Can Dalton calculate oscillator strength

Find answers or ask questions regarding Dalton calculations.
Please upload an output file showing the problem, if applicable.
(It is not necessary to upload input files, they can be found in the output file.)

xiongyan21
Posts: 300
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: Can Dalton calculate oscillator strength

Post by xiongyan21 » 21 Sep 2020, 09:10

Using the geometry found by spin-flip ORMAS, the vertical T1(MRPT2)- S0(MCQDPT, by spin-flip MRMP2) can be 0.09 eV deviating from 3.99 eV, so is the gap of T1(MCQDPT)- S0(MCQDPT, by spin-flip MRMP2) when the active orbitals and active electrons are all two. The triplet state energy calculation includes another kind of FCI created by Ames Lab, a full second order CI with the exact Hessian and a mcscf orbital optimization in advance, where the active orbitals and active electrons are all two. Perhap spin-flip MRMP2 only allows some of the methods. MCQDPT is only for CASSCF wavefunctions.

The spin-flip ORMAS optimization uses delocalized internal coordinates to accelerate its convergence. The basis set here is 6-31+G(d,p).

Using the geometry found by spin-flip TDDFT PBE0, also a planar one, the vertical T1(MRPT2)- S0(MCQDPT, by spin-flip MRMP2) can be about 0.29 eV deviating from 3.99 eV.


For furan, the quantum chemical calculation of the T1-S0 gap is challenging, but in the reference, professors perhaps use the triplet geometry from experiments or through their advanced chemical attainments. It seems perhaps we can say the vertical gap calculated by MRMP2 and spin-flip MRMP2 based on the triplet geometry found by spin-flip ormas can agree with the experimental data, i.e., 3.99 eV.


All the above use GAMESS, and natural MOs are used. I will use Dalton2018 to calculate its phosphorescence.

Perhaps this also applies to hemexazole, there being no reported phosphorescence and fluorescence.

Very Best Regards!
Last edited by xiongyan21 on 10 Oct 2020, 04:04, edited 4 times in total.

xiongyan21
Posts: 300
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: Can Dalton calculate oscillator strength

Post by xiongyan21 » 23 Sep 2020, 10:30

For the first excite state energy of furan, TDDFT b3lyp fails, but EOMCCSD, TDDFT TD-CIS, TDDFT CMQTP01 can describe it well. Surprisingly enough, cr-eomccsd(t) almost fail.

Multiconfiguration Pair-Density functional Theory, free of delocalization error, developed by Prof. D. G. Tuhlar, a PhD graduate from California Institute of Technology; now a member of American Academy of Arts and Sciences and National Academy of Sciences in US, an numerious prizers and awards winner, and a US university professor, et al, also can describle it, using tPBE as well as FTBLYP functionals,

I say this, because all the above have been tried, when compared with the results in the article of Prof. Holland in Daresbury Laboratory in UK, and other professors in Russian, Germany, Poland and France in Journal of Molecular Spectroscopy in 2015, dealing with normal and fully deuterated furan, but the results now refer to those from LR-CCSD/CC and EOMCCSD in the article.

TD-CIS can describe both excitation energies, but predicts both have zero oscillator strengths

All the above uses GAMESS and refers to vertical ones using preoptimized geometry having C1 symmetry with mp2 and 6-311++G(2d,2p), where Multiconfiguration Pair-Density functional Theory employs FCI created by Ames Lab, and all the excitation energy calculations use aug-cc-pvdz.

According to ten states GAMESS EOMCCSD (left-eigenstate to obtain oscillator strengths), perhaps some later degenerate states not appearring, the first ten states are
A 6.024 eV
A 6.458 eV
A 6.563 eV
A 6.752 eV
A 6.752 eV
A 7.260 eV
A 7.281 eV
A 7.432 eV
A 7.437 eV
A 7.744 eV
...
first state:
X-COMPONENT Y-COMPONENT Z-COMPONENT
----------- ----------- -----------
TRANSITION: 0.000 EV ==> 6.024 EV

RIGHT TRANSITION MOM. -0.00000000 -0.00000000 0.00000017
LEFT TRANSITION MOM. -0.00000000 -0.00000000 0.00000017
DIPOLE STRENGTH 0.00000000 0.00000000 0.00000000
OSCILLATOR STRENGTH 0.00000000 0.00000000 0.00000000

---------------------------------------------------------------------
EOMCCSD DIPOLE MOM. -0.150300087 0.298756479 -0.000000000
---------------------------------------------------------------------
second state
X-COMPONENT Y-COMPONENT Z-COMPONENT
----------- ----------- -----------
TRANSITION: 0.000 EV ==> 6.458 EV

RIGHT TRANSITION MOM. 0.90864944 0.45712963 0.00000000
LEFT TRANSITION MOM. 0.99324590 0.49968900 0.00000000
DIPOLE STRENGTH 0.90251234 0.22842264 0.00000000
OSCILLATOR STRENGTH 0.14279217 0.03614019 0.00000000

third state
X-COMPONENT Y-COMPONENT Z-COMPONENT
----------- ----------- -----------
TRANSITION: 0.000 EV ==> 6.563 EV

RIGHT TRANSITION MOM. -0.00000000 -0.00000000 -0.50808995
LEFT TRANSITION MOM. -0.00000000 -0.00000000 -0.52937653
DIPOLE STRENGTH 0.00000000 0.00000000 0.26897089
OSCILLATOR STRENGTH 0.00000000 0.00000000 0.04324835
...
the latter three are very weak. Several cr-eomccsd(t) result expressions make the 5 and 6 state excitation energy close to those of EOMCCSD in the article, e.g., delta-CR-EOMCC(2,3), D ones are 6.982 and 7.037 eVs, respectively. Perhaps you can found why eomccsd a little overestimates the two excitation energies through the common usages of the expressions(one not listed here). TDDFT CMQTP01 prediction of the first state’s oscillator strength and nature is quite different. Thus, previous excitation calculations using TDDFT CMQTP01 for pesticides should be examined about its nature, even the calculated can agree with the experimental for energies. Here, B3lyp, predicts similar oscillaor strengths except the second weaker and the third much stronger.Thus, so are B3lyp, etc. There is an article comparing the ocsillator strengths from EOMCCSD and those from TDDFT.

CR-EOMMCCSD(t) in GAMESS is developed by Prof. P. Pieccuch, a PhD from Poland, now a chemistry, also an adjunct physics professor in Michigan State University in US, holding many domestic and international academic positions, et al.

I will try MCSCF and coupled cluster methods, and use TDDFT to verify Prof. Tozer's diagnostics, of course an approximate one and having exceptions, in Dalton2018. I will try to change the FCI configuration with more functionals in Multiconfiguration Pair-Density functional Theory method.

Perhaps a weak absorption may be caused by different reasons.
Prof. Palmer in Edinburgh University in UK and other proferssors in UK and EMSL in PNNL in US published an article in Chemical Physics where the Rydberg states are at 5.91, 6.472, 6.752, 7.283, 7.377, 7.427 eVs, etc. and the singlet valence-excited states are 5.80, 6.04, 7.80 eVs, etc. This reference omits the T1-S0 of 3.99 eV. it seems cr-eomsscd(t) results can be compared with the experimental excitation energies although the third oscillator strength is too weak, perhaps a forbidden one. The cr-eomccsd(t) excitation energies, e.g., various corrected delta-cr-eomccsd(t) ones, not all of I, not all of A, can be compared with the first ten from the experimental values in the article, where DEL(IIA) 6.461 eV of the eomccd degenerate 6.752 eV matches the experimental 6.472 eV (4th of the combined list). GAMESS has absorbed many useful result expressions from various published couple-cluster references to meet different demands.

What is the usage of delta_R MO molecular overlap diagnostic in Dalton2018 and is there any reference?

Very Best Regards!
Last edited by xiongyan21 on 30 Sep 2020, 22:24, edited 13 times in total.

xiongyan21
Posts: 300
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: Can Dalton calculate oscillator strength

Post by xiongyan21 » 24 Sep 2020, 12:09

I have found the metric in Journal of Chemical Theory and Computation.

Very Bet Regards!

hjaaj
Posts: 394
Joined: 27 Jun 2013, 18:44
First name(s): Hans Jørgen
Middle name(s): Aagaard
Last name(s): Jensen
Affiliation: Universith of Southern Denmark
Country: Denmark

Re: Can Dalton calculate oscillator strength

Post by hjaaj » 01 Oct 2020, 14:05

FYI: the primary developer of pair-density functional theory is prof. Laura Gagliardi and not prof. Truhlar. Please give credit where credit is due!

xiongyan21
Posts: 300
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: Can Dalton calculate oscillator strength

Post by xiongyan21 » 01 Oct 2020, 23:12

Dear Prof. Jensen
Pair-density functional theory can be dated back earlier, e.g., in 1990s in journals of physics. Multifunctional pair-density functional theory in GAMESS is based on the recent papers coauthored by the two distinguished professors and their students, all in the same university in United States.

Very Best Regards!

xiongyan21
Posts: 300
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: Can Dalton calculate oscillator strength

Post by xiongyan21 » 06 Oct 2020, 11:38

For a patented pesticide (CAS:1070975-53-9), a 48 atom molecule developed by a group led by Prof. Guangfu Yang in our university, a famous professor.
The optimization, a little challenging, first uses Monte Carlo minimum search and then applies B3lyp and 6-31+G(d). The TD-CIS one employs 6-31+G(d,p). Something should be added to eliminate the linear dependence warning, which may be easy to be overcome especially when using HF, or it may not be very strickly treated in order to accelerate convergence speed. For this molecule, this should be treated, although perhaps not very strictly, because there is an eigenvalue significantly smaller than a thresold, which will cause a energy change which perhaps cannot be ignored. After changing the input, RHF converges one step further from the previous one, but it takes many more steps for a DFT optimization.

TD-CIS calcualtion exhibits the second largest absorption very close to (0.02 eV deviance) the second largest,the detection wavelength in a mixture of CH3CN and water with an Agilent 1200 in an article published in China in 2013, where the serrated peaks having largest absorptions, and the largest serrated one in the reference is about 0.24 eV deviating from that of the calculation where several absorptions around there. TD-CIS overestimates another peak significantly. Of course, the small eigenvalue problem has been treated. Hessian should be calculated to verify whether there is any imaginary frequencies. It will take long time to do it with b3lyp.

TD-CIS calculation using HF optimized geometry cannot give the experimental oscillator strength order and gives the 7th root the second largest, whose position close to the experimental largest.

Excitation energy calucaltion using b3lyp significantly undermines this two roots, but it seems it may agree with another root.

The above calculations use GAMESS for now.

Very Best Regards!

Post Reply

Who is online

Users browsing this forum: No registered users and 6 guests