The research in the Computational Soft Matter Group revolves around using the tools rooted in thermodynamics and statistical physics to understand phase transitions in nature. We apply these techniques to a wide range of systems, including ice formation in the atmosphere, hydrate nucleation, protein aggregation and colloidal crystallization. Here are some active projects that we are pursuing at the moment.

Desalination and Membrane Technology

Procuring widespread access to a clean, stable water supply is an elusive, yet paramount goal which will continue to play an increasingly important role in global development, health, and well-being. Water scarcity already affects billions of people. Ongoing trends such as a rising population and climate change will further exacerbate this issue, making the development of water-treatment infrastructure one of the key areas of focus for the coming decades. Alongside strategies to properly manage water usage, new technologies can allow for better water reuse and reclamation, as well as the ability to tap into previously untreatable sources such as brackish water and seawater.

Among these new technologies, reverse osmosis desalination has emerged as a key technique for making use of the world’s abundant supply of salt water. It is already employed in numerous plants across the globe. The central element of these reverse osmosis desalination systems is a semipermeable membrane which readily allows for the passage of water while rejecting the flow of solutes such as dissolved salts. An applied pressure gradient across a saltwater feed adjacent to the membrane will generate a stream of fresh water flowing through the membrane. While reverse osmosis desalination is already being deployed with excellent effects, further advancements in membrane technology can improve this process, making it feasible and cost-effective in more areas.

Our group uses molecular simulations to generate a better fundamental understanding of the mechanisms of solvent and solute passage through semipermeable membranes with the intention of informing the design of the next generation of reverse osmosis desalination membranes. We develop the computational techniques needed to study these systems and use those tools to accurately and efficiently examine a range of potential desalination membranes including nanoporous graphene, polyamides, and MOFs. Our main focus is to understand how a membrane’s molecular structure impacts its selectivity for different types of solutes.These simulations are advancing the broader understanding of how semipermeable membranes function and how they can be improved.


Crystallization is a first-order phase transition that is ubiquitous in the world around us. Examples include ice formation and its importance in cloud microphysics, crystallization of silicon and other Group IV elements used in electronics and solar cell industries, formation of molecular crystals of pharmaceuticals of organic semiconductors, clathrate hydrate formation in oil pipelines, and polymer crystallization, etc.

Crystallization can be favorable or problematic. For instance, it is usually desirable to form crystals from solutions of pharmaceuticals, while the formation of amyloid fibrils in the brain is linked to Alzheimer’s disease. Understanding crystallization is therefore crucial for developing strategies for its promotion, inhibition and control. Unfortunately, despite its omnipresence, the microscopic details underlying crystal formation are far from fully understood, and elucidating them is an active area of research.

In our lab, we use molecular simulations coupled with advanced sampling techniques like forward-flux sampling to explore the mechanism and rate of crystal nucleation, which is the rate-limiting step for crystallization in a wide-variety of situations with low to moderate thermodynamic driving force. Particularly, we are interested in understanding nucleation in non-classical environments such as in the vicinity of fluid-fluid interfaces (relevant to atmospheric freezing) or on chemically and topographically non-uniform surfaces, etc. We also develop rigorous heuristics to quantify and avoid artifacts in our nucleation simulations such as finite-size effects that arise as a result of simulating small systems.
Our work is aimed at broadening the understanding of both homogeneous and heterogeneous nucleation, and developing robust heuristics and tools that ensure efficient, fast and correct simulations.

Protein Aggregation

The second law of thermodynamics implies that the universe is constantly shifting to being less ordered and organized by inevitably increasing its entropy. However, life as we know it depends on a high degree of organization ranging from subcellular structures and biomolecular complexes to tissues and organs. The physical origin of such organization is, nevertheless, not fully understood. Although it is known that cells cannot maintain their integrity without consuming energy, there is a growing body of evidence suggesting that certain assembly processes can be thermodynamically driven and occur spontaneously due to changes in thermodynamic variables such as intermolecular interactions and constituent concentrations.

Some of these thermodynamically driven processes such as formation of membraneles organelles via liquid-liquid phase separation, reorganization of proteins and lipids in biological membranes, and tissue sorting are physiologically vital whereas others such as protein aggregation and denaturation, are pathological. For instance, Alzheimer’s disease (AD), Parkinson’s disease (PD), Amyotrophic Lateral Sclerosis (ALS), and cataract all occur as a result of the abnormal emergence of pathological assemblies in cells and tissues.

In our group we are interested in employing and developing computational techniques rooted in molecular thermodynamics and statistical physics to elucidate the aggregation kinetics of such biomolecular systems. We are particularly interested in the aggregation dynamics of γ-crystallin, a protein in the human lens that is associated with age-onset cataract. Revealing the range of timescales of such rare biological events can provide invaluable information for early disease detection and possible therapeutic interventions.