Has a double-slit experiment with detectors at each slit actually been done? My understanding of the famous double-slit experiment is that, when both slits are open, a wave-pattern interference arising from the two slits is observed. This is a different pattern than summing the pattern from each slit individually. This is not unusual as the same thing happens with water and sound waves, for example.
The odd/novel things (with regards to classical physics) are:

*

*the interference pattern is observed even if you only send one photon or electron etc through at a time (i.e. it appears the particle 'interferes with itself'). And

*If you detect which slit the photon or electron etc goes through, there is no more interference pattern - you observe the sum of the individual slits. i.e. the observation destroys the interference pattern because the position became known, so the photon could not 'interfere with itself' any more. And this holds even if you only have a detector on one slit but not the other.

My question is - has an experiment actually ever been performed to validate #2?
As per What is the *detector* in the double slit experiment and how does it work? , the top answer refers to "a long series of debates between Bohr and Einstein" -- which is not an experiment but rather a conversation.
Likewise Wikipedia states that Feynman "proposed (as a thought experiment) that if detectors were placed before each slit, the interference pattern would disappear." -- that is, a thought experiment, not an actual experiment.
None of the answers at this question Young's double-slit experiment with detectors cite an actual experiment that was performed.
I found a link to an article titled "Which-way detector unlocks some mystery of the double-slit experiment", but it doesn't seem to answer the question. It appears they simply change the setup of the experiment into one where there is no interference.
I'm presuming the answer is 'yes' but I haven't been able to find anything yet.
UPDATE: the answer here links to one experiment that uses a 'blocker' rather than a detector. The second link points to a YouTube video showing measurement of photon path by using polarization filters. But this doesn't demonstrate that it is knowledge of the photon's path that resulted in the wavefunction collapse, but rather that horizontally and vertically polarized photons do not cause an interference pattern.
Further the 'eraser' of a 45-degree polarized filter, doesn't demonstrate that erasing the knowledge of which slit the photon passed through re-creates the interference pattern, but rather simply that photons that have the same polarization do cause an interference pattern. i.e. the photons, after being horizontally and vertically polarized, then recombine to have the same polarization, which then results in an interference pattern again.
UPDATE 2: This part of the video describes exactly this experiment (placing an electron detector at the slit): https://youtu.be/U7Z_TIw9InA?t=243 , but I've been unable to find a paper about this or reference to an actual experiment being done.
 A: The experiment in the link you give, does answer your 2), but one has to follow the mathematics .
In this earlier experiment :

We give a measure of particle knowledge in a neutron interferometer that reflects one's ability to predict in which beam a neutron is located. We can measure wave knowledge by contrast of the interference pattern. Then one's simultaneous knowledge of both is determined by a single parameter (not an uncertainty relation), running from full particle to full wave knowledge. We extend the discussion to partially coherent beams. Our measure of information is much simpler than the conventional one.

It seems to me that you would like to see a simple experiment to be performed in class, but quantum mechanics needs complicated equipment and set ups. The experiment demonstrates that interference is crucial in studying the behavior of particles at the size of neutrons.
See for example the experiment for the single electron. It uses a filament instead of two slits, but the mathematics allows to make the statement of double slits. Due to the small dimensions of the filament it is not possible for the specific experiment to have which way detectors.
Once interference effects prove the probability-wave nature of quantum mechanics, the mathematics of quantum mechanics theories can translate the results of  the beam of neutrons quoted above, and translate the results of  your linked experiment to the double slit expectation/thought-experiment. It is not necessary to have two slits if the same mathematics can demonstrably be used in an analogous experiment.
Thus, the claim of the referenced experiments, of "detectors at slit" does relate to the double slit with interference thought experiments, through the mathematics.
A: Regarding point (2) when the photon's slit is known: in this case each slit produces a diffraction pattern, and they are summed incoherently.
In a diffraction pattern, the photon is still "interfering with itself", as one has to sum over all paths through the single slit, coherently. Otherwise, you would get a result consistent with geometric optics.
On the other hand, if there is no slit, you do get geometric optics. The photon just goes straight, so one could say that the diffraction pattern is produced by the photon no-longer interfering with the paths that are now blocked by that part of the apparatus that is not the slit.
tl;dr the behavior is always wave behavior, and you always have to add the amplitudes for all possible paths, coherently.
