Integrated Photonic Quantum Computing: From Silicon to Lithium Niobate

Abstract: Quantum technologies have surpassed classical systems by leveraging the unique properties of superposition and entanglement in photons and matter. Recent advancements in integrated quantum photonics, especially in silicon-based and lithium niobate platforms, are pushing the technology toward greater scalability and functionality. Silicon circuits have progressed from centimeter-scale, dual-photon systems to millimeter-scale, high-density devices that integrate thousands of components, enabling sophisticated programmable manipulation of multi-photon states. Meanwhile, lithium niobate, thanks to its wide optical transmission window, outstanding nonlinear and electro-optic coefficients, and chemical stability, has emerged as an optimal substrate for fully integrated photonic quantum chips. Devices made from this material exhibit high efficiency in in generating, manipulating, converting, storing, and detecting photon states, thereby establishing a basis for deterministic multi-photon generation and single-photon quantum interactions, as well as comprehensive frequency-state control. This review explores the development of integrated photonic quantum technologies based on both silicon and lithium niobate, highlighting invaluable insights gained from silicon-based systems that can assist the scaling of lithium niobate technologies. It examines the functional integration mechanisms of lithium niobate in electro-optic tuning and nonlinear energy conversion, showcasing its transformative impact throughout the photonic quantum computing process. Looking ahead, we speculate on the developmental pathways for lithium niobate platforms and their potential to revolutionize areas such as quantum communication, complex system simulation, quantum sampling, and optical quantum computing paradigms.

• The paper systematically reviews photonic quantum computing platforms, comparing silicon’s mature integration with LN’s breakthrough in high-bandwidth modulation and efficient SPDC.
• It details silicon’s achievements in high-confinement photonics and programmable circuits while noting limitations like two-photon absorption and slow, lossy modulation.
• The review emphasizes LN’s superior nonlinear and electro-optic properties, enabling deterministic photon generation, broadband operation, and scalable hybrid quantum networks.

Integrated Photonic Quantum Computing: Transitioning from Silicon to Lithium Niobate

Introduction and Motivation

This review systematically dissects the evolution of integrated photonic quantum computing (PQC) platforms, contrasting established silicon-based technologies with the emerging lithium niobate (LN) thin-film platform. As PQC targets quantum advantage in areas such as simulation, communication, and cryptography, the scalable integration of robust quantum resources on-chip—single photons, entanglement, squeezed states, and reconfigurable operators—becomes paramount. The analysis delineates the photonic quantum information pipeline across core layers: encoding, gate operations, measurement, and computational architecture, with a focus on the trade-offs and material-dependent limitations shaping each platform’s performance.

Silicon Photonics: Architecture, Achievements, and Bottlenecks

Silicon-Based Photonic Quantum Toolboxes

Silicon platforms, leveraging high-confinement SOI waveguides and mature CMOS fabrication, have enabled large-scale integration for PQC. Their strong third-order nonlinearity underpins efficient spontaneous four-wave mixing (SFWM) sources for heralded single-photon generation and frequency-bin entanglement. However, silicon suffers from severe two-photon absorption (TPA) and free-carrier absorption (FCA), fundamentally capping brightness and purity of photon-pair sources.

Programmable manipulation is realized through dense arrays of Mzus (Mach-Zehnder interferometers), thermo-optic and carrier-induced phase modulators, and a variety of on-chip beam splitters and couplers. SNSPDs (superconducting nanowire single-photon detectors) and SPADs (avalanche photodiodes) have been integrated with silicon, providing high-efficiency, low-jitter single-photon readout, albeit limited by cryogenic requirements and coupling losses.

Quantum Computing Models and Demonstrations

Three principal silicon photonic architectures are emphasized:

• Gate-based model: Implementation of reconfigurable gate sets (e.g. CNOT, high-dimensional qudit gates) and state teleportation, with up to eight-photon entanglement and arbitrary 6×6 unitary processors demonstrated.
• Measurement-based model (MBQC): Fabrication of cluster and hypergraph states with up to five encoded qubits, inclusion of redundancy for error correction, and high-fidelity fusion-type entangling gates.
• Unitary operator-based model: Large programmable optical meshes for boson sampling, quantum walks, and Hamiltonian simulation. Here, demonstrations include up to 18-photon Gaussian boson sampling, programmable graph processors, and simulation of molecular vibronic spectra.

Table-driven summary of boson sampling experiments highlights the scaling challenge, as source probabilism (≤10% emission probability per attempt) and cumulative propagation loss dominate system-level fidelity [(2601.16484), Table 2].

Limitations and Transition Drivers

Despite network complexity and software-defined versatility, Si remains limited by:

• Electro-optical modulation limited to slow thermal tuning or lossy carrier-based modulation due to lack of bulk Pockels effect.
• Nonlinear processes constrained to x(3), precluding efficient SPDC, and further hampered by TPA/FCA.
• Spectral transparency restricted to the telecom band. Silicon lacks the mid-IR/visible compatibility of LN, handicapping memory and detector integration.
• Frequency conversion requires hybrid integration with non-silicon materials.

Integrated quantum circuit depth is currently throttled by loss, nonlinearity, and modulation speed, affecting scaling towards universal/fault-tolerant PQC.

Lithium Niobate Photonics: Fundamentals and Quantum Integration Potential

Material and Fabrication Innovations

Lithium niobate, fortified by advances in thin-film transfer (LNOI) and periodic poling, offers:

• High χ(2) (d33 ∼ −25.2 pm/V) for efficient SPDC, sum/difference frequency generation, and highly squeezed state production.
• Bulk Pockels effect (r33 ∼ 31 pm/V) enabling GHz–THz arbitrary-phase, low-power, and cryo-compatible electro-optic modulation.
• Broadband transparency (UV–mid-IR) with ultra-low propagation loss (<0.027 dB/cm).
• Advanced waveguide fabrication via dry/chemo-mechanical etching and rib-loading significantly improves index contrast and device density.

Quantum Photonics Toolbox and Performance

LN's integrated quantum device suite includes:

• SPDC photon-pair sources in PPLN waveguides and microrings. Table 5 exhibits brightness several orders of magnitude higher than SFWM, with pair generation rates at μW pump power—fully compatible with high-rate, on-chip quantum protocols.
• Squeezed light sources in both OPA and OPO geometries, with measured squeezing levels up to −8 dB and inferred on-chip squeezing exceeding −11 dB (Table 6). These bandwidths span THz, enabling dense, time/frequency-multiplexed CV-MBQC.
• Programmable EO-modulated circuits: Monolithic phase/amplitude modulators, AWGs, Bragg/reflection structures, and high-extinction MZI meshes.
• Resonant structures: High-Q microdisks, microrings, and racetracks facilitating ultra-low threshold OPOs, high-efficiency frequency converters (>70%).
• Integrated single-photon detectors: Emerging waveguide-integrated SNSPDs with measured >45% on-chip efficiency (exceeding 90% anticipated with loss engineering), upconversion SPADs, and initial demonstrations of on-chip TES for photon-number resolution.
• Photon source multiplexing and delay lines: Sub-0.03 dB/cm true delay lines, EO-tunable delay, and all-integrated fast switching for photon multiplexing at repetition rates >100 MHz.
• Microwave-to-optics interface: State-of-the-art EO transducers achieving bidirectional microwave-optical conversion with >1% on-chip efficiency and added noise <10 photons per upconverted event—vital for interfacing with superconducting qubits.

Advantages over Silicon: Quantitative and Qualitative

Table 3 strongly demonstrates:

• Photon generation efficiency: PPLN microrings yield pair rates >2 MHz/μW—two orders beyond Si platforms.
• Squeezing: THz-bandwidth squeezing at >10 dB (essential for fault-tolerant CV computing) is within reach, surpassing the <2 dB bandwidth-limited squeezing in Si/SIN.
• Modulation: EO modulation in LN achieves >100 GHz bandwidth, 2.3 V·cm VπL, and extinction >53 dB, far exceeding thermally-limited Si tuning.
• Spectral bandwidth: 350–5000 nm enables photonic interfacing across diverse memory/detector wavelengths and hybrid quantum nodes.
• Quantum frequency conversion: 73% internal efficiency with low noise is practical on PPLN waveguides at mm scale, supporting quantum network interconnection.

Integrated Quantum Architecture and Next-Generation Pathways

Circuit-Level Integration and Programmable Processing

LN circuits, inheriting the architectural paradigms of Si photonics (e.g., arbitrary mesh unitary synthesis), achieve substantially faster modulation and lower loss, advancing the fidelity and scalability of MBQC, quantum walks, and boson sampling. Demonstrations include:

• Dynamically programmable unitary circuits (6×6 and beyond) using EO-tunable MZIs [299].
• Integrated quantum processors executing real-time simulation, high-fidelity state routing, and reconfigurable path/polarization manipulation [392, 393, 394].

Continuous-Variable Measurement-Based Computing

LN’s dominant x(2) nonlinearity and high-bandwidth squeezing are critical for universal CV-MBQC. The review traces protocols leveraging time-multiplexed, broadband squeezed state generation via cluster states, where >10 dB squeezing and >THz bandwidth unlock viable error-corrected, fault-tolerant architectures. Recent silicon nitride results—eight-mode CV entanglement [414]—are anticipated to scale even further with optimized LN integration.

Multiplexed Source Architecture and Low-Loss Feedforward

The physical deterministic photon generation rate is fundamentally limited by source probabilism. LN delay lines and fast EO switches enable aggressive spatial/temporal multiplexing, with total-system efficiencies projected to cross the deterministic threshold required for scalable MBQC. Integrated feedforward—via proximity-bonded electronics—promises sub-nanosecond latency, enabling active resource state switching and fast quantum feedback at scale.

Hybrid Quantum System Interconnects

Bidirectional, efficient, and low-noise microwave-to-optical conversion on TFLN leverages the Pockels effect and resonant field enhancement, facilitating scalable interfaces to superconducting circuits and microwave quantum networks—an essential enabler of modular quantum computing and the quantum internet.

Implications, Challenges, and Future Trajectory

Practical Impact

• Silicon photonics remains optimal for ultra-large-scale, digitally reconfigurable architectures—especially given foundry-scale CMOS compatibility, and for specific computational primitives where gate rates and loss are non-critical.
• Lithium niobate is demonstrably superior for:
  • High-rate, high-purity photon sources for boson sampling, CV MBQC, and quantum communication.
  • Cryogenic-compatible, GHz–THz EO modulation for in-situ quantum feedback and error correction.
  • Ultra-broadband, high-level squeezing for large-scale, fault-tolerant CV quantum computing.
  • Broadband frequency conversion and spectral translation for hybrid quantum networking.

Emergent capabilities in monolithic integration, ultra-low loss, and hybrid photonic-electronic interposers indicate that LN PQC will soon match the density and complexity of established silicon systems, while offering unique quantum capacity.

Theoretical and Technological Prospects

• Universal scalable MBQC in continuous variables, with on-chip error-protection and >10 dB squeezing, is likely to become practical only on TFLN.
• Hybrid quantum networks leveraging direct microwave-optical links will drive the transition toward distributed quantum computation and repeater networks.
• Photonic neural networking and optical ML: Large reconfigurable unitary meshes in LN, with on-chip nonlinearity and detection, are likely to become foundational ML accelerators.
• Challenges: Remaining integration issues concern further reduction of propagation/combination loss, scalable cryogenic-compatible packaging, ultra-efficient on-chip photon-number-resolving detectors, and manufacturability of high-Q, large-format PPLN with sub-μm domain control.

Conclusion

This comprehensive review establishes that thin-film lithium niobate is poised to play a central role in the next generation of integrated photonic quantum computing. Its superior electro-optic, nonlinear, and optical properties enable deterministic high-rate quantum state generation, ultrafast programmable quantum gates, low-noise, broadband quantum interface, and scalable continuous-variable resource states. Translation and extension of silicon circuit models to LN will yield immediate system-level performance gains. The roadmap toward fault-tolerant, universal PQC will likely be realized on or in conjunction with LN-based photonic circuits, with broad implications across quantum information science.

Reference: "Integrated Photonic Quantum Computing: From Silicon to Lithium Niobate" (2601.16484)