[Smeagol-discuss] gold contact error

[Ivan Rungger]

Dear Tao K,

 First the second question: the Mulliken print statements are caused by
a small bug in siesta. There is an option MullikenInSCF that if set to
true causes the Mulliken population to be printed at each scf step. The
option is read in on the mother node, but it is not broadcast to the
other nodes when run in parallel, so that the value of the option is
undefined on the other nodes. To remove this bug just add the line
"call MPI_Bcast(muldeb,1,MPI_logical,0,MPI_Comm_World,MPIerror)" after
the line
"call MPI_Bcast(initdmaux,1,MPI_logical,0,MPI_Comm_World,MPIerror)" in
siesta.F.

For the question about the transmission: I went through your input files
and they seem OK to me. Since the numbers for the transmission depend on
the Fermi energy E_F, can you please write the value of E_F, and maybe
send in the TRC calculated over a range of a few eV around E_F?

Cheers,

 Ivan


Tao K wrote:
> Dear all,
>
> I want to repeat the transmission calculation for the gold contact (Fig
> 5 in PRB 73, 085414, 2006).  The transmission coefficient what I have
> got is very small, such as:
>
>             E (Ry)                        Tran                       Tran
>   -0.3000000D+00   0.2380674D-03      0.2380674D-03       53
>   -0.2941748D+00   0.2149487D-03      0.2149487D-03       49
>   -0.2883495D+00   0.2688537D-03      0.2688537D-03       49
>   -0.2825243D+00   0.7296415D-03      0.7296415D-03       57
>   -0.2766990D+00   0.3890249D-02      0.3890249D-02       57
>   -0.2708738D+00   0.9409617D-03      0.9409617D-03       57
>
> Moreover, in my output file, I found a strange problem. When I ran the
> program in parallel, for each step, I got:
>
> siesta: iscf   Eharris(eV)      E_KS(eV)   FreeEng(eV)   dDmax  Ef(eV)
> siesta:    1  -307340.2957  -307047.3737  -307047.3737  0.5539  0.0000
> timer: Routine,Calls,Time,% = IterSCF        1  354969.421  99.54
> siesta: Mulliken populations before mixing
> siesta: Mulliken populations before mixing
> siesta: Mulliken populations before mixing
> siesta: Mulliken populations before mixing
> siesta: Mulliken populations before mixing
> siesta: Mulliken populations before mixing
> siesta: Mulliken populations before mixing
> siesta: Mulliken populations after mixing
> siesta: Mulliken populations after mixing
> siesta: Mulliken populations after mixing
> Siesta: Mulliken populations after mixing
> siesta: Mulliken populations after mixing
> siesta: Mulliken populations after mixing
> siesta: Mulliken populations after mixing
> siesta:    2  -315860.2739  -306903.7767  -306903.7767  1.9262  0.0000
> ... ...
>
> However, when I ran the same input file in sequential, I did not find
> this problem.
>
> I think many of us have repeated this calculation already, any
> suggestion would be appreciated. Thanks in advance!
>
> Best wishes,
> Kun Tao
>
>