UT theoretical chemistry code forum

Dear All,

I found that the energy did not go down on some images. why is that happened? is it reasonable?

Many thanks in advance,
Ibrahim

I can't really make much sense of this from the three files. I can see that the forces are dropping nicely, which is good. And if this is a climbing image NEB calculation, you can certainly have the energy of an image increase as it climbs up towards the saddle. I don't see anything wrong, but like I say, it is hard to see what's going on with just forces and energies on the images.

Thank you very much prof. Henkelman.

Yes, it was a run with climbing image NEB.

So now I got it.

Thanks,
Ibrahim

Dear Prof. Henkelman,

1- I have attached the mep1 image for one of my systems with 5 images, which related to my previous question where the energy of some images after some ionic steps, is not going down until convergence is reached.

2- For another structure (mep2), I see the energies of images 00 and 05 are minimum. I searched the forum and followed your suggestions to check the compatibility of input files and everything is OK, also I turned off symmetry, changed convergence criteria, but had similar results. I also relaxed those images and they give minimum energy than endpoints and do not relax back to the endpoints.

So, is the migration barrier from TS to that minimum image? or what should I consider?


Many thanks in advance,
Ibrahim

Dear Prof. Henkelman,

Are these results reasonable?

Many thanks in advance,
Ibrahim

No, the drop in energy between the band images and the endpoints (likely) indicates that you have different settings in the two calculations. If you post the full calculations of both the endpoints and the NEB, I can take a look and see what's going on.

Dear Prof. Henkelman,

So no problem in neb1?

Please find enclosed the files for neb2.

Thanks,
Ibrahim

Try turning symmetry off for your endpoints. (ISYM=0)

Yes I tried and used ISYM=0 for images and endpoints with the same results

Take image 01 and 05 and rerelax them with ISYM=0 to make sure that you have the lowest energy endstates.

Yes they give minimum energy than endpoints and do not relax back to the endpoints.

ok, so use those endpoints.

Thanks Prof. Henkelman,

Could I continue with that NEB calculation and consider the barrier relative to those minimum images?
Or should I use those images as endpoints then construct again 5 images and do neb calculation?

I am calculating the migration barrier by removing only one atom from the ideal optimized structure. 
 
Thanks

Dear Prof. Henkelman,

I have used the relaxed files of 01-opt and 05-opt and rerun the NEB calculation.

The mep figure and spline file are attached. Is it reasonable?
If it is OK, I will rerun the calculation with the aggressive parameters.

Best regards,
Ibrahim

it looks reasonable