Zero-field nuclear magnetic resonanceMicah P. Ledbetter and Dmitry Budker Citation: Phys. Today 66(4), 44 (2013); doi: 10.1063/PT.3.1948 View online: http://dx.doi.org/10.1063/PT.3.1948 View Table of Contents: http://www.physicstoday.org/resource/1/PHTOAD/v66/i4 Published by the American Institute of Physics. Additional resources for Physics TodayHomepage: http://www.physicstoday.org/ Information: http://www.physicstoday.org/about_us Daily Edition: http://www.physicstoday.org/daily_edition

Downloaded 02 Apr 2013 to 193.198.74.30. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://www.physicstoday.org/about_us/terms

44 April 2013 Physics Today www.physicstoday.org

Progress in nuclear magnetic resonance(NMR) spectroscopy and magnetic reso-nance imaging (MRI) is conventionally as-sociated with working at higher and highermagnetic fields. Although such fields cer-

tainly have their virtues, those virtues come with aprice, literally and figuratively. A typical high-fieldsetup costs on the order of $1 million, requires con-stant cryogenic maintenance, is immobile, and can’tbe used on samples with metallic inclusions or im-planted devices. MRI, for example, has long beenoff-limits to patients with pacemakers, which canmalfunction under strong magnetic fields. Also,larger magnets often require extensive magneticfield shimming to produce the spatially uniformfields needed to obtain high-resolution spectra.

Starting with the pioneering work of AlexanderPines and coworkers in the 1980s,1 however, a some-what unexpected trend has developed toward us -ing ultralow, submicrotesla fields—or even no ex-ternal field at all. That trend has been enabled bynew techniques and technologies that have relaxedthe requirement for large magnetic fields in each of the three main stages of an NMR experiment—polarization, encoding, and detection (see figure 1).

In traditional, high-field NMR, a sample’s nuclear spins are polarized via thermal equilibra-tion in a magnetic field. The degree of spin polar-ization is proportional to the magnitude of the field,inversely proportional to temperature, and usuallysmall. Spin-1⁄2 hydrogen nuclei (protons) at roomtemperature in a 1-T magnetic field—typical condi-tions for an MRI experiment—attain polarizationson the order of 10 parts per million.

Next, information about the sample’s molecularstructure is encoded in the form of spin- resonancefrequencies—the frequencies at which spins precessabout an applied magnetic field. The resonance fre-quency of a given nuclear spin depends on its gyro-magnetic ratio and the local magnetic field. In MRI, spatial information is encoded by applying aspatially dependent magnetic field; in NMR spec-troscopy, information about a spin’s chemical envi-ronment is inherently encoded by way of the so-called chemical shift—the difference, due to thediamagnetic screening effects of neighboring elec-trons, between a nucleus’s actual resonance fre-quency and the resonance frequency it would havein isolation. The magnitude of the chemical shift is proportional to the applied magnetic field and isnormally a tiny fraction—on the order of parts permillion—of the resonance frequency itself.

Detection often occurs concurrently with en-coding and is traditionally accomplished with Fara-day induction coils. An RF magnetic field pulse res-onantly excites a specific subensemble of thesample’s nuclear spins, which in turn induce an ACvoltage—the signal—in a pickup coil. Because theinduced voltage is proportional to the spin polariza-tion and the rate of precession, each of which areproportional to the magnetic field, the strength ofan NMR signal has approximately quadratic overalldependence on the applied field. Hence, the drivetoward bigger and stronger magnets. State-of-the-art magnets now used in NMR spectroscopy canproduce fields near 21 T. (See the article by SteveVan Sciver and Ken Marken, PHYSICS TODAY, August2002, page 37.)

Such magnets, however, are not always a pre-requisite for obtaining high-quality NMR spectra. Inthe absence of strong magnetic fields, large nuclearspin polarization can be achieved through tech-niques such as dynamic nuclear polarization, spin- exchange optical pumping,2 and parahydrogen-

Counter to intuition, one doesn’t necessarily need a strong magnet—orany magnet, for that matter—to extract richly informative spectra from nuclear spins.

Micah Ledbetter, a physicist at AOSense Inc in Sunnyvale, California, was an assistant researcher in the department of physics at the University of California, Berkeley. Dmitry Budker is a professor in the department ofphysics at UC Berkeley and a faculty scientist in the nuclear science division at Lawrence Berkeley National Laboratory.

nuclear magnetic resonance

Micah P. Ledbetter and Dmitry Budker

Downloaded 02 Apr 2013 to 193.198.74.30. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://www.physicstoday.org/about_us/terms

induced polarization;3 encoding can be accom-plished through what is known as scalar coupling,or J- coupling, between spins;4,5 and spin resonancesin weak fields can be detected with exquisite sensi-tivity using atomic magnetometers or supercon-ducting quantum interference devices (SQUIDs).6,7

Recently, zero-field techniques for polarization,encoding, and detection have been combined in acomplete “NMR without magnets” experiment8 andin experiments with tiny, 100-nT fields.9,10 In addi-tion to the economy and portability that comes withgetting rid of the large magnets, the narrow spectrallinewidths associated with low- and zero-fieldNMR allow one to detect certain low-frequencyspin–spin couplings that would be challenging toobserve in traditional high-field NMR. Indeed, onecan gather a surprising amount of informationabout molecular structure by turning down, or turn-ing off, the magnetic field.

Exploiting Earth’s fieldPerhaps the most common low-field NMR experi-ments are those carried out in Earth’s ambient magnetic field. That field, around 40 μT at Earth’ssurface, is some five orders of magnitude weakerthan that of a typical high-field NMR magnet. How-ever, Earth’s field has high spatial homogeneity,which significantly reduces the magnetic resonancelinewidth. Because homogeneity is important dur-ing the encoding and detection stages but not thepolarization stage, Earth-field NMR experimentscan be performed using prepolarization in astronger field; once the sample is removed from thatfield, the experimenter typically has a few secondsto detect spin resonance frequencies before the spinssuccumb to relaxation.

Since the demonstration of Earth-field NMR byMartin Packard and Russell Varian in the 1950s,11 thegroups of Paul Callaghan in New Zealand and Bern-hard Blümich in Germany have shown that the tech-nique can yield considerable information about mo-lecular conformation and chemical environment.12

An important advantage of abandoning high-fieldmagnets is that the NMR system becomes portable,and can be used to do research in situ or in the field,with no need to deliver samples to the laboratory.

SQUIDs and atomic magnetometersFaraday coils have long been the traditional meansfor detecting NMR signals, but in recent years,SQUIDs and atomic magnetometers have emergedas viable alternatives. SQUIDs, which detect mag-netic flux through a superconducting loop inter-rupted by a pair of Josephson junctions, can meas-ure magnetic fields with sensitivities of better than1 fT with one-second averaging. Similar sensitivitycan be achieved with atomic magnetometers, whichuse an atomic spin—normally the spin-1⁄2 valenceelectron of an alkali atom—to measure the Zeemaneffect, and thereby infer field strength, in the vicinity of a polarized sample. (See PHYSICS TODAY,July 2003, page 21.)

For optimal performance, SQUIDs are usuallyoperated at liquid-helium temperatures, whereasatomic magnetometers work best between room

temperature and 170 °C. Both are well suited for de-tecting the low-frequency magnetic resonance sig-nals typically encountered in low magnetic fields: Incontrast to detection by Faraday- induction coils, thesensitivity of SQUID and atomic-magnetometermeasurements is for the most part independent ofprecession frequency.

Nuclear quadrupole resonanceOne variation of NMR that does not rely on magnetsat all is nuclear quadrupole resonance (NQR). If anucleus has spin 1 or greater, then in addition to amagnetic moment, it can possess an electric quadru-pole moment. In a crystalline solid, the electricquadrupole moment interacts with the gradient ofthe electric field produced by the crystal lattice,splitting the energy of the Zeeman sublevels of the nucleus. The splitting produces a weak nuclear spinpolarization on the order of 1 ppm in the absence of any external fields. Transitions between the sublevels can be resonantly excited with MHz- frequency pulses and detected as oscillations in thesample magnetization.

Whereas in high-field NMR the chemical envi-ronment only weakly affects resonance frequencies,

www.physicstoday.org April 2013 Physics Today 45

Figure 1. The three stages of a nuclear magnetic resonance experi-ment, with representative techniques. In the traditional approach, asample’s nuclear spins are polarized via thermal equilibration in astrong magnetic field; information about molecular structure or spa-tial distribution of nuclei is encoded in the form of spin resonances—for instance, by way of chemical shifts or an applied magnetic fieldgradient; and the resonances are detected with Faraday inductioncoils. Each stage benefits from a strong magnetic field. In the absenceof such a field (techniques in red), spin polarization can be achievedby dynamic nuclear polari zation, spin-exchange optical pumping, orparahydrogen-induced polarization; encoding can take the form of J-coupling; and spin resonances can be detected using superconduct-ing quantum interference devices or atomic magnetometers.

Downloaded 02 Apr 2013 to 193.198.74.30. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://www.physicstoday.org/about_us/terms

46 April 2013 Physics Today www.physicstoday.org

Zero-field NMR

NQR frequencies are completely determined by theelectronic environment around the nucleus, a perfectfeature for chemical identification. Conveniently, oneof the most abundant nuclei in explosives and nar-cotics, nitrogen-14, is spin 1 and quadrupolar, whichhas enabled their detection at bor-der crossings and airports. BothSQUIDs and atomic magnetome-ters have been successfully appliedto NQR detection.13

J-couplingAlthough NQR has found suc-cess, many nuclei of interest inNMR have spin 1⁄ 2. In recent dec -ades, several groups have begun toinvestigate ultralow-field NMR ofsuch nuclei by exploiting a partic-ular form of spin–spin coupling.

To illustrate how it works, weconsider a simple, representativemolecular system: CH−, consistingof a carbon-13 atom and a proton.The extra electron is added so thatthe entire set of electrons has zerototal angular momentum. (We willnot be further directly concerned

with the electrons throughout this article.) The 13Cnuclear spin S and the 1H nuclear spin I are both 1⁄2.

Although there is magnetic dipole–dipoleinter action between the 13C and 1H nuclear spins, itvanishes when averaged over all the possible rela-tive locations of the spins. Therefore, in a liquid orgas, where molecules are rapidly tumbling, the effects of dipole–dipole coupling are unobservable.The molecular tumbling does, however, producerandom fluctuations of the dipole–dipole inter -action that contribute to nuclear spin relaxation.

It thus came as a surprise 60 years ago whenCharles Slichter and coworkers and a team compris-ing Erwin Hahn and Donald Maxwell indepen -dently discovered that a spin–spin interaction canexist even when the dipole–dipole interaction van-ishes due to molecular tumbling.14,15 Known asscalar coupling, or J-coupling, and described by theHamiltonian HJ = hJI · S, the interaction is invariantwith respect to the relative positions of the spins.Here, J, in frequency units, characterizes thestrength of the interaction and h is Planck’s constant;note that the Hamiltonian is formally identical tothat of hyperfine coupling between electrons andnuclei as encountered in atomic spectroscopy.

In a 1952 paper,16 Norman Ramsey and EdwardPurcell explained how such an interaction couldoccur via indirect means: The magnetic field of onenucleus perturbs the molecular electrons, which inturn “carry” the interaction to the other nucleus.Typically, J is in the range of 100–200 Hz for 13C and1H nuclei separated by a single covalent bond, andit rapidly decreases as the number of intermediatebonds grows. The high resolution of zero-fieldNMR, which for proton–13C systems can deliverlinewidths below 10 mHz, facilitates the detectionand characterization of multi-bond J-coupling. Forcomparison, one has to work extremely hard to re-alize linewidths narrower than 0.5 Hz in high-fieldproton NMR.

In the absence of other interactions, the quan-tum numbers corresponding to conserved quanti-ties at ultralow fields are I, S, F, and MF, where F is

Figure 2. In the zero-field experiment depicted here, the sample(green) is initially polarized in a permanent magnet and then shut-tled into a magnetically shielded region (bottom). A set of coils isused to null residual magnetic fields and to apply short magneticfield pulses, which excite coherences between the nuclear spinstates and set into motion spin oscillations at the sample’s J-coupling frequency. Those oscillations are detected by an atomic magnetometer, whose key component is a cell containinga vapor of alkali atoms: A pump laser polarizes the atoms, and aprobe laser detects their precession, from which one can infer the magnetic field due to the sample.

110 240115 250120 260125 270130

* *

FREQUENCY (Hz)

b

a

SIG

NA

L

Figure 3. The zero-field spectrum of acetonedoubly labeled with carbon-13. (a) The molecule’snuclear spins reside in its side methyl groups: C atoms are shown in black (with 13C indicated byasterisks), hydrogen-1 is in white, and oxygen is inred. (b) The resulting spectral pattern is uniquelydetermined by the J-couplings within and between the methyl groups and yields a distinct molecular “fingerprint.”

Downloaded 02 Apr 2013 to 193.198.74.30. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://www.physicstoday.org/about_us/terms

www.physicstoday.org April 2013 Physics Today 47

the quantum number associated with the total an-gular momentum F = I + S and MF is the projectionof F onto the quantization axis. For a pair of spin-1⁄2nuclei, F can take the values 0 and 1, correspondingto singlet and triplet states, respectively. Evaluatingthe matrix elements of the Hamiltonian, one findsthat the singlet state (for which ⟨HJ⟩ = −3hJ/4) andtriplet state (⟨HJ⟩ = hJ/4) are separated in energy by hJ.

Over the past few years, in collaboration with thePines group, we have been investigating J-coupling asa means for conducting NMR in zero magnetic field.Although it may sound impossible—without amagnetic field, the spins seemingly have nothing toprecess about—zero-field J-coupling spectroscopyworks and can deliver spectral resolution on theorder of millihertz and reveal rich informationabout molecular identity and structure.

Generating spectraA prototypical zero-field NMR experiment is de-picted schematically in figure 2. The spins are pre-

polarized by thermal equilibration in the field of a permanent magnet, just as in high-field NMR experiments. After polarization, the sample—assumed to be a liquid—is moved into a zero-fieldregion inside nested layers of magnetic shielding.

Provided the sample contains at least two nu-clei with distinct gyromagnetic ratios, one can putthe nuclei into coherent superpositions of the singletand triplet states by applying a short magnetic fieldpulse. The sample’s magnetic field then oscillateswith frequency J (see the box above), and those os-cillations are detected by an atomic magneto meterlocated in close proximity to the sample.

The technique is not limited to two-spin sys-tems; it can also be used to analyze more complexmolecules. To illustrate, we show in figure 3 an ex-perimental spectrum of acetone molecules in whichboth methyl groups are labeled with 13C. To under-stand the spectrum, first consider just one of themethyl groups. The three protons can have total spin1⁄2 or 3⁄2. Coupling of the 13C nucleus to a combined

Consider a sample consisting of two nuclei with spinsI = S = 1⁄2, polarized by a large magnetic field. Adiabati-cally moving the sample to zero field results in twotypes of nuclear spin polarization: a dipole moment in the triplet state—that is, an excess population in the ∣F = 1, MF = 1⟩ = ∣↑↑⟩ state relative to the ∣F = 1,MF = −1⟩ = ∣↓↓⟩ state—and an excess population in thesinglet state ∣F = 0, MF = 0⟩ relative to the ∣F = 1, MF = 0⟩state. (This can be seen by following the curves in fig-ure 4a from high to low field.) Here, F is the total spinangular momentum and MF is its projection on thequantization axis z.

Panel a illustrates the response of the dipole polar-ization to a magnetic field pulse. For simplicity, we assume that all molecules start off in the state ∣↑↑⟩. The effects of the applied magnetic field B are most easilyunderstood in the uncoupled basis ∣MI, MS⟩. The panel’sblack and red arrows indicate the expectation values of the spin vectors I and S, both of which start off polar-ized in the z-direction. Assuming the spins have distinctgyromagnetic ratios γi = 2πgiμN/h—where g is the nuclear g-factor, μN is the nuclear magneton, and h isPlanck’s constant—a magnetic pulse oriented in thex-direction causes the spins to rotate through differentangles. For simplicity of illustration, we assume here thatgI = 2gS; the ratio of the g-factors of carbon-13 nuclei

and protons is roughly four to one. By tuning the dura-tion and strength of the magnetic field pulse, we canplace the spin system in the state ∣ψ(0)⟩ = ∣↑↓⟩ = (∣F= 1,MF = 0⟩ + ∣F = 0, MF = 0⟩)/√2‾, a superposition of the coupled- basis singlet and triplet eigenstates.

The triplet and singlet states are separated in energyby hJ, so after the magnetic pulse, quantum beats occurbetween those two states at precisely the frequency J.One can show that the expectation values of the zcomponents of the spins are ⟨Iz⟩ = 1⁄2 cos 2πJt and⟨Sz⟩ = −1⁄2 cos 2πJt. The evolution of ⟨Iz⟩ and ⟨Sz⟩, scaled bygI and gS, is depicted by the black and red arrows, respectively, in panel b. Their sum yields the total mag-netization, represented by the smooth blue curve.

The picture is only slightly different if we assumethat all molecules start in the singlet state ∣F = 0,MF = 0⟩ = (∣↑↓⟩ − ∣↓↑⟩)/√2‾. We apply a pulse of magneticfield, now in the z-direction. During the pulse, thestates ∣↑↓⟩ and ∣↓↑⟩ each acquire different phases. By judicious choice of pulse strength and duration, onecan arrange for the final state, expressed in the coupled basis, to be ∣ψ(0)⟩ = (∣F = 1, MF = 0⟩ + i∣F = 0,MF = 0⟩)/√2‾. Immediately following such a pulse, theexpectation values ⟨Iz⟩ and ⟨Sz⟩ are both zero. Evolutionunder J- coupling leads to ⟨Iz(t)⟩ = 1⁄2sin2πJt and⟨Sz(t)⟩ = −1⁄2 sin 2πJt.

Origin of the zero-field signal

Downloaded 02 Apr 2013 to 193.198.74.30. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://www.physicstoday.org/about_us/terms

48 April 2013 Physics Today www.physicstoday.org

Zero-field NMR

proton spin of 1⁄2—essentially, the same problem asdescribed above—yields a single spectral line at1JHC ≈ 125 Hz, where the superscript indicates thatone bond separates the nuclei. Coupling to a com-bined proton spin of 3⁄2 yields a spectral line at2 × 1JHC ≈ 250 Hz.

When both methyl groups are considered, onefinds that the spins can be coupled through two,three, or four covalent bonds. Thus the 125- and250-Hz lines are split, which results in the richmulti plet structure seen in figure 3b. In that way,J-coupling produces unique spectra for dozens ofsmall molecules and provides a new, zero-fieldmethod for chemical identification.

Little fields mean a lotAlthough zero-field J-coupling spectroscopy canshed valuable light on a sample’s chemical makeup,even more information can be gleaned if a tiny, nano -

tesla magnetic field is applied.10 To understand how,consider figure 4a, which shows energy levels as afunction of magnetic field for a 13C–proton systemwith J-coupling frequency J = 220 Hz. In zero field,the triplet states are degenerate, separated in energyfrom the singlet state by hJ. A small field, however,lifts the degeneracy. As seen in figure 4b, the spectralline associated with the singlet-to-triplet transitionsplits into two, and the zero-frequency line associ-ated with transitions between triplet states shifts up-ward and eventually splits. (In more complex mole-cules, the zero-field eigenstates span larger values ofangular momentum and therefore lead to more com-plex splitting patterns.) When the magnetic field issufficiently strong that the Zeeman effect is muchlarger than the J-coupling, the spectrum reduces toa set of two doublets, each split by J. Such near-zero-field splitting patterns can be used as a diagnostictool to determine the types of transitions and degen-eracies associated with zero-field NMR lines.

Parahydrogen-induced polarizationThe spins in the previous section were polarized viathermal equilibration in a permanent magnet. How-ever, there are several alternative means for polar-izing nuclear spins, and some of them can yield polarizations much higher than one could achievewith even the strongest magnets. With dynamic nuclear polarization, for instance, spin polarizationis transferred from easily polarizable electrons tonuclei by way of a process known as cross relax-ation; with spin-exchange optical pumping, polar-ization is transferred from photons to nuclei by wayof light–atom interactions and interatomic colli-sions. In theory, both techniques can be used toachieve nuclear spin polarizations approaching100%; spin-exchange optical pumping is generallymost effective on nuclei of noble gases such as helium-3 and xenon-129.

Recently we have begun conducting zero-fieldNMR experiments8 using what is known as parahydrogen-induced polarization (PHIP).3 Below30 K, molecular hydrogen in equilibrium is almostexclusively parahydrogen, the nuclear spin- singletstate (∣↑↓⟩ − ∣↓↑⟩)/√2‾. That is a consequence of the spin- statistics theorem, which demands that the totalwavefunction be antisymmetric for fermions (andsymmetric for bosons). Because a large energy—equivalent to about 170 K—separates the molecule’slowest and second lowest rotational levels, the cooledmolecular hydrogen is confined to the lowest rota-tional wavefunction, which has even symmetry. To sat-isfy spin statistics, the molecules must have an anti-symmetric nuclear spin wavefunction. Con veniently,parahydrogen has an extremely long lifetime, on theorder of weeks, even at room temperature.

The parahydrogen state itself has zero mag-netic moment and does not produce an NMR signal.How then can one take advantage of its high degreeof spin order? The symmetry can be broken by thecatalytic addition of parahydrogen to a substratemolecule containing a nonzero nuclear spin such asthat of 13C or fluorine-19. Consider, for example, thereaction H2 + R=CH2 → H−R−CH3, where R is somechemical group. The parahydrogen-derived pro-

−40 4020−20 0

0

500

−500

1000

−1000

EN

ER

GY

/(H

z)h

Bz (µT)

0

0 0 0 BzγS BzγI

0 0J

J J

J J

SIG

NA

L

FREQUENCY

ζ = 0

ζ = 0

ζ = 0.01

ζ = 1

ζ = 0.05

ζ = 10.2

∣↓ ⟩↓∣↓ ⟩↓

∣↓ ⟩↑∣↓ ⟩↑

∣↑↓⟩

∣↑↑⟩

b

a

Figure 4. In a two-spin system with spins I and S, energy levels (a)and magnetic resonance spectral lines (b) shift and split as an ex-ternal magnetic field Bz is applied. When the normalized magneticfield ζ = Bz(γI − γS)/J = 0, the system consists of a singlet state anddegenerate triplet states; the system produces a spectral line at Jcorresponding to singlet-to-triplet transitions and a line at zero fre-quency corresponding to the static component of magnetization.(Here γ is the gyromagnetic ratio, and J is the J-coupling frequency.)As the magnetic field increases, the line at J splits into a doubletand the line at zero frequency moves to the right and eventuallysplits. For ζ ≫ 1, the spectrum is dominated by the Zeeman effectand consists of two doublets centered at BzγI and BzγS, each split by J. Note the change in the x-axis scale in the last two panels.

Downloaded 02 Apr 2013 to 193.198.74.30. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://www.physicstoday.org/about_us/terms

www.physicstoday.org April 2013 Physics Today 49

tons occupy chemically distinct sites on the productmole cule and therefore have different couplingswith the 13C spin.

Figures 5a and 5b depict a model system con-sisting of a parahydrogen and a 13C nucleus. Priorto the reaction, the parahydrogen protons occupythe singlet state. When the two protons attach to the substrate molecule, one couples strongly to the13C nucleus, forming a two-spin system with singletand triplet manifolds. The other proton is weaklycoupled to the strongly coupled system, and splitsthe singlet and triplet into a quartet and two dou-blets. In the limit where the strong coupling is muchgreater than the other couplings in the system, thepopulation of the upper doublet is three times thatof the lower one. The doublet and quartet states arestationary with respect to the J- coupling Hamilton-ian. However, owing to the differing gyromagneticratios of the 13C and 1H nuclear spins, coherent superpositions of eigenstates with equal values ofMF can be prepared by applying a magnetic fieldpulse, as described in the box. As the simulatedspectrum in figure 5c illustrates, the resulting oscil-lations between doublet and quartet states yieldthree distinct spectral lines.

Room to growLow- and zero-field NMR are best viewed as comple-ments to traditional high-field spectroscopy and im-aging. Over the years, high-field NMR has evolvedinto an incredibly powerful and versatile tool in re-search and industry, but low- and zero-field methodsare likely to find important niche applications: Be-cause they are portable, relatively inexpensive, andminimally invasive, the methods may enable mobiletools for real-time chemical analysis in microfluidiclab-on-a-chip devices. In that context, atomicmagneto meters and SQUIDs might eventually giveway to magnetic sensors based on nitrogen- vacancycenters in diamond.17 Such sensors are small andhighly sensitive and can operate in a broad temper-ature range. They’ve recently been used to measure

magnetic fields generated by nanoscale samples con-taining just a few thousand atoms (see page 12 of thisissue). In a different vein, hyperpolarized moleculesundergoing J-coupling-induced quantum beats mayserve as a new type of polarized target for nuclear-physics experiments.18

Low- and zero-field NMR are still relativelyyoung fields, and plenty of work lies ahead. In ad-dition to the universal and perpetual challenges of boosting the sensitivity and shrinking the size of detectors, multidimensional zero-field spectros -copy techniques need to be developed to analyzethe exceedingly complex spectra of biologically rel-evant molecules.We gratefully acknowledge support from the NSF. We also thank Alexander Pines, Max Zolotorev, Brian Patton,Bogdan Wojtsekhowski, Bernhard Blümich, SzymonPustelny, Louis Bouchard, and Yasuhiro Shimizu for stim-ulating discussions and useful feedback.

References1. D. P. Weitekamp et al., Phys. Rev. Lett. 50, 1807 (1983);

D. Zax et al., Chem. Phys. Lett. 106, 550 (1984). 2. T. G. Walker, W. Happer, Rev. Mod. Phys. 69, 629 (1997).3. C. R. Bowers, D. P. Weitekamp, Phys. Rev. Lett. 57, 2645

(1986). 4. R. McDermott et al., Science 295, 2247 (2002). 5. M. P. Ledbetter et al., J. Magn. Reson. 199, 25 (2009). 6. Y. S. Greenberg, Rev. Mod. Phys. 70, 175 (1998). 7. I. Savukov, S. Seltzer, M. Romalis, J. Magn. Reson. 185,

214 (2007).8. T. Theis et al., Nat. Phys. 7, 571 (2011). 9. J. Bernarding et al., J. Am. Chem. Soc. 128, 714 (2006).

10. M. P. Ledbetter et al., Phys. Rev. Lett. 107, 107601 (2011). 11. M. Packard, R. Varian, Phys. Rev. 93, 941 (1954). 12. J. N. Robinson et al., J. Magn. Reson. 182, 343 (2006);

S. Appelt et al., Nat. Phys. 2, 105 (2006).13. S.-K. Lee et al., Appl. Phys. Lett. 89, 214106 (2006). 14. E. L. Hahn, D. E. Maxwell, Phys. Rev. 84, 1246 (1951). 15. H. S. Gutowsky, D. W. McCall, C. P. Slichter, Phys. Rev.

84, 589 (1951). 16. N. F. Ramsey, E. M. Purcell, Phys. Rev. 85, 143 (1952). 17. J. M. Taylor et al., Nat. Phys. 4, 810 (2008). 18. D. Budker et al., Nucl. Instrum. Methods Phys. Res. A

694, 246 (2012). ■

cba

0 1 2 3 4 138 139 140 141 142FREQUENCY (Hz)

SIG

NA

L

2

31

EN

ER

GY

EN

ER

GY

H2 Three-spin system

2

1

3

Figure 5. Parahydrogen-induced polarization can be used to obtainnuclear magnetic resonance signals in the absence of a magnetic field, as depicted here for a hypothetical three-spin system consisting of a carbon-13 nucleus and the nuclei of a parahydrogen molecule. (a) In iso-lation, the antiparallel spins in the parahydrogen molecule correspond tothe singlet state. (b) If the molecule is catalytically added to a substrate

molecule containing 13C, and if one of the C–H couplings is much stronger than the other couplings in the system, the symmetry ofthe parahydrogen spins is broken and in the newly formed three-spin system, the population of the upper doublet is about threetimes that of the lower one. (Here, we ignore the rotational energies that may be correlated with the nuclear state.) The horizontallines represent magnetic sublevels and the red rectangles represent the expected populations in each sublevel. (c) The simulatedspectrum of a system with strong C–H coupling JCH = 140 Hz, weak C–H coupling JCH = −5.2 Hz, and H–H coupling JHH = 7.7 Hz yieldsthe three peaks shown here, which correspond to the three allowed transitions indicated by the dashed arrows in panel b.

Downloaded 02 Apr 2013 to 193.198.74.30. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://www.physicstoday.org/about_us/terms