ORIGINAL PAPER/PEER-REVIEWED

Fiberglass Vs. Synthetic

Air Filtration Media
By Edward Vaughn, Professor, and Gayetri Ramachandran, Graduate Student,
Clemson University, Clemson, SC

Abstract Ultrafiltration, Microfiltration, and General Particle Filtration

This paper deals with ASHRAE filters used in air filtration [9]. Depending on the requirements, different filtering sys-
applications and summarizes, on the basis of current research tems can be used. If surface filters are used, then the contam-
work, some of the reasons for the disagreement that exists inants are trapped and held on the surface of the media. On
among filter manufacturers concerning the properties and the other hand, if depth filters are used, larger particles are
performance of various types of air filter media in general, caught on the filter’s surface and finer contaminants are
glass and synthetic media in particular. Attention is also trapped in the media’s fibers within. Accordingly, the most
drawn to some important items that need to be incorporated suitable filter media that is disposable or reusable needs to be
in test methods and to some factors that could be affecting fil- selected. A wide variety of filter media are available. Some of
ter performance them include membranes, microporous plastics, sand,
. diatomaceous earth, perlite, paper, woven metal wire, woven
Introduction and nonwoven fibrous media. Nonwoven fibrous media are
The filtration industry today is a diverse and technically made of synthetic fibers, fiberglass and paper [1].
sophisticated business with annual sales reported to be in
excess of $100 billion [1]. Filter producers supply various Filter Media Classification
types of filtration systems and filter media designed to meet The classification of filter media depends on the test
a wide variety of liquid, air and gaseous fluid filtration needs. method used. As most testing is performed in the laboratory
The filtration business has evolved over time to become a with synthetic dust, the classification does not always provide
complex industry with very specific requirements for each a reliable basis for the estimation of a filter’s life or its per-
area of use. Performance standards for the media used in vir- formance in actual application. According to the European
tually every application have become very stringent. Recent Classification, particle filters are categorized into four types:
studies and ongoing research in the area of air filtration sug- Course, Fine, High Efficiency Particulate Air Filters (HEPA),
gest that the present test methods may be inadequate in pre- and Ultra Low Particulate Air Filters (ULPA). Filters pro-
dicting the most economical choice of media, with specific duced from glass fibers or synthetic fibers like polyester,
efficiency levels for different end-use applications. Also, acrylic and polyamide fibers, which separate particles that are
some test methods have been criticized for their inability to 5mm and larger, are designated as course filters. Fine filters
reflect the true performance of filters in real-life environment are made mainly from glass fibers with an average diameter
over their full lifetime. of 0.5 - 5.0 mm or from coarse plastic fibers, often in combi-
nation with an electrostatic charge. European Test Standards
classify filters according to their Arrestance (Am) and Dust
The Filter Spectrum Spot Efficiencies (Em). Table 1 gives the classification of fil-
The filtration spectrum covers a wide size range from ionic ters on this basis. For example, a filter with an average
particles measured in angstroms to larger solids up to several arrestance value between 65% and 80 % is designated as an
hundred microns in size. The spectrum is divided into four EU 2 filter by Eurovent Classification and G 2 filter by EN
broad categories as shown in Figure 1: Reverse Osmosis, 779 Classification. A filter with an average dust spot efficien-

INJ Fall 2002 41

 through the filter media. Depending on the total level of sep-
aration and leakage, a filter is classified as H10, H11, H12,
H13 or H14 and U15, U16 or U17.
Chemical filters are adsorption filters impregnated with
chemical substances that contain activated carbon. By means
of chemical reactions, these filters adsorb and retain gases
that are difficult to remove [4].

The Fiberglass Media Versus Synthetic Media Issue

High efficiency fiberglass filter media have been an indus-
try standard for air filtration applications. These media are
characterized by a dense structure of fine glass fibers, typi-
cally in the one-micron range. More recently, synthetic fiber
filter media with a more open structure formed from relative-
ly coarse fibers - mainly electrostatically charged polypropy-
lene fibers - have been introduced.
Short-term tests performed in the laboratory on the basis of
ASHRAE and European test procedures show that initial and
average efficiencies of these two types of filter media are
comparable. But extensive field-testing and real-life tests on
the two media show that laboratory tests do not predict the
performance of filters over their whole service life. Also, life-
time tests reveal the fact that there is a great difference in the
filtration performance of the two types of filtration media.
Glass media maintains its efficiency, while synthetic media
loses its efficiency over its service life [2].
Figure 1 To understand this disagreement between real-life test
FILTRATION GEOMETRY SCALE AND TYPICAL results and laboratory results generated from ASHRAE and
PARTICLES CAPTURED THEREIN (9) European test standards, one needs to look keenly into the
various factors that affect the filtration capability of filter
cy of 95% or less is designated as an EU 9 filter by Eurovent media. The standard tests do not cover the entire gamut of fil-
Classification and F 9 filter by EN 779 Classification. tration media types and challenge environments, and their
To meet current demands of clean air for specialized appli- results must be interpreted in light of knowledge of the char-
cations, such as the military, the nuclear power industry, hos- acteristics and properties of the filter medium in question and
pitals and the electronic industry, HEPA and ULPA filters are the conditions to which it is subjected.
used. Based on the CEN EN 1822:1998 test method, a filter’s
efficiency is determined by the Most Penetrating Particle Size
Filtration Mechanisms
A popular misconception regarding how a filter works is
(MPPS) value. The Most Penetrating Particle Size is defined
that fibrous filters behave like sieves, where particles above a
as the most frequently occurring particle size that penetrates
certain size are trapped
and smaller particles
Table 1
pass through. While this
EUROVENT CLASSIFICATION OF FILTERS (4).
is the case with some
Eurovent EN 779 Average Average Final Pa Filter membranes filtering liq-
Filter Filter Synthetic Atmospheric Classification uids, fibrous air filters
Class Class Dust Weight Dust Spot and Filter defy common sense by
Arrestance in % Efficiency in % Designation actually trapping small-
EU 1 G1 Am< 65 - 250 PRE-FILTER er and larger particles
EU 2 G2 65 < Am< 80 - 250 PRE-FILTER more effectively than
EU 3 G3 80 < Am < 90 - 250 PRE-FILTER mid-sized particles.
EU 4 G4 90 < Am - 250 PRE-FILTER Four mechanisms act
EU 5 F5 - 40 < Em < 60 450 FINE AIR FILTER to separate a particle
EU 6 F6 - 60 < Em < 80 450 FINE AIR FILTER from a fluid stream and
EU 7 F7 - 80 < Em < 90 450 VERY FINE AIR FILTER retain it on a filter medi-
EU 8 F8 - 90 < Em < 95 450 VERY FINE AIR FILTER um, namely,
EU 9 F9 - 95 < Em 450 VERY FINE AIR FILTER Interception, Inertial
Impaction, Brownian

42 INJ Fall 2002 2

 Figure 2 Figure 4
DIRECT INTERCEPTION MECHANISM BROWNIAN DIFFUSION

ter and hence enhance the mechanical filtration capability of

the filter medium.
Fiberglass filter media with their finer fibers utilize the first
three filtering mechanisms and are enhanced by the number
of fibers present per unit volume of the media. With the
decrease in fiber diameter, the number of fibers per unit area
increases. Also, the path that the contaminate particle must
take through the filter media becomes much more intricate,
Figure 3 thus dramatically increasing the chances that the particle will
be captured on the fiber surface by one of the physical filter-
INERTIAL IMPACTION MECHANISM
ing mechanisms.
Because of larger diameter fibers, large voids are present in
Diffusion, and Electrostatic Capture. synthetic media. This characteristic reduces the possibility of
Interception occurs when a particle following a gas stream- a particle colliding with a fiber. Application of electrostatic
line comes within one particle radius of a filter fiber. As charge to the fibers in synthetic media results in a greater
shown in Figure 2., the particle touches the fiber and is cap- attraction of dust particles on to the fiber. Atomic forces sub-
tured, thus being removed from the gas flow. Streamlines far- sequently hold the particles onto the media [2].
ther than one particle radius from the filter fiber will not con-
tribute to the interception mechanism. The size of the particle Characteristics of Fiberglass and Synthetic
determines how close it moves to the fiber. Filter Media
Inertial Impaction generally occurs with larger particles Fibrous filter media made from glass fibers or synthetic
that are unable to quickly adjust to streamline direction near a fibers are widely used, primarily in disposable filters, due to
filter fiber. Due to its inertia, the particle will continue along the favorable characteristics of low cost, depth filtration, good
its original path and hit the filter fiber and fall down in the dust holding capacity and variety of constructions. These
media, as shown in Figure 3. products are used in industrial as well as consumer applica-
The very fine particles in the air stream collide with the gas tions such as engine air filters, furnace filters, building venti-
molecules and create a random path through the media. The lation filters, cleanroom air filters, and gas cleaning filters in
smaller the particle and the lower the gas velocity, the longer nuclear installations [9].
the particle will zigzag around. The resulting random motion, Porosity: Fibrous filter media can be manufactured with
called Brownian Diffusion, will increase the probability of the uniform porosity or with a gradient density through the filter
particle impacting the fiber surface and adhering to it. This is depth. A variation in spacing between adjacent fibers results
shown in Figure 4. in a non-uniform undefined interconnected porous structure.
Electrostatic Capture requires imparting an electrostatic Particle Capture through the Filter Depth: In both filter
charge to a synthetic fiber during its formation. The filter types, particles are collected on the surface, as well as
media so formed have charges on the fibers and hence are throughout the interior of the filter media, this characteristic
able to attract dust particles. By this method of particulate results in their being classified as depth filters.
capture, the small particles initially adhere to the fibers and Pressure Drop: Fiberglass filter media are made from flame
form the nucleus for progressive attachment of more dust par- attenuated glass fibers of about 1.0 - 1.3 mm in diameter.
ticles, which finally results in the formation of conglomerate Typical synthetic media consists of one or more layers of
clumps or protuberances on the fibers [8]. This phenomenon fibers depending on the requirements. One of the layers pro-
is called dendrite formation. Continual attachment of the con- vides particle-capturing efficiency of the product. This layer
taminants onto each other results in the development of den- consists of 3.0 - 4.0 µm diameter fibers that capture particles
drite colonies which load the filter, reduce the spacing utilizing electrostatic enhancement.
between adjacent fibers, reduce the size of the voids in the fil- The fibers in the synthetic filter media, being coarser than

INJ Fall 2002 43

those in the glass media, are not able to pack as closely as the (about 1.0 µm in diameter) as compared to those used in syn-
glass media. The fibers are not able to come close together thetic media (about 3-4 µm). As fiber diameter decreases, for
and so, leave large void spaces between them. Hence, syn- the same mass of filter media, the number of fibers per unit
thetic filter media show lower pressure drop than fiberglass area increases and hence the surface area increases. Also, the
media [2]). path that a particle must travel through the media becomes
Loading: Synthetic media with coarser fibers have a more more intricate. Accordingly, the capability of particle capture
open structure as compared to the glass media. This charac- by physical mechanisms of Brownian Diffusion, Inertial
teristic yields a greater void volume per unit area and a high- Impaction and Interception improves greatly.
er fluid permeability and dust holding capacity.
Associated Cost: One objective of a filter manufacturer is Relationship Between Fiber Diameter And
to provide the most economical solution to a filtration prob- Surface Area
lem. A number of factors influence the total cost associated The effect of fiber diameter for typical glass and synthetic
with using a filter unit. The filter media needs to be designed fiber media is illustrated by an analysis of the data outlined in
so that it will achieve its performance requirements satisfac- Table 2 comparing the parameters of competitive glass and
torily while keeping the costs incurred to a minimum. Life meltblown filter media.
cycle analysis and life cycle cost studies are extremely useful The diameter-denier relationship for circular cross-sections
tools in assessing the costs of a filter function. Life cycle is given by,
analysis considers the environmental effect with reference to
ecological effects, health effects and consumption of (1)
resources. The LCA protocol provides a cost analysis of the
effect of a filter on the environment. Cost of raw material,
refining, manufacturing and transportation corresponds to From which,
approximately 20-30% of the environmental load, while filter
operation accounts for up to 80%. Energy returned by burn- (2)
ing the filter can reduce the environmental load by 0.5 - 1%.
A decrease of 10 Pa in the pressure loss reduces the load by
5%. Life cycle cost takes into account the economic aspects For the glass media,
of filter usage. The costs of investment, energy, maintenance
and dumping the final waste product throughout the lifetime (3)
of the plant are evaluated in LCC. One study shows that the
costs of the filter, investment, and maintenance correspond to
20% of the total cost. The energy cost for operation of the fil- Similarly, for the synthetic media,
ter plant accounts for 80%. Used filter disposal costs account
for 0.5%. (4)
Accordingly, it can be concluded that operation and low-
pressure loss are absolutely decisive in determining the cost
of a filtration system. LCA and LCC are excellent tools that According to the definition of denier and assuming that the
help in designing filters to minimize the cost of filtration [4]. fiber media is continuous and laid side by side,
For glass media with a denier of 0.01437,
Filter Media Properties 0.01437 is the weight of a 9000 meter length of fiber.
Physical parameters such as fiber diameter, fiber geometry, Therefore, the glass media basis weight of 49.5156 grams
fiber specific surface, fiber density, filter thickness, packing is equivalent to a total length of fiber of
density, porosity and pore size distribution are major factors,
which influence filtration efficiency.
Fiber Diameter: Scanning electron microscopy studies
show that fibers used in fiberglass filter media are finer
meters in 1 square meter of glass media.
For synthetic media with a denier
of 0.1904,
Table 2 0.1904 is the weight of 9000
PARAMETERS OF COMPETITIVE GLASS AND SYNTHETIC FILTER meter length of fiber.
MEDIA Therefore, the synthetic media
basis weight of 77.5027 grams is
µm)
Diameter (µ Specific Gravity Basis Weight (g/m2) equivalent to a total length of
Glass 0.9 2.5 49.5156 fiber of
Synthetic 5.4 0.92 77.5027

44 INJ Fall 2002

meters in 1 square meter of synthetic media.
By definition, the surface area of a cylindrical rod is equal
to its circumference times its length, or
(5)
Accordingly, for glass media,

(6)

Thus, 1 square meter of glass media, with a weight of

49.5156 grams has a surface area of 87.6395 square meters.
And, for synthetic media,

(7) Figure 5
Thus, 1 square meter of synthetic media, with a weight of CROSS-SECTIONAL VIEW OF 4DGTM FIBERS
77.5027 grams has a surface area of 62.1178 square meters. (COURTESY OF FIBER INNOVATION TECHNOLOGY)
Defining specific surface area as the surface area per
weight of the media, enhance particle capture and compensate for smaller surface
For the glass media, area. But within a few weeks of service, as electrostatic
charges are neutralized, the smaller surface area of synthetic
media yields lesser efficient performance, as compared to the
(8) glass media.
Fiber Geometry: Fiber geometry as well as surface area can
significantly affect the filtration capability of the media.
For the synthetic media, These characteristics have not been fully utilized by the fil-
tration industry. Figure 5. illustrates a fiber, designated
4DGTM , with both an unusual geometry and high surface
area. This fiber was introduced by the Eastman Chemical
(9)
Company a few years ago.
Earlier studies indicate that existing crenulated fibers had
The above analysis is summarized in Table 3. limitations with respect to surface geometry and had insuffi-
The ratio of specific surface areas of glass and synthetic cient channels for trapping and holding particulates. To
media is respond to this challenge, the Eastman Chemical Company
developed a deep-grooved polyester fiber with a very novel
(10)
eight-legged cross-sectional shape. Expansion of the fiber
perimeter results into high surface area of the fiber. The
This shows that surface area per gram of glass media is grooves in 4DG™ fibers are large enough to hold many types
more than two times that of synthetic media. of substances, whether they accumulate in use or are inten-
Glass fiber has a higher specific gravity than synthetic tionally placed there for release while being used.
fiber, but its smaller diameter more than compensates for this The photomicrograph in Figure 6. shows that in addition to
and results in a greater surface area. High surface area particulate matter being deposited in the interstices between
enhances the filtration capability of the media. the fibers, it also collects between the grooves. The grooves
Initially, for the synthetic media, electrostatic charges provide areas where eddy currents will preferentially deposit
particles without blocking the
pore of the fabric. This results in
Table 3 longer life and reduced weight of
CALCULATED SURFACE AREA VALUES FOR GLASS the filter [10]. Figure 7 shows
AND SYNTHETIC MEDIA carbon particles placed in the
Denier Total Filament Surface Area) Specific grooves for odor absorbency.
Lenngth in 1m2 m2 Surface Surface Area: Specific surface
of media (m) (m2/g) area of a fiber has a direct impact
Glass 0.01437 31,011,858.04 87. 6395 1. 7699 on the filtration performance of a
Synthetic 0.1904 3,663,467.962 62.1178 0.8015 filter. The cross-section illustra-
tions in Figure 8. show the 4DG
geometry compared to that of

INJ Fall 2002 45

 round fibers. Both configurations, by
definition, have the same cross-section-
al area when the polymer type and
denier per filament (dpf) are the same.
But, the measured surface area of the
4DG fiber is 2.3-2.8 times that of a
round cross-sectioned polyester fiber of
the same denier. The size of a single
round fiber needed to match the perime-
ter or surface area of the 4DG fiber is
also shown. .
As shown in Figure 9., when specific
surface areas are compared, 6 d/f 4DG
is found to be equivalent to round poly-
Figure 6 Figure 7 esters of approximately 0.8 d/f [10].
DUST LOADED IN GROOVES 4DG FIBERS DEPOSITED WITH With greater specific surface area of the
OF 4DG FIBER (10) CARBON (10) fibers, the possibility of particles of
interest colliding onto the media fibers
increases, thereby improving the filtra-
tion capability.
Porosity: Filtration performance also
depends on the porosity of the medium.
If the medium is highly porous, it will
allow particles to pass through it easily
and not perform the filtration function
satisfactorily. In glass media the fine
fibers can pack closely, hence the
porosity is less. In synthetic media the
coarser fibers cannot pack as closely
together. The pores in the glass media
being smaller can capture the particles
Figure 8 better by entrapment. Pore size distribu-
COMPARISON OF SURFACE AREAS OF 4DG™ AND tion is an important factor that deter-
mines which particles will be allowed to
ROUND FIBERS OF THE SAME DPF (10)
pass through and which particles will be
retained.

Design of Filter Media

Nonwoven filter media are designed to accommodate the
environment to be endured and to be functional either from a
structural support or effective surface area availability stand-
point. Parameters such as media pore size distribution and
the relationship between fiber surface area per unit weight or
per unit volume can also be used effectively in designing fil-
ter media structures.
The design goal for filter media is to maximize the space
available for filtration in order to remove large amounts of
undesirable contaminants, while not allowing them to pass
through the filter, and to keep the operating pressure differ-
ential at rated air flow as low as possible to achieve a long
service life. The filter design engineer must have an in depth
knowledge of the application, type of fluid to be filtered
Figure 9 /separated, acceptable power usage allowed to generate flow,
RELATIONSHIP BETWEEN DENIER AND and an understanding of the type and nature of contaminant
SPECIFIC SURFACE AREA FOR ROUND AND to be removed in order to maximize filter performance at
SHAPED FIBER CROSS-SECTIONS minimum cost [3].
Accordingly, the three design criteria that need to be con-

46 INJ Fall 2002 2

sidered in a filter product design are: Filtration Efficiency, Manufacturers of air-handling equipment, is leading the
Dust Holding Capacity, Filter Resistance to Air Flow and development of new and modern test methods that can meet
Environmental Conditions. today’s requirements for ventilating air filter installations in
Filtration Efficiency is defined as the efficiency of the fil- indoor environments and other areas. The European
ter product in capturing and removing the contaminants of Committee for Normalization (CEN) is working to establish
interest. Different applications require different levels of fil- common standards in Europe (11).
tration efficiency. In some processes, the specified level of The ANSI/ASHRAE 52.1–1992 test standard,
filtration efficiency is vital for normal operation. An under- “Gravimetric and Dust-Spot Procedures for Testing Air-
standing of the filtration mechanisms working in the applica- Cleaning Devices used in General Ventilation for Removing
tions and their interaction with the filter media structure Particulate Matter,” is a useful method for measuring the dust
selected can lead to creative and marketable solutions to fil- spot efficiency of filters, the relative arrestance values of low
tration problems. efficiency filters and dirt holding capacities (DHC) of all
Dust Holding Capacity characterizes the life of the filter types of filters. This standard defines Dust Spot Efficiency,
and hence, to a degree, the cost associated with operation. Arrestance and Dust Holding Capacity.
Fiberglass filter media are typically twice as thick as synthet- The Dust Spot Efficiency of a filter is its capacity to
ic media. The additional volume gives extra dust holding remove smaller diameter particles from the atmosphere. This
capacity. Also, the higher stiffness of glass gives greater is measured by comparing the opacity of glass media target
structural stability to the glass media. Glass filter media paper, upstream and downstream, of the filter under evalua-
retains its three-dimensional structure even as the pressure tion.
drop increases during the filter’s use. In contrast, the less stiff Arrestance is a parameter that measures the ability of a fil-
synthetic fibers are not able to resist the higher-pressure drop; ter to remove synthetic dust from the atmosphere. Arrestance
consequently the filter media collapses giving a more two- is indicative of the filter’s capability of removing coarse par-
dimensional structure. These two factors give glass media a ticles from the atmosphere. It is measured by feeding a
greater dust holding capacity and higher service life. known amount of ASHRAE Test Dust upstream of the target
The Filter Resistance to Air Flow is a measure of the ener- filter and comparing that with the weight gained by a HEPA
gy requirement and cost associated with using the product. filter placed downstream of the filter being characterized.
Synthetic filter media is more open and the, pressure differ- Dust Holding Capacity is a measure of the amount of
ential across it is lower when compared to glass media. ASHRAE Test Dust that a filter will capture until a specified
Initially, synthetic media requires less energy to maintain a pressure drop is reached. The basic test sequence is as fol-
particular airflow. Glass media renders greater resistance to lows:
the flow of air, and hence the fan consumes greater energy in 1) The pressure drop of a clean filter is measured at 50%,
maintaining equivalent airflow rates [5]. 75%, 100% and 125% of rated airflow.
Environmental condition is another important factor to be 2) Initial atmospheric dust spot efficiency is tested on a
considered in the design of filter media. The filter must be clean filter
designed to survive the temperature and chemical conditions 3) The filter is loaded with ASHRAE test dust at various
it will see in actual use. In most HVAC applications, the fil- intervals until a final pressure drop is reached or other condi-
ter is not exposed to temperatures much above that of ambi- tions are met. The dust spot efficiency and arrestance are
ent air. But in some installations, when the furnace cycles off, measured for each level of loading.
very high temperature air travels out of the furnace’s hot heat 4) At the end of the test, the average efficiency, arrestance
exchanger by convection and can reach the filter. The filter and dust holding capacity are calculated [5].
can melt if this temperature exceeds the melting point of the ASHRAE 52.1 test standard measures dust spot efficiency
fibers in the media. using atmospheric air, an uncontrolled test aerosol that does
not allow accurate repeatable comparisons among different
Laboratory and In Situ Testing Of Fiberglass laboratories and different manufacturers. The test also
and Synthetic Air Filters requires prolonged sequences that are influenced by outside
Over the past few decades, a number of laboratory test weather conditions. This test does not provide information
methods have been developed to measure and characterize air on the filter’s capability to remove particles of a particular
filters using synthetic dust. Initially, different countries tend- size range, information that is critical in some applications.
ed to develop their own test methods using different mea- This standard measures the average efficiency of an air filter
surement principles and synthetic test dusts. Today the ten- over its service life. In actual working, a filter’s performance
dency is more towards international or worldwide standards. is lower initially and then the efficiency improves over its ser-
In the U.S., the American Society for Heating, vice life. This test standard overestimates the performance of
Refrigerating and Air Conditioning Engineers (ASHRAE) a filter when it is new.
adopted test standards to characterize filter media. ASHRAE ASHRAE 52.2, “Method Of Testing General Ventilation
test standards have obtained approval from the American Air Cleaning Devices For Removal Efficiency by Particle
National Standards Institute (ANSI) as an American National Size”, is a significant step forward in filter testing and Indoor
Standard. Eurovent, the European Association for Air Quality (IAQ) control. It is designed not to replace the
INJ Fall 2002 47
earlier standard but to supplement it. This standard
provides a repeatable method for testing and measur- Table 4
ing air filter performance under controlled laboratory COMPOSITION AND PARTICLE SIZE DISTRIBUTION
conditions in terms of its fractional efficiency [6]. OF VARIOUS DUST PARTICLE TYPES.
This standard introduces the concept of fractional effi-
ciency. Fractional efficiency is defined as the charac-
teristic of a filter to remove known particle size frac- Type of Dust Particle Size Percentage by
tions from the atmosphere. Knowledge of fractional Ranges (µ µm) Weight
efficiency is vital in critical operations, such as the Atmospheric Dust 10.0-30.0 28%
manufacture of microelectronic devices, in order to 5.0-10.0 52%
facilitate the proper selection of filter media. Also, 3.0-5.0 11%
with increasing IAQ requirements, the ability of a fil- 1.0-3.0 6%
ter to remove the respirable particle size portions of 0.0-1.0 3%
atmospheric contaminants is becoming increasingly
important. Standard Air Cleaner Test Dust
The ASHRAE 52.2 test method uses laboratory- 0-5 39%
generated potassium chloride dispersed in air as the 5-10 18%
challenge aerosol, which yields more consistent 10-20 16%
results than the atmospheric test dust. After an initial 20-40 18%
efficiency measurement, the target filter is loaded 40-80 9%
with the test dust in five different cycles. Particle
counters both upstream and downstream of the target ASHRAE Test Dust SAE 72%
filter count particles in 12 different size ranges from Powdered Carbon 23%
0.3 µm-10 µm for each level of loading to give the No. 7 Cotton Linters 5%
fractional efficiencies for different particle size ranges
for different loading levels. Pressure drop across the
filter is also measured each time. From the above informa- mally be exposed to in real-life situations. The ASHRAE dust
tion, fractional efficiency curves for each particle size range is made up of much larger particles than those present in the
are obtained for incremental loading. From these sets of atmosphere; hence they load the filters rapidly, enhancing the
incremental loading fractional efficiency curves, a composite filter’s mechanical filtration capability. This loading is not
curve is developed that gives the filter’s minimum efficiency seen in actual performance in the atmosphere and hence the
at each particle size range. electrostatic synthetic filter performs poorly. In ASHRAE
The minimum efficiency composite values are averaged 52.2 test standard, potassium chloride particles are used
into three size ranges to group filters into three simple effi- which are different in nature from atmospheric particles. A
ciency classes: high, medium and low efficiency filters des- comparison of challenge dust particle distributions is given in
ignated as E1, E2 and E3 respectively. To target particles in Table 4.
the 0.3-1.0 µm range, an E1 efficiency filter would be For the test standards to closely reflect actual performance
required. To capture particles 1.0-3.0 µm in size, E2 efficien- of a filter, a challenge test dust similar to the particles in
cy filter will be needed. An E3 efficiency filter will trap par- atmosphere in terms of particle size distribution, shape and
ticles 3.0-10 microns in size (5). density is needed. Also, the tests need to take account of the
The main feature of the new ASHRAE 52.2 performance fact that atmospheric dust changes with time (season or even
standard is the Minimum Efficiency Reporting Value time of the day) and location (urban or rural). Thus, the chal-
(MERV) system. The minimum efficiency composite values lenge dust must be representative of all conditions.
averaged into three size ranges are used to determine the Another reason for the disagreement between laboratory
MERV, which ranges from MERV 1 (typically a low-effi- and real-life results was discovered from a fractional efficien-
ciency, throw-away filter) to MERV 16 (an over 95% effi- cy test performed on a synthetic charged media in the labora-
ciency filter under ASHRAE standard) (6). The MERV sys- tory. The media were loaded independently with a matrix of
tem makes it easy to compare filters at a glance and provides synthetic dusts of different size ranges with the objective of
a better yardstick in decisions involving choice of filters for determining the dependence of contaminant particle sizes on
different applications. filtration efficiency and to identify the mechanism of dendrite
The ASHRAE 52.2 test procedures and European test pro- formation. Test results showed that particles 1 to 3 µm in
cedures conducted in the laboratory have some limitations, diameter were the major particle size range, which built sig-
which need to be overcome for them to reflect the true per- nificant mechanical efficiency due to dendrite formation.
formance of filters in real life. The foremost difference is in SEM analysis of the filter media showed that the one loaded
the test dust used for the testing. The ASHRAE test dust used with 1 to 3 µm dust particles had dendrite formation very sim-
in the ASHRAE 52.1 test standard is of an entirely different ilar to one seen in real life operational filters [5]. A photomi-
nature from atmospheric dust, the dust that a filter would nor- crograph of ASHRAE test dust used in the laboratory tests is

48 INJ Fall 2002

 Health Concerns
In selecting a filter for a particular end-use,
one must look not only at performance aspects
but also health effects of the product. One con-
cern over the use of glass fibers in filter media
is their possible carcinogenic nature. Asbestos-
related lung diseases revealed the possible dan-
gers of inhaling foreign matter into the deep
lung, resulting in apprehension about using
Figure 10 Figure 11 synthetic vitreous fibers, including glass fibers
ASHRAE TEST DUST (2) ATMOSPHERIC TEST DUST (2) in filter media.
Kern and Harding [12] report that a great deal
shown in Figure 10. [2]. As can be seen, this dust has a very of scientific research has been done to investigate the injuri-
large particle size that clogs the media rapidly. In the synthet- ous effect of glass fibers. A study in the 1940’s on 27,000
ic media, loss in efficiency due to neutralization of charge is fiberglass workers exposed to fiberglass for more that 40
compensated by the large particles that block the voids in the years showed no cause and effect relationship between expo-
media, thereby giving a misleading picture of performance. sure and disease. Since the fiber is inhaled into the lungs,
As shown in Figure 11, atmospheric dust has a greater inhalation studies were considered to be more appropriate.
number of smaller size particles than the ASHRAE test dust. Studies on animals exposed to building insulation showed no
Since the finer particles in the 1 to 3 µm range are the main symptoms of lung-related disease.
contributors to dendrite formation, a period of time is The mechanism by which fibers in the lungs might cause
required for mechanical filtration efficiency to be improved disease is complex, but nevertheless, three key factors called
by loading. As aerosols neutralize electrostatic charges on the the three D’s have been identified that strongly influence the
media within a few weeks of service, there is a loss in effi- process. They are dimension, dose and durability. The fiber
ciency in synthetic charged filters in the early part of their needs to be of a minimum diameter to be respirable. A com-
service life. monly accepted figure is 3 mm. Glass fibers with diameter
Another factor that must be considered is that test condi- greater than 3 mm are considered to be harmless. Research
tions in the laboratory are controlled and are conducted for a has indicated that the longer and finer the fibers, the more dif-
reasonably short period of time. They are not subjected to the ficult it is to remove them out of the lungs by the natural body
environmental and temporal conditions that a filter would be mechanisms. So, fibers with a higher length-to-diameter ratio
subjected to in actual testing conditions. are more likely to cause lung disease. Also, the probability of
SINTEF Refrigerating and Air Conditioning, The Research fibers being inhaled into the lungs depends on the concentra-
Council of Norway and five filter manufacturers collaborated tion of the air being breathed. The greater the number of res-
on a project called “Long Term Tests of Air Filters in Real pirable fibers in the air and the longer a person is exposed to
Environment” [7]. In this work, the tests were done on three such air, the greater the risk factor. The European Union and
types of glass media and two types of synthetic charged the German Government have established standard tests to
media. Two filters were selected from each filter type. So in classify fibers as either carcinogenic or irritants. In North
all, ten filters were selected. All the filters were of EU 7 type. America, control steps and research work are underway to
The filters were mounted in a specially built test rig that was minimize the health risks of synthetic vitreous fibers (SVF’s)
mounted on the roof of the laboratory building. The filters and fiberglass. Since there could always be respirable fibers
were equipped with individual volume regulators to ensure in the media, the effort is to have fibers that are less durable
that identical and constant volumes of air flowed through all in the lung environment. A new biosoluble glass fiber, AF
filters for the entire test period of one year. Pressure drop, fil- 902, has been introduced. This material has passed the
ter efficiency, and the amount of dust accumulated were mea- European Union Biopersistance test and German
sured for the entire period of the test. Fractional efficiency Intratracheal test and has been reported to be performing well
testing was done in a separate test rig using both atmospher- in filtration applications.
ic dust and dioctyl sebacete (DOS) aerosols generated by a With regard to health concerns, the phenomenon of fiber
Laskin Nozzle aerosol generator. shedding from filter media has been studied. A scientifically
The long-term test results show that glass fiber filters main- based method has been developed to give quantitative results
tained a more or less constant degree of fractional efficiency on fiber shedding from organic fiber and fiberglass filter
throughout the test period. Efficiency of electrostatically media [4]. Early methods had problems with the detection of
charged synthetic media fell significantly right from the start the shed fibers and with the introduction of variability and
and did not improve much with service. From this work it can contamination in the sample by uncontrolled ambient air.
be stated that lifetime and in situ testing of filters clearly point This method overcomes these problems by using very clean
out the fact that the challenge dust and testing conditions air for the test by passing the test air through two 99.9% effi-
must correlate more closely to the environment to which a fil- cient High Efficiency Particulate Air (HEPA) filters in series,
ter would be exposed in actual use. before entering the test chamber. In order to provide for worst

INJ Fall 2002 49

 Table 5
PARTICLE COUNT ANALYSIS
FOR AMBIENT AIR (13)
Particle Size Particles/ft3 Particles/cm3
>5 3384 0.12
>3 6961 0.25
>1 19126 0.68
>0.5 152572 5.4
>0.3 437612 15.5
>0.19 1411000 49.8
Figure 12
CONFIGURATION OF TEST EQUIPMENT FOR
SHEDDING CHARACTERISTICS OF FILTER Table 6
MEDIA (13) PARTICLE COUNT ANALYSIS FOR HEPA
FILTERED AIR (13)
case release of shed particles, the test is performed under high Particle Size Particles/ft3 Particles/cm3
air flow rates across the test filter, about 35 cubic feet per >5 0.03 0.000001
minute, which is about 50% more than that encountered in >3 0.1 0.000004
commercial systems. A square foot area of the test filter is >1 1 0.0004
chosen as sufficient for minimizing the variation within the >0.5 4 0.00015
air filter medium. Scanning Electron Microscopy (SEM) >0.3 7 0.00025
Analysis and Particle Count Method are used to evaluate the 0.19 17 0.00060
system.
As shown in Figure 12, air enters the duct and passes
through the two HEPA filters, then passes through a sufficient Table 7
chamber length before it passes through the test filter in order SCANNING ELECTRON MICROSCOPY
to ensure laminar flow. The blower maintains a constant flow ANALYSIS FOR FIBER CHARACTERIZATION (13)
rate of 35 cubic feet per minute throughout the six-hour sam-
pling period. The air then passes through an eight inch mix- Product Manville AFS-3B2 Viledon MF95
ing orifice to ensure homogenous air for samples collected on Fiber Diameter
a 0.4 µm pore size nucleopore filter for microscopic sampling Average (µm) 1.0 2.0
and subsequent SEM analysis. As the concentration of the Fiber Diameter
fibers in the nucleopore filter is low, the test sample must be Range (µm) 0.2-8.0 0.3-24
concentrated in order to carry out an appropriate analysis. Fiber Length
Multi-channel Climet laser based airborne particle counters Average (µm) 25 22
are used to determine concentrations both in front and behind Fiber Length
the filter. Table 5. shows the number of particles in unfiltered Range (µm) 2-110* 3-65*
test air obtained by particle counts. * Some fiber lengths were greater than the field of view.
The HEPA filters removed 99.9% of the particles in the
ambient air to provide very clean air for the test as shown in present in ambient air, the number of fibers shed by the filter
Table 6. media is negligible.
SEM Analysis of the Manville fiberglass media (AFS-3B2)
and the synthetic (Polycarbonate/polyester) Viledon media Conclusions
(MF95), on the basis of viewing of 200 separate fields, yield- Test standards form the basis of selection of filters for dif-
ed the results tabulated in Table 7. ferent end-use applications. It is important to have standards
SEM analysis showed that both products shed an extreme- that will test products under controlled conditions and report
ly small number of fibers. Some of them were respirable on their performance so that both users and specifiers can
fibers with a diameter less than 3 mm and length-to-diameter compare products, predict their performance in operating
ratio greater than 3:1. Tables 8. and 9 summarize data on the conditions with reasonable certainty, and determine appropri-
average number of fibers shed per cubic foot and cubic cen- ate air cleaner efficiencies for specific situations. The studies
timeter of monitored air for the six-hour test period. and research work done on filter performance and testing
The analyses show that both types of filter shed fibers, with conclude that present standards need to be modified for labo-
some of them being respirable. Also the difference in fiber ratory-generated test results on filters to predict more closely
shedding between the two filter media is insignificant. The their real-life, in-use performance. Several factors need to be
test results also show that fibers shed from the media taken into account, such as variation in the nature of dust in
decrease with time. As compared to the contaminate particles different environments in terms of composition, particle size

50 INJ Fall 2002

 Results Versus Real Life Performance,” INTC-2000, INDA,
Table 8 Cary, NC, 2000.
FIBER SHEDDING EVALUATION BY 3. Homonoff, E., “Multilayered Materials For Filtration
SEM ANALYSIS (13) and Separation,” Nonwovens World, 2000, Vol 9, No. 6, pp
Manville AFS-3B2 Viledon MF95 43-47.
Fibers/ft3 20 9 4. Gustavsson, J., “Air Filters For Better Indoor Air
Fibers/cm3 0.0007 0.0003 Quality,” INDA, Cary, NC, 1998.
5. Gross, M. L., Arnold, B. D., “Air Filter Performance
Characterization,” INDA, Cary, NC, 1999.
Table 9 6. Thornburg, D., “Filter Testing and IAQ Control - Move
AVERAGE PARTICLE COUNT FOR Forward,” Proceedings of Heating, Piping and Air
AIR FILTRATION MEDIA (13). Conditioning, 1999, pp 54 -56, 95.
Particle size Manville Viledon 7. STF11-A95052, “Long - Term Tests Of Filters in a Real
AFS-3B2 MF95 Environment”, SINTEF Refrigerating and Air Conditioning,
Particles/ft3 Particles/cm3 June 1995.
>5 0.000021 0.0000013 8. Rodman, C.A., “Mechanisms of Filtration/Separation
>3 0.000077 0.000033 Which Take Place in Fibrous Nonwoven Media,” Index 87
>1 0.00033 0.00019 Congress, EDANA, Geneva, 1987.
9. Gregor E., Chiang, S-H., Smith, G., Rubow, K., Forster,
D., “AFS Short Course: An Introduction to Filtration,” AFS
and particle density and the effect of differences in testing Society, Falls Church, VA, 2001.
conditions on performance results of the filter media. Filter 10. www.clemson.edu/cucsm
media manufacturers should evaluate the use of fibers with 11. Gustavsson, J., “Air Filters for Ventilating Systems-
high specific surface area and deep-grooved channels. These Laboratory and In Situ,” International Nonwovens Journal,
irregular cross-section fibers have high shape factors and 1999, Vol 8, No. 2.
very high capillary surface areas, which provide greater par- 12. Harding, F. L., Kern, C. F., “A Biosoluble Glass fir Air
ticle capture and accumulation and hence improved filtration Filter Media”, INTC-2000, INDA, Cary, NC, 2000.
properties. 13. Shumate, M. W., Wilhelm, J. E., “Fiber Shedding
Characteristics Of Commercial Air Filtration Media,”
Bibliography Proceedings Of The Fifth International Conference On
1. Butler, I., Filtration Technology Handbook, INDA, Cary, Indoor Air Quality And Climate, Canada, 1990.
NC, 2000. 14. http://www.eurofiltec.com/diction.html
2. Kern, F. C., Jackson, L .F., “Review Of Ashrae Test 15. http://www.universal-silencer.com/PDFs/94-1368.pdf

Glossary of Terms

Absolute Rating (14): A term used to describe or define the The carbon or charcoal is produced by destructive distillation
degree of filtration. Various methods are used to determine of wood, peat, lignite, nut shells, bones, vegetable or other
absolute ratings, which are not necessarily interchangeable. carbonaceous matter, but must be activated by high tempera-
Generally absolute means 100% removal of solids above a ture steam or carbon dioxide, which creates a porous partic-
specified micron size ulate structure.

Absorption (15): The taking up of bulk material by another Adsorption (14): A natural phenomenon of a gas, liquid,
matter. Absorbent material extracts one or more substances vapor or fine particles being attracted and held on to the mol-
for which it has an affinity, and is altered physically or chem- ecular surface structure of a material. Not normally a
ically throughout the process. During absorbency one sub- reversible phenomena as absorption is.
stance penetrates into another.
Aerosol (15): A quasi-stable dispersion of small solid or liq-
Activated Carbon (14): Any form of carbon characterized by uid particles in air.
high absorptive capacity for gases, vapors or colloidal solids.

INJ Fall 2002 51

 American Society for Testing and Materials – ASTM (14): Direct Interception (14): The capture of relatively large par-
An organization in the United States setting standards for ticles near the surface of a filter material. The particles col-
quantitatively testing and measuring. lide with the filter structure without moving from streamline,
laminar flow.
ANSI (14): American National Standards Institute
DOS aerosols: Dioctyl Sebacete aerosols.
Angstrom (14): A unit of length abbreviated as A. Equals one
hundred millionth (10-8) of a centimeter or 0. 0001 micron. Downstream (14): Portion of the product stream, which
already passed through the system, or the portion of a system
Arizona (SAE, ISO) test dust (15): Standardized air cleaner located after separator/filter etc.
test dusts classified from natural Arizona dust and generally
referred to as SAE or ISO test dusts (old A.C. fine and A.C. Dust Spot Efficiency (9): The capacity of a filter to clean nor-
coarse test dusts) mal outdoor air. Average dust spot efficiency of the filter is
one of the factors used for filter classification.
Arrestance (9): The capacity of the filter to separate synthet-
ic dust. The average value of arrestance of the filter is one of Efficiency (14): Degree to which a filter will perform in
the factors used for filter classification. removing solids and/or liquids.

Brownian Movement (14): A natural phenomenon caused by Extractables (14): Substances that can and will leak out of a
small particles of similar mass to fluid molecules that are cartridge during filtration.
being bombarded by these molecules. In a liquid stream this
causes a random spiraling motion thus enhancing the filter's Fiber shedding (9): Particulate matter, which is flushed from
chances of removing the particle. the filter during the filtration process, which contaminates
the filtered fluid.
Capillary (14): A very thin tube. In filtration, the term is to
describe pores in a membrane. Filter (14): A term generally applied to a device used to
remove solid contaminate from a liquid or gas, or separate
CEN (14): European Committee for Normalization one liquid from another liquid or gas. A filter, as referred to
in the industry today, is limited to a device which removes
Chemical filters (9): Chemical filters are mainly adsorption solid contaminates only. A filter may be one of a number of
filters based on activated carbon, which, by means of chem- such types as replaceable cartridge, cyclone, edge, leaf, baf-
ical reaction, adsorb and retain gases, which are very diffi- fle, plate and frame, precoat, centrifuge. The term filter is
cult to separate. sometimes erroneously used to describe the media used
inside the vessel or filter case, but the correct use should be
Coarse filters (9): Filters made out of glass or synthetic plas- filter element, cartridge etc.
tic fibers like polyester, acrylic and polyamide and used for
separating mainly particles 5mm or larger in size with very Filtration (14): Removal of particles, normally solids, from
less influence of outdoor air. a fluid. These can be contaminants or valuable products.

Depth Filtration (14): Filtration of a fluid by passing it Fine Filter (9): Filters that are made mainly from glass fibers
through a deep filter material, providing a tortuous path with with an average diameter of 0.5-5.0 µm or of coarse plastic
many points for impingement of particles to occur. fibers, often in combination with an electrostatic charge.
Traditionally used in 'Packed Tower' type filters. Fine filters are defined according to the EN 779 as filters
which, when new, have a dust spot efficiency greater than
Diatomaceous Earth (14): Pre-historic sedimentary deposits 20%.
of fossilized diatoms. Used as a pre-coat material because
diatoms are non-compressible. Fractional Efficiency (15): The ability of a filter to remove
particles of a specified size, expressed as a percentage.
Differential Pressure (14): See Pressure Drop. The differ- Fractional efficiency is expressed as EFi = [(C1i-
ence in pressure between the upstream and downstream sides C2i)/C1i)]*100, where C1i = number of particles of the speci-
of a filter. fied size i in the upstream and C2i = the number of particles
of the specified size i in the downstream.
Diffusion (14): A natural phenomena of gas passing through
a liquid film in a membrane from the high pressure to the HEPA (high efficiency particulate air) filter (15): High effi-
low-pressure side. ciency normally refers to air filters that will remove more

52 INJ Fall 2002

than 99% of airborne particles that are in the size range of Particle Size Distribution (14): The size range and quantity of
0.1–0.3 µm in diameter. These particles are known as the particles, which are measurable in a fluid sample. Used
most penetrating contaminants. HEPA filters are sometimes to determine the micron rating of filters for a specific process.
described as HESPA (High efficiency submicron particulate
air) filters. Perlite (14): A siliceous volcanic glass, containing 2-5%
combined water, which allows for shattering by heat or pres-
HVAC (14): Heating, Ventilation and Air Conditioning. sure into a fine powder suitable for a pre-coat.

HVACR (14): Heating, Ventilation, Air Conditioning and Pores (14): A term used to describe the openings in a filter
Refrigeration. material normally a membrane.

IAQ (4): Indoor Air Quality. Pore size distribution (15): This is a measure of number of
pores in various groups of sizes.
IARC (12): International Agency for Research on Cancer.
Porosity (14), (15): A term used to describe a filter material’s
Inertial Impaction (14): The capture of medium sized parti- structure - sometimes known as a void volume. The ratio of
cles within the structure of a filter material. The particles col- voids to the total volume of material, for example, the ratio of
lide with the filter structure because they fail to negotiate the void volume to total cake volume. Also given as the ratio of
tortuous path and move out of laminar flow. the apparent to the true density and expressed as a percentage.

Interception (15): Dust particle deposition on a fiber or other Pressure Drop (14): Loss in applied pressure across a filter
collecting surface due to the size of the particles. This filtra- system or process.
tion mechanism is characterized by a dimensionless parame-
ter: particle size/fiber diameter. Re-entrainment (15): The process of rendering particles air-
borne again after they have been deposited from an air
IPCS (12): International Program on Chemical Safety. stream. For example, particles captured in a filter may be re-
entrained if the velocity through the filter is increased slight-
Isokinetic Sampling (15): Any technique for collecting air- ly or if the filter is subjected to increased vibration.
borne particulate matter in which the velocity of the air
stream entering the sampling probe is equal to that of the air Retention (15): The ability of a filter medium to retain parti-
passing around and outside that sampling probe. cles of a given size.

LCA (9): An LCA of a filter analyses the environmental effect Surface Filtration (14): Removal of particles on the outside
with reference to ecological effects, health effects and con- surface of a filter material.
sumption of resources.
SVF: Synthetic Vitreous Fiber.
Mechanisms of Filtration (14): The physical methods of
removing particles from a fluid. They are Direct Interception, ULPA (ultra efficiency particulate) (15): HEPA filters with an
Inertial Impaction and Diffusion. efficiency greater than 99.997% are termed as ULPA (ultra
efficiency particulate) or UHESPA (ultra high efficiency sub-
Medium (14): A term used to generally describe a filter mate- micron particulate air) filters
rial.
Ultrafiltration – UF (14): A pressure driven membrane fil-
Microfiltration (14): Filtration of particles between approxi- tration system operating in crossflow mode. Used to sepa-
mately 10 and 0.1 micron. rate macromolecules such as proteins and organic com-
pounds of molecular weight of 300 and over. UF operates at
MPPS (9): Most Penetrating Particle Size. This statistic is pressures between 1 and 8 atmospheres and generally sepa-
used as a measure of filtration efficiency in the CEN EN rates coarser materials than those removed by a Reverse
1822:1998 test method. MPPS is the particle that most fre- Osmosis system. — INJ
quently penetrates a filter medium.

Particle Size (15): The magnitude of some physical dimen-

sion of the particle. Unless the particle is a sphere it is not
possible to give its size uniquely by a unit of length. For non-
spherical particles the method of measurement must be spec-
ified.

INJ Fall 2002 53