusting process temperature, flow or reagent concentrations.

Within its own chemical synthesis needs, Bayer first concentrated on highly exothermic reactions or situations where rapid mixing was required, and has moved forward from there. Allowing this to happen was its 2004 acquisition of Ehrfeld Mikrotechnik (Wendelsheim, Germany), one of the first microreactor companies to exploit the range of microfabrication, microactuation and microsensing technologies that had just emerged from European academic and government labs.

At the time, BTS believed that microreactors would help its chemical manufacturing capabilities become more competitive; the company soon realized that it was onto something. “We’re not talking about 500 kilograms of product any longer,” Daszkowski says. “Reactors with eight liters per hour throughput can process 80 tons of material in a year.”

Customizing and controlling reactions

Reactors’ high temperatures and short reaction times allow development scientists to tweak reactivity and side product formation as never before. For example, using traditional reactors, alkylations of multi-functional nucleophiles often proceed beyond the mono-alkylation stage. Chip-based processes can, instead, stop after one alkylation, minimizing side products as well.

“Mixing and heat exchange are easier, especially for exothermic reactions, and it’s possible to obtain higher selectivity and yield, leading to fewer downstream steps,” notes Barbara Pieters, an analyst with Yole (Lyons, France), a market analysis firm specializing in micro- and nanotechnology. For complex, multi-step processes, porting the entire process to microreactor format is unnecessary. “Just one step can make a difference,” Pieters observes.

Typical microreactor channels measure between 50 and 500 micrometers across. Millimeter-sized channels are possible, but mixing and heat exchange advantages dissipate as the fluid path widens.

The location of material flowing through the microchannel represents a time point in a reaction’s progress. Operators can therefore control the amount of heat pumped into the system with great precision, and either quench by cooling as soon as the reaction is complete or by changing the pump speed.

In sharp contrast, in batch reactors, reaction time equals how long the heat is left on. Product sits around “waiting,” often at reflux temperatures, for reagents to be consumed. “Everything is happening at the same time,” observes Mark Gilligan, managing director at Syrris (Herts, UK), a five-year-old company that specializes in productivity tools for R&D chemists and financed its growth through partnerships with pharmaceutical companies, especially GlaxoSmithKline, a shareholder.

Syrris’ Africa microsynthesis modules come in two-channel, four-channel, and HPLC-integratable analysis configurations. These microreactors can interface with a liquid-liquid extraction module for rapid purification.

Syrris focuses on development-worthy microreactors, preferring to leave systems for large-scale manufacturing to other companies. Nevertheless, the Africa systems pump out 100 grams per day of product, which can easily support preclinical and early clinical testing of pharmaceutical compounds.

According to managing director Mark Gilligan, the top 10 pharma companies all use Syrris microreactors for discovery, development, compound library generation, and whenever they need milligrams to grams of material quickly. Since Syrris microreactors target R&D, and development chemists prefer to use the same platform as they scale up, one might ask why anyone would wish to develop a process in a microreactor, only to switch to another firm’s production reactor or even to a batch process. Speed is one reason. “A microreactor allows you to get material quickly,” Gilligan observes. Process conditions and yields from microreactions represent the best-case scenario for batch reactors – something to shoot for.

The Fraunhofer Institute for Chemical Technology (ICT; Pfinstal, Germany) employes about a dozen scientists in the development and scaleup of highly exothermic reactions in microreactors. ICT collaborates with industry on scaleup, process analytics and process intensification through microchemistry. Photo courtesy of ICT.

Since heat transfer and mixing don’t scale linearly with reaction volume, batch scaleup requires continual tweaking and optimization. Where batch processes tend to be dilute to prevent mishaps like runaway reactions and formation of side products, microreactors can handle much more intense, concentrated conditions. “We run reactions at up to five molar, and still achieve excellent control over reactor parameters like heat dissipation,” says Paul Watts, Ph.D., chemistry lecturer at the University of Hull (Hull, UK). Care must be taken not to push concentrations too high lest product precipitate within the channels, a problem Dr. Watts described as “massive” for some reactions.

Reactor chips developed at Hull are made from glass channels, a few centimeters long by one centimeter wide and two millimeters thick, with a reactor volume of less than 100 microliters. Glass has become the preferred material since it is inert, reusable and most closely mimics flasks. Throughput is about 1g/hr of finished product. Yield and purity are enhanced by the use of solid-phase reagents built into the channels and stoichiometric quantities of reagent whenever possible. Users can often avoid purification between steps if everything works perfectly. Watts claims his group has never failed to find a microreactor process for any reaction, but admits some transformations require fundamental chemistry changes. Simply transferring a batch reaction to microreactors is tempting, but often fails.

In-line process control, a feature in which industry is keenly interested, is possible by making channels optically transparent. Hull uses Raman spectroscopy, which is less frequently used in process analytics than infrared, but makes sense considering these microreactor channels are made of IR-opaque glass.

Watts believes that the engineering aspects of early-stage microreactor work are overblown. He prefers that microreactors function as black boxes, similar to analytic instrumentation. “If you put chemicals into reservoirs and pump them through, you expect to get your product. Engineering should be the domain of microreactor manufacturers, not end-users,” he says.

And microreactors’ engineering benefits for process, reaction and plant engineering are the “major drivers” for increased uptake of microreactor technology, even overshadowing the chemistry benefits, says Volker Hessel, vice director of R&D for IMM, a state laboratory that has become a worldwide leader in microreactor technology. IMM specializes in microfluidics, microstructured reactors, microprocess engineering, microfabrication (mostly through dry/wet silicon etching), and microsensors.

Hessel estimates that microreactor costs for a large-scale manufacturing facility would amount to no more than a few percent of the plant’s capital and operating costs. Microreactors safely handle reactions that might be explosive in larger batches, yet can still turn out many kilograms of product per hour.

The Fraunhofer Institute for Chemical Technology (Pfinstal, Germany) is also working on bioreactor technology, employing about a dozen scientists in the development and scaleup of highly exothermic reactions in microreactors. Fraunhofer became interested in this approach from a safety standpoint, and for the last 10 years has been transferring its explosives work to microfluidic platforms. “Most of our clients come from fine chemicals or pharmaceuticals,” says Stefan Loebbecke, Ph.D., vice director for energetic materials. Many projects at Fraunhofer begin as feasibility studies but increasingly, when an economic advantage becomes apparent, clients ask for a scaled-up process. Often this involves moving to a reactor with higher throughput, running several reactors in parallel, or integrating microreactors with batch reactors.

'Small' may mean micro-, meso- or mini-

The principal drawback with microreactors is the propensity of solids to clog the microchannels, and the smaller they are, the more vulnerable they are to this tendency. It is possible to generate very small particles in microreactors, or to take a process through nucleation and crystallize outside the reactor.

However, clogging is becoming less of an issue as developers turn to meso- or mini-scale reactors, and learn how to deal with the small dimensions by redesigning chemistry to fit into small places. A mesoscale reactor is about the size of a shoebox, with an active reactor volume of up to about one liter. Mesoreactors consume between five and 10 kilowatts of energy, and churn out hundreds of liters per hour of process fluid. Throughput is even higher, on a volumetric if not mass basis, for gas-phase reactions.

Adapting a process for microreactor format takes good communication between chemists and chemical engineers, and at a much earlier stage than for macro-scale batch reactors. “You need an engineer even at the lab scale, because it’s easy to make serious mistakes when operating a microflow device. And sometimes you simply have to change the chemistry,” says IMM’s Hessel.

MicroInnova (Graz, Austria) is yet another central European startup that exploits European microfluidics expertise for chemical synthesis. Two teams focus, respectively, on process development and engineering in microreactors, principally from IMM.

According to CEO Dirk Kirschnek, Ph.D., fine chemical companies often turn to microtechnology for process intensification and greater plant efficiency, while pharmaceutical companies seek more efficient production and higher yield due to the high value of pharma intermediates. Moreover, processes that are already efficient are the least likely to switch from batch to continuous microprocessing. “Fine chemical manufacturers, especially, are looking for significant improvements. Those who already enjoy good yields would probably not investigate microreactors," says Dr. Kirschnek.

MicroInnova recently completed a project for a U.S. fine chemicals manufacturer, where the goal was plant capacity and energy savings. This client presented a two-step batch process: the first reaction was exothermic and the second was endothermic. By porting the first step into a microreactor, MicroInnova achieved greater control and could harness the released energy for the second step, which was retained in batch format.

Moving from a large-footprint batch reactor led to essentially doubling the plant’s capacity. The process also scaled directly, without requiring a long bench-to-pilot-plant development time. Throughput was about three tons per hour using a single IMM reactor with channels 250 micrometers in diameter and a total internal volume of less than one liter. “This reactor wasn’t as small as possible, but as small as necessary to get the desired performance,” Dr. Kirschnek observes.

Breadth of chemical reactions

Built on expertise in microchanneled devices, microactuators and microelectromechanical systems (MEMS), NanoSciences (Aliso Viejo, Calif.) provides “small” technology services like design, fabrication, and, if necessary, manufacturing of microfluidic devices. NanoSciences offers microreactor technologies based on silicon, glass, quartz, and polymeric lab-on-a-chip platforms popularized for analytic devices during the early and mid-1990s. The company integrates microreactors directly to analytic or preparative HPLC, which aids in reaction optimization as well as isolation and analysis of microreactor products. Or, a conductivity detector can provide in situ process control.

NanoSciences’ business development director Jim Clements says the company’s “most exciting” microreactor work so far has been in collaboration with LioniX BV (Enschede, the Netherlands) and Chemtrix (Hull, UK), a spinout company from the University of Hull. LioniX specializes in integration of micro-optics, micropumping and microfluidics within glass-based chips; Chemtrix’s specialty is microreactors proper.

In one specialty pharma application, NanoSciences demonstrated nanofluidics separation of radioactive fluorine-18, used in nuclear medicine and research, from proton-irradiated water. Miniaturizing the synthesis cell to small chip size greatly relieves space constraints in “hot” cells where high specific-activity radiopharmaceuticals are manufactured.

The company claims to have validated an “enormous number” of reactions and processes with the Chemtrix microreactor, including dehydrations, enzyme-based reactions, aromatic nitration, amide synthesis, peptide synthesis, Diels-Alder, Grignard and Suzuki reactions, diazo couplings, carbanion chemistry, synthesis of enol ethers, Michael additions and pyrazole syntheses. “Microreactors provide much more flexibility in design of experiment, and extremely attractive economics for manufacturing and production engineers,” Clements adds.

Slow uptake?

Observers believe pharmaceutical manufacturers will adopt microreactors eventually, but not as soon or as readily as materials and polymers industries did. “As long as chemistry continues to be ‘thrown over a fence’ from discovery to manufacturing, microreactors will not have much impact or will advance slowly in pharmaceuticals,” says Thomas Schwalbe, Ph.D., CEO of MRSP and Acclavis (both Boston, Mass.). MRSP supplies proprietary microreactor hardware as well as distributing microreactors from CPC (Mainz, Germany), while Acclavis, a microtechnology consulting and “knowledge” company, provides process development expertise.

However, the concept of throwing technology over the fence is already starting to appear outdated. Given the industry’s growing emphasis on manufacturing flexibility, the next few years, and the pace of microreactor adoption in scaleup and manufacturing, could be quite interesting to watch.

Microreactor Materials: Glass Still Dominates

In the early days of nanofabrication and microfluidics, perfectly flat silicon was less expensive than glass of similar quality, due to semiconductor industry spinoff technologies.

This borosilicate glass microreactor features integrated platinum electrodes, and was manufactured by Chemtrix, Ltd., a University of Hull spinoff company. Photo courtesy of Chemtrix.

Eventually, glass and polymers became accessible at extremely high quality and flatness. Currently the material of choice for microreactors, glass is mechanically strong, chemically resistant, optically transparent, and etched conveniently with photolithography.

For several years, microfabricators have expected polymers to replace glass, silicon and quartz chips, but that has not yet occurred, except for some biochemical microdevices — for example, chips running polymerase chain reactions.

Still, plastics are gaining, and several optically pure, chemically inert polymers have been tested in microreactors. NanoSciences has developed a laser welding technology for bonding two “slices” of the polymeric nanochannel sandwich for analytical chips and synthesis reactors. Specialty polymers are not cheap, but the fluid contact surfaces are so small that cost hardly matters. The problem is convincing polymer manufacturers, who are used to delivering railcar loads of material, to commercialize small quantities of these materials.

About the Author

Angelo De Palma | Ph.D.