News1 min ago
Changing the past
7 Answers
Brian Greene describes an experiment (The Fabric of the Cosmos p194) in which a photon beam passes through a beam-splitter. Each beam then passes through a down-converter (which emits two half-energy photons for each input photon). One photon from each down-converter is sent into space for ten years, the others shine on a screen on Earth giving, or not giving, interference effects.
If the photons which have been sent into space are detected separately, no interference is seen between the photons back on Earth (because they are "tagged"). If the paths of the photons in space are merged, so that they are detected by the same detector, and are thus indistinguishable, then interference occurs back on Earth.
Suppose we use a mirror on Mars to reflect the photons which were sent into space. When they arrive back on Earth a few minutes later we can decide to combine their paths or detect them separately. So we can choose whether or not we have just seen interference from the other photons!
Please explain.
If the photons which have been sent into space are detected separately, no interference is seen between the photons back on Earth (because they are "tagged"). If the paths of the photons in space are merged, so that they are detected by the same detector, and are thus indistinguishable, then interference occurs back on Earth.
Suppose we use a mirror on Mars to reflect the photons which were sent into space. When they arrive back on Earth a few minutes later we can decide to combine their paths or detect them separately. So we can choose whether or not we have just seen interference from the other photons!
Please explain.
Answers
1 to 7 of 7
I think there is a problem with this experiment. When your photons enter the "down converters" they are interacting, in effect being measured and are no longer necesarilly in the same quantum state. The same thing happens when you manipulate the photons to merge them or not. However let's assume we could fix these problems and put that on one side this seems a...
09:55 Wed 25th Mar 2009
You don't need to split photons into two, just use a photon beam on a beam splitter into an interferometer. (Two arms at 90 degrees with mirrors at the end.) When the photons come back down the arms and recombine they form interference effects, fairly sensible.
But if you send one photon at a time through it, i.e. wait for a photon to come out before sending another in, then the photon has a 50/50 chance of going up either arm. If a sensor is in one of the arms then the arm each photon goes through can be determined. If this is so a simple bright dot appears when the photon comes back out, and subsequent photons just intensify the dot. Again sensible, a photon goes in, definitely goes through one arm and comes out as a single particle every time.
But if no sensor is present and the path of the photon is unknown then subsequent photons land in different places and eventually build up an interference pattern as expected from putting a continuous wave into the interferometer. In other words each single photon seems to travel down both arms and interfere with itself, taking a random position on the interference pattern, subsequent photons interfere with themselves and slowly build the pattern up.
Final bit of weirdness- if the sensors are on a time delay i.e. they say which arm the photon went through after it has come out then the single bright spot is formed as if the photons knew they would be measured in the future.
As for an explanation it is tricky, probs not best left to the mere efforts of ABers, I would recommend the excellent "Quantum -a Guide For the Perplexed" by Jim Al-Khalili. Not an insanely sciency book (for that see Quantum Physics by Eisberg and Resnick) but a brillaint explanation of the whole thing (complete with brilliant pictures).
But if you send one photon at a time through it, i.e. wait for a photon to come out before sending another in, then the photon has a 50/50 chance of going up either arm. If a sensor is in one of the arms then the arm each photon goes through can be determined. If this is so a simple bright dot appears when the photon comes back out, and subsequent photons just intensify the dot. Again sensible, a photon goes in, definitely goes through one arm and comes out as a single particle every time.
But if no sensor is present and the path of the photon is unknown then subsequent photons land in different places and eventually build up an interference pattern as expected from putting a continuous wave into the interferometer. In other words each single photon seems to travel down both arms and interfere with itself, taking a random position on the interference pattern, subsequent photons interfere with themselves and slowly build the pattern up.
Final bit of weirdness- if the sensors are on a time delay i.e. they say which arm the photon went through after it has come out then the single bright spot is formed as if the photons knew they would be measured in the future.
As for an explanation it is tricky, probs not best left to the mere efforts of ABers, I would recommend the excellent "Quantum -a Guide For the Perplexed" by Jim Al-Khalili. Not an insanely sciency book (for that see Quantum Physics by Eisberg and Resnick) but a brillaint explanation of the whole thing (complete with brilliant pictures).
I think there is a problem with this experiment.
When your photons enter the "down converters" they are interacting, in effect being measured and are no longer necesarilly in the same quantum state. The same thing happens when you manipulate the photons to merge them or not.
However let's assume we could fix these problems and put that on one side this seems a variation of the EPR paradox.
I'm guessing you've come across this but for those who've not you have the emission of two photons in the same quantum state, when one is measured the other appears to change state simultaneously despite apparent light years of separation.
http://en.wikipedia.org/wiki/EPR_paradox
Because this appears to be faster than light travel it can be seen as time travel (the two concepts are deeply linked)
The Photons are in what is known as a quantum entangled state
http://en.wikipedia.org/wiki/Quantum_entanglem ent
Now we come to what you're not going to like.
When you talk of explaining and Quantum Mechanics you're suddenly on very shakey ground QM doesn't explain what is going on, it gives you very powerful tools to predict the outcome.
There are numerous attempts to describe the underlying "nature" of such things these are known as Interpretations Such as the Copenhagen, the Many worlds and the Bohm
I'll let you investigate these because there's far too much for an AB reply
When your photons enter the "down converters" they are interacting, in effect being measured and are no longer necesarilly in the same quantum state. The same thing happens when you manipulate the photons to merge them or not.
However let's assume we could fix these problems and put that on one side this seems a variation of the EPR paradox.
I'm guessing you've come across this but for those who've not you have the emission of two photons in the same quantum state, when one is measured the other appears to change state simultaneously despite apparent light years of separation.
http://en.wikipedia.org/wiki/EPR_paradox
Because this appears to be faster than light travel it can be seen as time travel (the two concepts are deeply linked)
The Photons are in what is known as a quantum entangled state
http://en.wikipedia.org/wiki/Quantum_entanglem ent
Now we come to what you're not going to like.
When you talk of explaining and Quantum Mechanics you're suddenly on very shakey ground QM doesn't explain what is going on, it gives you very powerful tools to predict the outcome.
There are numerous attempts to describe the underlying "nature" of such things these are known as Interpretations Such as the Copenhagen, the Many worlds and the Bohm
I'll let you investigate these because there's far too much for an AB reply
Jake, I think you are correct that there is a problem with the experiment, and it is probably with the �down converters�.
I can�t find a description of this type of �down converter� on the web. Brian Greene implies that it is a device which effectively gives two �half photons� (full photons with half the energy). So, instead of the normal two-slit experiment, which gives interference if we don�t peek en-route and doesn�t give interference if we peek, we can have our cake and eat it, i.e. we obtain interference from �half� the photon, then detect its other �half� (which will cancel the interference that we have already seen).
The resolution of the EPR �paradox� was non-locality and in this experiment we have non-locality of time, although not of space, but you surely do not believe that the future detection of �half� the photon can destroy previously-observed interference of the first �half�?
I don�t even think that the particular interpretation of collapse of the wave function matters. This is a perfectly feasible experiment where the result is guaranteed, but guaranteed to be wrong.
I�ll assume that Brian Greene has mistaken the properties of �down converters�.
I can�t find a description of this type of �down converter� on the web. Brian Greene implies that it is a device which effectively gives two �half photons� (full photons with half the energy). So, instead of the normal two-slit experiment, which gives interference if we don�t peek en-route and doesn�t give interference if we peek, we can have our cake and eat it, i.e. we obtain interference from �half� the photon, then detect its other �half� (which will cancel the interference that we have already seen).
The resolution of the EPR �paradox� was non-locality and in this experiment we have non-locality of time, although not of space, but you surely do not believe that the future detection of �half� the photon can destroy previously-observed interference of the first �half�?
I don�t even think that the particular interpretation of collapse of the wave function matters. This is a perfectly feasible experiment where the result is guaranteed, but guaranteed to be wrong.
I�ll assume that Brian Greene has mistaken the properties of �down converters�.
Sym, yes Brian Greene agrees that the photons will form a bright spot if we perform the measurement in the future on the other �halves� (i.e. on the accompanying photons emerging from the down-converters), but he says we will obtain interference fringes if the paths of the other �halves� are merged so that it is impossible to detect which of the original paths was taken. We can decide in the future whether to merge the paths or not.
So let�s decide that we are going to merge the paths. Great, we see interference fringes. Then we change our minds and don�t merge the paths. The interference fringes, which we have already seen, go away.
So let�s decide that we are going to merge the paths. Great, we see interference fringes. Then we change our minds and don�t merge the paths. The interference fringes, which we have already seen, go away.
No that's not what I'm saying
Let's assume we can somehow magically create 4 quantum entangled photons - down-converters, whatever it's not important there they are.
Surely what we already know about quantum entanglement tells us that the moment thet the first two interfere (and hence are observed) the other two change state into an arbitary phase.
This is just the same as the experimental proof of the EPR paradox done in '82 by Alain Aspect in Paris, just doubled up.
Let's assume we can somehow magically create 4 quantum entangled photons - down-converters, whatever it's not important there they are.
Surely what we already know about quantum entanglement tells us that the moment thet the first two interfere (and hence are observed) the other two change state into an arbitary phase.
This is just the same as the experimental proof of the EPR paradox done in '82 by Alain Aspect in Paris, just doubled up.
So, changing your mind as to whether to merge the paths or not should necessarily create or destroy fringes classically. You are not merging and de-merging the same photon/s. Your reply suggested merging some photons and seeing interference, then keeping the paths separate for a second bunch of photons and not seeing the fringes as predicted by classical wave theory.
As for the original question, observing interference patterns or not from the first pair of photons will allow you to make measurements on the second pair assuming the photons produced by the down converter are exactly like the input photon but with half energy (hmmm). By looking at the patterns you can decide the intensity and hence the energy of the photons plus by studying interference patterns you can say something about the photons relative to each other.
I have to say the experiment doesn't sound very remarkable as with a beam of photons input then it sounds like an exercise in long range wave theory. Split one wave into two then each wave into another two, any two of these four waves will interfere in the same way if the path lengths are equal. By sending two straight to screen and another two into space you're just using two sets of waves with different path lengths. Unsurprisingly both sets either interfere or neither do.
" So we can choose whether or not we have just seen interference from the other photons! " No you can't the pattern will either be there straight away or never appear, not both.
As for the original question, observing interference patterns or not from the first pair of photons will allow you to make measurements on the second pair assuming the photons produced by the down converter are exactly like the input photon but with half energy (hmmm). By looking at the patterns you can decide the intensity and hence the energy of the photons plus by studying interference patterns you can say something about the photons relative to each other.
I have to say the experiment doesn't sound very remarkable as with a beam of photons input then it sounds like an exercise in long range wave theory. Split one wave into two then each wave into another two, any two of these four waves will interfere in the same way if the path lengths are equal. By sending two straight to screen and another two into space you're just using two sets of waves with different path lengths. Unsurprisingly both sets either interfere or neither do.
" So we can choose whether or not we have just seen interference from the other photons! " No you can't the pattern will either be there straight away or never appear, not both.
Sorry Jake, I�m being thick: I don�t see this.
Aspect effectively says that if we measure the second pair then the first pair can�t interfere. But Brian Greene assumes that if we merge the paths of the second pair, so that we can�t tell which is which, we will obtain interference from the first pair. However, if the first pair interferes we have time to decide not to merge the paths of the second pair.
Sym, I�d agree with you unreservedly but Brian Greene isn�t a complete duffer. Perhaps he�s mistaken this time though.
Aspect effectively says that if we measure the second pair then the first pair can�t interfere. But Brian Greene assumes that if we merge the paths of the second pair, so that we can�t tell which is which, we will obtain interference from the first pair. However, if the first pair interferes we have time to decide not to merge the paths of the second pair.
Sym, I�d agree with you unreservedly but Brian Greene isn�t a complete duffer. Perhaps he�s mistaken this time though.
1 to 7 of 7
