I have only tried by changing the initial geometry structures. I don’t know how to displace it along the normal mode corresponding to the imaginary frequency.
Well, the output file gives you the coordinates of the normal mode, so just add a small multiple of one of the imaginary modes to the final geometry and use that as your new geometry. (Normally, one of displacing in the “+” or “-” direction would be better, but it doesn’t matter for this case. The imaginary modes correspond to perturbing a linear to a c2v structure.)
I tried displacing the co-ordinates like you said but I am still recieving imaginary frequencies. Here is the final co_ordinate and the output file I am recieving even after displacement
C–4_output-co-ordinates.txt (231 Bytes)
C–4.txt (805.7 KB)
Your SCF occupations change between the geometry optimization and the calculation of the Hessian via finite differences of gradients. During the optimization, the occupations are (by irrep a’, a’') (11, 1), but then it switches to (10, 2) when Psi4 starts the displacements for the Hessian. (Compare lines 8164 and 8903, for example.) Try specifying the occupations explicitly in the input file using the DOCC keyword. (I don’t recall if that array gets projected when the point group drops or not.)