[Wien] a strange question about GGA+U

[Stefaan Cottenier]

Such behaviour is -- unfortunately -- typical for LDA+U. For sure you  
know that if you perform a regular GGA calculation, the final result  
might depend on the starting density (e.g., start with a non-magnetic  
density and you find a nonmagnetic result, start with a ferromagnetic  
density and you find a ferromagnetic result). Similarly, with LDA+U  
the final result depends on the starting density *and* on the density  
matrix (dmatup/dn). And the dependence is much stronger than for  
LDA/GGA: even small changes in the density matrix might lead a very  
different self-consistent solution. The total energy tells you which  
solution is the better one, but there is no systematic procedure that  
guarantees you have found all possible solutions, and hence you're  
never sure you've found the real ground state.

Stefaan

> Dear users,
>     When I run the calculation about GGA+U, there appears a strange problem.
> First I run the usual sp calculation with GGA(without U) in   
> directory A, then I do two kinds of calculations:
> 1: make a new directory B and copy the input files and   
> case.clmsum/up/dn files from        directory A to B, and write   
> case.inorb/indm, then
>     runsp_lapw -p -orb -cc 0.001 -ec 0.0001.
>      we got the result is MMTOT =1.00
> 2: make a new directory C and copy the input files from directory A   
> to C, then run
>            x lstart
>            x dstart
>            x dstart -up/dn
>            runsp_lapw -p -orb -cc 0.001 -ec 0.0001
>      we got the result is MMTOT =5.00
> I don't know what lead the results, the difference between two   
> methods should be the charge density,  In order to improve this i do  
>  the third kind calculation:
> 3: In directory B
>                          clean_lapw
>                          rm *clm*
>                          x dstart
>                          x dstart -up/dn
>                          runsp_lapw -p -orb -cc 0.001 -ec 0.0001
>       we got the result is MMTOT =1.00
> From calculation 3 the charge density may be not the reason.
> So I really can't understand the big difference, can you  lighten   
> me? Any suggestions
> will be appreciated.
>                                                                       
>                      yours sincerely,
>                                                                       
>                         hongxia
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
>



-- 
Stefaan Cottenier
Instituut voor Kern- en Stralingsfysica
K.U.Leuven
Celestijnenlaan 200 D
B-3001 Leuven (Belgium)

tel: + 32 16 32 71 45
fax: + 32 16 32 79 85
e-mail: stefaan.cottenier at fys.kuleuven.be


Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm