In my last entry on Hamilton’s method for systems with constraints I left some unfinished work: The general derivation that I gave is not completely correct; although the method does work, as one can see clearly in the example. As soon as I have some time I will add the explanations.

Now I want to add some comments on certain “exotic” quantum states that keep on appearing in the literature over the last years giving rise to a certatin amount of surprise. These states are basically attached to photons, but very similar arguments are presumably valid for particles satisfying the Schrödinger equation, as the paraxial approximation of optics is formally equivalent to a Schrödinger equation.

These states are known as Laguerre-Gaussian, Hermite-Gaussian, etc. Basically it consists in the existence of quantum states representing possible evolution modes for particles freely propagating along one direction and with an orbital angular momentum packaged within the beam, identifyable with an orbital motion in the direction perpendicular to the one of free motion. Thus, the component, say, is a free-propagating Gaussian packet, while the and components (the factor corresponding to the wave function, that is), are represented by, e.g., a Laguerre or Hermite polynomial. This is paradoxical because we have been taught that free evolution always leads to dispersive or spreading wave packets, and yet these states seem to be carrying along a confined, orbital motion in their evolution.

The reason that we find this paradoxical is that free propagation is incorrectly accounted for in quantum mechanics books. I will try to explain this. The usual procedure of axiomatically building quantum mechanics is based on Von Neumann, and it tells us that there are complete sets of compatible observables. Once these compatible observables expanding any state and containing all the possible statistical information on the state are identified, they are instrumental to represent the evolution of an arbitrary quantum state. In the case of linear momentum, there are three compatible operators (commuting in pairs), , , .

This assumes that, in a laboratory, it is feasible to guarantee for a free particle to be an eigenstate of , *and* , *and* . That’s not the way it works: First, it has to be an eigenstate of , where is the filtering direction chosen, but not necessarily of and . In fact. *diaphragms and collimators, slits, etc., that are used are essentially physical obstacles conditioning in position coordinates* and , and consequently representable by “obstacle potential functions” , so they will have produced a certain profile that neither is a free-propagating packet nor does it have to be.

Thus the natural thing to say is that a free particle with a *selected linear momentum* is an eigenstate of , where is the chosen direction of filtering, and that in principle we haven’t the slightest idea what it is in and . People generally do not see this, because in general they consider that the distinction between filtering or non-filtering state preparations are a nicety or unnecessary erudition, when actually it is absolutely crucial. Let us go directly to the relaxation of the statement “we haven’t the slightest idea”, as the truth is we can postulate how the state in and looks like in a very natural way.

If after a certain relaxation time the particle is indeed free, we can posit that it must be an eigenstate of the free Hamiltonian (kinetic energy),

We add now the demand that it be an eigenstate of the filtered linear momentum in the chosen direction. We have,

This amounts to,

Thus the state that represents a particle free-propagating with momentum determined in , perhaps not the most general one, but definitely general enough, is a function that is harmonic in , multiplied by an eigenstate of .

## Leave a Reply