1: \documentclass[pra,showpacs,showkeys,amsfonts,amsmath,twocolumn]{revtex4}
2: \usepackage{bm}
3: %\usepackage[cp1250]{inputenc}
4: \usepackage{graphicx}
5: \RequirePackage{mathptm}
6: \newtheorem{thm}{Theorem}
7: \newtheorem{cor}{Corollary}
8: %\newtheorem{lem}[thm]{Lemma}
9: %\newtheorem{defn}[thm]{Definition}
10: \newtheorem{rem}{Remark}
11: \numberwithin{equation}{section}
12: %\newcommand{\norm}[1]{\left\Vert#1\right\Vert}
13: %\newcommand{\abs}[1]{\left\vert#1\right\vert}
14: %\newcommand{\set}[1]{\left\{#1\right\}}
15: %\newcommand{\Real}{\mathbb R}
16: %\newcommand{\eps}{\varepsilon}
17: %\newcommand{\To}{\longrightarrow}
18: %\newcommand{\BX}{\mathbf{B}(X)}
19: %\newcommand{\A}{\mathcal{A}}
20: \newcommand{\si}[1]{\sigma_{#1}}
21: \newcommand{\sa}[2]{\sigma_{#1}^{#2}}
22: \newcommand{\ip}[2]{\langle \,{#1},\,{#2}\,\rangle}
23: \newcommand{\W}[4]{\begin{cases}
24: #1 ,\\
25: #3 ,
26: \end{cases}}
27: \newcommand{\ro}{\rho}
28: \newcommand{\la}{\lambda}
29: \newcommand{\La}{\Lambda}
30: \newcommand{\al}{\alpha}
31: \newcommand{\be}{\beta}
32: \newcommand{\g}{\gamma_{0}}
33: \newcommand{\ga}{\gamma}
34: \newcommand{\Ga}{\Gamma}
35: \newcommand{\om}{\omega_{0}}
36: \newcommand{\re}{\mathrm{Re}\,}
37: \newcommand{\I}{\mathbb I}
38: \newcommand{\ket}[1]{|{#1}\rangle}
39: \newcommand{\bra}[1]{\langle {#1} |}
40: \newcommand{\cH}{{\mathcal H}}
41: \newcommand{\C}{\mathbb C}
42: \newcommand{\R}{\mathbb R}
43: \newcommand{\fA}{\mathfrak A}
44: \newcommand{\fE}{\mathcal E}
45: \newcommand{\tr}{\mathrm{tr}\,}
46: \newcommand{\ptr}[1]{\mathrm{tr}_{#1}}
47: \newcommand{\tl}[1]{\boldsymbol #1}
48: \newcommand{\mr}[1]{\mathrm{#1}}
49: \newcommand{\kw}{a^{2}}
50: \newcommand{\aaa}{1-a^{2}}
51: %\newcommand{\begin{equation}}{\begin{equation}}
52: %\newcommand{\end{equation}}{\end{equation}}
53: \newcommand{\DS}{\displaystyle}
54: \begin{document}
55: \title{CHSH violation and entropy -- concurrence plane}
56: \author{{\L}ukasz Derkacz}
57: \affiliation{Institute of Theoretical Physics\\ University of
58: Wroc{\l}aw\\
59: Pl. M. Borna 9, 50-204 Wroc{\l}aw, Poland}
60: \author{Lech Jak{\'o}bczyk\footnote{
61: E-mail addres: ljak@ift.uni.wroc.pl}}
62: \affiliation{Institute of Theoretical Physics\\ University of
63: Wroc{\l}aw\\
64: Pl. M. Borna 9, 50-204 Wroc{\l}aw, Poland}
65: \begin{abstract}
66: We characterize violation of CHSH inequalities for mixed two-qubit
67: states by their mixedness and entanglement. The class of states
68: that have maximum degree of CHSH violation for a given linear
69: entropy is also constructed.
70: \end{abstract}
71: \pacs{03.65.-w; 03.65.Ud} \keywords{entanglement; mixedness; CHSH
72: inequalities} \maketitle
73: \section{Introduction}
74: Physically allowed degree of entanglement and mixture for two -
75: qubit mixed states were investigated by Munro \textit{et al.} in
76: terms of concurrence $C(\ro)$ and normalized linear entropy
77: $S_{\mr{L}}(\ro)$ \cite{MJWK}. These authors characterize the subset
78: $\La$ on $(C,\, S_{\mr{L}})$ plane corresponding to possible states
79: of the system and in particular identify maximally entangled mixed
80: states $\ro_{\mr{MEMS}}$ laying on the boundary of that subset. The
81: states $\ro_{\mr{MEMS}}$ have maximal allowed entanglement for a
82: given degree of mixedness. They obtained this result analytically
83: for some class of states. Numerical results suggests correctness of
84: the picture for general two-qubit states.
85: \par
86: In this note, we consider the problem of violation of Bell -- CHSH
87: inequalities \cite{Bell, CHSH} for mixed states. It is well known
88: that quantum states violating these inequalities have to be
89: entangled \cite{E, S}, but on the other hand, CHSH violation is
90: not necessary for mixed state entanglement \cite{Werner}. In the
91: context of the results of Ref.\cite{MJWK}, we address the
92: following question: what are the subsets of $\La$ which correspond
93: to states violating Bell -- CHSH inequalities ? In our previous
94: publication \cite{DJ}, we have studied the structure of such
95: subsets in the case of specific class of quantum states. The
96: results show that $\La$ is a sum of disjoint subsets
97: $\La_{\mr{V}},\, \La_{\mr{NV}}$ and $\La_{0}$ with the following
98: properties: states belonging to $\La_{\mr{V}}$ violate CHSH
99: inequalities, whereas states from $\La_{\mr{NV}}$ fulfil all CHSH
100: inequalities. The subset $\La_{0}$ has somehow unexpected
101: property: for any pair $(S_{\mr{L}},\, C)\in \La_{0}$ there are
102: two families of states with the same entropy and concurrence such
103: that all states from one family violate CHSH and at the same time,
104: all states from the other family fulfil all CHSH inequalities. In
105: the present paper, we continue these investigations for general
106: class of two-qubit states. First we consider larger class of
107: states still admitting explicit formulas for linear entropy,
108: concurrence and degree of CHSH violation. Unfortunately,
109: analytical analysis of the relation between these functions is not
110: possible. Numerical investigations lead to some modifications of
111: the picture from Ref. \cite{DJ}, but the general structure is not
112: changed. Finally, this problem is studied using numerically
113: generated arbitrary density matrices. The results indicate that
114: the structure of $\La$ for general two-qubit states seems to be
115: the same.
116: \par
117: We consider also the problem of maximal violation of CHSH
118: inequalities. For a class of mixed states we obtain counterpart of
119: the result of Ref.\cite{MJWK}, namely we find the form of mixed
120: states with maximal degree of CHSH violation for given linear
121: entropy. All that states lie on specific curve on entropy --
122: concurrence plane. Numerical results suggest also that general two
123: -- qubits states with this property satisfy the same relation
124: between entropy and concurrence. As we show, maximal violation of
125: Bell inequalities with fixed linear entropy is not equivalent to the
126: maximal entanglement under the same conditions.
127: \section{CHSH inequalities}
128: Let $\tl{a},\,\tl{a}^{\prime},\,\tl{b},\,\tl{b}^{\prime}$ be the
129: unit vectors in $\R^{3}$ and
130: $\tl{\sigma}=(\si{1},\,\si{2},\,\si{3})$. Consider the family of
131: operators on two - qubits Hilbert space $\cH_{AB}=\C^{4}$
132: \begin{equation}
133: B_{CHSH}=\tl{a}\cdot\tl{\sigma}\otimes
134: (\tl{b}+\tl{b}^{\prime})\cdot\tl{\sigma}+\tl{a}^{\prime}\cdot\tl{\sigma}\otimes
135: (\tl{b}-\tl{b}^{\prime})\cdot\tl{\sigma}\label{B}
136: \end{equation}
137: Then Bell - CHSH \cite{CHSH} inequalities are
138: \begin{equation}
139: |\tr (\ro\, B_{CHSH})|\leq 2\label{CHSH}
140: \end{equation}
141: If the above inequality is not satisfied by the state $\ro$ for
142: some choice of $\tl{a},\,\tl{a}^{\prime},\, \tl{b},\,
143: \tl{b}^{\prime}$ , we say that $\ro$ \textit{violates Bell-CHSH
144: inequalities.} In the case of two-qubits, the violation of Bell -
145: CHSH inequalities by mixed states can be studied using simple
146: necessary and sufficient condition \cite{HHH,H}. Consider real
147: matrix
148: \begin{equation}
149: T_{\ro}=(t_{nm}),\quad t_{nm}=\tr
150: (\ro\,\si{n}\otimes\si{m})\label{Tmatix}
151: \end{equation}
152: and real symmetric matrix
153: \begin{equation}
154: U_{\ro}=T_{\ro}^{T}\,T_{\ro}\label{Umatix}
155: \end{equation}
156: where $T_{\ro}^{T}$ is the transposition of $T_{\ro}$. Let
157: \begin{equation}
158: m(\ro)=\max_{j<k}\; (u_{j}+u_{k})\label{m}
159: \end{equation}
160: and $u_{j},\, j=1,2,3$ are the eigenvalues of $U_{\ro}$. As was
161: shown in \cite{HHH,H}
162: \begin{equation}
163: \max_{B_{CHSH}}\,\tr (\ro\, B_{CHSH})=2\,\sqrt{m(\ro)}
164: \end{equation}
165: Thus (\ref{CHSH}) is violated by some choice of
166: $\tl{a},\tl{a}^{\prime},\tl{b},\tl{b}^{\prime}$ if and only if
167: $m(\ro)>1$.
168: \par
169: We need also the measures of degree of entanglement and mixture for
170: given state. In the case of two qubits, the useful measure of degree
171: of entanglement is concurrence $C(\ro)$
172: \begin{equation}
173: C(\ro)=\max\, (0,\,
174: 2\lambda_{\mr{max}}(\widehat{\ro})-\tr
175: \widehat{\ro}\,)\label{concurrence}
176: \end{equation}
177: where $\lambda_{\mr{max}}(\widehat{\ro})$ is the maximal eigenvalue
178: of $\widehat{\ro}$ and
179: $$
180: \widehat{\ro}=\sqrt{\sqrt{\ro}\,\widetilde{\ro}\,\sqrt{\ro}},\quad
181: \widetilde{\ro}=(\si{2}\otimes\si{2})\overline{\ro}(\si{2}\otimes\si{2})
182: $$
183: with $\overline{\ro}$ denoting complex conjugation of the matrix
184: $\ro$. It is known that $C(\ro)$ can be used to obtain entanglement
185: of formation, which is natural measure of entanglement for mixed
186: states \cite{HW,Woo}. To measure degree of mixture, or deviation
187: from pure state, we use linear entropy
188: \begin{equation}
189: S_{\mr{L}}(\ro)=\frac{4}{3}\,(1-\tr \ro^{2}\,)\label{linentr}
190: \end{equation}
191: which is normalized such that its maximal value equals $1$.
192: \section{CHSH inequalities and entropy -- concurrence plane}
193: To study the subset of entropy -- concurrence plane corresponding
194: to violation of CHSH inequalities, consider first the following
195: class $\fE_{0}$ of states
196: \begin{equation}
197: \ro=\begin{pmatrix} 0&0&0&0\\ 0&a&\frac{1}{2}ce^{i\theta}&0\\
198: 0&\frac{1}{2}ce^{-i\theta}&b&0\\ 0&0&0&1-a-b
199: \end{pmatrix} \label{class}
200: \end{equation}
201: where
202: $$
203: c\in [0,1],\: a,b\geq 0,\: \theta\in [0,2\pi]
204: $$
205: and
206: $$
207: ab\geq\frac{c^{2}}{4},\quad a+b\leq 1
208: $$
209: Notice that for $\ro\in \fE_{0}$
210: $$
211: C(\ro)=c
212: $$
213: Define the subset $\La\subset \R^{2}$
214: \begin{equation}
215: \La=\{ (\,S_{L}(\ro),\,C(\ro)\,)\,:\, C(\ro)>0\quad\text{and}\quad
216: \ro\in \fE_{0}\}\label{obszar}
217: \end{equation}
218: In the paper \cite{DJ}, we analysed the set (\ref{obszar}) and we
219: have shown that it is a sum of disjoint subsets $\La_{\mr{V}},\,
220: \La_{0}$ and $\La_{\mr{NV}}$ with properties:
221: \begin{enumerate}
222: \item[\textbf{1.}] If $(s,c)\in \La_{\mr{V}}$, then every state $\ro\in
223: \fE_{0}$ such that $S_{L}(\ro)=s$ and $C(\ro)=c$ satisfies
224: $m(\ro)>1$. \vskip 2mm\noindent
225: \item[ \textbf{2.}] If $(s,c)\in
226: \La_{0}$, then there exist states $\ro_{1},\, \ro_{2}\in \fE_{0}$
227: such that
228: $$
229: S_{L}(\ro_{1})=S_{L}(\ro_{2})=s,\quad C(\ro_{1})=C(\ro_{2})=c
230: $$
231: and $m(\ro_{1})>1$, but $m(\ro_{2})<1$.
232: \vskip 2mm\noindent
233: \item[\textbf{3.}] If $(s,c)\in \La_{\mr{NV}}$, then every state $\ro\in
234: \fE_{0}$ such that $S_{L}(\ro)=s$ and $C(\ro)=c$ satisfies
235: $m(\ro)<1$.
236: \end{enumerate}
237: Detailed analytic description of regions $\La_{\mr{V}},\, \La_{0}$
238: and $\La_{\mr{NV}}$ as well as the proof of the above properties,
239: can be found in Ref. \cite{DJ}.
240: \par
241: In the present paper, we try to extend this analysis to the larger
242: class of two - qubit states. For general mixed two-qubit states
243: there is a bound on concurrence that guarantees violation of CHSH
244: inequalities irrespective of linear entropy. It follows from the
245: result of Verstraete and Wolf \cite{VW} that minimal violation of
246: CHSH inequality for given concurrence $C$ is equal to
247: \begin{equation}
248: m_{\mr{min}}=\max \,(1,\, 2C^{2}\,)\label{minimal}
249: \end{equation}
250: So if
251: \begin{equation}
252: C(\ro)>\frac{1}{\sqrt{2}}\label{cbound}
253: \end{equation}
254: then minimal value of $m$ is greater then $1$, and every state
255: satisfying (\ref{cbound}) violates CHSH inequality. On the other
256: hand, there is a bound on linear entropy that guarantees fulfilling
257: CHSH inequalities irrespective of concurrence. It is given by the
258: result of Santos \cite{Santos} that all states with (normalized)
259: linear entropy
260: \begin{equation}
261: S_{\mr{L}}(\ro)>\frac{2}{3}\label{sbound}
262: \end{equation}
263: satisfy all CHSH inequalities. We see that these general bounds are
264: compatible with our previous analysis (FIG. 1) and possible
265: modifications can occur in region $\Lambda_{\mr{V}}$ below the line
266: $C=\frac{1}{\sqrt{2}}$ and in region $\Lambda_{\mr{NV}}$ on the left
267: hand side of the line $s=\frac{2}{3}$.
268: \begin{figure}[h]
269: \centering {\includegraphics[height=56mm]{obszar.eps}} \caption{The
270: structure of the set $\Lambda$ and Verstraete -- Wolf and Santos
271: bounds (dotted lines) }
272: \end{figure}
273: Consider now the larger class $\fE_{1}$ of states of the form
274: \begin{equation}
275: \ro=\begin{pmatrix} \ro_{11}&0&0&\ro_{14}\\
276: 0&\ro_{22}&\ro_{23}&0\\
277: 0&\ro_{32}&\ro_{33}&0\\
278: \ro_{41}&0&0&\ro_{44}
279: \end{pmatrix}\label{class1}
280: \end{equation}
281: One can check that for that class
282: \begin{equation}
283: C(\ro)=\max\, (0,\, C_{1},\, C_{2}\,)
284: \end{equation}
285: where
286: \begin{equation}
287: \begin{split}
288: &C_{1}=2\,(\, |\ro_{14}|-\sqrt{\ro_{22}\ro_{33}}\,)\\[2mm]
289: &C_{2}=2\, (\, |\ro_{23}|-\sqrt{\ro_{11}\ro_{44}}\,)
290: \end{split}
291: \end{equation}
292: and
293: \begin{equation}
294: S_{\mr{L}}(\ro)=1-\ro_{11}^2-\ro_{22}^2-\ro_{33}^2-\ro_{44}^2-2|\ro_{14}|^2-2|\ro_{23}|^2
295: \end{equation}
296: Moreover,
297: \begin{eqnarray}
298: m(\ro)=&&\max\,\big[\,4(|\ro_{41}|+|\ro_{23}|)^2,
299: 4(|\ro_{41}|-|\ro_{23}|)^2, \nonumber\\
300: &&\hspace*{7mm}(\ro_{11}-\ro_{22}-\ro_{33}+\ro_{44})^2\,\big]
301: \end{eqnarray}
302: For the class (\ref{class1}), analytic description of regions of
303: $\Lambda$ where $m(\ro)>1$ or $m(\ro)<1$ is not possible, but it
304: can be done numerically. Results of numerical analysis of the
305: class $\fE_{1}$ are presented on FIG. 2. We see that the region
306: $\Lambda_{\mr{V}}$ where all states violate CHSH inequalities is
307: not changed, but there are states with $m(\ro)>1$ for some points
308: in $\Lambda_{\mr{NV}}$, so the region $\Lambda_{0}$ is slightly
309: enlarged.
310: \begin{figure}[h]
311: \centering {\includegraphics[height=56mm]{obsznum.eps}}
312: \caption{Numerical analysis of the set $\Lambda$ for the class
313: $\fE_{1}$: $m(\ro)>1$ (dark grey), $m(\ro)>1$ and $m(\ro)<1$ (grey),
314: $m(\ro)<1$ (light grey)}
315: \end{figure}
316: To study this problem for general two-qubit density matrices, we
317: numerically generate $3\cdot 10^{6}$ randomly chosen density
318: matrices. For such two-qubit states the structure of the set of
319: pairs $(S_{\mr{L}},C)$ is very simple. There are only points
320: corresponding to $m(\ro)>1$ or $m(\ro)<1$ (FIG. 3). Notice that by
321: the method of random choice of states, not all points of $\La$ are
322: achieved (the boundary of generated set correspond exactly to the
323: class of Werner states), but the obtained structure is compatible
324: with previous results. To have some insight into the properties of
325: the remaining part of the set $\La$, we modify the method of
326: generation of states and consider density matrices lying close to
327: the boundary of the set of all states i.e. such $\ro$ that one of
328: its eigenvalues is almost equal to zero. For these randomly
329: generated states, the pairs $(S_{\mr{L}},C)$ cover the whole set
330: $\La$, and its structure is the same as for the class $\fE_{1}$
331: (FIG. 4). The results suggest that the picture obtained using the
332: class $\fE_{1}$ should be correct also for all two-qubit density
333: matrices, although for most of randomly chosen density matrices
334: $\ro$ with fixed mixedness and linear entropy, either $m(\ro)>1$ or
335: $m(\ro)<1$. Unfortunately, we do not know analytic description of
336: the boundary of the enlarged region $\La_{0}$.
337: \begin{figure}[h]
338: \centering {\includegraphics[height=56mm]{wszystkie.eps}}
339: \caption{ The set $(S_{\mr{L}},\, C)$ for randomly chosen
340: two-qubit states: $m(\ro)>1$ (dark grey), $m(\ro)<1$ (light grey).
341: The boundary corresponds to the family of Werner states}
342: \end{figure}
343: \begin{figure}[h]
344: \centering {\includegraphics[height=56mm]{wszystkiebrzeg.eps}}
345: \caption{ The set $\Lambda$ for numerically generated states lying
346: close to the boundary: $m(\ro)>1$ (dark grey), $m(\ro)>1$ and
347: $m(\ro)<1$ (grey), $m(\ro)<1$ (light grey)}
348: \end{figure}
349: \section{Maximal CHSH violation}
350: Consider now the values of $m(\ro)$ for states violating Bell --
351: CHSH inequalities. We are especially interested in maximal values
352: of $m(\ro)$. For the class (\ref{class})
353: \begin{equation}
354: m(\ro)=\max (\,2c^{2},\, (2(a+b)-1)^{2}+c^{2}\,)\label{m}
355: \end{equation}
356: We see that (\ref{m}) is maximal iff $a+b=1$, and then
357: \begin{equation}
358: m(\ro)=1+c^{2}\label{mmax}
359: \end{equation}
360: By general result of Verstraete and Wolf \cite{VW}, for any
361: two-qubit state, (\ref{mmax}) is the maximal degree of CHSH
362: violation for given concurrence $c$. But we ask another question:
363: what is the maximum of (\ref{mmax}) \textit{for fixed linear
364: entropy}, and which states realize that maximum? We can simply
365: answer this question for the class of states (\ref{class}). Since
366: \begin{equation}
367: S_{\mr{L}}(\ro)=\frac{4}{3}\,\left(1-a^{2}-(1-a)^{2}-\frac{c^{2}}{2}\,\right)\label{entropy}
368: \end{equation}
369: so fixing $S_{\mr{L}}(\ro)=s$, we obtain
370: \begin{equation}
371: m(\ro)=1+4\, (a-a^{2})-\frac{3}{2}s\label{msfixed}
372: \end{equation}
373: Maximum of (\ref{msfixed}) is achieved at $a=\frac{1}{2}$ and
374: equals to $2-\frac{\DS 3}{\DS 2}s$ for $s\in [0,\frac{2}{3}]$. In
375: this way we obtain
376: \begin{thm}
377: In the class (\ref{class}), states maximizing degree of violation of
378: CHSH inequalities for fixed linear entropy, lie on the curve
379: $$
380: s=\frac{2}{3}\,(1-c^{2}\,)
381: $$
382: and have the form
383: \begin{equation}
384: \ro_{\mr{MVB}}=\frac{1}{2}\begin{pmatrix}0&0&0&0\\
385: 0&1&\sqrt{\beta-1}\,e^{i\theta}&0\\
386: 0&\sqrt{\beta-1}\,e^{-i\theta}&1&0\\
387: 0&0&0&0
388: \end{pmatrix}\label{mvb}
389: \end{equation}
390: with
391: $$
392: \beta\in [1,2],\: \theta\in [0,2\pi]
393: $$
394: Moreover,
395: $$
396: m(\ro_{\mr{MVB}})=\beta
397: $$
398: \end{thm}
399: \begin{figure}[h]
400: \centering {\includegraphics[height=60mm]{krzywa.eps}} \caption{
401: States with maximal degree of CHSH violation (dotted curve) on
402: $(S_{\mr{L}},\, C)$ plane}
403: \end{figure}
404: It is instructive to compare the value of $m(\ro_{\mr{MVB}})$ with
405: degree of CHSH violation for some other classes of states. Let $W$
406: be the family of Werner states
407: \begin{equation}
408: W=(1-p)\frac{\I_{4}}{4}+p\,\ket{\Psi^{-}}\bra{\Psi^{-}}\label{Werner}
409: \end{equation}
410: where $\Psi^{-}$ is a singlet state of two-qubits. Then
411: \begin{equation}
412: m(W)=2-2s\label{mW}
413: \end{equation}
414: For maximally entangled mixed states $\ro_{\mr{MEMS}}$ introduced in
415: Ref. \cite{MJWK}, the corresponding value of $m$ is given by
416: \begin{equation}
417: m(\ro_{\mr{MEMS}})=1-\frac{3}{4}s+\sqrt{1-\frac{3}{2}s}\label{mMEMS}
418: \end{equation}
419: We see that for a fixed linear entropy
420: \begin{equation}
421: m(\ro_{\mr{MVB}})\geq m(\ro_{\mr{MEMS}})\geq m(W)
422: \end{equation}
423: although
424: \begin{equation}
425: C(\ro_{\mr{MEMS}})\geq C(W)\geq C(\ro_{\mr{MVB}})
426: \end{equation}
427: So
428: \begin{cor}
429: The states with maximum amount of entanglement for a given linear
430: entropy do not maximize degree of Bell -- CHSH violation.
431: \end{cor}
432: \begin{figure}[ht]
433: \centering \rotatebox{270}
434: {\includegraphics[height=90mm]{maksi.eps}} \caption{ Plot of $\max\,
435: (m(\ro)-1\,)$ as the function of $S_{\mr{L}}$ and $C$ for the class
436: $\fE_{1}$. The curve corresponds to $m(\ro_{\mr{MVB}})-1$}
437: \end{figure}
438: The family of states with maximal degree of violation of Bell --
439: CHSH inequalities has another remarkable property. It is known that
440: fidelity of state $\ro$ defined as
441: \begin{equation}
442: F(\ro)=\max\, \ip{\psi}{\ro\psi}\label{fidelity}
443: \end{equation}
444: where the maximum is taken over all maximally entangled pure states
445: $\psi$ is bounded above by \cite{VV}
446: \begin{equation}
447: F(\ro)\leq \frac{1+C(\ro)}{2}\label{fidelity bound}
448: \end{equation}
449: By direct computation, one can check that
450: $$
451: F(\ro_{\mr{MVB}})=\frac{1+C(\ro_{\mr{MVB}})}{2}
452: $$
453: Thus
454: \begin{cor}
455: The states $\ro_{\mr{MVB}}$ maximize fidelity for given
456: concurrence.
457: \end{cor}
458: For a larger class $\fE_{1}$ we have studied $m(\ro)$ as a
459: function of $S_{\mr{L}}$ and $C$ numerically. Again the results
460: agree with those obtained analytically for the class $\fE_{0}$
461: (FIG. 6).
462: \begin{acknowledgments}
463: L.J. acknowledges financial support by Polish Ministry of
464: Scientific Research and Information Technology under the grant
465: PBZ-Min-008/PO3/2003.
466: \end{acknowledgments}
467: \begin{thebibliography}{99}
468: \bibitem{MJWK} W.J. Munro, D.F.V. James, A. G. White, P.G. Kwiat,
469: Phys. Rev. \textbf{A 64}, 030302(2001).
470: \bibitem{Bell} J.S. Bell, Physics (N.Y.)\textbf{1}, 195(1965).
471: \bibitem{CHSH} J.F. Clauser, M.A. Horne, A. Shimony and R.A. Holt,
472: Phys. Rev. Lett. \textbf{23}, 880(1969).
473: \bibitem{E} A. Einstein, B. Podolsky, N. Rosen, Phys. Rev.
474: \textbf{47}, 777(1935).
475: \bibitem{S} E. Schr\"odinger, Naturwissenschaften \textbf{23},
476: 812(1935).
477: \bibitem{Werner} R.F. Werner, Phys. Rev. A\textbf{40}, 4277(1989).
478: \bibitem{DJ} {\L}. Derkacz, L. Jak{\'o}bczyk, Phys. Lett.
479: A\textbf{328}, 26(2004).
480: \bibitem{HHH} R. Horodecki, P. Horodecki, M. Horodecki, Phys.
481: Lett. \textbf{A 200}, 340(1995).
482: \bibitem{H} R. Horodecki, Phys. Lett. \textbf{A 210}, 223(1996).
483: \bibitem{HW} S. Hill, W.K. Wooters, Phys. Rev. Lett. \textbf{78},
484: 5022(1997).
485: \bibitem{Woo} W.K. Wooters, Phys. Rev. Lett. \textbf{80},
486: 2254(1998).
487: \bibitem{VW} F. Verstraete, M.M. Wolf, Phys. Rev. Lett.
488: \textbf{89}, 170401-1(2002).
489: \bibitem{Santos} E. Santos, Phys.Rev. A\textbf{70}, 059901(E)
490: (2004).
491: \bibitem{VV} F. Verstraete, H. Verschelde, Phys. Rev. A\textbf{66},
492: 022307(2002).
493: \end{thebibliography}
494: \end{document}
495: