Chemshell and orca combined for QM / MM calculation. Atomic number in QM is 34.The total number of atoms is 11900.Optimizing calculation of triplet state is OK.Frequency calculation shows:ORCA finished by error termination in SCF Hessian.The frequency keywords is the same as the keyword of the optimization calculation.I tried different keywords.But failed.These are the keywords I tried:

1.# Orca input file generated by ChemShell

! O3LYP def2-SVP RIJCOSX def2/J Grid4 TightSCF freq Engrad

%pointcharges "pointcharges.xyz"

%maxcore 3000

%scf MaxIter 500 end

%pal nprocs 14 end

%scf Guess MORead MOInp "save.gbw" end %coords CTyp xyz

Charge 0 Mult 3 Units bohrs coords

2.O3LYP def2-SVP RIJCOSX def2/J gridx4 Grid4 TightSCF freq

3.O3LYP def2-SVP RIJCOSX def2/J Grid5 FinalGrid6 GridX8 NoFinalGridX TightSCF freq

4.O3LYP def2-SVP RIJCOSX def2/J gridx4 Grid6 NoFinalGrid VeryTightSCF freq

FInnally, I want to ask: Must the grid points for optimization and frequency calculation be the same? I am afraid that grid points is inconsistent thus the frequency calculation result is not reliable, so I use the same keywords in the frequency and optimization calculations.