1: \documentclass[showpacs,preprint,aps]{revtex4}%
2: \usepackage{amsfonts}
3: \usepackage{amsmath}
4: \usepackage{amssymb}
5: \usepackage{epsfig}
6: \usepackage{graphicx}%
7: \setcounter{MaxMatrixCols}{30}
8: %TCIDATA{OutputFilter=latex2.dll}
9: %TCIDATA{Version=5.00.0.2552}
10: %TCIDATA{CSTFile=revtex4.cst}
11: %TCIDATA{Created=Saturday, May 15, 2004 16:06:49}
12: %TCIDATA{LastRevised=Wednesday, August 04, 2004 22:18:38}
13: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
14: %TCIDATA{<META NAME="SaveForMode" CONTENT="3">}
15: %TCIDATA{<META NAME="DocumentShell" CONTENT="Articles\SW\REVTeX 4">}
16: %TCIDATA{Language=American English}
17: \newtheorem{theorem}{Theorem}
18: \newtheorem{acknowledgement}[theorem]{Acknowledgement}
19: \newtheorem{algorithm}[theorem]{Algorithm}
20: \newtheorem{axiom}[theorem]{Axiom}
21: \newtheorem{claim}[theorem]{Claim}
22: \newtheorem{conclusion}[theorem]{Conclusion}
23: \newtheorem{condition}[theorem]{Condition}
24: \newtheorem{conjecture}[theorem]{Conjecture}
25: \newtheorem{corollary}[theorem]{Corollary}
26: \newtheorem{criterion}[theorem]{Criterion}
27: \newtheorem{definition}[theorem]{Definition}
28: \newtheorem{example}[theorem]{Example}
29: \newtheorem{exercise}[theorem]{Exercise}
30: \newtheorem{lemma}[theorem]{Lemma}
31: \newtheorem{notation}[theorem]{Notation}
32: \newtheorem{problem}[theorem]{Problem}
33: \newtheorem{proposition}[theorem]{Proposition}
34: \newtheorem{remark}[theorem]{Remark}
35: \newtheorem{solution}[theorem]{Solution}
36: \newtheorem{summary}[theorem]{Summary}
37: \newenvironment{proof}[1][Proof]{\noindent\textbf{#1.} }{\ \rule{0.5em}{0.5em}}
38: %BeginMSIPreambleData
39: \ifx\pdfoutput\relax\let\pdfoutput=\undefined\fi
40: \newcount\msipdfoutput
41: \ifx\pdfoutput\undefined\else
42: \ifcase\pdfoutput\else
43: \msipdfoutput=1
44: \ifx\paperwidth\undefined\else
45: \ifdim\paperheight=0pt\relax\else\pdfpageheight\paperheight\fi
46: \ifdim\paperwidth=0pt\relax\else\pdfpagewidth\paperwidth\fi
47: \fi\fi\fi
48: %EndMSIPreambleData
49: \begin{document}
50: \title{Stability of Phase-modulated Quantum Key Distribution System}
51: \author{Zheng-Fu Han}
52: \email{zfhan@ustc.edu.cn}
53: \author{Xiao-Fan Mo}
54: \author{You-Zhen Gui}
55: \author{Guang-Can Guo}
56: \email{gcguo@ustc.edu.cn}
57: \affiliation{Key Laboratory of Quantum Information, University of Science and Technology of
58: China (CAS), Hefei, Anhui 230026, People's Republic of China.}
59:
60: \begin{abstract}
61: Phase drift and random fluctuation of interference visibility in double
62: unbalanced M-Z QKD system are observed and distinguished. It has been found
63: that the interference visibilities are influenced deeply by the disturbance of
64: transmission fiber. Theory analysis shows that the fluctuation is derived from
65: the envioronmental disturbance on polarization characteristic of fiber,
66: especially including transmission fiber. Finally, stability conditions of
67: one-way anti-disturbed M-Z QKD system are given out, which provides a
68: theoretical guide in pragmatic anti-disturbed QKD.
69:
70: \end{abstract}
71:
72: \pacs{03.67.Pp, 03.67.Dd, 42.65.Lm}
73: \maketitle
74:
75:
76: Quantum cryptography (QC) or quantum cryptographic key distribution (QKD) has
77: advanced for twenty years since Bennett and Brassard proposed the idea in
78: $1984$ \cite{Bennett1}. It has approached practicability to date, and is
79: applied in the field of information security \cite{Stucki,www}. Many protocols
80: for QC are proposed up to the present, including BB84 \cite{Bennett1}, B92
81: \cite{Bennett2} and improved protocols \cite{Bruss,Bourennane} based on single
82: photons, and EPR protocols \cite{Ekert} based on entangled photon pairs. It
83: has been proved that BB84 protocol and EPR protocol are consistent essence
84: \cite{Bennett4}. Considering current experimental technique, single-photon
85: protocols are more practical for commercial QKD.
86:
87: A general challenge for QKD is to choose a coding scheme free of disturbance
88: by the dominant noise sources. In the case of single-photon protocols, quantum
89: information is encoded by the states of single photons that travel from Alice
90: to Bob, and two kinds of schemes are utilized: polarization coding and phase
91: coding. Polarization coding is adopted in free-space QKD because atmospheric
92: density fluctuations only reduce the collection efficiency but don't degrade
93: the contrast between the polarization basis states. C. Kurtsiefer \textit{et
94: al.} have realized the secure exchange of keys over a free space path of
95: $23.4$ kilometres between Zugspitze and karwendespitze of the Alps
96: \cite{Kurtsiefer}. This marks a pioneering step towards accomplishing key
97: exchange with a near-Earth orbiting satellite and hence a global
98: key-distribution system \cite{Rarity}. On the other hand, optical fibres are
99: appropriate for communication on the ground without influence of weather and
100: atmospheric pollution. However, it is a calamity to use polarization as basis
101: for encoding quantum information due to significant birefringence. Hence,
102: phase coding is widely used in fibre-optic QKD-prototype systems.
103:
104: The most typical fiber QKD prototype was designed in Ref. \cite{Bennett3}. In
105: that prototype, two photon pulses pass through the same transmission fiber
106: (quantum channel), and the same disturbance, from transmission processes, is
107: expected to be counteracted on Bob's side. But in fact, this prototype gives
108: us a bad systemic stability. In order to improve the stability, two groups
109: \cite{Muller,Bethune}\ developed a new prototype independently, in which the
110: two pulses transmit a round trip with a Faraday reflection in mid-course.
111: Unfortunately, this prototype leaves a chance to Eavesdroppers in the key
112: exchange. Eve can send Trojan-horse photons to tail signal photons thus pass
113: in and out of Bob's secure office, collecting Bob's coded information without
114: being discovered \cite{Boileau}. Furthermore, any go-and-return QKD protocol
115: \cite{Nishioka} is essentially unsecured, and true secure QKD should be based
116: on photons propagating along one way. Therefore, for\ practical use of QKD, it
117: is very important to investigate the stability of double M-Z interferometers.
118: This paper investigates the stability in experiments and establishes a
119: theoretical model. Two kinds of perturbation motion, i.e., phase drift and
120: random fluctuation of interference visibility, are observed and discussed.
121: Formerly, it was ever believed that the dominant disturbance arises from
122: different environments of interferometers on Alice and Bob's sides, and the
123: disturbance in quantum channel is equivalent to both pulses (with $%
124: %TCIMACRO{\unit{ns}}%
125: %BeginExpansion
126: \operatorname{ns}%
127: %EndExpansion
128: $-scale time interval) which can be counteracted in the final interference.
129: But our experiments show that interference visibility depends intensively on
130: length of transmission fiber ($L$) and systemic stability is also influenced
131: by the way disturbance. Theoretical analysis, here, deduces the conditions for
132: systemic stabilization of double unbalanced M-Z interferometers. It will offer
133: important reference for stable one-way QKD systems.
134:
135: Typical QKD prototype is described in Fig. 1. Two uniform unbalanced M-Z
136: interferometers were built with common single mode fiber (SMF-28), which is
137: respectively the coder of Alice or Bob. Quantum channel between Alice and Bob
138: and beamsplitters ($50/50$) are also made of the same fiber. Experimental
139: result is depicted in Fig. 2. Fig. 2a shows the received optical power of
140: single detector D1 for $L=2%
141: %TCIMACRO{\unit{m}}%
142: %BeginExpansion
143: \operatorname{m}%
144: %EndExpansion
145: $ without any phase modulation. Apparently, the fluctuation is derived from
146: phase drift, which is relevant to the environment of interferometers. The main
147: reason is that environmental temperature has a small fluctuation, and is also
148: influenced by other perturbation motion, for instance, fiber vibration. Fig.
149: 2b describes the interference fringe of the QKD system when $L=75$ $%
150: %TCIMACRO{\unit{km}}%
151: %BeginExpansion
152: \operatorname{km}%
153: %EndExpansion
154: $, and here the phase modulator is driven by long-periods saw-tooth wave. The
155: fluctuation of envelope of interference fringe exceeds $50\%$ and shows full
156: randomness within six hours. The random fluctuation should be closely
157: associated with the environment of the system.
158:
159: Summarizing the results of Fig. 2, systemic instability resulted from
160: environmental fluctuation behave as phase drift and random fluctuation of
161: interference visibilities. The former leads to work-point destabilization for
162: QKD\ system and can be corrected by instantaneous calibration. But the latter
163: results in qubit-error-rate rising, which can not be artificially controlled
164: due to randomness of environmental fluctuation. In principle, rigorous
165: anti-disturbance methods can be used in Alice and Bob's secure offices, so the
166: effect of the disturbance can decrease to as weaker as possible. However,
167: coupling between quantum channel and its environmental disturbance is
168: uncontrolled. Therefore, disturbance from environment of transmission fiber is
169: discussed as follows.
170:
171: Fig. 3 shows practical interference visibilities for different transmission
172: fiber length $L$. For $L=0$, the fluctuation of the visibility does not exceed
173: $5\%$ within $280$ hours. For $L=25,50,75$ $%
174: %TCIMACRO{\unit{km}}%
175: %BeginExpansion
176: \operatorname{km}%
177: %EndExpansion
178: $, this fluctuation is more than $80\%$. It is most interesting, that the
179: longer the transmission fiber is, the faster\ the visibilities fluctuate.
180: These results indicate that visibilities are affirmatively influenced by way
181: disturbance. Additional consideration is that, temperature fluctuation only
182: leads to phase drift but not interference visibilities fluctuation. Therefore,
183: fluctuation of interference visibilities should arise from polarization
184: birefringence in fiber induced by perturbation motions, such as fiber bend and
185: distortion. In order to explore the relationships between interference
186: visibility and polarization characteristic in M-Z systems, a simple but
187: reasonable theoretical model is built here.
188:
189: For the sake of clarity and concision, we assume that there is no fiber
190: nonlinearity and no polarization-dependent loss, and that the usual loss terms
191: of the fiber have been factored out so that we can deal with unitary Jones
192: matrix \cite{Jones} to describe polarization-transport character of each part
193: in the system (interferometers of Alice and Bob, transmission fiber).
194: $A_{1},A_{2},C,B_{1},B_{2}$ are respectively Jones matrixes to each part of
195: fiber in the QKD apparatus (see Fig. 1). Two paths leading to single-photon
196: interference on $BS_{4}$ are: $P_{1}\left( A_{1}\rightarrow C\rightarrow
197: B_{2}\rightarrow PM_{B}\right) $ and $P_{2}\left( A_{2}\rightarrow
198: PM_{A}\rightarrow C\rightarrow B_{1}\right) $. Here $PM_{A}$ and $PM_{B}$ are
199: two phase modulators. So single-photon optical transformation matrixes through
200: two paths are respectively described by:%
201: \begin{align}
202: B_{2}e^{i\beta_{2}}\cdot e^{i\varphi_{B}}\cdot Ce^{i\phi}\cdot A_{1}%
203: e^{i\alpha_{1}} & =B_{2}CA_{1}e^{i\left( \alpha_{1}+\beta_{2}+\phi
204: +\varphi_{B}\right) },\\
205: B_{1}e^{i\beta_{1}}\cdot Ce^{i\phi}\cdot A_{2}e^{i\alpha_{2}}\cdot
206: e^{i\varphi_{A}} & =B_{1}CA_{2}e^{i\left( \alpha_{2}+\beta_{1}+\phi
207: +\varphi_{A}\right) },
208: \end{align}
209: where $\alpha_{i},\beta_{i}$ are common phases through $i$th fiber of Alice or
210: Bob and $\varphi_{A}\left( \varphi_{B}\right) $ is modulated phase of
211: $PM_{A}\left( PM_{B}\right) $.
212:
213: Supposing input Jones vector of field \cite{Jones} is $E_{in}$, and each
214: effective Jones vector arriving at $BS_{4}$ is $E_{in}/4$. Hence input power
215: $I_{in}=\left\vert E_{in}\right\vert ^{2}$ and the output Jones vector can be
216: written:%
217: \begin{align}
218: E_{out} & =\left[ B_{2}CA_{1}e^{i\left( \alpha_{1}+\beta_{2}+\phi
219: +\varphi_{B}\right) }\right. \nonumber\\
220: & \left. +B_{1}CA_{2}e^{i\left( \alpha_{2}+\beta_{1}+\phi+\varphi
221: _{A}\right) }\right] \frac{E_{in}}{4}.
222: \end{align}
223: On the consideration that $A_{i},B_{i},C$ are unitary, output power can be
224: expressed as:%
225: \begin{align}
226: P_{out} & =E_{out}^{+}\cdot E_{out}\nonumber\\
227: & =\frac{P_{in}}{8}+\frac{1}{16}E_{in}^{+}\left[ A_{1}^{+}C^{+}B_{2}%
228: ^{+}B_{1}CA_{2}e^{-i\left( \Delta\alpha+\Delta\beta+\Delta\varphi\right)
229: }\right. \nonumber\\
230: & \left. +A_{2}^{+}C^{+}B_{1}^{+}B_{2}CA_{1}e^{i\left( \Delta\alpha
231: +\Delta\beta+\Delta\varphi\right) }\right] E_{in},
232: \end{align}
233: where $\Delta\alpha=\alpha_{1}-\alpha_{2},\Delta\beta=\beta_{1}-\beta
234: _{2},\Delta\varphi=\varphi_{B}-\varphi_{A}$. Note that the polarized
235: fluctuation, is random, and consequently the matrixes $A_{i},B_{i},C$ are not
236: independent on time. Obviously, $P_{out}$ is a function of $A_{i}%
237: ,B_{i},C,\Delta\alpha,\Delta\beta,\Delta\varphi$, which means interferential
238: output is dependent not only on both M-Z interferometers but also on
239: transmission fiber. In fact, disturbance from M-Z interferometers and
240: transmission fiber are not independent, which supports our experimental results.
241:
242: In Eq. 4, if we make $B_{1}^{+}B_{2}=I$ on Bob's side, then:%
243: \begin{align}
244: P_{out} & =\frac{P_{in}}{8}+\frac{1}{16}E_{in}^{+}\left[ A_{1}^{+}%
245: A_{2}e^{-i\left( \Delta\alpha+\Delta\beta+\Delta\varphi\right) }\right.
246: \nonumber\\
247: & \left. +A_{2}^{+}A_{1}e^{i\left( \Delta\alpha+\Delta\beta+\Delta
248: \varphi\right) }\right] E_{in},
249: \end{align}
250: where we have considered $C$ is unitary and thus $C^{+}C=I$. Now $P_{out}$ no
251: longer depends on transmission fiber, i.e., $P_{out}$ is independent on\ any
252: polarized disturbance in transmission\ fiber. Furthermore, if $A_{1}^{+}%
253: A_{2}=I$ on Alice's side, Eq. 5 is simplified as%
254: \begin{equation}
255: P_{out}=P_{in}\frac{1+\cos\left( \Delta\alpha+\Delta\beta+\Delta
256: \varphi\right) }{8}.
257: \end{equation}
258: This means that interferential output power $P_{out}$ does not rely on any
259: polarized perturbation motion in whole QKD system. In the ideal case,
260: $\Delta\alpha,\Delta\beta$ are invariable, and hence interference fringes are
261: only modulated by $\Delta\varphi$ brought by both phase modulators. However,
262: in the actual case, the fluctuation of environmental temperature will bring
263: some drift of $\Delta\alpha,\Delta\beta$, corresponding to Fig. 2a. In fact,
264: the fluctuation of environmental temperature is so slow that one can calibrate
265: it by instantaneous compensation.
266:
267: Aforementioned anti-disturbance conditions $A_{1}^{+}A_{2}=I,B_{1}^{+}B_{2}=I$
268: can be equivalently written as:%
269: \begin{equation}
270: A_{1}=A_{2},B_{1}=B_{2},
271: \end{equation}
272: because of unitarity of matrixes $A_{i},B_{i}$. These mean that if we can
273: assure uniform polarization character for two arms of each interferometer,
274: then the QKD\ system resists polarized disturbance not only from transmission
275: fiber (quantum channel) but also from the interferometers. A typical prototype
276: to satisfy upper conditions is that the inner light paths of interferometers
277: are polarization-maintaining. In this case, the Jones matrixes are%
278: \begin{equation}
279: A_{1}=A_{2}=I,B_{1}=B_{2}=I.
280: \end{equation}
281: Obviously, Eq. 8 is a special case to Eq. 7, and it points out that each arm
282: must maintain input photon's polarization. In experiment, free space or
283: polarization-maintaining fibers of interferometers can fulfill Eq. 8.
284:
285: We have observed and distinguished two kinds of unstable phenomena in double
286: unbalanced M-Z QKD system, i.e., phase drift and random fluctuation of
287: interference visibilities. We have found that the interference visibilities
288: can be influenced deeply by the disturbance of transmission fiber. Through
289: theoretic analysis, it is pointed out that the influences are derived from the
290: environmental disturbance on polarization characteristic of fiber, and give
291: out the stability condition of anti-disturbance M-Z QKD system with one-way
292: photons transmission. The conditions contribute to the researches on pragmatic
293: no round trip anti-disturbed QKD system.
294:
295: We thank associate Prof. Bing Zhu, for his helpful advice. This work was
296: funded by National Fundamental Research Program of China (2001CB309301), also
297: by National Natural Science Foundation of china (60121503) and the Innovation
298: Funds of Chinese Academy of Sciences.
299:
300: \begin{thebibliography}{99} %
301:
302:
303: \bibitem {Bennett1}C. H. Bennett and G. Brassard, \textit{Proc. IEEE Int.
304: Conf. Computers, Systems and Signal Processing}, Bangalore, India, 175 (1984).
305:
306: \bibitem {Stucki}D. Stucki, N. Gisin, O. Guinnard, G. Ribordy and H. Zbinden,
307: \textit{New Jour. of Phys.}, \textbf{4}, 41.1 (2002).
308:
309: \bibitem {www}http://www.magiqtech.com.
310:
311: \bibitem {Bennett2}C. H. Bennett, \textit{Phys. Rev. Lett.}, \textbf{68}, 3121 (1992).
312:
313: \bibitem {Bruss}D. Bruss, \textit{Phys. Rev. Lett.}, \textbf{81}, 3018 (1998).
314:
315: \bibitem {Bourennane}M. Bourennane, A. Karlsson, \textit{et al}.,
316: \textit{Phys. Rev. A}, \textbf{64}, 012306 (2001).
317:
318: \bibitem {Ekert}A. K. Ekert, \textit{Phys. Rev. Lett.}, \textbf{67}, 661 (1991).
319:
320: \bibitem {Bennett4}C. H. Bennett, G. Brassard, and A. K. Ekert,
321: \textit{Scientific American}, \textbf{267}, 50 (1992).
322:
323: \bibitem {Kurtsiefer}C. Kurtsiefer, P. Zarda, \textit{et al.}, \textit{Nature}%
324: , \textbf{419}, 450 (2002).
325:
326: \bibitem {Rarity}J. G. Rarity, P. R. Tapster, \textit{et al.}, \textit{New
327: Jour. of Phys.,} \textbf{4}, 82.1 (2002).
328:
329: \bibitem {Bennett3}C. H. Bennett, US Patent 5307410.
330:
331: \bibitem {Muller}A. Muller, T. Herzog, B. Huttner, W. Tittel, H. Zbinden and
332: N. Gisin, \textit{Appl. Phys. Lett.,} \textbf{70}, 793 (1997).
333:
334: \bibitem {Bethune}D. S. Bethune and W. P. Risk, \textit{Appl. Opt. LPEO.},
335: \textbf{41}, 1640 (2002).
336:
337: \bibitem {Boileau}J. C. Boileau, \textit{et al.}, \textit{Phys. Rev. Lett.},
338: \textbf{92}, 017901 (2004).
339:
340: \bibitem {Nishioka}T. Nishioka, H. Ishizuka, \textit{et al.}, \textit{IEEE
341: Photonics Tech. Lett.}, \textbf{14}, 576 (2002).
342:
343: \bibitem {Jones}R. C. Jones, \textit{J. Opt. Soc. Am.} \textbf{31}, 488
344: (1941).\pagebreak
345: \end{thebibliography}
346:
347: \begin{figure}[ptb]
348: \includegraphics[scale=0.5]{fig1.eps}\caption{QKD system based on double
349: unbalanced Mach-Zehnder interferometers.}%
350: \end{figure}
351:
352: \begin{figure}[ptb]
353: \includegraphics[scale=1.0]{fig2a.eps}\end{figure}
354:
355: \begin{figure}[ptb]
356: \includegraphics[scale=1.0]{fig2b.eps}\caption{(a) Output power of single
357: detector D1 without any phase modulation, here $L=2\operatorname{m}$, and
358: shows phase excursion for different time. (b) Interference fringe using the
359: QKD system for different time, here $L=75$ $\operatorname{km}$. Phase
360: modulator is controlled by long-periods saw-tooth wave.}%
361: \end{figure}
362:
363: \begin{figure}[ptb]
364: \includegraphics[scale=1.0]{fig3.eps}\caption{Interference visibilities based
365: on double unbalanced M-Z interferometers at different time. a: $L=0$ ; b:
366: $L=25$ $\operatorname{km}$; c: $L=55$ $\operatorname{km}$; d: $L=75$
367: $\operatorname{km}$.}%
368: \end{figure}
369:
370:
371: \end{document}