Hi,

My system is GNR with a wire. My electrodes are GNR with three layers of
unit cells(42 atoms, in your Braz J. Phys article, you explain about the
PL layers). These electrodes are connected by a wire. Initially, I have
attached my electrodes directly to scattering region, but the calculations
do not converged (scf transiesta cycle). Then I read the manual and I had
the idea to include more layers of electrodes(42+42+mol+42+42) to
scattering region, the calculation converged with the zero bias value.
Increasing the bias, the calculation does not converge. I tried to raise
the temperature and the calculation converges for 0.1 0.2 0.3 values, but
with value of 0.4 for bias, it not converges again. Such problems occurs
with the DZP basis, with SZ basis the convergence is better. The 50 meV is
too high value to the temperature? I think this is a ~580 K.

Yoshi


> Hi Yoshi,
>
> On 06/24/2010 04:57 PM, [email protected] wrote:
>> Hi all,
>>
>> The option OccupationFunction      MP
>> can affect the scf cicle of transiesta?
>
> No, it wouldn't.
>
> I have a convergence problem which
>> value does not decrease below of the 0.0014 value. In the scf cicle of
>> siesta, the convergence was reached, but the same not occurs for
>> transiesta scf cicle(0.0014). There is any way(change of DM option
>> values)
>> to achieve the convergence in transiesta? I'm running a parallel
>> calculation, this can affect the convergence in transiesta scf?
>>
>
> If you want to send your input files (scattering and electrodes), maybe
> we could help you.
>
> Best regards,
>
> Frederico.
>
>

Attachment: zx6_dzp_elec_hyd.fdf
Description: application/vnd.fdf

Attachment: dzp_f13at+0090_trans.fdf
Description: application/vnd.fdf

Responder a