Imagine a particle in a momentum eigenstate $\vec p$ such that its wavefunction is a plane wave (or a wave packet with some width $\Delta p$ if we want this to be more realistic). Now imagine that this plane wave is elastically colliding with a thin rigid plate that is exactly perpendicular to $\vec p$. What will be the state of the plate after the collision?
If we view the particle's wave function as a superposition of localized states, each of those has some angular momentum relative to the center of the plate, and thus will cause the plate to rotate. So by the linearity of the Schrödinger equation, the plate will end up in a superposition of rotating states with different angular momenta (which will average to zero, but when measuring it we'll get some non zero value).
On the other hand, if we think of the time evolution of pure momentum states, it also seems feasible that there will only be a transfer of linear momentum, so that the plate will just move forward without rotating.
Which is correct?