Dear all,
I plan to calculate a broken‑symmetry state, that is, a spin‑polarized singlet state. My system involves a diradical case, in which the overall charge and spin are 0 1, while one fragment has charge and spin 0 2, and the other fragment has charge and spin 0 –2. I could not find specific instructions for setting up a broken‑symmetry state in the manual. Could you please provide some examples? Thank you very much for your help!
For what type of calculation?
1 Like
Dear jherbert,
Thank you for your reply.
I would like to perform a broken‑symmetry UKS calculation, specifically a spin‑polarized singlet, on a Cu(acetate)₂·H₂O dimer.
I plan to carry out a single-point energy calculation based on the spin‑polarized singlet wave function.
Cu 5.639839000000 0.749760000000 0.544940000000
O 4.884807000000 0.852000000000 -1.250884000000
O 6.759368000000 0.426000000000 2.167373000000
O 6.787478000000 2.130000000000 -0.210545000000
O 4.710989000000 -0.920160000000 1.077494000000
O 3.993068000000 1.704000000000 1.634819000000
C 5.196060000000 0.213000000000 -2.377918000000
C 4.405849000000 0.068160000000 -3.665957000000
C 7.932509000000 2.130000000000 -0.780254000000
C 8.686744000000 3.408000000000 -1.176574000000
Cu 7.510161000000 -0.749760000000 -0.544940000000
C 7.953940000000 -0.213000000000 2.377918000000
C 5.217491000000 -2.130000000000 0.780254000000
O 6.390632000000 -0.426000000000 -2.167373000000
O 8.439011000000 0.920160000000 -1.077494000000
O 8.265193000000 -0.852000000000 1.250884000000
O 6.362522000000 -2.130000000000 0.210545000000
O 9.156932000000 -1.704000000000 -1.634819000000
C 8.744149101322 -0.068160348014 3.665953905167
C 4.463256000000 -3.408000000000 1.176574000000
H 9.150921820136 0.930929157196 3.718452627138
H 9.595823822292 -0.731430162333 3.632437892968
H 8.114999700774 -0.348379298796 4.497848352445
H 8.153780408470 4.264639036729 -0.791191889809
H 8.767279908053 3.449320536753 -2.252774065395
H 9.664526679395 3.395708211497 -0.718109231659
H 4.816342574432 0.737531855688 -4.407466503757
H 4.522315242211 -0.938810984842 -4.038576175715
H 3.367806717936 0.291627475716 -3.468648845216
H 5.156217060552 -4.236308803107 1.187037997980
H 3.755749385166 -3.654286252337 0.398642465817
H 3.942056538262 -3.351309667008 2.120786543461
H 3.418243606186 1.168793631226 2.220934226811
H 3.260246956844 2.128641405650 1.141814947973
H 9.255197266056 -2.625579345806 -1.316329680130
H 9.116242476445 -1.631034596479 -2.611251492578
Best regards.
I look forward to your further reply. Thank you very much for your help, I truly appreciate it!
I think what you want is one of the following two options.
(1) The usual way to generate a broken-symmetry guess is to use SCF_GUESS = FRAGMO. Then you need to divide the molecule into fragments and specify charge and multiplicity for each, along with overall charge and multip. at the top:
$molecule
TotalCharge Total Multip
--
Frag1Charge Frag1Multip
[coordinates]
--
Frag2Charge Frag2Multip
[coordinates]
...
$end
The rest of the $rem section should look like normal SCF. This will perform an SCF on each fragment first, with the fragment charges and multiplicities, then use a superposition of those fragments as a guess for the full system. It’s not guaranteed that the full-system calculation maintains the electron configurations on the fragments and you may need to experiment with SCF_ALGORITHM to find the SCF solution that you want.
(2) Alternatively, you can use constrained DFT to force the system to maintain charge and spin constraints on fragments. See the manual for setup instructions:
https://manual.q-chem.com/latest/topic_cdft.html
Thank you very much for your constructive reply. We first tried using the option 1 you provided, and the input file is as follows. However, ‘SCF failed to converge’ occurred. We suspect that this may be because the structure we calculated corresponds to a roaming reaction transition state. For single‑point calculations on transition state structures, how should the convergence threshold be set?We look forward to your reply, and I sincerely appreciate your help!
$molecule
0 1
0 2
C 2.78936999 -0.96732046 0.04528319
C 3.16752399 0.37160654 0.00014119
C 2.20971599 1.38143254 -0.02634681
C 0.84498899 1.06522054 -0.00529681
C 0.54629899 -0.27001446 0.03949919
C 1.43282699 -1.31325646 0.06505719
H 3.53992899 -1.74710846 0.06427519
H 4.21769999 0.62990054 -0.01596181
H 2.51285799 2.41988554 -0.06462581
H 0.08685899 1.83709254 -0.02871681
H 1.11697199 -2.34691446 0.09699119
0 -2
N -2.36505601 -1.08161646 -0.27403881
O -1.92923101 -2.18295546 -0.18990481
O -2.54719001 -0.20785946 0.52240619
0 1
O -2.10027101 2.75259654 -0.20374381
H -2.46838201 1.89917554 0.04857719
H -2.63282301 3.40751054 0.25455619
$end
$end
$rem
BASIS = aug-cc-pvDZ
JOB_TYPE = SP
EXCHANGE = xygjos
AUX_BASIS= rimp2-cc-pVDZ
UNRESTRICTED true
SCF_CONVERGENCE 5
THRESH 8
MAX_SCF_CYCLES 5000
PURECART =1111
TIME_MP2= true
SYM_IGNORE true
scf_algorithm DIIS
MEM_TOTAL = 20000
MEM_STATIC= 2000
SCF_GUESS FRAGMO
symmetry false
$end
$rem_frgm
scf_algorithm DIIS
$end
Thank you sincerely for your valuable support. I look forward to your continued guidance.
Two minor things: (1) when posting input files, you can use the formatted text button </> to avoid having this GUI gobble up special characters. (2) There is an extra "$end$ in your input.
Those things aside, I just ran your (corrected) input with Q-Chem v. 6.4 and it converges w/o problem in 7 SCF iterations. If you still have a problem after fixing your input, we will need more information about the nature of the failure (e.g., what SCF guess is it using?) and the version of Q-Chem that you are running. Since Q-Chem v. 6.3 you can use SCF_ALGORITHM = ROBUST to try different convergence options automatically, though I did not have to do that to get this to work.