1:
2: \documentclass[prl,preprint,12pt,onecolumn]{revtex4}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: \usepackage{amsmath}
5:
6: %TCIDATA{OutputFilter=LATEX.DLL}
7: %TCIDATA{Created=Tuesday, October 24, 2000 13:05:13}
8: %TCIDATA{LastRevised=Monday, October 09, 2000 19:56:10}
9: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
10: %TCIDATA{<META NAME="DocumentShell" CONTENT="Articles\SW\REVTeX 4 (Test Version)">}
11: %TCIDATA{Language=American English}
12: %TCIDATA{CSTFile=revtex4.cst}
13:
14: \newtheorem{theorem}{Theorem}
15: \newtheorem{acknowledgement}[theorem]{Acknowledgement}
16: \newtheorem{algorithm}[theorem]{Algorithm}
17: \newtheorem{axiom}[theorem]{Axiom}
18: \newtheorem{claim}[theorem]{Claim}
19: \newtheorem{conclusion}[theorem]{Conclusion}
20: \newtheorem{condition}[theorem]{Condition}
21: \newtheorem{conjecture}[theorem]{Conjecture}
22: \newtheorem{corollary}[theorem]{Corollary}
23: \newtheorem{criterion}[theorem]{Criterion}
24: \newtheorem{definition}[theorem]{Definition}
25: \newtheorem{example}[theorem]{Example}
26: \newtheorem{exercise}[theorem]{Exercise}
27: \newtheorem{lemma}[theorem]{Lemma}
28: \newtheorem{notation}[theorem]{Notation}
29: \newtheorem{problem}[theorem]{Problem}
30: \newtheorem{proposition}[theorem]{Proposition}
31: \newtheorem{remark}[theorem]{Remark}
32: \newtheorem{solution}[theorem]{Solution}
33: \newtheorem{summary}[theorem]{Summary}
34: \newenvironment{proof}[1][Proof]{\textbf{#1.} }{\ \rule{0.5em}{0.5em}}
35: \input{tcilatex}
36:
37: \begin{document}
38:
39: \title{Practical creation and detection of polarization Bell states using
40: parametric down-conversion}
41: \author{K. J. Resch, J. S. Lundeen, and A. M. Steinberg}
42: \address{Department of Physics, University of Toronto\\
43: 60 St. George Street, Toronto ON M5S 1A7\\
44: CANADA}
45:
46: \begin{abstract}
47: The generation and detection of maximally-entangled two-particle states,
48: `Bell states,' are crucial tasks in many quantum information protocols such
49: as cryptography, teleportation, and dense coding. \ Unfortunately, they
50: require strong inter-particle interactions lacking in optics. \ For this
51: reason, it has not previously been possible to perform complete Bell state
52: determination in optical systems. \ In this work, we show how a recently
53: developed quantum interference technique for enhancing optical
54: nonlinearities can make efficient Bell state measurement possible. \ We also
55: discuss weaknesses of the scheme including why it cannot be used for
56: unconditional quantum teleportation.
57: \end{abstract}
58:
59: \maketitle
60:
61: \section{Introduction\qquad}
62:
63: The new science of quantum information builds on the recognition that
64: entanglement, an essential but long underemphasized feature of quantum
65: mechanics, can be a valuable resource. \ Many of the headline-grabbing
66: quantum communication schemes (including quantum teleportation \cite%
67: {bennbrass,zeil,demartini}, dense coding \cite{densetheory,denseexp}, and
68: quantum cryptography \cite{cryptbennbrass,cryptekert}) are based on the
69: maximally-entangled two-particle quantum states called Bell states. \ Using
70: the polarization states of a pair of photons in different spatial modes, the
71: four Bell states are written as:%
72: \begin{eqnarray}
73: \left| \psi ^{\pm }\right\rangle &=&\frac{1}{\sqrt{2}}\left( \left|
74: V\right\rangle _{1}\left| H\right\rangle _{2}\pm \left| H\right\rangle
75: _{1}\left| V\right\rangle _{2}\right) \notag \\
76: \left| \phi ^{\pm }\right\rangle &=&\frac{1}{\sqrt{2}}\left( \left|
77: H\right\rangle _{1}\left| H\right\rangle _{2}\pm \left| V\right\rangle
78: _{1}\left| V\right\rangle _{2}\right) ,
79: \end{eqnarray}%
80: where $\left| H\right\rangle $ and $\left| V\right\rangle $ describe
81: horizontal- and vertical-polarization states, and the subscripts 1 and 2 are
82: spatial mode labels. \ These four states form a complete, orthonormal basis
83: for the polarization states of a pair of photons. \ In each Bell state, a
84: given photon is completely unpolarized but perfectly correlated with the
85: polarization of the other photon. \ Photon Bell states were produced in
86: atomic cascades for the first tests of the nonlocal predictions of quantum
87: mechanics \cite{clauserbell}. \ Since that time, parametric down-conversion
88: sources \cite{othersources,steinberg,kwiat1,kwiat2,shapiro} have replaced
89: cascade souces due to their ease of use, high brightness, and the
90: high-purity states they produce. \ However, down-conversion sources do not
91: deterministically prepare photon Bell states, but rather states in which the
92: Bell state component is in a coherent superposition with a dominant vacuum
93: term; coincidence detection of photon pairs projects out only the two-photon
94: component of the state.
95:
96: While optical Bell state source technology has shown marked improvement,
97: methods of distinguishing these states has proven a difficult challenge. \
98: Perhaps the most well-known example of why distinguishing Bell states is
99: important comes from quantum teleportation. \ A general projective
100: measurement is required for unconditional teleportation; experimental
101: teleportation was originally limited to a maximum efficiency of 25\% since
102: only the singlet state, $\left| \psi ^{-}\right\rangle ,$ could be
103: distinguished from the triplet states \cite{zeil}. \ The challenge for
104: measuring Bell states stems from the requirement for a strong inter-particle
105: interaction, which is usually nonexistent for photons. \ Without such a
106: nonlinearity, only two of the four states can be distinguished\cite%
107: {circuits2}. \ It was realized that a strong enough optical nonlinearity,
108: typically $\chi ^{(3)}$, could be used to mediate a photon-photon
109: interaction. \ Unfortunately, even the nonlinearities of our best materials
110: are far too weak. \ An experiment using standard nonlinear materials to
111: demonstrate a scheme for unconditional teleportation was limited to
112: extremely low efficiencies (on the order of 10$^{-10}$) by the tiny
113: nonlinearities involved \cite{shihtelep}. \ Proposals for extending optical
114: nonlinearities to the quantum level include schemes based on cavity QED \cite%
115: {qed}, electromagnetically-induced transparency \cite{eit}, photon-exchange
116: interactions \cite{pei}, and quantum interference techniques\ \cite%
117: {switch,phase}. \ Using the latter, we have recently demonstrated a
118: conditional-phase switch \cite{phase} which is similar to the
119: controlled-phase gate in quantum computation. \ In this work, we show how to
120: apply the conditional-phase switch to the problem of Bell state detection. \
121: It should be noted that if recently published schemes for performing quantum
122: computing with linear optics \cite{KLM,otherlin} could be experimentally
123: realized, then the problem of distinguishing all four Bell states could be
124: performed without the need for strong optical nonlinearities. \ Theoretical
125: work has also shown that if the Bell state is embedded appropriately in a
126: higher-dimensional Hilbert space, all of the Bell states can be
127: distinguished \cite{embedded}.
128:
129: Strong optical nonlinearities are desired so that one can construct a
130: controlled-$\pi $, a specific case of the controlled-phase gate for photons.
131: \ Such a gate and all one-qubit rotations form a universal set of gates for
132: the more general problem of quantum computation -- just as the NAND gate is
133: universal for classical computation. \ The controlled-$\pi $ transformation %
134: \cite{cphiref} is described by:%
135: \begin{eqnarray}
136: \left| 0\right\rangle _{1}\left| 0\right\rangle _{2} &\longrightarrow
137: &\left| 0\right\rangle _{1}\left| 0\right\rangle _{2} \notag \\
138: \left| 0\right\rangle _{1}\left| 1\right\rangle _{2} &\longrightarrow
139: &\left| 0\right\rangle _{1}\left| 1\right\rangle _{2} \notag \\
140: \left| 1\right\rangle _{1}\left| 0\right\rangle _{2} &\longrightarrow
141: &\left| 1\right\rangle _{1}\left| 0\right\rangle _{2} \notag \\
142: \left| 1\right\rangle _{1}\left| 1\right\rangle _{2} &\longrightarrow
143: &-\left| 1\right\rangle _{1}\left| 1\right\rangle _{2}, \label{cpi}
144: \end{eqnarray}%
145: in which the two qubit states are $\left| 0\right\rangle $ and $\left|
146: 1\right\rangle $ and the subscript is the qubit label. \ This transformation
147: does nothing to the input state unless both qubits have a value of $\left|
148: 1\right\rangle ,$ in which case it applies a phase-shift of $\pi .$ \ On the
149: surface this transformation appears to do nothing since an overall phase in
150: quantum mechanics is meaningless. \ However, it is clearly nontrivial when
151: applied to superpositions of states. \
152:
153: The polarization of the photon makes an ideal two-level system for encoding
154: a qubit largely due to its relative immunity to environmental decoherence. \
155: A\ large enough $\chi ^{(3)}$ nonlinearity could be used to effect the c-$%
156: \pi $ transformation on a pair of photons. \ Given a polarization-dependent $%
157: \chi ^{(3)}$, or through the use of polarizing beam-splitters, only photon
158: pairs with, say, horizontal polarization would experience the nonlinear
159: interaction and pick up the additional phase shift. \ Such a gate could then
160: be incorporated into the optical implementation of the quantum circuits
161: shown in Fig. 1a. and 2a. (similar circuits are discussed in \cite%
162: {circuits2,circuits}). \ The circuit in Fig. 1a. converts, through unitary
163: transformation, a state in the rectilinear product state basis (i.e. $\left|
164: 0\right\rangle _{1}\left| 0\right\rangle _{2},\left| 0\right\rangle
165: _{1}\left| 1\right\rangle _{2},\left| 1\right\rangle _{1}\left|
166: 0\right\rangle _{2},$ and $\left| 1\right\rangle _{1}\left| 1\right\rangle
167: _{2}$) to the Bell basis. \ The circuit in Fig. 2a. performs the opposite
168: function converting a Bell state via unitary transformation to the
169: rectilinear basis. \ In essence, these circuits allow for the creation and
170: removal of entanglement between pairs of qubits. \ If the qubit states $%
171: \left| 0\right\rangle $ and $\left| 1\right\rangle $ are encoded into the
172: polarization states $\left| H\right\rangle $ and $\left| V\right\rangle $ in
173: two different spatial modes 1 and 2, then an optical realization of the
174: circuit in Fig. 2a. allows for the conversion of a photon pair in a Bell
175: state to a rectilinear basis state. \ These four rectilinear basis states
176: are easily distinguishable using the simple optical setup shown in Fig. 3. \
177: Thus, after passing the photon pair in a Bell state through the optical
178: realization of the circuit in Fig. 2a., the subsequent detection of the
179: rectilinear state is equivalent to determination of the Bell state.
180:
181: The conditional-phase switch we propose is related to the controlled-phase
182: gate of quantum computation and is described in the theory section of this
183: work. \ The switching effect occurs in a $\chi ^{(2)}$ nonlinear material
184: that is pumped by a strong, classical beam. \ This pump beam is capable of
185: creating pairs of down-converted photon pairs into a pair of output modes. \
186: Pairs of photons, in a coherent superposition with the vacuum, pass through
187: the crystal into those same output modes. \ It is the interference between
188: the amplitudes for multiple paths leading to a photon pair that greatly
189: enhances the effective nonlinearity; since the down-converted light is only
190: created in pairs, the interference only affects the amplitude for photon
191: pairs. \ However, since the switching effect is based on an interference
192: effect, it is intrinsically dependent on the phase and amplitude of the
193: incoming beams. \ This has two consequences. \ First, the switch requires an
194: input which is in a coherent superposition with the vacuum. \ In this way,
195: the input has the required \emph{uncertain} number of photons, since photon
196: number and phase are conjugate quantities. \ And second, the switch works as
197: described only for states in the correct superposition with the vacuum, not
198: a general input state. \ As we will show, these conditions do allow for one
199: to distinguish between the four Bell states provided they are in the correct
200: superposition with the vacuum. \ Nonetheless, the conditions are too
201: stringent to allow for unconditional teleportation using this method.
202:
203: First, we describe the effective nonlinearity. \ Then we show how the
204: nonlinearity can be used to construct optical devices analogous to the
205: quantum computation circuits shown in Fig. 1a. and Fig. 2a. \
206:
207: \section{Theory}
208:
209: \subsection{Effective Nonlinearity}
210:
211: The general down-conversion state can be written as%
212: \begin{equation}
213: \left| \psi \right\rangle =\left| 0\right\rangle +\varepsilon \left(
214: \begin{array}{cccc}
215: \left| H\right\rangle _{1}\left| H\right\rangle _{2}\text{ \ \ } & \left|
216: H\right\rangle _{1}\left| V\right\rangle _{2}\text{ \ \ } & \left|
217: V\right\rangle _{1}\left| H\right\rangle _{2}\text{ \ \ } & \left|
218: V\right\rangle _{1}\left| V\right\rangle _{2}%
219: \end{array}%
220: \right) \left(
221: \begin{array}{c}
222: \alpha \\
223: \beta \\
224: \gamma \\
225: \delta%
226: \end{array}%
227: \right) , \label{dcstate}
228: \end{equation}%
229: where the part of the state describing photon pairs has been written as an
230: inner product. \ The amplitudes for the polarization states $\left|
231: H\right\rangle _{1}\left| H\right\rangle _{2}$, $\left| H\right\rangle
232: _{1}\left| V\right\rangle _{2}$, $\left| V\right\rangle _{1}\left|
233: H\right\rangle _{2}$, and $\left| V\right\rangle _{1}\left| V\right\rangle
234: _{2}$ are $\varepsilon \alpha $, $\varepsilon \beta ,$ $\varepsilon \gamma $%
235: , and $\varepsilon \delta $, respectively. \ Again, the subscripts 1 and 2
236: describe two different spatial modes. \ Throughout this theory section, we
237: adopt a 4-dimensional vector representation to describe the polarization
238: state of the photon pairs. \ In this more compact notation, the general
239: state is written
240: \begin{equation}
241: \left| \psi \right\rangle =\left| 0\right\rangle +\varepsilon \left(
242: \begin{array}{c}
243: \alpha \\
244: \beta \\
245: \gamma \\
246: \delta%
247: \end{array}%
248: \right) ,
249: \end{equation}%
250: In both cases, we have suppressed the normalization factor for clarity, and
251: for the discussion here we will restrict ourselves to the case where the
252: probability of having a photon pair at any given time is small, i.e. $\left|
253: \varepsilon \right| ^{2}\ll 1$ (as is always the case in real
254: down-conversion experiments).
255:
256: The effective nonlinearity \cite{phase} can be described as follows. \ Modes
257: 1 and 2 are of frequency $\omega $ and pass through a $\chi ^{(2)}$
258: nonlinear crystal that is simultaneously pumped by a strong classical laser
259: beam of frequency $2\omega $ in mode p. \ The modes are so chosen such that
260: the nonlinear crystal can create degenerate horizontally-polarized photon
261: pairs in spatial modes 1 and 2 via spontaneous parametric down-conversion,
262: as shown in Fig. 4. \ The nonlinear process is mediated by the interaction
263: Hamiltonian,
264:
265: \begin{equation}
266: \mathcal{H}=ga_{1,H}^{\dagger }a_{2,H}^{\dagger }a_{p,V}+g^{\ast
267: }a_{1,H}a_{2,H}a_{p,V}^{\dagger },
268: \end{equation}%
269: where g is the coupling constant and $a_{i}^{(\dagger )}$ is the field
270: annihilation (creation) operator for the $i^{th}$ mode, and the subscripts $%
271: H $ and $V$ are the polarizations of the relevant modes for the type-I\
272: phase-matching. \ The pump laser is intense enough that we treat it
273: classically by replacing its field operators with c-number amplitudes, $%
274: \zeta $ and $\zeta ^{\ast }$:%
275: \begin{equation}
276: \mathcal{H}=g\zeta a_{1,H}^{\dagger }a_{2,H}^{\dagger }+g^{\ast }\zeta
277: ^{\ast }a_{1,H}a_{2,H}.
278: \end{equation}%
279: Due to phase-matching constraints, the nonlinear crystal can only produce
280: horizontally-polarized photon pairs. \ In the weak coupling regime, we can
281: use first-order perturbation theory to propagate our state under the
282: interaction to,%
283: \begin{eqnarray}
284: \left| \psi (t)\right\rangle &=&\left( 1-\frac{it}{\hbar }\mathcal{H}\right)
285: \left| \psi \right\rangle \\
286: &=&\left| 0\right\rangle +\varepsilon \left(
287: \begin{array}{c}
288: \alpha \\
289: \beta \\
290: \gamma \\
291: \delta%
292: \end{array}%
293: \right) -\frac{it}{\hbar }g\zeta \left(
294: \begin{array}{c}
295: 1 \\
296: 0 \\
297: 0 \\
298: 0%
299: \end{array}%
300: \right) \\
301: &=&\left| 0\right\rangle +\varepsilon \left(
302: \begin{array}{c}
303: \alpha -\frac{it}{\hbar }\frac{g\zeta }{\varepsilon } \\
304: \beta \\
305: \gamma \\
306: \delta%
307: \end{array}%
308: \right) .
309: \end{eqnarray}%
310: To first order, this Hamiltonian simply creates an amplitude for a
311: horizontally-polarized pair of photons. \ This new down-conversion amplitude
312: interferes with the preexisting amplitude for the $HH$ term. \
313:
314: The transformation, as described here, does not appear unitary. \ This is
315: due to a few approximations. \ We assume that the vacuum term in our state
316: is unchanged, and neglect terms describing more that one pair of photons. \
317: These approximations are only valid in the relevant limit where $\left|
318: \varepsilon \right| $ $\ll 1$, where we can also suppress the normalization
319: term for clarity. \ However, the exact propagator follows from a hermitian
320: Hamiltonian and is of course unitary.
321:
322: As was shown in the ``railcross experiment'' \cite{railcross} and in our
323: subsequent work with photon pairs from coherent state inputs \cite{switch},
324: interference between the amplitudes for existing pairs and for
325: down-conversion can modulate the rate of pair production. \ Given the
326: phase-matching scheme presented here, only the amplitude for $HH$ pairs is
327: affected. \ Accompanying this modulation of the photon pair production rate
328: is a shift in the phase of the horizontally-polarized photon pair term. \
329: The down-conversion crystal impresses a $\pi $ phase-shift on the $HH$ term
330: if the down-conversion amplitude, $-itg\zeta /\hbar $ to be $-2\varepsilon
331: \alpha $. \ To implement a tranformation analogous to the c-$\pi $ (Eq. \ref%
332: {cpi}) in the coincidence basis, this is the only condition that must be
333: enforced; the values for the coefficients $\alpha $, $\beta $, and $\gamma $
334: are free. \ This condition takes the place of the more usual normalization
335: condition on $\alpha $, $\beta $, $\gamma ,$ and $\delta $ to describe our
336: state space. \ It can be enforced experimentally by controlling the
337: amplitude and phase of the pump laser and/or the overall pair amplitude $%
338: \varepsilon $. \ Unfortunately, this means that the gate cannot be utilized
339: on arbitrary inputs without some prior information. \ Under these
340: conditions, the crystal implements
341: \begin{equation}
342: \left| 0\right\rangle +\varepsilon \left(
343: \begin{array}{c}
344: \alpha \\
345: \beta \\
346: \gamma \\
347: \delta%
348: \end{array}%
349: \right) \longrightarrow \left| 0\right\rangle +\varepsilon \left(
350: \begin{array}{c}
351: -\alpha \\
352: \beta \\
353: \gamma \\
354: \delta%
355: \end{array}%
356: \right) .
357: \end{equation}%
358: If horizontal polarization is used to represent a logical `0', this performs
359: a transformation analogous to a c-$\pi $ within the state space defined by
360: our constraint on $\alpha .$ \ We do not use the conventional c-$\pi $ so
361: that we can use the common convention for the Hadamard gate later on without
362: the need for additional quantum gates. \ We will now describe how this
363: operation can be used to perform Bell state creation under certain
364: conditions. \
365:
366: \subsection{Bell state creation}
367:
368: The circuit in Fig. 1a. is capable of converting each rectilinear basis
369: state to a different Bell state. \ To give a concrete example, we begin with
370: the qubit pair in the state $\left| 0\right\rangle _{1}\left| 0\right\rangle
371: _{2}\ $represented as the 4-vector%
372: \begin{equation}
373: \left| \psi \right\rangle =\left(
374: \begin{array}{c}
375: 1 \\
376: 0 \\
377: 0 \\
378: 0%
379: \end{array}%
380: \right) ,
381: \end{equation}%
382: where the rows now contain the amplitudes for the states $\left|
383: 0\right\rangle _{1}\left| 0\right\rangle _{2}$, $\left| 0\right\rangle
384: _{1}\left| 1\right\rangle _{2}$, $\left| 1\right\rangle _{1}\left|
385: 0\right\rangle _{2}$, and $\left| 1\right\rangle _{1}\left| 1\right\rangle
386: _{2}.$ \ The circuit contains one-qubit Hadamard transformations which are
387: defined by the 2$\times $2 matrix,
388:
389: \begin{equation}
390: H=\frac{1}{\sqrt{2}}\left[
391: \begin{array}{cc}
392: 1 & 1 \\
393: 1 & -1%
394: \end{array}%
395: \right]
396: \end{equation}%
397: and the two-qubit c-$\pi $ gate whose operation has already been discussed.
398: \ The circuit then takes the input state, $\left| \psi \right\rangle $, to
399: the output state $\left| \psi ^{\prime }\right\rangle $ given by%
400: \begin{eqnarray}
401: \left| \psi ^{\prime }\right\rangle &=&\left( H_{1}\otimes I_{2}\right)
402: \left( c\text{-}\pi \right) \left( H_{1}\otimes H_{2}\right) \left| \psi
403: \right\rangle \\
404: &=&\frac{1}{2\sqrt{2}}\left[
405: \begin{array}{cccc}
406: 1 & 0 & 1 & 0 \\
407: 0 & 1 & 0 & 1 \\
408: 1 & 0 & -1 & 0 \\
409: 0 & 1 & 0 & -1%
410: \end{array}%
411: \right] \left[
412: \begin{array}{cccc}
413: 1 & 0 & 0 & 0 \\
414: 0 & 1 & 0 & 0 \\
415: 0 & 0 & 1 & 0 \\
416: 0 & 0 & 0 & -1%
417: \end{array}%
418: \right] \left[
419: \begin{array}{cccc}
420: 1 & 1 & 1 & 1 \\
421: 1 & -1 & 1 & -1 \\
422: 1 & 1 & -1 & -1 \\
423: 1 & -1 & -1 & 1%
424: \end{array}%
425: \right] \left(
426: \begin{array}{c}
427: 1 \\
428: 0 \\
429: 0 \\
430: 0%
431: \end{array}%
432: \right) \\
433: &=&\frac{1}{\sqrt{2}}\left(
434: \begin{array}{c}
435: 1 \\
436: 0 \\
437: 0 \\
438: 1%
439: \end{array}%
440: \right) .
441: \end{eqnarray}%
442: This final state is the Bell state $\left| \phi ^{+}\right\rangle $. \ Each
443: different rectilinear state input will produce a different Bell state output
444: through this circuit.
445:
446: The conditional-phase operation can be incorporated into the optical device
447: schematically represented in Fig. 1b that can perform a very similar
448: transformation. \ Instead of using a state describing a pure photon pair as
449: input, this device requires the input pair to be in a coherent superposition
450: with the vacuum. \ As discussed previously, this is merely the output from a
451: parametric down-conversion source (Eq. \ref{dcstate}). \ Here we assume the
452: coefficients are normalized according to $\left| \alpha \right| ^{2}+\left|
453: \beta \right| ^{2}+\left| \gamma \right| ^{2}+\left| \delta \right| ^{2}=1,$
454: such that $\left| \varepsilon \right| ^{2}$ is the probability of a photon
455: pair of any polarization being present. \ The photons have been created into
456: spatial modes 1 and 2 by an initial down-conversion crystal (not shown) to
457: serve as input to the optical device in Fig. 1b. \ Hadamard operations are
458: accomplished via half-wave plates at 22.5 degrees, and the c-$\pi $ has been
459: replaced by the conditional-phase switch. \ The intial state will evolve as
460: follows through the device. \ The pair of Hadamard gates changes the general
461: state, $\left| \psi _{1}\right\rangle ,$ to $\left| \psi _{2}\right\rangle $%
462: ,
463: \begin{eqnarray}
464: \left| \psi _{2}\right\rangle &=&\left( H_{1}\otimes H_{2}\right) \left|
465: \psi _{1}\right\rangle \\
466: &=&\left| 0\right\rangle +\frac{\varepsilon }{2}\left[
467: \begin{array}{cccc}
468: 1 & 1 & 1 & 1 \\
469: 1 & -1 & 1 & -1 \\
470: 1 & 1 & -1 & -1 \\
471: 1 & -1 & -1 & 1%
472: \end{array}%
473: \right] \left(
474: \begin{array}{c}
475: \alpha \\
476: \beta \\
477: \gamma \\
478: \delta
479: \end{array}%
480: \right) \\
481: &=&\left| 0\right\rangle +\frac{\varepsilon }{2}\left(
482: \begin{array}{c}
483: \alpha +\beta +\gamma +\delta \\
484: \alpha -\beta +\gamma -\delta \\
485: \alpha +\beta -\gamma -\delta \\
486: \alpha -\beta -\gamma +\delta
487: \end{array}%
488: \right) .
489: \end{eqnarray}%
490: This state passes through the conditional-phase shift, which is
491: phase-matched to contribute an amplitude of $-\varepsilon $ for
492: horizontally-polarized photon pairs. \ It will evolve to $\left| \psi
493: _{3}\right\rangle ,$%
494: \begin{eqnarray}
495: \left| \psi _{3}\right\rangle &=&\left| 0\right\rangle +\frac{\varepsilon }{%
496: 2}\left(
497: \begin{array}{c}
498: \alpha +\beta +\gamma +\delta \\
499: \alpha -\beta +\gamma -\delta \\
500: \alpha +\beta -\gamma -\delta \\
501: \alpha -\beta -\gamma +\delta
502: \end{array}%
503: \right) -\varepsilon \left(
504: \begin{array}{c}
505: 1 \\
506: 0 \\
507: 0 \\
508: 0%
509: \end{array}%
510: \right) \\
511: &=&\left| 0\right\rangle +\frac{\varepsilon }{2}\left(
512: \begin{array}{c}
513: \alpha +\beta +\gamma +\delta -2 \\
514: \alpha -\beta +\gamma -\delta \\
515: \alpha +\beta -\gamma -\delta \\
516: \alpha -\beta -\gamma +\delta
517: \end{array}%
518: \right) .
519: \end{eqnarray}%
520: The final Hadamard gate acts only on mode 1, and converts $\left| \psi
521: _{3}\right\rangle $ to the output state $\left| \psi ^{\prime }\right\rangle
522: $,
523:
524: \begin{eqnarray}
525: \left| \psi ^{\prime }\right\rangle &=&\left( H_{1}\otimes I_{2}\right)
526: \left| \psi _{3}\right\rangle \\
527: &=&\left| 0\right\rangle +\frac{\varepsilon }{2\sqrt{2}}\left[
528: \begin{array}{cccc}
529: 1 & 0 & 1 & 0 \\
530: 0 & 1 & 0 & 1 \\
531: 1 & 0 & -1 & 0 \\
532: 0 & 1 & 0 & -1%
533: \end{array}%
534: \right] \left(
535: \begin{array}{c}
536: \alpha +\beta +\gamma +\delta -2 \\
537: \alpha -\beta +\gamma -\delta \\
538: \alpha +\beta -\gamma -\delta \\
539: \alpha -\beta -\gamma +\delta%
540: \end{array}%
541: \right) \\
542: &=&\left| 0\right\rangle +\frac{\varepsilon }{\sqrt{2}}\left(
543: \begin{array}{c}
544: \alpha +\beta -1 \\
545: \alpha -\beta \\
546: \gamma +\delta -1 \\
547: \gamma -\delta%
548: \end{array}%
549: \right) .
550: \end{eqnarray}
551:
552: If, for example, the input state to this device had only an amplitude for a
553: horizontally-polarized photon pair (i.e. $\alpha =1$ and $\beta ,\gamma
554: ,\delta =0$), then the output state would be,%
555: \begin{eqnarray}
556: \left| \psi ^{\prime }\right\rangle &=&\left| 0\right\rangle +\frac{%
557: \varepsilon }{\sqrt{2}}\left(
558: \begin{array}{c}
559: 0 \\
560: 1 \\
561: -1 \\
562: 0%
563: \end{array}%
564: \right) \\
565: &=&\left| 0\right\rangle -\varepsilon \left| \psi ^{-}\right\rangle .
566: \end{eqnarray}%
567: The other 3 possible rectilinear basis inputs would each evolve to a
568: different Bell state in a coherent superposition with the vacuum state. \
569: The resulting transformations on four possible rectilinear input states are
570: \
571: \begin{eqnarray}
572: \left| 0\right\rangle +\varepsilon \left| H\right\rangle _{1}\left|
573: H\right\rangle _{2} &\longrightarrow &\left| 0\right\rangle -\varepsilon
574: \left| \psi ^{-}\right\rangle \notag \\
575: \left| 0\right\rangle +\varepsilon \left| H\right\rangle _{1}\left|
576: V\right\rangle _{2} &\longrightarrow &\left| 0\right\rangle -\varepsilon
577: \left| \psi ^{+}\right\rangle \notag \\
578: \left| 0\right\rangle +\varepsilon \left| V\right\rangle _{1}\left|
579: H\right\rangle _{2} &\longrightarrow &\left| 0\right\rangle -\varepsilon
580: \left| \phi ^{-}\right\rangle \notag \\
581: \left| 0\right\rangle +\varepsilon \left| V\right\rangle _{1}\left|
582: V\right\rangle _{2} &\longrightarrow &\left| 0\right\rangle -\varepsilon
583: \left| \phi ^{+}\right\rangle .
584: \end{eqnarray}%
585: \
586:
587: \subsection{Bell state detection}
588:
589: The method just described for creating polarization Bell states is much more
590: experimentally difficult than the elegant methods of doing so in a
591: cleverly-oriented crystal or crystal pair \cite{kwiat1,kwiat2}. \ What is
592: unique about this method is that this device performs a one-to-one
593: transformation between rectilinear basis states and Bell basis states. \
594: This device for creating the Bell states can, in fact, be run in reverse to
595: distinguish between the four Bell states provided, again, that they are in a
596: superposition with vacuum. \ Fig. 2a. shows a quantum circuit for
597: transforming Bell states to the rectilinear basis, that is very similar in
598: structure to the circuit shown in Fig. 1a. \ To give a concrete example, we
599: can trace the evolution of the singlet state, $\left| \phi ^{-}\right\rangle
600: $, through the device. \ The singlet state can be written in 4-vector
601: notation as,%
602: \begin{equation}
603: \left| \psi ^{-}\right\rangle =\frac{1}{\sqrt{2}}\left(
604: \begin{array}{c}
605: 0 \\
606: -1 \\
607: 1 \\
608: 0%
609: \end{array}%
610: \right) .
611: \end{equation}%
612: The circuit transforms the input state to the output $\left| \psi ^{\prime
613: }\right\rangle $ in the following way,%
614: \begin{eqnarray}
615: \left| \psi ^{\prime }\right\rangle &=&\left( H_{1}\otimes H_{2}\right)
616: \left( c\text{-}\pi \right) \left( H_{1}\otimes I_{2}\right) \left| \phi
617: ^{-}\right\rangle \\
618: &=&\frac{1}{2}\left[
619: \begin{array}{cccc}
620: 1 & 1 & 1 & 1 \\
621: 1 & -1 & 1 & -1 \\
622: 1 & 1 & -1 & -1 \\
623: 1 & -1 & -1 & 1%
624: \end{array}%
625: \right] \left[
626: \begin{array}{cccc}
627: 1 & 0 & 0 & 0 \\
628: 0 & 1 & 0 & 0 \\
629: 0 & 0 & 1 & 0 \\
630: 0 & 0 & 0 & -1%
631: \end{array}%
632: \right] \frac{1}{\sqrt{2}}\left[
633: \begin{array}{cccc}
634: 1 & 0 & 1 & 0 \\
635: 0 & 1 & 0 & 1 \\
636: 1 & 0 & -1 & 0 \\
637: 0 & 1 & 0 & -1%
638: \end{array}%
639: \right] \frac{1}{\sqrt{2}}\left(
640: \begin{array}{c}
641: 0 \\
642: -1 \\
643: 1 \\
644: 0%
645: \end{array}%
646: \right) \\
647: &=&\left(
648: \begin{array}{c}
649: 0 \\
650: 0 \\
651: 0 \\
652: 1%
653: \end{array}%
654: \right) .
655: \end{eqnarray}%
656: The output state is the product state $\left| 1\right\rangle _{1}\left|
657: 1\right\rangle _{2}$. \
658:
659: The optical device that performs the analogous transformation is shown in
660: Fig. 2b. \ The device, again, uses half-wave plates to implement the
661: Hadamard transformations, and the conditional-phase switch which is set to
662: contribute an amplitude of +$\varepsilon $ for a horizontally-polarized
663: photon pair. \ The input state to this device, $\left| \psi
664: _{1}\right\rangle ,$ is again described by the general down-conversion state,%
665: \begin{equation}
666: \left| \psi _{1}\right\rangle =\left| 0\right\rangle +\varepsilon \left(
667: \begin{array}{c}
668: \alpha \\
669: \beta \\
670: \gamma \\
671: \delta%
672: \end{array}%
673: \right) .
674: \end{equation}%
675: This state passes through the polarization rotator in mode 1 and will evolve
676: to the state $\left| \psi _{2}\right\rangle $,%
677: \begin{eqnarray}
678: \left| \psi _{2}\right\rangle &=&\left( H_{1}\otimes I_{2}\right) \left|
679: \psi _{1}\right\rangle \\
680: &=&\left| 0\right\rangle +\frac{\varepsilon }{\sqrt{2}}\left[
681: \begin{array}{cccc}
682: 1 & 0 & 1 & 0 \\
683: 0 & 1 & 0 & 1 \\
684: 1 & 0 & -1 & 0 \\
685: 0 & 1 & 0 & -1%
686: \end{array}%
687: \right] \left(
688: \begin{array}{c}
689: \alpha \\
690: \beta \\
691: \gamma \\
692: \delta%
693: \end{array}%
694: \right) \\
695: &=&\left| 0\right\rangle +\frac{\varepsilon }{\sqrt{2}}\left(
696: \begin{array}{c}
697: \alpha +\gamma \\
698: \beta +\delta \\
699: \alpha -\gamma \\
700: \beta -\delta%
701: \end{array}%
702: \right) .
703: \end{eqnarray}%
704: This state is subsequently passed through the conditional-phase switch where
705: the pump laser is set to the appropriate amplitude and phase to add an
706: amplitude of $+\varepsilon $ for a vertically-polarized photon pair. \ The
707: state evolves to $\left| \psi _{3}\right\rangle $ where%
708: \begin{eqnarray}
709: \left| \psi _{3}\right\rangle &=&\left| 0\right\rangle +\frac{\varepsilon }{%
710: \sqrt{2}}\left(
711: \begin{array}{c}
712: \alpha +\gamma \\
713: \beta +\delta \\
714: \alpha -\gamma \\
715: \beta -\delta%
716: \end{array}%
717: \right) +\varepsilon \left(
718: \begin{array}{c}
719: 1 \\
720: 0 \\
721: 0 \\
722: 0%
723: \end{array}%
724: \right) \\
725: &=&\left| 0\right\rangle +\frac{\varepsilon }{\sqrt{2}}\left(
726: \begin{array}{c}
727: \alpha +\gamma +\sqrt{2} \\
728: \beta +\delta \\
729: \alpha -\gamma \\
730: \beta -\delta%
731: \end{array}%
732: \right) .
733: \end{eqnarray}%
734: Finally, this state passes through a pair of half-wave plates. \ The final
735: state, $\left| \psi ^{\prime }\right\rangle $, is%
736: \begin{eqnarray}
737: \left| \psi ^{\prime }\right\rangle &=&\left| 0\right\rangle +\frac{1}{2}%
738: \left[
739: \begin{array}{cccc}
740: 1 & 1 & 1 & 1 \\
741: 1 & -1 & 1 & -1 \\
742: 1 & 1 & -1 & -1 \\
743: 1 & -1 & -1 & 1%
744: \end{array}%
745: \right] \frac{\varepsilon }{\sqrt{2}}\left(
746: \begin{array}{c}
747: \alpha +\gamma +\sqrt{2} \\
748: \beta +\delta \\
749: \alpha -\gamma \\
750: \beta -\delta%
751: \end{array}%
752: \right) \\
753: &=&\left| 0\right\rangle +\sqrt{2}\varepsilon \left(
754: \begin{array}{c}
755: \alpha +\beta +\frac{1}{\sqrt{2}} \\
756: \alpha -\beta +\frac{1}{\sqrt{2}} \\
757: \gamma +\delta +\frac{1}{\sqrt{2}} \\
758: \gamma -\delta +\frac{1}{\sqrt{2}}%
759: \end{array}%
760: \right) .
761: \end{eqnarray}%
762: If, for example our input state has $\alpha =\delta =-1/\sqrt{2}$ and $\beta
763: =\gamma =0$ (i.e. the input is $\left| 0\right\rangle -\varepsilon \left|
764: \phi ^{+}\right\rangle $ -- one of the outputs of the previous device), then
765: the output state would be,%
766: \begin{eqnarray}
767: \left| \psi ^{\prime }\right\rangle &=&\left| 0\right\rangle +\sqrt{2}%
768: \varepsilon \left(
769: \begin{array}{c}
770: 0 \\
771: 0 \\
772: 0 \\
773: \sqrt{2}%
774: \end{array}%
775: \right) \\
776: &=&\left| 0\right\rangle +\varepsilon \left(
777: \begin{array}{c}
778: 0 \\
779: 0 \\
780: 0 \\
781: 1%
782: \end{array}%
783: \right) .
784: \end{eqnarray}%
785: That is, the output contains only an amplitude for a photon pair in the
786: product state $\left| V\right\rangle _{1}\left| V\right\rangle _{2}$. \ The
787: results for all of the input states are simply stated:%
788: \begin{eqnarray}
789: \left| 0\right\rangle -\varepsilon \left| \psi ^{-}\right\rangle
790: &\longrightarrow &\left| 0\right\rangle +\varepsilon \left| H\right\rangle
791: _{1}\left| H\right\rangle _{2} \notag \\
792: \left| 0\right\rangle -\varepsilon \left| \psi ^{+}\right\rangle
793: &\longrightarrow &\left| 0\right\rangle +\varepsilon \left| H\right\rangle
794: _{1}\left| V\right\rangle _{2} \notag \\
795: \left| 0\right\rangle -\varepsilon \left| \phi ^{-}\right\rangle
796: &\longrightarrow &\left| 0\right\rangle +\varepsilon \left| V\right\rangle
797: _{1}\left| H\right\rangle _{2} \notag \\
798: \left| 0\right\rangle -\varepsilon \left| \phi ^{+}\right\rangle
799: &\longrightarrow &\left| 0\right\rangle +\varepsilon \left| V\right\rangle
800: _{1}\left| V\right\rangle _{2},
801: \end{eqnarray}%
802: and are the inverse of the transformation the previous device
803: performed.\qquad
804:
805: In order to complete the measurement of the Bell state, the output of this
806: device is passed through an optical device like the one in Fig. 3. \ The
807: detection of a photon pair constitutes a successful measurement and will
808: occur with probability $\left| \varepsilon \right| ^{2}$ -- the probability
809: of having a Bell state in our input state. \ This probability ignores issues
810: of detector and path efficiency.
811:
812: \section{Discussion}
813:
814: We have proposed a way of implementing a transformation capable of
815: converting the polarization state of a pair of photons from the rectilinear
816: basis to the Bell state basis and vice versa provided the photon pairs are
817: in a known coherent superposition with the vacuum. \ This transformation
818: relies on a recently reported effective nonlinearity at the single-photon
819: level \cite{phase}. \ Requiring the photon pair to be in a superposition
820: with the vacuum seems unusual, but this type of superposition exists in all
821: down-conversion\ sources of entangled photons. \ It is only upon performing
822: a photon-counting coincidence measurement that the maximally-entangled
823: behaviour is projected out. \ While these down-conversion\ sources of Bell
824: states exist and are practical in the lab, the creation mechanism does not
825: suggest how one might try to measure those Bell states. \ In the device
826: discussed here, the Bell state creator and Bell state analyzer look very
827: similar. \ The creator can essentially be run in reverse to make the
828: analyzer. \
829:
830: This device cannot be used for performing unconditional quantum
831: teleportation. \ The device is only capable of distinguishing the four Bell
832: states; it is not capable of performing a general projective measurement in
833: the Bell basis. \ This is due to the conditional-phase shifter's dependence
834: on the magnitude and phase of the amplitude for the Bell state component in
835: the input state; the gate does not operate properly on arbitrary
836: superpositions of Bell states.\ Nevertheless, the device discussed herein
837: constitutes a novel way of manipulating the degree of entanglement between a
838: pair of photons, and may find a use in other quantum optics applications,
839: such as dense coding \cite{densetheory,denseexp}. \ The ability to entangle
840: and disentangle photon pairs is a crucial step toward building scalable
841: all-optical quantum computers.
842:
843: We would like to thank Andrew White and Ray Laflamme for valuable
844: discussions. \ We are grateful for the financial support of Photonics
845: Research Ontario, NSERC, and the US Air Force Office of Scientific Research
846: (F49620-01-1-0468).
847:
848: \bigskip
849:
850: \begin{thebibliography}{99}
851: \bibitem{bennbrass} C. H. Bennett, G. Brassard, C. Cr\'{e}peau, R. Jozsa, A.
852: Peres, and W. K. Wootters, \emph{Phys. Rev. Lett.}, \textbf{70}, 1895 (1993).
853:
854: \bibitem{zeil} D. Bouwmeester, J.-W. Pan, K. Mattle, M. Eibl, H. Weinfurter,
855: A. Zeilinger, \emph{Nature}, \textbf{390}, 575 (1997).
856:
857: \bibitem{demartini} D. Boschi, S. Branca, F. De Martini, L. Hardy, and S.
858: Popescu, \emph{Phys. Rev. Lett.}, \textbf{80}, 1121 (1998).
859:
860: \bibitem{densetheory} C. H. Bennett and S. J. Wiesner, \emph{Phys. Rev. Lett.%
861: }, \textbf{69}, 2881 (1992).
862:
863: \bibitem{denseexp} K. Mattle, H. Weinfurter, P.G. Kwiat, and A. Zeilinger,
864: \emph{Phys. Rev. Lett.}, \textbf{76}, 4656 (1996).
865:
866: \bibitem{cryptbennbrass} Bennett, C. H. \& Brassard, G., \emph{Proceedings
867: of the IEEE International Conference on Computers, Systems \& Signal
868: Processing, Bangalore, India }(IEEE, New\ York, 1984), p. 175-179.
869:
870: \bibitem{cryptekert} A. K. Ekert, \emph{Phys. Rev. Lett.},\textbf{67}, 661
871: (1991).
872:
873: \bibitem{clauserbell} S. J. Freedman and J. F. Clauser, \emph{Phys. Rev.
874: Lett.}, \textbf{28}, 938 (1972); J. F. Clauser, Phys. Rev. D, \textbf{9},
875: 853 (1974); A. Aspect, P. Grangier, and G. Roger, \emph{Phys. Rev. Lett.},
876: \textbf{47}, 460 (1981).
877:
878: \bibitem{othersources} Z. Y. Ou and L. Mandel, \emph{Phys. Rev. Lett.},
879: \textbf{61}, 50 (1988); Y. H. Shih and C. O. Alley, \emph{Phys. Rev. Lett.},
880: \textbf{61}, 2921 (1988); T. E. Kiess, Y. H. Shih, A. V. Sergienko, and C.
881: O. Alley, \emph{Phys. Rev. Lett.}, \textbf{71}, 3893 (1993).
882:
883: \bibitem{steinberg} Optical tests of quantum mechanics, R.Y. Chiao, P.G.
884: Kwiat, and A.M. Steinberg, in \emph{Advances in Atomic, Molecular, and
885: Optical Physics}, edited by B. Berson and H. Walther (Academic Press, New
886: York, 1994), Vol. 34.
887:
888: \bibitem{kwiat1} P.\ G. Kwiat, K. Mattle, H. Weinfurter, and A. Zeilinger,
889: \emph{Phys. Rev. Lett.}, \textbf{75}, 4337 (1995)
890:
891: \bibitem{kwiat2} P. G. Kwiat, E. Waks, A. G. White, and P. H. Eberhard,
892: \emph{Phys. Rev. A}, \textbf{60}, R773 (1999).
893:
894: \bibitem{shapiro} J. H. Shapiro and N. C. Wong, J. Opt. B:\ Quant. and
895: Semiclass. Opt. \textbf{2}, L1 (2000).
896:
897: \bibitem{circuits2} N. Lutkenhaus, J. Calsamiglia, and K.-A. Suominen, \emph{%
898: Phys. Rev. A}, \textbf{59}, 3295 (1999).
899:
900: \bibitem{shihtelep} Y.-H. Kim, S. P. Kulik, and Y. Shih, \emph{Phys. Rev.
901: Lett.}, \textbf{86}, 1370 (2001).
902:
903: \bibitem{qed} Q. A. Turchette, C. J. Hood, W. Lange, H. Mabuchi, and H. J.
904: Kimble, \emph{Phys. Rev. Lett.}, \textbf{75}, 4710 (1995); A.
905: Rauschenbeutel, G. Nogues, S. Osnaghi, P. Bertet, M. Brune, J. M. Raimond,
906: and S. Haroche, \emph{Phys. Rev. Lett.}, \textbf{83}, 5166 (1999).
907:
908: \bibitem{eit} S.\ E. Harris, and L.V. Hau, \emph{Phys. Rev. Lett.}, \textbf{%
909: 82}, 4611 (1999); M. M. Kash, V. A. Sautenkov, A. S. Zibrov, L. Hollberg, G.
910: R. Welch, M. D. Lukin, Y. Rostovtsev, E. S. Fry, and M. O. Scully, \emph{%
911: Phys. Rev. Lett.}, \textbf{82}, 5229 (1999).
912:
913: \bibitem{pei} J. D. Franson, \emph{Phys. Rev. Lett.}, \textbf{78}, 3852
914: (1997).
915:
916: \bibitem{switch} K. J. Resch, J. S. Lundeen, and A. M. Steinberg, \emph{%
917: Phys. Rev. Lett.}, \textbf{87}, 123603 (2001); K. J. Resch, J. S. Lundeen,
918: and A. M. Steinberg, \emph{Journal of Modern Optics}, \textbf{49}, 487
919: (2002).
920:
921: \bibitem{phase} K. J. Resch, J. S. Lundeen, and A. M. Steinberg, pre-print:
922: quant-ph/0205109, submitted to \emph{Phys. Rev. Lett.} (2002).
923:
924: \bibitem{KLM} E. Knill, R. Laflamme, and G. Milburn, \emph{Nature}, \textbf{%
925: 409}, 46 (2001).
926:
927: \bibitem{otherlin} D. Gottesman, A. Kitaev, and J. Preskill, \emph{Phys.
928: Rev. A}, \textbf{64}, 012310 (2001).
929:
930: \bibitem{embedded} P. G. Kwiat and H. Weinfurter, \emph{Phys. Rev. A},
931: \textbf{58}, R2623 (1998).
932:
933: \bibitem{cphiref} M. A. Neilsen and I. L. Chuang, Quantum Computation and
934: Quantum Information (Cambridge University Press, Cambridge, 2000), p. 294.
935:
936: \bibitem{circuits} M. A. Neilsen and I. L. Chuang, Quantum Computation and
937: Quantum Information (Cambridge University Press, Cambridge, 2000), p. 26.
938:
939: \bibitem{railcross} T. J. Herzog, J. G. Rarity, H. Weinfurter, and A.
940: Zeilinger, \emph{Phys. Rev. Lett.}, \textbf{72}, 629 (1994); H. Weinfurter,
941: T. Herzog, P. G. Kwiat, J. G. Rarity, A. Zeilinger, and M. Zukowski, \emph{%
942: Ann. N.Y. Acad. Sci.}, \textbf{755}, 61 (1995).
943:
944: \pagebreak
945:
946: \textbf{Figure Captions}
947: \end{thebibliography}
948:
949: \textbf{Fig. 1.\ a) A quantum circuit and b) its optical analogue for the
950: creation of Bell states from product states. \ a) The quantum circuit acts
951: on a pair of input modes 1 and 2. \ The circuit uses one-qubit Hadamard
952: gates, and a two-qubit controlled-}$\pi $\textbf{\ gate. \ This circuit
953: performs a unitary transformation on the inputs and takes each of the four
954: possible qubit product states to a different Bell state. \ b) The optical
955: analogue of the quantum circuit. \ In the diagram, }$\lambda /2$ \textbf{are
956: half-wave plates oreinted at 22.5 degrees and }$\chi ^{(2)}$ \textbf{is a
957: nonlinear material. \ \ The device is capable of converting the state of a
958: photon pair in a product state of polarization to one of the Bell states,
959: provided that the input is in the correct superposition with the vacuum. \ }
960:
961: \textbf{Fig. 2. a) A quantum circuit and b) its optical analogue for the
962: conversion of Bell states to product states. \ a) This quantum circuit takes
963: a pair of qubits in input modes 1 and 2 and performs a unitary
964: transformation that will convert a Bell state to a product state. \ b) The
965: optical analogue of the quantum circuit takes a photon pair in a Bell state
966: to a rectilinear product state, provided the photon pair is in the correct
967: superposition with the vacuum. }
968:
969: \textbf{\ \ }
970:
971: \textbf{Fig. 3. An optical device for distinguishing rectlinear basis
972: states. \ This simple device can distinguish between the product states for
973: the polarization of a pair of photons }$\left| H\right\rangle _{1}\left|
974: H\right\rangle _{2}$\textbf{, }$\left| H\right\rangle _{1}\left|
975: V\right\rangle _{2}$\textbf{, }$\left| V\right\rangle _{1}\left|
976: H\right\rangle _{2}$\textbf{, and }$\left| V\right\rangle _{1}\left|
977: V\right\rangle _{2}$\textbf{, where the subscripts 1 and 2 are mode labels.
978: \ The device consists of a pair of polarizing beam-splitters\ (PBS) and 4
979: photon counting detectors monitoring their outputs. \ For example, the
980: detection of a photon at detector 1 and detector 4 corresponds to the state }%
981: $\left| H\right\rangle _{1}\left| V\right\rangle _{2}$\textbf{.}
982:
983: \textbf{Fig. 4. Schematic for the conditional-phase switch. \ A strong,
984: classical, laser in mode p, of frequency 2}$\omega $\textbf{,} \textbf{pumps
985: a }$\chi ^{(2)}$ \textbf{nonlinear material such that it can create
986: down-conversion pairs in modes 1 and 2. \ A pair of input beams, of
987: frequency }$\omega $\textbf{, pass through the nonlinear material into modes
988: 1 and 2. \ Interference between the multiple paths leading to photon pairs
989: at the output can be used to introduce a large phase shift on the amplitude
990: for a photon pair.}
991:
992: \end{document}
993: