[Octopus-users] td run problem with absorbing boundaries

[Pavel Redkin]

Dear Octopus users!
I noticed a strange output of Octopus 3.0.1 when treating rather large molecules, for example, fullerens, with absorbing boundaries. When I used the following input file:

CalculationMode = gs
#CalculationMode = td
UnitsInput = ev_angstrom
UnitsOutput = atomic
BoxShape = parallelepiped
%LSize
100|6|6
%

#AbsorbingBoundaries=mask

spacing = 0.6
XYZCoordinates='C60.xyz'
#######LASER######
TDEvolutionMethod = exp_mid
TDMaximumIter = 10000
TDTimeStep = 0.0015
amplitude = 6.0
omega = 1.5
%TDExternalFields
 electric_field | 1 | 0 | 0 | envelope_constant | amplitude | omega 
%
OutputEvery=50
TDOutput = multipoles
OutputHow = gnuplot + plane_x + plane_z

with commented Absorbing boundaries I received usual oscillating dipole moment. I know this is not so exact and can introduce mistakes. But after I had uncommented #AbsorbingBoundaries=mask , I received an unphysical result for dipole with a sharp single peak during the first period and almost zero after it. I tried to use both larger grids and smaller spacings, but this did not change the outcome. I wonder whether this is due to some problems with absorbing boundaries or the reason is either box or spacings.

Thank you for attention. Pavel Redkin.