Travel time of observed VS unobserved object?

[pittsburghjoe]

Has someone tested the travel time of particles in the delayed choice experiment? I realize that timing an unobserved object isn't going to be easy, but could we, at least, get an average time?

[imatfaal]

Has someone tested the travel time of particles in the delayed choice experiment? I realize that timing an unobserved object isn't going to be easy, but could we, at least, get an average time?

 

Yes to an extent - lots of these delayed choice quatum eraser experiments (dcqe) work by coincidence matching; so the paths are measured to astonishing accuracy such that signal and idler photons can be matched by timing coincidence (you check this is working well before you start any interference / quantum effects testing). So whilst the travel time is not directly measured (spontaneous parametric down conversion is inherently random so we do not know which input photon will become two entangled output photons) whilst the dcqe is being run; it is measured before and checked in every result. If the travel time was not what was expected we could not run the experiment in the first place

[swansont]

Yes to an extent - lots of these delayed choice quatum eraser experiments (dcqe) work by coincidence matching; so the paths are measured to astonishing accuracy such that signal and idler photons can be matched by timing coincidence (you check this is working well before you start any interference / quantum effects testing). So whilst the travel time is not directly measured (spontaneous parametric down conversion is inherently random so we do not know which input photon will become two entangled output photons) whilst the dcqe is being run; it is measured before and checked in every result. If the travel time was not what was expected we could not run the experiment in the first place

 

Note that this is relative timing rather than comparison to some external clock. You don't know when the photons were created, and if the travel is of order a meter then nanosecond-level timing is a 33% error bar. Not very impressive. And nanosecond timing would not be easy, even if you had an indication of when the photons were created. (edit: see below)

 

OTOH, coincidence timing rejecting stray signals does a good job at rejecting spurious signals, and the timing requirements are not as stringent as long as the count rate is limited sufficiently — you're not getting 100 million counts per second, which would be an average of 10 ns between events. Not very good if you have a 10 ns coincidence window. But if you're down at 1000 cps, that's a millisecond between events, and coincidence will do a much better job of excluding bad events.

[imatfaal]

 

Note that this is relative timing rather than comparison to some external clock. You don't know when the photons were created, and if the travel is of order a meter then nanosecond-level timing is a 33% error bar. Not very impressive. And nanosecond timing would not be easy, even if you had an indication of when the photons were created.

 

OTOH, coincidence timing rejecting stray signals does a good job at rejecting spurious signals, and the timing requirements are not as stringent as long as the count rate is limited sufficiently — you're not getting 100 million counts per second, which would be an average of 10 ns between events. Not very good if you have a 10 ns coincidence window. But if you're down at 1000 cps, that's a millisecond between events, and coincidence will do a much better job of excluding bad events.

 

I see - I had assumed wrongly that timing accuracy would be greater than that.

 

Not sure why I would assume that - it is just that fineness of measurement always seems to me, as a layman, to be of unimaginable accuracy; so it is difficult to believe sometimes that you guys cannot just up the accuracy to almost any desired level (short of indeterminacy etc).

[swansont]

 

I see - I had assumed wrongly that timing accuracy would be greater than that.

 

Not sure why I would assume that - it is just that fineness of measurement always seems to me, as a layman, to be of unimaginable accuracy; so it is difficult to believe sometimes that you guys cannot just up the accuracy to almost any desired level (short of indeterminacy etc).

 

 

I should backtrack on that. You can do sub-ns precision interval timing if you have access to a good quartz crystal. Long-term timing wouldn't be good (which is what I'm used to thinking about), but short-term is quite good. Calibration of the equipment could still be an issue.

 

But you're still foiled by the fact that you don't know when the photons are created. Unless you're doing an experiment where one of the photons is used as the timing trigger.

[imatfaal]

 

 

I should backtrack on that. You can do sub-ns precision interval timing if you have access to a good quartz crystal. Long-term timing wouldn't be good (which is what I'm used to thinking about), but short-term is quite good. Calibration of the equipment could still be an issue.

 

But you're still foiled by the fact that you don't know when the photons are created. Unless you're doing an experiment where one of the photons is used as the timing trigger.

 

What I meant by the indirect measurement is that in some of these set-ups have deliberately skewed photon paths such that the idlers travel a significant distance more than the signal - thus the coincidence checking would need to be offset as well. If you know the time offset and the additional distance travelled then you know the speed that the extra distance was crossed at. My idea kinda all falls down if the timing accuracy is low and you are relying on very few events with large gaps to make your coincidence spotting easier. Will have to read-up an actual experiment rather than just the concept to see what is what.

[swansont]

 

What I meant by the indirect measurement is that in some of these set-ups have deliberately skewed photon paths such that the idlers travel a significant distance more than the signal - thus the coincidence checking would need to be offset as well. If you know the time offset and the additional distance travelled then you know the speed that the extra distance was crossed at. My idea kinda all falls down if the timing accuracy is low and you are relying on very few events with large gaps to make your coincidence spotting easier. Will have to read-up an actual experiment rather than just the concept to see what is what.

 

 

You can compensate for unequal optical paths with delay lines in your electronics. A 1m cable is something like 5n of delay, since the signal travels at around 2/3c.

 

Uncertainty in (non delay-line) electronics delays are probably what limits the coincidence window. I get the impression that they can be significantly longer (i.e. 100s of ns or even microseconds (depending on what you have), so a corresponding uncertainty of even a few % means a significant window for the coincidence.

[Sensei]

I should backtrack on that. You can do sub-ns precision interval timing if you have access to a good quartz crystal. Long-term timing wouldn't be good (which is what I'm used to thinking about), but short-term is quite good. Calibration of the equipment could still be an issue.

 

But you're still foiled by the fact that you don't know when the photons are created. Unless you're doing an experiment where one of the photons is used as the timing trigger.

 

Suppose so we have neutral pion meson,

after creation it could have some kinetic energy,

and it's flying through spherical vacuum chamber.. full of detectors...

 

After some time, randomly, it's decaying to two photons, it's the commonest decay mode of pion-0..

 

These photons, fly in two opposite directions, in FoR of decaying pion, prior decay, maintaining momentum (FoR)..

Then these photons hit two detectors on opposite sides of vacuum chamber..

Photon A from pion hits detector A,

then photon B from pion hits detector B,

from delay, difference of energy, there are calculated initial pion energy + kinetic energy,

and position were meson decayed..

 

So suppose so we have pion with 135 MeV/c^2 rest-mass and 100 MeV kinetic energy..

235 MeV/c^2 energy to conserve..

Pion will decay to two photons 135/2 = 67.5 MeV photons,

but they are in FoR of pion at rest,

but it's flying in our FoR (frame of vacuum chamber)...

The more energy one photon will take, then less energy 2nd photon will have,

but still sum of them will have 235 MeV..

red-shift of one photon, blue-shift of second photon..

Edited January 29, 2017 by Sensei

[swansont]

These photons, fly in two opposite directions, in FoR of decaying pion, prior decay, maintaining momentum (FoR)..

Then these photons hit two detectors on opposite sides of vacuum chamber..

Photon A from pion hits detector A,

then photon B from pion hits detector B,

from delay, difference of energy, there are calculated initial pion energy + kinetic energy,

and position were meson decayed..

 

How do you determine the position of the decay?

[Sensei]

How do you determine the position of the decay?

 

I gave example of pion-0 with kinetic energy 100 MeV...

 

So total 135+100=235 MeV to conserve..

 

so one detector is detecting

E₁=h*f*(1+v)*gamma

 

while the other is detecting

E₂=h*f*(1-v)*gamma

 

We know E₁+E₂ = 235 MeV

 

We know one photon is red-shifted, while the other one is blue-shifted.

Yet another parameter is delay between arriving photon at detector A, and B...

The longer delay, the further distance photon had to fly in vacuum chamber.

 

 

[swansont]

I gave example of pion-0 with kinetic energy 100 MeV...

 

So total 135+100=235 MeV to conserve..

 

so one detector is detecting

E₁=h*f*(1+v)*gamma

 

while the other is detecting

E₂=h*f*(1-v)*gamma

 

We know E₁+E₂ = 235 MeV

 

We know one photon is red-shifted, while the other one is blue-shifted.

Yet another parameter is delay between arriving photon at detector A, and B...

The longer delay, the further distance photon had to fly in vacuum chamber.

Aren't there potentially two solutions to that? And how do you know the energy of the meson?