1: \documentclass
2: [floats,superscriptaddress,balancelastpage,prl,a4paper,twocolumn,tightenlines,showpacs,10pt]{revtex4}%
3: \usepackage{graphicx}
4: \usepackage{amsmath}
5: \usepackage{amsfonts}
6: \usepackage{amssymb}%
7: \setcounter{MaxMatrixCols}{30}
8: %TCIDATA{OutputFilter=latex2.dll}
9: %TCIDATA{Version=5.00.0.2552}
10: %TCIDATA{LastRevised=Wednesday, April 21, 2004 11:28:38}
11: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
12: %TCIDATA{<META NAME="SaveForMode" CONTENT="1">}
13: \begin{document}
14: \title{Experimental demonstration of a non-destructive controlled-NOT quantum gate
15: for two independent photon-qubits}
16: \author{Zhi Zhao}
17: \affiliation{Department of Modern Physics, University of Science
18: and Technology of China, Hefei, Anhui 230027, People's Republic of
19: China} \affiliation{Physikaliches Institut, Universit\"{a}t
20: Heidelberg, Philisophenweg 12, 69120 Heidelberg, Germany}
21: \date{\today }
22: \author{An-Ning Zhang}
23: \affiliation{Department of Modern Physics, University of Science
24: and Technology of China, Hefei, Anhui 230027, People's Republic of
25: China}
26: \author{Yu-Ao Chen}
27: \affiliation{Department of Modern Physics, University of Science
28: and Technology of China, Hefei, Anhui 230027, People's Republic of
29: China}
30: \author{Han Zhang}
31: \affiliation{Department of Modern Physics, University of Science
32: and Technology of China, Hefei, Anhui 230027, People's Republic of
33: China}
34: \author{Jiang-Feng Du}
35: \affiliation{Department of Modern Physics, University of Science
36: and Technology of China, Hefei, Anhui 230027, People's Republic of
37: China}
38: \author{Tao Yang}
39: \affiliation{Department of Modern Physics, University of Science
40: and Technology of China, Hefei, Anhui 230027, People's Republic of
41: China}
42: \author{Jian-Wei Pan}
43: \affiliation{Department of Modern Physics, University of Science
44: and Technology of China, Hefei, Anhui 230027, People's Republic of
45: China} \affiliation{Physikaliches Institut, Universit\"{a}t
46: Heidelberg, Philisophenweg 12, 69120 Heidelberg, Germany}
47: \date{\today }
48:
49: \begin{abstract}
50: Universal logic gates for two quantum bits (qubits) form an
51: essential ingredient of quantum information processing. However,
52: the photons, one of the best candidates for qubits, suffer from
53: the lack of strong nonlinear coupling required for quantum logic
54: operations. Here we show how this drawback can be overcome by
55: reporting a proof-of-principle experimental demonstration of a
56: non-destructive controlled-NOT (CNOT) gate for two independent
57: photons using only linear optical elements in conjunction with
58: single-photon sources and conditional dynamics. Moreover, we have
59: exploited the CNOT gate to discriminate all the four Bell-states
60: in a teleportation experiment.
61:
62: \end{abstract}
63:
64: \pacs{03.67.Lx, 42.50.Dv} \maketitle
65:
66:
67: The controlled-NOT (CNOT) or similar logic operations between two
68: individual quantum bits (qubits) are essential for various quantum
69: information protocols such as quantum communication
70: \cite{bennett92,bennett93,gisin02} and quantum computation
71: \cite{shor97}. In recent years, certain quantum logic gates have
72: been experimentally demonstrated, for example, in ion-traps
73: \cite{blatt03,wineland03} and high-finesse microwave cavities
74: \cite{haroche99}. These achievements open many possibilities for
75: future quantum information processing (QIP) with single atoms.
76: Another promising system for QIP is to use single photons. This is
77: due to the photonic robustness against decoherence and the
78: availability of single-qubit operation. However, it has been very
79: difficult to achieve the necessary logic operations for two
80: individual photon-qubits since the physical interaction between
81: photons is much too small.
82:
83: Surprisingly, Knill, Laflamme and Milburn (KLM) has shown that
84: nondeterministic quantum logic operations can be performed using
85: linear optical elements, additional photons (ancilla) and
86: postselection based on the output of single-photon detectors
87: \cite{klm01}. The original proposal by KLM, though elegant, is not
88: economical in its use of optical components. Various schemes have
89: been proposed to reduce the complexity of the KLM scheme while
90: improve its theoretical efficiency
91: \cite{koashi01,pittman01,ralph02}. Remarkably, a recent scheme
92: proposed by Nielsen \cite{nielsen04} suggests that without using
93: the elaborate teleportation and Z-measurement error correction in
94: the KLM scheme, any non-trivial linear optical gate that succeeds
95: with finite probability is sufficient to obtain efficient quantum
96: computation. Hence, this scheme significantly simplifies the
97: experimental implementation of linear optical quantum computation
98: (LOQC)
99:
100: A crucial requirement in the schemes of LOQC is the so-called
101: \emph{classical feedforwardability}, that is, it must be in
102: principle possible to detect when the gate has succeeded by
103: performing some appropriate measurement on ancilla photons
104: \cite{klm01,nielsen04}. This information can then be feedforward
105: for conditional future operations on the photonic qubits to
106: achieve efficient LOQC .
107:
108: %TCIMACRO{\FRAME{ftbpFU}{2.4267in}{2.9464in}{0pt}{\Qcb{(a) A nondestructive
109: %CNOT gate constructed by polarizing beam splitters (PBS), half-wave plates
110: %(HWP) and an ancilla entangled photon pair $|\Psi^{-}\rangle_{34}$
111: %\cite{pittman01}. Conditioned on detecting a $|-\rangle$ photon in mode
112: %$3^{\prime}$ and a $|H\rangle$ photon in mode $4^{\prime}$ one can implement
113: %the CNOT operation between the photons 2 and 5. (b) Quantum circuit for
114: %quantum teleportation based on a CNOT gate \cite{gotteman}. By using the CNOT
115: %operation, Alice can discriminate the four orthogonal Bell state
116: %simultaneously such that a complete teleportation can be achieved.}}%
117: %{}{figure1.eps}{\special{ language "Scientific Word"; type "GRAPHIC";
118: %maintain-aspect-ratio TRUE; display "USEDEF"; valid_file "F";
119: %width 2.4267in; height 2.9464in; depth 0pt; original-width 2.386in;
120: %original-height 2.9023in; cropleft "0"; croptop "1"; cropright "1";
121: %cropbottom "0"; filename 'figure1.eps';file-properties "XNPEU";}}}%
122: %BeginExpansion
123: \begin{figure}
124: [ptb]
125: \begin{center}
126: \includegraphics[
127: height=2.9464in, width=2.4267in
128: ]%
129: {figure1.eps}%
130: \caption{(a) A nondestructive CNOT gate constructed by polarizing
131: beam splitters (PBS), half-wave plates (HWP) and an ancilla
132: entangled photon pair $|\Psi^{-}\rangle_{34}$ \cite{pittman01}.
133: Conditioned on detecting a $|-\rangle$ photon in mode $3^{\prime}$
134: and a $|H\rangle$ photon in mode $4^{\prime}$ one can implement
135: the CNOT operation between the photons 2 and 5. (b) Quantum
136: circuit for quantum teleportation based on a CNOT gate
137: \cite{gotteman}. By using the CNOT operation, Alice can
138: discriminate the four orthogonal Bell state simultaneously such
139: that a complete teleportation can be
140: achieved.}%
141: \end{center}
142: \end{figure}
143: %EndExpansion
144:
145:
146: Recently destructive CNOT operations have been realized using
147: linear optical elements
148: \cite{pittman02,sanaka02,pittman03,white03}. However, as they
149: necessarily destroy the output state such logic operations are not
150: classically feed-forwardable and have little practical
151: significance. Fortunately, it has been suggested \cite{pittman01}
152: that a destructive CNOT gate together with the quantum parity
153: check can be combined with a pair of entangled photons to
154: implement a nondestructive (conventional) CNOT gate that satisfies
155: the feed-forwardability criterion.
156:
157: In this paper, we present for the first time a proof-of-principle
158: demonstration of a non-destructive CNOT gate for two independent
159: photons, realizing the proposal of Pittman, Jacobs, and Franson
160: \cite{pittman01}. The quality of such a CNOT gate is further
161: demonstrated by discriminating all the four Bell-states in an
162: experiment on quantum teleportation.
163:
164: In our experiment, we consider qubits implemented as the
165: polarization states of photons. We define the horizontal
166: polarization state $|H\rangle$ as logic 1, and the vertical one
167: $|V\rangle$ as logic 0. As shown in Fig. 1a, one can achieve the
168: desired nondestructive CNOT gate for photons 2 and 5 by performing
169: a quantum parity check on photons 2 and 3 and a destructive CNOT
170: operation on photons 4 and 5, where photons 3 and 4 are in the
171: state $|\Psi^{-}\rangle _{34}$, which is one of the four Bell
172: states
173: \begin{equation}%
174: \begin{array}
175: [c]{c}%
176: |\Phi^{\pm}\rangle_{ij}=\frac{1}{\sqrt{2}}(|H\rangle_{i}|H\rangle_{j}%
177: \pm|V\rangle_{i}|V\rangle_{j}),\\
178: |\Psi^{\pm}\rangle_{ij}=\frac{1}{\sqrt{2}}(|H\rangle_{i}|V\rangle_{j}%
179: \pm|V\rangle_{i}|H\rangle_{j}).
180: \end{array}
181: \label{epr}%
182: \end{equation}
183: Here $i$ and $j$ index the spatial mode of the photons. Then,
184: according to ref. \cite{pittman01} the nondestructive CNOT gate
185: for photons 2 and 5 can be accomplished conditioned on detecting a
186: $|-\rangle$ photon in mode $3^{\prime }$ and a $|H\rangle$ photon
187: in mode $4^{\prime}$. The logic table of the CNOT operation is
188: given by $|V\rangle|V\rangle\rightarrow|V\rangle|V\rangle$,
189: $|V\rangle|H\rangle\rightarrow|V\rangle|H\rangle$,
190: $|H\rangle|V\rangle \rightarrow|H\rangle|H\rangle$ and
191: $|H\rangle|H\rangle\rightarrow |H\rangle|V\rangle$.
192:
193: One immediate application of the proposed CNOT gate is that it can
194: be used to generate entanglement between the control qubit and
195: target qubit \cite{barenco95}. For example, by setting the control
196: bit to be in the state $|-\rangle$ and the target qubit in the
197: state $\left\vert H\right\rangle $, one can utilize the
198: non-destructive CNOT gate to prepare photons 2 and 5 in the
199: entangled state $|\Psi^{-}\rangle_{25}$.
200:
201: Another important application is that the non-destructive CNOT
202: gate can be used to simultaneously identify all the four Bell
203: states \cite{barenco95} in a quantum teleportation protocol
204: \cite{bennett93}. For example, suppose photon 5, which Alice wants
205: to teleport to Bob, is in an unknown polarization state
206: $|\Phi\rangle_{5}=\alpha\left\vert H\right\rangle
207: _{5}+\beta\left\vert V\right\rangle _{5}$, and the pair of the
208: photons $1$ and $2$ shared by Alice and Bob is in the entangled
209: state $\left\vert \Psi^{-}\right\rangle _{12}$ (Fig. 1b). It is
210: necessary to discriminate the four Bell-states of photons 2 and 5,
211: $\left\vert \Psi^{\pm}\right\rangle _{25}$ and $\left\vert
212: \Phi^{\pm }\right\rangle _{25}$ in order to realize the complete
213: quantum teleportation \cite{bennett93}.
214:
215: Under the CNOT operation the four Bell-states of photons 2 and 5
216: will evolve into one of the four orthogonal separable states
217: \begin{equation}%
218: \begin{array}
219: [c]{c}%
220: \left\vert \Psi^{\pm}\right\rangle
221: _{25}\longrightarrow|\pm\rangle_{2^{\prime
222: }}\left\vert H\right\rangle _{5^{\prime}},\\
223: \left\vert \Phi^{\pm}\right\rangle
224: _{25}\longrightarrow|\pm\rangle_{2^{\prime }}\left\vert
225: V\right\rangle _{5^{\prime}}.
226: \end{array}
227: \end{equation}
228:
229:
230: Therefore, the crucial Bell-state measurement can be accomplished
231: probabilistically by applying a nondestructive CNOT operation and
232: performing a subsequent polarization analysis on photons 2 and 5.
233: Depending on Alice's measurement results, Bob can then perform a
234: unitary transformation, independent of $|\Phi\rangle_{5}$, on
235: photon 1 to convert its state into the
236: initial state of photon 5.%
237:
238: %TCIMACRO{\FRAME{ftbpFU}{3.435in}{3.563in}{0pt}{\Qcb{The experimental setup.
239: %Femtosecond laser pulses ($\approx$ 200fs, 76MHz, 788nm) first pass through a
240: %doubler LBO crystal (LiB$_{3}$O$_{5}$), then the mixed ultraviolet (UV) and
241: %infrared components were further separated by a dichroic beamsplitter (DM).
242: %The reflected UV pulses passing through the BBO crystal twice generate two
243: %pairs of entangled photons, one in modes 1-2 and the other in modes 3-4, where
244: %both pairs are prepared in the $|\Psi^{-}\rangle$ state \cite{kwiat95,dik01}.
245: %The observed coincidence rate of entangled photon pairs is about
246: %$2.4\times10^{4}/$s. The photons in modes 3 and 4 are used as the ancilla pair
247: %while the photon in mode 2 as the control qubit. The arbitrary polarization
248: %state of the control photon can be readily prepared by performing a
249: %polarization projection measurement on photon 1. In the experiment, in order
250: %to prepare the target qubit, i.e. photon 5, the transmitted near-infrared
251: %pulses are attenuated by penetrating a ultra-fast laser output coupler and two
252: %polarizers (i.e. the Atten) to a weak coherent beam such that there is only a
253: %very small probability of containing a single photon for each pulse (about
254: %0.05 photon per pulse). The quarter wave plate ($\lambda/4$) in modes 1 and 5
255: %are used to prepare and analyze the input and output circular polarized
256: %states, respectively. The five polarizers P1, P2, ... P5 are used for
257: %polarization analysis.}}{}{figure2.eps}{\special{ language "Scientific Word";
258: %type "GRAPHIC"; maintain-aspect-ratio TRUE; display "USEDEF";
259: %valid_file "F"; width 3.435in; height 3.563in; depth 0pt;
260: %original-width 3.3892in; original-height 3.5163in; cropleft "0";
261: %croptop "1"; cropright "1"; cropbottom "0";
262: %filename 'figure2.eps';file-properties "XNPEU";}}}%
263: %BeginExpansion
264: \begin{figure}
265: [ptb]
266: \begin{center}
267: \includegraphics[
268: height=3.563in, width=3.435in
269: ]%
270: {figure2.eps}%
271: \caption{The experimental setup. Femtosecond laser pulses
272: ($\approx$ 200fs, 76MHz, 788nm) first pass through a doubler LBO
273: crystal (LiB$_{3}$O$_{5}$), then the mixed ultraviolet (UV) and
274: infrared components were further separated by a dichroic
275: beamsplitter (DM). The reflected UV pulses passing through the BBO
276: crystal twice generate two pairs of entangled photons, one in
277: modes 1-2 and the other in modes 3-4, where both pairs are
278: prepared in the $|\Psi ^{-}\rangle$ state \cite{kwiat95,dik01}.
279: The observed coincidence rate of entangled photon pairs is about
280: $2.4\times10^{4}/$s. The photons in modes 3 and 4 are used as the
281: ancilla pair while the photon in mode 2 as the control qubit. The
282: arbitrary polarization state of the control photon can be readily
283: prepared by performing a polarization projection measurement on
284: photon 1. In the experiment, in order to prepare the target qubit,
285: i.e. photon 5, the transmitted near-infrared pulses are attenuated
286: by penetrating a ultra-fast laser output coupler and two
287: polarizers (i.e. the Atten) to a weak coherent beam such that
288: there is only a very small probability of containing a single
289: photon for each pulse (about 0.05 photon per pulse). The quarter
290: wave plate ($\lambda/4$) in modes 1 and 5 are used to prepare and
291: analyze the input and output circular polarized states,
292: respectively. The five polarizers P1, P2,
293: ... P5 are used for polarization analysis.}%
294: \end{center}
295: \end{figure}
296: %EndExpansion
297:
298:
299: A schematic of the experimental setup used to demonstrate both the
300: CNOT gate and the identification of the four Bell states required
301: for quantum teleportation is shown in Fig. 2. To realize the CNOT
302: gate, it is necessary to overlap photons 2 and 3 at the PBS$_{1}$
303: and photons 4 and 5 at the PBS$_{2}$. In the experiment the
304: PBS$_{2}$, i.e. the desired 45-degree oriented polarizing
305: beamsplitter, is accomplished by inserting one half-wave plate
306: (HWP) in each of the two inputs and two outputs of an ordinary
307: polarizing beamsplitter. Note that, all the four HWP were oriented
308: at $22.5^{0}$ with respect to the horizontal direction, which
309: corresponds to a 45$^{0}$ polarization rotation. The good temporal
310: overlap was achieved by adjusting the two delay mirrors, Delay 1
311: and Delay 2. Experimentally, we first adjust the position of Delay
312: 1 such that photons 2 and 3 arrive at the PBS$_{1}$
313: simultaneously, and then adjust the position of Delay 2 to achieve
314: the temporal overlap of photons 4 and 5 at the PBS$_{2}$.
315: Furthermore, the use of narrow-band interference filters (F) with
316: $\Delta\lambda_{FWHM}=3$ nm for all five photons makes the photons
317: at the same PBS indistinguishable \cite{marek}. The temporal
318: overlap between photons 2 and 3 was verified by observing a
319: four-particle interference visibility of $0.82$ among photons 1,
320: 2, 3 and 4 by removing the PBS$_{2}$; and the temporal overlap of
321: photons 4 and 5 was verified by observing a three-particle
322: interference visibility of $0.68$ among photons 3, 4 and 5 by
323: removing the PBS$_{1}$ \cite{ghz1,ghz2}.
324:
325: To experimentally demonstrate that the CNOT gate has been
326: successfully implemented, we first prepare the input control and
327: target qubits in the following specific states
328: $|H\rangle_{2}|H\rangle_{5}$, $|H\rangle
329: _{2}|V\rangle_{5}$, $|V\rangle_{2}|H\rangle_{5}$ and $|V\rangle_{2}%
330: |V\rangle_{5}$. If the CNOT gate works properly, then conditioned
331: on detecting a $|-\rangle$ polarized photon in mode $3^{\prime}$
332: and a $|H\rangle$ polarized photon in mode $4^{\prime}$ the two
333: qubits in modes $2^{\prime}$ and
334: $5^{\prime}$ would be, respectively, in the states $|H\rangle_{2^{\prime}%
335: }|V\rangle_{5^{\prime}}$,
336: $|H\rangle_{2^{\prime}}|H\rangle_{5^{\prime}}$,
337: $|V\rangle_{2^{\prime}}|H\rangle_{5^{\prime}}$ and $|V\rangle_{2^{\prime}%
338: }|V\rangle_{5^{\prime}}$. After the non-destructive CNOT operation
339: the output components corresponding to the above specific input
340: states, which were measured in the $H/V$ basis, are shown in Fig.
341: 3a, respectively. The experimental fidelity of achieving the CNOT
342: logic table is estimated to be
343: $0.78\pm0.05$.%
344:
345: %TCIMACRO{\FRAME{ftbpFU}{2.6654in}{4.1987in}{0pt}{\Qcb{(a) Experimental results
346: %for the CNOT operation in the $H/V$ basis. When the input control and target
347: %qubits are in the following specific states $|H\rangle_{2}|H\rangle_{5}$,
348: %$|H\rangle_{2}|V\rangle_{5}$, $|V\rangle_{2}|H\rangle_{5}$ and $|V\rangle
349: %_{2}|V\rangle_{5}$, the output two qubits in modes $2^{\prime}$ and
350: %$5^{\prime}$ would be, respectively, in the states $|H\rangle_{2^{\prime}%
351: %}|V\rangle_{5^{\prime}}$, $|H\rangle_{2^{\prime}}|H\rangle_{5^{\prime}}$,
352: %$|V\rangle_{2^{\prime}}|H\rangle_{5^{\prime}}$ and $|V\rangle_{2^{\prime}%
353: %}|V\rangle_{5^{\prime}}$. The experimental fidelity of achieving the CNOT
354: %logic table is estimated to be $78\%$. (b) Experimental results for the CNOT
355: %operation when the control qubit was in the state $\frac{1}{\sqrt{2}}\left(
356: %\left\vert H\right\rangle -\left\vert V\right\rangle \right) $ and the target
357: %qubit in the state $\left\vert H\right\rangle $. After the CNOT operation the
358: %two output qubits would be in the Bell state $\left\vert \Psi^{-}\right\rangle
359: %$. We perform a conditional measurement of polarization as a function of the
360: %orientation of polarizer P2 when polarizer P5 was fixed at $+45^{0}$. The
361: %experimental results of the polarization correlation exhibit an interference
362: %visibility of $58\%$.}}{}{figure3.eps}{\special{ language "Scientific Word";
363: %type "GRAPHIC"; maintain-aspect-ratio TRUE; display "USEDEF";
364: %valid_file "F"; width 2.6654in; height 4.1987in; depth 0pt;
365: %original-width 2.623in; original-height 4.1485in; cropleft "0";
366: %croptop "1"; cropright "1"; cropbottom "0";
367: %filename 'figure3.eps';file-properties "XNPEU";}}}%
368: %BeginExpansion
369: \begin{figure}
370: [ptb]
371: \begin{center}
372: \includegraphics[
373: height=4.1987in, width=2.6654in
374: ]%
375: {figure3.eps}%
376: \caption{(a) Experimental results for the CNOT operation in the
377: $H/V$ basis. When the input control and target qubits are in the
378: following specific states $|H\rangle_{2}|H\rangle_{5}$,
379: $|H\rangle_{2}|V\rangle_{5}$, $|V\rangle _{2}|H\rangle_{5}$ and
380: $|V\rangle_{2}|V\rangle_{5}$, the output two qubits in modes
381: $2^{\prime}$ and $5^{\prime}$ would be, respectively, in the
382: states
383: $|H\rangle_{2^{\prime}}|V\rangle_{5^{\prime}}$, $|H\rangle_{2^{\prime}%
384: }|H\rangle_{5^{\prime}}$,
385: $|V\rangle_{2^{\prime}}|H\rangle_{5^{\prime}}$ and
386: $|V\rangle_{2^{\prime}}|V\rangle_{5^{\prime}}$.(b) Experimental
387: results for the CNOT operation when the control qubit was in the
388: state $\frac{1}{\sqrt{2}}\left( \left\vert H\right\rangle
389: -\left\vert V\right\rangle \right) $ and the target qubit in the
390: state $\left\vert H\right\rangle $. After the CNOT operation the
391: two output qubits would be in the Bell state $\left\vert
392: \Psi^{-}\right\rangle $. We perform a conditional measurement of
393: polarization as a function of the orientation of polarizer P2 when
394: polarizer P5 was fixed at $+45^{0}$.}%
395: \end{center}
396: \end{figure}
397: %EndExpansion
398:
399:
400: Second, to show the CNOT gate also works for an arbitrary
401: superposition of the
402: control qubit, we now prepare the control qubit in the state $\frac{1}%
403: {\sqrt{2}}\left( \left\vert H\right\rangle -\left\vert
404: V\right\rangle \right) $ and the target qubit in the state
405: $\left\vert H\right\rangle $ to entangle these two independent
406: photons. Then, after the CNOT operation the two output qubits
407: would be in the state $\frac{1}{\sqrt{2}}\left( \left\vert
408: H\right\rangle \left\vert V\right\rangle -\left\vert
409: V\right\rangle \left\vert
410: H\right\rangle \right) $, i.e. the Bell state $\left\vert \Psi^{-}%
411: \right\rangle $. To verify the expected Bell state has implemented
412: successfully, we first measured the four possible polarization
413: combinations of the control and target qubits in the $H/V$ basis.
414: The signal-to-noise ratio between the desired
415: ($|H\rangle|V\rangle$ and $|V\rangle|H\rangle$) and unwanted
416: ($|H\rangle|H\rangle$ and $|V\rangle|V\rangle$) were measured to
417: be $4.2:1$. This confirms that the $|H\rangle|V\rangle$ and
418: $|V\rangle|H\rangle$ terms are the dominant components.
419: Furthermore, to prove the two terms are indeed in a coherent
420: superposition, we also perform a conditional coincidence
421: measurement as a function of the orientation of polarizer P2 as
422: polarizer P5 was fixed at $+45^{0}$. As shown in Fig. 3b, the
423: experimental results of the polarization correlation exhibit an
424: interference fringe with a visibility of $0.58\pm0.09$, which is
425: in consistent with the prediction of the interference fringe
426: for the Bell state $\left\vert \Psi^{-}\right\rangle $.%
427:
428: %TCIMACRO{\FRAME{ftbpFU}{2.7916in}{3.8908in}{0pt}{\Qcb{The experimental results
429: %for quantum teleportation with complete Bell-state analysis. The data clearly
430: %confirm the expected $\pi$ phase shift, hence demonstrating that the four
431: %Bell-state have been identified successfully. Moreover, the obtained
432: %teleportation fidelity is as high as $0.79$, which is well beyond the
433: %classical limit of $2/3$, hence demonstrating quantum teleportation with
434: %complete Bell-state identification.}}{}{figure4.eps}%
435: %{\special{ language "Scientific Word"; type "GRAPHIC";
436: %maintain-aspect-ratio TRUE; display "USEDEF"; valid_file "F";
437: %width 2.7916in; height 3.8908in; depth 0pt; original-width 2.7492in;
438: %original-height 3.8424in; cropleft "0"; croptop "1"; cropright "1";
439: %cropbottom "0"; filename 'figure4.eps';file-properties "XNPEU";}}}%
440: %BeginExpansion
441: \begin{figure}
442: [ptb]
443: \begin{center}
444: \includegraphics[
445: height=3.8908in, width=2.7916in
446: ]%
447: {figure4.eps}%
448: \caption{The experimental results for quantum teleportation with
449: complete Bell-state analysis. The data clearly confirm the
450: expected $\pi$ phase shift, hence demonstrating that the four
451: Bell-states have been identified successfully in the teleportation experiment.}%
452: \end{center}
453: \end{figure}
454: %EndExpansion
455:
456:
457: The CNOT gate can be used not only to entangle two independent
458: photons, it can also be used to disentangle two entangled photons.
459: To demonstrate the latter, let us now exploit the CNOT gate to
460: simultaneously discriminate all the four Bell-states in an quantum
461: teleportation experiment. As described in equation (2),
462: conditioned on detecting a $|-\rangle$ photon in mode $3^{\prime}$
463: and a $|H\rangle$ photon in mode $4^{\prime}$ the required joint
464: Bell-state measurement can be achieved by performing a
465: polarization measurement both on photon $2^{\prime}$ in the $+/-$
466: basis and on photon $5^{\prime}$ in the $H/V$ basis. For example,
467: registering a $|+\rangle_{2^{\prime}}\left\vert H\right\rangle
468: _{5^{\prime}}$ coincidence implies a projection onto the
469: Bell-state $|\Psi^{+}\rangle_{25}$. In this way, we can identify
470: all the four Bell-states. According to teleportation protocol
471: \cite{bennett93} it is obvious that, if the photons 2 and 5 are
472: measured to be in the state $|\Psi^{-}\rangle_{25}$, then photon 1
473: will be projected into the state $\alpha|H\rangle+\beta|V\rangle$;
474: if the photons 2 and 5 are measured to be in the state
475: $|\Psi^{+}\rangle_{25}$, then photon 1 will be left in the state
476: $\alpha|H\rangle-\beta|V\rangle$. In these two cases, the two
477: corresponding states of photon 1 would, in general, have a
478: relative phase shift of $\pi$
479: \cite{shih01}. Similarly, for the projections onto the state $|\Phi^{+}%
480: \rangle_{25}$ or $|\Phi^{-}\rangle_{25}$, photon 1 will be
481: correspondingly left in the state $\alpha|V\rangle-\beta|H\rangle$
482: or $\alpha|V\rangle +\beta|H\rangle$. Again, the two states of
483: photon 1 have a relative phase shift of $\pi$.
484:
485: To experimentally verify the above analysis, we decided to
486: teleport the left-hand circular polarization state
487: $\frac{1}{\sqrt{2}}(|H\rangle -i|V\rangle)$ from photon 5 to
488: photon 1. The output circular polarization states of photon 1 are
489: analyzed by inserting a QWP and a polarizer in front of the
490: detector D1. As shown in Fig. 4a and b, the five-fold coincidences
491: are recorded as polarizer 1 was rotated. The experimental results
492: are consistent with the prediction of the $\pi$ phase shift, hence
493: confirming that the four Bell-states have been successfully
494: discriminated. Following the same definition of ref.
495: \cite{shih01}, our experimental average fidelity of teleportation
496: was estimated to be $F\approx0.79\pm0.05$, which clearly surpasses
497: the classical limit of $2/3$. Note that throughout the whole
498: experiment the fidelities were obtained without performing any
499: background subtraction.
500:
501: Finally, we emphasize that in the above teleportation experiment
502: we only verified the corresponding relation between the initial
503: state of photon 5 and the final state of photon 1 after the
504: Bell-state analysis is complete, but did not perform a conditional
505: operation on the photon 1 to convert its final state into the
506: original state of photon 5. We plan to address this challenging
507: task in a forthcoming experiment.
508:
509: In summary, we have for the first time experimentally demonstrated
510: a probabilistic nondestructive CNOT gate for two independent
511: photons using only linear optics. Furthermore, we demonstrated
512: that such a device can be used not only to entangle two
513: independent photons, but also to discriminate all the four
514: Bell-states for quantum teleportation. We believe that the methods
515: developed for this experiment would have various novel
516: applications in quantum information processing with linear optics.
517:
518: This work was supported by the NSF of China, the CAS and the
519: National Fundamental Research Program (under Grant No.
520: 2001CB309303), and the Alexander von Humboldt Foundation.
521:
522: \begin{thebibliography}{99} %
523: \bibitem {bennett92}C. H. Bennett and S. J. Wiesner, Phys. Rev. Lett.
524: \textbf{69}, 2881 (1992).
525:
526: \bibitem {bennett93}C. H. Bennett \textit{et al}, Phys. Rev. Lett.
527: \textbf{70}, 1895 (1993).
528:
529: \bibitem {gisin02}N. Gisin, G. Ribordy, W. Tittel, \& H. Zbinden Rev. Mod.
530: Phys. \textbf{74}, 145 (2002).
531:
532: \bibitem {shor97}P. W. Shor, SIAM J. Comput. 26, 1484 (1997).
533:
534: \bibitem {blatt03}F. Schmidt-Kaler \textit{et al}., Nature \textbf{422}, 408 (2003).
535:
536: \bibitem {wineland03}D. Liebfried \textit{et al}., Nature \textbf{422}, 412 (2003).
537:
538: \bibitem {haroche99}A. Rauschebeutel \textit{et al}, Phys. Rev. Lett.
539: \textbf{83}, 5166 (1999).
540:
541: \bibitem {klm01}E. Knill, R. Laflamme, and G. J. Milburn, Nature \textbf{409},
542: 46 (2001).
543:
544: \bibitem {koashi01}M. Koashi, T. Yamamoto, and N. Imoto, Phys. Rev. A
545: \textbf{63}, 030301(R) (2001).
546:
547: \bibitem {pittman01}T. B. Pittman, B. C. Jacobs, and J. D. Franson, Phys. Rev.
548: A \textbf{64}, 062311 (2001).
549:
550: \bibitem {ralph02}T. C. Ralph, A. G. White, W. J. Munro, and G. J. Milburn,
551: Phys. Rev. A \textbf{65}, 012314 (2002).
552:
553: \bibitem {nielsen04}M. A. Nielsen, arXiv:quant-ph/0402005.
554:
555: \bibitem {pittman02}T. B. Pittman, B. C. Jacobs, and J. D. Franson, Phys. Rev.
556: Lett. \textbf{88}, 257902 (2002).
557:
558: \bibitem {sanaka02}K. Sanaka, K. Kawahara, \& T. Kuga, Phys. Rev. A
559: \textbf{66}, 040301(R) (2002).
560:
561: \bibitem {pittman03}T. B. Pittman, M. J. Fitch, B. C Jacobs, and J. D.
562: Franson, Phys. Rev. A \textbf{68}, 032316 (2003).
563:
564: \bibitem {white03}J. L. O'Brien \textit{et al}, Nature \textbf{426}, 264 (2003).
565:
566: \bibitem {gotteman}D. Gotteman and I. L. Chuang, Nature \textbf{402}, 390 (1999).
567:
568: \bibitem {barenco95}A. Barenco, D. Deutsch, \& A. Akert, Phys. Rev. Lett.
569: \textbf{74}, 4083 (1995).
570:
571: \bibitem {kwiat95}P. G. Kwiat \textit{et al}, Phys. Rev. Lett. \textbf{75},
572: 4337 (1995).
573:
574: \bibitem {dik01}D. Bouwmeester \textit{et al}, Nature \textbf{390}, 575 (1997).
575:
576: \bibitem {marek}M. Zukowski, A. Zelinger \& H. Weinfurter, Ann. NY Acad. Sci.
577: \textbf{755}, 91 (1995).
578:
579: \bibitem {ghz1}D. Bouwmeester \textit{et al.}, Phys. Rev. Lett. \textbf{82},
580: 1345 (1999).
581:
582: \bibitem {ghz2}J.-W. Pan \textit{et al.}, Phys. Rev. Lett. \textbf{86}, 4435 (2001).
583:
584: \bibitem {shih01}Y.-H. Kim, S. P. Kulik, \& Y. Shih, Phys. Rev. Lett.
585: \textbf{86}, 1370 (2001).
586: \end{thebibliography}
587:
588:
589: \end{document}
590: