Learning Forum

[Kyle]

CHARGE is designed to calculate carrier transport in semiconductors. It uses a complex model for semiconductor materials to achieve this. However, contacts (metals) are always treated as perfect electrical conductors (PEC) in CHARGE.

Even if you stack multiple metals in one structure group, the whole stack is going to behave like a single block of perfect conductor. The metal layers will be treated as a perfect metal when solving Poisson’s equation. The only non-ideal effect you will get is if you enable the ‘Schottky’ contact mode then the work function of the metal in contact with the semiconductor will be used to calculate the band bending at the (metal-semiconductor) interface.

[Jonath]

Hi there, thanks for the reply!
I understand that metals should be treated this way in CHARGE, but to me it seems that it doesn't treat them as simply as you put it, even if forced ohmic (where the work function shouldn't matter).
CHARGE still checks the surfaces of this block (structure group) and differentiates between surfaces of different metals with the same semiconductor or Insulator.
After meshing, it then gives the error: "Electrical contact at surface(s) [...] has more than one type of conductor specified, and the work function is undefined."
The easy way out would be to just set all metals identical or only choose one of the metals as contact (and delete the others)*, but there must be another way.
It doesn't seem to be related to the Surface Recombination BCs either, whereas here it would make sense to differentiate surfaces since different metals have different SRVs with the same semiconductor.

*PS: btw, setting contact to geometry solids doesnt work if the solid is in a group, ie.
setnamed("CHARGE::boundary conditions::electrical_contact", "solid", "::model::geometry::group::solid");
will not work, but not produce an error either.

[kghaffari]

Hi
Like the error suggests, the issue is that one of the surfaces of the specified contact, is shared between different materials. This will be an issue since the work function for surface has to be determined. As you mentioned to address this you can use a single material for all of the contacts. ItÔÇÖs worth noting that itÔÇÖs not possible to use metals in the simulation without them being defined as a contact.
Regarding choosing a geometry for contact, structure groups as a whole would be available for selection, however not the sub-components within them. If you are interested in selecting a particular geometry as contact, consider moving them out of the group. If this is for the metal components mentioned earlier, it would make sense to apply the contact to all the adjacent geometries.