Environmental Protection Technology Series
FILTRATION CHARACTERISTICS OF
GLASS FIBER FILTER MEDIA AT
ELEVATED TEMPERATURES
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-192
July 1976
FILTRATION CHARACTERISTICS OF GLASS FIBER
FILTER MEDIA AT ELEVATED TEMPERATURES
by
Dale A. Lundgren and Thomas C. Gunderson
Department of Environmental Engineering Sciences
University of Florida
Gainesville, Florida 32611
Grant No. 803126-01-0
Project Officer
Dr. Kenneth T. Knapp
Emissions Measurement and Characterization Division
Environmental Sciences Research Laboratory
Research Triangle Park, NC 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences
Research Laboratory, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S.
Environmental Protection Agency, nor does mention of trade names
or commerical products constitute endorsement or recommendation
for use.
li
-------
ABSTRACT
Particle collection characteristics of a newly developed,
high-purity "Microquartz" fiber filter media and a Gelman Type A
glass fiber filter media were evaluated over a range of tempera-
tures 20°C to 540°C (68°F to 1004°F), particle sizes 0.05 ym to
26 ym, gas velocities 0.5 cm/sec to 51 cm/sec, and particle vola-
tilities. Both types of high efficiency filters proved adequate
(>99.9% efficiency) for sampling nonvolatile particles over the
above variable ranges. Nonvolatile particle penetration decreas-
ed with increasing temperature and increasing filter loading.
The effect elevated temperature had on particle collection
characteristics was not a determining factor in application of
high efficiency filters. The main problems encountered in the
high temperature environment were filter holder leakage and vola-
tilization of gas-borne particles which passed through the filter
media.
111
-------
CONTENTS
Abstract i i i
Tables vii
Figures v i i i
I. General Introduction 1
II. Literature Review: Filtration of Nonvolatile
Aerosol s 3
A. Theory 3
B. Experimental Data of Others 10
III. Literature Review: Filtration of H2$04 13
A. Theory 13
IV. Description of Experimental Appartus and
Methods 20
A. General Experimental Arrangement 20
B. Aerosol Generation 22
C. Determination of Aerosol Size Distributions.24
D. Temperature and Velocity Measurements 25
E. Techniques to Determine Filter Efficiency...26
F. Filters and Filter Holders 29
V. Results and Analysis 31
A. Aerosol Size Distributions 31
B. Loading Effects on Filtration Efficiency....31
-------
C. Velocity Effects on Filtration Efficiency...34
D. Paricle Size Effects on Filtration
Efficiency 35
E. Temperature Effects on Filtration 35
Efficiency 35
F. Filter Holder Leakage 36
G. HUSO. Filtration at Elevated Temperature....37
H. Pinhole Effects on Filtration Efficiency 38
I. Comparison of Theory with Experimental Data.38
J. Experimental Errors 39
VI. Conclusions 41
Symbol s 43
References 45
Appendices
A. Conversion of SO. to H2SO. 50
B. "Microquartz" filter at 5000x magnification
(scanning electron microscope photo) 52
C. Gelman Type A filter at 5000x magnification
(scanning electron microscope photo) 53
VI
-------
TABLES
Number Page
1 Typical exhaust-gas composition from
coal-fired boiler. 54
2 Amount of H2S04 found in particulate
matter by various stock sampling
methods. 55
3 Size distribution of aerosols produced
in Collison atomizer with impactor
(water solvent). 56
4 Lui's data on performance of Collison
atomizer with impactor (water solvent). 56
5 Size distribution of experimental aero-
sols. 57
6 Effect of temperature on filtration
efficiency. 58
7 H2S04 distribution in sampling train. 59
8 Theoretical and experimental penetra-
tions of "Microquartz" filter media. 60
vii
-------
FIGURES
Number Page
1 Structure of "tree" on a metal fiber. 61
2 Effect of temperature on penetration (theoretical). 62
3 Effect of temperature on penetration (experimental). 63
4 Effect of temperature on collection mechanisms of
impaction, interception and diffusion. 64
5 Equilibrium conversion of S02 to $63. 65
6 Equilibrium conversion of S03 to H2S04 at 8.0
vol% H20 in flue gas. 66
7 Dewpoint and condensate composition for vapor mix-
tures of H20 and H2S04 at 760 mm Hg total pressure. 67
8 H2S04 dewpoint for typical flue gas moisture con-
centrations. 68
9 Variation of dewpoint with H2S04 content for gases
having different H20 contents. 69
10 H2S04 dewpoint obtained by various investigators. 70
11 Dewpoint as a function of H2S04 concentration. 71
12 Relation of dewpoint and S03 content of combustion
gases to sulfur content of oil. 72
13 General arrangement of test apparatus. 73
14 Clamp assembly for glass filter holder. 74
15 Effect of filter loading on penetration. 75
16 Method to determine penetration at 7.0 yg/cm^ loading. 76
17 Effect of velocity on penetration
("Microquartz" Filter). 77
-------
Number Page
18 Effect of velocity on pentration
(Gelman Type A Filter). 78
19 Pressure drop vs., velocity. 79
20 Effect of particle size on penetration
("Microquartz" Filter). 80
21 Effect of particle size on penetration
(Gelman Type A Filter). 81
22 Effect of temperature on penetration 82
23 Pressure drop vs. velocity at elevated tempera-
tures. 83
24 Effect of pinholes on penetration. 84
IX
-------
I. GENERAL INTRODUCTION
Glass fiber filter media is widely used for the
collection of participate matter from stack effluents.
This filtration media is both inexpensive and highly effi-
cient. It does, however, have a fairly high and variable
background of extractable impurities, which often inter-
feres with chemical analysis of collected particulate
matter. To overcome this problem of a high extractable
background, EPA supported development of a high-purity
filter media made from Oohns-Manvi11e 99.2% "Microquartz"
fibers. This development project, performed by A. D.
Little, Inc. , produced a filter with a low extractable
background suitable for stack gas sampling at temperatures
in excess of 500°C (932°F).
When hot stack gases are sampled with EPA Method 5
2
sampling train , a significant fraction of what may be
considered particulate matter is often found in the impin-
gers downstream from the filter. The efficiency of glass
fiber filter media has, therefore, been questioned. Further-
more, doubts about the effect of high temperature on filter
performance and about filter efficiency for volatile parti-
cles have arisen. To answer these questions EPA supported
this program to evaluate glass fiber filters, particularly
the "Microquartz" filter, over a reasonable range of tempera-
tures, velocities, particle sizes, and particle volatilities:
1
-------
Temperature Range: 20°C to 540°C (68°F to 1004°F)
Velocity Range: 0.5 cm/sec to 51 cm/sec
(=1 ft/min to 100 ft/min)
Particle Diameter Range: 0.05 ym to 26 ym
Particle Composition: Volatile and Nonvolatile
-------
II. LITERATURE REVIEW: FILTRATION OF NONVOLATILE AEROSOLS
A. Theory
1 . Introduction
The three factors which affect nonvolatile aero-
sol filtration are the dispersed particles, dispersing
gas, and fibrous filter. Characterizing the dispersed
particles are their size or size distribution, shape,
mass, electrical charge, and concentration. Charac-
terizing the dispersing gas are its velocity, density,
absolute temperature, pressure, viscosity, moisture
content, and composition. Finally, the fibrous filter
is characterized by its surface area, thickness, fiber
size or distribution of fiber sizes, filter porosity,
specific fiber surface area, fiber composition, electri-
cal charge, and surface characteristics.
Two phases may be distinguished in the filtration
process—the primary phase and the secondary phase. In
the primary phase, aerodynamic capture of particles occurs
in a clean filter, so the following assumptions are usu-
ally made for models of this phase:
a) Deposition of individual particles does not
influence filter efficiency.
b) Filter efficiency is time independent.
c) Any particle that touches a filter fiber is
retained.
-------
d) Particles are spherical.
e) All fibers have the same diameter.
f) Filter porosity is uniform.
g) Electrostatic, thermal, and gravitational
effects are negligible.
h) No phase transitions occur in the aerosol
during temperature changes.
In the secondary filtration phase,structural changes
occur as a result of particle deposition; therefore efficiency is
3
time dependent. So-called "secondary effects" appear,
consisting of deposition of particles upon one another,
dendrite formation, fusion and flushing of drops on the
fiber surfaces, capillary effects, loss of electrical
charge, clogging,and so on.
The solution to the basic problem of predicting
filtration efficiency is easier when regarded from the point
of view of the first phase. Relatively good results have
been obtained using primary'phase assumptions, while pub-
lished work including "secondary effects" has been empirical
in nature. The section which follows discusses the basic
collection mechanisms from the primary phase viewpoint
approach.
2. Filtration Efficiency Equations
Aerosol filtration by fiber filters involves the
three main collection mechanisms of diffusion, direct inter-
ception, and inertia! impaction. Diffusive capture of
-------
particles is described by the dimensionless parameter N^:
ND = D/DV [1]
where ND = Dimensionless diffusion parameter
"D = Diffusivity
D = Fiber diameter
V = Filtration velocity
Particle diffusivity is so related to absolute tempera-
ture and several other temperature dependent terms, that
diffusive capture increases with increasing temperature
(particularly for submicron particles). The following
equations represent these relations:
D" = CkT/Sirnd [2]
where I) = Diffusivity
C = Cunningham correction factor
k = Boltzmann's constant
T = Absolute temperature
n = Gas viscosity
d = Particle diameter
C = 1 + (X/d)[2.514 + (0.8)exp(-0.55d/X)] [3]
where C = Cunningham correction factor
X = Mean free path of gas molecules
d = Particle diameter
-------
X = n/(0.499pGc) [4]
where A = Mean free path of gas molecules
n = Gas viscosity
PG = Gas density
c = Average velocity of gas molecules
n = n0[(273.1 + G)/(T + GJKT/Z/S.l)' [5]
where n = Gas viscosity at temperature T
n0 = Gas viscosity at temperature 0°C
T = Absolute temperature, °K
G = Constant = 114 for air at 1 atm
PG = 1.293 X 10"3/[(1 + 3.67 X 10"3)H] [6]
where p~ = Air density, gm/cm
H = Temperature, °C
c = (8RT/TTM)0'5 [7]
where c = Average velocity of gas molecules
R = Universal gas constant
T = Absolute temperature
M = Gas molecular weight
Direct interception is defined by the temperature indepen-
dent parameter NR:
NR = d/D [8]
where NR = Dimensionless interception parameter
-------
d = Particle diameter
D = Fiber diameter
Inertial impaction is described by the Stokes1 number:
STK = Cppd2V/18nD [9]
where STK = Stokes1 number
C = Cunningham correction factor
p
p = Particle density
d = Particle diameter
V = Filtration velocity
n = Gas viscosity
D = Filter diameter
Gas viscosity is the factor most affected by increasing
temperature, rising with it. Hence, effectiveness of
inertial impaction as a collection mechanism decreases
with increasing temperature.
Efficiency of a filter mat is a function of NR, ND> and STK
and is usually calculated from collection efficiency of the
individual fibers comprising the mat. Many individual
equations have been developed and recently were reviewed
7 8
by Yen. One representative equation by Davies is:
Es = (0.16 + 10.9a - 17a2)[NR + (0.5 + 0.8NR)(NR + STK)
- 0.105NR(ND + STKT] [TO]
where E = Single fiber efficiency for diffusion,
interception and inertial impaction
NR = Dimensionless parameter for interception
Nn = Dimensionless parameter for diffusion
-------
STK = Stokes' number
a = Packing density = W/Lp_
F
W = Mass per filter face area
L = Filter thickness
p,., = Fiber density
r
Filter mat efficiency can then be related to individual
Q
fiber efficiency by the following equation :
EM = l-exp[-4EsW/TrpFD] [11]
where E^ = Filter mat efficiency
E = Single fiber efficiency for diffusion,
interception, and inertial impaction [1 0]
W = Mass per filter face area
PF = Fiber density
D = Fiber diameter
3. Secondary Processes in Aerosol Filtration
Filtration efficiency theories so far discussed are
based on particle capture by clean fibers. Actually, a
deposit builds up which may reduce the filter pore size and
increase filter pressure drop. Often, deposited particles do
not distribute themselves evenly over the surface of the
fibers, but form chain aggregates which act as collection
bodies and may capture particles more effectively than the
filter itself. This process has been described by Watson
and Leers and is illustrated in Figure 1.
Time variation of penetration and pressure drop of
8
-------
a filter during use depends upon the filter structure, fiber
shape, and nature of the aerosol. Several investigators,
3
including Radushkevich, have suggested that penetration
descreases exponentially with time according to the equation
P = P0e-bt [12]
where P = Penetration at time t=t
PQ = Penetration at time t=0
b = Constant under given conditions
4. Summary
The filtration theory presented explains the tempera-
ture effect on filter efficiency, which is defined as
efficiency dependence on temperature of the filtered gas
while all other conditions remain constant. However, the
theory rests on very limiting assumptions (Section IIA1),
which do not always hold, and does not allow for any "secon-
dary effects." Nevertheless, filtration theory does illustrate
both qualitatively and quantitatively that with only tempera-
ture increasing Figure 2):
1. Inertial deposition decreases due to the increase
in gas viscosity.
2. Direct interception is essentially
indenpendent of temperature.
3. Diffuse deposition increases due to greater
particle diffusivity at higher temperatures
(particularly for submicron particles).
-------
B. Experimental Data of Others
First e_t aK measured the efficiency of ceramic fiber fil-
ters [capable of withstanding temperatures up to 1093°C (2000°F)]
at temperatures of 21°C (70°F) and 760°C (1400°F). Values of
EM = 85% at 21°C (70°F) and EM = 82% at 760°C (1400°F) for fibers
with a D = 20 ym were determined. For D = 8 ym, the correspond-
ing values were EM = 98% and EM =91%. For all experiments, the
M rl
3
same conditions [d = 1 ym, V = 178 cm/sec, p = 6.4 g/cm ] were
maintained to keep the inertial mechanism of particle deposition
dominant. The efficiency decreased for the three filters with
increasing temperatures as predicted by theory.
Pich and Binek reported measurements of temperature char-
acteristics of a filter with D =1.2 ym from 20°C (70°F) to 200°C
(392°F)(Figure 3). NaCl cubic particles with each edge 0.2 ym
and p = 2.16 g/cm at V = 0.6 cm/sec were used. Under these
conditions, diffusion deposition predominated, and efficiency
increased from 94% [20°C (70°F)] to 98.5% [200°C (392°F)], re-
flecting filtration theory.
Dyment reported results of some glass fiber filter tests
with NaCl aerosol at temperatures up to 500°C (932°F). He
found glass paper shrinks at these temperatures, so he
10
-------
preheated the samples to avoid cracking. His paper does
not mention D, d, p , and V. However, at ambient tempera-
ture, he found penetration too low (<0.001%) to be regis-
tered. The temperature was gradually raised to 440°C
(824°F) when a penetration of 0.001% was measured. At
520°C (968°F) this increased to 0.09%. His conclusions
were as follows:
"In practice the effect of temperature and pressure
on filtration mechanism and performance has not been found
a determining factor in the application of high efficiency
filters. The main problems are the physical and chemical
effects of a high temperature environment on the materials
of construction of the filter which are manifested by
reduced mechanical strength and resilience of loss of
adhesion, leading to mechanical leakage and loss in
efficiency."
More recent high temperature filter efficiency tests
were reported by First. Heat-shrunk quartz fiber filters
were tested at temperatures up to 510°C (950°F) with a
polydisperse NaCl aerosol of 0.14ym mass median diameter
(mmd) at velocities (V) of 15 cm/sec to 30 cm/sec.
Average penetrations ranged from 0.032% to
0.078% with no consistent trend with respect to temperature.
Individual readings of NaCl penetration varied as much as
+ 50% about the mean during each test series on an identical
filter, so a trend would be difficult to detect.
11
-------
Thring and Strauss carried out filter efficiency
calculations for a filter with D = lOym, V = 25 cm/sec,
and a temperature range of 0°C (32°F) to 1600°C (2912°F).
Results for d = O.Olym, d = O.lym, d = 5pm (Figure 4)
show effect of temperature on inertial impaction, inter-
ception, and diffusion collection mechanisms.
12
-------
III. LITERATURE REVIEW: FILTRATION OF' H2S04 AEROSOLS
A. Theory
1. Introduction
A major problem in stack sampling is how to clas-
sify pases at duct condition which condense and/or react
in the filtering train and medium to form what may be con-
sidered particulate matter. One substance which falls into
this category is sulfur trioxide (SO^).formed in equilib-
rium with sulfur dioxide (SO^) when sulfur-containing
fossil fuels are burned. Up to 5% of the total sulfur in
the fuel is converted to S03, yielding from 5 to 50 ppm
1819
S03 in the flue gas. ' The S03 is in equilibrium with
water (H^O) vapor in the flue gas and, depending on tempera-
ture and gas component concentrations, various amounts of
sulfuric acid (H^SO.) vapor will be formed. This H2SO.
can be collected on filters and weighed as particulate.
The importance of establishing whether or not con-
densed SOo/HpSO. is to be considered particulate matter is
pointed out by a 36% average contribution of this material
to total measured particulate grain loading (oil-fired
20
boiler emissions) as reported by Jaworowski. As fly ash
emission levels are reduced by air pollution control equip-
ment, this amount of condensed SO^/hLSO. may equal or
surpass dry particulate contribution and could prevent
compliance under existing regulations.
13
-------
2. SOX, HgO and HgSO^ Equilibria in Flue Gas
Most sulfur in power plant flue gases appears as
S02 (Table I),21 with typical S03 levels ranging from 1.0% to
2.5% of the S02- However, as Figure 5 shows, the equilib-
rium constant for the reaction:
S02+l/2 02JS03
strongly favors the formation of SO., at temperatures below
about 540°C (1000°F). This graph was calculated from data
22
cited by Hedley. Kinetics of the reaction are unfavor-
able in the absence of a catalyst, but thermodynamical ly
the SO- concentrations could exist at levels much
greater than those normally encountered. Ratios of SO., to
23
SOo as high as 0.1 have been reported. Since formation
of SO., is controlled by catalytic effects as well as amount
of excess air present, concentration of S03 resulting from
combustion of a particular fuel can only be estimated in
absence of direct measurements.
Reaction between H^O vapor and SO, is given by:
Figure 6 shows equilibrium conversion of S03 to
H^SO, as a function of temperature for a typical flue gas
H20 vapor concentration of 8 vol%. Appendix A illustrates
the calculation method used in obtaining this curve. At
temperatures below 204°C (400°F), essentially all S03
present is converted to H2SO. at equilibrium. In contrast
14
-------
to formation of SO-, formation of H2SOa occurs rapidly in
the thermodynamically feasible temperature range.
\
3. Determination of H2$0. Dewpoint
Fly ash particles can influence the apparent dew-
point (saturation temperature of H^O^ in flue gas), but
one commits practically no error by neglecting the presence
25
of other gases and considering only the HpSO.-HpO system.
Thermodynamic analysis of the H^SO.-H^O-flue gas system,
ignoring fly ash effects, provides a theoretical basis for
predicting acid dewpoints and condensate composition from
vapor-liquid equilibria data.
26
Abel was the first to derive a relationship ena-
bling calculation of H2S04, H20 and S03 partial pressures
from enthalpy, entropy, free energy, and heat capacity
values. From his H2SO« partial pressures and Greenwalt's
H20 partial pressures over H2$04 solutions, H2$04-H20
dewpoint charts were prepared, Figures 7 and 8 . The range
of uncertainty indicated by Abel is on the order of 5°C
(9°F) at 10 vol% H20 vapor.
Information contained in Figure 7 can be used to
predict dewpoint temperature from an analysis of H2SO. and
H20 vapor content. If gas is cooled below its dewpoint,
1 condensate equilibrium concentration and mass can be ob-
tained. Condensate mass predicted from use of the dewpoint
chart is actually a prediction of the amount avai Table for
condensation. The actual amount of condensate depositing
15
-------
on a fiber or metal surface may differ from the chart pre-
diction because of mass transfer considerations.
As an example of the use of the chart, consider a
flue gas containing 10 ppm H^SO. and 10 vol% H^O vapor.
Condensation would occur at about 135°C (275°F), and conden-
sate composition at that point would be about 79 wt% H^SO
If the gas were cooled to 121°C (250°F), 85% of the H2$04
would be removed from the gas phase and an insignificant
amount of H20 vapor would also condense. Condensation,
therefore, follows the 10 vol% water line, resulting in a
condensate which would be the equilibrium composition of the
condensate at 121°C (250°F), assuming the vapor phase is in
equilibrium with the total liquid condensed. Composition
change of the liquid is small over the temperature interval
given in this example, ranging from 79 vol% at 135°C (275°F)
to 75 vol% at 121°C (250°F).
Large changes in HpO vapor content of flue gases
cause only slight changes in acid dewpoint. Variation of
dewpoint with HLSO. content of gases having different H^O
vapor concentrations is shown in Figure 9, where the range
from 0.5 vo~\% to 15 vol% H^O vapor changes dewpoint only
17°C (30°F) to 22°C (40°F) for the medium-to-high acid con-
centrations indicated.
In addition to the procedure based on calculated
partial pressure, a number of efforts have been made to
determine H^SO- dewpoints from instrumental and chemical
procedures. Figures 10 and 11 present results obtained for
flue gas dewpoints as a function of H2S04 content by various
16
-------
investigators. To make an exact comparison, all curves
should be for a gas of the same vol% H20 vapor. However,
reference to Figure 7 will indicate that a variation in H20 vapor
concentration from 7 vol% to 10 vol% can cause only about
1°C (2°F) to 2°C (4°F) change in dewpoint. Taylor's 29
results were obtained from an electrical dewpoint meter
which is inaccurate at low acid partial pressures. Lisle
30
and Sensenbaugh's data were obtained with a spiral con-
•3-1 or p c
denser. Dewpoint curves of Gmitro, Muller, Abel and
27
Greenewalt were based on calculated partial pressures.
In view of the difficulties with calculation based
on liquid phase thermodynamic properties and inaccuracy of
dewpoint meters at low acid partial pressure, the most reli-
able method of correlating H2$04 dewpoints with H^O and
H?SO. vapor concentration is the experimental condensation
method employed by Lisle and Sensenbaugh. Their data corre-
late.best with Muller's calculated dewpoints and are the
basis for ASME Power Test Code 19.10.
32
Rendle and Wilsdon have also published some data
on relation of the SO^ content of combustion gas and of gas
dewpoint to sulfur content of fuel oils> Figure 12 . Results
of several other investigators have also been plotted. The
type of oil, ash content, and combustion conditions differ
for the various sets of points. Although the plot of SO-
content shows considerable scatter, it is apparent that
with more than 0.5% sulfur in the oil, SO, content of the
gas does not increase proportionately to the fuel sulfur %.
17
-------
Figure 12 indicates the following:
1. There is a rapid initial rise in dewpoint
with the first increment of sulfur in the fuel.
For an estimated dewpoint of 38°C (100°F) with
no sulfur, an increase to 127°C (260°) (H2$04
dewpoint) is found with 1% sulfur.
2. There is a relatively small rise in dew-
point as sulfur in the fuel oil increases
from 1% to 6%.
B. Experimental Data of Others
21
Hillenbrand e_t a^L sampled flue gas from a coal-
fired boiler with quartz filters maintained at 205°C (400°F)
and 138°C (280°F) and found that filter, temperature significantly
affected the amount of H2$0. found on the filter. At 138°C
(280°F), 45% and 41% of the total H2S04 catch (total H2$04
mass in probe, filter, and impingers) were found on the
filter in two trials. At 205°C (400°F), 24% and 8% of the
total t^SO. catch were found on the filter in two trials.
The greater amount of H2$04 found on the cooler filter was
interpreted by them to mean:
1. A considerable portion of H2SO, collected
on the filter resulted from both condensation
and reaction of particulate with the S02 and
so3.
2. Condensation and consequent reaction is
favored at lower temperature.
20
Jaworowski sampled flue gas from several oil-fired
18
-------
boilers with three different sampling methods: EPA
Method 5 sampling train, ceramic thimble apparatus, and a
high-volume sampling system. In all three sampling meth-
ods, temperature of the filter was kept between 120°C (250°F)
and 150°C ( 300°F). His results (Table 2) show the magni-
tude of HpSO^ contribution to total particulate grain
loading ranged from 18% to 78%, and averaged 36% of the
total measured emissions.
19
-------
IV. DESCRIPTION OF EXPERIMENTAL APPARATUS AND METHODS
A. General Experimental Arrangement
The experimental apparatus provided capabilities
for: controlled aerosol generation, filter efficiency meas-
urement (by fluorescent, gravimetric, atomic absorption,
titrametric or microscopic counting techniques) and aerosol
size distribution determination.
1 . Nonacid Aerosols
The general arrangement of the experimental equipment
is shown in Figure 13. Air was supplied from the building's
compressed air system, filtered, and the stream split: one
portion was used to operate an aerosol generator and the
other to flow through an ion generator. The ion generator
produced a high concentration of ions to accelerate attain-
ment of a Boltzmann charge distribution on the aerosol in
the conditioning chamber. Part of the aerosol leaving the
conditioning chamber was exhausted and part heated in a
stainless steel coil located in an oven. Aerosol temperature
was measured with an iron-constantan (Type J) thermocouple,
whose output was registered on a potentiometer (an ice point
reference juction was used with the thermocouple). The
heated aerosol then was split into two equal streams. The
reference stream was piped out of the oven and consecutively
through a coil immersed in cooling water, a high efficiency
20
-------
particulate filter, a volumetric gas meter, a flow regulating
valve, and a vacuum pump. The test stream first passed
through the test filter, also in the oven, and then followed
a route identical to that of the reference stream. Test
filter efficiency was determined by using fluorescent, gravi-
metric, titrametric, and microscopic counting or atomic
absorption techniques, depending on the type of test aerosol,
to compare collected aerosol masses on the reference filter
and backup filter following the test filter.
2. H2S04 Aerosols
A special sampling train was used to experiment with
hLSCL aerosol. The train consisted of a heated box, to
keep a stainless steel coil and test filter at the desired
temperature; two Greenburg-Smith impingers filled with 100 ml of
80% isopropanol - 2Q% deionized water and immersed in an
ice bath (to collect any HpSO, that passed through the test
filter); a backup filter to catch any carry-over mist from
the impingers; a gas volumetric meter; a flow regulating
valve; and a vacuum pump. Details of how H^SO. was generated
are in the next section.
21
-------
B. Aerosol Generation
1. The Collison Atomizer
A Collison atomizer, as described by Green and Whitby
37-39
e_t a_l_. was used to generate aerosols in the 0.05 urn to
p
0.14 ym mmd range. The atomizer was operated at 2.1 kg/cm
(30 psig) pressure, which produced an 11 &/min aerosol stream
with =3 ym mmd particles. Flow was increased to 76 &/min. upon
dilution by the charge neutralizing stream from the ion generator.
Placing an external impactor with a cut size of = 2 ym in the line
reduced particle size to about 1 ym mmd.
Solute-solvent combinations of uranine in distilled water
and dinonyl phthalate in ethanol were used in the Collison atom-
izer. If generated droplets contained a dissolved solute, then
upon evaporation, the diameter of the aerosol is:
d = C1/3Dd [13]
Whrre: d = Particle diameter
C = Ratio of solute volume to solvent volume
plus solute volume
D. = Droplet original diameter
22
-------
Thus.'an 8 vol % solution nebulized with a mean droplet dia-
meter of 1 ym would produce a mean particle diameter of
0.43 ym.
Aerosol particles resulting from evaporation of the
liquid phase sometimes carried significant excess static charge,
which was effectively and quickly reduced by contact with
a stream of bipolar ions. '""^ The ionizer used to produce
the ion stream was operated at 3000 volts A.C. with an air
flow of 61 £/min. A conditioning chamber of approximately
60£ capacity allowed achievement of an equilibrium charge
distribution. Effectiveness of charge reduction was easily
observed by measuring aerosol transmission loss in a short
section of 0.63 cm O.D. Tygon tubing. For 0.14 ym
mmd uranine aerosol flowing at 1 &/min losses
amounted to about 20% per meter without the ionizer and less
than 5% per meter with it.
2. Spinning Disk Aerosol Generator
A spinning disk aerosol generator 36-38 was used
to generate aerosols in the 0.5 ym to:6.0 ym mmd range.
The disk's air motor speed, ionizer voltage, satellite air
flow, liquid feed rate, and all other variables were held
constant to produce a main droplet diameter of about 35 ym.
Equation 13 was again used to predict aerosol size. Differ-
ent size aerosols were generated by varying only solute
23
-------
concentration. Uranine ethanol was the solute-solvent com-
bination for all aerosol sizes generated with the spinning
disk.
3. ^$04 Aerosol Generation
aerosol was generated by evaporating 0.10 N
H2S04 in a stainless steel coil located in a 370°C (618°F)
oven. The ^$04 solution was metered at a constant rate
with a 50 ml syringe driven by a constant speed syringe
drive. A constant flow of ambient air was metered into
the coil used for evaporation of the ^504. It was essen-
tial to add the acid uniformly to prevent undue fluctuations
in the gas composition.
4. Other Methods of Aerosol Generation
For particles greater than 6.0 ym, Paper Mulberry
Pollen (= 13 vm mmd) and Ragweed Pollen (- 26 ym mmd) were
mechanically dispersed near the inlet to the filters with
a rubber squeeze bulb. Fly ash with a 5.6 ym mmd (ag =
2.09) was also dispersed with compressed air near the inlet
to the filters.
C. Determination of Aerosol Size Distribution
An electrical aerosol size analyzer was used to
measure aerosol size distribution in the 0.01 ym to 0.50 ym
24
-------
diameter range. Details of its operating theory and use
are given in two papers by Liu ert aJK , and in Thermo
Systems Operating Service Manual. ^2 Since the electrical
aerosol size analyzer gave a number median diameter (nmd),
the mmd was calculated from the usual formula:
Log(mmd) = Log(nmd) + 6.9 (log a )2 [14]
In this equation the standard geometric deviation (og) was
found by plotting particle size versus number data on loga
rithmic cumulative probability graph paper. Then a is
calculated from:
nmd at 84. 1% value
a
g nmd at 50.0% value -'
Aerosols larger than 0.50 pro diameter were sized
with a photomicroscope with incident illumi
nation, again from the above two equations to obtain mmd's.
D. Temperature and Velocity Measurements
Air stream temperature was measured with an iron-
constantan (Type J) thermocouple, with an ice bath refer-
ence junction and a potentiometer to record the thermocou-
ple millivolt output. Clean, filtered air was passed
through the system until the thermocouple reached equilib-
rium, so its response time was of minimal concern.
25
-------
Gas velocity through the filter was determined by
recording trial time, gas volume, temperature, ambient at-
mospheric pressure, vacuum pressureof the two volumetric
gas meters,and pressure drops across all three filters.
Gas velocity at the elevated temperature was then calcu-
lated from the following equation based on the Ideal Gas
Laws :
VF = (QM Tp PM)/(AF TM PF) [16]
where: Vp = Velocity through filter at T
Tp = Temperature of air stream
TjY| = Temperature of volumetric gas meter
PM - Atmospheric pressure - vacuum pressure
PP = Atmospheric pressure + filter(s) pressure
drop(s)
Ap = Filter area
Q|Y| = Volume at TM registered by volumetric gas
meter.
E. Techniques to Determine Filter Efficiency
1. F1uorescence
43 44
Fluorometric analysis ' was used for most of the
efficiency determinations, because it has excellent sensi-
tivity (* 0.5 yg/fc) and is easy to use. The procedure
followed to determine the % penetration of a test filter
was:
26
-------
a) Wash in distilled water and dry several Gelman
Type A (47 mm) filters to reduce their fluores-
cence background to an average of 0.5 yg/£.
b) Use these washed filters in the parallel filter
arrangement previously described.
c) Soak the uranine aerosol laden filters in 20 ml
distilled water for 12 hours to insure all the
uranine goes into solution.
d) Analyze the uranine leachate in a fluorometer
to determine the uranine concentration.
e) Calculate the % penetration:
% Pcnctration-Uran'"ie mass on backuP filter
Uranine mass on reference filter
x 100% [17]
The fluorescence of the uranine decreased to about
75% of its original value when exposed to 260°C(500°F).
However, since the decrease was the same for the reference
line and the test line, it did not affect the results.
Uranine was not used for temperatures over 260°C (500°F),
since its fluorescence is greatly reduced at these temperatures
2. Atomic Absorption
A Nad aerosol was used in place of uranine for
temperatures from 260°C (500°F) to 538°C (1000°F). First45
has demonstrated that the vaporization of NaCl at temperatures
27
-------
less than or about equal to 538°C (1000°F) does not repre-
sent a serious error in the determination of filter pene-
tration, especially since any error is on the conservative
side.
The NaCl leachates from the filters were analyzed
with, an atomic absorption spectrophotometer. The cali-
bration and operating procedures for the instrument were
46 47
carefully followed as decribed in the instruction manuals. '
For Na the practical working sensitivity proved to
be about 3.0 yg/£, much poorer than the 0.5 yg/£ sensi-
tivity of the uranine fluorometric analysis. Consequently,
longer test runs were needed to accumulate enough Na on the
backup filter for accurate analysis. Teflon filters (5 ym
pore size) were used in the reference and backup filter
holder because they had undetectable Na backgrounds.
Gelman Type A, Gelman Spectro Grade Type A, and Millipore
Membrane (0.8 ym pore size, white, 47 mm) filters all proved
to have Na backgrounds too high to be useful.
3. Mass Measurement
For the dinonyl phthalate aerosol, penetrations
at elevated temperatures were high enough for the aerosol to be
detectedby its mass alone. By weighing the filters before
and after the experimental runs, it was possible to determine
the mass of accumulated aerosol. A microbalance with a sensi-
tivity of 0.0001 g was used for the weighings.
28
-------
4. Microscopic Counting
The larger pollen aerosols were easily visible under
a stereo microscope, so penetrations could be determined by
counting the individual pollen particles. Millipore Membrane
Filters (0.8 ym pore size, black, 47 mm) were used as
backups to collect the few pollen particles that penetrated
the filter under test.
5. Ti tration
HpSO. masses in washings of the stainless steel coil,
leachates of the filters, and the impinger contents
were determined by titration with 0.01 NNaOH.
The NaOH titrant was standardized against 0.01 Npotas-
sium acid phthalate. Phenolphthalein indicator (0.05 wt %
in deionized water) was used.
F. Filters and Filter Holders
Two types of glass fiber filters suitable for high
temperature sampling were used as test filters. One was
the binderless Gelman Type A Glass Fiber Filter (47 mm),
manufactured .from microsize filaments of glass and treated
in a muffle furnace to remove trace amounts of organic
fiber content. This filter is acceptable for use in EPA
Method 5. It withstood 538°C (1000°F) but became brittle
at that temperature. The other filter used was a
29
-------
prototype "Microquartz" filter of 99.2% silica fibers,
by Johns-Manvi11e Company. It maintained its struc-
tural strength and flexibility even at 538°C (1000°F).
Two filter holders with different effective filter-
ing areas enabled the coverage of a wide velocity range
of 0.50 cm/sec to 51.0 cm/sec. A stainless
holder (Gelman No. 2220, 47 mm) was used for the higher
2
velocities. Its effective filtration area was 9.6 cm ,
and its Viton 0-ring and Teflon captive thrust
ring proved capable of withstanding 260°C (500°F). Remov-
ing these two parts for higher temperatures did not cause
the holder to leak. Two holes were drilled and tapped into
the holder for threaded stainless steel capillary tubing
for pressure taps. This holder style was also
used to hold the reference and backup filters.
A 10 cm Pyrex glass filter holder with
2
an effective filtration area of 64 cm was
used for the lower velocity range. The filter holder con-
sisted of two glass halves, fritted glass filter backing,
and a two-piece metal clamp with four bolts. The low melting
point of the Neoprene gasket prevented the filter holder
from being used above 260°C (500°F). Also, expansion of
the bolts at 260°C (500°F) caused the holder to leak ex-
cessively (Section VF). Therefore, small springs were
placed on each bolt between the locking nut and the metal
clamp (Figure 14) to prevent leakage.
30
-------
V. RESULTS AND ANALYSIS
Particle collection efficiency tests with Gelman
Type A, Gelman Type E, Gelman Spectrograde Type A, Mine
Safety Appliance Type 1106BH, and "Microquartz" glass fiber
filters all produced very similar results at room tempera-
ture. These filter media all have similar pressure drops
vs. flow rates filter masses per unit area, and fiber size
distribution (see Appendices B and C). Therefore, only the
Gelman type A filter was extensively tested for comparison
with the new "Microquartz" filter at high temperatures.
A. Aerosol Size Distribution
The aerosols generated with the Collison atomizer
with impactor were sized with an electrical aerosol size
analyzer. Their distributions (Table 3) were slightly more
AQ
polydisperse than those of LiJ (Table 4), who used a more
similar Col 1ison-Impactor generator. The size distributions
of the aerosols larger than 0.50 ym diameter were determined
with a photomicroscope using incident illumination (Table
5).
B. Loading Effects on Filtration Efficiency
Penetration of both the Gelman Type A and Micro-
31
-------
quartz filters declined significantly with aerosol loading
of only several micrograms per square centimenter. This
effect was noted at both 20°C (70°F) and 260°C (500°F)
(Figure 15) and was accompanied by a fi1ter-pressure-drop-
increase of less than 5%. A test to determine if the uran-
ine aerosol had an excessive charge which might have contri-
buted to this phenomenon proved negative. If the filter
was initially charged, the contribution to total filter
efficiency (due to filter charge) should have
dissipated with time as the charge was lost, thus increas-
ing penetration. This did not happen. So, the penetration
decrease was attributed to plugging of microsized holes by
the first few micrograms of aerosol and/or tree-like branch-
es formed by the initial aerosol deposit serving as parti-
cle collection surfaces. The latter proposition is supported
49
by Tomaides' findings that particles can form a rather
sturdy bridge about 10 particle diameters long.
A similar increase in efficiency well before an ap-
preciable filter-pressure-drop-increase was reported by Dorman.
He reported that a filter with an initial efficiency
of 99.50% on a heterodisperse aerosol of O.Gymmmd had a
final efficiency of 99.98%, while the filter pressure drop
changed from 2.50 cm to 2.75 cm HoO.
The time variation of penetration of a filter dur-
ing use naturally depends on the filter structure, fiber
32
-------
material and the nature of the aerosol. As previously dis-
cussed, Radushkevich suggested penetration decreases expon-
entially with time according to the equation:
P = PQe-bt [18]
Where: P = Penetration at time t=t
P = Penetration at time t=0
b = Constant under given conditions
From the data in Figure 15, the following b values were
calculated:
Filter Type t(min) Temperature (°C) b(min"')
Type A 100 21 (70°F) 0.0233
Type A 180 260 (500°F) 0.008
"Microquartz" 100 21 (70°F) 0.0240
"Microquartz" 180 260 (500°F) 0.005
With these b values, plots on semi-log graph
paper of P=P e"bt between the initial penetration and the
final penetration measurements compared favorably with the
experimental data.
To negate the decreasing penetration when comparing
filter penetrations, two consecutive tests were run on each
filter. The two measured penetrations were then plotted on
the normal axis of semi-logarithmic graph paper (Figure 16),
while the respective average loadings were plotted on the
logarithmic axis. Penetration at the arbitrary reference
33
-------
2
loading of 7yg/cm was then found by interpolation between
the two test penetration values.
C. Velocity Effects on Filtration Efficiency
Penetration of submicron aerosols was found to
increase with increasing velocity (Figures 17 and 18).
Similar penetrations for the 0.09vimmmd and 0.14pmmmd
aerosolswere not expected but may indicate the presence of
microsized holes in the filter media. Penetrations of less than
0.01% were found for aerosols greater than about O.Symmmd
for all filtration velocities tested. These penetrations
were so low that the aerosol measurement sensitivity was in-
adequate to determine larger particle velocity effects on
filter efficiency. This, in general, was also true for all
particle sizes at filtration velocities of less than 5 cm^ec.
The filter pressure drops increased linearly with
velocity (Figure 19), resembling the findings of Benson,
et al.. and obeying D'Arcy's Law:
Ap = XLnV [19]
Where Ap = Pressure drop
X = Permeability coefficient (a constant)
L = Filter thickness
n = Gas viscosity
V = Velocity
34
-------
D. Effect of Particle Size on Filtration Efficiency
Submicron aerosols were found to produce the great-
est filter penetrations (Figures 20 and 21). Similar pene-
trations for particles less than 0.14ymmmd cannot be easily
explained except for the possible presence of microsized
holes. Uranine aerosols greater than 0.8ym mmd, fly ash
aerosol of 1 .10ym mmd, a dinonyl phthalate oil aerosol of
about l.Oyrommd, and pollen aerosols of 13.1ymand 26.7 ym
mmd's all evidence penetrations of less than 0.01% at am-
bient temperatures.
E. Effect of Temperature on Filtration Efficiency
The effect of increased temperature was to decrease
aerosol penetration for nonvolatile submicron particles
whose penetration could be measured (Table 6 and Figure
22). Since submicron particles are collected mainly by dif-
fusion, which is more effective with increasing temperature,
the experiemental results qualitatively confirm the theory
previously outlined.
Volatile particles were not collected as effectively
at elevated temperatures as were the nonvolatile particles.
For example, a Gelman Type A filter was tested at 21°C
(70°F), 150°C (374°F), and 260°C (500°F) with a dinonyl
phthalate (DNP) aerosol, which vaporizes at 232°C (450°F).
The results at a 51 cm/sec filtration velocity
35
-------
were:
Le_!HE_.llCl ^Penetration
20 < 0.01
150 - 25
260 =100
Obviously,at 260°C (500°F) the DNP vaporized, passed through
the filter, and condensed upon being cooled below its dew-
point. DNP was found both in the cooling coil following the
test filter and on the backup filter.
These results imply that the definition of particu-
late matter is a function of the sampling method (i.e., the
temperature at which the sample is collected). Consequently,
if compliance tests are to be compared from source to source,
the same sampl ing method (i.e., sampling temperature) must
be employed if volatiles or condensables are involved.
Filter pressure drop also increased with increasing
temperature, due to the increasing air viscosity values at
elevated temperatures. This effect (Figure 23) reflects
D'Arcy's Law (Equation 19) in which pressure drop is direct-
ly proportional to both viscosity and velocity.
F. Filter Holder Leakage
A stainless steel filter holder (Gelman No. 2220,
47 mm)was used for most of the experiments and was air-tight
at temperatures to 538°C (1000°F). However, a Pyrex glass
filter holder held together by a two-piece metal clamp with
36
-------
four bolts leaked significantly at 260°C (500°F). The metal
bolts expanded, deforming the clamping device and creating
a relatively large air leak. This resulted in an apparent
50% to 60% aerosol penetration. This leakage problem was
corrected by spring loading the bolts of the metal clamp.
2
In EPA Method 5, a leakage check of the sampling
train, including the filter holder, is usually done with the
system at ambient temperature. From our findings, it would
seem necessary to bring the filter unit to the recommended
temperature of 120°C before making the leakage check. This
precaution would become even more necessary if the proposed
Method 5 filter box temperature maximum of 160°C (320°F)
is accepted.
G. H^SO. Filtration Efficiency at Elevated Temperatures
Experiments with FUSO. aerosol were conducted at
two temperatures [ 120°C (248°F), 205°C (401°F)],25 cm/sec
filtration velocity, 8.5 vol % HgO vapor,and 140
ppm (jJ5 ppm) F^SOa. At these concentrations of H^SO. and
H20 vapor, the acid dewpoint is about 170°C (338°F). The
results in Table 7 show at 120°C (248°F), below the H SO
dewpoint, most of the f^SO^was found in the coil and on
the test filter. At 205°C (401°F), above the H2SO. dew-
point, most of the H?SO. was found after the test filter
in the impinger contents and on the backup filter. Simple
calculations based on typical stack sampling data from oil-
37
-------
fired boilers and on this experimental data indicate HUSO,
could account for more than 50% of the total particulate
catch at a 120°C (248°F) sampling temperature (or at a tem-
perature below the HUSO, dewpoint in the stack gas), but
only for about 9% at a 205°C (401°F) sampling temperature
(or at a temperature above the I^SO^ dewpoint in the stack
gas).
H. Pinhole Effects on Filter Efficiency
Figure 24 shows how the penetration of a high ef-
ficiency glass fiber filter was affected by punching two
0.75 mm diameter pinholes through the filter mat. Note that
penetration is greater for the small aerosol than for the
large aerosol. The pinholes were clearly visible when
the filter was examined against an illuminated background,
so it is doubtful whether defective filters with pinholes
as large as these would pass unnoticed and be used for
sampling. These penetrations are small enough not to sig-
nificantly affect the outcome of a stack sampling test.
I. Comparison of Theory with Experimental Data
Equations 10 and 11 were used to calculate the
theoretical efficiency of the "Microquartz" filter, and
the results are in Table 8. Both the theory and data show
penetration to decrease with rising temperature. The
theoretical penetrations are only within two orders of
38
-------
magnitude of the experimental penetrations, being very de-
pendent on the filter fiber diameter and particle size and/
or size distribution. Work is presently being done by us
to improve the accuracy of several theoretical filtration
equations.
J. Experimental Errors
The largest variables in the experiments were the
changes in penetration from filter to filter. As an illus-
tration, the penetrations of ten Gelman Type A filters were
tested under identical conditions [T=21°C (70°F),
v=18 cm/sec, d=0.14ym mmd] using fluorometric analysis.
o
The penetrations ranged from 0.02% to 0.05% at 7.0 yg/cm
loading; the average penetration was 0.03% +_ 0.01%. Al-
lowing for the sensitivity of the fluorometric analysis
(0.5 ng/£ or 0.005% penetration for these trials), the fil-
ter-to-filter penetrations varied as much as 67% from the
average penetration for the ten trials.
The atomic absorption analytic method which was
used for the analysis of NaCl was not as sensitive as the
fluorometric technique. The sensitivity of the former was
3.0 yg/£, while that of the latter was 0.5 jjg/fc.' Hence, more
mass had to be accumulated on the test filter so that enough
aerosol penetrated to be detected accurately. The accuracy
of the atomic absorption method was j^ 0.01% penetration,
which was adequate, considering the Targe variability from
39
-------
filter to filter.
The smallest detectable H SO mass, with titration
of 0.01 NNaOH was about 45 yg. Since the H SO. mass col-
lected in the impingers, coil, or filters was about 6,000yg,
the error due to the titration was less than about 0.75%.
However, completely flushing all the H SO from inside the
coil with deionized water was difficult, so about a +_ 10%
error in the ^SO^ masses determined for the coil can be
assumed.
The accuracies of temperature and velocity measure-
ments were +_10% and +5% respectively. The sensitivity of
the mass determinations, using a microbalance, was +_0.0002 g.
This translates into an accuracy in penetration of +0.67%.
The mass method of determining filter efficiency was used
only for the dinonyl pthalate aerosol, which penetrated
100% at some high temperature conditions. Thus, +_0.67%
penetration was sufficient accuracy.
Due to the inherent errors in particle size measure-
ment by electrical mobility, the mmd's determined with an
electrical aerosol size analyzer were not better than +_20%.
About the same accuracy can be ascribed to the size distri-
butions made by counting the particles with a light micro-
scope. The size of the aerosols generated with the Collison
atomizer with impactor compared favorably with the size of
those generated by others (See Tables 3 and 4).
40
-------
VI CONCLUSIONS
It was experimentally demonstrated that Gelman
Type A and "Microquartz" high efficiency glass fiber
filters are adequate for sampling nonvolatile particles
at temperatures to 538°C (1000°). Submicron particles pen-
etrated more than did larger particles, and they penetrated
the most at the highest filtration velocity tested (51 cm/sec)
In all tests, however, the aerosol penetration was never
more than about 0.10%. Nonvolatile particles penetrated
less with increasing temperature and increasing filter
1oadi ng.
Particles with vaporization points below the samp-
ling temperature, including H^SO., can vaporize, pass
through the glass fiber filters, and then recondense when
cooled below their dewpoints. Therefore, the definition
of "particulate matter" must be based upon a prescribed
temperature. Hot stack gases sampled at different filter
temperatures should not necessarily be comparable. Partic-
ulate emission standards must involve a suitable reference
temperature to allow proper enforcement.
Filtration efficiencies calculated by theoretical
equations change dramatically with small changes in the
assumed average filter fiber diameter and/or particle size
(or size distribution) used in the calculations.
41
-------
Pinholes not visible to the naked eye do not
appear to affect the penetration of glass fiber filters
enough to significantly alter stack sampling results.
The effect of temperature on filtration of non-
volatile particles was an increase in the col-
lection of submicron particles with increasing tempera-
ture. The main problems encountered at elevated tempera-
tures were vaporization of volatile particles and mechani
cal leakage of the filter holder.
42
-------
SYMBOLS
Ap = Filter area
b = Constant (In Equation 12)
C = Ratio of solute volume to solvent volume plus
solute volume
C = Cunningham correction factor
c = Average velocity of gas molecules
d = Particle diameter
D, = Droplet original diameter
D = Fiber diameter
"D = Diffusivity
E» = Filter mat efficiency
E = Single fiber efficiency for diffusion, inter-
ception and inertia! impaction
G = Constant (In Equation 5)
H = Temperature, °C
k = Boltzmann's constant
L = Filter thickness
M = Molecular weight of gas
N = Dimension!ess interception parameter
R
ND = Dimensionless diffusion parameter
P Penetration at time t = t
PQ = Penetration at time t = 0
QM = Volume at TM registered by volumetric gas meter
R = Universal gas constant
43
-------
STK = Stokes1 number
T.. = Temperature of volumetric gas meter
T = Absolute temperature
Tp = Temperature of air stream
V = Filtration velocity
Vp = Velocity through filter at Tp
M = Mass per filter face area
X = Permeability coefficient
a = Packing density
Ap = Pressure drop
Pp = Fiber density
PQ = Gas density
Pp = Particle density
n = Gas viscosity
n0 = Gas viscosity at 0°C
A = Mean free path of gas molecules
PM = Atmospheric pressure - vacuum pressure
Pp = Atmospheric pressure + filter(s) pressure drop(s)
44
-------
References
1. Benson, A. L., Levins, P. L., Massucco, A. A., and
Valentine, J. R., Development of a High-Purity Filter
for High Temperature Participate Sampling and Analysis.
EPA-650/2-73-032, by Arthur D. Little, Inc., Cambridge,
Mass., Nov. 1973.
2. "EPA Standards of Performance for New Sources,"
Federal Register, 36. (247):24876 (1971).
3. Radushkevich, L. V., Izv. Akad. Nauk SSSR. ser. khim.
nauk, 3.:407 (1963).
4. Chen, C. Y., "Filtration of Aerosols by Fibrous Media,"
Chem. Rev., 55^:595-623 (1955).
5. Perry, J. H., and Chilton, C. H., Chem. Engineer's
Handbook, 5th Ed., McGraw Hill Book Co., Inc., New York,
p. 3-248 (1973).
6. Hodgman, C. D., Handbook of Chemistry and Physics,
Chemical Rubber Publishing Co., Cleveland, Ohio, p.
2205 (1962).
7. Yeh, H. C., A Fundamental Study of Aerosol Filtration
by Fibrous Filters. Univ. of Minn., Ph.D. Thesis (1972).
8. Davies, C. N., "The Separation of Airborne Dust and
Particles," Proc. Inst. Mech. Engng., ]_B:185 (1952).
9. Whitby, K. T. and Lundgren, D. A., "Mechanics of Air
Cleaning," Trans, of ASAE, 8(3):342-344 , 351-352 (1965).
10. Watson, J. H. L., "Filmless Sample Mounting for the
Electron Microscope," J. Appl . Phys. , ]_7:121-127 (1946).
11. Leems, R., "Die Abscheidung Von Schwebstaffen in
Fasernfiltern," Staub. 50:402-417 (1967).
12. Davies, C. N., Aerosol Science, Academic Press, New
York, p. 270 (1966).
13. Pich, J., and Binek, B., "Temperature Characteristics
of Fiber Filters," in Aerosols. Physical Chemistry and
Applications, Proc. First Nat. Conference on Aerosols,
p. 257-264, Czech. Akad. Sciences, Prague (1965).
45
-------
14. First, M. W. , Graham, J. B., Butler, G. M., Walworth, C.
B. , and Wanen, R. P., "High Temperature Dust Filtration,"
Ind. and Eng. Chem. . 48(4):696-720 (1956).
15. Dyment, J., "Assessment of Air Filters at Elevated
Temperatures and Pressures," Filtration and Separation.
p. 441-445, July/August (1970~T
16. First, M. W., "Performance of Absolute Filters at Temp-
eratures from Ambient to 1000°F," 12th AEC Air Cleaning
Conference, p. 677-702.
17. Thring, M. W., and Strauss, W., "The Effects of High
Temperature on Particle Collection Mechanisms, "Trans.
Instn. Chem. Engrs.. 4J_:248-259 (1963).
18. Danielson, J. A., Air Pollution Engineering Manual, Los
Angeles County Air Pollution Control District, Los
Angeles, Cal., p. 536, 1967.
19. Hemeon, W. C. L., and Black, A. W. J., "Stack Dust Sam-
pling: In-Stack Filter or EPA Train," J. Air Pol 1.
Control Assoc.. Vol. 22, No. 7, p. 516, 1972.
20. Jaworowski, R. J., "Condensed Sulfur: Trioxide Partic-
ulate or Vapor?" J. Air Poll. Control Assoc.. Vol. 23,
No. 9, p. 791, 1973.
21. Hillenbrand, L. J., Engdahl , R. B., and Barrett, R. E.,
Chemical Composition of Particulate Air Pollutants
From Fossil-Fuel Combustion Sources, Battelle Columbus
Laboratories, p. II-2, March 1, 1973.
22. Hedley, A. B., in The Mechanism of Corrosion by Fuel
Impurities (H. R. Johnson and D. L. Littler, editors),
Butterworth, London, p. 204, 1963.
23. Cuffe, S. T., Gerstle, R. W., Orning, A. A. and Schwartz,
C. H., J. Air Poll . Control Assoc. , Vol. 14, p. 353,
1964.
24. Snowden, P. N. and Ryan, M. H., "Sulfuric Acid Conden-
sation from Flue Gases Containing Sulfur Oxides," J.
Inst. Fuel, Vol. 42, p. 188, 1969.
25. Mueller, P., "Study of the Influence of Sulfuric Acid on
the Dew Point Temperature of the Flue Gas," Chemie-Ing.-
Tech. , Vol 31, p. 345, 1959.
26. Abel, E., "The Vapor Phase Above the System Sulfuric
Acid-Water," J. Phys. Chem., Vol. 50, p. 260, 1946.
46
-------
27. Greenewalt, C. H., "Partial Pressure of Water Out of
Aqueous Solutions of Sulfuric Acid," Ind. and Eng.
Chem., Vol. 17, pp. 522-523.
28. Matty, R. E., and Diehl, E. K., "New Methods for Deter-
mining S09 and S03 in Flue Gas," Power Engineering, Vol.
57, p. 87, Dec. , T953.
29. Taylor, A. A., "Relation Between Dew Point and the Con-
centration of Sulfuric Acid in Flue Gases," J. Inst.
Fuel , Vol. 16, p. 25, 194,2.
30. Lisle, E. S. and Sensenbaugh, J. D., "The Determination
of Sulfur Trioxide and Acid Dew Point in Flue Cases,"
Combustion, Vol. 36, No. 1, p. 12, 1965.
31. Gmitro, J. I., and Vermuelen, T., "Vapor-Liquid Equili-
bria for Aqueous Sulfuric Acid," Univ. of Cal. Radiation
Lab. Report 10866, Berkeley, Cal., June 24, 1963.
32. Rendle, L. K., and Wilson, R. D., "The Prevention of
Acid Condensation in Oil-Fired Boilers," J. Inst. Fuel ,
Vol. 29, pp. 372-380, 1956.
33. Taylor, R. P., and Lewis, A., "Sulfur Trioxide Formation
in Oil Firing," Proc. Fourth Inst. Congress on Industrial
Heating, Group II, Sec. 24, No. 154, Paris, France, 1952.
34. Flint, D., Lindsay, A. W., and Littlejohn, R. F., "The
Effect of Metal Oxide Smokes on the $03 Content of Com-
bustion Gases from Fuel Oils," J. Inst. Fuel, Vol. 26,
pp. 122-127, 1953.
35. Corbett, P. F., and Fireday, F., "The S03 Content of the
Combustion Gases from an Oil-Fired Water-Tube Boiler,"
J. Inst. Fuel, Vol. 26, pp. 92-106, 1953.
36. Green, H. L., and Lane, W. R., Particulate Clouds: Dusts
Smokes and Mists, E. & F. M. Spon. Ltd., London, 1957.
37. Whitby, K. T., Lundgren, D. A., and Jordan, R. C., "Hom-
ogeneous Aerosol Generators," Technical Report No. 13,
Cooperative Research Project, Univ. of Minn., Dept. of
Mech. Eng. and USPHS (USPHS Grant No. S-23 (C-4), Jan.,
1961.
38. Whitby, K. T., Lundgren, D. A., and Peterson, C. M.,
"Homogeneous Aerosol Generators," J. Air and Water Pol-
lution, Vol. 9, p. 263, 1965.
47
-------
39. Whitby, K. T., Generator for Producing High Concentra-
tions of Small Ions," Rev. Sci. Inst. . Vol. 32, p.
1351, 1961.
40. Liu, B. Y. H., and Pui, D. Y. H., "On the Performance
of the Electrical Aerosol Analyzer," Particle Technology
Lab. Publ. No. 237, Particle Tech. Lab., Mech. Eng.
Dept., Univ. of Minn., Oct., 1974.
41. Liu, B. Y. H., Whitby, K. T., and Pui, D. Y. H., "A
Portable Electrical Analyzer for Size Distribution Meas-
urement of Submicron Aerosols," J. Air Pol 1. Control
Assoc. . Vol 24. No. 11, Nov., 1974, p. 1067.
42. TSI Model 3030 Electrical Aerosol Size Analyzer Operat-
ing and Service Manual, Thermo-Systems, Inc., St. Paul,
Minn.
43. Manual of Fluorometric Clinical Procedures, G. K. Turner
Assoc. , Nov. , 1971.
44. Operating and Service Manual Model 110 Fluorometer, G.
K. Turner Assoc., April, 1971.
45. First, M. W., "Performance of Absolute Filters at Tem-
peratures from Ambient to 1000°F," 12th AEC Air Clean-
ing Conference, pp. 677-702.
46. Instruction Manual for Models 1100 and 1200 Atomic
Absorption Spectrophotometers. Varian Techtron Pty.,
Ltd., Melbourne, Australia, Sept., 1973.
47. Analytical Methods for Flame Spectroscopy. Varian Tech-
tron Pty., Ltd., Melbourne, Australia, Sept., 1972.
48. Liu, B. Y. H., "Methods of Generating Monodisperse Aero-
sols," Pub. No. 104, Particle Technology Lab., Mech.
Eng. Dept., Univ. of Minn., Feb. 8, 1967.
49. Tomaides, M., U. of Minn., Personal Communication (1967)
in: Billings, C. E., Wilder, J., Fabric Filter Systems
Study. Vol. I: Handbook of Fabric Filter Technology,
GCA Corporation, Report No. NTIS No. PB 200-648, 1970,
pp. 200-648, 1970, pp. 2-83.
48
-------
50. Dorman, R. G. , "Filtration," in Aerosol Science by C.
N. Davies, Academic Press, N.Y., p. 218, 1966.
51. Federal Register. Vol. 39, No. 177, Sept. 11, 1974, p
32853.
52. Yost, D. M. and Russell, H.,Jr., Systematic Inorganic
Chemistry. Prentice-Hall, Inc., N.Y., 336 pp. (1946).
49
-------
APPENDIX A
CONVERSION OF S03 TO
To answer the question of how much of the S03 in. flue gas is
present as H2SOi», consider the following equilibrium:52
H2SOi,(g)J S03(g)+H20(g)
* 0-ole. vapor/I)
+ 0.75 logioT - 0.00057T + 4.086
(over the temperature range 598
to 756*10
First calculate position of the above equilibrium at the highest
temperature of interest, 400 F (477 °K) assuming the above expression
for logioK is valid:
logioK=!0.480 + 2.009 - 0.272 + 4.086
= -4.657
so at 400"F K=0.22 x HT" =("A°^S?3)
(H2SOi»)
at 1 atm, 0°C, ideal gas = -0 ... m/£ = 0.0446 m/i
at 1 atm, 477°K ideal gas = (477) (22 40) = °'0256 m/Jl
In the typical exhaust-gas composition from a coal fired boiler
(Table 1) , H20 = 4%,
so H20 = 0.04 x 0.0256 = 1.024 x 10~3 m/5,.
Similarly, S03 = (3.0 x 10~5 x 2.56 x 10~2 - X) = (7.7 x 10~7 - X) moles/*
HaSOit = X moles/liter
0.220 x 10- = d.02 x 10-3)(7.7 x IP"7 - X)
A
X = 7.51 x 10~7; or 97.8% of the S0 is H$0.
50
-------
00
Similarly, at 300 F (422 K):
log K = 11.846 + 1.970 - 0.241 + 4.086
10
log K = -6.031
10
-6
K = 0.935 x 10
o 273
at 1 atm, 422 K, ideal gas is 0.446 x 4T2" = 0.289
-3
H 0 = 0.40 x 0.0289 = 1.16 x 10 m/Jl
2
H SO = X m/i
24
-5 -7
SO =3x10 x 0.0289 - X = (8.66 x 10 - X) m/l
3
-3 -7
-6 (1.16 x 10 ) (8.66 x 10 - X)
0.035 x 10 X
-6 -9 -3
(0.935 x 10 )X = 1.04 x 10 - (1.16 x 10 )X
-9
1 .004 x 10 -6
X = -3 = 0.866 x 10
1.16x10
-7
at 300 F H SO = 8.66 x 10 ; or * 100% of the SO is H SO .
24 324
Hence, SO essentially exists as H SO over the entire range
3 24
of conditions of interest, if the time is sufficient to allow
equilibrium of the hydration.
21
(Above from Hillenbrand, et al. )
51
-------
1 ym
"Microquartz" filter at BOOOx magnification (scanning
electron microscope photo).
52
-------
1 ym
Gelman Type A filter at 5000x magnification (scanning
electron microscope photo).
53
-------
Concentration
Component
H20
CO 2
Fly ash before
precipitator
Fly ash after
Volume
percent
4.0
15.0
(a)
g/m3
30
273
9.16
(4 gr/ft3)
0.458
precipitator
NO
S02
SO 3
Hydrocarbons
0.050
0.20
0.0030
(30 ppm)
(0.0010)
(0.2 gr/ft3)
0.63
5.3
0.10
(a) At 21°C (70°F), 1 atmosphere.
Table 1. Typical exhaust-gas composition from coal-fired
boiler. (From Hillenbrand, Engdahl, and Barrett21)
54
-------
Location
Plant A
Plant A
Plant A
Plant A
Plant B
Plant C
Plant C
H2S04
Filter ppm
Thimble1 8.1
8.1
10.8
9.5
9.9
9.5
9.5
9.5
Hi-volume1 6.9
6.0
8.8
8.1
EPA/APCO1 14.9
13.8
7.5
11.6
6.1
9.5
9.5
8.8
8.8
EPA/APCO1 11.0
9.9
9.9
EPA/APCO2 5.0
5.0
6.7
5.7
5.9
5.4
Hi-volume1 4.7
2.8
EPA/APCO2 3.5
4.2
H2S04
gr/SCF
0.0147
0.0147
0.0196
0.0172
0.0180
0.0172
0.0172
0.0172
0.0125
0.0109
0.0147
0.0050
0.027
0.025
0.0135
0.021
0.011
0.0172
0.0172
0.0159
0.016
0.020
0.018
0.018
0.0092
0.0092
0.0123
0.0104
0.0104
0.0098
0.0085
0.0050
0.0064
0.0076
Total
gr/SCF
0.0694
0.0344
0.107
0.0366
0.0645
0.0329
0.0688
0.0405
0.0548
0.0540
0.0255
0.0292
0.151
0.0321
0.0308
0.0388
0.0242
0.0403
0.0643
0.0659
0.033
0.033
0.077
0.0757
0.022
0.023
0.036
0.030
0.031
0.033
0.0111
0.0076
0.0212
0.0200
H2SO«.
%of
total
21
43
18
47
28
52
25
42
23
20
62
50
18
78
44
54
45
43
27
24
48
61
23
24
42
40
34
35
34
30
77
66
30
38
]BaCls precipitation.
*NaOH titration.
Table 2. Amount of H SO found in particulate matter by
24 20
various stock sampling methods. CFrom Jaworowski ).
55
-------
Solute Cone. (Wt.%) NMD(ym) MMD(ym)
1.00% Uranine 0.050 0.140 1.80
O'.l% Uranine '0.028 'o.089 1.86
0.01% Uranine 0,017 0.050 1.82
1.0% NaCl 0.023 0.061 1.77
Table 3. Size distribution of aerosols produced in Collison atomizer
with impactor (water solvent).
Solute Cone. (Wt.%) NMD(ym) MMD(ym)
1.0% Uranine 0.054 0.103 1.41
0.1% Uranine 0.028 0.049 1.49
0.01% Uranine 0.016 0.028 1.45
Table 4. Liu's data on performance of Collison atomizer with
impactor (water solvent). (From Liu48)
56
-------
Aerosol Description Generation Method NMD(ym) MMD(ym)
0.002% Uranine by Spinning Disk 0.75 0.79 1.14
wt. in ethanol
0.005% Uranine by Spinning Disk 1.06 1.10 1.12
wt. in ethanol
0.02% Uranine by Spinning Disk 1.43 1.49 1.12
wt. in ethanol
0.1% Uranine by Spinning Disk 2.68 2.84 1.15
wt. in ethanol
1.0% Uranine by Spinning Disk 5.82 5.98 1.10
wt. in ethanol
Paper Mulberry Mechanical 12.9 13.1 1.07
Pollen Dispersion
Ragweed Pollen Mechanical 26.2 26.7 1.08
Dispersion
Fly Ash Mechanical 1.10 5.61 2.09
Dispersion
Table 5. Size distributions of experimental aerosols.
57
-------
Aerosol
Uranine
en
oo
Uranine
Uranine
Uranine
NaCl
NaCl
NaCl
Uranine
mmd (ym)
0.09
0.09
0.14
0.14
0.06
0.06
0.06
0.79,1.10,
1.49,2.84,
5.98
Velocity(cm/sec)
51
51
18
*
18
6
6
0.5
0
51
51
18
18
6
6
0
0
51
51
Temp(°C)
21
120
260
21
120
260
21
120
260
21
120
260
21
120
260
21
260
21
260
21
260
21
260
21
260
537
21
% Pen. @ 7.0
yg/cm2 loading Filter Type
0.10
0.05
0.03
0.14
0.05
0.02
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.10
0.01
0.03
<0.01
<0.01
<0.01
<0.01
<0.01
0.02
<0.01
<0.01
<0.01
<0.01
Microquartz (Q)
Q
Q
Type A (A)
A
A
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Q
Table 6. Effect of temperature on filtration efficiency.
-------
vo
S.S. Coil at Test
Temperature
Filter at Test
Temperature
Two Impingers +
Backup Filter
120°C(1)
% of Total
H2SOi» Catch (2)
61
32
100 Total
205°C(1)
% of Total
H2SCK Catch (2)
11
81
100 Total
(1) HaSO.* Dewpoint =170°C
(2) Average of two trials
Table 7.
distribution in sampling train.
-------
Vel(cm/sec)
5.1
18.3
51.0
5.1
18.3
51.0
T(°C)
21
21
21
260
260
260
Theoretical - % Penetration
Experimental-
% Penetration
D=0.4 (1)
d=0.14 (2)
a =1.0
g
<0.01
<0.01
-------
Figure 1. Structure of "tree" on a metal fiber. (From
Davies }.
61
-------
direct interception
Figure 2. Effect of temperature on penetration (theoretical)
13
CFrom Rich ).
62
-------
£%
I wo
95
20 40 90
100
150
200
Figure 3. Effect of temperature on penetration (experimental)
13
CFrom Rich ). "
63
-------
100
50
20
10
o 1-0
g>
iH 0-5
•«-
UJ
0-2
0*1
C-05
0-02
I
400
800 1200
Temperature, °C
1600
— — Inertial impaction
Interception
Diffusion
Figure 4. Effect of temperature on collection mechanisms of
impaction, interception and diffusion. (From Thring and
17
Strauss }.
64
-------
I I I I I I
O
in
c
o
•H
a
H
o
O
u
W
100
do
80
70
60
50
40
* 30
20
10
600
1 I I I I ! I I
700
800 900 1000 1100
Temperature, °F
12'
Figure 5. Equilibrium conversion of SO to SO . (From
21 2 3'
Hillenbrand, Engdahl , and Barrett )
65
-------
100
o
CO
B*
90
80
a 70
c 60
5
a
8 so
>
40
I"
H
1, 20
M
10
I I I I
200
j I
j I
300
400 500
Taicperature, *?
600
700
Figure 6. Equilibrium conversion of $0 to H SO at 8.0
3 2 4
vol% H 0 in flue gas. (From Hillenbrand, Engdahl , and
2 21
Barrett )
66
-------
100
JIMttiMJV!!!;i!ii'!!!:il\*?!l!i!!!'\l-^^^
-Av. ^:.rr.!:v-..-:-t:::::..v^:/M.'i.\ ::p
.;\^T-I .;!itXinriH;i-!i:\"/;^:aTr:V:!!
6 8 10 20
Water Vapor, Vol %
40
60 60 100
Figure 7. Dewpoint and condensate compositi'on for vapor
mixtures of H 0 and H SO at 760 mm Hg total pressure.
26 2 2 4 27
[From Abel and Greenewalt )
67
-------
100
80
60
40
20
I 10
"
« 6
o
CO
220
0|p
<0
7
240
260 280
Dew Point, °P
300
320
Figure 8. H SO dewpoint for typical flue gas moisture
24 26 27
concentrations. (From Abel and Greenewalt )
68
-------
500
c
'6
o.
I
400
300
Water -vapor -
content of. gases,
200
QOOO 0.005 QOIO 0.015
H2 S04 in Dry Gas, %
Figure 9. Variation of dewpoint with H SO content for
24 28
gases having different H 0 contents. (From Matty, and Diehl )
2
69
-------
Figure 10. H SO dewpoint obtained by various investigators
2 4 26,27. 31 30 25
Abel and Greenewalt , Gmitro , Lisle , Muller , and
29
Taylor .
70
-------
i
3fU
350
330
310
290
270
250
230
210
ISq,
— — — — A.
p&
r.
0 E)
1
t
•
•
-
•
»
^i
t i i i
A. Toyior dewpoinl
Muller calculated
cperimental partia
pressure measurei
,/•
K
• iii
X
o _x
X
1 III
t meter
1
1
Tients
i
/
/
*j . r
/
i
t
i
i
i
i
i
i
i . 1 1 I
•
f
4
•
-
•
-
|
).OI 0.1 1.0 10 100
S03(H2S04) in Flue Gas, ppm
Figure 11. Dewpoint as a function of H SO concentration
29 25 24
CFrom Taylor and Mueller )
71
-------
m U.UUD
0)
3
O
J? 0.004
o
19 0.003
O
*o
| 0.002
"c
o
O
g 0.001
0.000
<
300
U.
•§ 200
Q.
0>
O
100
O
X
x^
X'
y\
>
f
I
1
%
V xX*1
x^^x^
X ^
\x E
Xvx
><^^r
•^^^
H Cor bet t
® Flint
^ Taylor c
All othe
Wilsdc
X
X*^-
\
^-
and Fired
ind Lewis
r points R
>n
v ^
<
^^ ^
X X
A 1M--f
•LJV" ^
ay
endle and
<•
C
"x
<
234
Sulfur Content of Oil,%
Figure 12. Relation of dewpoint and SO content of combustion
3 35
gases to sulfur content of oil. (From Corbett and Fireday
34 33 32
Flint et\aj_. , Taylor and Lewis , and Rendle and Wilson )
72
-------
CO
AEROSOL
GENERATOR
r
COIL
:N
'""'"I
i
r \
FILTER}
^ n i
U 1
l
WATER
BATH
rcoTlTI
1 AA '
i VV i
i i
1 COIL ]
1 AA !
• VV i
i i
REFERE
FILTEF
n >
U >
n >
u >
BACKUP
FILTER
TO
GAS METERS
AND
PUMPS
Figure 13. General arrangement of test apparatus.
-------
TOP VIEW
METAL t C
PLATES
BOLTS
NUTtC±=l
4 SPRING
FRONT VIEW
Figure 14. Clamp assembly for glass filter holder.
74
-------
en
0.20
0.18
0.16
| 0.14
£ 0.12
LJ
g o.io
Q.
S5 0.08
0.06
0.04
0.02
0.00
VELOCITY=5I cm/sec
AEROSOL MMD = 0.
URANINE DYE AEROSOL
A- GELMAN TYPE A-2I°C
B- GELMAN TYPE A-260°C
C-MICROQUARTZ-2I°C
D- MICROQUARTZ-260°C
TYPICAL ACCURACY FOR 95%
CONFIDENCE INTERVAL
I i I I I
10
FILTER LOADING - pg /cm2
100
Figure 15. Effect of filter loading on penetration.
-------
O.I2r
0.10
o.oe
or
LU
Z
UJ
CL
0.06
en
0.04
0.02
0.00
TEST I
I I llli
TEST 2
7 10
FILTER LOADING pg/cm2
100
Figure 16. Method to determine penetration at 7.0 pg/cm loading
-------
<
oc.
LJ
Q_
0.10
0.08
o.oe
0.04
0.02
0.00
FILTER -MICROQUARTZ
TEMPER AT URE-21 °C
FILTER LOADING-7.0 pg/cm2
A-AEROSOL MMD^ 0.09pm
B-AEROSOL MMD » 0.14 pm
C-AEROSOL MMD ^ 0.76 pm
TYPICAL 95% C.I
100
VELOCITY-cm/sec
Ptgure 17. Effect of velocity on penetration(nMicroquartz"Filter)
-------
00
o
6
or
LJ
Q.
0.14
0.12
0.10
0.08
0.06
0.04
002
0.00
FILTER-GELMAN TYPE A
TEMPERATURE-21°C
FILTER LOADING-7Opg/cm2
A-AEROSOL MMD* 009pm
B -AEROSOL M M D * 0.14 pm
C-AEROSOL M MD » 0.76pm
TYPICAL95% C.I
j I
I
10
VELOCITY-cm/sec
I I
i i i i
100
Figure 18. Effect of velocity on penetration (Gelman Type A Filter)
-------
or
LU
E
o
50
45
40
35
30
Q_
O
QC
0 25
LL)
QC
^
CO 2O
CO ^U
LU
a:
CL
15
10
A - GELMAN TYPE A
B- MICROOUARTZ
C-GELMAN TYPE A (I POINT)
D-MSA-II06BH (I POINT)
A
B
TYPICAL 95% C.I. FOR
MICROQUARTZ FILTER
10 20 30
VELOCITY cm/sec
40
_J
50
Figure 19. Pressure drop vs. velocity.
79
-------
FILTER-MICROQUARTZ
TEMPERATURE- 21° C
FILTER LOADING-7.0>ug/cm2
oo
o
0.10
O 0.08
0.06
LJ
LJ
CL
O.04
0.02
O.OO
A-VELOCITY =51 cm/sec
B- VELOCITY = 18 cm/sec
C- VELOCITY = 5 cm/sec
D- VELOCITY = 0.5 cm/sec
O.OI
TYPICAL 95% C.I
0.10
AEROSOL MMD-jum
i i I
I.O
Figure 20. Effect of particle size on penetration ("Microquartz'Tilter)
-------
00
I
UJ
z
UJ
Q.
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
FILTER-TYPE A
TEMPERATURE - 21 ° C
FILTER LOADING - 7.0pg/cm2
A - VELOCITY = 51 cm/sec
B - VELOCITY = 18 cm/sec
C- VELOCITY = 5 cm/sec
D - VELOCITY = 0.5cm/sec
TYPICAL 95 % C.I.
0.01
I I I
0.10
AEROSOL MMD-pm
1.0
Ftgure 21. Effect of particle size on penetration (Gelman Type A Filter)
-------
01
0.10
0.09
0.08
0.07
a: o.oe
LJ
z
UJ
0_ 0.05
0.04
0.03
0.02
FILTER - MICROOUARTZ
FILTER LOADING = 7.0 jug/cm2
AEROSOL MMD~0.09/im
URANINE DYE AEROSOL
A- VELOCITY" 51 cm/sec
8- VELOCITY =18 cm/sec
C- VELOCITY = 0.5 cm/sec
TYPICAL 95% C.I.
100 200
TEMPERATURE °C
300
Figure 22. Effect of temperature on penetration
82
-------
80
70
60
50
o:
UJ
E 4O
o
i
CL
O
cr
o
iu 30
a:
CO
LJ
cr
20
10
FILTER-MICROQUARTZ
A-538°C
B- 260°C
C- 120° C
D- 20°C
TYPICAL 95%C.I
10 20
VELOCITY
30
cm/sec
40
. 50
Figure 23. Pressure drop vs. velocity at elevated temperatures
83
-------
2.2
2.0
1.8
1.6
1.4
- A«l
>
—
—
_ O.IC
PEh
_ SAG
UJ
1.2
1-0
0.8
0.6
0.4
0.2
0.0
0
Fl LTER - GELMAN TYPE A
TEMPERATURE-2I°C
2 HOLES-0.75mm each
VELOCITY = 51 cm/sec
A-AEROSOL MMD*O.I4pm
B-AEROSOL MMD*2.6*jm
TYPICAL 95% C.I.
t
PENETRATION UNDER
SAME CONDITIONS
EXCEPT NO PINHOLES
0.75mm
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
i. REPORT NO.
EPA-600/2-76-192
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
FILTRATION CHARACTERISTICS OF GLASS FIBER
FILTER MEDIA AT ELEVATED TEMPERATURES
5. REPORT DATE
July 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Dale A. Lundgren and Thomas C. Gunderson
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Environmental Engineering Sciences
University of Florida
Gainesville, Florida 32601
10. PROGRAM ELEMENT NO.
1AA010
11. CONTRACT/GRANT NO.
R803126-01
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory
Office of Research and Development
United States Environmental Protection Agency
Triangle Park. North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 7/74 - 7/75
14. SPONSORING AGENCY CODE
EPA - ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Particle collection characteristics of a newly developed, high-purity "Micro-
quartz" fiber filter media and a Gelman Type A glass fiber filter media were evalu-
ated over a range of temperatures (20°C to 540°C), particle sizes (0.05 ym to 26 ym),
gas velocities (0.5 cm/sec to 51 cm/sec), and particle volatilities. Both types of
high efficiency filters proved adequate (>99.9% efficiency) for sampling nonvolatile
particles over the above variable ranges. Nonvolatile particle penetration decreased
with increasing temperature and increasing filter loading.
The effect elevated temperature had on particle collection characteristics was
not a determining factor in the application of high efficiency filters. The main
problems encountered in the high temperature environment were filter holder leakage
and volatilization of gas-borne particles that passed through the filter media.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
*Air pollution
*Aerosols
*Filtration
Filter materials
*Ceramic fibers
*Temperature
13 B
07 D
13 K
11 E
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
95
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
85
------- |