year of physics review articles

Happy centenary, photon

Anton Zeilinger
1,2
, Gregor Weihs
1
, Thomas Jennewein
2
& Markus Aspelmeyer
1
1
Institute for Experimental Physics, University of Vienna, Boltzmanngasse 5, A-1090 Vienna, Austria (e-mail: [email protected])
2
Institute for Quantum Optics and Quantum Information (IQOQI), Austrian Academy of Sciences, Boltzmanngasse 3, A-1090 Vienna, Austria
One hundred years ago Albert Einstein introduced the concept of the photon. Although in the early years
after 1905 the evidence for the quantum nature of light was not compelling, modern experiments
especially those using photon pairs have beautifully confirmed its corpuscular character. Research
on the quantum properties of light (quantum optics) triggered the evolution of the whole field of quantum
information processing, which now promises new technology, such as quantum cryptography and even
quantum computers.
O
f the papers written by Einstein in his annus
mirabilis (1905), it was not the one where he
introduced the special theory of relativity
1
, but
the one where he proposed the idea of quanta
of light
2
, later called photons
3
, that received the
acclaim of the Nobel committee. This paper is often pre-
sented as if Einstein, having analysed the photoelectric
effect, arrived at the idea of the photon. Yet, as is so often the
case, the real story is much more interesting (see Box 1).
Since 1905, the photon has come a long way, consider-
ing that it was first regarded only to be a mathematical
trick or a concept without any deeper meaning (Box 1).
But what exactly do we mean by a photon today and what
experimental evidence do we have to support the concept
of the photon?
Single photons as particles and waves
A basic meaning of the term photon is that radiation only
exists in quantized energy packets. This contrasts with semi-
classical radiation theories (see Box 2), which propose that
matter is ruled by quantum physics, while the radiation field
is classical.
One essential experiment that discriminates the quantum
theory of light from a semiclassical one uses a stream of single
photons incident on a beam splitter (Fig. 1). A semiclassical
theory predicts that the two detectors in the output beams
sometimes register in coincidence; according to this theory,
the probability of registering a count is proportional to the
square of the electric field. In contrast, full quantum theory
predicts that the two detectors never register in coincidence.
The quantum mechanically predicted statistics were experi-
mentally confirmed by Clauser in 1974 (ref. 4), who used
sources that emitted photons in pairs. Here, the registration
of one of the two photons in a trigger detector indicates that
a second, single photon is available for the experiment (Fig. 1).
In Clausers and in other early experiments the source was
an atomic cascade where two photons are emitted, one after
the other, within the lifetime of the intermediate state,
which in general is very short.
Today, the source of choice for photon paircreation is the
process of spontaneous parametric down-conversion
(SPDC), the inverse process of frequency doubling. Both
SPDC and frequency doubling are nonlinear optical
processes. Whereas in frequency doubling two photons are
converted into one photon of higher energy, in SPDC one
photon from a pump laser beam is spontaneously converted
into two photons, which emerge simultaneously
5
. Here also,
registration of one of the two photons can serve as a trigger
to indicate that the second photon has been generated. This
results in single-photon states to a good approximation,
because higher-order emission processes are negligible.
Very early on, Einstein criticized the new nature of ran-
domness in quantum physics, most unforgettably by stating:
God does not play dice. In the light of this randomness, he
also said that he would prefer to be an employee in a casino
than a physicist. How would he comment on the later finding
that one can construct random number generators on the
basis of a single photon and a beam splitter
6
, as just
described? Such a quantum random number generator
could well be used in a casino because of the high quality of
its random sequences.
Clausers experiment
4
contains the first demonstration
of sub-poissonian photon counting statistics, which can
only be understood within a quantum theory of light.
Further experiments showed other purely quantum-based
effects, such as the observation of photon antibunching in a
resonance-fluorescence experiment
7
. These early experi-
ments used beams of atoms as sources, where fluctuations in
the atom number, and thus in the emission statistics, are
unavoidable. Later, Walthers group in Munich realized
such experiments using single atoms in traps
8
.
Single-photon interference
One of the most fascinating phenomena is quantum inter-
ference with individual photons. The interference pattern is
observed by sending particles, one by one, through, say, a
double slit assembly; many particles are then collected at the
observation plane. In the simplest of such experiments, the
light intensity can be dimmed down far enough that only
one photon at a time is inside the apparatus. This was first
demonstrated by Taylor
9
. In his experiment Taylor simply
had a very dim light source together with a double slit
assembly and a photo plate inside a box. But the results of
such experiments can easily be understood semiclassically
without having to assume the existence of photons; that is,
without having to quantize the electromagnetic field as
discussed above.
A single-photon interference experiment was performed
by Grangier et al.
10
, who also used photon pairs emitted
by atomic cascades. He and his colleagues employed a
MachZehnder interferometer to observe real single-photon
interferences. Figure 2 shows the results of a single-photon
double-slit experiment
11,12
where the photon source was
parametric down-conversion. The intensities are extremely
low; nevertheless, the interference pattern accumulated
photon by photon shows perfect interference fringes.
Experiments of this kind clearly confirm that the quantum
state is not just a statistical property of an ensemble of
particles; indeed, it makes very precise predictions even for
individual particles. Following Feynman
13
, the fact that the
predictions of quantum mechanics hold for individual
particles and not just for ensembles is best illustrated by the
230 NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature
Zeilinger new 13/1/05 1:56 pm Page 230
Nature PublishingGroup 2005
year of physics review articles
The way that Einstein arrives at the photon concept in his seminal
paper ber einen die Erzeugung und Verwandlung des Lichtes
betreffenden heuristischen Gesichtspunkt (On a heuristic
aspect concerning the production and transformation of light)
2
is, contrary to widespread belief, not through the photoelectric
effect. Instead, Einstein compares the entropy of an ideal gas
filling a given volume with the entropy of radiation filling a cavity.
The logarithmic dependence on the volume of the entropy of the
gas can easily be understood by referring to the connection
between entropy and probability suggested by Boltzmann.
Because it is less probable that the gas particles will occupy
a smaller volume, such a state has a higher order, and hence
lower entropy. Interestingly, for the case of radiation filling a
cavity, Einstein merely uses the Wien black-body radiation
density, which is known to be correct only for high radiation
frequencies.
Einsteins crucial insight comes when he observes that
the entropy of light in a cavity varies in exactly the same way
with the volume of the cavity as the entropy of a gas. On the
basis of this observation, he suggests that light also consists of
particles which he calls light quanta. He clearly states that this
is only a heuristic point of view and not a logically binding
conclusion. Only in the last chapter (of eight) of the paper does
Einstein finally get to the photoelectric effect by asking where
quanta of light might have implications. He notes that it would
naturally explain why the wavelength of light emitted in photo-
luminescence is always larger than that of the absorbed light.
This is because a single particle of light is absorbed and unless
additional energy is supplied, the energy of the emitted particles of
light in general is lower.
When coming finally to the photoelectric effect Einstein
observes that the energy of the emitted electrons, as measured
by Lenard
84
, can be understood quantitatively by means of his
light quanta (see also Box 2). The only, but crucial, prediction he
makes is that the maximum energy of the electrons must vary
linearly with the frequency of the incident light. This prediction
was confirmed experimentally to high precision ten years later by
Millikan
85
. Millikan could extract from the slope of his measured
curve a value for Plancks constant h that precisely agreed with
the number found in earlier measurements of the black-body
radiation. This is one of the most convincing confirmations of
the idea of quanta. Millikan recalled: I spent ten years of my life
testing that 1905 equation of Einsteins and, contrary to my
expectations, I was compelled in 1915 to assert its unambiguous
experimental verification in spite of its unreasonableness since it
seemed to violate everything that we knew about the interference of
light
86
. Indeed, while Millikan proved the validity of Einstein's
equation beyond doubt, he categorically rejected Einstein's light-
quantum hypothesis as an interpretation of it. Only after the
discovery of the Compton effect in 1923 (ref. 87) and subsequent
experiments
88
did Millikan, like many other physicists, accept
Einstein's light-quantum hypothesis. These experiments
established the conservation of energy and momentum of individual
light quanta for the specific case of elastic scattering from electrons,
and so finally made it clear that Einsteins 1905 conception was
more than simply heuristic
89
.
However, the apparent conflict between a corpuscular theory
and interference could not be resolved before quantum mechanics
itself was fully developed. Once Schrdingers and Heisenbergs
formulations of quantum mechanics were known, it was obvious
that these should be applied to the electromagnetic cavity
oscillators and eventually to the field itself
79
. QED accommodates
both interference and quantization. Its fields are built on Maxwells
equations and populated with integral numbers of photons. QED
gives us the fundamental properties of the photon: the photon has
no rest mass, or, equivalently, moves at the vacuum speed of light.
Any finite rest mass would make the vacuum dispersive and modify
Coulombs law. Experiments put an upper limit of about 10
50
kg on
the photon mass
90
. The photon is also predicted to have zero
electric charge. The experimental upper limit is approximately
10
17
e (ref. 91).
Einsteins observations about the entropy of radiation are
intimately connected to the quantum statistics of photons, and
from Boses and Einsteins work we know that the photon is a
boson. In accordance with the vector character of the
electromagnetic field it must therefore have spin 1. Further,
equivalent to the transversality of electromagnetic radiation and
as a consequence of the the photons zero rest mass, we know
that only the spin eigenstates of +1 and 1 along its linear
momentum are allowed. These give rise to the two orthogonal
polarizations of light. Finally, in modern field theory the photon
is the exchange particle of electromagnetic interaction and it was
the first example of a gauge boson, eventually leading to the gauge
theories that form todays standard model
92
.
Box 1
A heuristic concept
A fascinating irony is that the photoelectric effect, as it was known
in Einsteins time, can be understood without having to assume that
light, or, in modern terms, the radiation field, is quantized. It suffices
to assume that the surface can only absorb or emit light in energy
quanta. More generally speaking, many phenomena thought to be
due to the quantum nature of light can actually be explained by
using a classical electromagnetic field and by assuming that only
the processes of absorption and emission are quantized. In the
simplest way, this is done by assuming that the absorbers consist of
oscillators which can absorb and emit radiation in quantized
packets only. Among the initial advocates of this semiclassical
theory, in which only the atoms are quantized while the
electromagnetic field remains as classical waves, were Planck
and Bohr themselves.
Bohr even pursued a wave-theoretic explanation of the
Compton effect within the later refuted BohrKramersSlater (BKS)
theory. It turns out, however, that semiclassical ideas cannot
account for all experimental observations. Examples are
experiments showing that there is no lower limit on the
accumulation time of radiation energy in the photoelectric effect,
which suggests an instantaneous energy transfer, such as would be
expected from a particle-like interaction
93,94
. Other examples are
correlation experiments related to quantum entanglement, which
can in principle not be modelled by local classical theories.
Curiously, Einstein together with Podolsky and Rosen first
discussed such correlations in 1935 (ref. 22) for completely different
reasons and apparently without perceiving that this famous paper
would eventually deliver another independent proof of the photon
hypothesis. A detailed discussion of the quantum versus the
semiclassical approach in the light of quantum non-locality has
been given by Clauser
95
. The interested reader is referred to the
accounts of Klein
96
, Stachel
97
, Pais
98
or Clauser
99
.
Box 2
Semiclassical radiation theories are a dead end
NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature 231
Zeilinger new 13/1/05 1:56 pm Page 231
Nature PublishingGroup 2005
finding that each individual photon knows it should never end up in
the minimum of an interference fringe.
We emphasize that the conceptual questions arising for photon
interference are the same as those arising for interference of massive
particles. In both cases we see particle-like and wave-like properties.
An inequivalence arises for certain interference experiments
14
because the photon has no rest mass.
Two-photon interference
An interesting consequence of the bosonic character of photons is
their bunching behaviour. This is seen most directly when two
photons one from each input port are incident on a beam splitter
(Fig. 3). If the two photons do not arrive simultaneously, each has a
50% chance of going either way after the beam splitter, independently
of the other photon. This results in the coincidences shown. But if the
photons arrive simultaneously, they become indistinguishable and
end up together randomly in either beam. In the experiment the rate
of coincident photon detections at the beam splitter outputs is
monitored. The resulting dip in the coincidence rate is called the
HongOuMandel dip
15
.
What happens is a quantum interfer-
ence effect. The only way that one photon
can arrive at each detector is if both pho-
tons are either reflected or transmitted.
The detection probability in quantum
physics is given by the square of the
probability amplitude, which is different
from squaring the actual electromag-
netic field. Curiously, the probability
amplitudes for these two possibilities
destructively interfere with each other.
This results from the well-known phase
jump of 90 degrees that each photon
experiences upon reflection. This implies
a total phase of 180 degrees of the state
|both photons reflected relative to the
state |both photons transmitted.
Fermions would behave differently
because their quantum state is antisym-
metric, as reflected by a negative sign in
their initial state. In this case the two
amplitudes introduced above interfere
constructively and the two particles are
always found in separate outputs. Inter-
estingly, this fermionic behaviour can
also be observed for two photons if the
photons are prepared in an antisym-
metric state with respect to their spin
(Fig. 3). This latter observation turned
out to be crucial for many quantum
information applications, specifically
for quantum dense coding
16
.
Complementarity, information
and quantum physics
Complementarity, the mutually exclusive
nature of the wave and particle concepts,
has led to intense discussions. Of these,
the early ones between Einstein and
Bohr raised the key issues. Whereas
Einstein thought that it should be possible
to observe an interference pattern and at
the same time know for each photon
which slit it went through, Bohr was
always able to show that an apparatus
capable of determining the particles
path was by necessity constructed such
year of physics review articles
Figure 1 Principle of Clausers experiment with correlated pairs of photons
(simplified). The source emits two photons. Registration of a photon on the left
detector provides the information that one and only one photon at the right side
encounters a 50/50 beam splitter where it is either reflected or transmitted.The fact
that only one of the two detectors behind the beam splitter registers and never both
can easily be understood using the photon concept, and is in clear conflict with
semiclassical theories of radiation
4
.
Figure 2 Single-photon double-slit interference. A pair of momentum-entangled photons is created by type-I parametric
down-conversion. Photon 2 enters a double-slit assembly and photon 1 is registered by a detector D1 placed at distance f
in the focal plane of the lens. This projects the state of photon 2 into a momentum eigenstate which cannot reveal any
positional information and, hence supplies no information about slit passage. Therefore, in coincidence with a registration
of photon 1 in the focal plane, photon 2 exhibits the interference pattern shown. On the other hand, when the detector is
placed in the imaging plane, it does reveal the path photon 2 takes through the slit assembly, which therefore does not
show the interference pattern. The observed count rate of at most two photons per second implies that the average spatial
distance between photons registered would be of the order of 100,000 km or more. Therefore, most of the time the
apparatus is empty (from refs 11 and 12). The error bars (s.d.) show the statistical errors of photon counting.
8,000 6,000 4,000 2,000 0 2,000 4,000 6,000 8,000
R
e
g
i
s
t
e
r
e
d

c
o
i
n
c
i
d
e
n
t

p
a
i
r
s

i
n

6
0

s
Visibility
97.22%
Position of the scanning slit in front of detector D2 (m)
Lens
Detector
Scanning
slit
Double slit
Coincidence
logic
Crystal
Pump
f
D2
0
20
40
60
80
100
120
140
D1
Photon 1
Photon 2
232 NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature
Zeilinger new 13/1/05 1:56 pm Page 232
Nature PublishingGroup 2005
that no interference pattern could arise and vice versa
17
. Today, it is
thought that a perfect interference pattern arises only when there is
no possible way of finding out which path the particle took. Evidently,
intermediate cases are also possible, of partial path information
together with non-perfect interference fringes
18
. In so-called delayed
choice experiments
19,20
the decision of whether to observe path
information can be delayed to when the particle is already inside the
interferometer setup and even until after it has been registered. This
again supports the view that the quantum state may be seen as a
representation of the information about the probabilities of possible
measurement results, which may include mutually exclusive, that is,
complementary ones.
An experiment supporting the information aspect of quantum
interference
21
has been performed by Mandels group at Rochester as
part of a series of ground-breaking experiments on the quantum
nature of light. In their experiment (Fig. 4), they used the emission of
one photon pair from two down-conversion crystals. One photon
passed a modified MachZehnder interferometer, and what happened
to the other photon decided whether the first photon showed inter-
ference or not. Thus, the still widespread view that the act of deter-
mining the path taken by the particle disturbs its state enough to
destroy the interference is untenable. The key factor is whether path
information is available: it does not matter if someone takes care to
read it out or not. Indeed, there have even been experiments where
the path information carried by the second particle is destroyed after
the particle passing through the double slit has already been regis-
tered. Here, the interference pattern is still observed. The double-slit
diffraction pattern shown in Fig. 2 was obtained in this way.
When analysing quantum interference we can fall into all kinds of
traps. The general conceptual problem is that we tend to reify to
take too realistically concepts like wave and particle. Indeed if we
consider the quantum state representing the wave simply as a calcula-
tional tool, problems do not arise. In this case, we should not talk
about a wave propagating through the double-slit setup or through a
MachZehnder interferometer; the quantum state is simply a tool to
calculate probabilities. Probabilities of the photon being somewhere?
No, we should be even more cautious and only talk about probabilities
of a photon detector firing if it is placed somewhere. One might be
tempted, as was Einstein
2
, to consider the photon as being localized at
some place with us just not knowing that place. But, whenever we talk
about a particle, or more specifically a photon, we should only mean
that which a click in the detector refers to.
Nonlocality, Bell and GHZ
Einstein did not only criticize quantum mechanics for the new role of
randomness mentioned above. His criticism went much further in
his insistence on the existence of a real factual world and on the role of
physics to describe that reality. This criticism forms the basis of his
famous article published in 1935, together with Podolsky and Rosen.
The EinsteinPodolskyRosen (EPR) paper
22
makes use of correla-
tions shown by entangled quantum states. Bell discovered in 1964
(ref. 23) that the predictions of quantum physics for these correla-
tions are at variance with a local realistic world view. Such a classically
intuitive view holds that the outcome of a measurement on a physical
system is determined by physical properties of the system prior to and
independent of the measurement (realism), and that the outcome
cannot depend on any actions in space-like separated regions (Einstein
locality). The quantum correlations
are too strong to be reproduced by any
such model. After the initial experi-
ments by Freedman and Clauser
24
, and
the much refined experiments by
Aspect
25
(Fig. 5) confirming the quan-
tum predictions, two loopholes
remained open. Thus, a local realistic
viewpoint remained at least logically
possible. The first loophole, the so-
called communication loophole, used
the fact that even before the two pho-
tons are registered, the apparatus
settings could be communicated to
detectors and/or to the source. A
beautiful experiment to close this
loophole using periodically switched
time-dependent polarizer settings was
performed by Aspect in 1982 (ref. 26).
The loophole was more decisively
ruled out by an experiment of Weihs et
al.
27
where the measurement settings
year of physics review articles
Figure 3 Bunching (left) or antibunching (right) behaviour of photon pairs. One
photon each is incident from each input of a 50/50 beam splitter. Coincidences
between detectors in the two output beams are registered as a function of the flight
time difference of the incident photons. No coincidences are observed for zero flight
time difference (left) for the usual symmetric spatial state of the two photons. This is
because the probability amplitudes for the transmission of both photons and for the
reflection of both photons destructively interfere: the latter one picks up a minus
sign owing to the phase shift of the photons upon reflection. Interestingly, the two
incident photons can also be in an antisymmetric spatial state (which occurs if the
two-photon spin state is also antisymmetric). In this case, the two amplitudes
interfere constructively. This results in the two photons always exiting in separate
beams for zero flight time difference. The observed coincidence peak (right)
confirms this expected antibunching.

Flight time difference (ps)

C
o
i
n
c
i
d
e
n
c
e
s

i
n

1
0

s
1.0 0.5 0.0 0.5 1.0
Flight time difference (ps)
1.0 0.5 0.0 0.5 1.0
10,000
7,500
5,000
2,500
0
+
Figure 4 A mind-boggling interference experiment (from ref. 21). Two nonlinear crystals NL1 and NL2 are pumped by the
same laser and produce one pair of photons in the superposition of being in beams s
1
and i
1
, or in beams s
2
and i
2
. After
transmission through NL2, the photons from i
1
are indistinguishable from photons in i
2
and thus the original distinguishability
of the path of s
1
and s
2
disappears, and quantum interference is observed for pairs of photons that are detected in
coincidence between detectors D
i
and D
s
. Insertion of an absorber (neutral density filter, NDF) can be used gradually to
introduce distinguishability and thus to make the interference disappear. Note that no disturbance whatsoever acts on the
interfering photon in beams s
1
and s
2
.
From argon laser
NL2
NDF
D
s
D
i
s
1
s
2
i
1
i
2
Counter
Counter
Coincidence
NL1
NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature 233
Zeilinger new 13/1/05 1:56 pm Page 233
Nature PublishingGroup 2005
were changed randomly and fast with respect to the distance between
the stations. This left only the so-called detection loophole open,
which implied that although all photon pairs would obey local realism
and hence be at variance with quantum physics, the small detected
subset confirmed quantum mechanics. Since the detection efficiencies
were far from ideal, such reasoning could not be ruled out. This latter
loophole was closed in an experiment on entanglement between two
ions in a cavity
28
, in which it was possible to detect nearly all entangled
pairs. So, although both remaining loopholes have now been closed
in separate experiments, the logical possibility exists that nature
tricks us and makes use of different loopholes in different experi-
ments. Although no one reasonably assumes nature to be so capricious,
a future experiment that closes both loopholes together would still
be interesting.
The conflict of local realism with quantum mechanics, first
exposed by Bell for entangled pairs, is even more striking for
three or more entangled particles. For the so-called GHZ (Green-
bergerHorneZeilinger) states
29,30
, situations exist for which a
local realist and a quantum mechanic make completely opposite
predictions, even for individual measurement results on one pho-
ton. Needless to say, experiments
31
have confirmed the quantum
prediction (Fig. 6).
Photons, atoms and beyond
As mentioned above, the initial question of whether it is only matter
or also radiation that is quantized has finally been settled in favour of
the photon. It has now become possible to investigate the interaction
between photons and atoms in great detail. For example, Kimbles
group at Caltech
32
showed that it was possible to observe the phase
shift experienced by an atom while it interacted with a field of, on
average, less than one photon. Haroche and his group at the cole
Normale in Paris
33
were able to construct entangled states between
single photons trapped in a high-finesse cavity and atoms passing
through (Fig. 7). Such experiments have also been used to demon-
strate several interesting aspects, such as time-resolved quantum
interference phenomena
34
, trapping of atoms with single
photons
35,36
or quantum non-demolition measurements, in which
the presence of single photons can be determined without
destroying the photon
37
.
Quantum information processing with photons
In the emerging field of quantum information technology the two
basic subfields are quantum communication and quantum compu-
tation. The photon has been put to work in recent years, particularly
in new concepts of communication. In quantum cryptography the
complementarity of different measurements on a quantum system
3840
is used to establish a secure key between two partners. In quantum
teleportation
41,42
it is possible to transfer the quantum state of inde-
pendent particles from one system to another by employing entangled
states as a quantum information channel.
When turning to future challenges and developments, it is likely
that the photon will have a significant role in quantum communica-
tion. The currently most advanced source for entanglement, an
important resource for quantum information processing, is photon
year of physics review articles
234 NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature
Figure 5 Bell tests violating local realistic predictions. a, Setup of the experimental
test of Bells inequality with correlated photon pairs produced by atomic cascade
relaxation (from ref. 25). The two photons
1
and
2
are analysed in separate
polarizers (I in direction a and II in direction b), and detected with photomultipliers
(PM). A dedicated electronic circuit is used to determine the coincidences, which
are included in evaluating the correlation terms in Bells inequality. This and several
other similar experiments proved with high significance that real particles follow the
prediction of quantum mechanics, and violate the limits imposed by local realistic
theories. b, Correlation curves (from ref. 27) in a Bell inequality experiment. A
characteristic of such experiments is that the correlation depends on the difference
angle of the analysers only and not on any of their absolute values.
R
e
g
i
s
t
e
r
e
d

c
o
i
n
c
i
d
e
n
t

p
h
o
t
o
n

p
a
i
r
s

i
n

5

s
1,000
800
600
0,50 0,25 0,00 0,25 0,50
400
200
0
PM
I (a)
Singles Singles
Coincidences
PM
P
M
P
M
S

1

2
II (b)
a
b
Angle of polarizer I
Polarizer II at /2
Polarizer II at 3/4
Figure 6 Sketch of an experimental test of the GHZ argument against local realistic
theories (from ref. 31). A source (S) emits polarization entangled photon triplets in a
GHZ state (bottom left). Measurements in three different combinations of circular
and diagonal polarization (not shown) show perfect correlation. These correlations
define local realistic elements of reality, which lead to definite predictions (a) for a
measurement in all diagonal polarization bases. The quantum predictions (b) are
exactly opposite. c, The experimental relative frequencies of triple coincidences
behind polarizers 1, 2 and 3. For example, for photons 2 and 3, which are both
horizontally (H) polarized, local realism (a) predicts the first photon to be vertically (V)
polarized, quantum mechanics (b) predicts it to be H polarized, and experiment (c)
confirms the predictions of quantum mechanics within experimental accuracy.
S
H V
H
V
V
H
1
2 3
VHH HVH HHV VVV
Local realistic prediction
HHH
VHH
VVH
HVH
VHV
HHV
HVV
VVV
Experiment
HHH VVH VHV HVV
Quantum prediction
F
r
a
c
t
i
o
n
F
r
a
c
t
i
o
n
F
r
a
c
t
i
o
n
0.25
0.20
0.15
0.10
0.05
0.00
0.25
0.20
0.15
0.10
0.05
0.00
a
0.25
0.20
0.15
0.10
0.05
0.00
c
b
Zeilinger new 13/1/05 1:56 pm Page 234
Nature PublishingGroup 2005
pairs generated in SPDC (ref. 43). Moreover, only with photons is it
possible to cover large distances outside the protected environment
of the laboratory. There have been experiments transporting entangled
states over more than 10 km using glass fibres
44
and across the river
Danube in free space
45
. Quantum cryptography, with faint laser
pulses containing less than one photon on average, has been tested in
free space for distances of more than 20 km (ref. 46) even in day-
light
47
and in optical fibres for a physical separation of 67 km (ref.
48). Furthermore, quantum communication through satellites is the
only possible way to cover global distances. Satellite-based quantum
communication may very well be realized within the next decade.
In terms of technical applications of the photon idea, the most
advanced is quantum cryptography
49,50
(Fig. 8). Prototype devices are
already on the market and the development of systems that are suitable
for the security industry is well under way. Quantum teleportation
might one day provide useful communication links between yet-to-
be-developed quantum computers. Initial tests of long-distance
quantum teleportation have recently been performed in Geneva
51
and in Vienna
52
. The latter was a real field test that even included
active feed-forward of measurement results (Fig. 9). An important
extension of teleportation in this sense will be quantum repeaters
53
, a
combination of entanglement swapping (that is, the teleportation of
an entangled state)
54,55
and local atomic memories of quantum infor-
mation, which also exploit the atomphoton interface
56,57
. The
development of these applications is intimately connected to the
development of the quantum computer itself.
Although for quantum communication the obvious choice is
photons, for quantum computation, implementations in localized
systems like atoms, ions or solid-state devices seem to be preferable.
Yet, surprisingly, even for implementing quantum computation
algorithms, photons offer interesting possibilities, despite the con-
siderable difficulty of storing them for a long time. After the discovery
that some gates could be realized
through teleportation
58
, an impor-
tant breakthrough was the sugges-
tion by Knill et al.
59
that even with
linear optical elements, universal
quantum computation could be
realized. Following these sugges-
tions, various quantum compu-
tation primitives have been
demonstrated with photons alone
including conditional phase shift
operations
60
, and destructive
61,62
and
even non-destructive controlled
NOT (CNOT) gates
63
. All these
schemes use unentangled states as
inputs on which the quantum gates
operate. A new and probably more
practical approach is the concept of a
one-way quantum computer
64
that
realizes universal quantum compu-
tation in a way that is totally different
from that used by existing quantum
computing schemes. Here, the idea
is to start with a general, highly
entangled multi-qubit state. The
computation is then performed by
applying a sequence of simple one-
particle measurements, specific to
the algorithm implemented. This
new approach uses highly entangled
cluster states, which recently have
been realized with photons and
applied to demonstrate elementary
quantum gates
65
. Entangled multi-
particle states also have a significant
role in other new protocols of quantum information. For example,
quantum error correction is based on such states
66, 67
.
Many laboratories all over the world are working towards devel-
oping many different physical implementations both of quantum
communication devices and of quantum computers, and it will be
interesting to see which technology will be the best. Yet, we are
convinced that some day in the future, the present classical informa-
tion technology will be replaced by a quantum one, even if this is only
because of the continuing miniaturization of switching elements in
computer chips.
For future technological developments, new sources for single-
photon states will be needed. The most basic of such sources would be
a single-photon source that, on demand, produces one, and only one,
photon at a specific time and not at random. There has been impor-
tant progress over the past few years in this field from various
directions, including atoms in cavities
68,69
and solid-state devices
such as cavity-coupled quantum dots
70
. An extensive account of such
activities has recently been collected by Grangier et al.
71
. More
generally, it would be good to have sources that produce any specific
multiphoton state, even entangled ones, on demand. Promising
experiments along this line have been performed in the context of
cavity quantum electrodynamics (QED; ref. 72). Also, more efficient
photon detectors are needed that operate over a broader wavelength
range than those currently available. A particularly interesting devel-
opment would be a detector that is able to discriminate clearly
between 1, 2, 3 Nphotons. Initial results discriminating between
one and two photons have been reported, for example, by
Yamamotos group
73
, by using a visible-light photon counter
(VLPC). In the far future, a detector that identifies deterministically
an arbitrary N-photon state would be useful.
We have been able to give only a glimpse of the vast expanse of
implications and applications of the photon concept, and of quantum
year of physics review articles
NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature 235
Figure 7 Photonatom interaction. a, The strong coupling of light modes in a cavity, through which excited atoms pass, is used
to perform a controlled quantum phase gate between the state of an atom and a photon (i, e and g are the internal states of the
atom where e and g are coupled by means of the cavity photon). This operation is required to create entanglement between the
atom and the photon (from ref. 82). b, Caesium (Cs) atoms passing through a high finesse optical cavity formed by two mirrors,
M
1
and M
2
, lead to a phase shift on a probe beam. The experimental data clearly show that even with a mean photon number
below one, the probe beam will pick up a significant phase shift (from ref. 32).
Probe
a
M
1
M
2
Pump
b
Optical pumping
Cs beam
Heterodyne
Intracavity photon number m
b
P
r
o
b
e

p
h
a
s
e

s
h
i
f
t

|


a
|
Cavity
Detector
14
12
10
8
6
4
2
10
4
10
3
10
2
10
1
1 10
T
a
m
a
/2
Local
oscillator

,g

+
,g
+
e
g
i
a
b
Atom
source
Cavity
Zeilinger new 13/1/05 1:56 pm Page 235
Nature PublishingGroup 2005
optics in general. Also deserving a mention is the wide field of experi-
ments on squeezed states initialized by Slusher et al.
74
and rapidly
expanded by others. These include continuous-variable demonstra-
tions of quantum teleportation
75
, of quantum optics in phase
space
76,77
, and of quantum cryptography
78
. Here, phenomena are
studied that are a consequence of the quantization of the electromag-
netic field
79
. But in general, the concept of the photon as an individual
particle is less important here.
Conclusion
Evidently, Einsteins 1905 proposal of the photon
concept has had tremendous impact. But Einstein
should also be highly credited for his various criticisms
of quantum physics that were part of the early debate
with his contemporaries (including Bohr). They trig-
gered a body of both theory and experiments concerned
with individual quantum systems. In this context,
experiments with photons have had a pioneering role.
Although such experiments now rule out Einsteins
point of view, they gave rise to the new fields of
quantum information processing. But the conceptual
problems are not fully settled. This is signified by the
wide spectrum of different interpretations of quantum
physics that compete with each other. In our view, a
common trait of many interpretations is that entities
are taken to be real beyond necessity. This is most
obvious for the case of the many-worlds interpreta-
tion
80
where the coexistence of parallel worlds is
year of physics review articles
236 NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature
Figure 8 Quantum cryptography in practice. Entangled pairs of photons are used to
generate secure keys at Alices and Bobs distant stations. The local measurement
outcomes at Alice are completely random but correlated to the results at Bob. Thus,
a cryptographic key is created whose security rests only on the principles of quantum
physics. a, The secure transmission of an image of the Venus of Willendorf over
350 m (ref. 83). b, Experimental setup of a prototype system for entangled photon
quantum key distribution (QKD), which has even been used in a real world
demonstration to securely transfer money into a bank account (from ref. 50).
Bitwise
XOR
Bitwise
XOR
Sync. laser
1.45 km
Entangled state source Detection module Detection module
QKD
electronics
QKD
electronics
Alice Bob
Ethernet bridge Ethernet bridge
Ethernet
port
Ethernet
port
Ethernet
port
Ethernet
port
Original Encrypted Decrypted
Alice's key Bob's key
QKD
client
QKD
client
a
b
Fibre
Fibre
Fibre
Figure 9 Schematic setup of quantum state teleportation across the river Danube (from ref. 52). The
polarization state of a photon is transferred from Alice to Bob through the use of a quantum channel
(entangled photons) and a classical channel (microwave pulses). Bob was able to regain the original photon
state with a fidelity as high as 90%, which is clearly above the limits imposed by classical concepts.
Alice logic
Classical channel

+
/

+
Alice
Input
UV-pulse
Trigger
Source
Quantum channel
River Danube
F
600 m
Bob
Bob
logic
R
F
-
u
n
i
t
R
F
-
u
n
i
t
Zeilinger new 13/1/05 1:56 pm Page 236
Nature PublishingGroup 2005
claimed without compelling evidence, but it also holds, for example,
for the Bohm interpretation
81
where, again without compelling
evidence, each particle is given a well-defined position and
momentum at any time. We suggest that these are simply attempts
to keep, in one way or other, a realistic view of the world. It may well
be that in the future, quantum physics will be superseded by a new
theory, but it is likely that this will be much more radical than anything
we have today.
doi:10.1038/nature03280
1. Einstein, A. Zur Elektrodynamik bewegter Krper. Ann. Phys. 17, 891921 (1905).
2. Einstein, A. ber einen die Erzeugung und Verwandlung des Lichtes betreffenden heuristischen
Gesichtspunkt. Ann. Phys. 17, 132148 (1905).
3. Lewis, G. N. The conservation of photons. Nature 118, 874875 (1926).
4. Clauser, J. F. Experimental distinction between the classical and quantum field-theoretic predictions
for the photoelectric effect. Phys. Rev. D9, 853860 (1974).
5. Burnham, D. C. & Weinberg, D. L. Observation of simultaneity in parametric production of optical
photon pairs. Phys. Rev. Lett. 25, 8487 (1970).
6. Jennewein, T., Achleitner, U., Weihs, G., Weinfurter, H. & Zeilinger, A. A fast and compact quantum
random number generator. Rev. Sci. Instrum. 71, 16751680 (2000).
7. Kimble, H. J., Dagenais, M. & Mandel, L. Photon antibunching in resonance fluorescence. Phys. Rev.
Lett. 39, 691695 (1976).
8. Diedrich, F. & Walther, H. Nonclassical radiation of a single stored ion. Phys. Rev. Lett. 58, 203206
(1987).
9. Taylor, G. I. Interference fringes with feeble light. Proc. Camb. Phil. Soc. Math. Phys. Sci. 15, 114115
(1909).
10. Grangier, P., Roger, G. & Aspect, A. Experimental evidence for a photon anticorrelation effect on a
beam splitter: a new light on single-photon interference. Europhys. Lett. 1, 173179 (1986).
11. Dopfer, B. Zwei Experimente zur Interferenz von Zwei-Photon Zustnden: Ein Heisenbergmikroskop
und Pendellsung. PhD thesis, Universtt Innsbruck (1998).
12. Zeilinger, A. Experiment and the foundations of quantum physics. Rev. Mod. Phys. 71, S288S297
(2000).
13. Feynman, R. P., Leighton, R. B. & Sands, M. L. The Feynman Lectures on Physics Vol. 3 (Addison-
Wesley, Massachusetts, 1965).
14. Brukner, C. & Zeilinger, A. Nonequivalence between stationary matter wave optics and stationary
light optics. Phys. Rev. Lett. 79, 25992603 (1997).
15. Hong, C. K., Ou, Z. Y. & Mandel, L. Measurement of subpicosecond time intervals between two
photons by interference. Phys. Rev. Lett. 59, 20442046 (1987).
16. Mattle, K., Weinfurter, H., Kwiat, P. G. & Zeilinger, A. Dense coding in experimental quantum
communication. Phys. Rev. Lett. 76, 46564659 (1996).
17. Bohr, N. in Albert Einstein: Philosopher-Scientist VII (ed. Schilipp, P. A.) Vol. 7, 201241 (Library of
Living Philosophers, Evanston, 1949).
18. Englert, B.-G. Fringe visibility and which-way information: An inequality. Phys. Rev. Lett. 77,
21542157 (1997).
19. Wheeler, J. A. in Problems in the Foundations of Physics . Proc. Intl. School of Physics Enrico Fermi,
Course LXXII. (ed. di Francia, G. T.) 395492 (North-Hollands, Amsterdam, 1979).
20. Scully, M. O. & Drhl, K. Quantum eraser: A proposed photon correlation experiment concerning
observation and delayed choice in quantum mechanics. Phys. Rev. A25, 22082213 (1982).
21. Zou, X. Y., Wang, L. J. & Mandel, L. Induced coherence and indistinguishability in optical
interference. Phys. Rev. Lett. 67, 318321 (1991).
22. Einstein, A., Podolsky, B. & Rosen, N. Can quantum mechanical description of physical reality be
considered complete? Phys. Rev. 47, 777780 (1935).
23. Bell, J. S. On the Einstein Podolsky Rosen paradox. Physics 1, 195200 (1964).
24. Freedman, S. J. & Clauser, J. F. Experimental test of local hidden-variable theories. Phys. Rev. Lett. 28,
938941 (1972).
25. Aspect, A., Grangier, P. & Roger, G. Experimental realization of Einstein-Podolsky-Rosen-Bohm
Gedankenexperiment: A new violation of Bells inequalities. Phys. Rev. Lett. 49, 9194 (1982).
26. Aspect, A., Dalibard, J. & Roger, G. Experimental test of Bells inequalities using time-varying
analyzers. Phys. Rev. Lett. 49, 18041807 (1982).
27. Weihs, G., Jennewein, T., Simon, C., Weinfurter, H. & Zeilinger, A. Violation of Bells inequality under
strict Einstein locality conditions. Phys. Rev. Lett. 81, 50395043 (1998).
28. Rowe, M. A. et al. Experimental violation of a Bells inequality with efficient detection. Nature 409,
791794 (2001).
29. Greenberger, D. M., Horne, M. A. & Zeilinger, A. in Bells Theorem, Quantum Theory, and Conceptions
of the Universe (ed. Kafatos, M.) 7376 (Kluwer, Dordrecht, 1989).
30. Greenberger, D. M., Horne, M. A., Shimony, A. & Zeilinger, A. Bells theorem without inequalities.
Am. J. Phys. 58, 11311143 (1990).
31. Pan, J.-W., Bouwmeester, D., Daniell, M., Weinfurter, H. & Zeilinger, A. Experimental test of
quantum nonlocality in three-photon GreenbergerHorneZeilinger entanglement. Nature 403,
515518 (2000).
32. Turchette, Q. A., Hood, C. J., Lange, W., Mabuchi, H. & Kimble, H. J. Measurement of conditional
phase shifts for quantum logic. Phys. Rev. Lett. 75, 47104713 (1995).
33. Rauschenbeutel, A. et al. Step-by-step engineered multiparticle entanglement. Science 288,
20242028 (2000).
34. Legero, T., Wilk, T., Kuhn, A. & Rempe, G. Time-resolved two-photon quantum interference. Appl.
Phys. B 77, 797802 (2003).
35. Pinkse, P., Fischer, T., Maunz, P. & Rempe, G. Trapping an atom with single photons. Nature 404,
365368 (2000).
36. Hood, C. J., Lynn, T. W., Doherty, A. C., Parkins, A. S. & Kimble, H. J. The atom-cavity microscope:
Single atoms bound in orbit by single photons. Science 287, 14471453 (2000).
37. Nogues, G. et al. Seeing a single photon without destroying it. Nature 400, 239242 (1999).
38. Wiesner, S. Conjugate coding. SIGACT News 15, 7888 (1983).
39. Bennett, C. H. & Brassard, G. in Proc. IEEE Intl.Conf. on Computers, Systems and Signal Processing
175179 (IEEE, Bangalore, 1984).
40. Ekert, A. K. Quantum cryptography based on Bells theorem. Phys. Rev. Lett. 67, 661663 (1991).
41. Bennett, C. H. et al. Teleporting an unknown quantum state via dual classical and Einstein-Podolsky-
Rosen channels. Phys. Rev. Lett. 70, 18951899 (1993).
42. Bouwmeester, D. et al. Experimental quantum teleportation. Nature 390, 575579 (1997).
43. Kwiat, P. G. et al. New high-intensity source of polarization-entangled photon pairs. Phys. Rev. Lett.
75, 43374341 (1995).
44. Tittel, W., Brendel, J., Zbinden, H. & Gisin, N. Violation of Bell inequalities by photons more than
10 km apart. Phys. Rev. Lett. 81, 35633566 (1998).
45. Aspelmeyer, M. et al. Long-distance free-space distribution of quantum entanglement. Science 301,
621623 (2003).
46. Kurtsiefer, C. et al. A step towards global key distribution. Nature 419, 450 (2002).
47. Hughes, R. J., Nordholt, J. E., Derkacs, D. & Peterson, C. G. Practical free-space quantum key
distribution over 10 km in daylight and at night. New J. Phys. 4 43, 143 (2002).
48. Stucki, D., Gisin, N., Guinnard, O., Ribordy, G. & Zbinden, H. Quantum key distribution over 67 km
with a plug&play system. New J. Phys. 4 41, 141 (2002).
49. Gisin, N., Ribordy, G., Tittel, W. & Zbinden, H. Quantum cryptography. Rev. Mod. Phys. 74, 145195
(2002).
50. Poppe, A. et al. Practical quantum key distribution with polarization entangled photons. Opt. Express
12, 38653871 (2004).
51. Marcikic, I., de Riedmatten, H., Tittel, W., Zbinden, H. & Gisin, N. Long-distance teleportation of
qubits at telecommunication wavelengths. Nature 421, 509513 (2003).
52. Ursin, R. et al. Quantum teleportation across the Danube. Nature 430, 849 (2004).
53. Briegel, H.-J., Dr, W., Cirac, J. I. & Zoller, P. Quantum repeaters: The role of imperfect local
operations in quantum communication. Phys. Rev. Lett. 81, 59325935 (1998).
54. Z

ukowski, M., Zeilinger, A., Horne, M. A. & Ekert, A. K. Event-ready-detectors Bell experiment via
entanglement swapping. Phys. Rev. Lett. 71, 42874290 (1993).
55. Jennewein, T., Weihs, G., Pan, J.-W. & Zeilinger, A. Experimental nonlocality proof of quantum
teleportation and entanglement swapping. Phys. Rev. Lett. 88, 017903 (2002).
56. Blinov, B. B., Moehring, D. L., Duan, L.-M. & Monroe, C. Observation of entanglement between a
single trapped atom and a single photon. Nature 428, 153157 (2004).
57. Julsgaard, B., Sherson, J., Cirac, J. I., Fiurasek, J. & Polzik, E. S. Experimental demonstration of
quantum memory for light. Nature 432, 482486 (2004).
58. Gottesman, D. & Chuang, I. L. Demonstrating the viability of universal quantum computation using
teleportation and single-qubit operations. Nature 402, 390393 (1999).
59. Knill, E., Laflamme, R. & Milburn, G. A scheme for efficient quantum computation with linear
optics. Nature 409, 4652 (2000).
60. Sanaka, K., Jennewein, T., Pan, J.-W., Resch, K. & Zeilinger, A. Experimental nonlinear sign shift for
linear optics quantum computation. Phys. Rev. Lett. 92, 017902 (2004).
61. Pittman, T. B., Fitch, M. J., Jacobs, B. C. & Franson, J. D. Experimental controlled-NOT logic gate for
single photons in the coincidence basis. Phys. Rev. A68, 032316 (2003).
62. OBrien, J. L., Pryde, G. J., White, A. G., Ralph, T. C. & Branning, D. Demonstration of an all-optical
quantum controlled-NOT gate. Nature 426, 264267 (2003).
63. Gasparoni, S., Pan, J.-W., Walther, P., Rudolph, T. & Zeilinger, A. Realization of a photonic controlled-
NOT gate sufficient for quantum computation. Phys. Rev. Lett. 93, 020504 (2004).
64. Raussendorf, R. & Briegel, H. J. A one-way quantum computer. Phys. Rev. Lett. 86, 51885191 (2001).
65. Walther, P. et al. Experimental one-way quantum computing. Nature (in the press).
66. Steane, A. Multiple particle interference and quantum error correction. Proc. R. Soc. Lond. A452,
25512577 (1995).
67. Shor, P. in Proc. 37th Symp. on Foundations of Computer Science 1565 (IEEE Computer Society Press,
Los Alamitos, 1996).
68. Kuhn, A., Hennrich, M. & Rempe, G. Deterministic single-photon source for distributed quantum
networking. Phys. Rev. Lett. 89, 679011679014 (2002).
69. McKeever, J. et al. Deterministic generation of single photons from one atom trapped in a cavity.
Science 303, 19921994 (2002).
70. Santori, C., Fattal, D., Vuckovic, J., Solomon, G. S. & Yamamoto, Y. Indistinguishable photons from a
single-photon device. Nature 419, 594597 (2002).
71. Grangier, P., Sanders, B. & Vuckovic, J. (eds) Focus on Single Photons on Demand New J. Phys. (Spec.
Edn) 6, 85100 (2004).
72. Brattke, S., Varcoe, B. T. H. & Walther, H. Generation of photon number states on demand via cavity
quantum electrodynamics. Phys. Rev. Lett. 86, 35343537 (1999).
73. Kim, J., Takeuchi, S., Yamamoto, Y. & Hogue, H. H. Multiphoton detection using visible light photon
counter. Appl. Phys. Lett. 74, 902904 (1999).
74. Slusher, R. E., Hollberg, L. W., Yurke, B., Mertz, J. C. & Valley, J. F. Observation of squeezed states
generated by four-wave mixing in an optical cavity. Phys. Rev. Lett. 55, 24092412 (1985).
75. Furusawa, A. et al. Unconditional quantum teleportation. Science 282, 706709 (1998).
76. Schleich, W. P. Quantum Optics in Phase Space (Wiley-VCH, Berlin, 2001).
77. Bachor, H.-A. & Ralph, T. C. A Guide to Experiments in Quantum Optics 2nd edn (Wiley-VCH,
Weinheim, 2004).
78. Grosshans, F., Assche, G. V., Wenger, J., Cerf, R. B. J. & Grangier, P. Quantum key distribution using
gaussian-modulated coherent states. Nature 421, 238241 (2003).
79. Dirac, P. A. M. Emission and absorption of radiation. Proc. R. Soc. Lond. A114, 243265 (1927).
80. Everett, H. Relative state formulation of quantum mechanics. Rev. Mod. Phys. 24, 454462 (1957).
81. Bohm, D. A suggested interpretation of the quantum theory in terms of hidden variables. Phys. Rev.
85, 166179, 180193 (1952).
82. Rauschenbeutel, A. et al. Coherent operation of a tunable quantum phase gate in cavity QED. Phys.
Rev. Lett. 83, 51665169 (1999).
83. Jennewein, T., Simon, C., Weihs, G., Weinfurter, H. & Zeilinger, A. Quantum cryptography with
entangled photons. Phys. Rev. Lett. 84, 47294732 (2000).
84. Lenard, P. ber die lichtelektrische Wirkung. Ann. Phys. 8, 149198 (1902).
85. Millikan, R. A. A direct photoelectric determination of Plancks h. Phys. Rev. 7, 355388 (1915).
86. Millikan, R. A. Albert Einstein on his seventieth birthday. Rev. Mod. Phys. 21, 343345 (1949).
87. Compton, A. H. The spectrum of scattered X-rays. Phys. Rev. 22, 409413 (1923).
88. Bothe, W. & Geiger, H. ber das Wesen des Comptoneffekts; ein experimenteller Beitrag zur Theorie
year of physics review articles
NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature 237
Zeilinger new 13/1/05 1:56 pm Page 237
Nature PublishingGroup 2005
der Strahlung. Z. Phys. 32, 639663 (1925).
89. Stuewer, R. The Compton Effect: Turning Point in Physics (Science History, New York, 1975).
90. Tu, L.-C., Luo, J. & Gillies, G. T. The mass of the photon. Rep. Prog. Phys. 68, 77130 (2004).
91. Cocconi, G. Upper limits on the electric charge of the photon. Am. J. Phys. 60, 750751 (1992).
92. Weinberg, S. Light as a fundamental particle. Phys. Today 28, 3537 (1975).
93. Lawrence, E. O. & Beams, J. W. The element of time in the photoelectric effect. Phys. Rev. 32, 478485
(1928).
94. Forrester, A. T., Gudmundsen, R. A. & Johnson, P. Photoelectric mixing of incoherent light. Phys. Rev.
90, 16911700 (1955).
95. Clauser, J. F. Experimental limitations to the validity of semiclassical radiation theories. Phys. Rev. A
6, 4954 (1972).
96. Klein, M. J. Einsteins first paper on quanta. Nat. Phil. 2, 5986 (1963).
97. Stachel, J. et al. in The Collected Papers of Albert Einstein (eds Stachel, J. et al.) Vol. 2, 134148
(Princeton Univ. Press, Princeton, 1989).
98. Pais, A. Einstein and the quantum theory. Rev. Mod. Phys. 51, 863914 (1979).
99. Clauser, J. F. in Quantum [Un]speakables: From Bell to Quantum Information (eds Bertlmann, R. A.
& Zeilinger, A.) 6198 (Springer, Heidelberg, 2002).
Acknowledgements We acknowledge wonderful collaborations and challenging
discussions with many colleagues and friends in the worldwide quantum optics
community over the years.
Competing interests statement The authors declare that they have no competing financial
interests.
year of physics review articles
238 NATURE| VOL 433| 20 JANUARY 2005| www.nature.com/nature
Zeilinger new 13/1/05 1:56 pm Page 238
Nature PublishingGroup 2005