total energy of TiO2 surface cannot converge

Total energy, geometry optimization, DFT+U, spin....

Moderator: bguster

Locked
andreichibisov
Posts: 17
Joined: Thu Jan 21, 2010 2:03 am

total energy of TiO2 surface cannot converge

Post by andreichibisov » Sat Sep 01, 2012 5:20 am

Hello all,
I want to calculate total energy of TiO2 relaxed surface (101) (anatase).
However, the total energy cannot converge! Can you help me?
My input file:

ecutsm 0.5
iscf 7
diemix 0.5
diecut 0.8
enunit 1
nline 20
#######################################################
#Definition of the k-point grids
kptopt 1

nshiftk 1
shiftk 0.0 0.0 0.0 # These shifts will be the same for all grids
ngkpt 2 1 1
#######################################################

#Definition of the unit cell
optcell 0
# dilatmx 1.1

acell 10.34510 7.64900 19.99550 angstrom
angdeg 90.00 90.00 90.00

spgroup 1
#chkprim 0
#nsym 1

#Definition of the atom types
ntypat 2
znucl 22 8
occopt 1
# prtdos 1
# tsmear 9.5d-5
nsppol 2

#Definition of the atoms
natom 60
typat 10*1 20*2 10*1 20*2

xcart
1.43343844254076E+01 2.69580333380496E-07 5.03343371608116E+00
1.69662740209039E+01 8.46404925763033E-07 1.21647166188274E+01
4.55968852473211E+00 3.61362823460934E+00 5.03342199021174E+00
7.19157371420752E+00 3.61362801136723E+00 1.21647123058032E+01
7.31441018654851E+00 3.61362879761057E+00 2.89440904698280E-01
9.98742252330993E+00 3.61362897745794E+00 6.82981919589423E+00
1.24651537893554E+01 3.61362834764113E+00 1.34600740887262E+01
1.70891130952061E+01 1.77231438740531E-17 2.89440904698280E-01
2.12718631113798E-01 6.95189851650336E-08 6.82982542265703E+00
2.69045438395992E+00 1.00871181086130E-07 1.34600714753553E+01
8.04164808295264E+00 0.00000000000000E+00 0.00000000000000E+00
1.08649165614958E+01 6.01771777113559E-09 6.61994094648844E+00
1.34439005799143E+01 -2.96032546080459E-07 1.38622304386218E+01
1.78163509916103E+01 3.61362879761057E+00 0.00000000000000E+00
1.09020901198938E+00 3.61362939822935E+00 6.61993552558032E+00
3.66919958084168E+00 3.61362942258526E+00 1.38622369065170E+01
4.14191466995886E+00 3.61362902541267E+00 1.34906788061143E+00
6.62605933567496E+00 3.61362944291807E+00 8.21232253680122E+00
9.31338788529943E+00 3.61362870068001E+00 1.49647596933833E+01
1.39166335427233E+01 -1.28855142294138E-07 1.34906826772146E+00
1.64007740803687E+01 -7.17718343267842E-07 8.21231857667026E+00
1.90880914378050E+01 9.06451153771625E-08 1.49647578701347E+01
3.77579145543586E+00 3.82270617404195E-07 5.29800684691173E+00
6.34095080781924E+00 4.82736240165097E-07 1.22108338465582E+01
1.35504963868376E+01 3.61362868373400E+00 5.29800714120382E+00
1.61156625850426E+01 3.61362951147424E+00 1.22108369137877E+01
8.18118874532401E+00 3.61362851349310E+00 3.70738619499109E+00
1.07661429425310E+01 3.61362861643512E+00 1.05111308426644E+01
1.79559003742658E+01 1.99703961135348E-07 3.70734909943217E+00
9.91436647991401E-01 3.35888481250905E-07 1.05111306557143E+01
1.43343843814803E+01 7.22725787155480E+00 5.03343300702326E+00
1.69662737895381E+01 7.22725715660818E+00 1.21647168214017E+01
4.55968889837583E+00 1.08408863708942E+01 5.03342183138184E+00
7.19157348837883E+00 1.08408867740737E+01 1.21647119437097E+01
7.31441018654851E+00 1.08408863928317E+01 2.89440904698280E-01
9.98742124437308E+00 1.08408860174045E+01 6.82981841863814E+00
1.24651528721212E+01 1.08408867635948E+01 1.34600743501230E+01
1.70891130952061E+01 7.22725759522114E+00 2.89440904698280E-01
2.12718667197281E-01 7.22725772360109E+00 6.82982416012863E+00
2.69045406858049E+00 7.22725757764749E+00 1.34600720647483E+01
8.04164808295264E+00 7.22725759522114E+00 0.00000000000000E+00
1.08649166740121E+01 7.22725718932632E+00 6.61994128930658E+00
1.34439002308600E+01 7.22725770509386E+00 1.38622304900752E+01
1.78163509916103E+01 1.08408863928317E+01 0.00000000000000E+00
1.09020915664931E+00 1.08408862010953E+01 6.61993486058291E+00
3.66919933784446E+00 1.08408859543010E+01 1.38622371519371E+01
4.14191219392588E+00 1.08408863968844E+01 1.34906809561172E+00
6.62606125023480E+00 1.08408866084722E+01 8.21232086175710E+00
9.31338930002522E+00 1.08408862399799E+01 1.49647598034450E+01
1.39166332470959E+01 7.22725752047987E+00 1.34906864381792E+00
1.64007737849231E+01 7.22725745571668E+00 8.21231896685104E+00
1.90880914991083E+01 7.22725775467010E+00 1.49647569478138E+01
3.77579183535234E+00 7.22725768356059E+00 5.29800698185230E+00
6.34095150649301E+00 7.22725729303523E+00 1.22108334958883E+01
1.35504969267165E+01 1.08408860502374E+01 5.29800746805281E+00
1.61156622026821E+01 1.08408854925696E+01 1.22108368311507E+01
8.18118830838424E+00 1.08408860663307E+01 3.70738838721905E+00
1.07661436118862E+01 1.08408859170943E+01 1.05111311772026E+01
1.79559007972578E+01 7.22725798492616E+00 3.70734923135279E+00
9.91436061088312E-01 7.22725792646910E+00 1.05111319794943E+01

natfix 8
iatfix 11 35 41 5 8 44 38 14

#Definition of the planewave basis set
ecut 30.0

#Definition of the SCF procedure
nstep 500 # Maximal number of SCF cycles
toldfe 1.0d-6
tolmxf 1.0d-4
ionmov 3
ntime 500

############## Molecular Dinamic
# ionmov 9
# mditemp 100
# mdftemp 298
Attachments
101.in
(4.88 KiB) Downloaded 281 times

ilukacevic
Posts: 271
Joined: Sat Jan 16, 2010 12:05 pm
Location: Dept. of Physics, University J. J. Strossmayer, Osijek, Croatia
Contact:

Re: total energy of TiO2 surface cannot converge

Post by ilukacevic » Sat Sep 01, 2012 9:42 am

Dear Andrei,

did you try other iscf options? Maybe some other algorithm works better for this kind of system.
Also, you might want to include npulayit and nnsclo variables if you go with iscf = 7.

Igor L.

andreichibisov
Posts: 17
Joined: Thu Jan 21, 2010 2:03 am

Re: total energy of TiO2 surface cannot converge

Post by andreichibisov » Sun Sep 02, 2012 5:46 am

Dear Igor,

Thank you very much for your reply...

Andrey

Boris
Posts: 128
Joined: Tue Feb 16, 2010 10:13 am
Location: France

Re: total energy of TiO2 surface cannot converge

Post by Boris » Mon Sep 03, 2012 10:47 am

And if you're using the KGB 3-level parallelization, you might want to play on npband and bandpp. Maximizing npband will greatly improve the convergence, but each scf step will be slower
----------------------------------------------------------
Boris Dorado
Atomic Energy Commission
France
----------------------------------------------------------

Locked