1:
2: \documentclass[twocolumn,showpacs,preprintnumbers]{revtex4}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: \usepackage{amssymb}
5: \usepackage{amsmath}
6: \usepackage{graphicx}
7: \usepackage{dcolumn}
8: \usepackage{bm}
9:
10: \setcounter{MaxMatrixCols}{10}
11: %TCIDATA{OutputFilter=LATEX.DLL}
12: %TCIDATA{Version=5.50.0.2953}
13: %TCIDATA{<META NAME="SaveForMode" CONTENT="1">}
14: %TCIDATA{BibliographyScheme=BibTeX}
15: %TCIDATA{LastRevised=Friday, March 31, 2006 13:54:08}
16: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
17: %TCIDATA{ComputeDefs=
18: %$\rho _{ap}=c_{0}\rho _{0}+c_{1}\rho _{1}$
19: %}
20:
21:
22: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
23: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
24: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
25: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
26: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
27: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
28: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
29: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
30: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
31: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
32: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
33: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
34: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
35: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
36: \def\CTeXPreproc{Created by ctex v0.2.5, don't edit!}
37:
38:
39: \begin{document}
40:
41: \preprint{APS/123-QED}
42: \title{Decoherence-free quantum memory for photonic state using atomic
43: ensembles\\
44: }
45: \author{Feng Mei}
46: \author{Ya-Fei Yu}
47: \email{yfyuks@hotmail.com}
48: \author{Zhi-Ming Zhang}
49: \email{zmzhang@scnu.edu.cn}
50: \affiliation{Laboratory of Photonic Information Technology, School for Information and
51: Photoelectronic Science and Engineering, South China Normal University,
52: Guangzhou 510006, PR China}
53: \date{November 4, 2008}
54:
55: \begin{abstract}
56: Large scale quantum information processing requires stable and long-lived
57: quantum memories. Here, using atom-photon entanglement, we propose an
58: experimentally feasible scheme to realize decoherence-free quantum memory
59: with atomic ensembles, and show one of its applications, remote transfer of
60: unknown quantum state, based on laser manipulation of atomic ensembles,
61: photonic state operation through optical elements, and single-photon
62: detection with moderate efficiency. The scheme, with inherent
63: fault-tolerance to the practical noise and imperfections, allows one to
64: retrieve the information in the memory for further quantum information
65: processing within the reach of current technology.
66: \end{abstract}
67:
68: \pacs{03.65.Ud,03.67.-a,42.50.Gy}
69: \maketitle
70:
71: % Force line breaks with \\
72:
73: % It is always \today, today,
74: % but any date may be explicitly specified
75:
76: % PACS, the Physics and Astronomy
77: % Classification Scheme.
78: %\keywords{Suggested keywords}%Use showkeys class option if keyword
79: %display desired
80:
81: Quantum information science involves the storage, manipulation and
82: communication of information encoded in quantum systems. In the future, an
83: outstanding goal in quantum information science is the faithful mapping of
84: quantum information between a stable quantum memory and a reliable quantum
85: communication channel. The quantum memory is a key element of quantum
86: repeaters \cite{1} that allow for long-distance quantum communication over
87: realistic noisy quantum channel, which is also necessary for scalable linear
88: optics quantum computation put forward by Knill et al. \cite{2}. Atomic
89: systems are excellent quantum memories, because appropriate internal
90: electronic states can coherently store qubits over very long timescales.
91: Photons, on the other hand, are the natural platform for the distribution of
92: quantum information between remote qubits, by considering their ability to
93: transmit large distance with little perturbation \cite{3}. Recently, various
94: quantum memory schemes for storing photonic quantum states have been
95: proposed, involving employing one atom or two atoms in a high-Q cavity \cite%
96: {4}, all optical approaches \cite{5}, and coupling the light into the atomic
97: ensembles \cite{6}. However, the above schemes have several disadvantages.
98: It is hard to efficiently couple a photon with a atom in a high-fineness
99: cavity, all optical approaches have large transmission loss. Given these
100: drawbacks it is of interest to explore the alternatives.
101:
102: In this paper, different from the quantum memory schemes using
103: electromagnetically induced transparency (EIT) \cite{6}, we propose a scheme
104: to achieve the decoherent-free quantum memory with atom ensembles and linear
105: optics by quantum teleportation. By using the photonic qubits as the
106: information carriers and the collective atomic qubits as the quantum memory,
107: the information can be encoded into the remote decoherence-free subspaces
108: (DFS) \cite{7,8} of quantum memory via teleportation of an arbitrarily
109: prepared quantum state. Compared with the above proposals, our protocol has
110: the following significant advances: (i) It is not necessary to employ a
111: cavity which works in the strong coupling regime. Instead, sufficiently
112: strong interaction is achieved due to the collective enhancement of the
113: signal-to-noise ratio \cite{9,10,11}. (ii) Different from all optical
114: approaches, the fidelity of our atomic memory is insensitive to photon loss.
115: The photon loss only influences the probability of success. (iii) We use two
116: atomic ensembles to encode a single logic qubit, the information is stored
117: in a decoherent-free subspace (DFS) which can protect quantum information
118: against decoherence effectively. (iv) Raman transition technique provides a
119: controllable and decoherence-insensitive way of coupling between light and
120: atoms, so the information storing in the quantum memory can be retrieved
121: with ease for further quantum information processing. (v) Finally, by
122: introducing the photon an additional degree of freedom with spatial modes,
123: we get complete Bell state measurement (BSM) and improve the preparation
124: efficiency a lot.
125:
126: Before describing the detailed model, first we summarize the basic ideas of
127: the scheme, which consists of four steps shown in Fig. 1. (A) Firstly, the
128: entanglement between atomic ensembles and Stokes photon is generated. (B)
129: The spatial mode is employed as additional degree of freedom of photon to
130: encode the information we want to imprint. (C) Next, complete Bell state
131: measurement is performed on the polarization and spatial qubits of photon.
132: (D) Via the BSM results, the information will be faithfully encoded into the
133: decoherent-free subspace (DFS) of atomic ensembles.
134:
135: \textit{Step }(A) --- Entanglement between atom ensembles and photon \cite%
136: {12} is crucial to achieve this task. In more detail, the basic element is a
137: cloud of $N$ identical atoms with the relevant level structure shown in Fig.
138: 1(a). The metastable lower states $\left\vert g\right\rangle $ and $%
139: \left\vert s\right\rangle $ can correspond to hyperfine or Zeeman sublevels
140: of the ground state of alkali-metal atoms, which have an experimentally
141: demonstrated long lifetimes \cite{13,14}. To achieve effectively enhanced
142: coupling to light, the atom ensembles should be preferably placed with
143: pencil-shape. Initially, all the atoms are prepared in the ground state $%
144: \left\vert g\right\rangle $. Shining a synchronized short, off-resonant pump
145: pulse\emph{\ }into the atomic ensemble $j\left( j=L\text{ or }R\right) $
146: induces Raman transitions into the state $\left\vert s\right\rangle $. The
147: emission of single Stokes photon results in the state $S_{j}^{+}\left\vert
148: 0_{a}\right\rangle _{j}$ of atomic ensembles, where the ensemble ground
149: state $\left\vert 0_{a}\right\rangle $ $=\otimes _{i}\left\vert
150: g\right\rangle _{i},$the symmetric collective mode $S=$ $\left( 1/\sqrt{N_{a}%
151: }\right) \sum_{i}\left\vert g\right\rangle _{i}\left\langle s\right\vert $
152: \cite{15}. By the selection rules and the conservation of angular momentum,
153: the pump pulse and the Stokes photon have the left $\left( L\right) $ and
154: right $\left( R\right) $ circular polarization. Assume that the interaction
155: time is short, so, the mean number of the forward-scattered Stokes photon is
156: much smaller than 1. We can define signal light mode bosonic operator $a$
157: for the Stokes pulse with its vacuum state denoted by $\left\vert
158: 0_{p}\right\rangle $, where $a^{+}\left\vert 0_{p}\right\rangle =\left\vert
159: R\right\rangle $. The symmetric collective mode $S$ and the signal light
160: mode $a$ are correlated with each other, which means, if the atomic ensemble
161: is excited to the symmetric collective mode $S^{+}$, the accompanying
162: emission photon will go to the signal light mode $a^{+}$, and vice versa.
163: The whole state of atomic ensemble and the Stokes photon can be written as \
164: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
165: \begin{equation}
166: \left\vert \phi \right\rangle _{j}=\left\vert 0_{a}0_{p}\right\rangle _{j}+%
167: \sqrt{p_{cj}}S_{j}^{+}a_{j}^{+}\left\vert 0_{a}0_{p}\right\rangle
168: _{j}+o(p_{cj}), \label{1}
169: \end{equation}%
170: where $p_{cj}=4g_{c}^{2}N_{j}L_{j}/c\left\vert \Omega \right\vert ^{2}/%
171: \mathbf{\bigtriangleup }^{2}t_{p}$ is the small excitation probability of
172: single spin flip in the ensemble $j$ \cite{15,16}. Here $g_{c}$ is
173: atom-field coupling constant, $N_{j}$ and $L_{j}$ are the linear density and
174: the length of the atomic ensemble $j$. The probability can be controlled by
175: adjusting the light-atom interaction time and pulse duration $t_{p}$. $%
176: o(p_{cj})$ represents its more excitations whose probabilities are equal to
177: or smaller than $p_{cj}^{2}$ \cite{15}. We should note that the scattered
178: photon goes to some other optical modes other the signal mode. However, when
179: $N$ is large, the independent spontaneous emissions distribute over all the
180: atomic modes, whereas the contribution to the signal light mode will be
181: small \cite{10}. So the use of atomic ensembles will result in a large
182: signal-to-noise ratio \cite{9,10,11} and improve the efficiency of the
183: scheme.
184:
185: \begin{figure}[tbp]
186: \includegraphics [width=9cm,height=4cm]{generation}
187: \caption{Schematic setup for quantum memory scheme. (a) The relevant atomic
188: level structure in the ensembles with $\left\vert e\right\rangle $ the
189: excited state, $\left\vert g\right\rangle $ the ground state and $\left\vert
190: s\right\rangle $ the metastable state for storing a qubit of information.
191: The transition $\left\vert g\right\rangle \rightarrow \left\vert
192: e\right\rangle $ can be coupled through the left circular classical pump
193: pulse with the Rabi frequency\textit{\ }$\Omega $ and a detuning $\mathbf{%
194: \bigtriangleup }$, and the forward-scattered Stokes light comes from the
195: transition $\left\vert e\right\rangle \rightarrow \left\vert s\right\rangle $%
196: , which has a right circular polarization. (b) The forward-scattered Stokes
197: fields are collected after the filters which are polarization- and
198: frequency- selective to filter the pumping light, and interfere at PBS. The
199: output Stokes photon is coupled into a single-mode optical fiber and guided
200: to the setup for preparing the information and subsequent
201: polarization-spatial Bell-state measurement. The phase setting $\left(
202: \protect\alpha ,\protect\beta \right) $ allows to prepare any superposition
203: of the spatial-mode qubit. Based on the PBSs and single-photon detectors, we
204: can get the BSM in the polarization-spatial-mode Hilbert space of the Stokes
205: photon.}
206: \end{figure}
207:
208: We can make $p_{cL}=p_{cR}$, the two emitted Stokes pulses are interfered on
209: a polarized beam splitter (PBS) after transmitting a quarter wave plate
210: (QWP) shown in Fig. 1(b). The PBS transmits only H and reflects V
211: polarization component, the QWP transforms the circularly polarized photon
212: into linearly polarized photon by the operator $P_{H}^{+}=\left\vert
213: H\right\rangle \left\langle R\right\vert $ and $P_{V}^{+}=\left\vert
214: V\right\rangle \left\langle L\right\vert $. The small fraction of the
215: transmitted classical pulses can be easily filtered through the filters. For
216: the Stokes pulse from the atomic ensemble $R$, a polarization rotator
217: \textbf{R} is inserted after the QWP. The function of the rotator is defined
218: by the operator $\overset{\sim }{R}=\left\vert H\right\rangle \left\langle
219: V\right\vert +$ $\left\vert V\right\rangle \left\langle H\right\vert $. By
220: selecting orthogonal polarization, conditional on a single-photon detector
221: click \cite{17}, the whole state of the atomic ensembles and the Stokes
222: photon evolves into a maximal entangled state
223:
224: \begin{equation}
225: \left\vert \Psi \right\rangle _{ap}=\left( S_{L}^{+}+S_{R}^{+}\overset{\sim }%
226: {R}\right) /\sqrt{2}\left\vert vac\right\rangle _{ap}\text{.} \label{2}
227: \end{equation}%
228: Because during the Stokes photons from the ensemble $L$ and $R$ interfere in
229: the input of the PBS, the information of their paths is erased. It denotes $%
230: \left\vert vac\right\rangle _{ap}\equiv \left\vert 0_{a}\right\rangle
231: _{L}\left\vert 0_{a}\right\rangle _{R}\left\vert H\right\rangle $. Then, the
232: entangled state can be rewritten as
233: \begin{equation}
234: \left\vert \Psi \right\rangle _{ap}=\left( \left\vert H\right\rangle
235: \left\vert 1\right\rangle _{a}+\left\vert V\right\rangle \left\vert
236: 0\right\rangle _{a}\right) /\sqrt{2}\text{,} \label{3}
237: \end{equation}%
238: where $\left\vert 0\right\rangle _{a}=\left\vert 0_{a}\right\rangle
239: _{L}\left\vert 1_{a}\right\rangle _{R}$, $\left\vert 1\right\rangle
240: _{a}=\left\vert 1_{a}\right\rangle _{L}\left\vert 0_{a}\right\rangle _{R}$
241: denote one spin flip in one of the ensembles. In fact, the generated
242: entangled state will be mixed with a small vacuum component if we take into
243: account the detector dark counts. However, the vacuum component is typically
244: much smaller than the repetition frequency of the Raman pulses. Here, we can
245: neglect this small vacuum component and the high order terms.
246:
247: In the following, we will use the atomic ensembles qubit of the atom-photon
248: entanglement as our quantum memory. Long-lived quantum memory is the
249: keystone of quantum repeater. Unfortunately, due to environmental coupling,
250: the stored information can be destroyed, so-called decoherence. Here,
251: decoherence-free subspace (DFS) has been introduced to protect fragile
252: quantum information against detrimental effects of decoherence. To establish
253: the DFS, we utilize the state of a pair of atomic ensembles to encode a
254: single logic qubit, i.e., $\left\vert 0\right\rangle _{a}=\left\vert
255: 0_{a}\right\rangle _{L}\left\vert 1_{a}\right\rangle _{R}$, $\left\vert
256: 1\right\rangle _{a}=\left\vert 1_{a}\right\rangle _{L}\left\vert
257: 0_{a}\right\rangle _{R}$, so that the phase noise can be effectively
258: suppressed.
259:
260: \textit{Step }(B) --- After the generation of\ the entanglement, the emitted
261: Stokes photon is coupled into a fiber and guided to the setup illustrated in
262: Fig. 1(b), where the state we want to imprint into the quantum memory is
263: prepared. The Hilbert space of the photon is extended by using two spatial
264: modes as an additional degree of freedom \cite{18,19}. The photon is
265: coherently splitted into two spatial modes $\left\vert a\right\rangle $ and $%
266: \left\vert b\right\rangle $, remains in the state $\left\vert \varphi
267: \right\rangle =\alpha \left\vert a\right\rangle +\beta \left\vert
268: b\right\rangle $ by a polarization independent Mach-Zehnder interferometer.
269: The phase setting $\left( \alpha ,\beta \right) $ is determined by the
270: optical path-length difference.
271:
272: \textit{Step }(C) --- Critical to the third step of our scheme is the
273: following identity:%
274: \begin{align}
275: \left\vert \varphi \right\rangle \left\vert \Psi \right\rangle _{ap}&
276: =\left( \alpha \left\vert a\right\rangle +\beta \left\vert b\right\rangle
277: \right) \otimes \left( \left\vert H\right\rangle \left\vert 1\right\rangle
278: _{a}+\left\vert V\right\rangle \left\vert 0\right\rangle _{a})/\sqrt{2}%
279: \right) \notag \\
280: & =\frac{1}{2}(\left\vert \Psi ^{+}\right\rangle \left\vert \widetilde{%
281: \varphi }\right\rangle +\left\vert \Psi ^{-}\right\rangle \hat{\sigma}%
282: _{z}\left\vert \widetilde{\varphi }\right\rangle \notag \\
283: & +\left\vert \Phi ^{+}\right\rangle \hat{\sigma}_{x}\left\vert \widetilde{%
284: \varphi }\right\rangle +\left\vert \Phi ^{-}\right\rangle \left( i\hat{\sigma%
285: }_{y}\right) \left\vert \widetilde{\varphi }\right\rangle )\text{,}
286: \label{4}
287: \end{align}%
288: where $\left\vert \Psi ^{\pm }\right\rangle =$ $\left( \left\vert
289: H\right\rangle \left\vert b\right\rangle \pm \left\vert V\right\rangle
290: \left\vert a\right\rangle \right) /\sqrt{2}$ and $\left\vert \Phi ^{\pm
291: }\right\rangle =\left( \left\vert H\right\rangle \left\vert a\right\rangle
292: \pm \left\vert V\right\rangle \left\vert b\right\rangle \right) /\sqrt{2}$
293: denote the four polarization-spatial Bell states, $\left\vert \widetilde{%
294: \varphi }\right\rangle =\alpha \left\vert 0\right\rangle _{a}+\beta
295: \left\vert 1\right\rangle _{a}$. We make a joint Bell state measurement on
296: the polarization-spatial qubits. To achieve the BSM, the two spatial modes
297: are combined in PBS1, and the polarization of photon will be analyzed in
298: each output (see Fig. 1(b)). The half-wave plate (HWP) performs a Hardmard
299: rotation $\left\vert H\right\rangle \rightarrow \left( \left\vert
300: H\right\rangle +\left\vert V\right\rangle \right) /\sqrt{2},\left\vert
301: V\right\rangle \rightarrow \left( \left\vert H\right\rangle -\left\vert
302: V\right\rangle \right) /\sqrt{2}$ on the polarization modes. If the detector
303: $D_{1}$ has a click, then it denotes identification of one of the Bell state
304: $\left\vert \Psi ^{+}\right\rangle =\left( \left\vert H\right\rangle
305: \left\vert b\right\rangle +\left\vert V\right\rangle \left\vert
306: a\right\rangle \right) /\sqrt{2}$. Because after the combining of the two
307: spatial modes on PBS1, the photonic orthogonal polarization can be
308: coherently superposed into $\left( \left\vert H\right\rangle +\left\vert
309: V\right\rangle \right) /\sqrt{2}$, which is rotated to $H$ by HWP1 and then
310: triggers the detector $D_{1}$. Accordingly, the click of the detector $D_{2}$%
311: , $D_{3}$ and $D_{4}$ corresponds to $\left\vert \Psi ^{-}\right\rangle
312: ,\left\vert \Phi ^{+}\right\rangle $ and $\left\vert \Phi ^{-}\right\rangle $%
313: , respectively. After the BSM, i.e. the confirmation of only one click from
314: the single photon detectors, the encoded information is transferred and
315: encoded into the DFS of atomic ensembles.
316:
317: \textit{Step }(D) --- According to the outcome of BSM, in the standard
318: teleportation, a proper local Pauli unitary operation should be carried out
319: on the atomic quantum memory to recover the original information. However,
320: it is worth noticing the difficulty of laser manipulation to get the unitary
321: operation of atomic ensembles. Thanks to the ease of performing precise
322: unitary transformation on photon, we take the manner of marking instead of
323: the recovering operation in the standard teleportation technique. The
324: quantum memory can be made four different marks in a classic way, depending
325: on the relevant BSM results. When there is a need to retrieve the
326: information from the quantum memory, we can simultaneously shine a weak
327: retrieval pulse with suitable frequency and polarization \cite{20} into the
328: atomic ensembles. The emitted anti-Stokes fields are then combined on PBS.
329: As a result, the atomic qubit is converted back to single photon qubit. The
330: efficiency of the transfer is close to unity at a single quantum level owing
331: to the collective enhancement. Finally, we just apply corresponding unitary
332: operation in agreement with the prior mark to recover the information. In
333: fact, the information storing in the atomic ensembles is just the original
334: one up to the local Pauli operation. The real formation $\left( \alpha
335: ,\beta \right) $ we stored has not been changed. The recovering operation is
336: just suspended to facilitate the experimental realization within our
337: technique.
338:
339: \begin{figure}[tbp]
340: \includegraphics[width=5cm]{instead}
341: \caption{(a) The relevant atomic level structure in the ensembles with $%
342: \left\vert g\right\rangle $ the ground state, $\left\vert e\right\rangle $
343: the excited state, and $\left\vert 0\right\rangle $ and $\left\vert
344: 1\right\rangle $ the metastable state for storing a qubit of information.
345: The transition $\left\vert g\right\rangle \rightarrow \left\vert
346: e\right\rangle $ can be coupled through the classical laser pulse with the
347: Rabi frequency\textit{\ }$\Omega (t)$ and a detuning $\mathbf{\bigtriangleup
348: }$, and the forward-scattered Stokes photons come from the transition $%
349: \left\vert e\right\rangle \rightarrow \left\vert 0\right\rangle $ and $%
350: \left\vert e\right\rangle \rightarrow \left\vert 1\right\rangle $, which are
351: left- and right-circularly polarized. (b) The forward-scattered Stokes
352: photon is collected after the filter which is frequency selective to
353: separate the pumping light, and transmits the quarter wave plate (QWP). Then
354: the entanglement of atomic ensemble and photon is generated.}
355: \end{figure}
356:
357: On the other hand, the two atomic ensembles $L$ and $R$ can also be replaced
358: by one ensemble, but with two metastable states $\left\vert 0\right\rangle $
359: and $\left\vert 1\right\rangle $ to store the quantum information shown in
360: Fig. 2. The states $\left\vert g\right\rangle $, $\left\vert 0\right\rangle $
361: and $\left\vert 1\right\rangle $ correspond to the hyperfine or the Zeeman
362: sublevels of alkali atoms in the ground-state manifold, and $\left\vert
363: e\right\rangle $ corresponds to an excited state. The $N$ atoms are
364: initially prepared in the ground state $\left\vert G_{a}\right\rangle $ $%
365: =\otimes _{i}\left\vert g\right\rangle _{i}$. The transition $\left\vert
366: g\right\rangle \rightarrow \left\vert e\right\rangle $ is driven
367: adiabatically by a weak classical laser pulse with the corresponding Rabi
368: frequency denoted by $\Omega (t)$. With the short off-resonant driving
369: pulse, only one atom is transferred nearly with unit probability to the
370: excited state $\left\vert e\right\rangle $. The excited state will transit
371: into the metastable states $\left\vert 0\right\rangle $ or $\left\vert
372: 1\right\rangle $ with equal probabilities by emitting a left- or
373: right-circularly polarized Stokes photon in the forward direction. Such
374: emitting events are uniquely correlated with the excitation of the symmetric
375: collective atomic mode $S_{h}$ which is given by $S_{h}=$ $\left( 1/\sqrt{%
376: N_{a}}\right) \sum_{i}\left\vert g\right\rangle _{i}\left\langle
377: h\right\vert $ $(h=0,1)$. The emission of single Stokes photon will result
378: in the state of atomic ensembles by $\left\vert h\right\rangle
379: _{a}=S_{h}^{+}\left\vert 0_{a}\right\rangle .$We also can define single mode
380: bosonic operator $a_{h}$ for the Stokes pulse with its vacuum state denoted
381: by $\left\vert 0\right\rangle _{p}$. The emitting process can be defined by $%
382: \left\vert L\right\rangle =a_{0}^{+}\left\vert 0_{p}\right\rangle $ and $%
383: \left\vert R\right\rangle =a_{1}^{+}\left\vert 0_{p}\right\rangle $, $%
384: \left\vert L\right\rangle $ and $\left\vert R\right\rangle $ denote the
385: polarization of single Stokes photon. Before the emitted Stokes photon is
386: coupled into the fiber, it transmits a quarter wave plate (QWP). By
387: neglecting the small vacuum component and the high order terms, the
388: entangled state of the composite of atomic ensemble and photon can be
389: written into the form
390:
391: \begin{equation}
392: \left\vert \Psi \right\rangle _{ap}=\left(
393: P_{V}^{+}S_{0}^{+}a_{0}^{+}+P_{H}^{+}S_{1}^{+}a_{1}^{+}\right) /\sqrt{2}%
394: \left\vert vac\right\rangle _{ap}, \label{5}
395: \end{equation}%
396: with $\left\vert vac\right\rangle _{ap}=\left\vert G_{a}\right\rangle
397: \left\vert 0_{p}\right\rangle $, i.e. $\left\vert \Psi \right\rangle
398: _{ap}=\left( \left\vert H\right\rangle \left\vert 1\right\rangle
399: _{a}+\left\vert V\right\rangle \left\vert 0\right\rangle _{a}\right) \sqrt{2}
400: $. Via the above teleportation, the information will be stored into the
401: atomic memory. Due to the collective enhanced coherent interaction, the
402: atomic metastable states $\left\vert 0\right\rangle _{a}$ and $\left\vert
403: 1\right\rangle _{a}$ can be transferred to optical excitations with high
404: efficiency. So the information can be read out for further quantum
405: information processing.
406:
407: \begin{figure}[tbp]
408: \includegraphics[width=8cm]{transfer}
409: \caption{Schematic setup for remote transfer of unknown quantum state. The
410: unknown quantum state $(\protect\alpha S_{I_{2}}^{+}+\protect\beta %
411: S_{I_{1}}^{+})\left\vert 0_{a}0_{a}\right\rangle _{I_{1}I_{2}}$ is prepared
412: in the atomic ensemble pair $I$. The atomic ensemble pair $L$ and $R$ are in
413: a long-distance entangled state $\left\vert \Psi \right\rangle _{LR}$. After
414: shined simultaneously by the repump pulse, the collective atomic excitations
415: in the atomic ensemble $I$ and $L$ are transferred to the optical
416: excitations, which are registered by the sigle-photon detectors.}
417: \end{figure}
418:
419: After the atomic quantum memory has been established, it is very necessary
420: to consider its use in quantum communication protocols. Take remote transfer
421: of unknown quantum state as an example. Suppose that a long-distance
422: entangled state $\left\vert \Psi \right\rangle _{LR}$ between the double
423: pairs of atomic ensembles $L_{1}$, $L_{2}$ and $R_{1}$, $R_{2}$ is generated
424: by the quantum repeater \cite{21,22}, $\left\vert \Psi \right\rangle
425: _{LR}=(S_{L_{1}}^{+}S_{R_{2}}^{+}+S_{L_{2}}^{+}S_{R_{1}}^{+})/\sqrt{2}%
426: \left\vert vac\right\rangle _{LR}$, where $\left\vert vac\right\rangle
427: _{LR}=\left\vert 0_{a}0_{a}0_{a}0_{a}\right\rangle _{L_{1}L_{2}R_{1}R_{2}}$.
428: We denote the entangled logical state by $\left\vert \Psi \right\rangle
429: _{LR}=\left( \left\vert 0\right\rangle _{L}\left\vert 1\right\rangle
430: _{R}+\left\vert 1\right\rangle _{L}\left\vert 0\right\rangle _{R}\right) /%
431: \sqrt{2}$, where $\left\vert 0\right\rangle _{L(R)}=\left\vert
432: 0\right\rangle _{L_{1}(R_{1})}\left\vert 1\right\rangle _{L_{2}(R_{2})}$, $%
433: \left\vert 1\right\rangle _{L(R)}=\left\vert 1\right\rangle
434: _{L_{1}(R_{1})}\left\vert 0\right\rangle _{L_{2}(R_{2})}$. The unknown
435: quantum state storing in the two atomic ensembles $I_{1}$ and $I_{2}$ is $%
436: \left\vert \widetilde{\varphi }\right\rangle _{I}=(\alpha
437: S_{I_{2}}^{+}+\beta S_{I_{1}}^{+})\left\vert 0_{a}0_{a}\right\rangle
438: _{I_{1}I_{2}}=\alpha \left\vert 0\right\rangle _{I}+\beta \left\vert
439: 1\right\rangle _{I}$. Then shined simultaneously by the repump pulse which
440: is near-resonant with the atomic transition $\left\vert s\right\rangle
441: \rightarrow \left\vert e\right\rangle $, the collective atomic excitations
442: in the ensembles $I_{1}$, $L_{1}$ and $I_{2}$, $L_{2}$ are transferred to
443: the optical excitations, which, interfere respectively in a 50\%-50\% beam
444: splitter, and are detected by two single-photon detectors on each output. If
445: one click in $D_{1}$ or $D_{2}$, and one click in $D_{3}$ or $D_{4}$ have
446: been confirmed, our protocol succeeds in transferring the unknown quantum
447: state in the ensemble pair $I$ into the ensemble pair $R$ up to a local $\pi
448: $-phase rotation. The fidelity of the remote transfer protocol is nearly
449: perfect.
450:
451: \begin{figure}[tbp]
452: \includegraphics[width=8cm]{Fidelity}
453: \caption{(color online) \ (a) The improvement of quantum memory fidelity
454: with the average preparation time. Here we choose the overall efficiency of
455: photon detector, the optical appratus and fiber $\protect\eta ^{\prime }=1/3$%
456: . (b) The fidelity imperfection versus the overall efficiency $\protect\eta %
457: ^{\prime }$ with the average preparation time $T=20$ $\protect\mu s,$ $30$ $%
458: \protect\mu s,$ $40$ $\protect\mu s$. }
459: \end{figure}
460:
461: Finally, we briefly discuss the influence of practical noise and
462: imperfections on the memory. The fidelity of the information we store in the
463: atomic ensembles is mainly depend on the purity of entangled state. In
464: practice, the entanglement between the atomic ensembles and the Stokes
465: photon is in fact a mixed entangled state. The dominant noise arises from
466: the channel attenuation, spontaneous photon scattering into random
467: directions, coupling inefficiency for the channel, and inefficiency of
468: single-photon detectors. Taking into account these noises, as will be shown
469: below, the entangled state can be modified to
470:
471: \begin{equation}
472: \rho _{ap}=p_{0}\left\vert vac\right\rangle _{ap}\left\langle vac\right\vert
473: +p_{1}\left\vert \Psi \right\rangle _{ap}\left\langle \Psi \right\vert
474: +p_{o}\rho _{o}\text{,} \label{6}
475: \end{equation}%
476: where $p_{0}$, $p_{1}$ and $p_{o}$ are the probabilities of the vacuum state
477: that the two atomic ensembles are in the ground state, one-excited state and
478: the states with more excitations. Then we will give the probability an
479: analysis. $\left( i\right) $ We assume the probability of creating a Stokes
480: photon behind PBS is $2p_{c}$, the detection efficiency $\eta $ \cite{23} is
481: determined by the finite photon-collection (coupling) efficiency of the
482: optical apparatus, $\chi $, and by the quantum efficiency of the detector
483: itself, i.e. the efficiency of distinguishing single photon click event from
484: more photons, $\eta _{d}$, so the success probability of getting a
485: single-photon detector click is determined as $p_{1}\approx 2p_{c}\chi \eta
486: _{d}e^{-L_{0}/L_{att}}$, where we have considered the channel attenuation
487: factor as $e^{-L_{0}/L_{att}}$ \cite{15}, $L_{att}$ is the channel
488: attenuation length, and the other noise is independent of the communication
489: length $L_{0}$. $\left( ii\right) $ In the entanglement generation step, we
490: have neglected the influence of the detector dark counts which is denoted by
491: $p_{dc}$ in each Raman round. In fact, it contributes to the vacuum
492: component with the probability $p_{0}\approx p_{dc}/(p_{c}\eta ^{\prime }),$
493: $\eta ^{\prime }=\eta e^{-L_{0}/L_{att}}$ denotes the overall efficiency of
494: the single-photon detector, the optical apparatus and channel in the scheme.
495: Now, we take it into account. However, this vacuum component is typically
496: very small since the normal dark count rate (100 Hz) is much smaller than
497: the repetition frequency (10 MHz) of the Raman pulses \cite{15}. In
498: addition, finally this component will be automatically eliminated in our
499: scheme since its effect can be included by the detector inefficiency in the
500: application measurements. $\left( iii\right) $ If more than one atom is
501: excited to the collective mode S, due to the inefficiency of the photon
502: detector, there is only one click. The probability of this event is given by
503: $p_{o}\sim p_{c}^{n}\chi (1-\eta _{d})e^{-L_{0}/L_{att}}$ (decays
504: exponentially with the number of excited atoms $n$). So, the fidelity
505: between the generated state $(6)$ and the ideal state $(3)$ is decreased by
506: the high-order component which is proportional to $p_{c}$, we can simply
507: estimate the fidelity imperfection of our quantum memory $\Delta
508: F=1-F\approx p_{c}$. Decreasing the small controllable excitation
509: probability $p_{c}$ for each driving Raman pulse, we can make the fidelity
510: of the information close to one, but the longer preparation time $T$ is
511: cost. We have to repeat the process about $1/p_{1}$ times, with the total
512: average preparation time $T\sim 1/\left( p_{1}f_{p}\right) $, where $f_{p}$
513: is the repetition frequency of the Raman pulses. As evident from Fig. 4(a),
514: the fidelity of quantum memory can exceed $0.99$ for the average preparation
515: time $T>15$ $\mu s$. Furthermore, the fidelity is insensitive to the
516: variation of the overall efficiency $\eta ^{\prime }$ caused by the
517: single-photon detector, the optical apparatus and channel shown in Fig.
518: 4(b). For instance, the change of the fidelity is about $10^{-2}$ for $\eta
519: ^{\prime }$ varying from $0.1$ to $1$. The Fig. 4(b) also shows that the
520: fidelity is more insensitive to the overall efficiency $\eta ^{\prime }$
521: with the increase of the average preparation time.
522:
523: In summary, resorting to the idea of quantum teleportation, we have proposed
524: a quantum memory scheme to imprint the quantum state of the photon into the
525: DFS of atomic ensembles with high fidelity, and show one of applications
526: with the memory. Moreover, based on the current technology of laser
527: manipulation, photonic state operation through optical elements, and
528: single-photon detection with moderate efficiency, the scheme is inherently
529: resilient to the noise. Due to the long coherent time of atomic ensemble
530: \cite{13,14}, the teleported information can finally be read out for further
531: quantum information applications.
532:
533: \bigskip This work was supported by the National Natural Science Foundation
534: of China under Grant. Nos. 10404007 and 60578055, and the State Key
535: Development Program for Basic Research of China (Grant No. 2007CB925204 and
536: 2009CB929604). \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
537: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
538: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
539: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
540: \ \ \ \ \ \ \ \ \ \ \
541:
542: \begin{thebibliography}{99}
543: \bibitem{1} H. J. Briegel, W. D, J. I. Cirac, and P. Zoller, Phys. Rev.
544: Lett. \textbf{81}, 5932 (1998).
545:
546: \bibitem{2} E. Knill, R. Laflamme, and G. J. Milburn, Nature (London)\textbf{%
547: \ 409}, 46 (2001).
548:
549: \bibitem{3} D. L. Moehring \textit{et al}., Nature (London)\textbf{\ 449},
550: 68 (2007).
551:
552: \bibitem{4} J. I. Cirac, P. Zoller, H. J. Kimble, and H. Mabuchi, Phys. Rev.
553: Lett. \textbf{78}, 3221 (1997); H. Wei \textit{et al}., Phys. Rev. A \textbf{%
554: 76}, 054304 (2007).
555:
556: \bibitem{5} T. B. Pittman and J. D. Franson, Phys. Rev. A \textbf{66},
557: 062302 (2002).
558:
559: \bibitem{6} E. Cornell, Nature (London) \textbf{409}, 461 (2001); M.
560: Fleischhauer and M. D. Lukin, Phys. Rev. A \textbf{65}, 022314 (2002); S. A.
561: Moiseev and S. Kroll, Phys. Rev. Lett. \textbf{87}, 173601 (2001).
562:
563: \bibitem{7} P. Zanardi and M. Rasetti, Phys. Rev. Lett. \textbf{79}, 3306
564: (1997); D. A. Lidar, I. L. Chuang, and K. B. Whaley, \textit{ibid}. \textbf{%
565: 81}, 2594 (1998); L. M. Duan and G. C. Guo, Phys. Rev. A \textbf{57}, 737
566: (1998).
567:
568: \bibitem{8} D. Kielpinski, \textit{et al}., Science \textbf{291}, 1013
569: (2001); P. Xue and Y. F. Xiao, Phys. Rev. Lett. \textbf{97}, 140501 (2006).
570: N. Sangouard, Phys. Rev. A \textbf{73}, 022304 (2006); E. Brion \textit{et al%
571: }., Phys. Rev. A \textbf{75}, 032328 (2007).
572:
573: \bibitem{9} M. Fleischhauer and M. D. Lukin, Phys. Rev. Lett.\textbf{\ 84},
574: 5094 (2000).
575:
576: \bibitem{10} L.-M. Duan, J. I. Cirac, and P. Zoller, and E. S. Polzik, Phys.
577: Rev. Lett.\textbf{\ 85}, 5643 (2000); L.-M. Duan \textit{et al}., Phys. Rev.
578: A \textbf{66}, 023818 (2002).
579:
580: \bibitem{11} A. Kuzmich, L. Mandel, and N. P. Bigelow, Phys. Rev. Lett.
581: \textbf{85}, 1594 (2000); D. F. Phillips, A. Fleischhauer, A. Mair, and R.
582: L. Walsworth, \textit{ibid}. \textbf{86}, 783 (2001).
583:
584: \bibitem{12} D. N. Matsukevich and A. Kuzmich, Science \textbf{306}, 663
585: (2004); D. N. Matsukevich \textit{et al}., Phys. Rev. Lett. \textbf{95},
586: 040405 (2005); S. Chen \textit{et al}., \textit{ibid}. \textbf{99}, 180505
587: (2007).
588:
589: \bibitem{13} C. Liu, Z. Dutton, C. Behroozi, and L. Hau, Nature (London)
590: \textbf{409}, 490 (2001); J. Hald, J. L. S\o ensen, C. Schori, and E. S.
591: Polzik, Phys. Rev. Lett. \textbf{83}, 1319 (1999).
592:
593: \bibitem{14} A. Kuzmich \textit{et al}., Nature (London) \textbf{423}, 731
594: (2003); T. Chaneli\`{e}re \textit{et al}.,\ \textit{ibid}. \textbf{438}, 833
595: (2005).
596:
597: \bibitem{15} L.-M. Duan \textit{et al}., Nature (London) \textbf{414}, 413
598: (2001); L.-M. Duan, Phys. Rev. Lett. \textbf{88}, 170402 (2002).
599:
600: \bibitem{16} M. D. Lukin, Rev. Mod. Phys. \textbf{75}, 457 (2003).
601:
602: \bibitem{17} Here, we prepare the etanglestate by the application
603: measurement. If in the final measurement, we keep only the results for which
604: an excitation appears on one of the esembles $L$ and $R$, i.e. postselect
605: the case when only one detector registers a click, the whole of our scheme
606: is successful since only the state $(2)$ has contributions to the
607: measurement.
608:
609: \bibitem{18} S. Popescu, quant-ph/9501020.
610:
611: \bibitem{19} D. Boschi, S. Branca, F. De Martini, L. Hardy, and S. Popescu,
612: Phys. Rev. Lett. \textbf{80}, 1121 (1998).
613:
614: \bibitem{20} M. D. Lukin, S. F. Yelin, and M. Fleischhauer, Phys. Rev. Lett.
615: \textbf{84}, 4232 (2000).
616:
617: \bibitem{21} Z.-S. Yuan, Y. A. Chen, B. Zhao, S. Chen, J. Schmiedmayer, and
618: J. W. Pan, Nature (London) \textbf{454, }1098 (2008). N. Sangouard \textit{%
619: et al}., Phys. Rev. A \textbf{77}, 062301 (2008).
620:
621: \bibitem{22} L. Jiang, J. M. Taylor, and M. D. Lukin, Phys. Rev. A \textbf{76%
622: }, 012301 (2007); Z.-B. Chen \textit{et al}., \textit{ibid}. \textbf{76},
623: 022329 (2007).
624:
625: \bibitem{23} Stefano Zippilli et. al., quant-ph/08061052.
626: \end{thebibliography}
627:
628: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
629: \ \ \ \ \ \ \ \ \ \ \ \
630:
631: \bigskip\
632:
633: \bigskip
634:
635: \end{document}
636: