Effects of filtration mechanisms on particles of various sizes
10/01/2003
By Mike Fitzpatrick and Ken Goldstein, Ph.D.
Last month, we discussed filters and how they work. We reviewed the three primary filtration mechanisms—interception, impaction and diffusion. We also briefly described sieving (the mechanism represented by a fly against a window screen) as a special case of interception.
We found that the determination of which filtration mechanism dominates a filtration process depends solely on the mechanical properties of the system, including:
- Particle diameter;
- Particle velocity;
- Filter fiber diameter;
- Fiber spacing.
And although a particular filtration mechanism may dominate, all three mechanisms are simultaneously at work at all times, hopefully contributing to the filter's performance. This month, we'll take a look at how the three mechanisms work together, and introduce the concept of the Most Penetrating Particle Size (MPPS).
The ability of a filter to "filter" is often expressed in terms of its Efficiency or its Penetration and by its MPPS. Recall from our previous article that:
Efficiency (E)—Fraction (or percentage) of particles captured by filter, for given particle size and velocity;
Penetration (P)—Fraction (or percentage) of particles that pass through filter, for given particle size and velocity.
Note that each of these definitions requires end users to specify a particular particle size. Thus, we find that the definition of a HEPA filter (from IEST RP-CC001.3) describes a HEPA as having "a minimum particle collection efficiency of 99.97% (that is, a maximum particle penetration of 0.03%) for 0.3-µm particles..."
Observe that this definition defines the filter's performance only on particles that are 0.3-µm in size and says nothing about particles that are larger or smaller. But what if we're concerned with 1.0-µm particles or with 0.1-µm particles?
To better understand how filters perform over a range of particle sizes, we need to consider a few graphs. Graph 1 plots a particular filter's efficiency—at a specified velocity—in "filtering" particles over a range of particle sizes. We see that the filter in Graph 1 has an efficiency of 0.999995 with particles 0.15-µm in size and, as the particle size increases, its efficiency increases to almost 1.0.
To our surprise, we also find that as the particle size decreases below 0.15-µm the efficiency of the filter increases, which is certainly counter-intuitive. This window screen stops all the big flies and, as we expect, some of the medium-sized flies get through. But this same screen also stops all the tiny flies. Is this some kind of magic screen?
If you read our September column, you'll quickly realize that there's no magic at play here. It's merely time to throw out our "window screen" concept and take a more realistic look at the forces at play—our window screen (sieving/interception) is only one of the mechanisms at work.
Larger particles are more likely to be removed by interception because their size puts them in close proximity to the filter fibers that capture them. And since size and mass are usually related, interception also works well on high-mass particles; however, as particle mass increases, impaction plays a significant role.
 |
null
We find that large (heavy) particles or those with high velocities are more likely to be removed by impaction, while small (light) particles, or those with low velocities, are less likely to be trapped by impaction. But as mass and velocities decrease, diffusion effects begin to dominate.
 |
null
When we take into consideration the effects of the different filter mechanisms on particles of different size, our graph of filtration efficiency versus particle diameter might look like Graph 2—typical for ULPA-grade ceiling filters.
 |
null
We see that for small particles, diffusion capture dominates, giving us near-perfect particle capture. And for large particles, interception and impaction dominate, again giving us near-perfect capture. But in the center part of Graph 2, things are not quite so rosy. These particles are small and light enough so that inertial impaction fails to capture them. At the same time, they are large and massive enough so that diffusion effects also fail to capture them all.
 |
null
These in-between sized particles are the most difficult to capture. Think of this as another example of the "Goldilocks phenomena." These mid-sized particles are too small for impaction, too large for diffusion but "just right" to make it through our filter.
Now, let's take a similar graph (Graph 3) that indicates the relative effects of the filtration mechanisms on particles of various sizes. Here we substitute penetration for efficiency on the vertical axis, making our numbers easier to deal with and flipping our chart by rotating it 180 degrees horizontally.
In Graph 3 we find that the maximum penetration (equivalently, the minimum efficiency) point is approximately in the center and indicates the Most Penetrating Particle Size (MPPS) for some given velocity and some given filter. The MPPS as defined in IEST-RP-CC007.1 is "that particle size for which penetration is a maximum" for a given filter at a specified velocity.
Graph 3 illustrates that interception and impaction capture are the dominant mechanisms for the larger and heavier particles resulting in low penetrations (high efficiencies) for these particles. Likewise, it shows that capture by diffusion is the dominant mechanism for the smaller, lighter particles resulting in low penetrations (high efficiencies) for these particles.
The total capture curve shows the MPPS to be slightly under 0.1-µm for this particular filter. Particles larger than this have lower penetrations (higher efficiencies), as we would expect. But particles smaller than this also have lower penetrations (higher efficiencies), which is counter-intuitive to anyone who thinks that our filters work just in the same manner as our window screens.
Filters are often rated and purchased based upon their MPPS. All other things being equal, a lower MPPS indicates a better filter; however, these other things are not always equal, as Graph 4 shows.
This graph illustrates the fractional penetration curves for two different filters. Which filter Is better? Looking only at the MPPS, one might conclude that Filter A is better since the MPPS of Filter A is lower than the MPPS of Filter B. But this conclusion is clearly incorrect.
Although the MPPS of Filter A is lower, the penetration for Filter B is less than that of Filter A for all particle sizes. And after all, the purpose of a filter is to capture particles. So, the filter that does a better job of capturing all particle sizes is superior to one with higher penetrations (lower efficiencies).
It's best to avoid evaluating filters based solely upon MPPS without taking into consideration all other relevant information.
 |
MICHAEL A. FITZPATRICK has participated in the design and construction of semiconductor facilities for more than 24 years and is a Senior Member of the Institute of Environmental Sciences and Technology (IEST). Mike can be reached at: [email protected].
 |
KEN GOLDSTEIN is principal of Cleanroom Consultants Inc. in Phoenix, Ariz., and is a member of the CleanRooms Editorial Advisory Board. He can be reached at: [email protected]