Re: One Photon Double Slit Reinterpretations

From: Franz Heymann (notfranz.heymann_at_btopenworld.com)
Date: 01/22/05 

Date: Sat, 22 Jan 2005 22:58:26 +0000 (UTC)

"Zigoteau" <zigoteau@yahoo.com> wrote in message
news:1106389663.993473.72970@z14g2000cwz.googlegroups.com...
> Hi, Franz and Uncle Al,
>
> > By the way, the term "single photon source" usually just means
that
> > the source is feeble enough for one photon to have passed through
the
> > equipment before the next one is sent in.
>
> While I fully appreciate that Landle is not the full Bose
condensate,
> and that the opinions of the two of you on any physical question are
> more generally to be trusted, this last statement of yours, Franz,
is
> the nub of the issue, and shows that Landle does have a point here.

Not at all. Landle has never had any point worth anything except for
lampooning.

Before we go any further, I have to make it clear that my single
photon source was not made for the purpose of investigating single
photon interference, but specifically to investigate the single photon
response of a batch of photomultipliers. However, I do understand the
ins and outs of single photon interference experiments sufficiently
well to feel quite confident about commenting on them.

> There is no way to determine what is going on during the flight of
the
> photon.

One does not have to. All one has to know is that there is
guaranteeably only zero or one photon at the plane of the slits at any
moment

> In your experiments, Franz, you are not directly observing
> photons but bunches of electrons at the final stage of your PMT.
Some
> of these bunches are false positives, and originate from the dark
> current of your cathode.

That is quite incorrect in all respects. What makes you think that I
did not test for such artefacts?
My light source was a LED pulsed with 5ns long pulses. Each pulse
sent usually 0, but occasionally 1 photon through the equipment. That
was guaranteed by the fact that the counting rate at the detector was
very small compared to the pulse rate of the LED. From the data
collected, an application of Poisson statistics made it possible to
determine what fraction of the detected photons was accompanied by a
second one. It was easy to keep this to a negligible fraction.
The detector was gated (apropriately delayed) with the source pulses.
There were no spurious coincidences of any consequence, as was proved
by deliberately mis-timing the gate signals.
Furthermore, the dark current pulse height distribution of a
photomultiplier is quite different from the single photoelectron pulse
height distribution, because there are in general more electrons
produced per second from the succession of dynodes than from the
cathode alone. It was therefore rather easy to discriminate against a
substantial fraction of the dark current pulses even before the gating
was applied.
Quoting from a poor memory, the following gives you an order of
magnitude feel:
Ungated, undiscriminated dark current counting rate ~ 100,000 per
second.
Ungated discriminated dark current counting rate ~ 20,000 per second
Gate length 10 ns
Therefore number of dark current pulses per gate pulse ~ 0.0002
Number of true single photon pulses per gate pulse ~ 0.1
So the typical signal to noise ratio was ~500 : 1.
I call that pretty clean.

> There are false negatives, too, because the
> number of bunches you get is less than the number you get by
dividing
> the beam energy by the photon energy (the cathode has less than 100%
> efficiency). Hence from an observation of your PMT output you can
never
> say categorically what the photons are doing, just vaguer statements
> about photon statistics.

That is utterly irrelevant if what you are trying to build up is a
spatial distribution to verify that the interference pattern is one
which is caused only by photons interfering with themselves. The only
requirement is that there must be only one photon at a time at the
slit position. Whether you detect or miss it is irelevant, as long as
the detection efficiency is constant throughout the experiment.
>
> You quite correctly state that single-photon experiments are just
> experiments at very low light intensities. However the "duration" of
a
> photon is a purely notional quantity, typically calculated from the
HUP
> the time as Planck's constant divided by the bandwidth. If I decree
> that my photons each have a bandwidth of a microhertz, you cannot
> contradict me because that is a purely notional thing.

Yes, I will indeed contradict you.

The photons I used were emitted from a LED to which 5 ns pulses were
applied. At the very most, the length of the photons in the time
domain must therefore have been less than 5 ns, since LED emission is
not from a forbidden transition. This translates to a wave train with
a physical length of 150 cm. My equipment was longer than that. By
the way, this is only a crude upper limit. If you were to make use of
a typical natural linewidth and use the uncertainty principle, you
would find that the typicsl wavetrain length of a photon is much less
than that. For example, the temporal duration of the wavetrain of a
single photon emitted in an incoherent process is just the lifetime of
the excited state. In the case of a gas this leads to photon
wavertrains which are typically a metre or so in length. I am not
quite certain, but I think that the lifetimes of the excited states
involved in LED emission are considerably shorter than those in a gas.
If so, the wavetrains of a LED photon will be even shorter.

> However that
> means it takes a week for each photon so defined to pass through the
> equipment.

You have made a mistake somewhere. Even the coherent light of a
typical laser has a coherence time measured in seconds only.

> Your PMT cathode would have to be pretty sensitive to
> extract an interference pattern under those circumstances.

The actual typical length of the wave train associated with a photon
in the optical region is measured in tens of centimetres, as I have
said before.

> There is no conceptual problem about having a laser source with a
> linewidth of a microhertz. However the duration of each of the
electron
> avalanches from your PMT is a function of the PMT design, not one
week.
> Again, the "duration of a photon" is a purely notional quantity.

The light source of which I spoke did not produce coherent light.

I had thought you knew more physical optics than you are displaying in
this note.
I invite you to have another thought about the lifetimes of excited
states which relax by optical emission.

> If an
> electron is considered to be a "point particle", then logically a
> photon should also be so considered.

To the best of our knowledge that is indeed so. Some of the groups
working on LEP have been studying the photon form factor, but
regrettably I don't know the present state of play.

> It then has zero duration, in
> which case any optical experiment, even using a terawatt laser, is a
> single-photon experiment.

You really have descended into raising strawmen. If the photon is
point like and they are spaced at a few cm in the direction of travel,
there will be many of them in the equipment at the same time.
However, when talking about interference matters, we are concerned
with the length of the wavetrain which determines the dynamics of the
photon and not with the actual shape of the photon.
>
Sorry, Zigoteau, but you have spoken so much nonsense up to here that
I simply cannot be bothered to read any more of what you have to say,
so I snip it.

[snip]

Franz

Relevant Pages

  • Re: Math/Physics Gurus Help - SR probability problem
    ... The photon emitting device can have a random speed (let's restrict ... sending out N pulses where N is a very large number. ... For pulses passing the device in the positive direction I ... measure that in k/N trials a photon crossed the x-axis between the ...
    (sci.physics.relativity)
  • Re: Single photon emission - absorbtion time delay experiments ?
    ... That's a misguided view of what a photon is, ... time for the measurement to finish as well as accept a large ... wave fronts or pulses. ...
    (sci.physics.research)
  • Re: One Photon Double Slit Reinterpretations
    ... to exactly one photon at a time. ... Can anyone let me know what equipment ... > Landle wrote: ... > Perseverative idiot. ...
    (sci.physics)
  • Re: One Photon Double Slit Reinterpretations
    ... > In the double slit where you shoot an individual photon ... > one photon at a time. ... When I last made a single photon source I calibrated it by pointing it ... equipment before the next one is sent in. ...
    (sci.physics)