THE TRUTH ABOUT HEPA FILTERS

Andrew F. Oberta, MPH, CIH

The Environmental Consultancy

www.asbestosguru-oberta.com

            Which of the following definitions and statements is the most correct?

A) High-efficiency particulate air (HEPA) refers to a filtering system capable of trapping and retaining at least 99.97 percent of all monodispersed particles 0.3 μm in diameter or larger. (EPA AHERA regulations, § 763.83 Definitions, 1987)

B) HEPA Filter: A High Efficiency Particulate Air (HEPA) filter capable of trapping and retaining 99.97% of all mono-dispersed particles of 0.3 microns in diameter. (Model Asbestos Abatement Guide Specification, National Institute of Building Sciences 1996)

C) HEPA Filter: High Efficiency Particulate Air filter. Such filters are rated to trap at lease 99.97% of all particles 0.3 microns (0.3μm) in diameter or larger. (Guidance Manual: Asbestos O&M Work Practices, National Institute of Building Sciences, 1996)

            If you picked B), congratulations. A) and C), while not exactly wrong, are incomplete at best and misleading at worst. The most complete and accurate definition of a HEPA filter I have found is the following:

“A throwaway, extended-media, dry type filter with a rigid casing enclosing the full depth of the pleats. The filter shall exhibit a minimum efficiency of 99.97% when tested at an aerosol of 0.3 μm diameter.”

(Specification for HEPA Filters Used By DOE Contractors, DOE-STD-3020-97, January 1997)

            What is incomplete and misleading about A) and C) above, variations of which appear in countless regulations, training manuals and other sources? HEPA filters do not act like a window screen, catching all particles >0.3μm and letting all others pass through. Filter performance depends on the size of the particles in the airstream, and the implication of A) and C) is that a HEPA filter is less that 99.97% efficient for particles <0.3μm in diameter. One would expect such a filter to perform as shown in Figure 1. The fact that this is not true comes as a surprise to many people, and I will explain why it is so.

Figure 1. Incorrect interpretation of HEPA filter performance

Before I do, let me stress that the definition applies to the filter and not to the device – respirator, vacuum cleaner, “negative air machine,” lab hood, clean room, etc. – that it is part of. Also, the performance rating of a HEPA filter is based on capturing spherical particles, not asbestos or other fibers and odd-shaped particles. Therefore, the fact that the filter meets this test standard may not accurately reflect the performance of the device under field conditions.

PRINCIPLES OF FILTRATION

            HEPA filters belong to a class of devices known as depth filters, as opposed to membrane filters. The “pleats” in a depth filter (see DOE definition) are plainly evident in the cutaway respirator cartridge in Figure 2. In a large filter for a ventilation system, including the negative air machine in Figure 3, the filter can be nearly a foot thick.

Figure 2. HEPA respirator cartridge filter	Figure 3. HEPA "negative air machine" filter

            The filter media is made up of densely packed fibers. Each fiber traps particles in the airstream passing through the filter by the three physical mechanisms shown in Figure 4. This diagram shows the cross-section of one fiber in the filter media (forget asbestos fibers for this explanation), the air flowing past the fiber as represented by the curved streamlines, and three airborne particles of different sizes.

Figure 4. Capture mechanisms for a depth filter

The capture mechanisms are:

            Interception – A particle following an imaginary airflow streamline is intercepted by the fiber because the diameter of the particle is more than twice the distance from the fiber surface to the streamline that passes through the center of the particle.

            Impaction – The trajectory of the particle departs from the imaginary airflow streamline due to its inertia, which is a function of the particle’s mass and velocity, and the particle impacts the fiber.

            Diffusion – The particle’s trajectory oscillates about the imaginary airflow streamline in a random manner known as Brownian motion, which can cause the particle to impact the fiber.

            The larger the particle, the greater the role of interception and impaction in capturing particles, which is a way of saying that these mechanisms are more efficient in capturing larger particles than smaller ones. Diffusion works in the opposite direction: smaller particles act more like the gas molecules we associate with Brownian motion, so it is a more efficient particle capture mechanism for smaller particles.

CALCULATION OF FILTRATION EFFICIENCIES

            Solving the equations for the efficiency of a single fiber in capturing particles of a range of diameters yields the values in Table 1 and the curves in Figure 5. [1] Note that both scales in the figure are logarithmic.

 Figure 5. Capture efficiencies for single fiber

    The overall capture efficiency due to all three mechanisms for a single fiber can be determined by solving the following equation for each particle size:

            η_S = 1 - (1 - η_I)(1 - η_P)(1 - η_D)

where

            η_S = the total capture efficiency

            η_I = the capture efficiency due to interception

            η_P = the capture efficiency due to impaction

            η_D = the capture efficiency due to diffusion

            Doing so yields the values in Table 2 and the bottom curve in Figure 6, which is for a single fiber. Filters, of course, are composed of many fibers, and the density of the filter media is represented by S in the following equation, which gives the overall efficiency of the filter for each particle size:

            η_F = 1 - e ^{- η}_S^/S

            Figure 6 shows two curves in addition to the one for a single fiber. The middle curve is for a low-efficiency filter with a density S = 30 and the top curve is for a HEPA-rated filter with a density S = 300. In Table 1, the minimum efficiency in the last column is in the range of 99.97% at which HEPA filters are rated dnd tested.

Dp, µm	One fiber	Low η filter (S=30)	HEPA filter (S=300)
0.010	26.19%	99.96%	100.00%
0.014	18.30%	99.59%	100.00%
0.021	11.77%	97.07%	100.00%
0.030	8.16%	91.36%	100.00%
0.043	5.73%	82.05%	100.00%
0.062	3.95%	69.40%	100.00%
0.100	2.54%	53.35%	99.95%
0.144	2.12%	47.02%	99.83%
0.208	2.10%	46.77%	99.82%
0.300	2.62%	54.39%	99.96%
0.433	4.13%	71.07%	100.00%
0.624	7.14%	88.26%	100.00%
1.000	15.30%	98.98%	100.00%
1.442	27.48%	99.97%	100.00%
2.080	47.96%	100.00%	100.00%
3.000	78.84%	100.00%	100.00%
Table 2. Total fiber and filter collection efficiencies				Figure 6. Collection efficiencies for one fiber and filters

            Notice that the curves in Figure 6 are not monotonic. Because diffusion is the dominant capture mechanism for small particles, the overall efficiency increases for particles below a certain size. Table 2 and Figure 6 show a minimum capture efficiency for a single fiber and both filter media occurring in the 0.1μm to 0.3μm range. Expanding the vertical scale between 90% and 100% for the conditions under which HEPA filters are tested would show a minimum capture efficiency of 99.97% at 0.3μm, which is the criterion that the filter has to pass according to the DOE specification. This is the most stringent measure of performance for this type of filter media, assuring that it will exhibit greater capture efficiency for larger and smaller particles.

[1] I am indebted to Dr. Klaus Willeke for his lectures in Particle Technology at the University of Minnesota where I obtained my MPH degree. Many years later, his explanations and examples of these mechanisms are still understandable.

© 2004 The Environmental Consultancy. All rights reserved.

RETURN TO HOME PAGE