Anonymous @ on
Effective mass variation in pcpbt is not working correctly for tight binding
For a double barrier structure doubling the barrier height or doubling the effective mass should roughly have the same effect (for low lying energy states, say when the barrier is already high and the well wide). The overall effect is interestingly enough that the resonances move UP in energy when the barrier heights are increased. Tight binding and transfer matrices give roughly the same result.
However if the effective mass is made inhomogeneous and the effective mass is doubled in the barrier only, then the TB resonance moves down in energy and the transfer matrix element-based resonance moves UP in energy.
I would suspect that there is a bug in the computation in the tight binding Hamiltonian. There are some notes in a book edited by Prof. Frensley from around 1994 that gives details on the appropriate discretization of the effective mass Hamiltonian when the effective mass changes spatially.
Please fix this problem.