Date of Award
Spring 2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Chemical and Environmental Engineering (ENAS)
First Advisor
Haji-Akbari, Amir
Abstract
Solute and solvent transport through porous membranes is critical to important processes in biology and chemical separation. This is readily apparent in reverse osmosis desalination where nanoporous membranes with high permeability to water and excellent rejection of salt ions are utilized to procure clean, safe water from saline sources. Procuring sustainable worldwide access to an abundant supply of properly treated water is an immediate global concern for human health, economic development, and environmental preservation. Reverse osmosis desalination can play a central role in this pursuit by allowing access to the ample reserves of water in oceans and inland brackish sources, but significant improvements are necessary to enable more widespread adoption of this promising technology. Many of these improvements are centered around the development of next-generation membranes which leverage a fundamental understanding of the mechanisms of ion and water transport to inform effective design. Experiments to study nanoporous membranes are challenging due to the small spatial and temporal scales involved as well as limitations in nanofabrication techniques. Therefore, computer simulations can serve as a powerful tool to be used in conjunction with experiments to probe the detailed behavior of nanoporous membranes and to explore novel designs. This thesis investigates the ability of computational methods including molecular dynamics and forward-flux sampling to accurately characterize the performance and mechanisms of action of desalination membranes in pressure-driven operation, and these tools are used to inform membrane design. Due to modern limitations in computing power, molecular simulations are generally confined to examining systems with ∼104 − 105 atoms in a nanoscale simulation box with periodic boundary conditions to avoid artificial interfacial effects. However, it is often desirable to understand the behavior of the relevant system in the thermodynamic limit. There can exist significant discrepancies between observed properties in systems of different sizes which are known as finite size effects. It is shown herein that simulations of ion transport through porous membranes are subject to extreme finite size effects which, for commonly-used system sizes, introduce orders of magnitude error in estimated rates of passage. These finite size effects can be understood by treating the saltwater feed compartment as an ideal conductor. The presence of a passing ion in the membrane or filtrate region pulls induced charge onto the conducting surface which generates a restraining pullback force on the leading ion. The periodic replicates of this induced charge, which are unphysical artifacts of periodic boundary conditions, also generate a restraining force. Classical electrostatic theory can be used to predict the magnitude of this excess restraining force and to correct the observed free energy profile so that it matches the behavior of the system in the thermodynamic limit. The so-called ideal conductor model demonstrates remarkably consistent predictions of ion passage times in the thermodynamic limit for simulations of finite systems across a range of sizes. While the ideal conductor model gives good estimates of ion passage times in the thermodynamic limit, the corrected free energy profiles exhibit some discrepancies, particularly in the second half of the membrane near the pore exit. To understand this phenomenon, we consider that the membrane and filtrate are dielectric regions, and just as charge is induced at the surface of the conducting feed, charge is also induced in dielectric regions. Incorporation of the effects of dielectric boundaries leads to the ideal conductor/dielectric model (ICDM) which offers substantial improvements over the original model and excellent agreement of corrected free energy profiles at all locations of the leading ion. The ICDM model is also extended the case of subsequent ion passage in the presence of other ions which have already entered the filtrate. This extension is critical to understanding co-ion transport and overall rates of salt transport through desalination membranes. Finally, molecular dynamics simulations with forward-flux sampling and the ICDM model are used to systematically explore the scaling of ion and water flux through multi-pore systems that have charged interiors. Direct interactions between dipoles at the pore interior and passing entities near or within nearby pores results in sub-linear scaling of the flux of counter-ion and water passage with the number of pores and super-linear scaling of the flux of co-ions. A simple, physics-based model is employed to assess the direction and magnitude of these effects to inform rational membrane design.
Recommended Citation
Shoemaker, Brian Andrew, "Computational Investigations of Solute and Solvent Transport in Nanoporous Desalination Membranes" (2023). Yale Graduate School of Arts and Sciences Dissertations. 1424.
https://elischolar.library.yale.edu/gsas_dissertations/1424
