A Noiseless Quantum Optical Memory at Room Temperature

Abstract

A quantum optical memory (QM) is a device that can store and release quantum states of light on demand. Such a device is capable of synchronising probabilistic events, for example, locally synchronising nondeterministic photon sources for the generation of multi-photon states, or successful quantum gate operations within a quantum computational architecture [1], as well as for globally synchronising the generation of entanglement over long distances within the context of a quantum repeater [2]. Desirable attributes for a QM to be useful for these computational and communicational tasks include high end-to-end transmission (including storage and retrieval efficiency), large storage-time-bandwidth product, room temperature operation for scalability and, of utmost importance, noise free performance for true quantum operation.Impressive realisations of QMs have materialised based on optical transitions in atomic systems [3][4][5][6][7]. However, several issues remain that prevent these devices from being used within large-scale networks. These include: elaborate cold atom [3,4] or cryogenic [5] experimental setups, complex preparation of the atomic system [5], additional loss through filtering required for noise suppression, and noise photons being induced in the same frequency, spatial or temporal mode as the output thus reducing the quality of the QM readout [3][4][5][6][7].Here we present a new QM protocol that addresses the above issues, the quantum ladder memory (QLAD). This protocol is based on a two-photon 'ladder' transition between the 6S 1/2 ground state and 6D 5/2 excited state of a caesium (Cs) ensemble at room temperature (see Fig. 1a). We characterise the memory with weak coherent states, storing GHz-band pulses with η=22% storage and retrieval efficiency and a characteristic storage time of τ=5.3ns. We measure the noise at the output to be 8×10 -6 photons per pulse, giving rise to a noise-to-efficiency ratio of 3.6×10 -5 , three orders of magnitude better than any memory to date. We generate GHz-band heralded single photons from a spontaneous parametric down conversion source that are matched to the QLAD, coincidence traces seen in Fig. 1b. A heralded second-order autocorrelation function of g (2) =0.02±0.005 is measured for the input photon, and true quantum operation is confirmed by storing and retrieving this photon with g (2) =0.03±0.01 for a 3.5ns storage time. This result is the lowest g (2) on the output of an on-demand memory to date, and represents a significant step toward large-scale quantum technologies. Fig. 1 (a) QLAD protocol in Cs level structure. (b) Coincidence histogram of storage and retrieval of heralded single photons. The blue (yellow) histogram represents the memory off (on), labelled Input (Output). (c) Log scale with Output and Noise (dark red histogram).

Field-deployable Quantum Memory for Quantum Networking

“…Successful products have already been developed with widespread deployment, such as chip-scale atomic clocks [46], atomic magnetometers [47], atomic interference gravimeters [48], etc. There have been a plethora of research activities on warm-vapor based QM, and very promising memory performances have been demonstrated, such as 87% efficiency [49], 1 second storage time [50], and close-to-unity fidelity [51,52].…”

“…Although they each work under very distinct conditions, the inherent physics and performance can be obtained in a similar way by modeling a three-level system [25,59]. While the high quantum performance of the individual parameters of {η, F, T } have been demonstrated in these systems [49][50][51][52], simultaneous high performance of all three parameters has not been realized. This often-overlooked gap in the research space is due to the conflicting physical processes which arise during the optimization of the experimental system.…”

“…For example, getting high fidelity requires exquisite filtering of the noise photons due to the strong control field (an external optical field modifies the optical response of the atomic medium). This can be achieved by angling the control field in a Λ-system [60] or having two fields counter-propagating in a laddersystem [51,52]. However, the storage time is severely limited in both of those approaches, due to either the short spin-wave (SW) wavelength (Λ-system) or rapid dephasing due to the excited state (ladder-system).…”

Variance-based sensitivity analysis of Λ -type quantum memory

“…In the following sections, we restrict our discussion to resonant Λ-type memory protocols, but the tools developed in this work are readily applicable to off-resonant protocols, as well as other level systems and a wide range of related techniques [46][47][48]. In Section II, we provide definitions for several quantitative aspects of memory sensitivity.…”