The spectrum of a magnetic Schrödinger operator can be discrete
due to the growth of the electric potential alone, but also due to a regular
growth of magnetic field because of noncommutativity of magnetic translations
which can be used in an uncertainty principle type argument. I will
explain new conditions for the discreteness of spectrum which combine the
influence of electric and magnetic fields by introducting effective potentials.
(These new results are joint with V. Kondrat'ev.)
Here are some open problems concerning qualitative behavior of the spectrum of this operator.
1. Establish simple necessary conditions for the spectrum to be discrete in terms of the electric potential V and magnetic field B.
An example of a necessary condition: if V
0 and the spectrum of HA,V is discrete,
then for any fixed r > 0
![]()
as |x| ![]()
.
Here B(x,r) is a ball of the radius r with
the center at x.
However this condition is not good because it is reasonable to expect that both V and B enter with the same power. This motivates the following more concrete question:
1a. Can we replace |B|2 by |B| in the necessary condition above?
2. Find a necessary and sufficient condition for the spectrum of HA,V to be discrete so as to extend the result of A. M. Molchanov (1953) who established such a condition in case when V is semi-bounded below and B = 0 (on Rn).
The Molchanov condition is formulated in terms of the Wiener capacity. So the following question naturally arises:
2a. Is there a good notion of magnetic capacity?
3. Relate a localization of classical particles moving in an
electromagnetic field, with a quantum localization which means a decay
of eigenfunctions of HA,V at infinity.
For example, the quantum localization may mean the property of the eigenfunctions
to be in L2 or, stronger, the discreteness of spectrum
of HA,V.