Polina Acheva

Quantum key distribution (QKD) promises information-theoretic security, yet practical implementations remain vulnerable to device-level attacks such as detector blinding. In such attacks, an eavesdropper can apply optical pulses of increased optical power to manipulate single-photon detectors, gaining control over the receiver’s electrical output and compromising the key. While superconducting nanowire single-photon detectors (SNSPDs) offer high efficiency, low dark counts, and excellent timing, they are not immune to these attacks. Existing countermeasures either require additional hardware or complex protocols, often reducing performance. A key challenge is therefore to design intrinsically secure detectors that provide direct detection of blinding attempts without adding system complexity or relying on statistical analysis. In this work, we demonstrate that waveguide-integrated SNSPDs (WI-SNSPDs) with engineered series inductances act as an intrinsic countermeasure. By varying the series inductance of an SNSPD between ≈38 nH and 442 nH, we show that detectors with low kinetic inductance (<172 nH) enter a latched resistive state immediately when exposed to optical pulses of elevated optical power, preventing blinding entirely. This approach embeds hardware-level security directly into the detector design and leverages the flexibility of waveguide integration for low-inductance configurations. Beyond attack prevention, such designs can be integrated into scalable photonic circuits, combining high-performance quantum detection with robust security features for practical QKD systems.

Quantum attacks to single-photon detectors with bright-light are known for more than a decade. Many countermeasures were suggested to protect detectors, but the most of them can close some attacks with given parameters but not a whole attack group. To solve the problem we are developing automated testbench that emulates attacks by an eavesdropper Eve. It combines emission of pulse laser and continuous-wave laser and observes detector's response. In future we hope to automatically prepare reports on detectors' safety or show bright-light attacks that were not covered by detectors' countermeasures.