I am studying nano-scale constrictions in semiconductor materials. Imagine two nano-particles joined together by a single point of crystalline contact to look like an hour-glass filled with sand. Since this structure consists of around 3000-4000 atoms for a unit cell, running DFT based calculations is not feasible with the computational resources available at hand.
The study requires the comparison of electron scattering between several of such structures, all with similar shape but with different cross sectional area at the constriction. Also the variation of surface area and volume between the individual structure configurations is relatively insignificant compared to the cross section of the constriction.
In this case, would it be acceptable to simply take the cross section of the constriction (modeled as an infinitely long section), and evaluate the electron relaxation time while neglecting the larger sections?
I am aware that looking at the complete structure would be the best way to go, but it is not feasible for the study. If there is an alternative feasible approach, I'd be happy to try it out.