Light storage in a room-temperature atomic vapor based on coherent population oscillations

We report the experimental observation of coherent population oscillation (CPO-) based light storage in an atomic vapor cell at room temperature. Using the ultranarrow CPO between the ground levels of a $\Lambda$ system selected by polarization in metastable ${}^4\mathrm{He}$, such a light storage is experimentally shown to be phase preserving. As it does not involve any atomic coherences it has the advantage of being robust to dephasing effects such as small magnetic field inhomogeneities. The storage time is limited by the population lifetime of the ground states of the $\Lambda$ system.

Figure 1

Experimental setup for EIT or ultranarrow CPO storage in metastable ${}^4\mathrm{He}$. The orthogonally polarized coupling and probe beams of optical frequencies ${\omega }_c$ and ${\omega }_p$ and Rabi frequencies ${\Omega }_c$ and ${\Omega }_p$, respectively, are separated or recombined with polarizing beam splitters (PBS). They are controlled in frequency and amplitude by acousto-optic modulators $({\mathrm{AO}}_c$ and ${\mathrm{AO}}_p)$. $\lambda /4$ plates can be added to generate circular polarizations (EIT configuration). A $\mu$-metal shielding protects the cell from stray magnetic fields. A solenoid can provide a longitudinal $B$ field. A piezoelectric transducer is used for homodyne detection. Inset: Recorded leak and retrieved signals after data processing.

Figure 2

(a) $\text{circ}\perp \text{circ}$: ${\sigma }^{+}$ coupling beam of optical and Rabi frequencies ${\omega }_c$ and ${\Omega }_c$ and ${\sigma }^{-}$ probe beam of optical and Rabi frequencies ${\omega }_p$ and ${\Omega }_p$; (b) $\text{lin}\perp \text{lin}$ configurations with a magnetic field that lifts the Zeeman degeneracy by a quantity ${\Delta }_Z$. (c) $\text{circ}\perp \text{circ}$ configuration: For a zero magnetic field, EIT occurs for coupling and probe beams of the same frequency (in black); for a 17 mG longitudinal magnetic field, the EIT window is shifted by $2{\Delta }_Z=100$ kHz (in red). (d) $\text{lin}\perp \text{lin}$ configuration: At zero magnetic field, the resonance is due both to EIT and CPO (in black); with a longitudinal magnetic field, EIT resonances occur for $\pm 2{\Delta }_Z$ coupling and probe frequency detunings while the central transmission resonance (same coupling and probe frequency), where CPO storage is performed, is due to ultranarrow CPO (in red) [10].

Figure 3

Transmission profiles (a) inside the $\mu$-metal and (b) at the edge of the $\mu$-metal shielding in (i) the $\text{lin}\perp \text{lin}$ configuration and a 0.7 G longitudinal magnetic field (in red); (ii) the $\text{circ}\perp \text{circ}$ configuration and a 0.7 G longitudinal magnetic field (in gray); (iii) the $\text{circ}\perp \text{circ}$ configuration and no longitudinal magnetic field (in black). In case (ii), there is no EIT resonance for coupling and probe beams of the same frequency: In case (i) with linear polarizations, the resonance is thus only due to CPO.

Figure 4

Measured storage efficiency as a function of the storage time, for a 4 $\mu \mathrm{s}$ rise-time exponential pulse, when the cell is inside the $\mu$-metal shielding. CPO storage efficiencies (red dots) are recorded in the $\text{lin}\perp \text{lin}$ configuration with a $B=0.7$ G longitudinal magnetic field. EIT storage (open black squares) measurements are performed in the $\text{circ}\perp \text{circ}$ configuration. The full line is an exponential fit with a 10 $\mu \mathrm{s}$ decay time constant.

Figure 5

Measured storage efficiency as a function of the storage time, for a 4 $\mu \mathrm{s}$ rise-time exponential pulse when the cell is at the edge of the $\mu$-metal shielding. CPO storage efficiencies (red dots) are recorded in the $\text{lin}\perp \text{lin}$ configuration with a $B=0.7$ G longitudinal magnetic field. EIT storage (open black squares) measurements are performed in the $\text{circ}\perp \text{circ}$ configuration with a 1 GHz optical detuning, which gives better efficiencies [12]. Continuous lines are exponential fits: When compared to Fig. 3, the decay time for CPO storage remains the same (about 10 $\mu \mathrm{s})$ but it strongly decreases down to 0.6 $\mu \mathrm{s}$ for EIT storage.

Figure 6

Interference signal for two different positions of the piezoelectric transducer. The leak and the retrieval interfere in the same way: The black curve shows a constructive interference for both pulses and the gray curve a destructive interference for both pulses.