(340bk) Multiscale Modeling for Design of Responsive Soft Material Interfaces

During my PhD, I have incorporated Density Functional Theory (DFT) calculations, atomistic molecular dynamics (MD) simulations, coarse-grained MD simulations, and continuum-scale elastic models to understand how the behavior at the nanoscale can influence material design. I have applied these modelling frameworks in two areas of research: functionalized nanoparticle (NP) interactions with lipid bilayers and molecular design of chemoresponsive liquid crystal sensors.

In the first area of research, I studied how the nanoparticle ligand chemistry influences the interactions with model lipid bilayers. Using coarse-grained MD simulations, I first found that the binding of simple cationic NPs to lipid bilayers was highly dependent on bilayer composition. Near lipid phase boundaries, a feature commonly observed in biological contexts, I found that NPs showed enhanced adsorption. The molecular scale information from these simulations showed that this enhanced adsorption to bilayers was due to an optimization of the favorable contacts between the NP and the bilayer and the curvature imposed by adsorption, where the latter is minimized due to the curved nature of the phase boundary. Additionally, I found that multiple cationic NPs tend to self-assemble into linear aggregates when adsorbed to a bilayer, where no aggregation is shown in solution. I determined that minimization of bilayer curvature deviations drove NPs together upon adsorption, while the coulombic repulsion between NPs led to a linear aggregation state. Finally, I investigated more complex NP designs, which include a hydrophobic core, PEG oligomers, and a cationic end group with varying alkyl chain lengths. As opposed to previous measurements, which were performed as a function of the z-distance between the NP and bilayer, I implemented advanced free energy calculations to measure the adsorption free energy as a function of both the z-distance between the NP and the bilayer and the number of contacts between the NP and the bilayer. From these calculations, I was able to determine that longer alkyl end groups resulted in the lowering of the adsorption barrier by allowing for intercalation of the alkyl end group into the bilayer as a means of initiating adsorption. This information allows for better design NP end groups to promote adsorption.

In the second area of research, along with experimental and DFT collaborators, I studied how the molecular structure of liquid crystals and analytes influence response dynamics. Using DFT calculations I parameterized an atomistic model to reproduce experimental properties, such as the phase behavior and diffusion coefficients. Using this model, I had predicted the diffusion coefficient and partition coefficient of analytes in bulk liquid crystal and near interfaces. I also developed a mass-transport model to connect these coefficients to the sensor activation time and found good experimental agreement with systems our collaborators had tested. I also found that the molecular structure of the binding mesogens, which are the species directly displaced by the analyte, can influence the anchoring energy of a given system, and I predicted the transition coverage of the sensor and thereby the activation time of the sensor. By further using these anchoring energies as an input into a continuum-scale model, we were able to make connections between the response optical patterns and molecular design parameters, allowing us to design sensors with better selectivity.

In the future, I want to apply my skills in DFT calculations, atomistic and coarse-grained MD simulations, free energy calculations, continuum-scale modeling, and extensive experience on collaborative projects to solve many of the problems faced in modern drug design/discovery and biosensing applications. In general, these skills are valuable for research endeavors where nanoscale interactions are driving the leading-order effects of the design process. In areas such as drug design/discovery and biosensing applications, molecular details are critical for the design process and in many cases, the information required to understand how to better design new materials is difficult to determine experimentally. Using some of these tools, I am particularly interested in screening large set of drug candidates, using combinations of artificial neural networks and physics-based models to determine optimal drug candidates, and understanding complex mechanisms of ligand-protein interactions to aid in the design process.