Gary S. Grest, Sandia National Laboratory, tel 505-844-3261,fax 505-844-9761, email gsgrest@sandia.gov
The role of polymer composites and coatings in industrial applications is rapidly growing due to their strength, lightweight, chemical stability and tailor-ability. One of the barriers to further applications is a better understanding of the interface between polymers and ceramics, polymers and oxides, polymers and metals, as well as between different polymers. That is, what is the nature of the interface/interfacial region and how does it effect mechanical and chemical properties. The interface is the major point for initiation of mechanical and chemical failure - the weak link in the chain. The interaction mechanisms between dissimilar materials, particularly when one of the components is polymeric, is difficult to quantify and hence to control. Given the advances in materials modeling for both hard materials, such as metals, ceramics and oxides, and soft materials, such as polymers and foams, a project to develop a better understanding of the interface between the two will be very valuable. The enabling science would come from building multidisciplinary teams between groups who traditionally work on only one of these two areas. By combining the expertise from different groups, it is now possible to address a number of critical issues, such as what are the chemical and physical processes which control adhesion and how does one relate the failure process to material and external properties. Another important issue is degradation of the interface due to diffusion of penetrant molecules, such as water, to ceramic or metal surfaces. How to mitigate the effect of penetrant molecules and other effects of aging are critical to long term stability of the interface. The ultimate goal of the project is to develop accurate and reliable computational models which can be used to understand and control the the interface between polymers and a wide range of dissimilar materials such as ceramics, semiconductors, metals, fibers, composites, clays as well as other polymers.
SCIENTIFIC ISSUES
Understanding the nature of the interface between dissimilar materials poses a number of scientific challenges. Not only do the specific interactions between the constituents play an important role but also the conformation of the polymer. The chemical and morphological structure of the interfacial region differs from that of the bulk. Other processes which maybe important depending on the specific system include, but are not limited to, selective absorption of polymer matrix components, penetration of the polymer components into the second phase, such as would occur in polymer/fiber composites, diffusion of low molecular weight components from the interface into the polymer matrix, surface induced or surface modified crystallization of the polymer, catalytic effects of the surface on the polymer matrix. Accounting for these effects and reliability determining their relative importance for each of the relevant types of interfaces of interest is an unsettled issue, particularly since the types of polymers of technological and scientific importance are varied, including charged or uncharged, elastomic or glassy, homogeneous or heterogeneous, as well as the surfaces that they are bonded to. Broadly speaking, the type of interface/interfacial region of interest falls into one of two categories, those in which the polymer does not inter mix with the second surface and those for which it does. In the latter, such as for polymer/polymer and polymer/fiber composites, the interface is strongly affected by miscibility on the microscopic scale. Miscibility is a desirable property as it is crucial for obtaining high tensile strength materials. To improve the adhesion between these two surfaces, copolymers or co-surfactants are often added. How they affect the morphology of the polymer matrix and interactions between the constituents is unknown. To improve the bonding between polymeric materials and impenetrable surfaces, adhesion promoters are used to improve the coupling. The details of the microscopic interactions strongly effect whether the material fails at the interface, adhesive failure, or in the polymer matrix leading to cohesive failure. The challenge is to develop interaction potentials between the constituents which are accurate enough to be used in establishing a failure criterion which can be used in continuum, finite elements calculations of stress and strains in the composite material. For most systems of interest, very good interaction potentials exist. The challenge is to extend these studies, often generated using a variety very different approaches from empirical classical force fields to tight binding models, to describe the mixed interactions at the polymer interface. Since one cannot possibly simulate all the systems of interest, density functional and polymer reference isomeric state methods will also be very important to generalize and extend the results. Simulations will be used to validate the theoretical calculations of the interfacial structure and surface free energy, which can then be applied more quickly and easily then repeating the simulations for each new case.
COMPUTATIONAL ISSUES
Because of the long length and time scales associated with polymer melts and cross linked networks, a range of approximate models have been used depending on the particular problem. Almost all polymer simulations have used classical, empirical potentials. In some models the hydrogen is included explicitly while in other models the hydrogen atoms are subsumed into a united atom model with the carbon. In both cases, the basic time step for a molecular dynamics simulations is at most a few femtoseconds, limiting the simulations to chain lengths of a hundred monomers and time scales of a few nanoseconds. For very long chains, coarse grained bead-spring model have been used. One computational challenge is to use information from the more realistic models in the less realistic, but more computationally faster coarse grained models. For all of these models, the computational challenge is to develop faster algorithms so that the range of chain lengths and time scale which are amenable to simulations can be extended. Present parallel algorithms on distributed memory machines work treat very large systems for relatively short times very well. Because of the slow relaxation times for polymers, the opposite type of algorithm is needed: one which works well for moderate size systems (from a few thousand to a few hundred thousand atoms or monomers) for very long times. Efficient algorithms of this type do not presently exist. Efficient parallel molecular dynamics codes have been developed to model most atomic systems. How to combine simulation codes, for say a ceramic or oxide, with those for the polymer matrix is an area which needs to be focused on. The ability to simulate long times is particularly important when studying the response of the composite material or coating to either external forces such as an applied stress or internal forces due to mis-match in thermal expansion coefficients during thermal cycling. This is because the polymer matrix is a viscous/elastic fluid in which the stress is very sensitive to the strain rate. One needs to go beyond the brute force approach in which one follows all the degrees of freedom in both the polymer and ceramic or oxide, for example. What will be needed is an algorithm which can vary the strain rate over several orders of magnitude. Finally, there is the issue of length scales. For many systems, such as polymer clay composites, one needs to both large systems and long times.
For success, a multidisciplinary team in essential to bring together the required expertise in material science, physics, chemistry, computational science, and computer science. Expertise in quantum mechanical calculations will be important for developing new potentials which can used in simulating the interface between dissimilar materials. Expertise in simulating polymers and soft systems needs to be complimented with those more familiar with hard materials. The large scale nature of the simulations, necessitates the need for computational scientist with expertise in parallelization of molecular dynamics codes, particularly to develop new algorithms which can be used to study moderately sized systems for very long times. Close contact with experimentalist, particularly those with expertise in adhesion and surface characterization, especially neutron and x-ray scattering, will essential to validate the theoretical models and simulations. No single laboratory or University has the expertise in all of these topics. Regular communication amongst the members of the group is vital for success. Periodic meeting, extended visits between the principals, shared students and postdocs are critical for success.
The overarching scientific goal of this project is to obtain a more quantitative relationship between the interface processes and mechanical properties and chemical stability of the material. A consequence of achieving this goal would be modeling tool capable of integrating atomic level understanding into reliable constituent relations which can be used in continuum, finite-element calculations for stress/strain. Such capability would provide predictions of technologically relevant mechanical properties in a wide range of applications. The enabling tools would also provide needed guidance in interpreting a host of experimental observations. The development of new computer algorithms more suited to long time simulations will be applicable for a range of other problems which require long time simulations of moderate sized systems, such as the slow relaxation as one approaches the liquid to glass transition. The identification of a few prototypical projects where there is a large and evolving set of experimental data on well characterized model systems will play an important role in the success of the project. One possible project is the study the interface between two immiscible polymers, particularly the effect of added copolymer on the mechanical response and strength of the interface. There is extensive amount of well characterized experimental data on the interface width, stress/strain and failure as a function of added copolymer. A second complimentary project is on the interface between a polymer matrix and a ceramic or oxide. A workshop bringing together those interested in polymers at interfaces would be critical to identify the appropriate system to assure success.
Success would be the development of accurate
models describing the interface/interfacial region of polymer
composites and polymer coatings for a wide range of industrial
applications such as electrical devices and composites and structured
materials. Success would be the development of inter atomic potentials
which can be used to describe the interactions between hard materials
(such as ceramics, oxides and metals) and soft materials (such
as polymers). Success would be new, efficient parallel codes for
modeling long times for moderate sized systems. Success would
be development of robust predictive models for the mechanical
behavior (stress/strain relations, thermal shock, crack propagation,
fraction, and ultimate failure) of complex systems when exposed
to thermal or mechanical stresses.