EPA-650/2-74-108
OCTOBER 1974
Environmental Protection Technology Series
• .- •_• •. .'. .*.*.•.v.v.*.' . . - •
S^li-:?^^??^
-------
EPA-650/2-74-108
WET SCRUBBER LIQUID UTILIZATION
by
Kunrad Semrau and Clyde L. Witham
Stanford Research Institute
333 Ravenswood Avenue
Menlo Park, California 94025
Contract No. 68-02-1079
ROAP No. 21ADJ-005
Program Element No. 1AB012
EPA Project Officer: Leslie E. Sparks
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
October 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
A bench-scale experimental investigation was made of the ways in
which energy consumption, liquid-to-gas ratio, and condensation of water
vapor affect the collection efficiency of a conventional cocurrent-
contact particulate scrubber of the orifice type. The test aerosol was
composed of spherical particles of ammonium fluorescein with a mass-
median diameter of approximately 0.6 micron. The collection efficiency
of the scrubber was determined and correlated as a function of the effec-
tive friction loss across the scrubber. In the absence of condensation
or evaporation effects, the scrubber efficiency was dependent only upon
effective friction loss, with no independent influences of gas velocity,
liquid-to-gas ratio, or contactor orifice size, except possibly in the
range of very low gas velocities, liquid-to-gas ratios, and pressure
drops. Evaporation of large amounts of water reduced collection effi-
ciencies, but scrubbing hot, humid gas with cold water to produce con-
densation significantly increased the collection efficiency. The most
favorable results were obtained by presaturating the hot, humid gas
stream before contacting it with cold water.
This report was submitted by SRI in fulfillment of Contract No.
68-02-1079, under the sponsorship of the Environmental Protection Agency.
Work was completed as of July 13, 1974.
iii
-------
CONTENTS
ABSTRACT lii
LIST OF FIGURES vi
LIST OF TABLES ix
ACKNOWLEDGMENTS x
I SUMMARY AND CONCLUSIONS 1
II RECOMMENDATIONS 5
III INTRODUCTION 7
IV BACKGROUND 9
V METHOD OF APPROACH 14
VI EXPERIMENTAL EQUIPMENT AND PROCEDURES 19
Bench-Scale Scrubber System 19
Aerosol Generation 22
Aerosol Sampling 28
Aerosol Analysis 30
General Test Procedure 31
VII RESULTS AND DISCUSSION 33
Aerosol Generation and Characterization 34
Ambient Scrubbing Tests 41
Adiabatic Saturation Tests . . . . 54
Condensing Tests 62
Vaporizing Tests 69
General Discussion 74
REFERENCES 86
GLOSSARY 87
CONVERSION FACTORS 90
iv
-------
Contents (concluded)
APPENDIX A LITERATURE REVIEW 91
REFERENCES . „ 99
APPENDIX B ANALYSIS OF EXPERIMENTAL DATA 102
APPENDIX C POTENTIAL COMMERCIAL APPLICATION OF CONDENSATION
SCRUBBING 109
TECHNICAL REPORT DATA 116
-------
FIGURES
1 Flowsheet of Experimental Scrubber System 20
2 Electron Micrograph of Aerosol 27
3 Scrubber Performance at Fixed Operating Conditions
(Tests EA-26 through EA-47) 38
4 Scrubber Performance Curves for Ambient Scrubbing Tests
EA-6 through EA-22 (Aerosol Generator at 35 C) 39
5 Scrubber Performance Data for Ambient Scrubbing Tests
EA-48 through EA-128 46
6 Scrubber Performance Curve for Ambient Scrubbing Tests—
Aerosol Generation Rates 350-570 Mg/min 47
7 Scrubber Performance Curve for Ambient Scrubbing Tests—
Aerosol Generation Rates 750-1000 |jg min 48
8 Scrubber Performance Curve for Ambient Scrubbing Tests—
Aerosol Generation Rates 1000-1200 pg/min 49
9 Scrubber Performance Curve for Ambient Scrubbing Tests—
Aerosol Generation Rates 1200-1500 |jg/min 50
10 Scrubber Performance Curve for Ambient Scrubbing Tests—
Aerosol Generation Rates 1500-1900 (Jg/min 51
11 Summary of Scrubber Performance Curves for Ambient
Scrubbing Tests 52
12 Correlation of Scrubber Pressure Drop Data for Ambient
Scrubbing Conditions 55
13 Scrubber Performance Curve for Adiabatic Saturation
o
Scrubbing—Aerosol A. Gas Temperatures: Inlet = 128.9 C;
Outlet = 71.7 C 60
vi
-------
Figures (continued)
14 Scrubber Performance Curve for Adiabatic Saturation
o
Scrubbing—Aerosol B. Gas Temperatures: Inlet = 128.9 C;
Outlet = 71.7°C 61
15 Scrubber Performance Curves for Condensation Scrubbing
(Series I). Gas Temperatures: Inlet = 130.6 C; Outlet =
54.4°C 65
16 Scrubber Performance Curves for Condensation Scrubbing
(Series II). Gas Temperatures: Inlet = 129.4 C; Outlet =
37.8°C 66
17 Scrubber Performance Curve for Condensation Scrubbing
(Series I and II)—Aerosol A ..'...' 67
18 Scrubber Performance Curve for Condensation Scrubbing
(Series I and II)—Aerosol B 68
19 Scrubber Performance Curve for Condensation Scrubbing
(Series III). Gas Temperatures: Inlet = 80.0 C; Outlet =
37.8°C 71
20 Scrubber Performance Curves for Condensation Scrubbing of
o o
Gases at Inlet Temperatures of 80 and 130 C—Aerosol B . . 73
21 Scrubber Performance Curve for Vaporization Scrubbing.
Gas Temperatures: Inlet = 125.6 C; Outlet = 58.3 C . . . . 75
22 Summary of Scrubber Performance Curves for Adiabatic
Saturation and Condensation Scrubbing—Aerosol A 76
23 Summary of Scrubber Performance Curves for Adiabatic
Saturation, Condensation, and Vaporization Scrubbing—
Aerosol B 77
24 Performance of Condensation Scrubbers—Literature Data
(Figure 12, Ref. 6A). Range of Present Investigation
Added 98
25 Analysis of Residual Deviations—Ambient Scrubbing Tests—
Aerosol Generation Rates 750-1000 pg/min . . 103
vii
-------
Figures (concluded)
26 Analysis of Residual Deviations — Ambient Scrubbing Tests —
Aerosol Generation Rates 1000-1200 (j,g/min ......... 104
27 Analysis of Residual Deviations — Ambient Scrubbing Tests —
Aerosol Generation Rates 1200-1500 Vg/min ......... 105
28 Analysis of Residual Deviations — Ambient Scrubbing Tests —
Aerosol Generation Rates 1500-1900 M-g/min ......... 106
29 Analysis of Residual Deviations — Adiabatic Saturation
Scrubbing — Aerosol A .................... 108
viii
-------
TABLES
1 Summary of Average Scrubber Test Conditions 16
2 Scrubber Performance Data for Ambient Scrubbing Tests—
o
Aerosol Generator Temperature 35 C 35
3 Scrubber Performance Data for Ambient Scrubbing Tests at
Fixed Scrubber Operating Conditions 37
4 Scrubber Performance Data for Ambient Scrubbing Tests ... 43
5 Scrubber Pressure Drop Data for Ambient Scrubbing Tests . . 56
6 Scrubber Performance Data for Adiabatic Saturation
Scrubbing Tests 59
7 Scrubber Performance Data for Condensation Scrubbing Tests—
Series I 63
8 Scrubber Performance Data for Condensation Scrubbing Tests—
Series II 64
9 Scrubber Performance Data for Condensation Scrubbing Tests—
Series III 70
10 Scrubber Performance Data for Vaporization Scrubbing Tests. 73
11 Selected Comparisons of Scrubber Performances at Different
Operating Conditions 78
ix
-------
ACKNOWLEDGMENTS
Mr. Charles E. Lapple, Senior Scientist, provided major assistance
in the development of the aerosol generation system and the design and
procurement of the bench-scale scrubber system, and valuable advice on
the conduct of the investigation. Mr. William W. Kerlin, Senior Chem-
ical Engineering Technician, provided valuable assistance throughout
the design, procurement, and assembly of the system, and in the sub-
sequent conduct of the experimental work.
The authors wish to express their appreciation for the assistance
and cooperation of the Project Officer, Dr. Leslie E. Sparks of the
Particulate Technology Section, Control Systems Division of the
Environmental Protection Agency.
-------
SECTION I
SUMMARY AND CONCLUSIONS
An experimental investigation was made under laboratory conditions
of the performance of a bench-scale orifice scrubber that was generally
a replica of many commercial scrubbers of the orifice or modified
venturi types. Studies were made of the effects of gas pressure drop,
gas velocity, liquid-to-gas ratio, and vaporization and condensation of
water vapor. Tests were made at both ambient and elevated temperatures.
All the tests were made with nominally the same test aerosol,
which was prepared by nebulizing a solution of ammonium flucrescein and
drying the mist particles in a stream of heated air. This procedure
produced a spherical, nonhygroscopic aerosol composed almost entirely of
submicron particles and having a mass median particle size on the order
of 0.6 micron. The performances of the scrubber under various operating
conditions were compared on the basis of gas pressure drop (or, more
strictly, of "effective friction loss"), and the influences—if any—of
other operating variables were inferred from differences in the relation-
ship of collection efficiency to the gas pressure drop.
In fact, the particle-size characteristics of the aerosol generated
varied sufficiently to introduce another variable into the experiments,
complicating the interpretation of the results. Fortunately, it was
found that a mass generation rate of the aerosol generator was a reason-
ably good correlating factor for the apparent particle-size character-
istics of the aerosol, and it was therefore possible to correlate the
scrubber collection efficiency data and to interpret their significance.
It appears that the functional relationship between efficiency
(actually expressed in terms of transfer units) and the effective friction
1
-------
loss is probably a more sensitive indicator of relative aerosol
particle-size characteristics than any of the available methods for
measurement of particle size in the submicron region.
Scrubber efficiency tests were carried out under the following
conditions:
» Scrubbing of ambient air with water at ambient temperatures
Scrubbing of hot, humid air (32-35% by volume of water vapor)
with hot water at the temperature of adiabatic saturation of
the gas stream. This condition simulated the common practice
of scrubbing gases with recirculated water, resulting in
essentially adiabatic saturation of the gas streams
. Scrubbing of hot, humid air (32-35% by volume of water vapor)
o o
with cold water to reach final gas temperatures of 54 C and 38 C,
resulting in condensation of most of the water vapor initially
present
. Scrubbing of hot, humid, nearly saturated air (32-35% by volume
of water vapor) with cold water to reach a final gas temperature
of 38°C
. Scrubbing of hot, dry air with hot water, resulting in flashing
and evaporation of part of the water and humidification of the
air
Results of the foregoing tests yielded the following conclusions:
. Over very wide ranges of gas velocity, liquid-to-gas ratio, and
gas pressure drop, the collection efficiency of the scrubber was
dependent only upon effective friction loss (essentially equal
to the gas pressure drop) and independent of gas velocity and
liquid-to-gas ratio (except as these latter factors affected
gas pressure drop). (This conclusion did not apply in cases where
-------
vaporization or condensation effects were involved.) There may still
be some independent effects of gas velocity or liquid-to-gas ratio
at very low values of these variables, but if so, the operational
region is not one of much practical significance.
The size of the orifice gas/liquid contactor had no effect on
collection efficiency, independent of effective friction loss.
• Condensation scrubbing of the test aerosol produced significant
but not very large increases in collection efficiency at a
given effective friction loss. The results were qualitatively
similar but quantitatively much less than indicated by earlier
pilot-plant studies of condensation scrubbing of Kraft recovery
furnace fume under similar operating conditions.
o
• Cooling the outlet gas below 54 C did not increase the condensa-
tion scrubbing effect measurably, presumably because only a
small additional quantity of water vapor was condensed•
Adjusting the inlet gas to a. nearly saturated condition before
contacting with cold water markedly improved the effectiveness
of condensation scrubbing.
• Vaporization of scrubbing water tended to reduce scrubber col-
lection efficiency, particularly in the lower range of effective
friction loss.
• In practical cases, condensation collection is not an alternative
to conventional high-energy scrubbing, but is a supplement that
can reduce the scrubbing energy requirements to a greater or
lesser degree. Promoting diffusional mass transfer (such as
vapor condensation) in practical equipment also requires energy
for gas/liquid contacting.
-------
Because of the requirements for water cooling and possible gas
humidification, the net saving of energy in condensation scrub-
bing may not often outweigh the other associated costs.
Condensation scrubbing may be, and probably is, warranted under
the following conditions:
(a) Where the waste gas already has a high water vapor content
and there is use for low-level heat in the associated process
system (e.g., Kraft pulp mills).
(b) Where the gas must be cooled and dehumidified anyway, for
process reasons (e.g., cleaning of blast furnace gas; purifica-
tion and conditioning of feed gas to contact sulfuric acid
plants).
-------
SECTION II
RECOMMENDATIONS
We recommend that priority be given to the following topics for
extended study of scrubber performance characteristics:
* A study of the performance characteristics of the orifice
scrubber with aerosols of particle sizes ranging from about
2.0 microns down to about 0.1 or 0.2 micron. The aerosols
can be produced by the same general technique used in the
present investigation: producing a mist of ammonium fluo-
rescein solution and drying the droplets to leave solid
particles. Even though the aerosols will not be monodis-
perse, the median particle sizes can be varied over a
fairly wide range by changing the concentration of the solu-
tion. The practical lower particle size is set by the
coincident lowering of the quantity of aerosol. This pro-
gram should provide badly needed information about the
relationship of scrubber collection efficiency to both
contacting power (effective friction loss) and aerosol
particle size. We expect that it should also elucidate,
or help to elucidate, some currently unexplained phenomena
that sometimes appear in the scrubbing of very fine
aerosols—most specifically, a rapid falling off of collec-
tion efficiency in the low contacting power range.
. An investigation of the comparative performances of dif-
ferent classes of gas/liquid contactors as functions of
contacting power. The same test aerosol, or aerosols,
should be used throughout.
-------
Particular attention should be given to two classes of
contactors: (1) those deriving contacting power from the
energy of the liquid stream, and (2), those deriving contact-
ing power from a mechanically driven rotor. The tests should
aim at establishing unequivocally whether contacting power
so derived is fully equivalent to that derived from the gas
stream in the form of friction loss (or pressure drop).
It is possible that with mechanically driven rotors, a
larger part of the total power input can be effective as con-
tacting power than in the other classes of contactors. This
should afford saving in total power input if contacting power
is fully equivalent for all the types of contactors.
If a field demonstration of condensation scrubbing in a potential
commercial application is desired, we recommend that a study be
made of the scrubbing of Kraft recovery furnace fume. The test
3
scrubber unit should have a gas flow capacity of about 850 m /hr
o •> 3
(500 ft /min) or, at most, of 1700 m /hr (1000 ft /min), which
is small enough to permit ease of experimental operation, or
modification without inordinate costs or time requirements. The
fan capacity should be sufficient to handle scrubber pressure
drops up to 1800 mm WC (70 inches WC). Heat exchangers should
be included for cooling recirculated scrubbing water.
Two scrubber units should be provided in series, each with
gas/liquid contactor and liquid entrainment separator. The first
unit should be designed for low-energy contacting, primarily to
cool and saturate the gas by scrubbing with recirculated water.
The second scrubber should be equipped for high-energy contacting
with cold water or cold recirculated brine. The scrubbers should
be of the variable-orifice type with cyclone entrainment separa-
tors. However, there should be provisions for substituting a
pipeline contactor for the variable orifice.
6
-------
SECTION III
INTRODUCTION
GENERAL
In the operation of particulate wet scrubbers, the collection of
progressively finer particles and an increase in the efficiency of col-
lection of particles of any size are both associated with increased
power consumption. Current trends in the requirements for control of
particulate air pollutants emphasize both the collection of fine parti-
cles (primarily those in the submicron range) and higher efficiencies
in the collection of particles of all sizes. It can therefore be anti-
cipated that future particulate scrubber installations will have increas-
ing power requirements. Hence, it is important to investigate any
possible means for minimizing those power requirements.
OBJECTIVES
The broad objective of this investigation was to study the ways in
which energy consumption, liquid-to-gas ratio, and condensation of water
vapor affect the collection efficiency of particulate scrubbers. The
original contract called for emphasis on the effects of liquid rate and
energy consumption. However, at the request of EPA, the emphasis of the
program was shifted to the effects of water vapor condensation on the
collection efficiency of conventional scrubbing equipment.
The revised scope of work included the following tasks:
(1) Review and assessment of the published literature and avail-
able unpublished information on the effects of liquid-to-gas
ratio; method of water injection, energy consumption, and
water condensation and evaporation on the efficiency of
particulate scrubbers.
7
-------
(2) Conduct of a bench-scale experimental study of particle
collection of a conventional coccurrent-contact particu-
late scrubber of the modified venturi or orifice type.
Emphasis was to be placed on enhancing the collection effi-
ciency by use of condensation techniques.
(3) Recommendation of a potentially suitable commercial appli-
cation for condensation-scrubbing techniques, using basic-
ally conventional scrubbers of the venturi or similar types.
(4) Development of a plan for a pilot-plant study of the scrub-
bing process in the recommended commercial application,
including preparation of a process design, an experimental
program, and a cost estimate.
8
-------
SECTION IV
BACKGROUND
It is now widely, though not universally, accepted that an increase
in the efficiency of particulate scrubbing is associated with an in-
crease in power consumption, and that the power consumption associated
with attainment of a given collection efficiency increases with decrease
in the particle size of the particulate matter. Recognition and accept-
ance of these facts, and use of them in design procedures, have formed
the basis for the recent greatly increased use of scrubbers for col-
lection of fine particles. Relatively high power consumption is thus
characteristic of scrubbers, although they may exhibit definite ad-
vantages in other respects for numerous applications. As requirements
grow for higher collection efficiencies on finer aerosols, the associated
increases in scrubber power consumption are a growing economic burden,
and it becomes increasingly important to seek ways to decrease the power
consumption if such can be found.
The number and diversity of scrubber designs is such as to defy
any entirely self-consistent or truly descriptive system of classifi-
cation, and it is even more difficult to try to describe the various
devices in terms of performance mechanisms. However, since power con-
sumption is a major characteristic of scrubbers, it is possible to
characterize various units according to the way in which power is
applied to the scrubbing process.
There are basically three ways of applying power in scrubbers:
(1) From the energy of the gas stream
(2) From the energy of the liquid stream
(3) From a mechanically driven rotor.
-------
When the energy is derived from the gas stream, the amount consumed
takes the form of gas pressure drop, and must be supplied by a fan.
Energy may be supplied to a liquid stream by three methods: (1) apply-
ing pressure to the liquid stream, (2) applying to the liquid the
energy of a compressed gas (e.g., compressed air, stream), and (3) apply-
ing to the liquid energy from a mechanically driven spray generator.
The net effect in each of the three cases is that kinetic energy is
imparted to the liquid stream being injected into the gas stream being
scrubbed.
In a mechanical scrubber, both the gas and liquid streams pass
through the rotating member. This member, which may be a modified
fan, frequently serves also to provide draft for the scrubber. In such
case, the power input to the drive shaft can be considered to have
two components, only one of which is related to scrubber efficiency.
It should be noted that the power consumption that is directly
related to scrubber efficiency is not the gross power input to the
scrubber, which may include power consumption due to motor, pump, or
fan inefficiencies, or power input that leads to increases in the kin-
etic energy or pressure head of the gas stream. The criterion of
scrubber efficiency is that portion of the power consumption that is
consumed in the gas-liquid contacting process and is ultimately con-
verted to heat through turbulence; this is termed "contacting power".
(A review of the pertinent background literature is presented in Appen-
dix A.)
The "contacting power rule" proposes that energy applied to the
scrubbing process by any of the three basic methods is equivalent,
and that the efficiency obtained on a given aerosol at a given contact-
ing power should be essentially independent of the scrubber design and
of the way, or combination of ways, in which the contacting power is
applied. A very substantial body of information that supports this
10
-------
hypothesis has now been accumulated, particularly for the types of
scrubbers in which contacting power takes the form of gas pressure
drop. There is considerable evidence to indicate that contacting power
derived from spray nozzles is also equivalent. The equivalence of
contacting power from mechanically driven rotors has not been fully
demonstrated, but is mostly inferred by analogy.
A scrubber design that could attain significantly higher effi-
ciency at a given power consumption than other competitive designs would
have a clear advantage. However, no such advantages for specific de-
vices have so far been clearly demonstrated, although they are frequently
claimed. A large part of the remaining uncertainty about the validity
of such claims results from the lack of test results for different de-
vices taken under precisely the same conditions, which are currently
the only available unequivocal basis for comparison.
One of the uncertainties in scrubber performance is the degree
to which liquid-to-gas ratio may affect efficiency, independent of
contacting power. It must be assumed that some lower liquid rate will
exist below which scrubbing efficiency will be lost. Some investi-
gators have reported that liquid-to-gas ratio has important—or, at
least, significant—effects on the performance of venturi scrubbers.
Others have found little or no independent effect. (See Appendix A.)
In the case of spray scrubbers using hydraulic spray nozzles,
the presumption is that a given contacting power and collection effi-
ciency may be attained alternatively by use of a relatively small
volume of spray at high pressure, or of a larger volume of spray at
correspondingly lower pressure. The possible limitations of this
hypothesis have not been systematically studied, although such data
as are available tend to confirm it.
The possibilities of channeling of liquid and gas flows appear to
11
-------
be much more serious in the case of some spray scrubbers, than in the
case of the venturi, orifice, or pipeline types. In the latter three
types, good liquid dispersion appears generally to be inherent, except
possibly at very low liquid rates or gas velocities. In the case of
spray scrubbers, it is possible to visualize arrangements in which
channeling effects might be severe.
The actual mechanisms through which power consumption leads to
particle collection can, for the present, only be inferred in very
general terms. It seems reasonable that the actual deposition of
particles involves primarily the collision of the particles with
spray droplets formed in the scrubber, and that the primary deposition
is inertial. The correlation of collection efficiency with power dis-
sipation strongly indicates that fluid turbulence governs the droplet
and particle motions that present the opportunities for collisions.
The observation that the contacting power-collection efficiency relatioi
ship is so similar even for radically different types of devices suggest
in addition that the scrubbing process is essentially the same in all
of the scrubbers. The dissipation of the power is evidently a gross,
overall measure of the atomization of the liquid and the turbulent
circulation of drops and particles in the gas stream.
The experience to date suggests that there is no way to increase
scrubber collection efficiency significantly at given levels of con-
tacting power so long as only the inertial mechanism can be used. It
may be practical under some circumstances to increase the scrubber col-
lection efficiency by superimposing additional particle collection
mechanisms on the process through use of condensation phenomena. One
such phenomenon is the condensation of moisture on the particles as
nuclei, which will increase the effective particle size and, hence,
the effectiveness of inertial deposition. The other known phenomenon
in this category is Stefan flow. Both of these phenomena have been
12
-------
demonstrated (see Appendix A). The major uncertainty has been in
the Quantitative increase in collection efficiency attainable by their
use, particularly in conventional scrubbing equipment. Improvements
in the efficiency of conventional equipment resulting from condensation
effects have been reported fairly frequently on a qualitative basis.
However, it has not often, been demonstrated clearly that part or all
of the reported effects may not have resulted from other, coincident
factors.
It has also been known from theoretical considerations and labora-
tory experiments that where a liquid is evaporating at a surface, the
resulting Stefan flow tends to inhibit particle deposition. Hence,
there should be a corresponding tendency to reduction of collection
efficiency in scrubbers in which the scrubbing liquid is vaporizing.
Apparently there has been no effort made prior to the present investi-
gation to determine how important this phenomenon may be in practical
scrubbing systems.
13
-------
SECTION V
METHOD OF APPROACH
The basic method of approach used in this investigation was em-
pirical, although it was guided qualitatively by theoretical considera-
tions. The particulate scrubbing process is obviously extremely com-
plex, and presenting an adequate and detailed picture of it is a highly
intractable problem from both theoretical and experimental viewpoints.
A number of workers have attempted to develop practical models of
scrubber operation, but have necessarily been obliged to simplify
the models and to introduce assumptions and estimates that are of more
or less questionable validity. In general, it has not yet been possible
to subject the models to adequately critical experimental verification.
So far, the correlation of the experimentally determined collectiol
efficiency as a function of contacting power has proved to be the most
reliable as well as the most practical approach to design of actual
installations. There has been no satisfactory experimental determinatii
of the relationship between collection efficiency and contacting power
for particles of specific, discrete sizes. Attempts to do this have
been largely frustrated by the problems of measuring the size and
quantity of particulates in a mixture of particles of different sizes,
and by the difficulties of generating particles of discrete and control*
led sizes. In fact, the collection efficiency/contacting power correlai
curve for a given scrubber appears to be a much more sensitive indicator
of the particle size and particle-size distribution of an aerosol—or
of changes in these quantities—than are the available methods for measii
ing low-micron and submicron particles. Of course, the scrubber does
not "measure" the particle size. However, if scrubber performance
curves can be established experimentally for discrete particle sizes,
14
-------
it should then be possible to make a reasonable quantitative or semi-
quantitative interpretation of the size characteristics of an aerosol
from the corresponding scrubber performance curve.
In the present investigation, the approach was to use a single
aerosol in tests of a bench-scale scrubber over a range of gas pressure
drops at various controlled operating conditions (see Table 1). For
each such operating condition, a correlation (graphical) was developed
between the collection efficiency (actually expressed as the number of
transfer units) and the contacting power. (See Glossary.) From compari-
son of the different correlation curves, the incremental effects of
condensation and vaporization were deduced.
In the general case, contacting power is best expressed in terms
4
of power per unit of volumetric gas flow rate. However, in many
scrubbers of the venturi or orifice types, including the one used in
this investigation, the contacting power is essentially equivalent to
the gas pressure drop, and can be expressed conveniently in the same
units (e.g., millimeters of water). However, the contacting power is
not in principle (or in fact, in some cases) necessarily equal to
4
the total pressure drop across a scrubber. To emphasize the distinction,
the quantity equivalent to contacting power has been termed the "effective
friction loss" in this report, even though it has been taken to be equal
to the gas pressure drop and has been expressed in the same units; mm
WG (millimeters of water column).
The average operating conditions used for the various scrubber
test series are summarized in Table 1. In the "Ambient" tests, ambient
air was in most cases scrubbed with water from the supply line, without
recirculation. In a few tests, the scrubbing water was recirculated.
The entering air was actually heated slightly above the ambient tempera-
ture by its mixing with the small flow of heated air containing the
aerosol (see Section VI). When the scrubbing water was recirculated,
15
-------
Table 1
SUMMARY OF AVERAGE SCRUBBER TEST CONDITIONS
Operating
Condition
Ambient
Adiabatic
saturation
Condensing
I
II
III
Vaporizing
Gas Temperatures
<°C and °F)
Inlet
25-40° C
77-1 04° F
128. 9° C
26 4° F
130. 6° C
267° F
129. 4° C
265° F
80.0°C
176°F
125. 6° C
258° F
Outlet
19-38° C
67-100°F
71.7°C
161° F
54.4°C
130° F
37.8°C
100°F
37 . 8° C
100°F
58. 3° C
137°F
Water Temperatures
(°C and °F)
Inlet
11-39°C
51-102°F
73. 9° C
165°F
13.3°C
56° F
12.2°C
54° F
12. 8° C
55° F
73. 3° C
164° F
Outlet
11-38° C
51-100°F
73.9°C
165° F
54.4°C
130° F
33 . 9° C
93° F
33.9°C
93° F
58,3°C
137°F
Gas Moisture Content
(g/g dry air)
In
0.006-0.010
0.317
0.327
0.315
0.322
0.007
Out
0.008-0.043
0.369
0.117
0.0404
0.0481
0.149
-------
it also was heated by contact with the warmer air as well as by the
turbulence generated in the water circulating pump. The gas and water
temperatures and the gas moisture contents are shown as ranges rather
than as averages.
For the test series designated as "Adiabatic Saturation", "Condensing ,
and "Vaporizing", the temperatures and gas moisture contents shown are
averages. The temperatures were actually read to 1° F, and the Fahren-
heit temperatures were converted into Centigrade values to the nearest
0.1° C. The tabulated Centigrade temperatures do not imply a precision
of measurements.
In the Adiabatic Saturation tests, an attempt was made to feed the
scrubbing water at a temperature as nearly equal as practical to the
adiabatic saturation temperature of the entering gas stream; the scrub-
bing process was thus intended to approximate an adiabatic humidification
(and saturation) of the gas stream, simulating the scrubbing of the gas
with recirculated water. The inlet water vapor content was selected
as representative of that in a considerable number of waste gases of
industrial importance.
In the Condensing tests, Series I and II, the inlet gas was nominally
the same as in the Adiabatic Saturation tests, but it was scrubbed with
sufficient cold water to condense water vapor and bring the exit gas
temperature down to desired levels. In Series III, the procedure was
the same except that the inlet gas was only of the order of 5 C
above the adiabatic saturation temperature, so that the tests approx-
imated scrubbing of a presaturated gas.
In the Vaporizing test series, the hot, dry inlet gas was scrubbed
with hot water. Upon injection into the hot gas, some of the water
flashed to vapor because of its level of sensible heat content. More
of it was thereafter evaporated by the hot gas. The Vaporizing tests
17
-------
were designed to indicate the possible extent to which particle col-
lection may be inhibited by Stefan flow effects.
18
-------
SECTION VI
EXPERIMENTAL EQUIPMENT AND PROCEDURES
BENCH-SCALE SCRUBBER SYSTEM
The bench-scale experimental scrubber system used in the investiga-
tion is portrayed schematically in Figure 1. This system was designed
to accommodate inlet gas flows of up to about 4200 liters/min (150 CFM)
o o o o
at temperatures of 20 -150 (70 -300 F), and with gas pressure drops
across the scrubber of up to 3000 mm WC (120 inches WC). The scrubber
itself consisted of a simple orifice contactor followed by a cyclone
entrainment separator. The orifice contactor was mounted in a horizontal
plane between flanges in a vertical line through which the gas and scrub-
bing water flowed downward. The gas and water leaving the orifice
contactor passed through a mitre elbow before flowing horizontally into
the cyclone separator. Interchangeable orifice plates were used to vary
the diameter of the contacting orifice. The scrubbing water was intro-
duced through a single tap in the wall of the inlet line, two pipe-
diameters above the orifice plate.
The scrubber design thus followed on a small-scale the general con-
figuration currently being used for most commercial scrubbers of the
orifice or modified venturi types.
Gas was drawn through the system by a rotary positive-displacement
blower of the Roots type. The main air flow through the system was drawn
directly from the room, metered, heated by a direct-fired gas burner,
humidified by steam injection, then drawn through the scrubber. The air
flow through the system was controlled by bleeding air from the room
directly to the inlet of the positive displacement blower. The air bypass
intake was fitted with a filter-silencer and the blower with a discharge
19
-------
Air
Natural
§ Ga,
Natural
Gas
Inlet Gas
Sample
Main
Air Flow
fit
[fll
Steam
I
T
BURNER
Aerotol
BURNER Orving
Air
AEROSOL
DRIER AND
MIXER
Compressed
Air
L
COARSE
SPRAY
TRAP
ULTRASONIC
NEBULIZER
Aerosol
Sweep
Air
Compressed
Air
Contacting
Orifice
MANOMETERS
w
Air
Bypass
ng
— 1 1 —
!» !
r i
\ 2
i ^ — *^*
M"M Outlet
^-^ Gas
Sample
CLONE (
_ Room
' Air
To
Stack
BLOWER
r—XI——-
Steam
STEAM
EJECTOR
To Water
Recycle Bypass
SA-2745-22
FIGURE 1 FLOWSHEET OF EXPERIMENTAL SCRUBBER SYSTEM
-------
silencer. However, the main air stream entering the system was not
filtered; since the test aerosol was measured by a specific fluorometric
method, the presence of atmospheric dust in the room air did not inter-
fere with the determinations of scrubber efficiency.
The scrubbing water system was so arranged that water could be sent
through the system in a single pass; in such a case the water supply
line pressure was usually sufficient to overcome line resistance without
use of the pump. Alternatively, the scrubbing water could be recircu-
lated by use of the water holding tank and the pump. When hot water
was used in scrubbing, the steam ejector could be used to heat the water
whether it was used in a single pass or was recirculated. The water in
the holding tank could be heated by recirculation through the steam
ejector and also by another steam siphon not indicated in Figure 1.
The scrubber and cyclone were elevated so that the water could be
discharged by gravity at the highest gas pressure drops used in the
tests. The water was discharged through a seal loop either directly to
the sewer or to the holding tank for recycle.
Contactor orifices of three different diameters were used in the
investigation: 2.54, 3.81, and 4.45 cm (1.0, 1.5, and 1.75 inches).
The line in which the contactor orifice was located was made from 2j-inch
stainless steel tubing. The cyclone separator, which was originally
built and used in an earlier investigation, was 45.7 cm (18 inches) in
diameter, with an inlet 7.6 cm (3 inches) in diameter and an outlet
10.2 cm (4 inches) in diameter. It was actually larger than necessary
to handle the gas flows used in the present investigation. The general
3
design was the same as that used by Lapple and Kamack for a semiworks
scrubber unit. The main inlet air header, including the orifice meter
section was made from 15.2-cra (6-inch) stainless steel tubing.
The main air flow was measured with a standard-design orifice meter
equipped with square-edged orifices and fitted with radius taps located
21
-------
1 pipe diameter upstream and \ pipe diameter downstream of the orifice
plate). Steam flows were measured with critical-flow orifice meters,
with Bourdon gages used to measure the steam pressures. The water flowi
were measured with rotameters of appropriate ranges.
System pressures and pressure drops were measured with U-tube mam*
meters filled with xylene or mercury, as demanded by the ranges of pres-
sures involved. The gas and water temperatures were measured with dial
thermometers of the bimetal type.
AEROSOL GENERATION
The aerosol generation system is also indicated schematically in
Figure 1. The original cloud of droplets was generated from a solution
with an ultrasonic nebulizer. A small flow of air swept the mist upward
through the chimney above the nebulization chamber and through a trap fd
removal of any residual coarse spray droplets. The fine mist then rose
straight upward through a section of 1-inch IPS pipe and entered the lot
end of a conical diffuser section which was located concentrically with!
the 15.2-cm (6-inch) riser that carried the main airflow upward to the
scrubber. Another small flow of heated drying air also entered the lowd
end of the conical diffuser section through the annular space around thd
pipe carrying the mist. The mist and heated air mixed with the diffuse*
section, where the water was evaporated from the droplets, leaving solid
aerosol particles. The resulting aerosol stream emerging from the uppei
end of the diffuser was then mixed with the main air stream, which flowe
through the annular space between the upper end of the diffuser and the
6-inch riser.
The sweep air to the nebulizer was taken from a compressed air line
3
and metered to a constant 14.2 liters/min (0.5 ft /min) in all tests
with a critical-flow orifice. The aerosol drying air was also metered
22
-------
with a critical-flow orifice to a constant rate of 84.9 liters/rain
3
(3 ft /rain). It was heated by a small direct-fired gas burner to a
o
temperature in the range of 170-200 C. The aerosol generation system
was developed in separate but parallel and continuing effort supported
by Stanford Research Institute. An ultrasonic type of nebulizer was
chosen to produce the original mist of solution droplets, because previous
experience at Stanford Research Institute and elsewhere had indicated
that the ultrasonic device produced a narrower band of droplet sizes
(with a median diameter of about 6 microns) than other types of atomizers,
and should therefore produce a more uniform aerosol. The ultrasonic
nebulizer is used commercially to prepare mists of therapeutic agents in
i fi
a particle-size range for effective inhalation therapy. ' It has also
been employed to prepare sodium chloride aerosols for testing respirator
2
filters, a use comparable to that in the present investigation.
The basic element of the generator system was a commercial ultra-
sonic nebulizer (DeVilbiss Model 35A). The original plastic nebuliza-
tion chamber and chimney were replaced by one made from stainless steel.
The height of the chimney was increased from that of the original to
provide a greater space in which oversized spray droplets could disengage
from the air stream and settle back into the liquid sump. The solution
to be nebulized was supplied to the nebulization chamber from an over-
head reservoir fitted with a gravity feed system so vented as to maintain
a constant liquid level in the chamber (not indicated in Figure 1).
Reports by other users and some preliminary experiments had indica-
ted that increases in the temperature of the liquid being nebulized also
increased the output of the generator. After some of the preliminary
experiments with the scrubber had been carried out, it was decided to
stabilize the temperature of the liquid at a temperature somewhat above
any room temperatures expected to be encountered. The metal chimney of
the nebulizer was therefore wrapped with heating tape, and the current to
23
-------
the tape was regulated by a controller operating off the output of a
thermistor immersed in the liquid in the nebulizer. The liquid tempera-
ture was set at 35 C. It was then found that operating at the higher
fixed temperature neither stabilized the output of the generator nor
clearly increased it (see Section VII). The heating system was there-
fore removed and the generator was operated thereafter at room temperature,
In another attempt to stabilize the generator output, a harmonically
compensated voltage regulator was inserted in the electrical power supply
line to the generator. The voltage regulator was continued in use, but
apparently had no effect in stabilizing the generator output.
Despite the attempts to stabilize generator output by fixing the
temperature and the level of the liquid in the generator and by regula-
ting the voltage of the electrical supply, fluctuations and shifts in
the generator output continued. The nature and consequences of these
changes are discussed below in Section VII.
The maximum liquid nebulization rate for the generator was about
3
3 cm /min. Since it was desired to produce a well-dispersed aerosol
with a mass-median size of about 0.5 - 0.8 micron, it was necessary to
use a dilute solution of the aerosol material. Hence, the concentration
of the aerosol in the scrubber air stream was necessarily low. This
condition was desirable for minimizing flocculation of the aerosol, but
also required a sensitive method of analysis if the necessity of long
sampling tiroes to collect large samples was to be avoided. For this
reason, a fluorescent dye was highly advantageous. Reports in the
literature Indicated that ammonium fluorescein produced a spherical,
7 8
nonhygroscopic, and readily analyzed aerosol. '
In a preliminary investigation a 10% solution of ammonium fluores-
cein was nebulized in the ultrasonic nebulizer and the droplets were
dried in an air stream to produce dry particles which were collected on
24
-------
a membrane filter and examined under a light microscope. The particles
were easily discerned to be spherical, and showed no sign of hygro-
scopicity when left exposed to atmospheric air. The particle sizes fell
mostly in the range of 1 to 5 microns with an apparent mass-median
diameter of about 3 microns (as judged by visual observation in the
microscope). From these observations it was judged that nebulizing a
0.1% ammonium fluorescein solution should produce an aerosol with a mass-
median diameter of about 0.6 - 0.8 micron that would be suitable for the
scrubber testing program.
The ammonium fluorescein solution was first prepared at a concentra-
tion of 1% by weight. The appropriate amount of fluorescein (acid form)
was weighed out and dissolved in a quantity of ammonium hydroxide solu-
tion equivalent to four moles of ammonium ion per mole of fluorescein
(acid form). Water was added to make up the solution to the appropriate
total weight (and volume). This stock solution was stored in glass and
in darkness. The 0.1% solution for experimental use was made up from
the stock solution by dilution.
One possible factor in the variation in aerosol generation rates,
which would also affect particle size, is concentration of the solution
in the nebulizer by evaporation. The sweep air to the nebulizer, taken
from a compressed air source and not humidified after expansion, was dry.
Although it had been expected that most of the moisture evaporated into
the sweep air would be absorbed from the aerosol mist being formed, it
was later found that the concentration of ammonium fluorescein in the
nebulizer solution did increase, demonstrating that at least part of the
water was evaporated from the liquid remaining in the nebulizer. It was
estimated that the concentration in the nebulizer should reach an equi-
librium level about 30% above that of the feed solution, assuming that
the liquid remained at room temperature. When analyses revealed that
the nebulizer liquid was being concentrated, some values approximating
25
-------
the estimated equilibrium concentration were obtained. Thereafter,
each day's operation was started with fresh feed solution in the
nebulizer.
Some of the variation in the aerosol generation rate may have been
related to changes in the solution concentration, but the pattern of
the variations did not correspond to the reasonable course of gradual
concentration of the solution. A series of measurements of aerosol
generation rate was also made with known nebulizer solution concentra-
tions ranging up to 16% above the nominal feed concentration. (This
work was done as part of the Institute-supported program on aerosol
generation.) The observed variations in generation rate were much
greater than could be accounted for by the changes in the nebulizer
solution concentration and did not, in fact, appear to be significantly
correlated with them.
Overall, it does not appear likely that variations in the concen-
tration of the nebulizer solution were responsible for more than a minor
part of the observed variations in the aerosol generation rate.
Near the end of the investigation, when the generator was producing
"Aerosol B" (see Section VII), samples were taken of the aerosol col-
lected on Nuclepore membrane filters with 0.4-micron pores. A gold-
palladium film (less than 100 Angstrom units thick) was vacuum deposited
on the filter to provide a conducting surface, and photomicrographs were
obtained with a scanning electron microscope. Photomicrographs of one
sample taken at magnifications of 5,000 X and 10,000 X are shown in
Figure 2. These show a range of particle diameters from about 0.2 to
0.8 micron, with a probable mass median of about 0.6 micron. Examina-
tion of the same membrane under a light microscope, covering a very much
larger field, indicated the presence of a small number of larger Particles
up to perhaps 1.5 microns in diameter. These few large particles appa-
rently contributed only a small fraction to the total size distribution
by mass.
26
-------
Id)
(b)
SA-2 746-20
FIGURE 2 ELECTRON MICROGRAPH OF AEROSOL
27
-------
At the time that the photomicrographs of Figure 2 were taken, it
was known that the particle size of the aerosol produced by the genera-
tor was variable, but the variations were probably too small to be
readily detectable by any of the available methods of size measurement.
They were, however, revealed by the performance of the scrubber. The
photomicrographs do illustrate that the particle size of the aerosol was
in the expected range.
AEROSOL SAMPLING
The aerosol was sampled at the inlet and outlet of the scrubber
(see Figure 1), using 25-mm Nuclepore membrane filters in Nuclepore
holders. The membrane holders were made of polycarbonate plastic, for
which the upper limit of operating temperature was about 140 C (285°F).
The inlet piece of the filter holder was machined out to expose almost
the entire effective diameter of the membrane directly to the incoming
gas. The holders were turned to face directly into the gas stream
being sampled, but no effort was made to sample isokinetically, because
the aerosol was too small for sampling to be appreciably affected by
inertial effects.
Initially, tests were made with two filter discs placed in series
in each holder, with both discs being analyzed. Tests were made with
membranes with 1.0, 0.8, 0.6, and 0.4-micron pores. The standard pro-
cedure finally adopted was to use a single disc with 0.6-micron pores to
provide an acceptable combination of efficiency and resistance to flow.
The filter holders were immersed in the gas stream to be sampled
so that the filter would be at the gas stream temperature and water vapor
condensation would be avoided until the gas sample had passed through
the filter. The gas stream was then drawn through a condenser for removal
of water vapor and cooled to ambient temperature before being metered in
28
-------
a critical flow orifice. Vacuum was supplied by a mechanical vacuum
pump. Sampling rates in the range of 2-10 liters/min were used, with
the highest rates at the scrubber outlet where the aerosol concentra-
tion was lowest.
The precision of aerosol sampling was checked by operating two
samplers in parallel at each of the sampling points. At the scrubber
outlet, 11 pairs of runs were made; the deviations between the paired
samples ranged from 1.1 to 8.9%, and averaged 3.5%. At the scrubber
inlet, only three pairs of runs were made; the deviations ranged from
1.6 to 12.5% and averaged 6.1%. Another four paired samples were taken
with one sampler operated at each position, inlet and outlet, and with
no water fed to the scrubber. In two of these paired runs, the measured
outlet aerosol concentration was higher than the inlet concentration.
The deviations ranged from 1.9 to 5.6%, with an average of 2.9%.
Evidently, the scrubber, operated dry, did not collect any of the
aerosol, at least within the precision of measurement.
Subsequently, material balance runs were made as part of an Insti-
tute-supported program for development and evaluation of the aerosol
generation system. The inlet and outlet air streams were sampled in the
usual manner, and integrated samples of the outlet scrubbing water were
also obtained. The quantity of inlet aerosol measured by sampling the
inlet air was compared with the sum of the quantities leaving the scrub-
ber in the exit air and collected in the water. This comparison involved
not only the precision of sampling but that of the measurement of the air
and water flow rates as well. Positive and negative deviations were
about equal in number, ranging from 0.9 to 16.7% with an average of 5.0%.
From these results, it appears that the average error in aerosol
sampling did not exceed 5 to 6% and that individual errors seldom exceeded
10%.
29
-------
AEROSOL ANALYSIS
Fluorlmetric analyses of the ammonium fluorescein aerosol material
were made with an Aminco-Bowraan spectrophotofluorometer, with a xenon
lamp (Catalog Numbers 4-8202 and 4-8202B). The unit was equipped with
monochrotneters for both exciting and emitted light. Instrument settings
used in the analyses were as follows:
Excitation wave length—482 nm
. Emission wave length—540 nm
. Primary slits—full open
. Secondary slits—5 mm
Photomultiplier slit—0.1 mm
. Meter multiplier setting—to give transmittance reading
between 30 and 9O*;
. Filter used—Corning 3-69 (to eliminate interferences
with excitation wave length)
The transmittance of the ammonium fluorescein solutions followed
_7 3 —10 3
Beers' Law at concentrations between 4 x 10 gn/cra and 1 x 10 gm/cm .
The fluorescence was markedly a function of pH, increasing by a factor
of 25 as the pH was increased from 4 to 10. Samples were dissolved in
a solution of ammonium hydroxide having an initial pH of 11.
3
Samples for analysis (1.5 to 2 cm ) were placed in a clean, dry
cuvette (fused quartz) of rectangular cross section.
It was found that a single washing of a N'uclepore filter was suffi-
cient to remove the sample deposit. Therefore, filters were simply
3
immersed and agitated in 10 cm of the ammonium hydroxide solution having
a pH of 11, If the resulting concentration of ammonium fluorescein was
above the preferred range for analysis, the sample was diluted 10 to 1
with the ammonium hydroxide solution of pH 11.
Experience with replicated analyses indicated that the error in
analysis did not exceed 1" except for the very smallest samples, for
which the error was somewhat larger.
30
-------
GENERAL TEST PROCEDURE
Scrubber performance tests were commonly made in groups, with the
scrubber and aerosol generator being allowed to operate continuously
between tests while the scrubber operating conditions were being changed.
The aerosol generator commonly required a warm-up period of about
15-20 minutes before the aerosol output (observed in the glass spray
trap) appeared visually to reach its maximum level. Except in a few
exploratory tests, the generator was adjusted to give its maximum output.
In general, it was relatively easy to adjust the scrubber operating
conditions except when it was desired to operate with the inlet gas near
saturation with water vapor. When hot water was used in scrubbing, the
water was frequently used on a once-through basis, with the temperature
being adjusted by addition of steam at the ejector. However, control of
the water temperature was somewhat more satisfactory when the water was
recirculated and its temperature was adjusted at the holding tank.
Before the start of a test, the sampling filter holders were pre-
heated, then inserted into the gas stream at the inlet and outlet of
the scrubber and permitted to come up to temperature with the filter
faces turned out of the gas stream. The holders were then turned to face
the filters into the gas stream, and the gas flow was started and stopped
simultaneously. Once the procedure was reduced to a routine, the sampling
period was usually five minutes. During this period, the quantities of
aerosol collected on the filters were mostly in the range of 0.1 to 15 ^g,
which could be analyzed with precision. Aerosol concentrations in the
inlet gas to the scrubber ranged from about 0.2 to 0.8 ^/liter of dry
air (25°C and 1 atm.), depending on the gas flow in the scrubber and on
the generation rate. All gas samples were referred to the selected
standard gas conditions: 25 C, 1 atm. pressure and the dry basis. The
corresponding aerosol concentrations measured at the inlet and outlet
of the scrubber were used to calculate the collection efficiency.
31
-------
During the period of the sampling run, readings were made of the
flowmeters and system temperatures and pressures.
Analyses of the ammonium fluorescein in the scrubbing water were
apparently interfered with by residual chlorine from the water purifi-
cation. Recirculating water through the scrubber with a flow of ambient
air apparently stripped out the interfering chlorine. Hence, where some
check runs were made at Ambient scrubbing conditions to determine
material balances on the ammonium fluorescein, a supply of water that
had been air-stripped to remove chlorine was first accumulated in the
holding tank and then pumped through the scrubber in a single pass during
the scrubbing tests.
Except in a few instances, no direct measurement was made of the
water vapor content of the air at the outlet of the scrubber. If the
gas was sampled for a sufficient period, the moisture content could be
determined from the amount of condensate collected in the condenser.
However, the standard sampling period was too short for collection of a
volume of condensate that could be measured with reasonable precision.
Hence, the outlet gas moisture content was calculated by assuming satura-
tion at the outlet gas temperature if the outlet water temperature was
equal to or greater than the outlet gas temperature; if the outlet water
was colder than the outlet gas, the gas was assumed to have a water vapor
partial pressure equal to the vapor pressure of the water at its outlet
temperature. Allowance was made in the calculation for the total pres-
sure (subatmospheric) at the outlet of the scrubber.
32
-------
SECTION VII
RESULTS AND DISCUSSION
As soon as the experimental equipment had undergone a preliminary
shakedown and appeared to be in suitable operating condition, a series
of "screening" tests was made to provide a quick indication of the range
of collection efficiencies that might be obtained under adiabatic satura-
tion, condensing, and vaporizing conditions. The scrubber was operated
at a single level of effective friction loss throughout the tests,
although setting the other operating conditions made it necessary to use
varying liquid-to-gas ratios in the different experiments. Two or three
replicate tests were made at each operating condition. (The results are
presented under the appropriate headings below.)
Although the aerosol generation rates observed during the tests
varied by as much as +35% of the mean value, the agreement between
replicate tests was generally very good or excellent, and it therefore
was considered that the variations in aerosol generator output probably
did not coincide with any substantial variations in the particle size of
the aerosol. However, it was recognized that such variations might take
Place. As additional data were obtained, difficulties in replicating
results appeared, and strong evidence of variations in the aerosol particle
size did emerge. Since the interpretation of scrubber performance at
various conditions was dependent on constancy of the aerosol character-
istics, it became necessary to devote a major effort to attempts to
stabilize the generator operation (see Section VI) and characterize the
aerosol. In the end, it was not actually found possible to prevent the
variations in the aerosol characteristics. Nevertheless, the observation
that these variations were correlated with the aerosol generation rate
33
-------
made it possible to select comparable collection efficiency data for
different scrubber operating conditions, and thus to interpret the
effect of the operating conditions on scrubber performance. However,
it was necessary to obtain many more data to establish the correlations
than was originally anticipated.
AEROSOL GENERATION AND CHARACTERIZATION
The measures taken in the effort to stabilize the performance of
the aerosol generator (see Section VI) apparently had essentially no
effect. There was a preliminary indication that raising the tempera-
o
ture of the liquid in the nebulizer to 35 C increased the aerosol
generation rate. Later, it became evident that the generation varied
over about the same, wide range whether the nebulizer was operated at
o o
35 C or at ambient temperatures (about 25 C). Surveyed over a period of
days, the generation rate appeared to follow no consistent or systematic
trends. The rate showed apparently random short-term variations about
the average value, and within periods of from several hours to a day or
more, shifts in the average generation rate appeared.
o
After the aerosol generator was equipped to operate at 35 C, a
series of scrubber efficiency tests was made at ambient operating con-
ditions (see Table 2), all In the course of one day. Tests were made at
a number of liquid and gas rates and gas pressure drops. All the aerosol
generation rates were high, ranging from 1747 to 4097 jjg/min., the latter
value being the highest observed during the entire investigation. Never
on. any other occasion were the generation rates uniformly so high, nor
wore so many very high individual values observed. The collection effi-
ciency data, studied in the relation to the effective friction loss,
showed a wide scatter.
34
-------
Table 2
SCRUBBER PERFORMANCE DATA FOR AMBIENT SCRUBBING TESTS
(Aerosol Generator Temperature 35 C)
Test
Code
No.
EA
6
7
8
9
10
11
12
13
14
15
16
17
IS
19
20
21
22
Aerosol
Generation
Rate
(tig/min)
1864
1747
1784
2743
2634
1898
1926
1889
2803
4097
2348
1988
2799
2138
3444
1749
1783
Contactor
Orifice
Size
(cm)
3.81
Inlet Gas
Flow Rate
(std llters/min)
1140
1130
1130
1130
2110
2120
2110
2110
2940
2770
2770
2770
2770
2770
3530
3520
3S20
Gas Moisture
Content
(g/g dry air)
In
0.01b
Out
0.008
0.010
0.013
0.008
0.008
0.008
0.008
0.009
0.010
0.009
0.008
0.008
0.008
0.008
0.008
0.009
0.010
Gas
Temperature
<°0
In
33.3
32 C
32 c
32 c
32 c
29.4
38.9
32.8
31.1
31.7
32.2
30.6
30,6
30.6
29.4
30.0
29.4
Out
25.6
27.8
28.9
26.1
22.8
22.2
22.2
23.3
26.7
23.3
22.8
22.2
21.1
20.6
20.6
22.2
22.8
Water
Temperature
C°0
In
10.6
14.4
18.9
11.7
11.7
10.6
11.1
11.1
13.9
13.3
11.7
11.7
11.1
10.6
11.1
12.2
13.9
Out
10.6
15.0
18.3
11.7
11.7
10.6
11.1
12.2
15.0
12.2
11.7
11.7
10.6
10.6
11.7
13.3
15.0
Gas
Velocity
in Orifice*
(m/sec )
16.7
16.5
16.5
16.5
30,8
31.0
30.8
30.8
43.0
40.5
40.5
40.5
40.5
40.5
51.6
51.5
51.5
Liquid-to-
Gas Ratio
/liters\
(•*)
25.5
42.8
42.8
57.2
30.7
19.1
9.53
5.40
1.29
2.05
5.60
5.52
10.3
16.9
4.50
1.56
0.699
Effective
Friction
Loss
(mm WC)
274
535
535
810
1480
870
394
272
259
279
535
535
810
1380
843
429
325
Collection
Efficiency
(%)
56.3
69.7
72.3
90.63
96.95
88.9
65.6
52.1
65.2
78.3
82.9
76.5
92.01
96.19
93.90
68.9
57.3
Transfer
Units
(Nt)
0.828
1.19
1.28
2.37
3.49
2.20
1.07
0.736
1.06
1.53
1.77
1.45
2.53
3.27
2.80
1.17
0.851
w
en
Calculated at standard gas conditions.
b
Estimated.
Estimated. Inlet scrubbing water apparently splashed on thermometer.
-------
The scrubber system was then used to study changes in the aerosol.
The scrubber was operated on ambient air at a fixed set of conditions—
air rate, water rate, and gas pressure drop. Repeated determinations
were made of the collection efficiency, with the aerosol generator
operated at ambient temperatures and at 35 C (see Table 3). The measured
aerosol generation rates ranged from 662 to 1705 pg/min., and the collec-
tion efficiencies from 67.3 to 77.5% (1.12 to 1.49 transfer units). The
plot of transfer units vs. aerosol generation rate (Figure 3) suggested
a trend to increasing collection efficiency with increase in the aerosol
generation rate, which inferred that the increase in the generation rate
coincided with an increase in the average particle size of the aerosol.
The data shown in Figure 3, together with other observations in the
o
same period, indicated that operating the generator at 35 C neither
stabilized its output nor clearly increased the output above that obtained
during operation at ambient temperature. Therefore, use of the higher
temperature was abandoned, and all subsequent operation of the generator
was carried on at ambient temperatures.
Subsequent studies of scrubber performance at various operating
conditions (see below) developed additional strong evidence that the
variations in aerosol generation rate did coincide with changes in the
average particle size of the aerosol, and possibly—if not probably—in
the particle-size distribution as well. With the aid of this background,
the data from the early tests (Table 2) were re-examined and plotted in
groups according to relatively narrow ranges of aerosol generation rate
(see Figure 4). When treated in this manner, the data show correlations
that are actually quite striking in their excellence and are in agreement
with earlier experience that indicated that scrubber performance data for
a consistent aerosol can yield very close correlations. The performance
curves for the generation rates of 1864-2138 and 2348-2803 M€/min are
generally consistent with the trends in the performance data later obtained
36
-------
Table 3
SCRUBBER PERFORMANCE DATA FOR AMBIENT SCRUBBING
TESTS AT FIXED SCRUBBER OPERATING CONDITIONS3
Test
Group
No,
1
2
3
4
5
6
Test
Code
No.
EA
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Aerosol
Generator
Temperature
(°C)
Ambient
35
35
Ambient
Ambient
35
Aerosol
Generation
Rate
(jig/min)
710
662
688
715
1041
1024
1051
1029
1149
1168
1002
1023
1044
1276
1178
1136
1085
1128
1462
1705
1521
1514
Collection
Efficiency
(%)
70.4
67.3
67.5
68.4
69.5
68.9
69.3
69.3
69.9
70.4
70.9
70.2
68.8
74.6
77.5
72.5
71.5
71.9
71.8
77.0
74.0
73.9
Transfer
Units
(Nt>
1.22
1.12
1.12
1.15
1,19
1.17
1.18
1.18
1.20
1.22
1.24
1.21
1.17
1.37
1.49
1.29
1.25
1.27
1.27
1.47
1.35
1.35
Contactor orifice diameter = 3.81 cm.
Gas rate = 2800 std. liters/min.
Water rate = 15.1 liters/min.
Liquid-to-gas ratio = 5.39 liters/Sm
Effective friction loss = 520 mm WC.
37
-------
1.8
1.7
1.6
z"
I
1
v> 1.5
2
D
£T
in
& 1.4
1
u_
O
NUMBER
w
1.2
1.1
1.0
4(
1 1 1 1 1 I !
TEST GROUP GENERATOR
NUMBER TEMPERATURE (°C)
~ O 1 Ambient ~~
A 4 Ambient
D 5 Ambient
— • 2 and 3 35 —
• 6 35
~ O
•
— —
A
Q ~
a •
A °
3P *
o
00
i i i i I i i
X) 600 800 1000 1200 1400 1600 1800 2000
AEROSOL GENERATION RATE — pg/min
SA-2745-1
FIGURE 3 SCRUBBER PERFORMANCE AT FIXED OPERATING CONDITIONS
(Tests EA-26 through EA-47)
38
-------
10
S
z
2348-2803
—— Q 3444-4O97
t 1 I 1 1 I I I
100
1000
EFFECTIVE FRICTION LOSS — mm WC
torooo
SA-2745-23
FIGURE 4 SCRUBBER PERFORMANCE CURVES FOR AMBIENT SCRUBBING TESTS
EA-6 THROUGH EA-22
(Aerosol generator at 35°CJ
39
-------
in ambient scrubbing tests with lower aerosol generation rates. Since
only two data points were obtained in the range 3444-4097 pg/min, the
performance curve they yield is merely indicative, but is consistent with
the other observed trends.
Over the remaining course of the investigation, the generator output
at first continued to show not only random variations over relatively
short periods, but shifts in the average rate from day to day, or some-
times from one period of several hours to another. The corresponding
scrubber efficiency data showed a wide scatter. However, the generator
gradually reached a point at which the band of variation of the average
generation rates became relatively narrow, although random fluctuations
about the average values continued. There was a corresponding reduction
of the scatter in the scrubber efficiency data. When the generator set-
tled into a mode producing about 750 to 1100 ug/rain of aerosol, scrubber
efficiency data of generally high consistency were obtained. The corres-
ponding aerosol is termed "Aerosol A" for convenience.
Still later, the nebulizer apparently malfunctioned briefly, and
the aerosol generation rate thereafter dropped abruptly to about half
that observed previously, or about 350 to 700 pLg/roin. The "new" aerosol,
designated "Aerosol B," again gave scrubber efficiencies that were con-
sistent, but that were systematically lower than those obtained with
Aerosol A. These results, combined with further analysis of the results
obtained earlier, indicated clearly that the nature of the aerosol, as
well as the generation rate, had been shifting.
Some additional observations, made with Institute support since
termination of this experimental program, Indicated a continuing drop in
the average aerosol generation rate. It appears likely that this behavior
was associated with deterioration of the transducer, or aging of the trans-
ducer crystal, of the nebulizer. Replacement of the transducer restored
40
-------
the aerosol generation rate to the previous "normal" range of about 900
to 1300 pg/min,
AMBIENT SCRUBBING TESTS
The tests of scrubbing ambient air with ambient water had two major
objectives:
(1) to establish the contacting power/efficiency relationship for
the test aerosol in the absence of appreciable condensation or
evaporation of water
(2) To determine the degree to which liquid-to-gas ratio might
have an influence on collection efficiency, independent of
the contacting power.
Scrubbing at ambient temperatures involves only small absolute
changes in the humidities of the air. Strictly, the attainment of
"neutral" conditions with respect to vaporization and condensation phe-
nomena requires that saturated gas should be scrubbed with water at the
same temperature. In practice, this is experimentally difficult to carry
out without the danger of bringing in other, undesired phenomena. If the
incoming gas stream is saturated, even a small temperature drop may result
in some condensation of water vapor and possible growth of the test aerosol
upstream of the gas/liquid contacting section, with consequent changes in
the contacting power/efficiency relationship.
o o
Attempts to make neutral tests at 54 C and 74 C were made during the
initial series of "screening" tests. The inlet gases were actually brought
in at temperatures several degrees above the saturation temperature to
avoid the possibilities of premature condensation. For those few "neutral"
tests that were not actually abortive, the efficiencies obtained were in
the same range as those obtained at adiabatic saturation and other operating
conditions, but presented no consistent, interpretable picture. In retro-
spect, it appears that the results reflected variations in the aerosol,
41
-------
which were not fully appreciated at the time. Ambient scrubbing gave
results that were at least similar to those obtained with adiabatic
saturation scrubbing with a comparable aerosol. There was no clear indi-
cation that temperature as such had any appreciable effect on scrubber
o
performance over the range of about 25-80 C in which most of the scrubbing
action presumably took place under the various scrubbing test conditions.
Had the liquid-to-gas ratio been found to have independent effects
on scrubber efficiency, it would have made interpretation of test results
especially difficult, since securing different amounts of condensation
involved use of different liquid-to-gas ratios. The first ambient scrub-
bing tests were made systematically to display the effects of orifice
contactor size, orifice gas velocity, and liquid-to-gas ratio (see
Table 4). The evidence of aerosol particle-size variations (previously
discussed) appeared speedily in the form of a wide scatter in the col-
lection efficiency data (see Figure 5). Although the data points showed
a general correlation of transfer units with effective function loss, the
straight-line relationship suggested by previous experience was not in
evidence. However, there was no indication—within the uncertainty posed
by the scatter in the data—that there were independent effects of gas
velocity or liquid-to-gas ratio on the number of transfer units.
When evidence mounted that the apparent variations in aerosol particle
size were not purely random, but were correlated with the aerosol genera-
tion rate, the ambient scrubbing data of Table 4 were grouped according
to convenient ranges of aerosol generation rate and replotted (Figures 6
through 10). (The correlation curves from Figures 6-10 are presented
together without data points in Figure 11, for purposes of comparison.)
The ranges of generation rate were chosen on a largely arbitrary basis,
and slightly different choices might produce somewhat better correlations.
Nevertheless, it is evident that the data are generally quite well ration-
alized by this device. The poorest correlation is given by the data for
42
-------
Table 4
SCRUBBER PERFORMANCE DATA FOR AMBIENT SCRUBBING TESTS
Test
Code
No.
EA
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
Aerosol
Generation
Rate
1370
1366
1503
1497
1396
846
1358
1292
1280
1687
1148
1179
1569
1397
1503
1642
1440
1234
1063
1316
857
885
1771
1029
1115
1080
1004
978
1008
1317
Contactor
Orifice
Size
(cm)
3.81
2.54
Inlet Gas
Flow Rate
(std liters/in in)
1230
1740
2650
3640
1790
2640
3580
2650
1390
1770
850
864
1220
1770
2390
Gas Moisture
Content
(g/g dry sir)
In
O.009
0.009
0.010
0.010
O.008
0.010
0.010
0.010
O.010
0.010
Out
O.OO8
0.008
0.008
O.008
0.009
0.008
0.008
0.009
0.010
0.009
0.008
0.008
0.012
0.010
0.009
0.008
0.008
0.008
0.009
0.010
0.009
0.009
0.008
0.008
0.008
0.008
0.009
0.008
0.009
0.010
Gas
Temperature
In
33.3
31.1
32 b
32 b
31 b
31 b
31.1
32.2
35.0
31.7
26.7
26.1
27.8
27.8
27.8
26.7
2S.6
23.9
25.6
27.8
31.1
30.6
27.8
28 b
29 b
28.9
34.4
32.2
31.7
30.0
Out
25.6
25.6
25.0
25.6
25 .0
23.3
23.3
26.1
27.2
22.8
21.1
20.6
23.9
22.8
20.0
19.4
22.2
20.6
20.6
23.9
23.9
23.3
26.1
25.0
23.9
23.3
23.9
23.3
22.8
22.8
Water
Temperature
In
11.7
11.1
11.1
12.8
12.2
11.1
11.1
11.7
16.7
11.7
11.1
10.6
18.3
16.7
11.1
11.1
10.6
10.6
11.1
13.3
11.7
11.7
11.1
11.1
11.1
11.1
11.7
11.1
11.7
12.8
Out
11.1
11.1
11.1
12.2
12.2
11.1
11.1
11.7
14.4
11.7
10.6
10.6
16.7
14.4
11.7
11.1
10.6
11.1
12.2
14.4
11.7
11.7
11.1
11.1
11.1
11.1
11.7
11.1
11.7
14.4
Gas
Velocity
in Orifice
(m/sec)
18.0
25.4
38.7
53.2
26.2
38.6
52.6
38.7
20.3
25.9
28.0
28.4
40.5
41.0
4O.8
58.2
78.7
Liquid-to-
Gas Ratio
fl±tera\
( So.3 J
18.7
26.8
36.6
52.8
37.4
25.9
19.0
8.62
1.43
5.66
12.5
17.0
0.52
1.04
4.12
9.07
25.3
12.4
2.63
1.43
10.9
8.6
38.8
53.2
37.1
26,8
15.7
14.8
8,47
1,59
Effective
Friction
Loss
(mm WC)
228
345
516
833
1217
787
516
239
191
442
864
1176
282
345
767
965
802
870
517
191
151
223
1092
1707
2680
1722
963
2300
1247
1087
Collection
Efficiency
(5)
44.3
57.1
72.9
85.6
93.7
78.4
72.7
40.3
35.6
77.7
88.1
93.5
60.9
67.5
88.9
93.1
86.7
88.9
74.1
60.3
12.0
36.9
95.7
97.2
99.24
97.4
90.7
98.7
94.3
94.1
Transfer
Units
(N't)
0.585
0.846
1.30
1.94
2.77
1.53
1.30
0.515
0.441
1.50
2.13
2.74
0.940
1.12
2.20
2.67
2.02
2.20
1.35
0.923
0.304
0.458
3.14
3.57
4.88
3.66
2.37
4.31
2.87
2.83
CO
-------
Table 4 (Continued)
Test
Code
No.
EA
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
1O4
105
106
107
108
109
Aerosol
Generation
Rate
860
1053
1160
1150
1146
1159
976
1487
1165
1176
1212
1243
1278
52 5d
1866
1795
1585
1817
1675
1674
1044
984
1021
937
847
965
967
973
1052
797
789
766
Contactor
Orifice
Size
-------
Table 4 (Concluded)
Test
Code
No,
EA
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
Aerosol
Generation
Rate
(yg/min)
765
762
784
766
789
831
868
824
374
355
404
421
443
398
553
541
548
565
510
Contactor
Orifice
Size
(cm)
2.54
3.81
2.54
Inlet Gas
Flow Rate
(std liters/rain)
548
607
767
789
1473
1713
1967
2235
2031
2031
2031
2036
2025
3294
1150
1150
1153
1153
2111
Gas Moisture
Content
(g/g dry air)
In
0.01C
0.01C
0.01C
0.006
Out
0.009
O.OO9
0.010
0.010
O.010
0.010
0.010
0.010
0.010
0.010
0.010
0.009
0.009
O.O10
0.009
0.009
O.OO9
0.009
0.010
Gas
Temperature
C°0
In
38.9
40.0
39.4
38.9
32.8
32.2
32.2
31.7
28.9
28.9
28.3
27.2
25.0
27.2
31.1
31.7
25.0
27 b
29.4
Out
26.1
26.1
26.1
26.1
26.1
26.1
26.1
26.1
24.4
24.4
22.2
21.7
21.7
21.1
23.3
23.3
23.3
23.9
22.8
Water
Temperature
In
12.2
12.2
12.8
12.8
12.8
12.8
12.8
12.8
13.9
—
12.8
12.8
12.2
12.8
12.8
12.8
12.2
12.2
13.3
Out
12.8
12.8
13.3
13.3
13.3
13.3
13.3
13.3
14.4
13.3
13.3
12.8
12.2
13.3
12.8
12.8
12.2
12.2
13.3
Gas
Velocity
in Orifice
(m/sec)
18.0
19.9
25.2
26.0
21.6
25.0
28.8
32.7
29.7
29.7
29.7
29.8
29.6
48.1
37.8
37.8
37.9
37.9
69.4
Liquid-to-
Gaa Ratio
/liters\
V Sffi3 /
10.4
9.36
7.41
7.20
3.84
3.32
2.89
2.54
1.87
5.59
8.29
11.8
16.0
4.54
16.9
22.5
28.1
38.2
1.80
Effective
Friction
Loss
(mm WC)
81.3
97.7
137
161
77.8
103
130
164
115
216
293
444
588
572
897
1271
1587
2440
866
Collection
Efficiency
17.8
11.6
17.8
21.7
6.2
12.0
16.7
22.9
6.5
21.7
33.4
52.1
62.5
64.0
85.5
91.60
94.82
98.25
84.8
Transfer
Units
(Nt>
0.196
0.123
0.196
0.244
O.O64
0.127
0.180
0.260
O.O67
0.244
0.406
0.736
0.982
1.02
1.93
2.48
2.96
4.04
1.88
en
Calculated at standard gas conditions.
b
Estimated. Inlet water apparently splashed on thermometer.
Estimated.
Generator output deliberately attenuated.
-------
QC
Ul
IT
£
m
IU
1
0.1
001
- 1 1 1 1 1 1 1 1 | I 1 1 1 1 1 1 1
i * *S*
| 1 1 1 II
»
1 U
:
~ V^s^ ORIFICE ORIFICE ~
— M.7 O DIAMETER VELOCITY —
— o ^
So >
^> f^-A &
V
A<& O
E" 0
~ ° O A
I O
V
D
a
A
\ 1 1 1 1 1 1 ll 1 1 1 1 1 1 1 1
(cm) (mine)
2.54 35-38
46
58
18-28
40-79
3.81 18
26
30
39
53
20-48
4.45 21-29
I I 1 1 I
—
—
_
~
~^
1 1 1
10 100 1000 10,0i
EFFECTIVE FRICTION LOSS — mm
we
SA-2746-2
FIGURE 5 SCRUBBER PERFORMANCE DATA FOR AMBIENT SCRUBBING TESTS
EA-48 THROUGH EA-128
46
-------
10
in
z
3
cc
ui
<
X
CC
UJ
ea
5
0.1
100
/
/
/>
ORIFICE
DIAMETER
(cm)
> 2.54
0
3.81
1000
EFFECTIVE FRICTION LOSS — mm WC
ORIFICE
VELOCITY
(m/sec)
38
46
69
30
48
J L_i
10,000
S A-27 45-3
FIGURE 6 SCRUBBER PERFORMANCE CURVE FOR AMBIENT SCRUBBING TESTS
Aerosol generation rates 350-570 /jg/min.
47
-------
10
z
D
£
u.
O
(C
LU
CO
5
z
0.1
I I
I I I
ORIFICE
DIAMETER
(cm)
2.54
O
0
A
D
O
A 4.45
I I ] I I I I
ORIFICE
VELOCITY -H
35-38
58
26
67
26
53
20-33
27-29
100
1000
EFFECTIVE FRICTION LOSS — mm WC
10.000
SA-2745-4
FIGURE 7 SCRUBBER PERFORMANCE CURVE FOR AMBIENT SCRUBBING TESTS
Aerosol generation rates 750-1000 fig/min.
48
-------
10
tc
111
IL
« 1
u.
O
85
00
0.1
1 I I I I I |
I I I I I L
I I I I I I I
ORIFICE
DIAMETER
(cm)
> 2.54
O
0
O 3.81
A
V
D
I
ORIFICE
VELOCITY
(m/sec)
35-38
58
26-28
40-50
18
26
39
53
100
1000
EFFECTIVE FRICTION LOSS — mm WC
10.000
SA-2746-5
FIGURE 8 SCRUBBER PERFORMANCE CURVE FOR AMBIENT SCRUBBING TESTS
Aerosol generation rates 1000-1200 /^g/min.
49
-------
10
I
t
z
3
OC
111
IL
z
oc
u.
o
£
o
0.1
I I I I
TTT
1 I IT
I I I I I I I
ORIFICE
DIAMETER
(cm)
A 2.54
0
O 3.81
A
V
D
I
100
1000
EFFECTIVE FRICTION LOSS — mm WC
ORIFICE
VELOCITY
(m/sec)
46
58
79
18
26
39
53
10.000
SA-2745-6
FIGURE 9 SCRUBBER PERFORMANCE CURVE FOR AMBIENT SCRUBBING TESTS
Aerosol generation rates 1200-1500 /jg/min.
50
-------
10
s
1 '
cc
&
00
2
Z
0.1
III I I_
O
ORIFICE
DIAMETER
Jem)
[> 2.54
O
0
O 3.81
V
D
I
ORIFICE
VELOCITY
(m/sec)
35-38
25-28
70
18
39
53
i 1 I 1 1 I 1
100
FIGURE 10
1000
EFFECTIVE FRICTION LOSS — mm WC
10.000
SA-2745-7
SCRUBBER PERFORMANCE CURVE FOR AMBIENT SCRUBBING TESTS
Aerosol generation rates 1500-1900 jug/min.
51
-------
1C
n
z
I
B
UJ
LL
I '
B
ii
CD
ID
I
O.I
I I
100
AEROSOL
GENERATION
RATE
350-570
750-1000
1000-1200
1200-1500
1500-1900
I I I I I I 1
1000
EFFECTIVE FRICTION LOSS — mm WC
10,000
SA-2745-8
FIGURE 11 SUMMARY OF SCRUBBER PERFORMANCE CURVES FOR AMBIENT
SCRUBBING TESTS
52
-------
the generation rate of 750-1000 pg/min (Figure 7), which most nearly cor-
responds to the range producing the aerosol defined as "Aerosol A." The
data of Figure 6, which were actually the last obtained (except for one
point) were taken from tests with "Aerosol B." The one point from Test
No. EA-91 was obtained by deliberately attenuating the output of the
nebulizer at a time when it was producing about 1200 (jg/min of aerosol
at full power (Table 4). The efficiency for that run checked very closely
the efficiency obtained in Test EA-90, made at the same conditions except
with full generator output.
In Figures 6-10, the straight-line relationship of transfer units to
effective friction loss (in a log-log plot) appears as suggested by expe-
rience in earlier investigations. However, another phenomenon, not pre-
viously observed, also appeared in Figures 6, 7, and 8. This was the
appearance of a sharp drop-off of transfer units at low values of effec-
tive friction loss. It appears that in this region there is an essentially
separate correlation of transfer units and effective friction loss (see
Figure 6) that may give a second straight line with a slope exceeding unity.
In Figures 7 and 8, the data in the low-efficiency region are too few and
too imprecise to establish clear correlations.
The correlations of Figures 6-10, as well as the other data correla-
tions obtained during the investigation, demonstrate clearly that the
relationship of transfer units to effective friction loss for the orifice
scrubber was independent of gas velocity, liquid-to-gas ratio, and con-
tactor orifice size, at least over the range of importance (See Appendix B).
There may be exceptions in the cases where separate data correlations appear
in the regions of low transfer units and low effective friction losses. How-
ever, in the principal operating range, orifice velocities were varied from
18 to 79 m/sec <59 to 259 ft/sec) and liquid-to-gas ratio from 0.52 to
80 liters/Sm3 (3,9 to 60O gal/1000 Sft3) without showing independent
influences on collection efficiency. Few data were taken at liquid-to-gas
53
-------
3 3
ratios under 2 liters/Sm (15 gal/1000 Sft ), even in tests at low
effective friction losses.
A correlation of the scrubber pressure drop data taken during the
ambient scrubbing tests is shown in Figure 12. The pressure drop is
expressed as the number of velocity heads based on the gas velocity
through the contactor orifice (without allowance for the presence of the
water), and is presented as a function of the liquid-to-gas ratio expressed
3
in liters/Am . In this correlation, the gas volumes, velocities, and
densities were computed at the measured scrubber inlet temperature and a
pressure of one atmosphere (see Table 5). Since the pressure drop measured
was the total across the contactor orifice, the mitre elbow ahead of the
cyclone, and the cyclone itself, deviations were to be expected between
results obtained at different orifice sizes and gas flow rates. The cor-
relation curve is no more than a rough average for the data as presented.
ADIABATIC SATURATION TESTS
The Adiabatic Saturation tests were started, fortuitously, at a time
when the aerosol generator first reached a relatively stable operating
mode. Most of the data (see Table 6) can therefore be considered appro-
priate to Aerosol A. A few additional tests (No. AS-23 to AS-27) were
taken after the generator reached its second stable mode and are generally
appropriate to Aerosol B. The tests No. EN-2 to EN-4 were part of the
series of preliminary screening tests.
The greater stability of the aerosol generator during these testing
periods was reflected in the close correlation of the number of transfer
units with the effective friction loss (Figures 13 and 14); the character-
istic straight-line relationship was clearly defined. However, the tests
with Aerosol A and Aerosol B each yielded a separate curve, with Aerosol B
showing systematically lower collection efficiencies. Once again, gas
54
-------
100
1
I
3
C
Q.
O
DC
Q
111
DC
Q-
ORIFICE
DIAMETER
(cm)
O 2.54
A
D
O
V
3.81
A
•
*
4.45
10 15 20 25 30 35 40
LIQUID-TO-GAS RATIO — liters/Am
45 50
3
55
ORIFICE
VELOCITY
(m/sec)
18-23
25-34
36-43
48-55
5&-61
68-81
18-23
25-34
36-43
48-55
18-23
25-34
1 1 1 1 I 1
1 1 1 1 1 1 1
60
65
70
SA-2745-21
FIGURE 12 CORRELATION OF SCRUBBER PRESSURE DROP DATA FOR AMBIENT SCRUBBING CONDITIONS
-------
Table 5
SCRUBBER PRESSURE DROP DATA FOR AMBIENT SCRUBBING TESTS
Test
Code
No.
EA
•18
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
Orifice
Diameter
(cm)
3.81
2.54
3.81
Liquid-to-Gas
Ratio
(liters/Am3)
17.9
26.2
35.7
51.3
36.4
25.4
18.5
8.49
1.38
5.59
12.4
16.9
0.52
1.03
4.12
9.00
25.3
12.4
2.63
1.42
10.7
8.50
38.4
53.0
36.9
26.4
15.2
14.4
8.26
1.56
0.505
1.41
8.38
18.7
18.5
18.6
Gas Velocity
in orifice8
(m/sec)
18.5
18.4
18.4
18.4
26.0
26.0
26.0
26.1
39.9
39.6
39.0
38.7
53.6
53.6
53.6
53.6
26.2
38.4
52.7
39.0
20.7
26.4
28.2
28.7
40.5
40.5
41.5
59.7
59.4
80.2
54.9
39.3
26.4
18.4
18.5
18.5
Velocity
Head
(nun WC)
20.2
20.1
20.2
20.2
40.1
40.1
40.1
40.4
94.0
93.0
91.4
91.2
173
173
173
173
41.7
90.2
168
91.7
25.4
41.4
47.8
49.3
98.8
98.6
100
211
210
384
177
92.5
41.4
20.1
20.3
20.2
Scrubber
Pressure Drop
mm WC
228
345
516
833
1217
787
516
239
191
442
864
1176
282
345
767
965
802
870
517
191
151
223
1092
1707
2680
1722
963
2300
1247
1087
302
200
217
217
217
217
Velocity Heads
11.3
17.2
25.6
41.3
30.3
19.6
12.8
5.92
2.03
4 75
9.44
12.9
1.63
1.99
4.42
5.59
19.3
9.63
3.07
2.08
5.94
5.37
22.8
34.6
27.1
17.5
9.58
10.9
5.93
2.84
1.71
2.16
5.24
10.8
10.7
10.8
56
-------
Table 5 (Continued)
SCRUBBER PRESSURE DROP DATA FOR AMBIENT SCRUBBING TESTS
Test
Code
No.
EA
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
Orifice
Diameter
(cm)
2.54
2.54
4.45
2.54
3.81
Liquid-to-Gas
Ratio
(liters/Am3)
8.22
9.04
10.7
12.6
35.2
25.8
17.7
17.7
38.3
24.0
12.4
3.38
9.37
6.93
78.2
42.4
21.1
9.37
12.8
38.5
9.24
3.17
21.7
3.88
2.98
2.73
9.92
8.92
7.06
6.87
3.74
3.24
2.82
2.49
1.84
5.52
8.20
12.3
16.0
4.51
Gas Velocity
in Orifice
(m/sec)
60.7
59.1
59.1
59.1
40.8
41.1
42.1
42.4
38.7
36.3
39.3
38.7
26.6
72.8
27.2
37.5
50.3
68.0
39.0
38.7
26.9
39.3
39.0
20.9
27.3
29.8
18.8
20.9
26.5
27.2
22.2
25.7
29.5
33.5
30.1
30.1
30.0
30.0
29.6
48.5
Velocity
Head
(mm WC)
214
209
209
209
102
103
105
105
88.4
78.2
89.7
87.6
40.9
307
43.9
83.8
154
277
89.4
89.4
42.4
90.4
89.7
26.4
44.5
53.3
20.5
25.3
40.6
42.9
29.2
39.1
51.6
66.3
54.1
54.1
54.1
54.1
53.1
142
Scrubber
Pressure Drop
mm WC
1290
1413
1633
2000
2532
1768
1224
1224
2590
1237
663
292
209
1854
2946
2642
2291
1902
673
2565
224
272
1250
82.1
136
162
81.3
97.7
137
161
77.8
103
130
164
115
216
293
444
588
572
Velocity Heads
6.02
6.75
7.81
9.59
24.8
17.2
11.7
11.6
29.3
15.8
7.39
3.33
5.12
6.05
67.1
31.5
14.9
6.87
7.52
28.7
5.27
3.00
13.9
3.12
3.04
3.04
3.96
3.85
3.39
3.75
2.67
2.64
2.52
2.48
2.11
3.99
5.40
8.20
11.1
4.03
57
-------
Table 5 (Concluded)
SCRUBBER PRESSURE DROP DATA FOR AMBIENT SCRUBBING TESTS
T*»at-
Code
No.
EA
124
125
126
127
128
Orifice
Diameter
(cm)
2.54
Liquid-to-Gas
Ratio
(liters/Am3)
16.5
22.0
28.1
38,0
1.77
Gas Velocity
in Orifice
(in/sec)
38.7
38.7
37.8
38.1
70.4
Velocity
Head
(mm WC)
88.6
88.6
87.1
87.9
297
Scrubber
Pressure Drop
mm WC
897
1271
1587
2440
866
Velocity Heads
10.1
14.3
18.2
27.8
2.92
Calculated at actual inlet gas conditions.
58
-------
Table 6
SCKUBBER PERFORMANCE DATA FOR AD1ABATIC SATURATION SCRUBBING TESTS
Test
Code
No.
EN
2
3
4
AS
6
7
9
10
11
12
13
14
IS
16
17
18
19
20
21
22
23
24
25
26
27
Aerosol
Generation
Rate
(M-g/mln)
1060
1245
1370
1319
1010
1225
1371
1176
2063
1226
1074
1190
1007
1083
1072
997
8S4
685
942
704
665
537
661
703
Averages
Contactor
Orifice
Size
(cm)
3.81
3.81
2.54
2.54
3.81
2.54
3.81
Inlet CBS Flow Rate
(std liters/mill)
Air
1670
1670
1670
1580
1580
1580
1580
1580
1240
1230
119O
1190
1190
1150
1120
1120
1580
1580
1580
1250
1240
1590
1560
1570
Water
Vapor
923
923
923
800
8OO
800
800
80O
659
659
573
573
573
542
542
542
783
7S3
783
682
682
787
801
787
Total,
2590
2590
2590
2380
2380
2380
2380
2380
1900
1890
1760
1760
1760
1690
1660
1660
2360
2360
2360
1930
1920
2380
2360
2360
Gas Moisture
Content
(g/B dry air)
In
0.341
0.341
0.341
0.315
0.315
0.315
0,315
O.315
0.329
0.334
0.300
0.300
0.300
0.292
0.301
0.301
0.308
0.308
0,308
0.338
0.343
0.308
0.319
0.312
0.317
Out
0.384
0.384
0.384
0.390
1.369
0.362
0.350
0.342
0.485
0.467
0.372
0.386
0.384
0.369
0.355
0.335
. 0.327
0.325
0.318
0.389
0.371
0.344
0.331
0.323
0.369
Gas
Temperature
<°C)
In
126.1
126.1
126.1
128.9
130
130.6
130.6
127.8
125
129.4
130
138.9
123.3
128.9
121.7
125
128.3
12$. 9
130
12S.9
12S.9
130
130
129.4
128.9
Out
73.9
73.9
73.9
72.2
72.2
72.8
72.8
72.8
71.1
70.6
70.6
71.1
71.7
70.6
71.1
70.6
71.7
71. 7
71.7
71.1
71.1
71.7
71.7
71.7
71.7
Water
Temperature
<°C>
In
74.4
74.4
74.4
73.3
73.3
73.9
73.3
75.6
73.3
72.8
73.3
73.9
75.6
73.3
73.3
73.3
73.9
73.3
73.9
73.9
73.9
73.9
73.9
74.4
73.9
Out
74.4
74.4
74.4
73.9
73.9
73.9
73.9
73.3
73.3
73.3
73.3
73.9
74.4
73.3
73.3
72.8
73.9
73.9
73.3
73.3
72.8
73.9
73.9
73.9
73.9
Gas
Velocity
In
Orifice
(ra/sec)
50.7
50.7
50.7
46.9
47.0
47.1
47.1
46.8
83.5
83.9
78.3
80.0
77.0
75.0
72.3
72.9
46.5
46,5
46.7
85.9
85,3
47.2
46.8
46.4
Liquid- to-
Gas Ratio
AitersX
\ Sm3 /
4.78
4.78
4.78
19.0
13.5
8.16
4.69
1.54
10.2
10.3
11.0
11.0
8.59
11.5
9.02
5.71
4.81
4.01
1.39
8.36
6.24
11.4
6.33
3.13
Effective
Friction
Loss
{mm WC)
432
432
432
1240
837
571
353
197
1750
1750
1620
1620
1370
1550
1150
1000
371
319
186
1660
1390
688
439
253
Collection
Efficiency
(%)
69.05
68.61
70.23
95.98
88.28
77.81
64.86
41.67
98.64
97.67
97.78
97.14
96.06
97.65
95.42
93.21
62.18
53.02
36.62
97.35
96.01
79.23
64.15
44.64
Transfer
Units
fNt>
1.17
1.16
1.21
3.21
2.14
1.51
1.05
0.539
4.30
3.76
3.81
3.56
3.23
3.75
3.08
2.69
0.972
0.756
0.456
3.63
3.22
1.57
1.03
0.591
en
to
-------
10
z
5
ffi
O.1
100
I I I I I I I I
1 I I I I I L
J I 1 I I I t I I
ORIFICE
DIAMETER
(em)
O 2-54
A 3.81
' I i
1000
EFFECTIVE FRICTION LOSS
10.000
— iwn WC
FIGURE 13 SCRUBBER PERFORMANCE CURVE FOR ADIABATIC SATURATION
SCRUBBING—AEROSOL A
Gat temperatures: Inlet * 128.9°C; Outlet - 71,7°C.
60
-------
s
£
0.1
100
1—I—I I I I 11
' I I I I 11
1000
T 1 I I I I U
ORIFICE
DIAMETER
(cm)
V 2.54
<> 3.81
J L
EFFECTIVE FRICTION LOSS — mm WC
10,000
SA-2746-10
FIGURE U SCRUBBER PERFORMANCE CURVE FOR ADIABATIC SATURATION
SCRUBBING—AEROSOL B
Gas temperatures: Inlet « 128.9°C; Outlet = 7t.7°C.
61
-------
velocity and liquid-to-gas ratio were found to exert no independent
influences on collection efficiency (See Appendix B).
CONDENSING TESTS
In the Condensing tests of Series I and II, the initial tests were
made with Aerosol A, and separate groups of tests were subsequently made
with Aerosol B when it became evident that the aerosol generator mode had
again shifted (Tables 7 and 8). The inlet gas conditions were regulated
to be the same as for the Adlabatic Saturation tests, but the gas was then
scrubbed with sufficient cold water to bring the exit gas to a selected
final temperature; 54.4°C (130°F) in Series I, and 37.8°C (100°F) in
Series II, For each series of tests, the liquid-to-gas ratio was there-
fore fairly constant. However, in Series II there was a trend toward
reduction of the required liquid-to-gas ratio as the scrubber pressure
drop increased, indicating that the efficiencies of heat transfer and
dehumidification, like the efficiency of particle collection, were
increased by increases in the effective friction loss. In the tests of
Series I, the average outlet gas and water temperatures were equal, but
in the tests of Series II, the outlet water stream was always substantially
colder than the exit gas, and colder at lower effective friction loss than
at high.
The scrubber performance curves for Test Series I and II are presented
in Figures 15 and 16, respectively. In both cases, good correlations were
obtained, and the curves for Aerosol A and Aerosol B were systematically
different, as in the Adiabatic Saturation tests. The corresponding
curves for Test Series I and II were actually identical; this is obvious
from Figure 17, in which the data points for collection of Aerosol A in
both Series I and Series II are presented together. The data points for
collection ol Aerosol B in Series I and II are similarly presented together
in Figure 18. It appears that the coincidence of the performance curves
62
-------
Table 7
SCRUBBER PERFORMANCE DATA FOR CONDENSATION SCRUBBING TESTS, SERIES I
Test
Code
No.
EC
4
6
7
27
28
29
30
31
32
33
46
47
48
49
Aerosol
Generation
Rate
(Jig/rain)
867
807
892
808
850
843
801
773
917
932
5S4
555
607
776
Contactor
Orifice
Size
(cm)
3.81
3.81
2.54
3.81
2.54
Inlet Gas Flow Rate
(std liters/min)
Air
2140
2140
2140
1370
1660
2190
2680
1220
1300
1650
1860
2340
2920
1760
Water
Vapor
1230
1230
1230
691
849
1080
1380
629
739
849
960
1170
1450
912
Total
33 7O
3370
3370
2060
2510
3270
407O
1850
2040
25OO
2820
3510
4370
1970
Averages
Gas Moisture
Content
(g/g dry air)
In
O.358
0.358
0.358
0.313
0.317
O.3O6
0.321
O.319
0.353
0.319
0.320
0.310
0.308
0.321
0.327
Out
0.104
0.104
0.104
0.113
0.115
0.118
0.122
0.119
0.122
0.131
0.119
0.119
0.122
0.139
0.117
Gas
Temperature
(°C)
In
130.6
130.6
130.6
130
131.1
131.1
133.3
127.2
127.8
130.6
131.1
131.7
133.3
131.1
130.6
Out
52.2
52.2
52.2
54.4
54.4
54.4
54.4
53.9
53.9
53.9
55.0
54.4
54.4
54.4
54.4
Water
Temperature
(°C)
In
12.2
12.2
12.2
12.8
12.8
12.8
12.8
13.3
12.8
12.8
13.9
13.3
13.9
13.9
13.3
Out
51.1
51.1
51.1
52.2
52.8
54.4
55.0
53.9
54.4
54.4
55.0
55.0
54. 4
55.0
54.4
Gas
Velocity
in
Orifice
(m/sec)
66.7
66.7
66.7
40.7
49.8
64.8
81.1
81.7
9O.2
111
55.9
69.7
87.1
87.9
Liquid- to-
Gas Ratio
/liters\
U3 j
3.37
3.37
3.37
2.94
3.09
2.90
2.41
2.78
2.97
2.73
2,75
2.75
2.25
2.62
Effective
Friction
Loss
(mm WC)
444
444
444
121
234
398
553
798
983
1550
282
467
645
1710
Collection
Efficiency
(%>
74.57
72.94
72.66
45.76
58.79
74.75
83.14
93.85
95.46
97.40
56.04
75.49
84.04
97.71
Transfer
Units
(Nt)
1.37
1.31
1.30
0.612
0.886
1.38
1.78
2.79
3.09
3.65
0.822
1.41
1.83
3.78
0>
w
-------
Table 8
SCRUBBER PERFORMANCE DATA FOR CONDENSATION SCRUBBING TESTS, SERIES II
Teit
Code
Mo.
EC
a
9
10
11
12
13
14
13
18
17
18
19
30
21
33
33
23A
24
29
28
43
44
45
50
Aeroaol
Generation
Rate
957
888
1316
1132
1197
111S
801
690
3340
1083
1258
855
915
981
787
789
854
879
875
850
340
430
510
370
Contactor
Orifice
SUe
td Uteri Bin)
Air
1740
1740
1390
1680
2210
2740
1140
1140
1480
1880
1250
1210
1300
1400
1520
1680
1340
1490
1690
1890
1910
2150
2350
1530
Water
Vapor
923
923
690
850
1000
13TO
573
567
739
838
641
603
660
728
7T8
832
634
739
849
960
960
1080
1180
838
Total
2660
2660
1980
2510
3300
4110
1710
1710
2200
2510
1890
1810
1960
2130
2300
2490
1970
2230
2540
2850
2870
3230
3530
2370
Averages
Gu Moisture
Content
(g/f. dry air)
In
0.33O
O.330
0.325
0.317
0.305
0.312
0.313
0.310
0.315
0.311
0.319
0.310
0.315
0.321
0.317
0.311
0.294
0.308
0.312
0.315
0.313
0.312
0.312
0.339
0.315
Out
0.0484
0.0481
0.0440
0.0449
0.0458
0.0470
0.0502
0.0487
0.0516
0.0535
0.0488
0.0471
0.0494
0.0508
0.0926
0.0955
0 .0453
0.0471
0.0447
0.0451
0.0446
0.0454
0.0479
0.0531
0.0404
Cu
Temperature
(9C)
In
129.4
129.4
128.3
128.3
129.4
131.7
127.8
128.3
128.9
128.9
128.3
128.7
127.2
128.3
128.3
129.4
128.3
128.9
130
130.6
134.4
133.9
131.1
130.6
129.4
Out
38,9
38.9
37.8
37.8
37. 8
37.8
38.3
37.8
37.8
37.2
37.8
37.2
37.8
37.8
37. 8
37.8
38.3
38.9
37.8
37.8
37.2
37.2
37.8
37.8
37.8
Water
Temperature
<°C>
In
12.2
12.2
12.2
11.7
11.7
11.7
12.2
11.7
11.7
11.7
11.7
12.2
12.2
12.2
12.2
12.2
12.2
12.2
12.2
12.8
12,8
12.8
12.8
13.3
12.2
Out
32.8
32.8
27.8
32.2
34.4
35.6
36.1
36.1
36.1
35.6
3S.6
34.4
39.0
38.1
35.6
35.8
27.8
—
30
32.2
32.2
33.9
34.4
36.1
33.9
Gai
Velocity
in
Orifice
(a/aec)
52.5
92.5
28.6
36.3
47.9
59.9
75.6
75.7
97.6
111
83.7
79.8
86.6
94.3
102
111
28. 6
32.3
36.9
41.5
57.4
64.5
70.0
106
U quid- to-
Gai Ratio
/lltera\
(~^~)
1.04
7.04
9.78
7.73
6.88
5.68
6.54
6.32
6.64
6.31
6.83
7.31
6.95
6.83
6.67
6.82
9.52
8.S5
8.78
8.26
8.25
7.50
7.13
6.S4
Effective
Friction
Lou
(mm WC)
447
447
183
278
480
666
938
925
1557
2120
1120
1050
1210
1450
1710
2120
176
227
301
379
587
726
866
1790
Collection
Efficiency
«)
79.27
79.65
58.83
66.62
80.41
87.37
95.55
94,98
99.16
99.24
96.46
95.84
96.90
97.87
98.48
99.00
51.02
58.98
68.15
74.22
80.00
85.35
89.17
97.92
Tranifer
Unlti
(Nt>
1.58
1.60
O.S87
1.10
1.63
2.07
3.11
2.99
4.78
4.88
3.34
3.18
3.17
3.85
4.18
4.61
0.714
0.891
1.14
1.36
1.81
1.92
2.22
3.87
-------
10
5
u.
O
CO
5
D
Z
0.1
100
1 I I I ! I I
II I 1 I 1 l_
O
I I I I I I I I 1
ORIFICE
DIAMETER
(cm)
O 2.54
V
A 3.81
O
I
AEROSOL
A
B
A
B
_L 1 I I I I
1000
EFFECTIVE FRICTION LOSS — mm WC
10,000
SA-2745-11
FIGURE 15 SCRUBBER PERFORMANCE CURVES FOR CONDENSATION SCRUBBING
(SERIES I)
Gas temperatures: Inlet = 130.6°C; Outlet = 54.4°C.
65
-------
10
z
D
C
tu
z
5
oa
0.1
100
T 1 1 I I I I I |
T 1—I I I I I I
I I I I It I
ORIFICE
DIAMETER
(cm)
O 2.54
V
A 3.81
<0
D 4.45
I
1000
EFFECTIVE FRICTION LOSS — mm WC
AEROSOL
A
B
A
B
A
I I I I I I I
10.000
SA-2746-12
FIGURE 16 SCRUBBER PERFORMANCE CURVES FOR CONDENSATION SCRUBBING
(SERIES II)
Gas temperatures: Inlet = 129.4°C; Outlet = 37.8°C.
66
-------
10
t
z
Z
<
DC
£
m
5
0.1
I I I I I I
1IT Till.
100
GAS TEMPERATURES (CC)
Inlet Outlet —
O 130.6 54.4
A 129.4 37.8
I
1000
EFFECTIVE FRICTION LOSS — mm WC
10.000
SA- 2745-13
FIGURE 17 SCRUBBER PERFORMANCE CURVE FOR CONDENSATION SCRUBBING
(SERIES I AND II)—AEROSOL A
67
-------
10
z
I
CO
85
I
£
CD
3
0.1
100
1—MMI
1 I I I I IJ
GAS TEMPERATURES CO
Inlet Outlet
O 130.6 54.4
A 129.4 37.8
I I I I I I I
1000
EFFECTIVE FRICTION LOSS — mm WC
10,000
SA-27 45-14
FIGURE 18 SCRUBBER PERFORMANCE CURVE FOR CONDENSATION SCRUBBING
(SERIES I AND II)—AEROSOL B
68
-------
for Series I and II results from the fact that only a relatively small
amount of additional water vapor condensed as the gas temperature was
reduced from 54.4° to 37.8°C (130° to 100°F).
There was no indication of independent effects of orifice size or
gas velocity on the collection efficiency.
In Test Series III the inlet gas was adjusted to the same moisture
content as in Series I and II, but the dry bulb temperature was only about
8 Centigrade degrees above the saturation temperature (Table 9). The
tests were carried out over the period in which the aerosol generator
shifted from producing Aerosol A to producing Aerosol B. Not enough
data points were obtained with Aerosol A to establish a performance curve,
but it was possible to obtain a curve for Aerosol B (Figure 19), which
could then be compared with the similar performance for Test Series I
and II(Figure 20). The substantial improvement in performance obtained
with the nearly saturated inlet gas probably indicates that condensation
started almost immediately in the gas/liquid contacting zone, whereas
superheated inlet gas had to be cooled substantially before condensation
could start.
VAPORIZING TESTS
The conditions for the Vaporizing tests were chosen arbitrarily to
be as similar as possible to those employed in the other test series.
The inlet air and water temperatures were adjusted to approximately the
same values as in the Adiabatic Saturation tests (Table 10). The water
rate was adjusted to yield the same outlet gas temperature in all the
runs. The three initial tests (EV-1, -2, and -3) were made as part of
the original series of screening tests, on an aerosol that was pre-
sumably equivalent to that later designated as Aerosol A. The remainder
of the tests were made with Aerosol B, but the data points from Tests
69
-------
Table 9
SCRUBBER PERFORMANCE DATA FOR CONDENSATION SCRUBBING TESTS, SERIES III
Test
Code
Ho.
EC
34
35
36
37
38
39
41
42
Aerosol
Generation
Rate
918
1015
1011
936
481
484
524
586
Contactor
Orifice
Size
2.54
3.81
Inlet Gas Flow Rate
(atd liters/mm)
Air
1160
1230
1240
1660
1450
1930
1670
1310
Water
Vapor
629
629
623
845
739
1060
838
682
Total
1790
1860
I860
2510
2190
2990
2510
1990
Averages
Ga.3 Moisture
Content
In
0.335
0.318
0.311
O.319
0.317
0.34D
0.312
0.323
0.322
Out
0.0477
0.0481
0.0480
O.OS39
0.0517
0.0460
0.0450
0.0443
0.0481
Gas
Temperature
In
77.2
77.2
80.0
80.0
80.6
82.8
82.8
80.6
80.0
Out
37.B
37.8
37.8
37.8
38.3
37.8
37.8
37.8
37.8
Water
Temperature
In
12.8
12.8
12.8
12.8
12.2
12.8
12.8
12.8
12.8
Out
33.3
34.4
34.4
35.6
37,2
33.9
32.2
27.8
33.9
Gas
Velocity
In
Orifice
la/sec)
69.2
71.9
72.5
97.8
85.5
52.2
43.8
34.5
Ll quid- to-
Gaa Ratio
/liters\
6.86
6.32
6.29
6.12
5.80
7,02
7.23
8.74
Effective
Friction
Loss
(ma WC3
922
980
978
1910
1360
549
371
242
Collection
Efficiency
93.04
97.95
97.23
99.23
97.55
87.53
77.61
67.11
Transfer
Units
(Nt)
3.93
3.89
3.59
4.93
3.71
2.08
1.50
1.11
-J
o
-------
10
D
EC
EC
CD
D
0.1
100
I I I I
1IIII 1 l_
ORIFICE
DIAMETER
(cm)
O 2.54
A
D 3.81
AEROSOL
A
B
B
I
I
1000
EFFECTIVE FRICTION LOSS — mm WC
10,000
SA-2745-15
FIGURE 19 SCRUBBER PERFORMANCE CURVE FOR CONDENSATION SCRUBBING
(SERIES III)
Gas temperatures: Inlet = 80.0°C; Outlet = 37.8°C.
71
-------
10
z
-------
Table 10
SCRUBBER PERFORMANCE DATA FOR VAPORIZATION SCRUBBING TESTS
Test
Code
No.
EV
1
2
3
4
5
6
7
8
9
Aerosol
Generation
Rate
920
990
1050
447
535
495
558
559
653
Contactor
Orifice
Size
(cm)
3.81
2.54
3.81
Inlet Gas Flow Rate
(std liters/rain)
Air
2250
2280
2280
1720
1570
1320
2170
1940
16 4O
Water
Vapor
__
—
—
—
--
—
—
—
—
Total
2250
2280
2280
1720
1570
1320
2170
1940
1640
Averages
Gas Moisture
Content
(g/g dry air)
In
0.01a
0.006
0.006
0.006
0.006
0.006
0.006
0.007
Out
O.150
0.147
0.147
0.178
0.159
0.138
0.147
0.145
O.132
0.149
Gas
Temperature
In
126.1
125.6
125.6
128.3
125
122.8
126.1
126.7
123.9
125.6
Out
58.9 ,
58.3
58.3
58.9
58.3
57.2
58.3
58.3
57.2
58.3
Water
Tempera ture
In
71.1
72.8
72.8
73.9
73.9
73.9
74.4
74.4
74.4
73.3
Out
58.9
58.9
58.9
60.0
60.0
58.3
60.0
60.0
59.4
58.3
Gas
Velocity
in
Orifice
(m/sec)
44.0
44,6
44.6
76.4
69.0
57.6
42.2
38.1
32.0
Liquid-to-
Gas Ratio
/liter s\
V sm3 )
5.04
4.99
4.99
7
6.75
4.30
5.01
4.77
4.38
Effective
Friction
Loss
(mm WC)
430
431
431
1610
1090
609
390
281
159
Collection
Efficiency
55.75
53.69
55.56
96.04
90.29
76.80
46.30
27.43
13.82
Transfer
Units
0.815
0.770
0.811
3.23
2.33
1.46
0.622
0.321
0.149
Estimated.
Estimated from pressure drop.
-------
EV-1, -2, and -3 appeared to fit the correlation curve fairly well
(Figure 21).
The scrubber performance curve for Vaporizing conditions is notable
for the very marked "break" and the appearance of a quite distinct and
separate correlation curve in the low range of effective friction loss,
similar to that obtained for Ambient scrubbing of Aerosol B (Figure 6).
This phenomenon did not appear in any of the Adiabatic Saturation or
Condensation tests.
GENERAL DISCUSSION
The scrubber performance curves for tests at Adiabatic Saturation,
Condensing, and Vaporizing conditions are summarized, without data points
in Figures 22 and 23. The curves for scrubbing of Aerosol A are presented
in Figure 22, and those for scrubbing of Aerosol B in Figure 23. In
Table 11, the numbers of transfer units and corresponding collection
efficiencies are tabulated for selected values of effective friction
loss, for convenient reference.
Although condensation scrubbing is seen to have afforded a signifi-
cant increase in the collection efficiency at a given level of effective
friction loss, the improvement was not extremely marked. The maximum
improvement obtained (comparing condensation scrubbing of the nearly
saturated gas with the adiabatic saturation scrubbing of the hot gas)
was equivalent to reductions of about 25 to 50% in the effective friction
loss required to attain a given collection efficiency. This is much less
favorable than indicated by previous experience in pilot plant studies of
4,5,9
scrubbing fume from Kraft recovery furnaces, where condensation
scrubbing of the hot gas appeared to reduce the required contacting power
by about 50 to 70% from that required in adiabatic saturation scrubbing
of the same gas. However, the qualitative trends in the results from the
74
-------
10
u>
DC
IU
u.
O
GS
00
0.1
\ I IMF)
I I 1 I II I
ORIFICE
DIAMETER
(on)
V 2.54
A 3,81
100
FIGURE 21
1000
EFFECTIVE FRICTION LOSS — mm WC
AEROSOL
B
A
B
10,000
SA-27 45-17
SCRUBBER PERFORMANCE CURVE FOR VAPORIZATION SCRUBBING
Gas temperatures: Inlet = 125.6°C; Outlet = 58.3°C.
75
-------
10
I
3
I
g
m
0.1
I I I I I |
OPERATING
CONDITION
- Condensing
• Adiabatic
Saturation
GAS TEMPERATURES <°C)
Inlet Outlet
130
128.9
37.8 and 54.4
71.7
I I I I I M I
100
1000
EFFECTIVE FRICTION LOSS — mm WC
10.000
SA-2745-18
FIGURE 22 SUMMARY OF SCRUBBER PERFORMANCE CURVES FOR ADIABATIC
SATURATION AND CONDENSATION SCRUBBING—AEROSOL A
76
-------
10
1—I I
1—I I
12
z
S
u.
O
S
I
OPERATING
CONDITION
__ _• Condensing
« — «— Condensing
i Adiabatic Saturation
.—— Vaporizing
GAS TEMPERATURES (*C)
Inlet Outlet
80.0 37.8
130.0 37.8 and 54.4 —
128.9 71.7
125.6 58.3
0.1
100
1000
EFFECTIVE FRICTION LOSS — mm WC
10.000
SA-2746-19
FIGURE 23 SUMMARY OF SCRUBBER PERFORMANCE
SATURATION, CONDENSATION, AND VAPORIZATION SCRUBBING—
AEROSOL B
77
-------
Table 11
SELECTED COMPARISONS OF SCRUBBER PERFORMANCES AT DIFFERENT OPERA!ING CONDITIONS
Test Condition
Gas Temp ( °C)
Inlet
Outlet
Aerosol
A
D
Effect ivo
Friction
(mm WC)
250
IOOO
1500
'250
IOOO
1500
Adiaba t ic
S a t u r:i i i o n
129
71.7
Trruis iv r
Unit*
O.G85
2.52
:«.68
2 . 'M
.'3.43
Co 1 Ice t ion
K J 1 if l«MH-y
'19. G
91.95
97.48
14.2
yo.o
96. 7G
Condons inp
Series 1 .md 11
I HO
54.4 »mi :V7.»
Tranclcr
Units
1 .01
2.9:1
•1 . 00
2 . >G
C'o 1 left loll
K I I u ii-ijcy
GM.:t
y'l.oi)
«H . 1 7
Mi . K
91, -Hi
9G.50
Condensing
Series III
80. O
:n.H
Transfer
I'lll Is
\ .14
-H . 1 2
'1.1H
C'o 1 Ice I ion
Ei f icluncy
0^7
Vapor i zmj;
126
&».a
Trims f».ir
Unl vs>
n.ruvj
Col loft ion
K 1 lie icii<-y
c:)
KK.9
<).'». 2 :H
00
-------
different investigations were similar, and there were wide differences
in the natures of the aerosols that might account for the quantitative
differences in the performances that were observed.
The scrubber tests clearly demonstrated that liquid-to-gas ratio,
gas velocity, and contactor orifice size had no influences on collection
efficiency that were independent of effective friction loss, at least
over the operating range of major practical interest (see Appendix B).
Exceptions to this conclusion may exist in the instances where a separate,
steep performance curve appears in the region of low effective friction
1oss and low collection efficiency. No clear exceptions appeared even
*« that operating region, but since the pertinent data obtained were
relatively imprecise and few in number, and covered relatively narrow
ranges of variables, some exceptions may have been concealed. However,
the operating region is not often one of much practical importance.
This observation concerning liquid-to-gas ratio is consistent with
the experience obtained in. other investigations with both similar and
some dissimilar types of scrubbers, although it conflicts with reports
fr°m some other investigators (see Appendix A). Where condensation and
VaPorization phenomena are controlled by liquid-to-gas ratio, this factor
Cai» be expected to have independent effects on efficiency. However, in
c°«densation tests, Series I and II, where additional gas cooling did not
effect substantial additional condensation of water, liquid-to-gas ratio
^ain appeared to have no independent effects.
The results of the investigation have thus constituted an additional
Verification of tne "contacting power rule" as applied to a small replica
°f the generalized class of commercial "venturi" scrubbers. (Most current
venturi scrubbers are actually orifice scrubbers employing some form of
Va*iable-area orifice.) Deviations between the various performance curves
8116 therefore ascribable to introduction of additional mechanisms—
79
-------
in the present case, vaporization and condensation.
Where diffusion and heat transfer effects come into play, as in con-
densation scrubbing, it is not to be expected that performance should be
dependent only on contacting power and be independent of scrubber con-
figuration. To the extent that particle scrubbing is an inertial phe-
nomenon, efficiency should be—and apparently is—independent of contact
time. However, where diffusion and heat transfer are involved, contact
time and scrubber configuration should play roles. Power consumption
does, of course, affect mass transfer as well as particle collection.
A number of investigators have now produced empirical correlations of gas
absorption data for cocurrent contactors (particularly pipeline contactors)
in terms of energy dissipation (see Appendix A). There should thus be
relationships between contacting power and particle collection by Stefan
flow and probably also between contacting power and the growth of particles
by condensation of water vapor, but these relationships are presumably
quite different from that between contacting power and particle collection
by inertial deposition. The scrubber performance curve for condensation
scrubbing thus represents the net effects of at least two major particle
collection mechanisms, each being a different function of contacting power.
Qualitatively, at least, it does not appear that condensation scrub-
bing should be regarded as a possible way to eliminate use of energy in
contacting, since utilization of condensation phenomena will also demand
contacting power. Rather, it should probably be regarded as a way merely
of reducing contacting power requirements. Of course, condensation scrub-
bing will, in practice, incur some energy consumption elsewhere (in water
cooling, for example) that will partly counteract the saving.
Where diffusional phenomena, such as Stefan flow, are involved, it
may be advantageous to use a type of gas/liquid contactor that provides
a relatively long contact time rather than one (such as a venturi or
80
-------
orifice) that gives an intense but short contact. A pipeline contactor
operating in the spray-flow regime appears to be a promising device
meeting this objective.
Since the results of this investigation clearly indicate a major
improvement from presaturating the gas before condensation scrubbing, a
high-energy condensation scrubber should be preceded by a. low-energy
Sas/iiquid contactor designed to operate as an adiabatic saturator by
^circulation, of the water. The pres&turator would not be designed to
collect fine particles.
It was also indicated that in condensation scrubbing there is minimal,
o
if any, benefit from reducing the final gas temperature below about 50 C,
since the additional condensation of water vapor is small. Hence, it
should not be necessary to use very cold scrubbing water unless there
are other reasons to dehumidify the gas. On the other hand, there may
an advantage in increasing the saturation temperature of the inlet gas,
since the quantity of water vapor present in the gas increases rapidly with
temperature. This approach incurs the extra costs not only of humidifying
the inlet gas but of removing the increased load of low-level heat from the
scrubbing water. Such economic considerations are further discussed in
Appendix C.
A number of phenomena appearing in this investigation may be of
wajor significance, theoretical or practical, but at this point they lack
explanation or—in some instances—complete confirmation of their reality.
The straight-line relationship in log-log plots of scrubber perfor-
mance was generally well confirmed by the results of this investigation.
However, the relationship is empirical, and at present there is no basis
other than experience for assuming that the performance curves should be
straight lines in the general case. Calvert's model for the venturi
scrubber (see Appendix A) leads to an approximation that, for a monodis-
Perse aerosol of given diameter, the number of transfer units is directly
81
-------
proportional to the pressure drop; hence, the predicted curve of transfer
units versus pressure drop on a log-log plot is a straight line with a
slope of unity. For a polydisperse aerosol, the overall collection effi"
ciency would have to be determined by integrating the collection effici-
encies over the range of particle sizes represented in the particle-size
distribution. Presumably, then, the form of the scrubber performance
curve should be dependent on the nature of the particle-size distribution,
and it is not obvious that the curve should necessarily be straight.
For the present, there is no theoretical basis for the applicability
of the "contacting power rule" to various types of scrubbers that do not
fit the models of Calvert and other investigators. If it can be assumed
on the basis of experience to date that scrubber performance curves are
straight lines on log-log plots, then it also appears logical that for a
monodisperse aerosol, the slope of the curve should be unity, with succes-
sive equal increments of contacting power producing equal increments in
the number of transfer units. There is no experimental confirmation of
this conjecture, since no appropriate data are available for the scrubbing
of a truly monodisperse aerosol. In scrubbing of a polydisperse aerosol,
larger particles will be collected with higher efficiency than smaller
ones, and successive increments of contacting power should yield succes-
sively smaller increments of transfer units, leading to a performance
curve with a slope of less than unity. In fact, it has been found that
the performance curves for most aerosols are straight lines with slopes
less than unity. However, some data developed from the literature have
yielded curves having slopes greater than unity.4 If the collection
mechanism is assumed to remain the same, such a performance curve
that the aerosol increases in average particle size in passing through the
scrubbing process. This interpretation, in turn, implies either that the
finer particles are collected selectively, which is contrary to experience
and reasonable expectation, or that the aerosol undergoes a coagulation
82
-------
Process in the scrubber that Is also related to contacting power, which
would seem possible but unlikely.
The present investigation is apparently the first to indicate that
the observed straight-line relationship in scrubber performance curves
may sometimes break down in some part of the contacting power range—
specifically, the low range. No indication of comparable deviations in
the high contacting power range has yet appeared. If it is accepted that
the performance curve is a straight line, then it appears reasonable that
it should remain so over the entire contacting power range, unless dif-
ferent particle collection mechanisms predominate in different parts of
the range.
In the present investigation, there were four occasions in which
separate correlation curves appeared in the low contacting power region
es 6, 7, 8, 21). In two of these cases, the curves were reasonably
defined (Figures 6, 21), and appear to be straight, or nearly
straight lines. Also, the slopes of these curves exceeded unity.
However, in each case, the principal correlation curve in the upper
contacting power range was a straight line with a slope less than unity.
^Is behavior suggests that in the cases reported in the literature where
correlation curves with very steep slopes were encountered, data may have
be«m taken only over a range defining the lower curvej extending the data
to a higher contacting power range might have revealed a second curve of
more familiar character.
It also appears that the steep curves in the low-energy region were
most pronounced in those cases in which the aerosol was presumably the
finest, at least in average particle size.
These observations suggest that the appearance of two segments in
scrubber performance curve for an aerosol may indicate a change in
83
-------
the dominant particle-collection mechanism over the contacting power
range in which the steeper correlation curve appears—for instance, a
shift from Brownian diffusion to inertia! deposition. At present, this
hypothesis is—at best—highly speculative, but it does suggest experi-
mental methods of attack on the problem.
A less rarified explanation for the low-energy-range performance
would be channeling of liquid in the contactor. Such channeling might
indeed explain behavior with very low water and air rates, but it does
not seem adequate to explain behavior such as shown in Figures 6 and 21,
or the absence of comparable behavior in other test series.
In cases where scrubbing tests on evidently different aerosols are
involved (Figure 11), crossing of the performance curves appears readily
explainable in terms of differences in particle-size distributions as
well as average particle sizes. However, in cases where the same aerosol
is being scrubbed at different operating conditions (Figure 23), it is
difficult to offer a satisfying physical interpretation for a crossing of
performance curves. Differences in collection mechanisms, as in the
cases of Adiabatic Saturation and Condensation scrubbing, can offer a
reason—if not a detailed explanation—why the performance curves have
different slopes. Since the curves do have different slopes, they must
at some point either intersect or merge. If they merge, the straight-
line relationship for at least one of the curves must break down. So f&r
there has been no evidence of such breakdowns in performance curves at
high contacting powers and high efficiencies.
The apparent intersection of the performance curves for Adiabatic
Saturation scrubbing and Condensation scrubbing in Figure 23 occurs near
the upper end of the range in which data were obtained, so that a merger
of the curves, rather than an intersection, may have been concealed by
lack of experimental precision. Nevertheless, there is no unequivocal
evidence that this was, in fact, what happened.
84
-------
To illuminate some of these questions, the performance of the scrubber
should be investigated with aerosols of different particle sizes, generated
by atomizing ammonium fluorescein solutions of different concentrations.
Aerosols should be used that are both larger and smaller than the one
used in the present investigation. Any trends appearing in the observed
phenomena as a result of known changes in particle size should provide
much better bases for interpretation.
85
-------
REFERENCES
1. Christoforidis, A. J., J. F. Tomashefski, and R. I. Mitchell, Use of
an Ultrasonic Nebulizer for the Application of Oropharyngeal, Laryn-
geal, and Tracheobronchial Anesthesia. CHEST 59 (6), 629-633 (June
1971).
2. Ferber, B. I., F. J. Brenenborg, and A. Rhode, Penetration of Sodium
Chloride Aerosol Through Respirator Filters. Auer. Ind. Hyg. Assoc.
J. 33 (12), 791-796 (December 1972).
3. Lapple, C. E., and H. J. Kamack, Performance of Wet Dust Scrubbers.
Chem. Eng. Prog. 51 (3), 110-121 (March 1955).
4. Semrau, K. T., Dust Scrubber Design—a Critique on the State of the
Art. J. Air Poll. Control Assoc. 13 (12), 587-594 (December 1963).
5. Semrau, K. T., C. W. Marynowski, K. E. Lunde, and C. E. Lapple,
Influence of Power Input on Efficiency of Dust Scrubbers. Ind. Eng.
Chem. 50 (11), 1615-1620 (November 1958).
6. Stevens, H. R., and H. B. Albregt, Assessment of Ultrasonic Nebuli-
zation. Anesthesiology 27 (5), 648-653 (September-October 1966).
7. Stober, W. and H. Flachsbart, An Evaluation of Nebulized Ammonium
Florescein as a Laboratory Aerosol. Atmos. Environ. 7_ (7), 737-748
(July 1973).
8. Stober, W., H. Flachsbart, and C. Boose, Distribution Analyses of
the Aerodynamic Size and the Mass of Aerosol Particles by Means
of the Spiral Centrifuge in Comparison to Other Aerosol Precipi-
tators. J. Coll. Interface Sci. 39 (1), 109-120 (April 1972).
9. West, P. H., H. P. Markant, and J. H. Coulter, New Venturi Scrubber
Developments. The Steam-Atomized Venturi. Tappi 44 (10), 710-715
(October 1961).
86
-------
GLOSSARY
Actual Conditions (gas) - The actual temperature and pressure of a gas
at the point of reference.
Collection Efficiency - The fraction (IJ) or percentage (100TJ) of an
aerosol entering a collection device that is removed from the gas
stream by the device.
Contacting Power - The power that is dissipated in fluid turbulence
^and ultimately as heat) in the gas and liquid phases during gas/liquid
contacting in a scrubber. It is commonly expressed in units of power
per unit of volumetric gas flow rate:
3 3
KwClOOO m /hr) or Kwh/1000 m Metric units
hp/1000 (ft3/min> English units
Diffusiophoresis - Movement of particles through a nonuniform gas
resulting from the existence of concentration gradients. It is produced
by the uneven molecular bombardment of the particles due to gradient in
gas composition.
Affective Friction Loss - The equivalent of contacting power, expressed
in terms of equivalent gas pressure drop (due to friction loss only)
across the gas/liquid contactor. In most cases, it is approximately
equal to the actual gas pressure drop across a device that takes all its
contacting power from the gas stream. It is commonly expressed in the
same units as gas pressure drop:
mm WC (millimeters of water column) Metric units
inches WC (inches of water column) English units
Liquid-to-Gas Ratio - The ratio of the volumes of liquid and gas brought
into contact in a gas/liquid contactor. It may be expressed on the basis
°f the gas volume either at actual or at standard conditions:
87
-------
3
liters/Am (liters of water per actual Metric units
cubic meter of gas)
[liters of water per
cubic meter of gas)
[gallons of water
actual cubic feet of gas)
3
litors/Sm (liters of water per standard
3
gal/1000 Aft (gallons of water per 100O English units
3
gal/1000 Sft (gallons of water per 10OO
standard cubic feet of gas)
Scrubber Performance Curve - A correlation curve expressing the functional
relationship between the number of transfer units and the contacting power
(or effective friction loss). This is most conveniently presented on a
log-log plot.
Scrubber Pressure Drop - The total drop in gas pressure across a gas
scrubber, whether due to friction loss or to kinetic energy changes.
It is commonly expressed as the equivalent fluid head loss:
mm WC (millimeters of water column) Metric units
inches WC (inches of water column) English units
It may also be expressed in dimensionless form as the number of velocity
heads, based on the gas velocity at the entrance to the gas/liquid con-
tactor or some other suitable entry point to the scrubber.
Standard Conditions (gas) - The selected reference temperature and pres-
sure for specification of gas volumes. In this report, the standard con-
o o
dltions are 25 C (77 F) and 1 atm pressure.
Stefan Flow (also, Stephan Flow) - The net convective gas flow produced
by the diffusion of vapor away from a surface where evaporation of liquid
is taking place, or by diffusion of vapor toward a surface where con-
densation is taking place. (Definition specific to circumstances of
interest to aerosol deposition.)
Transfer Unit - Defined by the relationship
N -
where N = number of transfer units and n = fractional collection
efficiency.
88
-------
Velocity Head - The kinetic energy per unit mass of a flowing fluid,
expressed in terms of fluid head:
mm
WC (millimeters of water column) Metric Units
inches WC (inches of water column) English Units
89
-------
CONVERSION FACTORS
Gas Flow
3 3
1 m /hr = 0.58857 ft /rain
1 liter/min = 0.035316 ft /min
Gas Velocity
1 m/sec = 3.2808 ft/sec
Liquid Flow
1 liter/min = 0.26418 gal/min
1 liter/hr = 0.0044030 gal/min
1 m /hr = 4.4029 gal/min
Liquid-to-Gaa Ratio
1 liter/ra3 = 7.4808 gal/1000 ft3
Contacting Power
3 3
1 Kw/1000 (ra /hr) = 2.2784 hp/1000 (ft /nin)
Effective Friction Loss
1 mm WC = 0.039370 inch WC
1 mm WC = 2.724 x 10~3 Kw/1000 (ra /hr)
3
1 inch WC = 0.1575 hp/1000 (ft /min)
Aerosol Concent rat ion
3 3
1 g/m = 0.43699 grain/ft
3 —7 3
1 M,g/m = 4.3699 x 10 grain/ft
3
1 g/llter = 436.97 grains/ft
—4 3
1 tig/liter = 4.3697 x 10 grain/ft
90
-------
APPENDIX A
*
LITERATURE REVIEW
Although particulate scrubbers went into use long ago, sometimes
with marked success, their wider application was long hampered by the
lack of any systematic procedure for designing—or even for selecting—
units for a specific application and specific duty. In a general way, a
number of investigators had noted that high scrubber collection effici-
ency generally coincided with high power consumption, but this was regarded
mainly as a disadvantage to be circumvented in some way if possible. The
14
recognition by Lapple and Kamack that the relationship of scrubber
collection efficiency to power consumption was actually a functional one
did not eliminate the disadvantage of high power consumption, but did
provide a highly useful if empirical method for designing scrubbers for a
19,20,21
given service. Later development of this method for correlating
scrubber performance data led to the "contacting power concept" which has
become, over the past 20 years, a widely accepted basis for practical
scrubber design, employed by many manufacturers as well as users of scrubbers.
The current demand for higher collection efficiencies for fine parti-
cles has led to increased interest in possible additional particle collec-
tion mechanisms that might accentuate the efficiencies of scrubbers without
further raising the power requirements. The use of water vapor condensation
phenomena was also suggested many years ago, but there has been very little
information presented to show what might be the magnitude of such effects
in practical scrubbing systems. At least two quite different mechanisms
5,19
are involved. One is the buildup of particles that act as nuclei for
the condensation of water vapor into droplets; this action increases the
The references cited in this appendix are collected at the end.
91
-------
effective size of the particles, making their deposition by the inertial
mechanism easier. This mechanism has, until recently, received the most
attention from research workers and others. The other mechanisms, diffu-
siophoresis and Stefan flow, are directly diffusional in nature, and have
received appreciable attention only in the last 25 years.
The available literature on scrubbers to 1962 was reviewed from the
19,20,21
standpoint of the contacting power correlation by Semrau et al,
14 27
following the pioneering paper by Lapple and Kamack. Walker and
28
Walker and Hall also used the contacting power basis in correlating
their own experimental data on scrubber efficiency as well as some data
from the literature.
Since 1963, most of the papers on scrubbers (other than those that
were purely descriptive) have fallen into two classes. The first consists
of those papers that treat applications and present some efficiency data,
or correlations of efficiency data, on the empirical basis of gas pressure
drop or some other expression of contacting power. Such papers have con-
tributed additional confirmation of the validity of the contacting power
concept but have not presented extensions of the earlier work. A few
empirical or semiempirical studies have attempted more systematic approaches,
and have reported either general agreement with the contacting power concept^
or some exceptions to it.
The second class of papers in this period has consisted mostly of
theoretical approaches to the problem of scrubber operation and design,
not usually accompanied by much, if any, experimental data for confirma-
tion of the theory. Most of these theoretical studies have treated the
case of the venturi scrubber.
Studies of condensation effects to date have been largely theoretical
or speculative in nature, but a number of laboratory investigations have
been carried out in recent years. Most of the latter have dealt with basic
92
-------
phenomena and did not attempt to treat applications to practical scrub-
bing equipment. There has been almost no quantitative information reported
on the magnitude of the effects attainable in actual scrubbing equipment.
POWER CONSUMPTION AND SCRUBBER EFFICIENCY
27
In experimental studies, Walker observed that the flooded-disc
(annular orifice) scrubber gave efficiencies on black liquor recovery
19.20,29
furnace fume that agreed well with those previously obtained
with venturi scrubbers. He also demonstrated improvements in the effi-
ciences obtained by electrostatic pre-agglomeration of the fume. In sub-
28
sequent studies of the flooded-disc scrubber, Walker and Hall concluded
that collection efficiency was dependent on the gas pressure drop and was
independent of gas velocity and liquid-to-gas ratio. They also concluded
that there was little effect of scrubber size on the efficiency/pressure
drop relationship.
23
Taheri and Haines studied the performance of a pilot-plant venturi
scrubber on a test aerosol generated by drying a spray of a solution of
methylene blue. They investigated the effects of three methods of water
injection: (l) injection through spray nozzles at the throat, (2) injection
through continuous slots at the throat, and (3) injection through a weir
at the beginning of the venturi convergence. They reported that injection
of the water through the weir produced a given efficiency for about 10%
less pressure drop than did the bther two methods of water injection.
They did not report any effects of gas velocity or liquid-to-gas ratio.
Their data on outlet aerosol concentrations show considerable scatter that
may have resulted from variations in the aerosol particle size.
24
Taheri et al. studied the efficiency of a pilot plant scrubber using
a butterfly valve as an adjustable orifice contactor. They used a methy-
lene blue aerosol, measuring the efficiencies in different particle-size
93
-------
ranges by sampling at the inlet and outlet of the scrubber with cascade
impactors. For particles in two size ranges, 1.0-3.0 microns and under
0.6 micron, efficiencies were shown as functions of pressure drop, with
no significant independent effects of liquid-to-gas ratio or butterfly
valve opening. For the intermediate-sized particles, 0.6-1.5 microns,
there were apparent, marked independent effects of liquid-to-gas ratio
and angle of opening of the valve. The authors do not appear to suggest
a clear explanation for this behavior, although they report an interaction
of inertial and diffusional collection mechanisms in the intermediate
particle-size range.
In another study with the same experimental system, but using fixed
25
orifice contactors, Taheri et al. reported that a 1.5-inch contactor
orifice gave higher efficiencies at given pressure drops than did a
1.0-inch orifice, except in the high pressure drop range. They also
reported that there were independent effects of liquid-to-gas ratio, as
well as of orifice size, in the lower pressure drop range. However, all
the tests in the lower pressure drop range were apparently made with low
or very low liquid-to-gas ratios, which might well show independent effects
and lead to Independent effects of orifice size as well.
A number of investigators have made theoretical analyses of the
venturi scrubber, but few of them have attempted experimental verifica-
2
tion of the theoretical models. Boll developed a model for the venturi
g
scrubber, and attempted to fit the data of Glintheroth, but acknowledged
that the agreement was only fair. Among major experimental investigators,
Guntheroth appears to make the strongest claims that venturi geometry,
gas velocity, and liquid-to-gas ratio have major independent effects on
10
efficiency.
Behie and Beeckmans developed a mathematical model of venturi scrub-
ber performance and attempted an experimental verification, but without
94
-------
attempting to correlate efficiency as a function of gas pressure drop.
Plots of their data suggest that the actual particle sizes of their test
aerosols may have varied considerably during some tests.
3 5
Calvert et al.' developed a mathematical model for the operation
of venturi and other atomizing scrubbers. They later extended this
analysis to present the efficiency of the scrubber as a function of the
gas pressure drop. Their design equations for particle penetration
through the scrubber depend on experimental evaluation of one of the
paraaeters, which may then absorb the effects of various actual deviations
from the theoretical development. It does not appear that the Calvert
model has yet received extensive experimental verification or evaluation.
CONDENSATION SCRUBBING
The possibilities of using condensation phenomena to assist in scrub-
bing of particulates have long been considered and were discussed at length
by Harmon in 1938 with specific reference to the cleaning of blast fur-
nace gas. The earlier suggestions were made almost entirely with reference
to the buildup of aerosol particle size by using the aerosols as nuclei
for the condensation of water vapor. However, more recent theoretical
7 ft 15 1_S 22
and experimental studies ' ' ' ' have treated the phenomena of dlffu-
siophoresis and Stefan flow. Many studies have been entirely theoretical,
faut in recent years there have been a number of laboratory investiga-
tions7'8'15'16 that have clearly demonstrated the existence and nature of
the phenomena. However, there has been very little information to indicate
the magnitude of possible effects in practical or conventional scrubbing
equipment. Apparently, the first published data of this sort were those
°f Semrau et al, who found different relationships between contacting
Power and collection efficiency in the scrubbing of Kraft recovery furnace
29
fume with hot and cold water. West et al<- later obtained very similar
data, although they did not ascribe the differences in scrubber performance
95
-------
to condensation phenomena. Other reports of the improvement of industrial
scrubber performance by scrubbing with cold water have been qualitative.
Calvert et al» have recently presented a theoretical analysis of the
use of condensation scrubbing for collection of fine particles. They
also obtained some data on the efficiency of collection of a submicron
aerosol by a single sieve-plate. The sieve-plate collector had a low pres-
sure drop and only small amounts of water vapor were condensed during the
tests. The measured collection efficiencies, which ranged from 11 to 31%,
were compared with the values predicted by the theory developed by the
investigators. Apparently, no data were obtained on the efficiency of
the device in the absence of water vapor condensation.
12 13 17
It is noteworthy that in recent years a number of investigators ' '
have successfully correlated mass-transfer data in cocurrent contactors as
functions of energy dissipation. Most such correlations have been devel-
oped for various forms of pipeline contactors but at least one of the
studies was made of a packed column with cocurrent gas and liquid flows.
The similarities of these correlations to the contacting power correlation
for particle collection is obvious. The existence of these correlations
strongly suggests that in practical scrubbing equipment, particle collec-
tion by diffusiophoresis and Stefan flow should also be a function of
contacting power. However, there is no present reason to assume that the
functions will be the same for different types of contactors .
In both the steel and pulp and paper industries, scrubber systems have
been fitted with cooling towers or other contactors in which the gases are
26,30
cooled and dehumidified. These systems were installed for reasons
other than attempts to increase collection efficiency, and apparently no
effort has been made to determine whether there have been increments of
efficiency resulting specifically from the condensation of water vapor.
96
-------
In a very recent paper, Calvert et al, report an extension of
f*
the earlier work on condensation scrubbing by a single sieve-plate
collector. The investigators studied the collection of iron oxide
aerosols by a three-plate bench-scale scrubber and a five-plate pilot-
scale scrubber. The units were operated with countercurrent flow of
water (entering cold), and the entering gas was saturated with water
vapor at its inlet temperature. Each scrubber was operated at essen-
tially the same gas pressure drop in all tests. As the quantity of water
vapor condensed (g/g dry air) was increased, collection efficiency
increased markedly.
a A
Calvert et al, also presented a plot (reproduced here as Figure 24)
that summarizes the results of a number of condensation scrubbing studies
reported in the literature; penetrations are shown as functions of the
amount of water vapor condensed in or injected into the respective scrub-
ber systems (expressed as g/g dry gas). Where investigators carried out
tests over ranges of condensation ratio, there are indicated reductions
in the particulate penetration with increasing condensation ratio. The
operating range observed in the present investigation has been added to
the figure, appearing as the shaded vertical band. The condensation
ratio used varied only slightly in the tests, but the range of values of
the penetration was determined by the effective friction loss across the
scrubber, which was the other major variable determining the penetration.
97
-------
1.0
o
I-
o
<
o:
LJ_
O.I -
o:
\-
u
z
u
a.
0.01
PRESENT INVESTIGATION
8
7,Na2CO, \ \
,np=I
LOW VELOCITY x
j
4, HIGH VELOCITY
i t i i i—
J L
0.01 0.05 O.I 0.5 1.0 5.0
CONDENSATION OR INJECTION RATIO (g H20/g DRY GAS)
FIGURE 24 PERFORMANCE OF CONDENSATION SCRUBBERS—LITERATURE DATA (FIGURE 12,
REF 6A) — RANGE OF PRESENT INVESTIGATION ADDED
(Number on curves refer to references in Ref. 6A.)
98
-------
REFERENCES
APPENDIX A
1. Behie, S. W., and J. M. Beeckmans, On the Efficiency of a Venturi
Scrubber, Can. J. Chem. Eng. 5JL (4), 430-533 (Aug. 1973).
2. Boll, R. H., Particle Collection and Pressure Drop in Venturi
Scrubbers, Ind. Eng. Chem. Fundam. 12 (1), 40-50 (1973).
3. Calvert, Seymour, Venturi and Other Atomizing Scrubbers Efficiency
and Pressure Drop, AIChE Journal 16 (3), 392-396 (May 1970).
4. Calvert, S., J. Goldshraid, D. Leith, and D. Mehta, Scrubber Handbook,
Vol. I, A.P.T., Inc., Riverside, California (July 1972), EPA Contract
No. CPA-70-95, NTIS: PB-213 016.
5. Calvert, S., J. Goldshmid, and David Leith, Scrubber Performance
for Particle Collection, AIChE Symposium Series 70 (137), 357-364
(1974).
6. Calvert, S., J. Goldshmid, D. Leigh, and N. C. Jhaveri, Feasibility
of Flux Force/Condensation Scrubbing for Fine Particulate Collection,
EPA Rept. No. EPA-650/2-73-036 (Oct. 1973), NTIS: PB-227 307.
6A. Calvert, S., and N. C. Jhaveri, Flux Force/Condensation Scrubbing,
Proc. EPA-APT Fine Particle Scrubber Symposium (May 28-30, 1974),
pp. 86-118.
7. Fuchs, N., and A. Kirsch, The Effect of Condensation of a Vapor on
the Grains and of Evaporation from Their Surface on the Deposition
of Aerosols in Granular Beds, Chem. Eng. Sci. 20, 181-185 (1965).
8. Goldsmith, P., and F. G. May, Diffusiophoresis and Thermophoresis
in Water Vapor Systems, Aerosol Science, Davies, C. N. (ed.),
Academic Press, New York (1966), pp. 163-194.
9. Guntheroth, H., Schwebstoff-Nassabscheidung aus Gasen mit dem Venturi
Scrubber, Fortschritt-Ber. VDI-Z Series 3, Issue 13 (1966).
10. Guntheroth, Hans, An Artificially Generated Aerosol for Wet Scrubbing
Tests on a Semi-Industrial Scale, Staub-Reinhalt Luft (in English)
28 (11), 82-87 (Nov. 1968).
99
-------
REFERENCES (continued)
11. Harmon, Robert Rogers, Removal of Suspended Matter from Industrial
Gases, J. Inst. Fuel 11_ (60), 514-527 (Aug. 1938).
12. Jepson, John C., Mass Transfer in Two-Phase Flow in Horizontal
Pipelines, AIChE Journal 16_ (5), 705-711 (Sept. 1970).
13. Kulic, E., and E. Rhodes, Chemical Mass Transfer in Co-Current Gas-
Liquid Slug Flow in Helical Coils, Can. J. Chem. Eng. 52 (1),
114-116 (Feb. 1974).
14. Lapple, C. E., and H. J. Kamack, Performance of Wet Dust Scrubbers,
Chem. Eng. Prog. 51 (3), 110-121 (Mar. 1955)
15. Meisen, A., A. J. Bobkowicz, N. E. Cooke, and E. J. Farkas, The
Separation of Micron-Size Particles from Air by Diffusiophoresis,
Can. J. Chem. Eng. 19 (4), 149-157 (Aug. 1971).
16. Melandri, C., V. Prodi, 0. Rimondi, and G. Tarroni, Submicron
Particle Precipitator Based on Water Vapor Condensation, Treat.
Airborne Radioactive Wastes, Proc. Symp. 1968, 541-550
17. Reiss, L. Philip, Cocurrent Gas-Liquid Contacting in Packed Columns
I & EC Proc. Des. Develop. 6 (4), 486-599 (Oct. 1967).
18. Rozen, A. M., and V. M. Kostin, Collection of Finely Dispersed
Aerosols in Plate Columns by Condensation Enlargement, Int. Chem.
Eng. 7 (3), 464-467 (July 1967).
19. Semrau, Konrad T., Correlation of Dust Scrubber Efficiency, J. Air
Poll. Control Assoc. 10 (3), 200-207 (June 1960).
20. Semrau, Konrad T., Dust Scrubber Design—A Critique on the State of
the Art, J. Air Pollution Control Assoc. ^3 (12), 587-594 (Dec. 1963).
21. Semrau, K. T., C. W. Marynowski, K. E. Lunde, and C. E. Lapple,
Influence of Power Input on Efficiency of Dust Scrubbers, Ind. Eng.
Chem. 50 (11), 1615-1620 (Nov. 1958).
22. Sparks, L. E., and M. J. Pilat, Effect of Diffusiophoresis on
Particle Collection by Wet Scrubbers, Atmos. Environ. 4, 651-660 (1970),
23. Taheri, M., and G. F. Haines, Optimization of Factors Affecting
Scrubber Performance, J. Air Poll. Control Assn. 19_ (6), 427-431
(June 1969).
100
-------
REFERENCES (concluded)
24. Taheri, M., S. A. Beg, and M. Beizaie, Gas Cleaning in a Wetted
Butterfly Valve, J. Air Poll. Control Assn. 22 (10), 794-798
(Oct. 1972).
25. Taheri, M., S. A. Beg, and M. Beizaie, The Effect of Scale-Up on
the Performance of High-Energy Scrubbers, J. Air Poll. Control Assn.
23 (11), 963-966 (Nov. 1973).
26. Vegeby, Anders, Hot Water from Recovery Boiler Stack Gas: SF
Scrubber—MODO System, Pulp Paper Mag. Can. 69 (9), 68-74 (May 3, 1968).
27. Walker, A. B., Enhanced Scrubbing of Black Liquor Boiler Fume by
Electrostatic Pre-Aggloraeration: A Pilot Plant Study, J. Air Poll.
Control Assn. 13 (12), 622-627 (Dec. 1963).
28. Walker, A. B., and R. M. Hall, Operating Experience with a Flooded-
Disc Scrubber—a New Variable Throat Orifice Contactor, J. Air Poll.
Control Assn. 18 (5), 319-323 (May 1968).
29. West, P. H., H. P. Markant, and J. H. Coulter, New Venturi Scrubber
Developments. The Steam-Atomized Venturi, Tappi, 44 (10), 710-715
(Oct. 1961).
30. Willett, H. P., and D. E. Pike, Venturi Scrubber for Cleaning Oxygen
Steel Process Gases, Iron & Steel Engr., 38 (7), 126-131 (July 1961).
101
-------
APPENDIX B
ANALYSIS OF EXPERIMENTAL DATA
The scrubber performance curves drawn through the data points in
the various correlations were placed by eye, and not all points were
accorded equal weight. Rather, judgments were made of the probable
reliability of individual points and some points were essentially
ignored. In general, the curve was selected to give the best appear-
ing straight line over the widest possible range of pressure drop.
When the data points in the low pressure drop range appeared to fall on
a different curve, they were fitted (so far as possible) by separate
straight lines.
This procedure involved relatively little uncertainty in those cases
where little scatter appeared in the data points and a straight line was
clearly defined. For the Ambient Scrubbing tests, considerable scatter
usually remained in the data even after they were rationalized by plot-
ting according to ranges of aerosol generation rate. In these instances,
the bias injected during curve fitting was substantially greater.
Although the Ambient Scrubbing tests were originally made with the
intention of displaying systematically the effects (if any) of orifice
gas velocity, liquid-to-gas ratio, and orifice size, the effort was
largely frustrated by the variations in the aerosol. Thus, the individual
correlations according to aerosol generation rates (Figure 6-10) were
composed of data having somewhat random values of gas velocity and liquid-
to-gas ratio. The transfer unit data were well rationalized by plotting
them as functions of effective friction loss, without reference to ori-
fice diameter, gas velocity, or liquid-to-gas ratio. This point is
illustrated more clearly and forcefully by the plots in Figures 25-28,
102
-------
3.0
2.0
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
a
o
-OD-
Cl
I I I I
4 5678910
LIQUID-TO-GAS RATIO
20
liters/m3
40 60 80 100
UJ
a
o
3
3.0
2.0
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
1
I I I I I I I I
. — —
0 D
0
a
n-l PI .. _,
Q
— —
— —
— —
— —
_ —
I I I I I I I I
0 20 40 60 80 10
GAS VELOCITY IN ORIFICE — m/«c
S A-2745-26
FIGURE 25 ANALYSIS OF RESIDUAL DEVIATIONS—AMBIENT SCRUBBING TESTS
Aerosol generation rates 750-1000
103
-------
z
z
LU
o
g
to
LU
DC
3.0
2.0
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
1
3.0
2.0
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
I I II IN
1 1 1 Mill
„ —
V Z
V ^ v v V _
— —
_ —
— —
_ —
I I I I 1 1 1 1 1 1 1 1 1 1 1 1
2 3 4 5 6 7 8 9 10 20 40 60 80 10
LIQUID-TO-GAS RATIO — liters/m3
1 1 1 1 1 1 1 1
I ^ * V" I
— —
— —
— —
_ —
1 1 1 1 1 1 I 1
10 20 40 60 80 100
GAS VELOCITY IN ORIFICE — m/sec
SA-2745-27
FIGURE 26 ANALYSIS OF RESIDUAL DEVIATIONS—AMBIENT SCRUBBING TESTS
Aerosol generation rates 1000-1200 jig/min.
104
-------
g
<
>
Q
_J
<
g
-------
g
<
UJ
D
CO
UJ
tc.
3.O
2.0
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
1
3.0
2.0
1.0
0.9
0.3
0.7
0.6
0.5
0.4
0.3
I 1 I 1 1 1 1 i 1 1 1 1 1 1 1 I I
_ —
A * * A A
-
_ A —
— —
— —
— —
— —
1 1 1 f 1 1 1 1 1 1 1 1 1 1 1 1
2 3456789 10 20 40 60 80 10
LIQUID-TO-GAS RATIO — liters/m3
1 1 1 1 1 1 1 1
A jU .
<> /j
~~ A
— A —
— —
— —
_ —
— —
1 1 1 1 1 1 1 1
10 20 40 60 80 100
GAS VELOCITY IN ORIFICE — m/sec
SA-2745-29
FIGURE 28 ANALYSIS OF RESIDUAL DEVIATIONS—AMBIENT SCRUBBING TESTS
Aerosol generation rates 1500-1900 jig/min.
106
-------
where the residual deviations of the data points from the correlation
curves are plotted as functions of liquid-to-gas ratio and gas velocity
in the contractor orifice.
The residual deviations were determined graphically, and laid off
from the reference axis 1.0 on the same log-log grid used for the origi-
nal plot. Any trends with liquid-to-gas ratio or gas velocity that are
not reflected in the effective friction loss should thus be made readily
visible.
The dispersion of the residuals in Figures 25-28 is exaggerated by
the bias introduced in placing the original correlation curves. Never-
theless, no significant independent trends in performance with varia-
tions in liquid-to-gas ratio or gas velocity are indicated. The effects
of these variables are evidently completely absorbed in the effective
friction loss.
No such plots were made for the separate correlations in the low
pressure drop regions of Figures 7 and 8, since the data in these regions
were inadequate to define good correlation curves in the first place.
No residuals were plotted for the correlations of Figure 6, since it was
obvious by inspection that the residuals would be very small except for
the one data point presumed to represent a systematically different
aerosol.
In Figure 29, similar plots are presented for the residual devia-
tions of the data points for the Adiabatlc Saturation scrubbing tests
with Aerosol A (Figure 13). In this case, the original correlation was
generally better than those for the Ambient scrubbing tests. The absence
of any significant trends in the residual deviations is again clear, even
though the liquid-to-gas ratio—in particular—was varied over a range
of nearly fourteenfold.
107
-------
Ul
Q
Q
in
UJ
tr
3.0
2.0
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
3.0
2.0
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
I I I I I
cP
I I I I M I
456789 1O 20
LIQUID-TO-GAS RATIO — liiers/m3
40
60 80 100
I
ORIFICE DIAMETER Icm)
O 2.54
A 3.81
10 20 4O 60 80 100
GAS VELOCITY IN ORIFICE — m/sec
SA-2746-3O
FIGURE 29 ANALYSIS OF RESIDUAL DEVIATIONS—ADIABATIC SATURATION
SCRUBBING—AEROSOL A
108
-------
APPENDIX C
POTENTIAL COMMERCIAL APPLICATION
OF CONDENSATION SCRUBBING
The results of the present investigation as well as of earlier ones
have indicated that the contacting power requirement for scrubbing of a
given aerosol may be substantially reduced if condensation phenomena can
be employed to enhance the more conventional scrubbing mechanisms. In
addition to the reduction in contacting power, there is an additional
saving on power that results from the cooling of the gas and the conse-
quent reduction in the volume of the gas that must be handled by the
induced-draft fan. Particularly in cases where the original gas has a
high water vapor content, the condensation of the vapor as well as the
reduction in temperature can combine to produce a major reduction in the
gas volume that must be handled.
The use of gas cooling towers following some scrubbers on steel making
furnaces (see Appendix A) has apparently been intended primarily to gain
public relations advantages by eliminating steam plumes from stacks, but
the saving of fan power may be of equal weight. Apparently it was (and
perhaps still is) a common practice to discharge the cooling water from
the towers directly to sewers, since the water contained relatively little
particulate matter. However, if such discharges are stopped by water
pollution control regulations (as they probably will be), the cooling water
will itself have to be treated and recirculated after being cooled by direct
or indirect means. In such a case, the economics of cooling and dehumidi-
fying the scrubber exit gas will present a different and less favorable
picture.
109
-------
On the other hand, the cleaning and dehumidification of blast
furnace gas for immediate or delayed use as fuel is a general necessity.
Hence, the conditions for potential use of condensation scrubbing already
exist in this case. A similar potential application exists where basic
oxygen furnaces are equipped with noncombustion gas collection systems.
When the offgas is accumulated for later use as fuel, it will—like blast
furnace gas—have to be both cleaned and dehumidified. Yet another simi-
lar situation exists where gases from smelters, pyrite roasters, or other
similar sources are used to make sulfuric acid by the contact process
(see Appendix A). The feed gas to the contact plant must be thoroughly
cleaned and dehumidified, and requirements for final gas cleanliness and
dryness are more critical than those for blast furnace gas.
Gas cleaning and conditioning systems for blast furnace gas and
acid-plant feed gas are already well developed and reasonably well stan-
dardized after long experience, even though they may still be far from
optimum. Gas scrubbing systems for unburned EOF offgas are in commercial
use, but apparently are not yet as well developed or standardized as in
the other two applications. In all three of these applications of scrub-
bers, condensation mechanisms of particle collection may already be oper-
ative, but possibly not in optimum ways. If condensation phenomena can
be used more effectively, it may require no more than relatively minor
modifications of the systems to accomplish this.
A fourth possible application of condensation scrubbing is in the
collection of fume from a Kraft black liquor recovery furnace (see Appen-
dix A). This collection problem is a more conventional one than the others
mentioned above. The gases already have high moisture contents (25 to 35%
by volume), and in this respect are representative of a considerable num-
ber of industrial waste gases (e.g., gases from wet-process cement kilns
and from lime kilns burning precipitated lime; tail gases from thermal
phosphoric acid furnaces). Scrubbers are already extensively used on
110
-------
black liquor recovery furnace gases, either alone or in combination with
electrostatic precipitators. Pulp mills utilize substantial amounts
of low-level heat, particularly for heating water for pulp washing.
Considerable use is already being made of scrubbers—particularly in
Scandinavia—to heat water by contact with the hot gas from recovery
furnaces, and this results in condensation of water from the gases. In
such cases, condensation effects may be brought into play inadvertently,
but apparently there are no operating installations that were designed
deliberately to take advantage of condensation.
Still another possible application for condensation scrubbing is
the cleaning of the flue gases from sewage sludge incinerators. Because
of the high moisture contents of the sludge filter cakes being burned,
the water vapor content of the flue gas is commonly equivalent to a
o
saturation temperature of about 80 C. Scrubbing with sufficient water
at ambient temperature will thus produce a large amount of condensation,
which will increase the collection efficiency. However, the fly ash from
the incinerators apparently contains little material that is fine enough
in particle size to present a difficult collection problem. Therefore,
there may not actually be a strong economic incentive to employ condensa-
tion scrubbing.
RECOMMENDED APPLICATION FOR TESTING
Scrubbing of black liquor recovery furnace gases appears to be by
far the most advantageous application for a field demonstration of con-
densation scrubbing in a potential commercial application. It does not
have the highly specialized aspects of the cleaning of blast furnace and
EOF gases. The fume is of a relatively uniform nature and composition
and is generated on a comparatively steady basis, and considerable infor-
111
-------
mation is already available on its scrubbing characteristics, including
preliminary indications of possible improvements in performance realizable
from use of condensation. Also, the use of low-level heat now wasted in
pulp mills is becoming more attractive as the costs of fuel rise. There
is already extensive experience with the use of commercial scrubbers,
including observations that the performances of small and large units at
the same contacting power level are essentially the same. Hence, it
should be possible to apply system modifications developed on a pilot-
plant scale to a full-scale commercial unit with minimum risk.
PILOT-PLANT SYSTEM
If a field demonstration is carried out it is recommended that the
investigation be carried out on a pilot-plant basis with a scrubber unit
3 3
(or units) having a gas flow capacity of about 850 m /hr (500 ft /min),
measured at inlet conditions. Such a pilot plant is small enough to be
operated with relative ease on a slip stream of gas from a commercial
recovery furnace without interfering with, or being unduly interfered
with, by the operational requirements of the furnace. Modifications of
the experimental system can be made quickly and at modest costs. At the
3
largest, the pilot plant gas capacity should not exceed 1700 m /hr
3
(1000 ft /min). The fan capacity should be adequate to handle scrubber
pressure drops of up to at least 1800 mm WC (70 inches WC).
A minimum program can be elected, or an optional expanded program
can be undertaken at higher cost. If the expanded program is adopted,
the following major system components will be required:
A. A simple baffle chamber with a spray nozzle located at the inlet.
The chamber should be designed for a maximum superficial gas
velocity of about 1.0 m/sec (3 ft/sec) and to give a nominal gas
residence time of up to three seconds.
B. A variable-orifice scrubber with a cyclone entrainment separator.
112
-------
C. A scrubber with alternative variable-orifice and pipeline contactors
and a cyclone entrainment separator.
D. A countercurrent packed tov/er with provisions for installation of
up to 4 m (13 ft) of grid packing. The tower should be designed
for a maximum superficial gas velocity of 1.5 m/sec (5 ft/sec).
It should be equipped with a louvre- or herringbone-type of entrain-
ment separator, designed for a maximum superficial gas velocity of
1.5 m/sec (5 ft/sec).
E. Induced-draft fan.
F. Water-circulating system, including pumps and shell-and-tube heat
exchangers for cooling recirculated scrubbing water.
If the minimum program is elected, then A, the baffle chamber, and
D, the packed tower, can be eliminated. The collectors in the system
should be so equipped that they can be fitted together in any desired
combination, although only certain combinations are contemplated. Pro-
visions should be made to sample the gas not only at the inlet and outlet
of the system but between any units placed in series.
MINIMUM PROGRAM
The basic, or minimum program will make use of the two cocurrent-
contact scrubbers B and C in the following steps:
The reference operating condition will be adiabatic saturation
scrubbing of the gas, using the variable-orifice scrubber C. Sufficient
data will be taken to establish a scrutbar performance curve for a wide
range of effective friction loss (up to at least 1500 mm WC).
Scrubber C will then be operated with cold water as a
condensation scrubber, to give an exit gas temperature of
50 C. Scrubber performance curves will be established for
both the orifice and pipeline contactors, if the two devices
should give different curves.
113
-------
• Scrubbers B and C will be used in series, with B operated
as a low-energy adiabatic saturator and with C operating
on the presaturated gas as a condensation scrubber with an
o
exit gas temperature of 50 C. The efficiencies of each
scrubber individually and of the two in combination will
be determined. Sufficient data will be taken to establish
scrubber performance curves over the range of effective
friction loss up to at least 1500 mm WC total (for both
scrubbers).
EXPANDED PROGRAM
The expanded program will incorporate the minimum program, together
with additional experiments using the baffle chamber A and the counter-
current packed tower D.
The baffle chamber will be used as a gas quencher rather
than as a scrubber. It will be followed by Scrubber C,
used in both adiabatic saturation scrubbing and condensation
scrubbing. The baffle chamber will be fed with only enough
cold water to bring the inlet gas to saturation and—perhaps—
to cause fume particles to nucleate droplet formation. If
significant buildup of the particles takes place, it may be
revealed by the performance curves for the scrubber.
The countercurrent packed tower D should represent a
low-energy device that nonetheless will produce a large
amount of water vapor condensation. It will be operated in
series with, and downstream of, Scrubber C, which will be
operated as an adiabatic saturation scrubber at two levels
of effective friction loss but at the same gas flow rate.
The overall efficiency and the separate efficiencies of each
114
-------
device will be determined. For each level of prescrubber
effective friction loss, the efficiency of the packed tower
will be studied as a function of water rate and amount of
water vapor condensed.
PROGRAM REQUIREMENTS
If the sizes of the pilot-plant units are kept within the capacity
3
range of 850 m /hr, the cost of the foregoing program can be kept within
a relatively modest figure. For the minimum program, the costs of the
equipment and experimental work could probably be kept within about
$250-275,000. Administrative and other costs involved in carrying on a
program at the plant site of a cooperating Kraft mill could add sub-
stantially to this estimate. The expanded program would perhaps add
$50-75,000 to the cost.
115
-------
4. TITLE AND SUBTITLE
Wet Scrubber Liquid Utilization
7. AUTHOR(S)
Konrad Semrau and Clyde L. Witham
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-650/2-74-108
2.
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
October 1974
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Stanford Research Institute
333 Ravens wood Avenue
Menlo Park, California 94025
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADJ-005
11. CONTRACT/GRANT NO.
68-02-1079
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
t3. TYPE OF REPORT AND PERIOD COVERED
Final; Through 7/13/74 .
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
. ABSTRACT
report gives results of bench-scale experiments on the ways in whicn
energy consumption, liquid-to-gas ratio (L/G), and water vapor condensation affect
the collection efficiency of a conventional, co cur rent -contact, orifice-type particu-
late scrubber. The test aerosol consisted of spherical particles of ammonium fluo-
rescein with a mass-median diameter of about 0. 6 micron. Scrubber collection
efficiency was determined and correlated as a function of the effective friction loss
across the scrubber. Without condensation or evaporation effects, scrubber efficient
was dependent only upon effective friction loss , with no independent influences of gas
velocity, L/G. or contactor orifice size, except possibly in the range of very low gas
velocities, L/G, and pressure drops. Large amounts of water evaporation reduced
collection efficiencies; but scrubbing hot humid gas with cold water (to produce con-
densation) significantly increased collection efficiency. The most favorable results
were obtained by presaturating the hot humid gas stream before contacting it with
cold water.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air Pollution
Scrubbers
Efficiency
Aerosol
Liquid Saturation
Condensing
8. DISTRIBUTION STATEMENT
Unlimited
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution Control
Stationary Sources
Liquid Utilization
Particulates
Energy Consumption
Liquid-to-Gas Ratio
19. SECURITY CLASS (ThisReport)
Unclassified
20. SECURITY CLASS (Thispage/
Unclassified
13B
07A
14A
07D
21. NO. OF PAGES
126
22. PRICE
EPA Form 2220-1 (9-73)
116
------- |