I have prepared a rectangular amorphous carbon (a-C) sample using molecular dynamics (LAMMPS) simulation. How can I measure the partial charges in each atom of this large system having ~20,000 carbon atoms? Thanks in advance.
Hi Shahriar,
I am not quite sure I understand the question. Since the entire system is neutral, and each atom is equivalent, wouldn't the partial charges all be 0?
Best,
Csaba
Dear Csaba Daday ,
As the amorphous carbon consists of carbon atoms with 0, 1, 2 and 3 dangling bonds (specially on the surface, different atoms would have different coordination numbers, i.e. dangling bonds), that is why I was hoping that all the atoms in the box would have positive and negative charges. And the summation of all the atomic charges would be 0 to make the system neutral. My understanding might be totally wrong. Would you kindly share more of your thoughts regarding this?
And thank you very much for responding to my question. I really appreciate it.
Oh right. I would do one of the following:
1) some Hartree-Fock calculations on variable chain lengths, but more likely:
2) look at the literature at how people modelled this before. I'd be shocked if no one had simulated a-C before.
I still suspect that if there are no hydrogen atoms, the model will have no partial charges on the carbons, just Lennard-Jones. There is no real reason why a particular atom would have a positive or negative charge, so if the charge is to be constant (as it is in classical MD), then it should be 0.
The partial charges used in molecular mechanics are fixed numbers that you provide. As the partial charge on an atom is not definable within quantum calculations, each force field has its own strategy as to how to define them. CHARMM uses interactions with water molecules. Amber, I believe, integrates the charge density over some specified volume. I'm not sure how LAMPSS develops its partial charges.
If you think that partial charge is somehow related to dangling bonds, that information will not be available from a molecular mechanics simulation. If your amorphous carbon model is not bonded, then all of the interactions will be the non-bonded terms: Lennard-Jones and electrostatic. For the simulation to run, you will have to provide the partial charge information.
If you perform a quantum (dynamics or optimization) simulation, the electron density will distribute itself accordingly but partial charges are not uniquely defined. You can, however, measure distances between carbon atoms. That is an observable that correlates to bond order.
Dear Csaba Daday and Mark A Cunningham thank you for your valuable answers. I will look into all the points you mentioned. Hope I can find a solution soon.
Dear Csaba Daday , I have a short follow up question. Let's say, I have prepared a system consisting of hydrogen atoms randomly distributed in amorphous carbon (~20000 atoms altogether). In that case what would be the procedure or technique I should follow to get the partial charges in each atom if there is any?
Thank you.
In a normal MD simulation, one would have maybe 5-10 different types of carbons, even though there would be thousands of individuals. But these 5-10 types would not be defined randomly, but due to their chemical environment. In this case, almost all carbons would have the same environment, so they would have 0 charge.
If you really want to get 1000 different charges for the 1000 carbon atoms (say),
a) these numbers would be small and based on (mostly) noise
b) the numbers would be wrong! after just a few ps, the atoms would move so their charges would need to be updated.
Hello Csaba Daday , yes now I understood the scenario. Thank you very much for your valuable suggestions.
Imposible to tell without knowing thr force field you used.
Doing HF or DFT in 20000 atoms is crazy.
I would use a reactive force field such as ReaxFF or COMB to compute the partial charges. I believe ReaxFF has parameters for carbon and hydrogen.
ReaxFF is parameterized using mulliken populations.
Best,
David
Thank you all for sharing yours thoughts, now I have a clearer image of the problem I am dealing with. I really appreciate that.
Since you have only Carbon atoms in your system, partial charges will be zero. Only atoms with different electronegativities should results in partial charges.
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics