UK scientists have extended the bounds of solid-state NMR to quickly solve a problem plaguing the pharmaceutical industry: how to spot unwanted crystal packing forms in a drug tablet. 

Polymorphism - when compounds adopt more than one crystal structure ('polymorph') in the solid state - is a thorn in the side of drugs manufacturers. A tablet holding an unwanted polymorph can behave unexpectedly by, for instance, releasing its active ingredient too quickly in the body. Different polymorphs count as separate patentable forms of drugs, so weeding them out can be an expensive irritant for pharmaceutical companies - or a lucrative opportunity for competitors to seize a share of the market. 

Solid-state NMR (nuclear magnetic resonance) is already used to pick out polymorphs. Just like liquid NMR, the technique works out molecular structure by detecting the response of nuclear spins to radiowave pulses. But there is an added complication: in the solid state, nuclear interactions are so strong that an NMR spectrum becomes swamped with information. The spectrum no longer appears as familiar narrow spikes for each nucleus, but as a series of broad peaks that are hard to untangle. 

So-called magic angle spinning (MAS) - rotating the solid sample to imitate the effects of a molecule tumbling in a liquid - gets rid of some of this information overload, so that solid-state NMR on carbon-13 nuclei can identify molecules and how they pack in a crystalline array. 

But to detect the subtler forms of polymorphism it is more useful to look at hydrogen, not carbon nuclei; for hydrogen bonds between molecules are usually the driving force determining which polymorph appears. Sadly, MAS doesn't cut out proton nuclei interactions, so decoding proton structures has been considered a specialised - and difficult - task. 

Now, John Griffin and Steven Brown from the University of Warwick, together with Dave Martin from pharmaceutical company AstraZeneca, have simplified the problem, taking a high resolution 2D proton solid-state NMR spectrum of a tablet in just two hours. The researchers easily managed to distinguish between two forms of an active ingredient in a tablet: particularly impressive, because the drug in a tablet is mixed up with other inactive compounds such as fillers and glues. 

'This new approach should be adopted as a routine tool in the characterisation of pharmaceuticals,' Brown's team recommend. 

Their technique, known as CRAMPS (combined rotation and multiple-phase spectroscopy) was conceived in the 1970s. The idea was to cancel out confusing hydrogen interactions by choreographing the phases and timings of the radiowave pulses used in NMR with the rotation of the sample. Thanks to improvements in the consoles that control the radiowave pulse sequences, it is now possible to put that theory into practice, Brown said. 

Robin Harris, an NMR expert at Durham University, said the team's work was very neat and would add to the NMR toolbox. But he pointed out that the detection limits of the technique for spotting unwanted polymorphs were not clear. 

However, polymorph-spotters in the pharmaceutical industry were immediately impressed. 'This approach will extend our capabilities for analysing complex mixtures, and we will certainly take advantage of it,' said Stephen Byard, head of physical and molecular characterisation at French-based pharmaceuticals giant Sanofi-Aventis. 

Richard Van Noorden