The ABINIT Users Mailing List ( CLOSED )

[<chengdaojian@gmail.com>]

Dear All Users,

I have a problem with "PAW SPHERES ARE OVERLAPPING!". It happens only in the
beginning of the SCF cycle, because some atoms have moved a lot randomly. So:
1) Should I make make PAW data sets with radii small enough that they do not
overlap?  Can this action work when the atoms move a lot in the beginning of
the SCF cycle?
2) Should I use a supercell geometry or other ionmov setting? Here I used only
one atom per layer and ionmov = 3.

The ERROR:
 Cartesian coordinates (bohr)
  0.00000000000000E+00  0.00000000000000E+00  0.00000000000000E+00
 -2.35349888466629E-08  2.71758625030793E+00  3.84324748168752E+00
  2.35349887474799E+00  1.35879310477207E+00  7.68649496337505E+00
  6.36427513530181E-04 -3.64574119790697E-04  1.15316906583828E+01
  5.35980233565532E-02  2.68663712421063E+00  1.54673211838139E+01
 -2.18889392228279E-02  1.26325808216285E-02  1.89255486877382E+01 
 fconv : at Broyd/MD step   2, gradients have not converged yet.
  max grad (force/stress) = 4.2691E-02 > tolmxf= 5.0000E-04 ha/bohr (free
atoms)
 BROYDEN STEP NUMBER   3 
------------------------------------------------------ 
 findmin : WARNING -
  The 2nd degree equation has no positive root (choice=4). 
 findmin : COMMENT -
  Will continue, since the new total energy is lower
  than the old. Take a larger step in the same direction.
   line minimization, algorithm   4
                        lambda      etotal            dedv        d2edv2
   old point         :  0.0000E+00 -2.0458222678E+02 -4.4545E-03  5.4768-315
   new point         :  1.0000E+00 -2.0458657309E+02 -4.3814E-03  2.0072E+10
   predicted point   :  2.5000E+00 -2.0459268701E+02  0.0000E+00  6.7989-259 
 Cartesian coordinates (bohr)
  0.00000000000000E+00  0.00000000000000E+00  0.00000000000000E+00
 -2.35349888466629E-08  2.71758625030793E+00  3.84324748168752E+00
  2.35349887474799E+00  1.35879310477207E+00  7.68649496337505E+00
  2.22720503993542E-02 -1.27882177834460E-02  1.15514582429084E+01
  1.56557949979702E+00  1.81359878370057E+00  1.41769829818673E+01
 -6.43183869046093E-01  3.71215698548800E-01  2.02198822354872E+01
     iter   Etot(hartree)     deltaE(h) residm    nres2    magn 
 FFT (fine) grid used in SCF cycle: 
 getcut: wavevector=  0.0000  0.0000  0.0000  ngfft=  30  30 225
         ecut(hartree)=     50.000   => boxcut(ratio)=   2.00377 
 chkpawovlp : ERROR -
  PAW SPHERES ARE OVERLAPPING !
   Distance between atoms   4 and   5 is  :   3.55118
   PAW radius of the sphere around atom   4 is:   2.31080
   PAW radius of the sphere around atom   5 is:   2.31080
   This leads to a (voluminal) overlap ratio of  7.43 %

Another ERROR:
 chkpawovlp : ERROR -
  PAW SPHERES ARE OVERLAPPING !
   Distance between atoms   5 and   6 is  :   3.04748
   PAW radius of the sphere around atom   5 is:   2.31080
   PAW radius of the sphere around atom   6 is:   1.11202
   This leads to a (voluminal) overlap ratio of  5.43 %

My input: t11.in
rprim
   .70710678  .0000000  .000000
 -0.35355339 0.6123724  .000000
   .000000    .000000  5.1961524 
occopt 7 tsmear 0.01  kptopt 1  
ntypat 2    
znucl 28 6   
ngkpt  4 4 1
nshiftk 1
shiftk  0.5 0.5 0.5 
acell 3*6.6567
#chkprim = 0 
natom 6  
typat 1 1 1 1 1 2  
xred
   .00000000   .00000000   .00000000
   .33333333   .66666667   .11111111
   .66666667   .33333333   .22222222
   .00000000   .00000000   .33333333
   .33333333   .66666666   .44444444
   .00000000   .00000000   .55000000 
natfix 3
iatfix 1 2 3 
ionmov 3
tolmxf 5.0d-4
ntime 200 
ixc 11   
ecut 26 
nsppol 2
spinat 0.0 0.0 0.5
       0.0 0.0 0.5
       0.0 0.0 0.5
       0.0 0.0 0.5
       0.0 0.0 0.5
       0.0 0.0 0
pawecutdg 50.
getwfk -1
prteig 0 prtden 0
optforces 2
diemac  100
diemix  0.2
nline  16
npulayit  30
nstep  200
toldfe 1.0d-6

My input: t1x.files
t11.in
t1x.out
t1xi
t1xo
t1x
../../Psps_for_tests/Ni.atompaw.gga.paw
../../Psps_for_tests/C-GGA-hard.paw