Theoretical Foundations

Comprehensive theory of continuous-variable quantum key distribution

Contents

1. Introduction to CV-QKD

Continuous-Variable Quantum Key Distribution (CV-QKD) enables two parties (Alice and Bob) to establish a shared secret key with information-theoretic security guaranteed by quantum mechanics. Unlike discrete-variable QKD that uses single photons, CV-QKD encodes information in the continuous quadratures of the electromagnetic field.

Key Advantages of CV-QKD:

  • Compatible with standard telecom infrastructure
  • Higher key rates at metropolitan distances
  • Room-temperature operation (no cryogenics)
  • Potential for photonic integration
  • Efficient reconciliation with low-density parity-check codes

The security of CV-QKD stems from the Heisenberg uncertainty principle: any eavesdropper (Eve) attempting to measure the quantum states will inevitably introduce detectable noise.

2. Gaussian Quantum States

In CV-QKD, information is encoded in coherent states |α⟩, which are minimum-uncertainty states displaced from the vacuum. The quadrature operators X̂ and P̂ satisfy the commutation relation:

[X̂, P̂] = 2i

For a coherent state |α⟩ with α = x + ip:

⟨X̂⟩ = 2x, ⟨P̂⟩ = 2p
Var(X̂) = Var(P̂) = 1 (Shot Noise Unit)

The Shot Noise Unit (SNU) is the fundamental reference for all noise measurements in CV-QKD. It represents the quantum vacuum fluctuations and equals 1 by convention.

Covariance Matrix

Gaussian states are fully characterized by their first moments (mean values) and covariance matrix γ. For a two-mode state shared between Alice and Bob:

γ = [[V·I, C·σz], [C·σz, V'·I]]

3. GMCS Protocol

The Gaussian-Modulated Coherent State (GMCS) protocol is the most widely implemented CV-QKD scheme:

1

State Preparation (Alice)

Alice draws random values x, p from Gaussian distribution N(0, V_A) and prepares coherent state |α = x + ip⟩

2

Quantum Transmission

State transmitted through quantum channel (fiber) with transmittance T and excess noise ξ

3

Measurement (Bob)

Bob performs homodyne (one quadrature) or heterodyne (both quadratures) detection

4

Classical Post-Processing

Parameter estimation, error correction (reconciliation), and privacy amplification

Modulation Variance V_A

The modulation variance V_A (in SNU) determines the signal strength. Typical values range from 1-20 SNU. Higher V_A increases mutual information I_AB but also increases Eve's accessible information. Optimal V_A depends on channel parameters.

4. Coherent Detection

Homodyne Detection

  • • Measures single quadrature (X or P)
  • • Bob randomly selects measurement basis
  • • Half of data discarded (basis mismatch)
  • • No 3 dB vacuum noise penalty
  • • Shot-noise limited with balanced detection

Heterodyne Detection

  • • Measures both quadratures simultaneously
  • • No basis selection needed
  • • All data used for key generation
  • • 3 dB noise penalty per quadrature
  • • Simpler implementation

Balanced Detection

Balanced (differential) detection uses two photodiodes to subtract common-mode noise, achieving shot-noise limited performance. The signal is interfered with a strong Local Oscillator (LO) at a 50:50 beamsplitter. Key parameters: Common-Mode Rejection Ratio (CMRR), phase matching, and amplitude balance.

5. Quantum Channel Model

The quantum channel is modeled as a thermal-loss channel characterized by transmittance T and excess noise ξ:

Transmittance T

T = η_PIC × 10^(-αL/10) × η_det

α ≈ 0.2 dB/km for standard fiber at 1550 nm

Channel-Added Noise

χ_line = (1 - T) / T

Noise referred to channel input

Total Noise

χ_tot = χ_line + ξ/T + χ_el/(T·η)

Including excess noise and electronic noise

Critical: Excess Noise

Excess noise ξ is the most critical parameter limiting CV-QKD range. Sources include: laser phase noise, timing jitter, polarization drift, and imperfect modulation. A change of 0.005 SNU can halve the achievable distance.

6. Security Analysis

CV-QKD security is proven against various attack models with increasing power:

Individual Attacks

Eve measures each signal independently. Weakest attack model.

Collective Attacks

Eve performs identical operations on each signal but may delay measurement until reconciliation. Security bounded by Holevo information χ_BE.

Coherent (General) Attacks

Eve can perform arbitrary joint quantum operations on all signals. Strongest attack model. Security proven via de Finetti theorem reduction.

Composable Security

Modern CV-QKD proofs provide composable security, meaning the key can be safely used in any cryptographic application. The security is characterized by small failure probabilities ε_sec (secrecy), ε_cor (correctness), with total security ε_tot = ε_sec + ε_cor.

7. Secret Key Rate

The asymptotic secret key rate (in reverse reconciliation) is:

K = β · I_AB - χ_BE

where β is reconciliation efficiency (typically 0.90-0.98)

Mutual Information I_AB

For homodyne detection:

I_AB = ½ log₂[(V_A + 1) / (1 + χ_tot)]

Holevo Bound χ_BE

Eve's accessible information:

χ_BE = g(λ₁) + g(λ₂) - g(λ₃) - g(λ₄)

λᵢ: symplectic eigenvalues

Finite-Size Effects

For practical block sizes N, corrections reduce the key rate:

K_finite = K_asymptotic - Δ_PE - Δ_PA - Δ_EC

Δ_PE: parameter estimation, Δ_PA: privacy amplification, Δ_EC: error correction

8. PIC Implementation

Photonic Integrated Circuit (PIC) implementation offers compact, stable, and potentially mass-manufacturable CV-QKD systems. Key considerations:

Typical PIC Loss Budget

Grating coupler (per facet) 3-6 dB
Waveguide propagation 0.5-2 dB/cm
Modulator 1-3 dB
Splitters/combiners 0.1-0.5 dB
LO routing 2-4 dB

Design Trade-offs

  • • 1 dB PIC loss ≈ 5% effective detector QE reduction
  • • Integration benefits: stability, compactness, cost reduction
  • • Challenges: loss minimization, thermal management, packaging
  • • Platform options: Silicon photonics (high integration), InP (active devices), SiN (low loss)

References

  1. Grosshans et al., "Continuous variable quantum cryptography," Phys. Rev. Lett. 88, 057902 (2002)
  2. Weedbrook et al., "Gaussian quantum information," Rev. Mod. Phys. 84, 621 (2012)
  3. Leverrier, "Composable security proof for CV-QKD," Phys. Rev. Lett. 114, 070501 (2015)
  4. Diamanti et al., "Practical challenges in CV-QKD," npj Quantum Information 2, 16025 (2016)
  5. Zhang et al., "Integrated silicon photonic CV-QKD transmitter," Nature Photonics (2019)