Transcript The UW Nonlocal Communication Experiment

The UW Nonlocal Quantum
Communication Experiment
John G. Cramer
Professor of Physics
Physics 324A
August 14, 2007
1
At a News Stand Near You …
New Scientist
September 30, 2006
Seattle Post Intelligencer
November 15, 2006
2
Quantum Nonlocality
“Spooky Action-at-a-Distance”
Albert Einstein
3
Entanglement and Nonlocality
Entanglement: The separated but
“entangled” parts of the same
quantum system can only be
described by referencing the state
of other part.
The possible outcomes of
measurement M2 depend of the
results of measurement M1, and vice
versa. This is usually a consequence
of conservation laws.
Nonlocality: This “connectedness”
between the separated system parts
is called quantum nonlocality. It
should act even of the system parts
are separated by light years.
Einstein called this “spooky actions at
a distance.”
Measurement 1
M1
Entangled
Photon
Source
Entangled
photon 1
Nonlocal
Connection
Entangled
photon 2
M2
Measurement 1
4
EPR Experiments
The Freedman-Clauser Experiment, Phys. Rev. Letters 28, 938-942 (1972).
A series of EPR experiments, beginning with the 1972 FreedmanClauser experiment, have demonstrated convincingly that measurements
performed on one of a pair of polarization-entangled photons affect the
outcome of measurements performed on the other entangled photon.
5
Can Quantum Nonlocality
be Used to Send Signals?
 It is now well established that quantum nonlocality really
does “connect” the separated parts of the same quantum
mechanical system (c.f. Freedman-Clauser, Aspect, etc.)
 There are several “No-Signal Theorems” in the literature
(c.f. P. Eberhard, A. Shimony, …) showing that quantum
nonlocal signaling is impossible, e.g., a change on one
measurement has no observable effect on the other, in
the absence of coincidence links.
 However, Peacock and Hepburn have argued that these
“proofs” are tautological and that certain key assumptions
(e.g., measurements are local) are inconsistent with the
quantum formalism (e.g., Bose-Einstein symmetrization).
 Therefore, the question of nonlocal signaling remains
“open” (at least a crack) and should be tested.
6
Interference and
Entanglement
7
Interference of Waves
8
One-Slit Diffraction
9
Two-Slit Interference
10
Turning Interference
On and Off
Two-slit
Interference
Pattern
H
V
No Two-slit
Interference
Pattern
11
“Ghost” Interference
In their 1994 “Ghost Interference”
experiment, the Shih Group at the
University of Maryland in Baltimore
County demonstrated that causing
one member of an entangled-photon
pair to pass through a double slit
produces a double slit interference
pattern in the position distribution of
the other member of the pair also.
If one slit is blocked, however, the
two slit interference pattern is
replaced by a single-slit diffraction
pattern in both detectors.
Note that a coincidence was
required between the two photon
detections.
12
Can We Use
Quantum Nonlocality
for Communication?
13
Down-Conversion with LiIO3
LiIO3 Crystal
Entangled Photons:
 
Beam
Stop
1
 k '1 k "2  k "1 k '2 

2
Energy and Momentum Conservation:
ELaser 
hc
Laser
1 1
 E1  E2  hc    ; kLaser  k1  k2
 1 2 
14
Starting Point: Dopfer’s PositionMomentum EPR Experiment
LiIO3
Down-Conversion
Crystal
“Heisenberg” Lens
f = 86 cm
“Heisenberg”
Detector D1
28.2o
Laser Beam
Stop
f
Auxiliary
Lens
Double Slit System
a = 75 mm, d = 255 mm
Momentum
Position
Double-Slit
Detector D2
Coincidence
Circuit
or
Birgit Dopfer
PhD Thesis
U. Innsbruck, 1998.
2f
f
2f
Note the use
of coincidence.
15
Dopfer’s Results
Receive: Observe Interference at D1?
Send: Move Detector D2 Position
16
What’s Going On?
17
Testing
Nonlocal
Communication
18
Can We Eliminate the
Coincidence Requirement?
University of Washington test of
nonlocal quantum communication.
Differences
from Dopfer
Experiment:
1. Use collinear Type 2 downconversion
in BBO.
2. Separate entangled beams with
polarizing splitter.
3. Image slit pairs with an upstream lens
at distance f.
4. Use fiber optics in to switch on/off
the which-way measurement.
19
Interference Fringes Visible?
Dealing with transverse variations
may require extra distance or some
compensation for the thick-source
effect (e.g., a diverging lens) so that
the wave fronts from the crystal
arriving at the slits are parallel to the
slit system, with a minimum of phase
variation introduced by differences in
the point of production within the
crystal.
BBO
V
351 nm
IR Pass
Filter
Polarizing
Splitter
V
702 nm
H
702 nm
IR Pass
Filter
Argon Ion Laser
The first issue to be addressed
experimentally is whether a 2-slit
interference pattern can be observed
with the entangled photons from the
“thick” BBO crystal down-conversion
source. Using collinear downconversion
eliminates longitudinal thick-source
variations.
Slits
Camera
20
Demonstrating Nonlocal
Quantum Communication
The University of Washington test
of nonlocal quantum communication.
Send
To demonstrate nonlocal quantum
communication, one simply changes the
switch and observes a change in the
interference pattern at the camera.
That would constitute a breakthrough
discovery.
Receive
21
Faster than Light
&
Backwards in Time
22
A Demonstration of
Superluminal Signaling
In this test, we would string
equal lengths of fiber optics
cables to separate the two
ends of the experiment by a
line-of-sight distance of ~1.4
km.
We would then send bits at
a photon rate of 10 MHz over
this link. Assuming a 10-photon
decoding “latency”, this would
demonstrate a signal
transmission speed of about 5
times the speed of light.
1 km
Send
1 km
Receive
23
A Demonstration of
Retrocausal Signaling
BBO
V
351 nm
10 km
D
D
C
V
702 nm
Polarizing
Splitter
Switch
D
S
S2
S1
H
702 nm
10 km
IR Pass
Filter
Receive
Argon Ion Laser
Send
IR Pass
Filter
S
IR Pass
Filter
Camera
In this test, we leave 10 km of optical fiber coiled up in the corner of the
laboratory, and pass the entangled “Transmitter” photons through this path.
The “Receiver” photons have no such optical delay, and the signal is
received as soon as these photons are detected at D1, which is about 50 ms
before the signal is transmitted, when the twin entangled photons arrive at D2.
24
Time-Travel
Paradoxes
25
The Bilking Paradox
Suppose that we constructed a million
connected retrocausal links of the type just
shown (or used 107 km of fiber optics). Then
the transmitted message would be received 50
seconds before it was sent.
Now suppose that a tricky observer
receives a message from himself 50 seconds in
the future, but then he decides not to send it.
This produces an inconsistent timelike loop,
which has come to be known as a “bilking
paradox”. Could this happen? If not, what
would prevent it?
26
Anti-Bilking
Discussions of such bilking paradoxes have been
published in the physics literature in the 1940s by
Wheeler and Feynman (advanced waves) and in the 1990s
by Kip Thorne and colleagues (timelike wormholes).
The consensus of both discussions is that Nature will
forbid inconsistent timelike loops and will instead require
a consistent set of conditions. Thorn and coworkers
showed that “nearby” to any inconsistent paradoxical
situation involving a timelike wormhole, there is selfconsistent situation that does not involve a paradox.
As Sherlock Holmes told us several times, “When
the impossible is eliminated, whatever remains, however
improbable, must be the truth.”
27
Bilking & Probability Control
These speculations suggest that equipment failure
producing a consistent sequence of events is more likely
than producing an inconsistency between the send and
receive events. The implications of this are that bilking
itself is impossible, but that very improbable events could
be forced into existence in avoiding it.
Thus, using the threat of producing an inconsistent
timelike loop, one might “bilk” Nature into producing an
improbable event. For example, you might set up a highly
reliable system that would produce an inconsistent
timelike loop unless the number for the lottery ticket you
had purchased was the winning number.
28
The “Immaculate
Conception” Paradox
The other issue raised by retrocausal signaling might be called
the “immaculate conception” paradox. Suppose that you are using
the setup described above, and you receive from yourself in the
future the manuscript of a wonderful novel with your name listed as
the author. You sell it, it is published, it becomes a best-seller, and
you become rich and famous.
When the time subsequently comes for transmission, you duly
send the manuscript back to yourself, thereby closing the timelike
loop and producing a completely consistent set of events. But the
question is, just who actually wrote the novel?
Clearly, you did not; you merely passed it along to yourself. Yet
highly structured information (the novel) has been created out of
nothing. And in this case, Nature should not object, because there
are no inconsistent timelike loops.
29
Present Status
 The experiment has been in testing phases since
mid-January. Our initial attempt to detect the
down-converted photons was with a cooled CCD
camera. We have demonstrated that this detetor
lacked the needed sensitivity.
 The experiment has been moved from B063 (the
UW Laser Physics Facility) to B055, to make room
for the arrival of Prof. Gupta, the newest member of
the UW Atomic Physics Group.
 The experiment is presently being rebuilt, using
avalanche photodiodes as the primary detectors. It
will continue this Fall.
30
Conclusions
 There are no obvious “show stoppers” that would
prevent the proposed measurements. Nevertheless,
because of their implications, they have a low
probability of success.
 My colleague Warren Nagourney and I have been
working on this experiment since January, and
because of the publicity, we now have $40k in
contributions from foundations and individuals to
support the work.
 This experiment is a rare opportunity to push the
boundaries of physics with a simple tabletop
measurement. We intend to push hard.
31
The
End
32
Transactions with D2 at 2f
33
Transactions with D2 at f
34
The Thick-Source Effect (1)
Possible Fix 1:
Insert a
converging lens in
path to slits.
Volume of LiIO3
down-conversion
crystal illuminated
by UV laser beam.
Path length differences at
slit positions can be greater
than l/2, shifting and washing
out the interference pattern
“signal”.
35
The Thick-Source Effect (2)
Lxc= 1 m
Lxc= 3 m
Lxc= 10 m
Fix 2: Lengthen the crystal-to-slit distance, flattening the wave fronts.
Fix 3: Srikanth suggests
spatial filtering by crossing
over the light from the
crystal with two lenses, and
using an aperture to eliminate
non-parallel wave front
components.
Conclusion: There seem to be several
ways of solving the Thick-Source
problem.
36