David Y. H. Pui, Associate Fellow

Nanoparticle Collision With a Surface

Filters are routinely used to remove toxic nanoparticles from industrial processes. Wang and Kasper (1991), however, argued that 2–3 nanometer particles may penetrate the filter by bouncing off the filter internal surfaces. If this controversial prediction is true, it could mean a serious problem.

To investigate the possibility of such bounce in more detail, the researchers used molecular dynamics (MD) simulations that track the motion of the particle and surface atoms by integrating their equations of motion. The simulations sought to solve the above controversy by providing detailed collision information, which is difficult to obtain using other approaches.

Another aspect of the group’s work concerned the collisions of nanoparticles with surfaces. This research bears on many applications, including material synthesis, microcontamination, and health effects. The properties of nanoparticles were expected to be different from those of large particles (e.g., micrometer range) due to the size effects or high proportion of atoms exposed to the surface. Accordingly, the interaction of nanoparticles with surfaces should be understood considering such effects. Classical MD methods were used in this study to understand particle-surface collisions from a microscopic viewpoint.

The main part of this project studied collision dynamics as a function of the incident velocity of particles. Materials of particles and surfaces were chosen so that they could simulate actual situations in filtration processes. The simulations in the research, which comprised the main part of this study, used computationally- intensive classical MD methods.

Visualization of particle trajectories is important both for understanding collision dynamics and presenting the results. The use of supercomputers, with algorithms suitable for multiple processors, dramatically reduced the computation time of the simulation; specifically, the resources of the Basic Sciences Computing Laboratory were used to help in this visualization process.

The researchers deposited particles in a tube with an abrupt contraction at low pressure. Prior to this, deposition mechanics such as this had been studied both experimentally and numerically: the main consideration of deposition mechanism here was the inertial impaction. In the experiment, the particle deposition efficiencies onto an orifice plate (orifice diameter: 1.16 cm) placed in a tube (inner diameter: 3.48 cm) were measured. The system pressure was controlled in a narrow range from 0.2 to 0.28 torr and the flow Reynolds number for the tube was 3.0. Spatially uniform aerosols were produced at low pressure by the method developed by Sato et al. (2000). The observed deposition curve, as a function of the particle Stokes number, was shown to be different from that obtained at atmospheric pressure for a Reynolds number greater than 1,000 (Chen and Pui, 1995). The difference was further confirmed by the numerical results for the cases with a Reynolds number of 3.0. The simulation also showed that the deposition curve had shifted to a smaller Stokes number as the Reynolds number increases. Empirical expressions for the deposition efficiency as a function of the Stokes number, the Reynolds number, and the contraction ratio (the ratio of the tube inner diameter to the orifice diameter), R, were given in order to estimate the particle deposition efficiency for any combination of these parameters for future applications.

Research Group

Da-Ren Chen, Faculty Collaborator
Hee-Siew Han, Graduate Student Researcher
Poshin Lee, Graduate Student Researcher
Shintaro Sato, Graduate Student Researcher