How to make a ZPE correction for saddle point?

[cenwanglai]

For a freqence-aimed run by IBRION = 5 with several involved atoms movable, we can get real several frequences and an intensive imaginary freqence.
Each of the frequences (fi):

ZPEi=0.5hfi

ZPE(total)=sum of all the ZPEi

E(corrected)= E+ZPE(total)

Is it right? And how to treat about the imaginary freqence?

For whole a MEP, ZPE correction is done by the same way for all the images? Each of all the images are corrected with all related freqencies included?

Thanks for any advice.

[attachment=0]cen1.png[/attachment]
[attachment=1]cen.png[/attachment]

[graeme]

The simplest ZPE correction is to add 1/2 hbar * omega for each positive mode at any geometry.

This is a good approximation at low temperature, but at temperatures closer to the Debye temperature, equation (4) with the sinh functions provides a better approximation. Our script called dymeffbar.pl can do these calculations.