Tunneling times through a barrier with inelasticity

Tunneling delay times of wavepackets in quantum mechanical penetration of rectangular barriers have long been known to show a perplexing independence with respect to the width of the barrier. This also has relevance to the transmission of evanescent waves in optics. Some authors have claimed that in the presence of absorption or inelastic channels (which they model by taking a complex barrier potential) this effect no longer exists, in that the time delay becomes proportional to the barrier width. Taking the point of view that complex potentials imply non-Hermitian Hamiltonians and are as such fraught with conceptual pit-falls particularly in connection to problems involving time evolution, we have constructed a two-channel model which does not suffer from such limitations in order to examine this issue. We find that the conclusions arrived at by the earlier authors need to be more precisely specified.