----------------------------------------------------------

!PBE def2-TZVP VERYTIGHTOPT DEFGRID3 NUMFREQ

%pal nprocs 20

nprocs_group 4

end

%scf

MAXITER 2500

end

*XYZFILE 0 34 Co01Ni37-034.xyz

----------------------------------------------------------

Input Orca file.

I am looking for a configuration to get all the NON imaginary frequencies, the system I am handling AnBn-38 with n≤38 and A, B = Fe, Co and Ni, ie the binary systems plus Fe38, Co38 and Ni38. In total I have 113 clusters. The method for optimization is as follows: 1. Optimize the clusters by the semiempirical method using Gupta as potential. 2. Reoptimize with DFT with the data provided in the .inp file, plus the calculation of frequencies to ensure that they are minima within the potential energy surface. So far I have obtained 82 clusters with zero imaginary frequencies, however, I have 31 clusters with imaginary frequencies, most with 1 imaginary frequency, followed by 2, some with 3, two frequencies of 4 and one of 5. As a last option, the following recommendation in the ORCA tutorial on page https://www.faccts.de/docs/orca/6.0/tut ... /freq.html suggests the following: “If such measures fail, displacing the respective mode manually and restart the optimization can help”, however, as too many atoms are moving, I fear that the structure or geometry of the principle cumulus may be lost. I have performed 3 reoptimizations with DFT, the last one, with the configuration I show at the beginning. It is worth mentioning that only the frequency analysis takes between 4 and 5 days, in high performance computer equipment, with 20 processors assigned per cumulus. Any comments, suggestions are welcome.