Do we not know that momentum is certainly not conserved at single electron double slit experiments? I am aware that tons of experiments are performed on this topic and the scientists try to be very careful with the experiments and take every possibility into account.
In a single electron double slit experiment, we know the limits to the direction of the electron's momentum since we know the path it travels from the gun to the slits. After the slits, we observe that many electrons go beyond these limits/range in direction, producing an interference pattern on the screen. This phenomena is explained by probability waves.
But, isn't it weird that a particle changes direction (therefore momentum) by itself? It looks like it would be a more robust explanation if something manipulated the electron or transferred energy to it. Ofc, I can say 'well, in the quantum world things are different; follow the formula; think out of the box; etc.', but that is how I feel at the moment.
In summary, what is the possibility that the interaction pattern on the screen is just due to the electrons interacting with the atoms at the edges of the slits? I don't know what causes the periodicity/pattern on the screen but the explanation  might be inside the atom, impossible?
 A: The slit is an aperture in a material absorber which is much more massive than the electron. Momentum is conserved: the electron's new momentum is balanced by a reaction in the absorber. The larger the absorber, the harder it is to detect its change in momentum.
 [source]
A: The two slits experiment gives the expected result regardless of whether the particles being sent through the slits are photons, electrons, neutrons, etc, and regardless of whether the barrier into which the slits are let is made of a conductor or an insulator etc.
If the interference pattern on the screen were due to some kind of interaction between the particles being defracted and the particles that formed the boundary of the slits, then you would need to have a convincing theory to account for the fact that the interference pattern seems only to depend on the wavelength associated with the particles being sent through the slits and not on any other factor.
A: Two things:

*

*The particles do not interact with the edges of the slits. If they do, they lose coherence and that's that. This has been tested experimentally. Rather, they interact (so to speak) with the paths that are not available to them because of the barrier.

That a path not not-taken, but not available, can affect the the evolution of a wave function may sound like quantum woo...it isn't. It's just linear QM. Further insight can be gathered from the Elitzur-Vaidman bomb-tester, where the lack of an available path (or presence thereof), affects the evolution elsewhere.


*In the Schrödinger equation set up of this, one would model a slit as a small region of $V(x)=0$, $V(x)=\infty$ elsewhere ($x$ is the transverse direction). With that, the potential lacks symmetry under translations in $x$, so that there should be no expectation that $p_x$ is conserved.

A: It's not a bug, it's a feature
It would be indeed a weird behavior for a particle. On the other hand, it is quite usual behavior for a wave - indeed, double slit experiment corresponds to rather usual diffraction phenomena that one can observed for the electromagnetic waves, waves on water, acoustic waves, etc. Thus, double slit experiment demonstrates the wave-like properties of electron (aka wave-partcile duality), which is precisely the point.
