LETTERS
NATUREPHYSICS
back to microscopic
(1), and characterizing it by homodynetomography
.Here, we implement state(2) and test it for the two salientfeatures of Schrödinger’s cat: macroscopicity and entanglement.First, we verify that, by changing the conditions of a microscopicmeasurementinAlice’schannelandconditioningonspecificresultsofthatmeasurement,weobtainstateswithmacroscopicallydistinctphoton number statistics in Bob’s channel. Second, we performhomodyne tomography on the undisplaced state and verify thattheentanglementhasbeenpreservedthroughthedisplacementandundisplacement operations.The principal scheme of the first part of the experiment isshown inFig. 1a,b. A heralded single photon from a parametricdownconversion setup propagates through a symmetric beamsplitter to generate the nonlocal singlephoton state. We performa phasespace quadrature measurement in Alice’s mode by meansofabalancedhomodynedetector
.Atthesametime,Bob’smodeissubjectedtophasespacedisplacementwith
α
2
∼
1
.
6
×
10
8
photons,afterwhichitsphotonnumber
N
B
ismeasured.These energy measurements exhibit macroscopic quantum fluctuationswhosestatisticsarecorrelatedwithAlice’smeasurementsof thefieldquadrature(Fig. 2).Thiscanbequalitativelyunderstoodasfollows.Alice’smeasurementofthepositionobservable
X
A
collapsestheentanglement,projectingBob’smodeontostate

ψ
B
=
1
√
2
ψ
0
(
X
A
)
ˆD
(
α
)

1
B
+
ψ
1
(
X
A
)
ˆD
(
α
)

0
B
(3)where
ψ
0
,
1
(
X
) are the wavefunctions of the zero and onephotonstates in the position basis. If
X
A
is close to zero, we have

ψ
1
(
X
A
)

ψ
0
(
X
A
), so the state in Bob’s channel is close to
ˆD
(
α
)

1
and its photon number noise variance is about
N
2
∼
3
α
2
. Onthe other hand, if Alice observes a high quadrature value
X
A
1,Bob’smodeisprojectedontoastatecloseto
ˆD
(
α
)

0
so
N
2
∼
α
2
.In this way, projecting onto different values of a microscopicobservable at Alice’s end leads to macroscopically different photonnumber statistics at Bob’s.Althoughideallytheratiobetweenthephotonnumbervariancesin these two situations equals 3, in our experiment this number isreduced to about 1.35, primarily owing to two effects. First, the observed data are influenced by the imperfection in the preparation of thesinglephotonstateandlinearlosses,whichmanifestthemselvesas an admixture of the vacuum state

0
A
⊗
0
B
to the ideal state(1)(refs17,19). In this work, the vacuum fraction is 1
−
η
=
0
.
46.Second, we measure the photon number by means of a balancedphotodetector
. Bob’s mode is incident onto the sensitive area of one of its photodiodes while the other photodiode is illuminatedby a reference laser pulse of exactly the same mean energy. Thesubtraction signal is then proportional to
N
B
−
N
R
, where
N
R
isthe number of photons in the reference pulse. This technique isnecessary because the photon number fluctuations of the displacedfield are on the scale of
α
, whereas its mean is much higher:
N
B
≈
α
2
. Subtraction of the reference pulse permits eliminationof this background along with its classical noise. As a tradeoff, itleads to addition of the shot noise
N
2R
=
α
2
to the signal, thereby reducingtheobservedratioofthephotonnumbervariances.The experimental results (Fig. 2)exhibit different behaviourdependent on the relative phase of Alice’s quadrature measurementand Bob’s displacement. If the two are the same, we observe thatnot only the variance but also the mean of the photon numberobserved in Bob’s channel is correlated with Alice’s results. On theother hand, if the phases are orthogonal, the mean photon numberis almost constant. Therefore, by choosing which quadrature tomeasure, Alice can influence the state prepared in Bob’s channel.This is a consequence of the entangled nature of state(2); similarphenomena have been observed in discrete
, continuous
and
¬5,000
×
10
8
5,0003.23.43.63.84.04.2II II IIIII IIIAlice’s quadrature measurement resultCoherent state
40,0000¬40,000 40,0000
¬40,00040,0000¬40,000
〈
N
B
¬
N
R
〉〈Δ
(
N
B
¬
N
R
)
2
〉
¬1 1 32¬3 ¬2¬1¬3 ¬2 321
abc
Figure2

Photon number statistics of the state in Bob’s channel that isconditionally prepared by Alice’s quadrature measurement.
a
,
b
, Mean (
a
)and variance (
b
) of the difference
N
B
−
N
R
between the photon numbersand the reference beam. Filled circles correspond to the displacement inBob’s channel along the same quadrature as Alice’s measurement; for opencircles the displacement and measurement are in orthogonal quadratures.The dashed line in
b
corresponds to 2
α
2
; that is, the variance that would beobserved if Bob’s channel contained a coherent state of amplitude
α
.
c
, Histograms of
N
B
−
N
R
conditioned on Alice’s measurement result withinintervals I, II, III shown in
a
,
b
by shaded areas. All histograms correspond tothe displacement and measurement in the same quadrature. Solid anddashed lines in
c
show theoretical predictions, respectively with andwithout taking experimental imperfections into account. The statisticsrepresented by histograms I and III, corresponding approximately to states1
/
√
2
ˆD
(
α
)
[
0
±
1
]
, can be distinguished by a single energy measurementwith a 68% certainty. They are reminiscent of the dead and alive states ofSchrödinger’s cat.
hybrid
systems, but not yet on a macroscopic level. In particular,this behaviour explicitly shows absence of decoherence of the twotermsin(2).Ifsuchdecoherencewerepresent,wewouldobservenodependenceonAlice’schoiceofquadratures.An interesting interpretation of our results arises if one rewritesstate (2) in the superposition basis:

D
=
12
√
2
[
(

0
+
1
)
A
⊗
ˆD
(
α
)(

0
+
1
)
B
−
(

0
−
1
)
A
⊗
ˆD
(
α
)(

0
−
1
)
B
]
2
NATURE PHYSICS

ADVANCE ONLINE PUBLICATION
