[Smeagol-discuss] Problems about eigenvalues

[Simon Dubois]

Hello Xiaoyan,

I do not well understand why you wand the eigenvalues in order to
calculate the PDOS and the MPSH. Moreover, if you include the
self-energies in your calculation you have no more insurance that the
effective hamiltonian of the system is hermitian and so you will possibly
obtain complex eigenvalues....

In order to calculate de DOS and the DOS projected on the orbitals, you
should compute the matrix GS (where G is the retarded green function and S
the overlap matrix). Then the trace of GS gives you the DOS and the
diagonal element are the DOS projected on the orbitals. In order to
compute the MPSH, an additional step is needed. You should express the GS
matrix in a new basis set containing the Fock eigenstates on which you
want to project the DOS.
I do not know if there is a public version of smeagol including these
features.... However, I have modified the transm.F routines in order to
compute it (but I do not know if I am authorized to distribute it) and it
appeared to be quite an easy task.


Good work,

Simon

PS: let me know if my answer is completely unappropriate or if I
misunderstood the question ;-)

> Dear all,
>   I try to get the PDOS and the MPSH of the molecular junction systems
> using
> smeagol, so I must get the eigenvalues of the Hamiltonian first.  Firstly,
> I set
> "WriteEigenvalues   T", then run smegaol. After the calculation, no
> eigenvalues
> exported. I can't get the .EIG file and there are no eigenvalues exported
> in the
> .out file. Then set "LongOutput   T", recalculate it, nothing changed.
> What's the
> matter? Thanks a lot!
>   Regards!
> Xiaoyan
>
>
> _______________________________________________
> Smeagol-discuss mailing list
> Smeagol-discuss at lists.tchpc.tcd.ie
> http://lists.tchpc.tcd.ie/listinfo/smeagol-discuss
>