Dear board members,
I am trying to run a DFT calculation on the Oxygen dimer using an initial guess from atomic fragments. The code should stop after the first SCF step and report the results. The following input runs smoothly and does the job:
start
geometry dimer
O 0.00 0.0 0.0
O 1.21 0.0 0.0
end
geometry atom
O 0.00 0.0 0.0
end
basis
O library STO-3G
end
title "Oxygen atom"
set geometry atom
dft
xc b3lyp
vectors output atom.movecs
end
task dft
title "Oxygen dimer"
set geometry dimer
dft
xc b3lyp
vectors input fragment atom.movecs atom.movecs
maxiter 1
convergence energy 1.d4
convergence density 1.d4
convergence gradient 1.d4
end
task dft
Now I would like to do the calculation with spherical input densities, e.g. p_x/p_y/p_z equally occupied. Unfortunately the following input:
start
geometry dimer
O 0.00 0.0 0.0
O 1.21 0.0 0.0
end
geometry atom
symmetry oh
O 0.00 0.0 0.0
end
basis
O library STO-3G
end
title "Oxygen atom"
set geometry atom
dft
smear
xc b3lyp
vectors output atom.movecs
end
task dft
title "Oxygen dimer"
set geometry dimer
dft
xc b3lyp
vectors input fragment atom.movecs atom.movecs
maxiter 1
convergence energy 1.d4
convergence density 1.d4
convergence gradient 1.d4
end
task dft
produces the desired atomic fragments, but the dimer calculation crashes with the error message:
"movecs_fragment: open shell mismatch 0", presumably because the fragments have 3 open shells each, but the dimer has none. Is there any way to tweak nwchem to get the desired calculation?
Thanks a lot!
- Thomas
|