Super-resolving frequency metrology via mode-selective quantum memory

High-precision optical frequency measurement is essential for probing atomic and molecular structures, as well as for applications in astronomy and medical imaging. Conventional spectroscopic techniques such as diffraction grating spectrometers are limited by the Rayleigh criterion, which is manifested as a vanishing Fisher information when spectral features become closely spaced. Recent advances in quantum metrology have challenged these classical limits by projecting the light field onto quantum-optmized mode bases. Here we introduce a mode-selective atomic Raman quantum memory implemented in warm cesium vapor to achieve super-resolving estimation in spectral domain. To estimate the separation between two spectral lines, we experimentally measure the mean squared error of the frequency estimate, achieving a sensitivity of 1/20 of the linewidth with a (34±4)-fold enhancement in precision over direct intensity measurements. This enhanced ability to resolve closely spaced spectral features paves the way for memory-based time-frequency sensors and their integration within quantum networks.