The folding of proteins and spontaneous assembly of intracellular components provides an existence proof of a viable nanotechnology. Making nanotechnology practical, however, is a key research challenge. Self-assembly is clearly an essential tool to realize practical and economical nanoscale structures. One of the primary challenges in modeling self-assembly is the multiple length and time scales involved, from those associated with electronic degrees of freedom and molecular motions, to aggregate formation and shape fluctuations, to the macroscopic continuum/hydrodynamic response (Figure 1).1 My research uses the tools of statistical mechanics and large-scale computer simulations to study surfactant organization, polymer and biopolymer assembly, and structure-function relationships with the overall objective of advancing self-assembly as the ideal tool for building tailored nanostructured materials. This research is inherently multiscale and utilizes a range of techniques from molecular simulations, to mesoscale simulations, to mean-field/continuum theoretical descriptions in a bottom-up line of attack. My work benefits from the interdisciplinary collaborations with theoreticians and experimentalists I have established during my research career to study topics including: The development of multiscale polymer simulation methods that bridge molecular scale correlations to mesoscopic structure with Prof. Shekhar Garde (Rensselaer Polytechnic Institute) and Prof. Sanat Kumar (RPI); the implementation of novel free energy simulation techniques for examining cooperative biomolecular conformational changes and aqueous solution behavior with Dr. Lawrence Pratt (Los Alamos National Laboratory), Dr. Christopher Jarzynski (LANL), and Prof. Tom Woolf (Johns Hopkins University); and the examination of the interactions of polyelectrolytes with surfactants as nanostructured materials and the interaction of polymer assemblies with paraffins to enhance the recovery of waxy crude oils with Prof. Björn Lindman (Lund University - Sweden), Prof. Robert Prud'homme (Princeton University), and Dr. Lewis Fetters (Cornell University). Specific research projects to establish a graduate research program are described below.
Water: A stage for self-assembly. The aqueous milieu in which supramolecular assembly takes place invariably plays an intimate role in organization and function - water is an active and essential participant. Despite its ubiquity, water and water mediated forces are among the most complex and poorly understood. Surfactant aggregation and protein folding are governed by a delicate balance of hydrophobic and hydrophilic interactions. Phenomenological descriptions of these forces often defer to the persistence of tetrahedral ice-like structures in liquid water. Experimental conformation of this picture, however, is tenuous due to the difficulty of probing these fleeting aqueous structures and the subjective nature of what constitutes "ice-like." We are interested in the boundary between molecular representations of water and macroscopic mean-field/continuum descriptions of hydration, for which molecular details are averaged to simplify and, in most cases, permit computations at the length scales at which assembly occurs. This research provides a theoretically sound framework for mesoscale simulations and tests molecularly founded approximations for predicting supramolecular structure and phase behavior. Challenges to be addressed include:
Tailoring of thermosensitive polymers. Elastin, an unstructured polypentapeptide, undergoes a counterintuitive collapse with increasing temperature (at ~25°C) from a coil to globule in water over a narrow temperature range. This transition has found application for molecular muscles converting thermal energy into mechanical work.2 Similarly, entangled aqueous solutions of synthetic polymers containing N-isopropyl acrylamide (NIPA) and chemically related monomers collapse with increasing temperature with a concomitant decrease in viscosity (Figure 2). This "viscosity switch" has been exploited in developing sieving matrices for DNA sequencing that can be rapidly injected into microfluidic devices.3 The collapse transition temperature and efficacy of a sieving matrix at separating long DNA fragments has been shown to depend on the polymer's hydrophobic and hydrophilic composition. Indeed the inverse temperature collapse transition parallels the solubility minimum of hydrophobic species in water and cold denaturation of proteins. Our research on thermoresponsive polymers will examine the roles of water and polymer sequence on the collapse. Poly(NIPA) will serve as a starting polymer for simulation investigations due to its relatively simple and homogeneous composition. The collapse transition will be studied using state-of-the-art techniques, such as parallel replica exchange, which permit the rapid enhanced of polymer conformations and thermodynamics over a wide range of temperatures. In addition to characterizing poly(NIPA), hydrophilic units, such as acrylamide, will be systematically incorporated into the polymer backbone to gain insights into the interplay between hydrophobic/hydrophilic monomers ratio on the collapse transition. The goal of this research is to develop molecularly based rules for the specification of thermosensitive aqueous polymers and speed up the trial and error process currently used in screening new materials.
Length scales of hydrophobic hydration and colloidal forces. The adage "oil and water do not mix" dominates thinking about hydrophobic effects and is upheld as the primary impetus for a variety of biological assembly processes. Enigmatic temperature signatures - such as the fact that many soluble proteins unfold upon heating and cooling - offer primary puzzles of hydrophobic effects. Moreover, it has been recognized that molecular hydrophobic effects, characterized by hydrocarbon-to-water transfer solubilities, and the driving forces for self-assembly on larger length scales (Figure 3) differ distinctly in their magnitude and temperature dependence, indicative of mechanistic changes in hydration with increasing solute dimension. This distinction is even more dramatic for measurements of long-range attractive forces between colloidal surface which are inexplicable on the basis of molecular hydrophobic effects.4 Recent theoretical treatments, which emphasize the phase behavior of water confined between surfaces, predict hydrophobic substrates can nucleate a vapor-liquid phase transition that overwhelms colloidal stability.5 The predicted interaction, however, is considerably longer ranged than observed experimentally. Moreover, experiments on the effects of salt on the range of surface forces and degassing on the stability of surfactant free emulsions challenge theoretical predictions.6,7 These shortcomings suggest a number of important questions to be addressed theoretically and with simulations. Preliminary calculations indicate that ubiquitous attractive dispersion interactions can affect rewetting of hydrophobic surfaces, suggesting long-ranged hydrophobic interactions are suppressed and depend intimately on the surface chemistry.8 Quantifying saline solution structure, energetics, and fluctuations near molecularly realistic surfaces as well as the effect of adding trace dissolved gases, e.g., nitrogen, will yield important clues for interpreting the stability of colloidal suspensions. The simulation challenges include the development of novel sampling methodologies to analyze large-scale density fluctuations in confined geometries, which can dominate surface interactions. Our calculations, in turn, will guide the development of molecularly founded mean-field theories which capture the unique equation-of-state properties of water, linked to hydrophobic interactions on molecular and macroscopic lengths scales, to extend our results to colloidal stability and interactions beyond the range obtainable in simulations.
Building bridges from molecular to mesoscopic length and time scales for self-assembly. The interplay between multiple length scales in self-assembly processes, from an individual surfactant or polymer/peptide monomer to the supramolecular aggregate, underlines the necessity of a multiscale approach. Mesoscale models are juxtaposed between the atomic scale, where the detailed molecular interactions play an explicit role, and the continuum scale, where only the averaged response is of consequence. Coarse-grained interactions are typically averaged over multiple atoms to form a collective super-atom and enhance computational performance. The loss of microscopic features is believed to make mesoscale models best suited for examining general trends of assembly and phase behavior associated with changes in properties such as molecular topology and solvent quality rather than the detailed chemistry. It may be anticipated, however, that mesoscale interactions are intimately related to the underlying chemical identity. Our research efforts then focus not only on general self-assembly trends, but also on building seamless relationships between molecular and mesoscale interactions to achieve a genuine mapping between scales. Challenges to be addressed include:
Mapping molecular correlations to mesoscale interaction potentials. Polyolefins have a wide range of physical properties and are typically blended to produce inexpensive materials with tailored properties. Unfortunately, many of these polymers do not mix and the desired properties cannot be readily achieved. It is therefore critical that a good a-priori knowledge of the miscibility of polyolefin mixtures be available. There are no predictive theories in this area, however, and the computational expense for performing explicit molecular simulations is prohibitive to be useful from an experimental standpoint (>1000 monomer units in length). While mesoscale simulations of polymers are routine, a significant difficulty in comparison with experiment is relating the coarse grained potentials to the subtle microstructural differences between the polyolefins. We have recently developed a technique, referred to as Empirical Potential Interaction Coarsening (EPIC), that maps polymers of a defined molecular structure and identity to the appropriate coarse-grained interaction (Figure 4).9 This methodology successfully extends polyethylene oligomer (<100 monomer units in length) liquid simulations to truly polymeric melts with molecular weights greater than 100,000 g/mole (8000 monomer units in length and longer). Ongoing efforts are extending EPIC to a variety of polyolefins, such as atactic and isotactic polypropylene, to benchmark the effects of chain microstructure on the coarse grained interactions. Subsequently we plan to examine the phase behavior of the polyethylene/atactic polypropylene blend, which has been documented as the most immiscible mixture10 and provides the simplest test bed for our methods. To further increase the range of polymer molecular weights obtainable, we will integrate EPIC with the lattice paradigm, developed by Kumar and coworkers,11 which enhances computational performance through algorithmic simplifications. The ultimate goal of this work will be to develop a predictive toolkit for polyolefin miscibility based on polymer architecture and chemistry.
A second issue towards the mesoscale modeling of polymers is the proper treatment of entanglement and stress relaxation. In particular, the diffusion coefficient for polymers less than ~100 monomer units is inversely proportional to the molecular weight, while that for polymers longer than ~100 monomer units is inversely proportional to the molecular weight. This change in scaling is indicative of the growing importance of engtanglement. Depending on the level of coarse graining employed, however, the polymer beads can relax by allowing the beads pass through one another (see for example Figure 4) resulting in unphysical dynamics. While this shortcoming can set practical limits on the choice of bead size, a route for introducing bond uncrossibility constraints has been proposed that permits a more liberal choice of bead size while effectively capturing polymer dynamics on multiple scales.12 A subsequent challenge to developing EPIC for the modeling of polyolefin miscibility then is the incorporation of dynamical constraints into the methodology.
Surfactant and polymer networks as novel soft-materials. Polyelectrolytes can electrostatically associate with oppositely charged micellar complexes well below the critical micelle concentration of the neat surfactant and also form, at higher surfactant concentrations, synergistic structural polymorphs related to those of the surfactant from cubic to hexagonal to lamellar.13,14 Additional interactions can be affected by chemical addition of hydrophobic units along a soluble polymer backbone. These hydrophobically-modified polymers can gel micellar solutions and stabilize nonequilibrium liposomes/vesicles against structural decomposition.15,16 As such, polymer-surfactant complexes have applications in rheological control, as nanocomposite materials, and in drug delivery. While equilibrium molecular modeling of these labile assemblies is beyond present computational abilities, these systems are well suited for mesoscale simulations to understand the general patterns of their phase behavior, organization, and dynamics. The recently developed dissipative particle dynamics (DPD) technique, in particular, has found application in modeling assembly.17,18 DPD can be thought of as an off-lattice extension of the Flory-Huggins model for polymers that permits large time steps to access time scales on the order of milliseconds and is consistent with the hydrodynamic equations of motion, in difference to related coarse-grained techniques. Moreover, the EPIC methodology described above provides a molecular rationale for the interactions utilized in DPD and is fully integrable within this framework. Preliminary calculations for a nonionic surfactant demonstrate that DPD captures the structural characteristics of a typical aqueous surfactant (Figure 5). Our continuing research will examine the gelation of micellar structures using hydrophobically modified polymers. An issue of particular interest is the mechanism by which the percolating network forms - do the individual polymers tie the micelles together through association with multiple micelles or do the polymers associate with a single micelle to form a steric boundary that effective increases packing and forms a jammed system?15 The fidelity of DPD to the hydrodynamic equations of motion will also allow us to monitor network formation through readily observable experimental variables like the viscosity and correlate these variables with the network structure. In addition to gelation, we plan to examine the interaction between surfactant lamellar sheets moderated by hydrophobically modified polymers. Of particular interest will be to determine how the polymers stabilize lamella against rupture after quenching into nonequilibrium conditions either by a simulated temperature jump or dilution by solvent addition.
B.S., North Carolina State University, 1992
Ph.D., University of Delaware, 1998
Hank Ashbaugh was born in Chicago, IL in 1968 and grew up in Charlotte, NC. After graduating from high school, he attended NC State University and graduated in 1992 with a BS in Chemical Engineering. He subsequently attended graduate school in Chemical Engineering at the University of Delaware, where he worked with Professors Michael Paulaitis and Eric Kaler on the computer modeling of surfactant solution interactions. After defending his PhD in 1998, Hank went on to post-doctoral assignments at Lund University in Sweden (working with Bjorn Lindman) and Princeton University (working with Robert Prud'homme). At those two positions he performed experimental studies of self-assembly in aqueous and non-aqueous environments with applications including the design of nanostructured gels and the enhanced recovery of oil from deep sea reservoirs. Hank subsequently joined Los Alamos National Laboratory in 2001 as a Director's Fellow in the Theoretical Division, focusing once again on modeling the fundamental interactions that drive self-assembly. In July of 2004, Hank joined the Department of Chemical Engineering at Tulane University as an Assistant Professor. His current research interests include the multiscale simulation and theory of self-assembly and hierarchical organization in complex fluids including surfactant solutions, polymer melts and solutions, and biopolymer gels and networks to advance self-assembly as a labile tool for building tailored nanostructured materials.
300 Lindy Boggs Center, Tulane University, New Orleans, LA 70118, T: 504-865-5772, F: 504-865-6744 firstname.lastname@example.org