Entangled photon polarization thought experiment disproof request

[Freeborn]

I was trying to search for a similar topic, but as I haven't found one here it goes.

Thought experiment: Periodically emitted chain of single photons of non-polarized light is split into two entangled pairs and they are traveling apart from each other. One going toward point A, the other toward point B.

After traveling a distance of d the signal photon is polarized in point A either with an alpha OR an alpha + 90° rotated polar filter.

After traveling a distance of d + epsilon the idler photon (the other one) goes through another polar filter in point B which is positioned in the same alpha angle respective to the one at A.

Desired (probably false) consequence:
By periodically detecing or not detecting consectuive photons in point B the observer is able to determine wheteher the polar filter in point A was rotated to alpha OR alpha + 90° position.
Successful detection -> bot filters are in the same pos,
Failure at detection -> filters are ortogonal respective to each other.

As we don't move the filter in point B we can deduce the angle of the filter at point A based on detecting or not detecting photons in point B.

I am pretty sure that the conclusion is wrong, still I don't see how.
If it could happen information could be transferred instantaneously between huge distances.
I would be very grateful for a brief explanation of where the mistake is.

EDIT: I came upon this thing while reading the Dancing Wu-Li Masters.

 

[unusualname]

50% of the photons will be detected at point B regardless of how the polarizers are set.

To corellate the ploarizer orientations you would need to do coincidence matching with photons detected at points A and B, and that would require subluminal (classical) communication to compare the detection data at either end.

 

[DrChinese]

Ditto what unusualname said.

 

[Freeborn]

First of all:
I am grateful for the answers,
Thank You!

What happened if the polarizers were oriented once via regular communication? And then the orientation would be kept, and only one of them would rotate between 2 positions.
I'm affraid I didn't grasp your answer to a full extent. Could you elaborate a bit more?
Thx once more!

 

[unusualname]

I don't see why that makes any difference. Polarizers don't enable you to influence individual photons, they just filter the incoming photons according to Malus's law, and for an unpolarized sample of photons ~50% (less in practice) will pass through a polarizer. You can't know anything about the behaviour of the entangled partners without doing coincidence counting.

 

[yuiop]

As we don't move the filter in point B we can deduce the angle of the filter at point A based on detecting or not detecting photons in point B.

Let us say initially A is at the 0° position. 50% of the time the photon arriving at A is orientated at more than 45° resulting in a non detection at A. B has the same orientation as A, so B detects a photon 50% of the time (Same as A). Now A is rotated to 90 degree so that A and B are orientated at 90 degrees to each other. When the two detectors are orientated like this respective to each other, they agree 50% of the time. A continues to detect 50% of the arriving photons randomly. 25% of the time B now gets a detection when A gets a detection, but 25% of the time B gets a detection when A does not get a detection. The result is that after A has rotated his detector, B continues to detect photons 50% of the time and is completely unaware that A has rotated his polariser. It is only later when they come together and analyse past results that they observe the correlations. Therefore A cannot transmit useful information to B by manipulating his polariser and certainly entangled photons can not be used to transmit useful information any faster than the speed of light.

 

[Freeborn]

Thank both of you for further details. (it makes sense for me as well now)