nphys566 Budker Review.indd

It has been nearly half a century since optical pumping techniques pioneered by Kastler1, Dehmelt2, and Bell and Bloom3,4 have been applied to sensitive measurements of magnetic fi elds5,6. Th e general idea of the method is that light that is near-resonant with an optical transition creates long-lived orientation and/or higher-order moments in the atomic ground state, which subsequently undergo Larmor spin precession in the magnetic fi eld. Th is precession modifi es the optical absorptive and dispersive properties of the atoms, and this modifi cation is detected by measuring the light transmitted through the atomic medium. Recent reviews of resonant magneto-optics have been given in refs 7,8. Th e fi elds of resonant magneto-optics and atomic magnetometry have been experiencing a new boom driven by technological developments, specifi cally by the advent of reliable, small, inexpensive, and easily tunable diode lasers, and by the refi nement of the techniques of producing dense atomic vapours with long (in some cases ~1 s) ground-state relaxation times. Th ese advances have enabled atomic magnetometers to achieve sensitivities rivaling9–11 and even surpassing12 that of most superconducting quantum interference device (SQUID)-based magnetometers that have been leading the fi eld of ultrasensitive magnetic fi eld measurements for a number of years13. As a result, optical magnetometers are starting to explore some of the applications that have previously been in the exclusive domain of SQUID magnetometers. Atomic magnetometers have the intrinsic advantage of not requiring cryogenic cooling, and off er a signifi cant potential for miniaturization. In contrast to SQUIDs, which measure magnetic fl ux through a pick-up loop, atomic magnetometers measure magnetic fi eld directly and can be used to detect other spin interactions. Th ey can be confi gured so that their output is directly related to the absolute magnitude of the magnetic fi eld through fundamental physical constants, so that no calibration is required. In contrast, SQUID magnetometers are relative fi eld sensors that can also be confi gured for direct measurement of magnetic fi eld gradients. Currently, the most sensitive atomic optical magnetometer is the spin-exchange relaxation-free (SERF) magnetometer, whose demonstrated sensitivity exceeds 10–15 T Hz–1/2, with projected fundamental limits below 10–17 T Hz–1/2 (ref. 12). SERF magnetometers also off er a possibility of spatially resolved measurements with millimetre resolution. Th e present-day interest in optical magnetometers is driven by numerous and diverse applications, a partial list of which includes tests of the fundamental symmetries of nature, search for man-made and natural magnetic anomalies, investigation of the dynamics of the geomagnetic fi elds (including attempts at earthquake prediction), investigation of the magnetic properties of rocks, detection of magnetic microparticles at ultralow concentrations, detection of signals in NMR and MRI, direct detection of magnetic fi elds from the heart and the brain, magnetic microscopy, and measuring magnetic fi elds in space. In this review we describe the basic principles and fundamental limits of the sensitivity of optical atomic magnetometers, and discuss several specifi c applications.