A leap forward for diamond-defect nuclear magnetic resonance

NV centers—point defects in diamond in which two neighboring carbon atoms are substituted with a nitrogen atom and a vacancy—are exquisitely sensitive magnetometers. That’s because an NV center’s fluorescence response depends on its spin state, which, in turn, is sensitive to external magnetic fields. Embedded a few nanometers deep inside a diamond, the defect can detect fields as small as a few nanotesla at the surface.

Five years ago, researchers showed that the diminutive magnetometers could be used to perform nuclear magnetic resonance spectroscopy. (See Physics Today, April 2013, page 12 .) In NMR, the nuclear spins of molecules in a target sample—say, a soup of proteins—are polarized and manipulated with RF pulses to produce a faint magnetic signal. Typically, that signal is detected with induction coils. But due to their limited sensitivity, such coils require micron-scale or larger samples. With NV-center magnetometers, the researchers could detect NMR signals from nanometer-scale samples. Unfortunately, however, the technique’s spectral resolution wasn’t good enough to identify the key markers of molecular structure: scalar couplings and chemical shifts.

Now a Harvard University team led by Mikhail Lukin , Hongkun Park , and Ronald Walsworth has improved the resolution of NV-center NMR a hundredfold. The trick was to synchronize the NMR pulses and measurements to an external clock, which made it possible to average thousands of repeat measurements on the same sample—and thereby achieve effective measurement times orders of magnitude longer than the NV center’s millisecond spin-coherence time. Combined with a noise-minimizing device design, the scheme delivered spectral resolution sharp enough to detect scalar couplings and chemical shifts in all three organic samples the team tested. (The accompanying image shows the device in action.) The group anticipates that the technique could one day be used to perform NMR spectroscopy on single living cells. (D. R. Glenn et al., Nature 555, 351, 2018 .)