Is it possible to run calculations (energy, optimization, etc) with background charge? I have tried to do this with External potentials with HF, but it cannot reproduce the result yielded with Gaussian charge function.
I don’t recall which, but the external potentials process you linked to had been checked against another qc program. You’ll have to supply the Gaussian result if you want us to try to match it.
Hi Lori,
Thanks for replying. I will give my G09 and Psi4 input files below.
PSI4 input:
molecule water { 0 1 O -0.778803000000 0.000000000000 1.132683000000 H -0.666682000000 0.764099000000 1.706291000000 H -0.666682000000 -0.764099000000 1.706290000000 no_reorient no_com }
Chrgfield = QMMM() Chrgfield.extern.addCharge(-0.834,1.649232019048,0.0,-2.356023604706) Chrgfield.extern.addCharge(0.417,0.544757019107,0.0,-3.799961446760) Chrgfield.extern.addCharge(0.417,0.544757019107,0.0,-0.912085762652) psi4.set_global_option_python('EXTERN',Chrgfield.extern)
set { scf_type direct reference rhf d_convergence 12 basis cc-pvdz }
energy('scf')
Gaussian Input:
#p HF/cc-pvdz charge
Comment
0 1 O -0.778803000000 0.000000000000 1.132683000000 H -0.666682000000 0.764099000000 1.706291000000 H -0.666682000000 -0.764099000000 1.706290000000
1.649232019048 0.0 -2.356023604706 -0.834 0.544757019107 0.0 -3.799961446760 0.417 0.544757019107 0.0 -0.912085762652 0.417
Results:
PSI4:
@RHF Final Energy: -76.03498602087275
=> Energetics <=
Nuclear Repulsion Energy = 9.4264780387548672 One-Electron Energy = -123.3805014956179633 Two-Electron Energy = 37.9190374359903402 DFT Exchange-Correlation Energy = 0.0000000000000000 Empirical Dispersion Energy = 0.0000000000000000 PCM Polarization Energy = 0.0000000000000000 EFP Energy = 0.0000000000000000 Total Energy = -76.0349860208727648
G09:
SCF Done: E(RHF) = -76.9673137369 A.U. after 11 cycles
Convg = 0.2117D-08 -V/T = 2.0117
KE= 7.607945147638D+01 PE=-2.003825938821D+02 EE= 3.807995635848D+01
I have tried running calculations without background charges in both programs, and the results agree with each other.
I wonder if it is the way I call calculations with background charge in PSI 4 is wrong.
Thanks a lot.
Chester
I understand where wnet wrong: the computation results with Gaussian include the self-energy of background charges, while PSI4 does not consider that; also a nosymm keyword should be added in Gaussian input file to avoid further reorientation.