1: % LaTeX source file! %
2: % %
3: % %
4: % Title: Exciton entanglement in two coupled semiconductor %
5: % microcrystallites %
6: % %
7: % Authors: Yu-xi Liu, Sahin K. Ozdemir, Adam Miranowicz, %
8: % Masato Koashi, and Nobuyuki Imoto %
9: % %
10: % submitted to J. Phys. A %
11: % %
12: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
13: \documentclass{iopart}
14: \input epsf
15:
16:
17: \def\labela{\def\theequation{\arabic{equation}a}}
18: \def\labelb{\addtocounter{equation}{-1}\def\theequation{\arabic{equation}b}}
19: \def\labelc{\addtocounter{equation}{-1}\def\theequation{\arabic{equation}c}}
20: \def\labeld{\addtocounter{equation}{-1}\def\theequation{\arabic{equation}d}}
21: \def\labelz{\def\theequation{\arabic{equation}}}
22:
23: \def \binom#1#2{{#1\choose #2}}
24:
25: \def \omit#1{{{\bf OMIT:} {\em #1}}}
26:
27: \begin{document}
28:
29: %\pagenumbering{roma}
30:
31: \title{Exciton entanglement in two coupled semiconductor microcrystallites}
32:
33: \author{Yu-xi Liu$^1$, \c{S}ahin K. \"Ozdemir$^{1,2,5}$,
34: Adam Miranowicz$^{1,2,3}$, Masato Koashi$^{1,2,5}$, and Nobuyuki
35: Imoto$^{1,2,4,5}$}
36:
37: \address{$^1$ The Graduate University for Advanced Studies
38: (SOKENDAI), Shonan Village, Hayama, Kanagawa 240-0193, Japan}
39:
40: \address{$^2$ SORST Research Team for Interacting Carrier
41: Electronics, Hayama, Kanagawa 240-0193, Japan}
42:
43: \address{$^3$ Nonlinear Optics Division, Physics Institute, Adam
44: Mickiewicz University, 61-614 Pozna\'n, Poland}
45:
46: \address{$^4$ NTT Basic Research Laboratories, 3-1
47: Morinosato-Wakamiya, Atsugi, Kanagawa 243-0198, Japan}
48:
49: \address{$^5$ CREST Research Team for Photonic Quantum Information,
50: Hayama, Kanagawa 240-0193, Japan}
51:
52: \begin{abstract}\\
53: Entanglement of the excitonic states in the system of two coupled
54: semiconductor microcrystallites, whose sizes are much larger
55: than the Bohr radius of exciton in bulk semiconductor but smaller
56: than the relevant optical wavelength, is quantified in terms of
57: the entropy of entanglement. It is observed that the nonlinear
58: interaction between excitons increases the maximum values of the
59: entropy of the entanglement more than that of the linear coupling
60: model. Therefore, a system of two coupled microcrystallites can
61: be used as a good source of entanglement with fixed exciton
62: number. The relationship between the entropy of the entanglement
63: and the population imbalance of two microcrystallites is
64: numerically shown and the uppermost envelope function for them is
65: estimated by applying the Jaynes principle.
66:
67: \vspace{3mm} \noindent {\sc published in J. Phys. A {\bf 37},
68: 4423-4436 (March 2004)}
69:
70:
71: \noindent {PACS numbers 03.67.-a, 71.10.Li, 71.35.-y, 73.20Dx}
72:
73: \end{abstract}
74:
75: \pagenumbering{arabic}
76: \section{Introduction}
77:
78:
79: Recently, quantum computation and quantum information excite large
80: enthusiasm among theoretical and experimental physicists. Many
81: theoretical and experimental researches have been devoted to the
82: preparation and measurement of the entangled states that have been
83: considered to be a key ingredient in the realization of the
84: quantum computation and information.
85:
86: We can quantitatively understand the role of quantum
87: entanglement as a resource for communication by studying the
88: quantum teleportation, which was first realized for discrete
89: variables~\cite{Zeilinger} and then for continuous
90: variables~\cite{B}. The possibility for the generation of the
91: maximally entangled states with a fixed photon number from
92: squeezed vacuum states or from mixed Gaussian continuous entangled
93: states by the quantum non-demolition measurement~\cite{mmm} has
94: been theoretically studied. Quantum teleportation using an
95: entangled source of fixed total photon number has also been
96: theoretically investigated in~\cite{c}. However, it has been shown
97: that a generation of the maximally entangled states with a fixed
98: finite photon number is a great challenge to modern technology. So
99: the exploration of new maximally entangled bipartite source with
100: fixed particle number is very interesting and important from both
101: experimental and theoretical view points.
102:
103: Several schemes based on coupled quantum dots have been proposed
104: for fabricating quantum gates~\cite{4,44}. The entanglement of the
105: exciton states in a single quantum dot or in a quantum dot
106: molecule has been demonstrated experimentally \cite{5,6}. The
107: authors of references \cite{7,99,yu-xi2,wang} theoretically
108: investigated the entanglement of excitonic states in the system of
109: the optically driven coupled quantum dots and propose models to
110: prepare Bell, Greenberger-Horne-Zeilinger (GHZ) and $W$ states.
111: The above investigations mainly focus on the case in which the
112: sizes of quantum dots are smaller than the Bohr radius of exciton
113: in bulk semiconductor, so that the quantum dots have discrete
114: energy levels and one energy state can only be occupied by one
115: exciton because of the Pauli principle. In such quantum dot
116: systems, it is easy to define a qubit by different spins of
117: excitons, or by single-exciton and no-exciton states. So when we
118: want to encode quantum information in systems of quantum dots by a
119: finite-dimensional quantum state space (called qudit system), we
120: need to consider higher energy levels. However, if we can enlarge
121: the size $R$ of the quantum dot so that it is larger than the Bohr
122: radius $a_{B}$ of the exciton in the bulk semiconductor but
123: smaller than the relevant light wavelength $\lambda$, in such a
124: system of quantum microcrystallites (which can also be called
125: large quantum dots system~\cite{Han,Ban,Eng}), the excitons can
126: occupy the same level when the average number of excitons per Bohr
127: radius volume is less than one. Then we can encode quantum
128: information in the finite Hilbert space built by the different
129: exciton number states. Quantum information properties of
130: semiconductor microcrystallites coupled by a cavity field were
131: investigated by us in Refs. \cite{yu-xi} where, in particular, we
132: have proposed a realization of symmetric sharing of entanglement
133: for the exciton states in semiconductor microcrystallites and also
134: studied the environment effect on the qubit.
135:
136: Motivated by these considerations, in this paper we will study a
137: system of two semiconductor microcrystallites coupled by Coulomb
138: interaction. We will study the entanglement of the qudit excitonic
139: states in these two coupled semiconductor microcrystallites with
140: fixed exciton number.
141:
142: The paper is organized as follows: In Sec. II, we
143: first give a theoretical description of two coupled
144: semiconductor microcrystallites. We obtain the analytical solution
145: of the system by virtue of the Schwinger representation of the
146: angular momentum. In Sec. III, the entanglement of the two
147: subsystems is discussed using the von Neumann entropy under
148: certain initial conditions. We discuss and numerically analyze
149: the oscillation of the exciton population imbalance between two
150: quantum dots. The relationship between the entanglement and
151: oscillation of the imbalance population of excitons is
152: demonstrated in detail in Sec. IV. Finally, we give our
153: conclusions and some comments.
154:
155: \section{The model and its analytical solution}
156:
157: We consider a system which consists of two semiconductor
158: microcrystallites coupled by Coulomb interaction. We assume that
159: two semiconductor microcrystallites are completely symmetric.
160: Their sizes are larger than the Bohr radius $a_{B}$ of an exciton
161: in a bulk semiconductor but smaller than the relevant light
162: wavelength $\lambda$. Using two-band approximation to model these
163: two coupled semiconductor microcrystallites, we have the
164: Hamiltonian
165: \begin{equation}
166: H=\hbar\sum_{j,k=1}^{2}\chi_{jk}a^{\dagger}_{j}a_{k}+\sum_{j,k,l,m=1
167: }^{2} \chi_{jklm}a^{\dagger}_{j}a^{\dagger}_{k}a_{l}a_{m},
168: \label{eq:Ham}
169: \end{equation}
170: where $a^{\dagger}_{j}\,(a_{j})$ are creation (annihilation)
171: operators of excitons, which are electron-hole pairs bound by the
172: Coulomb interaction. We assume that the density of the excitons is
173: so low and the external confinement potential to microcrystallite
174: is so weak that the exciton operators $a^{\dagger}_{j}\,(a_{j})$
175: can be described approximately as bosonic operators, that is they
176: satisfy the commutation relations of the ideal bosons, $[a_{j},
177: a^{\dagger}_{k}]=\delta_{jk}$, which is somewhat different from
178: that in Ref.~\cite{d}. The label $j$=1 (2) denotes the
179: microcrystallite $A\,(B)$. In the Hamiltonian (\ref{eq:Ham}), the
180: deviation of the exciton operators from the ideal bosonic model is
181: corrected by introducing an effective nonlinear interaction between
182: the hypothetical ideal bosons. In general, the parameters
183: $\chi_{jk}$ are different from each other, however in this study
184: we consider two completely equivalent microcrystallites, which
185: have nearly the same Bohr radius of the excitons and transitional
186: dipole moment. For the sake of simplicity, we assume the parameters
187: $\chi_{jk}=\omega
188: >0$ for $j=k$, which means that the two semiconductor
189: microcrystallites have the same transition frequency from the
190: valence band to the conduction band, and assume positive real
191: numbers $\chi_{jk}=g$ for $j \neq k$, which correspond to a linear
192: coupling constant of the two semiconductor microcrystallites. The
193: parameters $\chi_{jklm}$ of the nonlinear terms are also taken as
194: positive constant and set to be $\chi_{jklm}=\chi$. Under these
195: assumptions, the Hamiltonian (\ref{eq:Ham}) can be simplified as
196: follows
197: \begin{eqnarray}
198: H&=&\hbar\Omega {\bf N}+\hbar\chi {\bf N}^{2}
199: +\hbar{\bf G}(a^{\dagger}_{1}a_{2}+a^{\dagger}_{2}a_{1})\nonumber \\
200: &&+\hbar\chi(a^{\dagger}_{1}a_{2}+a^{\dagger}_{2}a_{1})^{2},
201: \label{eq:ham}
202: \end{eqnarray}
203: where $\Omega=\omega-2\chi$, and ${\bf G}=g-2\chi+2\chi {\bf N}$.
204: It is obvious that ${\bf N}=a^{\dagger}_{1}a_{1}
205: +a^{\dagger}_{2}a_{2}$ is a constant of motion, which means $[{\bf
206: N},H]=0$ and the total exciton number of the coupled semiconductor
207: microcrystallites is conserved. In such a situation, we find that
208: it is more convenient to get the solution of the Schr\"odinger
209: equation governed by Hamiltonian (\ref{eq:ham}) by virtue of
210: Schwinger representation of the angular momentum
211: \cite{Sch65,Sak94}. That is, we can introduce the angular momentum
212: operators \labela
213: \begin{eqnarray}
214: J_{x}&=&\frac{1}{2}(a^{\dagger}_{1}a_{2}+a^{\dagger}_{2}a_{1}),\label{eq:J1}
215: \end{eqnarray}
216: \labelb
217: \begin{eqnarray}
218: J_{y}&=&\frac{1}{2i}(a^{\dagger}_{1}a_{2}-a^{\dagger}_{2}a_{1}),
219: \label{eq:J2}
220: \end{eqnarray}
221: \labelc
222: \begin{eqnarray}
223: J_{z}&=&\frac{1}{2}(a^{\dagger}_{1}a_{1}-a^{\dagger}_{2}a_{2})\label{eq:J}
224: \end{eqnarray}
225: \labelz%
226: from the bosonic annihilation and creation operators of
227: the two exciton modes. The operators of
228: (\ref{eq:J1})--(\ref{eq:J}) satisfy the commutation relations for
229: the Lie algebra of SU(2):
230: \begin{equation}
231: [J_{j}, J_{k}]=i\varepsilon_{jkl} J_{l}, \hspace{0.5cm}
232: j,k,l=x,y,z,
233: \end{equation}
234: where the Levi-Civit\`a tensor $\varepsilon_{jkl}$ is equal to
235: $+1$ and $-1$ for the even and the odd permutation of its indices,
236: respectively, and zero otherwise. From
237: (\ref{eq:J1})--(\ref{eq:J}), the total angular momentum operator
238: can be expressed as
239: \begin{equation}
240: J^{2}=\frac{\bf N}{2}(\frac{\bf N}{2}+1).
241: \label{eq:Ca}
242: \end{equation}
243: In fact, ${\bf N}$ itself commutes with all the operators
244: (\ref{eq:J1})--(\ref{eq:J}) and (\ref{eq:Ca}). For a fixed total
245: excitonic number $\cal{N}$, the common eigenstates of $J^{2}$ and
246: $J_{z}$ are the two-mode Fock states \cite{Sch65,Sak94}
247: \begin{equation}
248: |j,m\rangle_s=|m_{1},m_{2}\rangle=
249: \frac{(a_{1}^{\dagger})^{j+m}(a_{2}^{\dagger})^{j-m}}{\sqrt{(j+m)!(j-m)!}}
250: |0\rangle \label{eq:schwinger}
251: \end{equation}
252: with eigenvalues $j={\cal N}/2$ and $m=-{\cal N}/2, \cdots, {\cal
253: N}/2$, where $|m_{1},m_{2}\rangle$ is the Fock state with
254: $m_{1}=j+m$ excitons and $m_{2}=j-m$ excitons in microcrystallite
255: $A$ and microcrystallite $B$ respectively. Although $j\pm m$
256: $(m_{1}, m_{2})$ must be integers, $j$ and $m$ can both be
257: integers or half-odd integers. For consistency, $j$ is replaced by
258: ${\cal N}/2$ in the following expressions. For clarity, the
259: subscript $s$ is used to indicate the Schwinger angular momentum
260: basis in (\ref{eq:schwinger}) and the following equations. By using
261: equation (\ref{eq:schwinger}), we take $j=1$ and $m=0$ as a
262: detailed example to show the relation between the Schwinger angular
263: momentum basis and Fock state basis. That is,
264: $|1,0\rangle_{s}=a_{1}^{\dagger}a_{2}^{\dagger}|0\rangle=|1,1\rangle$
265: which means that when the quantum number of the total angular
266: momentum $j$ is $1$ and its $z$ component $m$ is $0$, there is one
267: exciton in each microcrystallite respectively.
268:
269: In terms of an ${\rm SO(3)}$ rotation $e^{i{\pi}/{2}J_{y}}$ of
270: $2\hbar{\bf G}J_{x}+4\hbar \chi J_{x}^{2}$, equation (\ref{eq:ham})
271: can be simplified into :
272: \begin{eqnarray}
273: H&=&\hbar \Omega {\bf N}+\hbar\chi{\bf N}^{2}+2\hbar {\bf
274: G}J_{x}+4\hbar \chi
275: J_{x}^{2}\nonumber \\
276: &=&\hbar\Omega {\bf N}+\hbar\chi{\bf N}^{2}+2\hbar {\bf G}
277: e^{-i({\pi}/{2})J_{y}}J_{z}e^{i({\pi}/{2})J_{y}}\nonumber \\
278: &&+4\hbar \chi e^{-i({\pi}/{2})J_{y}}J_{z}^{2}
279: e^{i({\pi}/{2})J_{y}}. \label{eq:h2}
280: \end{eqnarray}
281: The eigenfunctions $\Psi_{{\cal N}/2,m}$ and the eigenvalues
282: $E_{{\cal N}/2,m}$ of Hamiltonian (\ref{eq:h2}) can be obtained
283: easily as
284: \labela%
285: \begin{eqnarray}
286: \label{eq:8a}|\Psi_{\frac{\cal N}{2},m}\rangle_{s}&=&e^{(-i\pi/2)
287: J_{y}}|\frac{\cal N}{2},m\rangle_s
288: =\sum_{m^{\prime}=-\frac{\cal
289: N}{2}}^{\frac{\cal N}{2}}{\cal D}^{\frac{\cal N}{2}}_{m^{\prime}m}
290: ({\textstyle\frac{\pi}{2}})|\frac{\cal N}{2} ,m^{\prime}\rangle_s,
291: \end{eqnarray} \labelb%
292: \begin{eqnarray} E_{{\cal N}/2,m}&=&\hbar\Omega {\cal
293: N}+\hbar\chi{\cal N}^2 +2\hbar {\cal G} m+ 4\hbar \chi m^{2},
294: \end{eqnarray}
295: \labelz%
296: where ${\cal N}$ denotes the total excitonic number of two
297: microcrystallites; ${\cal G}=g-2\chi+2\chi {\cal N}$, and Wigner's
298: formula for ${\cal D}^{\frac{\cal
299: N}{2}}_{m^{\prime}m}(\frac{\pi}{2})$ reads as
300: \begin{eqnarray}
301: {\cal D}^{\frac{\cal
302: N}{2}}_{m^{\prime}m}(\frac{\pi}{2})&=&\sum_{k}(-1)^{k-m-m^{\prime}}
303: (\textstyle\frac{1}{2})^{\frac{\cal N}{2}} \nonumber \\
304: &&\times \frac{\sqrt{(\frac{\cal N}{2}+m)!(\frac{\cal
305: N}{2}-m)!(\frac{\cal N}{2}+m^{\prime})!(\frac{\cal
306: N}{2}-m^{\prime})!}} {(\frac{\cal N}{2}+m-k)!k!(\frac{\cal
307: N}{2}-k-m^{\prime})!(k-m+m^{\prime})!}, \label{eq:wg}
308: \end{eqnarray}
309: where we take the sum over $k$ so that none of the arguments of
310: factorials in the denominator is negative. The wave function
311: $|\Psi (t)\rangle$ of the system for the initial condition
312: $|\Psi(t=0)\rangle$ is given by
313: \begin{eqnarray}
314: |\Psi(t)\rangle&=& \exp(-i t H/\hbar) |\Psi (0)\rangle=\sum_{{\cal
315: N}=0}^{\infty}\sum_{m=-{\cal N}/2}^{m={\cal N}/2}\nonumber\\
316: &&\times\exp\left(\frac{E_{\frac{\cal N}{2}, m}}{i\hbar}t \right)
317: \langle\Psi_{\frac{\cal N}{2}, m}|\Psi(0)\rangle |\Psi_{\frac{\cal
318: N}{2}, m}\rangle_{s}. \label{eq:T}
319: \end{eqnarray}
320: The coefficients $\langle\Psi_{\frac{N}{2}, m}|\Psi(0)\rangle$ are
321: rotating matrix elements that can be determined by Wigner's
322: formula. Equation (\ref{eq:T}) is basic and will be used in our
323: further discussions. In the following two sections, we will
324: discuss the entanglement of two exciton modes and demonstrate the
325: oscillations of population imbalance for excitons between two
326: quantum microcrystallites.
327:
328: \section{Entanglement of the excitonic states}
329:
330: Quantum entanglement plays the key role in the quantum information
331: and quantum computation. In general, for any pure composite state
332: $|\psi\rangle_{AB}$ of a bipartite system whose state space is
333: $H_{A}\otimes H_{B}$, the entanglement can be measured by von
334: Neumann's entropy of any reduced density operator $\rho_{A}={\rm
335: Tr}_{B}(|\psi\rangle\!_{AB}\,_{AB}\!\langle\psi|)$ or
336: $\rho_{B}={\rm Tr}_{A}(|\psi\rangle\!_{AB}
337: \,_{AB}\!\langle\psi|)$, where the reduced density operator of
338: system $A$ is obtained by tracing out system $B$ and that of
339: system $B$ by tracing out system $A$. In the context of quantum
340: computation and information, in particular in studies of quantum
341: entanglement, the reduced von Neumann entropy of a bipartite pure
342: state $|\psi\rangle_{AB}$ is usually referred to as the {\em
343: entropy of entanglement} or simply the {\em entanglement} (see,
344: e.g., \cite{Ben96a,Ben96b,Ved98,Pho91}):
345: \begin{eqnarray}
346: E(\rho)&=&-{\rm Tr}(\rho_{A}\ln\rho_{A})=-{\rm
347: Tr}(\rho_{B}\ln\rho_{B})
348: =-\sum_{n}\lambda_{n}\ln\lambda_{n} \label{eq:ss}
349: \end{eqnarray}
350: where $\lambda_{n}$ are the eigenvalues of either $\rho_{A}$ or
351: $\rho_{B}$. They form the (square of the) coefficients of the
352: Schmidt decomposition of the bipartite pure state. That is, the
353: bipartite pure state $|\psi\rangle_{AB}$ can be expressed by a
354: set of biorthogonal vectors using the Schmidt decomposition as
355: \begin{equation}
356: |\psi\rangle\equiv
357: |\psi\rangle_{AB}=\sum_{n}\sqrt{\lambda_{n}}|\alpha_{n}\rangle_{A}
358: |\beta_{n}\rangle_{B} \label{eq:ab}
359: \end{equation}
360: where we have chosen the phases of our basis states so that no
361: phases appear in the coefficients $\lambda_{n}$ in the sum of
362: (\ref{eq:ab}), $\{|\alpha_{n}\rangle, n=0 \cdots \}$ and
363: $\{|\beta_{n}\rangle,n=0\cdots \}$ are orthonormal states of the
364: two subsystems $A$ and $B$ respectively.
365:
366: The total number of excitons in the whole system is fixed by the
367: initial condition, such as $L$. Based on such a condition, the
368: maximally entangled state of the system is
369: \begin{equation}
370: |M\rangle=\frac{1}{\sqrt{L+1}}\sum_{l=0}^{L}|L-l, l\rangle,
371: \label{eq:M}
372: \end{equation}
373: where $|L-l,l\rangle$ represents the state with $L-l$ excitons in
374: microcrystallite $A$ and $l$ excitons in microcrystallite $B$.
375: According to definition (\ref{eq:ss}), the entropy of the
376: entanglement for the maximally entangled state (\ref{eq:M}) is
377: $\ln(1+L)$.
378:
379: Experimentally, the easier way is to excite some number of
380: excitons in one microcrystallite, or both of microcrystallites.
381: Without loss of generality, we first assume that microcrystallite
382: $A$ is initially excited with $L$ excitons, and no excitons
383: initially exist in microcrystallite $B$. So the modes of the two
384: microcrystallites are disentangled at the initial time $t=0$. That
385: is, the state of the whole system is a tensor product of the
386: states of two subsystems $A$ and $B$, i.e. $|\Psi(t=0)\rangle
387: =|L\rangle_{A}|0\rangle_{B}$ with the Schwinger realization of
388: this initial state as $|\Psi(t=0)\rangle
389: =|\frac{L}{2},\frac{L}{2}\rangle_s$. The system and each mode of
390: the system are in pure states and have zero entropies.
391:
392: It is well known that any pure state remains pure with the unitary
393: evolution, but it is not true for each subsystem. With the
394: evolution, the initially pure state of each subsystem can be
395: transformed into mixed states. If the von Neumann entropy
396: $E(\rho)$, given by Eq. (\ref{eq:ss}), is increased, then the two
397: subsystems become more entangled. Thus, the degree of
398: entanglement between the two subsystems in the system at any time
399: can be described by (\ref{eq:ss}).
400:
401: For the initial state $|\frac{L}{2},\frac{L}{2}\rangle_s$ of the
402: system, the total wave function of the system can be obtained from
403: (\ref{eq:T}) as
404: \begin{equation}
405: |\Psi(t)\rangle=\sum_{l^{\prime}=-L/2}^{l^{\prime}=L/2}
406: \exp\Big(\frac{E_{\frac{L}{2},l^{\prime}}}{i\hbar}t\Big)
407: \langle\Psi_{\frac{L}{2},l^{\prime}}|\frac{L}{2},\frac{L}{2}\rangle_s
408: |\Psi_{\frac{L}{2},l^{\prime}}\rangle_{s}, \label{eq:13}
409: \end{equation}
410: where the normalized coefficients
411: $\langle\Psi_{\frac{L}{2},l^{\prime}}|\frac{L}{2},\frac{L}{2}\rangle_s$
412: are determined by (\ref{eq:8a})-(\ref{eq:T}). We will use equation
413: (\ref{eq:13}) to discuss the degree of the entanglement of the two
414: subsystems for the following several concrete examples. First, we
415: consider that there is initially one exciton in the
416: microcrystallite $A$, i.e., $L=1$. In this case, we can obtain the
417: wave function of the whole system from (\ref{eq:13}) and
418: (\ref{eq:schwinger}) with the initial state
419: $|\frac{1}{2},\frac{1}{2}\rangle_s$ as
420: \begin{eqnarray}
421: |\Psi^{(1)}(t)\rangle&=& \cos(gt)|\frac{1}{2},\frac{1}{2}\rangle_s
422: -
423: i\sin(gt)|\frac{1}{2},-\frac{1}{2}\rangle_s\nonumber \\
424: &=&\cos(gt)|1\rangle_{A}|0\rangle_{B}- i
425: \sin(gt)|0\rangle_{A}|1\rangle_{B}, \label{eq:nn}
426: \end{eqnarray}
427: where subscripts $A$ and $B$ denote dots $1$ and $2$,
428: respectively, and we have omitted the global phase factor of time
429: dependence, $e^{-i(\Omega+2\chi)t}$. The entropy of entanglement
430: can be calculated as
431: \begin{equation}
432: E^{(1)}(t)=-{\cos}^{2}(gt)\ln[\cos^{2}(gt)] -{\sin}^{2}(gt) \ln
433: [\sin^{2}(gt)]. \label{eq:nnn}
434: \end{equation}
435: It is seen that the entropy of the entanglement periodically
436: evolves with zero values at times $gt=k \frac{\pi}{2}$ where $k$
437: is an integer. The entropy of the entanglement reaches its maximum
438: values $\ln2$ at times $gt=(2k+1)\frac{\pi}{4}$ with integer $k$,
439: then the maximally entangled states of the system is
440: \begin{eqnarray}
441: |\Psi^{(1)}(t)\rangle&=& \frac{|\frac{1}{2},\frac{1}{2}\rangle_s
442: +e^{-i 2gt}|\frac{1}{2},-\frac{1}{2}\rangle_s}{\sqrt{2}}
443: =\frac{|1\rangle_{A}|0\rangle_{B} +e^{-i
444: 2gt}|0\rangle_{A}|1\rangle_{B}}{\sqrt{2}}.
445: \end{eqnarray}
446: If there are initially two excitons in microcrystallite $A$, then
447: the wave function of the whole system with the initial condition
448: $|\Psi(t=0)\rangle=|1,1\rangle_s$ becomes
449: \labela%
450: \begin{eqnarray}
451: |\Psi^{(2)}(t)\rangle &=& \alpha_{1}|1,-1\rangle_s
452: +\alpha_{2}|1,0\rangle_s+\alpha_{3}|1,1\rangle_s\nonumber
453: \nonumber \\
454: &=&\alpha_{1}|0\rangle_{A}|2\rangle_{B}+\alpha_{2}
455: |1\rangle_{A}|1\rangle_{B}+\alpha_{3}|2\rangle_{A}|0\rangle_{B}
456: \end{eqnarray}
457: and
458: \labelb%
459: \begin{eqnarray}
460: \alpha_{1}&=&\frac{1}{4}[e^{i2gt}+e^{-i2(g+4\chi)t}-2],
461: \label{eq:x1}
462: \end{eqnarray}
463: \labelc%
464: \begin{eqnarray}
465: \alpha_{2}&=&\frac{\sqrt{2}}{4}[e^{-i2(g+4\chi)t}-e^{i2gt}],
466: \end{eqnarray}
467: \labeld%
468: \begin{eqnarray}
469: \alpha_{3}&=&
470: \frac{1}{4}[e^{i2gt}+e^{-i2(g+4\chi)t}+2]\label{eq:x3}
471: \end{eqnarray}
472: \labelz%
473: where the global phase factor of time dependence $e^{-i 2(\Omega
474: +2\chi)t}$ has also been neglected. We can obtain the entropy of
475: the entanglement as
476: \begin{equation}
477: E^{(2)}(t)=-\sum_{n=1}^{3}|\alpha_{n}|^{2}\ln|\alpha_{n}|^{2},
478: \end{equation}
479: where $\alpha_{n}$ determined by (\ref{eq:x1})--(\ref{eq:x3}).
480: There is no solution for the time $t$ in which the system exactly
481: evolves into a maximally entangled state such that the entropy of
482: entanglement is $\ln 3$. But numerically we find that the
483: difference between the maximum value of the entropy of the
484: entanglement and the entropy of the maximally entangled state is
485: of the order $10^{-5}$. At this point, we say we can nearly
486: obtain the maximally entangled state when the total number of the
487: excitons in the system is equal to two. Figures \ref{fig1} and
488: \ref{fig2} demonstrate the evolution of the entropy of the
489: entanglement as a function of $gt$ with two different sets of
490: parameters for given different initial states. They show that when
491: the second-order coupling between excitons is weak, the period for
492: reaching the maximum values of the entropy of the entanglement
493: becomes longer. It is worth pointing out that if the second-order
494: coupling constant is equal to zero and only linear coupling term
495: remains, then the exciton density is lower and the entanglement is
496: diminished (see solid curve with dots in comparison to other
497: curves in figure 3). So, in order to obtain high entanglement of
498: excitons, we should increase the exciton density in the coupled
499: microcrystallites. The above discussion also shows that the
500: entanglement between two microcrystallites depends on both the
501: ratio $\chi/g$ and initial conditions. By numerical calculations,
502: we can find that the survival time of higher entanglement also
503: depends on $\chi/g$ and initial conditions. The survival time of
504: higher entanglement is longer for the larger ratio $\chi/g$ when
505: there are two excitons initially excited in one of the
506: microcrystallites, but it almost the same for different ratios
507: $\chi/g$ when there is one exciton initially excited in each
508: microcrystallite, respectively.
509: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
510: %figure 1.
511: %---------------------------------------------------------------------------
512: \begin{figure}
513: %\hspace*{4.8cm}{\large $(a)$}\hspace{5.8cm}{\large $(b)$}
514: \hspace{15mm} \epsfxsize=12cm\epsfbox{fig1.eps}
515: \caption[]{$E^{(2,0)}$ is plotted as a function of $gt$ for (a)
516: $\chi/g=0.34$, (b) $\chi/g=0.01$ when two excitons are initially
517: excited in one of the microcrystallites, and $E^{(1,1)}$ is
518: plotted as a function of $gt$ for (c) $\chi/g=0.34$, (d)
519: $\chi/g=0.01$ when there is one exciton initially in each
520: microcrystallite.}\label{fig1}
521: \end{figure}
522:
523: For an arbitrary number of excitons $L$, the wave function of the
524: whole system is described by (\ref{eq:13}). Then the entropy of the
525: entanglement can be calculated as
526: \labela%
527: \begin{equation}
528: E^{(L)}(t)\equiv E^{(L,0)}(t)=
529: -\sum_{m^{\prime}=-\frac{L}{2}}^{\frac{L}{2}}
530: |\beta_{m^{\prime}}|^{2}\ln|\beta_{m^{\prime}}|^{2} \label{eq:g1}
531: \end{equation}
532: with
533: \labelb%
534: \begin{equation}
535: \beta_{m^{\prime}}=\sum_{m=-\frac{L}{2}}^{\frac{L}{2}} {\cal
536: D}^{\frac{L}{2}}_{\frac{L}{2},m}(\frac{\pi}{2})
537: e^{-\frac{i}{\hbar}E_{\frac{L}{2},m}t} {\cal
538: D}^{\frac{L}{2}}_{m^{\prime},m}(\frac{\pi}{2}), \label{eq:g2}
539: \end{equation}
540: \labelz%
541: where ${\cal D}^{\frac{L}{2}}_{m^{\prime},m}(\frac{\pi}{2})$ is
542: determined by (\ref{eq:wg}).
543:
544: Equations(\ref{eq:g1}--\ref{eq:g2}) are confined to the case of the
545: excitons initially excited in one microcrystallite only.
546: Experimentally, we can also initially excite excitons in both
547: microcrystallites. If $P$ and $Q$ excitons are initially excited in
548: microcrystallites $A$ and $B$ then we have $j=(P+Q)/2$ and
549: $m=(P-Q)/2$, and so the initial state can be expressed as
550: $|\Phi(t=0)\rangle=
551: |P\rangle_{A}|Q\rangle_{B}=|\frac{P+Q}{2},\frac{P-Q}{2}\rangle_{s}$.
552: At time $t$, the wave function of the whole system can be written
553: as
554: \begin{eqnarray}
555: |\Phi(t)\rangle&=&\sum_{l^{\prime}=-(P+Q)/2}^{(P+Q)/2}
556: \exp\Big(\frac{t}{i\hbar}E_{\frac{P+Q}{2},l^{\prime}}\Big)\nonumber\\
557: &&\times
558: \langle\Psi_{\frac{P+Q}{2},l^{\prime}}|\frac{P+Q}{2},\frac{P-Q}{2}\rangle_s
559: |\Psi_{\frac{P+Q}{2},l^{\prime}}\rangle_{s}. \label{eq:pq0}
560: \end{eqnarray}
561: We can also obtain the entropy of entanglement for the above
562: initial condition as
563: \labela%
564: \begin{equation}
565: E^{(P,Q)}(t)=-\sum_{n^{\prime}={-(P+Q)}/2}^{(P+Q)/2}
566: |\beta_{n^{\prime}}^{\prime}|^{2}{\rm
567: ln}|\beta_{n^{\prime}}^{\prime}|^{2} \label{eq:pq1}
568: \end{equation}
569: with
570: \labelb%
571: \begin{equation}
572: \beta^{\prime}_{n^{\prime}}=\sum_{n=-(P+Q)/2}^{(P+Q)/2}{\cal
573: D}^{\frac{P+Q}{2}}_{\frac{P-Q}{2},n}(\frac{\pi}{2})e^{-\frac{t}{i\hbar}E_{\frac{P+Q}{2},n}}{\cal
574: D}^{\frac{P+Q}{2}}_{n^{\prime},n}(\frac{\pi}{2}). \label{eq:pq2}
575: \end{equation}
576: \labelz%
577: \vspace*{0mm}
578: \begin{figure}
579: \hspace{15mm} \epsfxsize=12cm\epsfbox{fig2.eps}
580: \caption[]{$E^{(5)}\equiv E^{(5,0)}$, $E^{(4,1)}$, and $E^{(3,2)}$
581: are plotted as a function of $gt$ for $\chi/g=0.34$ (left panel)
582: and $\chi/g=0.01$ (right panel).}\label{fig2}
583: \end{figure}
584: Equations (\ref{eq:pq1}--\ref{eq:pq2}) for the entanglement were
585: obtained from the wave function (\ref{eq:pq0}) being a complete
586: solution of the model. Thus, the contributions of all elements
587: (also those off-diagonal) of the density matrix are included. In \
588: figure~\ref{fig2}, the entropies of the entanglement $E^{(P,Q)}$
589: are depicted as a function of $gt$ for different exciton numbers
590: $P$ and $Q$ in the first and second microcrystallite,
591: respectively, but for the same total number of five excitons in
592: both microcrystallites. Numerical results show that the evolution
593: behavior of the entropy of the entanglement for $L=5$ is similar
594: to that for $L=2$. However, when we compare the maximum value of
595: the entropy of entanglement with the entropy of the maximally
596: entangled state, we see that a time $t$ at which the generated
597: entangled state is very close to a maximally entangled state
598: cannot be found for any $\chi/g$. The survival time of the maximum
599: entanglement is different for different initial conditions and
600: $\chi/g$. The smaller difference of initial populations
601: corresponds to longer time in a larger entanglement area for fixed
602: $\chi/g$ when the total number of excitons in two
603: microcrystallites is five. We also find that the maximum
604: entanglement for fixed $\chi/g$ is the same for different initial
605: conditions for the total number $L$ of excitons in two
606: microcrystallites $L=2,3$ only, i.e.
607: $E_{\max}^{(2,0)}=E_{\max}^{(1,1)} =E_{\max}^{(0,2)}=\ln3$ and
608: $E_{\max}^{(3,0)} =E_{\max}^{(2,1)} =E_{\max}^{(1,2)}
609: =E_{\max}^{(0,3)} = 2 \ln2$. However, if $L>3$ then
610: $E_{\max}^{(L-n,n)}\neq E_{\max}^{(L-m,m)}$ for $n\neq m \leq
611: L/2$. Nevertheless, the differences are relatively small - only
612: at the second digit after the decimal point, as we have checked
613: numerically up to ten excitons. Concluding, we find that the
614: maximum values of the entropy of entanglement depends both on the
615: initial conditions and on $\chi/g$ when $L>3$.
616:
617: We can use equations (\ref{eq:g1}--\ref{eq:g2}) as an example to
618: further discuss the entanglement of the excitons between the two
619: subsystems $A$ and $B$ of the system for any initially given
620: exciton number. Here the entropy of the entanglement of two
621: subsystems is discussed up to ten excitons. An exact time $t$,
622: which corresponds to a maximally entangled state, can be obtained
623: when the system has one exciton. Numerically up to a evolution time
624: of $gt\leq 600$, when the total number of excitons in our system is
625: two or three, we find that the difference between the maximal
626: entropy of entanglement and the entropy of the maximally entangled
627: state is of the order $10^{-5}$ for $\chi/g\neq 0$. So, the
628: maximally entangled states can be approximately obtained. The
629: maximum values $E^{(L)}_{\rm max}$ of the entropy of the
630: entanglement for the two subsystems are also numerically calculated
631: by using (\ref{eq:g1})--(\ref{eq:g2}) up to the time $gt\leq 600$.
632: Figure \ref{fig3} shows a comparison of the maximal values
633: $E^{(L)}_{\rm max}$ of the entropy of the entanglement with the
634: values of the entropy of the entanglement for maximally entangled
635: states up to ten excitons for $\chi/g=0, \ 0.01, \ 0.34, \ 0.8$.
636: Figure \ref{fig3} shows that the entropies of the entanglement for
637: $\chi/g=0$ are smaller than those for $\chi/g=0.01, \ 0.34, \ 0.8$
638: for all exciton numbers, which means our system can achieve a
639: larger entanglement than that of the linear coupling model. At this
640: point, this coupled microcrystallites system can be considered as a
641: better source of entanglement with a fixed exciton number than that
642: of the linear coupling model with the same particle number. For a
643: fixed exciton number, we find that the difference in the maximum
644: values of the entropies of the entanglement between any two
645: different parameter ratios among $\chi/g=0.01, \ 0.34, \ 0.8$ is
646: very small. Every square in the solid curve of figure \ref{fig3}
647: actually denotes the maximum values of the entropies of the
648: entanglement for three different values of $\chi/g$. By comparing
649: plots for a small ($\chi/g=0.01$) and a higher ($0.34$) values of
650: the nonlinearity in figures \ref{fig1}, \ref{fig2}, and
651: \ref{fig3}, it is clearly seen that the global maxima of
652: entanglement are practically independent of the ratio of $\chi/g$
653: (except $\chi/g$), but dependent of the initial conditions.
654: However, higher nonlinearity corresponding to a larger $\chi/g$
655: enables generation of globally higher entanglement for shorter
656: evolution times.
657: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
658: %figure 3.
659: \begin{figure}
660: \hspace*{2cm} \epsfxsize=60mm\epsfbox{fig3.eps}
661: \caption[]{The maximal values $E^{(L)}_{max}$ of the
662: time-dependent von Neumann entropy $E^{(L)}(t)$ are plotted as a
663: function of exciton number $L\leq10$ for $\chi/g=0$ (dots in solid
664: curve), $\chi/g=0.01$, $0.34$, and $0.8$ (squares in solid curve).
665: The values of the von Neumann entropy of the maximally entangled
666: states are marked by circles in the dotted curve for different
667: exciton numbers $L$.}\label{fig3}
668: \end{figure}
669:
670:
671:
672:
673: \section{Oscillations of exciton imbalance}
674:
675: In the previous section, we show that the excitonic state of two
676: microcrystallites system has different degree of entanglement with
677: the evolution. The population of the excitons in the
678: microcrystallites $A$ and $B$ is not at equilibrium. The
679: nano-technology opens up a new direction of research for direct
680: measurements of exciton dynamics. Probing one exciton at a time
681: has been demonstrated by using the nonlinear
682: nano-optics~\cite{NH}. It is possible for experimentalists to
683: observe the population of the excitons in the microcrystallites.
684: Considering that experimentally, it is easier to initially
685: excite excitons in one microcrystallite, in this section our
686: discussions are the case when there are $L$ excitons excited
687: initially in one of the microcrystallites. We will show that the
688: excitonic population imbalance between two coupled
689: microcrystallites exhibits a collapse and revival phenomenon. The
690: time-dependent population difference of excitons between two
691: microcrystallites can be written as
692: \begin{equation}
693: \Delta N^{(L)}(t)=N^{(L)}_{1}(t)-N^{(L)}_{2}(t),
694: \end{equation}
695: where the average exciton number of each microcrystallite
696: $N^{(L)}_{l}(t)=\langle a^{\dagger}_{l}a_{l}\rangle$. The
697: evolution of the population difference of the excitons is similar
698: for different initial conditions. We can discuss the
699: population imbalance of the two microcrystallites using equation
700: (\ref{eq:T}), but here we only consider the initial condition at
701: which there are $L$ excitons in microcrystallite $A$ and no exciton
702: in microcrystallite $B$. From (\ref{eq:13}), we can calculate the
703: number of excitons in each microcrystallite as
704: \begin{eqnarray}
705: &&N^{(L)}_{l}(t)=\frac{L}{2}-(-1)^{l}\sum_{m^{\prime}=
706: -\frac{L}{2}}^{\frac{L}{2}} m^{\prime} \left |\beta_{m^{\prime}}
707: \right |^{2},
708: \end{eqnarray}
709: where $l=1$ represents the microcrystallite $A$ and $l=2$ denotes
710: microcrystallite $B$; time-dependent functions $\beta_{m^{\prime}}$
711: are given by equation (\ref{eq:g2}). Then we can obtain the
712: population difference of the excitons between the two
713: microcrystallites as
714: \begin{equation}
715: \Delta N^{(L)}(t)=2\sum_{m^{\prime}=-\frac{L}{2}}^{\frac{L}{2}}
716: m^{\prime} \left |\beta_{m^{\prime}} \right |^{2}.
717: \end{equation}
718: When only one exciton is initially excited in the microcrystallite
719: $A$, the exciton population difference is
720: \begin{equation}
721: \Delta N^{(1)}(t)=\cos (2gt),
722: \label{eq:n1}
723: \end{equation}
724: which is a simple sinusoidal oscillation. The population difference
725: of the exciton is zero at times $gt=(2k+1)\frac{\pi}{4}$, and
726: $\Delta N^{(1)}$ reaches a maximum at times $gt=k \frac{\pi}{2}$,
727: where $k$ is an integer. Comparing this result to equation
728: (\ref{eq:nnn}), we know that the two subsystems reach the maximal
729: entanglement when the exciton population difference of the two
730: microcrystallites is zero, but when the entropy of the entanglement
731: of the system is minimum, the population imbalance is maximum.
732:
733: When there are two excitons initially excited in
734: microcrystallite $A$, we obtain
735: \begin{equation}
736: \Delta N^{(2)}(t)=2\cos (4\chi t)\cos[2gt +4\chi t]. \label{eq:N}
737: \end{equation}
738: Equation (\ref{eq:N}) indicates that the population difference
739: periodically oscillates with time. We numerically give the
740: evolution of the population difference as a function of $gt$ in
741: figure \ref{fig4}.
742: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
743: %figure 4.
744: \begin{figure}[ht]
745: \hspace{15mm} \epsfxsize=12cm\epsfbox{fig4.eps} \caption[]{The
746: evolution of the population difference $\Delta N^{(2)}$ between two
747: microcrystallites is shown as a function of $gt$ for (a)
748: $\chi/g=0.34$ and (b) $\chi/g=0.01$.}\label{fig4}
749: \end{figure}
750: Figure \ref{fig4} shows that the exciton population difference of
751: the two microcrystallites exhibits the beat effect. If the second
752: interaction between the excitons becomes weak, the period of the
753: envelope becomes longer than that of the stronger second
754: interaction between excitons.
755:
756: The general formula for the exciton population difference of two
757: microcrystallites can be given as
758: \begin{equation}
759: \Delta N^{(L)}(t)= L \cos^{L-1}(4\chi t)\cos [2gt +4(L-1)\chi t].
760: \end{equation}
761: With increasing exciton number, the numerical results
762: show that exciton population difference evolution in time
763: resembles a collapse and revival phenomenon. As an example, the
764: evolution of the exciton population difference in a system of
765: the coupled microcrystallites with a total number of $L=5$ excitons
766: is plotted in figure \ref{fig5} as a function of $gt$. The smaller
767: is the ratio between the second coupling constant and the first
768: coupling constant, the clearer is this phenomenon.
769:
770: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
771: %figure 5.
772: \begin{figure}
773: \hspace{15mm} \epsfxsize=12cm\epsfbox{fig5.eps} \caption[]{The
774: evolution of the population difference $\Delta N^{(5)}$ between two
775: microcrystallites is shown as a function of $gt$ for (a)
776: $\chi/g=0.34$, and (b) $\chi/g=0.01$.}\label{fig5}
777: \end{figure}
778: Figure \ref{fig6} shows a relationship between the entropy of the
779: entanglement in our system and the exciton-population difference.
780: To estimate the uppermost envelope functions of the curves in
781: figure \ref{fig6}, we apply the Jaynes principle of maximum
782: entropy~\cite{Jaynes}. In the Jaynes formalism, the entropy can
783: exclusively be expressed in terms of the mean values. The method
784: corresponds to averaging over the generalized grand canonical
785: ensemble of states satisfying given constraints. Thus, the Jaynes
786: formalism of calculating the maximum entropy strongly resembles the
787: standard problems of entropy maximization in statistical mechanics
788: and thermodynamics. It is worth noting that, although the maximum
789: entropy principle ``has nothing to do with the laws of physics''
790: \cite{Jaynes}, it is a useful and widely applied tool in physics,
791: e.g., in the quantum state reconstruction from data obtained in a
792: measurement of a physical process \cite{Buzek}. By applying the
793: Jaynes principle we find the maximum entropy to be expressed as a
794: function of the population difference as
795:
796: \begin{eqnarray}\label{eq:Jay1}
797: &&E_{J}^{(L)}(\Delta N^{(L)})=-\sum_{n=0}^{L}p_{n}\ln
798: p_{n}=\frac{1}{2}(L+\Delta N^{(L)})\lambda ^{(L)}+\ln Z^{(L)},
799: \end{eqnarray}
800: where $p_{n}=\exp (-n\lambda ^{(L)})/Z^{(L)}\equiv x^{n}/Z^{(L)}$
801: are given in terms of Lagrange multiplier $\lambda ^{(L)}$ or
802: $x=\exp (-\lambda ^{(L)}) $ and the generalized partition function
803:
804: \begin{eqnarray}\label{eq:Jay2}
805: Z^{(L)}=\frac{x^{L+1}-1}{x-1}.
806: \end{eqnarray}
807: The Lagrange multiplier $\lambda ^{(L)}$ can be calculated from
808:
809: \begin{eqnarray}\label{eq:Jay3}
810: \Delta N^{(L)}=f(x)=L+\frac{2}{1-x}-2\frac{1+L}{1-x^{L+1}}.
811: \end{eqnarray}
812: To obtain an explicit formula for $E_{J}^{(L)}$ one has to find the
813: relation inverse to $\Delta N^{(L)}=f(x)$. By closer inspection of
814: (\ref{eq:Jay3}), we conclude that the analytical solutions of
815: $x=f^{-1}(\Delta N^{(L)})$ exist only for $L$=2, 3, and 4. For
816: example, the desired solution of $x=f^{-1}(\Delta N^{(2)})$ for
817: $N=2$ is given by
818:
819: \begin{equation}\label{eq:Jay4}
820: x\equiv \exp (-\lambda ^{(2)})=\frac{\Delta
821: N^{(2)}+\sqrt{16-3(\Delta N^{(2)})^{2}}}{2(2-\Delta N^{(2)})}.
822: \end{equation}
823: For brevity, we do not present our analytical, but rather
824: complicated, solutions for $L=3$ and $4$. The solutions for $L\geq
825: 5$ are to be found numerically. The curves marked by circles in
826: figure 6 correspond to our analytical ($L=2, 3, 4$) and numerical
827: ($L=5$) estimations of the maximum entropy of entanglement as a
828: function of $\Delta N^{(L)}$. It is seen, in agreement with our
829: analysis of figure 3, that with the increasing number $L$ of
830: excitons, the time-maximized entanglement $E^{(L)}_{\max}(t)$,
831: which can be generated in our system, increases in comparison to
832: the lower-exciton systems $E^{(L)}_{\max}(t)> E^{(L-1)}_{\max}(t)$
833: for the same population difference $\Delta N\equiv \Delta
834: N^{(L)}=\Delta N^{(L-1)}\neq \pm $ $L$ (with any $L$) for a given
835: $\kappa/g$. The states described by $E_{J}^{(L)}(\Delta N^{(L)})$
836: for $\Delta N=0$ correspond to the maximally entangled states. The
837: numerical results depicted in figure \ref{fig6} show that the
838: maximum entropy of the entanglement always corresponds to the
839: points of the exciton-population balance. However, the condition of
840: the exciton-population balance does not always imply the maximum
841: entropy of the entanglement. For any population-imbalance points,
842: the entropy of the entanglement is smaller, and when the population
843: difference reaches its maximum, the entropy of the entanglement
844: vanishes. \vspace{0.5cm}
845: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
846: %figure 6.
847: \begin{figure}
848: \hspace*{2cm} \epsfxsize=80mm\epsfbox{fig6.eps}
849: \caption[]{The entropy of entanglement is plotted as a function of
850: the population difference of microcrystallites for $\chi/g=0.34$
851: when the total number $L$ of excitons is 2, 3, 4, or 5. The
852: uppermost circles represent values estimated by the Jaynes
853: principle.}\label{fig6}
854: \end{figure}
855:
856: \section{Conclusions}
857:
858: We have discussed the entanglement of the excitonic states in the
859: system of the coupled microcrystallites with a fixed total exciton
860: number by the entropy of the entanglement. When the total number
861: of excitons is one, the maximally entangled state can exactly be
862: obtained. The maximally entangled states can approximately be
863: generated for two or three excitons. However, when the exciton
864: number is more than three, we cannot obtain the maximally
865: entangled state, but it is observed that the nonlinear interaction
866: between excitons increases the maximum values of the entropy of
867: the entanglement to higher values than those of the other models
868: based on the linear coupling. So this exciton system can be used
869: as a good source of entanglement. We also find that the
870: entanglement between the two microcrystallites depends on the
871: initial conditions, the different initial conditions will result
872: in different entanglements.
873:
874:
875: The oscillations of exciton population between two
876: microcrystallites have also been investigated analytically and
877: numerically. If the system is prepared with one exciton, the
878: population imbalance between two microcrystallites reaches the
879: maximum when the entanglement of the two microcrystallites is the
880: minimum; and the population reaches the balance when the entropy of
881: the entanglement of the system is maximum. But when the exciton
882: number is more than two, the numerical results showed that we could
883: find the maximum entropy of the entanglement at the points of the
884: exciton population balance. At some larger population imbalance
885: points, the entropy of the entanglement was found to be smaller.
886: The reason is that when the population imbalance becomes the
887: largest, all excitons are concentrated on one microcrystallite,
888: then the entanglement between two microcrystallites decreases. But
889: in the collapse area, the average population of each
890: microcrystallite is equal, so it is possible to find the maximal
891: entropy of the entanglement. We hope these phenomena can be
892: observed by experimentalists with the development of the
893: nano-technology.
894:
895: The main conclusion of our paper is that nonlinear interactions,
896: resulting from high exciton density between microcrystallites, can
897: increase exciton entanglement in comparison to a model of
898: linearly-coupled microcrystallites with lower exciton density. We
899: believe that this result could be important for
900: semiconductor-based quantum information processing. Our conclusion
901: is not only valid for the studied system, but can also be
902: generalized to other Kerr nonlinear vs linear interaction models.
903:
904:
905: \section*{Acknowledgments}
906: Authors are grateful to Makoto Kuwata-Gonokami and Yuri P. Svirko
907: for helpful discussions. Yu-xi Liu acknowledges the support of
908: Japan Society for the Promotion of Science (JSPS).
909:
910: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
911: \section*{References}
912:
913: \begin{thebibliography}{99}
914:
915: \bibitem{Zeilinger} Bouwmeester D, Pan J-W, Daniel M,
916: Weinfurter H and Zeilinger A 1997 {\em Nature (London)} {\bf 390}
917: 575
918:
919: \bibitem{B} Furusawa A, S\o rensen J I, Braunstein S L, Fuchs C A,
920: Kimble H J and Polzik E S 1998 {\em Science} {\bf 282} 706
921:
922: \bibitem{mmm} Duan L M, Giedke G, Cirac J I and Zoller P
923: 2000 {\em Phys. Rev. Lett.} {\bf 84} 4002
924:
925: \par\item[] Duan L M, Giedke G, Cirac J I and Zoller P
926: 2000 {\em Phys. Rev. A} {\bf 62} 032304
927:
928: \bibitem{c} Cochrane P T, Milburn G J and Munro W J
929: 2000 {\em Phys. Rev. A} {\bf 62} 062307
930:
931: \bibitem{4} Barenco A, Deutsh D, Ekert A and Jozsa R 1995
932: {\em Phys. Rev. Lett.} {\bf 74} 4083
933:
934: \par\item[] Loss D and DiVincenzo P 1998
935: {\em Phys. Rev. A }{\bf 57} 120
936:
937: \par\item[] Imamo\v{g}lu A, Awschalom D D, Burkard G, DiVincenzo D P,
938: Loss D, Sherwin M and Small A 1999 {\em Phys. Rev. Lett.} {\bf
939: 83} 4204
940:
941: \bibitem{44} Oh J H, Ahn D and Hwang S W 2000 {\em Phys. Rev. A} {\bf 62}
942: 052306
943:
944: \bibitem{5} Chen G, Bonadeo N H, Steel D G, Gammon D,
945: Katzer D S, Park D and Sham L J 2000 Science {\bf 289} 1906
946:
947: \bibitem{6} Bayer M, Hawrylak P, Hinzer K, Fafard S,
948: Korkusinski M, Wasilewski Z R, Stern O and Forchel A 2001 Science
949: {\bf 291} 451
950:
951: \bibitem{7} Quiroga L and Johnson N F 1999
952: {\em Phys. Rev. Lett.} {\bf 83} 2270
953:
954: \par\item[] Reina J H, Quiroga L and Johnson N F 2000 {\em Phys. Rev. A}
955: {\bf 62} 012305
956:
957: \bibitem{99} Yi X X, Jin G R, Zhou D L 2001 {\em Phys. Rev. A} {\bf 63}
958: 062307
959:
960: \bibitem{yu-xi2}
961: Miranowicz A, \"Ozdemir \c{S} K, Liu Yu-xi, Koashi M, Imoto N and
962: Hirayama Y 2002 {\em Phys. Rev. A} {\bf 65} 062321
963:
964: \bibitem{wang}Wang X, Feng M and Sanders B C 2003 {\em Phys.
965: Rev. A} {\bf 67} 022302
966:
967: \bibitem{3} Livermore C, Crouch C H, Westervelt R M,
968: Campman K L and Gossard A C 1996 Science {\bf 274} 1332
969:
970: \bibitem{hirayama} Fujisawa T, Oosterkamp T H, van der Wiel W G,
971: Broer B W, Aguado R, Tarucha S, Kouwenhoven L P 1998 Science {\bf
972: 282} 932
973:
974: \par\item[] Oosterkamp T H, Fujisawa T, van der Wiel W G, Ishibashi K,
975: Hijman R V, Tarucha S and Kouwenhoven L P 1998 Nature {\bf 395}
976: 873
977:
978: \bibitem{Han} Hanamura E 1988 Phys Rev B {\bf 38} 1228
979:
980: \bibitem{Ban} Belleguie L and B\'anyai L 1991 {\em Phys. Rev. B}
981: {\bf 44} 8785
982:
983: \par\item[] B\'anyai L and Koch S W 1993 {\em Semiconductor Quantum Dots}
984: (Singapore: World Scientific)
985:
986: \bibitem{Eng} Engelmann A, Yudson V I and Reineker P 1998
987: {\em Phys. Rev. B} {\bf 57} 1784
988:
989: \bibitem{yu-xi}
990: Liu Yu-xi, \"Ozdemir S K, Koashi M and Imoto N 2002 {\em Phys.
991: Rev. A} {\bf65} 042326
992:
993: \par\item[] Liu Yu-xi, Miranowicz A, Koashi M and Imoto N 2002 {\em Phys.
994: Rev. A} {\bf 66} 062309
995:
996: \par\item[] Liu Yu-xi, Miranowicz A, \"Ozdemir S K, Koashi M and Imoto N
997: 2003 {\em Phys. Rev. A} {\bf 67} 034303
998:
999: \bibitem{d} Chernyak V and Mukamel S 1996 {\em J. Opt. Soc. Am. B}
1000: {\bf 13} 1302
1001:
1002: \bibitem{Sch65} Schwinger J 1965, US Atomic Energy Commission Report
1003: No. NYO-3071 (US GPO, Washington, D.C.); reprinted in Biedenharn L
1004: C and van Dam H (eds) 1965 {\em Quantum Theory of Angular Momentum}
1005: (New York: Academic) p 229
1006:
1007: \bibitem{Sak94}Sakurai J J 1994 {\em Modern Quantum Mechanics}
1008: (New York: Addison-Wesley)
1009:
1010: \bibitem{Ben96a}Bennett C H, Bernstein H J, Popescu S, and Schumacher B 1996 {\em
1011: Phys. Rev. A} {\bf 53} 2046
1012:
1013: \bibitem{Ben96b}Bennett C H, DiVincenzo D P, Smolin J A, and Wootters W K 1996
1014: {\em Phys. Rev. A} {\bf 54} 3824
1015:
1016: \bibitem{Ved98}Vedral V and Plenio M B 1998 {\em Phys. Rev. A} {\bf 57} 1619
1017:
1018: \bibitem{Pho91}Phoenix S J D and Knight P L 1991 {\em Phys. Rev. A} {\bf 44}
1019: 6023; 1991 {\em Phys. Rev. Lett.} {\bf 66} 2833; 1988 {\em Ann.
1020: Phys.} {\bf 186} 381
1021:
1022: \bibitem{NH} Bonadeo N H, Chen Gang, Gammon D, Katzer D S,
1023: Park D and Steel D G 1998 {\em Phys. Rev. Lett.} {\bf 81} 2759
1024:
1025: \bibitem{Jaynes}
1026: Jaynes E T 1957 {\em Phys. Rev.} {\bf 106} 620; {\bf 108} 171
1027:
1028: \bibitem{Buzek} Bu\v{z}ek V, Adam G and Drobn\'{y} G 1996
1029: {\em Phys. Rev. A} {\bf 54} 804
1030:
1031: \end{thebibliography}
1032: \end{document}
1033: