RSC - Advancing the Chemical Sciences

Chemistry World


Column: In the pipeline

Chemists are finally going with the flow, says Derek Lowe

Organic chemists are no more immune to new gadgetry than any other consumers. Machines that promise to make our laboratory lives easier turn up all the time, but (just as in the home kitchen) many of them eventually end up under distant counters, far from any electrical socket and usually without their instruction manual. The usual downfall, in both cases, is the 'more trouble than it's worth' feeling that comes over you after you've spent two hours setting up something that you could have done faster the old way. 

But there's a recent lab fashion that might well claim some permanent bench space: flow chemistry. Instead of running things in flask-sized batches, the standard style of chemistry from undergraduate labs on up, these machines pump the starting material through a reaction zone and out the other side. Depending on the reaction, this could involve a passage through high heat or pressure, over a solid catalyst or reagent, or some combination of all of these.  

What may save flow reactors from the fate of the adjustable electric carrot peeler is the amount of time and effort that goes into the workup of many reactions. It's always a lot easier to set up reactions than it is to finish them off. Even experienced organic chemists can underestimate the time it takes to extract their products from a milkshake of metal salts, or to evaporate all the solvent they used while doing so. One promise of flow chemistry is that some of the messier reagents might stay put as the reaction moves through them, rather than being mixed into the pot and then cleaned out again. Solid-supported reagents have been around for years now, of course, but they haven't been used as often as they might be. In traditional batch chemistry, they still need to be filtered out at the end of the reaction, but in a flow apparatus the filtration step is the reaction. The increasing number of metal-catalysed coupling reactions would seem to be a natural fit for this technology. 

If at first you don't succeed . 

Another potential advantage is the chance to do chemistry at inconvenient temperatures and pressures. Since the flow reaction zone is relatively small (generally a short length of stainless steel tubing), it's easier to generate more extreme conditions there than it is in a larger reactor. Rather as with cooking fish, many reactions benefit from short exposure to very high temperatures. And as in the kitchen, if the first run doesn't take things to completion, the process can be repeated as needed. 

Much of the published flow research has come from academic labs, many of them in the UK, where Cambridge University's Steve Ley, among others, has been a visible and active proponent of the technology. Meanwhile, in the US, Novartis recently gave MIT an eyebrow-raising grant to develop flow apparatus suitable for large-scale process work. Their $65 million should remove any remaining doubts about taking the field seriously, although process chemists themselves have known better for some 
time now. 

What's the next step? Commercial apparatus is already appearing, with more surely in development, and this will allow people (at least in theory) to do flow chemistry without a supply of wrenches and electrician's tape on hand. It may take a while for these to lose their 'let's try tightening this part' aura, but that will come in time. And as with microwave reactors before them, turning them into tools rather than ends in themselves will be essential for broader success.  

For benchtop use, there are many potential refinements that haven't really been implemented outside of home-made rigs - microwave heating of the flow compartment being one of them. You'd figure that supercritical carbon dioxide might be a useful flow medium, in an attempt to get around one potential weakness of the technique (usage of solvent), and the first reports of this are already appearing. Another fix for this problem would be a recirculating flow reactor with some sort of in-line spectroscopic monitor. Such a machine would send the reaction mixture through a loop until the spectrum showed completion, and then divert the product into a receiver, no doubt freeing up industrial chemists to attend HR training courses or the like. But then there are disadvantages to every new technology, aren't there? 

Derek Lowe is a medicinal chemist working in preclinical drug discovery in the US