AAAR 35th Annual Conference October 17 - October 21, 2016 Oregon Convention Center Portland, Oregon, USA
Abstract View
Effect of Translational and Rotational Brownian Motions on the Drift Velocity of Non-Spherical Aerosol Particles
CHARLES HAGWOOD, George Mulholland, National Institute of Standards and Technology, Gaithersburg
Abstract Number: 130 Working Group: Aerosol Physics
Abstract A one dimensional model (Mulholland et al., accepted for publication in the J. Aerosol Science) for the effect of the rotation rate on the drift velocity of a nanorod in the free molecular regime was previously studied over the parameter range from fast rotation to slow rotation relative to the linear aerosol relaxation time. In this case the dependence of the friction coefficient on the orientation was proscribed as a trigonometric function of the product of the rotation velocity and the time. In reality, the rotation of the nanorod is driven by rotational Brownian motion. In the current study, we assume a Langevin equation (Ornstein-Uhlenbeck process) for the rotational dynamics. Because the friction tensor for translational motion will depend on the particle's orientation, the stochastic differential equation describing the translational velocity is coupled with the rotational Langevin equation. We solve this pair of equations numerically using the Ermak- Buckholz numerical integration scheme. The numerical integration is challenging since the time step must on the one hand be small enough to resolve the change in the orientation angle. On the other hand, it must be at least a factor of ten longer than both the translation aerosol relaxation time and the Brownian rotation time scale to obtain an average value of the drift velocity relevant to a measured drift velocity. A second solution method based on the analysis of the conditional probability distribution for the translational velocity distribution given the orientation distribution function was also used. This approach results in a Feynman- Kacs path integral. The results for these two methods will be compared. The results of Langevin approach will also be compared with the previously published results for the one dimensional model.