A technique known as magic-angle spinning has helped make nuclear magnetic resonance spectroscopy as sensitive for solids as it is for solutions. Inductive thinking leads to even better signal detection.
The great strength of nuclear magnetic resonance (NMR) spectroscopy is that it can determine, non-invasively and at atomic resolution, the chemistry, structure, dynamics and overall architecture of samples in solid, liquid or even gaseous forms. The liquid version of the technique, solution NMR, is used routinely to identify small molecules, study protein structures and dynamics, and probe intermolecular interactions. Solid-state NMR teases out the structure and properties of materials, surfaces and biological solids such as human tissue. But compared with many other analytical techniques, NMR has extremely poor sensitivity. A great deal of research has sought to improve this situation: on page 694 of this issue1, Sakellariou et al. describe a potential leap forward for solid-state NMR.
When atomic nuclei with non-zero spin are placed in an external magnetic field, they become polarized, precessing rather as a gyroscope does in Earth's gravitational field. When electromagnetic radiation of a frequency (energy) that corresponds exactly to that of the energy gap between two states of different polarization is applied to the sample, the nuclei resonate, jumping between those states. The accompanying gyroscopic precession of the spins induces a current in a conducting coil placed around the sample. This basic principle is both NMR's blessing and its bane as a spectroscopic technique: the small energies make the approach non-destructive, but they also make it difficult to distinguish the characteristic polarization (or signal) from thermal noise.
The signal-to-noise ratio in NMR measurements can be improved by either one of two general routes. The first of these is enhancing the starting polarization. NMR resonant energies are proportional to the strength of the magnetic field. Therefore, stronger magnetic fields improve the initial polarization and lead to more signal. But generating strong, uniform magnetic fields throughout a sample is expensive and requires considerable infrastructure, posing serious practical limitations. Dynamic nuclear polarization2,3, in which the relatively large polarization of electrons compared with nuclei is transferred to nuclei, is rapidly gaining popularity and applicability, but requires specialized equipment and substantial manipulation of the sample. Furthermore, it might not work for all samples and experiments.
The second general route to more sensitive NMR is to design detection schemes that make better use of the polarization signal. Several research groups are developing procedures that rely on mechanical coupling of the polarization to very sensitive cantilevers4, or optical rotation of a probe beam running through the sample5, to improve sensitivity. But these technologies, too, have limited practical application.
Sakellariou and colleagues1 build on what has proved to be one of the most general and cost-effective ways to improve the sensitivity of solution NMR: detecting the voltages induced in a coil that is optimized for and is closer to the sample. Such a coil is by its nature more efficient, because the signal-to-noise per unit mass of sample scales inversely with the diameter of the coil6. Furthermore, the 'filling factor' — the volume within the coil that is taken up by the sample — is an important variable. In solution NMR, solenoidal microcoils have been used to analyse liquid sample volumes of a few nanolitres7 and to perform magnetic resonance microscopy of individual neurons8. Systems for analysing volumes of 1–10 microlitres are available commercially, and the ability to reduce the sample size has also allowed for the collection of many NMR spectra simultaneously9. As well as being highly sensitive, solenoidal coils are quite easy to construct on a very small scale.
Until now, however, solid-state NMR had not enjoyed as much benefit from microcoils as had solution NMR. Unlike molecules in solutions, those in solid samples do not tumble rapidly or isotropically on the NMR timescale. The anisotropic interactions provide important structural information, but they also lead to broad, nondescript NMR spectra that are intractable to analysis. This problem can be countered, and solid-state spectra can achieve a resolution similar to that of their solution NMR counterparts10 by a trick known as magic-angle spinning (MAS), in which the sample is rotated at a speed of several kilohertz and at an angle of 54.7° relative to the magnetic field. In traditional MAS NMR, the sample is spun in a rotor within a static assembly containing a fixed coil that is some distance from the sample and therefore has a poor filling factor.
Sakellariou and colleagues' simple advance1 is to wind a solenoid microcoil directly around the sample — greatly improving the filling factor — and to spin the sample and coil together (Fig. 1). The spinning microcoil couples inductively to a coil that, just as in the conventional approach, remains static in the surrounding assembly. This 'magic-angle coil spinning' (MACS) technique uses existing commercial MAS solid-state NMR probe technology, while offering the advantages of small coil size and excellent filling factor that have been the province of solution NMR for over a decade7.
The authors' set-up can improve the signal-to-noise ratio by about an order of magnitude, and so allows smaller samples to be studied. The microcoil can also significantly increase the radio-frequency fields for a given current within the static coil, allowing more efficient manipulation of the spin polarization with radio-frequency pulses. The MACS technique has many conceivable applications, including structural measurements of very small protein samples, 'metabolomics' studies of the biochemistry of microscopic tissue extracts, and NMR measurements of radioactive materials that must be contained by specialized barriers1.
As with any technology, not all samples will be ideal candidates for the approach. This is especially true of samples in which the signal of interest is present in limited concentrations, such as those for trace amounts of metabolites in tissue, or membrane proteins that aggregate at higher concentrations. At low concentrations, the amount of material will still need to be increased. But for many applications in chemistry, biology and materials science, Sakellariou and colleagues' advance opens up new opportunities simply by reducing the amount of material required for solid-state NMR studies, without needing to invest substantially in new technologies.
About this article
Nature Methods (2009)