The ABINIT Users Mailing List ( CLOSED )

["Xiulin Ruan" <ruan@purdue.edu>]

Thanks for your comments right to the point. The TDDFT calculations on Ti3+
doped finite clusters do tell some physics and predict energy gaps well, but
they are not good enough when we ask tougher questions. For example, how
would the energy levels (and optical properties) of Ti3+ be affected by
different host, like Al2O3, Y2O3, YAG, etc? I assume the cluster
calculations are incapable since they are not able to reproduce the crystal
fields exactly. 

Also, we can use a minimum unit cell with one Ti dopant. Although this means
an unreasonably high doping concentration, but it is still acceptable as a
first attempt, better than the cluster model I guess. Would this reduce the
computation time significantly? 
    
 

-----Original Message-----
From: Nuno A. G. Bandeira [mailto:nuno.bandeira@ist.utl.pt] 
Sent: Wednesday, June 20, 2007 9:51 PM
To: forum@abinit.org
Subject: Re: [abinit-forum] TDDFT for periodic system

Xiulin Ruan wrote:

> Thanks for your reply. Our specific problem is to calculate the energy 
> levels of transition metal doped solids, like Ti3+ doped Al2O3, which is 
> an important laser material. The free-standing Ti3+ ion has 5 degenerate 
> d orbitals. After we put the ion in the Al2O3 host, the crystal field 
> splits these d orbitals so that d-d transitions become possible. These 
> transitions are utilized to make Ti:Al2O3 an excellent laser material. 
> We have used TDDFT (with Gaussian and CPMD) to calculate the excited 
> states of a Ti3+ doped cluster, and the energy levels are very good. 
> However we couldn't figure out a way to implement the periodic host with 
> these packages. If we wanna extend our calculations by replacing the 
> finite cluster with the more realistic periodic solids, could the TDDFT 
> capability of abinit help?

What you're asking is almost unfeasible. The only software I know that 
can perform TDDFT on solids is ADF-Band, but for your specific case you 
have to build a very large unit cell with a large number of atoms to 
reproduce the doping.

TDDFT is hard enough as it is but with such a unit cell it would be a 
monumental task!

Not wanting to discourage you but if your numbers are good for the 
finite cluster why not use them instead ? It seems that since only t2g-> 
eg transitions are of importance to you, the exact reproduction of the 
doped Al2O3 crystal should not be imperative. This would only be an 
issue I think if you were looking for transitions based on the Al-O bond.

But then if money or resources are no object you can purchase as many 
licenses of Band as you wish and install them on as many machines.

-- 
Nuno A. G. Bandeira, AMRSC
Graduate researcher and molecular sculptor
Inorganic and Theoretical Chemistry Group,
Faculty of Science
University of Lisbon - C8 building, Campo Grande,
1749-016 Lisbon,Portugal
http://cqb.fc.ul.pt/intheochem/nuno.html
Doctoral student @ IST,Lisbon
--