Author Topic: Convergence problem  (Read 29 times)

eliseo

  • Newbie
  • *
  • Posts: 2
  • Karma: +0/-0
Convergence problem
« on: February 18, 2019, 12:06:11 pm »
Hi all,

  I am trying to perform some riper calculations of the unit cell of a metal complex, Fe(II) mononuclear system, with two molecules in the unit cell.
I was trying PBE or SCAN functionals wit  basis = pob-TZVP and jbas  = dhf-TZVP but always I have convergence problems and after 8-9 scf cycles the energy explodes...

              | TOTAL ENERGY         =       -6433.0146972943    |
              | TOTAL ENERGY         =       -6047.3950058257    |
              | TOTAL ENERGY         =       -6377.5263506254    |
              | TOTAL ENERGY         =       -6623.9976990770    |
              | TOTAL ENERGY         =       -6518.7293471434    |
              | TOTAL ENERGY         =       -6574.3035042539    |
              | TOTAL ENERGY         =       -6627.6332493572    |
              | TOTAL ENERGY         =       -6606.1255709050    |
              | TOTAL ENERGY         =       -6294.1403523448    |
              | TOTAL ENERGY         =       -6545.2222578067    |
              | TOTAL ENERGY         =       -6586.3762207693    |
              | TOTAL ENERGY         =       -4334.0223948992    |
              | TOTAL ENERGY         =       -4101.3356070603    |
              | TOTAL ENERGY         =       -4514.5430856829    |
              | TOTAL ENERGY         =       -2667.6707907154    |

  I was trying to increasing the scfdamp parameter but I obtain more or less the same result.
Any advice will be welcome, I add the control file at the end

     best wishes, thanks a lot

            Eliseo




$title
s13 hs scan periodic D
$operating system unix
$symmetry c1
$user-defined bonds    file=coord
$coord    file=coord
$optimize
 internal   off
 redundant  off
 cartesian  on
 global     off
 basis      off
$atoms
fe 1-2                                                                         \
   basis =fe pob-TZVP                                                          \
   jbas  =fe dhf-TZVP
n  3-6,19-20,37-40,49-52,69-70,87-90                                           \
   basis =n pob-TZVP                                                           \
   jbas  =n dhf-TZVP
c  7-8,11-12,15-16,21-24,27-28,31-32,35-36,41-42,45-46,53-54,57-58,61-62,65-66 \
   71-74,77-78,81-82,85-86,91-92,95-96,99-100                                  \
   basis =c pob-TZVP                                                           \
   jbas  =c dhf-TZVP
h  9-10,13-14,17-18,25-26,29-30,33-34,43-44,47-48,55-56,59-60,63-64,67-68,     \
   75-76,79-80,83-84,93-94,97-98,101-102                                       \
   basis =h pob-TZVP                                                           \
   jbas  =h dhf-TZVP
b  103-104,113-114                                                             \
   basis =b pob-TZVP                                                           \
   jbas  =b dhf-TZVP
f  105-112,115-122                                                             \
   basis =f pob-TZVP                                                           \
   jbas  =f dhf-TZVP
$basis    file=basis
$rundimensions
   dim(fock,dens)=1815888
   natoms=122
   nshell=844
   nbf(CAO)=1904
   dim(trafo[SAO<-->AO/CAO])=2096
   rhfshells=2
   nbf(AO)=1808
$uhfmo_alpha   file=alpha
$uhfmo_beta   file=beta
$uhf
$alpha shells
 a       1-332                                  ( 1 )
$beta shells
 a       1-324                                  ( 1 )
$scfiterlimit       230
$thize     0.10000000E-04
$thime        5
$scfdump
$scfintunit
 unit=30       size=0        file=twoint
$scfdiis
$maxcor    500 MiB  per_core
$drvopt
   cartesian  on
   basis      off
   global     off
   hessian    on
   dipole     on
   nuclear polarizability
$interconversion  off
   qconv=1.d-7
   maxiter=25
$coordinateupdate
   dqmax=0.3
   interpolate  on
   statistics    5
$forceupdate
   ahlrichs numgeo=0  mingeo=3 maxgeo=4 modus=<g|dq> dynamic fail=0.3
   threig=0.005  reseig=0.005  thrbig=3.0  scale=1.00  damping=0.0
$forceinit on
   diag=default
$energy    file=energy
$grad    file=gradient
$forceapprox    file=forceapprox
$dft
   functional pbe
   gridsize   5
   radsize    50
$scfconv   6
$scfdamp   start=0.700  step=0.050  min=0.050
$scforbitalshift  closedshell=.05
$jbas    file=auxbasis
$ricore      500
$rij
$periodic 3
$cell
8.4098 8.4731 18.2909 90.  98.575 90.
$kpoints
nkpoints 2 2 1
$riper
lenonly on
sigma 0.01
$optcell
$disp3 bj
$actual step      riper
$end

« Last Edit: February 18, 2019, 12:44:34 pm by eliseo »

antti_karttunen

  • Sr. Member
  • ****
  • Posts: 205
  • Karma: +1/-0
Re: Convergence problem
« Reply #1 on: February 18, 2019, 08:40:36 pm »
Hi,

which TM version you are using? Periodic calculations are much more robust with TM 7.3.

Best,
Antti

eliseo

  • Newbie
  • *
  • Posts: 2
  • Karma: +0/-0
Re: Convergence problem
« Reply #2 on: Today at 08:32:20 am »
I was using TURBOMOLE V7.3 ( 22142 )

     thanks,

         Eliseo

uwe

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 433
  • Karma: +0/-0
Re: Convergence problem
« Reply #3 on: Today at 12:18:06 pm »
Hi,

from what I have heard (not just rumors), convergence problems can be a hint that the input structure is not reasonable. Some popular builder for crystal structures export xyz files which are not defined in a way riper expects it. Especially if you build in symmetry and then export to xyz.

Did you check the structure by visual inspection within Turbomole, i.e. by running TmoleX and view the unit cell plus replicates? Does that look reasonable?

You could try to import the structure directly from a cif file (using TmoleX) and check if you see a similar behavior.

If the structure is OK, and if you have tried changing the other typical parameters already, I'd recommend to contact the Turbomole support.

Regards,

Uwe