Because i have performed all my simulations at a time step of 0.50 fs (also varied the time step between 0.25-0.50 fs). After analyzing the thermal decomposition results, I was not able to match the exact chemical structure with my simulation results structure. Whereas the molecular weight can be matched with some slight difference. For example C8H8 - styrene. I have the same compound in my MD results , but the structure is not the same when i compare with the original chemical structure , were the molecular weight is same as the original data. Do i need to include that in my results or i have to skip the compound? What i have seen in all ReaxFF MD simulations is they only discuss about the initial fragments such a water, methane , carbondioxide etc . No one has discussed about the major products. Why is it so ? Need a clarification.
As all potentials in MD simulations are somehow 'parametrized' one can not expect, that this potential will be exact or 'general' for every structure. The problem in this (and also many other cases) is that your ReaxFF is parametrized only fore some hydro-carbon structures at some specific conditions. There can be many reasons why your simulation fails to obtain correct styrene structure, and it is hard to come up here with specific answers, as you did not provided all parameters of your simulation (and even then it could be hard). For example, did you tried to simulate only C8H8? Is then it;s structure correctly predicted? I mean, there are so many factors that can influence your simulation that guessing the reason is almost impossible. How this molecule looks like now?
If used potential in other cases gives good predictions for C8H8, then one can only try to find the reason for your results concerning the impact of conditions and/or environment. So in conclusions - try locate the cause before deciding is your simulation correct, or not.
Not sure if I understand your question(s) correctly, but indeed you would need to think about the ReaxFF parameters used and it's a good check to run it on some known stable compounds and perhaps check a few points on a reaction pathway. Furthermore you could consider a smaller time-step (0.2 fs) and perhaps pre-equilibrate your system with a local minimization and/or a non-reactive MD run at lower T.
You can follow your reaction network with ChemTraYzer so you can see which reactions take place and how compounds are formed during your reactive MD run.
Hope this helps,
Fedor
https://www.scm.com/doc/Tutorials/ReaxFF/Burning_methane.html#step-5-analyze-it-create-a-reaction-network
you need to optimize the ReaxFF parameters for your structure when the structure deviates especially this heavily. In addition, your timestep is pretty large as well.
Thank you all for answering me. And sorry for the delay in reply . Actually Styrene is one among the products which is formed during the thermal decomposition of lignin at a specific temperature using ReaxFF. I just gave styrene as an example , were there are many notable compounds such as guaiacol, catechol etc doesint have the similar chemical structure if its related with the original structure. I have chosen Alder's softwood lignin model for my simulation. The structure is well optimised, my aim is to the study the thermal decomposition of single lignin molecular were it has 435 atoms in it . Performing simulations with different time steps, I can only see the thermal decomposition occurs over 0.30 fs with a simulation time of 1000Ps. Also when i try to compare my simulation data with the experimental results , i can still say some products such as guaiacol, vanallin, 2 methoxy phenol etc are still achievable with a time step over 0.40fs . But in the real case , they dont have the identical structure when compared with original structure.
Dear Dr. Goumas,
I'm have tried with small time step as well with respect to your suggestion. But still i couldint find any decomposition taking place at that time step. The simulation time i used was 1000Ps. Hope you remember me , we both have discussed this already in a seminar held at imperial college london. You taught me how to perform energy minimisation for the lignin model which i had. Also i came to know they is going to be a one day seminar which is going to held at Cambridge university on Feb 20,2017. I would like to meet you in person and discuss this once again with you. Thank you very much for answering me the question.
Dear Dr.Konrad Gruszka,
Thank you very much for answering my question.
Dear Dr.Orkid Coskuner,
Thank you very much for answering my question.
- Badges
- Science topic