(15b) Impact Behavior of Microparticles on Component Surfaces Considering the Micro-Topography: Experiment and DEM Simulation

Authors: 
Krull, F., University of Kaiserslautern
Antonyuk, S., Technische Universität Kaiserslautern
The Discrete Element Method (DEM) is widely used to simulate particulate processes. The walls of apparatuses and tolls in these simulations are often calculated as smooth surfaces, and the micro-topography, for example the surface roughness, is not considered. However, with decreasing particle size the influence of the surface asperities is becoming more and more important, because the deformation of single asperities and adhesion forces in the contact area can become dominant compared with the entire deformation and adhesion of the particle. On the micro scale the contacts with asperities take place often oblique whereas the impact on the macro scale is in normal direction to the wall surface. Furthermore there is no continuous contact area and the contact force during impact is distributed over a many single asperities. This leads to an additional energy dissipation and influences the rebound angle and rotation of the particle. Therefore it is necessary to describe the influence of the surface micro-topography on the particle impact behavior depending on the particle size.

In this work the impact behavior of microparticles with component surfaces was investigated by the simulation using DEM and experiments. The main focus was on the influence of the surface micro-topography on the rebound angle and the coefficient of restitution, which depends on the ratio between kinetic energies of the translational movement after and before particle impact.

The impact tests of single microparticles with different surfaces were performed. For these experiments a novel setup was developed. The collision of spherical particles with a diameter varied in the range of 100-700 µm was captured three-dimensionally with three high-speed cameras coupled with the microscope. The particle was dropped vertically to the surface to obtain a normal impact. The trajectory, angles and velocities of the particle were measured directly during the impact and rebound. The tests were performed with a polished and an unpolished titanium surface and smooth polystyrene microparticles. The experiments were repeated several times for each particle size and surface to measure the distribution of rebound angle and coefficient of restitution depending on the micro-topography.

To study the micromechanisms of the collision the DEM simulation was used. The titanium samples investigated by experiments were measured by Scanning Probe Microscopy (SPM) imaging with a nanoindenter (TI Premier, Hysitron) to obtain the micro-topography of surfaces. The SPM imaging was performed with a diamond Berkovich tip (with a tip radius of 100 nm) at a constant normal force of 2 µN. A total area of 2 mm x 2 mm was scanned by 40 x 40 consecutive areas with a size of 50 µm x 50 µm. The total area was reconstructed by merging the measured areas of the SPM imaging. A CAD model of this area with the resolved micro-topography was obtained by reverse engineering for the polished and unpolished surfaces. The point cloud of the surface obtained from the SPM imaging was meshed in an unstructured triangulated surface and implemented directly in the DEM simulation environment. The mesh resolution of the surfaces was varied to investigate the influence of the triangle size on the simulation results. With coarsening and refinement of the triangulated grid a mesh convergence could be obtained. Additionally an ideal smooth surface was simulated. The roughness of the polystyrene particles was neglected in the simulations, because it is by a factor of three to five smaller than the roughness of the investigated titanium surfaces. The required parameters for the contact model, like the coefficient of restitution and elastic modulus, were obtained by the impact tests and nanoindentation of the samples and particles. The impact point of the particles on the microstructured surface was randomly varied in the simulation to obtain the distribution of the parameters which describes the influence of the surface roughness. The simulation results for the different surfaces (unpolished, polished and ideal smooth) were compared with the experiments using the mean values and distributions of the coefficient of restitution and the rebound angle.