Differential-Phase-Shift Quantum Key Distribution Kyo Inoue Osaka University NTT Basic Research Laboratories JST CREST Collaboration with H. Takesue, T. Honjo (NTT Basic Res. Labs.) Yamamoto group (Stanford Univ.)
Contents (1)DPS-QKD Setup & Protocol Eavesdropping Experiments (2) Modified protocol with decoy pulses (3) Entanglement-based schemes (4) DPS-QKD using macroscopic coherent light (5) DPS quantum secret sharing
DPS (Differential-Phase-Shift) QKD coherent pulse source {0, π} phase mod. att. 0.1-0.2 photon/pulse T T Bob DET-2 Protocol (1) Signal transmission (2) Bob : photon detection time (3) knows which detector clicked at Bob. (4) Key bits are created as DET-1 = 0 DET-2 = 1 Features Simple configuration Efficient usage of the time domain No photon discarded Robustness against photon number splitting attack DET-1 time Δθ = 0 DET-1 Δθ = π DET-2
Eavesdropping - intercept & resend - 0.1-0.2 photon/pls. Eve Bob Transmitter A photon is detected once in 10 slots. She sends a photon over two pulses with measured phase difference. She sends nothing for unmeasured slots. error 1 1 1 = 2 2 4 Eavesdropping!
Eavesdropping - photon number splitting - Eve.... photon number measurement (QND) split gate lossless Bob key creation rate (log) DPS BB84 with laser transmission loss (db) error Eavesdropping!
DPS-QKD Experiment cw-laser rep. rate: 10 GHz intensity mod. {0, π} phase mod. att. fiber waveguide interferometer pulse pattern generator data generator SSPD low noise non-gating low jitter logic unit time interval analyzer Takesue et al., Nature Photon., 1, 343 (2007) collaborating with NIST
Result 17 kbit/s at 100 km. 12 bit/s at 200 km. SSPD QE=1.4 % d.c.=50 cps Secure key against general individual attack based on Edo, Takesue, Yamamoto, PRA 73, 012344 (2006).
Field Transmission NTT Yokosuka Office pulse rate: 1 GHz 17.8 km NTT Yokosuka R&D Center Up-conversion photon detector pump (0.98μm) photon (1.55μm) PBS λ/2 BS PPLN filter 0.6μm Si-APD delay (200ps) for polarization independency QE: 2 %, d.c.: 2.8 kcps
Result 6 QBER (%) 5 4 3 Rate 120.26kbps (Avr) QBER 3.14% (Avr) 150000 100000 50000 Sifted key generation rate (bps) 2 0 100 200 300 Time (min) Sifted key:120 kbit/s with a QBER of 3.14 %.
(1)DPS-QKD Setup & Protocol Eavesdropping Experiments (2) Modified protocol with decoy pulses (3) Entanglement-based schemes (4) DPS-QKD using macroscopic coherent light (5) DPS quantum secret sharing
Conventionally Eavesdropping is found by bit error rate Eve Bob bit error eavesdropping! Eve gets some key bits, utilizing system errors. modified version DPS-QKD with Decoy Pulses Bob μ μ 10μ μ μ μ 5.5μ 5.5μ μ
Intercept & Resend against DPS-QKD with Decoy Eve Trans. Bob decoy click Eve does not know whether a click is decoy or not. count rate: 1 : 2 : 1 Eavesdropping! Intercept & Resend attack is prohibited.
Simulation 10-1 key creation rate (/pulse) 10-2 10-3 10-4 10-5 10-6 conventional DPS suffering from general individual attack decoy DPS fiber loss: 0.25 db/km dark count: 10-5 /gate detection efficiency: 0.1 20% fluctuates in detection rate. 10-7 0 20 40 60 80 100 120 distance (km) Transmission length can be extended.
(1)DPS-QKD Setup & Protocol Eavesdropping Experiments (2) Modified protocol with decoy pulses (3) Entanglement-based schemes (4) DPS-QKD using macroscopic coherent light (5) DPS quantum secret sharing
future scheme for long distance DPS-QKD utilizing Entanglement pump pulse source Entanglement generation < 1 photon/pls. parametric medium < 1 photon/pls. Bob signal idler correlation secret key
Quantum Relaying DPS-QKD signal 1 idler 1 idler 2 signal 2 Bob Entanglement generation Entanglement generation Charlie
Experiment: Entanglement Transmission LD IM EDFA FBG PPLN (SHG) 1.5μm suppression Filter 1551nm 1 GHz 100 ps PPLN (parametric down conversion) waveguide interferometer 1547,1555nm separation DSF(50km) Filter Filter pump (0.7-μm) Suppression DSF(50km) waveguide interferometer frequency up-conversion detector QE: 4.4 1.4 % d.c.: 4kcps 1547nm time interval analyzer 1555nm Frequency up-conversion detector
Result Average number of photon pair: 0.07/pilse. Coincidence rate per signal count 0.0001 5e 05 two-photon interference 15000 10000 5000 Count rate for idler (Hz) data on TIA 0 0 24.0 24.5 PLC temperature for idler (deg.) waveguide temperature for idler (deg) Visibility of 81.6% without removing background noise. Time-bin entangled photons are successfully transmitted over 50 x 2 km.
(1)DPS-QKD Setup & Protocol Eavesdropping Experiments (2) Modified protocol with decoy pulses (3) Entanglement-based schemes (4) DPS-QKD using macroscopic coherent light (5) DPS quantum secret sharing
Conventionally, photon counting is needed. DPS-QKD using Macroscopic Coherent Light Coherent source Phase mod. {-δ, δ} θ i = {δ, δ} θ 4 θ 3 θ 2 θ 1 Bob dec. Im[E] δ quantum noise Re[E] θ 4 θ 3 θ 2 θ 1 θ 4 θ 3 θ 2 θ 1 Δθ = {δ, 0, δ} Probability signal level, I
Protocol (1) Bob:Signal transmission -I d I d (2) Bob creates bit 1 when I > I d bit 0 when I < - I d 0 1 (3) Bob :Time slot at which bit was created (4) creates bit 1 in case θ i θ i+1 = 2δ bit 0 in case θ i θ i+1 = 2δ for the time slot at which Bob created bit. (5) Bob:Time slots for which θ i θ i+1 = 0 (6) Bob discards the bits for θ i θ i+1 = 0 Secret key Conventional photodetectors are available.
Simulation (1) error correction Bob Final key creation rate: R s (I AB max{i AE, I BE }) R s : sifted key rate I AB : mutual information between & Bob I AE : mutual information between & Eve I BE : mutual information between Bob & Eve 1 Key creation rate (/slot) 10-1 10-2 10-3 10-4 k = 0.25 k = 0.5 k = 1 k is a parameter indicating performance of Bob s detector relative to Eve s. k α α B E β β E B α:detection efficiency β:noise factor 10-5 0 10 20 30 40 50 Fiber length (km)
Simulation (2) error correction Bob Final key creation rate: R s (I AB I BE ) 1 key creation rate (/slot) 10-1 10-2 10-3 10-4 10-5 k = 1 k = 0.5 k = 0.25 0 50 100 150 200 fiber length (km)
(1)DPS-QKD Setup & Protocol Eavesdropping Experiments (2) Modified protocol with decoy pulses (3) Entanglement-based schemes (4) DPS-QKD using macroscopic coherent light (5) DPS quantum secret sharing
Quantum Secret Sharing (QSS) Function and Bob have fractions of a secret key shared with Charlie. (or Bob) cannot decipher message from Charlie by her (or him) alone. partial key partial key secrete key secrete key Bob Charlie Previous scheme - Entanglement based scheme - BB84 based scheme
DPS Quantum Secret Sharing (QSS) Charlie ~ 1ph./pls Bob att. coherent phase phase pulse mod. mod. source {0, π} {0, π} Δθ Δθ a monitor b 0 1 Charlie s data are XOR of s and Bob s. Δθ b Δθ a 0 π 0 0 1 π 1 0 Charlie s data are recovered in collaboration of and Bob. QSS
Eavesdropping against DPS-QSS Eavesdropping by dishonest Bob Bob ~ 0.1 photon/pls. measure coherent pulse source phase mod. Charlie Bob cannot fully know s data. Eavesdropping by dishonest Bob ~ 0.1 photon/pls. phase mod. measure coherent pulse source phase mod. Charlie Bob s monitoring forces to send 0.1 ph/pls. cannot fully know Bob s data.
Experiment cw-laser pulse pattern generator intensity mod. rep: 1 GHz pulse width: 125 ps 1 Gbps 1 Gbps data generator phase mod. att. 0.1 ph./pls fiber(5km) 10:1 coupler APD data generator phase mod. fiber(5km) waveguide interferometer APD (4MHz) ΔL = 20 cm time interval analyzer QBER: 6.4 % sifted key rate: 3.9 kbps error correction privacy amplification final key rate: 1.5 kbps
Summary DPS-QKD is presented. (1) Setup & protocol, eavesdropping, experiments Simple configuration, no photon discarded. Robust against photon-number-splitting attack 12 bit/s at 200 km, 17 kbit/s at 100 km for secure key (with SSPD) (2) Modified protocol with decoy slots Intercept-resend attack is prohibited. (3) Entanglement-based schemes Experiment utilizing fiber four-wave mixing for entanglement generation. (4) DPS-QKD using macroscopic coherent light Conventional photodetectors are available. (5) DPS quantum secret sharing Simple configuration