1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: \documentstyle{article}
3: \input epsf
4: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
5: \setlength{\textwidth}{14.0cm}
6: \setlength{\textheight}{20cm}
7: \setlength{\topmargin}{-5mm}
8: \setlength{\evensidemargin}{0cm}
9: \setlength{\oddsidemargin}{0cm}
10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
11: \begin{document}
12: \parbox{13 cm}
13: { \begin{flushleft} \vspace* {1.2 cm}
14: {\Large\bf {Engineering QND measurements for continuous
15: variable quantum information processing}}\\
16: \vskip 1truecm {\large\bf { Matteo G A Paris } }\\ \vskip 5truemm
17: {Quantum Optics $\&$ Information Group, INFM Udr Pavia, Italy\\
18: {\tt E-mail:paris@unipv.it}, {\tt URL: www.qubit.it/\,$\tilde{}\,$paris}
19: } \end{flushleft} } \vskip 0.5truecm {\bf Abstract:\\} {
20: \noindent A novel scheme to realize the whole class
21: of quantum nondemolition (QND) measurements of a field quadrature is
22: suggested. The setup requires linear optical components and squeezers,
23: and allows optimal QND measurements of quadratures,
24: which minimize the information gain versus state disturbance
25: trade-off.} \vskip 0.1 cm
26: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27: \section{Introduction}\label{s:intro}
28: There is a growing interest for continuous variable (CV) quantum
29: information processing, in particular for implementations based
30: on manipulations of Gaussian states of light in optical
31: circuits. Several quantum protocols \cite{furu,geza},
32: including teleportation, error
33: correction, cloning and entanglement purification have been extended
34: to CV systems, which may be easier to manipulate than quantum bits
35: in order to accomplish the desired tasks \cite{plz}. \par
36: In optical implementations, quantum information is encoded in values
37: of a single-mode field quadrature, say $x=1/2(a^\dag+a)$, $[a,a^\dag]$=1 being the
38: mode operators. Therefore, in principle, the most relevant measurement
39: for quantum protocols is provided by homodyne detection. However,
40: the usual implementation of homodyning corresponds to a destructive
41: detection, such that after the measurement we have no longer at disposal
42: a quantum signal for further manipulations and/or measurements.
43: It is thus of interest to devise a scheme for quantum nondemolition (QND)
44: measurements of a field quadrature, in particular for tunable QND
45: measurements, in which the trade-off between information and disturbance may
46: adjusted according to different needs. \par
47: In this paper we suggest an all-optical scheme to realize the whole class
48: of QND measurements of a field quadrature \cite{xqnd}, from Von Neumann projective
49: measurement to fully non-demolitive, non-informative one. The setup involves
50: only linear optical components (including squeezers) and also allows an
51: optimal QND measurement, which minimizes the information gain
52: versus state disturbance trade-off. \par
53: The next section is devoted to the abstract description of a quantum
54: measurement, and to introduce two fidelities in order to quantify the state
55: disturbance and the information gain due to a measurement. In Section \ref{s:qnd}
56: we analyze with some details the setup for QND measurements of a field
57: quadrature. Section \ref{s:outro} closes the paper with some concluding remarks.
58: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
59: \section{Quantum measurements}\label{s:qmeas}
60: A generic quantum measurement is described by a set of {\em measurement
61: operators $\{M_k\}$}, with the condition $\sum_k M^\dag_k M_k=
62: {\rm I}$. The POVM of the measurement is given by $\{E_k\equiv M_k^\dag M_k\}$
63: whereas its quantum operation is expressed as $\varrho \rightarrow
64: \sum_k M_k\varrho M_k^\dag$. This means that, if $\varrho$ is the initial
65: quantum state
66: of the system under investigation, the probability distribution of the outcomes
67: is given by $p_k=\hbox{Tr}[\varrho\: E_k]= \hbox{Tr}[\varrho\:
68: M^\dag_k M_k]$, whereas the conditional output state, after having
69: detected the outcome $k$, is expressed as $\sigma_k= \: M_k\varrho
70: M_k^\dag/p_k$, such that the overall quantum state after the measurement is
71: described by the density matrix $\sigma=\sum_k p_k \: \sigma_k=\sum_k
72: M_k\varrho M_k^\dag$. \par
73: Suppose you have a quantum system prepared in the state $\varrho$, and that
74: you are interested in measuring the observable $K$. The so-called Von-Neumann
75: (VN) measurement of $K$ is described by the operators $\{M_k=|k\rangle\langle
76: k|\}$, where the $|k\rangle$'s are the eigenstates of $K$. Following the above
77: prescription we have for VN measurements $p_k=\langle k|\varrho|k\rangle$,
78: $\sigma_k=|k\rangle\langle k|$ and $\sigma= \sum_k p_k |k\rangle\langle k|$.
79: As a matter of fact, VN measurement (also called {\em projective} measurement)
80: provides the maximum accessible information about the quantity $K$, at the price
81: of {\em erasing} the quantum information of the state being investigated, which
82: is no longer at disposal for further investigations or manipulations.
83: \par
84: In opposition to projective measurement one may conceive a {\em nondemolitive}
85: measurement of $K$, which preserves the quantum state. This kind of measurement
86: is described by the operators $\{M_k \propto I\}$, proportional to the identity
87: operator. We have uniform $p_k$ and $\sigma=\varrho$, {\em i.e.}
88: the quantum state is preserved, however the measurement is completely
89: {\em non informative}. Overall, such kind of measurement may be viewed as a {\em
90: blind} quantum repeater, which re-prepares any quantum state received at the
91: input, without giving any information on its characteristics.
92: \par
93: Between these two extrema there is a complete class of intermediate cases,
94: {\em i.e.} quantum measurements providing only partial information about the
95: distribution $\{p_k\}$ while partially preserving the quantum state of the system.
96: These schemes are sometimes referred to as QND
97: measurements of the quantity $K$.
98: \par
99: Let us now consider a generic quantum measurement $\{Q_k\}$ aimed
100: to provide information about the quantity $K$. Two questions naturally arise
101: about the characterization of its operation: \\
102: i) How much information is provided by the measurement ? Or, in other words,
103: how close are the probability distributions $q_k=\hbox{Tr}[\varrho\: Q_k^\dag
104: Q_k]$ and $p_k=\langle k|\varrho |k\rangle$ ? In order to quantify this
105: resemblance we remind that the space of probability distributions
106: $\{p_k\}_{k=1,...,M}$ is the M-simplex, where a privileged metric (the Fisher
107: metric) exists and induces a distance between probabilities \cite{ks} given by
108: $G=\left( \sum_k \sqrt{p_k\: q_k}\right)^2$
109: which represents a measure of the statistical distinguishability
110: between the two distributions. \\ ii) How destructive is the measurement ? {\em I.e.}
111: how far is the output state $\sigma$ to the input state $\varrho$ ?
112: The {\em geometric} distance between two density matrices is given by
113: $F=\left(\hbox{Tr}\left[\:\sqrt{\sqrt{\varrho}\:\sigma\sqrt{\varrho}}
114: \:\right]\right)^2$. This quantity is the proper generalization to mixed states
115: of the standard quantum overlap used to quantify
116: the (statistical, {\em i.e.} by measurements) distinguishability of pure states \cite{br}.
117: F is also characterized by Uhlman theorem, which states that if $\varrho$ and
118: $\sigma$ are density matrices on a given Hilbert space ${\cal H}$, then
119: $F=\max_{\psi,\varphi}\: \left| \langle \psi | \varphi \rangle \right|^2$,
120: where $|\varphi\rangle$ and $|\psi\rangle$ are generic purification of
121: $\varrho$ and $\sigma$, {\em i.e.} pure states on ${\cal H} \otimes {\cal H}$
122: which have $\varrho$ and $\sigma$ as partial traces. If either $\varrho$ or
123: $\sigma$ is pure F reduces to the standard overlap. For example, if
124: $\varrho=| \Psi\rangle\langle\Psi|$ then we have
125: $F=\langle\Psi|\sigma|\Psi\rangle$.
126: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
127: \section{QND measurements of a field quadrature}\label{s:qnd}
128: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
129: Our proposal to realize the whole class of QND measurements of a field
130: quadrature is depicted in Fig. 1. At first, the signal beam is mixed with a probe
131: beam (which will be excited in squeezed vacuum state) in beam splitter of tunable
132: transmissivity $\tau_1$ ($BS_1$ in the figure). A tunable beam splitter
133: can be easily implemented by a Mach-Zehnder interferometer. In the following we will write the
134: transmittivity in the form $\tau_1=\cos^2\phi$. For the sake of simplicity, we consider
135: the measurements of the zero-phase quadrature $x=1/2(a^\dag+a)$; however, the same analysis is
136: valid for the generic $\theta$-quadrature
137: $x_\theta=1/2(a^\dag e^{i\theta}+a e^{-i\theta})$. \\
138: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
139: \begin{minipage}{6cm}
140: \begin{center}
141: \epsfxsize=6cm \epsfbox{setup.ps}
142: \end{center}
143: \end{minipage}
144: \begin{minipage}{8cm} {\small
145: {\bf Fig. 1}: Schematic diagram of the setup for QND measurements of a field
146: quadrature. The signal is mixed with a squeezed vacuum probe in a beam
147: splitter ($BS_1$) of tunable transmittivity $\tau_1$. The probe beam is then
148: revealed by homodyne detection, whereas the signal beam is firstly displaced
149: according to the value of the homodyne outcome, and then squeezed according to
150: the transmittivity $\tau_1$ (see text). The local oscillator and the pumps
151: for the displacing and squeezing stages are provided by a common laser
152: source.}
153: \end{minipage} \\ $ $ \\
154: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
155: After the beam splitter, the probe beam is revealed by homodyne detection.
156: Taking into account the reflectivity amplitude of $BS_1$, from an outcome
157: $X$ from the homodyne we infer a value $x=-X/\sin\phi$ for the quadrature
158: of the signal beam. The signal is then displaced by an amount
159: $\alpha^*=-X\tan\phi=x\tan\phi\sin\phi$, by feedback of the outcome of the
160: homodyne detector, and finally squeezed by an amount $r^*$ such that
161: $\exp\{r^*\}=\sqrt{\tau_1}=\cos\phi$.
162: The displacement transformation $D(\alpha^*)=\exp\{\alpha^* a^\dag -
163: \bar\alpha^* a\}$ can be obtained by mixing the signal with a strong pump beam
164: excited in strong coherent state $|z\rangle$ $|z|\rightarrow\infty$ (which
165: may be by the local oscillator of the homodyne detector) in a
166: beam splitter ($BS_3$ in the figure) with transmittivity approaching unit
167: value $\tau_3 \rightarrow 1$ such that $\alpha^*=z\sqrt{1-\tau_3}$.
168: The squeezing transformation is obtained with a degenerate optical parametric
169: amplifier (DOPA), where the pump mode is again provided by the common laser source
170: providing the homodyne LO. Since the cosine is smaller than one, $r^*$ is negative,
171: and this means that we are squeezing the signal in a direction orthogonal to
172: the quadrature we are going to measure. \par
173: The probability density for the inferred values of the signal
174: quadrature is given by
175: \begin{eqnarray}
176: p(x)=-\sin\phi \: q(X) = \tan\phi \int dy\: |\psi_{\sc s}(y)|^2 \:
177: \left|\psi_{\sc p}\left[\tan\phi(y-x_0)\right] \right|^2
178: \label{probx0}\;,
179: \end{eqnarray}
180: where $q(X)=\hbox{Tr}[|\psi_{\sc s},\psi_{\sc p}\rangle
181: \rangle\langle\langle\psi_{\sc p},\psi_{\sc s}|\:
182: I \otimes |X\rangle\langle X|]$ is the probability density for the homodyne outcomes, and
183: $\psi_j(x), \: j={\sc s,p}$ are the signal and probe wave-functions in the
184: quadrature representation, {\em i.e.} $|\psi_j\rangle=\int dx\; \psi_j(x)\: |x\rangle$.
185: The conditional output state, after having inferred the value
186: $x$ for the signal quadrature, is given by
187: \begin{eqnarray}
188: |\psi_x\rangle = \sqrt{\frac{\sin\phi}{p(x)}}\: S(r^*)\: D(\alpha^*)\: \langle
189: -x \sin\phi | V_\phi |\psi_{\sc s},\psi_{\sc p}\rangle\rangle
190: \label{condx}\;,
191: \end{eqnarray}
192: where $V_\phi=\exp\{i\phi(a^\dag b+ b^\dag a)\}$ is the evolution operator of
193: the beam splitter $BS_1$.
194: For a probe mode excited in a squeezed vacuum we
195: have $\psi_{\sc p}(x)=(2\pi\Sigma^2)^{-1/4}\exp \{-\frac{x^2}{4\sigma^2_{\sc p}}\}$,
196: where the variance is given by $\sigma_{\sc p}^2=1/4 \exp\{\pm 2r\}$ according to the
197: direction of squeezing. We refer to as a {\em squeezed} probe for the minus sign
198: (squeezing in the direction of the quadrature to be measured) and to as an {\em
199: antisqueezed} probe for the plus sign (squeezing in the orthogonal direction).
200: The average number of photons carried by the probe is given by $N_{\sc
201: p}=\sinh^2 r$ in either cases. After minor algebra we get
202: \begin{eqnarray}
203: p(x)&=&|\psi_{\sc s}(y)|^2 \star G(y,x,\sigma^2_{\sc p}/\tan^2\phi) \\
204: |\psi_x(y)|^2&=&\frac{1}{p(x)}\:|\psi_{\sc s}(y)|^2\:G(y,x,\sigma^2_{\sc p}/\tan^2\phi)
205: \label{sqcond}\;,
206: \end{eqnarray}
207: where $\star$ denotes convolution and $G(y;x,\sigma^2)$ a Gaussian of
208: mean $x$ and variance $\sigma^2$. \par
209: For $\sigma^2_{\sc p}/\tan^2\phi \rightarrow 0$ {\em i.e.} either for strongly
210: squeezed probe or for an almost transparent $BS_1$ we have $p(x)\rightarrow
211: |\psi_{\sc s}(x)|^2$ and $|\psi_x (y)|^2 \rightarrow \delta(y-x)$, which means
212: that we are approaching a projective VN measurement of the quadrature.
213: On the other hand, for $\sigma^2_{\sc p}/\tan^2\phi \rightarrow \infty$ (strongly
214: antisqueezed probe or an almost opaque $BS_1$) we may write $p(x)$ as
215: a very broad Gaussian, and we have $|\psi_x (y)|^2 \rightarrow |\psi_{\sc s}(y)|^2$,
216: that is we are approaching a non-informative blind quantum repeater.
217: By tuning either the probe squeezing parameter or the transmittivity of $BS_1$
218: we may also realize the whole class of intermediate QND measurements of the quadrature.
219: \par
220: For Gaussian signals the two fidelities $F$ and $G$, which measure that state
221: disturbance and the information gain respectively, may be easily evaluated in
222: terms of the single variable $x=\sigma_{\sc p}/(\sigma_{\sc s}\tan\phi)$,
223: $\sigma_{\sc s}^2$ being the variance of the signal' wave-function.
224: We have $$ F=\frac{\sqrt{2}x}{\sqrt{1+2x^2}} \quad \hbox{and} \quad
225: G=2\:\frac{\sqrt{1+x^2}}{2+x^2}\:.$$
226: Of course we have $F\rightarrow 0$ and $G\rightarrow 1$ for $x \rightarrow 0$,
227: and vice-versa for $x\rightarrow \infty$. However, in general
228: the quantity $F+G$ is not constant, and this means that by
229: varying the squeezing of the probe we obtain different trade-off between information
230: gain and state disturbance. An optimal choice of the probe, corresponding to
231: maximum information and minimum disturbance, maximizes $F+G$.
232: The maximum is achieved for $x\equiv x_{\sc m}\simeq 1.2$, corresponding
233: to fidelities $F[x_{\sc m}]\simeq 86 \%$ and $G[x_{\sc m}]\simeq 91\%$.
234: Notice that for a chosen signal, the optimization of the QND measurement can
235: be achieved by tuning the internal phase-shift of the interferometer, without the
236: need of varying the squeezing of the probe. For a nearly balanced interferometer we have
237: $\tan\phi\simeq 1$: in this case the optimal choice for the probe is a state slightly
238: anti-squeezed with respect to the signal, {\em i.e.} $\sigma_{\sc p} \simeq 1.2\:
239: \sigma_{\sc s}$. Finally, the fidelities are equal for $x\equiv x_{\sc e}\simeq 1.3$,
240: corresponding to $F[x_{\sc e}]=G[x_{\sc e}]\simeq 88 \%$.
241: For non Gaussian signals the behavior is similar though no simple analytical
242: form can be obtained for the fidelities. In this case, in order to find the
243: optimal QND measurement, one should resort to numerical means.
244: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
245: \section{Conclusions}\label{s:outro}
246: In conclusions, we have suggested a novel scheme assisted
247: by squeezing and linear feedback to realize an arbitrary QND
248: measurement of a field quadrature. Compared to previous QND
249: proposals \cite{pre} the
250: main features of our setup can be summarized as follows: i) it involves
251: only linear coupling between signal and probe, ii) only single mode
252: transformations on the conditional output are needed.
253: \par
254: The present setup permits, in principle, to achieve both a projective and a
255: fully non-destructive quantum measurement of a field quadrature. In practice,
256: however, the physical constraints on the maximum amount of energy that can be
257: impinged into the optical channels pose limitations to the precision of the
258: measurements. This agrees with the facts that both an exact repeatable
259: measurement and a perfect state preparation cannot be realized for observables
260: with continuous spectrum \cite{oza}.
261: \par
262: Compared to a vacuum probe, the squeezed/anti-squeezed
263: meters suggested in this paper provide a consistent noise reduction in the desired
264: fidelity figure already for moderate input probe energy. In addition, by
265: varying the squeezing of the probe an optimal QND measure can be achieved,
266: which provides the maximum information about the quadrature distribution of
267: the signal, while keeping the conditional output state as close as possible to
268: the incoming signal.
269: \par
270: In order to tune the setup and achieve the whole
271: class of QND measurements we have two independent parameters at disposal:
272: the probe squeezing parameter and the transmittivity of $BS_1$. This is a
273: another relevant feature of the scheme, since a too large squeezing would increase too
274: much the energy impinged into the apparatus, whereas a tuning based only on
275: the transmittivity would largely affect the detection rate.
276: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
277: \section*{Acknowledgments} This work has been cosponsored by the INFM through
278: the project PRA-2002-CLON, and by EEC through the TMR project IST-2000-29681 (ATESIT).
279: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
280: \begin{thebibliography}{99}
281: \bibitem{plz} B. Julsgaard, A. Kozhekin, E. S. Polzik, LANL ArXive quant-ph/0106057
282: \bibitem{furu} A. Furusawa et al, Science {\bf 282}, (1998) 706.
283: \bibitem{geza} L. Duan, G. Giedke, J. I. Cirac, and P. Zoller, Phys. Rev.
284: Lett. {\bf 84}, (2000), 4002.
285: \bibitem{xqnd} M. G. A. Paris, Phys. Rev. A, {\bf 65} (2002), 012110.
286: \bibitem{ks} R. E. Kass, Stat. Sci. {\bf 4} (1989), 188.
287: \bibitem{br} S. Braunstein and C. M. Caves, Phys. Rev. Lett. {\bf 72} (1994),
288: 3439.
289: \bibitem{pre} M. D. Levenson et al, Phys. Rev. Lett. {\bf 57} (1986), 2743;
290: A La Porta et al, Phys. Rev. Lett. {\bf 62} (1989), 28;
291: S. Pereira et al, Phys. Rev. Lett. {\bf 72}, (1994), 214;
292: J. Poizat, P. Grangier, Phys. Rev. Lett. {\bf 70}, (1993), 217;
293: F. X. Kartner, H. A. Haus, Phys. Rev. A {\bf 47}, 4585 (1993).
294: \bibitem{oza} M. Ozawa, Publ. RIMS Kyoto {\bf 21}, 279 (1985).
295: \end{thebibliography}
296: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
297: \end{document}
298:
299:
300:
301:
302: