Source-linked AI summary
Hybrid integration methods for on-chip quantum photonics
Je-Hyung Kim, Shahriar Aghaeimeibodi, Jacques Carolan, Dirk Englund, Edo Waks
TL;DR
Integrated quantum photonics seeks to combine high-performance quantum emitters with scalable, phase-stable photonic circuits despite their incompatible material requirements. This review surveys emitter and PIC platforms, hybrid integration methods, and on-chip control techniques. It identifies hybrid integration as a route toward combining generation, manipulation, and detection, while emphasizing interface-related challenges.
Problem
Fully integrating quantum-light generation, manipulation, and detection is difficult because the components require different materials and silicon generation is not generally deterministic or wavelength-compatible with solid-state emitters.
Method
The paper reviews solid-state emitters, photonic integrated circuits, hybrid techniques including pick-and-place, wafer bonding, and epitaxial growth, and chip-based coherent control.
Results
Hybrid integration provides a potential route to combine quantum emitters, memories, coherent optical operations, and single-photon detection in one integrated platform.
Takeaways & Limitations
Integrated quantum photonics can combine deterministic quantum-light sources with PIC routing and processing while retaining the need to address material-interface challenges.
Takeaways & Limitations
Solid-state emitter environments introduce limited extraction, random position and frequency, and dephasing from interactions with charges and phonons.
Abstract
from arXiv · showhide
The goal of integrated quantum photonics is to combine components for the generation, manipulation, and detection of non-classical light in a phase stable and efficient platform. Solid-state quantum emitters have recently reached outstanding performance as single photon sources. In parallel, photonic integrated circuits have been advanced to the point that thousands of components can be controlled on a chip with high efficiency and phase stability. Consequently, researchers are now beginning to combine these leading quantum emitters and photonic integrated circuit platforms to realize the best properties of each technology. In this article, we review recent advances in integrated quantum photonics based on such hybrid systems. Although hybrid integration solves many limitations of individual platforms, it also introduces new challenges that arise from interfacing different materials. We review various issues in solid-state quantum emitters and photonic integrated circuits, the hybrid integration techniques that bridge these two systems, and methods for chip-based manipulation of photons and emitters. Finally, we discuss the remaining challenges and future prospects of on-chip quantum photonics with integrated quantum emitters.
1. INTRODUCTION
Integrated quantum photonics seeks to combine quantum-light generation, photonic manipulation, and detection on one phase-stable chip. Hybrid integration addresses incompatible material requirements by combining complementary emitter and PIC technologies.
- Motivation: Photonic integrated circuits enable chip-scale manipulation of quantum light with improved component density, loss, and phase stability over bulk optics.These capabilities support demonstrations including quantum-mechanics tests, quantum simulation, and quantum machine learning.
- Motivation: A complete quantum photonic processor requires generation, linear and nonlinear propagation, and single-photon readout.Bringing these functions together could support processors for applications such as quantum chemistry and inference.
- Platform mismatch: III-V quantum dots provide high-purity, bright, indistinguishable single photons, whereas silicon offers dense, scalable, CMOS-compatible photonic circuitry.The two platforms therefore provide complementary strengths for integrated quantum photonics.
- Platform mismatch: Fully integrating photon generation, manipulation, and detection remains difficult because each component has distinct material requirements.Silicon-based photon generation is spontaneous or incompatible with deterministic emitters at visible and infrared wavelengths.
- Hybrid solution: Hybrid integration combines disparate photonic technologies that may not be compatible within a single fabrication process.The review covers emitters, PICs, integration techniques, and coherent on-chip control of photons and emitters.
2. SOLID‐STATE QUANTUM EMITTERS
Solid-state quantum emitters provide deterministic, high-quality single photons across broad wavelength ranges, but their host materials introduce extraction, positioning, spectral, and dephasing challenges. Photonic structures and integrated circuits improve source performance while providing compact platforms for routing and control.
- Emitter properties: Solid-state emitters can deterministically produce single photons or entangled pairs with high purity and indistinguishability.Reported platforms include quantum dots, crystal defects, two-dimensional materials, and perovskite nanocrystals.
- Emitter properties: Emitter wavelengths span ultraviolet to near-infrared, including telecom wavelengths, with potential for room-temperature and electrically driven operation.
- Emitter challenges: Solid-state environments cause limited extraction efficiency, random emitter position and frequency, and dephasing from charges and phonons.These issues arise despite the absence of the complicated trapping setups required for cold atoms and trapped ions.
- Emitter challenges: Photonic crystals, nanowires, microdisks, and micropillars improve brightness, collection efficiency, and generation rate through extraction engineering and Purcell enhancement.Researchers also develop methods to control emitter position, frequency, and dephasing.
- Photonic integrated circuits: PICs provide compact, phase-stable, high-bandwidth platforms for transmitting, manipulating, and detecting light on-chip.Their integration with emitters aims to create internal, deterministic quantum-light sources, while manipulating multiple emitters remains challenging.
A. Material Platforms
Quantum-photonic material platforms trade off loss, wavelength range, confinement, nonlinear modulation, and device footprint. Hybrid integration and complementary materials can address these trade-offs while supporting cryogenic operation and scalable PIC development.
- Silicon and silicon nitride: Silicon photonics combines CMOS compatibility, thousands of optoelectronic components, high index contrast, and low component loss.Its reported propagation loss can be as low as 2.7 dB/m.
- Silicon and silicon nitride: Silicon-on-insulator has a ~1.1 µm bandgap that causes significant loss for many emitters operating below this wavelength.Alternatives include telecom-compatible emitters or higher-bandgap materials such as Si3N4.
- Silicon and silicon nitride: Si3N4 is transparent above 400 nm and can provide 0.1 dB/m ultra-low-loss waveguides, at the cost of larger bends and device footprints.Its low loss is important for on-chip delay lines.
- Emerging platforms: LiNbO3 offers strong electro-optic and acousto-optic properties with a 350–4500 nm transparency window, while processing advances enable high-confinement nanophotonic waveguides.
- Cryogenic modulation: Cryogenic PIC modulation is important for multiplexing, high-fidelity linear operations, and wavepacket engineering.LiNbO3 and AlN can enable Pockels-effect switching that is, in principle, not limited by cryogenic temperatures.
- Cryogenic modulation: Progress depends on both advances in PIC technology and deeper understanding of quantum-technology requirements.
4. HYBRID INTEGRATION TECHNOLOGY
Hybrid integration methods connect quantum emitters with photonic circuits while balancing crystal quality, optical coupling, positioning, and scalability. Random nanoparticle dispersion is simple but does not provide the deterministic placement needed for demanding multi-emitter applications.
- Integration requirements: Hybrid integration of emitters onto PICs must maintain crystal quality, efficient optical coupling, and scalability.The review compares techniques according to these requirements.
- Random dispersion: Nanoparticles such as colloidal quantum dots and diamond nanoparticles can be integrated simply by dispersing them near photonic structures.Their lack of a bulk dielectric host avoids total internal reflection during single-photon emission.
- Random dispersion: Random dispersion does not precisely control emitter position, limiting deterministic, high-efficiency coupling of multiple emitters.Lithographic masking or atomic-force-microscope tip manipulation can improve positioning.
- Epitaxial growth: Epitaxial growth directly forms hybrid heterostructures by growing quantum materials on photonic platforms.The method supports emitters embedded in high-crystalline-quality media.
- Integration methods: Figure 4 surveys random dispersion, epitaxial growth, wafer bonding, and two pick-and-place approaches for placing emitter-containing structures on photonic circuits.
B. Epitaxial growth of hetero‐structures
Epitaxial growth directly forms hybrid heterostructures by growing quantum materials on photonic platforms, while wafer bonding joins separately optimized materials with greater material flexibility.
- B. Epitaxial growth of hetero‐structures: Epitaxial growth can produce hybrid heterostructures for emitters and photonic circuits on a single wafer.
- B. Epitaxial growth of hetero‐structures: Wafer bonding preserves high crystal quality by growing each material separately under its own optimized conditions.
- B. Epitaxial growth of hetero‐structures: Wafer bonding offers more material options than monolithic epitaxial growth and enables electronic and photonic structures fabricated with micro/nanolithography.
- B. Epitaxial growth of hetero‐structures: The bonded emitter membrane sits above the photonic wafer, with capping-layer thickness controlling emitter-to-circuit distance.
D. Pick‐and‐place
Pick-and-place transfers individually selected quantum devices onto photonic circuits, enabling pre-characterization, positional selectivity, and independent design, but alignment and transfer-induced damage remain practical challenges.
- D. Pick‐and‐place: Pick-and-place transfers small-scale quantum devices individually, allowing emitters to be pre-characterized and selectively positioned on photonic circuits.
- D. Pick‐and‐place: Independent emitter and circuit designs provide greater flexibility in materials, dimensions, and structures than pre-integrated wafers.
- D. Pick‐and‐place: Transparent stamps enable real-time optical alignment, but accuracy is limited by visible-light diffraction to around a few hundred nm.
- D. Pick‐and‐place: Stamp transfer generally prevents repositioning after integration because adhesion to the integrated structures exceeds adhesion to the stamp.
- D. Pick‐and‐place: Micro-probes can transfer emitters under optical or electron microscopy, with electron-microscope transfer improving alignment accuracy and permitting post-integration position changes.
- D. Pick‐and‐place: Hybrid integration must also support low-loss photon routing, modulation, and detection alongside control of mutually identical emitters.
A. Coherent control of quantum emitters
Coherent control addresses dephasing and linewidth broadening that undermine photon indistinguishability, using resonant or quasi-resonant excitation alongside on-chip strategies for suppressing excitation-laser background.
- A. Coherent control of quantum emitters: Hong-Ou-Mandel interference requires highly coherent, indistinguishable photons and a coherence time τ_2 approximately twice the spontaneous decay time τ_1.
- A. Coherent control of quantum emitters: Above-band excitation broadens InAs quantum-dot linewidths to over a few tens of μeV, despite radiative decay times as short as 1 ns corresponding to sub-μeV homogeneous linewidths.
- A. Coherent control of quantum emitters: Resonant and quasi-resonant excitation have produced near transform-limited linewidths, while increased indistinguishability need not reduce emitter brightness.
- A. Coherent control of quantum emitters: Resonant excitation enables coherent emitter-state control and reveals atom-photon effects including Rabi oscillations and the Mollow triplet.
- A. Coherent control of quantum emitters: Resonant excitation creates strong scattered-laser backgrounds that cannot be removed by spectral filtering alone.
- A. Coherent control of quantum emitters: On-chip waveguides can separate resonant excitation from emitted photons by acting as polarization filters, while spatially separated excitation and collection further reduce laser scattering.
B. Generation of multiple, indistinguishable single photons
Hybrid photonic systems pursue multiple indistinguishable single photons using either deterministic temporal multiplexing from one emitter or frequency- and position-controlled arrays of emitters. Matched emitters also enable collective many-body emission on chip.
- Parametric down-conversion is intrinsically probabilistic, motivating deterministic strategies for supplying multiple single photons to photonic circuits.
- Temporal-to-spatial demultiplexing distributes photons from one bright emitter across multiple circuit channels using delay lines and beam splitters or electro-optic routing.Indistinguishability decreases with temporal separation, while integrating delay lines tens of nanoseconds long is challenging.
- Multiple integrated emitters require local frequency tuning because frequency randomness limits quantum interference between photons from individual emitters.Wafer-bonded quantum dots in multiple SiON channels provide an example of independently tunable emitters.
- Matched-frequency emitters provide multiple indistinguishable photons and a platform for many-body quantum physics, including Dicke-state superradiance.Two resonant emitters coupled to a photonic-crystal waveguide show superradiant emission, and three-emitter superradiance has been achieved with local strain tuning.
- Deterministic positional control uses patterned substrates, pyramidal nanostructures, or vertical nanowires to place quantum dots for efficient circuit coupling.Nanowires can control dot position and number and can be transferred into photonic circuits.
- Local frequency and positional control demonstrate strong potential for photonic integrated circuits containing multiple identical emitters and indistinguishable photons.
C. On‐chip manipulation of photons
On-chip manipulation combines integrated linear and nonlinear photonic components to route, filter, and frequency-convert quantum light. These functions extend the control available from photonic integrated circuits and nanophotonic structures.
- Integrated waveguides, beam splitters, phase shifters, and delay lines can form tunable Mach-Zehnder interferometers for reconfigurable photonic circuits.
- Micro-ring add-drop filters can spectrally isolate narrow emission lines from quantum emitters and unwanted background light.The resonant frequency can be tuned by changing the ring condition.
- Frequency conversion through four-wave mixing Bragg scattering provides a tuning range from a few tens to a few hundred nanometers, exceeding typical emitter tuning below 10 nm.This can match emissions from different emitter types, including near-infrared quantum dots and visible diamond defects.
D. Spin‐photon quantum interface
Integrated emitters can couple photonic and spin degrees of freedom, enabling spin-photon interfaces and nonlinear optical operations on chip. Hybrid assembly broadens these capabilities beyond the limits of many monolithic platforms.
- Quantum memories, gates, and coherent nonlinear elements are needed because photons alone lack storage time and deterministic photon-photon interactions.
- Strongly coupled nanophotonic cavities and waveguides can control spin states through photon polarization and photons through spin states.Demonstrations include conditional phase shifts and strong photon-photon interactions.
- Waveguide QED allows multiple emitters at different positions to interact through real and virtual photons, enabling long-range connectivity.Slow-light propagation modes mediate the interaction without requiring localized cavity modes.
- Integrated emitter light-matter interactions support applications including quantum repeaters, logic gates, photon-photon gates, single-photon transistors, and photon-number filters.
- Hybrid integration can scale these systems by post-assembling independently optimized emitters, cavities, and photonic circuits.Most prior demonstrations were monolithically integrated, limiting emitter number and function.
E. On‐chip detection of photons
On-chip photon detection can reduce coupling losses by integrating detectors directly with photonic circuits. Superconducting nanowire detectors offer high-performance detection and can also be post-integrated onto waveguides.
- Photons propagating in photonic circuits must be extracted and coupled to an objective lens or fiber for detection when detectors are external.Suggested coupling methods include gratings, evanescent coupling, tapered waveguides, and lensed-fiber end coupling.
- Superconducting nanowire detectors can be fabricated directly on photonic circuits or post-integrated onto waveguides using pick-and-place.
- Over 90% detection efficiency, below 50 ps response time, and over 100 MHz operation are reported for superconducting nanowire detectors across a broad spectral range including telecom wavelengths.
6. REMAINING HURDLES AND OUTLOOK
Integrated quantum photonics has advanced toward combining solid-state emitters with photonic circuits, but practical large-scale devices remain constrained by emitter control, quantum interactions, and room-temperature coherence. Hybrid integration may help combine complementary sources, spins, nanophotonic structures, and control capabilities, while applications such as Boson sampling and quantum networks remain prospective.
- Solid-state emitters now offer high-rate, pure, indistinguishable photons with controlled position, frequency, and spin, while photonic circuits provide multifunctional on-chip manipulation.These advances include spin–photon interaction and combined couplers, phase shifters, and linear/nonlinear components.
- The principal practical obstacle is generating multiple indistinguishable single photons from independently controlled emitters.The number of simultaneously controlled emitters is increasing, but long coherence times remain lacking.
- Efficient on-chip quantum interactions still require improved interference visibility, single-dipole cooperativity, and entanglement fidelity for large-scale systems.Higher efficiency, scalability, stability, and controllability of emitters and photons are also required for deterministic processing.
- Hybrid integration may combine efficient sources, coherent spins, high-quality nanophotonic structures, and spectral and spatial control that may be difficult to realize in one material.The review identifies pick-and-place, wafer bonding, and epitaxial growth as hybrid techniques, while applications remain prospective.
- Room-temperature coherent single-photon operation is intrinsically difficult because phonon interactions broaden emission linewidths and limit indistinguishability.An integrated tunable laser has been suggested to retain electrical operation while resonantly exciting emitters optically.