Diffraction Unlimited All-Optical Recording of Electron Spin Resonances

We show stimulated emission depletion microscopy to break the diffraction limit in the all-far-field-optical detection of magnetic fields and resonances. Electron spin resonances from single nitrogen-vacancy centers in diamond located at subdiffraction proximities are fully discerned. Since diffraction is overcome by disallowing the signaling state through an optical transition such as stimulated emission, the spin state remains unaffected and amenable to microwave manipulation. Stimulated emission depletion presents a universal scheme for superresolving spin resonances detectable by fluorescence.

Figure 1

STED nanoscope for optical detection of magnetic resonances of NV centers. Excitation by a pulsed (100 ps) laser diode, synchronized with a laser for STED, providing down-chirped pulses at 775 nm of 200 ps duration; the laser intensity is modulated by an acousto-optical modulator (AOM); the STED focal doughnut is produced by imprinting a $0–2\pi$ helical phase ramp. The overlaid beams are guided to the objective lens by using dichroic mirrors (DM1 and DM2) and focused into the sample (inset panels show focal intensity spots). Scanning is performed with a 3D piezo scanner of nanometric precision. Fluorescence collected by the lens is focused onto a confocalized detector (APD).

Figure 2

Superresolution far-field ODMR and ESR with Zeeman splitting induced by magnetic field $\overrightarrow{B}$. (a) Confocal image fails to resolve densely packed NV centers. (b) Confocal ESR signal acquired at the position of maximum fluorescence exhibits 4 resonance lines indicating more than one spin within the diffraction-limited area of signal origin. (c) STED image (dwell time 1 ms per pixel) resolves 5 spins, and (d) STED-ODMR reveals their ESR traces individually. Three spin orientations are found with respect to the crystal orientation, as indicated by the numbers in cyan (2 spins, splitting of 174 MHz at ${\overrightarrow{B}}_{0,1}$ and 144 MHz at ${\overrightarrow{B}}_{0,2}$ corresponding to local fields of 30 and 25 G, respectively), red [2 spins, splitting of 200 (${\overrightarrow{B}}_{0,1}$) and 30 MHz (${\overrightarrow{B}}_{0,2}$)], and green [1 spin with a splitting of 108 (${\overrightarrow{B}}_{0,1}$) and 165 MHz (${\overrightarrow{B}}_{0,2}$)]. From the splitting asymmetries we derived $|{\overrightarrow{B}}_{0,1}|=(4.46\pm 0.03)\mathrm{mT}$ and $|{\overrightarrow{B}}_{0,2}|=(3.02\pm 0.17)\mathrm{mT}$. Far-field optical magnetometry with nanometric resolution is demonstrated by recording two STED-ESR traces with different magnetic fields applied (${\overrightarrow{B}}_{0,1}$ and ${\overrightarrow{B}}_{0,2}$); the ESR spectra change strongly with the field. (e) The pulsed ESR sequence. First the ${\mathrm{NV}}^{-}$ center is initialized to the $m_s=0$ state by exposing it to green light. Afterwards a microwave $\pi$ pulse of varying frequency $\nu (n)$ is applied. The spin signal is read out with high spatial resolution by simultaneously illuminating the sample with excitation and STED light. The fluorescence occurring during the first 300 ns of exposition is recorded and compared to the reference value recorded afterwards. To increase the signal-to-noise ratio, the sequence is repeated $r$ times (typically ${10}^4$) for each microwave frequency.

Figure 3

Preservation of spin population (Rabi oscillations) by STED. (a) Pulse sequence: (i) An undisturbed Rabi oscillation yields a reference fluorescence signal; (ii) an optical excitation pulse train (EXC, 200 ns exposure, $240\mu \mathrm{W}$, 532 nm, 100 ps pulse length) succeeding the microwave pulse (MW) repolarizes the spin to $m_s=0$, thus disturbing the spin prepared by the microwave. (b) While the reference displays a nearly perfect Rabi oscillation (yellow), the oscillation is disturbed by adding the second excitation pulse (dotted blue line). If a STED pulse is applied jointly with the disturbing pulse, the oscillations are preserved (solid blue line). Det indicates the detection window and $n$ the number of iterations.