FPGA控制隨機單雙波長次時序分割量子密鑰共網通信

Abstract

對於長距離量子密鑰分發傳輸上，低暗計數對於最終量子誤碼率影響甚大。而為了降低暗計數，單光子偵測器只會在特定時間開啟以接收量子密鑰並且在接收到後以長時間關閉偵測來降低後脈衝機率。在單光子計數器與傳送端達到完全同步時，這意味著單光子分布在時間上的隨機性也大幅減少且接收速率也大幅降低。 在第二章中，利用 FPGA 隨機位元產生器和 TTL 同步器控制單光子偵測器開啟時間且將單光子偵測器的休息時間設定為1us來使後脈衝機率降低。並提出了一種 6 時間倉 DPSK 編碼資料格式，在同個接收週期同時傳送以增加接收時間上的隨機性，並使用光子到達時間統計分布圖來確認光子位於光子偵測器開啟時間以進行解碼上的切割並估計相對延遲，然後利用HBT實驗得到了二階關聯函數來驗證我們的弱相干源。此外，在長距離傳輸上，將單光子量子通道與同步經典通道傳輸結合，為了觀察串擾效應，應用了不同的波長間距並獲得了串擾引起的暗計數率，然後演示了10公里的傳輸。在1551nm作為量子通道波長的情況下，使用1538nm觸發通道獲得了8 kbit / s的安全密鑰速率和2.9％的QBER。 1545nm 觸發通道，安全密鑰速率為 3 kbit/s，QBER 為 3.9%，1550nm觸發通道則因為暗計數過大而無法進行解調。 在第三章中，藉由使用雙波長切換的注入鎖定，來達到單時間倉但由於傳送端與解調端隨機的波長選擇來增加安全性。為了實現具有高自由光譜範圍延遲干涉儀的低調變電壓，已經採用MHz級線寬主雷射來直接調製相位資訊。對於2.3MHz線寬，主雷射的驅動電流將工作在6 Ith，當延遲干涉儀的自由光譜範圍為2GHz時，由洩漏率引起的理論計算QBER僅為0.14％。 主雷射調變電壓的π相移為280mV，調變時間為0.25ns，初始相位差設為延遲干涉儀最大輸出功率的一半，變化範圍為π至0。為盡量減小波形產生器內建直流阻隔或雷射偏壓器所造成的相位誤差，將調變電壓積分設計為0。對於雙波長傳輸。發送端和接收端選擇兩種波長基準的光子到達時間分布也分別在4*4的圖中顯示。當發射端與接收端選擇相同的基準時，干涉結果將會清楚地觀察到光子0和1的存在。但如果選擇不同的基，光子就會隨機出現在DLI的兩個臂上，最終通過l1-l1(1485nm)在13km城際網路鏈路的安全密鑰速率為861bit/s。由於QBER超過了糾錯極限，在傳輸18km時衰減為零；而l2-l2(1485nm)的安全密鑰速率在1485nm為831 bit/s。由於 QBER 較高，使用錯誤基礎（l2-l1和l1-l2）的安全密鑰速率將保持為 0 bit/s。

For long-distance quantum key distribution transmission, a low dark count rate has a great impact on the final quantum bit error rate. To reduce the dark count rate, single-photon detectors are turned on only at specific times to receive the quantum key and then hold off for a long time after receiving it to reduce the afterpulsing probability. When the single-photon counter is completely synchronized with the transmitting end, the temporal randomness of the single-photon distribution is greatly reduced, and the receiving rate is also greatly reduced. In Chapter 2, an FPGA random bit generator and a TTL synchronizer are used to control the single photon detectors on time, and the single photon detector's hold off is set to 1us to reduce the probability of afterpulses. Proposed a 6-time bin DPSK coded data format, which is transmitted simultaneously in the same receiving period to increase the randomness of the receiving time. Used the statistical distribution histogram of the photon arrival time to confirm that the photon is located at the photon detector turn-on time for decoding, cutting, and estimation of relative delay. Then used the HBT experiment to obtain the second-order correlation function to verify the weak coherence source. In addition, for long-distance transmission, the single-photon quantum channel is combined with the synchronous classical channel transmission. In order to observe the crosstalk effect, different wavelength spacings are applied and the dark count rate caused by the crosstalk is obtained, and then a 10-km transmission is demonstrated. With 1551nm as the quantum channel wavelength, the secure key rate is 8 kbit/s, and a QBER of 2.9% was obtained using a 1538nm trigger channel. The 1545nm trigger channel has a 3 kbit/s security key rate and a 3.9% QBER. The 1550nm trigger channel cannot be demodulated due to the high dark count rate. In Chapter 3, a single time bin is achieved by using injection locking with dual wavelength switching, but security is increased by random wavelength selection at the transmitter and demodulator. To achieve low-voltage modulation with high free spectral range delay interferometer, a MHz-level linewidth master laser has been used to directly modulate the phase information. For a 2.3MHz linewidth, the driving current of the Master laser will operate at 6 Ith, and when the free spectral range of the delay interferometer is 2GHz, the theoretically calculated QBER caused by the leakage rate is only 0.14%. The π phase shift of the Master laser modulation voltage is 280 mV, the modulation time is 0.25 ns, the initial phase difference is set to half of the maximum output power of the delay interferometer, and the variation range is π to 0. In order to minimize the phase error caused by the built-in DC block of the waveform generator or laser bias-Tee device, the modulation voltage integral is designed to be 0. For dual-wavelength transmission. The photon arrival time distributions of the two wavelength bases selected at the transmitter and receiver are also shown in the 4*4 graphs, respectively. When the transmitting and receiving ends select the same bases, the interference results will clearly show the existence of photons 0 and 1. But if a different basis is chosen, the photons will appear randomly on the two arms of the DLI. The security key rate over the 13km intercity network link is 861bit/s via l1-l1 (1485nm), and decays to zero over 18km since the QBER exceeds the error correction limit; and the security key rate of l2-l2 (1485nm) is 831 bit/s at 1485nm. Due to the high QBER, the security key rate using the wrong basis (l2-l1 and l1-l2) will remain at 0 bit/s.