Dear Stefano, Thank you for this quick response. I was following Xavier Gonze’s argument in PRB58, 6224 (1998) Eq. 20 where
Z*(u)=dp/du, where p is the dipole moment and u is the displacement. Since p=u.Z(u), we get Z*(u)=Z(u)+udZ(u)/du, where Z(u) is the static charge. So here the Born charge is given in terms of change in charge. I will try the projected DOS to see if the charge transfer is actually that large. Best, Vahid > On Apr 13, 2020, at 12:35 PM, Stefano de Gironcoli <[email protected]> wrote: > > CAUTION: The Sender of this email is not from within Dalhousie. > > I didnt follow completely your argument about the Bader charge but what > one should keep in mind is that the effective charge is the change in > polarization due to displacement not the change in charge... > > think of a core electron shell (as the d orbitals of Pb can > approximately be considered) ... as you move the atom they follow > rigidly (that would make a contribution of 10 not far from your estimate). > > from the PHONON result it looks like Pb gave away 4 all its valence > electrons to S. or rather they are so weakly bound to Pb that they don't > follow it, even if they can still belong its Bader volume. > > it looks a bit extreme but this appears to be the result. You could > compute the atomic projected density of states and see if this seems the > case. > > stefano > > On 13/04/20 17:11, Vahid Askarpour wrote: >> Dear QE Community, >> >> I have calculated the Born charges using the PHONON code for PbS. The only >> non-zero elements are the diagonal ones and are 4.122 and -4.168, >> respectively. >> >> In the zstar_eu.f90, Born charges consist of two terms as seen below: a part >> due to polarization calculation (dynamic) and the other is zv (static) which >> is the z_valence according to read_upf_v2.f90. >> >> do ipol = 1, 3 >> do na = 1, nat >> zstareu (ipol, ipol, na) = zstareu (ipol, ipol, na) + zv (ityp ( na) >> ) >> enddo >> enddo >> >> The zv values for Pb and S are 14 and 6 given in the PSP. If we subtract zv >> from the Born charges, we get the term due to polarization: -9.878 and >> -10.168. These values seem too large because of the argument below. >> >> To estimate the polarization term, I reduce the alat by 1% and relax the >> atoms. This shifts the atoms from the unstrained position. I calculate the >> Bader charges for the unstrained and the strained cases. The change in the >> Bader charge is related to the atomic displacement. I have also tried >> keeping alat fixed and moving the atoms by 1%. >> >> For unstrained PbS, the Bader charges are 12.998 and 7.001. >> For the strained PbS, they are 13.004 and 6.995. >> >> So a ~1% change in atomic positions results in a +/-0.006 change in Bader >> charge. From this calculation, I expect the contribution from polarization >> to be u(dZ/du), where u is interatomic distance, which amount to +0.6 for Pb >> and -0.6 for S. >> >> The contribution from polarization I get (0.6 and -0.6) are quite different >> from the those of the PHONON code (-9.878 and -10.168). I am assuming that >> the code is correct and my logic is flawed. I would appreciate any thoughts >> you may have on this discrepancy. >> >> Thank you, >> Vahid >> >> >> Vahid Askarpour >> Department of physics and atmospheric science >> Dalhousie University >> Halifax, NS >> Canada >> _______________________________________________ >> Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso) >> users mailing list [email protected] >> https://lists.quantum-espresso.org/mailman/listinfo/users > _______________________________________________ > Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso) > users mailing list [email protected] > https://lists.quantum-espresso.org/mailman/listinfo/users _______________________________________________ Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso) users mailing list [email protected] https://lists.quantum-espresso.org/mailman/listinfo/users
