[CPMD-list] Restarting interrupted TDDFT geometry optimization

[Eung-Gun Kim]

Dear List Subscribers,

I am running a geometry optimization on a conjugated molecule in the excited
state using TDDFT (v3.9.1).  When geometry optimization was conducted in two
runs (due to an interruption), the resulting excited state geometry is different
from a non-stop calculation which was started from the same ELECTRONIC SPECTRA
RESTART file.  Now I start to wonder if I am doing RESTART properly.

What I have observed so far is that the final geometry obtained from multiple
runs is very close to that of the preceding interrupted run.  For instance, the
final value of a dihedral angle from a run terminated earlier is 33 deg whereas
it's 25 deg in a run interrupted later.  Portions of my two restart files are
attached below for your reference.

Thanks very much for reading.

Regards,

EG Kim

###############################################################################

1) Restart from an ELECTRONIC SPECTRA run ------------------------------------

&CPMD
  RESTART COORDINATES WAVEFUNCTION LINRES
  CENTER MOLECULE OFF
  TDDFT
  OPTIMIZE GEOMETRY
  LBFGS
  HESSIAN UNIT
  ODIIS
    5
  CONVERGENCE ORBITALS
    1.D-6
  ELECTROSTATIC POTENTIAL
  STORE
    50
  RESTFILE
    2
&END

&TDDFT
  TAMM-DANCOFF
  TD_METHOD_A SAOP
  DIAGONALIZER DAVIDSON
  DAVIDSON PARAMETER
    200  1.D-7  50
  STATES SINGLET
    1
  FORCE STATE
    1
&END

2) Restart from an interrupted TD geometry optimization -----------------------

&CPMD
  RESTART LATEST COORDINATES HESSIAN WAVEFUNCTION LINRES
  CENTER MOLECULE OFF
  TDDFT
  OPTIMIZE GEOMETRY
  LBFGS
  ODIIS
    5
  CONVERGENCE ORBITALS
    1.D-6
  ELECTROSTATIC POTENTIAL
  STORE
    50
  RESTFILE
    2
&END

&TDDFT
  TAMM-DANCOFF
  TD_METHOD_A SAOP
  DIAGONALIZER DAVIDSON
  DAVIDSON PARAMETER
    200  1.D-7  50
  STATES SINGLET
    1
  FORCE STATE
    1
&END