This thread is locked.Only browsing is available.
Top Page > Browsing
DOS calculation by OpenMx
Date: 2011/06/01 15:33
Name: Sheelan SC   <sheelan@cms.unipune.ac.in>

I am calculating DOS of a Au-Cd-Se system. But I am not getting the correct values of DOS, sometimes getting -ve values!!..........

Please help

メンテ
Page: [1]

Re: DOS calculation by OpenMx ( No.1 )
Date: 2011/06/02 13:32
Name: T.Ozaki

Hi,

Unfortunately, it may be difficult to provide comments
without knowing details on what you performed.

Regards,

TO
メンテ
Re: DOS calculation by OpenMx ( No.2 )
Date: 2011/06/02 18:00
Name: Sheelan SC  <sheelan@cms.unipune.ac.in>

Dear TO,

We attach a Cd-Se tube with an Au cluster and calculate the DOS for the system. We generate pseudo-potentials and atomic orbitals of Cd, Se and Au using Adpack. Use the same pseudo potentials to calculate the electronic structure (using openmx command). Then calculate the Density of states by using DosMain command. We have checked our result using various combination of s,p,d orbitals. But didn't find any improvement.

NOTE: I am new to this type of numerical calculation. So, if I am still not providing you sufficient details of my calculations please let me know what else technical points I should mention.

Regards,
Sheelan SC
メンテ
Re: DOS calculation by OpenMx ( No.3 )
Date: 2011/06/02 21:34
Name: T.Ozaki

Hi,

Can I assume that you have already reproduced the example of DOS
calculation in the manual?

Regards,

TO
メンテ
Re: DOS calculation by OpenMx ( No.4 )
Date: 2011/06/03 14:32
Name: Sheelan SC  <sheelan@cms.unipune.ac.in>

Yes sir, and that gives us correct result.
メンテ
Re: DOS calculation by OpenMx ( No.5 )
Date: 2011/06/04 22:45
Name: T.Ozaki

Hi,

Have you checked several quantities calculated by the SCF calculations, such as
Mulliken population and eigenvalues in its *.out file?
If you obtain some erratic values for such quantities, this indicates that
your pseudopotentials may not be accurate.

Regards,

TO
メンテ
Re: DOS calculation by OpenMx ( No.6 )
Date: 2011/06/06 14:55
Name: Sheelan SC  <sheelan@cms.unipune.ac.in>

No, I didn't check. Okay, will check & let u know.

Thanks for your suggestion.

Regards,
Sheelan
メンテ
Re: DOS calculation by OpenMx ( No.7 )
Date: 2011/06/07 18:08
Name: Sheelan SC  <sheelan@cms.unipune.ac.in>

Sir,

I have checked the pseudo-potential calculations. Even I have used *.vps from the database supplied by openmx3.5/DFT_DATA06. But still getting -ve values of PDOS for p and d orbital momenta.

Regards,
Sheelan
メンテ
Re: DOS calculation by OpenMx ( No.8 )
Date: 2011/06/07 21:04
Name: T.Ozaki

Hi,

I did not know that you are calculating PDOS not the total DOS.
Although the total DOS should be alway positive, PDOS can be negative,
since it is calculated based on the Mulliken population analysis.

Especially for cases where there are delocalized electrons such as s-electrons
in alkaline atoms or too many basis functions are allocated to one atom, the negative
PDOS tends to appear. Of course, this cannot be related to physics, but come from the
way how DOS is decomposed to each contribution.

Best regards,

TO
メンテ
Re: DOS calculation by OpenMx ( No.9 )
Date: 2011/06/08 19:35
Name: Sheelan SC  <sheelan@cms.unipune.ac.in>

Sir,

Actually I have calculated both DOS and PDOS. First I was getting -ve values for both DOS and PDOS. I got rid off the first problem by using correct pseudo-potentials and correct method. But I didn't know the correctness of -ve PDOS.

Thanks for your suggestion/explanation. Hope I am getting correct results now.

Thanks again,

Regards,
Sheelan
メンテ

Page: [1]