1: \documentstyle[aps,prl,epsf]{revtex}
2:
3:
4: \def\be{\begin{equation}}
5: \def\ee{\end{equation}}
6: \def\bea{\begin{eqnarray}}
7: \def\eea{\end{eqnarray}}
8: \def\sp{{\mathcal H}_2}
9: \def\ep{{\vec\epsilon}\ }
10: \newcommand{\la}{\langle}
11: \newcommand{\ra}{\rangle}
12: \begin{document}
13: \draft
14: \title{Three-party entanglement from positronium}
15: \author{A. Ac\'{\i}n, J. I. Latorre and P. Pascual}
16: \address{Departament d'Estructura i Constituents de la
17: Mat\`eria, Universitat de Barcelona, Diagonal 647, E-08028 Barcelona, Spain.\\
18: e-mail: acin@ecm.ub.es
19: }
20: \date{\today}
21: \maketitle
22: %%%%%%%%%%%% Abstract %%%%%%%%%%%%%%%%%%%%%%%%%%%
23: \begin{abstract}
24: The decay of ortho-positronium into three photons produces a
25: physical realization of a pure state with three-party
26: entanglement. Its quantum correlations are analyzed using recent
27: results on quantum information theory, looking for the final
28: state which has the maximal amount of GHZ-like correlations. This
29: state allows for a statistical dismissal of local realism
30: stronger than the one obtained using any entangled state of
31: two spin one-half particles.
32: \end{abstract}
33: \pacs{PACS Nos. 03.65.Bz, 03.67.-a, 12.20.-m}
34: \bigskip
35:
36: \section{Introduction}
37: Entanglement or quantum correlations between many space-separated
38: subsystems has been recognized as one of the most intrinsic
39: properties of quantum mechanics and provides the basis for many
40: genuine applications of quantum information theory. It is, then,
41: quite natural to look for physical situations in which quantum
42: entangled states are obtained. Most of the theoretical and
43: experimental effort has so far been devoted to unveil physical
44: realizations of quantum states describing two quantum correlated
45: subsystems. The search for physical systems displaying clean
46: three-party entanglement is not simple. In this paper, we shall
47: analyze decays of particles as a natural scenario for fulfilling
48: such a goal. More precisely, we shall show that the decay of
49: ortho-positronium into three photons corresponds to a highly
50: entangled state. Let us now review what entanglement can be used
51: for and why it is interesting to look for quantum correlation
52: between more than two particles.
53:
54: In 1935 Einstein, Podolsky and Rosen \cite{EPR}, starting from
55: three reasonable assumptions of locality, reality and
56: completeness that every physical theory must satisfy, argued that
57: quantum mechanics (QM) is an incomplete theory. They did not
58: question quantum mechanics predictions but rather quantum
59: mechanics interpretation \cite{GHZ}. Their argument was based on
60: some inconsistencies between quantum mechanics and their
61: local-realistic premises (LR) which appear for quantum states of
62: bipartite systems, $|\psi\ra\in{\mathcal H}_{d_1}\otimes{\mathcal
63: H}_{d_2}$. It was in 1964 when Bell \cite{Bell} showed that any
64: theory compatible with LR assumptions can not reproduce some of
65: the statistical predictions of QM, using a gedankenexperiment
66: proposed in \cite{BA} with two quantum correlated
67: spin-$\frac{1}{2}$ particles in the singlet state
68: \be
69: |s\ra=\frac{1}{\sqrt 2}\left(|01\ra-|10\ra\right) .
70: \ee
71: In his
72: derivation, as it is well-known, quantum correlations or
73: entanglement have a crucial role. Actually, the singlet state is
74: known to be the maximally entangled state between two particles.
75: The conflict between LR and QM arises since the latter violates
76: some experimentally verifiable inequalities, called Bell
77: inequalities, that any theory according to the local-realistic
78: assumptions ought to satisfy. It is then possible to design real
79: experiments testing QM against LR (for a detailed discussion see
80: \cite{CS}). Correlations of linear polarizations of pair of
81: photons were measured in 1982 showing strong agreement with
82: quantum mechanichs predictions and violating Bell inequalities
83: \cite{ADR}. Nowadays, Bell inequalities have been tested thoroughly in favor
84: of QM \cite{exptest}.
85:
86: More recently, it has been pointed out that some predictions for
87: quantum systems having quantum correlations between more than two
88: particles give a much stronger conflict between LR and QM than
89: any entangled state of two particles. The maximally entangled
90: state between three spin-$\frac{1}{2}$ particles, the so-called
91: GHZ (Greenberger, Horne and Zeilinger) state \cite{GHZor}
92: \be
93: \label{GHZst} |{\rm GHZ}\ra=\frac{1}{\sqrt{2}}(|000\ra+|111\ra) ,
94: \ee
95: shows some perfect correlations incompatible with any LR
96: model (see \cite{GHZ} and also \cite{Mer} for more details). It
97: is then of obvious relevance to obtain these GHZ-like
98: correlations. Producing experimentally a GHZ state has turned out
99: to be a real challenge yet a controlled instance has been
100: produced in a quantum optics experiment \cite{BPDWZ}.
101:
102: Entanglement is then important for our basic understanding of
103: quantum mechanics. Recent developments on quantum information
104: have furthermore shown that it is also a powerful resource for
105: quantum information applications. For instance, teleportation
106: \cite{BBCJPW} uses entanglement in order to obtain surprising
107: results which are impossible in a classical context. A lot of
108: work has been performed trying to know how entanglement can be
109: quantified and manipulated. Our aim in this paper consists on
110: looking for GHZ-like correlations, which are truly three-party pure
111: state entanglement, in the decay of
112: ortho-positronium to three photons. The choice of this physical
113: system has been motivated mainly by several reasons. First, decay
114: of particles seems a very natural source of entangled particles.
115: Indeed, positronium decay to two photons was one of the physical
116: systems proposed long time ago as a source of two entangled
117: space-separated particles \cite{CHSH}. On a different line of
118: thought, some experiments for testing quantum mechanics have been
119: recently proposed using correlated neutral kaons coming from the
120: decay of a $\phi$-meson \cite{kaon}. In the case of
121: positronium, three entangled photons are obtained in the final
122: state, so it offers the opportunity of analyzing a quantum state
123: showing three-party correlations similar to other experiments
124: in quantum optics.
125:
126: The structure of the paper goes as follows. We first review the
127: quantum states emerging in both para- and ortho-positronium
128: decays. Then, we focus on their entanglement properties and
129: proceed to a modern analysis of the three-photon decay state of
130: ortho-positronium. Using techniques developed in the context
131: of quantum information theory, we show that this state allows in
132: principle for an experimental test of QM finer than the ones
133: based on the use of the singlet state. We have tried to make the paper
134: self-contained and easy to read for both particle physicists and
135: quantum information physicists. The first ones can found a
136: translation of some of the quantum information ideas to a
137: well-known situation, that is, the positronium decay to photons,
138: while the second ones can see an application of the very recent
139: techniques obtained for three-party entangled states, which allow
140: to design a QM vs LR test for a three-particle system in a
141: situation different from the GHZ state.
142:
143: \section{Positronium decays}
144:
145: \subsection{Positronium properties}
146:
147: Let us start reminding some basic facts about positronium.
148: Positronium corresponds to a $e^+\,e^-$ bound state. These two
149: spin-$\frac{1}{2}$ particles can form a state with total spin
150: equal to zero, para-positronium (p-Ps), or equal to one,
151: ortho-positronium (o-Ps). Depending on the value of its angular
152: momentum, it can decay to an even or an odd number of photons as
153: we shall see shortly.
154:
155: Positronium binding energy comes from the Coulomb attraction
156: between the electron and the positron. In the non-relativistic
157: limit, its wave function is \cite{IZ}
158: \be
159: \Psi(r)=\frac{1}{\sqrt{\pi a^3}}\,e^{-\frac{r}{a}}=
160: \int\frac{d^3p}{(2\pi)^{3/2}}\,e^{i\vec p\cdot\vec
161: r}\tilde\Psi(\vec p)=\int\frac{d^3p}{(2\pi)^{3/2}}\,e^{i\vec
162: p\cdot\vec r}\frac{\sqrt{8 a^3}}{\pi(1+a^2p^2)^2} ,
163: \ee
164: where $a=\frac{2}{m\alpha}$, i.e. twice the Bohr radius of atomic
165: hydrogen, and $m$ is the electron mass. Note that the wave
166: function takes significant values only for three-momenta such that
167: $p\lesssim\frac{1}{a}\ll m$, which is consistent with the fact
168: that the system is essentially non-relativistic.
169:
170: The parity and charge conjugation operators are equal to
171: \be
172: U_P=(-1)^{L+1}\qquad U_C=(-1)^{L+S}, \ee
173: where $L$ and $S$ are the
174: orbital and spin angular momentum. Positronium states are then
175: classified according to these quantum numbers so that the ground
176: states are $^1S_0$, with $J^{PC}=0^{-+}$, for the p-Ps and
177: $^3S_1+^3D_1$, having $J^{PC}=1^{--}$, for the o-Ps.
178:
179: Positronium
180: is an unstable bound state that can decay to photons. Since a
181: $n$-photon state transforms as $U_C|n\gamma\ra=(-1)^n|n\gamma\ra$
182: under charge conjugation, which is an exact discrete symmetry for
183: any QED process such as the decay of positronium, we have that
184: the ground state of p-Ps (o-Ps) decays to an even (odd) number of
185: photons \cite{WR}. The analysis of the decay of positronium to
186: photons can be found in a standard QED textbook \cite{IZ}.
187: Para-positronium lifetime is about 0.125 ns, while for the case
188: of ortho-positronium the lifetime is equal to approximately 0.14
189: $\mu$s \cite{Czarn}.
190:
191: The computation of positronium decays is greatly simplified due
192: to the following argument. The scale which controls the structure
193: of positronium is of the order of $\vert \vec p\vert\sim \alpha
194: m$. On the other hand, the scale for postrinomium annihilation is
195: of the order of $m$. Therefore, it is easy to prove that
196: positronium decays are only sensitive to the value of the wave
197: function at the origin. As a consequence, it is possible to
198: factor out the value of the wave function from the tree-level QED
199: final state computation \cite{IZ}. A simple computation of
200: Feymann diagrams will be enough to write the precise structure of
201: momenta and polarizations which describe the positronium decays.
202: Furthermore, only tree-level amplitudes need to be computed since
203: higher corrections are suppressed by one power of $\alpha$. Let
204: us now proceed to analyze the decays of p-Ps and o-Ps in turn.
205:
206: \subsection{Para-positronium decay}
207: Para-positronium ground state decays into two photons. Because of
208: the argument mentioned above, the determination of the two-photon
209: state coming from the p-Ps decay is simply given by the lowest
210: order Feynmann diagram of $e^+e^-\longrightarrow \gamma\gamma$.
211: Since positronium is a non-relativistic particle to a very good
212: approximation, the three-momenta of $e^+$ and
213: $e^-$ are taken equal to zero,
214: and the corresponding spinors are replaced by a two-component
215: spin. This implies that the tree-level calculation of the
216: annihilation of p-Ps into two photons is equal to, up to
217: constants,
218: \be \label{ampl}
219: {\cal M}(e^+e^-\longrightarrow \gamma\gamma)\sim\chi^{c\dagger}_+M_2\chi_- ,
220: \ee
221: where (see \cite{IZ} for more details) $\chi_{\pm}$
222: is the two-component spinor describing the fermions,
223: $\chi^{c\dagger}\equiv\chi^Ti\sigma_2$, and $M_2$ gives
224: \be M_2=\sum_{perm}(\ep_1^*\times\ep_2^*)\cdot\hat k\,I_{2\times
225: 2} \equiv A(\hat k_1,\lambda_1;\hat k_2,\lambda_2)\,I_{2\times 2}
226: , \ee where $\ep^*_i\equiv\ep^*(\hat k_i,\lambda_i)$ stands for
227: the circular polarization vector associated to the outgoing
228: photon $i$ and $I_{2\times 2}$ is the $2\times 2$ identity matrix. More
229: precisely, for a photon having the three-momentum vector
230: $\vec{k}=|\vec{k}|\hat{k}=|\vec{k}|
231: (\sin\theta\cos\phi,\sin\theta\sin\phi,\cos\theta)$, the
232: polarization vectors can be chosen
233: \begin{equation}
234: \label{epsilon}
235: \ep(\hat k, \lambda)= -\frac{\lambda}{\sqrt 2}
236: \left(\cos\theta \cos \phi- i \lambda
237: \sin\phi,\cos\theta\sin\phi+i\lambda\cos\phi,-\sin \theta\right) ,
238: \end{equation}
239: where $\lambda=\pm1$ and they obey
240: \begin{eqnarray}
241: \label{epsprop}
242: &&\hat k\cdot\ep(\hat k,\lambda)=0
243: \qquad \hat k\times \ep(\hat k,\lambda)=-i \lambda
244: \ep(\hat k,\lambda) \nonumber
245: \\
246: && \ep(\hat k_i,\lambda_i)\cdot\ep(\hat k_j,\lambda_j)
247: =-{1\over 2}\left( 1-\lambda_i\lambda_j \hat k_i \cdot \hat k_j\right) .
248: \end{eqnarray}
249: From the expressions of the polarizaton vectors and the
250: three-momentum and energy conservation, it follows that the
251: scalar term $A$ is \be A(\hat k,\lambda_1;-\hat
252: k,\lambda_2)=-\frac{i}{2}(\lambda_1+\lambda_2), \ee and it
253: verifies \bea \label{scalar}
254: && A(\hat k,+1;-\hat k,+1)=-A(\hat k,-1;-\hat k,-1) \nonumber\\
255: && A(\hat k,+1;-\hat k,-1)=-A(\hat k,+1;-\hat k,-1)=0 .
256: \eea
257: The two fermions in the para-positonium ground state are in
258: the singlet state, $|S$$=$$0,S_z$$=$$0\ra=\frac{1}{\sqrt{2}}
259: \left(|\frac{1}{2},-\frac{1}{2}\ra-|-\frac{1}{2},\frac{1}{2}\ra\right)$,
260: and then, using the previous relations for $A$ and (\ref{ampl}),
261: the two-photon state resulting of the p-Ps desintegration is
262: \be
263: \label{p-Psdes}
264: |\psi_p\ra=\frac{1}{\sqrt{2}}\left(|++\ra-|--\ra\right) .
265: \ee
266: The two-photon state resulting from p-Ps decay is thus equivalent
267: to a maximally entangled state of two spin-$\frac{1}{2}$ particles.
268: This is a well-known result and was, actually, one of the physical system
269: first proposed as a source of particles having the quantum
270: correlations needed to test QM vs LR \cite{CHSH}.
271:
272: \subsection{Ortho-positronium decay}
273: The ground state of ortho-positronium has $J^{PC}=1^{--}$ and,
274: due to the fact that charge conjugation is conserved, decays to
275: three photons. Repeating the treatment performed for the p-Ps
276: annihilation, the determination of the three-photon state
277: resulting from the o-Ps decay requires the simple calculation of
278: the tree-level Feynmann diagrams corresponding to
279: $e^+e^-\longrightarrow \gamma\gamma\gamma$. Its tree-level
280: computation gives, up to constants,
281: \be \label{ampl3}
282: {\cal M}(e^+e^-\longrightarrow \gamma\gamma\gamma)\sim\chi^{c\dagger}_+M_3\chi_- ,
283: \ee
284: and the $2\times 2$ matrix $M_3$ is equal to \cite{IZ}
285: \be
286: M_3=\sum_{cyclic\,perm.}
287: \left(\left(\ep^*_2 \cdot \ep^*_3-\vec\delta_2
288: \cdot \vec\delta_3\right) \ep^*_1 +
289: \left(\ep^*_2\cdot \vec\delta_3+
290: \ep^*_3\cdot \vec\delta_2\right) \vec\delta_1
291: \right) \cdot \vec\sigma ,
292: \ee
293: where
294: \begin{equation}
295: \vec \delta_i=\vec k_i\times \ep^*_i .
296: \end{equation}
297: Using (\ref{epsprop}) we can rewrite $M_3$ in the following way
298: \begin{equation}
299: \label{amplvec}
300: M_3\equiv \vec\sigma\cdot
301: \vec V(\hat k_1,\lambda_1;\hat k_2,\lambda_2;\hat k_3,\lambda_3) ,
302: \end{equation}
303: where
304: \begin{eqnarray}
305: \label{vector}
306: \vec V=
307: &&\left( (\lambda_1-\lambda_2)(\lambda_2+\lambda_3)\ \ep^*(\hat
308: k_1,\lambda_1) \left(\ep^*(\hat k_2,\lambda_2)\cdot
309: \ep^*(\hat k_3,\lambda_3)\right)\right. \nonumber
310: \\
311: &&+(\lambda_2-\lambda_3)(\lambda_3+\lambda_1)\ \ep^*(\hat
312: k_2,\lambda_2) \left(\ep^*(\hat k_3,\lambda_3)\cdot
313: \ep^*(\hat k_1,\lambda_1)\right) \nonumber
314: \\
315: &&\left. +(\lambda_3-\lambda_1)(\lambda_1+\lambda_2)\ \ep^*(\hat
316: k_3,\lambda_3) \left(\ep^*(\hat k_1,\lambda_1)\cdot
317: \ep^*(\hat k_2,\lambda_2)\right)
318: \right) .
319: \end{eqnarray}
320: Notice that the helicity coefficient
321: $(\lambda_i-\lambda_j)(\lambda_j+\lambda_k)$ for the cyclic permutations
322: of $ijk$ explicitly enforces the vanishing of the $(+++)$ and $(---)$
323: polarizations,
324: \begin{equation}
325: \vec V(\hat k_1,+;\hat k_2,+;\hat k_3,+)=
326: \vec V(\hat k_1,-;\hat k_2,-;\hat k_3,-)= 0 .
327: \end{equation}
328: On the other hand, the rest of structures are different from zero
329: \begin{eqnarray}
330: \label{nzterms}
331: && \vec V(\hat k_1,-;\hat k_2,+;\hat k_3,+)= 2\,
332: \ep^*(\hat k_1,-) (1-\hat k_2\cdot\hat k_3) \nonumber
333: \\
334: && \vec V(\hat k_1,+;\hat k_2,-;\hat k_3,-)= 2\,
335: \ep^*(\hat k_1,+) (1-\hat k_2\cdot\hat k_3) ,
336: \end{eqnarray}
337: and similar expressions for the other cyclic terms.
338:
339: The original $e^+\,e^-$ in the ortho-positronium could
340: be in any of the three triplet states. It can be shown,
341: using (\ref{ampl3}) and (\ref{amplvec}), that when the initial
342: positronium state is $|S$$=$$1,S_z$$=$$1\ra=|\frac{1}{2},\frac{1}{2}\ra$,
343: the decay amplitude is proportional to $V_1+iV_2$,
344: while the same argument gives $-V_1+iV_2$ for
345: $|S$$=$$1,S_z$$=$$-1\ra=|-\frac{1}{2},-\frac{1}{2}\ra$ and
346: $-\sqrt{2} V_3$ for
347: $|S$$=$$1,S_z$$=$$0\ra=\frac{1}{\sqrt{2}}(|\frac{1}{2},-\frac{1}{2}
348: \ra+|-\frac{1}{2},\frac{1}{2}\ra)$. Now, considering the explicit
349: expressions of the polarization vectors (\ref{epsilon}), with
350: $\theta=\frac{\pi}{2}$ without loss of generality, and
351: (\ref{nzterms}), it is easy to see that the three-photon state
352: coming from the o-Ps decay is, up to normalization, \bea
353: \label{finst}
354: |\psi_0(\hat k_1,\hat k_2,\hat k_3)\ra = &&(1-\hat k_1\cdot \hat k_2)(|++-\ra+|--+\ra) \nonumber \\
355: +&&(1-\hat k_1\cdot \hat k_3)(|+-+\ra+|-+-\ra) \nonumber \\
356: +&&(1-\hat k_2\cdot \hat k_3)(|-++\ra+|+--\ra) ,
357: \eea
358: when the third component of the ortho-positronium spin, $S_z$, is equal to zero, and
359: \bea
360: \label{finst2}
361: |\psi_1(\hat k_1,\hat k_2,\hat k_3)\ra = &&(1-\hat k_1\cdot \hat k_2)(|++-\ra-|--+\ra) \nonumber \\
362: +&&(1-\hat k_1\cdot \hat k_3)(|+-+\ra-|-+-\ra) \nonumber \\
363: +&&(1-\hat k_2\cdot \hat k_3)(|-++\ra-|+--\ra) , \eea when
364: $S_z=\pm 1$.
365:
366: The final state of the o-Ps decay is, thus, an
367: entangled state of three photons, whose quantum correlations
368: depend on the angles among the momenta of the outgoing three
369: photons. For the rest of the paper we will consider the first
370: family of states ($S_z=0$) although equivalent conclusions are
371: valid for the second one. In the next sections we will analyze
372: the entanglement properties of the states $|\psi_0(\hat k_1,\hat
373: k_2,\hat k_3)\ra$, using some of the quantum information
374: techniques and comparing them to the well-known cases of the
375: singlet and GHZ state.
376:
377: \section{Entanglement properties}
378: The quantum correlations of the three-photon entangled state
379: obtained from the o-Ps annihilation depend on the position of the
380: photon detectors, i.e. on the photon directions we are going to
381: measure. Our next aim will be to choose from the family of states
382: given by (\ref{finst}) the one that, in some sense, has the
383: maximum amount of GHZ-like correlations. In order to do this, we
384: first need to introduce some recent results on the study of
385: three-party entanglement.
386:
387: The set of states $|\psi_0(\hat
388: k_1,\hat k_2,\hat k_3)\ra$ form a six-parameter dependent family
389: in the Hilbert space $\sp\otimes\sp\otimes\sp$, so that each of
390: its components is equivalent to a state describing three
391: spin-$\frac{1}{2}$ particles or three qubits (a qubit, or quantum
392: bit, is the quantum version of the classical bit and corresponds
393: to a spin-$\frac{1}{2}$ particle). Two pure states belonging to a
394: generic composite system ${\mathcal H}_d^{\otimes N}$, i.e. $N$
395: parties each having a $d$-dimensional Hilbert space, are
396: equivalent as far as their entanglement properties go when they
397: can be transformed one into another by local unitary
398: transformations. This argument gives a lower bound for the
399: entanglement parameters a generic state $|\phi\ra\in {\mathcal
400: H}_2^{\otimes N}$ depends on. Since the number of real parameters
401: for describing it is $2^{N+1}$, and the action of an element of
402: the group of local unitary transformations $U(2)^{\otimes N}$ is
403: equivalent to the action of $U(1)\times SU(2)^{\otimes N}$, which
404: depends on $3N+1$ real parameters, the number of entanglement
405: parameters is bounded by $2^{N+1}-(3N+1)$. For our case this
406: counting of entanglement parameters gives six, since we have
407: $N=3$, and it can be proved that this is indeed the number of
408: nonlocal parameters describing a state in
409: $\sp\otimes\sp\otimes\sp$ \cite{Pop}.
410:
411: The above arguments imply that six independent quantities
412: invariant under the action of the group of local unitary
413: transformations will be enough, up to some discrete symmetry, to
414: describe the entanglement properties of any three-qubit pure
415: state. Given a generic state $|\phi\ra\in {\mathcal H}_2^{\otimes
416: 3}$
417: \be \label{genst} |\phi\ra=\sum_{i,j,k} t_{ijk}|ijk\ra\quad
418: i,j,k=1,2, \ee where $|i\ra,|j\ra,|k\ra$ are the elements of a
419: basis in each subsystem, A, B and C, the application of three
420: local unitary transformations $U^A$, $U^B$ and $U^C$ transforms
421: the coefficients $t_{ijk}$ into \be t_{ijk}'=\sum
422: U_{i\alpha}^AU_{j\beta}^BU_{k\gamma}^Ct_{\alpha\beta\gamma} . \ee
423: From this expression it is not difficult to build polynomial
424: combinations of the coefficient $t_{ijk}$ which are invariant
425: under local unitary transformations \cite{Pop,Sud}. These
426: quantities are good candidates for being an entanglement
427: parameter. For example, one of these invariants is \be \sum
428: t_{i_1j_1k_1}t_{i_1j_2k_2}^\ast
429: t_{i_2j_2k_2}t_{i_2j_1k_1}^\ast={\rm tr}(\rho_A^2), \ee where
430: $\rho_A={\rm tr}_{BC}(|\phi\ra\la\phi|)$ is the density matrix
431: describing the local quantum state of A (and the same happens for
432: B and C).
433: In \cite{Sud} the six linearly independent polynomial invariants of
434: minor degree were found (a trivial one is the norm) and a
435: slightly modified version of these quantities was also proposed
436: in \cite{nos}. In the rest of the paper we will not consider the
437: norm, so the space of entanglement parameters of the normalized
438: states belonging to $\sp\otimes\sp\otimes\sp$ has dimension equal
439: to five.
440:
441: A particularly relevant polynomial invariant is the so-called
442: tangle, $\tau$, introduced in \cite{CKW}. There is strong
443: evidence that somehow it is a measure of the amount of
444: ``GHZ-ness'' of a state \cite{nos,CKW,DVC,BC}. It corresponds to
445: the modulus of the hyperdeterminant of the hypermatrix given by
446: the coefficients $t_{ijk}$ \cite{GKZ}, which from (\ref{genst})
447: corresponds to
448: \be \label{tangle} \tau(|\phi\ra)=|{\rm
449: Hdet}(t_{ijk})|=\left|\sum \epsilon_{i_1i_2} \epsilon_{i_3i_4}
450: \epsilon_{j_1j_2} \epsilon_{j_3j_4} \epsilon_{k_1k_3}
451: \epsilon_{k_2k_4}t_{i_1j_1k_1}t_{i_2j_2k_2}t_{i_3j_3k_3}t_{i_4j_4k_4}\right|
452: ,
453: \ee
454: where $\epsilon_{00}=\epsilon_{11}=0$ and $\epsilon_{01}=-\epsilon_{10}=1$.
455: This quantity can be shown to be symmetric under
456: permutation of the indices $i,j,k$.
457:
458: Because of the interpretation
459: of the tangle as a measure of the GHZ-like correlations, we will
460: choose the position of the photon detectors, from the set of
461: states (\ref{finst}), the ones that are associated to a maximum
462: tangle. In the figure \ref{tangledet} it is shown the variation
463: of the tangle with the position of the detectors. It is not
464: difficult to see that the state of (\ref{finst}) with maximum
465: tangle corresponds to the case $\hat k_1\cdot \hat k_2=\hat
466: k_1\cdot \hat k_3=\hat k_2\cdot \hat k_3=-\frac{1}{2}$, i.e. the
467: most symmetric configuration, that we shall call ``Mercedes-star''
468: geometry. The normalized state obtained from (\ref{finst}) for
469: this geometry is
470: \be \label{state}
471: |\psi\ra=\frac{1}{\sqrt{6}}\left(|++-\ra+|--+\ra+|+-+\ra+|-+-\ra+|-++\ra+|+--\ra\right)
472: . \ee
473: Note that the GHZ state has tangle equal to $\frac{1}{4}$,
474: while the value of the tangle of (\ref{state}) is lower,
475: \be
476: \label{stangle}
477: \tau(|\psi\ra)=\frac{1}{12} .
478: \ee
479: It is arguable that a ``Mercedes-star'' geometry was naturally
480: expected to produce a maximum tangle state. Indeed, GHZ-like
481: quantum correlations do not singularize any particular qubit.
482:
483: Let us also mention that the
484: state we have singled out has some nice properties from the point
485: of view of group theory. It does correspond to the sum of two of
486: the elements of the coupled basis resulting from the tensor
487: product of three spin-$\frac{1}{2}$ particles,
488: $\frac{1}{2}\otimes\frac{1}{2}\otimes\frac{1}{2}$, \cite{Rai} \be
489: |\psi\ra=\frac{1}{\sqrt
490: 2}\left(|\frac{3}{2},+\frac{1}{2}\ra+|\frac{3}{2},-\frac{1}{2}\ra\right)
491: , \ee where \bea
492: &&|\frac{3}{2},+\frac{1}{2}\ra=\frac{1}{\sqrt 3}\left(|++-\ra+|+-+\ra+|-++\ra\right) \nonumber\\
493: &&|\frac{3}{2},-\frac{1}{2}\ra=\frac{1}{\sqrt 3}\left(|--+\ra+|-+-\ra+|+--\ra\right) .
494: \eea
495: The quantum correlations of (\ref{state}) will be now analyzed.
496:
497: \section{Useful decompositions}
498: In this section, the state (\ref{state}) will be rewritten in some
499: different forms that will help us to understand better its
500: nonlocal properties. First, let us mention that for any generic
501: three-qubit pure state and by performing change of local bases,
502: it is possible to make zero at least three of the coefficients
503: $t_{ijk}$ of (\ref{genst}) \cite{nos,HS}. A simple counting of
504: parameters shows that this is in fact the expected number of
505: zeros. This means that by a right choice of the local bases, any
506: state can be written with the minimum number of coefficients
507: $t_{ijk}$, i.e. we are left with all the non-local features of
508: the state, having removed all the ``superfluous'' information due
509: to local unitary tranformations. For the case of the state
510: (\ref{state}) it is easy to prove \cite{nos2} that it can be
511: expressed as
512: \be
513: \label{mindesc}
514: |\psi\ra=\frac{1}{2\sqrt{3}}\left(|001\ra+|010\ra+|100\ra\right)+\frac{\sqrt{3}}{2}|111\ra ,
515: \ee
516: which is the minimum decomposition in terms of product
517: states built from local bases (four of the coefficients $t_{ijk}$
518: are made equal to zero).
519:
520: An alternative decomposition, that will prove to be fruitful for
521: the rest of the paper, consists of writing the state as a sum of
522: two product states. This decomposition is somewhat reminiscent of
523: the form of the GHZ state, which is a sum of just two product
524: states, and is only possible when the tangle is different from
525: zero \cite{nos,DVC} as it happens for our state (see \ref{stangle}). The
526: state then can be written as
527: \bea
528: \label{prdesc} |\psi\ra && =\frac{2}{3}\left(\left(\matrix{1 \cr
529: 0}\right)\otimes\left(\matrix{1 \cr
530: 0}\right)\otimes\left(\matrix{1 \cr
531: 0}\right)+\left(\matrix{\frac{1}{2} \cr
532: \frac{\sqrt{3}}{2}}\right)\otimes\left(\matrix{\frac{1}{2} \cr
533: \frac{\sqrt{3}}{2}}\right)\otimes\left(\matrix{\frac{1}{2} \cr
534: \frac{\sqrt{3}}{2}}\right)\right)\nonumber\\
535: &&\equiv\alpha(|000\ra+|aaa\ra) , \eea
536: where $|0\ra\equiv\left(\matrix{1 \cr 0}\right)$ and
537: $a\equiv\left(\matrix{\frac{1}{2} \cr \frac{\sqrt{3}}{2}}\right)$.
538: We omit the details for the explicit computation of this
539: expression since they can be found in \cite{nos,DVC}. It is worth noticing
540: that o-Ps decay is hereby identified to belonging to an interesting type
541: of states already classified in quantum information theory \cite{DVC}.
542:
543: The above
544: decomposition allows for an alternative interpretation of the
545: initial state as an equally weighted sum of two symmetric product
546: states. Note that the Bloch vector, $\hat
547: n=(\sin\theta\cos\phi,\sin\theta\sin\phi,\cos\theta)$,
548: representing the first local spinor appearing in (\ref{prdesc})
549: is pointing to the $z$ axis, i.e. $\hat n_1=(0,0,1)$, while the
550: second is located in the $XZ$ plane with an angle of 120$^\circ$
551: with the $z$ axis, i.e.
552: $\hat{n}_2=(\frac{\sqrt{3}}{2},0,-\frac{1}{2})$. By performing a
553: new unitary transformation, (\ref{prdesc}) can be written as
554: \be
555: \label{ghzdesc} |\psi\ra=\frac{2}{3}\left(\left(\matrix{c \cr
556: s}\right)\otimes\left(\matrix{c \cr
557: s}\right)\otimes\left(\matrix{c \cr s}\right)+\left(\matrix{s \cr
558: c}\right)\otimes\left(\matrix{s \cr
559: c}\right)\otimes\left(\matrix{s \cr c}\right)\right) ,
560: \ee
561: where
562: $c=\cos15^\circ,s=\sin15^\circ$. Now, the two Bloch vectors are in
563: the $XZ$ plane, pointing to the $\theta=30^\circ$ and
564: $\theta=150^\circ$ directions. The GHZ state corresponds to the
565: particular case $c=1$ and $s=0$.
566:
567: \section{Quantum mechanics vs local realism}
568: The quantum correlations present in some three-qubit pure states
569: show, as it was mentioned in the introduction, a much stronger
570: disagreement with the predictions of a local-realistic model than
571: any two-qubit entangled state. In fact, contrary to the case of
572: the singlet state, no LR model is able to reproduce all the
573: perfect correlations predicted for the maximally entangled state
574: of three qubits \cite{GHZ}. The state (\ref{state}) emerging from
575: o-Ps decay is not a GHZ state, although it has been chosen the
576: one with the maximum tangle in order to maximize GHZ-like
577: correlations. In this section we will show how to use it for
578: testing quantum mechanics against local-realistic models, and
579: then we will compare its performance against existing tests for
580: the maximally entangled states of two and three spin-$\frac{1}{2}$
581: particles. We start reviewing some of the consequences derived
582: from the arguments proposed in \cite{EPR}.
583:
584: \subsection{QM vs LR conflict}
585: Given a generic quantum state of a composite system shared by $N$
586: parties, there should be an alternative LR theory which
587: reproduces all its statistical predictions. In this LR model, a
588: state denoted by $\lambda$ will be assigned to the system
589: specifying all its elements of physical reality. In particular,
590: the result of a measurement depending on a set of parameters
591: $\{n\}$ performed locally by one of the parties, say A, will be
592: specified by a function $a_\lambda(\{n\})$. The same will happen
593: for each of the space-separated parties and, since there is no
594: causal influence among them, the result measured on A can not
595: modify the measurement on B. For example, if the measurement is
596: of the Stern-Gerlach type, the parameters labeling the
597: measurement are given by a normalized vector $\hat n$ and
598: $a_\lambda(\hat n)\equiv a$ are the LR functions describing the
599: outcome.
600:
601: The LR model can be very general provided that some conditions
602: must be satisfied. Consider a generic pure state belonging to
603: $\sp\otimes\sp\otimes\sp$ shared by three observers A, B and C,
604: which are able to perform Stern-Gerlach measurements in any
605: direction. Since the outcomes of a Stern-Gerlach measurement are
606: only $\pm1$, it is easy to check that for any pair of
607: measurements on each subsystem, described by the LR functions $a$
608: and $a'$, $b$ and $b'$, $c$ and $c'$, and for all their possible
609: values, it is always verified
610: \be \label{LRconstr}
611: a'bc+ab'c+abc'-a'b'c'=\pm 2 . \ee
612: It follows from this relation
613: that
614: \be \label{mermineq} -2\leq\la a'bc+ab'c+abc'-a'b'c' \ra\leq
615: 2. \ee
616: This constraint is known as Mermin inequality
617: \cite{Mermin} and has to be satisfied by any LR model describing
618: three space-separated systems.
619:
620: Let us now take the GHZ state (\ref{GHZst}). It is quite simple to
621: see that if the observables $a$ and $a'$ are equal to $\sigma_y$
622: and $\sigma_x$ (the same for parties B and C), the value of
623: (\ref{mermineq}) is $-4$, so an experimental condition is found
624: that allows to test quantum mechanics against local realism. Note
625: that this is the maximal violation of inequality
626: (\ref{mermineq}). Moreover, the GHZ state also satisfies that
627: $a'bc=ab'c=abc'=-a'b'c'=-1$ and no LR model is able to take into
628: account this perfect correlation result because of
629: (\ref{LRconstr}) \cite{GHZ}. This is a new feature that does not
630: appear for the case of a two maximally entangled state of two
631: spin-$\frac{1}{2}$ particles. In this sense it is often said that
632: a most dramatic contrast between QM and LR emerges for
633: entanglement between three subsystems.
634:
635: Let us go back to the state given by the ortho-positronium decay
636: (\ref{state}). Our aim is to design an experimental situation
637: where a conflict between QM and LR appears, so we will look for
638: the observables that give a maximal violation of (\ref{mermineq}).
639: Such observables will extremize that expression.
640: Using the decomposition (\ref{ghzdesc}), the expectation value of
641: three local observables is \bea \label{expval}
642: \la abc \ra&=&\la\psi|(\hat n_a\cdot\vec\sigma)\otimes(\hat n_b\cdot\vec\sigma)\otimes(\hat n_c\cdot\vec\sigma)|\psi\ra \nonumber\\
643: &=&\frac{4}{9}\left(\prod_{i=a,b,c}(\tilde c\cos\theta_i+\tilde s\sin\theta_i
644: \cos\phi_i)+\prod_{i=a,b,c}(-\tilde c\cos\theta_i+\tilde s\sin\theta_i\cos
645: \phi_i)\right. \nonumber\\
646: &+&\left.\prod_{i=a,b,c}\sin\theta_i(c^2e^{-i\phi_i}+s^2e^{i\phi_i})+\prod_{i=a,b,c}\sin\theta_i(c^2e^{i\phi_i}+s^2e^{-i\phi_i})\right)
647: , \eea
648: where $\tilde c\equiv c^2-s^2$ and $\tilde s\equiv 2sc$.
649: Because of the
650: symmetry of the state under permutation of parties, the
651: Stern-Gerlach directions are taken satisfying $\hat n_a=\hat
652: n_b=\hat n_c=(\sin\theta\cos\phi,\sin\theta\sin\phi,\cos\theta)$
653: and $\hat n_{a'}=\hat n_{b'}=\hat
654: n_{c'}=(\sin\theta'\cos\phi',\sin\theta'\sin\phi',\cos\theta')$.
655: Substituting this expression in (\ref{mermineq}), we
656: get the explicit function $f(\theta,\phi,\theta',\phi')$ to be
657: extremized. For the case of the GHZ state described above, the
658: extreme values were obtained using two observables with
659: $\theta=\theta'=\frac{\pi}{2}$, i.e. in the $XY$ plane. Since
660: (\ref{ghzdesc}) is the GHZ-like decomposition of the initial
661: state, we take $\theta=\theta'=\frac{\pi}{2}$ and it is easy to
662: check that in this case $\frac{\partial
663: f}{\partial\theta}\big\vert_{\theta=\theta'=\frac{\pi}{2}}=\frac{\partial
664: f}{\partial\theta'}\big\vert_{\theta=\theta'=\frac{\pi}{2}}=0,\,\forall\phi,\phi'$.
665: Mantaining the parallelism with the GHZ case, it can be seen that
666: all the partial derivatives vanish when it is also imposed
667: $\phi=\frac{\pi}{2}$ and $\phi'=0$. In our case the calculation of
668: (\ref{mermineq}) gives $-3$, so a conflict between
669: local-realistic models and quantum mechanics again appears, and
670: then the three-photon state coming from the ortho-positronium
671: decay can be used, in principle, to test QM vs LR with the set of
672: observables given by the normalized vectors \be \label{test} \hat
673: n_a=\hat n_b=\hat n_c=(0,1,0)\qquad\hat n_{a'}=\hat n_{b'}=\hat
674: n_{c'}=(1,0,0) . \ee
675: There is an alternative
676: set of angles $\phi$ and $\phi'$ that makes zero all the partial
677: derivatives of $f$: the combination of local observables
678: (\ref{mermineq}) is equal to $\approx-3.046$ for \be
679: \label{minalt}
680: \phi'=\arctan\left(-\frac{\sqrt{17+27\sqrt{41}}}{10}\right)\approx
681: 126^\circ\quad\phi=\frac{1}{2}\arctan\left(\frac{2\sqrt{17+27\sqrt{41}}}{25}\right)\approx
682: 24^\circ. \ee This second set of parameters will be seen to
683: produce in the end a weaker dismissal of LR.
684:
685: Our next step will be to carry over the comparison of this QM vs
686: LR test against the existent ones for the maximally entangled
687: states of three and two spin-$\frac{1}{2}$ particles, i.e. the GHZ
688: and singlet state. It is quite evident that the described test
689: should be worse than the obtained for the GHZ state. It is less
690: obvious how this new situation will compare with the singlet case.
691:
692:
693: \subsection{Comparison with the maximally entangled
694: states of two and three spin-$\frac{1}{2}$ particles}
695:
696: We will now estimate the ``strength'' of the QM vs LR test
697: proposed above, being this ``strength'' measured by the number of
698: trials needed to rule out local-realism at a given confidence
699: level, as Peres did in \cite{Peres}. A reasoning anologous to the
700: one given in \cite{Peres} will be done here for the state
701: (\ref{state}) and the observables (\ref{test}).
702:
703: Imagine
704: a local-realistic physicist who does not believe in quantum
705: mechanics. He assigns prior subjective probabilities to the
706: validity of LR and QM, $p_r$ and $p_q$, expressing his personal
707: belief. Take for instance $\frac{p_r}{p_q}=100$. His LR theory is
708: not able to reproduce exactly all the QM statistical results of
709: some quantum states. Consider the expectation value of some
710: observable ${\mathcal O}$ with two outcomes $\pm 1$ such that
711: $\la{\mathcal O}\ra=E_q$ is predicted for some quantum state,
712: while LR gives $\la{\mathcal O}\ra=E_r\neq E_q$. Since the value
713: of the two possible outcomes are $\pm1$, the probablity of having
714: ${\mathcal O}=+1$ is $q=\frac{1+E_q}{2}$ for QM and
715: $r=\frac{1+E_r}{2}$ for LR. An experimental test of the
716: observable ${\mathcal O}$ now is performed $n$ times yielding $m$
717: times the result $+1$. The prior probabilities $p_q$ and $p_r$
718: are modified according to the Bayes theorem and their ratio has
719: changed to \be
720: \frac{p_r'}{p_q'}=\frac{p_r}{p_q}\frac{p(m|_{LR})}{p(m|_{QM})} ,
721: \ee where \be p(m|_{LR})=\left(\matrix{n \cr
722: m}\right)r^m(1-r)^{n-m} , \ee is the LR probability of having $m$
723: times the outcome $+1$, and we have the same for $p(m|_{QM})$,
724: being $r$ replaced by $q$. Following Peres \cite{Peres}, the {\sl
725: confidence depressing factor} is defined \be
726: D\equiv\frac{p(m|_{QM})}{p(m|_{LR})}=\left(\frac{q}{r}\right)^m\left(\frac{1-q}{1-r}\right)^{n-m}
727: , \ee which accounts for the change in the ratio of the
728: probabilities of the two theories, i.e. it reflects how the LR
729: belief changes with the experimental results. Like in a game, our
730: aim is to destroy as fast as we can the LR faith of our friend by choosing an
731: adequate experimental situation. It can be said, for example,
732: that he will give up when, for example, $D=10^{4}$.
733: Since the world is quantum, $m=qn$, and the number of
734: experimental tests needed to obtain $D=10^{4}$ is equal to \be
735: \label{ntests}
736: n_D(q,r)\equiv\frac{4}{q\log_{10}\left(\frac{q}{r}\right)+(1-q)\log_{10}\left(\frac{1-q}{1-r}\right)}=\frac{4}{K(q,r)}
737: , \ee being $K(q,r)$ the information distance \cite{KU} between
738: the QM and LR binomial distribution for the outcome $+1$. The
739: more separate the two probability distributions are, measured in
740: terms of the information distance, the fewer the number of
741: experiments $n_D$ is.
742:
743: Let us come back to the three-party entangled state
744: coming from the ortho-positronium decay (\ref{state}) under the
745: local measurements described by (\ref{test}). As it has been
746: shown above, a contradiction with any LR model appears for the
747: combination of the observables given by the Mermin inequality. In
748: our case quantum mechanics gives the following predictions
749: \be
750: \label{qmres} \la a'bc\ra=\la ab'c\ra=\la abc'\ra=-\frac{2}{3}
751: \qquad \la a'b'c'\ra=+1 , \ee
752: and this implies that
753: $q_1=prob(a'bc=+1)=prob(ab'c=+1)=prob(abc'=+1)=\frac{1}{6}$ and
754: $q_2=prob(a'b'c'=+1)=1$. This is the QM data that our LR friend
755: has to reproduce as well as possible. Because of the symmetry of
756: the state he will assign the same probability $r_1$ to the events
757: $a'bc=+1$, $ab'c=+1$ and $abc'=+1$ and $r_2$ to $a'b'c'=+1$.
758: However, his model has to satisfy the constraint given by
759: (\ref{mermineq}), so the best he can do is to saturate the bound
760: and then
761: \be
762: \label{constraint}
763: 3r_1=r_2\,\Longrightarrow\, 0\leq r_1\leq\frac{1}{3} .
764: \ee
765: Now, according to the probabilities $r_1$ and $r_2$ his LR
766: model predicts, we choose the experimental test that minimizes
767: (\ref{ntests}), i.e. we consider the event $a'bc=+1$
768: ($a'b'c'=+1$) when $n_D(q_1,r_1)<n_D(q_2,r_2)$
769: ($n_D(q_1,r_1)>n_D(q_2,r_2)$), and the experimental results will
770: destroy his LR belief after $n_D(q_1,r_1)$ ($n_D(q_2,r_2)$)
771: trials. The best value our LR friend can assign to $r_1$ is the
772: solution to
773: \be
774: \label{bestLR}
775: n_D(q_1,r_1)=n_D(q_2,r_2) ,
776: \ee
777: with the constraint (\ref{constraint}), and this condition means that
778: $r_1\approx0.315$ and $n_D\approx 161$ trials are needed
779: to have a depressing factor equal to $10^{4}$. Repeating the
780: same calculation for the observables giving by (\ref{minalt}),
781: the number of trials slightly increases, $n_D\approx 166$,
782: despite of the fact that the violation of the inequality is
783: greater than the obtained for (\ref{test}).
784:
785: In ref. \cite{Peres} the
786: same reasoning was applied to the maximally entangled state of
787: two and three spin-$\frac{1}{2}$ particles, showing that
788: $n_D\approx 200$ in the first case, and $n_D\approx 32$ for the
789: latter (see table \ref{summary}). Our result then implies that
790: the three-photon entangled state produced in the
791: ortho-positronium decay has, in some sense, more quantum
792: correlations than any entangled state of two spin-$\frac{1}{2}$
793: particles.
794:
795:
796: \subsection{Generalization of the results}
797: It is easy to generalize some of the results obtained for the
798: entangled state resulting from the o-Ps decay. As it has been
799: mentioned, this state can be understood as an equally weighted
800: sum of two symmetric product states, since it can be written as
801: (\ref{ghzdesc}). The Bloch vectors of the two local states
802: appearing in this decomposition form an angle of $120^\circ$. It
803: is clear that the conclusions seen above depend on the angle
804: between these vectors, i.e. with their degree of non-orthogonality.
805: The family of states to be analyzed can be parametrized in the
806: following way \be \label{famangle}
807: |\psi(\delta)\ra=\alpha_\delta\left(\left(\matrix{c_\delta \cr
808: s_\delta}\right)\otimes\left(\matrix{c_\delta \cr
809: s_\delta}\right)\otimes\left(\matrix{c_\delta \cr
810: s_\delta}\right)+\left(\matrix{s_\delta \cr
811: c_\delta}\right)\otimes\left(\matrix{s_\delta \cr
812: c_\delta}\right)\otimes\left(\matrix{s_\delta \cr
813: c_\delta}\right)\right) , \ee where $\delta$ is the angle between
814: the two local Bloch vectors,
815: $c_\delta\equiv\cos\left(\frac{\pi-\delta}{4}\right)$ and
816: $s_\delta\equiv\sin\left(\frac{\pi-\delta}{4}\right)$ and
817: $\alpha_\delta$ is a positive number given by the normalization of the state.
818: An alternative parametrization of this family is, using (\ref{mindesc}) and
819: defining $\delta'\equiv\frac{\delta}{4}$,
820: \be
821: \label{famangle2}
822: |\psi(\delta)\ra=2\alpha_\delta\left(\sin^2\delta'\cos\delta'
823: \left(|001\ra+|010\ra+|100\ra\right)+
824: \cos^3\delta'|111\ra\right) .
825: \ee
826:
827: The expectation value of three local observables for this set of
828: states follows trivially from (\ref{expval}). Using this
829: expression it is easy to see that the combination of the
830: expectation values of (\ref{mermineq}) has all the partial
831: derivatives equal to zero for the set of observables given in
832: (\ref{test}) independently of $\delta$. For these observables,
833: the dependence of expression (\ref{mermineq}) with the degree of
834: orthogonality between the two product states is given in figure
835: \ref{mermangle}. There is no violation of the Mermin inequality
836: for the case in which $\delta\lesssim 85^\circ$. In this
837: situation one can always found a LR model able to reproduce the
838: QM statistical prediction given by (\ref{mermineq}) and the
839: observables (\ref{test}). We can now repeat all the steps made in
840: order to determine the number of trials needed to rule out local
841: realism as a function of the angle $\delta$. In figure
842: \ref{ntrials} we have summaryzed the results. We have shown only
843: the cases where the number of trials is minor than two hundred,
844: since this is the value obtained for the singlet. Note that the
845: case $\delta=120^\circ$, which corresponds to (\ref{state}), is
846: very close to the region where there is no improvement compared
847: to the maximally entangled state of two qubits.
848:
849: All these results can be understood in the following way: the
850: smaller the angle between the two local states, $\delta$, the
851: higher the overlap of the state $|\psi(\delta)\ra$ with the
852: product state having each local Bloch vector pointing in the
853: direction of the $x$ axis, which corresponds to the state $|111\ra$ in
854: (\ref{famangle2}). This means that the quantum state we
855: are handling is too close to a product state \cite{HS}, and thus, no
856: violation of the Mermin inequality can be observed.
857:
858:
859: \section{Concluding remarks}
860: In this work we have analyzed the three-particle quantum
861: correlations of a physical system given by the decay of the
862: ortho-positronium into a three-photon pure state. After obtaining
863: the state describing the polarization of the three photons
864: (\ref{state}), some of the recent techniques developed for the
865: study of three-party entanglement have been applied. The
866: particular case where the three photons emerge in a symmetric,
867: Mercedes-star-like configuration, corresponds to the state with
868: the maximum tangle. We have shown that this state allows a priori
869: for a QM vs LR test which is stronger than any of the existing
870: ones that use the singlet state. In this sense, ortho-positronium
871: decays into a state which carries stronger quantum correlations
872: than any entangled state of two spin-$\frac{1}{2}$ particles.
873:
874: Bose symmetrization has played a somewhat negative role in
875: reducing the amount GHZ-ness of the o-Ps decay state. Indeed, the
876: natural GHZ combination $\vert ++-\rangle +\vert --+\rangle$
877: emerging from the computation of Feynmann diagrams has been
878: symmetrized due to the absence of photon tagging to our state
879: $\vert ++-\rangle +\vert +-+\rangle+\vert -++\rangle +\vert
880: --+\rangle+\vert -+-\rangle +\vert +--\rangle$, inducing a loss
881: of tangle. The quantum optics realization of the GHZ state does
882: avoid symmetrization through a geometric tagging \cite{BPDWZ}. It is, thus,
883: reasonable to look for pure GHZ states in decays to distinct
884: particles, so that tagging would be carried by other quantum
885: numbers, as e.g. charge. It is, on the other hand, peculiar to
886: note that symmetrization in the $K^0\bar K^0$ system is
887: responsible for its entanglement ($\vert +-\rangle +\vert
888: -+\rangle$) \cite{kaon}.
889:
890: Finally, let us briefly discuss the experimental requirements
891: needed for testing quantum mechanics as it has been described in
892: this paper. In order to do this, the circular polarizations of the
893: three photons resulting from an ortho-positronium decay have to be
894: measured. The positions of the three detectors are given by the
895: ``Mercedes-star'' geometry and their clicks have to detect the
896: coincidence of the three photons. The energy of these photons is
897: of the order of 1 Mev. Polarization analyzers with a good
898: efficiency would allow us to acquire statistical data showing
899: quantum correlations which would violate the Mermin inequality
900: discussed above. Unfortunately, as far as we know, no such
901: analizers exist for this range of energies. A possible way-out
902: might be to use Compton scattering to measure the photon
903: polarizations \cite{AHSHZ}. However, Compton effect just gives a
904: statistical pattern depending on the photon and electron
905: polarizations which is not a direct measurement of the
906: polarizations. Further work is needed to modify our analysis of
907: QM vs LR to accommodate for such indirect measurements.
908:
909: \bigskip
910: \section*{Acknowledgments}
911: We acknowledge J. Bernabeu for suggesting positronium as a source
912: of three entangled particles and reading carefully the paper. We
913: also thank A. Czarnecki, D. W. Gidley, M. A. Skalsey and V. L.
914: Telegdi for comments about the measurement of the photon
915: polarizations in the ortho-positronium decay. We acknowledge
916: financial support by CICYT project AEN 98-0431, CIRIT project
917: 1998SGR-00026 and CEC project IST-1999-11053, A. A. by a grant
918: from MEC (AP98). Financial support from the ESF is also
919: acknowledged. This work was concluded during the 2000 session of the Benasque
920: Center for Science, Spain.
921:
922: \begin{references}
923: \bibitem{EPR}
924: A. Einstein, B. Podolsky and N. Rosen, Phys. Rev. {\bf 47} (1935), 777.
925: \bibitem{GHZ}
926: D. M. Greenberger, M. A. Horne, A. Shimony and A. Zeilinger, Am. J. Phys. {\bf 58} (1990), 1131.
927: \bibitem{Bell}
928: J. S. Bell, Physics {\bf 1} (1964), 195.
929: \bibitem{BA}
930: D. Bohm and Y. Aharonov, Phys. Rev. {\bf 108} (1957), 1070.
931: \bibitem{CS}
932: J. F. Clauser and A. Shimony, Rep. Prog. Phys. {\bf 41} (1978), 1881.
933: \bibitem{ADR}
934: A. Aspect, J. Dalibard and G. Roger, Phys. Rev. Lett. {\bf 49} (1982), 1804.
935: \bibitem{exptest}
936: W. Tittel, J. Brendel, H. Zbinden and N. Gisin, Phys. Rev. Lett. {\bf 81} (1998), 3563; G. Weihs, T. Jennewein, C. Simon, H. Weinfurter and A. Zeilinger, Phys. Rev. Lett. {\bf 81} (1998), 5039.
937: \bibitem{GHZor}
938: D. M. Greenberger, M. A. Horne and A. Zeilinger, ``Going beyond Bell's theorem'', in {\sl Bell's Theorem, Quantum Theory and Conceptions of the Universe}, edited by M. Kafatos (Kluwer Academic, Dordrecht, The Netherlands, 1989), pp. 73-76.
939: \bibitem{Mer}
940: N. D. Mermin, Am. J. Phys. {\bf 58} (1990), 731.
941: \bibitem{BPDWZ}
942: D. Bouwmeester, J. Pan, M. Daniell, H. Weinfurter and A. Zeilinger, Phys. Rev. Lett. {\bf 82} (1999), 1345, quant-ph/9810035.
943: \bibitem{BBCJPW}
944: C. H. Bennett, G. Brassard, C. Cr\'epeau, R. Josza, A. Peres and W. K. Wootters, Phys. Rev. Lett. {\bf 70} (1993), 1895.
945: \bibitem{CHSH}
946: J. F. Clauser, M. A. Horne, A. Shimony and R. A. Holt, Phys. Rev. Lett. {\bf 23} (1969), 880.
947: \bibitem{kaon}
948: See for instance A. Di Domenico, Nucl. Phys. B {\bf 450} (1995), 293; B. Ancochea, A. Bramon and M. Nowakowski, Phys.Rev. D {\bf 60} (1999), 094008, hep-ph/9811404; F. Benatti and R. Floreanini, Eur. Phys. J. C {\bf 13} (2000), 267, hep-ph/9912348.
949: \bibitem{IZ}
950: C. Itzykson and J. Zuber, {\sl Quantum field theory}, McGraw-Hill.
951: \bibitem{WR}
952: L. Wolfenstein and D. G. Ravenhall, Phys. Rev. {\bf 88} (1952), 279.
953: \bibitem{Czarn}
954: Andrzej Czarnecki, Acta Phys. Polon. B {\bf30} (1999) 3837, hep-ph/9911455.
955: \bibitem{Pop}
956: N. Linden and S. Popescu, Fortsch. Phys. {\bf 46} (1998), 567,
957: quant-ph/9711016.
958: \bibitem{Sud}
959: A. Sudbery, ``On local invariants of pure three-qubit states'',
960: quant-ph/0001116.
961: \bibitem{nos}
962: A. Ac\' \i n, A. Andrianov, L. Costa, E. Jan\'e, J.I. Latorre and R. Tarrach,``Schmidt decomposition and classification of three-quantum-bit states'', quant-ph/0003050, to appear in Phys. Rev. Lett..
963: \bibitem{CKW}
964: V. Coffman, J. Kundu, W. K. Wootters, Phys. Rev. A {\bf 61} (2000), 052306,
965: quant-ph/9907047.
966: \bibitem{DVC}
967: W. D\"ur, G. Vidal and J. I. Cirac, ``Three qubits can be entangled in two inequivalent ways'', quant-ph/0005115.
968: \bibitem{BC}
969: T. A. Brun and O. Cohen, ``Parametrization and distillability of three-qubit entanglement'', quant-ph/0005124.
970: \bibitem{GKZ}
971: I. M. Gelfand, M. M. Kapranov and A. V. Zelevinsky, ``Discriminants,
972: resultants and multidimensional determinants'', Birkh\"auser Boston 1994.
973: Its explicit form is: ${\rm Hdet}(t_{ijk})=t_{000}^2t_{111}^2+t_{001}^2t_{110}^2+t_{010}^2t_{101}^2+t_{100}^2t_{011}^2$ $-2(t_{000}t_{111}t_{011}t_{100}+t_{000}t_{111}t_{101}t_{010}+t_{000}t_{111}t_{110}t_{001}$ $+t_{011}t_{100}t_{101}t_{010}+t_{011}t_{100}t_{110}t_{001}+t_{101}t_{010}t_{110}t_{001})$ $+4(t_{000}t_{110}t_{101}t_{011}+t_{111}t_{001}t_{010}t_{100})$.
974: \bibitem{Rai}
975: S. Rai and J. Rai, ``Group-theoretical Structure of the Entangled States of N Identical Particles'', quant-ph/0006107.
976: \bibitem{HS}
977: A. Higuchi and A. Sudbery, ``How entangled can two couples get?'', quant-ph/0005013; H. A. Carteret, A. Higuchi and A. Sudbery, ``Multipartite generalisation of the Schmidt decomposition'', quant-ph/0006125.
978: \bibitem{nos2}
979: A. Ac\'\i n, A. Andrianov, E. Jan\'e, J. I. Latorre and R. Tarrach, in preparation.
980: \bibitem{Mermin}
981: N. D. Mermin, Phys. Rev. Lett. {\bf 65} (1990), 1838.
982: \bibitem{Peres}
983: A. Peres, ``Bayesian analysis of Bell inequalities'', quant-ph/9905084.
984: \bibitem{KU} S. Kullback, {\sl Information theory and statistics}, Wiley, New York (1959).
985: \bibitem{AHSHZ}
986: B. K. Arbic, S. Hatamian, M. Skalsey, J. Van House and W. Zheng, Phys. Rev. A {\bf 37} (1988), 3189.
987:
988: \end{references}
989: \newpage
990:
991: \begin{figure}
992: \begin{center}
993: \epsffile{fig.eps}
994: \medskip
995: \caption{ Variation of the tangle with the position of the photon
996: detectors, that are represented by two angles (the third one has
997: to sum up to 360$^\circ$). We have taken $\tau=0$ when the
998: position of the detectors, i.e. the photon trajectories, are
999: incompatible with momentum conservation.} \label{tangledet}
1000: \end{center}
1001: \end{figure}
1002: \begin{figure}
1003: \begin{center}
1004: \epsffile{fig2.eps}
1005: \medskip
1006: \caption{ Violation of the Mermin inequality (\ref{mermineq})
1007: with the angle $\delta$ for the family of states
1008: (\ref{famangle}). We have substracted 2 to the combination of the
1009: expected values of (\ref{mermineq}), so a positive value means
1010: that a conflict between QM and LR appears.} \label{mermangle}
1011: \end{center}
1012: \end{figure}
1013: \begin{figure}
1014: \begin{center}
1015: \epsffile{fig3.eps}
1016: \medskip
1017: \caption{ Number of trials needed to rule out local realism as a
1018: function of the angle $\delta$ for the family of states
1019: (\ref{famangle}). Values greater than two hundred are not shown
1020: since in these cases there always exists a two-qubit entangled
1021: state which gives the same result, i.e. it has the same
1022: ``strength'' for ruling out local realism.} \label{ntrials}
1023: \end{center}
1024: \end{figure}
1025: \begin{table}
1026: \begin{tabular}{||c|lc||}
1027: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ State ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ &
1028: Number of trials & \\
1029: \hline
1030: GHZ & $\approx 32$ & \\
1031: Positronium state (\ref{state}) & $\approx 161$ & \\
1032: Singlet & $\approx 200$ & \\
1033: \hline
1034: \end{tabular}
1035: \medskip
1036: \caption{Comparison of the strength of the QM vs LR test which
1037: can be performed for the maximally entangled states of two and
1038: three spin-$\frac{1}{2}$ particles and for the three-photon
1039: entangled state reulting from the ortho-positronium annihilation.}
1040: \label{summary}
1041: \end{table}
1042:
1043:
1044: \end{document}
1045: