1: \documentstyle[12pt,aps]{revtex}%

2: \parskip=.4cm

3: \newcommand{\mc}{\multicolumn}

4: \newcommand{\bce}{\begin{center}}

5: \newcommand{\ece}{\end{center}}

6: \newcommand{\be}{\begin{equation}}

7: \newcommand{\ee}{\end{equation}}

8: \newcommand{\bea}{\vspace{0.25cm}\begin{eqnarray}}

9: %\newcommand{\eea}{\end{eqnarray}\newline\vspace{0.25cm}\noindent}

10: \newcommand{\eea}{\end{eqnarray}}

11: 

12: \def\NCA{{Nuovo Cimento } A }

13: \def\NIM{{Nucl. Instrum. Methods}}

14: \def\NPA{{Nucl. Phys.} A }

15: \def\PLA{{Phys. Lett.}  A }

16: \def\PRL{{Phys. Rev. Lett.} }

17: \def\PRA{{Phys. Rev.} A }

18: \def\PRC{{Phys. Rev.} C }

19: \def\PRD{{Phys. Rev.} D }

20: \def\ZPC{{Z. Phys.} C }

21: \def\ZPA{{Z. Phys.} A }

22: \def\JMO{Journ. of Mod. Opt.}

23: \def\PTP{{Progr. Th. Phys. }}

24: \def\LNC{{Lett. al Nuovo Cimento} }

25: 

26: 

27: %\input{tcilatex}

28: 

29: \begin{document} \draft

30: 

31: \vskip 1cm

32: PACS number: 03.67.Dd,  03.65.Bz, 03.67.Hk, 42.79.-e

33: \vskip 3cm

34: 

35: 

36: 

37: 

38: \begin{center}

39: {\bf {\LARGE Proposal of an experimental scheme for realising a translucent

40: eavesdropping on a quantum cryptographic channel}}

41: \end{center}

42: 

43: \vspace{ .25cm}

44: \begin{center}

45: {M.Genovese \footnote{ \small  genovese@ien.it}} 

46: \\[0pt]

47: 

48: Istituto Elettrotecnico Nazionale Galileo Ferraris \\[0pt]

49: Str. delle Cacce 91 \\[0pt]

50: I-10135 Torino, Italy

51: \end{center}

52: 

53: \vspace{ 3.5 cm} {\large  Abstract }

54: \vskip 0.5 cm

55: Purpose of this paper is to suggest a scheme, which can be realised with

56: today's technology and  could be used for entangling a probe to a photon

57: qubit based on polarisation.  Using this probe a translucent or a coherent

58: eavesdropping can be performed.

59: \vskip 1.7cm

60: 

61: 

62: In the last years quantum mechanics properties of states are assuming a

63: technological relevance.

64: Among different applications of quantum mechanics to technology the

65: possibility of transmitting absolutely confidential messages has been of

66: great interest. This is due to the possibility of creating a key for

67: encoding and decoding  secret messages by transmitting single quanta

68: between two parties (usually dubbed Alice and Bob).

69: The underlying principle of quantum key distribution (QKD) is that nature

70: prohibits gaining information on the state of a quantum system without

71: disturbing it. Thus possible eavesdropping by a third party (usually dubbed

72: Eve) can be identified.

73: Since the original proposal of quantum cryptography \cite{Wiesner}, many

74: different protocols for this kind of transmission have been suggested

75: \cite{cryp}.

76: 

77: For example in the BB84 scheme \cite{BB84} single photons are transmitted

78: from Alice to Bob, preparing them at random in four partly orthogonal

79: polarisation states (at $0^o$ and $90^o$, $45^o$  and $135^o$ for example).

80: Bob selects the bases for measuring the photons polarisation at random

81: too. Then Alice and Bob communicate on a classical channel the bases they

82: have used (but not the results of course): when they have used the same

83: base Bob knows what polarisation was selected by Alice and can build a key.

84: If a spy (Eve) tries to intercept the message, she will inevitably

85: introduce errors, which Alice and Bob can detect by comparing a subsample

86: of the generated key.

87: 

88: In Ekert's  protocol \cite{Ekert} entangled pairs are used. Both Alice and

89: Bob receive one particle of the entangled pair. Then they perform a

90: measurement choosing  among at least three different  directions.   Again,

91: Alice and Bob communicate on a classical channel the bases they have used:

92: if measurements were performed along parallel axes they are used for

93: generating the secret key. The other measurements can be used for a test of

94: Bell inequalities. If Eve tries  to eavesdrop, she inevitably affects the

95: entanglement between the two particles leading to a reduction of the

96: violation of the Bell inequalities, which allows  Alice and Bob to

97: recognise the presence of the spy.

98: 

99: Finally, in the B92 protocol \cite{B92}, Alice sends a state chosen between

100: two non-orthogonal states to Bob, who performs a measurement using a

101: projection on the subspaces orthogonal to  the two states. Then Bob

102: publicly informs Alice when he obtained positive results (but not, of

103: course, which measurement he made). Finally they retain only bits

104: corresponding to these results. 

105: 

106: Other schemes have also been proposed \cite{oth}. Common characteristics of

107: all these schemes are the presence of a quantum channel, where different

108: quantum states can be transmitted, and of a classic channel which is used

109: for selecting a subsample of the transmitted states and for testing the

110: presence of an eavesdropper.   If the classical channel is eliminated,  Eve

111: could simply cut the quantum channel and  substituting herself as Bob for

112: Alice and as Alice for Bob creating in this way two keys with the two other

113: parties.

114: 

115:  

116: Many different experiments have been realised using the former schemes,

117: demonstrating the feasibility of QKD up to a distance of many kilometers

118: \cite{lontano}.      

119: 

120: All of them are based on transmission of single photon states, where the

121: alphabet is based either on photon polarisation or on photon phase.

122: 

123: Concerning the strategy of Eve for eavesdropping, the simplest one is when

124: she simply intercepts the state, which Alice has sent. Then she performs  a

125: measurement on this state and finally she sends a new state to Bob

126: preparing it according to the result of the measurement.

127: 

128: However, more elaborated strategies have been proposed, where Alice does

129: not stop the  transmitted state, but causes it to interact with a second

130: state (an ancilla) and then obtains information about the transmitted qubit

131: thanks to the result of a measurement on the ancilla. This procedure is

132: known as translucent eavesdropping \cite{TEV}. 

133: 

134: Finally, Alice can use eavesdropping schemes where her probe interacts with

135: more than one of Alice's qubits. These schemes are known as coherent or

136: joint attacks \cite{CA}.

137: 

138: Many interesting studies \cite{ET} have been devoted to understanding

139: general conditions for obtaining a secure transmission between Alice and

140: Bob in the presence of an eavesdropper, even using non-ideal channels. 

141: 

142: However, as far as we know, no practically realisable scheme for

143: translucent or coherent eavesdropping has been proposed yet.

144: 

145: The purpose of this paper is to suggest a scheme which can be realised with

146: today's technology and  could be used for entangling a probe to a photon

147: qubit based on polarisation.  Using this probe Eve can then perform a

148: translucent or a coherent eavesdropping (causing the ancilla  interact with

149: more than one transmitted qubit).

150: 

151: The scheme of the eavesdropping apparatus, shown in the figure,  is  a

152: Mach-Zender interferometer with a  Kerr cell on one of the arms inserted on

153: the quantum channel (see the figure).

154: 

155: The input port of the first beam splitter is fed with a single photon of

156: vertical polarisation, which splits on the two interferometer arms. On arm

157: 1 it interacts with the transmitted qubit inside the Kerr cell.

158:   

159: This has no effect except when both the photons interacting in the cell

160: have vertical polarisation ($\vert V \rangle \vert V \rangle \rightarrow

161: \vert V \rangle \vert V \rangle e^{i \phi}$) (in the following $H$ will

162: denote the horizontal polarisation).

163: 

164: Let us suppose that the transmitted qubit is in the general form:

165: \be

166: \vert u \rangle = cos (\theta) \vert H \rangle + sin( \theta ) \vert V \rangle

167: \ee

168: 

169: If we denote the probe photon with $ \vert p \rangle$, the final state is

170: in the entangled form:

171: 

172: \be

173: \vert \Psi \rangle = { 1 \over \sqrt{2}} \left [cos (\theta) \vert H

174: \rangle \vert p \rangle_1 + sin( \theta ) \vert V \rangle \vert p \rangle_1

175: e^{(i \phi)}+

176: i \vert u \rangle \vert p \rangle_2 \right ]

177: \ee

178: where the suffixes after the probe $\vert p \rangle$ denotes the path

179: followed and where we have considered  a 50 \% : 50 \% beam splitter (which

180: allows Eve to obtain the largest information on Alice-Bob transmission).

181: 

182: We have thus obtained the desired entanglement between the probe and the

183: transmitted qubit, which can be used for  translucent or coherent

184: eavesdropping.

185: 

186: 

187: The possibility of realising this scheme, and thus the interest of it,

188: derives by the fact that, although admittedly very difficult,  the Quantum

189: Non Demolition (QND) detection of a single photon is at present possible

190: \cite{SI,SM}. QND measurements of { \it welcher Weg} (which path) have

191: already been achieved 

192: using 100 meter long optical fiber \cite{QNDexp}. Of course, the

193: implementation

194: of the present scheme using such devices would be, even though not impossible

195: in theory, almost impossible in practice. The

196: recent discovery of new materials with very high Kerr coupling, could

197: however allow an easier and more realistic implementation of this scheme.

198: Two candidates as Kerr cell with ultra-high susceptibility to be used for

199: this scheme are the Quantum Coherent Atomic Systems (QCAS)  \cite{QCAS,SI}

200: and the Bose-Einstein condensate of ultracold (at nanoKelvin temperatures)

201: atomic gas \cite{BEC}.

202: These are recent great technical improvements which could permit the

203: realisation of small Kerr cells, capable of large phase shift, even with a

204: single photon probe. In fact, both exhibit extremely high

205: Kerr couplings compared to more traditional materials. In particular, the

206: QCAS is rather a simple system to be realised (for a review see

207: \cite{Arimondo})  and thus represents an ideal candidate in this role.

208: Incidentally, one can notice that Kerr coupling can be further enhanced by

209: enclosing the medium in a cavity \cite{Agarwal}.

210: The scheme that we  propose in this paper could, in principle, be used with

211: relatively small phase shifts too. However, the maximal efficiency is

212: reached when a phase shift of $\pi$ is produced on the probe by a single

213: photon. Recently a Lukin and Imamoglu's paper has shown that this result

214: can be effectively reached \cite{LU}. Experiments addressed to single

215: photon QND, using a Kerr cell, are in progress \cite{G}.

216: 

217: This recent development of high coupling Kerr cell has already been applied

218: to the proposal of schemes for complete teleportation \cite{TVF},

219: for generating Schr\"odinger cats and modulating quantum interference

220: \cite{nos1,nos2}, for generating \cite{nos2} GHZ states \cite{GHZ} and for

221: realising quantum gates \cite{TVF,Chiara}. 

222: 

223: Incidentally, the use of a Bose condensate could also allow Eve to "stock"

224: her photon till when the other has reached Bob, using the very low

225: propagation velocity of light inside a suited Bose condensate (as low as

226: few tens of meters per second) \cite{BEC}. 

227: 

228: For the sake of exemplification in the following we consider the

229: application of this scheme to a simple procedure of eavesdropping on a

230: quantum channel where the BB84 protocol is used. 

231: 

232: Let us begin with the simple example where Bob performs measurement on the

233: base H,V or in the one at $45^o$ degrees.

234: 

235: If Alice sends $H$ ($\theta = 0$), Bob measures $H$ and Eve has the state 

236: \be

237: {\vert p \rangle _1 + i \vert p \rangle _2 \over \sqrt{2} }

238: \ee

239: and thus after the second beam splitter the photon will be detected by the

240: photodector D3. On the other hand, if Alice send $V$, Bob measures $V$ and,

241: for Eve, only photodetector D4 clicks.

242: In this case no transmission error is inserted on the quantum line by Eve

243: presence, while she obtains a perfect identification of the transmitted qubit.

244:  

245: On the other hand, if Alice sends a photon $ \vert + \rangle =  {\vert H

246: \rangle  +  \vert V \rangle  \over \sqrt{2} }$, after the Kerr cell the

247: entangled state is:

248: \be

249: \vert \Psi \rangle = { 1 \over 2 \sqrt{2}} \left [ \vert + \rangle ( [1 +

250: e^{i \phi}] \vert p \rangle_1 + 2 i \vert p \rangle_2 )+

251:  \vert - \rangle [1 - e^{i \phi}]  \vert p \rangle_1 \right ]

252: \ee

253: where

254: $ \vert - \rangle =  {\vert H \rangle  -  \vert V \rangle  \over \sqrt{2} }$.

255: 

256: If $\phi=\pi$, Eve has a 50 \% probability of observing the photon at D3 and 

257: a 50 \% probability of observing the photon at D4.

258: Also, a 50 \% error on Bob measurement is introduced.

259: 

260: The same situation happens when Alice sends a $\vert - \rangle $ photon.

261: 

262: Altogether, the probability of a successful eavesdropping for Eve is

263: $p=3/4$, which leads to a capacity of the channel (Alice-Eve) \cite{dover}

264: \be

265: I_{AE} = 1 + p log_2 p + (1-p) log_2 (1-p) = 0.189

266: \ee

267: As the error rate for Bob is $1/4$, the capacity of the channel Alice-Bob

268: is $I_{AB} = 0.189$ as well.

269: 

270: Of course, a careful spy will not use such a procedure which produces

271: asymmetric  errors for the two bases, which can be easily identified.

272: 

273: In general, if Alice sends a generic state $\vert u \rangle$ the

274: probability that Bob measures $\vert u \rangle$ as well and Eve sees a

275: photon at D3 is

276: \be

277: P^3_{uu} = 1 / 2 [ 1- cos^2 (\theta ) sin^2 (\theta) (1 -cos (\phi)) +

278: cos^2 (\theta) + cos( \phi) sin^2 (\theta)]

279: \label{P4uu}

280: \ee

281: Similarly, in the same situation Eve could observe a photon at D4 with

282: probability:

283: \be

284: P^4_{uu} = 1 / 2 sin^4 (\theta) (1- cos (\phi))

285: \ee

286: With similar notations, if Alice sends the orthogonal state :

287: \be

288: \vert v \rangle = cos (\theta) \vert V \rangle - sin( \theta ) \vert H \rangle

289: \ee

290: one has

291: \be

292: P^3_{vv} = 1 / 2 [ 1- cos^2 (\theta ) sin^2 (\theta) (1 -cos (\phi)) +

293: sin^2 (\theta) + cos( \phi) cos^2 (\theta)]

294: \ee

295: and

296: \be

297: P^4_{vv} = 1 / 2 cos^4(\theta) (1 - cos (\phi))

298: \ee

299: Finally, the probabilities corresponding to introducing an error on the

300: Alice-Bob communication are:

301: \be

302: P^3_{uv}= P^4_{uv}= P^3_{vu}= P^4_{vu}= sin^2 (\theta) cos^2 (\theta)

303: sin^2( \phi /2)

304: \label{P3uv}

305: \ee

306: 

307: From Eq.s \ref{P4uu}-\ref{P3uv} it follows that if Eve uses the base

308: bisecting the two used by Alice and Bob and intercepts a fraction $\alpha$

309: of the transmitted qubits (in order to reduce the errors introduced in Bob

310: measurements), higher information is obtained. In fact, being the error

311: rate on Alice-Eve channel

312: \be

313: q_{AE}=  ( P_3^{vu} + P_3^{vv} + P_4^{uu} + P_4 ^{uv}) / 2 

314: \ee

315: one obtains $ I_{AE}=  0.40 \alpha $ (when $ \phi = \pi $). Furthermore, as

316: this procedure generates symmetric errors for the two bases, it is much

317: more difficult for Bob to distinguish if these derive from noise in the

318: quantum channel or by Eve's presence. 

319: 

320: Let us now analyse the security threshold. The error rate in the Alice-Bob

321: channel is unsafe if the mutual information in the Alice-Bob channel does

322: not exceed the minimum of the mutual information in the Alice-Eve and

323: Eve-Bob channels \cite{Eck2}: 

324: \be I_{AB} \leq \min ( I_{AE} , I_{EB})

325: \ee

326: 

327: This means that whenever $I_{AB}$ is greater than either $I_{AE}$ or

328: $I_{EB}$, then at least in principle there is a way for Alice and Bob to

329: distribute a string of secret information. On the other hand, if $I_{AB}$

330: is smaller than $I_{AE}$ or $I_{EB}$, no sifting procedure can make the

331: transmission safe.

332: 

333: In \cite{Eck2} it was suggested that this condition may be overly cautious,

334: however in  \cite{Eck3} it has been shown that this is not the case for the

335: entangled translucent eavesdropping scenario.

336: 

337: In our case the information on the Eve-Bob and Alice-Eve channels are $I_{EB}=

338: I_{AE}= 0.4 \alpha  $, whilst the error introduced in the Alice-Bob channel

339: becomes $q_{AB} = \alpha ( P_3^{vu} + P_3^{uv} + P_4^{uv} + P_4 ^{vu}) / 2

340: = \alpha / 4 $.

341: 

342: Thus, the transmission cannot be considered safe if Eve intercepts a

343: fraction $\alpha=0.755$, or larger, of the transmitted photons.

344: This corresponds to an error rate on the Alice-Bob channel $ q_{AB} = 0.189$.

345: Of course, this error is relatively large, but one should not forget that

346: nowadays quantum channels, both in fibers or air, have huge losses. 

347: Furthermore, with other protocols this value could be smaller.

348: 

349: If Ekert or BB92 protocols are used, a similar analysis can be carried out.

350: Many papers with general results about eavesdropping and security of

351: quantum channels are already published \cite{ET}. We refer to them for

352: general discussion about application of translucent or joint eavesdropping

353: and  conditions for obtaining a safe communication in the presence of

354: eavesdropping.

355: 

356: In conclusion, we think that the proposed scheme represents an interesting

357: chance for an experimental realisation of eavesdropping on a quantum

358: channel and, as far as we know, this is the first proposal for a practical

359: realisation of translucent eavesdropping. Of course, this scheme could also

360: be used in any other situation where interaction with a qubit (represented

361: by polarisation properties of a photon) and an ancilla qubit is required.

362:  

363: \vskip 1cm

364: {\bf Acknowledgements}

365: 

366: \noindent We would like to acknowledge the support of MURST via special

367: programs "giovani ricercatori" Dip. Fisica Teorica Univ. Torino and of

368: Istituto Nazionale di Fisica Nucleare. Thanks are due to C. Novero for

369: useful discussions.

370: 

371: 

372: { \noindent {\bf References}

373: \begin{enumerate}

374: 

375: \bibitem{Wiesner} S. Wiesner, Sigact News 15 (1983) 78.

376: 

377: \bibitem{cryp} see for example S.J. Lomonaco, quant-ph 9811056 and ref.s

378: therein. See also "Special issue of Quantum Communication", Journ. of Mod.

379: Opt. 41, 12 (1994).

380: 

381: \bibitem{BB84} C.H. Bennet and G. Brassard, Proc. of  Int. Conf. Computer

382: Systems an Signal Processing, Bangalore (1984) 175.  

383: 

384: \bibitem{Ekert} A.K. Ekert, \PRL 67 (1991) 661.

385: 

386: \bibitem{B92} C. H. Bennet, \PRL 68 (1992) 3121.

387: 

388: \bibitem{oth} see for example: M. Czachor, quant-ph 9812030; A. K. Ekert et

389: al., \PRL 69 (1992) 1293; K. Shimizu and N. Imoto, \PRA 60 (1999) 157; W.

390: Tittel, H. Zbinden and N. Gisin, quant-ph 9912035.

391:  

392: \bibitem{lontano} W.T. Buttler et al., quant-ph 0001088; W. Tittel et al.,

393: quant-ph 9911109; H. Zbinden Appl. Phys. B 67 (1998) 743, W.T. Buttler et

394: al, Phys.Rev.Lett. 81 (1998) 3283; A.V. Sergienko et al., \PRA 60 (1999)

395: R2622 and ref.s therein.

396: 

397: \bibitem{TEV} see C.a. Fuchs and A. Peres, \PRA 53 (1996) 2038;  A.K.Ekert

398: et al., \PRA 50 (1994) 1047; H.E. Brandt et al., \PRA 56 (1997) 4456 and

399: ref.s therein. 

400: 

401: \bibitem{CA} see C. Bennet et al., quant-ph 9611006; J.I. Cirac and N.

402: Gisin \PLA 229 (1997) 1 and ref.s therein.

403: 

404: \bibitem{ET} S.M. Barnett et al., \JMO 40 (1993) 2501; S.M. Barnett and

405: S.J.D. Phoenix, \PRA 48 (1993) R5, \JMO 40 (1993) 1443; D. Deutsch et al.,

406: \PRL 77 (1996) 2818; B. Huttner and A.K. Ekert, \JMO 41 (1994) 2455; 

407: K.J. Blow and S.J.D. Phoenix, \JMO 40 81993) 33; G. Brassard et al,

408: quant-ph 9906074, quant-ph9911054; E. Biham and T. Mor, \PRL 78 (1997)

409: 2256, \PRL 79 (1997) 4034. 

410: 

411: \bibitem{QNDexp} M.D. Levenson et al., \PRL 57 (1986) 2473; N. Imoto et al.,

412: Opt. Comm. 61 (1987) 159. 

413: 

414: \bibitem{SI} H. Schmidt and A. Imamoglu, Opt. Lett. 21 (1996) 1936.

415: 

416: \bibitem{SM} B.C. Sanders and G.J. Milburn, \PRA 39 (1989) 694.

417: 

418: \bibitem{QCAS} U. Rathe et al., \PRA 47 (1993) 4994.

419: 

420: \bibitem{BEC} L. Vestergaard Hau et al., Nature 397 (1999) 594.

421: 

422: \bibitem{Arimondo} E. Arimondo, in Progress in optics XXXV, E. Wolf

423: editor, Elsevier Science 1996, pag. 257.

424: 

425: 

426: \bibitem{Agarwal} G.S. Agarwal, Opt. Comm. 72 (1989) 253.

427: 

428: \bibitem{LU} M.D. Lukin and Imamoglu, \PRL 84 (2000) 1419.

429: 

430: \bibitem{G} G. Brida et al, private communication.

431: 

432: \bibitem{TVF} D. Vitali et al., quant-ph 0003082.

433: 

434: \bibitem{nos1} M. Genovese and C. Novero, Phys. Rev. A 61 032102 (2000).

435:  

436: \bibitem{nos2} M. Genovese and C. Novero,  \PLA to appear.

437: 

438: \bibitem{GHZ} D.M. Greenberger et al., \PRA 82 (1999) 1345 and ref.s therein.

439: 

440: \bibitem{Chiara} G. M. D'Ariano, C. Macchiavello, and L. Maccone,

441: Fortschr. Phys. {\bf 48}, 573

442: (2000).

443:  

444: \bibitem{dover} R.B. Ash, Information Theory, (Dover, New York, USA 1990).

445: 

446: \bibitem{Eck2} A. K. Eckert et al, \PRA 50 (1994) 1047.

447: 

448: \bibitem{Eck3} H.E. Brandt et al, \PRA 56 (1997) 4456.

449: 

450: 

451: \end{enumerate} 

452: }

453: \vfill \eject

454: 

455: \newpage

456: {\bf Figure Caption}

457: 

458: The Alice - Bob cryptographic channel, with the eavesdropping apparatus of

459: Eve, which is constituted of a Mach-Zender interferometer and  a Kerr cell.

460:  A probe photon can follow the arm of the interferometer where the Kerr

461: cell is posed or the other. If the transmitted photon has a vertical

462: polarisation the probe photon phase is changed. The observation of the

463: probe photon at photodetector D3 or D4 gives Eve information on the

464: transmitted photon polarisation.

465: 

466: \end{document}

467:  

468: