Probably the best way to think about the itinerant case is to imagine a free electron gas (a simple fermi sea) except that the number of up spin electrons is not the same as the number of down spin electrons
The reason we use Hubbard model on a lattice to describe this case is because it is easier to think about electron interactions (which creates the ferromagnetism) when we are on the lattice - but the same physics holds even for interacting electrons that are not on a lattice.
And yes itinerant ferromagnets do have domains. Essentially the same type of argument holds... If you split the system into two parts, each of which is a small magnet, the two parts want to counteralign to minimize the long range magnetic forces.
Does this help at all?
« Back to index