Generating indistinguishable photons from independent nodes is an important challenge for the development of quantum networks. In this work, we demonstrate the generation of highly indistinguishable single photons from two dissimilar atomic quantum nodes. One node is based on a fully blockaded cold Rydberg ensemble and generates on-demand single photons. The other node is a quantum repeater node based on a Duan-Lukin-Cirac-Zoller quantum memory and emits heralded single photons after a controllable memory time that is used to synchronize the two sources. We demonstrate an indistinguishability of 94.6 ± 5.2 % for a temporal window including 90 % of the photons. This advancement opens new possibilities for interconnecting quantum repeater and processing nodes with high-fidelity Bell state measurement without sacrificing its efficiency. Published by the American Physical Society 2024
Generating indistinguishable photons from independent nodes is an important challenge for the development of quantum networks. In this work, we demonstrate the generation of highly indistinguishable single photons from two dissimilar atomic quantum nodes. One node is based on a fully blockaded cold Rydberg ensemble and generates on-demand single photons. The other node is a quantum repeater node based on a Duan-Lukin-Cirac-Zoller quantum memory and emits heralded single photons after a controllable memory time that is used to synchronize the two sources. We demonstrate an indistinguishability of 94.6 ± 5.2 % for a temporal window including 90 % of the photons. This advancement opens new possibilities for interconnecting quantum repeater and processing nodes with high-fidelity Bell state measurement without sacrificing its efficiency. Published by the American Physical Society 2024
Abstract Hybrid quantum systems play a crucial role in advancing scalable and versatile quantum networks as they combine the strengths of different quantum platforms. An important challenge for the development of hybrid quantum networks lies in interfacing heterogeneous quantum nodes and distributing entanglement among them. Single photons emitted from these dissimilar quantum nodes typically show distinct spectral and temporal properties. Therefore, they necessitate spectral filtering and temporal synchronization, which introduce significant photon losses and require additional resources. In this work, we successfully generate indistinguishable photons from two distinct quantum systems of a warm atomic ensemble and a solid-state quantum dot. Remarkably, quantum interference between dissimilar sources is achieved without additional spectral filtering and time synchronization, which enables autonomous quantum nodes for a hybrid quantum network. 133Cs atomic ensemble can efficiently generate heralded single photons at the wavelength of 917 nm of the 6P_(3/2)-6D_(5/2) transition, while the single photons emitted from an InAs/GaAs quantum dot can be tuned to match the 133Cs transition wavelength. Our dense warm atomic ensemble and cavity-coupled quantum dot can efficiently generate bright and resonant single photons at detection rates approaching MHz, respectively. More importantly, these single photons exhibit inherent spectral similarities not only in the wavelength but also in the spectral linewidth, achieving a high spectral overlap of 0.92. Such intrinsic compatibility between dissimilar quantum sources is essential to leverage the advantages of different quantum platforms, paving the way toward a large-scale and functional hybrid quantum network.
Remote entanglement of matter qubits plays an important role in quantum networks and quantum repeater. Collective excitations generated via Rydberg blockade in an atomic ensemble-known as Rydberg superatoms-are promising candidates due to collectively enhanced atom-photon coupling and Rydberg-enabled nonlinearity. Here, we experimentally realize remote entanglement between two Rydberg superatoms via single-photon interference. The two setups are separated by 3 m and linked with two 20-m fibers, operated with independent control lasers. We verify that the entanglement generated is free from high-order excitation noise that is a key drawback in the traditional Duan-Lukin-Cirac-Zoller scheme. This Letter paves the way for entangling long-distance separated Rydberg superatoms at a high rate, as the success probability of a single-photon-based entangling scheme scales with the square root of the channel transmittance.
With the accelerating developments in the field of quantum communication, there will be a need for sources of highly quantum single-photons. In addition to communication applications, the purity, brightness, and fidelity of a given single-photon source are intrinsically linked to the success rate and scalability of a range of photonic technologies in quantum computing, metrology, memory, and cryptography. Further, quantum repeater architectures for maintaining quantum information over long distances will become crucial. In these schemes, end-to-end entanglement is generated by establishing entanglement between the ends of several elementary segments, followed by successive entanglement-swapping operations to increase the entanglement distance until it spans the transmission channel. The most widely studied proposal for the practical implementation of a quantum repeater is the Duan-Lukin-Cirac-Zoller (DLCZ) protocol. This consists of off-resonant Raman excitation by a write pulse, which induces a spin-wave excitation in an atomic ensemble and emits a Stokes photon. At a later time, this spin-wave may be read with a high-power resonant pulse, resulting in the emission of an anti-Stokes photon into a well-defined spatial mode. The Stokes photon may be used for entanglement with another ensemble, while the anti-Stokes may be read out for an entanglement-swapping operation. Notably, in the DLCZ protocol, the atomic ensemble serves as both the quantum memory and a heralded single-photon source. Using atomic ensembles for single-photon sources enables the production of highly pure, indistinguishable photons with minimal spectral broadening; however, the generation rate remains a challenge. In the case of DLCZ, the Raman-scattering step is highly probabilistic, reducing the anti-Stokes generation rate considerably. Furthermore, DLCZ requires operation in the low-excitation-probability regime to avoid multiphoton noise contributions. Additionally, the profile of the induced spin wave is suboptimal for readout. To compensate for these drawbacks, in this thesis we report the experimental realization of a variant of DLCZ that leverages collectively enhanced superradiant (SR) emission. Superradiant emission is characterized by intense bursts of photons emitted in a shorter amount of time than independent atomic state decay. By switching to a resonant excitation and optimizing the write profile, the induced spin-wave is primed for readout, thus increasing retrieval efficiency and the level of quantum correlation. We report data for both the DLCZ and SR schemes under similar optical conditions and compare the two. At high write power, the SR protocol exhibits conditional readout probabilities as high as $\sim$1.2\%, compared to $\sim$0.6\% for the DLCZ protocol under the same conditions. We also analyze the temporal distribution of anti-Stokes emission and find a degree of maintenance of the spin-wave profile, even in the anti-Stokes direction.
In the famous quantum communication scheme developed by Duan et al. [L.M. Duan, M.D. Lukin, J.I. Cirac, and P. Zoller, Nature (London) 414 (2001) 413 ], the probability of successful generating a symmetric collective atomic state with a single-photon emitted have to be far smaller than 1 to obtain an acceptable entangled state. Based on strong dipole-dipole interaction between two Rydberg atoms, two simultaneous excitations in an atomic ensemble are greatly suppressed, which makes it possible to excite a mesoscopic cold atomic ensemble into a near-ideal singly-excited symmetric collective state accompanied by a signal-photon with near unity success probability.
Photons are good candidates for carrying quantum information because they are very stable particles: they interact weakly with the medium and barely with each other. However, this has drawbacks when you want to process the information because, in this case, it is preferable to have photon-photon interactions. For example, for applications in quantum repeaters, such interactions would allow deterministic Bell state measurements, increasing the entanglement distribution rate between two remote nodes. Getting two photons to interact with each other efficiently requires mapping them into a nonlinear medium at the single-photon level, that is, a medium that reacts differently when it interacts with a single photon than when it does with two. Such strong nonlinearity has been demonstrated with Rydberg atoms, which are atoms excited to a state with a high principal quantum number.
In this thesis we have performed nonlinear quantum optics experiments using an ensemble of cold Rydberg atoms, where we have studied the properties of the quantum light emitted by these atoms. First, we demonstrated nonlinearities at the single-photon level. To reach this stage, we made several improvements to the previous experimental setup available in the group, of which the implementation of a dipole trap was especially relevant. We evidence quantum nonlinearity by measuring photon antibunching for the transmitted light after interacting with the Rydberg state under electromagnetically induced transparency (EIT). We also showed the generation of single photons on-demand after storing weak coherent states of light pulses as collective Rydberg excitations.
Then, we studied the variation of the light statistic throughout the output pulse after propagating through the medium as Rydberg polaritons, which are superposition states of light and Rydberg excitations. We showed that the properties at the beginning and the end of the pulse were different from those of the steady state. In particular, the light detected after the input pulse was abruptly turned off gave much stronger suppression of two-photon events. Then, we investigated how to exploit this effect to generate single photons on demand. To do this, we analyzed the quality of the single photons detected at the end of the pulse as a function of the detection probability and compared the results with those obtained by storing the input pulse as collective Rydberg excitations. We showed that the photons were generated more efficiently when increasing the detection window at the cost of deteriorating the single photons statistics.
Finally, we investigated the indistinguishability of the photons emitted by our Rydberg atomic ensemble, a crucial property for using Rydberg atoms as nodes in quantum networks. We also compared the single photons generated after storage under EIT conditions with those obtained using a two-photon Raman excitation to the Rydberg state. We measured the indistinguishability by making them interfere with weak coherent states of light in a Hong-Ou-Mandel experiment. And we showed that, although we obtained better photon statistics under EIT conditions, the indistinguishability from those obtained with Raman excitation was significantly higher.