[Smeagol-discuss] Convergence with tPA lead

[Ivan Rungger]

Hi Lienheng,
> Thanks  for the reply.  I think I did include the two leads in the
> scattering region.  The device is benzene, i.e. 10 atoms, and each
> lead contribute to 8, giving total of 26.  I actually copy and pasted
> the lead coordinates from the lead calculations.  I have attached the
> lead files here.  Notice that left leads and right leads are
> different, they are mirror images of each other.
This might be the problem, in the current version of smeagol to treat
different leads you would have to use a separate utility that Alex
wrote. There is an email about this I think in the mailing list. But for
your system you might just use the same leads on both sides by making
the cell a bit longer. I would make it a bit longer anyway (i.e. add a
few layers of leads on each side), since it is quite a small system.
>
> One thing I have noticed though,  convergence worsens -- the energy
> fluctuates from >+2000 to < -2000 from time to time -- when I change
> the EnergyLowestBound value.   Also, since the scattering region
> usually contains quite a few atoms (including two lead unitcells), the
> lowest eigen energy is usually much lower than that of the leads,  in
> the region of -2000 or -3000 eV.  Is this the correct value I should
> put for EnergyLowestBound here?  I initially tried this, and got the
> large fluctuations in SC steps as mentioned above.
The lowest eigenvalue for the energy is usually of the order of minus a
few Rydbergs, -3000 eV seems very low to me. Can you send the siesta
output file, there we can just check the value of the lowest eigenvalue.

So I think that the main problem is due to the leads, but I suggest to
try out all of Alex suggestions as well, and see if you can fix the
problems.

Cheers,

 Ivan
>
> Cheers,
>
> Lianheng
>
>
>
>
>
>
> On 12 Dec 2007, at 13:10, Ivan Rungger wrote:
>
>> Hi Lienheng,
>>
>>    I had a look at your files. I think there is a problem in the atomic
>> coordinates. You need to have one leads unit cell on each side of the
>> scattering region. I think you leads unit cell has 8 atoms, therefore
>> the first and the last 8 atoms in your scattering region unit cell have
>> to have the same coordinates, except a shift along z. If you send the
>> leads input file we can check this.
>>
>> Cheers,
>>
>>  Ivan
>>
>> Lianheng Tong wrote:
>>> Hi all,
>>>
>>> I have been trying to do a calculation involving benzene sandwiched
>>> between two tPA leads.  However, I failed to get any convergence with
>>> DZP or DZ basis set.  The static calculations in SIESTA does
>>> converge... I read about the convergence issue with carbon based leads
>>> in the previous posts,  but I am not sure if this is not just simply
>>> due to my mistakes in input parameters.
>>>
>>> Could anyone check whether I have done something stupid or not?  Thank
>>> you very much in advance.
>>>
>>> I have attached the input, output and CHR files for my calculations.
>>>
>>> Regards,
>>>
>>> Lianheng
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>>
>>> _______________________________________________
>>> Smeagol-discuss mailing list
>>> Smeagol-discuss at lists.tchpc.tcd.ie
>>> http://lists.tchpc.tcd.ie/listinfo/smeagol-discuss
>>>
>>
>>
>> --=================================================
>> Ivan Rungger,
>>
>> School of Physics,
>> Trinity College Dublin,
>> Dublin 2,  IRELAND
>> Phone: +353-1-6088454
>> Email: runggeri at tcd.ie
>>
>> =================================================
>>
>


-- 
=================================================
Ivan Rungger,

School of Physics,  
Trinity College Dublin,  
Dublin 2,  IRELAND  
Phone: +353-1-6088454  
Email: runggeri at tcd.ie

=================================================