1: \documentclass[twocolumn,showpacs,superscriptaddress]{revtex4}
2:
3: \usepackage{amssymb}
4: \usepackage{amsfonts}
5: \usepackage{amsmath}
6: \usepackage{graphicx}
7: \usepackage{mathrsfs}
8:
9: \begin{document}
10:
11: \title{Entanglement oscillation and survival induced by non-Markovian decoherence dynamics of entangled squeezed-state}
12: \author{Jun-Hong An}
13: \email{phyaj@nus.edu.sg}
14: \affiliation{Department of Modern Physics, Lanzhou University, Lanzhou 730000, P. R. China}
15: \affiliation{Centre for Quantum Technologies and Department of Physics, National University of Singapore, 3 Science Drive 2, Singapore 117543, Singapore}
16: \author{Ye Yeo}
17: \affiliation{Centre for Quantum Technologies and Department of Physics, National University of Singapore, 3 Science Drive 2, Singapore 117543, Singapore}
18: \author{Wei-Min Zhang}
19: \affiliation{Department of Physics and Center for Quantum Information Science, National Cheng Kung University, Tainan 70101, Taiwan}
20: \author{C. H. Oh}
21: \email{phyohch@nus.edu.sg}
22: \affiliation{Centre for Quantum Technologies and Department of Physics, National University of Singapore, 3 Science Drive 2, Singapore 117543, Singapore}
23: \begin{abstract}
24: We study the exact decoherence dynamics of the entangled squeezed state of two single-mode optical fields interacting with two independent and
25: uncorrelated environments. We analyze in detail the non-Markovian effects on the entanglement evolution of the initially entangled squeezed state for
26: different environmental correlation time scales. We find that the environments have dual actions on the system: backaction and dissipation. In
27: particular, when the environmental correlation time scale is comparable to the time scale for significant change in the system, the backaction would
28: counteract the dissipative effect. Interestingly, this results in the survival of some residual entanglement in the final steady state.
29: \end{abstract}
30: \pacs{03.65.Yz, 03.67.Mn, 03.67.-a} \maketitle
31:
32: \section{Introduction}
33: Studies on the decoherence dynamics of open quantum systems are of great importance to the field of quantum information science \cite{Nielsen00}. Any
34: realistic analysis of quantum information protocols should take into account the decoherence effect of the environment. In many quantum communication
35: and computation schemes, information is transmitted using photons. For instance, the first experimental verification of quantum teleportation
36: \cite{Bennett} used pairs of polarization entangled photons to transfer the polarization state of one photon onto another \cite{Bouwmeester97}.
37: Within a year, unconditional quantum teleportation of optical fields was demonstrated experimentally using squeezed-state entanglement
38: \cite{Braunstein98, Furusawa98}. Given their central role in these schemes and many others, much work has been carried out on the decoherence dynamics
39: of optical fields. In particular, several authors have studied the continuous variable entanglement of optical fields (see, for instance,
40: Refs. \cite{Jakub04, An05, Rossi06, Ban06, Goan07, Paris07, An07}).
41:
42: Conventional approaches not only treat the interactions between the quantum system $S$ of interest and its environment $E$ perturbatively, they also
43: assume that the environmental correlation time $\tau_E$ is small compared to the time scale $\tau_0$ for significant change in $S$. These yield
44: approximate equations of motion, i.e. master equations, under the Born-Markov approximation \cite{Carmichael93, Breuer02}. Indeed, many studies on the
45: entanglement dynamics of continuous variable system relied on this approximation \cite{Jakub04, An05, Rossi06}. However, it is evident from recent
46: experiments (see, for instance, Refs. \cite{Dubin07, Koppens07, Mogilevtsev08}), that there are many physically relevant situations where the Markovian
47: assumption does not hold, and a non-Markovian treatment of the open system dynamics is necessary. So, there has been an increasing interest in the
48: understanding of the decoherence effect of open quantum system going beyond the Born-Markovian approximation in the last decades \cite{Breuer02, Weiss}.
49:
50: Very recently, some phenomenological models on non-Markovian entanglement dynamics of optical fields have been investigated \cite{Ban06, Goan07, Paris07}.
51: It was found that in contrast to the monotonic decrease of entanglement over time in Born-Markovian entanglement dynamics \cite{Jakub04, An05, Rossi06},
52: there are transient entanglement oscillations in non-Markovian ones. These oscillations are caused by the backactions of the environments on their
53: respective local quantum systems \cite{Goan07, Paris07}. The backaction, characteristic of non-Markovian dynamics, means that the environments with
54: their states changed due to interactions with the systems, in turn, exert their dynamical influences back on the systems.
55:
56: In this paper we consider the exact decoherence dynamics of the continuous variable entangled squeezed state of two single-mode optical fields, $S_1$
57: and $S_2$, that are spatially separated. Each optical field, $S_k$, interacts with its own environment $E_k$ ($k = 1, 2$). $E_1$ and $E_2$ are
58: independent and uncorrelated. We study the exact entanglement dynamics of the two optical-field system for different $\tau_E$'s in comparsion with
59: $\tau_0$, and analyze when the system dynamics will exhibit novel non-Markovian effects, and provide a detailed description of these. To this end, we
60: use the influence functional formalism \cite{Feynman63, Leggett87}, developed explicitly in Refs. \cite{An07, An072, An08}. Our results show that besides
61: the short-time oscillations, the non-Markovian effect can affect the long-time behavior of the system dynamics and the steady state as well. In
62: particular, when $\tau_E$ is comparable to $\tau_0$, we find that the backaction effects counteract the dissipative effects of $E_1E_2$ on $S_1S_2$
63: respectively. This leads to there being some nonzero residual entanglement in the steady state.
64:
65: Our paper is organized as follows. In Sec. II, we introduce a model of two single-mode optical fields in two independent and uncorrelated environments;
66: and outline the exact dynamics that was derived in detail in Ref. \cite{An08}. In Sec. III, using logarithmic negativity as an entanglement measure of
67: continuous-variable states, we discuss the entanglement dynamics of the entangled squeezed state. Sec. IV presents the numerical results of the
68: entanglement dynamics, where we analyze explicitly the non-Markovian effect of the environments on the system for different $\tau_E$'s in comparison
69: with $\tau_0$. Finally, we conclude in Sec. V.
70:
71: \section{The total Hamiltonian and exact reduced system decoherence dynamics}
72: The total Hamiltonian of the system $S_1S_2$ plus environment $E_1E_2$ is given by
73: \begin{equation}
74: H = H_{\mathrm{S}} + H_{\mathrm{E}} + H_{\mathrm{I}},
75: \end{equation}
76: where
77: \begin{eqnarray}
78: H_{\mathrm{S}} & = & \sum_{k = 1}^{2}\hbar\omega_ka_k^{\dag}a_k, \nonumber \\
79: H_{\mathrm{E}} & = & \sum_{k = 1}^{2}\sum_{l}\hbar\omega_{kl}b_{kl}^{\dag}b_{kl}, \nonumber \\
80: H_{\mathrm{I}} & = & \sum_{k = 1}^{2}\sum_{l}\hbar(g_{kl}a^{\dag}_kb_{kl} + g_{kl}^{\ast}a_kb_{kl}^{\dag}),
81: \end{eqnarray}
82: are, respectively, the Hamiltonian of the two optical fields, the two independent environments, and the interactions between them. The operators $a_k$
83: and $a^{\dag}_k$ ($k = 1, 2$) are respectively the annihilation and creation operators of the $k$-th optical mode with frequency $\omega_k$. The two
84: independent environments are modeled, as usual, by two sets of harmonic oscillators described by the annihilation and creation operators $b_{kl}$ and
85: $b^{\dag}_{kl}$. The coupling constants between the $k$-th optical field and its environment are given by $g_{kl}$. Currently, most quantum optical
86: experiments are performed at low temperatures and under vacuum condition. In this case, vacuum fluctuations are the main source of decoherence.
87: Therefore, we take the environments to be at zero temperature throughout this paper.
88:
89: Since we are only interested in the dynamics of $S_1S_2$, we like to eliminate the degrees of freedom of $E_1E_2$. The influence-functional theory of
90: Feynman and Vernon \cite{Feynman63} enables us to do that exactly. By expressing the forward and backward evolution operators of the density matrix of
91: the system $S_1S_2$ plus environment $E_1E_2$ as a double path integral in the coherent-state representation \cite{Zhang90}, and performing the
92: integration over the degrees of freedom of $E_1E_2$, we incorporate all the environmental effects on $S_1S_2$ in a functional integral named influence
93: functional \cite{Feynman63, An07, An08}. The reduced density matrix, which fully describes the dynamics of $S_1S_2$ is given by
94: \begin{eqnarray}\label{rout}
95: \rho (\boldsymbol{\bar{\alpha}}_{f},\boldsymbol{\alpha }_{f}^{\prime };t)
96: &=&\int d\mu (\boldsymbol{\alpha }_{i})d\mu (\boldsymbol{\alpha }%
97: _{i}^{\prime })\mathcal{J}(\boldsymbol{\bar{\alpha}}_{f},\boldsymbol{\alpha }%
98: _{f}^{\prime };t|\boldsymbol{\bar{\alpha}}_{i},\boldsymbol{\alpha }%
99: _{i}^{\prime };0) \notag \\
100: &&~~~~~~~~~\times \rho (\boldsymbol{\bar{\alpha}}_{i},\boldsymbol{\alpha }%
101: _{i}^{\prime };0),
102: \end{eqnarray}%
103: where $\rho(\boldsymbol{\bar{\alpha}}_f,\boldsymbol{\alpha}^{\prime}_f; t) =
104: \langle\boldsymbol{\alpha}_f|\rho(t)|\boldsymbol{\alpha}^{\prime}_f\rangle$ is the reduced density matrix expressed in coherent-state
105: representation and
106: $\mathcal{J}(\boldsymbol{\bar{\alpha}}_f, \boldsymbol{\alpha}^{\prime}_f; t|\boldsymbol{\bar{\alpha}}_i, \boldsymbol{\alpha }^{\prime}_i; 0)$
107: is the propagating function. In the derivation of Eq. (\ref{rout}), we have used the coherent-state representation
108: \begin{equation}
109: |\boldsymbol{\alpha}\rangle = \prod_{k = 1}^2|\alpha_k\rangle, ~|\alpha_k\rangle = \exp (\alpha _ka^{\dagger}_k)|0_k\rangle.
110: \end{equation}
111: which are the eigenstates of annihilation operators, i.e.
112: $a_k|\alpha_k\rangle = \alpha_k|\alpha_k\rangle$ and obey the
113: resolution of identity, $\int d\mu \left( \boldsymbol{\alpha
114: }\right) |\boldsymbol{\alpha }\rangle \langle \boldsymbol{\alpha }|
115: = 1$ with the integration measures defined as
116: $d\mu\left(\boldsymbol{\alpha}\right) =
117: \prod_{l}e^{-\bar{\alpha}_{l}\alpha_{l}}\frac{d\bar{\alpha}
118: _{l}d\alpha _{l}}{2\pi i}$. $\bar{\boldsymbol{\alpha}}$ denotes the
119: complex conjugate of $\boldsymbol{\alpha}$.
120:
121: The time evolution of the reduced density matrix is determined by the propagating function
122: $\mathcal{J}(\boldsymbol{\bar{\alpha}}_f, \boldsymbol{\alpha}^{\prime}_f; t|\boldsymbol{\bar{\alpha}}_i, \boldsymbol{\alpha}^{\prime}_i; 0)$. The
123: propagating function is expressed as the path integral governed by an effective action which consists of the free actions of the forward and backward
124: propagators of the optical-field system and the influence functional obtained from the integration of environmental degrees of freedom. After
125: evaluation of the path integral, the final form of the propagating function is obtained as follows
126: \begin{eqnarray}
127: & & \mathcal{J}(\boldsymbol{\bar{\alpha}}_{f},\boldsymbol{\alpha }%
128: _{f}^{\prime};t|\boldsymbol{\bar{\alpha}}_{i},\boldsymbol{\alpha }%
129: _{i}^{\prime };0)=\exp \Big\{\sum_{k=1}^{2}\big[u_k(t)\bar{\alpha}%
130: _{kf}\alpha _{ki} \notag \\
131: &&~~~~~+\bar{u}_k(t)\bar{\alpha}_{ki}^{\prime }\alpha _{kf}^{\prime
132: }+[1-\left\vert u_k(t)\right\vert ^2]\bar{\alpha}_{ki}^{\prime }\alpha _{ki}%
133: \big]\Big\}, \label{prord}
134: \end{eqnarray}
135: where $u_k(\tau)$ satisfies
136: \begin{equation}\label{ut}
137: \dot{u}_k(\tau) + i\omega_k u_k(\tau) + \int^{\tau}_0 \mu_k(\tau - \tau^{\prime})u_k(\tau^{\prime}) = 0
138: \end{equation}
139: with $\mu_k(x) \equiv \sum_le^{-i\omega_lx}\left\vert g_{kl}\right\vert^2$ being used. Combining Eq. (\ref{prord}), we can get the exact time-dependent
140: state from any initial state by the evaluation of the integration in Eq. (\ref{rout}).
141:
142: To compare with the conventional master equation description of such system, we now derive a master equation from the above results. After taking the
143: time derivative to Eq. (\ref{rout}) and recalling the explicit form of Eq. (\ref{prord}), we can derive an exact master
144: equation
145: \begin{eqnarray}
146: \dot{\rho}(t) &=&-\frac{i}{\hbar }[H^{\prime }(t),\rho (t)]+\sum_{k=1} ^2\Gamma_k
147: (t)[2a_{k}\rho (t)a_{k}^{\dag } \notag \\
148: &&-a_{k}^{\dag }a_{k}\rho (t)-\rho (t)a_{k}^{\dag }a_{k}] , \label{mas}
149: \end{eqnarray}
150: where
151: \begin{eqnarray}
152: H^{\prime }(t)=\sum_{k=1} ^2 \hbar \Omega_k (t)a_{k}^{\dag }a_{k}, \label{go}
153: \end{eqnarray}
154: is the modified Hamiltonian of the two optical modes and
155: \begin{equation}
156: \frac{\dot{u}_k(t)}{u_k(t)}\equiv -\Gamma_k (t)-i\Omega_k (t). \label{pa}
157: \end{equation}
158: Eq. (\ref{mas}) is the exact master equation for the optical-field system. $\Omega_k(t)$ plays the role of a time-dependent shifted frequency of the
159: $k$-th optical field. $\Gamma_k(t)$ represents the corresponding time-dependent decay rate of the field. We emphasize that the derivation of the
160: master equation goes beyond the Born-Markovian approximation and contains all the backactions between the system and the environments self-consistently.
161: All the non-Markovian character resides in the time-dependent coefficients of the exact master equation.
162:
163: The time-dependent coefficients in the exact master equation, determined by Eq. (\ref{pa}), essentially depend on the so-called spectral density, which
164: characterizes the coupling strength of the environment to the system with respect to the frequencies of the environment. It is defined as
165: $J_l(\omega)=\sum_{k}\left\vert g_{lk}\right\vert^{2}\delta (\omega -\omega _{l})$. In the continuum limit the spectral density may have the form
166: \begin{equation}\label{spectral}
167: J_k(\omega )=\eta_k \omega \Big( \frac{\omega }{\omega _{c}}\Big)^{n-1} e^{-\frac{\omega }{\omega_{c}}} ,
168: \end{equation}
169: where $\omega_{c}$ is an exponential cutoff frequency, and $\eta_k $ is a dimensionless coupling constant between $S_k$ and $E_k$. The environment is classified as Ohmic if
170: $n = 1$, sub-Ohmic if $0 < n < 1$, and super-Ohmic for $n > 1$ \cite{Leggett87, Hu92}. Different spectral densities manifest different non-Markovian
171: decoherence dynamics.
172:
173: We note that our exact master equation reduces to the conventional master equation under the relevant Markov approximation. The coefficients in the
174: master equation (\ref{mas}) become time-independent \cite{An08}
175: \begin{eqnarray}\label{mp}
176: \Gamma_k(t) & = & \pi J_k (\omega_k), \notag \\
177: \Omega_k (t) & = & \omega_k - \mathcal{P}\int^{+\infty}_0\frac{J(\omega)d\omega}{\omega-\omega_k},
178: \end{eqnarray}
179: where $\mathcal{P}$ denotes the Cauchy principal value. The coefficients in Eqs. (\ref{mp}) are precisely the corresponding ones in the Markovian master
180: equation of the optical system \cite{Carmichael93}.
181:
182: \section{The non-Markovian entanglement dynamics of the entangled squeezed state}
183: Initially at time $t = 0$, $S_1S_2$ is in an entangled squeezed state. The entangled two-mode squeezed state is defined as the vacuum state acted on by the two-mode squeezing operator
184: \begin{equation}\label{initial}
185: |\psi(0)\rangle = \exp[r(a_1a_2 - a^{\dag}_1a^{\dag}_2)]|00\rangle,
186: \end{equation}
187: where $r$ is the squeezing parameter. In the coherent-state representation, this initial state is given by
188: \begin{equation}
189: \rho(\boldsymbol{\bar{\alpha}}_i, \boldsymbol{\alpha}^{\prime}_i; 0) =
190: \frac{\exp[-\tanh r(\bar{\alpha}_{1i}\bar{\alpha}_{2i} + \alpha^{\prime}_{1i}\alpha^{\prime}_{2i})]}{\cosh^2r}.
191: \end{equation}
192: The state approaches the ideal Einstein-Podolsky-Rosen (EPR) state \cite{EPR35} in the limit of infinite squeezing ($r\rightarrow \infty$).
193: The traditional way to generate the entangled two-mode squeezed state is via the nonlinear optical process of parametric down-conversion \cite{Ou92}.
194: Recently, a microwave cavity QED-based scheme to generate such states has also been proposed \cite{Pielawa07}.
195: After generating the entangled state given by Eq. (\ref{initial}), the two cavity fields are then propagated, respectively, to the two locations
196: separated between the sender and the receiver. A quantum channel is thus established through the entangled two-mode squeezed state and is ready for
197: teleporting unknown optical coherent states \cite{Braunstein98,Furusawa98}.
198:
199: At $t > 0$, due to interactions with $E_1E_2$, $|\psi(0)\rangle$ evolves to a mixed state. A straightforward way to obtain the time-dependent
200: mixed state is by integrating the propagating function over the initial state of Eq. (\ref{rout}). Then the time-evolution solution of the reduced
201: density matrix can be obtained exactly,
202: \begin{equation}
203: \rho (\boldsymbol{\bar{\alpha}}_{f},\boldsymbol{\alpha }_{f}^{\prime
204: };t)=a\exp [\sum_{k\neq k^{\prime }}(\frac{b}{2}\bar{\alpha}_{kf}\bar{\alpha}%
205: _{k^{\prime }f}+c\bar{\alpha}_{kf}\alpha _{kf}^{\prime }+\frac{b^{\ast }}{2}%
206: \alpha _{kf}^{\prime }\alpha _{k^{\prime }f}^{\prime })], \label{final}
207: \end{equation}
208: where
209: \begin{eqnarray}
210: a & = & \frac{1}{\cosh ^{2}\left\vert r\right\vert [1-\tanh ^{2}\left\vert r\right\vert (1-\left\vert u(t)\right\vert ^{2})^{2}]}, \\
211: b & = & \frac{-\tanh \left\vert r\right\vert u(t)^{2}}{1-\tanh ^{2}\left\vert r\right\vert (1-\left\vert u(t)\right\vert ^{2})^{2}}, \\
212: c & = & \frac{\tanh ^{2}\left\vert r\right\vert (1-\left\vert u(t)\right\vert ^{2})\left\vert u(t)\right\vert ^{2}}{1-\tanh ^{2}\left\vert r\right\vert
213: (1-\left\vert u(t)\right\vert ^{2})^{2}}.
214: \end{eqnarray}
215:
216: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
217:
218: To measure the entanglement in continuous variable system, one generally uses the logarithmic negativity \cite{Werner02}. The logarithmic negativity of
219: a bipartite system was introduced originally as
220: \begin{equation}
221: E_{N} = \log_2\sum_i\left\vert \lambda^-_i\right\vert,
222: \end{equation}
223: where $\lambda^-_i$ is the negative eigenvalue of $\rho^{T_i}$, and $\rho^{T_i}$ is a partial transpose of the bipartite state $\rho$ with respect to
224: the degrees of freedom of the $i$-th party. This measure is based on the Peres-Horodecki criterion \cite{Peres96,Horodecki96} that a bipartite quantum
225: state is separable if and only if its partially transposed state is still positive.
226:
227: For the continuous-variable (Gaussian-type) bipartite state, its density matrix is characterized by the covariance matrix defined as the second moments
228: of the quadrature vector $X = (x_1, p_1, x_2, p_2)$,
229: \begin{equation}
230: V_{ij} = \frac{\langle\Delta X_i\Delta X_j + \Delta X_j\Delta X_i\rangle}{2},
231: \end{equation}
232: where $\Delta X_i = X_i - \langle X_i\rangle$, and $x_i = \frac{a_i + a^{\dag}_i}{\sqrt{2}}$, $p_i = \frac{a_i - a^{\dag}_i}{i\sqrt{2}}$. The canonical
233: commutation relations take the form as $[X_i, X_j] = iU_{ij}$, with $U = \left(
234: \begin{array}{cc}
235: J & 0 \\
236: 0 & J
237: \end{array}
238: \right)$ and $J = \left(
239: \begin{array}{cc}
240: 0 & 1 \\
241: -1 & 0%
242: \end{array}
243: \right)$ defining the symplectic structure of the system. The property of the covariance matrix $V$ is fully determined by its symplectic spectrum
244: $\nu = (\nu_1, \nu_2)$, with $\pm\nu_i$ ($\nu_i > 0$) the eigenvalues of the matrix: $iUV$. The uncertainty principle exerts a constraint on $\nu_i$
245: such that $\nu_i \geqslant \frac{1}{2}$ \cite{Adesso05}. Thus the Peres-Horodecki criterion for the continuous-variable state can be rephrased as
246: the state being separable if and only if the uncertainty principle, $V + \frac{i}{2}U \geqslant 0$, is still obeyed by the covariance matrix under
247: the partial transposition with respect to the degrees of freedom of a specific subsystem \cite{Simon00}. In terms of phase space, the action of
248: partial transposition amounts to a mirror reflection with respect to one of the canonical variables of the related subsystem. For instance,
249: $\tilde{V} = \Lambda V\Lambda$, and $\Lambda = diag(1,1,1,-1)$ is the partial transposition with respect to the second subsystem. If a Gaussian-type
250: bipartite state is nonseparable, the covariance matrix $\tilde{V}$ will violate the uncertainty principle and its symplectic spectrum
251: $\tilde{\nu} = (\tilde{\nu}_1, \tilde{\nu}_2)$ will fail to satisfy the constraint $\tilde{\nu}_i \geqslant \frac{1}{2}$.
252: The logarithmic negativity is then used to quantify this violation as \cite{Werner02}
253: \begin{equation}\label{measu}
254: E_N = \max \{0, -\log_2(2\tilde{\nu}_{\min})\},
255: \end{equation}
256: where $\tilde{\nu}_{\min}$ is the smaller one of the two symplectic eigenvalues $\tilde{\nu}_i$. It is evident from Eq. (\ref{measu}) that, if
257: $\tilde{V}$ obeys the uncertainty principle, i.e., $\tilde{\nu}_i \geqslant \frac{1}{2}$, then $E_N(\rho) = 0$, namely, the state is separable.
258: Otherwise, it is entangled. Therefore, the symplectic eigenvalue $\tilde{\nu}_{\min}$ encodes a qualitative feature of the entanglement for an
259: arbitrary continuous-variable bipartite state.
260:
261: With this entanglement measure at hand, we study now the entanglement dynamics of the squeezed-state quantum channel in our model.
262: From the time-dependent state, the covariance matrix for the optical field
263: can be calculated straightforwardly,
264: \begin{equation}
265: V=\left(
266: \begin{array}{cccc}
267: \frac{y(1+d)}{2(1-d)^{2}} & 0 & \frac{a\text{Re}[b]}{x} & \frac{a\text{Im}[b]%
268: }{x} \\
269: 0 & \frac{y(1+d)}{2(1-d)^{2}} & \frac{a\text{Im}[b]}{x} & \frac{-a\text{Re}%
270: [b]}{x} \\
271: \frac{a\text{Re}[b]}{x} & \frac{a\text{Im}[b]}{x} &
272: \frac{y(1+d)}{2(1-d)^{2}}
273: & 0 \\
274: \frac{a\text{Im}[b]}{x} & \frac{-a\text{Re}[b]}{x} & 0 & \frac{y(1+d)}{%
275: 2(1-d)^{2}}%
276: \end{array}%
277: \right) ,
278: \end{equation}%
279: where $x=[(1-c)^{2}-\left\vert b\right\vert ^{2}]^{2}$, $y=\frac{a}{1-c}$,
280: and $d=c+\frac{\left\vert b\right\vert ^{2}}{1-c}$. And the logarithmic negativity $E_N(t)$ can also be obtained
281: exactly from Eq. (\ref{measu}). It is easy to verify that the initial entanglement is $E_N(0) = \frac{2r}{\ln 2}$.
282:
283:
284:
285: \section{Numerical results and discussions}
286: \begin{figure}[tbp]
287: \scalebox{0.37}{\includegraphics{Fig1.eps}}
288: \caption{Damping rate $\Gamma(t)$ and logarithmic negativity $E_N(t)$ for the entangled squeezed state as function of dimensionless quantity
289: $\omega_0 t$ and their corresponding Markovian results (dashed lines). The parameters $\omega_c/\omega_0 = 50.0$, $\eta = 0.1$, and $r = 1.0$ are used
290: in the numerical calculation.}\label{f1}
291: \end{figure}
292:
293: \begin{figure}[tbp]
294: \scalebox{0.37}{\includegraphics{Fig2.eps}}
295: \caption{Damping rate $\Gamma(t)$ and logarithmic negativity $E_N(t)$ for the entangled squeezed state as function of dimensionless quantity
296: $\omega_0 t$ and their corresponding Markovian results (dashed lines). The parameters $\omega_c/\omega_0 = 1.0$, $\eta = 5.0$, and $r = 1.0$ are used
297: in the numerical calculation.}\label{f2}
298: \end{figure}
299:
300: \begin{figure}[tbp]
301: \scalebox{0.37}{\includegraphics{Fig3.eps}}
302: \caption{Damping rate $\Gamma(t)$ and logarithmic negativity $E_N(t)$ for the entangled squeezed state as function of dimensionless quantity
303: $\omega_0 t$ and their corresponding Markovian results (dashed lines). The parameters $\omega_c/\omega_0 = 0.2$, $\eta = 5.0$, and $r = 1.0$ are used
304: in the numerical calculation.}\label{f3}
305: \end{figure}
306:
307: In the following, we analyze explicitly the exact decoherence dynamics of the entangled squeezed state of $S_1S_2$ under the influence of $E_1E_2$. For
308: simplicity, we assume from here on that the two optical fields are identical, i.e., $\omega_1 = \omega_2 \equiv \omega_0$; and they interact with the
309: same strength, $g_{1l} = g_{2l} \equiv g_l$, with their individual environments. For definiteness, we consider both $E_1$ and $E_2$ to have Ohmic
310: spectral density. The environmental correlation time $\tau_E$ in this case is roughly inversely proportional to the cutoff frequency $\omega_c$ in
311: Eq. (\ref{spectral}), i.e., $\tau_E \simeq 1/\omega_c$ \cite{Weiss}. It is emphasized that the cutoff frequency $\omega _{c}$, which is originally
312: introduced to eliminate infinities in frequency integrations, therefore also determines if the dynamics of open system $S$ is Markovian or non-Markovian.
313: Our non-perturbatively derived exact results allow us to explore all these possibilities.
314:
315: In Fig. \ref{f1}, we plot the numerical results of the decay rate $\Gamma(t)$ and logarithmic negativity $E_N(t)$ when $\tau_E \ll \tau_0$. The
316: positivity of $\Gamma(t)$ throughout the whole evolution process guarantees the monotonic decrease of $E_N(t)$. Accordingly, the entangled squeezed
317: state eventually evolves to a product state, namely the ground state of the system: $\rho_g = |00\rangle\langle00|$. Clearly, in this case, the
318: backactions of $E_1E_2$ have a negligible effect on the dynamics of $S_1S_2$, and we say the system dynamics is mainly governed by the dissipative effect
319: of the environments. There is thus no qualitative difference between the exact entanglement dynamics and the Markovian results. Quantitatively,
320: however, we note that for $t < \tau_E$, the distinctive increase of $\Gamma(t)$ results in $E_N(t)$ decreasing rapidly. This non-Markovian effect only
321: shows up in a very short time scale. In fact, for $t > \tau_E$, $\Gamma(t)$ decreases and approaches gradually to a constant value as $t$ approaches
322: $\tau_0$; and the rate of decrease of $E_N(t)$ decreases.
323:
324: Fig. \ref{f2} shows $\Gamma(t)$ and $E_N(t)$ when $\tau_E = \tau_0$. In this case, the backactions of $E_1E_2$ have a considerable impact on the dynamics
325: of $S_1S_2$, and the Markovian approximation is not applicable. Firstly, we note that $\Gamma(t)$ can take negative values. Physically, this
326: corresponds to the systems reabsorbing photons from the environments, which leads to an increase in the photon number of the systems \cite{Breuer02}.
327: These negative decay rates provide evidences for backactions in non-Markovian dynamics \cite{Piilo08}. Secondly, we observe that $\Gamma(t)$ approaches
328: zero asymptotically. Both results clearly differ from the Markovian ones. Consequently, $E_N(t)$ presnts distinctive behaviors that are absent in the
329: Markovian results. Firstly, due to the negative decay rates, $E_N(t)$ shows oscillations. We must emphasize that these oscillations are fundamentally
330: different from the transient entanglement oscillations previously obtained when two optical-fields interact with a common environment \cite{An07}. They
331: are caused by the backactions of the environments on their respective local optical fields and are characteristic of non-Markovian dynamics. Similar
332: oscillations have been obtained in a system of two two-level atoms in two separated damping cavities \cite{Bellomo07}. Secondly, and more interestingly,
333: we find that there is some residual entanglement left in the steady state. From previous studies \cite{Ban06, Goan07, An07}, one would have concluded
334: that non-Markovian effects only show up in short-time dynamics. Our results, however, clearly show on the contrary that non-Markovian effects can also
335: have an influence on the long-time behavior of the system dynamics and the final steady state of the system. This counterintuitive behavior can be
336: explained with the fact that the dissipative influence on the entanglement dynamics by the environments is strongly counteracted by the effect due to
337: their backactions. Consequently, the decay of the entanglement ceases when the system evolves to some steady state, which is not the ground state
338: $\rho_g$.
339:
340: The results when $\tau_E \gg \tau_0$ are shown in Fig. \ref{f3}; a situation considered in Ref. \cite{Paris07}. Due to extremely long memory of the
341: environments, the backactions on the systems are so strong that they govern the decoherence dynamics. As a result, $\Gamma(t)$ and hence $E_N(t)$
342: oscillate over a very long duration. These oscillations persist even as the state approaches the ground state. The `equilibrium' position for the
343: oscillation of $\Gamma(t)$ is not at zero, but a small positive value. This positivity means the systems dynamics will experience a weak dissipation,
344: which is verified by the time evolution of $E_N(t)$ in Fig. \ref{f3}.
345:
346: In summary, we have studied the exact entanglement dynamics of the two optical-field system for different $\tau_E$'s in comparison with $\tau_0$.
347: Specifically, we have analyzed when the system dynamics will exhibit novel non-Markovian effects, and provided a detailed description of these.
348:
349: \section{Conclusions}
350: We have applied the influence-functional method of Feynman and Vernon to investigate the exact entanglement dynamics of two single-mode optical fields
351: $S_1S_2$ coupled to two independent and uncorrelated environments $E_1E_2$. From our analytical and numerical results, it is seen that $E_1E_2$ exert
352: two competing influences on our system. One effect, ${\cal D}$, is dissipative and is responsible for the decoherence of $S_1S_2$. The other,
353: ${\cal B}$, is due to the backactions of $E_1E_2$ on $S_1S_2$. The degree of manifestations of ${\cal D}$ and ${\cal B}$ in the dynamics of $S_1S_2$
354: depends on $\tau_E$ in comparison with $\tau_0$. For $\tau_E \ll \tau_0$, ${\cal D}$ dominates and ${\cal B}$ only gives rise to a transient coherent
355: oscillation of $S_1S_2$. The state of $S_1S_2$ evolves to the ground state $\rho_g$, which is coincident with the Markovian result. If
356: $\tau_E = \tau_0$, the near resonant interaction between $S_1S_2$ and $E_1E_2$ results in ${\cal D}$ and ${\cal B}$ being comparable and counteract each
357: other. These give rise to transient negative decay rates and asymptotical zero decay rate. The state of $S_1S_2$ thus evolves asymptotically to some
358: steady state, which is not the ground state $\rho_g$. Finally, when $\tau_E \gg \tau_0$, ${\cal B}$ dominates and governs the dynamics of $S_1S_2$. The
359: decay rates of the system oscillate about some non-negative equilibrium position over a very long duration. This slight positivity guarantees an
360: overall weak dissipative effect on the system dynamics. Therefore, the state of $S_1S_2$ eventually approaches the ground state with the entanglement
361: oscillation persisting on for a very long time.
362:
363: The theory we have established is a non-perturbative description of the exact decoherence dynamics of a system of two single-mode optical fields. It
364: can serve as a useful basic theoretical model in analyzing the non-Markovian decoherence dynamics of optical fields employed in practical quantum
365: information processing schemes. It should be noted that although only the Ohmic spectral density is considered here, it is straightforward to
366: generalize our discussion to the non-Ohmic cases.
367:
368: \section*{Acknowledgement}
369: The work is supported by NUS Research Grant No. R-144-000-189-305. J.H.A. also thanks the financial support of the NNSF of China under Grant
370: No. 10604025, and the Fundamental Research Fund for Physics and Mathematics of Lanzhou University under Grant No. Lzu05-02.
371:
372:
373:
374: \begin{thebibliography}{99}
375: \bibitem{Nielsen00} M. A. Nielsen and I. L. Chuang, \textit{Quantum Computation and Quantum Information} (Cambridge University Press, Cambridge, U.K., 2000).
376: \bibitem{Bennett} C. H. Bennett, G. Brassard, C. Crepeau, R. Jozsa, A. Peres, and W. K. Wootters, Phys. Rev. Lett. {\bf 70}, 1895 (1993).
377: \bibitem{Bouwmeester97} D. Bouwmeester, J.-W. Pan, K. Matter, M. Eibl, H. Weinfurter, and A. Zeilinger, Nature \textbf{390}, 575 (1997).
378: \bibitem{Braunstein98} S. L. Braunstein and H. J. Kimble, Phys. Rev. Lett. \textbf{80}, 869 (1998).
379: \bibitem{Furusawa98} A. Furusawa, J. L. Sorensen, S. L. Braunstein, C. A. Fuchs, H. J. Kimble, and E. S. Polzik, Science \textbf{282}, 706 (1998).
380:
381: \bibitem{Jakub04} J. S. Prauzner-Bechcicki, J. Phys. A: Math. Gen. \textbf{37}, L173 (2004).
382: \bibitem{An05} J.-H. An, S.-J. Wang, and H.-G. Luo, J. Phys. A: Math. Gen. \textbf{38}, 3579 (2005).
383: \bibitem{Rossi06} R. Rossi Jr., A. R. Bosco de Magalh\~{a}es, M. C. Nemes, Physica A \textbf{365}, 402 (2006).
384:
385: \bibitem{Ban06} M. Ban, J. Phy. A: Math. Gen. \textbf{39}, 1927 (2006).
386: \bibitem{Goan07} K.-L. Liu and H.-S. Goan, Phys. Rev. A \textbf{76}, 022312 (2007).
387: \bibitem{Paris07} S. Maniscalco, S. Olivares, and M. G. A. Paris, Phys. Rev. A \textbf{75}, 062119 (2007).
388: \bibitem{An07} J.-H. An and W. M. Zhang, Phys. Rev. A \textbf{76}, 042127 (2007).
389:
390: \bibitem{Carmichael93} H. J. Carmichael, \textit{An Open Systems Approach to Quantum Optics}, Lecture Notes in Physics, Vol. m18 (Springer-Verlag, Berlin, 1993).
391: \bibitem{Breuer02} H.-P. Breuer and F. Petruccione, \textit{The theory of open quantum systems} (Oxford University Press, Oxford, 2002).
392:
393: \bibitem{Dubin07} F. Dubin, D. Rotter, M. Mukherjee, C. Russo, J. Eschner, and R. Blatt, Phys. Rev. Lett. \textbf{98}, 183003 (2007).
394: \bibitem{Koppens07} F. H. L. Koppens, D. Klauser, W. A. Coish, K. C. Nowack, L. P. Kouwenhoven, D. Loss, and L. M. K. Vandersypen, Phys. Rev. Lett. \textbf{99}, 106803 (2007).
395: \bibitem{Mogilevtsev08} D. Mogilevtsev, A. P. Nisovtsev, S. Kilin, S. B. Cavalcanti, H. S. Brandi, and L. E. Oliveira, Phys. Rev. Lett. \textbf{100}, 017401 (2008).
396:
397: \bibitem{Weiss} U. Weiss, \textit{Quantum Dissipative Systems}, 2nd ed. (World Scientific, Singapore, 1999).
398:
399: \bibitem{Feynman63} R. P. Feynman and F. L. Vernon, Ann. Phys. (N. Y.) \textbf{24}, 118 (1963).
400: \bibitem{Leggett87} A. J. Leggett, S. Chakravarty, A. T. Dorsey, M. P. A. Fisher, A. Garg, and W. Zwerger, Rev. Mod. Phys. \textbf{59}, 1 (1987).
401:
402: \bibitem{An072} J.-H. An, M. Feng, W. M. Zhang, arXiv:0705.2472v2 [quant-ph].
403: \bibitem{An08} J.-H. An, Y. Yeo, and C. H. Oh, arXiv:0808.3178v1 [quant-ph].
404:
405: \bibitem{Zhang90} W. M. Zhang, D. H. Feng, and R. Gilmore, Rev. Mod. Phys. 62, 867 (1990).
406:
407: \bibitem{Hu92} B. L. Hu, J. P. Paz, and Y. Zhang, Phys. Rev. D \textbf{45}, 2843 (1992).
408:
409: \bibitem{EPR35} A. Einstein, B. Podolsky, and N. Rosen, Phys. Rev. \textbf{47}, 777 (1935).
410:
411: \bibitem{Ou92} Z. Y. Ou, S. F. Pereira, H. J. Kimble, and K. C. Peng, Phys. Rev. Lett. \textbf{68}, 3663 (1992).
412:
413: \bibitem{Pielawa07} S. Pielawa, G. Morigi, D. Vitali, and L. Davidovich, Phys. Rev. Lett. \textbf{98}, 240401 (2007).
414:
415: \bibitem{Werner02} G. Vidal and R. F. Werner, Phys. Rev. A \textbf{65}, 032314 (2002).
416:
417: \bibitem{Peres96} A. Peres, Phys. Rev. Lett. \textbf{77}, 1413 (1996).
418: \bibitem{Horodecki96} M. Horodecki, P. Horodecki, and R. Horodecki, Phys. Lett. A \textbf{223}, 1 (1996).
419:
420: \bibitem{Adesso05} G. Adesso and F. Illuminati, Phys. Rev. A \textbf{72}, 032334 (2005).
421:
422: \bibitem{Simon00} R. Simon, Phys. Rev. Lett. \textbf{84}, 2726 (2000).
423:
424: \bibitem{Bellomo07} B. Bellomo, R. LoFranco, and G. Compagno, Phys. Rev. Lett. \textbf{99}, 160502 (2007).
425:
426: \bibitem{Piilo08} J. Piilo, S. Maniscalco, K. H\"{a}rk\"{o}nen, and K-A. Suominen, Phys. Rev. Lett. \textbf{100}, 180402 (2008).
427:
428: \end{thebibliography}
429:
430: \end{document}
431: