UT theoretical chemistry code forum

Hi everyone,
I am a newer using NEB method to search TS. I choose vinyl alcohol (CH2CHOH) that transforms to acetaldehyde (CH3CHO) as a example to learn how to search TS. I have a few questions.
1、I insert 8 image between initial and final states calculating with CINEB, then using the highest energy structure usually regard as saddle point for the frequency calculation. It has 4 imaginary frequencies. I repeat the frequency calculation using other parameters, it still has more than one imaginary frequency. Maybe it is not saddle point, but i donot how to do get the saddle point. The kpoint is 1*1*1
2、I optimized the reactant and product, they all have 6 imaginary frequencies. But there are no imaginary frequency when optimized with Dmol3. I donot know why? I thougt The freedom of molecular is 3N-6, so it has 6 imaginary frequnecies. How to eliminate the imaginary frequency in the vasp.
I put all the geometries in a big cell that lattice constant is 10, using gamma point in the calcualtion. if the NEB is not suitable for single molecular TS calculation.

You can probably ignore the 6 modes that are near zero in your calculations. They can be slightly positive or negative.

For the saddle point, there should be one significantly negative mode. If you have trouble obtaining that, check on this forum for posts about calculating frequencies to see how parameters should be set to get converged values.

Thank you for your reply. I have another question ,In the CINEB calculation the parameter EDIFFG is must be set negtive?

It should be so that the convergence criterion is based upon the force.

I donot set the parameter EDIFFG that the default value is EDIFF*10 in the CINEB calculation. So the search TS process is wrong. I have a test with a molecular, the TS energy is nearly equal between EDIFFG=-0.01 and EDIFFG=EDIFF*10. I donot know why? If there is a default force convergence criterion in the calculation? If yes, the default value is ?