Session P12: Bridging Time and Length Scales in Polymers and Soft Materials: Computational Pathways to Accelerate the Lab to Fab Transition - Industry Day

Spanning From Atoms to Micrometers in Simulations of Contact, Adhesion and Friction

Improved understanding of the forces between realistic solid surfaces is needed to optimize adhesion and friction. Modeling these forces is challenging because they arise from interactions between atoms separated by less than a nanometer, but the number and spatial distribution of these contacting atoms depends on surface roughness and deformation on micrometer and larger scales. There are also strong scale effects in the role of elastic deformations along the surface. The talk will first describe a seamless Greens function (GF) method that allows a full treatment of elastic deformations and atomic contact for micrometer scale surfaces and multibody potentials. Next applications of the method to calculations of the contact area, contact stiffness, adhesion and friction for a range of geometries and interactions will be described. The results can be captured with simple analytic expressions and explain why most contacting surfaces do not adhere. Theoretical and experimental studies of single nanometer-scale asperities show that the frictional shear stress depends strongly on whether surfaces are commensurate. A large constant stress is obtained for identical, aligned crystalline surfaces, but the stress averages to zero in the more common case of incommensurate surfaces. The resulting ultralow friction is called superlubricity and is found in experiments and simulations of small contacts. Our simulations reveal dramatic changes in this behavior because different parts of the surface are able to advance independently as the contact radius increases towards micrometer scales. The friction between identical surfaces drops with increasing radius and then saturates at a low value. The force between incommensurate surfaces saturates at a similar value that can be related to the Peierls stress for dislocation motion at the interface. Studies of multiasperity contacts also show that incoherent motion along the interface can lead to pronounced changes in the macroscopic friction.   

Accelerating materials discovery through the development of polymer databases

Efficient materials discovery can be greatly aided by access to databases that tabulate material property measurements and that allow for the exploration of material-property relationships. Such databases are less prevalent for polymers than other materials such as metals, in part due to the variety of structures associated with a single polymer identifier. For example, polyethylene could be branched or linear; it could also have a narrow or broad molecular weight distribution. I will discuss initial efforts towards generating a polymer property database in collaboration with Prof. Juan de Pablo and colleagues at the University of Chicago. Specifically, we focused on tabulating the Flory-Huggins chi parameter, describing the miscibility of polymer blends, using a course-based approach coupled with specialty software. In the context of a class setting, the undergraduate students learned about the field of polymer physics and used the software to identify chi parameters and related quantities, such as the method of measurement, from previously identified articles from literature. Both successes and challenges of this approach are measured through metrics of the resulting database and feedback from the students.   

Complex Suspension Rheology Using High Performance Computing

In processing advanced ceramic materials, the properties of the final product depend on the process conditions and the interactions between the materials at the scale of the individual particles. Along with general bulk properties, more subtle properties including particle orientation, segregation, and pore structure must be established during processing to achieve the desired functionality. Accomplishing this requires a thorough understanding of the mesoscale interactions and how they influence the macroscale behavior. We conduct a series of large scale simulations of highly filled polymer-nanoparticle composites as analogs of ceramic pastes and reveal how the ceramic particle and binder properties determine the structure and rheology of the bulk material. As with real ceramic pastes, particle shape and size distribution along with composition determine the shear modulus, extent of segregation, and degree of particle alignment. These factors are influenced by the binder through the rheology of the binder phase and the interaction between binder and particles. This talk presents the results of this study of polymer-nanoparticle composites along with a brief overview of research and development at Corning showing the similarities and differences between research in industry and academia.   

Conduction and Narrow Escape in Dense, Disordered, Particulate-based Heterogeneous Materials

For optimal and reliable performance, many technological devices rely on complex, disordered heterogeneous or composite materials and their associated manufacturing processes. Examples include many powder and particulate-based materials found in phyrotechnic devices for car airbags, electrodes in energy storage devices, and various advanced composite materials. Due to their technological importance and complex structure, these materials have been the subject of much research in a number of fields. Moreover, the advent of new manufacturing techniques based on powder bed and particulate process routes, the potential of functional nano-structured materials, and the additional recognition of persistent shortcomings in predicting reliable performance of high consequence applications; leading to ballooning costs of fielding and maintaining advanced technologies, should motivate renewed efforts in understanding, predicting and controlling these materials' fabrication and behavior. Our particular effort seeks to understand the link between the top-down control presented in specific non-equilibrium processes routes (i.e., manufacturing processes) and the variability and uncertainty of the end product performance. Our ultimate aim is to quantify the variability inherent in these constrained dynamical or random processes and to use it to optimize and predict resulting material properties/performance and to inform component design with precise margins. In fact, this raises a set of deep and broad-ranging issues that have been recognized and as touching the core of a major research challenge at Sandia National Laboratories. In this talk, we will give an overview of recent efforts to address aspects of this vision. In particular the case of conductive properties of packed particulate materials will be highlighted. Combining a number of existing approaches we will discuss new insights and potential directions for further development toward the stated goal. Sandia National Laboratories is a multiprogram laboratory managed and operated by Sandia Corporation, a Lockheed-Martin Company, for the U. S. Department of Energy's National Nuclear Security Administration under Contract No. DE-AC04-94AL85000.   

Coarse-graining to the meso and continuum scales with molecular-dynamics-like models

Many engineering-scale problems that industry or the national labs try to address with particle-based simulations occur at length and time scales well beyond the most optimistic hopes of traditional coarse-graining methods for molecular dynamics (MD), which typically start at the atomic scale and build upward. However classical MD can be viewed as an engine for simulating particles at literally any length or time scale, depending on the models used for individual particles and their interactions. To illustrate I'll highlight several coarse-grained (CG) materials models, some of which are likely familiar to molecular-scale modelers, but others probably not. These include models for water droplet freezing on surfaces, dissipative particle dynamics (DPD) models of explosives where particles have internal state, CG models of nano or colloidal particles in solution, models for aspherical particles, Peridynamics models for fracture, and models of granular materials at the scale of industrial processing. All of these can be implemented as MD-style models for either soft or hard materials; in fact they are all part of our LAMMPS MD package, added either by our group or contributed by collaborators. Unlike most all-atom MD simulations, CG simulations at these scales often involve highly non-uniform particle densities. So I'll also discuss a load-balancing method we've implemented for these kinds of models, which can improve parallel efficiencies. From the physics point-of-view, these models may be viewed as non-traditional or ad hoc. But because they are MD-style simulations, there's an opportunity for physicists to add statistical mechanics rigor to individual models. Or, in keeping with a theme of this session, to devise methods that more accurately bridge models from one scale to the next.