Hi, I am trying to use cDFT to perform geometry optimization in NWChem 6.8, however, I kept having cdft convergence issue when optimization driver goes to around 30 iteration. I would appreciate if you could offer any recommendations! Thank you.
start alpha_single_neg
title "alpha_single_neg"
geometry noautoz noautosym
F 7.50413827 1.95064659 -5.87931388
H 4.94665183 1.66949276 -5.98679942
C 7.29468738 2.32810273 -4.59121494
C 5.98277190 1.67226798 -4.10534824
F 7.13321377 3.67613711 -4.55365685
C 8.57399311 1.98670716 -3.78860928
C 4.78529892 2.05170887 -4.98084810
F 6.16040104 0.32193221 -4.09423159
F 5.76525286 2.06192722 -2.81614771
F 8.83586195 0.67606040 -3.85572398
F 9.60872158 2.65960541 -4.31059437
F 8.43253074 2.33956793 -2.50504246
N 3.55440438 1.47718083 -4.44455512
H 4.68865499 3.13674122 -4.98606361
O 3.82268177 -0.32293882 -5.82626879
O 3.15014635 3.41722550 -3.28278567
Cl 2.47193995 -2.84818069 -5.62220249
H 1.08831471 3.41770289 -1.94144119
H 0.18673991 -3.33659031 -4.12841697
Cl -1.23387125 2.95721655 -0.51105331
C 3.17926223 0.18803030 -4.92277187
C 2.80207886 2.27564120 -3.56619601
C 2.00168537 -0.44667170 -4.26767149
C 1.57787260 1.64637447 -3.01844204
C 1.61092409 -1.75869239 -4.53004591
C 1.22256306 0.31276955 -3.34105077
C 0.79877491 2.39413294 -2.16088885
C 0.47911095 -2.31535742 -3.90890830
C 0.06715828 -0.23889260 -2.71553460
C -0.34897844 1.84575251 -1.56164328
C -0.28097203 -1.57648532 -3.02750761
C -0.72544958 0.52519886 -1.80492021
C -1.46282799 -2.23336018 -2.42344797
C -1.89405909 -0.11594581 -1.13399048
N -2.22689189 -1.43441609 -1.55232525
O -1.75983695 -3.39968879 -2.65013730
O -2.55481814 0.42020597 -0.25888044
C -3.40077227 -2.05247744 -0.93566957
C -4.66197574 -1.76067612 -1.75272328
H -3.24759485 -3.13086189 -0.91003747
H -3.52489158 -1.65962723 0.07168285
F -4.48963062 -2.17157854 -3.04139989
F -4.91853791 -0.42308112 -1.78824276
C -5.91049963 -2.47360703 -1.18199756
C -7.24688857 -2.19732575 -1.91290536
F -5.68855292 -3.81262062 -1.21711263
F -6.06583467 -2.09301524 0.11185123
F -7.15625305 -2.53277902 -3.20608422
F -8.21711650 -2.92645851 -1.34449163
F -7.57345221 -0.90272629 -1.81967866
end
basis "small"
* library 3-21G*
end
basis "large"
* library 6-31G*
end
set "ao basis" "small"
dft
xc b3lyp
convergence nolevelshifting
odft
mult 1
cdft 14 34 charge -1.0
maxiter 50
end
task dft
set "ao basis" "large"
dft
direct
xc b3lyp
convergence nolevelshifting
odft
mult 1
cdft 21 34 charge -1.0
maxiter 1000
vectors input atomic output alpha_single_neg.mos
end
set quickguess t
task dft optimize
A few things:
I restarted the optimization, so I am not sure whether restarting is causing the problem.
The final calculation will be two of this molecule with +1 and -1 charge constrained on the same atom groups. However, I was getting same cdft convergence issue on full system so I am trying to optimize one molecule first.
When I used 6-31+G* or 6-31++G*, it keeps printing nbf_ao .ne. nmo, and still generates cdft unable to improve multiplier error
|