Using different PP (LDA FHI, GGA FHI, HGH NC, GGA PBE) I optimized the geometry with (optcell=1 or optcell=2), but when I restart optimization from optimized acell and WFN (with irdwfk 1 keyword) I've got different energy from previously optimized ( higher one) and optimization process continued again to the previously optimized geometry.
graph demonstrate relationship between the x-dim of AgN3 unit cell and iteration number for first and second run.
the PP http://www.abinit.org/downloads/psp-links/lda_fhi
the input:
Code: Select all
# AgN3 : geometry optimization
#
# Definition of the unit cell
acell 5.600 5.980 5.998 angstrom
spgroup 72
rprim 1.0 0.0 0.0
0.0 1.0 0.0
0.0 0.0 1.0
chkprim 0
# Definition of the atom types
ntypat 2
znucl 7 47
# Definition of the atoms
natom 16 # There are 16 atoms
typat 4*2 12*1
xred # This keyword indicate that the location of the atoms
# will follow, one triplet of number for each atom
0.0 0.0 1/4 # Note the use of fractions (remember the limited.
1/2 1/2 1/4 # interpreter capabilities of ABINIT)
0.0 0.0 3/4 # Note the use of fractions (remember the limited.
1/2 1/2 3/4 # interpreter capabilities of ABINIT)
0.3422 0.1327 0.0 # 1-st N3 group (N1)
1/2 0.0 0.0 # 1-st N3 group (N2)
0.6578 -0.1327 0.0 # 1-st N3 group (N3)
0.1578 0.6327 0.0 # 2-nd N3 group (N1)
0.0 1/2 0.0 # 2-nd N3 group (N2)
-0.1578 0.3663 0.0 # 2-nd N3 group (N3)
0.1578 0.3663 1/2 # 3-rd N3 group (N1)
0.0 1/2 1/2 # 3-rd N3 group (N2)
-0.1578 0.6327 1/2 # 3-rd N3 group (N3)
0.6578 0.1327 1/2 # 4-th N3 group (N1)
1/2 0.0 1/2 # 4-th N3 group (N2)
0.3422 -0.1327 1/2 # 4-th N3 group (N3)
# Definition of the planewave basis set
ecut 20.0 # Maximal kinetic energy cut-off, in Hartree
# Definition of the k-point grid
kptopt 1 # Option for the automatic generation of k points, taking
# into account the symmetry
ngkpt 1 1 1 # This is a 1x1x1 grid based on the primitive vectors
nshiftk 1 # of the reciprocal space (that form a BCC lattice !),
# repeated four times, with different shifts :
shiftk 0.0 0.0 0.0
# In cartesian coordinates, this grid is simple cubic, and
# actually corresponds to the
# so-called 4x4x4 Monkhorst-Pack grid
# Definition of the SCF procedure
nstep 30 # Maximal number of SCF cycles
toldfe 1.0d-10 # Will stop when, twice in a row, the difference.
# between two consecutive evaluations of total energy.
# differ by less than toldfe (in Hartree).
diemac 3.0 # Although this is not mandatory, it is worth to
# precondition the SCF cycle. The model dielectric
# function used as the standard preconditioner
# is described in the "dielng" input variable section.
prtden 0 # No output of electron density
# add to conserve old < 6.7.2 behavior for calculating forces at each SCF step
# optforces 1
optcell 1
ntime 30
dilatmx 1.2
ecutsm 0.5
ionmov 2how to avoid this behavior?
Best.
Vladimir.
