1: \documentclass[aps,prl,a4paper,twocolumn,showpacs,superscriptaddress,floatfix]{revtex4}
2: \newcommand{\media}[1]{\langle #1 \rangle}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: \usepackage{graphicx,amsmath,bbm,mathrsfs,amssymb,pstricks,times,psfrag}
5: \begin{document}
6: \title{Full characterization of Gaussian bipartite entangled states by a
7: single homodyne detector}
8: %
9: \author{V.~D'Auria}
10: \affiliation{CRS Coherentia CNR-INFM, Napoli, Italia.}
11: \author{S.~Fornaro}
12: \affiliation{Dipartimento di Scienze Fisiche Universit\`a ``Federico II'',
13: Napoli, Italia.}
14: \author{A.~Porzio}
15: \affiliation{CRS Coherentia CNR-INFM, Napoli, Italia.}
16: \affiliation{CNISM UdR Napoli Universit\`a, Napoli, Italia.}
17: \author{S.~Solimeno}
18: \affiliation{Dipartimento di Scienze Fisiche Universit\`a ``Federico II'',
19: Napoli, Italia.}
20: \affiliation{CNISM UdR Napoli Universit\`a, Napoli, Italia.}
21: \author{S.~Olivares}
22: \affiliation{CNISM UdR Milano Universit\`a, Milano, Italia.}
23: \affiliation{Dipartimento di Fisica dell'Universit\`a di Milano,
24: Milano, Italia.}
25: \author{M.~G.~A.~Paris}
26: \affiliation{Dipartimento di Fisica dell'Universit\`a di Milano,
27: Milano, Italia.}
28: \affiliation{CNISM UdR Milano Universit\`a, Milano, Italia.}
29: \affiliation{ISI Foundation, Torino, Italia.}
30: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
31: \begin{abstract}
32: We present the full experimental reconstruction of Gaussian entangled
33: states generated by a type--II optical parametric oscillator (OPO) below
34: threshold. Our scheme provides the entire covariance matrix using a
35: single homodyne detector and allows for the complete characterization of
36: bipartite Gaussian states, including the evaluation of purity,
37: entanglement and nonclassical photon correlations, without a priori
38: assumptions on the state under investigation. Our results show that
39: single homodyne schemes are convenient and robust setups for the full
40: characterization of OPO signals and represent a tool for quantum
41: technology based on continuous variable entanglement.
42: \end{abstract}
43: \date{\today}
44: \pacs{03.67.Mn, 03.65.Wj, 42.65.Yj}
45: \maketitle
46: %%%%%%%%%%%%
47: \indent {\em Introduction}---In this letter we address the complete
48: experimental characterization of bipartite Gaussian entangled states.
49: In our experiment continuous-wave (CW) entangled light beams are
50: generated by a single type--II optical parametric oscillator (OPO) below
51: threshold, and then their covariance matrix (CM) is fully reconstructed
52: using a novel scheme \cite{sh05} that involves a single homodyne detector
53: \cite{ray99}.
54: To our knowledge this is the first {\em complete} characterization of
55: OPO signals without a priori assumptions and paves the way to a deeper
56: investigation of continuous variable entanglement without experimental
57: loopholes.
58: \par
59: Light beams endowed with nonclassical correlations \cite{eis03} are
60: crucial resources for quantum technology and find applications in
61: quantum communication \cite{vnl02}, imaging \cite{lug02} and
62: precision measurement \cite{dar01,dau05}. Their full characterization has
63: a fundamental interest in its own and represents a tool for the design of
64: quantum information processing protocols in realistic conditions.
65: Remarkably, entangled states produced by OPOs are Gaussian states
66: \cite{simXX,marYY} and thus may be fully characterized by the first two
67: statistical moments of the field modes. In turn, the CM contains the
68: complete information about entanglement \cite{sim00,dua00}, {\em i.e.}
69: about their performances as a resource for quantum technology.
70: \par
71: Bipartite entangled states may be generated by mixing at a beam splitter
72: (BS) two squeezed beams obtained by a degenerate OPO below threshold
73: \cite{bow03}. The beams exiting the BS are entangled \cite{theZZ} and a
74: partial reconstruction of the corresponding CM has been obtained
75: % by measuring the noise relative to the sum/difference variances of the modes
76: \cite{bow04}. The complete reconstruction of a CM has been obtained for
77: different entangled states, by varying single-mode squeezing
78: \cite{dig07}. In this configuration two OPOs are used and the
79: amount of entanglement critically depends on the symmetry between the two
80: squeezed beams. From the experimental point of view this requires an
81: accurate setting on the two squeezers and a strict control on the
82: relative phase.
83: The measured CM presents some unexpected deviations from a proper
84: form so leaving a question open on the reliability of double homodyne
85: schemes due to technical difficulties \cite{bow04,lau05}.
86: A more direct way to generate quadrature entanglement
87: is to use a single non-degenerate OPO \cite{dru90} which represents a
88: robust and reliable source of EPR-type correlation either below
89: \cite{zha99,kim92} or above threshold \cite{jin06,bre07,kel08}.
90: Partial reconstructions of the CM in the pulsed regime has been achieved for
91: the spectrally degenerate but spatially non-degenerate twin beams at the
92: output of a type--I parametric amplifier \cite{wen04}, whereas in the CW
93: regime cross polarized beams emitted by a self-locked type--II OPO have
94: been examined at frequency degeneracy \cite{lau05}.
95: Although the correlation properties of OPO signals have been widely
96: investigated, no proper CM reconstruction has been performed so
97: far. In turn, in previous proposals and experiments non-physical CM
98: entries \cite{bow03,bow04,wen04,lau05} or deviations from a proper CM
99: \cite{dig07} appeared, thus requiring a priori hypothesis on the measured
100: state to understand the experimental results.
101: \par
102: In this letter we report the first complete measurement of the CM for
103: the output of a single non-degenerate OPO. The two entangled beams are
104: emitted with orthogonal polarization and degenerate frequency by a CW
105: type--II OPO below threshold. In order to reconstruct the ten
106: independent elements of the CM, the beams are optically combined into
107: six auxiliary modes, whose quadratures are measured using a single
108: homodyne detector \cite{sh05}. The first two moments of the relevant
109: quadratures are obtained by tomographic reconstruction using the whole
110: homodyne data set, the CM is, then, fully reconstructed after assessing
111: the Gaussian character of the signal and compared with a general model
112: describing a realistic OPO. Entanglement is demonstrated using the
113: partial transpose method \cite{sim00}, the Duan inequality \cite{dua00}
114: and the stricter EPR criterion \cite{tre05}, and quantified upon
115: evaluating the logarithmic negativity and the entanglement of formation
116: (EoF). We also reconstruct the joint
117: photon number distribution and demonstrate nonclassical photon
118: correlations by evaluating the noise reduction factor.
119: %Our experimental
120: %setup makes use of a single OPO and a single homodyne detector and thus
121: %represent a compact and robust setup for entanglement generation and
122: %characterization.
123: In the following, after defining notation and a
124: brief summary of the reconstruction method, we describe in details the
125: apparatus and the experimental results.
126: %%%%%%%%%%%%%%%%%%
127: \par{\em Reconstruction method---}Upon introducing the vector
128: $\boldsymbol{R}=(x_{1},y_{1},x_{2},y_{2})$ of canonical operators, in
129: terms of the mode operators $a_{k}$, $x_{k}=(a_{k}^{\dag
130: }+a_{k})/\sqrt{2}$, $y_{k}=i(a_{k}^{\dag }-a_{k})/ \sqrt{2}$, $k=1,2$,
131: the CM $\boldsymbol{\sigma}$ of a bipartite state $\varrho$ is defined
132: as the block matrix
133: $\boldsymbol{\sigma }=\left( \begin{array}{c|c} A & C \\ \hline
134: C^{T} & B \end{array} \right)$,
135: $\sigma _{hk}=\frac{1}{2}\langle \{R_{k},R_{h}\}\rangle -\langle
136: R_{k}\rangle \langle R_{h}\rangle$
137: where $A$, $B$ and $C$ are $2\times 2$ real matrices, $\langle O\rangle
138: = \mathrm{Tr}(\varrho \,O)$ and $\{f,g\}=fg+gf$. In the following we
139: will use the notation $a\equiv a_{1}$ and $b\equiv a_{2}$ and also
140: consider the four additional auxiliary modes $c=(a+b)/\sqrt{2}$,
141: $d=(a-b)/\sqrt{2}$, $e=(ia+b)/\sqrt{2}$, and $f=(ia-b)/\sqrt{2}$
142: obtained by the action of polarizing beam splitters (PBS) and
143: phase-shifters on modes $a$ and $b$. Positivity of the density matrix
144: for physical states is written in terms of the uncertainty relation for
145: the minimum symplectic eigenvalue $\nu _{-}$ of the CM, {\em i.e.}
146: $\nu_{-}\geq 1/2$.
147: For Gaussian states, the state purity is given by $\mu (\boldsymbol{
148: \sigma })=(4\sqrt{\mathrm{Det[\boldsymbol{\sigma }]}})^{-1}$ whereas
149: separability corresponds to positivity of the partially transpose (PPT)
150: density matrix. A bipartite Gaussian state is separable iff
151: $\tilde{\nu}_{-}>1/2 $, where $\tilde{\nu}_{-}$ is the minimum
152: symplectic eigenvalue of $\boldsymbol{\Delta } \boldsymbol{\sigma
153: }\boldsymbol{\Delta }$, $ \boldsymbol{\Delta }=
154: \mathrm{Diag}[1,1,1,-1]$. A convenient measure of entanglement is thus
155: given by the logarithmic negativity $E_{\mathcal{N}}(\boldsymbol{ \sigma
156: })=\mathrm{max}(0,-\ln 2 \tilde{\nu}_{-})$ \cite{vid02} and
157: the EoF $E_{\cal F}(\boldsymbol{ \sigma})$ can be evaluated
158: following Ref.~\cite{MM:08}. In addition
159: to $E_{\mathcal{N}}(\boldsymbol{\sigma })$ the Duan criterion gives a
160: necessary condition for non-separability in terms of the noise
161: properties of $c$ and $d$ \cite{dua00} whereas a stricter condition
162: \cite{tre05}, referred to as EPR criterion, involves the conditional
163: variances on $x_{a}$ and $y_{a}$ obtained from a measurement of $x_{b}$
164: and $y_{b}$ for the explicit expressions in terms of the noise on modes
165: $a$, $b$, $c$ and $d$.
166: \par
167: In our experiment, the block $A$ of the CM is retrieved by measuring
168: the single-mode quadratures of mode $a$: the variances of $x_{a}$ and
169: $y_{a}$ give the diagonal elements, while the off diagonal ones are
170: obtained from the additional quadratures $z_{a}\equiv \left(
171: x_{a}+y_{a}\right) /\sqrt{2}$ and $\ t_{a}\equiv \left(
172: x_{a}-y_{a}\right) /\sqrt{2}$ as $\sigma _{12}= \sigma
173: _{21}=\frac12(\langle z_{a}^{2}\rangle -\langle t_{a}^{2}\rangle
174: )-\langle x_{a}\rangle \langle y_{a}\rangle$ \cite{sh05}. The block $B$
175: is reconstructed in the same way from the quadratures of $b$, whereas
176: the elements of the block $C$ are obtained from the quadratures of the
177: auxiliary modes $c$, $d$, $e$ and $f$ as follows
178: $\sigma _{13}= \frac12(\langle x_{c}^{2}\rangle -\langle x_{d}^{2}\rangle
179: )-\langle x_{a}\rangle \langle x_{b}\rangle$,
180: $\sigma _{14}= \frac12 (\langle y_{e}^{2}\rangle -\langle y_{f}^{2}\rangle
181: )-\langle x_{a}\rangle \langle y_{b}\rangle$,
182: $\sigma _{23}= \frac12 (\langle x_{f}^{2}\rangle -\langle x_{e}^{2}\rangle
183: )-\langle y_{a}\rangle \langle x_{b}\rangle$,
184: $\sigma _{24}= \frac12 (\langle y_{c}^{2}\rangle -\langle y_{d}^{2}\rangle
185: )-\langle y_{a}\rangle \langle y_{b}\rangle$.
186: Notice that the measurement of the $f$-quadratures is not mandatory,
187: since $\langle x_{f}^{2}\rangle =\langle x_{b}^{2}\rangle
188: +\langle y_{a}\rangle ^{2}-\langle x_{e}^{2}\rangle $ and $\langle
189: y_{f}^{2}\rangle =\langle x_{a}^{2}\rangle +\langle y_{b}^{2}\rangle
190: -\langle y_{e}^{2}\rangle $. Analogous expressions hold for $\langle
191: x_{e}^{2}\rangle $ and $\langle y_{e}^{2}\rangle $.
192: \par
193: In the ideal case the OPO output is in a twin-beam state
194: $\mathbf{S}(\zeta )|0\rangle$, ${\mathbf S}(\zeta) =
195: \exp\{\zeta a^\dag b^\dag - \bar\zeta ab\}$ being the entangling two-mode
196: squeezing operator: the corresponding CM has diagonal blocks
197: $A$, $B$, $C$ with the two diagonal elements of
198: each block equal in absolute value. In realistic OPOs, cavity and crystal
199: losses lead to a mixed state, {\em i.e.} to an effective thermal
200: contribution. In addition, spurious nonlinear processes, not perfectly
201: suppressed by the phase matching, may combine to the down conversion,
202: contributing with local squeezings. Finally, due to small misalignments of
203: the nonlinear crystal, a residual component of the field polarized along $a$
204: may project onto the orthogonal polarization (say along $b$), thus
205: leading to a mixing among the modes \cite{dau08}. Overall,
206: the state at the output is expected to be a zero amplitude Gaussian
207: entangled state, whose general form may be written as
208: $\varrho_{g} = {\mathbf U} (\beta) {\mathbf S}(\zeta)\,
209: {\mathbf{LS}}(\xi_1,\xi_2)\, {\mathbf T}\,
210: {\mathbf{LS}}^\dag(\xi_1,\xi_2)\, {\mathbf S}^\dag (\zeta)
211: {\mathbf U}^\dag (\beta)$, where
212: ${\mathbf T}= \tau_1 \otimes \tau_2$, with
213: $\tau_k = (1+ \bar n_k)^{-1} [\bar n_k/(1+\bar n_k )]^{a^\dag a}$
214: denotes a two-mode thermal state with $\bar n_k$ average photons
215: per mode, ${\mathbf{LS}}(\xi_1,\xi_2)= S(\xi_1) \otimes S(\xi_2)$,
216: $S(\xi_k)=\exp\{\frac12 (\xi_k a^{\dag 2} - \bar\xi_k a^2)\}$ denotes
217: local squeezing and ${\mathbf U}(\beta)= \exp\{\beta a^\dag b
218: - \bar\beta ab^\dag\}$ a mixing operator, $\zeta$, $\xi_k$ and $\beta$
219: being complex numbers.
220: For our configuration, besides a thermal contribution due to internal
221: and coupling losses, we expect a relevant entangling contribution
222: with a small residual local squeezing and, as mentioned above, a possible
223: mixing among the modes. The CM matrix corresponding to $\varrho _{g}$ has
224: diagonal blocks $A$, $B$, and $C$ with possible
225: asymmetries among the diagonal elements.
226: \par
227: {\em Experimental setup}---The experimental setup, shown in Fig.~\ref{f:setup},
228: relies on a CW internally frequency doubled Nd:YAG laser
229: pumping (@532~nm) a non degenerate OPO based
230: on a periodically poled $\alpha $--cut KTP (PPKTP) crystal
231: (\textit{Raicol Crystals Ltd}. on custom design) \cite{fei92}.
232: The use of the $\alpha $-cut
233: PPKTP allows implementing a type-II phase matching with cross polarized
234: signal ($a$) and idler ($b$) waves, frequency degenerate @1064~nm for a
235: crystal temperature of $\approx 53^{\circ }$C.
236: The OPO cavity is locked to the pump beam by Pound-Drever technique
237: \cite{dre83} and adjusted to work in triple resonance by finely tuning its
238: geometrical properties \cite{dau08}. The cavity output coupling @1064~nm is $%
239: \approx 0.73$, corresponding to an experimental line-width of $16$ MHz
240: @1064~nm. The measured oscillation threshold is $P_{th}\approx 50$ mW; during
241: the acquisition the system has been operated below threshold at 60\% of the
242: threshold power.
243: %%%%%%%%%%%%%%%
244: \begin{figure}[h]
245: \includegraphics[width=0.45\textwidth]{Fig1.eps}
246: \vspace{-0.3cm}
247: \caption{
248: Experimental setup: A type-II OPO containing a
249: periodically poled crystal (PPKTP) is pumped by the second harmonic of a
250: Nd:YAG laser. At the OPO output, a half-wave plate
251: ($\lambda /2_{\rm out}$), a quarter-wave plate ($\lambda /4_{\rm out}$) and a
252: PBS$_{\rm out}$ select the mode for homodyning.
253: The resulting electronic signal is acquired via a PC module.} \label{f:setup}
254: \end{figure} \\
255: %%%%%%%%%%%%%%%
256: In order to select mode $a$ and $b$ or their combinations $c$ and $d$,
257: the OPO beams are sent to a half-wave plate and a PBS.
258: Modes $e$ and $f$ are obtained by inserting an additional
259: quarter-wave plate \cite{sh05}. The PBS output
260: goes to a homodyne detector, described in details in \cite{dau05,opt05},
261: exploiting the laser output @1064~nm as local oscillator (LO).
262: %The homodyne photocurrent is formed from the output of two
263: %photodiodes \emph{Epitaxx ETX300}, both matched to a low--noise
264: %trans--impedance AC ($>$ few kHz) amplifier. The difference-photocurrent is
265: %further amplified by a low noise high gain amplifier \emph{Miteq AU1442}.
266: The overall homodyne detection efficiency is $\eta =0.88 \pm 0.02$. The LO
267: reflects on a piezo-mounted mirror (PZT), which allows varying its phase $%
268: \theta $. In order to avoid the laser low frequency noise, data sampling is
269: moved away from the optical carrier frequency by mixing the homodyne current
270: with sinusoidal signal of frequency $\Omega =3$~MHz \cite{opt05}. The
271: resulting current is low--pass filtered ($B=300$~kHz) and sampled by a PCI
272: acquisition board (Gage 14100, 1M--points per run, 14 bits resolution). The
273: total electronic noise power has been measured to be $16$~dBm below the
274: shot--noise level, corresponding to a signal to noise ratio of about $40$.
275: %%%%
276: \par{\em Reconstruction and experimental results}---Acquisition is triggered
277: by a linear ramp applied to the PZT and adjusted to obtain a $2\pi $
278: variation in $200$ ms. Upon spanning the LO phase $\theta $,
279: the quadratures $x\left( \theta \right) =x\cos \theta +y\sin \theta $
280: are measured. Calibration with respect to the noise of the vacuum state
281: is obtained by acquiring a set of data with the output from the OPO
282: obscured. All the expectation values needed to reconstruct ${\boldsymbol
283: \sigma }$ are obtained by quantum tomography \cite{qht03}, which
284: allows to compensate nonunit quantum efficiency and to reconstruct any
285: expectation value, including those of specific quadratures and their
286: variances, by averaging special pattern functions over the whole data set.
287: As a preliminary check of the procedure, we verified
288: that the CM of the vacuum state is consistent with
289: ${\boldsymbol\sigma}_0 = \frac12\,\mathbb{I}$ within the experimental
290: errors.
291: %%%%%%%%%%%%%
292: \begin{figure}[h]
293: \includegraphics[width=0.4\textwidth]{Fig2.eps}
294: \vspace{-0.3cm}
295: \caption{(Left): experimental homodyne traces and reconstructed CM;
296: (Right): reconstructed Wigner functions. Top plots are for modes $c$
297: and bottom ones for $d$. Arrows on the
298: homodyne plots show the positions of the maximum and minimum variances.
299: $\theta$ is the relative phase between the signal and the LO.
300: \label{f:wig:cd}}
301: \end{figure}
302: %%%%%%%%%%%%%
303: \par
304: We start our analysis by checking the Gaussian character of the OPO
305: signals, {\em i.e.} upon evaluating the Kurtosis of homodyne
306: distribution at fixed phase of the LO \cite{opt05}.
307: Besides, we checked that the mean
308: values of all the involved quadratures are negligible, in agreement with
309: the description of OPO output as a zero amplitude state. Then, we have
310: measured the quadratures of the six modes $a$-$f$. We found the modes
311: $a$ and $b$ excited in a thermal state, thus confirming the absence of
312: relevant local squeezing. Their combinations $c$, $d$, $e$ and $f$ are
313: squeezed thermal states with squeezing appearing on $y_{c}$,
314: $x_{d}\,$,$\ t_{e}$ and $z_{f}$, respectively. In Fig.~\ref{f:wig:cd}
315: we show the experimental homodyne traces for modes $c$ and $d$ as well
316: as the corresponding Wigner functions, obtained by reconstructing the
317: single-mode CM. As it is apparent from the plots both modes are
318: squeezed with quadratures noise reduction, corrected for nonunit
319: efficiency, of about $2.5$~dB. An analogue behavior has been observed
320: for modes $e$ and $f$. The CM of Fig. \ref{f:wig:cd} indeed reproduce
321: that of an entangled thermal state with small corrections due to local
322: squeezing and mixing. The relevant parameters to characterize the
323: corresponding density matrix $\varrho_g$ are the mean number of thermal
324: photons $\bar n_1 \simeq 0.67$, $\bar n_2 \simeq 0.18$ and entangling
325: photons $\bar n_s = 2 \sinh^2 |\zeta|\simeq 0.87$ \cite{pars}. The
326: errors on the CM elements for the blocks $A$ and $B$ are of the order
327: $\delta\sigma_{jk} \simeq 0.004$ and have been obtained by propagating
328: the tomographic errors. In this case phase fluctuations are irrelevant,
329: since the two modes are both excited in a thermal state. On the other
330: hand, in evaluating the errors on the elements of the block $C$ the
331: phase-dependent noise properties of the involved modes have to be taken
332: into account, and the tomographic error has to be compared with the
333: error due to the finite accuracy in setting the LO phase $\theta$. The
334: elements $\sigma _{13}$ and $\sigma _{24}$ are obtained as combinations
335: of squeezed/anti--squeezed variances, which are quite insensitive to
336: fluctuations of $\theta $. As a consequence the errors on these elements
337: are given by the overall tomographic error $\delta\sigma_{jk} \simeq
338: 0.004$. On the other hand, the elements $\sigma _{14}$ and $\sigma
339: _{23}$ depend on the determination of $x_{e,f}^{2}$ and $y_{e,f}^{2}$,
340: which are sensible to phase fluctuations. In order to take into account
341: this effect we evaluate errors as the fluctuations in the
342: tomographically reconstructed quadratures induced by a $\delta \theta
343: \simeq 20$~mrad variation in the LO phase, corresponding to the
344: experimental phase stability of the homodyne detection. The resulting
345: errors are about $\delta\sigma_{14} = \delta\sigma_{23}\simeq 0.03$ for
346: both CM elements. The off-diagonal elements of the three matrices $A$,
347: $B$ and $C$ are thus zero within their statistical errors, in agreement
348: with the expectation for an entangled thermal state. As mentioned
349: above, the experimental procedure may be somehow simplified exploiting
350: the relationships among modes, and expressing mode $e$ or $f$ in terms
351: of the others: only five modes are then needed.
352: Upon rewriting the off-diagonal terms of $C$ in terms of the five modes
353: we arrive at $\sigma _{14}=0.02\pm 0.03$ and $\sigma_{23}=0.04\pm 0.03$
354: when eliminating the mode $f$ and $\sigma_{14}=0.06\pm 0.03$ and
355: $\sigma_{23}=0.06\pm 0.03$ when eliminating the mode $e$. Both
356: procedures provide results in agreement with those obtained by using the
357: complete set of homodyne data for the six modes.
358: \par
359: Since the minimum symplectic eigenvalue of ${\boldsymbol\sigma }$ is
360: $\nu_{-}=0.68\pm 0.02\geq 0.5$, the CM corresponds to a physical state.
361: State purity is $\mu (\boldsymbol\sigma) =
362: 0.31\pm 0.01$. The minimum symplectic eigenvalue for the partial
363: transpose is $\tilde{\nu}_{-}=0.24\pm 0.02$, which corresponds to a
364: logarithmic negativity $E_{\mathcal{N}}(\boldsymbol{\sigma }%
365: )=0.73\pm 0.02$, {\em i.e.} the state is entangled, with EoF
366: $E_{\mathcal{F}}(\boldsymbol{\sigma })=1.46\pm 0.02$.
367: In turn, it satisfies the Duan inequality with the results $0.29\pm
368: 0.01<1/2$ and the EPR criterion with $0.21\pm 0.01<1/4$.
369: \par
370: Entangled Gaussian states as $\varrho_{g}$ may be endowed with
371: nonclassical photon number correlations, {\em i.e} squeezing in the
372: difference photon number. This may be checked upon evaluating the noise
373: reduction factor ${\cal R} = \hbox{Var}(D_{ab})/(\bar N_a+\bar N_b)$
374: where $\hbox{Var} (D_{ab})$ denotes the variance of the difference
375: photocurrent $D_{ab}= N_a - N_b$, $N=a^\dag a$ being the number
376: operator, and $\bar N_k = \langle N_k \rangle$ the average
377: photon number. A value ${\cal R}<1$ is a marker of nonclassical
378: correlations between the two modes. We obtained ${\cal R}=0.50\pm 0.02$,
379: in agreement with the theoretical description \cite{deg07} for the
380: values of thermal and entangling photons reported above. Starting from
381: the CM one can reconstruct the full joint photon distribution
382: $p(n,m)$ of the modes $a$ and $b$: the result is shown in Fig.~\ref{f:pnm}
383: where the correlations between the two modes are clearly seen. We have
384: also evaluated the single-mode photon distributions (either from
385: data or from the single-mode CM) for modes $a$-$d$. Results are
386: reported in Fig.~\ref{f:pnm}: distributions of $a$ and $b$ are thermal,
387: whereas the statistics of modes $c$ and $d$ correctly reproduces the
388: even-odd oscillations expected for squeezed thermal states.
389: %%%%%
390: \begin{figure}[h]
391: \centerline{\includegraphics[width=0.40\textwidth]{Fig3.eps}}
392: \vspace{-0.3cm}
393: \caption{(Left): Joint photon number distribution $p(n,m)$ for the entangled
394: state of modes $a$ and $b$ at the output of the OPO. (Right): single-mode
395: photon distributions $p(n)$ for modes $a$ and $c$ (top right) and $b$ and
396: $d$ (bottom right). The single-mode distributions of modes $a$ and $b$
397: are thermal and correspond to the marginals of $p(n,m)$.
398: The distributions for modes $c$ and $d$ are those of squeezed thermal
399: states. \label{f:pnm}}
400: \end{figure}
401: %%%%%
402: %%%%%%%%%%%%%%%%%%%%%%%%
403: \par{\em Conclusion}---We have presented the complete
404: reconstruction of the CM for the output of a CW type II non-degenerate
405: OPO, below threshold and frequency degenerate. The CM elements have
406: been retrieved as combinations of expectations and variances of suitable
407: mode quadratures, obtained by combining the entangled modes by linear
408: optics. The quantities of interest have been obtained tomographically,
409: processing the whole data set and thus reducing statistical
410: fluctuations. Upon exploiting a general model allowing local squeezing
411: and polarization cross-talking inside the crystal, we have very
412: precisely described the experimental CM with the theory underlying
413: parametric downconversion, thus providing a full explanation of
414: experimental findings. The reconstructed state is a Gaussian entangled
415: state close to a two-mode squeezed thermal state, the corresponding
416: entanglement and nonclassical photon number correlations have been
417: demonstrated. We conclude that single homodyne schemes are convenient
418: and robust setups for the full characterization of OPO signals and, in
419: turn, represent a relevant tool for quantum technology based on CV
420: entanglement, e.g., the full characterization of CV Gaussian
421: channels by input-output signals' characterization. Finally,
422: making use of a single OPO and a single homodyne detector, our setup
423: represents also a compact and robust tool for entanglement generation and
424: characterization.
425: %%%%%%%%%%%%%%%%%%%%%%%%
426: \par {\em Acknowledgments}---This work has been partially
427: supported by CNR-CNISM. MGAP thanks M.~Bondani and A.~Allevi
428: for discussions.
429: %%%%%%%%%%%%%%%%%%%%%%%%
430: \begin{thebibliography}{99}
431: \bibitem{sh05} V.~D'Auria \emph{et al.}, J. Opt. B {\bf 7}, 750 (2005);
432: A. Porzio \emph{et al.}, Int. J. Quant Inf. {\bf 5}, 63 (2007).
433: \bibitem{ray99} M. G. Raymer, A. Funk, Phys. Rev. A {\bf 61}, 015801
434: (1999); D. F. McAlister, M. G. Raymer, J. Mod. Opt. {\bf 44}, 2359
435: (1997).
436: \bibitem{eis03} J. Eisert \emph{et al.}, Int. J. Quant. Inf. \textbf{1}, 479 (2003); A.~Ferraro
437: \emph{et al.}, {\em Gaussian States in Quantum Information}, (Bibliopolis,
438: Napoli, 2005); G. Adesso \emph{et al.}, J. Phys. A 40, 7821 (2007);
439: S.~L.~Braunstein \emph{et al.}, Rev. Mod. Phys. \textbf{77},513 (2005).
440: %%
441: \bibitem{vnl02} P. van Loock, Fort. Phys. {\bf 50}, 1177 (2002).
442: %%
443: \bibitem{lug02} L. A. Lugiato \emph{et al.}, J. Opt. B {\bf 4}, S176 (2002).
444: %%
445: \bibitem{dar01} G. M. D'Ariano \emph{et al.}, Phys. Rev. Lett. {\bf 87}, 270404 (2001).
446: %%
447: \bibitem{dau05} V. D'Auria \emph{et al.}, J. Phys. B {\bf 39}, 1187 (2006).
448: %%
449: \bibitem{simXX} R.~Simon \emph{et al.}, Phys. Rev. A \textbf{36}, 3868 (1987);
450: Phys. Rev. A \textbf{49}, 1567 (1994).
451: %%
452: \bibitem{marYY} P.~Marian \emph{et al.}, Phys. Rev. A \textbf{47}, 4474 (1993); 4487 (1993);
453: Phys. Rev. A \textbf{68}, 062309 (2003).
454: %%
455: \bibitem{sim00} R.~Simon, Phys. Rev. Lett. \textbf{84}, 2726 (2000).
456: %%
457: \bibitem{dua00} L.-M. Duan \emph{et al.}, Phys. Rev. Lett. \textbf{84}, 2722 (2000).
458: %%
459: \bibitem{bow03} W.~P.~Bowen \emph{et al.}, Phys. Rev. Lett. \textbf{90},
460: 043601 (2003).
461: %%
462: \bibitem{theZZ} M. G. A. Paris, Phys. Lett. A {\bf 225}, 28 (1997); X. B. Wang,
463: Phys. Rev. A {\bf 66}, 024303 (2002).
464: %%
465: \bibitem{bow04} W.~P.~Bowen \emph{et al.}, Phys. Rev. A \textbf{69}, 012304 (2004).
466: %%
467: \bibitem{dig07} J.~Di Guglielmo \emph{et al.}, Phys. Rev. A \textbf{76}, 012323 (2007).
468: %%
469: \bibitem{lau05} J.~Laurat \emph{et al.}, J. Opt. B \textbf{7}, S577 (2005).
470: %%
471: \bibitem{dru90} P.~D.~Drummond \emph{et al.}, Phys. Rev. A \textbf{41}, 3930 (1990).
472: %%
473: \bibitem{zha99} Y.~Zhang \emph{et al.}, Phys. Lett. A \textbf{259}, 171 (1999).
474: %%
475: \bibitem{kim92} Z.~Y.~Ou\emph{et al.}, Phys. Rev. Lett. \textbf{68}, 3663 (1992).
476: %%
477: \bibitem{jin06} J.~Jing \emph{et al.}, Phys. Rev. A \textbf{74}, 041804(R) (2006).
478: %%
479: \bibitem{bre07} A.~S.~Villar \emph{et al.}, J. Opt. Soc. Am. B \textbf{24}, 249 (2007).
480: %%
481: \bibitem{kel08} G. Keller \emph{et al.}, submitted to (2008).
482: %%
483: \bibitem{wen04} J. Wenger \emph{et al.}, Eur. Phys. J. D \textbf{32}, 391 (2004).
484: %%
485: \bibitem{tre05} N. Treps \emph{et al.}, Las. Phys. \textbf{15}, 187 (2005).
486: %%
487: \bibitem{vid02} G.~Vidal \emph{et al.}, Phys. Rev. A \textbf{65}, 032314 (2002).
488: %%
489: \bibitem{MM:08} P.~Marian and T.~A.~Marian, arXiv:0809.0321v1 [quant-ph].
490: %%
491: \bibitem{dau08} V. D'Auria \emph{et al.}, Appl. Phys. B, \textbf{91}, 309 (2008).
492: %%
493: \bibitem{fei92} M.~M.~Fejer \emph{et al.}, IEEE J. Q. Electr. \textbf{28}, 2631 (1992).
494: %%
495: \bibitem{dre83} R.~W.~P.~Drever \emph{et al.}, Appl. Phys. B \textbf{31}, 97 (1983).
496: %%
497: \bibitem{opt05} V.~D'Auria \emph{et al.}, Opt. Express \textbf{13}, 948
498: (2005).
499: %%
500: \bibitem{qht03} G.~M.~D'Ariano \emph{et al.}, Adv. Imag. Electr. Phys. \textbf{128}, 205 (2003).
501: %%
502: \bibitem{pars} The other parameters are given by $\xi_1 \simeq 0.07 e^{i
503: 0.12 \pi}$, $\xi_2 \simeq 0.12 e^{-i 0.12\pi}$, $\beta \simeq 0.13\pi
504: e^{-i 0.17 \pi}$.
505: %%
506: \bibitem{deg07} I. P. Degiovanni \emph{et al.}, Phys. Rev. A {\bf 76}, 062309 (2007).
507: \end{thebibliography}
508: %%%%%%%%%%%%%%%%%%%%%%%%
509: \end{document}
510: