arXiv:2611.00805
Semi-classical analysis of noise squeezing in Kerr microresonators
Sara Persia, Lev Fandeev, Fuchuan Lei, Bastian Ruhnke, Alexander E. Ulanov, Vaishali Adya, Victor Torres-Company, Tobias Herr
Abstract
Kerr microresonators are integrated nonlinear devices widely used to generate classical optical frequency combs and nonclassical states of light. In particular, they provide a promising platform for squeezed light sources, which are important elements of emerging quantum sensing, communication, and computing technologies. Designing such sources requires numerical tools that efficiently incorporate realistic device physics and noise processes. Here, we show that established semi-classical frameworks based on the Lugiato-Lefever equation and the Ikeda map, instrumental in the development of classical light sources, can also predict the quadrature-noise properties of squeezed light when vacuum fluctuations are properly included. We consider three canonical configurations producing two-mode quadrature squeezing, single-mode quadrature squeezing, and intensity difference squeezing. In each case, the semi-classical simulations produce quadrature variances and noise spectra in close agreement with the quantum theoretical predictions. These frameworks therefore provide practical tools for designing squeezed light sources, with the capability to incorporate additional noise mechanisms, nonlinear effects, and complex resonator arrangements. More fundamentally, our results show that key experimental signatures of squeezing can be modeled as phase-sensitive amplification of properly normalized vacuum fluctuations within stochastic semi-classical field equations. This observation helps distinguish squeezing properties accessible through semi-classical field evolution from those intrinsically requiring a full quantum description.
AI Summary
The paper asks whether classical simulation frameworks for Kerr microresonators can reproduce squeezed-light noise observables when supplemented with vacuum fluctuations, providing numerical tools beyond analytically tractable configurations. It implements stochastic versions of the time-domain Ikeda map and the frequency-domain coupled-mode Lugiato–Lefever equation. Independent Gaussian vacuum fields enter through the bus-waveguide coupling and intrinsic-loss channels, normalized to half a photon per mode. Simulated output trajectories yield quadrature distributions and noise spectra, which are benchmarked against existing quantum-theoretical expressions for three configurations: single-pump two-mode quadrature squeezing below threshold, dual-pump single-mode quadrature squeezing, and above-threshold signal–idler intensity-difference squeezing. Both simulation methods agree with the corresponding quantum predictions in the examples studied. The single-pump cases use powers of $0.99P_{\mathrm{th}}$ and $1.03P_{\mathrm{th}}$, respectively, with external and intrinsic decay rates $\kappa_{\mathrm{ex}}/2\pi=200\,\mathrm{MHz}$ and $\kappa_0/2\pi=10\,\mathrm{MHz}$. The two-mode calculation also reproduces the mode-dependent squeezing across the generated spectrum. The agreement is established numerically through quadrature statistics and squeezing/anti-squeezing spectra, rather than experimentally. The single-mode benchmark deliberately restricts the resonant dynamics to two pumps and one central signal mode, suppressing parasitic four-wave-mixing processes rather than evaluating their full impact. Extensions to Raman effects, thermorefractive and technical noise, and complex resonator arrangements are proposed but not demonstrated. The conclusions concern the tested noise observables; the paper does not establish equivalence to a full quantum description for arbitrary observables or operating regimes.
Error found: The printed stochastic CME input–output prescription does not specify the directly transmitted bus vacuum and its correlation with cavity-driving noise, which are necessary to reproduce output squeezing spectra. The figure captions express δ₀ in radians although the equations define it as an angular frequency, leaving an unstated roundtrip-time conversion.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 12 pages, 4 figures