Mitigating Phase Correlations in Quantum Key Distribution Using Path-Selection Modulation
Amita Gnanapandithan, Li Qian, Hoi-Kwong Lo · 2025
Read Mitigating Phase Correlations in Quantum Key Distribution Using Path-Selection Modulation by Amita Gnanapandithan, Li Qian, Hoi-Kwong Lo (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-18490.txt.
Introduction
Quantum key distribution lets two parties share secret keys, but the tools that encode the keys can leak. Amita Gnanapandithan, Li Qian, and Hoi-Kwong Lo study one such leak: phase correlations from active modulators. Their 2025 paper measures these links in tests and simulations at rates up to billions of pulses per second. The fix they propose removes the risky part entirely.
Their source picks a path at random, and each path stands for one encoded state. No active phase modulator is needed at all. Random phase comes from gain switching, a laser trick. That design closes not just this leak but all side channels tied to active phase encoding. The team characterized the new source at a clock rate of one billion steps per second. The work hardens a core tool of quantum-safe communication.
Catalog record
| CREATOR | Amita Gnanapandithan, Li Qian, Hoi-Kwong Lo |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-18490
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Mitigating Phase Correlations in Quantum Key Distribution Using Path-Selection Modulation Preprint: APS/123-QED Amita Gnanapandithan, Li Qian, Hoi-Kwong Lo Abstract Phase correlations are an under-explored vulnerability in QKD. Here, we present an experimental and simulated characterization of correlations arising from electro-optic phase encoding, over repetition rates up to the GHz level. To mitigate this vulnerability (and all side channels arising from active phase modulators), we propose a “path-selection modulation” source that eliminates the need for active phase modulation altogether. Encoding is achieved by randomly selecting between multiple paths, each path corresponding to one of the desired encoded states. Phase randomization is achieved using gain-switching. We characterize this source at a clock rate of 1 GHz. I Introduction Quantum key distribution (QKD) is a technique enabling two parties to securely share a secret key by leveraging the postulates of quantum mechanics Xu et al. 2020 . This process involves encoding information into a sequence of quantum optical states, often utilizing polarization or phase states. Electro-optic phase modulators (PM) are commonly used in this encoding process. Despite the theoretical security guarantees of QKD, practical implementations often fall short due to various imperfections. These imperfections manifest as side channels that can potentially leak information to an eavesdropper Kurtsiefer et al. 2001 ; Makarov et al. 2006 ; Brassard et al. 2000 ; Sun et al. 2015 ; Fung et al. 2007 . While the measurement-device-independent QKD (MDI QKD) protocol eliminates detector side channels Lo et al. 2012 , source side channels continue to pose a significant risk Brassard et al. 2000 ; Sun et al. 2015 ; Fung et al. 2007 . The recent drive towards high-speed QKD sources operating at GHz clock rates has unveiled new source side channels that demand attention. Bandwidth-limited modulation has been shown to introduce hidden multidimensional modulation side channels, such as time-varying encoding Gnanapandithan et al. 2025 . Furthermore, bandwidth-limited modulation has been shown to produce source correlations ichiro Yoshino et al. 2018 ; Roberts et al. 2018 ; Lu et al. 2021 ; Kang et al. 2023 ; Pereira et al. 2020 ; Agulleiro et al. 2025 . Assuming a decoy state protocol with phase or polarization encoding, these correlations typically fall into one of the following categories – intensity correlations from decoy state generation, phase correlations from phase randomization, and phase/polarization correlations from phase/polarization encoding. Multiple studies regarding intensity correlations (commonly referred to as the pattern effect ) from the use of bandwidth-limited intensity modulators (IM) have been reported thus far ichiro Yoshino et al. 2018 ; Roberts et al. 2018 ; Lu et al. 2021 ; Kang et al. 2023 ; Agulleiro et al. 2025 . These studies have a) characterized intensity correlations in GHz-clocked sources and b) proposed/demonstrated various mitigation methods. These methods (based on filtering, precalibration, or improved hardware designs) largely take advantage of the sinusoidal voltage-intensity relationship (specifically, the zero-slope points) of IMs ichiro Yoshino et al. 2018 ; Roberts et al. 2018 ; Lu et al. 2021 ; Kang et al. 2023 . For example, the technique discussed in Lu et al. 2021 reduces correlations by over two orders of magnitude at a 1 GHz repetition rate through the use of dual IMs. Here, a correlation is defined as the percent difference in the intensity of a signal/decoy pulse with respect to the nominal intensity (signal/decoy/vacuum) of the preceding pulse. On the other hand, there has been relatively little focus on characterizing and mitigating phase correlations arising from the use of bandwidth-limited PMs. Unlike electro-optic IMs, electro-optic PMs have a linear relationship with voltage, necessitating a separate (albeit analogous) characterization procedure, as well as entirely new mitigation methods. Hence, we report a comprehensive experimental characterization of phase correlations resulting from active (electro-optic) PMs for polarization/phase encoding. We performed this experimental characterization at 500 MHz and 1 GHz, using equipment bandwidths typically used in GHz QKD experiments ichiro Yoshino et al. 2018 ; Lu et al. 2021 ; Kang et al. 2023 ; Roberts et al. 2018 . In addition, we performed a simulated characterization at various repetition rates up to 3 GHz. Fully passive QKD has been proposed as a solution for all modulator side channels at the source, including both intensity and phase/polarization correlations Wang et al. 2023 . Experimental demonstrations have also been reported, albeit at low source repetitions of 20 and 50 MHz Lu et al. 2023 Hu et al. 2023 . While intriguing, the fully passive QKD protocol requires a significant level of post-selection, resulting in an order of magnitude reduction in the secret key rate. In addition, fully passive QKD requires multiple detectors at the source, which limits the source repetition rate and may introduce vulnerabilities due to imperfections in the detectors Wang et al. 2023 . Hence, this work highlights an alternative option to fully passive QKD for mitigating source correlations and other side channels arising from PMs – a path-selection modulation QKD source which we demonstrate at a repetition rate of 1 GHz using 1550 nm single-mode fiber, requiring no additional post-selection or the use of detectors. In addition to performing passive polarization encoding using multiple optical paths, it uses gain-switching to generate phase-randomized pulses Kobayashi et al. 2014 ; Paraïso et al. 2019 . Although the source does not use phase/polarization modulators, it still uses IMs to select the path taken by each pulse, which determines its polarization state. Specifically, IMs are used only for pulse suppression and transmission in the passive polarization encoding setup. When operated in this on/off manner, any resulting intensity correlations have been shown to be negligible at GHz repetition rates ichiro Yoshino et al. 2018 ; Lu et al. 2021 , which we further verify through experimental characterization. This paper is organized as follows. We start by discussing our experimental and simulated characterization of encoding correlations resulting from PMs in Section II . We then proceed to introduce and discuss our schematic for a path-selection modulation source in Section III . Lastly, in Section IV , we experimentally characterize the source. Specifically, we characterize the a) generated polarizations, b) level of phase randomization achieved through gain-switching, c) indistinguishability of the generated polarization states, and d) intensity correlations from using IMs. II Characterization of Correlations from Active Phase Modulation Here, we present experimental and simulated characterizations of phase correlations (including non-nearest-neighbour correlations) when using electro-optic PMs to perform encoding. Experimental characterization (constrained by equipment bandwidth) was performed at 500 MHz and 1 GHz, while simulated characterization was performed at 500 MHz, 1 GHz, 2 GHz, and 3 GHz. II.1 Experimental Characterization The experimental setup shown in Figure S1 was used to obtain the characterization results. Phase modulation was achieved by employing an electro-optic IM – i.e. an electro-optic PM that uses a Mach-Zehnder interferometer to convert a phase change to an intensity change. In this way, modulated phase can be measured and characterized using a photodetector. When the modulator is biased at its minimum, its phase-intensity relationship can be expressed as per Equation S1 , where \phi represents phase and I represents normalized intensity. Note that \phi has a linear relationship with the voltage applied to the modulator. \phi=2cos^{-1}(\sqrt{1-I}) (1) Figure 1: Experimental setup for measuring phase correlations at