Reports: DNI651837-DNI6: Theoretical Considerations of Conjugated Polymer Self-Assembly

Jim Pfaendtner, University of Washington

This project report contains the activities from our first year of funding.  Progress was limited due to the lack of a PhD student working on the project until their classes were over in February or March.  Our major activities this year focused on: 

1) Establish a molecular dynamics simulation protocol for polythiophenes and thiophene oligomers in various solvents

2) Perform extensive characterization of the solvent effects on the flexibility and structural properties of polythiophenes in dilute solution. 

The project is ongoing, and a preliminary report detailing the findings is below:

Simulation Details:

Molecular dynamics simulations were conducted using Gromacs 4.5.5. The OPLS all-atom forcefield with modifications for use with conjugated polymer systems was used as outlined in existing literature. Each system was initialized by steepest descent energy minimization followed by 1 ns of NVT simulation. Next, 4 ns of NPT simulation at 300 K allowed the box size to adjust to equilibrate the solvent density. For production runs, each system was simulated for 150 ns with a time step of 2 fs. Each system was run at 300, 450 and 600K, in order to probe the effect of temperature on polymer structure and interactions with solvent, with temperature held constant using the velocity-rescaling thermostat. Pressure-coupling in the initialization steps used the Berendsen barostat.

Results and Discussion: 

In order to verify that the OPLS-AA force field accurately captures the solvent chemistry, 10 ns simulations of 1000 solvent molecules in a cubic box were conducted. From these relatively short trajectories density and heat capacity were calculated and compared to experimental values. The first 5% of data points were skipped to ensure averages were representative of equilibrated solvent, and the calculations were performed from three unique simulations to provide information about the error. The heat capacities predicted by the simulations were high by roughly a factor of two. This, however, is consistent with calculations from literature.

Persistence length is one way to characterize polymer conformation. This length, in essence, characterizes the rigidity of a polymer chain. Segments of a polymer chain shorter than its persistence length will act as a rigid rod. There are a number of different models by which persistence length can be calculated. One model involves fitting the decay in the angles between monomers by the following relation:

<cos theta(s) > = exp (-s/lp) [ 1]

where lp is the persistence length and cos θ(s) is the angle between tangent vectors separated by contour distance s along the polymer backbone. The persistence lengths calculated by [1] decreases at the higher temperatures which is consistent with expectations as the polymers should be more flexible when more energy is present in the system. The possible exception is dichlorobenzene, but with this system the calculated error in the 300 and 450 K cases are too high to truly call this an exception.

The P3HT simulations were performed in duplicate in order to provide additional statistics and allow for the calculation of error. However, further verification is needed in order to determine repeatability of the data since all 24 of the initial simulations (P3HT in 4 solvents at 3 temperatures each with 2 repeats) summarized elsewhere were started with single folded P3HT conformation surrounded by varying solvent configurations. In order to determine if this is valid select systems have been annealed to 800 K and then cooled back to 300 K to obtain new starting structures. The uniqueness of these new structures were verified using an RMSD calculation, and new simulations were set up identically to what has been previously described.


Solvent simulations have verified that the selected MD models accurately reproduce the chemistries of four solvents of interest and can thus be used in our polymer-solvent simulations. Additionally, 150 ns NVT simulations of P3HT have been run. The analysis of the resulting trajectories is in progress. Persistence length calculations have been performed using one of several calculation methods.  The results obtained from these calculations in general meet basic qualitative expectations of representing polymers of higher flexibility, corresponding to a lower persistence length, at higher temperatures.