Next-Generation Heralded Single Photon Sources

TL;DR

The paper addresses the lack of standardized benchmarking for heralded single-photon sources (HSPSs) on integrated photonics, highlighting how inconsistent reporting of spectral purity $P_S$, brightness $B$, and heralding efficiency $H$ impedes progress toward next-generation quantum applications. It analyzes trade-offs among these metrics, explains how the joint spectral intensity (JSI) and Schmidt decomposition constrain useful brightness, and introduces a framework with defined measurement points ($B_1,B_2,B_3$ and $H_1,H_2,H_3$) and pump considerations to enable fair comparisons across devices and platforms. The authors show that reported $P_S$ and $H$ have improved over time while brightness has stagnated or declined, and they stress the need for standardized reporting and pump-relevant definitions to disentangle source performance from system losses. The outlook advocates for rigorous, label-consistent benchmarking to guide the design of scalable HSPSs on integrated platforms, ultimately accelerating practical quantum computing, communication, and sensing capabilities by providing reliable performance guarantees expressed through $P_S$, $B$, and $H$.

Abstract

Scaling up quantum computers and building a quantum internet requires the development of ideal photon sources. Heralded single photon sources on integrated photonic platforms are the way forward to achieve this goal. Here we identify inconsistencies in source characterisation and propose methods to facilitate fairer comparison and better understanding of which sources could enable next-generation quantum applications.

Figures (5)

• Figure 1: The development of next-generation HSPSs will enable next-generation applications, such as satellite-based quantum communications yin, complex molecular simulations sparrow, and the sensing of individual bacteria spedalieridaher. A timeline of milestone correlated photon pair sources are depicted. These sources are: a) Compton scattering following electron-positron annihilation bleuler, b) bulk crystals burnham, c) fibre bonfrate and integrated waveguides fukuda, d) microstructure fibre sharping, photonic crystal waveguides suzuki, and microring resonators clemmen, e) photonic molecules zeng, interferometrically coupled resonators liu, and photonic molecule interferometrically coupled resonators integrateandscale.
• Figure 2: Attributes of HSPSs directly affecting the performance of quantum information processing. a) Imperfect spectral purity will result in exponentially decaying fidelity for subsequent gate operations. Near-$100\%$ spectral purity is therefore crucial. b) The trade-offs between brightness, heralding efficiency, and spectral purity must be overcome to optimise all parameters and achieve next-generation HSPSs for next-generation applications. High brightness and high heralding efficiency are essential for the speed of quantum operations. Values are taken from integrateandscale and paesani, as well as a theoretical near-ideal example. c) The fundamental relation between spectral purity and the useful brightness of a HSPS. Spectral purity, often described by the JSI (inset), can effectively be split into orthogonal Schmidt modes, $\lambda_k$. The total brightness is thus divided among the Schmidt modes according to their relative weights. Since Schmidt modes are orthogonal, they can be viewed as independent HSPSs which will not interfere in quantum information processing operations. The useful brightness is that of the fundamental Schmidt mode, which is equal to the total brightness for $100\%$ spectral purity.
• Figure 3: Defining three locations in an integrated HSPS where brightness, B, and heralding efficiency, H, can be measured for (a) a microring resonator and (b) a waveguide source.
• Figure 4: Reported HSPS parameters over time. The full list of references and corresponding parameter values can be found in the Appendix.
• Figure 5: a) Simulated spectral profiles for two different pulse linewidths within a side-coupled waveguide, the overlap of these with the MRR transmission function gives the spectral profiles within the MRR. b) Temporal profiles corresponding to the spectral profiles, and the relative average pump powers between them. c) Simulated plots of $R_{s,i}$ against $P^2_{avg}$, where a given value of $R_{s,i}$ (indicated by the black dotted line) is measured at the average power of each pump profile.