ABINIT Discussion Forums

Dear abinit users,
I encounter a problem with paw pseudopotentials calculations.
I performed volume optimisation of LaCrO3 cell with 5 atoms using the new paw dataset from abinit website. I obtained a lattice parameter in a good agreement with other theoretical works. Next I tried to performe surface calculcations on LaCrO3 using supercell with 22 atoms, but I encounter a problem (overllaping spheres !!!), here after the error message in the log file. I tried to ajust the values of the variable pawovlp but we are performing position relaxation using ionmov variable and we areb unable to ajust the pawovlp value for each atomic move.
Can someone help me to resolve this problem






----------------------------------------------------------------------------------------
--- !ERROR
message: |
Action: 1- decrease cutoff radius of PAW dataset
OR 2- ajust "pawovlp" input variable to allow overlap (risky)
src_file: chkpawovlp.F90
src_line: 183
...


leave_new : decision taken to exit ...

chkpawovlp : ERROR -
PAW SPHERES ARE OVERLAPPING !
Distance between atoms 13 and 20 is : 1.26108
PAW radius of the sphere around atom 13 is: 1.41465
PAW radius of the sphere around atom 20 is: 1.41465
This leads to a (voluminal) overlap ratio of 37.57 %

--- !ERROR
message: |
Action: 1- decrease cutoff radius of PAW dataset
OR 2- ajust "pawovlp" input variable to allow overlap (risky)
src_file: chkpawovlp.F90
src_line: 183
...
--------------------------------------------------------------------------------
input file :
surface lacrO3 terminée lao (nm)
iprcel 45
kptopt 1 # Option for the automatic generation of k points, taking
# tsmear 0.04 # into account the symmetry
#pawovlp 5

acell 7.322688 7.322688 51.258816

rprim 1 0 0 # FCC primitive vectors (to be scaled by acell)
0 1 0
0 0 1

chkprim 0
nsym 0
prtdensph 0
#nsppol 2
#nspden 2
#nspinor 1
#spinat



#Definition of the atom types
ntypat 3 # There is only one type of atom
znucl 57 24 8 # The keyword "znucl" refers to the atomic number of the
# possible type(s) of atom. The pseudopotential(s)
# mentioned in the "files" file must correspond
# to the type(s) of atom. Here, the only type is Silicon.
natfixx 22
natfixy 22
natfixz 2
iatfixx 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
iatfixy 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
iatfixz 1 10
nband 95
occopt 1
#Definition of the atoms
natom 22
typat 5*1 4*2 13*3
xred
0.00000000 0.00000000 0.50000000
0.00000000 0.00000000 0.64285714
0.00000000 0.00000000 0.78571429
0.00000000 0.00000000 0.21428571
0.00000000 0.00000000 0.35714286

0.50000000 0.50000000 0.57142857
0.50000000 0.50000000 0.71428571
0.50000000 0.50000000 0.28571429
0.50000000 0.50000000 0.42857143

0.50000000 0.50000000 0.50000000
0.00000000 0.50000000 0.57142857
0.50000000 0.00000000 0.57142857
0.50000000 0.50000000 0.64285714
0.00000000 0.50000000 0.71428571
0.50000000 0.00000000 0.71428571
0.50000000 0.50000000 0.78571429
0.50000000 0.50000000 0.21428571
0.00000000 0.50000000 0.28571429
0.50000000 0.00000000 0.28571429
0.50000000 0.50000000 0.35714286
0.00000000 0.50000000 0.42857143
0.50000000 0.00000000 0.42857143
ntime 40
ionmov 2
#dilatmx 1.05
toldff 1.0d-4
tolmxf 1.0d-5

ecut 45
pawecutdg 90

ecutsm 0.5 # Maximal kinetic energy cut-off, in Hartree

#Definition of the SCF procedure
nstep 40 # Maximal number of SCF cycles
#toldfe 1.0d-8 # Will stop when, twice in a row, the difference
# between two consecutive evaluations of total energy
# differ by less than toldfe (in Hartree)
# This value is way too large for most realistic studies of materials
diemac 14 # Although this is not mandatory, it is worth to
diemix 0.5d0 # precondition the SCF cycle. The model dielectric
# function used as the standard preconditioner
# is described in the "dielng" input variable section.
ixc 11

optforces 1

You can try to increase pawovlp to 50 % or so and see what happens - you don't have to change it at each step. If it ends up with a physically unreasonable situation you can check it yourself.

Your set of constraints may make it quite hard to converge though.

Also check out the recently developed ionmov 22 which is more efficient thant ionmov 2