Research



Accessing the inaccessible: Enabling rare event simulations with new methods and software



Our group is interested in efficiently simulating rare events. Rare events are transitions that occur infrequently during an observation time. That is, the waiting time for the event to occur is usually orders of magnitude larger than the timescale of the event itself. An example of this is volcano eruption -- it could take years to observe a volcano erupt but the eruption event itself is a few days long. In molecular simulations, examples of rare events include protein folding, permeation of a polymer through nanotube and nucleation of a new phase from a metastable phase. Using straightforward simulation techniques to study these events results in spending considerable simulation time (and hence computational times) waiting for the event to happen rather than sampling the event itself.
We are interested in developing software and methods that efficiently sample rare events in molecular simulations. To this end, we have developed SAFFIRE -- software to perform large scale forward flux sampling simulations in high performance computing environments. We are also developing forward flux sampling based methods that enable sampling along multiple order parameters simultaneously thereby, increasing the efficiency of sampling. Our group also integrates the power of machine learning with these methods to further improve rare event sampling.

Please email us (ssarupr@g.clemson.edu) if you are interested in collaborating with us to use SAFFIRE. It is in the development stage and will be released to the public soon.




 

Nucleation in aqueous systems: Ice and gas hydrate nucleation




Using the methods developed to study rare events, we study heterogeneous ice nucleation -- i.e. effect of surfaces on nucleation of ice. The key question we would like to answer is -- what physical and chemical properties of a surface facilitate or hinder the formation of ice? This is important for a broad range of applications -- understanding ice formation in clouds, preventing ice on airplanes, designing icephobic coatings, developing freezing based food preservation techniques and cryopreservation of tissues and organs. Through, detailed molecular simulations combined with our rare event methods we are developing molecular parameters that could be used to predict the ice nucleating ability of a surface.

Aqueous mixtures of molecules like methane, carbon dioxide and ethane form beautiful crystalline structures called gas hydrates when pressurized and cooled. Naturally occurring methane gas hydrates are expected to be promising source of energy. Further novel technologies to purify water, store energy, sequester carbon dioxide and separate gases have been based on gas hydrates. However, many molecular details on the formation of gas hydrates remain unknown. We use molecular simulations to answer these questions -- how does the aqueous mixture transform to hydrates, how is this transition affected by the presence of surfaces and how do additives affect such transitions.




 

Enzyme Catalysis: Rational engineering to deployment



Enzymes are the proteins that catalyze reactions. The powerful quality of enzymes are their specificity. This has made them an attractive catalyst to use for biotechnological applications. However, enzymes can work only in certain conditions and often these conditions are different from those in the industrial applications. Thus, there is a need to engineer enzymes and enzyme environments such that the enzymes are functional in the desired operating conditions. To enable such technologies, we use molecular simulations to understand the structure-function-dynamics relation in enzymes. We study the effect of mutations as well as immobilization of enzymes on this relationship.



Multiscale modeling of membranes for water purification



Clean water is a basic necessity of life. With increasing population the demand to produce clean water efficiently has become imminent. Membrane based reverse osmosis methods and membrane based water filtration methods are commonly used in water purification. One of the biggest challenges in these techniques is fouling -- the deposition of unwanted materials from the water being treated on the purification membrane. In our research, we are collaborating with Drs. Ladner, Husson and Battitio to develop multiscale models that predict the fouling behavior of a membrane. Using these models we hope to discover novel membrane chemistries and topologies that are fouling resistant.