((Sigh)) Go back to the top, re-read the thread and all of the instruction, advice, and suggestions that have already been given, read some papers (including some on the fundamentals of computational chemistry), do some work yourself, try a bunch of different basis sets, CIS/TD-DFT/etc. methods...Try to answer for yourself what you really mean by "suitable?" and remember that in any case you will only *model*, you will not *reproduce* any given experimental result.

I would suggest using cam-b3lyp or minnesota functionals for uv-vis spectra calculation.

Secondly, go to the EMSL basis set library, you will find suitable basis sets for I atom. I hope this helps.

Regarding recommendable benchmark sets, check out these papers:

Article Extensive TD-DFT Benchmark: Singlet-Excited States of Organi...

Article A comparative study of DFT calculated and experimental UV/Vi...

Article Benchmark study of UV/Visible spectra of coumarin derivative...

You will have to decide whether you want to improve those studies or whether you generate a new standard; in contrary to bond lengths and ionization potentials theres is no normed molecule set so far.

Dear Sir,

Please give us ( the answerers) the source of the value of the experimental data of Iodine (e.g. gas phase or solution).

Please read the papers of the George Long et al.

https://pdfs.semanticscholar.org/2dbf/bab4ad77eea819afbed8800fe7a5f26da86e.pdf

and this link too.

https://hal.archives-ouvertes.fr/jpa-00208967/document

On other part prior you try to compute the complex spectrum of iodine, try to compute the molecules cited with hints in the book (please order it in your library)(joined as files 1 and 2).

After this exercise, you can try to compute the iodine spectrum with an another basis set.

Good luck

JRG

Thanks a lot for your suggestion.But I want to compute UV-VIS spectra of iodine.In this book no such example regarding UV-VIS spectra is found so that I can make a benchmark study of UV-VIS spectra.@ Jean Rene Grezes

1. EMSL/BasisSetExchange website for basis sets (https://www.basissetexchange.org/)

2. Or Stuttgart/Cologne website for pseudopotentials/basis sets (http://www.tc.uni-koeln.de/PP/clickpse.en.html)

3. B3LYP is not good for most things...definitely not optimized for large atoms or UV spectra...try a recent Minnesota functional or a recent Head-Gordon functional (wB97X-D, for example) - you may also need a TD-DFT approach (see references given by an previous answerer)

4. I2 is such a small molecule - why the interest in modeling its UV spectrum??

Dear Sir,

It seems that you don´t answer my enquiry concerning the experimental data of iodine.

Please read this link:

https://pdfs.semanticscholar.org/6d5e/b201e6c214c93903f52c95247a9c6fe2cd97.pdf

here an abstract of this paper about Iodine computation:

III. Computational Methods The Gaussian program was used to create the potential energy curves for the ground state, the lowest three singlet excited electronic states, and the lowest four triplet excited electronic states. The program calculates the allowed electronic energy levels using numeric approximations. Gaussian uses basis sets and the principle of superposition to arrange the wavefunctions as linear combinations, called Gaussians3 . The program uses the Moller-Plesset second-order perturbation theory (MP2) to calculate the ground state potential energy curve. And it uses the Configuration Interaction-Singles (CIS) theory to calculate the excited state potential energy curves. CIS theory is an extension of the Hartree-Fock approximation3 .

On other part, if you want to obtain good results with Gaussian, please read my answer again and try to tame or win over Gaussian with small molecules other than iodine. According to your results, you ´ll earn good results with iodine in due course.

Read the conclusion of the cited paper:

VI. Conclusions The iodine molecule was found to be approaching its convergence limit near 500nm. The Gaussian program predicted this to occur at a much higher energy. Gaussian also predicted the term value and the energy gap ΔE(I*) from X( 1 Σg + ) to B( 3Π0u + ) to be much greater than accepted literature values. This indicates that Gaussian is insufficient at predicting relative energy levels. Gaussian was however able to predict reasonable values for the equilibrium bond lengths indicating that it may be suitable for determining the shape of potential energy curve. However, these values were slightly high due to the neglect of relativistic and spinorbit coupling effects. The Morse potential energy curves provided a more accurate representation of the iodine molecule than the harmonic oscillator model. The harmonic oscillator model is only valid very near the equilibrium bond length. The most intense transitions occurred from the v”=0 level to the v’=34-36 range as predicted by the Franck-Condon principle. The Stokes shift was not observed in I2. The comparative values of the force constant, equilibrium bond

Good luck and tame the Gaussian !!

JRG

Another useful link:

https://www.uni-ulm.de/physchem-praktikum/media/literatur/Experiments_in_Physical_Chemistry_S436-454_2.pdf

Subhanjan Majumder I have performed a geometry opt and freq on

I2 in the past and not show if that will answer your question in the choice of a basis set to use.

See the attached(output file).

Thanks a lot for your suggestion and valuable information Jean Rene Grezes Bryan Moloto Allen Clabo Jürgen Weippert . Actually I am very new in this field. So first I want to know however I able to do it correctly or not. I calculate the UV-VIS spectra of Iodine bt can not get a satisfactory result. For that very reason I need an example molecule for which UV-VIS calculation usuing gaussian was already done by someone and he/she get an satisfactory result i,e simillar to experimental literature value. I want to repeat that calculation usuing gaussian to see that am I able to do it or not. It also help me to establish a benchmark study of UV-VIS spectra usuing Gaussian.

It seems that you need to 1) read all of the references that have been given above, both about methods in general, basis sets in general, and the chemistry of I2; 2) look at the gaussian output file posted and see how it compares both to your results and the experimental value (which is what, by the way? and what is the error from your calculation and from the posted calculation and from the computational literature? You're not the first to have thought of this...); and 3) show us that you are actually doing some work yourself and not just looking here for a quick answer or for someone to do it for you. That's my 2 rupees for the day...

Thanks for your suggestion Allen Clabo I tried to calculate UV-VIS spectra of acrolein and compare it with some literature value. But I cant not find any satisfactory result which is coincide with that literature value. I first optimize the structure and then done energy calculation to get UV-VIS spectra.

I uploaded both the literature value and my work. Can you suggest me what I am doing wrong?

Can you please define what you define as a satisfactory result? TD-DFT is not a method with systematic "chemical precision", so there will always be deviations.

You do not appear to have done anything "wrong." You appeared to have followed the directions in the pdf file to do those computations. The calculations that you are doing are, of course, *models* - they cannot and will not exactly reproduce the results of an experiment. What is the absolute or percent error? What is the typical accuracy of various models and various basis sets? As Jürgen Weippert said, what do think is satisfactory (and why?)? What are the sources of error (geometry, level of theory)? Then ask yourself the same questions for I2 (as I suggested above...). This is not an exact method that gives you the "right" answer.

Thanks for your comment Allen Clabo Jürgen Weippert .Actually the input file format for acrolein structure optimization usuing a perticular basis set is given in that pdf document. Also the UV-VIS spectra was given. I repeat that calculation usuing the same basis set. But my input file format was different to that given value in pdf and the resulted UV-VIS spectra was also different. So I cannot understand wheather I am wrong in computing or the data given in that pdf file was not correct.

UV-VIS spectra of acrolein computrd by me usuing two different basis set is attached herewith.

basis sets are

1. CIS (nstates=6)/cc-PVDZ

2. B3LYP/ cc-PVDZ TD (nstates=6)

** satisfactory result means I have some literature data about UV-VIS spectra that was calculated experimentally usuing spectrophotometer. Now I want to do the same usuing computation methodology. In order to do so I have to corelate both the data that was found experimentally and computationally. When these two data are differ to a large extent then it is not satisfactory.

OK, let us repeat it: TD-DFT is not exact. You will never receive exactly the same spectrum from a TD-DFT calculation and an experiment. That is a fact you have to accept.

If you want to use TD-DFT, you have to check where your spectra are usually shifted because if you stay within the framework of one functional+basis set there is often a certain systematic redshift/blueshift in the experiment vs. theory comparison.

Probably the exercise was meant to show exactly that; what did your course instructor say about this?

OK, let me repeat one last time...no computation is exact - any result (energy, geometry, spectra, etc.) will depend on 1) the level of theory (HF, B3LYP, CIS, TD-DFT, etc.), 2) basis set (cc-pVDZ, etc.), 3) the combination of optimization and UV jobs (a) HF opt. followed by CIS or b) HF opt. followed by TD or c) B3LYP opt. followed by CIS or d) B3LYP opt. followed by TD or e) any other opt + UV combination that could be thought of), 4) lots of other variables...also, if you know experimental values, what are the absolute and percent differences between computation (your method(s) of choice) and experiment? and what do you consider "satisfactory"? I think Jürgen Weippert and I might be done answering the same questions over and over again...P.S. if you know that your "input file format was different to that given," then you are doing a different calculation and the results will be "also different"...(!)

Thanks for your valuable suggestions. Allen Clabo Jürgen Weippert Bryan Moloto Jean Rene Grezes

I cant find any suitable basis set to reproduce the UV-VIS plot of Iodine molecule and tri-iodide ion presented in J. Phys. Chem. A., 2016, 120, 2144. Please help. Allen Clabo Jürgen Weippert Jean Rene Grezes Bryan Moloto Prasanta Bandyopadhyay

((Sigh)) Go back to the top, re-read the thread and all of the instruction, advice, and suggestions that have already been given, read some papers (including some on the fundamentals of computational chemistry), do some work yourself, try a bunch of different basis sets, CIS/TD-DFT/etc. methods...Try to answer for yourself what you really mean by "suitable?" and remember that in any case you will only *model*, you will not *reproduce* any given experimental result.

Thanks for your suggestion Allen Clabo

I am unable to generate parameter for iodide ion in tleap of amber. Amber parameter database doesn't have parameter for it. The force field I am using is ff99SB. Kindly suggest how I can go about it? Thanks a lot in advance. Allen Clabo Jürgen Weippert Bryan Moloto Jean Rene Grezes