Abstract:
Viscosity-enhancing polymers such as Polyacrylamide (PAM) are gaining traction as candidates for Enhanced Oil Recovery (EOR). This popularity encourages the computational study of PAM to understand its behavior under various solution conditions. On account of the involvement of large length and time scales in the description of its properties, coarse-graining the atomistic models has become a major subject of interest.
This thesis is focused on developing mesoscale models of PAM in an aqueous environment,from an atomistic base. We have followed three approaches towards achieving a coarse-grained model: MARTINI, Iterative Boltzmann Inversion (IBI), and a hybrid scheme of integrating IBI and MARTINI. The objective was to evaluate the reliability of these approaches in representing the structure and thermodynamics of the target system.
We have reproduced the global structural properties (radius of gyration, RG, and
end-to-end distance, Ree) to a reasonable extent with the CG system developed within
the MARTINI framework, although the local structure could not be precisely captured.
The results with IBI and IBI+MARTINI of bonded distributions, and radial distribution functions (RDFs) show an absolute replication of the local structure of a single chain of PAM in water. The viscosity results show that the all-IBI method fails to mimic the
dynamical property, whereas IBI+MARTINI was successful in mimicking the trend exhibited by the atomistic system. We also show that the derived potentials fail to reproduce the structure beyond 4 wt% of concentration of solution, entailing the need for re-parameterization of potentials for higher concentrations.
