Particle-fluid separation without filters
10/01/2001
by Robert P. Donovan
|
Over the last few months we've been discussing the various filters commonly used for particle control in cleanrooms. These filters separate entrained particles from a fluidremove particles from recirculating ambient cleanroom air or from various process fluids used in cleanroom operations.
Using such fibrous or membrane filters, however, is not the only available method for achieving particle-fluid separations. For example, mechanisms that depend just on particle mass also separate large particles from smaller particles in a fluid and particles of any size from the molecules of the fluid.
Consider the following:
Gravitational settling: The gravitational force on a particle varies directly with particle mass. At Reynold's numbers < 1, the terminal velocity, VT, of a settling particle, obtained by equating the Stokes drag force on the particle to the gravitational force on that particle, is:
VT = ρd2g /18η
where:
Thus, in a static settling tank, large, dense particles sink to the bottom of the tank more rapidly than less massive particles, in effect diluting the concentration of massive particles at the top of the tank. This effect becomes graphic after filling a drinking glass with muddy water and letting it sit. While no filter or fluid flow is present, particle-fluid separation occurs.
Gravitational settling also occurs in dynamic fluid flows. For example, elutriators are devices that use gravitational settling to separate particles from flowing fluids. Elutriators can be either vertical or horizontal. Assuming laminar flow, vertical elutriators remove particles from the fluid whose settling velocity exceeds the fluid flow velocity, horizontal elutriators separate particles by settling along the direction of fluid flow. A horizontal elutriator continuously classifies particles according to mass.
In filter testing, a stable challenge fluid of constant particle size distribution and concentration is usually desired and gravitational separation of particles is unwanted. Here, gravitational settling will introduce undesired time-dependent changes in particle concentration and distribution unless continuous remixing by stirring, or other methods, is deliberately introduced.
Inertia: An impactor for measuring aerosol particle concentrations and collecting particle samples takes advantage of the dependence of inertial forces upon mass to separate particles from a fluid flow. This device consists of a capture surface placed perpendicular to the direction of the fluid flow. Low mass particles and fluid molecules flow around the capture surface; large, massive particles do not. Their inertia prevents them from following the streamlines and they collide with the capture surface where they stick and can be evaluateda useful particle-fluid separation effected without a traditional filter although utilizing one of the same particle capturing mechanisms.
Diffusion: Particle diffusion coefficients also depend on particle mass. Low-mass particles diffuse more rapidly than large-mass particles and gas molecules diffuse still more rapidly than particles. This mass dependence means that the size distribution of particles flowing through a long thin tube changes with distance, as the more rapidly diffusing particles in the fluid collide earlier with the tube walls and are captured.
A diffusion battery capitalizes on this property to fractionate aerosol particles of diameters smaller than 0.2 to 0.3 mm, analogous to the action of the horizontal elutriator which fractionates particles of diameter larger than 0.5 mm. Because large-mass particles diffuse more slowly than low-mass particles, the mass of the captured particles in a diffusion battery increases with distance from the entry point of the flow.
These mechanisms of particle-fluid separation contribute to particle control and/or measurement in cleanrooms, especially in environments inhospitable to or inappropriately sized for traditional filter media.
References
- Hinds, W. C., Aerosol Technology, John Wiley & Sons, 1982, p. 42.