The ABINIT Users Mailing List ( CLOSED )

[Gerbrand Ceder <gceder@MIT.EDU>]

Title: Re: [abinit-forum] Can Abinit investigate the valence stat

Tang,

That is not an Abinit issue.  The charge density around 3d transition metals

looks pretty much always the same independent of formal valence.  If you

are interested in formal valence it is much better to look at the

net spin density, which gives a much better indication of valence

as you filter most of the hybridization with the oxygen ions out.

Bill Mackrodt showed that this also occurs in Hartree Fock

Mackrodt WC, Williamson EA. First principles description of the valence states in manganese oxides. Berichte der Bunsen-Gesellschaft - Physical Chemistry, Chemical Physics 1997; 101:1215-21.

Several of the papers of my own group deal with the issue of valences in

transition metal oxides.  The charge density remains mostly the same when

adding electrons to TM-ion as they accommodate most of that extra charge

by rehybridizing with the oxygen ions.

If you need to see things such as charge ordering then standard DFT

will not get you there anyway, and you may need to do +U

gerd Ceder

At 10:59 AM +0200 10/17/05, tfl03@mails.tsinghua.edu.cn wrote:

Can Abinit investigate the valence state of the complex oxides?

Dear everyone:
   I want to use Abinit code to investigate a question of valence with the real physical background. It is well known that the valence states of La, Mn are 3+ in LaMnO3. When a Ca cation is substituted for La, Ca is 2+ and a Mn3+ ion near the Ca ion changes into Mn4+. However, after imaging the charge density given by Abinit, I find that there is no significant difference between the charge density of Mn ions near the Ca ions and that of the Mn ions near La ions.
        The input file is enclosed as below and Troullier-Martins pseudopotentials are used. The lattice parameters and the coordinates of ions came from atomic simulation method, which are in agreement with some experimental data approximately. Then, my questions are: i) Can Abinit code investigate such complex compound as manganite? ii) Can Abinit code investigate the valence state of elements in oxides? iii) Is there something wrong in my input file? or I used the wrong pseudopotentials? Any help or comment is welcomed.
        Thank you very much!
                                  F. L. Tang


The input file:
# Description
# To compute the charge density of Ca doped LaMnO3

 #Definition of the unit cell
 acell  30.54614509 7.117517053 10.19366793
 chkprim 0
 rprim  1.0  0.0  0.0
        0.0  1.0  0.0
        0.0  0.0  1.0

 #Definition of the atom types
 ntypat 4
 znucl  20 57 25 8

 #Definition of the atoms
 natom 30
 typat 4*1 2*2 6*3 18*4
 xred

0.315292496 0.499992718 -0.008456786
0.640264554 0.499997562 -0.012080163
0.480905516 0.499998037 0.466623656
0.796570507 0.499998892 0.483776997
0.131429075 0.499997087 0.47381952
-0.029729909 0.499997942 -0.037183826
0.162524214 -0.002453255 -0.025542832
0.642978223 -0.001745869 0.478231266
0.802045238 -0.000896629 -0.017568346
0.324384057 0.000890791 0.482435297
0.483450688 0.001737762 -0.021767894
-0.036094665 0.002445169 0.474456237
0.252337515 -0.082451313 0.206027825
0.077293439 -0.052378973 0.229870323
0.739903207 -0.052276898 0.280369744
0.417357252 -0.04684778 0.268495351
0.903759099 -0.045289295 0.162363442
0.579054307 -0.040021775 0.18279104
0.547374486 0.040014125 0.682789217
0.222669413 0.045285143 0.662367223
0.7090716 0.046837352 0.768492609
0.386525409 0.052272394 0.78037337
0.049136428 0.052371391 0.729869726
0.874092104 0.082447617 0.706024686
0.489799807 0.499994966 0.027902736
-0.022933813 0.499995346 0.404421367
0.332761907 0.499995979 0.42487798
0.169923818 0.499996486 0.027521008
0.645686522 0.499997056 0.432654719
0.797710508 0.499998449 0.042815201

 #Definition of the planewave basis set
 ecut 75

 #Definition of the k-point grid
 nkpt 3
 nband 120

 #Definition of the SCF procedure
 nstep 50
 toldfe 1.0d-10
 #Output
 occopt 3

 tsmear 0.05
 prtden 1