EPA-600/2-77-197
September 1977
Environmental Protection Technology Series
EVALUATION OF FOAM SCRUBBING
AS A METHOD FOR
COLLECTING FINE PARTICULATE
Industrial Environmental 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 Protec-
tion Agency, have been grouped into nine series. These nine broad categories were
established to facilitate further development and application of environmental tech-
nology. Elimination of traditional grouping was consciously planned to foster technology
transfer and a maximum interface in related fields. The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY
series. This series describes research performed to develop and demonstrate instrumen-
tation, equipment, and methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the new or improved tech-
nology required for the control and treatment of pollution sources to meet environmental
quality standards.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved for
publication. Approval does not signify that the contents necessarily reflect the views and
policies of the Government, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/2-77-197
September 1977
EVALUATION OF FOAM SCRUBBING
AS A METHOD FOR
COLLECTING FINE PARTICULATE
by
Geddes H. Ramsey
U.S. Environmental Protection Agency
Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, N.C. 27711
ROAP No. 21ADL-007
Program Element No. 1AB012
- Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
CONTENTS
Paqe
Figures iv
Tables v
Metric Conversions vi
Sections
I Conclusions 1
II Recommendations 3
III Introduction 4
IV Experimental Work 6
V Results and Discussion 17
VI Foam Scrubber Economics 29
VII References . 33
Appendices \
A The Metallizing Spray Gun as a Fine Particulate A-l
Generator
B Andersen and Brink Impactor Data B-l
C Entrainment Problem Due to Foam Destruction C-l
m
-------
FIGURES
No. . Page
1 Perforated Plate Foam Generator 5
2 Ssray Foam Generator 5
3 Foam Scrubber System 8
4 Nozzle Arrangement for Spraying Surfactant on Screen 10
5 Extra Chamber Added to Increase Residence Time 10
6 Foam Destruction Disk 12
7 Foam Destruction Chamber 12
8 Collection Efficiencies Based on Mass Filter Determinations 18
9 Inlet and Outlet Concentration Distributions for 40 Second 20
Residence Time
10 Inlet and Outlet Concentration Distributions for 53 Second • 21
Residence Time
11 Irlet and Outlet Concentration Distributions for 74 Second 22
Residence Time .
12 Irlet and Outlet Concentration Distributions for.112 Second ' 23
Residence Time , • ,'
13 Penetration Versus Particle Size for Various Residence 24
Times '
14 Grain Loading of Scrubber Outlet Versus Residence Time, 26
Based on Impactor Runs
15 Inlet and Outlet Concentration Distributions for Water . 27
Spray Only
16 Penetration Versus "Particle Size for.Water Spray:.Test 28
17 Comparison of Various Control Devices for Fine Particle 32
: Control . > .: •;••.-
Al Pa^ticulate;Generating System . - > A-2
A2 TOD View of Electric Arc Area A-2
A3 Distribution Curves for Metallizing Spray Gun Tests Based A-5
on Voltages
A4 Distribution Curves for Metallizing Spray Gun Tests Based A-6
on Air Flow Rates
A5 Distribution Curves for Metallizing Spray Gun Tests Based A-7
on Wire Feed Rate
A6 Outlet Grain Loading Versus Voltage for Iron Oxide A-8
A7 Outlet Grain Loading Versus Air Flow for Iron Oxide A-8
A8 Grciin Loading Versus Wire Feed for Iron Oxide A-9
IV
-------
TABLES
No. Page
1 Residence Times and Air Flows for the Foam Scrubber Tests 15
2 Screen Areas and Liquid Flow Rates for Various Size Foam 30
Scrubbers
3 Operating Cost for 50,000 ACFM Foam Scrubber at 99% Recycle 31
Al Experimental Conditions for Wire Spray Gun Tests for Iron A-3
Oxide
A2 Test Data for Aluminum Oxide and Iron Oxide for Spray Gun A-10
Tests
-------
METRIC CONVERSIONS
To Convert Fram To Multiply By
foot meter . 3.05 x 10"
foot2 meter2 . 9.29 xjO"2
foot3 ; meter3 • -., 2.83 x 10"2
o
foot/min (fpm) - , meter/sec 5.08 x 10
ft/min (cfm) meter3/min 2.83 x 10~2
gallon ,. , - , liter , ;3-78
inch cm 2.54 .
x 1(
,2
grains/ft3 mg/m3 2.29 x 10"3
Ib/min g/min 4,54 x 10
2 2
Ib/in (psi) newton/meter 6.89 )
gallon/min (gpm) liter/min 3.78
-------
SECTION I
CONCLUSIONS
Results of testing a foam scrubbing apparatus indicate that foam
scrubbing is a rather costly method for removing fine particulates from
a gaseous stream. The apparatus consisted of a foam generator, a residence
chamber, and a foam destructor. The foam was generated by passing a
particulate laden gas stream through a screen wetted by a spray containing
a surfactant liquid. The foam was destroyed by impacting it against the
chamber walls, using a high-speed rotating disk. The method of generating
and destroying foam was uncomplicated on a small scale, but problems
could arise in designing a large scale scrubber.
The rate at which foam could be produced was limited by the velocity
of the gasebus stream passing through the wetted screen. If the velocity
were too high, the gas stream would pass through the wetted screen
without generating foam, thus creating large voids or air pockets.
Since the voids contained the dirty gas, the overall collection efficiency
would decrease if the air pockets persisted. A good quality foam could
not be produced at an air velocity over 15.5 m/min on the foam scrubber
tested in this experiment.
The particle collection was experimentally verified for a fine
particulate of iron oxide (Fe203). The bubble size for the tests was
approximately 1.0 mm. To obtain a 95 percent collection efficiency in a
foam produced with Tergitol surfactant, a residence time approaching 2
minutes was required.
o
A process flow rate of 100,000*acfm (2832 m /sec), and a residence
time of 2 minutes would require a 1.5 million gallon (5.7 million liter)
chamber volume to contain the foam. This high residence time would
greatly increase the capital costs. In order for foam scrubbing to be
competitive, in terms of capital investment with other conventional
collection devices, the residence time must be less than 1 minute;
however, at 2 minutes the capital cost would be only slightly higher
than for an electrostatic precipitator, the most expensive conventional
device.
*EPA policy is to use SI units only or to list both the common British Unit
and its metric equivalent. For convenience and clarity, nonmetric units
are used in this report. Readers more familiar with metric terms may use
the conversion factors on page vi to convert to that system.
-------
Although the capital cost may be competitive, the operating costs
for foam scrubbing are likely to be much higher than for conventional
devices. The higher operating costs are due mainly to the cost of the
necessary surfactants. Even with a recycle of 99 percent of the surfactant
liquid, foam scrubbing is an order of magnitude higher in operating cost
than the most, expensive conventional device.
-------
SECTION II
RECOMMENDATIONS
Results of testing a small scale foam scrubber indicate that its
operating costs will not be economical on a large scale. Although the
capital cost of foam scrubbing will be close to that of conventional
devices, its operating costs will be too high for the collection efficiencies
obtained.
To be a viable method of fine particle control, the foam must be
generated at a much higher rate and the quantity of surfactant used must
be reduced considerably. Increasing the foam generation rate would
reduce the overall size of the screen and would reduce the capital cost;
however, any Ap increase will also increase operating costs. Increasing
the surfactant recycle or reducing the quantity of surfactant used would
reduce operating costs.
-------
SECTION III
INTRODUCTION
Although various air pollution control devices are available to
collect part'culates, many of the devices are inadequate for removing
submicron particles; those which do an adequate job consume large amounts
of energy, require special maintenance, or :are very large. Although
available control devices offer mass removal efficiencies of over 99
percent, even these do 'not capture much of the fine particulate which
passes into the atmosphere, creating air pollution problems.
One concept which shows promise in solving the problem of fine
particulates is foam scrubbing. Foam scrubbing has been shown to
effectively control submicron aerosol emissions on a small scale; however,
a foam scrubber scaled up for industrial applications has not been
tested. ' f ; :
Several different methods for producing foam for foam scrubbing are '
available, "he foam can be generated by forcing gas through a'liquid
phase on a perforated plate as shown in Figure 1. The foam can also be
generated by a nozzle arrangement such as used for producing foams with
firefighting equipment. For this experiment, the foam was produced by a
spray foam generator, with the foam generated in the gaseous phase. The
apparatus is shown in Figure 2.
This program attempted to experimentally determine the feasibility
of foam scrubbing as a viable method for removing submicron particles
from gaseous streams and to assess the preliminary economics of foam
scrubbing.
-------
d.
iam.
»'.' " '•'(' V '•'
I ,• «f '» •
rtl"1,
AIR
Figure 1. Perforated plate foam generator.
AIR
FOAM
SOLUTION
'SCREEN
Figure 2. Spray foam generator.
-------
SECTION IV
EXPERIMENTAL WORK
The three basic mechanisms for cpaturing particles inside a foam
bubble are: (1) sedimentation, (2) inertial impaction, and (3) diffusion.
For submicron particles captured by the foam, diffusional capture and
2
sedimentation (for particles greater than 0.1 ym) are the most evident.
Not accounting for variables which would increase the random motion
of the particles, the variables which would probably affect the chances
of the particle striking the liquid interface of the foam and being
captured are:
• Bubble size
• Residence time
• Foam stability
• Particle Size
Decreasing the bubble size should increase particle collection.
Reducing the bubble diameter by a factor of 2 reduces the area of the
2
liquid interface (Area = 4 irr ) by a factor of 4, but also decreases the
3
volume of ga; inside the bubble (V = 4/3 IT ) by a factor of 8. Thus,
halving the Dubble diameter will give the same volume of gas a collecting
area 6 times as great.
A = 4 Ttr
.V 4/3 nr3
As the bubble size decreases, the time for capture is reduced since
the average distance"from the particle to the bubble wall is reduced.
The particles inside the bubble are proportionally closer to the wall;
so if the bubble diameter is halved, the capture time is also halved.
Therefore, residence time is influenced strongly by bubble size and a
reduction in bubble size should result in a shorter residence time in
the scrubber.
-------
By increasing residence time, the particles have more time in which
to be captured by the foam. The foam must not degrade while in the
chamber or efficiency will decrease. Degradation of the foam results in
a larger bubble size and less efficiency.
Earlier work has shown that diffusional capture is theoretically
1 2
the predominant mechanism for particle capture inside a bubble. '
Basically, the diffusion of particles is caused by Brownian motion, the
random motion of particles colliding with surrounding molecules. If the
particle size decreases relative to the gas molecules, the velocity of
the particle after each collision with a gas molecule must increase. A
decrease in the particle size will result in an increase in the diffusion
rate and the particle should strike the wall sooner than a larger particle.
The diffusional coefficient of a 0.01 ym diameter is about 2,000 times
2
larger than that of a 1.0 ym particle.
The surfactant which was selected to produce the foam for this
experiment was Tergitol TMN-6 manufactured by Union Carbide. The
bubbles of the Tergitol-produced foam did not break appreciably during
the several minutes residence time that the foam was inside the scrubber.
The bubble size appeared to remain constant through the cycle.
Experimental Device
The experimental device to investigate the performance of foam
scrubbing is shown in Figure 3. The particulate was produced with a
metallizing spray gun (Wall Golmonoy Model VT-500). The colmonoy Electrpspray
Metallizer utilized an electric arc to melt two metal wires fed continuously
to the spray gun from two coils. A compressed air stream was directed
through the arc zone to atomize the molten metal and propel the particles
through the scrubbing system.
The particles were sprayed into a 55 gallon (208 liter) drum where
the large molten metal particles settled "but and the fine particulate as
iron oxide (Fe20.,) was carried to the foam scrubbing apparatus. The air
fed through the spray gun was sufficient to carry the particles to the
foam generator and stilT produce a good quality foam.
-------
FILTERED
AIR
1 ^
1 ^*
f
AEROSOL
GENERATION
SYSTEM
VACUUM
PUMP
-C^~\
«-i \
t
r J
SAMPLING
PROBE _
INLET
kLCMIV MIH ~"^
Q
^ FQAM SAMPLING PROBE
, CONCENTRATE OUTLET S
L. , i, ,..,;:., r ^__ , . • • • •' - '' J FC
/ SCREEN' I.
., .; . •••: .:-.',-' . . : ....... ^
^
IAIV
JCT
MR
EFFLUENT
(RECIRCULATED)
Figure 3. Foam scrubber system.
-------
The foam generator consisted of a fine mesh wire screen placed at a
45° angle inside a 15.25 cm (6 inch) diameter Plexiglas tube. Two spray
nozzles were used to spray a solution of surfactant onto the screen.
The foam was produced when the air from the metallizing spray gun passed
through the wetted screen. Various size screens were used but the
highest quality foam was produced with a 270 mesh stainless steel screen.
(A high quality foam is defined as a foam of uniform bubbles about 1.0
mm in diameter containing no air voids.)
Various spray nozzles were tested to determine which nozzle or
combination of nozzles sprayed the surfactant solution onto the screen
in the best pattern to completely wet the screen and produce the most
uniform foam. The combination of nozzles finally selected to spray the
surfactant solution was a 5 gph (0.315 1/min) 60° full spray nozzle and
a 10 gph (0.6301/min) 60° full spray nozzle. The 10 gph nozzle was
pointed to the center of the screen and located 12.0 cm (4.7 inch) from
the screen. The 5 gph nozzle was located 6.3 cm (2.5 inch) in front of
and 2.5 cm (1 inch) below the other nozzle but pointed toward the lower
portion of the screen as shown in Figure 4.
Since two nozzles were required to completely wet the screen and to
produce a good quality foam, more water was sprayed oh the screen than
was necessary to produce foam. The excess water prevented the screen
from plugging and maintained a low pressure drop across the screen by
continuously washing particles from the screen. The screen became
plugged in less than 1 minute when the surfactant spray was discontinued.
After the foam was generated on the wire screen, the bubbles traveled
down a 15.25 cm diameter Plexiglas tube where the particulate had sufficient
time to contact the bubble wall and be captured in the foam. The length
of the Plexiglas scrubber chamber was 5.8 m (19 ft) which was originally
designed to be shortened by 1.8 m (6 ft) lengths to decrease the residence
time in the chamber. During preliminary operation of the foam scrubber,
it was determined that a longer (not shorter) residence time was required
to obtain a high collection efficiency. To increase the residence time,
the chamber was modified by adding a 1.8m (6 ft) tube chamber as
9
-------
SURFACTANT
SOLUTION
SPRAY RUNOFF
:igure 4. Nozzle arrangement for spraying surfactant on screen.
Figure 5. Extra chamber added to increase residence time.
10
-------
shown in Figure 5 between the screen and destruction chamber. This
extra section was necessitated since space limitations prevented the
addition of extra sections lengthwise to the scrubbing chamber. The
double-tube section increased the residence time by approximately 32
percent at the same foam velocity as through the single-tube chamber.
The bubble size was unaffected by the modification in the residence
chamber, and the only effect on the foam was the lower velocity through
the double-tube section.
From the residence chamber, the foam entered the destruction chamber.
The destruction chamber was simply a Teflon disk on a shaft inside a
15.25 cm (6 inch) galvanized duct. The disk configuration is shown in
Figure 6. The disk was rotated at 5000 to 15,000 rpm impacting the foam
against the duct wall, thereby breaking the foam bubbles. This type of
foam destruction chamber was selected after trying many other foam
breaking methods without success. (Heated steam coils, a flame directly
on the foam, an air jet, and a water spray were tried: the method that
destroyed the foam most efficiently and economically was the high-speed
o
spinning disk. )
During preliminary testing, many modifications were made to the
foam destruction chamber. Originally the disk was powered by a 50,000
rpm electric motor, via a flexible shaft. The disk could not be operated
for more than several minutes without the disk separating from the
shaft. Eventually the flexible shaft was replaced by a solid metal rod,
centered in the destruction chamber. Unless the rod was perfectly
straight, the disk would vibrate and hit the chamber walls. To prevent
shaft vibration, the rod was centered in a 2.5 cm Teflon bushing, bolted
into the destruction chamber as shown in Figure 7. With only minor
problems, the foam destruction chamber performed satisfactorily during
the remaining tests.
During the experiments, the liquid from the destroyed foam ran out
of the bottom of the chamber and the cleaned air exited at the top of
the chamber. The liquid, with the collected particulate and the screen
runoff, was allowed to stand overnight to. allow the solids to settle.
All the liquid was recycled in this manner: the only liquid lost was
the entrainment from the foam destructor.
11
-------
Figure 6. Foam destruction disk.
ELECTRIC MOTOR
TEFLON
BUSHING
DISK
Figure 7. Foam destruction chamber.
12
-------
Foam Scrubbing Tests
The foam scrubbing tests were an extension of earlier tests on a
2
similar foam scrubber. OOP and polypropylene glycol aerosols were
used in the earlier tests to measure the efficiency, and the need still
existed to test foam scrubbing on a solid particulate.
Preliminary tests were run on the foam scrubber to determine the
optimum operating conditions for the particle generating system. These
tests on the metallizing spray gun were run to determine the effect of
spray pressure, voltage, and air flow on the size and number of particles
produced. The particles generated by the metallizing spray gun were
examined by X-ray diffraction and found to be essentially iron oxide
(Fe203). The metallizing spray gun tests are discussed more fully in
Appendix A.
After conditions for generation of an acceptable particulate were
set, tests began on the foam scrubber. Initially, various conditions
were tested to determine the operating limits for generation of a high
quality foam (a foam with all bubbles aproximately 1 mm or smaller). At
high air flows, the foam tended to have air pockets or voids in the foam
which were about 5 cm in diameter or larger. Obtaining a small bubble
size* was not as difficult as eliminating the large air pockets which
traveled through the scrubber with the foam. The position of the spray
nozzles was critical in controlling the air voids since the screen had
to be wetted evenly to ensure constant foam formation. Complete wetting
of the screen was necessary in order to prevent "holes" in the liquid on
the screen where the air could blow through without generating foam;
hence, the large air pockets.
The foam quality tended to improve after several minutes of operation
with the metal fume generator operating. The iron oxide apparently
plugged any unwetted areas of the screen, thereby forcing the air through
the less restrictive wetted areas of the screen. Once a good quality
foam was being produced, the foam quality remained constant until the
nozzles or screen was altered;, then the system had to be realigned.
*The bubble size was determined by collecting a sample of the foam and
counting the number of bubbles per unit length.
13
-------
The res'dence times which were tested for the foam are shown in
Table 1. (Residence time is defined as the length of time the foam is
in the scrubber—from the screen where the bubbles were generated to the
destruction disk wher,e the foam was destroyed.) Residence times were
determined by measuring the velocity of the foam in the clear Plexiglas '
pipe, converting to residence time per unit length, and correcting for
the total length. .
To measure the particulate loading of the air stream before and
after foam scrubbing, total mass determinations and impactor runs were
made on a}1 tests. The mass determinations were made with a Gelman Type
A Fiberglas filter sample. A Brink cascade impactor was used to measure
the high inlet loadings and an Andersen Impactor was used on the outlet.
To compare the results of the two types of impactors, the Brink impactor
was used on the ,outlet for: several tests.
The sampling probe nozzles were placed 8 duct diameters from an
upstream disturbance and greater than 5 duct diameters from a downstream
disturbance. The sampling, therefore, was done at an excellent location
in the duct, All of .the impactor and total mass determinations were
made at the center of-the duct.
For the impactor runs, the samples were pulled as isokinetically as
possible through the nozzle and probe to the impactor. Errors due to
deviations from isokinetic sampling should be of little consequence for
particles having aerodynamic diameters smaller than 5 um. Because of
the small duct dimensions and the fine particulate being sampled, single
point extractive sampling was utilized. The impactor sampling times
were 2 minutes at the inlet and 10 or 15 minutes at the outlet' to provide
sufficient impactor stage loadings. Either grease or glass fiber sub-
strates were used for the Brink impactor tests and glass fiber substrates
for the Ancersen impactor tests. The inlet Brink impactor runs were
conducted c.t ambient temperature, but the impactors were heated to 120°C
at the out'et to remove the water droplets carried over from the destruc-
tion chamber. All impactor data are given in Appendix B.
Since the captured particulate was removed from the used scrubber
liquid (destroyed foam), a significant particle buildup did not occur.
14
-------
Table 1. RESIDENCE TIMES AND AIR FLOWS FOR THE FOAM SCRUBBER TESTS
Rotameter
Reading
%
15
20 •
25
30
35
40b
Flow
3 Rate
in /min (cfm)
0,057 (2.0)
0.074 (2.6)
0.113 (4.0)
0.156 (5.5)
0.209 (7.4)
0.283 (10.0)
Residence
Single Tube
112
85
56
40
30
22
Time, sec
Double Tube
147
113
74
53
40
29
Foam
m/mi n
3.1
4.0
6.2
8.5
11.5
15.5
Velocity3
(ft/min)
(10.2)
(13.2)
(20.4)
(28.0)
(37.6)
(51.0)
aFoam velocity for single-tube section of modified system. The velocity will
be halved in the double-tube section.
Turbulence: large air pockets
15
-------
If the liqird had been recycled without particle removal, the grain
loading at the outlet would have been higher due to the extra solids
carried over- in the water droplets formed during foam destruction. Any
droplets formed during destruction of particulate laden foam would have
contributed to the exit grain loading to a substantial degree. A mist
eliminator was not utilized after the foam destruction chamber, but the
outlet concentration could have been reduced by adding some sort of mist
control device.
16
-------
SECTION V
RESULTS AND DISCUSSION
Particle collection is believed to occur in the foam scrubber at
the bubble surfaces as the foam travels the length of the scrubbing
chamber. By increasing the residence time, the number of particles
striking the bubble wall will be increased and the total efficiency
thereby increased. By decreasing the bubble size, the efficiency should
also be increased; however, the effect of bubble size was not tested
during these experiments. A 325 mesh screen was installed for testing,
but a good quality foam could not be generated. Large air pockets were
always present in the foam and all attempts to modify the system to
obtain a better quality foam with the 325 mesh screen were futile.
Whether the cause of the air pockets in the foam was due to the screen
size, the spray nozzles, or the system as a whole was not determined.
Collection determinations were observed with the foam generated by>.
the 270 mesh screen with different residence times. The addition of the
1.83 m section in the chamber allowed an increase in residence time
without an increase in air flow through the scrubber. The mass filter
determinations were used to calculate the overall efficiency of the foam
scrubber; the results are shown in Figure 8. At residence times approaching
2 minutes, the overall collection efficiency approached 95.0%.
The lowest residence time before turbulence developed in the foam
(35 sec) yielded a collection efficiency of about 58 percent. Lower
residence times were not attempted because of the poor quality foam due
to the large air pockets produced at the higher flow rates. The air
pockets may have been present at a residence time of 35 sec, but they
were not evident. When the pockets were visible, they were very noticeable
and appeared very suddenly. Reducing the air flow always eliminated the .
pockets or at least they were not visible near the scrubber walls. The
pockets may have occurred first in the center of the chamber where they
could not be seen.
17
-------
100
90
80
u
oe
UJ
Z
2 70
o
u
60
50
O SINGLE TUBE SCRUBBER
ODOUBLE TUBE ADDED
20
40
60 80 100
RESIDENCE TIME, seconds
120
Figure 8. Collection efficiencies based on mass filter determinations.
140
18
-------
Impactor Runs
3
The impactor data were reduced by means of a computer model. The
following parameters were calculated:
• Total mass collected (grams)
o
• Total volume of gas sampled (ft at NTP)
Particulate mass concentration (grains/dscf)
Flow rate through impactor (acfm)
50 percent cutoff diameter for each stage for unit
density spherical particles (ym)
• Mass percent on each stage (percent)
• Cumulative mass percent less than each stage cutoff
diameter (percent)
• Geometric mean diameter for each stage interval (vm)
• dM/d log D for each stage interval (grains/dscf)
The data sheets for all the impactor runs are given in Appendix B.
To obtain the fractional efficiency for a test run, dM/d (log D) versus
the geometric mean diameter is plotted for each stage interval for the inlet
and outlet impactor tests. The curves from these graphs yield the concentration
for any given size particle. The ratio of the outlet concentration to the
inlet concentration for any size particle is the penetration for that
particular size.
The inlet and outlet concentrations for various residence times are
shown in Figures 9, 10, 11, and 12 and the penetration is shown in Figure 13.
The Figure 13 curves show that the penetration reaches a minimum at an
aerodynamic particle diameter of approximately 2 to 3 microns. This
minimum point in the penetration curves may be the result of an entrainment
problem since foam scrubbing theoretically should work most effectively
with the very small particles.
19
-------
I I MINI
RESIDENCE TIME 40 SECONDS
I
0.1
M
Vt
O
5
INLET
BRINK
10
.!•
OUTLET
<
DC
u
0.01
ANDERSEN 2
ANDERSEN
6
BRINK
6
0.001
I
0.01 0.1 1
PARTICLE AERODYNAMIC DIAMETER, \m
Figure 9. Inlet and outlet concentration distributions for 40 second residence time.
10
20.
-------
1
<
cc
N
V)
o
H
OC
a.
z
o
0.1
0.01
CO
CO
0.001
= I I I I I I I I |
Z RESIDENCE TIME 53 SECONDS
I I I I I I
INLET
OUTLET
BRINK
14
ANDERSEN 8
ANDERSEN
4
TTF
0.01 0.1 1
PARTICLE AERODYNAMIC DIAMETER, fim
Figure 10. Inlet and outlet concentration distributions for 53 second residence time!
10
21
-------
1
<
oc
V)
o
oc
<
ca
0.1
u
ta
I
0.01
0.001
~r ii i IN
RESIDENCE TIME 74 SECONDS
INLET BRINK11'
OUTLET
0.01 0.1 1
PARTICLE AERODYNAMIC DIAMETER, urn
Figure 11. Inlet and outlet concentration distributions for 74 second residence time.
10
22
-------
_ RESIDENCE TIME 112SECONDS
<
>
DC
V)
u
oc
<
co
o
K
0.1
0.01
o
u
CO
CO
5
O.M1
OUTLET
0.01 0.1 1
PARTICLE AERODYNAMIC DIAMETER, jum
Figure 12. Inlet and outlet concentration distributions for 112 second residence time.
10
23
-------
100
90
80
70
60
50
40
30
20
$ '.'
cc 9
RESIDENCE TIMES
53
1 2 34 5
AERODYNAMIC PARTICLE DIAMETER, Mm
Figure 13. Penetration versus particle size for various residence times.
24
-------
If the predominant collection mechanisms for foam scrubbing were
sedimentation and interception, one would expect the penetration to
decrease with increasing particle size. If diffusion were predominant,
one would expect an improvement in collection efficiency at the smallest
particle sizes. At last one more consideration must be made with foam
scrubbing with no demister: some fraction of particulate material which
has been captured in the foam will escape the destruction chamber,
either because the droplets are too fine to be retained or because the
liquid is atomized when impacted against the chamber walls.
Figure 13 shows that the efficiency decreases with increasing
particle size for particles above 3 ym, so interception and sedimentation
are not important in foam scrubbing and diffusion does appear to be the
predominant collection mechanism. The increase in penetration for
particles less than 3 jam, as shown in Figure 13, is probably a result of
the entrainment problem. The entrainment problem due to foam destruction
is discussed in Appendix C.
The curves in Figures 9 through 12 show that the outlet concentration
for all particle sizes decreases as residence time increases. The
curves for two different residence times cannot be compared exactly
because each residence time has a different inlet concentration due to
different air flows to the particle generator.
The outlet grain loadings calculated by the computer and based on
the impactor data are plotted versus residence times in Figure 14. From
this plot and the penetration curves of Figure 13, the outlet grain
loading appears to reach a minimum at about 2 minutes.
A final test was run to determine the efficiency of the screen in
removing particles from the gas stream. The test involved spraying only
water onto the screen while the particulate laden air passed through so
no foam was produced. All other conditions were the same as the tests
with foam. The results of the test are shown in Figure 15 and the
penetration is shown in Figure 16. Only the larger particles were
removed at the screen; particles less than 1 micron were unaffected.
25
-------
250
200
CO
"
150
a
2 .
5
<
< 100
a:
(9
•50
0
OANDERSEN
DBRINK
20
40
120
140
60 80 100
RESIDENCE TIME, seconds
Figure 14. Grain loading of scrubber outlet versus residence time, based on impactor runs.
160
26
-------
oc
UJ
N
35
o
i
X
5
0.1
0.01
o
u
«t
0.001
0.1 1
PARTICLE AERODYNAMIC DIAMETER, /
10
Figure 1 5. Inlet and outlet concentration distributions for water spray only.
27
-------
100
90
80
70
60
50
*- 40
z
•3
1C
ts
IU
Ik
30
20
10
9
8
1 2 34 5
AERODYNAMIC PARTICLE DIAMETER, fim
Figure 16. Penetration versus particle size for water spray test.
28
-------
SECTION VI
FOAM SCRUBBER ECONOMICS
Based upon the data collected during this small scale investigation
of -foam scrubbing, a scale-up was attempted to determine the economic
feasibility of foam scrubbing. Since this experiment was representative
of very low flow rates and high residence times, the major limitation
may be the size of the unit. Foam scrubbing has been compared to other
conventional control devices: capital cost estimates indicate that foam
scrubbing costs are similar to those for electrostatic precipitation,
2
the most expensive conventional control method.
The foam scrubber as tested had a maximum air flow of 6 cfm (0.17
m /min) through the 15.24 (6 inch) diameter duct. When scaling up the
foam scrubber to much larger air flows, the size of the apparatus
becomes unmanageably large. Table 2 shows the screen areas required for
different flow sizes for foam scrubbers with an air velocity of 37
ft/min (11.28 m/min). The screen area at 1000 cfm (28.30 m3/min) would
2 2
be 26.6 ft (2.47 m ) which represents a rather large duct. Also shown
in Table 2 is the foam volume required for a residence time of 2 minutes.
The data from Table 2 tends to prove that, from the standpoint of total
area and volume, foam scrubbing would not be attractive for large flow
volumes.
The most costly expenditure in the operating cost for foam scrubbing
is the cost of the surfactant. Using a less costly surfactant, increasing
the recycle percentage, or reducing the liquid-to-air ratio at the
screen can reduce the surfactant cost. At 99 percent surfactant recycle,
the cost of the surfactant amounts to over 75% of the total operating
cost as shown in Table 3. The operating cost for foam scrubbing is not
dependent on residence time since the operating cost of the installed
equipment is minimal compared to the cost of the surfactant. Comparison
of foam scrubbing to conventional control devices, as presented in
Figure 17, shows, that the operating cost of foam scrubbing with 99
percent surfactant recycle is an order of magnitude greater than the
2
most expensive conventional control method (high energy wet scrubbing).
29
-------
Table 2. SCREEN AREAS AND LIQUID FLOW RATES FOR VARIOUS
SIZE FOAM SCRUBBERS3
Flow Rate (cfm)
2
Screen area (ft )
Liquid spray
7.5
0.20-;
1 0.25
1000
26.6 ,
33 ' '
10,000
266
330
.100,000
2660 •
3300
rate (
-------
Table 3. OPERATING COST FOR. 50,000 ACFM SCRUBBER AT 99% RECYCLE
Percent of Total
$/acfm
Labor
Materials
Surfactant
Other
Utilities
Depreciation
Capital Charges
76
1
13
1
2
Total
1.58
17.08
0.45
2.93
0.23
0.45
22.72
31
-------
1000
100
E
"o
u
u 10
1.0
0.1
FOAM SCRUBBER
.RECYCLE
\
FOAM SCRUBBER
99% RECYCLE .
HIGH ENERGY
WET SCRUBBER
ELECTROSTATIC
PRF.CIPITATOR
FABRIC FILTER
103 104 105
CAPACITY, acfm
Figjre 17. Comparison of various control devices for fine particle control.
32
-------
SECTION VII
REFERENCES
1. Calvert, Seymour, Jhuda Goldshmid, David Leith, and Dilip Mehta.
"Wet Scrubber Systems Study, Scrubber Handbook," A. P. T., Inc., EPA-R2-72-118a
(NTIS No. PB 213016), August 1972.
2. Ctvrtnicek, T. E., C. M. Moscowitz, T. F. Walburg, and H.H.S. Yu.
"Application of Foam Scrubbing to Fine Particle Control, Phase I," Monsanto
Research Corporation, EPA-600/2-76-125 (NTIS No. PB 261075/AS), May 1976.
3. Smith, W. B., K. M. Cushing, and J. D. McCain. "Particulate
Sizing Techniques for Control Device Evaluation," Southern Research Institute,
EPA-650/2-74-102 (NTIS No. PB 240670/AS), October 1974.
33
-------
APPENDIX. A
THE METALLIZING SPRAY GUN AS A FINE PARTICULATE GENERATOR
Preliminary experiments were run with the Wall Colmonoy Model VT-
500 metallizing spray gun to determine the conditions necessary to
generate a satisfactory particulate to be tested on the foam scrubber.
The metallizing spray gun generated fine particles by melting two metal
wires with an electric arc. The metal wires were fed continuously by
the spray gun and a compressed air stream was directed through the arc
zone to atomize the molten metal and propel the particles to the foam
scrubber. The spray was directed into a 55 gallon (2081) drum to allow
the larger particles to settle to the bottom. The drum and spray system
is shown in Figures Al and A2.
Actually very few of the particles settled in the drum because the
metal particles were impacted on the side of the drum opposite the wire
and compressed air entry point forming a very heavy layer of metal on
the side of the drum. The compressed air through the arc zone, which
was the only source of air to generate foam, probably was too high and
caused the metal buildup. A longer spray length probably would have
decreased the buildup and increased the grain loading of the stream out
of the drum.
A sample of the particles generated from the carbon steel wire was
analyzed by X-ray diffraction and found to be gamma Fe90~. The corresponding
CO' \
layer of metal on the side of the drum was analyzed by X-ray fluorescence
and found to be essentially metallic iron.
A complete set of impactor runs were made with the metallizing
spray gun to determine the effect of air flow, voltage, and wire feed
rate on the.size and number of particles generated using carbon steel
wire. The experimental conditions are given in Table Al. The particle
characteristics for each of the tests were determined, using the Brink
cascade impactor. The sample was collected isokinetically at the inlet
sampling port of the foam scrubber. The same sampling conditions were
utilized for the particle generation tests as for the foam scrubber
inlet tests. The distribution curves for the voltage, air flow, and
wire feed rate (in Figures A3, A4, and A5) resulted in an increase in
A-l
-------
METAL WIRE
i • FROM .
ELECTRIC SPRAY GUN
55 GALL ON DRUM
Figure /\,1. Participate generating system.
MOLTEN METAL
SPRAY
METAL WIRE
; TUBES
COMPRESSED AIR
Figure A2. Top view of electric arc area.
A-2
-------
Table Al. EXPERIMENTAL CONDITIONS FOR WIRE SPRAY GUN TESTS
FOR IRON OXIDE
Carbon Steel Wire
Feed Rate, Ib/min
0.0224
0.0224 .
0.0224
0.0706
0.0706
0.0706
0.1272
0.1272
0.1272
0.0224
0.0224
0.0224
0.0706
0.0706
0.0706
0.1272
0.1272
0.1272
0.0224
0.0224
0.0224
0.0706
0.0706
0.0706
0.1272
0.1272
0.1272
Air Flow, cfm
4.4
4.4
4.4
4.4
. . 4.4
4.4
4.4
4.4
4.4
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
8.8
8.8
8.8
8.8
8.8
8.8
8.8
8.8
8.8
Pressure, psi
10
10
10
10
10
10
10
10
10
20
20
20
20
20
20
20
20
20
30
30
30
30
30
30
30
30
30
Voltage
25
28
31
25
28
31
25
28
31
25
28
31
25
28
31
25
28
31
25
28
31
25
28
31
25
28
31
A-3
-------
the mean particle size—and to a minor extent, increasing the voltage
increased the mean particle size. The same results are obtained by
plotting voltage, air flow, and wire feed rate versus grain loadings!as
seen in Figures A6, A7, and A8.
A second set of'tests were run with the aluminum wire in which the
voltage was varied from 22 to 46 volts with all other conditions constant.
For comparison five runs were made with the carbon steel wire for voltages
from 34 to 46 volts. -The impactor samples were collected under the same
conditions as described previously; results are given in Table A2,. > As
the voltage increased, the grain loading increased with a higher percentage
of large particles tn the sample.
A-4
-------
cc
o
10
9
8
7
6
5
1.0
9
8
7
10
iirnif
VOLTAGE
• 25 VOLTS
A 28 VOLTS
• 31 VOLTS
20
I I I I I I I
1
1
30 40 50 60 70 80 90 95 98 99
CUM PERCENT OF MASS SMALLER THAN D50
Figure A3. Distribution curves for metallizing spray gun tests based on voltages.
99.8
A-5
-------
cc
o
10
9
8
7
6
1.0
9
8
7
6
10
AIR FLOW
• 4.4 CFM
A 6.5 CFM
• 8.8 CFM
20
I
30
99.8
40 50 60 70 80 90 95 98 99
CUM PERCENT OF MASS SMALLER THAN D50 ' .
Figure A4. Distribution curves for metallizing spray gun tests based on air flow rates.
A-6
-------
o
cc
u
in
a
10
9
8
7
6
5
1.0
9
8
7
6
10
WIRE FEED
• .0224 LBS/MIN
A .0706 LBS/MIN
• .1272 LBS/MIN,
99.8
20 30 40 50 60 70 80 90 95 98 99
CUM PERCENT OF MASS SMALLER THAN D50
Figure A5. Distribution curves for metallizing spray gun tests based on wire feed rate.
A-7
-------
32
30
28
LU
O
5 26
o
24
22
20
1 I 'I I T
I
0.2
0.8
0.4 0.6
GRAIN LOADING, grains/act
Figure A6. Outlet grain loading versus voltage for iron oxide.
1.0
10
_ 6
w
e
oe
< 4
I I
i i i n T
0.8
0 0.2 0.4 0.6
GRAIN LOADING, grains/act
Figure A7. Outlet grain loading versus air flow for iron oxide.
1.0
A-8
-------
0.14
0.12
E 0.10
E
UJ
U; O.OB
UJ
E 0.06
004
0.02
I I
0.2
II II I
0.8
1.0
0.4 0.6
GRAIN LOADING, grains/acf
Figure A8. Outlet grain loading versus wire feed rate for iron oxide.
A-9
-------
Table A2. TEST DATA FOR ALUMINUM OX1DI AND IRON 0X11)1 I OH SIMMY CUN
5.
Voltage
Grain Loading
grains/acf
Percent Less Than Stated Size
13.5 m
8.2 m
4.7 m
2.8 m
1.9 m
1 .0 m
0.7 m
Aluminum Oxide
22
25
28
31
34
37
40
43
46
Iron
34
37
40
43
46
0.755
1.012
1.646
1.631
1.667
3.676
2.881
2.637
2.816
Oxide0
1.942
1.659
2.148
1.299
2.688
1 00
100
78.58
64.29
60.19
75.20
65.18
74.15
68.95
80.80
80.63
81.44
79.21
55.99
m in
-// . 1 -/
92.10
77.24
57.15
59.26
70.23
59.83 .
64.88
65.30
77.49
75.97
76.05
72.28
52.64
or* f\ f^
71.19
64.74
51.79
55.56
59.55
51.34
57.08
57.08
68.88
67.45
65.27
59.41
43.07
73.24
68.37
48.67
39.29
47.23
47.71
41.97
47.32
46.12
43.71
57.37
51.50
47.53
31.58
61 . 95
54.81
38.84
36.91
51.67
33.21
32.15
37,08
35.62
34.44
45.74
37 . 73
33.67
23.45
49.30
47.46
31.26
27 . 39
31.49
19.47
21.88
24.40
23.75
26.50
34.89
27.55
26.74
18.67
js.22
35.60
23.67
20.24
28.71
11.84
13.40
16.59
15.53
19.21
20.16
.18.57
21.79
10.05
For all tests, spray pressure was 20 psi and air flow was 6.3 cfm.
Wire feed rate was 6.2 gm/min.
cWire feed rate was 60 gm/min.
-------
APPENDIX B
ANDERSEN AND BRINK IMPACTOR DATA
B-l
-------
ANDERSEN IMPACTOR
Date 6/7/76 Run # 4-1
Sample volume at STP (ft3) =2.. 362.
Impactor flow rate (acfm) = O
Concentratioi (grains/ft ) = 0
Location
Bar. Pressure (in Hg)
Temp in.duet ("F) =
Time, (minutes) =
3O> '3
75. O
J>. O
Stage
1
2
3
4
:-5
"«6
7
R
F
Total
Net weight
(>P)
.OOOO
.0000
.000°
.0002.
.000 2.
.OOO6
.OOO3
-00/4
.0035-
.^)068
% on
stage
2.^4
2.34
0.52,
/3.24
2^v®
5/-47
Size cutoff
(vm)
/3.B7
5.76
6.C3
4.26
1 2v79
" /.i2
ad£
A6/
5S - stated
size
/oo.oo
IQQ.OO
JOO.OO
97.06
54. /2-
85. ZO
7^.o6
5/.4ft
dm/d log D
.OOOO
.OOOO
.oo&s
.0971
.0115.-.--
.(72.77
.0713
.OUT7
Geo.^ Mean
(vm)
/6.^5
V 7.63
767
5".O7
J^4£T
/.-?2r
/.04
^?,70
0.22
74
4.0 cf
B-2
-------
ANDERSEN IMPACTOR
Date (o/7/7(o Run # A-2.
Sample volume at STP (ft3) = 4.02.7
Impactor flow rate (acfm) = 0.SO"3
Concentration (grains/ft3) = 0.072.&
Location
Bar. Pressure (in Hg)
Temp in .duct '("F) =
Time (minutes) =
3O.'3
75
8
Stage
1
2
3
4
5
6
7
ft
F
Total
Net weight
(gm)
.OOOO
.OOQO
.0002.
.0001
.0003
.001$
.002.0
.0062.
.0087
.Ot3O
% on
stage
LOB
/.SB
/.SB
6>W
10.53
32.63
45.7$
Size cutoff
(um)
73.44
J.45
5.84
*./3
2.70
/.2fi
0.73
0.59
% 4 stated
size
/oo.oo
100. 00
38.35
37.37
2S.79
&b3S
78.4-3
45.79
dm/d log D
.0034
.0075
.0O6>2.
.0/5/
*OZk>O
.1*47
.03&\
Geo.^Mean
(ym)
;6.39
11.27
7.43
4.51
5.34
/.84
1.00
6.6ft
0.21
40 sec
7.4
B-3
-------
ANDERSEN IMPACTOR
Date
Run #
Sample volume at STP (ft)
Impactor flow rate (acfm)
Concentration (grains/ft,3)
2.7O
O.542.
Location
Bar. Pressure (in Hg) 3O. <
Temp in duct ("F) = 75.0
Time (minutes) = _5. O
Stage
1
2
3
4
--5
6
7
R
F
Total
Net weight
(gir)
>Q60S
.000&
.oois,
.002.4
.0051.
% on
stage
S>.$2.
/5.3fi
25.55
4£./5
Size cutoff
(ym)
2.61
M^
. 76
. 56
$ -? stated
size
lOO.OO
30. 59
75.00
4L.II*
dm/d log D
.00P6
.01/4
.Ofck2.
.011*0
Geov Mean
(vm)
3.22.
/-75>
0.36
O.bG
0. 21
Residence. TI*I&, -
= 2.6
II 3 sec
B-4
-------
ANDERSEN IMPACTOR
Date 6-8-76 Run # A-4
Sample volume at STP (ft3) = 2.4.94
Impactor flow rate (acfm) = 0.439
Concentration (grains/ft.3) =
Location Puf/et
Bar. Pressure (in Hg) 30. 2Q
Temp in.duct ("F) = 75 Q
Time (minutes) = 5". 0
Stage
1
2
3
4
5
6
7
R
F
Total
Net weight
(gm)
.OOOl
.OOOl
,ooe>7
.ooo3
.oo^\
.0039
.007&
% on
stage
/,2.a
L2B
8.97
/1 .54
26.92
50.00
Size cutoff
(ym)
5*%7
4. /4
2.72.
MS
0.79
0.55
j£ - stated
size
IOO.OO
18.71
37.44
58.47
76.9^
50.01
dm/d log D
.0042.
.0033
.0131
.0161
.0953
.0274
Gep^Mean
(ym)
7. 4fc
4.93
3.35
/.67
AOI
0.63
0.2.1
im*.
Air
5. 5
B-5
-------
ANDERSEN IMPACTOR
Date
Run #
Sample volume at STP (ft3) = 3.00
Impactor flow rete (acfm) = O.S59
Concentration (crains/ft3) = 0. 06t>4
Location 0u.t"/e-"t
Bar. Pressure (in Hg) 3O.22-
Temp 1n.duGt"("F) = 75
Time (minutes)' = £
Stage
1
2
3
4
5
6
7
,q
-
Total
^et weicht
(gm)~
.ooos
.0€>14
.00 ^'L
.0035
.OOS3
.0^9
% on
stage
5,06
10. 85
V. 0£
27. /3
4/.OS)
Size cutoff
(vm)
Z.7&
2.47
i.n
0.72.
0.53
£ - stated
size
/oo.oo
Jt>.ll
£5.28
60. 2 Z
4/.OP
.
dm/d log D
.0/38
.02/fi
,O52d
J35/
.^328
Gep^ Mean
(vmJ
4.49
5,05
A69
031
0.61
0.20
5. 5
, 3-7 y
ore,
or
^ /o 4e.
B-6
-------
ANDERSEN IMPACTOR
Date 6-/0-76 Run #4-7
Sample volume at STP (ft3) = '
Impactor flow rate (acfm) = t
Concentration (grains/ft3) = 0.04B8
Location Oocfte-'f
Bar/Pressure (in Hg) 5O.2/2.
Temp in.duct'(PF) = 75
Time (minutes) = 4.0
Stage
1
2
3
4
5
6
7
fl
F
Total
Net weight
(gin)
.000/
.OOO!
.0003
.0006
.0007
,0005
.0007
.0033
% on
stage
3.03
3.03
$.0?
/a/8
2/.2I
24.24
2/.2.I
•.
Size cutoff
(ym)
5.50
UJ
4. /5
2.71
/.3L9
0.79
O.S2
% - stated
size
/OD.OO
^6.37
33.34
£*.85
66».4>7
45.4fc
2/.aa
dm/d log D
.0037
.005/
.01 xs
.0141
.02.54
,^/23
,00i>Z
Geo.^ Mean
(ym)
//.33
7.47
4.34
3.34
/. 87
1.01
o.t>&
6.11
4.V
B-7
-------
ANDERSEN IMPACTOR
Date 6-//- 7(9 Run # A'B
Sample volume at STP (ft3) = 2. SO
[mpactor flow rate (acfm) = 0.50
Concentration (grains/ft3) = 0. O&OS
Location
Bar. Pressure (in Hg) £9.95
Temp in .duct ("F) = 75.0
Time (nunutes) , = S. O
Stage
1
2
3
4
5
6
7
A
F
Total
Net weight
(gm)
'.cool
.OOQ9
.0011
.00.1&
.0&&&
.o@98
% on
stage
I.Ot
3. IB
12.14
2&.S7
40.98
Size cutoff
(ym)
4.14
2,11
i.za
0.79
0.59
% -. stated
size
100.00
3838
B9.BO
77.5-6
n.$B
dm/d log D
.OOJ3
.0170
.OZSO
J35t>
.0340
Geo .^ Mean
(ym)
4.22.
^.^5
/. 86
A 00
O.t>8
0.11
= 53
Air
5.5
B-8
-------
ANDERSEN IMPACTOR
Date b-W-lb Run # 4-9
Sample volume at STP (ft3) = 2.48
Impactor flow rate (acfm) = 0.496
Concentration (grains/ft3) = 0.099&
Location
Bar. Pressure (in Hg) 3O-O7
Temp in.duct (QF) = 7S.O
Time (minutes) = 5- O
Stage
1
2
3
4
5
€
7
ft
F
Total
Net weight
(gm)
.OOOO
.OOO \
.0001
.0004
.0003
.Ooi'L
.0035
.008L
.oibO
% on
stage
0. hi
/. 88
2.50
5.fcS
/3.7S
2/.07
53.75
Size cutoff
(ym)
/3.5^
5.51
S.fiP
4./fc
2.7^
A29
0.19
O.S9
% 4 stated
size
too. oo
39.3ft
3 7. SO
35.01
49.38
75.63
Sl.7k>
dm/d log D
,0030
.0/23
,01^5
.oni
.0644
JfeSl
.OU)
Gep.^Mean
(vm)
/^>.4t
/I. ^4
7.4^
4^5
3.37
1.87
l.o\
O.t&
0.11
R
77 me-
B-9
-------
ANDERSEN IMPACTOR
Date
Run # A- 10
= 2.47
Sample volume at STP. (ft)
Impactor flow rate (acfm)
Concentration (grains/ft3) = 0-02.B
Location
Bar. Pressure (inHg). 30.07
Temp in.duct (SF) = 75.0
Time (minutes) = 5-®
Stage
1
2
3
4
5
6
7
a
F
Total
Net woight
(gn)
.0000
.oaoi
.Q&05
.&CO&
.OQl9
0.12.
.
B-10
-------
ANDERSEN IMPACTOR
Date 6-l5-7fe Run # A- II
Sample volume at STP (ft3) = I. 57/
Impactor flow rate (acfm) = 0-53.4
Concentration (grains/ft3) = O • 11 SO
Location
Bar. Pressure (in Hg) ZS.9I
Temp in.duct (QF) = 75.0
Time (minutes) = H>.0
Stage
1
2
3
4
5
6
7
«
F
Total
Net weight
(gm)
.OOOO
.0003
.ooo&
,0011°
.0034
.00+5
.0117
% on
stage
4.0O
5.56
6.04
22.22
29.91
38.46
Size cutoff
(urn)
4.04-
2.6S
/.25
0.77
O.S7
% * stated
size
100.00
37.44
30.6>0
^>B.iB
35.47
dm/d log D
.0/6O
.02,40
.120 A
,2J*7*>
.05 IS
Geo.^Mean
(ym)
4.fil
3, a?
/.82
o.se
0.6^
o.zi
= 74
- 4.O dfm
B-ll
-------
ANDERSEN IMPACTOR
Date &-/4-7'6 Run # A- 12.
Sample volume at STP (ft3) = 2.52.
;!mpactor flow °ate (acfm) = O.SOS
Concentration (grains/ft,3) = 0.083
Location
Bar. Pressure (in Hg)
Temp in.duGt'C'F) = 75. O
Time (minutes) ,= 5. Q
Stage
1
2
3
4
5
6
7
R
F
Total
Net weight
M
.OQ&O
.0001
.OOI O
.0C'2Z
.0^>4 1
.^062-
.0/36
% on
stage
£.00
0. 7B
7.35
/<>./&
30. 13
15.59
Size cutoff
Um).
4. )Z
2.70
rza
0,73
0.S9
% - stated
size
100.60
91. 92.
75.74
45.55
dm/d log D
.0033
.01 & 7
.0637
J3*5>
.04-3^
Geo .^ Mean
(urn)
4.30
3.36
/.fiS
/.00
0.6fl
0.22.
- 40
XI. r
= 7-4
B-12
-------
ANDERSEN IMPACTOR
Date i-21'lt* Run # A-I-S
Sample volume at STP (ft3) = 2
Impactor flow rate (acfm) = O.S7/
Concentration (grains/ft3) = 0.
Location
Bar. Pressure (in Hg) SO. 14
Temp in.duct'(PF) = 75,0
Time (minutes) = 5.0
Stage
1
2
3
4
5
6
7
ft
F
Total
Net weight
(gm)
.OOOO
.OOOI
.OOO1.
.OOO b
.0014
.002J
.003J
.0015
% on
stage
/.33
2.^7
^.60
/fi.^7
25.00
4/33
Size cutoff
(ym)
5.4B
3.B7
2.54
/.2O
0.14
o.ss
% - stated
size
/oo.oo
3£.67
_9^.0O
<95.0I
63.34
4/.34
dm/d log D
.O035
.005B
.^>03B
.0354
.OBBO
.0138
Gep^ Mean
(ym)
6.57
-4.6O
3.13
/.7^
034
0,6*
0.11
v4»'r
B-13
-------
ANDERSEN IMPACTOR
Date
Run J >4-/6
Sample volume at STP (ft3) = /.
"mpactor flow rate (acfm) = O.S3O
Concentration (grains/ft. ) = O.33O
Location
Bar. Pressure (in Hg) JO. 14
Temp in-duct (''F) = 75.0
Time (minutes) = 3.0
Stage
1
2
3
4
5
6
7
R
F
Total
Net weight
(gin)
.00O3
.0003
.0005
.0004
.00\O
.001$
.O077
.O&B9
.OUIe
M4\
% on
stage
0.£&
^.Bfi
/.f7
/./7
2J2
).$7
22. SB
2L>JO
34.02.
Size cutoff
(vm)
11.09
9.2.O
5.(>9
4.02.
2.t>*>
1.24
0.77
0.57
% - stated
size
39.13
3£. 25
2t>.78
35.61
32.^7
£2.70
60.^
34.^
dm/d log D
.0/5&
.0/fift
.0130
.0254
.0525
./OO/
.3505
.^>5d4
./307
Gep.^Mean
(ym)
/4./B
/0.98
7.14
4.7A
3.25
/.8I
O.^B
o.(>(>
^.21
= 4.0
B-14
-------
•BRINK IMPACTOR
Date S-24-7£ Run # 8-3
Sample volume at STP (ft3) = 0.
Impactor flow rate (acfm) = 0.
Concentration (grains/ft3) = f> .
Location Outk.t
Bar. Pressure (in Hg) = 2L9.fiO
Temp in duct (°F) = &O>0
Time (min) =
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
Net weight
(gin)
.0 000
.0001
.0003
.0012,
.OOlb
% on
stage
0.00
&.ZS
/B.7S
75.00
Size cutoff
(ym)
1.7&
03b
0.k>T>
L— stated
size
IOO.OO
93.75
75.00
dm/d log D
.0091-
.0401
.0332.
Geo. Mean
(ym)
^.?4
2./4
/.30
0.77
0.11.
Time. * 8$
B-15
-------
BRINK IMPACTOR
C'ate 5-24-
# £-4
Sample volume at STP (ft) = O. ISO
Impactor flow rate (acfm) = 0.075
Concentration [grains/ft3) = O-3SI
Location
Bar. Pressure (in Hg)
Temp in duct (°F)
Time (min)
29.&0
, O
2.O
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
tet weight
(gn)
.0001
.000 1
.0004-
.OooS
.Oo&3
.000$
.0013
.QCSB
% on
stage
l.tol
l.bl
10. S 5
ll.lt>
13. U*
I2.it,
34.U
Size cutoff
Urn)
//.S3
7.0I
3JB
2.37
/.65
O.B&
O.S7
/0~ stated
size
/oo.oo
97.17
34.73
£4.22.
"71.0k
47.37
34. VL
dm/d log D
.GOOD
.04BS
,042.7
./&SS
.32Zfe
.350D
.280k
.1533
Geo. Mean
(ym)
/5./B
833
S.Z&
3.07
1.31,
1.19
0.71
0.1,1
B-16
-------
BRINK IMPACTOR
Date 5-2s-7t> Run # B'
Sample volume at STP (ft3) =
Impactor flow rate (acfm) =
Concentration (grains/ft3) =
O.
d>.
0,438
Location
Bar. Pressure (in Hg) = 25.69
Temp in duct (°F) = &$.O
Time (min) = 2.O
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
Net weight
(gm)
.O005
.0003
.0007
.00&*>
.0003
.0017
.005%
% on
stage
I3.lt>
7.89
18.42.
7.89
7.B9
*f.74
Size cutoff
(ym)
7. 42.
4.2Z
2.51
/. 73
0.21
O.b\
/o— stated
size
100.00
&<*.&$
78.33
60. S3
52 (04
44.74
dm/d log D
.OOOO
.2385
.155$
• 5051
./*(*--
JB90
.2261
Geo. Mean
(ym)
5.51
5. 6O
3.25
2.06
/.27
0.75
0.22.
B-17
-------
BRINK IMPACTOR
Date 5-25-7*. Run # 6- 4>
Sample volume at STP (ft3) = 0.576
Inpactor flow rate (acfm) = 0
Concentration (grains/ft3) = O.OS7
Location
Bar. Pressure (in Hg) = 2.9.77
Temp in duct.(°F) = &O*O
Time (min) ,= /£• 0
J'.tage
Cyclone
0
1
2
3
4
5
Filter
Total
Net weight
(gm)
.OOOO
.0001
.0003
.OOO4
.OOO5
.0008
.00 'Li
% on
stage
4.74
/4.18
13. OS
23. ft/
58.10
Size cutoff
(ym)
*S7
2.72
1.67
1.01
o.u
/o— stated
size
IQO-OO
35.24
80 .96
61. S>l
3B.IO
dm/d log D
.OOOO
.0/2.1
.£>508
.0^-03
.0742,
.0238
Geo. Mean
(ym)
£. 05
3.52
2.26
/.38
0.82
0.21
B-18
-------
BRINK IMPACTOR
Date 5-2.7-76 Run # B-&
Sample volume at STP (ft3) = O.b07
Impactor flow rate (acfm) = 0.fcl
Concentration (grains/ft3) = 0.62.57
Location ouf/et
Bar. Pressure (in Hg) = 3O.I3
Temp in duct (°F) = 8O. 0
Time (min) =
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
Net weight
(gin)
.OOOO
.OOO \
.0005
.0003
.000 1
.0001
.OOlO
% on
stage
/D.OD
3O.OO
. . -: !
30.00
10. OO
20.00
Size cutoff
(ym)
4.45
2.^5
/. 52L
6.39
0.M
k— stated
size
IOO. 00
BO. DO
ho.oo
30. OO
20.00
dm/d log D
.OOOO
.01 14
.O4~ib~
-aa-
.OIBA^
.0139
.0057
Geo. Mean
(vm)
5.90
1.4-3
2.19
1.34-
O.&O
0.23
ideate.
HI
B-19
-------
BRINK IMPACTOR
Date 5-25-76, Run # B-9
Sample volume at STP (ft3) = 0-II&
Inpactor flow rate (acfm) = 0.059
Concentration (grains/ft3) = 0.85O
Location
Bar. Pressure (in Hg) = 3O./0
Temp in duct (°F) = $5.0
Time (min) , . = 2.O
Stage
Cyclone
Q
1
2
3
4 '"
S
filter
Total
Net weight
(gin)
. oooo
.OOOb
.0011
.OOO3
.OOIO
.0009
.00 1 8
,00b5
% on
stage
2.23
20.00
/3,0S
/5.3ft
'3. 85
27.69
Size cutoff
(ym)
7.52.
4. SO
!.(>&
1. 85
-" i. oo "
0.fe5
/o- stated
size
•
/oo.oo
30.77
70.77
5b.9l
: "47:54
27.70
dm/d log D
.OOOO
.32.30
.7fcl/
.7346>
.4^29
.^427
.24)08
Geo. Mean
(ym)
/OJ4
^.97
3.47
2,22.
/.3fc
O.fll
0.23
•
B-20
-------
BRINK IMPACTOR
Date 5-26-76 Run # 8-/O
Sample volume at STP (ft3) = 0.109
Impactor flow rate (acfm) = O. 054
Concentration (grains/ft3) = 0-fc>43
Location
Bar. Pressure (in Hg)
Temp in duct (°F)
Time (min)
= 3O./O
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
Net weight
(gin)
.OOOO
.000+
.oo\ \
.0007
.0010
.000k
.0007
.0045
% on
stage
&W
24.44
IS.St>
22.22.
13.33
/S.54,
Size cutoff
(ym)
£.28
4.7J
2.50
/.S^
/.^>4
0.66
I,,- stated
size
IOO.OO
31; li
^A .^7
51. /£
20.89
/5.5^
dm/d log D
.0000
.2354
.7042.
.6244
.5^92
.4^75
JOBj
Geo. Mean
(ym)
/0.60
6.24
3.4,3
2.33
/.42.
^54
^.25
B-21
-------
BRINK IMPACTOR
Gate 5"- 26-76
#
3
Sample volume
20.07
Size cutoff
(urn)
755
4.44
2.&6
/.55
0.99
C.toS
/*— stated
size
100. OO
34.74
70.^5
^4.5Z
43.&t
25.0S
dm/ d log D
.d?ooo
.1590
.5/S7
.^422.
.5024
.^36^
.2258
Geo. Mean
(vim)
10. ot>
£.31
S.44
2.21
1.35
0.80
0.23
B-22
-------
BRINK IMPACTOR
Date 5-2fi-74 Run # £
Sample volume at STP (ft3) = O.9OS
Impactor flow rate (acfm) = 0.0feO
q
Concentration (grains/ft ) = 0.072-
Location
Bar. Pressure (in Hg) = 5O.07
Temp in duct (°F) = &O.O
Time (min) = IS- O
Stage
Cyclone
0
1
2
3
4
5 _j
Filter
Total
Net weight
(gin)
.OOOO
.0001
.Oool
.6007.
.000$
.0004
.0007
.0022.
.00+1
% on
stage
1.3ft
2.50
4.74
//.90
3.52.
lto.t.1
52.38
Size cutoff
(ym)
/2.85
7.82.
4.45
2.^5
/.B2
0..9fl
^).64
/^- stated
size
/oo.oo
97 62.
J5.2.4
3^.46
70. SB
b$.0S
51.19
dm/d log D
.OOOO
.0080
.O01O
.0/52.
.0530
.02,fc0
.Ofc>5£)
.04/7
Geo. Mean
(um)
7&.Ctf
10 .03
5.3^?
i.43
2.19
/.34
0.gd?
0.2}
B-23
-------
BRINK IMPACTOR
[late 5-28-lb Run # 6-/4
Jiample-.volume.it STP .(ft3) = O. 734
;:rnpactor flow rate (acfm) = O.Otol
Concentration (grains/ft3) = 0.040
Location
Bar. Pressure {in Hg)
Temp in duct (°F)
Time (min) .
= 8O.OO
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
Net vieight
(<)m)
.OOOO
.000 1
.0000
.lOoOl
.0017
.0013
% on
stage
S.lfe
6.00
5.2.4
89.47
Size cutoff
(ym)
2.£>3
1.81
0.98
0.H-
/»- stated
size
^
100.00
li. 74
34.74
£.9.48
dm/d log D
.OOOO
.0/30
.^OOO
.PII4
.053B
Geo. Mean
(ym)
3.40
2./B
/.3^
A7fi
0.2-S
B-24
-------
BRINK IMPACTOR
Date
2.4-74, Run #
Sample volume at STP (ft3) =
Impactor flow rate (acfm) =
Concentration (grains/ft3) =
O.S&b
0.056
0.O56
Location
Bar. Pressure (in Hg) = 3O./0
Temp in duct (°F) = 8O.O
Time (min) = IO.O
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
Net weight
(gm)
.OOOO
.000 1
.000/
.0003
.0002.
.0007
.0007
.00/5
.0036
% on
stage
2.7B
2.7fi
3.33
5.55
l$.44
19.44
4/.67
Size cutoff
(ym)
/Z.07
7.3^
4.53
2.^9
/. 86
1 00
O.blo
%- stated
size
/OO.OO
37.13
34.45
5fc. /Q.
50.56
^/./i
4/. 67
dm/d log D
.00OO
.0/^3
.01 OB
.0354
.032.9
.0^9^
J008
.0437
Geo. Mean
(ym)
ih.n
10.2.0
6.00
3A9
1.14
/.36
O.BI
0.21
B-25
-------
BRINK IMPACTOR
Date
#
Sample volume at STP (ft3) =
Impactor flow rate (acfm) =
Concentration (grains/ft ) =
0.057
0.03*7
Location
Bar. Pressure (in Hg)
Temp in duct (°F)'
Time (min) • . : •
#5.0
lO.O
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
Net weight
(gm) ;
.OOOO
.0002.
.Oool
.OOO2.
.0003
.0014
% on
stage '
11.23
7.14
ll.tf
(A.2&
Size; cutoff
(ym)
2.7 O
1. bl*
.1.01
O.lob
/a- stated
size
/OQ.OO
85.72.
78.S&
H.2.9
dm/d log D
-•••'-
.OOOO
.02*>Z
.OIOO
.0231
. ozw
(
Geo. Mean
(v.m) -
s.so
2.2-4
i.n
o.&\
0.2-5
B-26
-------
BRINK IMPACTOR
Date t- Zfi -76 Run # B
Sample volume .at STP (ft3) =
Impactor flow rate (acfm) =
Concentration (grains/ft ) =
0.056
0.048
Location
Bar. Pressure (in Hg)
Temp in duct (°F)
Time (min)
= IO . 0
Stage
Cyclone
0
1
2
3
4
5
Filter
Total
Net weight
(gin)
.0000
.0001
.0003
.0014
.00/8
% on
stage
5.55
/k.fc?
777 B
Size cutoff
(ym)
/.fit
/.0/
0.64
/£- stated
size
#>0.00
54.45
77. 7fl
dm/d log D
.OOOO
.0/00
.043&
.0+\l
_
Geo. Mean
(vm)
2.2.4
/.3fe
0.81
£> 23
= 40 sec
B-27
-------
APPENDIX C
ENTRAINMENT PROBLEM DUE TO FOAM DESTRUCTION
Foam destruction by the spinning disk method probably creates more
of an entrainment problem than any other method of foam destruction.
Even if the foam is allowed to destruct naturally, some water droplets
are formed every time a bubble breaks. Using a condensation nuclei
counter* to measure the particles from the destruction chamber at
various conditions showed that many particles were generated just by
foam breakage. After a test the chamber would remain about half full of
foam. If clean air was allowed to flow over the foam at the same conditions
as tested the foam destructed very slowly (2 to 3 hours). The particle
count with clean air through a dry tube was about 100 particles per cc;
with clean air over foam, the particle count was over 100,000 particles
per cc; and with the foam destruction disk operating and foam flowing
through the chamber, the particle count was over 1 million particles
per cc--in the non-linear range of the counter where concentration is
impossible to measure. (The particle count was in the non-linear range
with the metallizing spray gun on or off.)
Other than for the few tests run to determine if an entrainment
problem existed, the condensation nuclei counter was not a practical
measurement device for this experiment.
*Model Rich 100 manufactured by Environment/One Corporation
C-l
-------
TECHNICAL REPORT
(Please read liiantctiom on the reverse
DATA
before completing)
1. REPORT NO.
I3PA-600/2-77-197
2.
3. RECIPIENT'S ACCESSION'NO.
4. TITLE AND SUBTITLE
Evaluation of Foam Scrubbing as a Method for
Collecting Fine Particulate
5. REPORT DATE
September 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOfilS)
Goddes H. Ramsey
8. PERFORMING ORGANIZATION REPORT NO.
9. F'ERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-007
See Block 12.
11. CONTRACT/GRANT NO.
NA (Inhouse)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office o:: Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 1-12/76
14. SPONSORING AGENCY CODE
EPA/60.0/13,
i.>. SUPPLEMENTARY noTEsAuthorRamseyfs Mail DrQp & 61; his phone is 919/541-2298.
is. ABSTRACT
summarizes the knowledge and data obtained during an investi-
gation of foam scrubbing as a method for collecting fine particulate. The foam
scrubber was tested at room temperature, using iron oxide aerosols at concentrations
near 0.00137 mg/cu m. Inlet and outlet samples were taken with cascade impactors
sind toral maes filters. These tests were performed with different foam residence
times and flow rates. A residence time of 35 seconds yielded a collection efficiency
of 58%, while a residence time of 120 seconds yielded an efficiency of 95%. The
operating cost of foam scrubbing with 99% surfactant recycle is an order of magnitude
higher than that of the most expensive conventional method.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Held/Group
Air Pollution
Scrubbers
Foam
Dust
Iron Oxides
Aerosols
Surfactants
Air Pollution Control
Stationary Sources
Foam Scrubbers
Particulate
13B
07A
11G
07B
07D
11K
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PACKS
78
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
C-2
------- |