If a nanoscale gallium arsenide structure is excited with an oscillating magnetic field, superpositions of nuclear spin states can be created and detected electrically. Quantum computing could be the beneficiary.
Nuclear magnetic resonance (NMR) spectroscopy is a near-ubiquitous technique in the physical and biological sciences. Applications include the verification of molecular structures in synthetic chemistry, investigation of superconductivity in solid-state physics, the determination of protein conformations in structural biology, and diagnostic imaging in medicine. Typically, an adequate signal-to-noise ratio in NMR can be achieved only with a relatively large sample of 1015–1018 molecules or atoms. On page 1001 of this issue, Yusa et al.1 demonstrate an unconventional approach to NMR in a cleverly designed semiconductor nanostructure with which strong signals can be detected from around 108 atoms.
NMR depends on manipulations of the spin states of atomic nuclei that possess magnetic moments. When an external magnetic field is applied to such nuclei, the energy levels allowed by the rules of quantum mechanics split — in the simplest case of a nucleus with a spin of magnitude 1/2, they split into two levels close to one another. If resonant radio-frequency pulse sequences with precisely controlled amplitudes, phases, frequencies and timing are then applied, transitions can be induced between these energy levels, creating coherent superpositions of spin quantum states. The frequencies at which these transitions occur (‘NMR spectra’) correspond to the gaps between spin energy levels and are therefore characteristic of a particular nucleus and its environment.
It is the ability to control quantum mechanical states more precisely than in any other type of spectroscopy that makes NMR spectroscopy useful in many disparate disciplines. The ‘controllability’ of nuclear spin states has also made NMR of interest in quantum computation (for an example see ref. 2). Here, mathematical problems can be efficiently solved by algorithms that depend on the manipulation of coherent superpositions of quantum mechanical states.
The main weakness of conventional NMR, which uses metal coils with centimetre-scale dimensions to detect nuclear-spin magnetism from samples with similar dimensions, is its relatively low sensitivity. Not only have Yusa et al.1 significantly reduced the minimum sample needed to perform NMR spectroscopy, but they have also shown that, by exciting the nanostructure with radio-frequency magnetic fields, arbitrary superpositions of spin states can be created. These include ‘multiple quantum coherences’, superpositions of states that differ by more than one quantum of angular momentum3 in nuclei with a total spin greater than a half. Although this work is motivated by the goal of developing a solid-state electronic device for quantum computation, other applications and extensions of this nanotechnological approach to NMR may be possible.
A schematic representation of the device described by Yusa et al. is shown in Figure 1. Using a combination of semiconductor-growth, lithography and electrical-gating techniques, they created a nanometre-scale channel of electron-doped gallium arsenide through which a current I of around 7 nanoamperes passes in response to an applied voltage V. At a temperature of about 100 millikelvin and with a static magnetic field B0 of 5.5 tesla, the current generates large polarizations in the nuclear spin — that is, a large difference between the number of gallium and arsenic nuclei that have spin parallel to the static field and the number of nuclei with spin antiparallel to the field. As shown in earlier studies4,5,6,7, under these experimental conditions the polarized nuclear spins interact with flowing electrons in such a way that the electrical resistance R=V/I of the gallium arsenide channel increases by roughly 10%.
Simulations performed by Yusa et al. to fit their data indicate that this resistance change is directly proportional to the nuclear-spin polarization1. When an oscillating magnetic field B1 is applied with a frequency f1 close to an NMR frequency, the direction of the nuclear-spin polarization is inverted in a periodic, sinusoidal manner, producing sinusoidal variations in R. By ‘sweeping’ the frequency of the applied field, the complete NMR spectrum of the gallium arsenide channel can thus be mapped out — not just the single-quantum transitions of the 69Ga and 75As nuclei seen in conventional NMR spectra, but also transitions involving states of these nuclei with spin 3/2 that differ by two or three quanta of angular momentum. Moreover, the sinusoidal variations in R persist over timescales approaching 1 millisecond, showing that long-lived coherent superpositions are generated in these experiments, and that the creation of arbitrary superpositions would be possible with more elaborate radio-frequency irradiation schemes.
Could electrical detection using similar devices become a general method for performing NMR measurements? The electronic properties of gallium arsenide, and the technology to use these properties for creating complex nanostructures, are rather special. But nanostructures based on gallium arsenide are widely used for studying the generic characteristics of confined, interacting electrons8. Nanometre-scale NMR devices such as those described by Yusa et al. could therefore contribute to our understanding of such systems and of interacting many-body systems in general. Given the interest in the development of organic materials possessing electronic properties analogous to those of gallium arsenide, it also seems likely that we could soon see nanometre-scale NMR devices constructed from organic molecules. By depositing biological macromolecules on such a device, and taking advantage of NMR tricks that permit the exchange of nuclear-spin polarization between molecules in contact with one another9, it might be possible to record NMR spectra of picomolar quantities of proteins and nucleic acids by purely electrical measurements.
Yusa, G., Muraki, K., Takashina, K., Hashimoto, K. & Hirayama, Y. Nature 434, 1001–1005 (2005).
Cappellaro, P. et al. Phys. Rev. Lett. 94, 020502 (2005).
Vega, S. & Naor, Y. J. Chem. Phys. 75, 75–86 (1981).
Kane, B. E., Pfeiffer, L. N. & West, K. W. Phys. Rev. B 46, 7264–7267 (1992).
Wald, K. R., Kouwenhoven, L. P., McEuen, P. L., Vandervaart, N. C. & Foxon, C. T. Phys. Rev. Lett. 73, 1011–1014 (1994).
Machida, T., Yamazaki, T., Ikushima, K. & Komiyama, S. Appl. Phys. Lett. 82, 409–411 (2003).
Yusa, G., Hashimoto, K., Muraki, K., Saku, T. & Hirayama, Y. Phys. Rev. B 69, 161302 (2004).
Tycko, R., Barrett, S. E., Dabbagh, G., Pfeiffer, L. N. & West, K. W. Science 268, 1460–1463 (1995).
Pines, A., Gibby, M. G. & Waugh, J. S. J. Chem. Phys. 59, 569–590 (1973).
About this article
Miniaturization of NMR Systems: Desktop Spectrometers, Microcoil Spectroscopy, and “NMR on a Chip” for Chemistry, Biochemistry, and Industry
Chemical Reviews (2014)
Semiconductor Science and Technology (2009)
Spectrochimica Acta Part B: Atomic Spectroscopy (2008)
The Journal of Chemical Physics (2007)
Journal of Physics: Condensed Matter (2006)