SQUID-Detected Liquid State NMR in Microtesla Fields Andreas H. Trabesinger, §,² Robert McDermott, ⊥,‡ SeungKyun Lee, ⊥ Michael Mu 1 ck, ⊥ John Clarke, ⊥ and Alexander Pines* ,§ Departments of Chemistry and Physics, UniVersity of California, Berkeley, and Materials Sciences DiVision, Lawrence Berkeley National Laboratory, Berkeley, California 94720 ReceiVed: May 1, 2003; In Final Form: NoVember 12, 2003 Nuclear magnetic resonance (NMR) experiments performed in magnetic fields on the order of microtesla yield line widths comparable to the lifetime limit even in grossly inhomogeneous magnets. The potential loss in sensitivity is overcome by combining prepolarization in fields on the order of millitesla and signal detection with a Superconducting Quantum Interference Device (SQUID). The enhanced spectral resolution attainable in microtesla fields enables NMR studies of pure liquids and solutions without the need for strong magnets. We have investigated a variety of heteronuclear systems in both the weak and strong J-coupling regimes. Six different nuclear species have been detected with the same experimental apparatus. NMR signals of thermally polarized protons were obtained in fields as low as 554 nT. Introduction Throughout the history of nuclear magnetic resonance (NMR) spectroscopy and magnetic resonance imaging (MRI), there has been a drive to higher and higher magnetic field strengths. This drive has been fueled by the need for increased spin polarization, improved detection sensitivity, and broader chemical shift dispersion. Currently the majority of NMR and MRI studies utilize magnetic fields in the range of tesla to tens of tesla. Nevertheless, since the very early days of NMR, researchers have explored the possibility of performing NMR experiments in much lower magnetic fields, on the order of the Earth’s field (∼50 μT). In 1954, the Varian laboratories reported switched fields experiments with detection in the Earth’s magnetic field. This initial work encompassed high precision measurements of the strength of the Earth’s magnetic field, the simultaneous detection of 1 H and 19 F signals, 1 as well as relaxation time measurements in low field. 2 These early experiments inspired researchers to study geomagnetism, 3 to perform geophysical investigations of seawater, 4 groundwater, 5 and Antarctic ice, 6 and to evaluate sugar content in plants. 7 These rather exotic applications involved the investigation of (very) bulky samples that cannot be inserted into a magnet. But in addition they capitalized on instrumental simplicity (compared to high-field setups); it was realized that with relatively inexpensive and easy-to-maintain equipment one can study a variety of NMR phenomena, including J-coupling in pure liquids, 8 and relaxation of body fluids. 9 These experiments were considered a useful supplement to high-field NMR, and promised the perspective of making NMR more mobile, taking it out of the traditional laboratory environment. In the past two decades, low-field MRI applications have attracted particular attention. 10 Apart from the substantial reduction in cost and complexity, advantages here include improved T 1 (longitudinal relaxation time) contrast, as well as the elimination of distortions due to spurious gradients in the case of samples with inhomogeneous magnetic susceptibility. 11 Despite considerable interest and continued effort, however, low-field NMR remains more of a curiosity than a practical diagnostic tool. The principal obstacle to low-field studies is the inherently low sensitivity. In a conventional pulsed NMR experiment, the static magnetic field serves a dual purpose, as both polarizing field and detection field. This leads to a quadratic dependence of the NMR signal strength on the magnitude B 0 of the static field: for a nuclear moment μ, the thermal magnetization μB 0 /k B T scales linearly with the strength of the polarizing field, while the voltage induced in the receiver coil, via Faraday’s law, scales with Larmor frequency, and hence with the strength of the detection field. In the majority of previous low-field NMR studies, the low sensitivity often necessitated sample volumes on the order of liters, and thus severely limited the range of possible useful applications. In general, to perform practical NMR experiments in low field it is necessary to address both the problems of low thermal magnetization and the frequency-dependent response of the Faraday detector. Concepts In magnetic fields of the order of microtesla, thermal polarizations are extremely small, of the order of 10 -11 . Nuclear polarizations can be significantly enhanced, however, by pre- polarizing the spins in a strong transient field. 12 Prepolarization in a field on the order of millitesla leads to an enhancement of spin magnetization by 3 orders of magnitude. The enhanced magnetization is available in the detection field for a time comparable to T 1 . As slight variations in the local polarization over the sample volume have negligible effect on the detected signal, the demands on the homogeneity of the polarization field are insignificant. In comparison with other, nonthermal polariza- tion modalities such as dynamic nuclear polarization (DNP) 13 or optical pumping, 14 thermal prepolarization offers the advan- tage that all NMR-active nuclei are polarized, with Curie’s law * Address correspondence to this author. § Department of Chemistry. ⊥ Department of Physics. ² Present address: Physical Chemistry Laboratory, ETH Zurich, Zurich, Switzerland. ‡ Present address: National Institute of Standards and Technology, Boulder, CO 80305. 957 J. Phys. Chem. A 2004, 108, 957-963 10.1021/jp035181g CCC: $27.50 © 2004 American Chemical Society Published on Web 01/17/2004