The ABINIT Users Mailing List ( CLOSED )

[Masayoshi Mikami <mmikami@rc.m-kagaku.co.jp>]

Dear Xangdong:
(http://www.crystalstar.org/ShowSource.asp?Action=ShowUser&UserID=3567)

We have asked you to follow our netiquette many times...
        http://www.abinit.org/community/?text=netiquette

If you would like to know "how to ask questions in smart way,"
here it is:
        http://www.catb.org/~esr/faqs/smart-questions.html
The Chinese version is also available:
        http://www.lat30n.cn/doc/oss/smart-questions.html
(NB: translations are available: Bahasa Indonesian Brazilo-Portuguese  
Chinese Czech Danish Estonian Finnish French Georgian German Hebrew  
Hungarian Italian Japanese Polish Portuguese Romanian Russian Serbian  
Spanish Swedish Turkish.)

About your problem: my guess is as follows:
1) your pseudopotentials (Y, Ta) might lack semi-core electrons as  
valence.
   If so, please make them by yourself. FHI98PP/OPIUM may help.
        http://www.fhi-berlin.mpg.de/th/fhi98md/fhi98PP/
        http://opium.sourceforge.net/
   You could ask how to make PPs to the authors...
2) other things, e.g. bad choice of exchange-correlation functional ?
3) low ecut, ...
These are jut guesses, because your mail lacks all the information  
suggested by the Netiquette.

I may not reply to you again, while you do not follow our Netiquette.

"Adding oil" (Bon courage),
Masayoshi

On 2007/06/03, at 21:28, xd0533@163.com wrote:

> Dear users,
> I did YTaO4 acell geometry.I use TM Pseudopotentials.Fist I modify  
> nuclear
> positions,and use optcell 0 ,ionmov 2.Then I did acell  
> geometry.Part of the input file is:
> ionmov 3
> optcell 2
> tolmxf 5.0d-4
> dilatmx 1.2
> The experimental value of lattice constants is:9.6584000000E+00
> 1.0011770000E+01 1.0300900000E+01 Bohr
> The result I get is: 1.0700498247E+01 1.1096190361E+01 1.0438822477E 
> +01 Bohr
> Only the last one is in good agreement with experimental value.I  
> wonder
> whether this is correct? Why are the front two constants worse  
> agreement
> with experimental value?
> Best wishes!
> xd