Detector-device-independent quantum key distribution (DDI-QKD) was proposed to combine measurement-device-independent QKD's immunity to detector side channels with the linear detector-efficiency scaling and implementation simplicity of conventional prepare-and-measure QKD, by performing a single-photon Bell-state measurement with uncharacterized detectors inside a trusted receiver. Sajeed et al. (Phys. Rev. Lett. 117, 250505 (2016)) showed that this promise is unfulfilled: DDI-QKD's security is not based on post-selected entanglement, and blinding, trigger-threshold, time-shift, and device-imperfection attacks break it even when all apparatus is built by the trusted users. The open problem is to determine whether detector-device-independence is achievable at all: either exhibit an explicit set of benign assumptions on realistic uncharacterized detectors (nonlinear threshold response, detector-to-detector threshold variability, time-dependent efficiency mismatch) under which a complete security proof exists for a DDI-type protocol — or establish that no such assumption set can suffice. No security proof, validated countermeasure suite, or impossibility result is known for this problem.

Background

Quantum key distribution (QKD) lets two parties, Alice and Bob, grow a shared secret key whose security rests on quantum physics rather than computational assumptions; the canonical example is the BB84 protocol of Bennett and Brassard, in which Alice sends single photons encoded in one of two conjugate bases and Bob measures in a randomly chosen basis. In theory QKD is information-theoretically secure, but real devices deviate from the proof models, and unmodeled imperfections become side channels. The most damaging class in practice targets the single-photon detectors: in a detector blinding attack, the eavesdropper Eve shines bright continuous light onto Bob's avalanche photodiode detectors, forcing them out of the single-photon-sensitive Geiger mode into a classical linear mode in which a click is generated only by pulses above a controllable energy threshold, so Eve can decide which detector clicks and learn the key without raising the error rate (demonstrated against a commercial system by Lydersen et al., Nature Photonics 4, 686 (2010)). Because detectors respond differently — blinding thresholds and detection efficiencies differ from detector to detector and in time — Eve can also exploit detector-to-detector threshold variability and time-shift (time-dependent efficiency mismatch) attacks. Measurement-device-independent QKD (MDI-QKD), introduced by Lo, Curty, and Qi, removes all detector side channels by having Alice and Bob send states to an untrusted relay that performs a two-photon Bell-state measurement (BSM; a joint measurement projecting two photons onto a maximally entangled Bell state). Its security rests on post-selected entanglement: the relay's announced Bell results post-select entanglement between Alice and Bob, so the detectors need no assumption at all. The price is that the key is extracted from two-fold coincidences, so the rate scales quadratically with channel transmittance, and two independent single-photon sources must interfere at a distance. Detector-device-independent QKD (DDI-QKD), proposed by Lim et al. (Appl. Phys. Lett. 105, 221112 (2014)) with independent variants by several other groups, replaces the two-photon BSM by a single-photon BSM: Alice encodes a qubit in one degree of freedom (e.g., polarization) of each photon and Bob encodes his basis choice in another degree of freedom (e.g., spatial path and phase) of the same photon inside his receiver, so the BSM needs no interference of independent sources and the key rate scales linearly with detector efficiency — the claimed advantage over MDI-QKD. The detectors themselves are uncharacterized and need not be trusted, but the receiver's linear-optics network and state preparation must be trusted, and Eve must not influence Alice's or Bob's state preparation. Qi (Phys. Rev. A 91, 020303(R) (2015)) showed DDI-QKD is broken if Eve can substitute Bob's measurement apparatus; Sajeed et al. (Phys. Rev. Lett. 117, 250505 (2016)) then showed the breach is far deeper: the single-photon BSM does not post-select entanglement, and blinding combined with detector threshold variability, a time-shift attack, phase-modulator imperfections, or wavelength-dependent beamsplitter ratios lets Eve take the key even when Alice and Bob build the entire apparatus themselves — refuting the contemporaneous defense of Boaron et al. (J. Appl. Phys. 120, 063101 (2016)), who had argued that a complete four-detector BSM makes blinding detectable through double-click statistics and that security holds under the assumptions that Eve has no physical access to the interferometer outputs and that the detectors come from a non-malicious manufacturer. Boaron et al. did establish that for attacks restricted to single-photon-level inputs DDI-QKD is equivalent to MDI-QKD; the insecurity appears precisely in the realistic regime where Eve may send bright, multi-photon light through the quantum channel into Bob's trusted transmitter. As the review of Diamanti, Lo, Qi, and Yuan (npj Quantum Information 2, 16025 (2016)) put it, DDI-QKD's ability to completely remove detector side channels had yet to be proven, and either Trojan-horse countermeasures or trust in the BSM device is still required — where a Trojan-horse attack means Eve injecting light that traverses Bob's own state-preparation optics to probe or control them. What remains unresolved is whether this protocol family can be saved on terms consistent with its original motivation. A related practical consideration, not part of the completion criterion, is whether a repaired DDI-type protocol would retain a practical secret-key-rate or implementation-complexity advantage over standard prepare-and-measure QKD with trusted detectors and over MDI-QKD.

Problem Statement

Determine whether detector-device-independence is an achievable security notion for practical quantum key distribution — that is, whether the single-photon-Bell-state-measurement protocol family introduced by Lim et al. can be granted rigorous, composable security against eavesdroppers limited to the quantum channel and standard classical side channels, without assumptions that quietly restore full trust in the detectors. This single objective decomposes into scientifically linked components; components (2) and (3) pertain to the positive branch, in which a security proof is exhibited. (1) Assumptions and proof: either exhibit an explicit set of benign assumptions on the uncharacterized detectors — assumptions satisfied by trusted-manufacturer avalanche photodiodes exhibiting realistic nonlinear threshold response, detector-to-detector threshold variability, and time-dependent detection-efficiency mismatch — under which a complete security proof exists for a DDI-type protocol, covering the regime where Eve sends bright multi-photon light through the channel into Bob's trusted transmitter, or prove that no such assumption set suffices. (2) Countermeasures: any receiver modifications or protocol-level countermeasures on which the security argument relies must be shown to cover the demonstrated attack vectors (detector blinding, trigger-threshold control, time-shift, photon-number-adding/Trojan-horse influence on Bob's state preparation, and wavelength- or imperfection-based double-click evasion). (3) State preparation: within any such proof, prevent or bound Eve's influence on Bob's state-preparation process via the incoming signals (optical isolation, authenticated state preparation, or a proof bounding the influence). An answer that establishes security only under assumptions equivalent to fully trusted, characterized detectors does not meet the objective.

References

  1. Shihan Sajeed, Anqi Huang, Shihai Sun, Feihu Xu, Vadim Makarov, Marcos Curty, "Insecurity of Detector-Device-Independent Quantum Key Distribution", Physical Review Letters 117, 250505 (2016), DOI 10.1103/PhysRevLett.117.250505, arXiv:1607.05814, https://doi.org/10.1103/PhysRevLett.117.250505
  2. Charles Ci Wen Lim, Boris Korzh, Anthony Martin, Félix Bussières, Rob Thew, Hugo Zbinden, "Detector-Device-Independent Quantum Key Distribution", Applied Physics Letters 105, 221112 (2014), DOI 10.1063/1.4903350, arXiv:1410.1850, https://doi.org/10.1063/1.4903350
  3. Alberto Boaron, Boris Korzh, Raphael Houlmann, Gianluca Boso, Charles Ci Wen Lim, Anthony Martin, Hugo Zbinden, "Detector-device-independent quantum key distribution: security analysis and fast implementation", Journal of Applied Physics 120, 063101 (2016), DOI 10.1063/1.4960093, arXiv:1607.05435, https://doi.org/10.1063/1.4960093
  4. Bing Qi, "Trustworthiness of detectors in quantum key distribution with untrusted detectors", Physical Review A 91, 020303(R) (2015), DOI 10.1103/PhysRevA.91.020303, https://doi.org/10.1103/PhysRevA.91.020303
  5. Kejin Wei, Hongwei Liu, Haiqiang Ma, Xiuqing Yang, Yong Zhang, Yongmei Sun, Jinghua Xiao, Yuefeng Ji, "Feasible attack on detector-device-independent quantum key distribution", Scientific Reports 7, 449 (2017), DOI 10.1038/s41598-017-00531-y, https://doi.org/10.1038/s41598-017-00531-y
  6. Xiuqing Yang, Kejin Wei, Haiqiang Ma, Hongwei Liu, Zhenqiang Yin, Zhu Cao, Lingan Wu, "Detector-device-independent quantum secret sharing with source flaws", Scientific Reports 8, 5728 (2018), DOI 10.1038/s41598-018-23876-4, https://doi.org/10.1038/s41598-018-23876-4
  7. Eleni Diamanti, Hoi-Kwong Lo, Bing Qi, Zhiliang Yuan, "Practical challenges in quantum key distribution", npj Quantum Information 2, 16025 (2016), DOI 10.1038/npjqi.2016.25, arXiv:1606.05853, https://doi.org/10.1038/npjqi.2016.25
  8. Feihu Xu, Xiongfeng Ma, Qiang Zhang, Hoi-Kwong Lo, Jian-Wei Pan, "Secure quantum key distribution with realistic devices", Reviews of Modern Physics 92, 025002 (2020), DOI 10.1103/RevModPhys.92.025002, arXiv:1903.09051, https://doi.org/10.1103/RevModPhys.92.025002
  9. Hao Shu, "Measurement-Device-Independenization of Quantum Key Distribution Protocols", Chinese Journal of Physics 85, 135–142 (2023), DOI 10.1016/j.cjph.2023.06.019, arXiv:2110.01979, https://doi.org/10.1016/j.cjph.2023.06.019
  10. Hoi-Kwong Lo, Marcos Curty, Bing Qi, "Measurement-Device-Independent Quantum Key Distribution", Physical Review Letters 108, 130503 (2012), DOI 10.1103/PhysRevLett.108.130503, https://doi.org/10.1103/PhysRevLett.108.130503
  11. Lars Lydersen, Carlos Wiechers, Christoffer Wittmann, Dominique Elser, Johannes Skaar, Vadim Makarov, "Hacking commercial quantum cryptography systems by tailored bright illumination", Nature Photonics 4, 686–689 (2010), DOI 10.1038/nphoton.2010.214, https://doi.org/10.1038/nphoton.2010.214
  12. Charles H. Bennett, Gilles Brassard, "Quantum cryptography: Public key distribution and coin tossing", Theoretical Computer Science 560, 7–11 (2014), DOI 10.1016/j.tcs.2014.05.025, https://doi.org/10.1016/j.tcs.2014.05.025
  13. M. A. Ruhul Fatin, Shihan Sajeed, "Generalized efficiency mismatch attack to bypass the detection-scrambling countermeasure", Optics Express 29, 16073 (2021), DOI 10.1364/OE.419338, https://doi.org/10.1364/OE.419338
  14. Polina Acheva, Konstantin Zaitsev, Vladimir Zavodilenko, Anton Losev, Anqi Huang, Vadim Makarov, "Automated verification of countermeasure against detector-control attack in quantum key distribution", EPJ Quantum Technology 10, 22 (2023), DOI 10.1140/epjqt/s40507-023-00178-x, https://doi.org/10.1140/epjqt/s40507-023-00178-x