High-precision nuclear magnetic resonance spectroscopy generally requires the use of powerful magnets. But using Earth's magnetic field allows us to gain some of the same information on the cheap.
Are expensive superconducting magnets necessary to perform high-resolution nuclear magnetic resonance (NMR) spectroscopy? Not absolutely, say Stephan Appelt and his colleagues in the February issue of Nature Physics1. They use a far less pricey source of magnetism — Earth's own magnetic field — to distinguish the chemical structures of various molecules containing hydrogen, lithium and fluorine. The level of accuracy they achieve is an order of magnitude better than that possible with the most advanced superconducting magnets.
When an external magnetic field is applied to an atomic nucleus, it induces a polarization in the direction of the intrinsic rotation, or ‘spin’, of that nucleus' constituent protons and neutrons. These spins align either parallel or antiparallel to the field, causing the quantum-mechanically allowed energy levels of the nucleus to split. At a frequency corresponding exactly to the difference between these energy levels, the nucleus can absorb electromagnetic radiation: the phenomenon known as nuclear magnetic resonance or NMR, first observed by Felix Bloch and Edward Mills Purcell in 1946 (for which achievement they won the Nobel Prize in Physics in 1952).
Soon after the initial discovery, it became clear that the effective magnetic field on a nucleus — and consequently the observed NMR frequency — is subtly changed by the effects of both orbiting electrons (the ‘chemical shift’)2 and the spins of neighbouring nuclei (‘J-coupling’)3. This was the beginning of the triumphant success of NMR as a spectroscopic tool for exploring the composition and chemical environment of molecules in the liquid state. In the decades since, the need for higher sensitivity and lower spectral dispersion has demanded higher, more homogeneous magnetic fields, fuelling the devel- opment of powerful superconducting magnets. Nowadays, the highest-resolution NMR techniques, using magnets producing field strengths of between 1 and 10 tesla, reproduce the hydrogen spectrum with a broadening of spectral lines caused by the instrumentation of less than a tenth of a hertz.
Compared with the fields that can be attained with superconducting magnets, Earth's magnetic field is weak: it varies from about 25 microtesla (μT) at the Equator to 75 μT at the poles, with geomagnetic field lines inclined, in Europe and North America, at an angle of about 60° to the (horizontal) surface. The field is not constant: currents in the ionosphere and disturbances from Earth's interior produce slow daily variations in the field with amplitudes of some 25 nanotesla (nT), and superimposed on these are further oscillations with periods of a few seconds and amplitudes of about 1 nT. Far enough from electric installations and other sources of artificial magnetic perturbation, however, proper shielding can reduce these shorter variations to about 0.1 nT s−1, and local spatial gradients to below 1 nTm−1. These variations are comparable to those found in the fields of artificial magnets.
The first observation of a nuclear magnetic effect in Earth's magnetic field — the free precession of proton spin4, akin to the precession of a spinning top in Earth's gravitational field — came not long after the discovery of NMR. The weakness of the geomagnetic field is such that it causes only a slight natural polarization in proton and neutron spins. Before an NMR measurement in the Earth field can be made, therefore, these spins generally have to be polarized by a high magnetic field, or polarization has to be transferred from more-readily polarizable electrons using a method known as dynamic nuclear polarization. Such techniques have been used for measurements of the precise ‘Larmor’ precession frequency for protons, of the relaxation time that spins require to return to their normal states following polarization, and of the strength of J-coupling between nuclei5. The method has also been applied to the magnetic resonance imaging of pieces of fruit and of phantoms standing in for human tissue6, and for the detection of groundwater reservoirs7.
The disadvantage of low-field NMR for chemical analysis is that the chemical shift of spectral lines is smaller than the broadness that is due to the magnetic field. This renders effectively unobservable the direct information about a nucleus' chemical environment gained from the chemical shifts caused by the surrounding electrons' screening of the applied magnetic field. On the other hand, the J-coupling between nuclear spins caused by chemical bonds is almost independent of the magnetic field, so its effect is proportionately greater where the field is low. In J-coupling, the quantum states of shared electrons impose information about the chemical environment through the indirect dipolar coupling between nuclei that are relatively close to one another. The effect of the resultant J-splitting on the NMR frequency is at an observable level of between a few tenths and a few tens of hertz.
In their experiments, Appelt et al.1 measure the J-coupled spectra of various compounds containing hydrogen, lithium and fluorine. They first pre-polarized the nuclei of each sample in the field of a permanent magnet to a value about 10,000 times that attainable in the geomagnetic field, and transferred the sample within one second into the detection coil of an Earth-field NMR spectrometer. They investigated benzene, lithium chloride dissolved in water, tetramethylsilane, silicone oil, octamethylcyclotetrasiloxane and nonafluorohexene, demonstrating in each case that their spectrometer could image these compounds' molecular structure. The instrumental broadening of their spectrometer is only a few millihertz over a sample volume of 1 cm3 — less than the broadening of high-resolution NMR with advanced superconducting magnets. The spectral resolution is, however, limited by the shorter spin relaxation time to a few tenths of a hertz.
The authors predict that the combination of low-field NMR with advanced methods of spin hyperpolarization will allow time-resolved NMR spectroscopy of rare nuclei such as 6Li, 13C and 29Si, and the separation of the corresponding J-couplings with high precision. Many applications of this mobile low-field NMR technique are also likely in medicine and materials science: the tracking of hyperpolarized 6Li+ and 7Li+ migrating through ion channels or membranes, for example, or the characterization of mineral oil in well-logging and the online detection of chemical reactions.
Earth-field NMR cannot replace the diagnostic capabilities of the chemical shift measurements available with the high-field method. But particularly in combination with superconducting quantum interference device (SQUID) detectors8, the technique could give limited, but potentially very useful, information about the coupling of many molecules.
Appelt, S., Kühn, H., Wolfgang, F. & Blümich, B. Nature Phys. 2, 105–109 (2006).
Proctor, W. G. & Yu, F. C. Phys. Rev. 77, 717 (1950).
Ramsey, N. F. & Purcell, E. M. Phys. Rev. 85, 143–144 (1952).
Waters, G. S. & Francis, P. D. Sci. Instrum. 35, 88–93 (1958).
Bene, G. J. Phys. Rep. 58, 213–267 (1980).
Stepišnik, J., Eržen, V. & Kos, M. Magn. Reson. Med. 15, 386–391 (1990).
Shushakov, O. A. Magn. Reson. Imaging 14, 959–960 (1996).
McDermott, R. et al. Science 295, 2247–2249 (2002).
About this article
European Journal of Physics (2015)
Journal of Magnetic Resonance (2012)
Progress in Nuclear Magnetic Resonance Spectroscopy (2008)