Side Channel Exploitation in Transmitter Imperfections

System

No two laser diodes emit identical light, so a transmitter (Alice) that encodes by switching between several of them stamps its setting onto properties of the pulse it never meant to use: the spectrum, the shape in time, the spatial mode. A threat actor (Eve) who measures those properties on the link learns something about each bit without touching the quantum state that carries it. This is a passive side channel of the light source, and the quantity that decides how much it gives her is how distinguishable the emitted signal states are from one another [Babukhin2020,Duplinskiy2021].

What makes the weakness worth cataloguing is that on its strongest routes the leak is silent. Where the threat actor reads the side channel on its own and leaves the transmitted state alone, she adds no errors at the receiver (Bob), so the error-rate check that catches an ordinary intercept-and-resend attack cannot see her at all [Babukhin2021]. Not every route is free. The oldest one replaces each state and pays the usual error cost, and one later strategy has to disturb the signal to work at all; which route costs what is set out in the expert description.

What she gains on present-day hardware is a share of the key rather than the key. Every study of the attack concludes that the protocol survives: the secure key rate falls and the usable distance shortens, because real laser sources are already close enough to identical that the margin she wins is small [Babukhin2020,Babukhin2022a]. The mechanism is not benign in principle. The same model taken to fully distinguishable states describes a protocol completely compromised through side channels, so the reassuring verdict is a statement about today's sources and not about the weakness [Babukhin2022].

Literature 12
[Babukhin2020] D. Babukhin and D. Sych. "Intercept-resend attack on passive side channel of the light source in BB84 decoy-state protocol" In: J. Phys.: Conf. Ser. 1695, 012119. (2020) 10.1088/1742-6596/1695/1/012119.
[Babukhin2021] D. Babukhin and D. Sych. "Explicit attacks on passive side channels of the light source in the BB84 decoy state protocol" In: J. Phys.: Conf. Ser. 1984, 012008. (2021) 10.1088/1742-6596/1984/1/012008.
[Babukhin2022] D. Babukhin and D. Sych. "Joint eavesdropping on the BB84 decoy state protocol with an arbitrary passive light-source side channel" (2022) 10.1134/S1995080224602753. arXiv:2211.13669.
[Babukhin2022a] D. Babukhin, D. Kronberg, and D. Sych. "Explicit attacks on the BennettBrassard 1984 protocol with partially distinguishable photons" In: Phys. Rev. A 106, 042403. (2022) 10.1103/PhysRevA.106.042403.
[Duplinskiy2021] A. Duplinskiy and D. Sych. "Bounding passive light-source side channels in quantum key distribution via Hong-Ou-Mandel interference" In: Phys. Rev. A 104, 012601. (2021) 10.1103/PhysRevA.104.012601.
[Molotkov2020] S. Molotkov. "Trojan Horse Attacks, Decoy State Method, and Side Channels of Information Leakage in Quantum Cryptography" In: J. Exp. Theor. Phys. 130, 809–832. (2020) 10.1134/S1063776120050064.
[Molotkov2020b] S. Molotkov and K. Balygin. "Side channels of information leakage in quantum cryptography based on geometrically uniform coherent states" In: Laser Phys. 30, 065201. (2020) 10.1088/1555-6611/ab8298.
[Nauerth2009] S. Nauerth, M. F¨urst, T. Schmitt-Manderbach, H. Weier, and H. Weinfurter. "Information leakage via side channels in freespace BB84 quantum cryptography" In: New J. Phys. 11, 065001. (2009) 10.1088/1367-2630/11/6/065001.
[Sych2021] D. Sych, A. Duplinskiy, and D. Babukhin. "Practical security of quantum key distribution in the presence of side channels" In: J. Phys.: Conf. Ser. 1984, 012001. (2021) 10.1088/1742-6596/1984/1/012001.
[Tajima2026] S. Tajima, A. Tomita, A. Okamoto. "Hong-Ou-Mandel test to verify indistinguishability of the states emitted from a quantum key distribution transmitter implementing decoy Bennett-Brassard 1984 protocol". (2026) arXiv:2603.26488.
[Tamaki2026] Kiyoshi Tamaki, Marcos Curty, Akihiro Mizutani, Alvaro Navarrete. "Deterministic QKD source robust against side-channel attacks". (2026) arXiv:2607.28063.
[Zapatero2024] Víctor Zapatero, Álvaro Navarrete, Marcos Curty. "Implementation Security in Quantum Key Distribution" (2024) 10.1002/qute.202300380.
QID: A-0083
UUID: b7737bf1-dbb1-46f2-bab3-b537aa7074ad
Tier: T0
Type: Quantum
Domain(s):
  • Quantum Communication
Tactic: Execution
Created: 2024-04-15
Updated: 2026-11-14
Maturity: Theoretical
Technique properties

Attack type: Key extraction
Attack category: Degree-of-freedom coupling attack
Target party: Transmitter
Target function: Quantum layer
Quantum channel: Fibre optic, Free space
Security layers: L1 — Physical Layer (Quantum & Optical) L5 — Post-Processing & Cryptographic
Maturity: Theoretical
Operational properties

Technique → Countermeasures 7
Description Countermeasure

Selecting laser diodes with closely matching wavelengths, and ordering them in the transmitter module, reduces the spectral route of this side channel at its origin [Nauerth2009].

CM-0103 Advanced Laser Diodes

Duplinskiy2021 specifies it inside the transmitter: a beam splitter after state preparation feeds a testing arm whose intensity is set independently of the communication signal, with optical isolators at the output of Alice after the beam splitter so the test cannot be influenced from outside, and its stated purpose is to allow Alice to control the mismatch of her lasers on the fly and continuously monitor potential passive side channels [Duplinskiy2021]. Continuous operation is a requirement rather than a refinement, because even if the laser sources have been perfectly matched at some point, there is no guarantee that their behaviour does not change over a long time interval [Duplinskiy2021]. CM-0016's own scope, checking that each signal is correctly timed, appropriately polarized and emitted at the correct wavelength, is exactly the set of degrees of freedom this side channel lives in.

CM-0016 State Preparation Monitoring

Spectral filtering at the transmitter's output narrows the band the emitted light occupies, and the spectrum is half of what Duplinskiy2021 identifies as the dominant cause of the mismatch this technique reads, naming the combination of frequency chirp and emission time jitter [Duplinskiy2021]. CM-0013's own scope is filters that selectively allow specific signals to enter or exit the device in order to maintain the purity of the quantum states, which covers the outbound direction this control needs. Tajima2026 cites an earlier measurement in which narrowing the spectral filter raised the Hong-Ou-Mandel visibility, which shows that spectral mismatch lowers the visibility [Tajima2026]. The spatial and polarisation axes are closed by a different control, single-mode fibre coupling (CM-0055).

CM-0013 Employment of Optical Filters

Coupling every emitted state into a single-mode fibre before it leaves the transmitter closes the spatial axis of this side channel by construction, which is CM-0055's own stated mechanism: it confines the light within a single spatial mode of propagation, significantly reducing spatial distinguishability among the states. Tajima2026 reaches the same conclusion from the measurement side and extends it to polarisation, stating that transverse modes and polarisation measurement may be omitted because they can be defined in the transmitter simply by using single-mode polarization-maintaining fibers [Tajima2026]. Spatial shape is one of the three properties Duplinskiy2021 says always vary between lasers [Duplinskiy2021], and this control removes it from the list the threat actor can read, which is why it is the transmitter-side answer rather than the receiver-side spatial filter CM-0045.

CM-0055 Single-Mode Fiber Coupling

Device-independent QKD removes a source side channel, and it does so at a stated cost. Babukhin2022 is explicit: the device-independent QKD excludes side channels of the source and the receiver simultaneously, but at the cost of complicated experimental realization and low secret key rate at practically reasonable distances [Babukhin2022]. Switching to it is an instance of employing another quantum protocol, which is what this countermeasure covers.

Measurement-device-independent QKD does not close this channel, and neither does twin-field QKD. Three of this technique's own linked sources say so. Babukhin2022 states that a measurement-device-independent QKD protocol allows closing every possible quantum hacking on the receiver side [Babukhin2022]; Babukhin2022a that MDI and twin-field exclude the measurement devices, but this protocol does not close loopholes on Alice's transmitting side, which require specific characterization and countermeasures [Babukhin2022a]; and Duplinskiy2021 that the development of MDI-QKD allowed to secure the receiving part, however, the sender of the quantum states is still at risk [Duplinskiy2021]. In MDI-QKD both parties are transmitters, so the exposure is doubled rather than removed, and this technique carries a separate protocol-class junction to MDI-QKD for that reason. An MDI deployment is not immune to this technique.

No source analyses an entanglement-based protocol against a passive source side channel, so the catalogue claims neither that it closes the channel nor that it fails to.

CM-0006 Employing Another Quantum Protocol

The catalogue's own architectural answer to the carrier this technique is demonstrated against. CM-0056 Single Laser Diode Utilization removes the spatial, spectral and temporal differences between separate lasers by employing a single laser diode for state generation, and that is exactly the leak Babukhin2022 derives from the source's emitter count when it states that, where the states are prepared by different lasers, the number of side-channel basis states is equal to the number of lasers [Babukhin2022].

CM-0056 Single Laser Diode Utilization

Consider the attack in the security proof.

Including adjusting the security parameters based on observable data at the receiving end, ensuring that the final key remains secure despite any detected vulnerabilities [Molotkov2020b].

Molotkov2020b states exactly that, that an estimate of the secret key length is obtained which is expressed only in terms of the observed values at the receiving station and the parameters of quantum states in the side channels [Molotkov2020b]. The mechanism by which a side channel is made computable inside the ordinary protocol is Babukhin2022's effective-error construction: the key rate can be written either with the real receiver error and an inflated eavesdropper bound, or with the bound unchanged and an inflated effective receiver error, and the second form lets the standard decoy-state machinery consume the leak. That is also why the control is needed at all here, because the method captures eavesdropping of the side channel which is unnoticeable through a standard communication error calculation [Babukhin2022].

CM-0014 Full Device Imperfections Incorporation in Scientific Security Proof
Technique → Security Layers 2
Description Category Impact Type

Post-processing and cryptographic layer, as a secondary effect, and this is where the attack is actually paid for. Three of the four routes complete only here, because after Alice and Bob exchange their basis choices on all bit positions Eve makes a full register measurement and obtains a binary string correlated with the distributed bits [Babukhin2022]. It is also the layer every control on this row operates in: the leak is priced by rewriting it as an inflation of the receiver's effective error so the standard decoy-state machinery can consume it [Babukhin2022], and the secret key length is then estimated only in terms of the observed values at the receiving station and the parameters of quantum states in the side channels [Molotkov2020b]. The cost lands here as a reduced key rate and a shortened secure distance rather than as an observable at L1.

L5 — Post-Processing & Cryptographic Secondary Effect

Physical layer, quantum and optical, and the primary target: the leak is a property of the light the transmitter emits. Zapatero2024 places it there, holding that Alice's settings become encoded in other degrees of freedom associated with the emitted light, such as frequency, space and time [Zapatero2024], and Duplinskiy2021 gives the physical cause at the same layer, that even if two lasers seem identical at first sight the actual parameters of the emitted light always vary [Duplinskiy2021]. Tajima2026 makes the layer boundary a security criterion, since as long as the state originates from the light source it does not affect security provided it is independent of modulation [Tajima2026].

L1 — Physical Layer (Quantum & Optical) Primary Target
Related Techniques 2
Relation type Description Technique
Alternative to

Two ways to the same leak from the transmitter. This technique reads setting information that the light source puts into unused properties of its own light: passive generation of information leakage may be produced via mode dependencies in Alice's transmitter [Zapatero2024]. The Trojan-horse attack reaches a similar effect from outside: Eve could actively provoke it, for example by launching a Trojan-horse attack against the transmitter [Zapatero2024]. The two routes need different defences. Duplinskiy2021 notes that the modulator-free carrier of this attack is immune to conventional Trojan-horse attacks while still being insecure because of information leakage beyond the single-mode space [Duplinskiy2021].

A-0049 Quantum Trojan-Horse Attack (THA)
Refines

One route of this technique is an intercept-and-resend attack made cheaper by the side channel, which is what refines records. Babukhin2020 computes exactly that: the threat actor performs an optimal state discrimination over the enlarged signal space and resends what she found, and the side channel is what lowers her error cost, since applying the ordinary approach to BB84 leads to a critical error of 0.25 but if states used for transmission have additional distinguishability, the resulting critical error will be lower [Babukhin2020]. The relation is refines rather than variant_of because the two rows sit in different attack categories, A-0045 in intercept-and-resend and this row in degree-of-freedom coupling, and only one of this technique's four routes resends anything at all.

A-0045 Intercept-and-Resend Attack
Techniques Relating to This One 1
Relation type Description Technique
Alternative to

The same distinguishability analysis applied at the other end of the link. A-0086 exploits imperfections that make a receiver's detectors distinguishable from one another; A-0083 exploits imperfections that make a transmitter's emitted states distinguishable. The primary itself draws the connection in its closing paragraph, noting that if the states to be sent are prepared by different physical devices their temporal response needs to be characterised and the possible information leakage evaluated with a similar analysis [LamasLinares2007]. That analysis was later carried out on an eight-diode free-space transmitter, which measured 2.8×10−32.8 \times 10^{-3} bits per pulse of temporal leakage after the emitters had been overlapped with programmable delay lines [Nauerth2009]. Both are passive leaks of partial key that add no errors.

A-0086 Timing Attack Exploiting Detector Response Times
See Also
A-0081 Spatial Misalignment Exploitation in QKD
A-0082 Wavelength Manipulation Attack in Free-Space
A-0084 Passive Faraday Mirror Attack
A-0085 Dead Time Correlation Attack
A-0086 Timing Attack Exploiting Detector Response Times