While performing Amset calculations, I observed that at 600 K for p-type dopings, some of the Seebeck coefficients values are negative. The band gap of the system is 0.28 eV. I understand that experimentally system may show n-type to p-type transition and accordingly such values. However, in our calculations, we define p-type dopings, then why such negative values.
1.00e+18 -1.56e+02
2.00e+18 -9.29e+01
3.00e+18 -2.42e+01
4.00e+18 4.22e+01
5.00e+18 1.01e+02
6.00e+18 1.51e+02
7.00e+18 1.93e+02
8.00e+18 2.26e+02
9.00e+18 2.53e+02
1.00e+19 2.75e+02
Just like lbwnb123q3 said, this could happen when you have a material with relatively narrow band gap.
Both conduction band (CB) and valence band (VB) contribute to the Seebeck coefficient. When the temperature is high, the Fermi window will be enlarged. As a result, even if the Fermi level is close to VB, the contribution from CB can somethimes still be significant.
That being said, -156 uV/K at a carrier concentration of +1e18 is still unsusual. I would assume this is because your material have a rather flat CBM and a dispersive VBM or the DOS effective mass of CBM is way larger than VBM (i.e. large band degeneracy CBM). Under these circumstances, the contribution form CBM will be much larger than VBM causing the negative Seebeck at hole doping. Please check the band structure and see if this is the case.