I’ve been trying to use CDFT (with or without CI, Q-Chem 7.0) to calculate diabatic state energies for a proton-coupled electron transfer process without much success. Here’s a prototypical input file. I’m examining different positions for the transferring proton (atom 55), but the remainder of the molecule is fixed.
$molecule
1 6
Fe 0.61270957 0.48479155 0.23526882
N -0.06966848 2.35653181 -0.63835954
C -0.46870444 2.59260596 -1.93608927
H -0.29320258 1.86813012 -2.71550714
C -1.07838690 3.80955761 -2.01073730
H -1.52316093 4.34392896 -2.83360756
N -1.04434141 4.31603945 -0.73158033
H -1.41257153 5.20540929 -0.43390687
C -0.43191123 3.41388956 0.06145994
H -0.26234805 3.53749214 1.11936775
N 2.69340039 1.00329250 -0.07386111
C 3.22999571 2.10776276 -0.69684591
H 2.59997872 2.87471812 -1.11910226
C 4.59135291 2.03503556 -0.66051164
H 5.36059965 2.69450641 -1.02662869
N 4.87744101 0.86121824 -0.00230797
H 5.79663883 0.50322526 0.20226286
C 3.71127800 0.27208547 0.33342139
H 3.61933115 -0.66948123 0.85219694
N 0.52589125 -0.47231331 -1.70157963
C -0.32947836 -1.43645528 -1.98590155
H -1.18407715 -1.71878732 -1.38934149
N 0.01042207 -2.04485327 -3.13510356
H -0.45742745 -2.84545616 -3.52758715
C 1.16614531 -1.46448863 -3.60426841
H 1.64278458 -1.78361161 -4.51627413
C 1.47307963 -0.48810327 -2.70402273
H 2.30349541 0.20089661 -2.70510539
O 0.77697477 1.57159287 2.06570898
C 0.47693771 1.28189608 3.24382204
N 0.02383647 0.09105404 3.61745445
H -0.14012162 -0.67220685 2.94915418
H -0.20524689 -0.07819961 4.58271760
C 0.62643143 2.34210132 4.30547793
H -0.12208929 3.11953576 4.13012424
H 0.50068118 1.95884983 5.31928433
H 1.61174725 2.80113040 4.21098768
O 1.34930324 -1.29728206 0.95098522
C 0.63427757 -2.29777127 1.30653501
O -0.42031017 -2.20947674 1.96865005
C 1.15094416 -3.66395900 0.86710949
H 0.34782788 -4.38558092 1.04571599
C 2.37477789 -4.05201640 1.71088510
H 3.19402681 -3.34777511 1.53331880
H 2.13169569 -4.03812610 2.77730071
H 2.72124710 -5.05781142 1.46190557
N 1.37876082 -3.58523431 -0.57339563
H 2.05282004 -2.84547284 -0.74751155
C 1.79968619 -4.82543292 -1.20242141
H 2.75319645 -5.23561183 -0.83566640
H 1.03046233 -5.58954267 -1.04962140
H 1.89992687 -4.66306883 -2.27963296
O -1.22109611 0.04477974 0.68530482
H -1.33061388 -0.82636508 1.11189130
H -2.17590324 0.18465128 -0.16438037
C -3.13071037 0.32452282 -1.01406557
H -2.54796581 0.67806272 -1.87152849
C -3.85214653 1.44711683 -0.39941413
H -3.24907784 2.35629374 -0.39080046
C -5.07011891 1.55178398 0.16676143
H -5.31511247 2.52976904 0.57953238
C -6.16600101 0.53915449 0.28182625
H -7.11063410 0.97504976 -0.05975464
H -6.31535923 0.25071902 1.32935656
H -5.97202607 -0.36181535 -0.29699506
C -3.66833710 -1.04001435 -1.25198447
H -3.76056819 -1.33333016 -2.29807187
C -3.94526598 -1.97015179 -0.32599777
H -4.27786477 -2.94637604 -0.67441271
C -3.86019179 -1.79960307 1.16006851
H -3.68033896 -0.76020302 1.44456469
H -4.79350168 -2.12491661 1.63110557
H -3.05800467 -2.42104446 1.57680299
$end
$rem
BASIS = 6-31G**
GUI = 2
METHOD = wB97X
SCF_CONVERGENCE = 8
CDFT_BECKE_POP = TRUE
CDFTCI = TRUE
CDFT_PREDIIS = TRUE
CDFT_POSTDIIS = TRUE
CDFT_THRESH = 8
XC_GRID = 000099000590
MEM_TOTAL = 30000
MAX_SCF_CYCLES = 2000
$end
$cdft
-1
1 1 54
5
1 1 54 s
----------------
-1
1 1 55
4
1 1 55 s
$end
The central issue I’ve seen is that the Becke populations printed out just don’t satisfy the constraints as I’ve specified them, and for some proton positions the populations will match the requested charge for the fragments but not the net spin or vice-versa. I’ve verified that the total charge and net spin from the Becke populations is always +1 and 5. This occurs whether I use CDFTCI or two individual CDFT calculations on the same geometry. As such, I don’t really trust the final diabatic state energies. To further test this problem, I have tried using CDFT_POP = FBH, but this produces a segmentation fault every time I try to run the same input. I have tried increasing the total memory up to 50 GB with 8 threads and decreasing the basis set size down to 6-31G as a test and I still get a segmentation fault, suggesting that this is not a memory issue. If anyone has any suggestions for how to resolve these issues, I’d love to hear them. Thank you!