1:
2: \documentclass{article}
3: \usepackage{amsmath}
4: \usepackage{amssymb}
5: \input epsf
6: \textheight 22cm
7: \textwidth 15cm
8: \oddsidemargin 0cm
9: \evensidemargin 0cm
10: \newcommand{\si}[1]{\sigma_{#1}}
11: \newcommand{\sa}[2]{\sigma_{#1}^{#2}}
12: \newcommand{\ip}[2]{\langle \,{#1},\,{#2}\,\rangle}
13: \newcommand{\ro}{\rho}
14: \newcommand{\la}{\lambda}
15: \newcommand{\al}{\alpha}
16: \newcommand{\be}{\beta}
17: \newcommand{\g}{\gamma_{0}}
18: \newcommand{\ga}{\gamma}
19: \newcommand{\tet[1]}{\theta_{#1}}
20: \newcommand{\re}{\mathrm{Re}\,}
21: \newcommand{\I}{\mathbb I}
22: \newcommand{\ket}[1]{|{#1}\rangle}
23: \newcommand{\bra}[1]{\langle {#1} |}
24: \newcommand{\cH}{{\mathcal H}}
25: \newcommand{\C}{\mathbb C}
26: \newcommand{\R}{\mathbb R}
27: \newcommand{\fA}{\mathfrak A}
28: \newcommand{\fE}{\mathcal E}
29: \newcommand{\tr}{\mathrm{tr}\,}
30: \newcommand{\ptr}[1]{\mathrm{tr}_{#1}}
31: %\newcommand{\begin{equation}}{\begin{equation}}
32: %\newcommand{\end{equation}}{\end{equation}}
33: \newcommand{\DS}{\displaystyle}
34: \newtheorem{tw}{Theorem}[section]
35: \newtheorem{prop}{Proposition}[section]
36: \newtheorem{wn}{Conclusion}[section]
37: \newtheorem{df}{Definition}[section]
38: \newtheorem{uwaga}{Remark}[section]
39: \begin{document}
40: \begin{center}
41: \begin{LARGE}
42: \textbf{Entangling two-level atoms by spontaneous emission}
43: \end{LARGE}\\[12mm]
44: L. Jak{\'o}bczyk\\[2mm]
45: Institute of Theoretical Physics\\ University of
46: Wroc{\l}aw\\
47: Pl. M. Borna 9, 50-204 Wroc{\l}aw, Poland
48: \end{center}
49: \vskip 12mm
50: \noindent
51: {\sc Abstract}: It is shown that the dissipation due to
52: spontaneous emission can entangle two closely separated two-level
53: atoms.
54: \section{Introduction}
55: Analysis of various aspects of spontaneous emission by a system
56: of
57: two-level atoms, initiated in the classical paper of Dicke
58: \cite{Dicke} was further developed by several authors (see e.g.
59: \cite{dill, lemb1, lemb2}). In particular, in the case of
60: spontaneous emission by two trapped atoms separated by a
61: distance
62: small compared to the radiation wavelength, where is a
63: substantial probability that a photon emitted by one atom will be
64: absorbed by the other, there are states of the system in which
65: photon exchange can enhance or diminish spontaneous decay
66: rates.
67: The states with enhanced decay rate are called superradiant and
68: analogously states with diminished decay rate are called
69: subradiant \cite{Dicke}. It was also shown by Dicke, that the
70: system of two coupled two-level atoms can be treated as a single
71: four-level system with modified decay rates. Note also that such
72: model can be realized in a laboratory by two laser-cooled trapped
73: ions, where the observation of superradiance and subradiance is
74: possible \cite{DeVoe}.
75: \par
76: Another aspects of the model of the spontaneous emission
77: are studied in the present paper.
78: When the compound system of two atoms is in an entangled state,
79: the irreversible process of radiative decay usually destroys
80: correlations and the state becomes unentangled. In the model
81: studied here, the photon exchange produces correlations between
82: atoms which can partially overcome decoherence caused by
83: spontaneous radiation. As a result, some amount of entanglement
84: can survive, and moreover there is a possibility that this
85: process can entangle separable states of two atoms. The idea that
86: dissipation can create entanglement in physical
87: systems, was recently developed in several papers
88: \cite{plenio,kim,milb,kni}. In the present paper we show that the
89: dissipation due to spontaneous emission can entangle two atoms
90: that are initially prepared in a separable state. We study the
91: dynamics of this process. In the Markovian approximation it is
92: given by the semi-group $\{ T_{t} \}$ of completely positive
93: linear mappings acting on density matrices \cite{Alicki}. We
94: consider time evolution of initial state of the system as well as
95: the evolution of its entanglement, measured by so called
96: concurrence \cite{HW, W}, in the case when the photon exchange
97: rate $\ga$ is close to spontaneous emission rate of the single
98: atom $\g$ and we can use the approximation $\g=\ga$ (similar
99: model was also considered in \cite{bash}). We calculate
100: asymptotic stationary states $\ro_{\mathrm{as}}$ for the
101: semi-group $\{ T_{t} \}$ and show that they depend on initial
102: conditions (i.e. $\{ T_{t} \}$ is relaxing but not uniquely
103: relaxing). The concurrence of $\ro_{\mathrm{as}}$ also depends
104: on initial state and can be non zero for some of them. We discuss
105: in details some classes of initial states. In particular, we show
106: that there are pure separable states evolving to entangled mixed
107: states and such which remain separable during evolution. The first
108: class contains physically interesting initial state when one atom
109: in in excited state and the other is in ground state. The
110: relaxation process given by the semi-group $\{ T_{t} \}$ produces
111: in this case the states with entanglement monotonically
112: increasing in time to the maximal value. The class of pure
113: maximally entangled initial states is also discussed. Similar
114: "production" of entanglement is shown to be present for some
115: classes of mixed states. On the other hand, when the photon
116: exchange rate $\ga$ is
117: smaller then $\g$, the relaxation process brings all initial
118: states to the unique asymptotic state when both atoms are in its
119: ground states. As we show, even in that case the dynamics can
120: entangle two
121: separable states, but the amount of entanglement is decreasing to
122: zero.
123: \section{Pair of two-level atoms}
124: Consider two-level atom $A$ with ground state $\ket{0}$
125: and excited state $\ket{1}$. This quantum system can be described
126: in terms of the Hilbert space $\cH_{A}=\C^{2}$ and the algebra
127: $\fA_{A}$ of $2\times 2$ complex matrices. If we identify
128: $\ket{1}$ and $\ket{0}$ with vectors $\bigl( \begin{smallmatrix}
129: 1\\0
130: \end{smallmatrix}\bigr)$ and $\bigl( \begin{smallmatrix} 0 \\ 1
131: \end{smallmatrix}\bigr)$ respectively, then the raising and
132: lowering operators $\si{+},\;\si{-}$ defined by \begin{equation}
133: \si{+}=\ket{1}\bra{0},\quad \si{-}=\ket{0}\bra{1} \end{equation} can be
134: expressed in terms of Pauli matrices $\si{1},\; \si{2}$
135: \begin{equation}
136: \si{+}=\frac{1}{2}\,(\si{1}+i\,\si{2}),\quad
137: \si{-}=\frac{1}{2}\,(\si{1}-i\,\si{2}) \end{equation} For a joint system
138: $AB$
139: of two two-level atoms $A$ and $B$, the algebra $\fA_{AB}$ is
140: equal to $4\times 4$ complex matrices and the Hilbert space
141: $\cH_{AB}=\cH_{A}\otimes \cH_{B}=\C^{4}$. Let $\fE_{AB}$ be the
142: set of all states of the compound system i.e. \begin{equation}
143: \fE_{AB}=\{
144: \ro\in \fA_{AB}\, : \, \ro\geq 0\quad\text{and}\quad\tr \ro =1 \}
145: \end{equation} The state $\ro\in \fE_{AB}$ is \textit{separable}
146: \cite{Werner}, if it has the form \begin{equation}
147: \ro=\sum\limits_{k}\la_{k}\ro_{k}^{A}\otimes \ro_{k}^{B},\quad
148: \ro_{k}^{A}\in \fE_{A},\;\ro_{k}^{B}\in \fE_{B},\; \la_{k}\geq
149: 0\quad\text{and}\quad \sum\limits_{k}\la_{k}=1 \end{equation} The
150: set
151: $\fE_{AB}^{\,\rm sep}$ of all separable states forms a convex
152: subset of $\fE_{AB}$. When $\ro$ is not separable, it is called
153: \textit{inseparable} or \textit{entangled}. Thus \begin{equation}
154: \fE_{AB}^{\,\rm ent}=\fE_{AB}\setminus \fE_{AB}^{\,\rm sep}
155: \end{equation}
156: If $P\in \fE_{AB}$ is a pure state i.e. $P$ is one-dimensional
157: projector, then $P$ is separable iff partial traces $\ptr{A}P$ and
158: $\ptr{B}P$ are also projectors. For mixed states, the separability
159: problem is much more involved. Fortunately, in the case of $4$ --
160: level compound system there is a simple necessary and sufficient
161: condition for separability: $\ro$ is separable iff its partial
162: transposition $\ro^{T_{A}}$ is also a state \cite{HHH}. Another
163: interesting question is how to measure the amount of entanglement
164: a given quantum state contains. For a pure state $P$, the entropy
165: of entanglement \begin{equation} E(P)=-\tr[ (\ptr{A}P)\,\log_{2}\,
166: (\ptr{A}P)]
167: \end{equation} is essentailly a unique measure of entanglement
168: \cite{poproh}. For mixed state $\ro$ it seems that the basic
169: measure of entanglement is the entanglement of formation
170: \cite{Bennett} \begin{equation} E(\ro)=\min \,
171: \sum\limits_{k}\la_{k}E(P_{k})
172: \end{equation} where the minimum is taken over all possible
173: decompositions
174: \begin{equation} \ro=\sum\limits_{k}\la_{k}P_{k} \end{equation}
175: Again, in the case of
176: $4$ -- level system, $E(\ro)$ can be explicitely computed and it
177: turns out that $E(\ro)$ is the function of another useful
178: quantity $C(\ro)$ called \textit{concurrence}, which also can be
179: taken as a measure of entanglement \cite{HW, W}. Since in the
180: paper we use concurrence to quantify entanglement, now we
181: discuss
182: its definition. Let \begin{equation} \ro^{\dag}=(\si{2}\otimes
183: \si{2})\,\overline{\ro}\,(\si{2}\otimes \si{2}) \end{equation} where
184: $\overline{\ro}$ is the complex conjugation of the matrix $\ro$.
185: Define also \begin{equation}
186: \widehat{\ro}=(\ro^{1/2}\ro^{\dag}\ro^{1/2})^{1/2} \end{equation} Then
187: the
188: concurrence $C(\ro)$ is given by \cite{HW,W} \begin{equation}
189: C(\ro)=\max\;
190: (\,0, 2p_{\mathrm{max}}(\widehat{\ro})-\tr \widehat{\ro}\,)
191: \end{equation}
192: where $p_{\mathrm{max}}(\widehat{\ro})$ denotes the maximal
193: eigenvalue of $\widehat{\ro}$. The value of the number $C(\ro)$
194: varies from $0$ for separable states, to $1$ for maximally
195: entangled pure states.
196:
197: \section{Decay in a system of closely separated atoms}
198: We study the spontaneous emission of two atoms separated by a
199: distance $R$ small compared to the radiation wavelength . At such
200: distances there is a substantial probability that the photon
201: emitted by one atom will be absorbed by the other. Thus the
202: dynamics of the system is given by the master equation
203: \cite{Agar} \begin{equation} \frac{d\ro}{dt}=L\ro,\quad \ro\in \fE_{AB}
204: \end{equation}
205: with the following generator $L$ \begin{equation}
206: \begin{split}
207: L\ro=&\frac{\gamma_{0}}{2}\;[2\sa{-}{A}\;\ro\;\sa{+}{A}+2\sa{-
208: }{B}\;\ro\;\sa{+}{B}-
209: (\sa{+}{A}\;\sa{-}{A}+\sa{+}{B}\;\sa{-}{B})\;\ro-\ro\;
210: (\sa{+}{A}\;\sa{-}{A}+\sa{+}{B}\;\sa{-}{B})]+\\[2mm]
211: &\frac{\gamma}{2}\;[2\sa{-}{A}\;\ro\;\sa{+}{B}+2\sa{-
212: }{B}\;\ro\;\sa{+}{A}-(\sa{+}{A}\;
213: \sa{-}{B}
214: +\sa{+}{B}\;\sa{-}{A})\;\ro-
215: \ro \;(\sa{+}{A}\;\sa{-}{B}+\sa{+}{B}\;\sa{-}{A})]
216: \end{split}
217: \end{equation}
218: where
219: \begin{equation} \sa{\pm}{A}=\si{\pm}\otimes\I,\; \sa{\pm}{
220: B}=\I\otimes \si{\pm},\; \si{\pm}=\frac{1}{2}(\si{1}\pm i \si{2})
221: \end{equation}
222: Here $\gamma_{0}$ is the single atom spontaneous emission rate,
223: and $\gamma=g\gamma_{0}$ is a relaxation constant of photon
224: exchange. In the model, $g$ is the function of the distance $R$
225: between atoms and $g\to 1$ when $R\to 0$. In this section we
226: investigate the time evolution of the initial density matrix
227: $\ro$ of the compound system, governed by the semi - group
228: $\{T_{t}\}_{t\geq 0}$ generated by $L$. In particular, we will
229: study the time development of entanglement of $\ro$, measured by
230: concurrence.
231: \par
232: Assume that the distance between atoms is so small that the
233: exchange rate $\gamma$ is close to $\gamma_{0}$ and we can
234: use the
235: approximation $g=1$. Under this condition we study evolution of
236: the system and in particular we consider asymptotic states. Direct
237: calculations show that the semi - group $\{ T_{t} \}$ generated
238: by $L$ with $g=1$ is relaxing but not uniquely relaxing i.e.
239: there are as many stationary states as there are initial
240: conditions. More precisely, for a given initial state $\ro
241: =(\ro_{jk})$, the state $\ro(t)$ at time $t$ has the following
242: matrix elements
243: \begin{equation*}
244: \begin{split}
245: &\ro_{11}(t)=e^{-2\g t}\ro_{11}\\
246: &\ro_{12}(t)=\frac{1}{2}[e^{-2\g t}(\ro_{12}+\ro_{13})+e^{-
247: \g t}(\ro_{12}-\ro_{13})]\\
248: &\ro_{13}(t)=\frac{1}{2}[e^{-2\g t}(\ro_{12}+\ro_{13})+e^{-\g
249: t}(\ro_{13}-\ro_{12})]\\
250: &\ro_{14}(t)=e^{-\g t}\ro_{14}\\
251: &\ro_{22}(t)=\frac{1}{4}e^{-2\g t}(\ro_{22}+\ro_{33}+2\re
252: \ro_{23})+\frac{1}{2}e^{-\g t}(\ro_{22}-\ro_{33})+\g t e^{-2\g
253: t}\ro_{11}+\frac{1}{4}(\ro_{22}+\ro_{33}-2\re \ro_{23})\\
254: &\ro_{23}(t)=\frac{1}{4}e^{-2 \g t}(\ro_{22}+\ro_{33}+2\re
255: \ro_{23})+\frac{1}{2}e^{-\g t}(\ro_{23}-\ro_{32})+\g te^{-2\g
256: t}\ro_{11}-\frac{1}{4}(\ro_{22}+\ro_{33}-2\re \ro_{23})\\
257: &\ro_{24}(t)=-e^{-2\g
258: t}(\ro_{12}+\ro_{13})+\frac{1}{2}e^{-\g
259: t}(2\ro_{12}+2\ro_{13}+\ro_{24}+\ro_{34})+\frac{1}{2}(\ro_{24}-
260: \ro_{34})\\
261: &\ro_{33}(t)=\frac{1}{4}e^{-2\g t}(\ro_{22}+\ro_{33}+2\re
262: \ro_{23})-\frac{1}{2}e^{-\g t}(\ro_{22}-\ro_{33})+\g t e^{-2\g
263: t}\ro_{11}+\frac{1}{4}(\ro_{22}+\ro_{33}-2\re \ro_{23})\\
264: &\ro_{34}(t)=-e^{-2\g
265: t}(\ro_{12}+\ro_{13})+\frac{1}{2}e^{-\g
266: t}(2\ro_{12}+2\ro_{13}+\ro_{24}+\ro_{34})-\frac{1}{2}(\ro_{24}-
267: \ro_{34})\\
268: &\ro_{44}(t)=-\frac{1}{2}e^{-2 \g t}(1+\ro_{11}-\ro_{44}+2\re
269: \ro_{23})-2\g t e^{-2\g
270: t}\ro_{11}+\frac{1}{2}(1+\ro_{11}+\ro_{44}+2\re \ro_{23})
271: \end{split}
272: \end{equation*}
273: and remaining matrix elements can be obtained by hermiticity
274: condition $\ro_{kj}=\overline{\ro}_{jk}$. In the limit $t \to
275: \infty$ we obtain asymptotic (stationary) states parametrized as
276: follows \begin{equation} \ro_{\mathrm{as}}=\begin{pmatrix}
277: 0&\hspace{2mm}0&\hspace{2mm}0&\hspace{2mm}0\\
278: 0&\hspace{2mm}\al&-\al&\hspace{2mm}\be\\
279: 0&-\al&\hspace{2mm}\al&-\be\\
280: 0&\hspace{2mm}\overline{\be}&-\overline{\be}&1-2\al
281: \end{pmatrix}
282: \end{equation}
283: where
284: \begin{equation}
285: \al=\frac{1}{4}(\ro_{22}+\ro_{33}-2\re \ro_{23}),\quad
286: \be=\frac{1}{2}(\ro_{24}-\ro_{34})
287: \end{equation}
288: We can also compute concurrence of the asymptotic state and
289: the result is:\\[2mm]
290: \textit{
291: Concurrence of asymptotic state of the semi
292: - group $\{ T_{t} \}$ generated by $L$ with $g=1$ equals to
293: \begin{equation}
294: C(\ro_{\mathrm{as}})=2|\al|=\frac{1}{2}|\ro_{22}+\ro_{33}-2\re \ro_{23}|
295: \end{equation}
296: where $\ro_{jk}$ are the matrix elements of the initial
297: state.}\\[2mm]
298: \section{Some examples}
299: \noindent In this section we consider examples of initial states
300: and its evolution.\\[2mm]
301: \textbf{I. Pure separable states}\\[2mm]
302: Let
303: \begin{equation}
304: \ro=P_{\Psi\otimes \Phi}=P_{\Psi}\otimes P_{\Phi}
305: \end{equation}
306: where
307: $$
308: \Psi=\begin{pmatrix}
309: \Psi_{1}\\
310: \Psi_{2}
311: \end{pmatrix}
312: \in \cH_{A},\quad
313: \Phi=\begin{pmatrix}
314: \Phi_{1}\\
315: \Phi_{2}
316: \end{pmatrix}\in \cH_{B}
317: $$
318: are normalized. Then one can check that
319: \begin{equation}
320: \begin{split}
321: \al&=\frac{1}{4}(1-|\ip{\Psi}{\Phi}|^{2})\\
322: \be&=\frac{1}{2}(|\Phi_{2}|^{2}\Psi_{1}\overline{\Psi}_{2}-|\Psi_{2}|^{2}
323: \Phi_{1}\overline{\Phi}_{2})
324: \end{split}
325: \end{equation}
326: where $\ip{\cdot}{\cdot}$ is the inner product in $\C^{2}$. So
327: \begin{equation}
328: C(\ro_{\mathrm{as}})=\frac{1}{2}(1-|\ip{\Psi}{\Phi}|^{2})
329: \end{equation}
330: From the formula (21) we see that there are separable initial
331: states for which asymptotic states are entangled. In particular,
332: the asymptotic state has a maximal concurrence if vectors $\Psi$
333: and $\Phi$ are orthogonal and its concurrence is zero (the state
334: remains separable) if $|\ip{\Psi}{\Phi}|=1$.
335: \par
336: \noindent
337: Now we discuss some special cases.\\
338: \textbf{a.}When one atom is in excited state and the other is in
339: ground
340: state
341: $$
342: \Psi=\ket{1},\quad \Phi=\ket{0}
343: $$
344: the asymptotic (mixed) state is given by
345: $$
346: \ro_{\mathrm{as}}=\begin{pmatrix}
347: 0&\hspace{2mm}0&\hspace{2mm}0&0\\[2mm]
348: 0&\hspace{2mm}\frac{1}{4}&-\frac{1}{4}&0\\[2mm]
349: 0&-\frac{1}{4}&\hspace{2mm}\frac{1}{4}&0\\[2mm]
350: 0&0&0&0
351: \end{pmatrix}
352: $$
353: It can also be shown that in this case the relaxation process
354: produces the state $\ro_{t}$ with concurrence
355: $$
356: C(\ro_{t})=\frac{1-e^{-\g t}}{2}
357: $$
358: increasing to the maximal value equal to $1/2$. Thus two atoms
359: initially in separable state become entangled for all $t$ and the
360: asymptotic (steady) state attains the maximal amount of
361: entanglement.\\
362: \textbf{b.} When two atoms are in excited states
363: $$
364: \Psi=\Phi=\ket{1}
365: $$
366: the asymptotic state equals to
367: $$
368: \ket{0}\otimes \ket{0}
369: $$
370: Thus the relaxation process brings two atoms into ground
371: states.\\
372: \textbf{c.} The state $\ket{0}\otimes\ket{0}$ is stationary state
373: for semi - group $\{ T_{t} \}$.\\[4mm]
374: \textbf{II. Pure maximally entangled states}\\[2mm]
375: Let
376: \vskip 2mm
377: \noindent
378: $$
379: \ro=Q(a,\theta_{1},\theta_{2})=\begin{pmatrix}
380: \frac{a^{2}}{2}&
381: \frac{a\sqrt{1-a^{2}}}{2}e^{-i\theta_{1}}&
382: \frac{a\sqrt{1-a^{2}}}{2}e^{-i\theta_{2}}&
383: -\frac{a^{2}}{2}e^{-i(\theta_{1}+\theta_{2})}\\[2mm]
384: \frac{a\sqrt{1-a^{2}}}{2}e^{i\theta_{1}}&\frac{1-a^{2}}{2}&
385: \frac{1-a^{2}}{2}e^{i(\theta_{1}-\theta_{2})}&-\frac{a\sqrt{1-a^{2}}}{2}e^{-
386: i\theta_{2}}\\[2mm]
387: \frac{a\sqrt{1-a^{2}}}{2}e^{i\theta_{2}}&\frac{1-a^{2}}{2}e^{-i(\theta_{1}-
388: \theta_{2})}&
389: \frac{1-a^{2}}{2}&-\frac{a\sqrt{1-a^{2}}}{2}e^{-i\theta_{1}}\\[2mm]
390: -\frac{a^{2}}{2}e^{i(\theta_{1}+\theta_{2})}&-\frac{\sqrt{1-
391: a^{2}}}{2}e^{i\theta_{2}}&
392: -\frac{a\sqrt{1-a^{2}}}{2}e^{i\theta_{1}}&\frac{a^{2}}{2}
393: \end{pmatrix}
394: $$
395: \vskip 2mm
396: \noindent
397: where $ a\in [0,1],\; \theta_{1},\theta_{2}\in
398: [0,2\pi]$. Pure states $Q(a,\theta_{1},\theta_{2})$ are maximally
399: entangled and form a family of all maximally entangled states of
400: the $4$ - level system \cite{BJO1}. It turns out that
401: $\ro_{\mathrm{as}}$ is defined by
402: \begin{equation}
403: \begin{split}
404: \al&=\frac{1}{4}(1-a^{2})(1-\cos (\theta_{1}-\theta_{2}))\\
405: \be&=\frac{1}{4}a\sqrt{1-a^{2}}(e^{-i\theta_{1}}-e^{-i\theta_{2}})
406: \end{split}
407: \end{equation} and \begin{equation}
408: C(\ro_{\mathrm{as}})=\frac{1}{2}(1-a^{2})(1-\cos
409: (\theta_{1}-\theta_{2})) \end{equation} From the formula (23) we see
410: that
411: there are initial maximally entangled states which asymptotically
412: become separable ($a=1$ or $\theta_{1}-\theta_{2}=2k\pi$) and
413: such that the asymptotic concurrence is greater then $0$. States
414: with $a=0$ and $\theta_{1}-\theta_{2}=(2k+1)\pi$ remain maximally
415: entangled. For example the state \begin{equation}
416: \frac{1}{\sqrt{2}}(\ket{0}\otimes\ket{1}-\ket{1}\otimes\ket{0})
417: \end{equation} is stable. On the other hand, the concurrence of
418: \begin{equation}
419: \frac{1}{\sqrt{2}}(\ket{0}\otimes\ket{1}+\ket{1}\otimes\ket{0})
420: \end{equation} goes to zero faster then the concurrence of
421: \begin{equation}
422: \frac{1}{\sqrt{2}}(\ket{0}\otimes\ket{0}+\ket{1}\otimes\ket{1})
423: \end{equation} as shown on \textbf{Fig. 1.} below. In Dicke's theory
424: of
425: spontaneous radiation processes the state (24) is called
426: subradiant whereas the state (25) has half the lifetime of a
427: single atom and therefore is called superradiant \cite{Dicke}. We
428: see that the time-dependence of concurrence reflects the
429: relaxation properties of those states. \vskip 8mm \noindent
430: %\centerline{\epsffile{rys1.eps}}
431: \begin{picture}(300,300)
432: \put(50,70){\begin{picture}(180,180) \epsffile{rys1.eps}
433: \end{picture}}
434: \put(350,75){$\g t$} \put(60,255){$C(\ro_{t})$}
435: \end{picture}
436: \vskip -18mm \noindent \centerline{\textbf{Fig. 1.} Concurrence as
437: the function of time for initial states:
438: $\frac{1}{\sqrt{2}}(\ket{0}\otimes\ket{0}+\ket{1}\otimes\ket{1})$}
439: \hspace*{20mm} (dotted line)
440: %\hspace*{9mm}
441: and
442: $\frac{1}{\sqrt{2}}(\ket{0}\otimes\ket{1}+\ket{1}\otimes\ket{0})$
443: (solid line).
444: %\vskip 8mm
445: %\noindent
446: %\centerline{\epsffile{rys2.eps}}
447: \vskip 6mm
448: \noindent
449: \textbf{III. Some classes of mixed states}\\[2mm]
450: \textbf{a.} Bell - diagonal states. Let \begin{equation}
451: \ro_{\mathrm{B}}=p_{1}\ket{\Phi^{+}}\bra{\Phi^{+}}+p_{2}\ket{\Phi^{-
452: }}\bra{\Phi^{-}}+p_{3}
453: \ket{\Psi^{+}}\bra{\Psi^{+}}+p_{4}\ket{\Psi^{-}}\bra{\Psi^{-}}
454: \end{equation} where Bell states $\Phi^{\pm}$ and $\Psi^{\pm}$ are
455: given by
456: \begin{equation}
457: \Phi^{\pm}=\frac{1}{\sqrt{2}}(\ket{0}\otimes\ket{0}\,\pm\,
458: \ket{1}\otimes\ket{1}),\quad
459: \Psi^{\pm}=\frac{1}{\sqrt{2}}(\ket{1}\otimes\ket{0}\,\pm\,
460: \ket{0}\otimes\ket{1}) \end{equation} It is known that all $p_{i}\in
461: [0,1/2],\; \ro_{\mathrm{B}}$ is separable, while for
462: $p_{1}>1/2,\; \ro_{\mathrm{B}}$ is entangled with concurrence
463: equal to $2p_{1}-1$ (similarly for $p_{2},\,p_{3}\, p_{4}$)
464: \cite{HHHa}. Now the asymptotic state has the form
465: \begin{equation}
466: \ro_{\mathrm{as}}=\begin{pmatrix}
467: 0&\hspace{2mm}0&\hspace{2mm}0&0\\[2mm]
468: 0&\hspace{2mm}\frac{p_{4}}{2}&-\frac{p_{4}}{2}&0\\[2mm]
469: 0&-\frac{p_{4}}{2}&\hspace{2mm}\frac{p_{4}}{2}&0\\[2mm]
470: 0&\hspace{2mm}0&\hspace{2mm}0&1-p_{4}
471: \end{pmatrix}
472: \end{equation}
473: with concurrence $C(\ro_{\mathrm{as}})=p_{4}$. So even when the
474: initial state is separable, the asymptotic state becomes
475: entangled. \\[2mm]
476: \textbf{b.} Werner states \cite{BBPSSW}. Let
477: \begin{equation}
478: \ro_{\mathrm{W}}=(1-p)\frac{\I_{4}}{4}+p\ket{\Phi^{+}}\bra{\Phi^{+}}
479: \end{equation}
480: If $p>1/3,\; \ro_{\mathrm{W}}$ is entangled with concurrence equal
481: to $(3p-1)/2$. On the other hand
482: \begin{equation}
483: \ro_{\mathrm{as}}=\begin{pmatrix}
484: 0&0&0&0\\[2mm]
485: 0&\frac{1-p}{8}&\frac{p-1}{8}&0\\[2mm]
486: 0&\frac{p-1}{8}&\frac{1-p}{8}&0\\[2mm]
487: 0&0&0&\frac{3+p}{4}\\[2mm]
488: \end{pmatrix}
489: \end{equation}
490: has the concurrence $C(\ro_{\mathrm{as}})=\frac{1-p}{4}$, so the
491: asymptotic states are entangled for all $p\neq 1$. Notice that
492: even completely mixed state $\frac{\I_{4}}{4}$ evolves to
493: entangled asymptotic state.\\[2mm]
494: \textbf{c.} Maximally entangled mixed states . The
495: states
496: \begin{equation}
497: \ro_{\mathrm{M}}=\begin{pmatrix}
498: h(\delta)&0&0&\delta/2\\[2mm]
499: 0&1-2h(\delta)&0&0\\[2mm]
500: 0&0&0&0\\[2mm]
501: \delta/2&0&0&h(\delta)
502: \end{pmatrix},\quad h(\delta)=\begin{cases}
503: 1/3& \delta\in [0,2/3]\\
504: \delta/2& \delta\in [2/3,1]
505: \end{cases}
506: \end{equation}
507: are conjectured to be maximally entangled for a given degree
508: of inpurity measured by $\tr \ro^{2}$ \cite{Kwiat}. According to
509: (18) the concurrence of the asymptotic state is given by
510: \begin{equation}
511: C(\ro_{\mathrm{as}})=\frac{1}{2}(1-2h(\delta))
512: \end{equation}
513: \vskip 8mm
514: \noindent
515: %\centerline{\epsffile{rr3.eps}}
516: \begin{picture}(300,300)
517: \put(50,70){\begin{picture}(180,180) \epsffile{rr3.eps}
518: \end{picture}}
519: \put(350,75){$\tr \ro_{\mathrm{M}}^{2}$}
520: \put(60,255){$C(\ro_{t})$}
521: \end{picture}
522: \vskip -18mm \noindent \centerline{\textbf{Fig. 2.} Concurrence of
523: $\ro_{\mathrm{M}}$ (solid line) and $\ro_{\mathrm{as}}$ (dotted
524: line) as the function of $\tr \ro_{\mathrm{M}}^{2}$} \vskip 6mm
525: \noindent Even in that case, there are initial states (for
526: sufficiently small $\tr \ro_{\mathrm{M}}^{2}$) such that the
527: asymptotic state is more entangled (see \textbf{Fig. 2.}).
528: \section{Remarks on general case}
529: In the case of arbitrary distance between the atoms i.e. when
530: $g\in [0,1)$, semi-group generated by $L$ is uniquely relaxing,
531: with the asymptotic state $\ket{0}\otimes\ket{0}$. Thus, for any
532: initial state $\ro$, the concurrence $C(\ro_{t})$ approaches $0$
533: when $t\to \infty$. But it does not mean that the function $t\to
534: C(\ro_{t})$ is always monotonic. The general form of $C(\ro_{t})$
535: is rather involved, so we consider only some special cases.\\[2mm]
536: \textbf{1.} Let the initial state of the compound system be equal
537: to $\ket{0}\otimes\ket{1}$. This states evolves to \vskip 4mm
538: \noindent
539: \begin{equation}
540: \ro_{t}=\begin{pmatrix} 0&0&0&0\\[2mm]
541: 0&\frac{1}{2}e^{-\g t}\,(\cosh \ga t +1)&-\frac{1}{2}e^{-\g
542: t}\sinh \ga t&0\\[2mm]
543: 0&-\frac{1}{2}e^{-\g t}\sinh \ga t& \frac{1}{2}e^{-\g t}\, (\cosh
544: \ga t -1)&0\\[2mm]
545: 0&0&0&1-e^{-\g t}\cosh \ga t
546: \end{pmatrix}
547: \end{equation} \vskip 4mm \noindent with concurrence
548: \begin{equation} C(\ro_{t})=e^{-\g
549: t}\sinh \ga t \end{equation} In the interval $[0,t_{\ga}]$, where
550: $$
551: t_{\gamma}=\frac{1}{2\ga}\ln \frac{\g +\ga}{\g -\ga}
552: $$
553: the function (34) is increasing to its maximal value
554: $$
555: C_{\mathrm{max}}=\frac{\ga}{\g-\ga}\,
556: \left(\frac{\g+\ga}{\g-\ga}\right)^{-\frac{\DS\g+\ga}{\DS 2\ga}}
557: $$
558: whereas for
559: $t>t_{\ga}$, $C(\ro_{t})$ decreases to $0$. Thus for any nonzero
560: photon exchange rate $\ga$, dynamics given by the semi - group
561: $\{ T_{t} \}$ produces some amount of entanglement between two
562: atoms which are initially in the ground state and excited state.
563: Note that the maximal value of $C(\ro_{t})$ depends only on
564: emission rates $\g$ and $\ga$.\\[2mm]
565: %\vskip 8mm
566: %\noindent
567: %\centerline{\epsffile{rys3.eps}}
568: %\vskip 8mm
569: %\noindent
570: %\centerline{\textbf{Fig. 3.} $C(\ro_{t})$ for
571: %$\ga/\g=0.2,\,0.4,\,0.6,\,0.8,\,0.9,\,0.99$}
572: %\vskip 8mm
573: %\noindent
574: \textbf{2.} For the initial states
575: $$
576: \Psi^{\pm}=\frac{1}{\sqrt{2}}\,(\ket{0}\otimes\ket{1}\pm
577: \ket{1}\otimes\ket{0})
578: $$
579: the relaxation to the asymptotic state $\ket{0}\otimes\ket{0}$ is
580: given by density matrices \vskip 4mm\noindent
581: \begin{equation}
582: \ro_{t}^{\pm}=\begin{pmatrix} 0&0&0&0\\[2mm]
583: 0&\frac{1}{2}e^{-(\g \pm \ga)t}&-\frac{1}{2}e^{-(\g\pm
584: \ga)t}&0\\[2mm]
585: 0&-\frac{1}{2}e^{-(\g\pm\ga)t)}&\frac{1}{2}e^{-(\g\pm\ga)t}&0\\[2mm]
586: 0&0&0&1-e^{-(\g\pm\ga)t}
587: \end{pmatrix}
588: \end{equation}
589: \vskip 4mm\noindent with the corresponding concurrence
590: $$
591: C(\ro_{t}^{\pm})=e^{-(\g\pm\ga)t}
592: $$
593: The state $\Psi^{-}$ is no longer stable (as in the case of
594: $\ga=\g$), but during the evolution its concurrence goes to zero
595: slower than $C(\ro^{+}_{t})$ (\textbf{Fig. 3.}). For $\ga $ close
596: to $\g$, $\Psi^{-}$ is almost stable.
597: %\newpage
598: \vskip 8mm
599: \noindent
600: %\centerline{\epsffile{Rys4.eps}}
601: \begin{picture}(300,300)
602: \put(50,70){\begin{picture}(180,180) \epsffile{rys4.eps}
603: \end{picture}}
604: \put(350,75){$\g t$} \put(60,255){$C(\ro_{t})$}
605: \end{picture}
606: \vskip -18mm
607: \noindent
608: \centerline{\textbf{Fig. 3.}
609: $C(\ro^{+}_{t})$ (dotted line) and $C(\ro^{-}_{t})$ (solid line)
610: for $\ga/\g = 0.99$} \vskip 8mm \noindent
611: \begin{thebibliography}{99}
612: \bibitem{Dicke} R.H. Dicke, Phys. Rev. \textbf{93}, 99(1954).
613: \bibitem{dill} M. Dillard, H.R. Robl, Phys. Rev. \textbf{184},
614: 312(1969).
615: \bibitem{lemb1} R.H. Lehmberg, Phys. Rev. \textbf{181}, 32(1969).
616: \bibitem{lemb2} R.H. Lehmberg, Phys. Rev. \textbf{A 2}, 883(1970);
617: \textbf{A 2}, 889(1970).
618: \bibitem{DeVoe} R.G. DeVoe, R.G. Brewer, Phys. Rev. Lett.
619: \textbf{76}, 2046(1996).
620: \bibitem{plenio} M.B. Plenio, S.F. Huelga, Entangled light from
621: white noise, arXiv: quant-ph/0110009.
622: \bibitem{kim}M.S. Kim, J. Lee, D. Ahn, P.L. Knight, Phys. Rev.
623: A\textbf{65}, 040101(R)(2002).
624: \bibitem{milb} S. Schneider, G.J. Milburn, Phys. Rev. A\textbf{65},
625: 042107(2002).
626: \bibitem{kni}M.B. Plenio, S.F. Huelga, A. Beige, P.L. Knight,
627: Phys. Rev. A\textbf{59}, 2468(1999).
628: \bibitem{Alicki} R. Alicki, K. Lendi, \textit{Quantum Dynamical
629: Semigroups and Aplications}, Lecture Notes in Phys. Vol 286,
630: Springer, Berlin, 1987.
631: \bibitem{HW} S. Hill, W.K. Wootters, Phys. Rev. Lett. \textbf{78},
632: 5022(1997).
633: \bibitem{W} W.K. Wootters, Phys. Rev. Lett. \textbf{80},
634: 2254(1998).
635: \bibitem{bash} A.M. Basharov, JETP Letters, \textbf{75},
636: 123(2002).
637: \bibitem{Werner} R.F. Werner, Phys. Rev. A\textbf{40}, 4277(1989).
638: \bibitem{HHH}M. Horodecki, P. Horodecki, R. Horodecki, Phys.
639: Lett. \textbf{A 223}, 1(1996).
640: \bibitem{poproh} S. Popescu, D. Rohrlich, Phys. Rev. \textbf{A 56},
641: R3319(1997).
642: \bibitem{Bennett} Ch. Bennett, P.D. DiVincenzo, J. Smolin, W.K.
643: Wootters, Phys. Rev. \textbf{A 54}, 3824(1996).
644: \bibitem{Agar} G.S. Agarwal, \textit{Quantum Statistical Theories
645: of Spontaneous Emission and their Relation to Other Approaches},
646: Springer, Berlin, 1974.
647: \bibitem{BJO1} Ph. Blanchard, L. Jak{\'o}bczyk, R. Olkiewicz,
648: Phys.
649: Lett. A \textbf{280}, 7(2001).
650: \bibitem{HHHa} R. Horodecki, M. Horodecki, Phys. Rev. \textbf{A
651: 54}, 1838(1996).
652: \bibitem{BBPSSW} Ch. H. Bennett, G. Brassard, S. Popescu, R.
653: Schumacher,
654: J.A. Smolin, W.K. Wootters, Phys. Rev. Lett. \textbf{76},
655: 722(1996).
656: \bibitem{Kwiat}W.J. Munro, D.F.V. James, A.G. White, P.G. Kwiat,
657: Phys. Rev. A\textbf{64}, 030302(2001).
658: \end{thebibliography}
659: \end{document}
660:
661:
662: