Hi Chetan,

At 1st check this: does your input look like this?

...

$DATA

name of your molecule or empty

C1

Atomsymbol Atomicnumber X Y Z

O 8.0 0.0000000000 0.0639340000 -0.0001140000 this line should be like this

...

$END

1. It might happen that your metal is not correctly assigned to an atom symbol and/or atomic number, e.g. Mo is shortened to M or replaced with X and the atomic number is set to 0.0.

2. If this section is OK, then something is not correct in the basis function definitions. It is possible that your metal does not have a basis set for some elements at your calculation level. You can check it and may download if that is missing in GAMES from:

https://www.basissetexchange.org/

E.g. ANO-RCC has basis sets for all real elements (I mean a very few transuranium elements is missing only), LANL2DZ covers the majority of elements, but 6-311G have basis sets for some alkali and alkali-earth but not transition metals only. 6-311G++** has even less.

In the program documentation, you can find information about how to insert and use mixed basis sets or those which are not included in the most recent GAMESS version.

Good luck,

Laszlo

Laszlo Jicsinszky Thank you.

I will look into it sir.

Hi Chetan,

Finally, what was the problem? No basis set or element error in the input file?

Hi Laszlo Jicsinszky , I have tried using the 6-31 basis set and then got the error regarding the point group usage of C1. I changed it to the Cnv (as given in the GAMESS manual for the linear molecules). Still I am getting the error sir.

I am attaching the input and log files for your reference.

Sorry for the delayed response.

Here are the files sir

Hi Chethan,

I am sending you an inp file with some examples and a log file. The SCF did not converge, and I am almost sure that something is wrong with your molecule.

The problem was with the basis set (apart from the incorrect structure). You wanted to use 6-31G what contains the parameters for Zn but not the polarization and diffusion extended sets. When you put N?FUNCT and DIFFS? =true, you command GAMESS to use polarization and diffusion function(s). As I wrote to you before, you need to check your element availability in the (supposed to be) used basis set.

So first, I suggest you choose an appropriate basis set, because 6-31G+, +*, +**, ++, ++** do not contain parameters for Zn.

1. I corrected the BASIS line. It would have worked but SCF not converged.

2. I removed the instruction for parallel use. First, because in optimization is practically useless, and second, my experience is that using multiprocessor mode in geometry optimization, the calculation is considerably slower than using only one processor. But I am not sure how the linux version is working, I used it under linux a long time ago. Perhaps, that line is necessary.

3. The $Data axis ... is not necessary.

4. Cnv is written back to C1. It is working.

5. 6-31G** needs more memory, but for 6-31G the default is enough.

I also recommend performing some preliminary optimizations before you are moving to a higher basis set, using the previously got coordinates. Extract them from the log file because any conversion will result in loss of decimal digits.

1. molecular mechanics (Avogadro has this option).

2. A semiempirical method, perhaps MNDO/AM1/PM6, etc. what you want and implemented in GAMESS(RM1 can give false geometry)

3. STO-3G

4. 3-21G

5. 6-31G

then you can move to 6-31** etc.

6. You may use solvent, but sometimes the optimization not only takes for a long time, and even any convergence limits are not always reached. Then a single point calculation can help.

If your molecule is in solution, the final energy calculation should also use a solvent.

Besides the above, it seems that your molecule contains errors. I am attaching a picture after MM optimization. It seems there are two 'free' hydrogens. Is that OK?

I removed them, but the SCF still did not converge.

Thank you Laszlo Jicsinszky sir for the detailed explanation.

I will surely go through your suggestions and work on it sir.

regarding the image sir, the water molecule in the first image is coordinated with the Zn metal atom sir. and the hydrogen atoms in the second image must not be present sir.

The oxygen is in the special position and by the symmetry there are four hydrogen atoms.

I will look into the structure again to resolve that hydrogen issue and use the inputs give by you sir to run the optimization calculations sir.

Thank you for the suggestions.

With regards,

Chethan

You cannot run an optimization on structure what I corrected, because the SCF will not converge. So, first, find the error in the structure then you can test that molecule for running.

The composition of 1c is (Avogadro)

C12H10O5Zn4 is this correct?

Additionally, it is strange because Zn is double ionized, and the present structure should have 4+ charges. Is it true?ű

The next is that if only two Zn is in the "middle" between the carboxylic oxigens, then only one water is coordinated? Or, on both sides one-one water?

What is strange also that the formula is identical both in your original inp file and in my corrected one, but the MW is different!

1c=497.741, the corr=495.725, and the number of atoms is identical in both molecules (31)! Very strange.

Yes sir, I have checked the structure and as I told earlier that the water molecule attached to the metal atom is in the special position (oxygen). Due to the symmetry I got 4 hydrogen atoms attached to the oxygen atom.

I have corrected it now and given the calculations.

Hence due to the presence of the extra 2 hydrogen atoms (due to symmetry) in the oxygen atom, there is a difference in the MW by 2.

Hi Chethan,

Unfortunately, I am not a theoretical chemist and I am not sure that I can give you always the complete answers to your questions.

Regarding the use of symmetry, I would not recommend it. Although the calculation is faster, my opinion is that the C1 version always can lead to a more realistic picture.

I would use a simpler starting structure, like benzoic acid. I created some structures to help you.

I am attaching a zip file that contains some possible structures. Both with 1 and 2 Zn. They are the monomers, but I inserted a unit that can be used as a "polymer" template. When you could perform the correct calculations on the simple molecules, you can move forward to using $ELG to simulate polymers. I have not used it, so my knowledge is restricted to this mode of calculations.

Two notes on Avogadro:

1. It does not calculates correctly the charges. You need to set and change manually.

2. I am not sure that the multiplicity is correctly calculated as well.

So, what I can advise you:

- open the mol files with Avogadro, the if you need modifications, do it

- save your file

- perform MM optimization(s) (max step is 9999)

- create the GAMESS input file, BUT correct charges and multiplicity, and the potential basis sets with an external text editor.

- run GAMESS.

When the monomers are OK, go to dimers, then to the polymer.

I am almost sure that I did not take into account all possible molecules, sorry for it. YOu need t complete it.

If the SCF does not converge, in the control section ICUT can control it. Sometimes it helps if the SCF cut is temporarily higher (e.g. 1e-7), and when the geometry converged, you can re-optimize it with the default cutoff. But, you must use the coordinates from the log file.

The preliminary MM optimization is mandatory because otherwise, the SCF will not converge. The zip file contains an example.

I am afraid the further details are less interesting for the public, if you need some more information or help, send me a private gmail (ljicsinszky). I am rarely using the univ mail because it is full of unimportant information (for me).

Thanks a lot sir, I will work on the suggestions provided by you.

It has definitely helped me to increase the knowledge regarding the usage of the GAMESS.

Hi Chethan,

Meantime I could run your file (after the correction). I tried to use your original file but using +6 charge it did not converge. As I wrote before, the problem is the missing charges.

It gave a very strange structure after optimization at AM1/RM1 (similar), MNDO, and STO 3 level. These three are very different from each other.

If you stick to your molecule 1c (after the removal of two Hs) then I suppose you need to add 1-1 OH to Zns to reduce the effect of charges.

But, anyway, I advise you to calculate first simplified structures.

Good luck,

Laszlo