I was thinking about another experiment to try, to check the wave/particle nature of light. Basically it is the double slit experiment but with a prism added to the light source to split the source into two beams. One beam is sent through a dark filter and then through the double slit and hits the screen. The filter until no dots or interference fringes appear on the screen. Then the second, brighter beam from the prism which hasn't gone through the double slit, is shined onto the screen. This is coherent light with the other darkened beam, and so should add constructively to the undetected interference fringes that are theorized by wave theory to be on the screen. It is possible that this would then start to show an interference pattern again, as the extra energy from the other coherent light source could be enough to trigger the quantum orbital thresholds.
I was thinking about another experiment to try, to check the wave/particle nature of light. Basically it is the double slit experiment but with a prism added to the light source to split the source into two beams. One beam is sent through a dark filter and then through the double slit and hits the screen. The filter is darkened until no dots or interference fringes appear on the screen. Then the second, brighter beam from the prism which hasn't gone through the double slit, is shined onto the screen. This is coherent light with the other darkened beam, and so should add constructively to the undetected interference fringes that are theorized by wave theory to be on the screen. It is possible that this would then start to show an interference pattern again, as the extra energy from the other coherent light source could be enough to trigger the quantum orbital thresholds.
You are describing the classical geometrical optics with photons as billiard balls interpretation.
The photon isn't real until you detect it. There is only the time varying probability wavefunction of finding the photon at every position inside the experiment. Destructive interference cancels most of these contributions out and reduces to the classical geometrical path. The interference image on the screen is determined only by the paths that are able to contribute to it.
No and that is exactly the point. The wavefunctions that have taken both paths have to overlap at the recombination point or you do not see any interference fringes. If you can determine which path the photon must have travelled then the fringes do not occur.
That is pure classical optics. No quantum interference possible.
To illustrate my point lets consider an extreme experiment where one leg is 1m in the lab and the other is by moon bounce with a suitably large optically perfect retroreflector. The long leg photon will not have left the lab before the other one is detected. But neither of them will generate interference fringes because their wavefunctions never have a chance to overlap and interact at the target.
In Larkin world if I hold a top hat in the way of the returning moon bounce leg of the interferometer for the first two seconds it affects the outcome of the experiment. No fringes with the top hat in and fringes with the top hat out, but all the photons still detected.
In the real world you would in this case see two fairly clean pulses of photons that had travelled their respective distances. The ones that had moon bounced with some additional timing spread from atmospheric effects. It is only when the time delays in the multipath experiment are roughly comparable with the length of the non-zero part of the wavefunction that you can see interference patterns.
A femtosecond laser pulse is a broadband light source because of time gating even though the pulse is made from a single laser frequency. The pulse coming out behaves exactly as you describe but it does not produce interference fringes because of the wide range of frequencies to form the amplitude modulated pulse envelope.
By comparison a continuous carrier wave coherent laser in a long cavity is capable of delivering an output that is coherent over length scales of at least L*Q on a good day where L is the length of the lasing cavity and Q is the quality factor of the transition.
As Phil said some time ago. You can get yourself very very confused thinking about a single photons behaviour inside interferometers.
Fast light pulses have very uncertain frequency photons in them - that is the whole point and why your reasoning here fails.
Not at all. The photon is either detected or it isn't. If the other part of the wavefunction that has followed the second path arrives in time then interference fringes occur and the photon is found most often where they interfere constructively. If not then the two wavefunctions behave completely independently and you get two clean pulses detected but no interference fringes.
Yes. But it is you who doesn't seem to realise that!
OK, fire CW light from a laser into an interferometer with leg lengths that differ by 5 ns.
Is it possible to make fringes?
If yes, put a detector on the screen. Do the fringes disappear?
Now pulse the laser at some modest rate, 1 ns pulse width. Do the fringes disappear?
If not, connect an oscilloscope and measure the TOF. Do the fringes disappear?
If not, what does the oscilloscope waveform look like? What are the TOFs?
Are you arguing that one must wait for the full moon-and-back round trip before interference can be detected? The short-leg photons, half of the ones that hit the screen, will hit the screen in nanoseconds... but they have to wait seconds to make up their minds where to land?
I believe that exact experiment has been done. If the hat makes it impossible for photons to hit the moon, there is no interferance.
You can do interferance experiments with one photon per second. Each photon interferes with itself. It doesn't care about other photons.
John
As other people have siad, QM makes no sense. It's not causal.
Only in "Larkin World" which is so often detached from reality.
Of course.
Obviously not.
Yes. The frequency dispersion is too great. The short pulse contains a range of frequencies and if you run them over too long a baseline difference the phase errors at the output smear out the fringes.
OK. Lets try it with some numbers assume a NdYAG laser 1032nm ~ 1um Take c = 3 x 10^8 m/s so f = 3 x 10^14
A 1ns pulse contains 3x10^5 cycles and to obtain that pulse shape there is now a spread in the frequency of the original pure laser wavelength equivalent to roughly 1ppm. When you dim the source you could get any photon chosen from the timegated frequency distribution. So once you have a baseline difference of ~ 1m the fringes are gone. 5ns is around
1.5m so I think you would actually see some (about 33%) fringe amplitude at the maximum of the sinc(x) function sat on a higher baseline and as you go to ever longer baselines the fringes fade out. I could easily be wrong by a factor of two or so here but the basic description of what happens is OK.
When the rms phase error across the path difference in baselines reaches any multiple of pi the fringes all vanish.
It would be even more extreme with a femto second laser pulse as they are almost broadband radiation sources.
That changes nothing
If you can measure the TOF and can see two distinct independent peaks then it is pure classical ray optics and you do not see any interference fringes. It is only when the wavefunctions taking the two paths are able to at least partially overlap that you see interference patterns.
NO. I am pointing out that it has no knowledge of the other path and as such it does not form fringes. The short path photon hits the screen in a few nanoseconds and the long path photon hits just just over 2 seconds later. Plenty of time to remove the top hat from the return leg of the interferometer before the moon bounce photon returns.
The photons each follow one path or the other. You can tell which path each took and so there are no interference fringes at all. Closing the return path whilst the wavefunction at that position was still zero was meant to highlight how completely off the wall your reasoning was but it seems you are unwilling to think about this stuff at all.
Putting the hat in the return leg of the interferometer prevents photons from going to the moon? Your ideas are wackier than I ever thought possible. I don't seem to be making any progress here, but I hope that others reading this will perhaps be able to understand.
Yes. But unless you have a whole bunch of individual photons with pretty close to the same frequency you end up with smeared out interference fringes. It all comes back to the Heisenberg Uncertainty principle and the intimate Fourier relationship of the time domain and frequency domain. If you amplitude modulate the light beam you necessarily create uncertainty in the frequency domain.
It can if the path difference is an appropriate length for interference fringes to be seen. The physics involved here isn't that much different from Young's slit done with white light that you don't have to go too far off axis before the fringes all smear out to uniform grey.
The same is true for nearly monochromatic light. A CW laser is an incredibly good monochromatic light source - once you start pulsing it there is significant frequency domain uncertainty.
Heck you can even hear it in the dots vs dashes of Morse code!
OK, but how can you explain the fact that, at very low intensities, interferance can be observed literally one photon at a time?
Imagine a light source that is known to contain a very good, long-coherence CW laser and an attenuator. The attenuator is adjustable from 1 watt to 1 photon per minute average output. At one watt, we set up to see interferance fringes.
So turn down the attenuator. At 1 photon per second or per minute, we can run the light through the interferometer onto a piece of film and eventually accumulate an interferance pattern. There are examples on the web.
Now we can replace the film with a photocathode and an electron multiplier and a position-sensitive detector of some sort. We'll see those same photons and the same band pattern. But now we can time-stamp the position and arrival time of individual photons to picosecond resolution.
So what's keeping us from measuring TOF? All we need to know is the photon launch time. You are saying that, if we know this, we can determine which leg of the system each photon took, so the fringes must disappear.
Last step: use a shutter for all or part of the attenuator that gets us down to 1 photon average per minute. The shutter could be picoseconds wide, and allow, on average, well less than one photon through per attempt. The photons that emerge from the source are no different from any other photons, except that we know when they must have left. If we don't connect an oscilloscope to the shutter, we have no clue that the photons actually passed through a shutter, but somehow such photons are different from ordinary photons, in they they are incapable of interferance.
So, where in the chain do the fringes disappear? Can a black-box source have a switch on the side, such that it can manufacture ordinary low-rate, monochromatic photons that are capable/incapable of interferance depending on the switch setting?
You've got a perfect shutter that you can open and close. You send a 'monochromatic' pulse of light into your interferometer, let's say it's 10 ns long. (you open and close your shutter for 10 ns.) (I see a 'envelope' of light 3 meters long bouncing around.) Now there is a beam splitter, half the light goes one way, half the other.
(Now I see two envelopes of light, both still 3 meters long.)
Now we bring the light beams back together each travelling a different distance. (We have an interferometer with different path lenghts.)
If that path length difference is greater than 3 meters then the two light pulses 'miss' each other and there is no interference.
That's an easy experiment! No TOF, no single photons. If the path length difference is greater than the length of the light pulses no interference pattern! (at least that's my prediction)
'Photons' are a consequence of the detection process. Either the weak EM field of the strongly attenuated light kicks an electron out of the photo-cathode, or it doesn't. The probability of this happening at any specific location depends on the strength of the EM field there, and that is shaped by the interference pattern.
There is no need at all to think in terms of individual photons tracing their way through multiple paths at the same time.
Which leads one to conclude that, if your conjecture is true, photons interact with one another such as to prevent interference. The photons at the end of the 10 ns pulse somehow decohere the ones at the beginning, ones that have already hit the screen.
If you think about it that way, one photon, traveling over differing path lengths to the screen, always misses. So interferance is impossible.
" At first Einstein believed that the light-quantum hypothesis was merely "heuristic": light behaved only as if it consisted of discontinuous quanta. But in a brilliant series of subsequent papers in 1906 and 1907, Einstein used his statistical mechanics to demonstrate that when light interacts with matter, Planck's entire formula can arise only from the existence of light quanta?not from waves. Einstein considered that light quanta, together with the equivalence of mass and energy, might result in a reduction of electrodynamics to an atom-based mechanics. But in 1907 he discovered that atoms in matter are also subject to a quantum effect.5
Here he made use of another galling experimental problem. Experimentalists had found that when solid bodies were cooled, the amount of heat they lost failed to fit a simple formula that followed from Newtonian mechanics. Einstein showed that the experiments could be explained only on the assumption that the oscillating atoms of the solid lattice can have only certain, specific energies, and nothing in between. In other words, even the motions of atoms?which are continuous in Newtonian mechanics?exhibit a quantum structure. Mechanics and electrodynamics both required radical revision, Einstein now concluded: neither could yet account for the existence of electrons or energy quanta.6 "
From that it looks like quantization of light was assumed before quantization of matter was assumed, and then later on the quantization of matter was added to the model. I think looking back, it makes more sense to assume quantization of matter first, before inherent quantization of light is assumed.
A lot of people were, and still are, upset about the wave-particle duality, and the fact that the behavior of photons doen't seem to be sensible or causal. But that's the way they behave.
Personally, I'm still trying to figure out whether, if we measure the TOF through an interferometer, the fringes go away.
No, if the path length difference is less than the 'pulse length' then you get some interference. With zero path length difference you get maximum interference. Think about pulses of RF travelling different lengths and then interfering. If they don't overlap in time there is no interference.
(getting the spatial overlap 'correct' is usually the 'tweaky' part of an interferometer.)
I think the biggest variable is the "photon" source. I am skeptical of any instrument that claims to be able to emit a single time gated photon that can be detected a distance away. I have seen some lithography created "single photon" sources, but I'm not sure if they are time gated but if thats possible they might be a good thing to use for an experiment.
All you need is a low intensity source and a fast shutter. It doesn't guarantee exactly one photon per shutter opening, but it does guarantee that you know the launch time of any photons that do get out.
The other end, a single-photon-sensitive detector, is common enough. So whenever the detector senses a photon, you can measure the arrival time and, knowing the shutter timing, you have TOF.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.