Can this vibrational analysis be accepted as for a geometry optimization stationary point?


Forum Vet
I used DFTB2 to optimize C2H5SH with GAMESS, but the following vibrational analysis told me it was not a stationary point on PES. Yestoday, Dr. Schmidt told me why here there was a warning like this. I used NWCHEM6.8 to do the vibrational analysis by DFT and RHF, the inputs were
echo
start molecule
memory total 12000 stack 4000 heap 500 global 7500 mb

title "Title"
charge 0

geometry units angstroms print xyz autosym
  ...
end

basis
 * library 6-31G
end

dft
 xc b3lyp
mult 1
end

task dft freq
and
echo

start molecule
memory total 12000 stack 4000 heap 500 global 7500 mb

title "Title"
charge 0

geometry units angstroms print xyz autosym
  ...
end

basis
 * library 6-31G
end

SCF
RHF
END

task scf freq

I got in the DFT case:


Normal Eigenvalue ||    Projected Derivative Dipole Moments (debye/angs)
Mode [cm**-1] || [d/dqX] [d/dqY] [d/dqZ]
------ ---------- || ------------------ ------------------ -----------------
1 -0.000 || 0.001 -0.000 0.211
2 -0.000 || -0.089 -0.121 0.005
3 -0.000 || 0.070 0.094 0.001
4 -0.000 || -0.001 0.000 -0.163
5 0.000 || -0.020 -0.027 0.001
6 0.000 || 0.003 0.002 0.298
7 251.813 || -0.001 -0.001 -0.036
8 327.340 || -0.049 0.277 -0.000
9 455.218 || -0.002 -0.001 -0.758
10 717.307 || -0.034 0.188 0.000
11 859.934 || -0.004 0.001 0.365
12 904.596 || -0.333 -0.063 -0.005
13 1070.980 || 0.105 -0.371 0.000
14 1109.831 || 0.011 -0.014 -0.030
15 1168.240 || 0.203 0.203 -0.002
16 1328.353 || 0.015 -0.033 -0.170
17 1392.799 || 0.483 -0.868 0.011
18 1481.538 || 0.275 0.377 -0.002
19 1542.006 || -0.002 0.010 0.490
20 1552.169 || -0.173 0.287 -0.013
21 1568.989 || 0.200 0.247 -0.019
22 2120.546 || 0.906 0.775 -0.001
23 2892.624 || 0.738 -0.294 0.399
24 2965.138 || 0.438 -0.171 -0.595
25 3033.997 || -0.354 -0.643 -0.003
26 3104.707 || 0.595 -0.449 0.002
27 3116.689 || 0.016 -0.009 -0.799
----------------------------------------------------------------------------



 

----------------------------------------------------------------------------
Normal Eigenvalue || Projected Infra Red Intensities
Mode [cm**-1] || [atomic units] [(debye/angs)**2] [(KM/mol)] [arbitrary]
------ ---------- || -------------- ----------------- ---------- -----------
1 -0.000 || 0.001924 0.044 1.876 1.582
2 -0.000 || 0.000982 0.023 0.957 0.807
3 -0.000 || 0.000596 0.014 0.581 0.490
4 -0.000 || 0.001157 0.027 1.128 0.951
5 0.000 || 0.000049 0.001 0.048 0.041
6 0.000 || 0.003839 0.089 3.743 3.156
7 251.813 || 0.000057 0.001 0.055 0.046
8 327.340 || 0.003427 0.079 3.341 2.817
9 455.218 || 0.024917 0.575 24.291 20.480
10 717.307 || 0.001583 0.037 1.543 1.301
11 859.934 || 0.005771 0.133 5.626 4.743
12 904.596 || 0.004993 0.115 4.867 4.104
13 1070.980 || 0.006461 0.149 6.299 5.311
14 1109.831 || 0.000052 0.001 0.051 0.043
15 1168.240 || 0.003564 0.082 3.474 2.929
16 1328.353 || 0.001311 0.030 1.278 1.078
17 1392.799 || 0.042811 0.988 41.734 35.187
18 1481.538 || 0.009436 0.218 9.199 7.756
19 1542.006 || 0.010398 0.240 10.136 8.546
20 1552.169 || 0.004863 0.112 4.740 3.997
21 1568.989 || 0.004410 0.102 4.299 3.625
22 2120.546 || 0.061632 1.422 60.082 50.657
23 2892.624 || 0.034234 0.790 33.373 28.138
24 2965.138 || 0.024935 0.575 24.308 20.495
25 3033.997 || 0.023362 0.539 22.774 19.202
26 3104.707 || 0.024058 0.555 23.453 19.774
27 3116.689 || 0.027675 0.638 26.979 22.746
----------------------------------------------------------------------------

in the RHF case


Normal Eigenvalue ||    Projected Derivative Dipole Moments (debye/angs)
Mode [cm**-1] || [d/dqX] [d/dqY] [d/dqZ]
------ ---------- || ------------------ ------------------ -----------------
1 -0.000 || 0.003 0.002 0.346
2 -0.000 || -0.024 -0.032 -0.002
3 -0.000 || -0.096 -0.126 0.005
4 -0.000 || 0.001 0.000 0.173
5 0.000 || 0.076 0.100 -0.001
6 0.000 || -0.000 -0.001 0.166
7 362.271 || -0.001 0.005 -0.008
8 367.001 || -0.034 0.275 0.000
9 568.596 || -0.001 -0.002 -0.782
10 751.215 || -0.089 0.276 -0.000
11 962.902 || -0.005 -0.001 0.291
12 997.976 || -0.370 -0.179 -0.004
13 1131.567 || 0.225 -0.275 -0.001
14 1219.019 || 0.012 -0.007 0.008
15 1265.285 || 0.216 0.303 -0.002
16 1440.273 || 0.009 -0.022 -0.146
17 1518.070 || 0.525 -0.888 0.008
18 1617.833 || 0.271 0.336 -0.002
19 1662.843 || -0.002 0.009 0.469
20 1670.485 || -0.184 0.222 -0.011
21 1689.357 || 0.229 0.309 -0.018
22 2127.666 || 0.801 0.811 -0.002
23 2932.493 || 0.714 -0.325 0.414
24 3002.990 || 0.459 -0.202 -0.570
25 3073.851 || -0.342 -0.759 -0.006
26 3132.781 || 0.649 -0.531 0.003
27 3145.187 || 0.018 -0.012 -0.877
----------------------------------------------------------------------------



 

----------------------------------------------------------------------------
Normal Eigenvalue || Projected Infra Red Intensities
Mode [cm**-1] || [atomic units] [(debye/angs)**2] [(KM/mol)] [arbitrary]
------ ---------- || -------------- ----------------- ---------- -----------
1 -0.000 || 0.005175 0.119 5.045 3.999
2 -0.000 || 0.000068 0.002 0.066 0.053
3 -0.000 || 0.001090 0.025 1.063 0.842
4 -0.000 || 0.001305 0.030 1.272 1.008
5 0.000 || 0.000679 0.016 0.662 0.525
6 0.000 || 0.001194 0.028 1.164 0.923
7 362.271 || 0.000004 0.000 0.004 0.003
8 367.001 || 0.003334 0.077 3.250 2.576
9 568.596 || 0.026523 0.612 25.856 20.496
10 751.215 || 0.003640 0.084 3.549 2.813
11 962.902 || 0.003664 0.085 3.572 2.831
12 997.976 || 0.007306 0.169 7.123 5.646
13 1131.567 || 0.005482 0.126 5.344 4.236
14 1219.019 || 0.000011 0.000 0.011 0.008
15 1265.285 || 0.006003 0.139 5.852 4.639
16 1440.273 || 0.000952 0.022 0.928 0.736
17 1518.070 || 0.046084 1.063 44.925 35.612
18 1617.833 || 0.008057 0.186 7.854 6.226
19 1662.843 || 0.009525 0.220 9.286 7.361
20 1670.485 || 0.003608 0.083 3.517 2.788
21 1689.357 || 0.006425 0.148 6.263 4.965
22 2127.666 || 0.056353 1.300 54.936 43.547
23 2932.493 || 0.034098 0.787 33.240 26.349
24 3002.990 || 0.024977 0.576 24.349 19.301
25 3073.851 || 0.030016 0.692 29.261 23.195
26 3132.781 || 0.030497 0.704 29.730 23.567
27 3145.187 || 0.033326 0.769 32.488 25.753
----------------------------------------------------------------------------

Can this be accepted as a stationary point for geometry optimization.

I know this requires rich group knowledge, structural intuition and art, just as indicated in the GAMESS test for [Ni(NH3)6]2+, etc.

Very Best Regards!

Forum Vet
The OPTTOL in GAMESS is set to 1.0D-10 for DFTB, and also to 1.0D-11, and the related RHF
frequency analysis is
    1      22.790    A        4.789345    0.040791
2 0.084 A 6.890251 0.000000
3 0.023 A 6.885812 0.000000
4 0.034 A 6.890374 0.000000
5 37.846 A 3.929379 0.036886
6 172.729 A 2.086567 0.022109
7 363.748 A 3.410336 0.077421
8 434.011 A 1.023444 0.000317
9 615.593 A 1.038901 0.723858
10 751.224 A 4.094947 0.087465
11 979.847 A 1.089670 0.094492
12 1038.404 A 1.197033 0.161830
13 1134.356 A 1.978351 0.117989
14 1237.757 A 1.203437 0.002447
15 1264.022 A 1.567909 0.140356
16 1456.219 A 1.189515 0.011190
17 1530.772 A 1.269786 1.076701
18 1617.789 A 1.219209 0.186595
19 1662.150 A 1.039877 0.219738
20 1669.327 A 1.045953 0.090570
21 1686.310 A 1.098514 0.141698
22 2126.253 A 1.037980 1.311076
23 2985.475 A 1.057992 0.829015
24 3016.134 A 1.106556 0.438324
25 3073.863 A 1.036438 0.693348
26 3132.885 A 1.099075 0.716761
27 3145.120 A 1.101099 0.797947
No imaginary frequency existing in GAMESS.
According to an article published in JACS in 1995, the ab initio vibrational analysis of the minimum of D-H exchange of CD4 and Zeolite cluster had 5 imaginary frequencies -200.414, -127.723, -56.788, -22.541, -12,698cms-1, which was explained as a spurious consequence of the Cs symmetry of the cluster. The transition state also has 3 imaginary frequencies -1423.384, -155.086 and -42.834 cms-1.
I think the optimized geometry at this level makes this intrisinc and only a very good first guess may eliminate it, thus the warnings from GAMESS perhaps can be ignored with an low enough OPTTOL.

Special, sincere thanks should be given to Dr. Schmidt for his clarification of the technical meaning of this warning in GAMESS and instructions for me to beigin with GAMESS, especially geometry optimization, etc., and long time lucid explanations as well as useful advices on my ceaseless questions; and all those on this forum and NWCHEM Github to teach me to compile NWCHEM.

Forum Vet
It seems there is no obvious problem to identify the dftb2 optimized one is a minimum for some practical use, e.g., reorganization energy calculation for some molecules, and the first excited state parameters of TDDFT TPA calculation of thiol using a certain functional tried by me .
I used MP2 and a large basis set to optimize it and RHF for frequency analysis, but still got the warning.
The MP2 semi-numerical HESSIAN calculations using the large basis set eliminate the warning of the geometry from MP2 with the same basis set but still get it for that from DFTB2.
Please note the thermochemical composite G3MP2 in GAMESS using the MP2 and DFTB2 optimized initial geometries, respectively, all gave gaseous heats of formation of -47.82 kJ/mol-1 at 298.15K, in good agreement with that of -46.15kJ/mol-1 in NIST Chemistry Webbook.

Forum Vet
The following references are related to the imaginary frequencies of a geometry
1.In an article on BC3 honeycomb published in J. Phy.Chem.C in 2011
"Our fragment structure II is in fact a ninth-order saddle point. Geometry optimization following the imaginary frequencies led us to structure I, which is the most stable isomer found in our calculations. "
"However, the completely planar structure of the 1,2,3-C6H4(BH2)2 molecule has two imaginary frequencies. Thus, we proved that the repulsion between hydrogen atoms is responsible for the instability of isomer II in the C6(BH2)6 case. "
2.Prof. Gordon commented on an article published in JACS in 1991
"...The twisted group IVB carbenes (2b) have two imaginary frequencies-one which corresponds largely to rotation about the M-C bond leading to the planar structure (2a) and one which entails bending of the MH2 fragment such that the coordination about the metal goes from trigonal planar to pyramidal. ..."
"If all else(meaning the methods to eliminate the imaginary frequencies resulting from a flat PES and numerical errors) fails, it is probably better to pretend that the imaginary frequency is real and add the corresponding vibrational free energy contribution. However, this needs to be systematically tested".
4. In an article published in Journal of Organometallic Chemistry in 1995 on tetrafluorocyclodisilazanes on References and notes 18
"In the case of 8c a very small imaginary frequency ( ~ 7 cm- ’) corresponding to rotation of the exocyclic silyl groups was present. Because the effect of such a small imaginary frequency on the energy or the structure is neglegible, we have not reoptimized this structure."
5. In an article pubished in Chemical Physics in 1992
"However the C, keto form contains a planar secondary amine moiety, and it is therefore not surprising that it represents a transition state with an a” transition vector (imaginary frequency 96i cm- ’ )."(Singlet excited-state intramolecular proton transfer in 2- ( 2 ’ -hydroxyphenyl ) benzoxazole: spectroscopy at low temperatures, femtosecond transient absorption, and MNDO calculations)
6. In an article published on Computational Materials Science in 2015:
" The results, shown in Fig. 4, demonstrate that if the frequency is very low (<39.0 cm-1), a high-accuracy frequency calculation can remove up to 42% of the imaginary frequencies. However, high accuracy re-optimization is less helpful for larger imaginary frequencies. In contrast, the two molecular geometry perturbation strategies work extremely well: they remove at least 21 out of the 24 imaginary frequencies for the whole range of imaginary frequencies."
7. In an article published in J. Chem. Edu. in 2002
"Often, as in the case of (PH3)2Ir(H)2(H2)Cl discussed later in this paper, optimizations with a symmetry plane present (Cs in this example) result in a converged structure with one or more imaginary frequencies that correspond to motions that would break the symmetry plane. To converge a structure to an energy minimum (no imaginary frequencies), the computation must be carried out again in a lower symmetry group (here C1). "
8. In an article published in J, Phy. Chem. A in 2012
"the u-B3LYP/6-31G(d)-optimized 1B1 structure had one imaginary frequency, corresponding to rotation of the perpendicular CH2. "
9. In an article published in JACS in 1987 entitled A Theoretical Study of Thermal Reactions of Bicyclo[ 2.1.0 ]pent-2-ene.
"A similar analysis for 4 gave one imaginary frequency (134i cm"1). This vibrational mode of a" symmetry is mainly a twist of the CH2 group showing that 4 is a transition state for the conversion of 3 into itself."
10. In an article published on pyrimidine in JACS in 1992
"The coordinate associated with this imaginary frequency is an out-of-plane one of bl symmetry which lowers the symmetry of the molecule from C2v to Clh, the reflection plane being the zx plane. The subsequent optimization of the geometry within Clh symmetry leads to a minimum corresponding to a nonplanar geometry for pyrimidine in the 3Bl(n7r*) state at an energy 68 cm-I below that of the corresponding state in the C, geometry."
I think most of these comments are suitable.


Forum >> NWChem's corner >> Running NWChem