What's the reason cause the energy of last image increase after BAND optimization converged for transition state search?

[Jingyun Ye]

Hi, all

I am trying to use Climb NEB to search the transition state for a 
reactions. For instance, I have 7 images, first one is reactant, last one 
is product. After the CI-NEB done, the energy of last image goes up which 
is higher than the energy of product. See the picture below, blue line is 
the data from CP2K, every dot is the energy of each image, red line is same 
with blue line, except I use the energy of reactant and product to replace 
that of image 1 and image 7. 

I have specify the ALIGN_FRAMES F, OPTIMIZE_END_POINTS F. (Am I right, F 
means do not optimize the last image)  
  
<https://lh3.googleusercontent.com/-SKgCEki_qzA/U6BT5YYd7QI/AAAAAAAAAM0/qSu0mtcPeMA/s1600/Untitled.png>
In VASP, the first image (reactant) and last image (product) are not 
changed after the CI-NEB converged.  

Thanks very much!


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140617/ddeb3459/attachment.htm>