Mol_Fra

Dear All,

I wish to use the mol_fra program to analyze the fragments that result from my dynamics using ReaxFF.

Since my system contains the elements C,H,O,N and S, I needed to add the latter one both in the mol_fra.c programme and as well in the Cutoff.dic

I hardly have any programming experience but I added the information for S in the mol_fra.c.

After having compiled the mol_fra.c, the created executable complains on the test case that:

Can’t fine a certain pair for cut off dictionary!

However, I clearly added the bond order cut-off for sulphur (even in both ways), i.e, for example:

C S 0.3

and

S C 0.3

Somehow, it does not seem to be get recognized. Moreover, the test example does not even contain sulphur (it is: C6H6O6N6).

I’m therefore a bit puzzled.

Has anyone succeeded in adding new elements (other than C, H, O and N) and got the mol_fra program working?

Thanks in advance,

Best regards

Theo

Theo de Bruin
Ingénieur de recherche
IFP Energies nouvelles - Direction Chimie et physico-chimie appliquées
1 et 4 avenue de Bois-Préau
92852 Rueil-Malmaison Cedex - France
Tél. : 01 47 52 54 38 - Fax : 01 47 52 70 58
Email : [email protected]…3903…
Web : www.ifpenergiesnouvelles.fr

Ray might be able to help with this.

Steve

Hi Theo,

For starters, did you define bond order cutoffs between S and other
elements besides C? More comments below.

Thanks,
Ray

Since my system contains the elements C,H,O,N and S, I needed to add the
latter one both in the mol_fra.c programme and as well in the Cutoff.dic
I hardly have any programming experience but I added the information for S
in the mol_fra.c.

This may be obvious, but make sure where ever you find N or Nitrogen,
just add an additional S or Sulfur as appropriate. Also when ever you
see 3 or 4, you might need to change them to 4 or 5 accordingly.

After having compiled the mol_fra.c, the created executable complains on the
test case that:
Can't fine a certain pair for cut off dictionary!
However, I clearly added the bond order cut-off for sulphur (even in both
ways), i.e, for example:
C S 0.3
and
S C 0.3

For a certain pair, you only need to define its cutoff once. But you
need four more definitions of bond order cutoffs besides the C-S pair.
Which four?

Somehow, it does not seem to be get recognized. Moreover, the test example
does not even contain sulphur (it is: C6H6O6N6).
I'm therefore a bit puzzled.

Even your test sample contains no S, since you have modified the
mol_fra.c to include 5 elements, you need a complete definition of
Cutoff.dic; and it is irrelevant to how many atom types your test
sample has.

Dear Ray,

Many thanks for your quick and moreover useful reply.
I was now able to "repair" the .c

Just two more quick additional questions:
1) does mol_fra only need the reax.bonds file for its analysis?
2) is there a way to get more detailed molecular, i.e. structural information. Since I'm looking at the thermal decomposition of very large molecules (MW > 2500), the molecular formula (CxHyOz...) of the fragments is in my case not sufficient. Is it possible to write out the coordinates of the atoms of the fragments, so that the fragments can be visualized in VMD, Molden or any other program?

Thanks again for your help.
Best regards
Theo

-----Message d'origine-----

Hi Theo,

Just two more quick additional questions:
1) does mol_fra only need the reax.bonds file for its analysis?

Yes, but it should be "bonds.reax".

2) is there a way to get more detailed molecular, i.e. structural information. Since I'm looking at the thermal decomposition of very large molecules (MW > 2500), the molecular formula (CxHyOz...) of the fragments is in my case not sufficient. Is it possible to write out the coordinates of the atoms of the fragments, so that the fragments can be visualized in VMD, Molden or any other program?

I am afraid not, unless you modify fix reax/c/bonds source code.

Hope this helps, and cheers,
Ray

Hi Ray,

Thanks for your reply.

I have no idea how much work it needs (again, I’m not a programmer) to modify fix reax/c/bonds source code, but it would be a great (new) feature, if such additional information could be extracted from the simulations.

Again thanks for your rapid and constructive reply.

Bye, Theo

-----Message d’origine-----