A flawed experiment
In all experiments of the quantum eraser type that I know of, the underlying assumption seems to be that the (entangled) state wave collapses at the detector. Hence experimenters always configure the quantum information eraser – a beam splitter – before the detector.
In the experiment of Kim et al. – see picture above – the photon sent by the laser is split – after passing the double slit – into two polarized photons by a birefringent crystal (BBO). The Glan-Thompson prism ensures that those two photons go into two different directions, whereby the angle between those two directions is known. The photon traveling the upper path, the signal photon, is used to detect interference with movable detector D0. The other photon traveling the lower path, the idler, is used to detect which slit has been passed. The quantum eraser is implemented by a semi-transparent mirror – a beam splitter – that erases the path information by one of the two possibilities, reflection or passing through. For more extensive information read this WikiPedia article.
The experiment was flawed because it was realised later on that the observed interference was due to path length differences and not to the information erasure. The quantum wave of the idler photon travels two different paths from double slit to the eraser beamsplitter in the same way as in a Mach-Zehnder interferometer. The observed interference cannot be attributed to the information erasure. The path difference interference explains also the observed complementary interference patterns at D3 and D4 in the original setup. See figure below.
But why would the state wave actually collapse due to a detector that is ultimately a complex superpositioned quantum state wave itself? That assumption is an essential part of the decoherence, superselection, spontaneous collapse or macro hypothesis.
Information erasure áfter the detector – easier to do
But what if it is the information that can be gathered from the experimental set-up that is bringing the quantum collapse about? That is a hypothesis that is experimentally testable with the set-up above. It is an adaptation of the quantum eraser experiment performed by Kim et al. in 1999.
Whether the path information is recorded or not depends in this configuration no longer on a semi-transparent mirror – beamsplitter – placed in front of the detectors D3 and D4 but here depends on the position of electronic switches that are controlled by a QRNG (Quantum Random Number Generator) with zeros and ones. When the switches are in the upper position the detection signal from D3 and D4 is fused together, which means that the information about the selected slit has been erased. In the lowest position the information in the detection signal is retained. Actually a much simpler set-up than the original one. An added advantage is that the same detectors are used to detect the idler photon whether or not path information is being measured. And if the interference at D0 does indeed depend on the position of the switches – top position: interference, bottom position: no interference – then I think it is very clear that it is purely information that influences the interference and therefore decoherence, super selection, spontaneous collapse and macro collapse have been demonstrated as incorrect.
When mechanical switches are used here, this is certainly not a lightspeed delayed choice experiment a lá John Wheeler. Mechanical switches are simply not fast enough to switch in the short time between the moment when the photon signal reaches the D0 detector and the moment when the idler detection signal arrives at the switches. But when we replace those switches with optoelectronic ones that function at the speed of light, I think that could do the job. I am very curious about the result. I expect a confirmation of the role of information. What do you expect?
Proceed your explorational tour here: Ten hypotheses concerning the measurement problem