Impedance-matched cavity quantum memory

We consider an atomic frequency comb based quantum memory inside an asymmetric optical cavity. In this configuration it is possible to absorb the input light completely in a system with an effective optical depth of one, provided that the absorption per cavity round trip exactly matches the transmission of the coupling mirror (“impedance matching”). We show that the impedance matching results in a readout efficiency only limited by irreversible atomic dephasing, whose effect can be made very small in systems with large inhomogeneous broadening. Our proposal opens up an attractive route toward quantum memories with close to unit efficiency.

Figure 1

We consider a quantum memory (QM) based on an atomic frequency comb which is placed in an asymmetric optical cavity with reflectivity $R_1<R_2\approx 1$. The input and output fields ${\mathcal{E}}_{\mathrm{in}}$ and ${\mathcal{E}}_{\mathrm{out}}$ are separated by a quarter-wave plate ($\lambda /4$) and a polarization beam splitter (PBS). If the QM strongly absorbs only a particular linear polarization mode, one can also use a Faraday rotator and a half-wave plate as in Ref. [16]. The atomic comb memory is based on an inhomogeneously broadened transition, where the absorption depth $d$ is shaped into a comb structure as function of detuning $\delta$ with periodicity $\Delta$ and tooth width $\gamma$. The interaction between the atomic comb structure and a incoming light pulse leads to a coherent re-emission at $t=2\pi /\Delta$. Longer storage times can be achieved by using additional ground state levels [22, 25].

Figure 2

We here show the efficiency of an AFC-cavity quantum memory in an asymmetric cavity ($R_2=0.999$) as a function of the input mirror reflectivity $R_1$, based on Eq. (14). We show the result for different comb finesse $F_A=10$ (solid line), $F_A=6$ (dashed line), and $F_A=4$ (dashed-dotted line). The single-pass effective absorption depth was set to $\mathrm{d̃}=0.1$, which in a memory without cavity would bound the efficiency to $~1\%$ [by use of Eq. (12)]. In the figure one clearly observes the great enhancement of memory efficiency using an impedance-matched cavity, i.e., at $R_1=\exp (-2\mathrm{d̃})\approx 0.82$, reaching $\eta ~92\%$ for $F_A=10$. At this point the efficiency is only limited by irreversible atomic dephasing due to the finite comb finesse $F_A$. We also plot the reflectivity of the combined AFC-cavity system (dotted line), showing the complete absorption of light at the optimal point.