An ultralow-loss integrated photonic platform for discrete-variable quantum information processing

Abstract: Photonic integrated circuits offer a scalable and robust route toward quantum information technologies by consolidating photon sources and linear optical networks onto compact, wafer-manufacturable chips. Although silicon photonics has enabled diverse discrete-variable quantum breakthroughs -- spanning multiphoton entanglement, quantum networking, and photonic qubit fusion for quantum computing -- scaling these platforms beyond proof-of-principle demonstrations remains severely constrained by a critical system-level bottleneck. Optical loss compounds rapidly across photon generation, routing, and state analysis, causing multiphoton generation probabilities to plummet exponentially as circuit depth and complexity grow. Here we overcome this rate-loss barrier by demonstrating a monolithic, ultralow-loss silicon nitride (Si$_3$N$_4$) integrated photonic platform engineered for high-performance discrete-variable quantum information processing. Our architecture seamlessly integrates narrowband photon-pair sources with low-loss qubit-fusion circuits and reconfigurable state-analysis interferometers. The on-chip sources prepare Einstein-Podolsky-Rosen (EPR) states with a fidelity of 0.9875(3) and exhibit near-unity photon indistinguishability, yielding a heralded Hong-Ou-Mandel interference visibility of 0.990(6). By executing on-chip fusion of two EPR states, we synthesize and characterize four-photon Greenberger-Horne-Zeilinger states with a record fidelity of 0.943(8) and a fourfold count rate of 27 Hz -- more than two orders of magnitude higher than previous silicon-photonic implementations. Combined with standard CMOS-compatible fabrication on 150-mm-diameter wafers, these results establish ultralow-loss Si$_3$N$_4$ integrated photonics as a definitive, manufacturable platform for deployable, large-scale quantum information processors.

• The paper demonstrates a monolithic Si3N4 platform that integrates photon pair sources and ultralow-loss circuits to overcome the rate–loss trade-off in discrete-variable quantum processing.
• The method leverages cavity-enhanced spontaneous four-wave mixing and precise thermal tuning to yield near-ideal Hong–Ou–Mandel visibilities and narrow photon linewidths.
• The platform achieves record multiphoton entanglement with four-photon GHZ fidelities up to 0.943 and over 100× higher count rates compared to previous approaches.

Ultralow-Loss Integrated Photonic Platform for Discrete-Variable Quantum Information Processing

Introduction

The realization of large-scale, deployable quantum information processing (QIP) platforms critically depends on scalable, wafer-compatible photonic integrated circuits (PICs) that combine high-performance photon sources with ultralow-loss optical networks. While silicon photonics has enabled a series of proof-of-principle demonstrations such as multiphoton entanglement, quantum networking, and linear-optical qubit fusion, scaling such systems is fundamentally limited by compounding optical loss. This work establishes a monolithic silicon nitride (Si$_3$N$_4$) PIC platform with system-level loss minimized across the entire source–fusion–analysis stack, solving the rate–loss bottleneck intrinsic to previous architectures and advancing the state-of-the-art in discrete-variable (DV) quantum photonics (2606.26910).

Platform Architecture and Physical Implementation

The presented Si$_3$N$_4$ platform leverages fully foundry-compatible processes on 150-mm wafers, integrating cavity-enhanced spontaneous four-wave mixing (SFWM) photon-pair sources, low-loss qubit-fusion circuits, and reconfigurable Mach-Zehnder state analyzers within a single chip. This technological stack ensures that photon loss per component is suppressed, with measured insertion losses of $\leq 0.05\,\mathrm{dB}$ for all passive devices. Concomitantly, chip–fiber coupling, filtering, and detection employ matched commercial hardware to preserve system-level efficiency, yielding average single-photon end-to-end efficiencies of $>30\%$.

The platform uses path-encoded DV qubits, with the logical $\ket{0}$, $\ket{1}$ defined by occupation in spatially distinct waveguides. Integrated microresonator-based photon-pair sources are engineered for both high spectral purity and geometric uniformity across the wafer. Key architectural features include:

• Cavity-enhanced SFWM in high-Q Si$_3$N$_4$ rings yielding 100–160 MHz heralded photon linewidths.
• Mode-selective directional couplers with coupling idealities approaching 99.5%.
• MZI-based reconfigurable analysis units with thermal phase shifters allowing for arbitrary basis selection and precise phase control.
• Scalable inverse taper edge-couplers for high-efficiency chip/fiber interfacing.

Photon Source Characterization and Indistinguishability

The Si$_4$0N$_4$1 photon sources demonstrate near-ideal metrics: measured Hong–Ou–Mandel (HOM) interference visibilities up to $_4$2 and narrow spectral linewidths exceeding competing integrated and bulk-optical implementations. Crucially, these strong properties persist at high pair generation rates ($_4$3 up to $_4$4 per pulse) with HOM visibilities remaining above 0.9, setting a new benchmark for integrated source brightness and indistinguishability.

Pairwise photon indistinguishability is directly measured by on-chip, heralded HOM interference between photons from independent resonators. Temporal and spectral overlap is a direct consequence of the sub-nanometer fabrication precision and microresonator alignment enabled by on-chip thermal tuning.

Multiphoton Entanglement and Qubit Fusion

The platform executes on-chip fusion of two EPR states, generated from four independent microresonator sources, to synthesize four-photon Greenberger–Horne–Zeilinger (GHZ) states. The process involves path-exchange operations and post-selected parity projection, realized via passive, loss-minimized waveguide networks.

Key state generation and analysis results:

• EPR state fidelity $_4$5.
• Four-photon GHZ state fidelity $_4$6 at a fourfold count rate of $_4$7.
• At higher pump powers, a sustained GHZ fidelity of $_4$8 at $_4$9 count rate.

Population measurements in the computational basis and coherence characterization (via directly measurable parity oscillations in rotated bases) confirm the synthesized GHZ states display genuine four-photon coherence with visibilities and statistical robustness that substantially exceed previous integrated photonic records.

Comparison with Prior Work

The authors’ integrated Si$_3$0N$_3$1 platform establishes clear performance advantages compared to prior silicon, thin-film lithium niobate, and quantum dot-based PICs. Key contrasts:

Metric	This Work (Si$_3$2N$_3$3)	Best Prior Si Photonics	Comments
HOM visibility	0.990(6)	0.995(3)	Comparable to bulk optics
Four-photon GHZ fidelity	0.943(8)	0.854	System-level loss suppressed
Fourfold GHZ rate (Hz)	27	$_3$40.2	>100$_3$5 improvement
Photon linewidth (MHz)	160	500–2000	Quantum memory compatible

Prior platforms (e.g., [Llewellyn et al., Alexander et al., Chen et al., Lee et al.]) consistently suffered from exponential rate loss as circuit depth (and photon number) increased, due to non-negligible propagation and coupling loss at every stage. Here, the combination of bright, indistinguishable sources and ultralow-loss routing results in a two-orders-of-magnitude higher multiphoton count rate without fidelity degradation.

Practical and Theoretical Implications

The work provides an industrially manufacturable foundation for DV-QIP applications ranging from scalable photonic quantum computing and entanglement-enhanced metrology, to quantum communication with memory-compatible photons (narrow linewidth for efficient light–matter interfaces). The demonstrated loss-and-indistinguishability metrics remove the historical system-level rate–loss/fidelity trade-off that has limited further circuit scaling, and the measured uniformity and repeatability across a 150-mm wafer strongly portend reproducibility at scale.

From a theoretical perspective, the ultimate scalability of measurement-based photonic quantum computing, cluster state generation, and quantum repeaters all require exactly this combination of loss and indistinguishability. The current architecture supports straightforward extension to larger cluster/graph states by spatial multiplexing and fusion-gate stacking, with no fundamental architectural bottleneck identified.

Future Directions

Opportunities for further system optimization are clear:

• Increasing microresonator–bus coupling to approach loaded linewidths of $_3$610 MHz.
• Integration of waveguide-coupled SNSPDs and on-chip semiconductor pump lasers to eliminate all off-chip coupling.
• Improvements to chip–fiber coupling, spectrally flattened filters, and enhanced detection to reach single-photon end-to-end efficiencies $_3$7.
• Higher repetition rates and source brightness for kHz-level multiphoton experiments at fidelities suitable for distributed quantum applications.

The architecture’s intrinsic compatibility with standard CMOS processes allows direct translation of performance improvements to all-wafer, high-volume manufacturing, with applications extending to field-deployable, multiplexed quantum processors and quantum repeaters.

Conclusion

This work conclusively demonstrates a Si$_3$8N$_3$9 integrated photonic platform for DV-QIP that combines system-aligned, narrowband, indistinguishable photon-pair sources with an ultralow-loss routing and analysis network, all fabricated in a scalable foundry process. The results define new records for multiphoton entanglement generation (GHZ fidelity and count rate), while breaking the exponential rate–loss scaling that fundamentally constrained prior integrated platforms. The clear path toward further reductions in loss and increased integration density signifies this approach as a definitive candidate for large-scale, deployable quantum information processors as well as quantum networking nodes (2606.26910).