Abstract | Keywords | Export | Availability | Bookmark

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Simulations of atoms and molecules have a very wide range of applications. They’re used to study complicated biological processes and to design fantastic new materials such as self-healing plastics. Atoms, however, are very small and move rapidly. As a result, atomistic simulations require a lot of time. Say for example, that a scientist would like to simulate a human cell down to every last atom, for one second. Even a supercomputer would require millions of years to perform all the necessary calculations. For this reason, every possible technique for accelerating atomistic simulations is very valuable. It comes as no surprise then, that many techniques already exist. They are mostly focused on dealing with simple systems however, that are made of a single material that is nearly the same everywhere. Such systems are called homogeneous and their counterparts are heterogeneous systems. A system may be made heterogeneous by introducing a solid interface, that attracts a liquid material. This will cause the material to stick together and achieve a higher density near the interface. Therefore, the material is no longer homogeneous, since it is no longer the same everywhere. The work presented here serves as a basis to generalize well-known techniques for acceleration, such as coarse-graining, so that they may be used for heterogeneous systems. The basic idea of coarse-graining is to represent small groups of atoms by beads, letting them interact in an artificial way so that they effectively behave in the same way as the original groups of atoms they represent. This artificial interaction depends on the density of the material, so that it cannot simply be applied to a heterogeneous system. One technique for simplification that was shown to work reliably involves representing the interface by an external potential. The simplified explanation of this is that in stead of calculating every force from every interaction with every particle in the interface, only a single average force is calculated for each liquid particle. A simple technique was set up to calculate such a potential, namely by simulating an ideal gas next to the interface and deriving the potential from the resulting density profile. In this context, an ideal gas refers to particles that interact with the interface, but not with each other. The new technique was subsequently applied to more complex, realistic, heterogeneous systems in order to check if the results remain consistent. This was indeed the case. After successfully simplifying the interface, the next step was to simplify the simulated molecules themselves. A technique for coarse-graining homogeneous materials was applied to the same heterogeneous systems, with the goal of determining what kind of problems occur, and possibly finding solutions for them. The thesis was concluded with a perspective on a new adapted technique that may make coarse-graining for heterogeneous systems possible in the future.