Yu-Chih Tseng

simulating InAs in Schred V1 vs V2

In version 2, I tried the following parameters for InAs:

L-valley energy offset: 0.73
X-valley energy offset: 1.02

effective masses: m(gamma)=0.023
m(L)=0.29
m(X)=0.64

The C-V curves I got are very different from Schred version 1 which has InAs. Which version is more accurate? Please also verify if the InAs parameters I used are correct.

Thanks.

1. Hi Tseng, The values you are using seem to be right. But the reason for the difference is as follows: Schred V2.0 assumes a generalized 3 conduction band valley model for alls emiconductors, and uses principal effective masses and the principal directions for the different valleys to compute its own values for the transport,width and confinement masses. But other materials (other than silicon), cannot use the same approach to derive the transport,width and confinement masses from the principal effective masses and its direction (Imagine if this were possible their would be much less need to compute bandstructure calculations for different materials and obtain masses from them). So Schred V2.0, makes an approximation in that, it uses the principal effective masses and computes the geometric mean to calculate the transport,width and confinement masses. The advantage of this is that, you can treat any semcionductor other than Silicon that can be approximated with a 3-conduction band valley model, and it will be reasonably accurate(while for silicon/strained silicon it will be very accurate). But whereas Schred V1.0, treats only 3 materials,Si,GaAs,InAs. As a result of that, it also has an advantage, it does not make any approximations in the masses for these specific cases and assumes a 2 valley system for silicon, 4-valley system for GaAs and InAs (gamma,l,X(primed),X(unprimed)). So for the InAs/GaAs cases, since it uses the exact masses from the bandstructure caluclations for the InAs/GaAs, you will get a capacitance curve that is more accurate than Schred V2.0.

Hope it helps, thanks, gokul