Mode-resolved picosecond single-photon polarimetry maps modal dynamics in multimode fibers

Article https://doi.org/10.1038/s41467-026-72129-w Mode-resolved picosecond single-photon polarimetry maps modal dynamics in multimode fibers Harikumar K. Chandrasekharan & Ross Donaldson Polarization dynamics in multimode optical fibers (MMFs) play a key role in applications ranging from high-capacity communication to distributed sen- sing. However, most approaches probe only a single polarization state, limiting polarimetric characterization and leaving orthogonal dynamics hidden. Here, we show that time-resolved single-photon avalanche diode (SPAD) arrays enable quasi-real-time observation of dual-polarization mode dynamics in few- mode fibers. Simultaneous detection of orthogonal polarization channels provides picosecond time-of-flight polarimetric readout, revealing spatio- temporal correlations between LP01 and LP11 modes under stress. Building on this capability, we introduce a SPAD-based single-photon polarimetric plat- form that reconstructs Stokes vectors across ~ 1000 spatial channels with 55 ps resolution, visualizing complex modal dynamics in MMFs. Through Hilbert- transform analysis, the system provides mode-resolved Stokes retrieval with per-pixel minimum detectable modulation of 0.002 (0.2%, 3σ) and signal-to- noise ratios up to 33 dB. This scalable platform enables ultrafast Stokes polarimetry in MMFs, opening new opportunities in classical and quantum photonics. Polarization-dependent mode behavior in optical fibers underpins applications in telecommunications1,2, sensing3,4, and imaging5–7. Single-mode fibers (SMFs) offer high-bandwidth, long-haul links and precise sensing via fundamental-mode confinement8,9, but deploy- ment is costly due to alignment-sensitivity10, and performance remains bounded by nonlinear Shannon-limit considerations11. Multimode fibers (MMFs), with larger cores supporting many spatial modes, enable mode multiplexing and thus higher-capacity trans- mission at lower cost as well as easier installation12,13. However, their modal profiles vary with geometry, environment, and input polar- ization; perturbations induce mode mixing, dispersion, and inter- ference that shape propagation and signal integrity14–17. Properly exploited and in some cases corrected for, these multimodal photon dynamics can advance communications, sensing, and endoscopic or deep-tissue imaging18–20. Accordingly, characterizing such dynamics across polarization states is essential for high-precision transmission, high-dimensional encoding, and imaging through scattering media21,22. The high-fidelity spatial information of optical signals in MMF systems can be extracted through advancements in multiplexing, mode tomography, holography, single-photon detection, and machine learning techniques23–26. However, these techniques face several chal- lenges and limitations, including mode crosstalk, computational complexity, sensitivity to alignment, and system cost, which hinder their widespread adoption. Furthermore, methods like machine learning require significant data and computational resources, while techniques such as digital holography demand precise optical align- ment and often fail to resolve full polarization behavior at ultrafast timescales27. Critically, most existing approaches are limited in their ability to perform simultaneous, mode-resolved, time-resolved, and polarization-resolved measurements at the single-photon level —a capability essential for next-generation fiber-based quantum Received: 29 September 2025 Accepted: 1 April 2026 Check for updates Scottish Universities Physics Alliance, Institute of Photonics and Quantum Sciences, School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh, Scotland, UK. e-mail: hk47@hw.ac.uk Nature Communications| (2026) 17:6174 1 1234567890():,; 1234567890():,;