I see an example about calculating the dielectric constants of Nylon 6(n=20) in the Blog(https://www.materialssquare.com/blog/matsq-tip-lammps-polymer-md-dielectric-const-en). The dielectric is so accurate that it is nearly the same as I got from experimental method. It is amazing work. How can you calculate the dielectric constants of polymer so accurate? What is the key point when calculating the dielectric constants? I see the charge and atom type were obtained from Dreiding Force Field. As we know, the Dreiding FF is a universal FF, which is not as accurate as COMPASS or PCFF. So, I guess the secret is your dielectric constants module. Can you give me some suggestions?
Hello.
First of all, thank you for your acknowledgment. I'm glad to hear that.
'Modeling and thermalization' are the factors important to obtain similar properties to experimental values.
As you know, modeling involves two steps: Generating a single polymer chain in the Molecule builder; Creating an MD model in Structure Builder.
When preparing a single polymer chain, the better, the larger the repeating unit is set. In the example, n=20 is used. You can choose an appropriate value according to your polymer.
Also, it is better if there are many polymer chains in the MD model - that is, the more atoms. It can simulate better the real system.
Then you should optimize it appropriately in the thermalization module. In this time, better results may be obtained if thermalization is performed at the temperature of the amorphous region.
Please let me know if you have further questions.
Have a good day!
Thank you so much for your suggestions. I learn so much from your suggestion. I do have a question. Did the algorithm in the Dielectric constant module is using the fluctuations of the dipole moment as equation: ε=1+(<M*M>-<M><M>)/3ε0*V*k*T? Or you adopt other methods or more advanced methods to calculate the dielectric constants?
Matsq set that equation, ε=1+(<M*M>-<M><M>)/3ε0*V*k*T. Thank you~!
Many thanks for your response. Best wishes.