FTL signaling via frustrated total internal reflection

[Jay-qu]

But what Im saying is if you send say 1000 photons and the probablity of tunneling is just 1% then at least 10 photons should make it through. So sending 1000 photons with the same spin state as a single bit could be used to send a signal.

[CraigD]

Originally Posted by CraigD

AFAIK, unlike in the case of the ansible, there exists no thorough, compelling explanation of why tunneling such as that measured in the N&SFTIRE cannot be used to send a signal faster than light speed (but not instantaneously) across a specific optical apparatus.

Originally Posted by Erasmus00

I think the reason you can't send a signal this way is that the tunneling process is random, not all the photons will tunnel. It seems to me this will likely destroy any signal you try to pass.

Originally Posted by Jay-qu

I dont think its that easy to discount.. Im sure an protocol could be designed, such as sending enough photons with the one signal so that the probability becomes reasonable enough to assume transmission.

A single bit signal consisting of a large number of photons is adequate to exhibit a particular signal speed, can’t be destroyed by loss of fewer than nearly all of its photons, and requires the minimal “protocol” of simply detecting at least one photon.

I think we’re confusing the concepts of latency (distance / signal speed) and data rate (“throughput”). A signal can have a high speed, and thus a low latency, and at the same time a low data rate, or vice versa. For example, a distant supernova (suppose that such a single event requires 1 day to detect) has a signal speed of c, and a data rate of (1 event in one day), while a Netflix DVD has a signal speed (supposing, as is my case, a mailing distance of about 20 km, mailing time (latency) of about 36 hours, and about 2 hours of movie in 4 GB on DVD) of about and a data rate of (the wikipedia article “comparison of latency and throughput” discusses this at greater length).

Originally Posted by Erasmus00

Each photon can only carry 1 bit at a time (either spin up or spin down). Hence, any information carrying signal will require many photons. Removing a random number of random bits from that signal will destroy any signal.

This assumes that the signal must be of the maximum possible “density” of 1 bit per photon. Normally communication with photons (light, radio, etc) is much less dense – consider, for example, a 19th – 20th century ship to ship signal lamp, hand operated using Morse code. Such a signal can withstand the loss of many individual photons to random causes (fog, bugs on the lens, etc) without being destroyed. Such a signal is (at a very rough estimate) around 10^22 photons/bit. A modern photo-optic (ie: fiber optic) communication system typically requires a few hundred photons/bit, although systems requiring as few as 1 photon/bit exist (ie: those used for quantum encryption) and do use polarization to encode bits in single photons.

As with the previously described data rate, photon/bit ratios are unrelated to signal speed. Transmitting a single bit between a sender and receiver in less time that the distance between them divided by c qualifies a signal as “faster than light”, as I understand the usual meaning of the expression. N & S’s experiment appears to have demonstrated just such a signal – detection of a single or many photons from a precisely timed emission sooner than an ordinary light signal could do so.

PS: I must confess a personal, not-fully rational bias: I’m uncomfortable with the possibility of FTL signaling. My reason is that I’ve long counted on special relativity as a sort of “absolute limiter” on the amount of information that can exist in a given volume of space, a limit important to some of my “believe intuitively but can’t prove” suspicions about the ultimate nature of reality. That it can be, in principle, violated, “blasphemes” against this belief of mine. None the less, I still find myself without any compelling objection to the implications of N & S’s experiment.

[Erasmus00]

Originally Posted by CraigD

I must confess a personal, not-fully rational bias: I’m uncomfortable with the possibility of FTL signaling. My reason is that I’ve long counted on special relativity as a sort of “absolute limiter” on the amount of information that can exist in a given volume of space, a limit important to some of my “believe intuitively but can’t prove” suspicions about the ultimate nature of reality. That it can be, in principle, violated, “blasphemes” against this belief of mine. None the less, I still find myself without any compelling objection to the implications of N & S’s experiment.

The thing is: quantum mechanics and special relativity are consistent. (and when put together yield quantum field theory, which the standard model is built on). Hence, quantum mechanics cannot create ftl signaling. Unfortunately, building models of tunneling is rather difficult, but I do want to sit down and try to understand this experiment when I have a bit of time.
-Will

[Qfwfq]

Certainly, I don't expect a proper analysis would reveal propagation faster than c but I don't see it as a matter of photons getting lost. It's a matter of analysing an incoming wave packet and seeing how its Fourier transform is altered by the exponential factor of the evanescent wave. Anyone up to doing the integrals?

[Jay-qu]

Originally Posted by Qfwfq

Certainly, I don't expect a proper analysis would reveal propagation faster than c but I don't see it as a matter of photons getting lost. It's a matter of analysing an incoming wave packet and seeing how its Fourier transform is altered by the exponential factor of the evanescent wave. Anyone up to doing the integrals?

Maybe depends how hard they are if not I can take them to uni and get one of the professors to take a look

[Qfwfq]

Originally Posted by Jay-qu

Maybe depends how hard they are

Hard enough.

Actually I put it a bit sloppily yesterday. The idea is that, for a given incident wave packet, you need to multiply its Fourier transform by the exponential factor and then re-transform to get the emergent wave packet according to the distance between the two prisms.

[Jay-qu]

Ive just learned how to do fourier expansions in the last few weeks but somehow I think the rest will fly ominously above my head..

[LaurieAG]

Originally Posted by Qfwfq

Certainly, I don't expect a proper analysis would reveal propagation faster than c but I don't see it as a matter of photons getting lost. It's a matter of analysing an incoming wave packet and seeing how its Fourier transform is altered by the exponential factor of the evanescent wave. Anyone up to doing the integrals?

Hello Q,

I'm not too sure about it being a standing wave as a standing wave should only produce a photon at one of the detectors or the other not both. Multiple moving wave pulses (1 photon equivalent) could be recorded sequentially at both detectors if the back of the prism acted like a mirrored surface.

The vectors 'Tv' and 'Th' in the diagram can be obtained by mirroring the high and low points of a moving pulse at the back end of the prism. While the photon isn't exactly a wave the mirror vectors of these two wave points (top & bottom) will appear as they do in the diagram.

It would be good to see exactly how the experiment was undertaken and have copies of the raw data sent and captured, before going any further maths wise, than the basic vectors shown in the diagram.

Also, It would be interesting to know the density of the glass used in the prisms, their exact corner angles (i.e. are they 60 degrees?) and the atomic bonding angles of the atoms in the glass (and any impurities/doping).

[Qfwfq]

Where did I imply it being a standing wave, Laurie?

Originally Posted by Jay-qu

Ive just learned how to do fourier expansions in the last few weeks but somehow I think the rest will fly ominously above my head..

Actually, I imagine you're talking about a Fourier series, which is suitable for periodic functions. The Fourier transform I mean is tricky and entangled with distribution theory as well. You could, however, ask your professors if they can make anything out of it, I would need to do quite some looking up before drawing any reliable conclusions.

[Jay-qu]

right you are Q

I had a talk to one of my lecturers about this experiment and he wasnt terribly surprised.. he seemed to think that eventually someone would show that it was unable to be used for FTL communication - though he did still think it warranted a closer look