Dielectric tensor and susceptibility tensor mismatch in optical spectrum

[Heinz]

Dear all,

I performed a simple optical spectrum analyses of a GNR sample and got the dielectric and susceptibility data as attached. In the ATK manual page it says that Er=1+X where Er is the relative dielectric constant and X is the susceptibility. However as you can see from the data below there is no such a relation. Is the dielectric or susceptibility data correct? I'm using ATK 2014.1. Thanks in advance.

[Anders Blom]

There is logic in the madness, but I agree something is not completely right. Let me first check, is this a spin polarized or unpolarized calculation?

What you see is that the real part differs not by 1 but by 0.5, or more specifically,

eps_r = susc_r/2 + 0.5
eps_i = susc_i/2

Thus what we seem to have is eps = 0.5+susc/2, which gives an indication of what is wrong.

I'm relatively sure the suscpetibility is correct (because that is what we actually calculate), with the small (big!) reservation that perhaps it's wrong by x2 in which case eps is actually correct... We will investigate and let you know asap.

[Heinz]

Hi,

Could you manage which one is the true value? Thanks.

[Anders Blom]

EPSILON
The bug is fixed in 2014.2, you can verify all there.