Surface Dynamics and Hydrophobicity of High-Voltage Insulator Polymers

  • PDF / 354,650 Bytes
  • 3 Pages / 612 x 792 pts (letter) Page_size
  • 64 Downloads / 216 Views

DOWNLOAD

REPORT


Surface Dynamics and Hydrophobicity of High-Voltage Insulator Polymers Oliver Fritz1, Philip T. Shemella2, Teodoro Laino2, and Alessandro Curioni2 1 ABB Switzerland Ltd, Corporate Research, Baden-Dättwil, Switzerland 2 IBMResearch – Zurich, Rüschlikon, Switzerland

ABSTRACT Using large-scale, all-atom molecular dynamics simulations, we show that microscopic mechanisms found for molecules on the material surface of siloxane polymers can explain an important surface-hydrophobicity restoration process. In particular, a net orientation and polarization on the surface can be found which is the result of an augmented motion of certain molecules. Based on this result, surface hydrophobicity, its loss through oxidation, and its restoration through a unique interaction between cyclomethicone molecules, oxidized methyl groups, and counterions can be understood.

SYSTEM AND SIMULATION SETUP Our model system is based on a mixture of small and large siloxane molecules, similar to Sylgard-184©. The polymer blend consists of long PDMS molecules with vinyl termination (95 % of the mass), considerably shorter PMHS “cross-linker” molecules with a trimethyl termination, and very few small cyclomethicone molecules which are representative of depolymerization byproducts. Name PDMS PMHS D8 D4

MW [g/mol] 72562.3 1966.2 592.8 296.4

NA 9789 237 80 40

DP 978 32 8 4

Table 1: Constituents of the model system and their respective molecular weight (MW), number of atoms (NA), and degree of polymerization (DP)

Molecular dynamics calculations are performed on a system consisting of 106 atoms filling a cube of 22 nm edge length. For the sake of keeping periodic boundary conditions on all sides of the calculation cube, the surface is modeled by exposing the bulk to vacuum on two parallel sides at a distance of about 10 nm from each other. For the calculations on surface dynamics we can use both surfaces independently and thereby improve the statistics of the results. For the wetting calculations, only one surface is used. In both cases, particles on the two sides of the bulk do not interact and we therefore believe that our results are valid for arbitrary surfaces.

RESULTS Diffusion in the bulk and near the surface Our most general results stem from molecular dynamics calculations of the mean square displacement of PMHS, D8, and D4 molecules. When we evaluate these far from a surface we get results showing single-rate thermally activated, isotropic diffusion. Calculated diffusion constants are in the expected order of magnitude of 10-10 m2/s , but we see that the local structural and electronic environment has a strong influence: Certain trajectories show sudden deviations of up to a factor 2 from the average mobility due to a locally available increased free volume. Similar calculations performed in the immediate vicinity of the surface show accelerated diffusion as expected from the reduced local density.

Methyl groups The non-random alignment of methyl groups is seen by analysis of a geometric order parameter for the angle between the meth