We control using bright light an actively-quenched avalanche single-photon detector. Actively-quenched detectors are commonly used for quantum key distribution (QKD) in the visible and near-infrared range. This study shows that these detectors are controllable by the same attack used to hack passively-quenched and gated detectors. This demonstrates the generality of our attack and its possible applicability to eavsdropping the full secret key of all QKD systems using avalanche photodiodes (APDs). Moreover, the commercial detector model we tested (PerkinElmer SPCM-AQR) exhibits two new blinding mechanisms in addition to the previously observed thermal blinding of the APD, namely: malfunctioning of the bias voltage control circuit, and overload of the DC/DC converter biasing the APD. These two new technical loopholes found just in one detector model suggest that this problem must be solved in general, by incorporating generally imperfect detectors into the security proof for QKD.
We control using bright light an actively-quenched avalanche single-photon detector. Actively-quenched detectors are commonly used for quantum key distribution (QKD) in the visible and near-infrared range. This study shows that these detectors are controllable by the same attack used to hack passively-quenched and gated detectors. This demonstrates the generality of our attack and its possible applicability to eavsdropping the full secret key of all QKD systems using avalanche photodiodes (APDs). Moreover, the commercial detector model we tested (PerkinElmer SPCM-AQR) exhibits two new blinding mechanisms in addition to the previously observed thermal blinding of the APD, namely: malfunctioning of the bias voltage control circuit, and overload of the DC/DC converter biasing the APD. These two new technical loopholes found just in one detector model suggest that this problem must be solved in general, by incorporating generally imperfect detectors into the security proof for QKD.
We control using bright light an actively-quenched avalanche single-photon detector. Actively-quenched detectors are commonly used for quantum key distribution (QKD) in the visible and near-infrared range. This study shows that these detectors are controllable by the same attack used to hack passively-quenched and gated detectors. This demonstrates the generality of our attack and its possible applicability to eavsdropping the full secret key of all QKD systems using avalanche photodiodes (APDs). Moreover, the commercial detector model we tested (PerkinElmer SPCM-AQR) exhibits two new blinding mechanisms in addition to the previously observed thermal blinding of the APD, namely: malfunctioning of the bias voltage control circuit, and overload of the DC/DC converter biasing the APD. These two new technical loopholes found just in one detector model suggest that this problem must be solved in general, by incorporating generally imperfect detectors into the security proof for QKD.
We control using bright light an actively-quenched avalanche single-photon detector. Actively-quenched detectors are commonly used for quantum key distribution (QKD) in the visible and near-infrared range. This study shows that these detectors are controllable by the same attack used to hack passively-quenched and gated detectors. This demonstrates the generality of our attack and its possible applicability to eavsdropping the full secret key of all QKD systems using avalanche photodiodes (APDs). Moreover, the commercial detector model we tested (PerkinElmer SPCM-AQR) exhibits two new blinding mechanisms in addition to the previously observed thermal blinding of the APD, namely: malfunctioning of the bias voltage control circuit, and overload of the DC/DC converter biasing the APD. These two new technical loopholes found just in one detector model suggest that this problem must be solved in general, by incorporating generally imperfect detectors into the security proof for QKD.
In this study, we investigated several possible attacks that an eavesdropper could launch against cryptosystems equipped with yet another commonly used silicon photon counting detector module operating with active quenching and manufactured by Perkin Elmer (model SPCM-AQR). We showed that when the avalanche photo-diode (APD) of the detector gets illuminated with bright optical pulses (instead of single photons), the voltage at which the APD is biased drops by more than 12-14 V if the repetition rate of the bright pulses is large enough (above 70kHz here).
Single photon detectors (SPDs) based on passively quenched avalanche photodiodes can be temporarily blinded by relatively bright light, of intensity less than 1 nW. A bright-light regime suitable for attacking a quantum key distribution system containing such detectors is described in this paper. In this regime, all SPDs in the receiver Bob are uniformly blinded by continuous illumination coming from the eavesdropper Eve. When Eve needs a certain detector in Bob to produce a click, she modifies the polarization (or other parameters used to encode quantum states) of the light she sends to Bob such that the target detector stops receiving light, while the other detector(s) continue to be illuminated. The target detector regains single photon sensitivity and, when Eve modifies the polarization again, produces a single click. Thus, Eve has full control of Bob and can perform a successful intercept–resend attack. To check the feasibility of the attack, three different models of passively quenched detectors have been tested. In the experiment, I have simulated the intensity diagrams the detectors would receive in a real QKD system under attack. Control parameters and side effects are considered. It appears that the attack could be practically possible.
It has previously been shown that the gated detectors of two commercially available quantum key distribution (QKD) systems are blindable and controllable by an eavesdropper using continuous-wave illumination and short bright trigger pulses, manipulating voltages in the circuit [Nat. Photonics 4, 686 (2010)]. This allows for an attack eavesdropping the full raw and secret key without increasing the quantum bit error rate (QBER). Here we show how thermal effects in detectors under bright illumination can lead to the same outcome. We demonstrate that the detectors in a commercial QKD system Clavis2 can be blinded by heating the avalanche photo diodes (APDs) using bright illumination, so-called thermal blinding. Further, the detectors can be triggered using short bright pulses once they are blind. For systems with pauses between packet transmission such as the plug-and-play systems, thermal inertia enables Eve to apply the bright blinding illumination before eavesdropping, making her more difficult to catch.
Quantum key distribution (QKD) allows two users to communicate with information theoretic security by encoding information on single photons. This security is based on the laws of physics and as such can never be broken in theory. However, in practice, components do not always behave according to their theoretical models and these deviations can be exploited by an eavesdropper. In recent years, exposing loopholes in QKD systems, known as quantum hacking, has attracted significant attention. The components most susceptible to being hacked are the single-photon detectors, often avalanche photodiodes (APDs), as they are directly exposed to the optical channel. Whilst measurement-device-independent QKD removes detector vulnerability from the system, secure key rates with this technique can be much lower than point-to-point links. As such, mitigating attacks on QKD systems is a pressing challenge in QKD. In this thesis, the focus is on a special class of detectors, self-differencing APDs (SD-APDs), which have facilitated state-of-the art demonstrations of QKD. The susceptibility of SD-APDs to blinding attacks, the most explored and successful attack to date, was investigated and it was shown that by following best practice for their operation, such an attack would be unsuccessful. We have also proposed and developed a countermeasure such that the onus for appropriate operation could be removed from the user. We have also explored an arguably more dangerous attack, in the form of the after-gate attack. We have shown that delayed detection events, ordinarily considered detrimental in QKD, can provide inherent protection against this attack. Finally, backflashes in GHz-gated APDs were investigated for the first time and it was shown that threat they pose to QKD security is negligible. These results highlight the inherent protection to a number of attacks that self-differencing APDs possess. We stress that the findings presented in this thesis are also applicable to other types of fast-gated InGaAs APDs that don't possess self-differencing circuitry.