Dispersive Hong-Ou-Mandel Interference with Finite Coincidence Windows

Abstract

Hong-Ou-Mandel (HOM) interference is a fundamental tool for assessing photon indistinguishability in quantum information processing. While the effect of chromatic dispersion on HOM interference has been widely studied, the interplay between dispersion and the finite detection window of realistic measurement devices remains under-explored. In this work, we demonstrate that the rectangular coincidence window inherent to modern time-tagging modules, which effectively acts as a temporal filter, breaks the standard dispersion cancellation condition and restores sensitivity to symmetric group velocity dispersion. We derive an analytical model for type-II SPDC processes that predicts a modification of the HOM dip shape, specifically the emergence of characteristic oscillations and dip broadening. We experimentally validate this theoretical framework using a ppKTP source and transmission through optical fibers of lengths up to 29 km. The experimental data show excellent agreement with the model, confirming the presence of window-induced oscillations and allowing for the precise extraction of the fiber dispersion parameter. These findings underscore the importance of accounting for finite timing resolution in the design and characterization of dispersive quantum communication links.

Figures (4)

• Figure 1: Orthogonally polarized photons from a filtered ppKTP source are separated and coupled into the FUT via a polarization combiner, with one arm delayed by a piezo stage. After propagating through the FUT, a polarization controller and splitter restore and separate the modes. The photons then interfere at a 50:50 beam splitter and are detected by SNSPDs connected to a coincidence counter.
• Figure 2: Example fits showing predicted oscillations for different experimental configurations of delay $T$ and fiber length $L$.
• Figure 3: FWHM of the model (solid lines) and data (points) plotted against the fiber length for several values of the coincidence window. The model uses the extracted fit parameters, $\rho$ and $\beta_2$. While the fit allowed $\eta$ to vary, the model curves here are plotted with fixed $\eta=0.5$ to illustrate the theoretical trend.
• Figure 4: Average scaled residuals of a subset of the data with respect to fiber length and coincidence window.