1: \documentstyle[preprint,aps,pra,epsfig,cite]{revtex}
2: %\documentstyle[epsf]{article}
3: \begin{document}
4:
5: \draft
6: \title{Entangled Nonorthogonal States and
7: Their Decoherence Properties
8: }
9: \author{Osamu Hirota$^{1,2}$, Steven J. van Enk$^+$,
10: Kazuo Nakamura$^{\ast1,2}$,\\ Masaki Sohma$^\dagger$,
11: and Kentaro Kato$^{1,2}$\\
12: 1.Tamagawa University, Machida, Tokyo, Japan.\\
13: 2. CREST, JST$:$Japan Science and Technology.\\
14: $^+$Bell Labs, Lucent Technologies, Murray Hill, NJ, USA\\
15: $^\ast 1$ Fundamental Research Laboratory, NEC, Tsukuba, Japan.\\
16: $^\ast 2$ CREST, JST$:$Japan Science and Technology.\\
17: $^\dagger$ Matsushita Research Institute Inc. Tokyo, Japan.
18: }
19:
20: \maketitle
21:
22: \begin{abstract}
23: This paper presents properties of the so-called quasi-Bell states:
24: entangled states written as superpositions of nonorthogonal states.
25: It is shown that a special class of those states, namely entangled
26: coherent states, are more robust against decoherence due to
27: photon absorption than the standard bi-photon Bell states.
28: \end{abstract}
29:
30: \pacs{PACS numbers: 03.65.Bz}
31:
32: \section{Introduction}
33: Entanglement and its information-theoretic aspects have been
34: studied by many authors
35: \cite{Bennett,Bennett2,vedral,Wootters,Wootters2}.
36: Here we give a short survey of the theory of entanglement that
37: we will later apply to quasi-Bell states.
38: For a pure entangled state of a bipartite system
39: $|\Psi\rangle_{AB}$, the measure of entanglement defined as
40: \cite{Bennett,Barnett}
41: \begin{equation}
42: E(|\Psi\rangle_{AB})=-{\rm Tr}_{A}{\rho}_{A}\log {\rho}_{A},
43: \quad
44: {\rho}_{A}={\rm Tr}_{B}|\Psi\rangle_{AB}\langle\Psi|,
45: \end{equation}
46: is called the ``entropy of entanglement''.
47: This quantity enjoys two kinds of information-theoretic
48: interpretations.
49: One is that $E$ gives the entanglement of formation, which is defined as
50: the asymptotic number $k$ of standard singlet states required to
51: faithfully locally prepare $n$ identical copies of
52: a system in the bipartite state $|\Psi\rangle_{AB}$ for
53: very large $k$ and $n$.
54: The other is that $E$ gives the amount of distillable entanglement,
55: which is the asymptotic number of singlets $k$
56: that can be distilled from $n$ identical copies of
57: $|\Psi\rangle_{AB}$.
58: With either of these definitions of $k$ and $n$, $E$ satisfies
59: \begin{equation}
60: \lim _{n, k \rightarrow \infty} \frac{k}{n} = E(|\Psi\rangle_{AB}).
61: \end{equation}
62: For pure states we can rewrite
63: \begin{equation}
64: E(|\Psi\rangle_{AB})=
65: H\left(\frac{1}{2}(1+\sqrt {1 - C(|\Psi\rangle_{AB})^2})\right)
66: \end{equation}
67: where $H(x) = -x \log x -(1-x)\log (1-x)$ is the binary entropy function
68: and
69: $C(|\Psi\rangle_{AB})$ is the ``concurrence'' defined by
70: $C(|\Psi\rangle_{AB})=
71: |_{AB}\langle\Psi|\tilde\Psi\rangle_{AB}|$ with
72: $|\tilde\Psi\rangle_{AB}=\sigma|\Psi\rangle_{AB}^*$. The above expression
73: is valid for mixed states of two qubit systems as well \cite{Wootters}.
74:
75: For mixed states of qubits one may also define
76: an expression for the entanglement of formation
77: \cite{Bennett,Wootters,Wootters2}.
78: It is defined as the average entanglement of
79: the pure states of a decomposition
80: $\rho=\sum_ip_i|\psi\rangle_i\langle\psi_i|$
81: of the density operator $\rho$, minimized over
82: all decompositions \cite{Bennett}
83: \begin{equation}
84: E(\rho)= \min \sum p_i E(|\psi_i \rangle).
85: \end{equation}
86: In general, it is difficult to find the exact amount
87: of entanglement of formation except for special cases.
88: However, there is a lower bound which is expressed
89: in terms of a quantity called the ``fully entangled fraction'',
90: which we denote by $f(\rho)$ and is defined as
91: \begin{equation}
92: f(\rho)=\max \langle e |\rho | e \rangle,
93: \end{equation}
94: where the maximum is over all completely entangled states
95: $ | e \rangle $.
96: A lower bound on the entanglement of formation is \cite{Bennett}
97: \begin{equation}
98: E(\rho) \ge h[f(\rho)],
99: \end{equation}
100: where
101: \begin{equation}
102: h[f(\rho)] = \left\{
103: \begin{array}{ll}
104: H[ \frac{1}{2} + \sqrt{f(1-f)}] & (f \ge \frac{1}{2})\\
105: 0 & (f < \frac{1}{2})\\
106: \end{array}
107: \right.
108: \end{equation}
109: Usually in order to construct entangled states one writes
110: superpositions of orthogonal states.
111: For instance the standard Bell basis uses states like
112: $|\updownarrow\rangle$ and $|\leftrightarrow\rangle$,
113: and of course its properties are well known.
114: Our concern here is what kind of properties appear if we
115: have superpositions of nonorthogonal states.
116: In this paper, we will clarify properties of entangled states
117: of nonorthogonal states such as coherent states based on the above
118: basic theory.
119:
120: \section{Quasi-Bell states}
121: \subsection{General definition}
122: Let us consider entangled states based on two nonorthogonal states
123: $|\psi_1 \rangle$ and $|\psi_2 \rangle$ such that
124: $ \langle \psi_1 | \psi_2 \rangle = \kappa$ where $\kappa$ is real.
125: We can define a set of 4 entangled states as follows:
126: \begin{eqnarray}
127: \left\{
128: \begin{array}{lcl}
129: |\Psi_1 \rangle_{AB} &=& h_{1} (|\psi_1 \rangle_A|\psi_2 \rangle_B
130: +|\psi_2 \rangle_A|\psi_1 \rangle_B ) \\
131: |\Psi_2 \rangle_{AB} &=& h_{2} (|\psi_1 \rangle_A|\psi_2 \rangle_B
132: -|\psi_2 \rangle_A|\psi_1 \rangle_B )\\
133: |\Psi_3 \rangle_{AB} &=& h_{3} (|\psi_1 \rangle_A|\psi_1 \rangle_B
134: + |\psi_2 \rangle_A|\psi_2 \rangle_B )\\
135: |\Psi_4 \rangle_{AB} &=& h_{4} (|\psi_1 \rangle_A|\psi_1 \rangle_B
136: -|\psi_2 \rangle_A|\psi_2 \rangle_B)
137: \end{array}
138: \right.
139: \end{eqnarray}
140: where
141: $\{h_{i}\}$ are normalization constants:
142: $h_{1}=h_{3}=1/\sqrt{2(1+\kappa^{2})}$,
143: $h_{2}=h_{4}=1/\sqrt{2(1-\kappa^{2})}$.
144: We call these states ``quasi-Bell states''.
145: They are not orthogonal to each other.
146: In fact for $\kappa$ real their Gram matrix
147: $G_{ij}=|_{AB}\!\langle\Psi_i|\Psi_j\rangle_{AB}|$ becomes
148: \begin{equation}
149: G=
150: \left(
151: \begin{array}{cccc}
152: 1& 0& D& 0\\
153: 0& 1& 0& 0\\
154: D& 0& 1& 0\\
155: 0& 0& 0& 1\\
156: \end{array}
157: \right)
158: \label{ohmsgrammat}
159: \end{equation}
160: where $D=2 \kappa/(1 + {\kappa}^2)$.
161: If the basic states are orthogonal ($\kappa=0$), then these states
162: reduce to standard Bell states.
163: Let us discuss the entropy of entanglement for the above states.
164: We first calculate the reduced density operators of
165: the quasi-Bell states.
166: They are ${{\rho}_{A}}^{(1)}={{\rho}_{A}}^{(3)}$ and
167: ${{\rho}_{A}}^{(2)} = {{\rho}_{A}}^{(4)} $ with
168: \begin{equation}
169: {{\rho}_{A}}^{(1)}
170: = \frac{1}{2(1+\kappa^2)}
171: \{|\psi_1 \rangle\langle\psi_1| +
172: \kappa |\psi_1 \rangle\langle \psi_2|
173: +\kappa |\psi_2 \rangle\langle \psi_1|
174: + |\psi_2 \rangle\langle \psi_2\},
175: \end{equation}
176: \begin{equation}
177: {{\rho}_{A}}^{(2)}=\frac{1}{2(1-\kappa^2)}
178: \{|\psi_1 \rangle_A\langle \psi_1|
179: -\kappa |\psi_1 \rangle_A\langle \psi_2|
180: -\kappa |\psi_2 \rangle_A\langle \psi_1|
181: + |\psi_2 \rangle_A\langle \psi_2|\}.
182: \end{equation}
183: The eigenvalues of the above density operators
184: ${{\rho}_{A}}^{(1)}$(or ${{\rho}_{A}}^{(3)}$) are given
185: in terms of the Gram matrix elements $G_{ij}$ as follows,
186: \begin{equation}
187: \lambda_{1/1} = \frac{1+G_{13}}{2}=
188: \frac{(1+\kappa)^2}{2(1+\kappa^2)},
189: \quad
190: \lambda_{2/1} = \frac{1-G_{13}}{2}
191: =\frac{(1-\kappa)^2}{2(1+\kappa^2)},
192: \end{equation}
193: and for ${{\rho}_{A}}^{(2)}$(or ${{\rho}_{A}}^{(4)}$)
194: we have
195: \begin{equation}
196: \lambda_{1/2}= \frac{1+G_{24}}{2}=\frac{1}{2},
197: \quad
198: \lambda_{2/2}= \frac{1-G_{24}}{2}=\frac{1}{2}.
199: \end{equation}
200: Hence, the entropy of entanglement is
201: \begin{equation}
202: E(|\Psi_1\rangle_{AB})=E(|\Psi_3\rangle_{AB})
203: = - \frac{1+D}{2} \log \frac{1+D}{2}
204: - \frac{1-D}{2} \log \frac{1-D}{2},
205: \end{equation}
206: and
207: \begin{equation}
208: E(|\Psi_2\rangle_{AB})=E(|\Psi_4\rangle_{AB})=1,
209: \end{equation}
210: because $G_{13}=D$, and $G_{24}=0$.
211: Thus $|\Psi_2\rangle_{AB}$ and $|\Psi_4\rangle_{AB}$
212: are maximally entangled, even though the entangled states
213: consist of nonorthogonal states in each subsystem.
214: These results are true for arbitrary nonorthogonal states
215: with $\langle \psi_1 | \psi_2 \rangle$ $=$
216: $\langle \psi_2 | \psi_1 \rangle$ $=$ $\kappa$ and
217: do not depend on the physical dimension of
218: the systems. This property may be unexpected but can be
219: understood easily by noting that the states $|\Psi_2\rangle$
220: and $|\Psi_4\rangle$ are equivalent to
221: $(|+\rangle|-\rangle \pm |-\rangle|+\rangle)/\sqrt{2}$
222: in terms of the orthogonal basis
223: \begin{equation}
224: |\pm\rangle=(|\psi_1\rangle \pm |\psi_2\rangle)/\sqrt{N_{\pm}},
225: \end{equation}
226: with $N_{\pm}=2\pm 2\kappa$.
227:
228: \subsection{Mixtures of quasi Bell states}
229: We can construct a quasi-Werner mixed state based on
230: quasi-Bell states by
231: \begin{equation}
232: W = F|\Psi_2 \rangle_{AB} \langle \Psi_2 | +
233: \frac{1-F}{3}\{|\Psi_1 \rangle_{AB} \langle \Psi_1|
234: +|\Psi_3 \rangle_{AB} \langle \Psi_3 | +
235: |\Psi_4 \rangle_{AB} \langle \Psi_4 |\},
236: \end{equation}
237: where $0 \le F \le 1$.
238: If $|\Psi_1 \rangle_{AB}$,$|\Psi_2 \rangle_{AB}$,
239: $|\Psi_3 \rangle_{AB} $,
240: $|\Psi_4 \rangle_{AB}$ are Bell states, then
241: the above equation gives a standard Werner state \cite{werner}.
242: It is known that the fully entangled fraction of
243: the Werner state is $F$, and the entanglement of
244: formation of the Werner state is given by
245: \begin{equation}
246: E(W) = H\left(\frac{1}{2} + \sqrt{F[1-F]}\right).
247: \end{equation}
248: The fully entangled fraction of
249: the quasi Werner state is analogously given by
250: \begin{equation}
251: f(W) = _{AB}\!\langle \Psi_2 |W |\Psi_2 \rangle_{AB} = F,
252: \end{equation}
253: because the quasi Bell states are orthogonal to each other,
254: except for the pair of states
255: $|\Psi_1 \rangle_{AB}$ and $|\Psi_3 \rangle_{AB} $,
256: as one can see from the Gram matrix $G$.
257: However, the quasi Werner state and Werner state are
258: completely different states.
259: In particular, the eigenvalues of quasi Werner states are
260: different from those of Werner states.
261: The eigenvalues of the density operator are given by those of
262: the modified Gram matrix
263: \cite{hirota0}. For the quasi Werner state,
264: the Gram matrix is
265: \begin{equation}
266: G_W=
267: \left(
268: \begin{array}{cccc}
269: \frac{1}{3}(1-F)& 0& \frac{1}{3}(1-F)D& 0\\
270: 0& F& 0& 0\\
271: \frac{1}{3}(1-F)D& 0& \frac{1}{3}(1-F)& 0\\
272: 0& 0& 0& \frac{1}{3}(1-F)\\
273: \end{array}
274: \right)
275: \end{equation}
276: As a result, we have
277: $F$, $\frac{1}{3}(1-F)$, $\frac{1}{3}(1+D)(1-F)$,
278: $\frac{1}{3}(1-D)(1-F)$ as the eigenvalues of
279: the quasi Werner state.
280:
281: Thus the lower bound of the entropy of formation of
282: quasi Werner state is the same as that of Werner state.
283: For more general mixtures of quasi Bell states
284: we will need a more advanced analysis which is
285: reported on in a subsequent paper.
286:
287: \section{Quasi-Bell states based on bosonic coherent states}
288: Let us consider two coherent states of a bosonic mode
289: $\{|\alpha\rangle, |-\alpha\rangle\}$, e.g., let
290: $\pm \alpha$ be the coherent amplitude of a
291: light field. Using previous notation, we have
292: $\kappa=\langle\alpha|-\alpha\rangle=\exp\{-2|\alpha|^2\}$.
293: Then one can construct the quasi Bell states as follows:
294: \begin{eqnarray}
295: \left\{
296: \begin{array}{lcl}
297: |\Psi_1 \rangle_{AB} &=&
298: h_{1} (|\alpha \rangle_A|-\alpha \rangle_B
299: +|-\alpha \rangle_A|\alpha \rangle_B ) \\
300: |\Psi_2 \rangle_{AB} &=&
301: h_{2} (|\alpha \rangle_A|-\alpha \rangle_B
302: -|-\alpha \rangle_A|\alpha \rangle_B )\\
303: |\Psi_3 \rangle_{AB} &=&
304: h_{3} (|\alpha \rangle_A|\alpha \rangle_B
305: + |-\alpha \rangle_A|-\alpha \rangle_B )\\
306: |\Psi_4 \rangle_{AB} &=&
307: h_{4} (|\alpha \rangle_A|\alpha \rangle_B
308: -|-\alpha \rangle_A|-\alpha \rangle_B)
309: \end{array}
310: \right.
311: \end{eqnarray}
312: Since the coherent states are nonorthogonal,
313: we can apply the results of Section 2 to these states.
314: States of similar form were discussed by Sanders \cite{sanders},
315: and Wielinga \cite{wi}, who called these states entangled
316: coherent states.
317: From the results in the Section 2, we know that
318: $|\Psi_2 \rangle_{AB}$
319: and $|\Psi_4 \rangle_{AB}$ have
320: one ebit of entanglement independent of $\alpha$.
321: This is an interesting and potentially useful property
322: (see next Section).
323: On the other hand, $|\Psi_1 \rangle_{AB} $ and
324: $|\Psi_3 \rangle_{AB}$ are
325: maximally entangled states only in the limit
326: $\alpha\rightarrow\infty$.
327: In order to avoid confusion with continuous
328: variable states \cite{kimble},
329: we should mention here that, of course, the dimension of
330: the space spanned by the quasi Bell states is 4 even though
331: they are embedded in a vector space of infinite dimension.
332: This implies that the maximum value of
333: the von Neumann entropy for the quasi Bell states is unity.
334: Different amplitudes $\alpha$ just give the degree of nonorthogonality.
335: It also means that one cannot use coherent entangled states
336: for teleportation of continuous quantum variables \cite{kimble,akira}.
337: In a separate paper \cite{enk}, we will show how to use
338: entangled coherent states
339: for teleportation of Schr\"odinger cat states.
340: The average photon numbers of the reduced states of the
341: quasi Bell states read
342: \begin{equation}
343: \langle n_{A}^{(1)} \rangle =
344: \frac{(1-\kappa^2)}{(1+\kappa^2)} |\alpha |^2,
345: \quad
346: \langle n_{A}^{(2)} \rangle=
347: \frac{(1+\kappa^2)}{(1-\kappa^2)} |\alpha |^2.
348: \end{equation}
349: Thus the quasi Bell states can have arbitrary photon numbers.
350: As said above, however, the quasi Bell states are not truly
351: continuous variable states and in particular
352: do not belong to the class of Gaussian states
353: \cite{holevo},\cite{holevo2} in contrast to, e.g.,
354: the two mode squeezed state \cite{yuen}.
355: This is shown in the following way.
356: The characteristic functions of the quasi Bell states
357: are given by
358: \begin{eqnarray}
359: C(\xi, \eta) &=& {\rm Tr}\big[
360: |\Psi\rangle_{AB} \langle\Psi|
361: \exp({\xi} {\mit a}_A^\dagger)
362: \exp({-\xi^*} {\mit a}_A)
363: \exp({\eta} {\mit a}_B^\dagger)
364: \exp({-\eta^*} {\mit a}_B)
365: \big] \nonumber \\
366: &\times& \exp\{-(|\xi |^2 + |\eta |^2)/2\}
367: \end{eqnarray}
368: where $a$ and $a^\dagger$ are the annihilation and
369: creation operators, respectively.
370: They can be calculated, with the result
371: \begin{eqnarray}
372: C(\xi, \eta |i=1, 2)&=&{h_{i}}^2\exp\{-(|\xi |^2 +
373: |\eta |^2)/2\}\{\exp(A_1-B_1)\alpha\nonumber\\
374: &+&\exp(-A_1+B_1)\alpha \pm \exp(A_2 - B_2)\alpha \nonumber \\
375: &\pm& \exp(-A_2 + B_2)\alpha\}\\
376: C(\xi, \eta |i=3,4)&=&{h_{i}}^2\exp\{-(|\xi |^2 +
377: |\eta |^2)/2\}\{\exp(A_1+B_1)\alpha\nonumber \\
378: &+&\exp(-A_1-B_1)\alpha \pm \exp(A_2 +B_2)\alpha \nonumber \\
379: &\pm& \exp(-A_2 - B_2)\alpha\}
380: \end{eqnarray}
381: where $A_1 = (\xi - \xi^*), A_2 = (\xi + \xi^*),
382: B_1 = (\eta - \eta^*),
383: B_2 = (\eta + \eta^*)$.
384: The characteristic functions are indeed not Gaussian.
385: Finally let us explore one more property of a similar set
386: of entangled states.
387: If the amplitudes of the modes A and mode B
388: in $|\Psi_2 \rangle_{AB}$ (or $|\Psi_4 \rangle_{AB}$)
389: are chosen to be different, say, $\alpha$ and $\beta$,
390: respectively, the eigenvalues of the reduced density operator
391: are
392: \begin{equation}
393: \lambda_{1} =
394: \frac{(1+\kappa_A)(1-\kappa_B)}{2(1-\kappa_A \kappa_B)},
395: \quad
396: \lambda_{2} =
397: \frac{(1-\kappa_A)(1+\kappa_B)}{2(1-\kappa_A \kappa_B)},
398: \end{equation}
399: where $\kappa_A=\exp(-2|\alpha|^2)$ and
400: $\kappa_B=\exp(-2|\beta|^2)$.
401: We can then easily see that the entropy of entanglement
402: attains its maximum value of 1
403: only when the amplitudes of both modes
404: are the same.
405:
406: \section{Decoherence properties}
407: In this section, we will discuss
408: decoherence properties of the state $|\Psi_2\rangle_{AB}$.
409: We are concerned with the decoherence due to energy loss or
410: photon absorption.
411: In particular, we would like to demonstrate that entangled coherent
412: states possess a certain degree of robustness against decoherence
413: when compared to a standard bi-photon Bell state.
414: We assume that
415: Alice produces a coherent entangled state $|\Psi_2\rangle_{AB}$,
416: keeps one part ($A$) and transmits
417: the other part $B$ to Bob through a lossy channel. Bob will receive
418: an attenuated optical state.
419: Thus, Alice prepares
420: \begin{equation}
421: |\Psi_2\rangle_{AB}
422: =h_2( |\alpha\rangle_A |-\alpha\rangle_B
423: -|-\alpha\rangle_A |\alpha \rangle_B ) \equiv|\Psi_2(\alpha)\rangle_{AB}
424: \end{equation}
425: where $h_2=1/\sqrt{2(1-{\kappa_A}^2)}$,
426: $\kappa_A = \langle \alpha | -\alpha \rangle$.
427: When we employ a half mirror model for the noisy channel,
428: the effect of energy losses is described by a linear coupling
429: with an external vacuum field as follows:
430: \begin{equation}
431: U_{BE}|\alpha\rangle_B|0\rangle_E =
432: |\sqrt{\eta}\alpha\rangle_B|\sqrt{(1-\eta)}\alpha\rangle_E.
433: \end{equation}
434:
435: where the mode $E$ is an external mode responsible for the energy loss,
436: $\eta$ is the noise parameter, and $\alpha$ is taken as real.
437: If we use $|\Psi_2(\alpha)\rangle_{AB} $ as the initial state,
438: the final state entangled with the environment is
439: \begin{eqnarray}
440: &\hat I_A\otimes U_{BE}|\Psi_2(\alpha)\rangle_{AB}
441: \otimes |0\rangle_E & \nonumber\\
442: &=h_2(|\alpha\rangle_A
443: |-\sqrt{\eta}\alpha\rangle_B|-\sqrt{(1-\eta)}\alpha\rangle_E
444: -|-\alpha\rangle_A|\sqrt{\eta}\alpha\rangle_B|\sqrt{(1-\eta)}\alpha\rangle_
445: E ).&
446: \end{eqnarray}
447: The normalized density operator shared by Alice and Bob is given by
448: a super operator calculation \cite{barnett2},
449: \begin{eqnarray}
450: \rho_{AB}&=& h_2^2\{|\alpha \rangle \langle
451: \alpha |
452: \otimes |-\sqrt{\eta}\alpha \rangle \langle -\sqrt{\eta}\alpha |
453: + |-\alpha \rangle \langle -\alpha |
454: \otimes |\sqrt{\eta}\alpha \rangle \langle \sqrt{\eta}\alpha |\nonumber \\
455: &-& |-\alpha \rangle \langle \alpha |
456: \otimes L |\sqrt{\eta}\alpha \rangle \langle -\sqrt{\eta}\alpha |
457: -|\alpha \rangle \langle -\alpha |
458: \otimes L |-\sqrt{\eta}\alpha \rangle \langle \sqrt{\eta}\alpha |\}
459: \end{eqnarray}
460: where
461: $L = \exp {\{-2(1-\eta)|\alpha|^2\}}$.
462: Let us discuss the entanglement of the above density
463: operator. As discussed in Section 2, we can use the entangled
464: fraction to measure the entanglement of this mixed state.
465: The fully entangled fraction of $\rho_{AB}$ is given by
466: \begin{equation}
467: f(\rho_{AB}) =\max_{\beta} \mbox{}_{AB}\! \langle \Psi_2(\beta)
468: |\rho_{AB}|\Psi_2(\beta)\rangle_{AB}
469: \end{equation}
470: because $|\Psi_2(\beta) \rangle_{AB}$ is indeed maximally entangled.
471: The above is given by
472: \begin{equation}
473: f(\rho_{AB})=\max_{\beta}\frac{(1+L)({\kappa_1}^2(\beta){\kappa_2}^2(\beta) +
474: {\kappa_3}^2(\beta) {\kappa_4}^2(\beta)
475: -2 \kappa_1(\beta) \kappa_2(\beta) \kappa_3(\beta)\kappa_4(\beta))}
476: {2(1-{\kappa_A}^2)(1-{\kappa_0}^2(\beta))}
477: \end{equation}
478: where
479: $\kappa_0(\beta) = \exp \{-2|\beta|^2\}$,
480: $\kappa_1(\beta) = \exp \{-|\alpha - \beta|^2/2\}$,
481: $\kappa_2(\beta) = \exp \{-|\beta - \sqrt{\eta}\alpha|^2/2\}$,
482: $\kappa_3(\beta) = \exp \{-|\alpha + \beta|^2/2\}$, and
483: $\kappa_4(\beta) = \exp \{-|\beta + \sqrt{\eta}\alpha|^2/2\}$.
484: The maximum is attained for the value
485: \begin{equation}
486: \beta=\frac{\alpha+\sqrt{\eta}\alpha}{2}
487: \end{equation}
488: that is, exactly halfway between the original coherent amplitude
489: and the attenuated amplitude.
490: In Figure 1 we plot $f(\rho_{AB})$ as a function of $|\alpha|$
491: for various values of the noise parameter $\eta$.
492: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
493: \begin{figure}[h]\label{f1} \leavevmode
494: \centerline{
495: \epsfig{file=ent.eps,width=8cm}
496: %\epsfile{file=ent.eps,width=8cm}
497: }
498: %\epsfxsize=8cm \epsfbox{ent.eps}
499: \caption{Overlap of $\rho_{AB}$
500: with appropriately chosen fully entangled state (see text) as
501: a function of $|\alpha|$ for noise parameters
502: $\eta=0.9,0.7,0.5,0.3,0.1$ for top to bottom curves.}
503: \end{figure}
504: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
505: For comparison, we consider the biphoton Bell state:
506: \begin{equation}
507: |\Psi_{2(p)}\rangle_{AB} ={1\over\sqrt2}
508: ( |\updownarrow\rangle_A |\leftrightarrow\rangle_B
509: -|\leftrightarrow\rangle_A |\updownarrow\rangle_B ),
510: \end{equation}
511: where $\{|\updownarrow\rangle,|\leftrightarrow\rangle\}$ denotes
512: single photon polarization directions.
513: After passing through the same lossy
514: channel, Alice and Bob share the state
515: \begin{equation}
516: \rho_{AB}^{\rm pol}=
517: \eta|\Psi_{2(p)}\rangle_{AB} \langle\Psi_{2(p)}|
518: +(1-\eta)\frac{1}{2}I_A\otimes|0\rangle_B\langle0|
519: \label{biphoton_entangle}
520: \end{equation}
521: with
522: $I_A$ the identity on mode $A$.
523: As a result, the fully entangled fraction is
524: \begin{equation}
525: f(\rho_A^{\rm pol})=\eta
526: \end{equation}
527: in that case, which is clearly less than for entangled coherent
528: states with sufficiently small amplitudes (see Figure 1).
529: Finally, we note that in \cite{enk} we give the analogous
530: decoherence properties for a symmetric noise channel,
531: describing the case where both Alice's and Bob's mode suffer losses.
532: \section{Concluding remarks}
533: In this paper, our effort was devoted to clarify
534: several properties of entangled nonorthogonal states.
535: We constructed 4 entangled states that generalize the standard Bell states.
536:
537: Two out of these 4 ``quasi Bell states'' possess less than one unit of
538: entanglement, the other two, however, possess exactly one unit.
539: The latter two states were shown to be more
540: robust against decoherence due to photon absorption than are bi-photon Bell states.
541: The most important remaining problem is the physical realization of
542: such states, which is discussed in a forthcoming separate paper \cite{enk}.
543:
544: \section*{Acknowledgment}
545: We are grateful to C.H.Bennett, S.M.Barnett, C.A.Fuchs,
546: A.S.Holevo, R.Jozsa,\\
547: M.Sasaki, and P.Shor
548: for helpful discussions.
549: The first idea of quasi Bell state was presented in QCM
550: $\&$ C-2000 held in Capri. \\
551: This work was supported by
552: the project in CREST Japan.
553:
554: \begin{thebibliography}{99}
555:
556: \bibitem{Bennett} Bennett.C.H, Bernstein.H.J, Popescu.S,
557: and Schumacher.B,
558: \newblock{1996, Phys. Rev.,{\bf A-53}, 2046}.
559:
560: \bibitem{Bennett2} Bennett.C.H, DiVincenzo.D.P, Smolin.J.A,
561: and Wootters.W.K,
562: \newblock{1996, Phys. Rev.,{\bf A-54}, 3824}.
563:
564: \bibitem{vedral} Vedral.V, and Plenio.M.B,
565: \newblock{1998, Phys. Rev.,{\bf A-57}, 1619}.
566:
567: \bibitem{Wootters} Hill.S, and Wootters.W.K,
568: \newblock{1997, Phys. Rev. Lett.,{\bf 78}, 5022}.
569:
570: \bibitem{Wootters2} Wootters.W.K,
571: \newblock{1998, Phys. Rev.Lett.,{\bf 80}, 2245}.
572:
573: \bibitem{Barnett} Barnett.S.M, and Phoenix.S.J.D,
574: \newblock{1989, Phys. Rev.{\bf A-40}, 2404}.
575:
576: \bibitem{werner} Werner.R.F,
577: \newblock{1989, Phys. Rev.{\bf A-40}, 4277}.
578:
579: \bibitem{sanders} Sanders.B.C,
580: \newblock{1992, Phys. Rev.{\bf A-45}, 6811}.
581:
582: \bibitem{wi} Wielinga.B, and Sanders.B.C,
583: \newblock{1993, Journal of Modern Optics, {\bf 40}, 1923-1937}.
584:
585: \bibitem{hirota0} Hirota.O,
586: \newblock{2000, Applicable Algebra in Eng. Commun. and Computing,
587: {\bf 10}, 401-423}.
588:
589: \bibitem{holevo} Holevo.A.S,
590: \newblock{1975, IEEE.{\bf IT-21}, 533-543}.
591:
592: \bibitem{holevo2} Holevo.A.S, Sohma.M, and Hirota.O,
593: \newblock{1999, Phys. Rev.{\bf A-59}, 1820-1828}.
594:
595: \bibitem{yuen} Yuen.H.P, and Shapiro.J.H,
596: \newblock{1979, Optics Lett., {\bf 4}, 334}.
597:
598: \bibitem{kimble} Braunstein.S.L, and Kimble.H.J,
599: \newblock{1998, Phys. Rev. Lett., {\bf 80}, 869}.
600:
601: \bibitem{akira}A. Furusawa, J.L. S\o rensen, S.L. Braunstein,
602: C.A. Fuchs, H.J. Kimble, and E.S. Polzik, Science {\bf 282}, 706 (1998).
603:
604: \bibitem{enk} van Enk, S.J. and Hirota.O,
605: \newblock{2000, submitted to PRA}.
606:
607: \bibitem{barnett2} Barnett.S.M, and Radmore.P.M,
608: \newblock{1997, Methods in theoretical quantum optics, Oxford Press}.
609:
610: \bibitem{bennett3} Bennett.C,H, Brassard.H.G, Crepeau.C,
611: Jozsa.R, Peres.A, and Wootters.W.K,
612: \newblock{1993, Phys. Rev. Lett.,{\bf 70}, 1895}.
613:
614: \bibitem{hirota} Hirota.O, and Sasaki.M,
615: \newblock{Proc.QCM and C of 2000, in press}.
616:
617: \end{thebibliography}
618: \end{document}
619:
620: