[Pw_forum] something extremly strange for anatase 001 structure prediction

Tue Jul 15 10:36:01 CEST 2008

Dear all,

I'm doing anatase 001 surface relaxation using different pseudo cutoff. But the surface structure depending
strongly on pseudo cutoff.
For example, when I set ecutwfc=30 Ry and ecutrho=240 Ry for the calculation, the relaxed structure is
ATOMIC_POSITIONS (angstrom)
Ti       0.000019234   0.000016929   5.241338958
Ti       1.896866310   1.894973881   0.424606117
Ti      -0.002611826   1.894973568   7.609344136
Ti       1.894935792   0.000016645   2.794368221
O       -0.000497095   1.894979291   5.668409409
O        0.000892662   0.000020764   7.220228726
O        1.897822837   1.894975309   8.112610082
O       -0.002069692   1.894975762  -0.077265872
O        1.894316864   0.000021111   0.814634124
O        1.895351386   1.894979829   2.365218854
O        0.000036842   0.000033726   3.235875667
O        1.894936685   0.000033184   4.800631577

when I set ecutwfc=50 Ry and ecutrho=400 Ry for the same calculation, the relaxed structure is 

ATOMIC_POSITIONS (angstrom)
Ti       0.022799298   0.000081081   5.241509392
Ti       1.936096642   1.894995998   0.443063583
Ti      -0.037388794   1.894978067   7.592310260
Ti       1.868524560  -0.000008072   2.794849232
O       -0.102420723   1.894969978   5.654104817
O        0.193587382   0.000011079   7.215460406
O        2.107237211   1.894969616   8.171152812
O       -0.209955519   1.895019689  -0.141004735
O        1.700007345   0.000017676   0.820240249
O        1.998867848   1.894956901   2.381473163
O        0.007469691   0.000019075   3.237525157
O        1.885175059  -0.000011088   4.799315665

unfortunately this forum is not support pictures or pdf files, I can't display it in photos. I'll illustrate 
the big difference in structure of the calculations only by word.

In the case of ecutwfc=30 Ry and ecutrho=240 Ry, there is a the mirror plane symmetry along 
the [100] direction. The two Ti-O bonds on surface is about 1.964 Angstrom and the angle of 
surface O-Ti-O atoms is about 150.269 degrees. The structure is consistent well with the literature
(figure SI in supporting information of J. Phys. Chem. B 2006, 110, 2804-2811). But the my calculated s
urface energy is 1.06049 J/m^2, which is about 8% higher than the literature.

I think the surface energy difference may be because of the ecutwfc and ecutrho. So I increased it to 
higher value, ecutwfc=50 Ry and ecutrho=400 Ry. But the calculated structure with this cutoff is so surprising. 
As shown above, the mirror plane symmetry along the [100] direction is broken: the two O-Ti bonds on 
the surface become strongly inequivalent, with bond lengths of 1.74 and 2.22 angstrom. And the angle of 
surface O-Ti-O atoms is about 145.508 degrees. Even from the coordinate of each atom is quite different in 
the two cases.  I also find this kind of structure reported by literature Physical Review letters, 2001, 87, 266105. 
But my calculated structure parameters is about 1% errors, and energetic parameters is about 6% errors with 
respect to the literature.

For my two tested case the structure is so greatly depending on the cutoff. I think the higher cutoff must be more accurate.
So there may be something inaccurate with the former case and the literature J. Phys. Chem. B 2006, 110, 2804-2811, 
because of the smaller cutoff.
How do you think about it? 
And Why my calculated energy is ~6-8% different than the literature.

Here is my input file

 &CONTROL
                       title = 'Anatase 001' ,
                 calculation = 'relax' ,
                restart_mode = 'from_scratch' ,
                      outdir = '/home/vega/TiO2/Anatase/001/relaxtest/tmp/' ,
                      wfcdir = '/tmp/' ,
                  pseudo_dir = '/home/vega/espresso-4.0/pseudo/' ,
                      prefix = 'Anatase 001 1X1' ,
                     disk_io = 'none' ,
               etot_conv_thr = 0.0005 ,
               forc_conv_thr = 0.0011668141375 ,
                       nstep = 1000 ,
 /
 &SYSTEM
                       ibrav = 6,
                   celldm(1) = 7.1619,
                   celldm(3) = 4.7591,
                         nat = 12,
                        ntyp = 2,
                     ecutwfc = 30 ,   #ecutwfc=50 for later case
                     ecutrho = 240 ,  #ecutrho=400 for later case
 /
 &ELECTRONS
                    conv_thr = 1.0D-8 ,
 /
 &IONS
                ion_dynamics = 'bfgs' ,
 /
ATOMIC_SPECIES
   Ti   47.86700  Ti.pw91-sp-van_ak.UPF 
    O   15.99940  O.pw91-van_ak.UPF 
ATOMIC_POSITIONS angstrom 
   Ti      0.000000000    0.000000000    5.220000000    
   Ti      1.895000000    1.895000000    0.410000000    
   Ti      0.000000000    1.895000000    7.625000000    
   Ti      1.895000000    0.000000000    2.815000000    
    O      0.000000000    1.895000000    5.630000000    
    O      0.000000000    0.000000000    7.214000000    
    O      1.895000000    1.895000000    8.036000000    
    O      0.000000000    1.895000000    0.000000000    
    O      1.895000000    0.000000000    0.820000000    
    O      1.895000000    1.895000000    2.404000000    
    O      0.000000000    0.000000000    3.226000000    
    O      1.895000000    0.000000000    4.810000000    
K_POINTS automatic 
  4 4 1   1 1 1 

Thank you for reading. looking forward to your reply.

Vega Lew
PH.D Candidate in Chemical Engineering
State Key Laboratory of Materials-oriented Chemical Engineering
College of Chemistry and Chemical Engineering
Nanjing University of Technology, 210009, Nanjing, Jiangsu, China
_________________________________________________________________
Discover the new Windows Vista
http://search.msn.com/results.aspx?q=windows+vista&mkt=en-US&form=QBRE
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20080715/ff32ba96/attachment.htm 