Paper title: Atomistic simulation of water and salt transport in the reverse osmosis membrane FT-30
Journal: JMS 139 (1998) 1-16
Author/s: M.J. Kotelyanskii, N.J. Wagner, M.E. Paulaitis*
Center for Molecular and Engineering Thermodynamics, Department of Chemical Engineering, University of Delaware, Newark, DE 19716, USA
Summary:
The paper investigates the transport of water and NaCl across the polyamide (PA) layer of the FT-30 reverse osmosis (RO) membrane by using atomistic computer simulations.
Molecular dynamics (MD) simulations were performed using NVT and NPT ensemble. The force fields used were parameterized for the groups of organic molecules contained in the PA structure: aromatic, amide and carboxyl groups. Six systems of varying salt concentration and polymer content (wt% and linear/cross-linked) were studied. The buildup of ions, water and membrane models, and the procedure for conducting the simulations were detailed in the paper.
As observed from the simulations, self-diffusion in pure water proceeds by atomistic motion and allows water to sample the entire volume of the simulation box. On the other hand, self-diffusion in hydrated PA exhibits transport through localized sites, wherein water moves by “jumping” on these sites. This is similar to the movement of small dissolved gas molecules in amorphous polymer glasses.
Lower water mobility was obtained for higher polymer densities. When density increases because of matrix cross-linking, water mobility decreases. For the mobility of the salt ions, it was found that chloride has a lower mobility compared to sodium in the hydrated polymer because of the presence of more polar groups on the polymer chain during the solvation of the anion. Furthermore, it was confirmed that anion transport limits salt transport in the simulated FT-30 model.
The effects of salt concentration were also investigated. This work confirmed that water mobility decreases as the concentration of NaCl increases. However, the authors disagreed with the reasons for this phenomenon presented in previous studies. In their simulations, no change in volume was observed, thereby opposing the view that water mobility decreases because salt concentration causes contraction of the polymer matrix.
By estimating the salt partition coefficient as well as the diffusion coefficient, it was concluded that the high salt rejection of the FT-30 RO membrane is based on the significantly large difference of water and ion mobilities within the PA discriminating layer.
Contribution and application:
This paper gave a fairly detailed description of setting up simulations using water, NaCl and PA. This will be helpful when ESEL would venture into MD simulations of desalination and reverse osmosis processes, specifically when investigating water and ion transport.
By: Hannah Ebro
hannah@gist.ac.kr