AAAR 34th Annual Conference
October 12 - October 16, 2015
Hyatt Regency
Minneapolis, Minnesota, USA
Abstract View
Low-Pressure Particle Filtration Measurements
Meilu He, SURESH DHANIYALA, Matthew Wagner, Clarkson University
Abstract Number: 588 Working Group: Control Technology
Abstract Pall metal filters are widely used in semiconductor industry for inline control of contamination in high-purity precursor gases. These filters are usually deployed with gas lines operated at low pressures, but their performance characteristics have only been studied under atmospheric pressure because of the challenges of sub-atmospheric filter testing. In this project, the particle capture performance of the filter was studied experimentally and theoretically as a function of its operating pressure. The experiments were designed to minimize artifacts associated with multiple charging effect that can be amplified under low pressure conditions. The filter penetration results show that the penetration is lower with non-spherical particles compared to spherical particles and the measurement difference between the two type particles increases with increasing particle size and decreasing operating pressure. Measurements were made to a minimum pressure of 4.5 kPa with a commercial condensation particle counter (CPC; TSI 3010) modified for low pressure operation. The experimental results together with the classical filtration theory were used to obtain effective filter parameters. Using these effective filter parameters, acceptable agreement between theoretical predictions and experimental results under ambient and sub-ambient pressures when the inhomogeneous filtration model of Dhaniyala & Liu (2001) is applied for the prediction. In aerodynamic slip flow regime the slip boundary condition is applied for the prediction, while in the transition flow regime the prediction based on non-slip boundary condition has better agreement than that from the slip boundary condition. Using a universal geometric standard deviation of the packing density distribution, filter penetration performance can be calculated at any flowrate and operating pressure with high accuracy.