6
$\begingroup$

I am running Monte Carlo simulations of polymers with nearest neighbor interactions only. For benchmarking purposes, I decided to run some excluded volume simulations (no overlaps) allowed on a 3D simple cubic lattice (coordination number of 6).

These are the moves I implemented: enter image description here

The problem I am seeing is that the end-to-end autocorrelation function for my single polymer is not dying down to zero. The reason I call this a problem because I am getting incorrect Flory Exponents: $$\langle R_g^2\rangle \propto N^{2\nu}, \tag{1}$$ where I am getting $\nu = 0.8$, which is too high.

I am running 10^7 steps with Rosenbluth sampling.

E.g. If there are N possible kink jump spots, I will test each spot until there is a vacancy available. Same applies for the end rotations.

What could be the cause of this? Is it simply that my moves are too local? How do I decorrelate my system faster, as N increases?

Improve this question
$\endgroup$
4
  • $\begingroup$ Just to clarify, how close is the autocorrelation function getting to zero? What (roughly) are you expecting the Flory exponents to be? $\endgroup$
    – Tyberius ♦
    Commented Feb 1, 2022 at 17:56
  • $\begingroup$ From renormalization group theory, I expect $\nu$ to be 0.33 in bad solvent, and 0.6 in a solvent with only excluded volume interactions -- and higher if solvent quality increases further. But when I run simulations at a T=0.5, my scaling exponent hits 0.37 for bad solvent. $\endgroup$
    – megamence
    Commented Feb 1, 2022 at 20:06
  • $\begingroup$ As for the exact value of the auto-correlation function, it either stays very non-zero (0.2-0.4), or oscillates more wildly than expected $\endgroup$
    – megamence
    Commented Feb 1, 2022 at 20:09
  • $\begingroup$ Did you figure this out now? Please update us! $\endgroup$
    – Nike Dattani - No Free Time
    Commented Sep 6, 2022 at 20:03

1 Answer 1

Reset to default
1
$\begingroup$

The problem with this moveset is that it is too local. The best way to decorrelate polymer configurations is to run simulations with more aggressive moves, such as chain regrowth.

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .