Solid-state single-photon sources operating in the telecom wavelength range

Abstract

Solid-state quantum emitters operating in the telecom wavelength range are pivotal for the development of scalable quantum information processing technologies. In this review, we provide a comprehensive overview of the state-of-the-art solid-state emitters of single photons targeting quantum information processing in the discrete-variable regime and telecom wavelength range. We focus on quantum dots, color centers, and erbium ion dopants, detailing their synthesis methods and their applications. The review addresses the strategies for the integration of these quantum emitters into photonic devices alongside the associated challenges. We also discuss their applications in quantum technologies, examining current limitations, including performance constraints, decoherence, and scalability. Finally, we propose future directions for advancing photonic-based quantum technologies.

Keywords: color centers; erbium; quantum communication; quantum light sources; rare-earth dopants; semiconductor quantum dots.

Figure 1:

Photonic quantum networks operating at telecom. (a) Optical transmission of the atmosphere (dark red) and of a 20 km-long fused silica fiber (blue). Photons in the telecommunication bands (colors) exhibit low loss in both channels. (b) Scheme of a photonic quantum network with various quantum light sources: quantum dots (QDs), Er³⁺ dopants, and color centers such as the G-, T-, or W-center in silicon.

Figure 2:

Examples of quantum dot growth techniques described in this review. (a) Droplet epitaxy QDs: atomic-scale scanning transmission electron microscopy (STEM) image in the cross-sectional geometry of InAs(P) (bright area) QDs in the InP matrix. The red line is included as a visual guide. (b) Stranski–Krastanov QDs: atomic-scale STEM image in a cross-sectional geometry of InAs (bright area) QDs in InP matrix. The red line is included as a visual guide. (c) Site-controlled QDs: schematic illustration and SEM image of an InAs QD formed at the apex of the InP pyramidal template. (d) Local droplet etching: cross-sectional atomic force microscopy profiles and plan-view images of droplet-etched holes in InAlAs filled with different amounts of InGaAs to form QDs. (e) Nanowire QDs: STEM image of InAsP QD embedded into InP nanowire. (f) QD in a 2D material: optical microscope image of a TMD monolayer flake outlined by yellow dashed lines (with a small bilayer region outlined in black) on a nano-pillar array. Inset: SEM image of a nano-pillar coated with monolayer TMD forming a QD. The scale bar of the inset is 500 nm. The central panel is a schematic illustration of the optical relaxation of an exciton in the two-level approximation. Panel (a) is adapted from Ref. [89], panel (d) from Ref. [90], the scheme in panel (e) from Ref. [91], and panel (f) is adapted from Ref. [92] under the Creative Commons Attribution 4.0 International License, Panel (b) and the STEM image in panel (e) are adapted from Refs. [57], [93], respectively, under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC-BY-NC-ND 4.0). Panel (c) is reprinted from Ref. [94] with the permission of AIP Publishing.

Figure 3:

Creation of color centers by local implantation. A focused ion beam (FIB) is used to implant Si atoms (blue) at well-defined positions (red) into a carbon-rich layer of crystalline Si. Inset: The created interstitial Si atoms (blue) can form G-centers (red) together with a pair of substitutional C atoms (brown). At low Si implantation doses, single G-centers are produced probabilistically that can be used as single-photon sources in the telecom O-band. Figure adapted from Hollenbach et al. under a CC-BY license [169].

Figure 4:

Spectral selection and individual control of multiple erbium emitters. (a) Spectrally multiplexed single-photon source in a Fabry–Perot cavity. A laser with precisely controlled frequency is used for pulsed excitation of the emitters. The fluorescence after the pulses exhibits distinct peaks (top), most of which correspond to single emitters, as evidenced by the antibunching in the dark-count corrected photon correlation function (bottom). As the spectral diffusion linewidth is much narrower than the inhomogeneous distribution, hundreds of emitters can be spectrally resolved and multiplexed. Adapted from [24]. (b) Level scheme of erbium emitters. Each dopant is characterized by an effective electronic spin 1/2, with an anisotropic splitting that depends on the applied magnetic field direction. The isotope ¹⁶⁷Er further exhibits a 7/2-nuclear spin that can achieve second-long coherence when a large magnetic field freezes the electronic spins in their ground state [193].

Figure 5:

Different types of photonic resonators used with single-photon sources at telecommunications wavelength. (a) Fabry–Perot resonator. A standing-wave cavity mode (red) is formed between two Bragg mirrors of alternating high (dark blue) and low (light blue) refractive index. Photon emitters are integrated into a crystalline slab with a thickness of a few micrometers (yellow). A small mode waist of only a few μm and a large mirror reflectivity, up to 99.999 %, allow for Purcell factors of several hundred. To first order, this is independent of the crystal thickness, such that the distance from the interface can be chosen large enough to preserve the coherence of the emitters. Adapted from [23]. (b) Bulls-eye, or circular Bragg grating cavity. A set of concentric rings is used to confine the light in the plane of a photonic thin film and to shape the mode of the emitted light. With a metal mirror at the bottom, a highly efficient collection can be achieved. The Purcell enhancement in this approach is ≲ 10, making it ideally suited for fast photon emitters such as quantum dots. (c) Nanophotonic resonator. A photonic thin film is patterned to form a waveguide with holes that generate a photonic band gap. Thus, the light can be confined below a single cubic wavelength while keeping quality factors in excess of 10⁵. The resulting lifetime reduction, up to 1000-fold, is required to use otherwise slow emitters such as T-centers or erbium dopants. Adapted from [196].

Figure 6:

Colocalized fabrication of CBG cavities at the local implantation sites (color center creation sites). (a) The color W-centers are created in silicon-on-insulator (SOI) at well-defined positions by Si⁺ ion implantation through nanoholes in a PMMA mask (top), followed by thermal annealing (center), and fabrication of CBG cavities centered on the nominal coordinates of the nanoholes. (b) PL map of the unpatterned SOI wafer, taken at 10 K. Inset: PL map of a single W-center in an unpatterned SOI wafer. The arrow serves to indicate the idea and does not show the location of an emitter showcased in the inset. Figure adapted from Lefaucher et al. under a CC-BY license [170].

Figure 7:

Photoluminescence imaging of quantum emitters at telecom bands. Quantum dots are grown at random positions and are localized optically for the subsequent deterministic fabrication of cavities. (a) PL imaging in C-band using the wide-field imaging system (left), and CBG cavities are deterministically fabricated at the localized QD positions (right). (b) Comparison of the imaging quality for the wide-field PL imaging (left) and confocal laser scanning microscopy (right) in C-band. (c) Wide-field imaging and deterministic fabrication of cavities in the telecom O-band. Panels (a) and (b) are adapted from Refs. [102], [250], respectively, via a Creative Commons Attribution 4.0 International License. Panel (c) is adapted from Ref. [225] with permission of Chinese Laser Press.

Figure 8:

Fiber-optical links for QKD with static and dynamic polarization switching. (a) Test-experiments toward QKD in an intercity fiber-optical link using a QD single photon source emitting at 1,555.9 nm. (b) Experimental layout of a QKD field experiment based on a QD single-photon source quantum frequency converted from 942 nm to 1,545 nm using dynamic, random polarization-state switching on Alice side and a deployed optical fiber link in the Copenhagen metropolitan area serving as quantum channel. Panel (a) is adapted from Ref. [284], and panel (b) from Ref. [285] via a Creative Commons Attribution 4.0 International License.

Figure 9:

QKD testbed using a benchtop plug&play telecom-wavelength QD SPS providing single-photon pulses via an SMF28 optical fiber for polarization encoding. (a) The 19-inch rack module houses a compact Stirling cryocooler, including the fiber-pigtailed QD device, a pulsed diode laser, and a fiber-based bandpass filter. (b) Photon autocorrelation histogram g ⁽²⁾(τ) on the fiber-coupled QD-emission confirming the single-photon nature of the photon stream coupled to the quantum channel. (c) Asymptotic secret key rate S versus loss revealing the performance of the plug&play system (blue) compared to early bulky laboratory-scale QKD-implementations reported by Waks et al. (orange, Ref. [278]) and Takemoto et al. (green, Ref. [283]). Adapted with permission from Gao et al., Appl. Phys. Rev. 9, 011412 (2022), Ref. [287]. Copyright 2022, AIP Publishing LLC.

Figure 10:

Fiber-optical links for entangled photon transmission. (a) Illustration of a field experiment for entangled photon transmission in a fiber-optical network in Cambridge, England, using a QD-based entangled-light emitting diode emitting around 1,310 nm. Entangled photon pairs are generated by a spectrally tunable entangled-light emitting diode featuring on-chip optical pumping at West Cambridge. The quantum-state measurement at the Cambridge Research Laboratory (CRL), after propagation through 15 km of optical fiber, revealed a high entanglement fidelity. (b) Intracity connection in Stuttgart, Germany, used for entanglement distribution with a QD photon-pair source emitting at 780 nm based on the biexciton–exciton radiative cascade. While one photon was detected immediately, the second one was quantum frequency converted to 1,515 nm before propagating to Stuttgart Feuerbach and back to the research laboratory via a 35.8 km long fiber-optical loop (14.4 dB transmission loss). Panel (a) is adapted from Ref. [300], and panel (b) from Ref. [301] via a Creative Commons Attribution 4.0 International License.

Figure 11:

Optical excitation schemes typically applied for the excitation of quantum emitters. Coherent excitation schemes are marked by a framed tile. Left to right: above-band excitation, where the optical pump (p) photons have a frequency much larger than the emitting state (X), ω _p ≫ ω _X, and excite the bandgap states of the surrounding matrix material, resulting in an incoherent excitation; Resonant excitation, or resonance fluorescence, where the pump photon energy matches the energy of the probed optical transition, resulting in a strictly coherent driving of the optical transition, ω _p = ω _X; quasi-resonant excitation, where an excited emitter state is pumped, for example, p-shell for QDs; schematic illustration and simulation of the swing-up excitation (Swing-Up of Quantum Emitter Population, SUPER), based on protocol introduced in Ref. [239]. In this approach, the application of two red-detuned pulsed coherently drives the two-level system in analogy to resonant excitation but avoids the spectral overlap of the pump and signal photons. Phonon-assisted excitation, where the pump photon energy is relaxed via nonradiative transition, which includes the emission of optical or acoustic phonon, losing the pump coherence, and ω _p > ω _X; two-photon excitation, where the biexciton state is resonantly excited via a nonlinear process, using a virtual state so that 2ω _p = ω _XX + ω _X. The fine structure of the exciton results in two linearly polarized radiative recombination paths.

Figure 12:

Illustration of a future quantum network on the European continent comprising many long-haul quantum communication links connecting metropolitan areas.