EPA-650/2-73-023
September 1973
ENVIRONMENTAL PROTECTION TECHNOLOGY SERIES
\
ul
O
-------
EPA-650/2-73-023
MANUAL METHODS
FOR SAMPLING AND ANALYSIS
OF PARTICULATE EMISSIONS
FROM MUNICIPAL INCINERATORS
by
John T. Funkhouser, Edward T. Peters, Philip L. Levins,
Arnold Doyle, Paul Giever, and John McCoy
Arthur D. Little, Inc.
Cambridge, Massachusetts
Contract Number EHSD 71-27
Program Element No. 1AA010
EPA Project Officer: John O. Burckle
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, N.C. 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, B.C. f 20460
September 1973
-------
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
-------
ACKNOWLEDGMENT
This program was conducted by two organizations—Arthur D. Little,
Inc., having the primary responsibility with Walden Research Corpo-
ration being a sub-contractor. Although Walden had an explicit
assignment to carry out the field sampling program, their contri-
bution extended well beyond this role.
During the course of this program, there were two staff members
of the Environmental Protection Agency who were primarily involved
in the direction of this program. We are indebted to John Burckle
and Robert Larkin (now associated with the National Institute for
Occupational Safety and Health) for their support, cooperation and
direction to this program.
ill
-------
FOREWORD
Airborne participate matter is a major air pollutant having
significant effects on health, economics, ecology, visibility,
and aesthetics. Effective techniques and hardware systems for
source emission measurements are required for application to the
various emitting sources to achieve control of particulate emissions
and protect the environment.
This contract work was sponsored by the National Environmental
Research Center at the Research Triangle Park, North Carolina, to
review methods applied to the measurement of particulate emissions
emanating from incinerators and, in particular, investigate the
validity of the EPA sampling train.
While today's incinerators contribute only a small portion of
the total particulate burden in the Nationwide Emission Inventory,
this portion is emitted in close proximity to urban centers. Further,
recent studies on solid waste disposal practices indicate an increase
in incineration practices in urban centers. The facts leading to
this conclusion are given as:
...Urban areas now account for some 50% of the U.S.
population with the trend for centralization
increasing
...The rate of production of urban wastes is
estimated to be about 8.5 pounds per capita per
day and increasing annually at about 4% per
capita
...Landfill sites for major urban centers are
rapidly disappearing and alternatives to
volume reduction through thermal processing
are not practiced to any significant extent.
This report is included in the Environmental Protection
Technology series - the series, devoted to new or improved technology
required for control and treatment of pollution sources to meet
environmental quality goals, includes reports of work dealing with
research, development, and demonstration of instrumentation, equipment,
and methodology to repair or prevent environmental degradation from
point and non-point sources of pollution.
John 0. Burckle
Project Officer
U.S. Environmental Protection Agency
Office of Research & Development
IV
-------
TABLE OF CONTENTS
1.
1.1.
1.2.
1.3.
2.
2.1.
2.2.
2.3.
2.4.
2.5.
2.6.
3.
3.1.
3.2.
3.3.
3.4.
4.
4.1.
SUMMARY
PURPOSE AND SCOPE
APPROACH
FINDINGS AND CONCLUSIONS
1.3.1. Sampling Procedures
1.3.2. Nature of Particulate Catch
1.3.3. Potential Errors in Sampling and Analysis
BACKGROUND
REFUSE COMPOSITION
THE INCINERATOR
PARTICULATE CHARACTERISTICS AND EMISSION MECHANISMS
NON-PARTICULATE EMISSIONS
INFLUENCE OF AIR POLLUTION CONTROL EQUIPMENT
SAMPLING TRAINS
EXPERIMENTAL PROGRAM
COLLECTION OF FIELD SAMPLES
TABULATION OF SAMPLING RUNS
OPERATION OF A COMMERCIAL EPA SAMPLING TRAIN
CHARACTERIZATION TECHNIQUES
PROGRAM RESULTS
CHARACTER OF THE PARTICULATE CATCH FOUND IN THE "FRONT
Page
1
1
1
2
2
3
3
5
5
5
11
11
11
14
17
17
19
19
22
25
25
HALF" OF THE SAMPLING TRAIN
-------
Page
4.2. CHARACTER OF THE IMPINGER CATCH
4.3. DISTRIBUTION OF SULFATE AND CHLORIDE BETWEEN THE PROBE/
CYCLONE FILTER AND IMPINGERS
4.4. PROCEDURES FOR MEASUREMENT OF THE PARTICULATE CATCH
4.5. RELATION TO PRIOR INFORMATION
APPENDIX A COMPOSITION OF INCINERATOR EFFLUENTS
APPENDIX B THE NATURE OF THE INCINERATOR PROCESS
APPENDIX C EVALUATION OF EXISTING SAMPLING TRAINS
APPENDIX D TABULATION OF SAMPLING RUNS
APPENDIX E QUALITATIVE CHARACTERIZATION OF PARTICULATE
CATCH
APPENDIX F QUANTITATIVE ANALYSIS OF PARTICULATE CATCH
APPENDIX G HIGH-EFFICIENCY GLASS FIBER FILTER PERFORMANCE
CHARACTERISTICS
APPENDIX H CHARACTERIZATION OF PARTICULATES DIRECTLY ON
THE FILTER
APPENDIX I SULFUR OXIDE STUDIES
APPENDIX J SOME IMPORTANT CONSIDERATIONS FOR SAMPLING OF
PARTICULATES
APPENDIX K METHOD 5 - DETERMINATION OF PARTICULATE
EMISSIONS FOR STATIONARY SOURCES
29
2.9
34
34
39
112
141
155
169
185
209
221
233
257
277
vi
-------
1. SUMMARY
1.1. PURPOSE AND SCOPE
In response to growing public demands for clean air, the Federal Govern-
ment is expending considerable effort to identify sources of air pollu-
tion and appropriate abatement procedures to control these sources.
Municipal incinerators, and similar stationary sources, though contribut-
ing only a few percent of the total national air pollution load, are
important sources of pollution near population centers. The particulate
matter they emit has a significant adverse effect on health, on materials
of construction and on visibility; they are responsible for many com-
plaints. Therefore, the Federal Government, through the Environmental
Protection Agency, has promulgated standards that specify the permissible
levels of particulate matter emitted from newly constructed incinerators
operating at or above a charging rate of 50 tons per day.
In many cases, however, the chemistry of the particulate species present
in these emissions is not well known. Furthermore, because sampling and
analytical methods have not been scientifically documented, methods
commonly used may not adequately consider the physical and chemical
changes that can occur in the assay process and thus lead to the forma-
tion of what is expressed empirically as "false particulate." Therefore,
the results that are presently obtained may not truly reflect the parti-
culate burden to the atmosphere. Consequently, there is a need to define
more thoroughly the chemical nature of particulate emissions from incin-
erators and to gain a better understanding of how the sample collection
equipment used by the EPA influences the physical and chemical properties
of the particulate.
The primary goal of the program has been to help develop the data base
and the technology which will permit representative measurements of source
particulate emissions to be obtained from waste incineration sources, and
from the particulate pollution control devices associated with such
sources. One of the prime requirements for these measurements is that
they reflect, as accurately as possible, the particulate burden on the
ambient air. Thus, it will be necessary to develop standard methods
which are sufficiently well documented to be applied in the field to new
and unfamiliar situations. Our part in this overall process was to help
develop a practical working understanding of the complex physical and
chemical interactions during sampling that lead to the formation of
particulates.
1.2. APPROACH
To accomplish this goal, we:
• Prepared an information base describing incineration processes,
sampling methodology, and procedures;
-------
• Delineated the technical requirements necessary to meet the program
objectives;
• Evaluated existing methods in terms of these requirements to deter-
mine their strengths and weaknesses;
• Filled in gaps in knowledge and understanding by reviewing the
literature and speaking with people knowledgeable in the field;
• Prepared a plan to characterize the emissions through field tests
and laboratory evaluations; and
• Performed the required field tests, using the sampling train
designated by EPA, and evaluated results.
1.3. FINDINGS AND CONCLUSIONS
Our findings and conclusions can be grouped into categories relating to
sampling procedures, nature of the particulate catch, and potential errors.
They are summarized below along with unanswered questions associated with
each of these categories that remained unanswered at the completion of this
program.
1.3.1. Sampling Procedures
• In our sampling runs, the EPA sampling train caught more than 80%
of the particulate catch in the probe/cyclone and filter.
Question: How does process cycle variation influence this
distribution ?
• The probe/cyclone collects a significant proportion of the parti-
culate.
Question: How does particulate size distribution affect this?
What is the influence of having the cyclone-filter
arrangement inside the stack?
• For incinerator particulate catch, little change in mass seems to
occur with aging of the sample.
Question: Is this true for a wide variety of incinerators or
just the two studied under this program?
• Very little of the material caught in the impingers is related to
what is normally considered particulate.
Question: In the case of incinerator sampling, do the impingers
have any role in the EPA train except to protect the
meters and pumps?
-------
• The measurement of impinge'r catch by means of weight may be
inappropriate.
Question: If the impinger oatch is important, titration for acid
equivalent or determination of chloride and sulfate
should be considered as an alternative approach for
incinerator sampling.
1.3.2. Nature of Particulate Catch
• The "front half" catch is almost totally inorganic mineral parti-
culate.
Question: How does composition vary with particle size? How does
the composition of condensible fraction vary with
temperature of collection?
• The impinger catch is essentially mineral acid.
Question: Even though the mass of mineral particulate in the
impingers is low, do significant quantities (number
basis) of "fines" collect there?
1.3.3. Potential Errors in Sampling and Analysis
• The use of Drierite to "dry" the particulate catch may lead to
erroneous results.
Question: Iflhat are proper conditions for measurement of filter
catch?
• With the possible exception of sulfates, there seemed to be little
false particulate formed in the "front half" of the sampling train.
Question: What sampling train arrangements and materials of con-
struction are needed to ensure that this does not
become a problem?
• Based on our findings, it is not clear whether the oxidation of S02
in the sampling train is a potential source of error in the impinger
catch.
Question: To what extent is this a problem, and what needs to bo
done to eliminate, it or keep it to a minimum?
-------
• Particulate catch is defined as all material other than uncombined
water which exists as a finely divided liquid or solid at stand-
ard conditions.
Question: Does the EPA train meet this definition? Would the
use of a dilution, cooling system or other procedures
do a better job?
-------
2. BACKGROUND
When attempting to adequately sample a complex effluent stream such as the
emissions from a municipal incinerator, it is important to understand the
situation in which the measurements were taken. As a result of the com-
plexity of the feed material and the difficulties in obtaining complete
combustion, incinerator emissions themselves are very complex and variable.
This is true for both the particulate* and the gaseous emissions, as can
be seen from the data presented in Table 1 and the detailed discussion of
the nature of the emissions from incinerators given in Appendix A. Some
of the factors that influence the nature of these emissions are discussed
below.
2.1. REFUSE COMPOSITION
A generalized average composition for the waste that must be handled by a
municipal incinerator is shown in Table 2. As the composition of the waste
changes, the emissions can be affected. For instance, the percentage of
the ash carried over usually ranges between 10 and 20% of the total avail-
able in the refuse so the ash content of the refuse is a major factor in
determining emission rates. As shown in Table 3, the ash content of incin-
erated material will vary widely from a few percent (newspaper, polyethylene
film, etc.) to quite high (rubber tape, shoe leather, filled plastics,
etc.)- Other factors associated with the character of the refuse are vola-
tile metals content and volatile combustible content. As either of these
increases, more emissions may be observed. A large volume of combustible
rich pyrolysis gas generated during the incineration of refuse with a high
volatile content tends to result in particulates with a high fraction of
soot and other combustibles.
2.2. THE INCINERATOR
The incinerator furnace provides the environment for controlled combustion
of solid wastes in air. The furnace usually includes grates to support
the burning material within a refractory or other enclosure. The refuse
is processed by controlled oxidation, with liberation of heat, that pro-
duces flue, or combustion gases and a residue or ash. The refuse to be
processed in the furnace must be fed at controlled rates with suitable
material handling equipment. The ash residue must be removed and trans-
ported in bulk to a disposal area, while the effluent flue gases must be
removed and delivered to the stack or chimney after suitable treatment
for control of air pollution.
*Particulates are defined as any material, other than uncombined water,
which exist as a finely divided liquid or solid at standard conditions
(70°F and 1° atm.). See Federal Register, 36.. No • 247» Part u> 24876-
24895 (December 23, 1971).
-------
TABLE 1
Composition of Incinerator Emissions*
Concentration Levels
Component
Nitrogen
Water
Oxygen
Carbon Dioxide
Carbon Monoxide
Sulfur Oxides (S02)
Nitrogen Oxides (N02)
Hydrocarbons
Hydrochloric Acid
Particulate
1
60-65
10-30
10-15
3-5
Ib/ton refuse
40-500
30-60
15-50
15-50
10-40
1 grain/SCF
3-35
2-4
1-3
1-3
1-2
6-10
*A summary of literature values, see Appendix A.
-------
TABLE 2
Average Refuse Composition—-As-Discarded Basis*
Category
Glass
Metal
Paper
Plastics
Leather & Rubber
Textiles
Wood
Food Wastes
Miscellaneous
Yard Wastes
Weight (%)
8.3
8.2
35.6
1.1
1.5
1.9
2.5
23.7
1.7
15.5
100.0%
Description
Bottles (primarily)
Cans, Wire, Foil
Various Types, Some with Fillers
Polyvinyl Chloride, Polyethylene, Styrene, etc., as Found
in Packaging, Housewares, Furniture, Toys and Nonwoven
Synthetics
Shoes, Tires, Toys, etc.
Cellulosic, Protein, and Woven Synthetics
Wooden Packaging, Furniture, Logs, Twigs
Garbage
Inorganic Ash, Stones, Dust
Grass, Brush, Shrub Trimmings
*ADL estimate—from "Systems Study of Air Pollution from Municipal Incineration," a report to NAPCA (now EPA)
under Contract CPA-22-69-23.
-------
TABLE 3
Ash Content of
Selected Industrial and Municipal Wastes
Ash Content Material
0-2 Polyethylene film, draft paper, waxed milk cartons,
vegetable food wastes, semi-cured tubes, nylon fabric
and yarn, newspaper
2-5 Banbury rubber scrap, vinyl scrap, plastic-coated
paper, lawn grass, balsam spruce, ripe tree leaves,
cooked meat scraps
5-10 Vinyl-coated fabric, expanded Ensolite, uncured
frictional duck, cured flash and molded goods
10-15 Vinyl-coated felt, raw batch stock, junk mail, wire-
braid hose
20-25 Shoe leather, rubber-coated fabric, cloth uppers,
ensolex trim, trade magazines
25-30 Foam scrap, heel and sole composition
<30 Cloth-backed foam, die strip, rubber tape
^"Systems Study of Air Pollution from Municipal Incineration," a
report to NAPCA (now EPA) under Contract CPA-22-69-23.
-------
The feeding of the refuse may be either batch or continuous, although the
recent trend has been toward the use of continuous firing to improve pro-
cess control. In most cases, refuse is batch fed directly into the furnace
with a clam-shell bucket or grapple attached to a traveling crane; the
rate of feed is controlled by the time cycle and the degree of bucket load-
ing. In a few plants, a front-end loader operating on a paved floor charges
the furnace.
A ram-type feeding device is sometimes used for controlled feeding. With
such a system, either the ram can clear the hopper at each stroke or an
oversize hopper can be filled with refuse and the ram used to shear a
horizontal section of refuse at selected intervals. The ram feeder provides
an air seal at the feed to the furnace—an improvement over the bucket or
the front-end loader systems of batch feeding, which usually let in undesir-
able quantities of cold air, as well as releasing occasional puffs of
flames or hot gases, while the charging gate is open. The inrush of cold
air can be detrimental to the inside refractory walls of the furnace and
can cause smoke evolution by cooling and quenching the burning process.
Newer designs for incinerator systems nearly always specify continuous
feeding of refuse to the incineration furnace. Continuous feeding can be
accomplished by means of a hopper and a gravity chute; a mechanical feeder,
such as a pusher, ram, rotary feeder, or the like, which can be filled
directly from a hopper supplied with refuse by a bucket and crane; a front-
end loader from a feeding floor, a conveyor transporting the refuse from
the receiving area; or an air injection system (shredding with suspension
burning).
The most frequently used system is the hopper and gravity chute. A rect-
angular hopper receives the refuse delivered by the crane and bucket. The
bottom of the hopper terminates in a rectangular chute leading downward
to the furnace grate or other feeder conveyor at the entrance to the
furnace chamber itself.
If suspension burning is to be employed in the incinerator furnace, the
refuse should be prepared by suitable shredding or grinding, and the most
desirable feeding method is air injection. Suspension burning has been
used successfully in waterwall boilers for the generation of steam from
waste products such as wood bark, bagasse, and similar materials. In con-
formance with practices in the fossil-fuel-fired boiler industry, shredded
refuse can be air injected for corner firing, for spreader firing, or for
"cyclone" firing—all of which are in commercial use for the generation of
steam from powdered or crushed bituminous coal. With suspension burning,
a burn-out grate is provided at the bottom of the water-wall furnace
chamber to permit larger particles or slow-burning materials to burn com-
pletely.
A more comprehensive description of the incineration process is given in
Appendix B, but even from the above it should be clear that to permit
interpretation of particulate emission data, as many known and suspected
-------
incinerator variables as possible should be recognized. Some of these
variable are:
• Incinerator Type: Batch, continuous grate or rotary kiln.
• Grate Type: The grate type (plus charging method—batch or
continuous) often defines the general design of the total furnace
system. Stoking intensity (causing attrition of the residue)
varies widely between systems and would be expected to increase
the fraction of the ash material in the size range capable of
suspension and to expose more of this suspendable material to
the undergrate air flow. The percent open area of the grate
surface controls the amount of ash material dropping through the
grate (siftings). This grate parameter varies widely (2-30%).
Since the small-particle-size ash is "lost" from the bed by this
mechanism, less is available for suspension in the undergrate
air flow.
• Batch Size, Method, and Batch Frequency in Batch Units; Emissions
will vary greatly with these variables. Sampling duration and
timing will be critical considerations.
• Combustion Chamber Design: Furnace design will affect mixing
patterns, particle settling, etc.
• Incinerator Size: Increasing the size of incinerator units
from 3 to over 100 TPD has been shown to result in higher
emission rates, but the effect of size on emission factors has
not been established quantitatively over the more practical
size range for municipal furnace units (50-300 TPD).
• Burning Rate: It is expected that higher emission rates will
be encountered at higher burning rates.
• Undergrate Air Velocity; Ash particles may be entrained when
the velocity of the gases through the fuel bed exceed the terminal
velocity of the particles. Undergrate air velocities typically
vary from a minimum of 10 SCFM/ sq. ft. of grate area to
100 SCFM/ sq. ft. On the basis of the terminal velocity, parti-
cles up to 70p (equivalent diameter) are expected to be entrained
at the lowest velocities and up to 400^ at the highest. One
systematic study of the effects of underfire air, secondary air,
excess air, charging rate, stoking interval, and fuel moisture
content on the emission rate from an experimental incinerator
has led to the conclusion that the velocity of the underfire
air was the variable that most strongly influenced particulate
emission rate.
10
-------
2.3. PARTICULATE CHARACTERISTICS AND EMISSION MECHANISMS
Particulate incinerator emissions occur mainly by the following mechanisms:
• The mechanical entrainment of particles from the burning refuse
bed;
• The cracking of pyrolysis gases; and
• The volatilization of metallic salts of oxides.
The first of these mechanisms is favored by refuse with a high percent-
age of small-particle low-density ash, by high underfire air velocities,
or by other factors that induce a high gas velocity through the bed. The
second mechanism is favored by refuse with a high volatile content that
produces pyrolysis gases with a high carbon content, and by conditions
above the fuel bed that prevent the burnout of the coked particles formed
by the cracking of the volatiles. The third mechanism is favored by
high concentration of metals that form low-melting-point oxides and by
high temperatures within the bed.
Two major types of particles are:
• Mineral particulate —the incombustible fraction of fly ash; and
• Combustible particulates—the char and soot produced by the
thermal cracking, and condensed organic vapor (white smoke).
A generalized summary of the incinerator emission rates for a variety of
pollutants is given in Table 4.
2.4. NON-PARTICULATE EMISSIONS
The analytical methodology employed for determining particulates must take
into account the influence of gaseous emissions, such as condensible
hydrocarbons and organic acids. The values shown in Table 4 are "typical"
for some of these materials. These typical values, however, will not
remain constant. For example, hydrochloric acid content is expected to
quadruple by the year 2000 because of an increase in polyvinyl chloride
resins (Table 5).
2.5. INFLUENCE OF AIR POLLUTION CONTROL EQUIPMENT
Although the flue gases from incinerators contain a number of pollutants,
most, if not all, air pollution control (APC) equipment installed on
existing units addresses the problem of particulate removal. For this
purpose, a number of devices are in use, ranging in particulate removal
efficiency from 5-15% to upward of 95%.
11
-------
TABLE 4
Emission Rate Estimates
For a Variety of Pollutants
Pollutants
Particulates
Mineral
Combustible
Carbon Monoxide
Hydrocarbons
Specific Organics
Acids
Aldehydes
Polynuclears
Nitrogen Oxides
Sulfur Oxides (total)
Inorganic Acids
Hydrochloric Acid
Hydrofluoric Acid
Volatile Metals
Amount
(Ibs/ton of refuse)
15
4.6
35
2.7
0.1
0.2
0.005
3.9
1.5
0.01
0.03
*A summary of literature values, see Appendix A.
12
-------
TABLE 5
Anticipated Growth Patterns of PVS in
Average Refuse^ and Associated HC1 Furance Emission Factors
Year
1968
1970
1975
1980
1990
2000
Plastics
in Refuse
(%)
1.1
1.3
1.8
2.7
3.5
4.2
PVC Resin
in Plsstics
(%)
7.8
0.0
0.6
8.7
0.9
1.2
HCL Emission Factor
Ib HC1 per Ton
of Refuse
0.99
1.50
2.20
2.71
4.41
5.44
*ADL estimate—from "Systems Study of Air Pollution from Municipal
Incineration," a report to NAPCA (now EPA) under Contract CPA-22-69-23.
13
-------
Settling chambers or expansion chambers have been used in this breeching
and flue gas ducts and many of the older installations have employed
refractory baffles across the breechings to collect particles. In some
instances, a coarse spray of water is directed into the flue gases and
toward the baffles with most of the water falling to the floor of the
chamber without vaporization. The wet floor and baffles improve parti-
culate removal by preventing re-entrainment of settled ash into the flue
gas stream. Dry collectors are usually "cyclones" in which the flue
gas particles concentrate on the inside of a cylindrical wall (as a
result of centrifugal force) and solids are discharged at the lower end
and opposite to the cleaned gas outlet.
Other devices used for particulate removal from flue gases include
scrubbers, which may be open spray chambers, packed chambers, or high-
pressure drop units (e.g., Venturi scrubbers). Filters used for parti-
culate removal usually are high-temperature fabrics, such as silicone-
treated glass fiber cloth, arranged in bags or tubes. Electrostatic
precipitators are currently receiving increased attention for particulate
removal from incinerator flue gases.
2.6. SAMPLING TRAINS
Procedures and devices should be able to remove a measured representative
sample of stack gas and to collect the particulate matter (solid and
liquid) which exists in the stack gas plus the potential particulate
matter, i.e., material which would condense to an aerosol (free liquid
water excepted) as incinerator gases cool by dilution with the ambient
atmosphere. The ultimate method should be able to collect particulate
and potential particulate matter without introducing any extraneous
material that could be misinterpreted.
Classically, stack sampling for particulate pollutants has been directed
towards measuring the mass of dust emitted which contributed to dust
fall (solid particles in the gas stream >lym). At present there also is
interest in the smaller sizes of particulate (0.1-l.Oym) because they
are believed to influence cloud formation in the atmosphere, and atmos-
pheric visibility and haze.
All particulate sampling systems include the following units:
nozzle
probe
collecting device
metering unit
suction source
These units are not always independent components of the system; for
example, a cyclone is sometimes used to collect particles and meter gas.
14
-------
The arrangement, both as to location and sequence, is a primary differentiat-
ing feature among systems. The design of the sampling system, hoth as to
configuration and size, may depend on the sampling procedure contemplated
and the interests of the designer. Most state-of-the-art sampling systems
do not collect condensibles, but only particles which are in essence gas-
borne solid matter larger than one micron in diameter at conditions
existing inside the stack. For a more complete discussion of existing
sampling trains refer to Appendix C. A general presentation on sampling
for particulates is included as Appendix J.
15
-------
3. EXPERIMENTAL PROGRAM
Our basic approach to this program was to collect samples in the field in
a manner that would permit them to be fully characterized in the labora-
tory. Thus, in our field work, we endeavored to ensure that the samples
were of adequate size and were representative with respect to variations
in chemical and physical properties.
3.1. COLLECTION OF FIELD SAMPLES
The sampling format and procedure used during this program was similar to
the one published in the Federal Register on December 23, 1971. (A copy
is included as Appendix K.) The major difference between our procedure
and the published one were related to the fact that we did not sample to
determine emission rates but rather only to obtain a qualitatively
representative sample of the stack emissions. Therefore, we did not
sample the stack on a statistical basis, nor did we correct gas volumes for
water vapor, nor did we, therefore, relate emissions to incinerator per-
formance.
The EPA sampling train is shown in Figure 1. Basically, it consists of
a nozzle, probe, cyclone, filter and impingers. The material collected
in these various components of the sampling train were separated with an
eye towards exploring whether different types of particulate matter was
preferentially collected by the different parts of the train. Thus, Che
particulate catch was separated according to probe and cyclone catch,
probe washings, filter catch, filter holder washings, impinger catch,
and acetone rinse of the impingers.
The criteria for evaluating the EPA sampling train with regards to its
ability to adequately collect the particulate -matter in the effluent stack
gas from municipal incinerators centered on whether:
a. The particulate collected by the EPA train* matched the material
defined as particulate by EPA—namely, "anything that is a
solid or liquid (other than uncombined water) at 70°F and one
atmosphere."
b. The filter acted as an efficient collector of the particulate
matter generally classified as inorganic mineral particulate;
c. The impingers collected condensibles;
*The EPA train currently recommended does not include a cyclone, whereas
all of this work was done with a train that included a cyclone.
17
-------
00
FIGURE 1
EPA ?articulate Sampling Train (early version)
-------
d. There were interactions between the sampling train and the sample
during its collection; and
e. Interactions between the gas phase and the collected particulate
generated false particulate or caused weight losses.
Most of our sampling was conducted at a local 250-tons per day (TPD) triple
traveling grate incinerator, which operated continuously from Monday to
Friday (samples coded (NI- ). The sampling point selected was in the
rectangular duct leading from the wet scrubber to the base of the stack.
The duct was 6'6" wide, 16'6" high. The roof of a small shed served as
the sample platform and facilitated installation of rails for stabiliza-
tion and operation of the sample collecting unit. The control module was
located inside the shed. In addition, a few runs were made at a local
batch incinerator. (These samples were coded BI- . )
3.2. TABULATION OF SAMPLING RUNS
Eighty-two sampling runs were made during the field program. A complete
listing of these runs is given in Appendix D. In summary, these can be
categorized as follows:
Table in
Type of Experiment Appendix D No. of Runs
EPA System Familiarization D-l 10
Filter Leakage Studies D-2 6
Filter Leakage Studies D-3 19
Evaluation of Sampling System D-4 8
Collection of Samples for Microscopy
Studies D-5 16
Collection of Samples for Quantitative
Analysis and Chemical Characterization D-6,7 23
3.3. OPERATION OF A COMMERCIAL EPA SAMPLING TRAIN
In order to help in our evaluation of the EPA sampling train, we made
several operational tests on a commercially available unit. (Note: the
train manufactured by Research Appliance Company was used for this program,
but it is likely that other commercially available units provide equivalent:
performance. This study was not intended to act as a certification ol: t.he
particular train, but rather to evaluate the general arrangement of ttie
component parts,)
19
-------
Several tests were conducted to evaluate the apparent inconsistent develop-
ment of high pressure drop across the sample train during period of sampl-
ing which exceeded one hour (see Table 6). For this evaluation, pressure
drop was measured as a function of time, volume throughput, and filter
loading for several operating procedures.
In the first set of tests, a filter was operated for approximately 3-1/2
hours, with three changes in impinger fluid. In Test 39 (65 minutes),
the average sampling rate was 0.93 cfm (0.026 cmm); the initial vacuum was
6.5" Hg and increased to 7.4" Hg. The impinger samples were removed and
the impinger recharged. Test 40 (65 minutes) was then conducted at an
average rate of 0.92 cfm (0.026 cmm) starting at a pressure of 7" Hg
which increased to 16" Hg. The impinger samples were again removed and
the impinger recharged. Test 41 (87 minutes) was then carried out in
which the vacuum increased to 22" Hg. At the latter condition, the flow
was only 0.23 cfm (0.007 cmm). The total sample volume through the filter
was 5.12 nH. Dust loading for this sample as shown in Table 6, was
195.2 mg/m3.
Two additional tests, 42 and 43, were run at a collection rate of 0.9
to 1 cfm (0.025 to 0.028 cmm) until the vacuum at the pump reached a
nominal 20" Hg. Run 42 had an initial vacuum of 7" Hg which is approx-
imately double that usually indicated at the sampling outset. This built
up to 14" Hg at the end of the first 65 minutes. On examination, it was
observed that silica gel had carried over into the suction line. When
the silica gel was removed from the pump suction line and the accumulated
water was removed from the two wet impingers, the vacuum dropped back to
8" Hg. The run was continued for another 31 minutes, by which time the
pump pressure had reached 21" Hg. Run 43 was carried out in a similar
manner with a change in impinger solutions after 7 minutes. After a
total elapsed time of 105 minutes, the pump pressure reached 20" Hg. In
each of the two tests, 42 and 43, approximately the same amount of incin-
erator effluent was sampled with the final vacuum in each case being
around 20" Hg. Filter loadings were almost identical for these two tests.
Three additional tests, numbers 44, 45, and 46, were made using the same
sampling rate as in tests 42 and 43. The results of these tests were
similar to those of 42 and 43.
From a review of these results and those from earlier tests, the follow-
ing conclusions were drawn:
• Careful observations of several tests conducted over the course
of a month show no anomalies in the function or behavior of the
RAG sampling train.
• Minor variations of 0.5 to 1.0" Hg which may occur in starting
vacuums probably are due to variations in the pressure drop
across the filter media and the packed silica gel bed.
20
-------
TABLE 6
RAG Sampling Train Evaluation Data
Date
7-7
7-15
7-15
7-22
7-22
7-22
Run No.
(NI- )
39,40,41
42
43
44
45
46
Gas
Volume
(m3)
5.12
2.35
2.55
2.47
1.49
2.42
Time
(min)
217
96
105
100
60
93
Pressure
6.5
7
3
3
4
3
Change ("Hg)
to
to
to
to
to
to
22
21
20
20
20
16
Filter
(nig)
989
563
551
271
563
Weight
(mg/m3)
195
240
216
182
198
-------
• The pump vacuum can be directly related to the mass of particulate
matter collected on the filter and the sampling rate.
• A sampling rate of 0.9 to 1.1 cfm (0.025 to 0.031 cfm) can gen-
erally be maintained until pump vacuum increases to 18" Hg; it
then becomes difficult to maintain this flow.
• Longer periods of sampling can be achieved at lower sampling
rates if coverage of a long-time burning cycle is desired.
• For the "normal" particulate loading of the effluent at the
incinerator tested, a sampling rate of 1 cfm (0.028 cmm) can
usually be maintained for only about 90 minutes in which time
filter loading is about 5 mg/cm2 and pump vacuum has increased
to about 20" Hg.
3.4. CHARACTERIZATION TECHNIQUES
A variety of analytical techniques is available to characterize the col-
lected particulates with respect to their chemical and physical properties.
Figure 2 illustrates how these options were considered for this program.
Although all techniques were tried during this program (with the exception
of fluorescence and liquid column chromatography) because of the low
levels of organics, the most useful information was developed via those
techniques which were most applicable to inorganic materials. Qualitative
information was obtained utilizing x-ray techniques, emission spectroscopy,
and wet chemical techniques. Figure 3 illustrates the sequence followed
for quantitative data.
22
-------
Optical and Scanning
Electron Microscopy
S3
Filter or
Impactor
Plates
IE
Transmission
Electron Microscopy
Electron Diffraction
SEM Energy
Dispersive X-ray
Analysis
X-ray Diffraction and
Fluorescence Analysis
Elemental
Composition
(Inorganic)
Identification of
Compounds or
Functionality
(Inorganic)
Infrared Emission
and fnterferometry
Removal from Filter
(i.e.. Extract
High Resolution
Mass Spectromeiry
Legend:
. Preferred Route
—————Possible Options
(when only small
samples are available)
1 Emission SpectroscopyM —
te
I Fusion »
^ ,
\
Dissolu
Partic
, * ,
| Acid j
tion of
ulate
| Alkali |
Extraction of
Soluble Species
Determination of
Individual Cations
and Anions
Separate into Groups
by Liquid Column
Chromatography
Aromatic
Hydrocarbons
Saturated
Hydrocarbons
Oxygenates
Atomic Absorption
Wet Chemistry
Spectrophotometry
UV-Vis Fluorescence
Spectroscopy
Ultraviolet
Spectrophotometry
Gas Chromatography
High Resolution
Mass Spectrometry
FIGURE 2
Flow Chart for Examination of Collected Particulate
-------
FIGURE 3
Analysis of Solid Catch
Dried and Weighed Sample
I
Extract three times with 50 ml boiling water.
Combine extracts, evaporate, and dilute to 100 ml.
Filter and/cr Solids
Dry at 105°C and Reweigh
i
FUSION
Fuse with Sodium Carbonate
Dissolve in HC1-H20
Extract
Determine the following:
• Halides (titration)
• Acidity (titraCion)
• Sulfate (precipitation)
• Nitrate (spectrophotometry)
• Trace Metals (K,Na,Pb,Zn,Ti,Al,Fe,Sn)
(atomic adsorption spectrometry)
Filter
SOLIDS
Insoluble Portion
Dry at 105 °C~weigh
Ignite at 850°C—weigh
Dissolve in NaOH and determine
silica (spectrophotometric)
FILTRATE
Filtrate
Dilute to 100 ml and determine:
• Sulfntc (precipitation)
• :<2°3 (precipitaticn
• Me in Is (K,Ca,Mf>,Ai ,TL ,l"e,Pb,Zn,Sn) (afmic absorption spectrometry)
-------
4. PROGRAM RESULTS
Detailed discussions of the experimental findings obtained during this
program are presented in Appendices E, F, G and H. A condensed overview
of these findings is given in the following discussion.
4.1. CHARACTER OF THE PARTICULATE CATCH FOUND IN THE "FRONT HALF" OF
THE SAMPLING TRAIN
Table 7 shows the distribution of the particulate catch in the EPA sampling
train. As can be seen from this data and that in Table 8, the total
material caught in the probe/cyclone/filter was remarkably consistent,
both between incinerators and seasons. Only a small portion of the total
particulate catch was found in the impingers. (With but one exception,
20Z or less was caught in the impingers.) Because of program limitations,
we did not make any effort to explain the variations between incinerators
(NT vs. BI) or amount collected in the probe/cyclone. The fact that the
probe/cyclone collected almost twice as much material in October as in
the spring may be related to the composition of the refuse; that is, it
led to more particles greater than five microns, which were then collected
in the cyclone rather than the filter. Because the EPA train no longer
utilizes a cyclone, this type of difference does not appear important.
Small quantities of organic material could be found by extraction, but
total organic matter on the filter amounted to less than 5% of the filter
catch. (When coupled with the organic matter in the impingers, total
organics were estimated to be always less than 10% of the total catch.)
Overall summaries for the composition of the probe/cyclone and filter
catches are given in Table 9. The principal cation components (>5% of
total cations) of the mineral particulate common to both the probe/cyclone
and filter catch were: zinc, lead, potassium, and sodium. In addition,
iron, titanium, and aluminum were present in both but were relatively
much higher in the probe/cyclone catch. Interestingly, even though calcium
was a major component in the probe/cyclone catch, neither it nor magnesium
was found in the filter catch. The reasons for this difference are not
known at this time.
No evidence was found for the presence of acid in these materials. It
is possible that any free oxide present in the sample reacts with the HC1
or 803 to form a salt. To the extent that this happens, filter weight
would be artificially raised, but although not strictly, definitive, the
data from this program suggests that filter weight gain via this mechanism!
is not a serious problem. Details on the qualitative and quantitative com-
position of the probe/cyclone and filter catches are provided in Appendices
E and F. Studies made directly on the filter catch using scanning
electron microscopy are described in Appendix H.
25
Arthur I) i.iuiehu
-------
TABLE 7
Distribution of Incinerator
Particulate Emissions Using EPA Sampling Train
Collection Device
Percent Found
With Two Types of Incinerators
Continuous Grate Batch
Ranee
Ave.
Range
Ave.
Nozzle, Probe, and Cyclone 19-33 23 10-13 11
Filter 49-75 62 70-82 76
Combined Total 74-95 85 83-93 87
Impingers 15 13
26
-------
TABLE 8
Proportion of Particulate Caught in the Probe Plus Cyclone and
Probe Plus Cyclone Plus Filter as Function of Season and Incinerator Type
Percent of Catch
Sample Code
Date Sampled
Probe/Cyclone
Ave.
Filter
Ave.
Probe/Cyclone/Filter
Ave.
NI- 1
2
3
4
5
9
10
13.13A
22,23,24
25,26,27
77
78
79
80
BI-73
74
75
76
4/8/71
4/14
4/15
4/15
4/20
4/22
4/22
5/4
5/18
5/19
10/14
10/14
10/14
10/14
9/16
9/15
9/16
9/16
20
16
15
18
26
21
17
24
18
16
31
31
41
28
13
11
11
10
19
19
33
11
61
67
65
71
75
65
70
59
55
64
52
58
49
62
70
82
75
79
67
60
55
76
81
83
80
89
95
86
87
83
74
80
83
89
90
90
83
93
86
89
86
79
88
87
*NI represents continuous grate incinerator.
Bl represents batch incinerator.
-------
TABLE 9
Composition of Filter and Probe/Cyclone Catch
Relative Amounts (%)*
Component Probe/Cyclone Filter
Hot Water Solubles 24-60 75
Cations 40-55 40
Anions 24-50 60
Sulfate 10-40 25
Chloride 5-15 25
OtherA (oxides, ^IQ 10
silicates, etc.)
Total ,,. ...
25-50 60
+From a continuous grate incinerator.
*See Table 7 for distribution of total particulate catch between
ProLe/cycloue.
AEstimated.
'28
-------
4.2. CHARACTER OF THE IMPINGER CATCH
As shown in Table 7, the impinger catch amounted to less than 15% of the
total particulate catch. Moreover, very little mineral particulate or
organic matter was found to be present in the organic extract of the
impinger catch. The composition of the impinger catch by the original
EPA "dry down" procedure is shoxm in Table 10 where "dry down" refers to
equilibration with Drierite. In comparison, the composition of the
impinger catch on an "as received" basis is shown in Table 11. (Water,
which is actually most of the sample, is not included in this breakdown.)
It can be seen that a large portion of the particulate catch is mineral
acid (90-95%). (See Appendices E and F for complete details.)
A more detailed breakdown of the impinger catch is given in Table 12.
Recognizing that the impinger catch is only 15% of the total particulate,
this means that the non-mineral acid content of the impinger catch is less
than 2% of the total material caught in the sampling train—a relatively
insignificant sum. Thus, the impinger should not be viewed as a device
for collecting solid mineral particulate in this application.
Because the impinger contains only sulfuric and hydrochloric acids, the
only chance for false particulate in the impinger comes from the possible
oxidation of S02 to SO^ which is then trapped as sulfate. At most, this
represents less than 10% of the total particulate catch.
4.3. DISTRIBUTION OF SULFATE AND CHLORIDE BETWEEN THE PROBE/CYCLONE
FILTER AMD 1MPINGERS
The distribution of sulfate and chloride among the various components of
the sampling train was found to be different (see Table 13). This dif-
ference may be the result of several factors, for example:
• Higher volatility for HC1 over S03;
• Higher stability of sulfate salts;
• Two different SOX species present whereas chloride is present
only in one oxidation state.
A great deal of work was conducted both in the laboratory and in the field,
in an effort to develop an understanding of whether the sulfuric acid in
the impingers comes from collection of S03 originally present in the stack
gas or from oxidation of S02 by the train prior to the impingers with
subsequent collection of 803 in the latter (see Appendix I). Based on
calculations of its dew point at 250°F, only 2 ppm 803 could be present in
the vapor phase. This is equivalent to roughly 13 rag of sulfate. Higher
filter box temperature will raise this value but this does not appear to
be a sufficient mechanism to explain all of the sulfate in the impingers.
29
-------
TABLE 10
Composition of Impinger Catch
After Evaporation to "Dryness" via EPA Procedures
Amount Found (%)
Continuous
Component Grate Batch
Sulfuric Acid 46 32
Hydrochloric Acid 2-3 2-3
Ash 2-3 2-3
*Unaccounted components assumed to be mainly water with small
amount of organics.
30
-------
TABLE 11
Composition of Impinger Catch—As Received Basis
Amount Found (%)
Continuous
Component Grate Batch
Sulfuric Acid 10 8-9
Hydrochloric Acid 86 82
Ash <1 <1
Organic* 3 8-9
*Not including water, which is bulk of sample.
+Estimated.
31
-------
TABLE 12
Analysis of Impinger Catch
for Continuous Grate Incinerators
Component UnitA :
Cations meq.
H+(acidity)
Other*
Total (max.)
Anions meq.
Sulfate
Chloride
Total
Organic Extractible mg
Ash @ 850°C mg
*0ther cations were present as follows:
Calcium
Magnesium
Potassium
Sodium
Lead
Zinc
Total
Sample No.
37-3 41-3/43-3
15.0
<0.1
15.1
0.9
14.8
15.7
<2
<2
12.3
<0.1
12.4
1.2
12.9
14.1
<2
<2
45-3/36-3
12.0
<0.1
12.1
1.0
12.0
13.0
<2
<2
(milligrams)
0.2
0.04
0.1
0.3
0.06
0.05
0.8
0.5
0.07
0.1
0.4
0.1
0.05
1.2
0.2
0.03
0.1
0.2
0.2
0.05
0.8
Ameq = milliequivalents; mg = milligrams,
32
-------
Sample
(NI- )
77
78
79
80
Average
TABLE 13
Distribution of Chloride and Sulfate in Sampling Train
Distribution of Sulfate (%)
Probe/Cyclone
28
16
28
28
Filter
40
42
34
26
Imp ing er
32
42
38
46
Total Weight
Found (mg)
70
103
93
61
Distribution
Probe/Cyclone
9
(8)
11
(12)
Filter
32
—
29
—
of Chloride
Impinger
59
(208)
60
(140)
*
(%)
Total Weight
Found (mg)
80
220
L05
150
25
35
45
82
10
30
60
*Values in parenthesis are weights in milligrams on a 1.0 m^ sample basis.
-------
4.4. PROCEDURES FOR MEASUREMENT OF THE PARTICULATE CATCH
The EPA procedure calls for drying of the filter over Drierite before
weighing to determine weight gain and thus filter catch. In reality, this
approach is a mutual equilibration of water vapor between the filter catch
and calcium sulfate (Drierite) . If the filter contains materials whose
hydrates are more stable than calcium sulfate, the residual water in the
Drierite will transfer to the filter. If the reverse is true, the filter
catch will give up its water and become "dry."
To demonstrate this problem, we took three samples of impinger water
residue representing different total weights and dried them first in a
desiccator containing Drierite (calcium sulfate) , and then one containing
saturated calcium chloride solution. The results were as follows:
Weight (rag) in
Desiccator Containing*
Sample Code Drierite CaCl^H^ % Gain
NI-23-3 140(8.7) 193(6.4) 38
NI-24-3 37(3.3) 43(0.7) 19
NI-24A-3 17.5(1.0) 20(0.3) 14
*Values in parenthesis are standard deviation (la) .
Although not dramatically different at the lower levels of particulate,
the data does suggest that depending on the drying conditions, a varying
amount of water will be called particulate catch. This factor needs
more study.
4.5. RELATION TO PRIOR INFORMATION
Although the information developed under this program does not specifically
relate to emission levels for each pollutant, we can conclude that our
findings are supportive of what has been found in the literature (see
Appendix A) .
-------
APPENDIX A. COMPOSITION OF INCINERATOR EFFLUENTS
Table of Contents
Page
I. INTRODUCTION A-3
II. SUMMARY OF EMISSIONS A-4
A. Particulates A-4
1. Solid Particulate A-6
2. Condensible Particulates A-ll
B. Gaseous Emissions A-12
1. Carbon Oxides A-12
2. Sulfur Oxides A-13
3. Nitrogen Oxides A-13
4. Hydrocarbons A-14
5. Specific Organics A-14
6. Inorganic Acids A-14
7. Volatile Metals A-14
C. Summary A-14
III. ANNOTATED BIBLIOGRAPHY OF INCINERATOR EMISSIONS A-15
A. Particulates A-17
1. Bibliography A-17
2. Discussion A-36
B. Carbon Oxides (^,02) A~37
1. Bibliography A-37
2. Discussion A-44
C. Sulfur Oxides A-44
1. Bibliography A-44
2. Discussion A-47
D. Nitrogen Oxides A-47
1. Bibliography A-47
2. Discussion A-53
A-l
-------
Table of Contents (Continued)
Page
E. Hydrocarbons A-53
1. Bibliography A-53
2. Discussion A-60
F. Specific Organics A-60
1. Bibliography A-60
2. Discussion A-65
G. Inorganic Acids A-65
1. Bibliography A-65
2. Discussion A-65
H. Volatile Metals A-66
1. Bibliography A-66
2. Discussion A-66
I. List of References Cited A-67
A-2
-------
APPENDIX A
COMPOSITION OF INCINERATOR EFFLUENTS
I. INTRODUCTION
The chemical composition of incinerator stack emissions is principally
dependent upon incinerator design and operating conditions, the parti-
cular anti-pollution control devices that are employed and refuse
composition. The emissions exhibit continual variations in chemistry as
a result of moment-to-moment variations in refuse composition and combus-
tion efficiency; furthermore, seasonal variations occur that are mainly
due to the influence of yard wastes on average refuse composition.
In addition to variations due to the "incineration" factors given above,
chemical measurements are greatly affected by the particular sampling and
analysis methods that are employed. As a result, the values for incinerator
emissions that are reported in the literature cover a wide range.
This appendix will present a summary of what is known of the composition
of incinerator effluents (Section B) as determined by a comprehensive
search of the literature. Section C presents annotated biographies of
the pertinant literature for each pollutant. Much of the summary data
presented herein has appeared in a three volume report to the Environmental
Protection Agency (EPA) lender contract CPA 22-69-23 by Arthur D. Little, Inc.,
entitled, "Systems Study of Air Pollution from Municipal Incineration."
Sections of that report which are particularly pertinent to emission
chemistries are, "The Nature and Causes of Incinerator Air Pollution,"
Volume I, Sections B through I, and "Incinerator Emissions Data,"
Volume II, Appendix J. In addition to the above, reference sources have
included the open literature, an APTIC (Air Pollution Technical
Information Center) literature search based upon the title, "Incineration -
Measurement Methods and Combustion Products (Particulates and Particulate
Sampling)," and reports based on various studies sponsored by the Office of
Air Programs and the Office of Solid Waste Management Programs, EPA.
The following introductory comments serve to preface the discussion of
effluent compositions.
• The pertinent literature is mostly from the last decade.
Of the 40 references cited herein, 23 are from the period
1966-1970, 12 are from 1960-1965 and 5 are from 1951-1959.
• Many of the reported emissions data are presented with very
little discussion of incinerator design, operating parameters,
or sampling and analysis procedure. It is difficult to
determine the reliability of data in these cases.
• Emissions data for some municipal incinerators have been
reported as long as 15 to 20 years ago, with n,o follow-up
studies. These data may be questionable due to the develop-
ment of more reliable sampling and analysis methodology.
A-3
-------
• Very few studies have been performed in a systematic way,
that is, with control of refuse composition and incinerator
operating parameters. Most such studies have been carried
out on low capacity experimental incinerators for obvious
reasons.
• Calculated emissions from estimated refuse compositions and
combustion parameters are of questionable value due to the
large number of variable parameters involved.
• Comparisons of reported data are often difficult or impossible
due to the lack of convention in units of conversion. Various
units used in the literature for reporting emissions include:
Ibs/ton of refuse
lb/1000 Ib flue gas corrected to 50% excess air
lb/1000 Ib flue gas corrected to 12% CO,
grains/standard cubic foot (60°F, 1 atm) at 50% excess air
grains/standard cubic foot (60°F, 1 atm) at 12% C02
grains/standard cubic foot (60°F, 1 atm) at stack conditions
grains/cubic meter at NTP (32op, 1 atm) and 7% C02
The conversion of units depends on refuse composition and
incinerator operating parameters, i.e., stack gas flow rate,
percent excess air, refuse charging rate, etc. Direct con-
version of units is therefore precluded. A set of conversion
factors based on representative combustion conditions and
typical refuse has been presented by Niessen (Ref. A-l);
these factors are given in Table A-l. Summary data reported in
this Appendix are presented in terms of pounds per ton of refuse.
The following sections present specific discussions of particulate and
gaseous emissions. The solids, which are generally collected in a cyclone
or on a filter, are composed of incombustible (mineral) and combustible
fractions. The impingers, which collect the condensates, also include
fine organic and inorganic particulates which are not previously trapped
by a filter. The impinger may also include chemical species which are
the product of gas reactions with the impinger solutions.
The gaseous emissions include oxidation products of combustion as well
as reaction products of elements released from the refuse, such as
hydrogen chloride.
II. SUMMARY OF EMISSIONS
A. Particulates^
Particulate has been defined by EPA as "all solids and condensible
materials other than uncombined water which are liquid at standard condi-
tions of one atmosphere and 70°F. Particulate matter has almost always been
A-4
-------
Table A-l
Conversion Factors for Particulate Incinerator Emissions
Lbs/Ton Refuse
(As Received)
Lbs/1000 Ibs Flue Gas at
50% Excess Air
Lbs/1000 Ibs Flue Gas at
12% C02
Grains/SCF at 50% Excess Air
Grains/SCF at 12% C02
Grams/Nm3 at NTP, 7% C02
Lbs/Ton
Refuse
(As Received)
11.27
10.0
Lbs/1000 Lbs
Flue Gas at
50% Excess Air
0.089
0.89
Lbs/1000 Lbs
Flue Gas At
12% CO?
O.iO
1.12
Grains/SCF
at 50%
Excess Air
0.047
0.52
0.46
Grains/SCF
at 12%
CO?
0.053
0.585
0.52
Grams/Nm-'
At NTP
7% C02
0.067
0.74
0.66
21.31
18.85
15.0
1.93
1.61
1.36
2.16
1.82
1.53
1
0.89
0.704
1.12
1
0.79
1.42
1.26
1
:r
o
-------
equated to what is collected as particulate at stack conditions rather than
what is present in the effluent from the stack. Until recently, it has been
taken that, in the EPA train solid particulate was collected by a heated
cyclone and/or filter; the materials found in the impingers (and therefore
frequently referred to as the impinger catch) is generally referred to as the
"condensibles." These may be fine inorganic particulates carried through the
filter or organic matter that is trapped in the impinger. In addition, some
chemical interactions may occur in the impinger between gaseous species which
normally would not condense and the impinger solution.
1. Solid Particulate
The incombustible fraction of the solid particulate that is entrained in
the stack gases by the combustion air or combustion products is composed
mainly of mineral matter that was present in the waste. Volitalization
and condensation of metals also contributes to the mineral particulates
but, for mixed wastes, this usually contributes only a small fraction of
the total mineral particulate loading. The combustible fraction consists
of char (blackbirds) entrained from the grate, "black smoke" (soot)
produced by the thermal cracking of pyrolysis products before these have
been mixed with sufficient oxygen to complete their combustion, and
"white smoke" produced by the condensation of pyrolysis gases.
• Mineral Particulate
An appreciation for the factors that determine particulate emission and
for their relative importance may be gained by considering the two
mechanisms mainly responsible for the emissions:
• The mechanical entrainment of particles from the burning
refuse bed;
• The volatilization of metallic salts or oxides.
The first of these mechanisms is favored by refuse with a high percentage
of small particle size, low density ash; by residue or residue geometry
favoring entrainment (plates); by high underfire air velocities; and by
other factors that induce a high gas velocity through the bed. The
second mechanism is favored by a high concentration of metals that form low
melting point oxides and by high temperatures within the bed.
• Mechanical Entrainment
Ash particles may be entrained when the velocity of the gases through
the fuel bed exceeds the terminal velocity of the particles. Under-
grate air velocities typically vary from a minimum of lOSCFM/sq. ft.
of grate area to 100 SCFM/sq. ft. Based on the terminal velocity of
ash particles it is therefore expected that particles up to 70 ym
(equivalent diameter) will be entrained at the lowest velocities and
up to 400 pi at the highest. These data are shown in Figure A-l.
-------
10 F
10"
V
o
c
e
10''
10
Air Velocity
- at 2000°F
for Full-Scale
Incinerator
1111 i i i i 11111 i i i
\
1000
100 10
Particle Diameter (Microns)
FIGURE A-l
Particle Fluidization Velocities (Terminal Velocitiesl
A-7
-------
Overall findings show that the ash content of the refuse is a major
factor in determining emission rates, and that the percentage of the
ash carried over ranges mostly between 10 and 20% of the total,
corresponding to 15 pounds/ton of refuse on the average. The actual
percentage carried over for a particular incinerator will, of course,
depend on other factors such as underfire air rate. On the average,
a two-fold reduction in air velocity will produce an approximately
30 percent reduction in emission rate.
Entrainment of mineral particulate, however, cannot be eliminated
completely by the elimination of underfire air since the buoyancy of the
combustion gases will induce an air flow that will produce a minimum
particulate emission. From dimensional analysis, it can be shown that
for a buoyancy-driven flow the velocity V of th<; induced air is
proportional to VLAp/pg where L is a characteristic length of the
combustion chamber, Ap is the difference in density between the colder
entrained gases and the hot combustion products, P is the mean density
of the combustion products, and g is the gravitational acceleration.
Entrainment of particulate by buoyancy-driven flows will, therefore,
increase with furnace size and flame temperature. An additional factor
that enhances particulate emission is agitation (stoking) of the burning
waste but the magnitude of this factor has not been quantitatively
established.
Emissions to the atmosphere depend greatly upon the particular anti-
pollution control equipment that is employed. Mineral particulates are
mostly composed of Si02> CaO, A120_ and Fe2°3' Th686 a*& often present
as complex compounds, which also may contain appreciable amounts of
sodium and potassium- Variation in flyash chemistries, particularly
for iron and aluminum compounds, are largely due to differences in the
refuse composition.
Particle size distribution measurements of stack emitted flyash show
that over 90 percent by weight of the particles are less than 250 um
and about 10 percent are sub-micron. Recently reported studies of
particle size distribution give the following values for weight percent
smaller than indicated size.
Size
L 10 11 IP. 30 45_ 150
Ref. A-2 15 25 — 40 — 60
Ref. A-3 37 64 75 81 87
Ref. A-4 6 20 — 47 69 89
Ref. A-5 20 42 52 58 70 77 95
Ref. A-6 20 29 44 — 57 62
Ref. A-7 12 18 39 42 44 60 95
Average 18 33 52 53 65 70 95
A-B
-------
These data are compared to particle size distributions measured for
three typical U. S. incinerators, European practice and both coal- and
oil-fired boilers in Figure A-2. It is of interest to mention that
emissions of 3 to 5 Ibs particulate/ton of refuse correspond to about
10 individual particles, half of which are less than 0.05 urn and 95%
of which are less than 0.1 ym in size.
The flyash particulate can have an important influence on other chemical
emissions. Since a great many particles are in the sub-micron particle
size range, they could serve as condensation nuclei for many vapor
species. In addition, the relatively high alkaline earth contents of
flyash and the high moisture content in flue gases allows a certain level
of SO and S0« absorption, resulting in the formation of alkaline earth
sulfates. Finally, the large surface area of the particulates is
favorable for adsorbing certain gas species. These factors tend to
influence the total weight of particulate and should be taken into
account when performing a materials mass balance.
• Volatilization of Metallic Salts
The temperature within a burning refuse bed can reach 2500-3000°F. At
these temperatures, a number of metal salts (particularly those of the
alkali metals) will vaporize or sublime. Therefore, it would appear
that although volatilization of salts may be occurring in localized
hot spots within the burning bed, the low concentration of these salts
in refuse, the low frequency and/or intensity of the hot spots, con-
densation of the vapors within cooler regions of the bed, or other
considerations prevent or attenuate emission. The contribution of
metallic salt volatilization to the total furnace particulate emission
rate thus appears to be small in municipal-scale incinerators.
• Combustible Particulate
Combustible particulate consists of char (blackbirds) entrained from the
grate, soot (black smoke) produced by the thermal cracking of pyrolysis
products before these have been mixed with sufficient oxygen to complete
their combustion, and an aerosol (white smoke) produced by the quenching
of pyrolysis products before they have reacted. All three are highly
visible and have a nuisance value in excess of their contribution to the
total particulate emission rate.
Char formed from paper or sheets of other carbonizable material has a
high surface-to-volume ratio, and is, therefore, entrained at relatively
low velocities, in amounts that depend on the make-up and degree of
agitation of the refuse bed. Inert material in the bed has a beneficial
effect of keeping elements of char from being entrained. No quantitative
estimate of the emission rate of charred waste components is available,
but it is believed to be a small fraction of the total particulate
emission.
A-9
-------
0.1
1.0 10.0
Particle Diameter. Microns
100.0
Sources:
FIGURE A-2
Particle Size Distributions of Furnace Effluents
European: Andritzky, M., Brennstoff-Warme-Kraft 19(9), 436(1967); U.S.A: Walker, A.B., and
Schmitz, F.W., Proc. 1966 ASME Nat'l. Incin. Conf., pp 64-73.
-------
The second source of combustible material, soot, is usually the most
significant one. Soot particles, once formed, are relatively difficult
to burn, and as a consequence of their characteristically small dimensions,
are difficult to collect. The best form of control of black smoke is,
therefore, the prevention of soot formation. This can be achieved by the
rapid mixing of the pyrolysis gases with sufficient air to complete their
combustion since it has been shown that with perfect mixing little soot
is formed even under fuel-rich conditions. For example, Wright (A-8)
has found that for a wide range of hydrocarbons burned in a well-stirred
reactor, the atomic ratio of oxygen to carbon at which soot was first
detected ranged from 1.35 to 1.81, well below the stoichiometric ratio
of 2.00. The practical significance of these results is that soot
formation may be prevented by the proper design and placement of over-
fire air or steam jets.
The mechanism of soot formation is so complex, however, that the amounts
of combustible particulate that leave a furnace chamber cannot be
reliably estimated. The amounts of soot formed are a measure of the poor
mixing conditions in a chamber and are expected to be greatest when a
furnace is overcharged or when a waste having a high fraction of
volatile carbon (total minus fixed carbon) is burned. Under normal
operating conditions, the average combustible fraction of the particulates
leaving combustion chambers of municipal incinerators is found to be
about 15 percent (A-l); but because soot forms as sub-micron particles,
it is difficult to collect and the fraction of combustible in the
particulates leaving air pollution control devices can be considerably
higher. Contemporary, continuous-feed municipal incinerators, for example,
emit an average of 3.9 Ibs combustible/ton refuse, a figure which
corresponds to about 20 percent of the total particulate.
The third type of combustible particulate emission is the aerosol of
unreacted hydrocarbons which is formed when the pyrolysis products
evolved on the grate are cooled and condense before mixing with hot
furnace gases. The white smokes that are produced usually consist of fine
particles (<2um) which are difficult to collect. Such emission, however,
is relatively infrequent and can be eliminated by either afterburning or
the adjustment of the air flow to a furnace to provide better overbed
mixing.
2. Condensible Particulate (Impinger Catch)
The impinger catch includes vapor condensates and fine sub-micron ash
particles which have passed the primary particle collector (cyclone or
filter). Collected material that would have remained gaseous had it
been directly admitted to the atmosphere but instead has condensed in
the sampling apparatus clearly should not be included as part of the
condensible particulate catch.
Studies of condensible particulates have been quite limited. The
particulate caught after the filter corresponds to about 30 percent of
the total. Roughly two-thirds of this appears as inorganic (by definition
A-ll
-------
of the extraction treatment) and is thought to represent flyash "fines"
which have penetrated the filter. The other third, or about 10 percent
of the total particulate, is organic in nature. There has been an
insufficient amount of work reported in the literature to indicate the
mechanisms by which these materials are captured by the impingers.
Further it is not clear whether these materials enter the impingers as
solids or condensible vapors or if any portion is formed by combustion
gas interaction with the impinger liquids.
B. Gaseous Emissions
The major flue gas components that are routinely reported in measurements
of incinerator emissions include oxygen, nitrogen, water and CCL. Other
gas species, present in relatively small quantities, constitute the
major gaseous pollutants. These gases, which are formed directly by the
combustion process or indirectly by chemical interactions within the gas
stream, include carbon monoxide, sulfur oxides, nitrogen oxides, ammonia,
inorganic acids (HC1, HF, etc.) and various organics including acids,
aldehydes, esters, hydrocarbons, polynuclear hydrocarbons and others
(alcohol, ketones, mercaptains, organometallics, nitriles, amides,
sulfones, etc.) The organic species are identified and quantitated by
combined methods of infrared spectrometry, gas chromatography and mass
spectrometry.
The major gaseous species within incinerator emissions have composition
ranges of 10-13% CL, 3-5% CO-, and 10-30% H-0, or on a dry basis,
13-15% 0- and 4-8% CCL with the balance being N . Continuous monitoring
of these gases, either by Orsat analysis or by continuous gas sampling,
provides the necessary data to convert emissions to a common basis, i.e.,
50% excess air (STP) or 12% CO (STP).
1. Carbon Oxides
Carbon monoxide is the most significant pollutant (on a weight basis)
emitted from municipal incinerators. Although it is reported that the
level of CO can be reduced by increasing the amount of excess air, the
resulting increased stream velocities of the exit gases cause entrain-
ment of an additional fraction of the ash residue, leading to an
increase in flyash emissions. The reported levels of CO emissions vary
considerably, ranging from 0.3 to over 200 pounds/ton of refuse burned.
Carbon monoxide is generally measured by Orsat analysis. This method
has a detectibility limit of 0.1 volume percent which, for typical
incineration conditions of 200% excess air and stack gas temperature of
500 F, corresponds to 9 pounds CO/ton of refuse. The very poor
sensitivity of Orsat analysis leads to a wide range of reported emission
levels.
An additional difficulty is encountered in converting measured CO
volumes to Ibs/ton of refuse. Rigorous conversion requires knowledge of
the stack gas volume per unit time and rate of refuse burning, parameters
A-12
-------
which are seldom given. Approximate conversions can be made which are
based upon the weight of dry air required to burn a pound of typical
refuse or upon a carbon balance. The latter is computed on the basis
of 500 Ibs carbon/ton of refuse, yielding:
°f
' v/o C0 v/o C02
This computation yields a maximum level of CO, for it assumes that all
carbon in the refuse is converted to the oxide.
Stack gas analyses from 16 incinerators representing 122 individual
measurements average 23.8 lbs CO/ton of refuse with a standard deviation
of 34.2 Ibs/ton. A more comprehensive study of 33 units and 302 individual
data points carried out by Niessen (A-l) yields an average of 33.2 lbs
CO/ton of refuse with a standard deviation of 34. 1 Ibs/ton. The
estimated average emission factor for U. S. incinerators was 34.8 Ibs/ton.
The values given above represent maximum CO levels, as they were
calculated according to a carbon balance. Kaiser (A-9) suggests that
25% of the carbon (i.e., 125 Ibs/ton of refuse) is carried over to the
ash residue as unburned char. For this case, the reported CO average
emission would be 17.9 Ibs/ton of refuse or, according to Niessen
26.1 Ibs/ton of refuse.
2. Sulfur Oxides
The level of S0_ emissions is reported to be directly related to the
sulfur content of refuse, which is given in several reports as 0.1% by
weight. This concentration would provide a maximum S0_ level of
4 pounds/ton of refuse. Various reports of the S0_ content of incinerator
emissions range from 0.7 to 3.9 pounds/ton of refuse, with an average
of 2.2 pounds/ton of refuse. The remaining sulfur is believed to be
tied up as sulfate in the ash residue and flyash and as condensed E^SO .
Niessen (A-l) takes exception to the above and concluded that 95% of the
gaseous sulfur' is emitted as S02; he concludes that reports of high levels
of 803 are the result of sampling and analysis errors.
3. Nitrogen Oxides
The content of nitrogen oxides is small since the temperature required
for their copious production (> 2000°P) are generally not reached during
incineration. Typical levels are 90 ppm (3.1 Ibs/ton of refuse) at
temperatures of 2000°F. The reported range of nitrogen oxide emissions
is 0.3 to 5.7 Ibs/ton of refuse, with an average of 2.1 Ibs/ton of
refuse. In contrast to the above, a study of seasonal variation in
incinerator emissions in the New York City area (A-10) yielded average
values of about 0.05 Ibs/ton of refuse for both NO and NO .
A-l 3
-------
4. Hydrocarbons
Hydrocarbons represent a relatively small fraction of the total emitted
pollutants. The combustion of hydrocarbons is a two-step process: rapid
oxidation to CO followed by slow oxidation to C02. As a result, hydro-
carbon emissions can often be correlated with CO emissions. Literature
data on hydrocarbon emissions range from 0.3 to 2,7 Ibs/ton of refuse.
5. Specific Organics
Very limited data are available for the levels of specific organic
classes present in municipal incinerator effluents. Concentrations
reported are generally in the range of 1 ppm or less, corresponding to
approximately 0.01 Ibs/ton of_5refuse. Polynuclear hydrocarbons are
reported at levels of 5 x 10 to 10~ Ibs/ton or refuse. Numerous
organic compounds, all present in very small quantities, have been
identified as being present in incinerator emissions.
6. Inorganic Acids
HC1 emissions are of major concern. Although HC1 emissions in the past
were generally low (less than 1 Ib/ton of refuse), the increased usage
of chlorinated plastics for disposable packaging materials suggests a
greatly increased growth in HC1 emissions in the future. Anticipated
growth patterns developed by Niessen (A-l) indicate expected average
emissions (in pounds per ton of refuse) of 1.5 in 1970, 2.2 in 1975,
2.7 in 1980, 4.4 in 1990 and 5.4 in 2000.
HF is also probably present in incinerator emissions, most of which
results from combustion of fluorinated hydrocarbons such as "Teflon."
Limited studies indicate HF levels of 0.002 to 0.2 Ibs/ton of refuse.
It is anticipated that properly designed wet scrubber APC systems may
remove up to 99% of the HC1 and HF, thus reducing the pollution hazard.
7. Volatile Metals
Some metals, particularly Zn, Cd and Pb are volatile or have oxides that
are volatile at incineration temperatures. In general, very small
particles are formed. Little quantitative data are available for metal
emissions. Based upon the lead content of refuse, emissions are about
0.03 Ibs of lead/ton of refuse. Selenium emissions are of particular
interest due to its toxicity. The one published measurement of Se
reports 0.002 Ibs/ton of refuse.
C. Summary
In summary, the studies of municipal incinerator emissions are limited
in number as well as in scope. The reported results vary widely
depending upon refuse composition, incinerator type and operating
conditions, employment of various types of APC devices, and sampling
and analysis procedures. Although there have been about half a dozen
A-14
-------
comprehensive studies of incinerator emissions over the past fifteen
years, there is no specific agreement in the collected data. A
description of the composition of incinerator effluents is therefore
rather vague. Based upon the results of the literature survey, the
following estimates are made for incinerator emissions. The results
obtained fay Niessen (A-l) in a recent study are presented for comparison.
Component
Pounds/Ton of Refuse
Solid Particulate
CO
SO (as SCL)
x 2
NO (as NO )
X £-
Hydrocarbons
Polynuclear Hydrocarbons
HC1
HF
Pb
1.9-23.0
0-202
0.7-3.9
0.05-5.7
0.3-2.7
5xlO~5
Average
7.2
23.8
2.2
2.1
1.0
5x10
-5
Ref.A-l
13.6
34.8
3.9
2.6
2.7
3.2x10
0.8
-3
2x10
3x10
-2
-2
III. ANNOTATED BIBLIOGRAPHY OF INCINERATOR EMISSIONS
Annotations to literature references describing the chemistry of
incinerator emissions are presented chronologically for the following
categories: particulates, carbon oxides, sulfur oxides, nitrogen
oxides, hydrocarbons, specific organics (including acids, aldehydes and
polynuclear hydrocarbons), inorganic acids and volatile metals. These
annotations are sufficiently complete so as to provide a preliminary
information source to the literature for any of the above chemical
categories. The reader is directed to the cited references for more
specific discussions. A reference listing is presented at the end of
this section.
A summary sheet is presented in Table A-l which specifies the
availability of literature data for each of the chemical specie categories
in terms of effluent chemistry, sampling and analysis procedures,
incinerator operating practice, chemical interactions and so forth. The
references cited in this table are ranked according to the: value of the
data reported.
A-15
-------
TABLE A-2
Summary of Abstracted Incinerator Emissions Data
Effluent Chemistry
a. Caseous
b. Particulate
c. Fly Ash
d. Particle Size
Distribution
f. Influence from
APC device
Sampling Procedure
Analysis Procedure
Influence by Incin-
erator Operating
Parameters
Chemical Interations
in Effluent.
Chemical Correlation
General
References*
Carbon Sulfur Nitrogen Specific Inorganic Volatile
Particulate Oxides Oxides Oxides Hydrocarbons Organics Acids Metals
9, 11, 16, (17), 14,34,38, 10,20,22,25, 22,23,24,25,29, 17,18,20,22,30, 34 27
(29) (31), (36) 37, (30) (2) (2)
4,15,16,22,23,
24, 27, (2), (13),
(20), (26)
4,11,18,5,6,
7,9,14,27
2,3,4,5,6,7,15
2,13,15 11,16
11,12,18,20,21, IS.,9, 11,17, 18,20 18,17,20,23, 18,22
23.25 20,26 25,26
18,23,25 9,11,20,22, 18,20,33 18,20,23,25 22,23,25,39 18,20,22,40
23,24,25,28
22,23,24,25 18,23,24,25 (32) .22, ^5_, 23,24 24,25
14, (25) 9 14
12 9,18,20,23, 25
24,25
17, (19) (35) 20, 30, (13)
*References that are underlined are thorough and complete; references in parenthesis have sketchy information which is difficult to evaluate.
-------
A. Particulates
1. Bibliography
Reference A-ll (1970)
A comprehensive evaluation of the performance of seven incinerators is
presented which has determined: (1) the quality ah'3 quantity of solid
waste processed, residue, and gas borne particulate emissions, (2) the
quality of the flyash collected and the waste water produced, and (3) the.
economics involved in incineration. The sampling procedure employs the
PHS sampling train and the resulting chemical analyses are discussed in
detail. Particulate emissions range from 8.6 to 20.4 pounds per ton of
refuse, with an average of 12.5. In all cases, these emissions exceeded
the most lenient air pollution grain-loading emission standards.
The particulates caught after the filter average 30% of the total
particulate catch including (1) residue left after evaporation of the
acetone used to rinse the sampling train from after the filter to before
the impinger that contains the silica gel, (2) residue left after
evaporation of the chloroform and ether used to extract organic materials
from the impinger water wash, and (3) residue after evaporation of the
impinger water wash. The residues from the acetone wash and from the
chloroform-ether extracts average 30 percent of the material caught after
the filter; the inorganic residue from the impinger water wash is 70
percent. Approximately 0.05 percent of the residue is metal. The most
abundant metallic species were Ca (400 ppm), Zn (100 ppm), Al (25 ppm),
Mg (13 ppm) and Fe (6 ppm).
Several impinger residue samples were combined into two test lots for
wet-chemical analysis for inorganics and instrumental analysis for
organics. Approximately 28 and 43 percent, respectively, of these
residues were acetone soluble. Sulfates were present in approximately
32 and 20 percent, respectively. The acetone extract of both samples
showed carbonyl and aromatic bands in the infrared (presumably derived
from polynuclear compounds). No hydroxyl or aliphatic bands were noted.
The reported analyses indicate that perhaps some of the material caught
after the filter should be reported as particulates and some should not.
The organics and metals would clearly be classified as particulates. The
chlorides, sulfates, and phosphates, however, may be formed by gases
reacting with cations to form particulates while in close contact in the
impinger water. If so, they probably would not react if emitted to
the atmosphere and would not fall within the category of particulates.
Further work is needed on identifying composition of impinger water
residues and their origin since the cited work was "primarily a screening."
A-17
-------
Reference A-12 (1970)
This paper describes a laboratory scale program to evaluate source-
sampling equipment used to measure particulate emissions from incinerators.
Tests involving simultaneous sampling using the IIA T-6 train and the
PHS train exhibited results which varied appreciably for the two systems.
Reference A-13 (1968)
Typical estimates of particulate emissions are 1.6 grnins/SCF corrected
to 50 percent excess air. Current recommended standards range from
0.35 to 0.05 grains/SCF (50 percent excess air) which requires from 82
to 97 percent removal by APC devices. A comparison of the effectiveness
of various APC systems for particulate removal is discussed.
Reference A-14 (1968)
Although the article is principally concerned with calculating a sulfur
mass balance between refuse and incinerator emissions, data are given
for typical fly ash compositions collected in the hot flue gases, the
dust zone and in the stack. The S0_ content of the fly ash was found to
increase proportionate to combined alkaline earth contents. The author
concludes that about 25 percent of the sulfur introduced in refuse is
tied up in the fly ash as sulfate.
The table below presents typical chemical analyses of fly ash sampled
in the hot flue gases, in the dust collection zones, and in the stacks.
Chemical Analysis
A B C_
Organic 31 4 21
Inorganic 69 96 79
Inorganic Fraction
Si02 47.2 48.7 36.5
A120 10.2 23.4 25.9
Ti02 1.1 0.7
Fe202 15.6 6.5 7.1
CaO 18.4 9.2 8.9
MgO 2.9 2.3 2.8
Na20 + K20 4.5 5.8 10.5
S02 1.2 3.0 7.6
100.0 100.0 100.0
A. Suspended in hot flue gases.
B. Dust collection zones.
C. Collected in stack.
A-18
-------
Reference A-15 (1968)
This paper reviews and discusses the incinerator emissions problem,
methods of particulate emissions control and their relative cost.
Incinerator dust size measurements, determined in the BAHCO centrifugal
classifier using the methods and procedures of ASME Performance Test
Code No. 28, indicate that about 35 percent of the average dust leaving
the furnace is below ten microns in size. Furance dust emissions vary
from less than 10 pound/ to 60 pound/ton of refuse. High-performance
turbulent incinerators emit about 35 pounds/ton.
Various types of APC devices are described. Multi-cyclones are extremely
efficient for large particles, but performance drops off rapidly for
dust sixes smaller than 20 microns.
Reference A-2 (1968)
A compilation of air pollutant emission factors is presented, including
fly ash particle size distributions, collection efficiencies of various
APC devices and thrj composition of gaseous emissions. A sizable
bibliography is included.
Municipal incinerators are quoted as emitting 17 pounc's particulate/ton
of refuse with the following size distributions.
Size Range (yim) Weight Percent
less than 5 15
5-10 10
10-20 15
20-40 20
greater than 44 40
Reference A-16 (1967)
The maximum stack emission rates measured on the North Hemstead, New
York, 3-unit incinerator system averaged 2.43 pounds particulate/ton of
refuse for 8 tests carried out over 5 days. Other New York area
incinerators exhibit particulate emissions of 2.3 to 30 pounds/ton of
refuse for 10 incinerators cited.
Reference A-17 (1967)
This paper describes the initiation of incinerator testing in the
New York City area to (1) evaluate seasonal variations in refuse and
stack emissions, (2) identify and quantify specific inorganic and
organic emissions and (3) to compare emissions from four different types
of incinerators. The preparation of a sampling train and initial test
results are described.
A-19
-------
Reference A-19 (1967)
A specific procedure, for incinerator stack sampling is presented, using
experience in the Los Angeles area as a guide. Sampling and analysis
methods for specific pollutants are reviewed in detail. Spectrogrsphic
analysis of two samples of solid particulate yields the following
results:
Particulate - Weight Percent
Component
From
Multiple Chamber
Incinerators
Aluminum
Boron
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Potassium
Silicon
Silver
Sodium
Tin
Titanium
Vanadium
Zinc
Zirconium
1.2
0.32
0.92
0.14
0.0015
0.13
1.8
0.33
1.3
0.18
0.0093
0.16
Trace
12.0
0.016
Trace
0.031
0.022
0.0012
3.2
0.010
0.34
0.32
0.50
3.7
0.026
0.015
19.5
0.11
1.0
0.11
0.003
2.8
—
2.7
—
—
0.020
0.009!
0.014
0.51
Reference A-4 (1966)
A comprehensive series of tests were carried out on a new municipal
incinerator to establish performance data. Total particulate emissions
were in the range of 6 to 12 pounds/ton of refui.se, with an average fcr
17 tests of 7.8 pounds/ton of refuse. Spe.ctrographic and chemical
analysis results and particle size distributions of the fly ash are as
follows:
A-20
-------
Spectrographic Analyses
Elements Reported in Percent Ashed Material
Element^
Silicon
Manganese
Chromium
Nickel
Copper
Vanadium
Iron
Tin
Aluminum
Zinc
Magnesium
Titanium
Silver
Boron
Barium
Beryllium
Calcium
Sodium
Lead
Dust
Stack Effluent
5+
.1-1.0
.1-1.0
1-10
.1-1.0
.001-. 01
.5-5.0
.05-. 5
1-10
1-10
1-10
.5-5.0
.001-. 01
.Ol-.l
.l-.l
.001-. 01
10+
1-10
.Ol-.l
Test Run No. 6
Collector Catch
10+
.1-1.0
.Ol-.l
.001-. 01
.Ol-.l
.Ol-.l
.5-5.0
.-5-. 5
1-10
1-10
1-10
.5-5.0
.001-. 01
.Ol-.l
.1-1.0
.001-. 01
10+
.5-5.0
.1-1.0
Residue
10+
.1-1.0
.Ol-.l
.001-. 01
.Ol-.l
.Ol-.l
1-10
.1-1.0
1-10
.1-1.0
1-10
.5-5.0
.0001-. 001
.Ol-.l
.1-1.0
-.001
10+
.1-1.0
.1-1.0
Dust Test
Run No. 8
Stack Effluent
5+
.1-1.0
.1-1.0
10+
.1-1.0
.001-. 01
.1-1.0
.001-. 01
.1-1.0
1-10
1-10
.5-5.0
-.0001
.Ol-.l
.1-1.0
-.001
10+
1-10
.05-5
PH
Sulphur
Phosphorus
Silicate
Anions
Phosphates
Nitrates
Sulfates
Chlorides
8.3
1.460
.88
.62
.50
.02
Wet Chemistry Analyses
12.3 12.3
.620 .350
1.760 1.390
7.7
1.140
5.4
.77
.64
2.1
.22
Particle Size Analysis
Density: 1.85 grams per cubic centimeter
Microns
44
30
20
10
5
Percent by Weight Less
Than the Stated Size
89.2
68.7
47.2
20.5
6.0
A-21
-------
Reference A-19 (1966)
No chemical or sampling information given. However, this study reports
that limiting the excess air to a maximum of 25 to 50 percent minimizes
the entrainment of particulf-tes in the exhaust gas.
Reference A-6 (1966)
Total incinerator emissions are given for several Western European
incinerators. Representative fly ash compositions and particle size
distributions are given below. There is no discussion of sampling or
analysis methodology.
Chemical Analysis of Fly Ash
(Percent
Organic
Inorganic
Silicon as SiCL
Aluminum as Al_0»
Iron as FeoOo
Calcium as CaO
Magnesium as MgO
Sodium as NA20
Potassium as K-O
Sulfur as SO.
by Weight)
Collected
Portion
0.5
99.5
52.4
21.5
5.7
7.8
1.8
3.7
3.8
3.3
100.2
Discharged
Portion
15.0
85.0
36.1
25.1
7.1
8.6
2.4
6.0
5.9
8.9
100.0
A-22
-------
Size Distribution of Stack Dust Emission
Size Distribution Cumulative
Microns Percent _ Percent
+840 0.4 0.4
+250 3.2 3.6
+149 8.4 12.0
+ 74 17.6 30.4
+43 5.4 35.8
+40 5.3 41.1
+ 28 13.3 54.4
+ 15 14.1 68.5
+ 9 '9.6 78.1
+4 11.9 90.0
+2 6.7 96.7
Reference A-20 (1964)
Test data are presented for flue gas compositions and total particulate
emissions for two municipal incinerators. Particulate emissions from
the stack correspond to 2.6 Ibs/ton of refuse. The benzene soluble
fraction of the particulate is less than 1 percent. Sampling and
analysis procedures are as follows:
The gas stream is passed through a series of water bubblers kept at
32°F, then a series of freeze-out traps immersed in a dry-ice-alcohol
bath held at -98°F, and a high-efficiency glass-fiber filter (MSA type
1106-BH). Isokinetic sampling conditions range from 4-6 SCFM. Total
weight of particulates is determined by collecting material from the
sampling probe and from the sampling train filter; material filtered
from the condensate and liquids used to rinse the train; and material
recovered from evaporation of the condensate and water rinses. The
measured emissions from several municipal incentors are as follows:
A-23
-------
POLLUTANT EMISSION SUMMARY
Type of Unit
Sampling Point
Flue-Gas Conditions
CX>2 02
H70 Dry Easis
Total Particulates
Lbs. per
Ton of
Refuse
% Benzene
Soluble
Organics
Municipal Breeching (before
250-Ton/Day settling chamber)
Multiple Chamber
8.8
6.2 13.7
18
0.32
50-Ton/Day
Breeching (before
scrubber)
9.1
8.7 11.0
8.1
0.15
Stack (after
scrubber)
12.4
3.7 16.7
2.6
0.96
-------
Reference A-9 (1964)
In a study where measured incinerator stack emissions are compared to
calculated emissions based upon refuse composition, the following
compositions are given for fly ash (average of five tests)
S02 0.1 MgO 9.3
A1203 6.2 Na20 4.3
Fe03 2.6 K20 3.5
CaO 14.8 S03 0.1
Reference A-21 (1964)
This program resulted in the construction of a prototype instrument for
the analysis of fly ash emitted from municipal incinerators. The
instrument v/a* proven both in the laboratory and in the field. The
instrument samples a predetermined amount of stack effluent isokinetically,
separates the particulate material in a cyclone, collects a representative
sample of fly ash on filter paper, and measures the amount of sample by
a gravimetric technique using a beta-gauge principle. The entire concept
is designed to function on a routine basis and be operated by unskilled
personnel.
Reference A-22 (1962)
A sequence of field tests was made on full scale incinerators to
determine the relationship between underfire air and particulate
emissions in terms of incinerators of different sizes and designs.
Findings: There appears to be a critical level of underfire air for
large municipal incinerators beyond which particulate emissions do not
increase significantly. On the two units tested this change in trend
occurred at an underfire airflow between 35 and 40 SCFM/sq. ft. of grate
area. The use of 50% overfire and 50% underfire air approaches an ideal
distribution causing the maximum burning rate which also keeps the
particulate emissions well below the loadings which result when using
higher underair percentages. The sampling techniques are the same as
those described in Reference A-25. Results are as follows:
A-25
-------
Combustion Conditions Emissions
Underfire Excess Temp. Secondary Before After
Air - % Air - % Chamber - % Scrubber Scrubber
50 Ton/Day Batch Charge Incinerator
20 235 1080 0.78* 0.57*
50 110 1500 1.01 0.55
70 100 1500 1.79 0.61
20 215 — 6.8** 4.1**
55 110 — 8.6 4.7
60 100 — 15.5 5.7
250 Ton/Day Continuous Charge Incinerator
20 190 1750 3.8*
50 180 1790 2.8
80 190 1930 3.3
100 150 1960 4.6
*Pounds Particulate/1000 Pounds Dry Flue Gas (connected to 50% excess air)
**Pounds Particulate (Ton of Refuse Burned).
Reference A-7 (1961)
The composition and particle size distributions of fly ash are given for
tests carried out on the Gansevoort Incinerator, New York City, summarized
as follows:
Chemical Analysis of Fly Ash Emitted from
Gansevoort Incinerator, New York City
Component Percent by Weight
Organic 14.5
Inorganic 85.5
Silica as SiCL 36.0
Iron as Fe20 10.0
Alumina as AI90« 27.7
Calcium as Ca6 8.5
Magnesia as MgO 3.4
Sulfur as SO 9.7
Sodium and potassium oxides 4.7
Apparent specific gravity 2.58
A-26
-------
Reference A-7 (continued)
Particle Size of Fly Ash Emitted from
Gansevoort Incinerator, New York City
Size
Plus
Plus
Plus
Plus
Plus
Plus
Plus
Plus
120 microns
90
60
40
30
20
10
5
microns
microns
microns
microns
microns
microns
microns
minus
minus
minus
minus
minus
minus
minus
120
90
60
40
30
15
microns
microns
microns
microns
microns
microns
10 microns
Percent
by Weight
5.
6.
17.
13.
12.
1.
21,
5,
8
,5
,7
,2
,5
.9
,2
,8
Minus 5 microns
12.0
Reference A-23 (1960)
This reference is one of a series (Ref. A-23, A-24, A-25) describing the
effects of incinerator operating conditions on pollution emissions. The
present paper cites the results obtained with a highly volatile fuel
component (asphalt saturated felt roofing). Particulate loadings were
determined by collecting samples on a seven inch diameter glass fiber
filter; samples taken under isokinetic conditions at a representative
velocity point in the stack. The experimental multichamber incinerator
that was employed is described in detail. Particulate emissions were
higher with asphalt roofing material than with cellulose fuel. With
the asphalt roofing, minimum discharge was 2.6 Ib./ton. With worst
conditions of underfire air, excess air and fuel feed rate, particulate
With large batch charging added, particulates
Large batch charging with only 50% excess
emissions were 47.2 Ib./ton.
increased up to 65 Ib./ton.
combustions air resulted in up to 84 Ib./ton. Cellulose fuel tested
discharged from 0.8 Ib./ton to 9.0 Ib./ton maximum.
To separate solid particulate into noncombustible ( or nonvolatile) and
combustible (or volatile) components, particulate samples were subjected
to a temperature of 550°C for three hours and checked for weight loss.
In general, the noncombustible-nonvolatile component of particulate
amounted to less than 20% of the total weight. Only one test condition
(with 40% noncombustible content) was higher, and this occurred under
optimum combustion conditions. Tests made at the 50% excess air level
had only 5% noncombustible content.
A-27
-------
Results are summarized in the following Table:
Fuel
Excess Feed
Combus- Rate
tion Air (Ib/hr)
Air Distribution
Under- Over- Second-
fire fire ary
Particulate
Lb/Ton
Gr/Scf Fuel %Ash
50% 150
100% 100
150
150% 150
200% 100
150
300% 100
80%
80%
10%
22%
60%
90%
15%
60%
15%
60%
60%
15%
60%
60%
12%
48%
48%
15%
60%
15%
60%
10%
15%
40%
40%
60%
15%
60%
15%
60%
20%
20%
83%
77%
40%
10%
85%
40%
85%
40%
15%
85%
40%
15%
68%
12%
12%
85%
40%
85%
40%
90%
85%
60%
10%
40%
85%
40%
85%
40%
0
0
0
0
0
0
0
0
0
0
25%
0%
0
25%
20%
40%
40%
0
0
0
0
0
0
0
50%
0
0
0
0
0
1.31
1.94
0.09
0.24
0.27
0.33
0.44
1.11
0.04
0.23
0.25
0.09
0.59
0.02
0.00
0.28
0.36
0.40
0.44
0.41
0.59
0.14
0.19
0.47
0.39
0.54
0.22
0.41
0.42
0.64
58.4
84.0
5.3
14.4
15.6
19.8
19.2
65.0
2.6
13.3
14.4
5.9
34.0
35.6
6.3
19.6
27.6
34.4
36.4
38.0
52.8
11.2
14.8
41.8
32.4
47.2
18.6
34.0
47.0
78.2
5.7
3.9
20.2
7.9
11.7
14.2
6.7
7.0
38.4
12.6
9.4
18.5
7.8
4.8
19.4
9.1
7.4
8.4
11.5
0.7
7.7
19.1
17.2
12.7
6.9
19.9
13.2
17.6
9.4
11.9
Reference R-24 (1960)
In a continuation of experimental studies of the influence of incinerator
parameters on pollution emissions (Ref. A-23, and A-25), particulate
emissions were found to increase as underfire airflow was increased.
Particulate discharge is essentially the same on a dry weight basis for
both 25% and 50% moisture fuel levels at a given underfire air velocity.
Increasing the fuel charging rate increases the particulate discharge
per hour, but shows no effect on a Ib. of particulate per Ib. of fuel
basis.
A-28
-------
When the dry weight of fuel burned is used as the basis of comparison,
the maximum produced by 25% moisture fuel is only 1.4 times the maximum
resulting from 50% moisture fuel. The results are summarized below.
Test Conditions
Excess
Combus-
tion
Air
Fuel Air Distribution %
Burned Under- Second- Over-
Ib/hr fire ary fire
Stack
Dis- Particulates
charge, Lb/Ton
scfm gr/scf Burned
50% Moisture Fuel
50 140
240
300 140
240
25% Moisture Fuel
50 140
240
100 140
240
60
15
60
15
60
15
60
15
60
15
60
15
60
15
60
15
20
0
0
20
0
20
20
0
20
0
0
20
0
20
20
0
20
85
40
65
40
65
20
85
20
85
40
65
40
65
20
85
184
184
200
186
322
276
289
259
410
428
406
398
700
618
772
530
233
211
228
219
440
348
364
388
545
568
576
553
710
920
875
915
0.037
0.054
0.019
0.023
0.074
0.11
0.031
0.040
0.053
0.017
0.031
0.023
0.18
0.073
0.044
0.032
0.33
0.33
0.019
0.16
0.13
0.33
0.044
0.041
0.045
0.016
0.027
0.030
0.089
0.075
0.085
0.052
0.83
1.2
1.2
0.53
1.7
2.4
0.64
0.73
2.6
2.4
1.5
1.1
5.2
3.2
2.4
1.2
9.4
8.4
1.4
4.3
4.0
8.2
1.2
1.1
3.0
3.2
1.9
2.0
4.5
4.9
5.3
3.4
A-29
-------
Reference A-25 (1959)
This describes the first phase of a study carried out on an experimental
incinerator to determine the effects of variation in combustion air
distribution. (See Ref. A-23 and A-24) Particulates in the flue
gas were sampled isokinetically at a representative point eight diameters
downstream from the breeching and eight diameters upstream from the top
of the stack. Solids were collected on an MSA type 1106-BH glass-fiber
filter without organic binder. Mass determination of the solids was made
on an analytical balance, and the volume of gas sampled was measured
with a dry-type bellows meter. A calibrated orifice meter, preceding the
gas meter, was used to control the instantaneous sampling rate. Correc-
tions were applied for meter temperature, meter pressure, barometric
pressure, and humidity.
Tests were limited to a uniformly prepared rubbish of 25% moisture
content which was burned under closely controlled conditions of charging
and air distribution in a fixed-dimension, multiple chamber incinerator.
One of the common concepts regarding particulate discharge from municipal
incineration processes is that total solid emissions result solely from
physical entrainment of particles in the gas stream, with particulate
discharge increasing with increased gas velocity through the fuel bed.
This study concluded that the concept of physical entrainment of solids
as the sole mechanism for producing total solid emissions is not valid.
While physical entrainment probably contributes to the discharge of
particulates, other factors such as volatilization and condensation of
inorganic compounds and incomplete combustion also affect the degree
of this discharge. The following results were obtained.
A-30
-------
TABLE 3A
Air Distribution Stack
Fuel Burned Underfire Overfire Discharge
Ibs/hr 50% Excess Air Level (Scfm)
Particulates
(pr/SCF)(lbs/ton burned
140
155
305
155
305
205
205
250
240
230
230
220
270
270
250
330
320
320
300
300
310
320
330
0.11
0.044
0.034
0.068
0.22
0.13
0.17
0.21
0.049
0.10
0.068
0.057
0.27
0.086
0.15
0.072
1.6
1.5
0.96
1.9
5.9
4.1
5.7
6.4
1.6
5.2
2.1
1.6
8.6
2.5
4.4
2.5
TABLE 3B
Mr Distribution Stack
Fuel Burned Underfire Overfire Discharge Particulates
Ibs/hr 50% Excess Air Level (scfm) (gr/SCF) (ibs/ton burned)
140 15% 0% 400
390
380
440
30% 20% 400
370
350
370
15% 20% 490
480
510
490
30% 0% 500
530
490
450
0.047
0.042
0.037
0.037
0.040
0.035
0.038
0.024
0.036
0.035
0.029
0.014
0.058
0.0
0.035
0.036
2.5
2.0
1.7
2.0
2.0
1.6
1.6
1.6
1.7
1.6
1.4
1.6
2.8
2.1
2.6
2.6
A-31
-------
Reference A-26 (1957)
A general discussion of incinerator testing and emission results is
presented. Specific findings reported include: 1) incinerators lacking
special fly ash collecting equipment emit an average of 20 Ibs particu-
late/ton of refuse with a range of 10 to 26 Ibs particulate/ton of
refuse; 2) forced draft air systems double particulate emissions; and
3) there are approximately 16 Ibs of stack gas per Ib of refuse.
Reference A-27 (1953)
The results of tests made on municipal incinerators in Los Angeles County
indicate 3.34 to 17.8 pounds of solids per ton of refuse burned; the
average for 9 tests is 8.84 Ibs/ton of refuse. The percentage of com-
bustibles was 4 to 10 percent. 30 percent by weight of the solid particles
were less than 5 microns in size. Four of the collected samples were
analyzed, yielding the following:
Sample 1
This sampling train consisted of a precipitator,
six impingers and a Millipore filter. A spectrographic
analysis of the material collected in the precipitator
thimble showed as follows. The amount collected in the
precipitator thimble represented 77.6 percent of the
total sample.
Large Amount (xx%) Very small Amount (.x%)
Sodium Lead
Potassium Titanium
Calcium Zinc
Aluminum
Silicon
Small Amount (x%) Traces (.Ox to .000x%)
Magnesium Barium
Iron Strontium
Lithium
Copper
Chromium
Manganese
Tin
A-32
-------
A chemical analysis of'the cations and anions showed:
SiO 20.6% Cl 12.2
R2CL 10.4 SO 16.8
Ca 2.7 PCT 14.0
Mg 4.7 (max., cal. on NO^ 0.5
basis of SO.) Combustibles 4.2
Na 4.1 4
K 13.3
The precipitator condensate, or the material that was
collected in the impingers, representing 22.3 percent of
the total sample, was also analyzed chemically. The re-
sults of these analyses were:
R203 2.2% Ca 5.6
SI02 6.1 Na+K+S04 86.1 (by
difference
Sample 2
The second of the four samples for chemical identifi-
cation were taken using four impingers followed by a
Millipore filter. The dissolved and undissolved material
collected in the impinger train (84 percent of the total
sample) were concentrated by evaporating the free mois-
ture to dryness at 110°C. The analysis of the cations
and anions showed the following:
SiO 13.8% Cl 15.2
R20^ 5.7 SO 24.1
Ca 2.5 fO, 11.4
Mg 4.7 NO]! 0.4
Na 4.2 Combustibles 4.2
K 11.7
This analysis compares favorably with that of the
precipitator sample of the first test of this series.
Even though the SiO-and R-O are lower, they appear
to be in the same relationship to each other.
The Millipore filter which followed the impingers
collected the remaining 16 percent of the sample. The
spectographic analysis of this material was as follows:
A-33
-------
Large Amount (xx%) Very Small Amount (.x%)
Sodium Magnesium
Potassium Iron
Small Amount (x%) Traces (.Ox to .000x%)
Calcium Barium Copper
Lead Strontium Chromium
Zinc Lithium Manganese
Tin Silicon
Sample 3 and 4
The third and fourth tests of this series were made
using a Whatman thimble following four impingers. The
small amounts of material collected in the Whatman
thimbles in each test were not used in this analysis
because it was impossible to accurately remove the
material from the thimbles. The materials collected
in all the impingers were combined for the following
chemical analysis:
Analysis of Water Insoluble Portion
Water Insoluble Material....1.9608 grams
Acid Soluble 0.1509 grams
Acid Insoluble 1.8099 grams
Ash 1.2235 "
Combustibles 0.5864 "
Acid Soluble Analysis
Iron 0.0671 "
Calcium 0.0350 "
Magnesium 0.0047 "
Phosphorus Present
A-34
-------
Analysis of Water Soluble Portion
Water Soluble Material (after drying) 4.1512 grains
Total Acid, as Sulfuric Acid 1.2315
Total Sulfates, as Sulfuric Acid 1.9659 "
Combined Sulfates 0.7344 "
Elemental Analysis of Water Soluble Portion
Chlorine 1.1504grams
Sulfates, as SO 1.6041 "
Nitrates 0.1537 "
Sodium 0.1794 "
Potassium 0.6966 "
Phosphorus 0.0174 "
Calcium 0.0631 "
Magnesium 0.0881 "
Iron and Alum, as R 0 ...0.0475 "
Insoluble 7.. ...0.0664 "
The total weight of the complete sample would thus be
Total Dried Solids 4.1512grams
Volatilized Chlorides 1.1504 "
Volatilized Nitrates 0.1537 "
Water Insoluble 1.9608 "
Material in Thimble 0.3340 "
Total 7.7501grams
Conclusions are:
1. Approximately 20 percent of the discharge is condensable,
with approximately 5 to 15 percent as SO .
2. Approximately 80 percent is particulates containing silicon,
lead, aluminum, calcium and iron. Since these constituents exist in the
atmosphere in aerosol form, they contribute to the reduction in the
visibility. It has been shown that as the particles in the 0.5 to 0.8
micron range increase relative to other sizes, the visibility decreases
in Los Angeles County.
3. Approximately 2/3 of the total sample collected is water
soluble.
4. Approximately 30 percent by weight of the sample is below
5 micron in size.
5. Approximately 4 to 10 percent of the solid particulate
matter is combustible.
6. The solids emitted into the atmosphere from properly
designed municipal incinerators can be as much as an 80 percent improve-
ment over that from backyard types.
A-35
-------
2. Discussion
Although there are many references to particulate emission levels, there
does not seem to be very good agreement among the data. In particular,
there is very little discussion of particulate sampling procedures
employed for specific reported emissions. Additionally, there is a
consistent lack of descriptive data on incinerator operating parameters,
antipollution control devices employed, combustion practice, etc. A
second area of uncertainty relates to definitions or terminology of
"solid particulate", "fly ash," "combustibles," "condensibles," and
"residues." These terms are often used interchangeably which causes
confusion in data interpretation. A third area, in which only cursory
discussion is presented, concerns particulate analysis. Emissions are
reported on a mass basis and, in some cases, fly ash chemistries and
particle'size distributions are given. There seems to be no use of the
more sophisticated characterization methods such as scanning electron
microscopy to measure the size and shape of individual particles and to
identify particle agglomerates, electron probe microanalysis to measure
the chemistry of individual fly ash particles, or electron diffraction
to identify crystalline phases present in fly ash. As a result, the
reported emissions can only be treated in a qualitative way.
Another area of concern that is not discussed in the literature is
chemical interactions due to the residency of collected particles in
the sampling apparatus. For example, is the fly ash collected on a filter
altered to carbonates and sulfates by the passage of wet flue gases? Is
the particle size distribution altered by agglomeration or by coalescence
of particles?
Finally, there is essentially no discussion of condensed particulates.
The recent paper by Achinger and Daniels (Reference A-ll) and the early
paper by Chass and Rose (Reference A-27) give the only reports of impinger
catch chemistries. There is insufficient data to evaluate chemical
reactions which may possibly occur between impinger liquids and the gas
effluent.
There is specific need for more quantitative data on the chemistry and
particle size distributions of solid particulates so that the efficiencies
of anti-pollution control devices can be more carefully evaluated.
A-36
-------
B. Carbon Oxides (N2.02)
1. Bibliography
Reference A-ll (1970)
Integrated gas samples were collected in a flexible bag sampler and
analyzed with the use of an Orsat analyzer. Instantaneous grab samples
were also taken during each stack test and analyzed with a manual wet-
chemistry carbon dioxide indicator.
Sampling was done at a number of different types of municipal incinators
with various kinds of APC devices. The composition of the refuse burned
at each incinerator is detailed. CO. levels ranged from 2.8 to 5.0
percent with excess air values of 200 to 500 percent. No correlation
existed between CO- levels and percentage excess air.
Reference A-28 (1970)
This paper describes experimental incineration of widely-used disposable
container plastics, and describes the sample train used to test the
resulting effluent. High proportions of CO and CO- were reported. The
furnace used was a tightly-controlled experimental one. Volatile
combustion products were collected in Saran bags attached to the end of
the combustion tube, which protruded about 6" past the furnace. Analyses
of major products were done by infrared spectroscopy. Minor components
sometimes required gas chromatography.
Reference A-2 (1968)
Municipal incinerators are reported to emit one pound CO per ton of
refuse.
Reference A-17 (1967)
This paper describes parts of a program designed to compare the
composition of stack emissions from several New York City area incinera-
tors. The sampling train includes a probe with which embodies a filter
for particulates, two large volume and one small volume, specially-
designed coil traps. During operation, the first two coil traps are
maintained at 0°C and the third at -78°C (with dry ice-acetone). This
procedure was employed to freeze out and condense various organic species.
Representative samples of gaseous components are collected in flush-thru
gas sampling tanks.
The high collection efficiency of the trapping system was demonstrated
in experimental runs. Two tests yielded 5.73% CO- and 0.046% CO for
12.8% 0-, and 7.4% CO, and 0.090% CO for 11.6% 0-.
A-37
-------
Reference A-16 (1967)
This paper describes the results of studies at a municipal incinerator
in New York. All testing was conducted in accordance with the ASME
Test Code PTC-21 (1941) and PTC-27 (1957). Average Orsat analyses for
a series of 10 tests yielded: CO - 3.2 to 4.2%; 02 - 15.2 to 16.0%,
CO - 0.0 to 0.1%; and N2 - 79.7 to 81.0%; excess air levels ranged
from 242 to 313%.
As an example of the APC devices involved, #3 incinerator in this study
included a secondary combustion chamber, a low velocity expansion or
cooling chamber, 4 banks of high pressure water sprays, a secondary
baffle chamber, and cyclone collectors. Testing location: openings
constructed in the chimney walls at a point 129 ft. above the base of
the 265 ft. stack.
Reference A-29 (1966)
In a multiple chamber incinerator, there was emitted 0.7 Ib CO/ton of
refuse. However, no sampling or analysis data were given.
Reference A-20 (1964)
Selected tests on multiple chamber incinerators yielded 6.2 to 8.7
C02 and 11 to 13.7 0 in the breeching and 3.7% CO and 16.7% 0 in the
stack; CO levels were to 1 to 4 Ibs./ton of refuse and 2 Ibs./ton of
refuse (max.) respectively.
CO-, O-, CO and total gaseous hydrocarbon concentrations were measured
by collecting a 50 to 100 liter sample of the combustion gases in a
Mylar plastic bag and subjecting this sample to individual analyses.
Integrated samples were obtained by maintaining a flow of one liter per
minute or lower of sample gas to the bag over a one- to three-hour
period simultaneous with the poly-nuclear hydrocarbon sampling. Where
different modes of operation yielded more than one flue-gas flow rate in
the stack, the sample flow rate to the bag was adjusted proportionately.
CO- was analyzed by Orsat and checked with a nondispersive infrared
analyzer; 0. was measured by Orsat and a paramagnetic type 0. analyzer;
CO was measured by a nondispersive infrared analyzer and gas detector
tubes.
Emissions of total gaseous hydrocarbons were plotted against carbon
monoxide emissions; the least-squares line drawn through the points has
a correlation coefficient of 0.71 (significant at the 1.0% level). A
better correlation between CO and total gaseous hydrocarbons was not
expected in view of the variety of types of combustion processes involved.
A-38
-------
Hydrocarbons are more susceptible to oxidation and decomposition at high
temperatures than is carbon monoxide. Accordingly, the hydrocarbon
emission rates are lower than those for CO, and in general the ratio of
the total gaseous hydrocarbon emission to the CO emission is lower for
the larger incinerators.
Reference A-9 (1964)
Analyses of the components of municipal refuse were assembled, from
which an average composite analysis was calculated. The analyses
of the flue and stack gases from 18 municipal incinerators were also
tabulated and plotted. A marked difference was noted between the
analyses of the stack gases and those to be expected from the combustion
of the composite refuse. Much of the difference can be accounted for by
the carbon loss in the residue and oxidation of metallics. Suggestions
for future investigations include new refuse and residue analyses, and
chemical studies on the interaction of spray water and fly ash with
flue gas.
Calculated gas analyses based upon refuse compositions are:
Dry Volume, Per Cent
Air
Supplied
Theoretical
100
200
300
400
per
per
per
per
cent
cent
cent
cent
excess
excess
excess
excess
CO
19
9
6
4
3
•
.
•
.
•
2
62
73
47
85
87
0
10
14
15
16
°2
.00
.56
.08
.76
.81
N
80.
79.
79.
79.
79.
2
38
71
50
39
32
C/(H)
36.3
37.7
38.1
39.4
41.2
Assuming 20 percent carbon in the residue and partial oxidation of metals,
the calculated flue gas composition becomes:
Mr
Supplied
Theoretical
100
200
300
400
per
per
per
per
cent
cent
cent
cent
excess
excess
excess
excess
Dry Volume, Per
co2 o2
18
9
6
4
3
.80
.27
.15
.61
.68
0
10
14
15
16
.00
.61
.09
.80
.84
Cent Apparent
N2 C/(H)
81
80
79
79
79
.20
.12
.76
.59
.48
21.
21.
22.
22.
22.
2
6
1
3
7
A-39
-------
In practice, 35 tests on incinerators with dry gas systems yielded an
excess air level of 145 to 771 percent with an average of 324 percent.
The average gas analysis was 3.88 CO , 16.15 0 , 0.03 CO, 79.94 N and
C/(H)0f 10,8. Thirty nine tests on Incinerators with wet after chambers
had a range of 115 to 965 percent excess air, with an average of 266
percent. Gas analysis was 4.69 CO , 15.38 0 , 0.13 CO, 79.82 N0 and
C/(H) of 14.9. ^ 2
Precautions are given for gas sampling and analysis:
1. Use a water-jacketed sampling probe when withdrawing flue gas
from furnaces and other hot zones. A hot low carbon or stainless steel
or a cppper tube can be oxidized by the flue gas, which will reduce the
oxygen content appreciably. Also, the water cooling will help quench
combustion of CO and H_ that may be present in the gases.
2. When collecting gas samples over liquid, use a liquid that will
not react with or absorb any of the sample. Saturated NaCl brine, mercury
or saturated water solution of sodium sulfate acidulated with sulfuric
acid are satisfactory. Tap water absorbs CO^ readily, causing serious
error. Gas-saturated tap water readily evolves or absorbs gas with
changes of temperature and gas composition.
3. Gas compositions fluctuate rapidly with time, and will vary
across furnaces and passages. A gas sample represents only the gas
flowing past the end of the probe during the period of sampling. Strati-
fication is least in the stack and particularly after a dust collector
or induced-draft fan. However, mixing in passage does not eliminate
fluctuations based on time cycles, such as in batch furnaces. Numerous
gas samples are usually required, each analyzed carefully.
4. The Orsat apparatus works best at normal room temperatures.
Low readings for 0. are obtained at 40-60°F. The solutions must be kept
out of the manifoia piping, and must be washed out if accidentally
allowed to contaminate the bores. A gas sample should not be admitted to
the apparatus until the manifold is filled with nitrogen remaining from
a previous analysis of gas or air.
The following conclusions are reached:
1. The flue-gas analyses calculated for the probably refuse had an
apparent C/(H) ratio of 15.5 at 300 percent excess air. The drop from
24.1 to 15.5 was the result of an allowance for an estimate 20 percent
carbon in the total residue and a 54 percent oxidation of the metals in
the refuse. The result approximates the average of actual flue gases
from municipal incinerators.
A-40
-------
2. The analyses of gas samples taken at the inlets and outlets of
dry separation chambers indicated "the combustion of much carbon and some
carbon monoxide in the dry separation chambers. Char flakes are
deposited in a hot zone and continue to glow and burn over an extended
period of time. The median C/(H) ratios increased from 7.3 to 17.1 in
the plants from which data were available. Hydrocarbons may also have
burned in the dry chambers but to a minor degree in comparison with
carbon.
3. During passage through wet separation chambers the changes in
flue gases showed relatively more combustion of hydrogen and hydrocarbons
than of free carbon. Char flakes were probably burned more slowly and
were quenched before they had an opportunity to burn out. Between the
furnace outlets and the spray chamber outlets the median C/(H) of the
gases dropped from 17.5 to 11.8. Most of this change probably occurred
in the passage between the furnace exit and the inlet to the wet chamber.
4. While flue gases undoubtedly react chemically with the unevapo-
rated spray water and wet fly ash, the effect on the flue gas analysis
appears to be slight. Research in this area would be advisable.
Reference A-22 (1962)
Field tests were conducted on two different size full-scale incinerators
(50- and 250-tons/day). On both, the concentration of CO was too low
to be measured using the Orsat technique. The 50-ton/day incinerator
gave readings slightly above the limit of detectability on a nondispersive
infrared gas analyzer having a full-scale range of 0 to 5%. The 250-ton/
day incinerator CO was measured with a stain-in-tube type gas detector
with a detection range of 10 to 2000 ppm.
The 50-ton/day incinerator, at lowest average temperature, produced the
highest CO emission: 1.0 lb/1000 Ibs. dry flue gas.
The 250-ton/day incinerator, with high temperatures and high excess air
levels, produced 0.03 to 0.07 Ib. GO/1000 Ibs. dry flue gas.
Reference A-30 (1962)
Methods of sampling and analysis were developed by the APC District of
Los Angeles. The sampling train consisted of a miniature glass cyclone
separator, an alundum filter thimble, two or more wet impingers and a
dry impinger with a thermometer which was used for obtaining vapor
content of the water saturated effluent gas. An integrated gas sample
is taken during the test for carbon dioxide analysis.
A-41
-------
Reference.A-23 (1960)
In a special study measuring stack effluents resulting from incineration
of a high volatile fuel component (asphalt saturated felt roofing),
particulate and gas emissions were measured as a function of fuel feed
rate, charge size and percent excess air. The experimental setup is the
same as described in Reference A-25.
Sampling was done with continuous measurement by nondispersive infrared
gas analyzers supplemented with strip chart recorders for CO and CO,,.
Oxygen was measured on a continuous basis using an instrument operating
on the paramagnetic properties of oxygen.
All values reported herein are the averages of intermittent surges which
occurred for brief intervals (up to 50% of fuel charging cycle) imme-
diately after fuel was charged. Maximum peak concentrations ranged up
to five percent CO. Under the variety of conditions tested, this study
found 0 - 143.5 Ib. CO/ton of fuel and 3.9% - 10.9% average 02 - C02.
The CO levels were negligible for excess air levels greater than 100%.
Reference A-24 (1960)
The paper describes additional studies carried out on the experimental
incinerator set-up described in Reference A-25.
The thermal conductivity type analyzer used for CO. measurement was
replaced with a nondispersive infrared gas analyzer, with a range of
0 to 18% CO- and operated at a wave length of 4.29 microns. CO was
measured continuously with a positive type nondispersive infrared analyzer
with a range from 0 to 14%. 09 content was determined with an instrument
employing principles based on the paramagnetic properties of oxygen.
The CO content of the stack gases followed exactly the same pattern as
the hydrocarbons. With 50% moisture fuel, the CO concentration was
0.12% (30 Ibs per ton fuel) at 300% excess air and 0.11 (12 Ibs per ton
fuel) at 50% excess air. A maximum concentration of 2% CO was produced
when oxygen level dropped to 3%. With 25% moisture fuel, the CO con-
centration was 0.09% (29 Ibs per ton fuel) at 300% excess air and 0.13%
(19 Ibs per ton fuel) at 50% excess air. CO production occurred uniformly
at 300% excess air, but intermittant peaks of short duration occurred at
50% excess air immediately after introduction of the fuel.
Reference A-25 (1959)
A small, experimental, refractory-lined, multiple-chamber prototype
incinerator was employed to measure stack emissions under a variety of
experimental conditions. A detailed description of the incinerator and
sampling train is given: 0? and CO^ were continuously measured. CO
A-42
-------
content was measured by passing a continuous stream of effluent gas
through a positive type nondispersive infrared gas analyzer, employing
a diaphragm cell, with a range from 0 to 5% and operating at a wave length
of 4.4 4. Tests were limited to a uniformly prepared rubbish of 25%
moisture content.
Except for one unexplainable occurance of CO on a duplicate test, all
conditions of combustion at the 150% excess air level failed to produce
CO in excess of the 0.1% lower limit of detectability. Combustion at
the 50% excess air level produced intermittant surges of CO. (This
paralleled the occurance of hydrocarbons). CO discharges are a function
of proper employment of combustion air as it relates to distribution
within the ignition chamber and levels of oxygen available in the com-
bustion zone.
Reference A-18 (1957)
In two multiple chamber incinerators, CO was found at 6.3% and 6.4%
dry basis. An integrated gas sample for Orsat analysis is collected by
withdrawing gases continuously from the stack at a constant rate by a
liquid displacement method. A 5 liter bottle filled with a saturated
sodium sulfate solution, acidified with sulfuric acid, is used as the
collecting gas holder. As the solution siphons out the gas sample is
drawn into the bottle. The rate of siphoning is adjusted so that a gas
sample of about 4 liters is obtained during the course of the test. The
sample is drawn from the stack at the same cross section used for
collecting the combustion contaminants.
This reference contains specific discussions and explanations on the
location of testing equipment, number and variety of sampling holes in
the stack, isokinetic flow, sampling rate, and meter conditions. Data
from an actual test on a municipal incinerator is used to demonstrate
step-by-step the calculations made, with descriptions and equations.
Reference A-26 (1956)
A series of 6 tests on a unspecified incinerator yielded the following
Orsat analyses: 6.13% - 6.75% CO , 13.43% - 14.20% 0 , and 0.10%-
0.20% CO.
Measurements of the sampled gas volume involved the use of calibrated
pressure drop methods, Flowrotor, and velocity determinations. These
were always double or triple checked. This study was mainly concerned
with particulate emissions.
A-43
-------
Reference A-27 (1953)
In a summary of incinerator discharges in the Los Angeles area, the
CO. levels of 9 incinerators ranged from 5.5 to 8.9% (dry basis). The
major emphasis in this study was particulates, and not much specific
information is given on the sampling and analysis of gaseous effluents.
Reference A-31 (1967)
For a typical refuse composition (4500 BTU/lb.) and combustion with 200%
excess air, the author calculates the following stack gas concentrations:
CO, - 1738 Ibs./ton (6.05% dry volume), CO- lO'lbs./ton (0.06% or
600 ppm) and 0£ - 2980 Ibs./ton (14.32%).
2. Discussion
Most measurements of major combustion gases are made continuously by
Orsat analysis. These gases include C0_, N_, 02 and sometimes CO.
Orsat analysis has a typical sensitivity of + 0.1 percent, which is com-
parable to the reported range of CO emission levels. Thus, there is a
question of data credibility due to unintentional human bias in making
readings'.
Several authors have reported a change in CO emissions of up to 100
times increase immediately after charging in a batch type incinerator.
This effect must be kept in mind during sampling to insure a represen-
tative measurement of CO emissions.
C. Sulfur Oxides
1. Bibliography
Reference A-32 (1969)
It is reported that SO emissions are not affected by incinerator design;
they are a direct function of the sulfur content in the refuse burned.
Values given for emissions from multiple chamber incinerators are:
S0_ - 1.7 Ibs/ton of refuse burned; SO. - 0.1 Ibs/ton of refuse burned.
Reference A-2 (1968)
In a summary of emission data from municipal incinerators, SO- levels are
reported to be 2 pounds per ton of refuse.
A-44
-------
Reference A-33 (1968)
Determination of sulfur dioxide, sulfur troxide, or sulfuric acid mist
becomes difficult in the presence of other acid gases or where collected
samples are highly colored or turbid. A method is described by which
all oxides of sulfur are oxidized and precipitated as barium sulfate.
The barium in the precipitate is determined by measuring the intensity
of the L-alpha emission produced by x-ray excitation. The method is
useful for about 0.05 to 10 micromoles of oxides of sulfur in the
aliquot analyzed.
Reference A-34 (1968)
In Europe (West Germany), municipal incinerator refuse is generally
70% domestic and 30% industrial waste. Sulfur contents of the refuse
are typically 0.3 percent. SO emissions range from 0.122 to 0.286
grains/SCF; 15-20% of the sulfur is emitted as SO,^
Reference A-14 (1968)
Typical U. S. refuse (containing 28% moisture and 22% inerts) has a
sulfur content of 0.1%. The author has used test data from several
sources to evaluate sulfur emissions. Although there is a wide spread
of data, he concludes that 25% is emitted as S0~, 25% is tied up as
sulfate in fly ash and 50% is present as sulfate in the ash residue.
In a series of 27 tests carried out on four California incinerators,
SO levels were 1-100 ppm at 5% CO ; the median SCL levels was 17 ppm.
The sulfate content of fly ash is dependent on the percentage of alkaline
oxides (Na.O and K 0) . In a series of tests, the SO level of fly ash
increased from 1.2 percent (collected over the incinerator bed) to 7.2
percent (collected in the stack). The alkaline oxide levels increased
proportionately.
Reference A-31 (1967)
In a general discussion of incineration, the author claims that 0.1%S
is normal refuse (equivalent to 1 Ib. sulfur/ton refuse) results in
1 Ib/ton SO ; the remaining sulfur (75%) is fixed as sulfate in the fly
ash and residue.
Reference A-35 (1967)
Low SO emissions are due to low levels of sulfur in refuse (i.e., 0.1
percent) and competing reactions to form sulfate in fly ash and residue.
A-45
-------
Reference A-36 (1966)
The author gives typical SO (reported as SCO emissions from incinerators
as 20-460 ppm with an average of 250 ppm.
Reference A-20 (1964)
Sampling for SO analyses is described. The gas stream is passed through
a series of three 500 ml Greenburg-Smith impingers containing a 3%
solution of HO Sampling is at 1 cfm for 20 to 40 minutes. Analysis
is gravimetric (precipitating sulfate as barium sulfate) or by titration
of an aliquot with barium chloride with thorin as an indicator.
There is no discussion of incinerator SO emissions.
X
Reference A-37 (1962)
Data collected from several published sources is compiled. The following
SO levels (reported as S0_) are given: incinerator with spray chamber-
36 ppm; three incinerators with scrubber-0-32 ppm; three incinerators
without scrubber-11-60 ppm. There is no discussion of the data collec-
tion.
Reference A-18 (1957)
Provides sampling and analysis procedures for S0? and SO as follows:
The sample train consists of a paper thimble maintained just above the^
dew point of the stack gases followed by 3 series-connected impingers
immersed in an ice bath, a dry gas meter and a pump. The thimble acts
as a collector for the sulfuric acid aerosol formed from the sulfur
trioxide. The first 2 impingers contain 100 ml each of approximately
5% sodium hydroxide solution. The third impinger is dry. The sulfur
dioxide gas passes through the thimble and is collected in the impingers.
The thimble is extracted with hot water and the solution is titrated with
standard sodium hydroxide solution to determine the sulfur trioxide.
The sulfur dioxide collected in the impingers is determined by oxidation
with bromine, acidification, and precipitation as barium sulfate. For a
single chamber incinerator, the following test results are given:
SO- - 2.3 and 1.4 Ibs/ton of refuse burned; SO- - nil and 0.5 Ibs/ton of
refuse burned.
Reference A-38 (1954)
In 1950, a specific Los Angeles incinerator produced 1.9 Ibs S0?/ton of
refuse burned. They reported that 20 percent of the total emissions
were condensible, with 5 to 15 percent being SO .
A-46
-------
2. Discussion
Among the papers surveyed, there was no complete discussion of SO
emissions, including incineration parameters, sampling procedure and
analysis methodology. It is therefore difficult or impossible to judge
the reliability of the reported data. Most papers report SO levels in
ppm; these cannot be directly compared, however, for the basis of measure-
ment (i.e., percent excess air) is generally not stated. It appears that
there has been no attempt to measure sulfur mercaptans, such as H S. In
general, SO emissions do not seem to be considered important due to the
relatively low level of sulfur in refuse.
Based upon our review of the literature, we infer that about half of the
sulfur in the refuse is collected in the fly ash, the wet scrubbers and
the incinerator residue, the remainder being emitted as S09 and SO .
D. Nitrogen Oxides
3L. Bibliography
Reference A-13 - 1968
The author claims that nitrogen oxides are not of concern in refuse
burning since temperatures required for their formation (significantly
above 2000°F.) are not attained in municipal incinerators.
Reference A-29 - 1968
The author found 2.1 Ib. NO /ton of refuse. No sampling or analysis
methods are given.
Reference A-2 - 1968
In a summary of emission data from municipal incinerators, NO levels
are reported to be 2 Ibs./ton of refuse.
Reference A-31 - 1967
Typical nitrogen oxide levels in incinerator emissions correspond to
3 Ibs/ton of refuse (93 ppm, percent dry volume). There is no discussion
of sampling or analysis methods. Typical refuse analysis yields 0.5%
by weight nitrogen (10 Ib/ton of refuse).
Reference A-17 - 1967
Equipment was prepared for gas analysis of incinerator stack effluents
in the New York City area. Nitrogen oxides were to be analyzed according
to PHS Publication No. 99-AP-ll "Selected Methods for the Measurement of
Air Pollutants", May, 1965. No nitrogen oxides emissions data are given.
A-47
-------
Reference A-20 - 1964
In three tests, NCL was 2.5 Ibs/ton of refuse in a 250-ton/day incinerator
(sample taken before settling chamber). In a 50-ton/day incinerator,
there were 2.4 Ibs. N0?/ton of refuse before the scrubbers and 2.8 Ibs.
NCL/ton of refuse after scrubber. Oxides of nitrogen are produced by the
burning of all waste materials regardless of the quality of the combus-
tion process. (This is different from Ref. A-14's view on preceeding
page). NO concentrations were measured from two-liter integrated
samples, and analyzed either by the Saltzman or the phenoldisulfonic
acid technique. The Saltzman technique was not satisfactory for analyz-
ing samples from units that burn fuel with significant sulfur content
because of serious interference from sulfur dioxide.
Reference A-22 - 1962
A series of field tests were made on various-size full-scale incinerators.
Both sizes (50- and 250-ton/day) showed the same emission rate for NO ;
NO. emission increased at higher excess air levels. With low underfire
ana high overfire air, higher NO levels resulted (because oxygen con-
sumption is slower, and the air exists in the high temperature zone at
a comparatively high 0?/N2 ratio f°r a longer time.) The following data
were reported:
Combustion Conditions
Underfire Excess Temp. Secondary
Air -% Mr -% Chamber
Emissions
Before After
Scrubber Scrubber
50 Ton/Day Batch Charge Incinerator
20
50
70
20
55
60
235
110
100
215
110
100
1080
110
100
0.38*
0.28
0.16
3.2**
2.3
1.4
0.46*
0.25
0.23
3.3**
2.3
2.2
250 Ton/Day Continuous Charge Incinerator
20
50
80
100
190
180
190
.50
1750
1790
1930
1900
0.33* (3.0**)
0.27 (3.1)
0.22 (2.2)
0.20 (2.5)
Pounds NO (as NO )/1000 Pounds dry flue gas (corrected to 50%
excess air)
**
Pounds NO (as NO )/Ton of Refuse Burned.
X £•
A-48
-------
Reference A-3Q - 1962
This study used methods of sampling- and analysis developed by the Research
Division of the APC District, Los Angeles County: single chamber incinerator =
0.1 Ibs. N02/ton of fuel, multiple chamber = 2.1 Ibs. N02/ton. of fuel. Oxides
of nitrogen have little relation to the waste material composition, and tend
to be formed in proportion to incinerator operating temperatures. Thus,
multiple chamber incinerators, with higher temperatures, form more NOX.
Reference A-23 - 1961
For this study, these were done on an experimental incinerators. Incinerator
operating parameters, including excess air, fuel feed rate, Ib. fuel per
charge, and under- and over-fire air, were varied in a controlled way. NC^
ranged from 1.8 Ib. to 5.7 Ib./ton of fuel, in a series of thirty tests. NOX
were measured by Saltzman modification of the Griess colorimetric method and
expressed as equivalent NC>2. Samples were collected in evacuated flasks con-
taining the liquid reagent. The following results were obtained:
Fuel
Excess Feed
Combus- Rate
tion Air (Ib/hr)
50% 150
100%
150
150% 150
200% 100
150
300% 100
Air
Under-
fire
80%
89%
15%
22 1/2%
60%
90%
15%
60%
15%
60%
60%
15%
60%
60%
12%
48%
48%
15%
60%
15%
60%
10%
15%
40%
40%
60%
15%
60%
15%
60%
Distribution
Over- Second-
fire are
20%
20%
85%
77 1/2%
40%
10%
85%
40%
85%
40%
15%
85%
40%
15%
68%
12%
12%
85%
40%
85%
40%
90%
85%
60%
40%
40%
85%
40%
85%
40%
0
0
0
0
0
0
0
0
0
0
25%
0%
0
25%
20%
40%
40%
0
0
0
0
0
0
0
50%
0
0
0
0
0
N02
Lb/Ton
Fuel
1.8
2.1
5.2
4.2
4.0
3.8
4.7
3.3
4.1
4.5
3.6
3.8
3.9
3.6
4.2
4.4
4.4
4.4
4.4
4.3
4.7
5.2
4.7
5.6
4.5
3.0
4.4
5.7
3.0
3.4
A-49
-------
Reference A-24 - 1960
The same small experimental incinerator described in References A-23 and
A-25 was used in this study, which sought to determine the effects of fuel
moisture content on pollutant emissions, while varying the amount of fuel
charged per hour.
Oxides of nitrogen, expressed as N02, were determined by the Griess colori-
metric method as modified by Saltzman. On tests with 50% moisture fuel,
average NO^ concentrations were determined by taking up to ten instantaneous
grab samples over two or three burning cycles. On tests with 25% moisture
fuel, integrated samples were collected over two or more fuel charging
cycles. The two methods gave comparable average N02 values when used with
identical test conditions. The results are summarized as follows:
Excess Fuel
Combus- Burned
tion Air Ib/hr
50% Moisture Fuel
50 140
240
300 140
240
25% Moisture Fuel
50
100
140
240
140
240
Under-
fire
60
15
60
15
60
15
60
15
60
15
60
15
60
15
60
15
Stack
Second- Over- Discharge
ary fire (sefm)
20
0
0
20
0
20
20
0
20
0
0
20
0
20
20
0
20
85
40
65
40
65
20
85
20
85
40
65
40
65
20
85
184
184
200
186
322
276
289
259
410
428
406
398
700
618
772
530
233
211
228
219
440
348
364
388
545
568
576
553
710
920
875
915
101
108
121
128
100
139
132
144
61
68
86
77
67
73
66
83
89
98
102
90
106
106
119
47
42
46
52
50
47
58
61
Nitrogen Oxides
Ib/ton
ppm Burned
1.9
2.0
2.5
2.2
3.1
2.3
2.3
1.8
2.7
3.0
3.6
3.1
2.8
2.7
3.0
2.6
2.1
2.3
2.5
2.2
3.0
2.5
2.6
2.8
2.6
2.9
3.2
2.3
2.8
3.2
3.6
A-50
-------
Geometric Mean of Nitrogen Oxides for 25% and 50% Moisture Fuel Tests on
Both Wet and Dry Bases
Moisture Content of Fuel
Excess combustion air
Under fire air
Secondary air
Fuel feed rate, Ib/hr
Indicates statiscally
Reference A-25 -
50%
300%
15%
60%
0
20%
140
240
Lb NO /10 Lb Fuel Lb NO /10 Lb
asXFired Ratio DryxFuel
50% 25% 25/50% 50% 25%
1.40
1.40
1.28
1.26
1.32
1.22
1.28
1.26
significant
1959
1
1
1
1
1
1
1
1
.24
.45
.39
.29
.41
.27
.27
.41
difference
,
1.0
1.1
1.0
1.1
1.0
1.0
1.1
at 95%
a 2
2
2
2
2
2
2
3 2
.20
.92
.56
.52
.64
.44
.56
.52
confidence
1
1
1
1
1
1
1
1
.65
.93
.72
.88
.69
.69
.69
.88
Ratio
25/50%
0.
0.
0.
0.
0.
0.
0.
0.
8a
7a
7a
7a
7a
7a
7a
level .
This study was done on the same multiple chamber experimental incinerator
as described in References A-23 and A-24. Oxides of nitrogen, expressed
as NO., were determined by the Griess colorimetric method as modified by
Saltzman. For this determination, grab samples of the flue gases were
collected in evacuated flasks containing Saltzman reagent. An average
value of nitrogen oxides concentration, representative of operating
conditions, was obtained by integrating beneath the nitrogen-oxides time-
of-sampling curve. Preliminary studies indicated that the larger portion
of nitrogen oxides in the stack effluent were in the form of nitric oxide.
Variables found to be most significant in the formation of NO were
temperature, excess air and feed rate (grate loading); not significant
were pounds of fuel/charge, stoking interval, % secondary air, and %
underfire air.
Concentration of NO did not increase linearly with temperature, but as
a logarithmic function of temperature. Ignition-zone temperatures are
controlling. Other conditions being equal, increased operating tempera-
ture results in increased formation of NO . Typical levels of nitrogen
oxides (reported as NO ) are 30 to 40 ppmXat 1000°F. and 80 to 90 ppm
at 2000°F. Variations in excess air and refuse feed rate produced a
smaller effect, with NO values of 1.9 to 2.1 Ibs./ton of refuse (28 to
34 ppm). A summary of results is presented below.
A-51
-------
Fuel
Burned Air Distribution %
Ibs/hr Underfire Overfire
Stack
Discharge
Nitrogen Oxide
Emissions
ppm Ibs/ton burned
50% Excess Air
140
140
140
140
140
140
140
140
180
180
180
180
180
180
180
180
150%
140
140
140
140
140
140
140
140
180
180
180
180
180
180
180
180
15
15
15
15
30
30
30
30
15
15
15
15
30
30
30
30
Excess Air
15
15
15
15
15
15
15
15
15
15
15
15
30
30
30
30
20
20
20
20
0
0
0
0
0
0
0
0
20
20
20
20
0
0
0
0
0
0
0
0
20
20
20
20
0
0
0
0
250
240
230
230
220
270
270
250
330
320
320
300
300
310
320
330
400
390
380
440
400
370
350
370
490
480
510
490
500
530
490
450
80
83
66
74
69
65
74
65
62
79
78
85
79
95
91
75
51
53
63
64
48
54
63
52
75
60
47
66
60
55
54
58
2.1
2.0
1.6
1.7
1.6
1.8
2.1
1.6
1.6
2.0
2.0
2.0
2.1
2.3
2.3
2.0
2.1
2.1
2.4
2.9
1.9
2.0
2.2
2.0
2.9
2.2
1.9
2.6
2.4
2.3
2.1
2.1
A-52
-------
Reference A-18 - 1957
The author reports measured NO levels of 3.9 to 4.6 Ibs/ton of refuse
for a single chamber incinerator. Samples for oxides of nitrogen were
collected in evacuated bulbs. At least four samples were taken during
a 1 hour test period. The bulbs for oxides of nitrogen contain a
mixture of hydrogen peroxide and 0.1N sulfuric acid. Oxides of nitrogen
are determined by the phenol disulfonic acid method and reported as
nitrogen dioxide. Both the nitric oxide and nitrogen dioxide which may
be present are measured by this method.
2. Discussion
The literature contains numerous reports of nitrogen oxide emission
levels. Samples are generally collected in grab bottles and analysed
by the phenol disulfonic acid method. This method does not distinguish
between NO and N02. Emissions are believed to be quite constant with
time, showing reductions after charging in batch process incinerators
due to a lowering of the combustion temperature. High temperatures
(~2000°F) are required for rapid production of nitrogen oxides. Niessen
(A-l) reports 440 ppm nitrogen oxides are formed at 2000°F (100% excess
air) and 7 ppm at 1100°F (300% excess air). Although NO and NO are
unstable at ambient temperatures, the rates of dissociation are so slow
that there is little change in composition during residency in the
incinerator and stack.
Potential interactions of NO and N0« with other incinerator effluents
has not been treated in the literature.
E. Hydrocarbons
1. Bibliography
Reference A-2 - 1968
In a summary of emissions from municipal incinerators, the level of
hydrocarbons (reported as hexane) is reported as 0.3 pounds per ton
of refuse.
Reference A-22 - 1962
For comparison to laboratory studies (Ref. A-24 and A-25), a series of
field tests were made on full scale incinerators. The total organic
content of flue gases in a 250 ton/day incinerator was continuously
monitored using a nondispersive hexane sensitized infrared gas analyzer.
All tests showed hydrocarbon below the 15 ppm limit of detectability.
A-53
-------
On a 50 ton/day incinerator, integrated "bag samples were analyzed by a
flame ionization detector with a sensitivity of 4 ppm carbon (0.7 ppm
hexane). Hydrocarbon concentrations were below the 4 ppm carbon limit
of detectability. The total hydrocarbon emissions rate of this incinera-
tor for a variety of operating conditions was therefore less than 0.003
Ibs/ 1000 Ibs. dry gas. A 250 ton/day continuous feed incinerator
yielded less than 0.08 Ibs. hydrocarbons (as hexane) per 1000 Ibs. dry
gas.
Reference A-23 - 1961
Grab samples of incineration flue gases taken during hydrocarbon surges
were analyzed by gas-liquid partition chromatography. Components iden-
tified included acetylene, ethylene, ethane, propylene, and benzene.
Acetylene and ethylene predominated.
Tests were done on a multiple-chamber experimental incinerator (the same
used in Reference A-24 and A-25. Fuel used in this test was asphalt
saturated felt roofing cut into four-inch squares. The results are
presented on the next page.
A-54
-------
Fuel
Excess Feed
Combus- Rate
tion Air (Ib/hr)
50% 150
100% 100
150
150% 150
200% 100
150
300% 100
Air Distribution-%
Under- Over- Second-
fire fire ary
80
80
15
22 1/2
60
90
15
60
15
60
60
15
60
60
12
48
48
15
60
15
60
10
15
40
40
60
15
60
15
60
20
20
85
77
40
10
85
40
85
40
15
85
40
15
68
12
12
85
40
85
40
90
85
60
10
40
85
40
85
40
0
0
0
0
0
0
0
0
0
0
25
0
0
25
20
40
40
0
0
0
0
0
0
0
50
0
0
0
0
0
HC
Lb/Ton
Fuel
3.8
6.5
0.0
0.7
0.4
1.3
3.4
13.4
0.0
0.3
0.3
0.6
4.9
1.4
0.0
0.0
0.4
0.0
0.0
0.0
0.7
0.0
0.0
0.0
0.0
0.0
0.0
1.9
0.0
0.0
A-55
-------
Reference A-39 - 1960
Gas chromatography was employed to measure C~ to C, hydrocarbons for
comparison to other analysis methods and to a proposed regulation
limiting C? to C.. hydrocarbons to 50 ppm. The following data are re-
ported:
Gas Chromatographic Analysis of C^ to C Hydrocarbon in
Incinerator Effluents
Concentration, ppm
Test
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Typea
A
A
A
A
B
A
B
B
A
A
B
A
A
B
B
Acety
lene
0.8
1.5
0.9
2
711
—
247
341
2
—
59
—
800
600
Ethyl-
ene
8
9
8
3
2175
3
122
482
6
4
129
5
12
3100
2300
Ethane
0.3
0.3
0.3
0.0
425
—
—
32
1
—
3
—
—
500
300
C3
__
0.5
0.3
—
375
—
3
25
1
—
—
—
—
450
220
C4
__
—
—
—
120
—
3
4
—
—
—
—
—
175
120
C5
__
—
—
—
250
—
—
2
—
—
—
—
—
275
180
C6
__
—
—
—
75
—
—
—
—
—
—
—
—
120
90
Total
C9 to C
Hydro- 5
carbons
9
11
9
6
4100
3
375
886
10
4
191
5
12
5420
3810
Methaneb
15
10
15
30
9000
5
3000
3000
10
2
500
10
25
9000
7000
a - A + Adequately designed multiple chamber incinerator; B =
inadequately designed incinerator.
b - Infrared analysis.
A-56
-------
TABLE II Gas Chromatographic Analysis of Some C to C Hydrocarbons
Present in Incinerator Effluents
Test No.
Compound
Total C? compounds
n-butane
i-bu£ane
C4H3
3 methyl butene
Pentene-1
n-pentane
2 methyl butene- 2C
2 methyl pentane
3 methyl pentane
n-hexane
1
B
2
Type of
B
3
4
5
Incinerators3
B
Concentration »
3321
35
—
150
30
25
48
58
56
6
3
4400
70
20
160
50
40
85
75
77
15
12
3200
20
—
105
20
12
30
17
50
10
5
B
Ppm
855
4
5
2
B
369
2
1
a B = Inadequately designed incinerator.
This peak includes:
acetylene.
This peak includes 2 methyl pentene-1.
butene-1, isobutylene, butadiene 1,3, ethyl
The results indicate that, when the concentration of C? hydrocarbons is
below 10 ppm, the C_ to C, hydrocarbons are generally present in less
than 1 to 2 ppm. — "L ~ " 1~~~J L— — *" J~ t-J~T- '
tions, the C,
Wfien the C hydrocarbons are present in high concentra-
^, to C, compounds are generally present in significant con-
centrations. Some specific C, to Cfi compounds found in the effluents
from incinerators have been identified.
The standard of 50 ppm of C to C, hydrocarbons in incinerator effluents
appears to be a satisfactory emission standard since it adequately
differentiates between incinerators which are well designed and operated
and those which are not.
Reference A-24 - I960
This paper describes a continuation of the laboratory work described in
reference A-25. Substantial variations in excess air, percent under-
and over-fire air and refuse charge rate had little effect on total
hydrocarbons. The following results were found:
A-57
-------
Hydrocarbon Content (ppm)
300% Excess Air 50% Excess Air
50% moisture fuel 26 45
25% moisture fuel 13 37
Reference A-25 - 1959
In a series of laboratory experiments carried out to evaluate the in-
fluence of incinerator design and operation on atmospheric pollution,
hydrocarbon concentrations of the flue gases were monitored by a positive
type infrared gas analyzer sensitized with hexane and operating at a
wave length of 3.4y with a range from 0 to 750 ppm (expressed as hexane).
It is recognized that a hexane-sensitized instrument has zero sensitivity
to acetylene and very low sensitivity to ethylene, acetone, benzene and
other organic compounds which are normally encountered in incinerator
flue gases. The system cannot reliably detect hydrocarbon concentrations
of less than 30 ppm.
The findings relating to hydrocarbon and carbon monoxide concentrations
can be applied to industrial and municipal incinerator design and
operating practices. Since combustion at the 50% excess air level
produced these contaminants, discharges are controllable by proper
employment of combustion air as it relates to (1) distribution within
the ignition chamber, and (2) levels of oxygen available in the combus-
tion zone.
Tests were carried out on a multiple chamber experimental incinerator
employing a uniformly prepared rubbish of 25% moisture content, which
was burned under closely controlled conditions of charging and air
distribution. Hydrocarbon production was intermittent, each occurrence
lasting only one or two minutes depending on the combinations of condi-
tions responsible for its productions. With 50% excess air, maximum
concentrations ranged up to 400 ppm. With 150% excess air, there were
no hydrocarbons in excess of the 30 ppm lower limit of detectability of
the measuring system. The results of this work are given on the next
page.
A-58
-------
Fuel
Burned
Air Distribution
Under- Over-
fire fire
Stack
Discharge
(sefm)
Hydro-
carbons
50% Excess Air Level
140
140
140
140
140
140
140
140
180
180
180
180
180
180
180
180
150% Excess
140
140
140
140
140
140
140
140
180
180
180
180
180
180
180
180
15
15
15
15
30
30
30
30
15
15
15
15
30
30
30
30
Air Level
15
15
15
15
30
30
30
30
15
15
15
15
30
30
30
30
20
20
20
20
0
0
0
0
0
0
0
0
20
20
20
20
0
0
0
0
20
20
20
20
20
20
20
20
0
0
0
0
250
240
230
230
220
270
270
250
330
320
320
300
300
310
320
330
400
390
380
440
400
370
350
370
490
480
510
490
500
530
490
450
0.72
0.00
0.76
2.1
0.62
2.5
0.50
0.47
0.00
1.4
0.37
0.64
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
A-59
-------
2. Discussion
There is not very much information available in the literature concerning
hydrocarbon emissions. Most of the work reported dates back eight years
and longer. The hydrocarbon emissions appear to be quite low, generally
being less than 1 Ib/ton of refuse, and consequently have not been of
much concern as a pollution hazard.
_F. Specific Organics
1. Bibliography
Reference A-2 (1968)
In a summary of emission from municipal incinerators, the following
emission levels are reported: aldehydes - 0.3 Ibs/ton of refuse, organic
acids (reported as acetic acid) - 0.6 Ibs/ton of refuse.
Reference A-7 (1967)
In an evaluation of the specific organic compounds present in the stack
effluent of incinerators, the results of a partial analysis of a
representative sample (test run of sampling equipment) were:
33 g of water recovered from the 0°C traps contained about 10 ppm of
both acetone and acetaldehyde, and methanol, ethanol, methylethyl
ketone, a high molecular weight ketoalcohol, phenol, a high boiling
hydrocarbon, a solid carboxylic acid salt (possibly a formate) and a
solid high molecular weight alcohol, each of these in concentrations of
about 1 ppm.
6.7 g of water collected in the trap cooled to -78°C contained concentra-
tion levels of about 1 ppm of methanol, ethanol, methylethyl (2)
ketone, acetaldehyde, n-butyl alcohol and n-amyl alcohol; also a
relatively higher concentration of acetone and less than 1 ppm of benzene.
Reference A-20 (1964)
In a program designed to characterize particulate emission from heat
generation and incineration sources, the sampling system was specifically
designed to trap polynuclear hydrocarbons by condensation and to minimize
their loss when subjected to gas flow for extended periods of time. The
sampling train includes ice bath water bubbles (32°F) , freeze-out traps
(-98°F) and a high efficiency fiber glass filter. The system was esti-
mated to have removed 99% of the PNHs. After collecting total particu-
late weight, the benzene soluable fraction was analyzed by UV spectros-
copy. The PNH species analyzed were in the range 0.01 to 2 p grams per
pound of refuse.
A-60
-------
The initial step in determining polynuclear hydrocarbons involved benzene
extraction of the particulate matter, the condensate, and the rinse
liquids. Separation of the benzene-soluble fraction of the samples was
done by column chromatography, and the analysis was made by ultraviolet
visible spectrophotometry. The individual polynuclear compounds that
were quantitatively analyzed include benzo(a) pyrene, anthracene,
phenanthrene, pyrene, fluoranthene, benz(a) anthracene, benzo(e) pyrene,
perylene, benzo(ghi) perylene, anthanthrene, and coronene.
These compounds can be placed in two groups corresponding to the relative
reliability of the analytical determination for the individual compounds.
Collection efficiency of the sampling train was 99% or better for all of
the polynuclear compounds studied.
Benz(a) pyrene and benzo(e) pyrene were detected in the flue gases from
every incinerator studied. The concentrations of pyrene from each unit
were higher than those of any of the other polynuclear hydrocarbon com-
pounds detected, and ratios of pyrene to benzo(a) pyrene ranged from 6
to 120. A summary of data measured for several incinerators is as
follows:
A-61
-------
POLYNUCLEAR HYDROCARBON EMISSION SUMMARY—INCINERATOR SOURCES
NJ
Benzo(a)pyrene
Pyrene
Benzo(e)pyrene
Perylene
Benzo(g,h,i)perylene
Anthanthrene
Coronene
Anthracene
Phenanthrene
Fluoranthene
Benz(a)anthracene
Municipal
250-Ton/Dayb
0.075
8.0
0.34
34
0.24
9.8
0.37
50-Ton/Dayc
6.1
52
12
0.63
0.63
18
3.3
0.15
Commercial
5.3 Ton /Day d
53
320
45
3.1
90
6.6
21
47
140
220
4.6
3 Ton /Day
200
4200
260
60
870
79
210
80
50
3900
290
A blank in the table for a particular compound indicates it was not detected in the sample.
Breeching (before settling chamber)
•»
"Breeching (before scrubber)
Stack.
-------
Reference A-22 . (1.962)
For comparison to experimental incinerator tests, a series of field tests
were made on full-scale incinerators . Formaldehyde concentrations were
measured by a technique involving the bubbling of a sample gas stream
through chromatropic acid solution. Emissions from both the 50-ton and the
250-ton units were higher at the lower temperatures, and lower underfire
air levels.
Summary of Average Emissions from 50-Ton/Day Batch Charge Furnace
r. , ,. „ ' Formaldehyde0
linderrire Excess Secondary
Air, % Air % Chamber (°F) If. 2^
20 235 1080 9.9xlO~A 7.2xlO~4
50 110 1500 2.2x10-^ 0.0x10-4
70 100 1500 0.0x10-4
aFrom sampling point ahead of water spray scrubber.
''From sampling point after water spray scrubber.
cLbs/1000 Ibs dry flue gas (corrected to 50% excess air) .
Summary of Average Emissions from 250-Ton/Day Continuous
Feed Furnace
Combustion Conditions3
Underfire Air Excess Air
190
50 180
80 190
100 150
Temp. Sec.
Chamber
1750
1790
1930
1960
•u
Formaldehyde
3.9xlO-4
2.8x10-4
1.7xlO-4
1.7x10-4
sampling point in breeching.
bibs/1000 Ibs dry flue gas (corrected to 50% excess air.
A-63
-------
Reference A-30 (1962)
In this study, methods of sampling and analysis were those developed by
the APC District of Los Angeles. The following table details emission
findings.
Organic Emissions from Incinerators
Emissions, Ibs/Ton of Charged Material
Compound or Group Single Chamber Multiple Chamber
Methanol
Ethylene
Acetone
Methane
Acetylene
Alpha defines (as propylene)
Carbonyl Sulphide
Ben zone
Acids (as acetic)
Phenols (as phenols)
Aldehydes (as formaldehyde)
Ammonia
9-23
8-61
8
23-150
4-73
6
3
3
4
8
5-64
0.9-4
.05
.05
.05
.05
.05
.05
.05
.05
.05
.05
.3
.05
Reference A-40 (1962)
A variation of the chromotropic acid method of formaldehyde analysis was
investigated, which employs direct collection and color formation in a
0.1% chromotropic acid solution in concentrated sulfuric acid, instead
of the usual method involving preliminary collection in a bubbler con-
taining an aqueous bisulfite solution of just water. A brief study was
also made of the use of an aqueous solution of chromotropic acid. The
aqueous procedure is not useable in analyses of diesel or incinerator
effluents; the acid procedure is not applicable to raw and diluted auto
exhaust, but both can be used to analyze synthetic and actual photochemi-
cal smog. Because of its much higher sensitivity, the acid procedure is
convenient for formaldehyde analysis, even when the formaldehyde levels
are only a few parts per hundred million by volume. At these concentra-
tion levels and below, the use of optical cells of 5-cm path length are
advisable. Thus, for trace gas analyses, direct collection in acid
solution provides a more sensitive procedure than those various formal-
dehyde and aldehyde analytical methods that involve a l-to-10 dilution
step.
A-64
-------
Reference A-18 (1957)
Samples for the determination of hydrocarbons and aldehydes are collected
in evacuated bulbs. At least 4 samples are taken for each of the gases
during a 1 hr test period. The bulbs for aldehyde samples contain
sodium bisulfite absorbing solution. The aldehydes are determined by a
modified Ripper's method, and are expressed as formaldehyde. The follow-
ing emission levels (in Ibs/ton of refuse) were given for two single
chamber incinerators: organic acids (as acetic acid) - 2.0 and 3.9;
aldehydes (as formaldehyde) - 0.03 and 2.7; hydrocarbons - nil; and
acetylene - nil.
2. Discussion
As with total hydrocarbons, there is little literature data on the occur-
ance and amounts of various organic species in incinerator emissions.
Most reports lump all species of a functional group together. All
aldehydes are reported as formaldehyde and all organic acids as acetic
acid. As a result, the individual organic components have not been
analysed in any detail. Detection procedures in the past have generally
involved infrared spectroscopy. More recent studies combine gas chroma-
tography and high resolution mass spectrometry.
G. Inorganic Acids
1. Bibliography
Reference A-34 (1968)
The maximum HC1 content ranges around 0.02 vol. percent in wet flue
gases. A municipal incinerator shows emissions of HC1 from 0.122 to
0.204 grains/scf.
This study was done on European municipal incinerators. It states that
the emission values for gaseous material depend practically exclusively
upon the refuse composition, the proportion of plastics and industrial
refuse. No information was given about sampling or analysis methodology.
2. Discussion
There is essentially no data given on inorganic emissions. The increased
usage of chlorinated and fluorinated plastics as packaging materials has
caused some concern about future increases in HC1 and HF emissions. The
employment of appropriate APC methods can probably keep inorganic acid
emissions to very low levels.
A-65
-------
H. Volatile Metals
1. Bibliography
Reference A-27 (1953)
Emission spectrographic analysis of material collected in a precipitator
thimble (representing 77 percent of total particulate) yielded less than
one percent lead and zinc.
2. Discussion
There are no reports in the literature of specific attempts to measure
levels of volatilized metals. It is expected that low melting point
metals that are present in the refuse such as lead and tin will voli-
tilized and/or condensed metals will undoubtedly be emitted to the
atmosphere. Present sampling and analysis techniques are not suitable
for quantifying these emissions.
A-66
-------
I. List of References Cited
Ref. No. TITLE
A-l SYSTEM STUDY OF AIR POLLUTION FROM MUNICIPAL INCINERATORS
Walter R. Niessen
Arthur D. Little, Inc., Report to National Air Pollution
Control Administration, DHEW, March 1970, Volumes I and II.
A-2 REFUSE INCINERATION
R. L. Duprey
Pages 9-11, 53-67 in "Compilation of Air Pollutant Emission
Factors," U.S. DEPT. OF HEALTH, EDUCATION, AND WELFARE,
Public Health Service Publication.
A-3 PARTICULATE EMISSION AND BAFFLE WASHER PERFORMANCE AT TWO
INCINERATORS,
The Research Corporation of New England, Conducted for Metcalf
and Eddy, (November, 1970).
A-4 MUNICIPAL INCINERATION AND AIR POLLUTION CONTROL
Wilmer Jens and Fred R. Rehm
Proc. 1966 Nat. Incinerator Conf., New York, May 1-4, 1966
p. 74-83.
A-5 ELECTROSTATIC FLY ASH PRECIPITATION FOR MUNICIPAL INCINERATORS -
A PILOT PLANT STUDY
A.B. Walker
Proc., 1964 National Incinerator Conf., New York, p. 13-19
(May, 1964).
A-6 CONTROL OF AIR POLLUTION AND WASTE HEAT RECOVERY FROM INCINERATION
Casimir A. Rogus
Public Works £7 (6), 100-5 (1966)
A-7 MUNICIPAL REFUSE DISPOSAL
Comm. on Refuse Disposal, APWA
APWA Res. Found. Proj. 104, 182-9 (1961)
A-8 F.J. Wright
Twelfth Symposium (Internation) on Combustion,
Combustion Institute (1969).
A-9 REFUSE COMPOSITION AND FLUE GAS ANALYSES FROM MUNICIPAL
INCINERATORS
E. R. Kaiser
Proc. 1964 Nat. Incin. Conf., N.Y., May 18-20, 1964, p. 35-51.
A-10 "Air Borne Emissions from Municipal Incinerators."
A.A. Carotti
Draft copy of final report to HEW (Cont. No. PH86-67-72
and PH86-68-121).
A-6 7
-------
A-ll AN EVALUATION OF SEVEN INCINERATORS
W. C. Achinger and L. E. Daniels
Proc. 1970 Incinerator Conf., p. 32-61.
A-12 A COMPARISON OF NAPCA AND IIA SOURCE SAMPLING iffiTHODS
Gary D. McCutchen
National Air Pollution Control Administration, U.S. Dept.
of HEW, Durham, N.C.
A paper to be presented at a meeting of the Incinerator
Division, ASME. United Engineering Center, NYC, 30 Sept., 1970.
A-13 CONTROLLED COMBUSTION FOR SOLID WASTES DISPOSAL
James A. Fife
Heating, Piping and Air Conditioning 40_ (3), 140-7, March 1968.
A-14 THE SULFUR BALANCE OF INCINERATORS
Elmer R. Kaiser
Journal of the Air Pollution Control Assoc. 18 (3), 171-4,
March 1968).
A-15 INCINERATOR AIR POLLUTION CONTROL
J. H. Fernandes
Proc. 1968 Nat. Incinerator Conf., May 5-8, 1968, N.Y. , p. 101-16.
A-16 INCINERATORS CAN MEET TOUGHER STANDARDS
Herbert Mandelbaum
Reprint from the August 1967 issue of the American City
magazine, 3 pp.
A-17 ANALYSIS OF STACK EFFLUENT FROM MUNICIPAL INCINERATORS
R.A. Smith, J. Hornyak, and A.A. Carotti
Engineering Foundation Research Conference "Solid Waste
Research and Development, II"
Conference Preprint Number: C-8 (1967) 2 pp.
A-18 TECHNIQUES OF TESTING FOR AIR CONTAMINANTS FROM COMBUSTION
SOURCES
Carl V. Kanter, Robert G. Lunche and Albert P. Fudurich
Journal of the Air Pollution Control Assoc., 6^ (4), 191-9
February, 1957.
A-19 EUROPEAN DEVELOPMENTS IN REFUSE INCINERATORS
Casimir A. Rogus
Public Work, 97_ (5), 113-117 (May, 1966).
A-20 EMISSIONS OF POLYNUCLEAR HYDROCARBONS AND OTHER POLLUTANTS
FROM HEAT-GENERATION AND INCINERATION PROCESSES
R.P. Hangebrauch, D.J. vonLehmden, and J.E. Meeker,
Journal of the Air Pollution Control Assoc., 14 (7),
268-78 (1964).
A-68
-------
A-21 PROTOTYPE FLY ASH MONITOR FOR INCINERATOR STACKS (FINAL
REPORT, APRIL 15, 1963'- JULY, 1964)
M. J. Salkowski
IIT Research Institute, Chicago, 111., Technology Center
(Report IITRI-C8015-5.) July 1964, 54 p.
A-22 FIELD EVALUATION OF COMBUSTION AIR EFFECTS ON ATMOSPHERIC
EMISSIONS FROM MUNICIPAL INCINERATORS
Robert L. Stenburg, Robert P. Hangebrauck, Darryl J. Von
Lehmden
Journal of the Air Pollution Control Assoc. I2_ (2), 83-9 (1962).
A-23 EFFECTS OF HIGH VOLATILE FUEL ON INCINERATOR EFFLUENTS
Robert L. Stenburg, Robert P. Hangebrauck, Darryl Von Lehmden
Journal of the Air Pollution Control Assoc. _U (8), 376-83 (1961).
A-24 EFFECTS OF DESIGN AND FUEL MOISTURE ON INCINERATOR EFFLUENTS
Robert L. Stenburg, Ronald R. Horsley, Robert A. Herrick, and
Andrew H. Rose, Jr.
Journal of the Air Pollution Control Assoc. _10 (2), 114-20
(April 1960).
A-25 AIR POLLUTION EFFECTS OF INCINERATOR FIRING PRACTICES AND
COMBUSTION AIR DISTRIBUTION
Andrew H. Rose, Jr., Robert L. Stenburg, Morton Corn,
Ronald R. Horsley, Daniel R. Allen, and Paul W. Kolp
Journal of the Air Pollution Control Assoc. 8_ (4), 297-309
(February, 1959).
A-26 INCINERATOR TESTING AND TEST RESULTS
F. R. Rehtn
Journal of the Air Pollution Control Assoc. 6_ (4), 199-204
(February 1957).
A-27 DISCHARGE FROM MUNICIPAL INCINERATORS
Robert L. Chase and Andrew H. Rose, Jr.
Air Repair 3^ (2), 119-22 (November 1953).
A-28 COMBUSTION PRODUCTS OF PLASTICS AND THEIR CONTRIBUTION TO
INCINERATION PROBLEMS
Gwendolyn Ball and Edward A. Boettner
Presented before the Division of Water, Air and Waste
Chemistry, American Chemical Society, Chicago, 111., Sept. 15, 1970,
A-29 A RAPID SURVEY TECHNIQUE FOR ESTIMATING COMMUNITY AIR POLLUTION
EMISSIONS
Guntis Ozolins and Raymond Smith
U. S. Dept. of HEW, Public Health Service, Division of Air
Pollution, Publication 999 - AP 29, Cincinnati, Ohio (October
1966), pp. 72-3.
A-69
-------
A-30 INCINERATION PRACTICE & DESIGN STANDARDS
R. E. Williams
Clean Air Conference, Proc. 4th Technical Session, Paper 27,
University New South Wales, 19-21, February 1962, p. 1-23.
A-31 REFUSE REDUCTION PROCESSES
Elmer R. Kaiser
Proc. Surgeon General's Conf. on Solid Waste Management for
Metropolitan Washington, July 19-20, 1967, p. 93-104.
A-32 INCINERATION
S. Smith Griswold, Air Pollution Control Officer
Technical Progress Report, Volume 1, 69-83 (April 1960).
A-33 THE DETERMINATION OF OXIDES OF SULFUR BY X-RAY EMISSION
SPECTROMETRY
Janet Walkley Cares
Am. Ind. Hyg. Assoc. J. , 29_ (4), 386-9 (July-August 1968).
A-34 EXPERIENCES WITH REFUSE INCINERATORS IN EUROPE
PREVENTION OF AIR AND WATER POLLUTION, OPERATING OF REFUSE
INCINERATION PLANTS COMBINED WITH STEAM BOILERS, DESIGN &
PLANNING
H. Eberhardt and W. Mayer
Proc. 1968 Nat. Incinerator Conf., May 5-8, 1968, N.Y., p. 73-86.
A-35 COMPOSITION AND COMBUSTION OF REFUSE
Elmer R. Kaiser
Proc. MECAR Symp. Incineration of Solid Wastes, N.Y.,
March 21, 1967, p. 1-9.
A-36 CONTROL OF AIR POLLUTION FROM MUNICIPAL INCINERATORS
Fred R. Rehm
Proc. 3rd Nat. Conf. on Air Pollution, Washington, D. C.
Dec. 12-14, 1966, p. 327-31.
A-37 INCINERATORS
R. B. Engdahl
Pages 28-35 of "Air Pollution," Vol. 2, by Arthur C. Stern,
Academic Press, N. Y. (1962).
A-38 COMBUSTION AND SMOG - THE RELATION OF COMBUSTION PROCESSES TO
LOS ANGELES SMOG. IV. INCINERATION OF REFUSE
W. L. Faith
Air Pollution Foundation, Los Angeles, Calif., p. 50-63.
(September 1954).
A-39 GAS CHROMATOGRAPHIC ANALYSIS OF INCINERATOR EFFLUENTS
William N. Tuttle and Milton Feldstein
J. Air Poll. Control Assoc. 10 (6), 427-9, 467 (Dec. 1960).
A-70
-------
A-40 SOURCE AND ATMOSPHERIC ANALYSES FOR FORMALDEHYDE BY
CHROMOTROPIC ACID PROCEDURES
A. P. Altshuller, L. G. Lage, and A. F. Wartburg
Preprint, Robert A. Taft Sanitary Engineering Center,
Cincinnati, Ohio, Lab. of Engineering and Physical Sciences,
17 p., 1962, 7 Refs.
A-71
-------
APPENDIX B THE NATURE OF THE INCINERATOR PROCESS
Table of Contents
I. INTRODUCTION B_j_
II. THE SOLID WASTE PROBLEM B-2
III. INCINERATORS B_7
A. Types of Incineration B-7
B. Important Operation Factors B-10
1. Grates and Hearths B-IO
2. Furnace Enclosures B-ll
3. Combustion Air B-12
4. Flue Gas Conditioning B-12
5. Air Pollution Control B-13
6. Stacks B-19
7. Ash Removal B-19
IV. GASEOUS EMISSIONS B-19
V. PARTICULATE EMISSIONS FROM INCINERATORS B-20
A. Refuse Composition B-20
B. Method and Frequency of Feeding B-20
C. Underfire Air Rate B-21
D. Incinerator Size B-21
E. Burning Rate B-21
F. Combustion Chamber Design B-21
G. Grate Type B-21
H. Mixing Effects on Combustible Pollutants B-22
VI. REFERENCES B-23
B-l
-------
APPENDIX B
THE NATURE OF THE INCINERATOR PROCESS
I. INTRODUCTION
The purpose of operating an incinerator is to reduce the volume of solid
waste materials; converting the combustible matter to gaseous products
while producing a minimum of residual solid ash. Because of the moment-
to-moment variability of materials that are fed into an incinerator, the
accomplishment of this objective is a difficult task.
The processes involved in the operation of a municipal incinerator are:
accumulation of the refuse in a storage area or pit; feeding the refuse
to the furnace; burning; cleaning the flue gases of particulate collec-
tion and removal of residue; and discharge of the gaseous effluents.
Those systems with a system for air pollution control will also have a
need for a separate system handling the collected solids. Effective
emission testing of an incinerator requires an understanding of these
variabilities and that they can affect the outcome of the test.
This chapter will discuss types of waste encountered, the various types
of incinerators and the incineration processes which are currently being
used. A more in-depth discussion (from which almost all of the material
was taken) can be found in a three-volume report to NAPCA (now EPA) under
Contract CPA-22-69-23 by Arthur D. Little, Inc., entitled: "Systems
Study of Air Pollution from Municipal Incineration.4 The following
chapters of that report are particularly pertinent to this report, Volume
I, Chapter V, "The Nature and Causes of Air Pollution," Sections A and
B, and Appendix H, "The Incinerator Process."
II. THE SOLID WASTE PROBLEM
The increasing quantity of collected municipal solid waste, coupled with
changes in the refuse itself—composition, heating value, ash content,
and the like—point to the need for new or expanded incineration facilities
for the future and for new plant designs and standards to preclude air
pollution problems. Table B-l presents an analysis of the national annual
average composition of municipal refuse in 1968.^ These estimates were
computed from data from 23 sources, modified to provide an average composi-
tion of municipal waste. These averages can only be considered estimates
of the composition of average national refuse because of the limited
quantity of data available. Also, some variations can be expected, reflect-
ing localized mixes of commercial and domestic waste sources, industrial
wastes, and so forth. The differences between geographical locations can
be seen from the data in Table B-2.
Table B-3 presents an ultimate analysis of refuse categories by oxidizable
elements and inerts.
B-2
-------
TABLE B-l
Average Refuse Composition - As-Discarded Basis
Category
Wt (%)
Description
Glass
Metal
Paper
Leather & Rubber
Textiles
Wood
Food Wastes
Miscellaneous
Yard Wastes
8.3
8.2
35.6
1.5
1.9
2.5
23.7
1.7
15.5
100.0%
Bottles (primarily)
Cans, Wire, Foil
Various Types, Some with Fillers
Polyvinyl Chloride, Polyethylene,
Styrene, etc., as Found in Packaging,
Housewares, Furniture, Toys and
Nonwoven Synthetics
Shoes, Tires, Toys, etc.
Cellulosic, Protein, and Woven
Synthetics
Wooden Packaging, Furniture, Logs,
Twigs
Garbage
Inorganic Ash, Stones, Dust
Grass, Brush, Shrub Trimmings
*Taken from Reference 4, Vol. I, page IV-16.
B-3
-------
TABLE B-2
Estimated Annual Average of Municipal Refuse Composition (1968)'
("As-Discarded" Basis)
Category
Glass
Metal
Paper
Plastics
Leather ft Rubber
Textiles
Wood
Food Wastes
Yard Wastes
Miscellaneous
Un-Seasonal State Semi-Seasonal State Seasonal State
(e.g., Florida) (e.g. , Alabama) (e.g., Massachusetts')
7.6
7.5
32.6
1.0
1.3
1.8
2.3
18.2
26.1
1.6
8.1
8.1
35.1
1.1
1.4
1.9
2.4
19.5
20.7
1.7
8.8
8.7
28.2
1.1
1.5
2.0
2.7
21.1
14.1
1.8
*Taken from Reference 4, Vol. I, page IV-15.
B-4
-------
w
TABLE B-3
Estimate of the National Annual Average Composition of Municipal
Refuse Excluding Yard Wastes and Miscellaneous Categories'*"
(as-fired basis)
(1)
Component
Glass
Metal
Paper
Plastics
Leather &
Rubber
Textiles
Wood
Food Wastes
(2)
Data*
Samples
Utilized
23
23
23
9
9
17
22
23
(3)
Mean
Weight
Percent
9.7
10.0
50.3
1.4
1.9
2.6
2.9
18.8
(4)
% Mean
(100% basis)
9.9
10.2
51.6
1.4
1.9
2.7
3.0
19.3
(5)
Estimated
Standard
Deviation
S(X)
4.37
2.18
11.67
.96
1.62
1.80
2.39
10.95
(6)
95%
Confidence
Limits
1.89
.93
5.04
.74
1.25
.93
1.06
4.73
(7)
95%
Confidence
Range
8.0-11.8
9.3-11.1
46.6-56.6
0.7-2.1
0.7-3.2
1.8-3.6
1.9-4.1
14.6-24.0
97.6
+Taken from reference 4, Vol. I, p. IV-10
*Several data sets were not presented in a form suitable for extracting the weight fractions of all of the
above refuse components.
-------
The anticipated changes in the nature of this refuse over the next 30
years are particularly significant. Over this time span, study data
point to a more than 70% increase in the per capita refuse generation rate
over the current level which, when compounded by the effect of increas-
ing population, will increase by a factor of 2.65 the quantity of gen-
erated solid waste. Thus, a community with a stable population level may
well experience a refuse disposal requirement increase of 70% and an area
growing in population at the national average rate will require an incin-
eration capacity over 2.5 times that in 1968 by the year 2000.
In order to project the per capita waste loads and refuse composition for
each state, national indicators were developed for each of the 10 major
categories comprising municipal refuse. These indicators reflect the
national growth rates (through the year 2000) in the production of the
commodities comprising the major (but not necessarily all) sources of each
refuse component. The indicator for plastics, as an example, was developed
from weighted averages of the consumption projections for plastics used
in packaging, housewares, furniture, toys, and synthetic-nonwoven textiles.
These are major sources of plastics in municipal wastes. Projections of
the consumption of these segments of the plastics market should, therefore,
provide a realistic tool for estimating the growth of plastics in muni-
cipal refuse.
It should be noted that there are varying "lag times" between the purchase
and disposal of the various components of refuse. These lag times have
been estimated and incorporated into the development of these national
indicators for categories which are changing rapidly. Plastics packaging,
for example, has substantially no lag time associated with its disposal.
Plastics in furniture, however, may have lag times varying from a few years
to 30 years or longer, depending on the type of furniture and its use.
An average lag time of 10 years was estimated for plastics in furniture.
These indicators are not in themselves national projections of the quanti-
ties of the various components contained in the national refuse. Each
indicator merely projects the national growth rates of materials and goods
comprising that classification of refuse with appropriate adjustments for
lag time between consumption and disposal.
The projected national annual growth rates of each of the 10 major refuse
components as obtained from the indicator projections are shown in
Table B-4. For some components, such as food wastes, rubber and leather,
and wood, the annual growth rate is predicted to remain relatively con-
stant. The annual growth rate of plastics, metal, and glass are predicted
to vary considerably. Plastics, for example, are predicted to have signi-
ficant growth rates through 1980, but then to level off at a relatively
stable growth rate of about 5% per year.
B-6
-------
TABLE B-4
Time Period
From
1968
1970
1975
1980
1985
" 1990
1995
To
1970
1970
1980
1985
1990
1995
2000
Paper
Products
4.5
4.5
3.6
3.7
4.2
3.9
4.0
Miscellaneous
& Food Wastes
('/,
1.
1.
I.
1.
1.
1.
1.
:)
i
i
i
8
6
5
4
Metal
4.0
4.0
3.9
2.8
2.7
1.5
0.4
Glass
5.4
5.4
3.9
2.6
2.1
1.7
0.8
Plastics
11.9
11.9
11.4
5.6
5.6
5.0
4.7
Textiles
4.1
4.1
3.6
4.6
4.7
4.4
3.9
Rubber &
Leather
3.2
3.2
3.2
3.3
3.3
3.7
3.7
Wood
1.0
1.0
0.9
1.1
1.0
0.8
0.9
Yard
Wastes
2.6
2.6
2.6
2.6
2.6
2.6
2.6
*Taken from Reference 4, Vol. I, page IV-19.
-------
III. INCINERATORS
This discussion is presented in order to provide the reader with some idea
of the engineering aspects of incinerators. A more comprehensive discus-
sion can be found in Appendices H and I of Reference 4.
Figure B-l shows the annual incinerator construction over the past 30
years. It can be seen that the majority of the plants now operating in
this country were built in the late 1950's and the early 1960's.
A. Types of Incineration
Incinerators fall into two major classifications—batch feed and contin-
uous feed. Batch feeding of refuse can be done directly into the furnace
manually, or with a clamshell bucket or grapple attached to a travelling
crane; the rate of feed is controlled by the time cycle and the degree of
bucket loading. In a few plants, a front-end loader operating on a paved
floor charges the furnace; sometimes with a ram. One of the main character-
istics of a batch feed unit is that it operates in cycles—adding refuse
and burning (perhaps several times) and then dumping the grate—so that
each step of the operation recurs on a periodic basis. As may be anti-
cipated, a larger level of emissions from batch incinerators occurs during
or immediately following the charging or dumping steps (as compared to
"normal burning").
Within the continuous feed category, there are a large number of concepts—
but most are comprised of refractory chambers of rectangular construction.
A few units use a kiln following combustion on a grate to improve burn-out
and a very small number (in the U.S.) use waterwall boiler construction
(grate burning) followed by a convection boiler.
Most recent construction specifies continuous feeding of refuse to the
incineration furnace. Continuous feeding is most often accomplished by
means of a hopper and a gravity chute. A rectangular hopper receives the
refuse delivered by the crane and bucket. The bottom of the hopper termin-
ates in a rectangular section chute leading downward to the furnace grate
or other feeder conveyor at the entrance to the furnace chamber itself.
Intermittent feeding of refuse directly into the furnace is done, in most
cases, with a clamshell bucket or grapple attached to a traveling crane.
Figure B-2 presents a picture of the change in type of furnace from batch
to continuous as well as an indication of the furnace capacity. In
regards to the latter point, it should be recognized that an incinerator
facility frequently will have two or three furnaces so that incinerator
capacity varies from less than 100 tons per day (TPD) up to 1200 TPD.
B-8
-------
to
VO
Dashed Lines
May Reflect
Incomplete Data
1930 1934 1938
Source: Table 1-1
1966
FIGURE B-l: Total Annual Additions to United States
Incinerator Capacity (See Table 1-1, Vol. 1)*
1970
*Taken from Reference 4, Vol. II, page 1-26.
-------
bd
I
40Q
300-
o
X
*r
CN
O
~ 200h
«
O
s.
^
LL
100
BATCH
D
CONTINUOUS
' ' ' i ' I I i ' i
1945
1950
1955
1960
1965
1969
Plants Reporting 31 5 6 20 22 1 142 21 1 32 5 30 3 19 9 20 7 14 12 18 18 6 15 28
FIGURE B-2: Range of Furnace Capacities Batch and Continuous*
*Taken from Reference 4, Vol. II, p. 1-32.
-------
B. Important Operation Factors
1. Grates and Hearths
Nearly all incinerator furnaces employ either a stationary refractory
hearth to support the burning refuse or a variety of grate types which
stoke or mix the refuse during the combustion process in various ways
depending upon the type of grate or stoker. Suspension burning (not now
used in municipal incineration practice) would "be the only process which
does not necessarily require a hearth or grate, since most of the refuse
would be oxidized while in suspension in furnace gases. To obtain accept-
able burn-out with suspension burning, a burn-out grate should be pro-
vided at the bottom of the furnace chamber.
There are many different types of hearth or grate, each of which has its
own special features.
a. Rotary Kiln, Stationary Hearth, and Rotary Hearth
Incinerator furnace systems which operate without grates include the
stationary hearth, rotary hearth, and rotary kiln. The stationary hearth
is usually a refractory floor to the furnace. The rotary hearth, which is
seldom used for refuse, involves a rotating refractory table that turns
on a vertical shaft by means of a mechanical device operated from below.
The rotary kiln type of hearth has been used for several hundred years in
the pyroprocessing industry to move solids in and out of high-temperature
combustion zones and to mix them during combustion by rotating them. The
kiln is inclined toward the discharge end and the movement of the solids
being processed is controlled by the speed of rotation. There has been
no use of the rotary kiln for municipal incinerator furnaces except
after the burning of refuse on a multiple~grate system.
b. Grates
• Stationary Grates
Stationary grates are metal bars or rails supported in the masonry side
walls of the furnace chamber. They normally require manual stoking with
a slice bar to stir the burning bed of refuse in order to obtain reason-
ably complete combustion. This grate system is now found only in very
old or very small incinerators.
• Mechanical Grates
Mechanical grates are grates that are activated periodically to mix the
burning refuse and dump the ash. Mechanical grates are being used in
many of the newer batch fed incinerators. The grate moves the refuse
and residue from the refuse feed point through the incinerator furnace to
B-ll
-------
the discharge end of the grate, and does some stoking and mixing of the
burning materials on the grates.
c. Stokers
Incinerator stokers are mechanical devices to burn solid refuse. They
provide support and continuous or intermittent conveying of the refuse
bed (sometimes with stoking), often are supplied with most of air for
combustion, and act to discharge the non-combustible residue from the
furnace. Thus, the stoker is not only a conveyor, but a means to con-
trol the combustion process. In fact, the incinerator stoker is the
principal item for control in the incinerator furnace system. It
establishes or affects the rate of refuse feed (in the continuous flow
system), the rate of combustion, the residence time of the refuse in the
combustion chamber, the completeness of the combustion or burn-out, the
temperature of the combustion chamber, and the temperature of the ash
residue discharged from the system.
2. Furnace Enclosures
The furnace enclosure provides a controlled environment for the combus-
tion process in the incinerator system. Without the furnace enclosure,
the combustion process would, in effect, be "open burning." Materials
for conventional incinerator furnace enclosures include:
• Nonmetals (refractories, i.e., firebrick walls and roofs);
• Metal (plate, tubes, pipe, etc.);
• Refractory covered metal (castable or firebrick refractory
lining or coating, 1" to 9" thick).
There are at least three types of refractory enclosures: gravity walls
with sprung arch roofs; suspended walls and suspended roofs; and refract-
ory linings supported directly by a metal shell.
Either water or air may be used to cool these enclosure materials. Cool-
ing water can be either contained in tubes or pipes or uncontained in the
form of a film on the surface of the external metal plate. Air cooling
can be employed with forced convection; with forced air jets impinging
on the surface; or by radiation and natural convection to the ambient
environment. Warmed cooling air can be used as preheated air for com-
bustion or can be ducted for building heat.
The use of waterwall enclosures is of increasing interest. These
enclosures are characterized as waterwall boilers or waterwall incin-
erator furnaces. Such walls are cooled by forced circulation of hot
water or by natural convection as in steam boiler tubes.
B-12
-------
The physical shape of the furnace enclosure is a factor in controlling
the incineration process. For example, there is a tendency for hot com-
bustion gases to rise to the top of the chamber. If the furnace outlet
is at the top of the enclosure, the cooler gases from the discharge portion
of the grate will leave the furnace below the layer of hotter combustion
gases, often without mixing. Thus, a large fraction of the combustion
air is not used and the flue gases may contain excess air as well as
unburned combustible. This problem is reduced somewhat with chamber
designs which exhaust the gases from the incinerator furnace either at
the top near the refuse feed or at the lower portions of the furnace near
the ash discharge. The installation of overfire air jets, refractory
baffles, or bridge walls will also aid mixing.
3. Combustion Air
For incineration, the term combustion air usually includes underfire air,
overfire air, and secondary air. Underfire air is required to cool the
grates (to maintain their structural integrity and to avoid oxidation
corrosion of the grate metal), and to furnish oxygen for the combustion
reaction. It may be insufficient for complete combustion, yet sufficient
to release enough heat to pyrolyze the refuse and remove the volatile
components. Although it is common practice to pass a quantity of under-
fire air theoretically sufficient for complete combustion of the refuse
on the grates, poor air distribution vis-a-vis the distribution of air
demand, and poor gas mixing requires additional air as overfire or
secondary air for mixing and for dilution of the gases to maintain temp-
eratures below 1800°F to 2000°F, respectively.
Overfire air is usually admitted either in low- or high-velocity jets to
mix the combustible gases rising from the burning refuse with combustion
air. Secondary air, added for temperature control, may be admitted
through high-^velocity jets in the side walls and roof of the furnace
enclosure, either near the upstream end of the primary furnace or at the
transition between the primary and secondary furnace. Also, secondary
air can be added at low velocity through a slot or small openings in the
bridge wall separating the primary chamber from the secondary chamber.
In the latter case, mixing is dependent more on the shape of the chamber
and changes in direction of the main gas stream, than on the energy
carried by the air jets.
4. Flue Gas Conditioning
Inasmuch as the incineration of municipal refuse is a combustion process,
it is necessary to remove the heat that is generated. In the past, this
has usually been done with a gravity stack which, except for the heat
lost through the incinerator walls, discharged most of the combustion
energy to the atmosphere as sensible heat in the hot flue gases. The
heat of evaporation of moisture in the refuse and from cooling of flue
gases by water sprays is discharged as latent heat of the water vapor
in the flue gases.
B-13
-------
However, cooling of the flue gases to carbon steel operating temperatures
(below 600°F to 700°F) as may be necessary for several APC or stack systems,
is frequently required. It can be accomplished by one or a combination
of any of three methods:
• By direct injection and vaporization of water;
• With a heat exchanger, such as a waterwall or convection
boiler, air-cooled refractory, or an air preheater; and
• By dilution and mixing with cool atmospheric air.
5. Air Pollution Control
The emission of mineral particulates is controlled most effectively by
incineration Air Pollution Control (APC) systems which are the only means
of reducing stack concentrations of noncombustible gaseous pollutants.
APC devices are selected on the basis of a compromise between the initial
capital investment required, the cost of operation, and the desired effi-
ciency of collection. Operating characteristics, reliability, and water
pollution potential usually play lesser roles in the selection procedure.
Settling chambers are the simplest, least expensive, and least effective
of the APC systems now used to remove particulate pollutants. Basically,
this chamber slows the flue gases (no temperature reduction necessary)
to permit gravity settling of the coarse particulate. The calculated
performance of a "perfect" settling chamber (nominal dust size and
density, 30 feet high, and a 5-second residence time), is about 40%;
realistically it would probably be closer to 20%.9
Other devices used to collect particulate emissions include mechanical
collectors, such as cyclone devices, fabric filters, and electrostatic
precipitators which separate the dust particles from the effluent by
interaction of electrical charges placed upon the surface of the dust
particles. The latter is capable of very high collection efficiencies.
Fabric filters, properly designed and operated, promise to be more
effective in the removal of particulate matter than any of the conven-
tional APC systems. The particulate collection efficiency of these
devices often exceeds 99%. Also, some absorption of HCl may take
place on the alkaline fly ash material in the dust cake. By 1968, how-
ever, no filter units have been installed in U.S. incinerators.
Various types of scrubbing devices are currently used for contacting
liquids with gas streams to remove particulate matter and gaseous pollu-
tants. These include wetted surfaces, liquid sprays, and bubbling gas
devices. The wetted surface and liquid spray devices are more promin-
ently used in municipal incinerators. However, all such devices are
noted for their side effects in humidifying and cooling the gas, thus
producing a steam plume at the stack under certain atmospheric conditions.
B-14
-------
In general, gaseous pollutants will.pass unchanged through the dry control
devices, except the bag filters which have purportedly absorbed as much
as 90% of the t^SO^ values3 but little of the SC>2.6 Wet control devices,
although primarily used as particulate removal agents, will remove some
of the gaseous pollutants. However, most wet scrubber devices are less
effective than a single-equilibrium contact. Analyses have shown that
particulate scrubbers can be expected to remove—in addition to material
participates—a portion of the H2S04, HC1, S02, organic acids, and in
some cases C02. Other components—CO, NO, hydrocarbons, aldehydes, and
SO^—are essentially untouched except by scrubbing.
Figure B-3 shows the shifts in the choice of primary air pollution con-
trol equipment among incineration plants. The primary incineration control
equipment is that having the greatest effectiveness. In many plants, a
combination of sprays and wet-bottom expansion chambers are used, followed
by a higher efficiency system, such as close-spaced wetted baffles, or
a high-energy scrubber. There is a distinct trend away from the dry expan-
sion chamber systems (and "none") toward increased use of cyclones, close-
spaced wetted baffles, and scrubbers.
Modification of certain factors in the incineration process may alleviate
air pollution somewhat, for instance, the shift toward continuous units
and closing of older batch units. The batch units have considerably
higher combustible pollutant loadings, though total particulate emissions
tend to be slightly lower than the continuous plants. The charging
operation is related to air pollution and a number of other operating
problems associated with batch fed units. Figure B-4 shows, as a histo-
gram for cylindrical and rectangular units, the various responses from
operators to questions regarding charging frequencies. Although the
responses were widely variable, they appear to concentrate in the range
of 5-10 charges per hour.
Data on charging and dumping frequency for batch-feed units and on the
residence time for continuous-feed traveling grate units are summarized
in Table B-5. It can be seen that the batch-feed units require a con-
siderably longer grate residence time than continuous-feed units.
Table B-6 shows reported values for flue gas temperatures in various
parts of the furnaces, according to furnace type. Although some vari-
ability is shown, the reported values are quite similar. In all cases,
the temperatures shown are more than needed to assure complete burnout
of all combustible air pollutants. The fact that such materials do
appear at substantial concentrations in stack effluents indicates the
unreliability of the reported measurements as representative averages.
It is clear, from other data,13 that stratification occurs in furnaces,
permitting substantial amounts of colder gas to pass through the system.
The range of gas temperatures that can be expected in a stack on a muni-
cipal incinerator without wet scrubbers is shown in Figure B-5.
Clearly, in some cases the gas stream that must be sampled is still quite
hot.
B-15
-------
w
Dry Expansion Chamber
Wetted Walls
Year
Plants Reporting
1940
2 3
1945 1950
1 6 12 8 13 10 8 8
1955 1960 1965 1969
10 12 23 14 22 11 16 13 12 15 12 14 21 15 13 8 6 4
FIGURE B-3: Primary Air Pollution Control Equipment
*Taken from Reference 4, Vol, ,11, page 1^34.
-------
TABLE 5
Average Grate Residence Time
Furnace Type
Circular Batch
Rectangular Batch
Traveling Grate
Hearth (manual Stoked)
All Types
Average Grate Residence Time (min.)
116
103
61
180
94
TABLE 6
Reported Incinerator Temperatures
I. Primary Chamber
Average Temp (°F)
II. Secondary Chamber
Average Temp (°F)
III. Stack
Average Temp (°F)
Furnace Type
All
Units
1675
1288
636
Grate-
Kiln
1833
Circ. Rect.
Batch Batch Continuous Hearth
1581 1621
1725
1850
B-17
-------
I
M
00
1
8
7
S
I6
3
sr a
i- j
UL
1
i
.
"
-
-
OC
|$V
ii 1
1
18
1
i
&
1
^
—
W Cylindrical Units
Rectangular Cell Units
k
P75J
^9
i
8
XX
XX
oo
?vs
i
1 — 1
*x
Rx
11
, l, ,n, , 11 , , , , i
2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32
Charges Per Hour
FIGURE B-4: Frequency of Occurrence of Charging
Rates, Batch Feed Incineration Systems
*Taken from Reference 4, Vol, II, page 1-50.
-------
w
M
VO
20
15
o
Q.
S 10
o
100
8 Data Points > 900 F,
200
300
400 500 600
Temperature of Stack Gas (°F)
700
800
900
FIGURE B-5 HISTOGRAM OF AVERAGE GAS TEMPERATURES
IN STACK FOR MUNICIPAL INCINERATORS
(Based on Data from Appendix J, Reference 4)
-------
6. Stacks
Several types of stacks or chimneys are used to discharge incinerator
flue gases into the ambient atmosphere. Stub stacks are usually fabri-
cated of steel and extend a minimum distance upward from the discharge
of an induced draft fan. Stacks taller than about 5 diameters and less
than 100 feet high are often referred to as "short" stacks. These are
either constructed of unlined or refractory lined steel plate, or are
constructed entirely of refractory and structural brick. "Tall" stacks
are constructed of the same materials as the short stacks above, and are
used to provide greater draft than that resulting from the shorter stack,
and to obtain more effective diffusion of the flue gas effluent. Some
metal stacks are made with a double wall with an air space between the
metal sheets. This double wall provides an insulating air pocket to pre-
vent condensation on the inside of the stack and thus avoid corrosion
of the metal.
7. Ash Removal
After complete incineration of the refuse, the ash residue drops into an
ash chamber or chute from the end of the grate or kiln. Siftings that
have fallen through the grates (which may have been either partially or
completely oxidized) and collected fly ash also may be conveyed to this
ash chamber. The ash may be discharged directly into a container or onto
suitable conveyors for disposal, or into water for quenching and cooling.
The ash residue is then removed from the water with a drag conveyor,
pusher conveyor, or other means.
To prevent inleakage of air or outleakage of furnace gases at the point
where the gas is removed, an air seal is desirable. Dry mechanical seals
and seals made by covering the ash receptacle or container have been used
to control air leakage. With wet removal of the ash, a wet or hydraulic
(water) seal is utilized or a combination of a wet and mechanical seal
is used (Martin ash quencher seal).
IV. GASEOUS EMISSIONS
If combustion of the volatile fraction of the refuse is complete, the
composition of the flue gas will be principally nitrogen, oxygen, and
carbon dioxide. There will be small amounts of sulfur oxides, nitrogen
oxides, and traces of mineral acids (principally hydrochloric acid, which
will result from the combustion of halogenated plastics, particularly
polyvinyl chloride).
If combustion of the volatiles is not complete, the flue gases will con-
tain significant amounts of carbon monoxide and other uncombusted or
partly combusted organic materials. A more detailed description of the
composition of the effluent gas stream is given in Appendix A of this
report.
B-20
-------
V. PARTICULATE EMISSIONS FROM INCINERATORS
An appreciation for the factors that determine particulate emission and
for their relative importance may be gained by considering the three
mechanisms mainly responsible for the emissions:
• The mechanical entrainment of particles from the burning refuse
bed;
• The cracking of pyrolysis gases; and
• The volatilization of metallic salts or oxides.
The first of these mechanisms is favored by refuse with a high percentage
of small particle size, low density ash, by residue or residue geometry
favoring entrainment (plates), by high underfire air velocities, or by
other factors that induce a high gas velocity through the bed. The
second mechanism is favored by refuse with a high volatile content pro-
ducing pyrolysis gases with a high carbon content and by conditions above
the fuel bed that prevent the burnout of the coked particles formed by
the cracking of the volatiles. The third mechanism is favored by high
concentration of metals that form low melting point oxides and by high
temperatures within the bed.
Furnace emissions depend on many variables: refuse composition; method
and frequency of feeding; underfire air; incinerator size; burning rate
and temperature; combustion chamber design; grate type; and mixing.
A. Refuse Composition
Ash particles may be entrained when the velocity of the gases through the
fuel bed exceed the terminal velocity of the particles. Undergrate air
velocities typically vary from a minimum of 10 SCFM/ sq. ft. of grate area
to 100 SCFM/sq. ft. Based on the terminal velocity of ash particles, it
is expected that particles up to 70 y (equivalent diameter) will be en-
trained at the lowest velocities and up to 400 y at the highest.
The large volumes of combustible-rich pyrolysis gases generated during the
incineration of refuse with a high volatile content tend to result in
particulates with a high fraction of soot and other combustibles. Sooting
tendencies are particularly high for pyrolysis gases with a high carbon
content.
B. Method and Frequency of Feeding
Start-up of an incinerator is a time when particulate emissions are gen-
erally high. Therefore, a batch process, by definition, will have a
cyclic pattern with high emissions at the start of each cycle. The
amount of emissions will be highly dependent on the charge and manner of
feeding. The behavior of continuous systems also can be greatly influenced
by the rate of feeding of the refuse.
B-21
-------
C. Underfire Air Rate
A systematic study of the PHS of the effects of underfire air, second-
ary air, excess air, charging rate, stoking interval, and fuel moisture
content on the emission rate from an experimental incinerator led to the
conclusion that the velocity of the underfire air was the variable that
most strongly influenced particulate emission rate. The wide variations
in undergrate air use between different incinerators partially accounts
for the scatter seen in collections of particulata emission data.
Unfortunately, however, such engineering data are seldom measured or
reported during emission testing.
D. Incinerator Size
Increasing the size of incinerator units from 3 to over 100 TPD (tons per
day) has been shown to result in higher emission rates, but the effect
of size on emission factors has not been quantitatively established over
the more limited size range of municipal furnace units (50-300 TPD).
Part of the increase is due to the higher burning rates (Ib/hr-sq ft of
grate), and hence higher underfire air rates associated with large units.
E. Burning Rate
It is expected that higher emission rates will be encountered at higher
burning rates. Rhem2 cites that reductions in rate of burning to 70% of
rated capacity have shown as much as 30% reduction in furnace emission
from that at full capacity; however, insufficient data are available for
a quantitative relationship to be established as important variable
changes specified.
F. Combustion Chamber Design
Practically the only data in this area are provided by the studies on small
incinerators in Los Angeles County. From these data, it was concluded^
that the emission from small units could be reduced substantially by the
use of multiple chambers and by the use of a low arch. The emission in
these tests primarily consisted of particles under 5 u in size, and there
is uncertainty concerning the applicability of the results or the design
parameters developed to large municipal incinerators. The need for multiple
combustion chambers for large units, for example, will depend mainly on
the adequacy of mixing in the primary chamber.
G. Grate Type
Although the type of grate used in an incinerator does not in itself define
the emission rate of pollutants, it can be viewed as an integral variable
reflecting the effect of a number of parameters, including: specific
undergrate air flows, the stoking intensity, and the percent open air.
B-22
-------
H. Mixing Effects on Combustible Pollutants
The reduction in combustible solid and gaseous pollutants within an incin-
erator is primarily a function of the mixing processes within the system.
In all realistic cases, the bulk flue gas temperature and the calculated
average gas residence time within the furnace are more than adequate to
assure complete combustion. Therefore, insufficient mixing plays an
important role in generation of particulates.
B-23
-------
VI. REFERENCES
1. CONSTRUCTION AND TESTS OF THE FLUE GAS DUST-REMOVAL UNIT IN THE MUNICH
INCINERATOR PLANT NORTH I
M. Andritzky
Brennst.-Warms-Draft, 19 (9), 436-9 (1967).
2. INCINERATOR TESTING AND TEST RESULTS: DISCUSSION
R. L. Daugherty
J. Air Pollution Control Assoc. ]_ (2), 118-20 (1958).
3. PILOT-PLANT INVESTIGATION OF THE BAG FILTERHOUSE FOR CONTROL OF
VISIBLE STACK EMISSIONS FROM OIL-FIRED STEAM-ELECTRIC GENERATING
A. E. Gosselin, Jr.
Proc. Amer. Power Conf. 26^ 128-37 (1964).
4. SYSTEMS STUDY OF AIR POLLUTION FROM MUNICIPAL INCINERATION.
VOLUMES I AND II
Arthur D. Little, Inc.
Report to The Division of Process Control Engineering, National Air
Pollution Control Admin., U.S. Department HEW, under Contract
CPA-22-69-23 (1970).
5. MANUAL METHODS FOR SAMPLING AND ANALYSIS OF PARTICULATE EMISSIONS
FROM INCINERATORS
RFP (CPA-70-198), National Air Pollution Control Admin.
6. LIMESTONE-DOLOMITE PROCESSES FOR FLUE GAS DESULFURIZATION
A. E. Potter, R. E. Harrington, and P. W. Spaite
Air Eng. 10 (4), 22-7 (1968).
7. RESEARCH FINDINGS IN STANDARDS OF INCINERATOR DESIGN
A. H. Rose, Jr., and H. R. Carbaugh
In: "Problems and Control of Air Pollution," F.S. Mallette, Editor
Reinhold Publishing Corp., New York (1955).
8. AIR POLLUTION EFFECTS OF INCINERATOR FIRING PRACTICES AND COMBUSTION
AIR DISTRIBUTION
A. H. Rose, Jr., R. W. Stenburg, H. Corn, et. al.
J. Air Pollution Control Assoc. 8_, 297-309 (1959).
9. DUST EMISSION FROM THE HAMILTON AVENUE INCINERATOR
E. J. Schulz
Report to Sovereign Construction Co., Ltd., for the Dept. of Public
Works, New York City, by Battelle Memorial Institute (Sept. 1964).
10. EFFECTS OF HIGH VOLATILE FUEL ON INCINERATOR EFFLUENTS
R. L. Stenburg, R. P. Hangerbrauck, D. J. vonLehmden, and A. H. Rose, Jr.
J. Air Pollution Control Assoc. 11, 376-83 (1961).
B-24
-------
11. EFFECTS OF DESIGN AND FUEL MOISTURE ON INCINERATOR EFFLUENTS
R. L. Stenburg, R. R. Horsley, R. A. Herrick and A. H. Rose, Jr.
J. Air Pollution Control Assoc. 10, 114-20 (1966).
12. CHARACTERISTICS OF FURNACE EMISSIONS FROM LARGE, MECHANICALLY-
STOKED MUNICIPAL INCINERATORS
A. B. Walker and F. W. Schmitz
Proc. 1966 National Incinerator Conf., New York, p 64-73 (May 1-4, 1966)
13. COMBUSTION PROFILE OF A GRATE-ROTARY KILN INCINERATOR
P. H. Woodruff and G. P. Larson
Proc. 1968 National Incinerator Conference, New York, May 5-8, 1968,
p. 142-53.
B-25
-------
APPENDIX C EVALUATION OF EXISTING SAMPLING TRAINS
Table of Contents
I. INTRODUCTION C-l
II. DESCRIPTION AND EVALUATION OF EXISTING SAMPLING
SYSTEMS C-2
A. EPA System C-2
B. Los Angeles APCD Type System C-2
C. Incinerator Institute of America,
Procedure T-6 C-4
D. An ASME PTC Type System C-4
E. British Standard System C-6
F. WP-50 System C-6
III. DISCUSSION C-8
IV. REFERENCES c~12
C-l
-------
APPENDIX C
EVALUATION OF EXISTING SAMPLING TRAINS
I. INTRODUCTION
This appendix presents a discussion of the methodology currently in use
for sampling incinerators for particulates. In evaluating these methods,
it is important to recognize that particulate matter, as considered by
EPA for our study, was any material, except uncombined water, which
would be in the solid or liquid state at 70°F and a pressure of one
atmosphere. To be effective and useful as a monitoring method, applic-
able procedures and devices should remove a measured representative
sample of stack gas, collect the particulate matter (solid and liquid)
which exists in the stack gas, as well as the potential particulate
matter, i.e., material which would condense to an aerosol (free liquid
water excepted) as incinerator gases cool by dilution with the ambient
atmosphere. The ultimate method should make possible the collection of
particulate and potential particulate matter without the introduction of
any extraneous material which could be misinterpreted.
In addition to focusing attention on the mass of emitted particulate
matter as a measure of its effect as a pollutant, currently there is
interest in many other effects of air pollution, such as: cloud forma-
tion in the atmosphere and its relation to nuclei population; influence
on atmospheric visibility and haze by small particulates (0.1 to lym) ;
health effects of the "respirable" fraction (approximately 0.1 to 5ym) ;
and the mass of material emitted to the atmosphere in the larger size
range (>lym).
All particulate sampling systems include the following units:
• nozzle
• probe
• collecting device
• metering unit
• suction source
These units are not always independent components of the system, for
example, a cyclone is sometimes used for particle collecting and gas
metering. The design of the sampling system, both as to configuration
and size, may be dependent on the sampling procedure contemplated and
the interests of the designer. In any case, the arrangement of the
components, both as to location and sequence, is a primary feature in
differentiating systems.
C-2
-------
II. DESCRIPTION AND EVALUATION OF EXISTING SAMPLING SYSTEMS
Six systems were selected as being significant and state of the art in
their arrangement and/or application. These have been associated with
the following groups:
• Environmental Protection Agency - EPA (formerly PHS);
• Los Angeles Air Pollution Control District (Los Angeles APCD);
• Incinerator Institute of America, Procedure T-6;
• An ASME PTC Type System;
• British Standards System; and
• WP-50 System.
A. EPA System
The description of the EPA (formerly PHS) sampling system specified for
testing incinerators is presented as Appendix K in this report.
The most widely leveled criticism of this system has been directed toward
the nature of the material caught after the filter. In a series of
incinerator studies using this system, it was found that the weight of
the material collected after the filter varied from a few percent to 75%
of the toted, sample weight with an average of 30%. '•'•' In view of the low
reproducibility of the impinger catch, suspicion abounded that some of
the material found in the impinger train was produced by chemical action
in the aqueous scrubbing solution, hence was not particulate matter
under the definition of particles as utilized for our investigation.
Also, since the impinger train was in an ice bath (32°F), material may
have been formed physically (condensation) which was not consistent
with the present definition of particulate (particles which exist at 70°F).
Another potential problem with this system has been the use of a
heated glass probe in place of an all metal probe. By supporting the
glass probe in stainless steel, it is hoped to take advantage of the
fact that glass is less reactive chemically (except for HF) compared to
metal. Hopefully, continued field experience with the glass system will
reduce the problem of potential breakage to an acceptable level.
Finally, it should be noted that the earlier PHS system included the
impinger catch as part of total particulate whereas the procedure des-
cribed in Appendix K does not.
B. Los Angeles APCD Type System
The Los Angeles system^) shown in Figure C-l, is similar to the EPA system
and appears to be an earlier version of it. However, this system is
significantly different from the EPA system in that the materials of
C-3
-------
FIGURE C-l
Los Angeles APCD Type System
Schematic diagram of sampling train incorporating a miniature glass cyclone, alundum thimble
filter, and impingers. This equipment is frequently used when sampling incinerator effluent,
where particles larger than 5 microns are common. The components are: (1) sampling probe;
(2) cyclone; (3) electric heaters; (4) dry filter; (5) impinger (dust concentration sampler);
(6) ice bath container; (7) thermometers; (8) mercury manometer; (9) hose clamp; (10) vacuum
pump; and (11) Sprague dry gas meter.
*Taken from Reference 2.
-------
construction are not as well chosen from a chemical reactivity point of
view, i.e., stainless steel probe and filter holder plus rubber or tygon
tubing connections. These features could tend to decrease the accuracy
of this system.
Another Los Angeles APCD system'^) functions similarly to the one in
Figure C-l except that a filter downstream of the impingers is used to
capture particles which pass through the impingers. It is reported that
the filter collects a very small amount of material compared to the
impingers, but this would be expected since it collects only particles
which pass through the impingers plus carry-over from the impingers.
Backing up the impingers with a filter is a prudent technique in that it
prevents accumulation of carry-over in the sample lines and it protects
the pump and metering system from possible contamination. However, it
could lead to early filter plugging and necessitate shortened sampling
times.
C. Incinerator Institute of America, Procedure T-6
IIA Procedure T-6'3) is designed for large volume sampling (>5 cfm) of
dust (solid particles £ 1pm); therefore, as a system it only partially
fulfills the particle sampling requirement under the EPA definition.
The apparatus consists of a 3/4-inch diameter null (balance) nozzle,
a stainless steel probe with one section water-cooled (to protect the
filter downstream from the hot sample gas). The probe attaches to a
well-packaged stainless steel cyclone (mason jar pot) and low efficiency
bag filter. The pump is a large volume centrifugal type. This system
is shown in Figure C-2.
While the null nozzle allows rapid flow adjustment, it has small static
pressure ports which are subject to plugging. Also, a sensitive dif-
ferential pressure gauge (such as Hooke gauge) is necessary to determine
the null point. The most positive feature of this system is that the
collection stage (made by UOP Air Correction) is packaged in a convenient
package for field handling. The significant negative features are:
• Materials of construction (large opportunity for reaction with
the sample gas);
• Difficulty in recovering all the sample from the collection
stage; and
• No direct sample gas volume measurement.
D. An ASMS PTC Type System
The American Society of Mechanical Engineers performance test codes
(PTC 21-1941 and PTC 27-1957)!(4) do not specify particular equipment,
but they do outline the basic requirements for a sampling system.
C-5
-------
Pressure Tap
For Gas Metering
Y
Flow
Control
n
Centrifical
Pump
3/4" Dia
Null
Nozzle
Cyclone
FIGURE C-2
Schematic of T-6 System
-------
The comments on the IIA/T-6 system apply here with the exception that
this system does not appear to be as convenient to use because of its
increased bulk and overall poor packaging. In spite of these apparent
difficulties, many people feel that the system is reliable and easy to
use in the field.
E. British Standard System
This system^?) was developed by the British Coal Utilization Research
Association for sampling stack §3.3 from coal-fired power plants. Since
then, the system has been established as a British Standard and used for
general particle source testing in the United Kingdom and Commonwealth
countries. The apparatus consists of a nozzle leading directly to a
cyclone (used for sample gas flow measurement) which is closely coupled
to a packed glass filter. All of this equipment is located inside the
stack. A centrifugal pump outside the stack is used for suction. Stain-
less steel is the construction material throughout.
This system is designed for large volume sampling of dust; therefore, as
a system it only partially meets the requirements for particle sampling
under the EPA definition. (Large volume flow systems must sacrifice in
filter efficiency to avoid rapid plugging.) However, this system has
features that are worth discussing. The location of the collection
stages inside the stack reduces the effort in recovering sample material
from probes. It eliminates the need for an external heating source to
maintain the cyclone and filter above the dewpoint (hot stack gases).
This equipment appears to be convenient to transport, to set up, to
operate, and to recover the sample. In operation, it is simple to adjust
the flow for isokinetic conditions since the flow meter (cyclone) operates
at the same gas temperature and pressure condition as a pitot tube located
close to the nozzle. Since both the pitot tube and the flow meter are
square law devices, a simple linear calibration curve may be drawn for a
given size nozzle that relates AP across the meter to AP across the pitot
tube. The relationship is independent of the stack conditions. This
curve allows an operator to quickly and accurately adjust the flow to
isokinetic conditions.
F. WP-50 System
The WP-50 system^) ±s widely used in stack sampling and is shown
schematically in Figure C-3. In the conventional form, this system
consists of a buttom-hook-type nozzle connected directly to an alundum
thimble filter holder, all of which are located inside the stack. The
filtered sample gas is then drawn through a pipe (and frequently moisture
condenser) to a dry gas meter and a sunction source. The construction
material is stainless steel.
This system is designed for measuring dust, hence, only partially ful-
fills the requirement for particle sampling under the EPA definition.
The following positive features are noted about this system: sample
easily recovered (short sample line to collector ; and total sample gas
volume measures.
C-7
-------
Alumdum
Thimble
Filter
i-J
to
Dry Gas
Meter
o
oo
Nozzle
Vane
Pump
FIGURE C-3
WP-5Q System
-------
The potential negative features are: filter (thimble) plugging especially
in saturated gas stream; no monitor of flow rate for quick isokinetic
adjustment; reactive construction materials; and because of large weight
of thimble, large sampling volumes are required in order to collect suf-
ficient sample to overcome large tar weight of the filter.
There are two types of systems specified by the Industrial Gas Cleaning
Institute for wet scrubber testing (inlet and outlet) which are modifica-
tions of the WP-50 system, hence, most of the above discussion applies.
The significant modification to the outlet system is an electrically heated
filter thimble which reduced filter plugging due to condensed water. The
modification to the inlet system seems regressive in that a calibrated
orifice has been substituted for the dry gas meter.
III. DISCUSSION
Some general observations which are appropriate to this overall presenta-
tion are given below.
There is a wide variety of sampling systems reported in the literature
from which the above six systems have been selected and discussed individ-
ually. While they are all different in detail, they include the follow-
ing features:
• intake nozzle
• sample collection stage (s)
• sample gas metering stage (s)
• pump (includes any aspiration technique
Furthermore, they fall into only four general arrangements as shown in
Figure C-4. All systems in use today fit into the types of Figures
C-4a, b, c. It is difficult to conceive of a system of the type in Figure
C-Ad, where all the collected sample could be recovered. A system of
this type is depicted in the ASME PTC 21 and 27 but no actual device of
this type is known to be in use today.
The components which comprise the features of a sampling system are
based on what is to be measured plus features the designer deems a con-
venience. The systems which have been examined have been designed for
the following criteria:
(a) sample for solid particulate ^Ipra using large volume sampling
(M..5 to 20 cfm);
(b) sample for solid particulate :>lpm using small volume sampling
(^.5 to 1.5 cfm);
(c) sample for particles and condensible vapors using small volume
sampling .
C-9
-------
1
COLLECTION
METERING
PUMP
>_
"^
NOZZLE
NOZZLE
NOZZLE
COLLECTION
FIGURE C-4a
FIGURE C-4b
METERING
FIGURE C-4c
COLLECTION
METERING
PUMP
>_
*"
METERING
PUMP
COLLECTION
>_
-
NOZZLE
FIGURE C-4d
FIGURE C-4
Generalized Sampling Systems Configuration
C-10
-------
With the equipment designed for criteria (a), non-solid matter is ignored
and the gas metering system is designed for large flow rates (usually a
calibrated orifice or a cyclone). The flow rate must be accurately
recorded as a function of time to obtain the volume of gas sampled. The
pump must be a high volume type such as centrifugal blower or large
ejector. The equipment designed for (b) has a collection stage similar
to (a) but smaller in size, the gas metering may be the orifice/cyclone
technique (for flow rate monitoring) and/or a dry gas meter for sample
volume. The pump is usually a mechanical vane vacuum pump (>"20" Hg
static). The equipment designed for (c) is similar to (b) but has a
collection stage for condensible vapors, i.e., cooled Greenberg-Smith
impingers, and filter selection is very critical.
Sample methodology for particulate source sampling is undergoing change.
The modern definitions of particulate make different demands on sampling
apparatus and techniques than the traditional definition. Presently,
the best attempt to satisfy these new demands is the EPA system.
The EPA system is certainly not the definitive answer to modern, incin-
erator, particulate source sampling; however, it does have the following
positive features:
• Consideration of collecting total particulate; and
• A system design concept.
With the exception of the Los Angeles system, none of the other systems
evaluated attempted to collect condensible vapors; in fact, most of
these systems were not designed to collect particles less than one micro-
meter. None of the other systems showed overall engineering design.
Most sampling systems have the appearance of being assembled and not
designed. The EPA system has obviously been designed. Some consideration
has been given to the chemical reactivities of the construction materials
and packaging for ease of operation and installation. However, it is
not clear that this system is able to measure particulate matter under
the EPA definition.
The most significant questions to be resolved about the EPA and other
like systems are:
• Is any material being formed in the collecting system? When the
impinger residue is included in the particulate weight, there has
been an opportunity for chemical reaction in the impinger to
form solids after collection. Therefore, this could be considered
false particulate and lead to measurement errors.
• Is there an inconsistency between the measurement method and
the desired definition of particulate when the sample is cooled
to ice bath temperature, returned to room temperature and
weighed?
C-ll
-------
The pretreatment of impinger residue prior to weighing calls for extrac-
tion, evaporation under heat, and drying via desiccation. Such a com-
plicated procedure is subject to many potential errors.
Particulate matter present in the stack can be collected at high effi-
ciency with many of the present sampling systems as is or with appropriate
modification. However, the present methods advocated for the collection
of condensibles can not be expected to determine quantitatively the con-
densibles which would form in the atmosphere since the conditions are so
different. In fact, to simulate the atmospheric composition, temperature,
relative humidity and radiation in a sampling system is unreasonable.
Therefore, measurement of total particulate matter which includes con-
densibles should be interpreted as "potential" total particulate.
This review of the present state of the art and the instrumentation used
in incinerator sampling has indicated to us that alternate and more
sophisticated means of collecting and evaluating condensible material
merits further study.
C-12
-------
IV. REFERENCES
1. AN EVALUATION OF SEVEN INCINERATORS
W. C. Achinger and I.E. Daniels
Proceedings of the National Incinerator Conference (1970), p. 32-61.
2. SOURCE TESTING MANUAL
Air Pollution Control District, Los Angeles County, Los Angeles
(1963).
3. INCINERATOR TESTING
Bulletin T-6, Incinerator Institute of America, New York, New York
(August, 1968).
4. DETERMINING DUST CONCENTRATIONS IN GAS STREAMS—Performance Test
Codes, PTC 27-1957, The American Society of Mechanical Engineers.
5. SIMPLIFIED METHOD FOR MEASUREMENT OF GRIT AND DUST EMISSIONS
FROM CHIMNEYS
British Standards Institution, London, B.S. 3405 (1961).
6. BULLETIN WP-50
Western Precipitation Division, Joy Manufacturing Company, Los
Angeles, California.
C-13
-------
APPENDIX D
TABULATION OF SAMPLING RUNS
A complete list of all sampling runs conducted under this program is given
in the following tables. These tables are grouped by the objective of the
runs rather than sequentially by date of sampling. The relation of the
objective of the sampling runs to the remainder of this report is indicated
below. Even though all of the runs did not generate samples which were
used for detailed characterization, the total series is tabulated for
reference purposes. Also, material collected during individual sampling
runs sometimes were, used for several purposes, so that data collected dur-
ing the experimental program from that run may be reported in more than
one place.
Discussion of Data
Objective of Main Report
Table No. Sampling Run Section No. Appendix
D-l Become familiar with EPA sampling 4 E
train and collection of samples for
general use
D-2,3 Study filter efficiency under — E,G
freed conditions and collection
of samples for general use
D-4 Evaluation of a commercial sampl- 3 E,F
ing train under field conditions
and collection of samples for
general use
D-5 Examination of filter catch — E,H
directly on the filter using x-ray,
optical, and scanning electron
microscopy, and collection of
samples for general use
D-6,7 Collection of samples for 4 F
chemical characterization and
quantitative analysis
D-l
-------
Table D-l
EPA System Familiarization Experiments
Run No.
NI-1
NI-2
NI-3
NI-4
NI-5
NI-6
f NI-7
NJ
NI-8
NI-9
NI-10
Filter
Date Type
4/8/71 Glass
4/14
4/15
4/15
4/20
4/21
4/21
4/22 "
4/22 "
4/22
Dry Gas
Volume
(m3)
1.44
1.15
1.69
2.04
1.61
2.24
1.46
1.62
.1.65
1.67
Sampling
Time
(min.)
60
40
64
80
60
80
60
60
60
60
Impinger Filter
Stack Filter Outlet Weight
Temp. Temp. Temp. Gain
"Ply (°F) (°F) (mg)
191
336
174
391
283
260-280 260-280 70-100
320-280 280 77-84
300-380 280 75-130
380-400 280 75-80 263
380-400 280 84-102 359
-------
Table D-2
Filter Leakage Experiments
Run No. Date
NI-11 4/27/71
NI-11A
-12
-12A
-13 5/4/71
-14
-14A
-31 6/18/72
-31A
-32
-32A
Filter
Type
Glass
ii
ii
ii
ii
ii
11
ii
ii
ii
Dry Gas
Volume
(m-3)
1.65
2.78
1.50
2.65
0.91
0.86
0.89
0.86
0.82
0.82
0.84
0.88
Sampling
Time
(min.)
60
60
60
60
60
60
60
60
60
60
60
60
Stack
Temp.
(°F)
280-335
300-335
320-360
320-360
320
320
340-355
340-355
240-340
260-340
320-360
320-360
Filter
Temp.
(°F)
280
285-295
280
280-320
280
270-280
280
270-280
260-290
280-320
260-290
240-270
Impinger
Outlet
Temp.
(°F)
77-110
90-105
65
65
95
85
-------
Table D-3
Filter Leakage Experiments
a
Run No.
Test 1
NI-15
NI-15A
NI-16
NI-17
NI-17A
Test 2
NI-18
NI-18A
NI-19
Test 3
NI-20
NI-20A
NI-21
Test 4
NI-22
NI-22A
NI-23
NI-24
NI-24A
Test 5
NI-25
NI-25A
NI-26
NI-27
NI-27A
Date
5/17/71
Filter
Type
Glass
it
5/7/71
it
it
5/13/71
it
it
5/18/71
5/19/71
Dry Gas
Volume
0.91
0.85
5.55
0.91
0.85
0.90
0.83
2.45
0.91
0.86
3.43
0.85
0.83
4.27
0.88
1,02
0.85
0.89
3.44
0.44
0.80
Sampling
Time
(oin.)
60
60
90
60
60
180
Filter
Temp.
300
300
300-330
260-280
260-280
270-300
250-270
250-290
240-260
250-270
220-270
250-290
Impinger
Outlet
Temp.
60
60
240
60
60
300-340
300-340
300-340
320
320
280-320
260-280
350-330
250
250
70-75
65-90
65-70
65-70
85-90
60-65
60
60
175
60
60
280
280
220-310
280-310
280-310
250-260
230-265
240-280
250-260
240-260
65
65-80
75
60
60
180
45
45
300-340
320-340
300-340
280-380
280-320
240-270
260-290
265-275
270-290
260-290
65-71
75
270
-------
Table D-3
Filter Leakage Experiments (Continued)
Impinger
Filter Dry Gas Sampling Stack Filter Outlet
Run No. Date Type Volume Time Temp. Temp. Temp.
(m^) (min.) (°F) (°F) (°F)
Test 6
NI-33 6/22/71 Glass 0.87 60 300-380 260-290 75
NI-33A " 0.84 60 300-380 260-290
NI-34 " 3.03 180 300-380 260-290 70-80
NI-35 " 0.86 60 380-440 260-280 80
NI-35A " 0.90 60 380-400 260-280
Test 7
NI-36 6/23/72 " 0.84 60 320-350 270-275 65
NI-36A " 0.93 60 320-350 380-390
NI-37 " 3.87 150 300-380 270-290 75-80
NI-38 " 0.84 60 350-360 280-290 75
NI-38A " 1.03 60 350-380 270-300
-------
o
Table D-4
Sampling System Evaluation Tests
No. Date
NI-39 7-7
-40
-41
-42 7-15
-43
-44 7-22
-45
-46
Dry Gas
Volume
(meter^)
1.69
1.69
1.58
2.35
2.55
2.47
1.49
2.13
Sampling
Time
(min . )
65
65
87
96
105
100
60
93
Average
Stack
Temp.
(°F)
270
310
330
250
285
300
330
400
Outlet
Temp.
(6F)
95
85
95
85
85
85
80
75
Filter Weight
Gain
(8)
«._
—
989*
563
551
—
271
563
*This weight includes the filter catch from runs 39, 40 and 41.
-------
Table fl-5
Filter Catch Samples
Collected for Examination Via Microscopy
No.
NI-47
-48
-49
-50
-51
-52
-53
-54
-55
-56
-57
-58
-59
-60
-61
-62
*
Collection Media
Organic Membrane
n ii
ti M
Silver Membrane
M ii
n
Organic Membrane
n n
n n
n n
ii n
n
n ii
.. n
ii
Silver Membrane
Weight
of Catch
8.6 mg
4.0
7.0
AA
AA
**
0.1
32.6
31.6
32.7
59.7
2.0
1.6
1.1
A*
Sampling
Time
5 sec
10 sec
5 min
10 sec
+
+
+
+
5 sec
7 sec
15 sec
22 sec
5 min
Purpose
X-Ray Analysis
Optical Microscopy
Wet Chemical Analysis
Optical Microscopy
* Organic Membrane - esters of cellulose, Millipore, SSWP04700 (3um pore size)
Silver Membrane - Flotronics, MF-47-5 (5ym pore size).
** Silver membrane filters were not tared.
+ Sampling was continued until pump pressure reached 20" Hg.
D-7
-------
Table D-6
Samples Collectedfor Quantitative Analysis
o
oo
Run No.
Continuous
NI-69A
-69B
-69C
-69D
Date
Grate Incinerator
9/23/71
Dry Gas
Volume
0.17
0.06
0.08
0.25
Batch Incinerator
BI-70
-71
-72
-73
-74
-75
-76
8/31/71
9/14/71
9/16/71
9/21/71
1.28
1.46
1.50
1.50
1.44
1.42
1.71
Dry Gas
Volume
Sampling
Time
Average
Stack
Temp.
Average
Outlet
Temp.
Filter
Weight
Gain
11
4
8
22
70
70
70
70
60
70
80
>310
>310
>310
>310
460
390
395
405
425
390
410
90
70
75
75
85
75
80
278
168
191
200
293
197
248
-------
o
No.
NI-77
MI-78
NI-79
NI-80
BI-81
BI-82
BI-83
BI-84
NI-85
NI-86
NI-87
NI-88
Table D-7
Samples Collected for Chemical Characterization
Date
10/14/71
11/9/71
12/8/71
Filter*
Type
Membrane
Glass
Membrane
Glass
Membrane
Membrane
Glass
Glass
Glass
Glass
Glass
Glass
Dry Gas
Volume
(m3)
0,57
0,59
0.53
0.59
0.56
0.57
0.57
0.67
1.34
1.37
1.27
1.53
Sampling
Time
(min . )
33
28
44
32
15
28
27
29
60
60
60
60
Stack
Temp.
(°F)
320-450
400-720
620-700
425-710
280-350
360-420
420-460
360-420
425-485
400-425
410-430
420-430
Filter
Temp.
(°F)
195-205
230-295
190-210
245-300
210
180-210
180-240
215-240
220-248
210-250
200-240
210-270
Impinger
Outlet
Temp.
60-70
60-80
60-75
60-70
<50
<50
<50-60
50
50-120
50-70
50-65
60-75
Filter
Weight
Gain
(mg)
50
153
124
108
31
44
74
103
377
232
261
269
*Membrane - Millipore SSWP 3y.
Glass - 1106 BH.
-------
APPENDIX E
QUALITATIVE CHARACTERIZATION OF PARTICULATE CATCH
Table of Contents
Page
I. FILTER CATCH E-l
A. Elemental Analysis E-l
B. Inorganic Compound Identification E-4
C. Identification of Organic Material E-5
D. Development of Approach for Quantitative
Analysis Scheme E-8
E. Conclusion E-8
II. IMPINGER RESIDUE E-12
E-l
-------
APPENDIX E
QUALITATIVE CHARACTERIZATION OF PARTICULATE CATCH
A major focus of this program has been the development of a characteriza-
tion approach which will generate useful and meaningful data about the
character and composition of the particulate catch. A wide variety of
chemical and physical techniques were explored to learn whether they
could be of assistance in this task. Figure 2 in the main report (repro-
duced as Figure E-l here) provides an overview of these techniques. In
this appendix, a discussion is provided on the qualitative character for
each portion of the particulate catch. The results of our quantitative
studies to determine elemental composition is given in Appendix F.
I. FILTER CATCH
Although a few studies were made on the probe washings, most of our
qualitative studies of the mineral particulate was done on samples of
filter catch, therefore, this section covers results associated with both
types of particulate catch.
A. Elemental Analysis
Using x-ray fluorescence (XRF) and emission spectroscopic techniques, the
"classical" particulate catch was examined in an effort to learn the
identity of the components of that catch. The results of these studies
whereby the sample was collected on glass fiber filters is shown in
Tables E-l and E-2.
The x-ray fluorescence and emission spectrographic methods are compli-
mentary and yield good agreement for the major elemental species. The
particulate that collects in the nozzle, probe and cyclone has essential-
ly the same chemistry as the filter particulate.
Relative Concentration
Levels ^__ Element
High Pb,Zn,Cl,Si,Ca
Medium Sn,K,S,Pb
Low Br,Cd,Cu,Fe
Al,Mg and Na are not detectable by the XRF method employed, but emission
spectrographic analysis has shown the presence of these three elements.
E-2
-------
TABLE E-l
Results of X-Ray Fluorescence Analysis
(S - Strong; M
Run No.
A.
B.
C.
NI-3-1
1
2
3
4
9
10
Ave.
NI-3-2
1
2
3
4
9
10
Ave.
NI-3-3
1
2
3
9
10
Br
- Filter
__
M
W
W
—
—
W
Ca
and
M
W
M
M
M
M
M
Cd
Loose
W
W
—
W
W
W
W
- Acetone Washings
W
W
M
W
W
W
W
S
S
S
S
S
S
S
- Impingers
W
M
M
W
M
W
W
W
W
M
W
W
W
W
W
W
W
(Water
w^
—
—
—
—
Cl
Cu
Fe
- Medium; W - Weak)
K
Mo Pb
s_
S:
Particles
S
S
S
S
S
S
S
from
M
M •
S
S
M
M
M
__
—
—
W
W
W
W
Probe,
W
W
W
W
—
—
W
«•_
W
W
W
—
—
W
Nozzle
W
W
W
W
W
W
W
M
M
M
M
M
M
M
and
M
W
W
M
W
W
W
M
S
__ g
S
S
— — S
S
Cyclone
W S
W S
S
W S
**._ s
S
W S
M
M
M
M
M
M
M
W
W
W
M
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Residue)
__
W
M
—
—
__
W
—
—
—
—
—
—
—
W
W
W
W
W
.__
—
W
—
W
S
M
S
S
S
—
—
—
—
—
Sn
M
M
M
W
M
W
M
M
M
W
W
W
W
W
Sr
W
W
W
W
W
M
W
W
Ti
W
M
M
W
M
M
W
M
Zn
S
S
M
S
S
S
S
S
S
S
M
S
W
W
M
W
W
Ave
M
W
W
W
W
-------
TABLE E-2
Results of Qtialitative Emissions Spectrographic Analysis
w
Percent
Present
10
3
1
0.3
0.1
0.03
0.01
0.003
0.001
0.0003
0.0001
10
3
1
0.3
0.1
0.03
0.01
0.003
0.001
0.0003
0.0001
- 100
- 30
- 10
- 3
- 1
- 3
- 0.1
- 0.03
- 0.01
- 0.003
- 0.001
- 100
- 30
- 10
- 3
- 1
- 0.3
- 0.1
- 0.03
- 0.01
- 0.003
- 0.001
Sample Type and Number
Filter
(NI-3-1)
__
Ca,Si
Zn,Na
Al,Pb,Ti
Fe,K,Mg,P
Ag,B,Cu,Cr,Sb,Sn
Cd.Mn
Ba,Ni
Mo
Ga,Sr
Ge
NI 10-1
iur LJ
Ca,Si,Zn
Na,Ti,K
Fe,Pb
Al,Mg,P,Sb,Sn,Cu
Ag,Cr,Cd
B,Mn
Ba,Ni
Mo
Ga,Sr
Ge , Be
Probe Washings
(NI-3-2)
Ca
Si
Fe,K,Ti,Na,Zn
Al,Mg,Pb
P
Ag,Ba,Cu,Cr,Ni,Sn,Sb
B,Cd,Mn
Sr
Co, Mo
Ga,V
Ge
NI 10-2
Si,Ca
—
Fe,K,Ti,Na,Zn
Al,Mg,Pb
P
Ag,Ba,Cu,Cr,Ni,Sb
B,Cd,Mn,Sn
Sr
Co, Mo
Ga,V
Ge,Be
Impinger Residue
(NI-3-3)
__
Al
B,Ca
Cr,Si,Na
Zn
Fe,Ni,Pb
Cu,K,Mn,Sn,Ti
Ag,Ba,Cd
—
—
—
NI 10-3
Al
B,Ca
Cr,Si,Na
Zn
Fe,Ni,Pb,Sn
Cu,K,Mn,Ti
Ag,Ba,Cd
—
—
—
All elements below the .03-3 range should be
considered as minor constituents.
-------
Unfortunately, although more field engineering studies have been done with
glass fiber filters, they present a serious interference problem when
attempting to identify trace inorganic components. To overcome this back-
ground problem, samples were collected on cellulose ester fibers, which
have a very low background. The results of this analysis, including com-
parisons of x-ray line intensity, provides an indication of elements
present and relative metal concentration.
Line
Sample No. Intensity Element
NI-47-1 Strong Pb.Zn
Medium Sn,Br
Weak K.S.Fe.Cu.Cl.Cd
NI-47-1 Strong Zn
(Combusted) Medium Sn,K
Weak Ca,S,Fe,Cu,Cl,Pb
NI-49-1 Strong Pb,Zn
Medium Sn
Weak K,Si,Fe,Cu,Cd,Br
In comparison to the catch on glass fiber filters, there appears to be a
lower level of K, Ca, S and probably Cl. Combustion of the cellulose
ester filter apparently drives off most of the Pb and Br. The other
elements do not exhibit much change in intensity.
(Note: These x-ray fluorescence scans were obtained from 1/4-inch dia-
meter pieces cut from the center and periphery of filter Sample NI-43-1.
The two sample pieces yielded identical results, indicating a uniform
distribution of elements over the 4-inch diameter filter.)
Finally, to obtain a sense of the amount of organic matter present in the
sample, a 1" square sample of filter NI-27-1 was analyzed for total carbon
via microchemical determination. A maximum of 3% carbon (both organic and
inorganic) was found by this approach—indicating a low level of organic
matter.
B. Inorganic Compound Identification
X-ray diffraction (XRD) patterns were obtained from several of the heavily-
loaded filters, including NI-27, -32, -35, -38, -42 and -43. The patterns
were all quite similar. With the exception of NaCl, which was occasionally
present, the patterns, although rich in lines, were not solved with regard
to identity of compounds present. To aid in this study, a portion of NI-
43-1 was heated to 800°C for 15 minutes after which the sample changed color
from the original jet black to a pale straw color. A diffraction scan of
the heated sample yielded a few lines which were identified as Zn2SiC>4. How-
ever, because the sample was on a glass fiber filter, it was impossible to
determine whether the silicate phase came from a reaction with the glass
fiber filter or from silica in the sample. Therefore, for comparison to
these studies, a cellulose ester filter (NI-47-1) was analyzed by XRD. This
E-5
-------
yielded the presence of NaCl and the characteristic, unidentified diffrac-
tion pattern observed previously. An ashed sample from this filter exhibited
a pale yellow color and provided an x-ray diffraction pattern that was
identified as 65% Z^SiC^ - 35% Zn2TiO^. The formation of the silicate
phase upon heating to 800 C is thus confirmed and apparently was not re-
lated to the presence of a glass fiber filter. The presence of medium
levels of K and Sn in this ashed sample as determined by XRF (see above)
is not accounted for in the diffraction results; it is concluded that there
is sufficient silica to form a SiC^-I^O-PbO non-crystalline glassy phase.
To help in the identification process, infrared analyses were made for
organics on the filter catch and probe washings for several runs: NI-1,
2, 3, 9 and 10. The filter catch was sampled by lightly tapping the
filter to dislodge about 1 mg of particulate which was then analyzed using
KBr disc techniques. It is recognized that any condensed liquid droplets
on the filter would not have been included in the infrared sample by the
sampling procedure that was used. The results were consistent for all
samples and only gave an indication for the presence of sulfate in all of
the probe washings and filters. From these results, it is unlikely that
organic material was present in the samples at concentration levels
greater than 10-20%.
Incinerator samples 9-1, 9-2, 9-3, 2-1 and 2-2 were examined by TGA
and DTA (TGA - thermal gravimetric analysis; DTA - differential thermal
analysis) in an effort to provide information about the inorganic material
through thermal decomposition or dehydration at specific temperatures.
Ideally, one would hope to observe successive weight losses correspond-
ing to loss of water, C02, SC^, etc. The results of the thermal analyses
were essentially negative in terms of providing a really useful means for
providing quantitative data. In the case of filter sample 2-1, the TGA
results showed no weight change (when heated in nitrogen) over the range
50° to 1000°C. The probe and cyclone residue sample (2-2) showed a small
weight gain between 300-400°C followed by a gradual weight loss. There
was no net weight change after heating up to. 1000°C. The same sample
was run in air and showed a gradual weight loss from 50-450°C (correspond-
ing to approximately 30% of the sample weight); this loss could not be
specifically related to a phase transformation such as dehydration. The
DTA analyses on each of the samples showed only minor bumps and wiggles;
there was no reproducible endotherms or exotherms that one could inter-
pret in a meaningful manner. Therefore, it was not possible to identify
specific compounds in the samples by DTA or TGA.
C. Identification of Organic Material
To better understand the organic character of the particulate, two sets
of filter and probe washings were extracted with organic solvents and
examined for their organic material content by infrared spectrometric
analysis. To do this, the filter samples were extracted sequentially
with 20 ml portions of hexane, then chloroform, and finally methanol.
Each solvent extract was evaporated to dryness under nitrogen and sub-
jected to IR analysis via the KBr disc technique. (A similar approach
was used for the residues from the probe washings but due to smaller
sample sizes, the solvent volume used for extraction varied from 2 to
25 ml.) The results are shown in Table E-3. Clearly, only small portions
of these samples are organic in nature and soluble in these solvents.
E-6
-------
NI-5-1 Filter 283
w
NI-9-1 Filter 263
NI-5-2 Probe 23
NI-9-2 Probe 14
TABLE E-3
Examination of Filter and Probe Washings for Organics
lexane Chloroform Methanol Hexane Chloroform Methanol Solvent Extract Found
0.6 0.8 88 0.2 0.3 31
0.3 0.4 87 0.1 0.2 33
0.2 0.07 1.9 1 0.3 8
0.1 0.06 2 0.8 0.5 15
Hexane - Silicone
Chloroform - Silicone,
Hydrocarbon .
Carbonyl
Methanol - •Ammonium
Hexane - Hydrocarbon
Chloroform - Not Run
Methanol - Carbonyl
(possibly
acetone
residue) and
Hydroxyl
-------
The filter catch of samples NI-2-1 and NI-9-1 was examined by high
resolution mass spectrometry (HRMS) in an effort to gain more insight
into the composition of the material collected on the filters. The
samples were obtained by tapping the filters to obtain sufficient mater-
ial (substantially less than 1 mg) for testing. The samples were placed
in a small glass capillary and introduced into the direct insertion
probe of the mass spectrometer where the temperature of the sample was
gradually increased from about 70°C to approximately 250°C. In effect,
this procedure causes a fractional distillation of multi-component
samples. Spectra are recorded for anything present in the sample which
could exhibit a vapor pressure of at least 10-7 torr at these tempera-
tures .
For each of the samples, spectra were recorded for a low temperature
(80-150°C) and a high temperature (170-250°C) region. The photoplate
data were reduced by a computer system, and an elemental composition list-
ing of all the species observed in the spectra was generated. The com-
puter was allowed to search for all combinations of carbon and hydrogen,
including up to one nitrogen atom and four oxygen atoms.
Very little material was observed to vaporize from these samples, consis-
tent again with our other analyses. These results indicate that most of
the filter catch is inorganic material. The predominant vapor species
observed in all four spectra (two samples at two temperatures) were
typical hydrocarbons, both aliphatic and aromatic hydrocarbon species
being represented. At the higher temperature (170-250°C) ZnCl£ and
ZnClBr were observed for Sample 2-1. HBr, apparently due to decomposi-
tion of a bromide, was also observed at this temperature. There was an
occasional indication of some nitrogenous species in both samples.
In Sample 9-1, an occasional aliphatic nitrogen specie was observed
which could be indicative of the presence of some proteinacious-type
material on the filter. One of the prominent species in all instances
was diethylphthalate. Phthalates as a whole are incidious impurities
in most laboratory samples that can generally be attributed to plastic
containers, tubing, etc. In this instance, there is some difficulty in
ascribing the source of the phthalate since, to the best of our knowledge,
the filter paper was always handled so as not to be contaminated from
plastic containers and the sample was obtained by direct transfer from
the paper onto aluminum foil and into the glass capillary. Our experience
leads us to believe that the plasticizer was an impurity; however, the out-
side chance does exist that it was condensible species actually collected
on the filter from the incinerator. There certainly was no shortage of
phthalate plasticizers in the plastic materials being burned in the incin-
erator. Several trace aromatic oxygenated species were also observed in
the spectra.
One group of species expected to be present in the particulates, due to
their high boiling point, were the polynuclear aromatic hydrocarbons.
There was no evidence of these material in any of the samples. The
E-8
-------
experimental technique employed in these measurements had a detectability
for polynuclears in the ppra range. If the PNA's were present in the filter
catch, they existed in the ppm concentration range.
D. Development of Approach for Quantitative Analysis Scheme
The results of the preliminary approach to a quantitative analysis scheme
are discussed briefly in the following paragraphs.
Several samples were collected at the continuous grate incinerator facility
using organic membrane filters (millipore SSWP 04700). Because only 33 mg
of particulate was collected on each filter, two were combined and treated
as follows: (two blank filters were treated at the same time in an
identical manner).
The two filters (NI-54 and NI-56) were combined and leached with 50 ml of
hot water for 10 minutes. The leach liquor and subsequent hot water
washings were filtered through a dry filter into a 100 ml vol. flask.
This filtrate was cooled and diluted to the mark (the solutions were coded
FC for sample and FB for blank). The filters were fused in platinum
crucibles with 6 g sodium carbonate. Then the fusion was dissolved in
water and nitric acid, filtered* and diluted to 100 ml (these solutions
were coded FC/F for sample and FB/B for blank).
The two leachate solutions (FB and FC) were analyzed for R203 sulfate,
chloride, acidity and selected cations. The dissolved fusion mix was
analyzed for sulfate, R2^3» ari^ selected cations (see Table E-4). In
addition to this data, a high efficiency glass filter containing catch
material (NI-45-1) was analyzed for moisture and ash content of the
filter catch. The original sample was very hydroscopic. However, loss
at 105°C (moisture) was 2%, and loss after ignition at 800°C was 25% of
the drierite dried catch. This represents a residual ash content of 75%.
Decomposition of nitrates, halides and others could represent the dif-
ference, but the composition of the volatile is unknown. A blank filter
showed less than 1% weight loss at 850°C.
E. Conclusion
The filter catch seems to be almost totally mineral type particulate.
There does not appear to be much free acid (HC1 or I^SO^ nor is there
much sulfate. R203 type compounds (Fe, Al, Ti, Zr, Cr, etc.) and silica
seem to account for much of the sample. A composite picture of the filter
catch based on these preliminary studies is given in Table E-5.
*At this point, there was no residue in the blank but the sample con-
tained a trace of reddish brown residue similar to
E-9
-------
TABLE E-4
Chemical Analysis of Filter Catch
Nature of Sample
2 blank filters NI-54 32.6 mg
(millipore SSWP04700) NI-56 32.7 mg
Hot Water Leach
Acidity
Sulfate
Chloride
R203
Cations
Ca
Mg
K
Na
Pb
Zn
Units
meq
mg
mg
mg
<0.05
<0.01
<0.01
<0.01
<0.05
<0.01
<0.01
0.3
1.3 (0.03 meq)
10.6
0.2
0.03
20.5
3.0
0.3
9.3
Filter Fusion
Sulfate
R203
Cations
Pb
Zn
mg
mg
<0.05
<0.05
33
1.3
3.5
*Units - mg = milligrams; meq = milliequivalents.
E-10
-------
Volatile Matter
Loss @ 105°C
Loss @ 850°C
Acidity
Anions
Sulfate
Chloride
Non R£03Cations
Potassium
Sodium
Zinc
Lead
Other (Ca, Mg)
Total
TABLE E-5
Composition of Filter Catch
Glass
Membrane
SiO? (Acid insol)
Total
*Units - meq = milliequivalents;
45 47 54/56
Units*
-------
II. IMPINGER RESIDUE
Most of the data reflecting on the composition of the impinger catch is
given in Appendix F. However, a few qualitative examinations were con-
ducted early in the program to help guide our quantitative analysis. The
results of x-ray fluorescence analyses on the residue from the impinger
catch was given in Table E-l. It is clear that this material was quite
different in composition from the probe and filter catch. In fact, the
high sulfur values, along with the low pH, was an early clue that sulfuric
acid may be a predominant specie. Emission spectrographic examination
of the impinger residue also showed differences from the probe and filter
catch (see Table E-2). For the material collected, aluminum, boron,
calcium, chromium, silicon, sodium and zinc seemed to be most prevelant.
(This does not include HC1 trapped but lost during evaporation step
to obtain impinger residue.)
Using infrared techniques, the organic extract from the impinger solutions
was identified as containing hydrocarbons, oxidized hydrocarbons, car-
boxy lie acids, ammonium ion and sulfate. (As the infrared spectra was
obtained via the KBr disc technique, any free sulfuric acid would be
converted to KHSO^ or i^SO^.) The relative amounts of organic extract-
ibles from the impinger solutions was extremely small so that it was not
deemed necessary to follow-up on these findings and make positive identi-
fication of these trace organic components in the impinger catch.
In addition to the above qualitative examination, a number of samples of
impinger catch were analyzed quantitatively during the preliminary phases
of this program. A composite tabulation of these results is given in
Table E-6.
From this data, it was concluded that the impinger catch was probably
mineral acid with very little traditional mineral particulate.
-------
TABLE E-6
Analysis of Imptnger Solutions
a,b
Total
Calcium, Residual Solids (mg)
After
Ignition
3 800°C
Code
NI-12
-13A
NI-181
NI-18A
NI-6
NI-7
NI-20
-20A
pi
i
£ NI-30
Nt-34
NI-39
NI-40
Sample Ion Cone.
Description (men)
After evaporation
to dryness
After evaporation
to dryness
After evaporation
to dryness
After evaporation
to dryness
After concentration 5.4
to 100ml
After concentration 3.0
to 100ml
After concentration 1.7
to 100ml
After concentration 1.5
to 100ml
After concentration
to -v.!0ini
After concentration
to MOml
As Received 4.5
After concentration 4.05
to 100ml
As Received 6.54
After concentration 6.85
ppt'n Titration . Chloride
(»;>) (ma) (taeql ma
15 0.31 0.9
8.9 0.25 0.5
11.6 0.24 0.5
16.8 0.35 1.2
37.8 37.9 0.81 188
26.3 27.6 0.6 101
12.7 21.8 0.27-./.5 71
6.2 16.3 0.1 -0.3 50
—
—
—
15.0 — 0.32 173
—
23.3 -- 0.48 272
(roeq)
0.03
0.01
0.01
0.03
5.3
2.8
2.0
1.4
—
—
—
4.9
—
8.0
Total 23 3 Potassium Over
iSSal (mp) (raft) (m;0 Drierite
0.4 26
0.1 23
0.9 20
0.2 31
6.1 1.2
3.4 5.6
2.3-2.5
1.5-1.7
0.9
0.5
._
5.2 3.0 — — 35
—
8.5 3.4 — — 31
1.3
2.4
to lOOmJ
NOTES
a - By titration with 0.01N NaOH to phenolphthalein end-point.
b - Meq = milliequivalents. One meq is the weight in milligrams of the substance (chloride or sult'ate) wnicn will react wit,.
one mole of hydrogen ion. This is a convenient way to put cations and anions on an equivalent basis.
c - R203 is an indicator of trivalenc and tetravalent cations wuch as Al , Fe , Ti , etc.
d - XH3 determined by ticration with HC1 after distillation from NaOH.
-------
APPENDIX F
QUANTITATIVE ANALYSIS OF PARTICULATE CATCH
Table of Contents
Page
I. PRELIMINARY STUDY F-l
II. COMPREHENSIVE ANALYSIS OF PARTICULATE CATCH F-3
A. Experimental Procedure F-3
B. Results and Discussion F-8
1. Catch in Probe and Cyclone F-8
2. Filter Catch F-8
3. Comparison of Probe/Cyclone and
Filter Catches F-17
4. Impinger Catch F-20
III. GENERAL COMMENT F-20
F-l
-------
APPENDIX F
QUANTITATIVE ANALYSIS OF PARTICULATE CATCH
In order to confirm the overall description obtained from the qualitative
examination, tests were run with the objective of obtaining quantitative
data on the exact composition of the particulate catch.
The basic goal of these quantitative studies was to learn exactly the types
and relative amounts of the chemical substances that were being collected
in the sampling train. To do this required collecting as large a sample
as reasonably possible and then using analytical techniques which permitted
a maximum of information with a minimum consumption of the collected sample.
Unfortunately, many techniques, especially the so-called wet chemical
methods, require a fairly large amount of sample in order to obtain reliable
quantitative results. This is especially true when precipitation methods
must be used.
In any case, the samples were analyzed using a variety of techniques,
recognizing the trade-offs required for obtaining quantitative data on
major components at the sacrifice of not being able to analyze for other
substances present at lower concentrations.
I. PRELIMINARY STUDY
As a first step in this task, larger-than-normal samples of impinger catch
were obtained and analyzed. The impinger catch samples were:
Original Source Combined
Code Volume (ml) Volume (ml)
NI-37-3 930 930
NI-41-3 540
NI-43-3 690 JU
NI-45-3 500
NI-46-3 675 °
These solutions were evaporated in open beakers on a hot plate to slightly
under 100 ml and then diluted back to that volume with distilled water.
All subsequent analyses were made on aliquots of these solutions. Anal-
yses included: acidity (titration), sulfate (precipitation), R203 (precip-
itation), chloride (titration), selected cations (via atomic adsorption
spectrometry), ash @ 800°C, and material extractible with ether and
chloroform. The results are shown in Table F-l. It is clear from this
work that the impinger catch from the continuous grate incinerator was
almost totally mineral acid with little else present.
F-2
-------
TABLE F-l
Analysis of Impinger Catch
from Continuous Grate Incinerators
Component Units
Cations meq.
H+ (acidity)
Other*
Total (max.)
Anions meq.
Sulfate
Chloride
Total
Organic Extractible mg
Ash @ 850°C mg
*0ther cations were present as follows:
Calcium
Magnesium
Potassium
Sodium
Lead
Zinc
Total
37-3
15.0
<0.1
15.1
0.9
14.8
15.7
<2
<2
Sample No
41-3/43-3
12.3
<0.1
12.4
1.2
12.9
14.1
<2
<2
45-3/36-3
12.0
<0.1
12.1
1.0
12.0
13.0
<2
<2
(milligrams)
0.2
O.OA
0.1
0.3
0.06
0.05
0.8
0.5
0.07
0.1
0.4
0.1
0.05
1.2
0.2
0.03
0.1
0.2
0.2
0.05
0.8
Ameq = milliequivalents; mg - milligrams.
F-3
-------
While it is recognized that examination of the effluent from two sources
is not in any sense representative of all incinerators, it was felt to
be of value to study the emissions from another incinerator. Therefore,
several samples of impinger catch were collected via the EPA train pro-
cedure at a batch incinerator and quantitatively analyzed. The results
are presented in Table F-2. As was found previously, sulfuric and hydro-
chloric acid were the main components of the impinger catch. A compari-
son summary of the two incinerators where all of the data has been cal-
culated on the basis of a 1.5 m^ sample is given in Table F-3. It is
clear from this data that the collected "particulate" is similar from
both incinerators. The batch unit does have slightly more organic matter
and a little less sulfate in the impingers, but mineral acid is the major
component being collected by the impingers in both cases.
II. COMPREHENSIVE ANALYSIS OF PARTICULATE CATCH
To help complete our understanding of the composition of the particulate
catch, a detailed chemical characterization of four samples of particulate
catch from a continuous grate incinerator was carried out. The goal of
this part of our program was to obtain a thorough understanding of the
composition of the particulate catch collected in each portion of the EPA
sampling train.
A. Experimental Procedure
To do this, four samples were collected via the normal EPA sampling pro-
cedure, except that:
• The sampling was only continued for 28-44 minutes rather than
one hour so that the sample sizes were 35-40 percent of our
normal runs;
• Organic membrane filters were substituted for glass filters in
two of the runs. This was done to permit a chemical analysis
of the filter catch without being faced with a high blank level
for the cations and anions of interest. (Note: For details
on sampling conditions, see Runs Nos. NI-77 through 80 in
Table D-7 of Appendix D.)
The distribution of particulate for the runs with glass filters are. not
significantly different than those with the membrane filters, but the
amount of particulate collected has a wider variation and higher average
level than previous runs. The average percentage distribution between
probe plus filter versus impingers is consistent with our earlier find-
ings, but the probe plus cyclone collected more of the "traditional"
particulate catch (catch in probe, cyclone and filter) in this latest
series.
Once collected, the experimental procedures utilized for analyzing these
samples were as follows:
F-4
-------
TABLE F-2
Analysis of
Impinger Catch from Batch Incinerator
Component
Impinger Catch via
EPA Procedure
Water Soluble
Organo Soluble
Acetone Rinse
Total
S02 (via
Solids
(3 105°C
@ 850°C
R2°3
Anions
Sulfate
Sulfate
Chloride
Chloride
Total
Acidity
Units A
mg
mg
mg
mg
mg
meq
mg
meq
meq
meq
Sample
71-3 72-3 73-3
31
11
8
50
<2
22 30
0.8 2.4 <5
<1
10 13 12
0.2 0.3 0.3
108 136 127
3.1 3.9 3.5
3.3 4.2 3.8
3.5,3.4 4.3,4.0 4.4
Code (BI- )
74-3 75-3
11 16
9 11
6 9
26 36
<2 <2
__
<5
<1
13
0.3
112
3.2
3.5
3.8
76-3
24
3
9
36
<2
—
—
—
—
—
—
—
—
Units: meq = milliequivalent; mg = milligram.
F-5
-------
TABLE F-3
Comparison Summary for
Particulate Catch from Two Incinerators
Average Values for
Continuous Grate
Batch
407
77
272
349
45
4
8
271
Results via EPA Procedure;
Total Weight of Particulate
Distribution of Particulate
Catch Before Filter
Filter Catch
Combined Total
Impinger Catch
Water Soluble
Organic Extract
Acetone Rinse
Combined Total
3
Weight Per Volume Sampled (mg/m )
Chemical Analysis of Impinger Catch
As Received
Sulfate
Chloride
Ash
Acid (meq)
Organic (est)
Total (est)
After Evaporation to "Drvness" (Drierite)
Sulfate
Chloride
Ash
Organic (est. total)
Water (est)
Total (est)
*A11 values converted to basis of 1.5 m^ sample.
57
100
19
67
86
11
1
2
14
305
34
234
268
21
8
8
203
100
11
76
87
7
3
3
37
13
19
160
1+
(5.1)
6
186
10
86
—
—
3
—
12
121
1+
(4.0)
12
146
8+
82
—
—
8+
—
19
1
1
7
13
41
46
2+
2+
17
32
12
1
1
12
11
37
32+
2+
2+
32+
30
F-6
-------
• Acetone Washings of Nozzle .and Probe
The samples were transferred to platinum dishes, evaporated to dryness
with nitrogen, equilibrated in a drierite desiccator and weighed. They
were then dried at 105°C and reweighed. Following this, the sample was
extracted three times with 50 ml hot water. The extract was filtered
through a 2" millipore membrane filter and combined. The residual solids
were collected on the filter and dried. Following this, the extract
and solids were treated as shown in Figure F-l.
• Filter Catch
The membrane filters (NI-77 and -79) were dried at 105°C, weighed,
extracted three times with hot water, dried at 105°C and then treated
as shown in Figure F-l.
The glass filters (NI-78 and-80) were desiccated under drierite, weighed,
dried at 105°C and reweighed. These were fused with sodium carbonate
and then analyses made for &2®3 anc' sulfate.
• Impinger Catch
After measuring the volume of each solution, the impinger catch was
evaporated, transferred to a 100 ml volumetric flask and diluted to
volume with water. These solutions were then analyzed for: acidity,
halides, sulfate, nitrate, trace metals (K, Na, Zn, Pb, Ca, Mg, Fe, Sn),
organic extractibles, and "dried" solids.
The procedures employed for analyses of these samples were:
Acidity
Chloride
Sulfate
Nitrate
R203
- Titrate either potentiometrically, or to a phenol-
phthalein end point, with 0.01 N sodium hydroxide.
- Titrate with 0.05 N mercuric nitrate.
- Gravimetrically after precipitation with barium
chloride, and in some cases spectrophotometrically
via the barium chloranilate procedure.
- Spectrophotometrically using the phenol disulfonic
acid (PDS) procedure.
- Gravimetrically after precipitation with ammonia and
ignition to 800°C.
Trace Metals -Atomic absorption spectrometry using selected source
lamps for the elements of interest, including:
calcium, magnesium, sodium, potassium, iron, tin,
zinc, lead, aluminum and titanium.
F-7
-------
FIGURE F-l
Analysis of Solid Catch
Dried and Weighed Sample
i
CD
I
Filter and/or Solids
Dry at 105°C and reweigh
FUSION Fuse with sodium carbonate
Dissolve in HG1-H20
Filter
SOLIDS Insoluble Portion
Extract three times with 50 ml boiling water.
Combine extracts, evaporate and dilute to 100 ml.
Extract
Determine the following:
• Halides
• Acidity
• Sulfate
• Nitrate
• Trace Metals (K,Na,Pb,Zn,Ti,Al,Fe,Sn)
Dry at 105°C—weigh
Ignite at 850°C—weigh
Dissolve in NaOH and determine silica
FILTRATE
Filtrate
Dilute to 100 ml and determine:
• Sulfate
• R2°3
• Metals (K,Ca,Mg,Al,Ti,Fe,Pb,Zn,Sn)
-------
B. Results and Discussion
1. Catch in Probe and Cyclone
With these samples, we attempted to obtain as comprehensive an analysis
as possible so that a. proper material balance on the composition of the
catch could be derived. Unfortunately, the small size of the sample as
well as its chemical complexity made this a difficult goal and one that
was not completely attained. However, in spite'of these limitations, it
was possible to derive a good picture of the nature of the probe catches
(see Tables F-4 through F-6). Some observations on this data for the
probe catch are:
• As evidenced by an examination of Table F-5, the overall ratio
of the components of the probe catch did not vary appreciably
between runs even though there was variation between solubles
and insolubles. The major cation constituents (>10%) were
calcium, zinc, potassium, sodium, and aluminum.
• Distribution of the measured constituents is also quite con-
sistent (Table F-6) . Cations make up about 50% by weight of the
total probe catch. The quantity of anions determined is low in
NI-78-2 and possibly 79-2 and 80-2. Some of the cations may
have been present as oxides—an anion which was not determined
via analysis.
• Even though hot water solubles varied appreciable (24-60%),
their composition did not seem to change except that total
measured anions were lower with lower percentage of solubles.
This could be related to presence of ignited oxides—which
are insoluble and have oxygen as the anion.
2. Filter Catch
A similar effort to that for the material caught in the probe and cyclone
was devoted to the filter catch. The results of our analyses on the
filter catch are given in Tables F-7 through F-9. Because the glass
filters were known to contain impurities, complete data was obtained only
on NI-77-1 and NI-79-1 where membrane filters had been employed. As
shown in Table F-7, the materials balance for NI-77-1 is poor (almost
twice by combining the analyses for individual components as actually
collected). Possibly because of this, the agreement between the two
runs, NI-77 and NI-79, is not overly good. Thus, in Table F-8, which
gives the distribution of the filter catch between analysis categories,
there is poor agreement. However, Table F-9, which shows the relative
levels of the individual cation components, indicates reasonable con-
sistency except for lead, potassium and sodium.
F-9
-------
TABLE F-4
Analysis of Probe Plus Cyclone Catch
Code (MI- )
Weight of Catch
Dried over Drierite
Dried @ 105° C
Hot Water Solubles
Weight
Percent of Catch
Free Acidity
Cations
Zinc
Lead
Tin
Iron
Potassium
Sodium
Titanium
Aluminum
Total Cations
Anions
Analysis Results
mg
mg
rag
%
meq
mg
mg
Chloride
Sulfate
Nitrate
Total Anions
Total Measured Solubles mg
Ratio of Components
Determined by Measure-
ment of amount
Extracted
77-2
32
30
18
60
1.5
<0.05
<0.5
1.0
0.9
<0.05
<0.05
3.4
4
5
9
«
0.7
78-2
82
80
19
24
1.7
<0.05
<0.5
1.8
1.5
<0.05
<0.05
5
5
5
10
15
0.7
79-2
107
105
47
45
1.7
<0.06
<0.5
3.0
2.5
<0.05
<0.05
7.2
12
5
17.
24_
0.5
80-2
52
50
19
38
2.3
<0.06
<0.5
2.8
2.1
<0.05
<0.05
7.3
7
5
12
Ii
1
Blank
1
<0.1
<0.05
<0.05
<0.5
<0.02
<0.05
<0.05
<0.05
1
1
1
-------
TABLE F-4 (Continued)
Analysis Results
0.6
0.09
0.09
0.2
<0.5
0.6
0.2
0.4
0.9
3.1
3.2
0.9
1.3
0.8
<0.5
1.4
1.1
1.8
5.0
15.5
6.9
1.4
1.8
1.7
<0.5
1.6
0.9
1.2
4.2
19.7
4.4
0.5
0.7
1.2
<0.5
0.6
0.5
0.7
1.9
10.5
0.1
0.05
0.2
0.3
<0.5
0.05
0.1
0.05
0.1
Unit 77-2 78-2 79-2 80-2 Blank
Hot Water Insolubles
Weight (by difference) mg 12 61 58 31
Acid Insolubles After Fusion mg 2 7 14 3
Cations mg
Calcium
Magnesium
Zinc
Lead
Tin
Iron
Potassium
Titanium
Aluminum
Total Cations
mg 16 40 31 18 4
mg
Sulfate 0-8 3-6 3-13 1-8 0-17
Silicate <0.5 5 10 <0.5 0.9
Total Soluble and Insoluble mg
Hot Water Solubles
Cations by AA 3.4 5 7.2 7.3
Aiiions 9 10 17 12
Insolubles
Fusion Insolubles 2 7 14 3
Cations by AA 3.1 15.5 19.7 10.5
Anions 6 9 20 5
R203 Cations* 8 20 16 9
Combined Total 32^ 66. 94_ 47_
Percentage Recovered** 100 80. 87. 91_
*A11 values have been corrected for measured blank levels where possible
AUnits are: mg = milligram; m'eq = milliequivalents; 7, = weight percent.
+In calculating, assume 50% of R£03 is cations by weight.
-divided -by-weight of catetu
F-ll
-------
TABLE F-5
Relative Levels of Individual Components
of Probe and Cyclone Catch
Element
Cations
Calcium
Magnesium
Zinc
Lead
Tin
Iron
Potassium
Sodium
Titanium
Aluminum
Total
Amount in Sample mg (%)
77-2
0.6
0.09
1.6
0.2
<0.5
0.6
1.2
0.9
0.4
0.9
6.5
(9)
(1)
(25)
(3)
(9)
(18)
(14)
(6)
(14)
(100)
78-2
3.2
0.9
3.0
0.8
<0.5
1.4
3.9
1.5
1.8
5.0
21.5
(16)
(4)
(15)
(4)
(7)
(14)
(7)
(9)
(24)
(100)
79-2
6.9
1.4
3.5
1.7
<0.5
1.6
3.9
2.5
1.2
4.2
26.9
(26)
(5)
(13)
(6)
(6)
(14)
(9)
(4)
(16)
(100)
80-2
4.4
0.5
3.0
1.3
<0.5
0.6
3.3
2.1
0.7
1.9
17.8
(25)
(3)
(17)
(7)
(3)
(19)
(12)
(4)
(11)
(100)
Average Cc
mg
15.1
2.9
11.1
4.0
0.5
4.2
12.3
7.0
3.9
12.0
72.7
>mposition
%
21
4
15
5
1
6
17
10
5
17
* Listed numbers are milligrams found. Values in parenthesis are relative percent of element to
total cation levels in the probe catch.
-------
TABLE F-6
DistributionWithin Probe/Cyclone Catch
Amount in Sample mg (%)'
7
77-2
30
18
6.5
8
14.5
11
4
5
2
31.5
(100)
(60)
(22)
(26)
(48)
(37)
(13)
(50)
(7)
(105)
o total probe
78-2
80
19
20.5
20
40.5
9
5
19
7
66.5
catch.
(100)
(24)
(26)
(25)
(51)
(11)
(6)
(24)
(9)
(83)
79-2
105
47
27
16
43
15
12
37
14
94
(100)
(45)
(26)
(15)
(41)
(14)
('ID
(35)
(13)
(90)
80-2
50
19
18
9
27
10
7
17
3
47
(100)
(38)
(36)
(18)
(54)
(20)
(14)
(34)
(6)
(94)
Component
Total Collected Catch
Hot Water Solubles
Cations
Determined by AAS
Calculated from
Total
Anions
Sulfate
Chloride
Total (incl. silicate)
Acid Insolubles
Total by Analysis (cations, anions,
acid insolubles)
A Atomic absorption spectrometry.
+ RO^ represents cations precipitated with ammonia. For purposes of this table, 50% of measured R00- was
assumed to be due to cations.
-------
TABLE F-7
Analysis of Filter Catch
Code (NI- )
Filter Type
Weight of Catch
Dried over Drierite
Dried at 105°C
Hot Water Solubles
Weight (by difference)
Percent
Free Acidity
Cations via AAS
Calcium
Magnesium
Zinc
Lead
Tin
Iron
Potassium
Sodium
Titanium
Aluminum
Total
Anions
Chloride
Sulfate
Nitrate
Total
Total SolublesA
Measured
Ratio Measured
to Extracted
Unit*
mg
mg
meq
mg
mg
mg
77-1
Membrane
50
41
33
80
0.2
6.6
0.2
<0.5
<0.01
12
3.5
<0.05
<0.05
22.3
14
13
11
49
1.5
Results Found
78-1 79-1 80-1
Glass Membrane Glass
153 124 108
110
78
71
0.4
6.6
0.4
<0.5
0.01
3.7
7.3
<0.05
0.3
18.3
31
11
42_
60
0.75
Blank
— — —
10
0.1
<0.05
<0.05
<0.5
<0.01
<0.05
<0.1
<0.05
<0.05
-------
TABLE F-7 (Continued)
Results Found
Hot Mater Insolubles
Weight
Acid Insolubles After
Fusion
Cations
Calcium
Magnesium
Zinc
Lead
Tin
Iron
Potassium
Titanium
Aluminum
Total
Unit
mg
rag
mg
Anions
Sulfate
Silicate
Total Insolubles
Combined Total
Hot Water Insolubles
Cations by AAS
Anions
mg
mg
mg
mg
77-1
10
0.05
0.05
0.5
1.4
0.5
0.05
0.3
0.
0.
.1
.2
2.5
1-5
0.1
22
22
27
78-1
75
79-1
32
0.2
0.06
0.8
5.
0.
0.
0.
0.4
0.8
8.8
4-10
1
13
18
42
80-1
95
Blank
0.2
0.05
0.05
0.2
0.5
0.2
0.2
0.05
0.1
0.5
5
Insolubles
Fusion Insolubles
Cations by AA
Anions
R203
Total Found
Recovered
1
3
10
22
75_
180
2
9
11
13
£5
86
*Units are: mg - milligrams; meq = milliequivalents.
^Total solubles = total cations plus total anions.
~*~Ratio = Total solubles divided by weight (by difference of hot water solubles).
xTotal insolubles = fusion insoluvles + cations + anions + R203-
aAssume 0% by weight of R203 is cation.
F-15
-------
Component
Total Collected Catch
Hot Water Solubles
Cations
TABLE F-8
Distribution Within Filter Catch
Determined by AAS
Calculated from R-O™
Total
Anions
Sulfate
Chloride
Total (including Silicate)
Acid Insolubles
Total by Analysis
(Cations, Anions, acid Insol.)
41
33
24.
16
At
16
14
) 30
1
72
77-2
(100)
(80)
8 (61)
(27)
(88)
(39)
(34)
(73)
(2)
(175)
Amount in Sample -
78-2
135 (100) 110
78
27
10
37
25 (19) 18
31
50
— 7
89
mg (%)*
79-2
(100)
(71)
.1 (25)
(6)
(31)
(16)
(28)
(45)
(2)
(81)
80-2
98 (100)
9 (11)
* Values in parentheses are relative percent of total filter catch.
F-16
-------
TABLE F-9
Relative Levels of Individual Components of Filter Catch
Amount in Sample - mg (%)
Element 77-2 79-2 Average
Calcium
Magnesium
Zinc
Lead
Tin
Iron
Potassium
Sodium
Titanium
Aluminum
<.
<•
7.
1.
0.
0.
12.
3.
0.
0.
05
05
1 (29)
6 (6)
5 (<1)
5 (<1)
3 (50)
5 (14)
1 (
(
(1)
(32)
(20)
(1)
(2)
Total (mg) 24.8 (100) 27.1 (100)
*Values in parentheses are relative percent of element to total
cation levels in the filter catch.
F-17
-------
3. Comparison of Probe/Cyclone and Filter Catches
In order to obtain an indication as to whether the composition of the filter
catch was different than that found in the probe and cyclone, the averages
for each set of data were compared (see Table F-10). In looking at this
table, it must be remembered that the averages represent a wide range
between the individual results. This may represent a special problem for
the filter values because the values for NI-77-1 add up to 175% of the
original sample. However, in spite of these reservations, it is con-
structive to compare the available data.
• Each is primarily inorganic mineral type particulate;
• There is little free acid measured in either sample. Thus, if
collected by the solids, the acid reacts to form a more neutral
species;
• Sulfate levels seem consistent;
• Chloride is much higher in the filter catch. Although the dif-
ference may be due to a lower temperature on the filter, it also
may be an artifact because of the use of a solvent evaporation
step for preparing the probe catch for analysis. Hydrochloric
acid could be distilled out of the sample more easily at that
point than during drying of an essentially "dry" filter at 105°C.
• There are more acid insolubles (after sample dissolution) in
the probe catch than in the filter catch. Very little silicate
was found in the filter catch whereas in two cases (NI-78 and
NI-79) some silicate was found in the probe catch. This could
explain the difference in the acid insolubles. (Silicate acid
should be insoluble and reported as acid insolubles. However,
sometimes it remains solubllized and this is the reason it was
determined in the filtrate. When silica was found in the filtrate,
it almost certainly had to be a major portion of the value reported
as acid insolubles reported in Table F-6. However, this potential
error does not change our conclusions in any way.
• The relative proportions of individual cations to the total deter-
mined by atomic absorption spectrometry shows differences (Table
F-ll). In particular, a high proportion of calcium and magnesium
is in the probe/cyclone catch. (Unfortunately, these elements
were not determined in the water soluble portion in either case
so that these results may be distorted. However, until further
work is done, this possible difference has been noted.) Also,
there seems to be more zinc, lead, potassium and sodium on the
filter and more aluminum in the probe/cyclone.
F-18
-------
TABLE F-10
Comparison of Filter Catch
and Probe/Cyclone Catch
Results (Weight Percent of Total Catch)
Probe/Cyclone Filter
Component
Hot Water Solubles
Cations
Determined by AAS
Calculated from
Total Cations
Anions
Sulfate
Chloride
Total Anions
Acid Insolubles
Total by Analysis
Range
24-60
22-36
15-26
Average
42
28
21
49
11-37 20
6-13 11
31
6-13 9
89
Range
71-80
25-61
6-27
Average
75
44
17
61
11-39 21
28-34 31
52
2 2
115
*Range and Averages for all runs.
F-19
-------
TABLE r'-ll
Proportion of Individual Cations
in Filter and Probe/Cyclone Catches
Element
Calcium
Magnesium
Zinc
Lead
Tin
Iron
Potassium
Sodium
Titanium
Aluminum
Results (Percent of Total Cations Measured by AAS)
Probe/Cyclone
Range
9-26
1-5
13-25
3-7
3-9
14-19
7-14
4-9
11-24
Average
19
3
18
5
6
16
11
6
16
Filter
Range Average
1 1
1 1
27-29 28
5-23 15
-------
• A presentation of the distribution of sulfate and chloride between
the various parts of the sampling train is given in Table F-12.
As can be seen from this data, the sulfate was fairly evenly dis-
tributed between the probe/cyclone, filter, and impinger. The
total amounts of sulfate collected ranged between 92-155 ing which
based on a 1.5 m-* sample represents 15-25 ppm of sulfur oxides.
• Chloride, on the other hand, was found predominantly in the impin-
gers with half as much on the filter and only a small amount in
the probe/cyclone. This may be directly related to the tempera-
ture and amount of water available for trapping—i.e., since the
probe is hottest, there is less "available water." Also, sulfur
trioxide will probably displace any HC1 trapped on the filter.
4. Impinger Catch
The results of the detailed chemical analysis of the impinger catch is
given in Table F-13. As had been demonstrated in the preliminary quanti-
tative work, the impinger catch was primarily mineral acid (sulfuric
acid and hydrochloric acid) with practically no mineral particulate.
Thus, the amount (milliequivalents) of sulfate and chloride matches
almost exactly the total acidity. Even though a small quantity of sodium,
potassium and iron were found, mineral type particulate represented a
negligible amount compared to the total catch. Because of the extreme
difficulty of characterizing trace quantities of non-extractable organics,
the exact nature of the organics in this faction was not determined and
thus is unknown. (Typical organics which might have been present include
low molecular weight acids, alcohols, ethers, esters, aldehydes, and ketones.)
III. GENERAL COMMENT
By means of this very detailed quantitative analytical program, we were
able to generate a good understanding of the composition of the particulate
catch. However, the material balance aspects of the results, especially
in the case of the filter catch (Table F-7), are not as good as would be
desired. Whether this was due to the variation in sample homogeniety,
impression of the analytical methods, or presence of unidentified com-
ponents is not known. It is likely that operating with such small samples
the analytical error was sufficiently large to explain the observed dif-
ferences.
In summary, we concluded that the probe/cyclone catch and filter catch
were similar in composition and predominantly inorganic materials.
Although each had similar compositions, variations were noted between
the probe/cyclone and filter in relative concentrations of some cations
as well as sulfate and chloride. The significance of these differences
to the use of this sampling train are unknown, but presumed to be small.
As noted earlier, the impinger catch is primarily mineral acid, and thus
should not contribute to a traditional measure by weight of particulate
catch.
F-21
-------
to
TABLE F-12
Distribution of Chloride and Sulfate in Sampling Train
Sample
(NI- )
77
78
79
80
Distribution of Sulfate (%)
Probe/Cyclone
28
16
28
28
Filter
40
42
34
26
Impinger
32
42
38
46
Distribution of Chloride (%)
Probe/Cyclone Filter Impinger
9 32 59
—
11 29 60
— — —
Average
25
35
45
10
30
60
-------
TABLE F-13
Analysis of Impinger Catch
General Source (NI- )
Volume of Gas Sampled
Volume of Water in Impingers ml
Results via EPA Procedure
Water Solubles
Organic' Extractibles
Total Impinger Catch
As received
on 1 m3 basis
Chemical Analysis*
Acidity
Cations
Calcium
Magnesium
Zinc
Lead
Tin
Iron
Potassium
Sodium
Total Cations
Results for
UnitA
m3
rs ml
mg
meq
mg
mg
meq
77-3
0.57
335
16
2
16
28
0.9
0.05
0.05
0.05
0.05
1
0.05
0.03
0.07
0.2
13
26
0.5
39
0.1
0.9
0.9
0.1
0.7
0.1
0.8
78-3
0.59
330
— —
_^«.
3.7
0.05
0.05
0.05
0.05
1
0.05
0.1
0.6
0.8
25
122
0.7
148
0.1
3.7
3.7
0.3
3.4
0.1
3.7
79-3
0.53
345
24
2
24
45
2.0
0.05
0.05
0.05
0.05
1
0.02
0.08
3.4
3.4
20
64
1.3
85
0.1
2.0
2.1
0.2
1.8
0.1
2.0
Sample
80-3
0.59
325
17
2
17
29
2.5
0.05
0.05
0.05
0.05
1
0.02
0.04
0.3
0.4
16
83
0.9
100
0.1
2.5
2.5
0.2
2.3
0.1
2.5
Blank
350
„_
•••*•«.
0.1
0.05
0.05
0.05
0.05
1
0.01
0.02
0.02
0.2
2
1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Anions
Sulfate
Chloride
Nitrate
Total Anions
Materials Balance
Positive lons^
Cations
Acidity
Total
Negative Ions
Sulfate
Chloride
Nitrate
Total
+0n sample as received and before evaporation to dryness.
Am3 = gas sample in cubic meters; mg = weight in milligrams; meq = quantity of
component in milliequivalents.
F-23
-------
APPENDIX G
HIGH-EFFICIENCY GLASS FIBER FILTER
PERFORMANCE CHARACTERISTICS
I. INTRODUCTION
High-efficiency glass fiber filters are commonly used to collect parti-
culates from both ambient air and stack gases. The filter material was
developed for the Atomic Energy Commission^'2) for ventilation applica-
tions. It is a felt-like material composed of a combination of glass
fibers in mixed sizes (about 0.5 to 3pm).
There are specifications for this filter for processing use (MIL-F-
51079A) but not sampling applications. At the present time, two binder-
less filters (MSA 1106BH and Gelman Type A) have been found the most
acceptable for atmospheric sampling and analysis(3>*) y and are specified
for use in the EPA particulate stack sampling train.
Because significant fractions of particulates collected in the EPA sampl-
ing train have been found in the impingers downstream of the filter, the
efficiency of the filter has been questioned as follows:
• Is the filter efficiency adequate for aerosol particles smaller
than 0.Sum?
• Is the filter efficiency for solid aerosol particles of various
densities as good as the tested efficiency for liquid OOP
aerosol?
• What are the effects of high temperature on filter performance?
• Do electrostatic charges on the particles and/or filter reduce
filter efficiency?
We have evaluated the capability of the filter by a review of previous
studies and by experimental incinerator sampling.
Several of the more important properties of filters are described below:
• Efficiency is a measure of the ability of a filter to remove parti-
cles from an air stream. Initial efficiencies of new filters are
most commonly quoted, but efficiency as a function of filter load-
ing is equally important. Performance is frequently expressed in
terms of penetration, which is a measure of the ability of parti-
cles to pass through a filter. Percent efficiency equals 100%
minus percent penetration.
G-l
-------
• Flow resistance is measured by the static pressure drop across a
filter at a given flow rate.
• Particle collection capacity or filter loading is the amount of
particulate that a filter can hold before flow resistance or
penetration become excessive.
The filtration performance of these filters depends on filter composition,
particle size, particle loading, and many other factors. However, the
following supplier's data on MSA 1106BH with 0.3ym diameter dioctyl
phthalate (DOP) liquid aerosol is typical:
Initial Flow
Velocity Efficiency Resistance
(ft/min) (%) (in. H?0)
10 99.995 1.48 to 1.58
28 99.98 3.94 to 4.33
Linear filtration velocities in the EPA sampling train at 1 CFM are
approximately as follows:
Filter Diameter (in.) Velocity (ft/min)
2-1/4 36
3 24
4 12
II. EFFECT OF VELOCITY AMD PARTICLE SIZE ON FILTRATION PERFORMANCE
Figure G-l shows that flow resistance Increases linearly with velocity.
Figure G-l also shows that efficiency decreases with velocity up to
about 50 ft/min and then increases. The minimum efficiency point results
from the combined effects of particle diffusion and impaction, and
varies with filter construction (usually between 30 to 70 ft/min).
The efficiency on particles other than 0.3ym has been the subject of
much misunderstanding and controversy. It is commonly thought that the
filter collects particles only down to 0.3ym diameter. This misunder-
standing results from the common practice of stating 0.3ym DOP effi-
ciences without interpretation. The DOP test was designed during World
War II to test high-efficiency gas mask filters against what was thought
at that time to be the most penetrating particle size (0.3ym). The most
penetrating particle size of these filters is now believed to be between
0.1 and 0.3pm diameter; the peak penetration is about 2 to 3 times the
DOP penetration.(5)6,7,17) Particles smaller and larger than the most
penetrating are collected with greater efficiency, largely by diffusion
and impaction, respectively.
G-2
-------
100.0
o
99.99
99.98
U
(3
-------
III. EFFECT OF SOLID VS. LIQUID AEROSOLS ON FILTER EFFICIENCY
The efficiency of filters for solid vs. liquid aerosols is also not well
understood. However, there is reasonable agreement that HEPA filters
have equal initial efficiency for both solid and liquid aerosols.(8-10)
IV. FILTER LEAKAGE EXPERIMENTS
As the filter becomes loaded with solid particles, efficiency increases,
whereas the reverse is true of liquid particles. Loading with liquid
particles is not a common problem with stack sampling. However, where it
does occur, HEPA filters should be limited to particulate loadings of
about 20 mg/sq cm of surface area.(H) The use of two or more filters in
series will increase loading capacity but will also increase flow resis-
tance .
In an effort to learn whether the high-efficiency glass filter is prone
to leakage and thus creates an artificially high catch in the impingers,
the EPA sampling train was modified by incorporating a glass "T" after
the filter holder. The stack gas sample was split by the "T" with
one side of the "T" passing in a normal fashion to the EPA impingers
while the gas stream from the other side was passed through an auxiliary
filter and then to a set of EPA-type impingers. This modified sampling
system was then used for direct sampling of incinerator stack gases
using MSA 1106BH filters in each side of the train. The data for this
work are shown in Table G-l. Based on the fact that the amount of
particulate collected on the auxiliary filter was a small percentage of
the total collected by the train, it has been concluded that the primary
filter is not allowing a significant (>3%) amount of filterable parti-
culate to reach the impingers. Therefore, a back-up filter would have
little or no effect on the quantity or quality of the material caught
in the impingers.
Efforts to run the EPA sampling train in such a way that the filter
became overloaded and thus was more susceptible to leakage were unsuccess-
ful. This was true for two reasons; one being the difficulty in operat-
ing the train successfully for 5-6 hours, and the second being that we
were unable to reach liquid overloading of the filter by organic tars
and oils because of the low organic emissions from this incinerator.
The experiments did show, however, that no breakthrough was obtained
even after the filter collected up to four times the "normal" amount of
particulate. (Normal represents 200-300 mg.)
V. EFFECT OF PARTICLE DENSITY ON FILTER EFFICIENCY
The effect of increased particle density on filter efficiency is to
increase efficiency at high velocities, due to impaction. Reduced
efficiencies with heavier particles have been measured at low velocities,
but not understood.
G-4
-------
TABLE G-l
Distribution of Particulate
from Filter Leakage Experiments
Run No.
Dry Gas
Volume
(meter3)
Particulate Catch(mg)
On Filter
Probe and Cyclone Primary Back-up
Irapirger Catch (mg)
Collected After
Primary
Filter
Primary &
Back-Up Filter
Material on Back-Up
Filter as a Percent of
Probe/Cyclone/
Filter Catch Impinger Catch
Nl-11 Primary 1.65
-11 Auxiliary 2.78
Nl-12 Primary 1.49
-12 Auxiliary 2.64
Nl-13 Primary 0.91
-13 Auxiliary 0.86
129
154
100
785
563
246
4
2
15
49
28
32
52
80
33
0.3
1.2
20-50
12
*A11 runs 60 minutes, sample initially gous through primary filter, then is split so (.hat part goes directly to a set of impingers while
other part (aux.) goes through a second filter and then impingers.
-------
VI. EFFECT OF ELEVATED TEMPERATURE ON FILTER PERFORMANCE
This subject is discussed by Dyment^ ' as follows:
"At elevated temperature, gas viscosity will increase and therefore
increased pressure drops will be obtained for the laminar flow conditions
which prevail in high-efficiency paper media under normal ambient condi-
tions.
The effect of temperature on fiber filtration mechanisms has been sum-
marized by Thring and Strauss(14) ±n terms of inertial, interception and
diffusional collection. Because of the way in which the three mechan-
isms vary with conditions (i.e., fiber and particle sizes, fluid velocity)
a general rule cannot be formulated for the effects of temperature on
filtration performance. When the predominant mechanism is known, however,
the effects can be predicted qualitatively. Inertia and interception
efficiency are reduced by increased temperature. For the larger parti-
cles for which inertial impaction is more important, changes in particle/
fiber adhesion at high temperature may exaggerate or modify this effect.
The diffusion mechanism is of increasing importance as particle size
diminishes below one micron. For diffusion the effect of temperature
rise is to increase the fiber collection efficiency particularly for sub-
micron particles."
Dyment also reported the results of some glass fiber filter tests with
sodium chloride aerosol at temperatures up to 500°C. He found that glass
paper shrinks at these temperatures so he preheated the samples to
avoid cracking failure. His data, while extremely limited, suggest that
increasing temperatures to 400°C slightly decreased filter efficiency.
His conclusions were as follows:
"In practice the effect of temperature and pressure on filtration mech-
anism and performance has not been found a determining factor in the
application of high-efficiency filters. The main problems are the
physical and chemical effects of a high-temperature environment on the
materials of construction of the filter, which are manifested by reduced
mechanical strength and resilience or loss of adhesion, leading to
mechanical leakage and loss in efficiency."
More recent high-temperature filter efficiency tests were reported by
First.(15) Heat-shrunk quartz fiber filters were tested at tempera-
tures up to 950°F with sodium chloride aerosols of 0.14vjm mass median
diameter at velocities of 15 to 30 cm/sec. No effect of temperature
on efficiency was found.
VII. ELECTROSTATIC EFFECTS ON FILTER EFFICIENCY
The effects of electrostatic charges on filtration efficiency are poorly
understood. The most commonly reported effects are relatively small
increases in efficiency when the particles are charged, and greater
increases when the filter is charged.
G-6
-------
The high-efficiency glass fiber filter is reported to collect at least
99% of charged particles of all sizes in the atmosphere(18,19). Adverse
electrostatic effects on filter efficiency may be possible, but appear
unlikely.
G-7
-------
VIII. REFERENCES
1. Arthur D. Little, Inc., Report No. NYO 4527 to USAEC, Development of
a High-Temperature, High-Efficiency Air Filter (August 18, 1953).
2. W.J. Smith, "Concerning the Absolute Filter," J. de Recherches
Atmospheriques, 205-209 (1966).
3. J.P. Pate and E.G. Tabor, "Analytical Aspects of the Use of Glass
Fiber Filters for the Collection and Analysis of Atmospheric Partic-
ulate Matter," Am. Ind. Hyg. Assoc. J., 145-150 (March-April 1962).
4. D.N. Kramer and P.W. Mitchel, "Evaluation of Filters for High-Volume
Sampling of Atmospheric Particulates," Am. Ind. Hyg. Assoc. J.,
224-228 (May-June 1967).
5. J. Dyment, Penetration of Glass Fibre Media by Aerosols as a Function
of Particle Size and Gas Velocity, AWRE Report No. 05/69.
6. R.G. Dorman, "The Sodium Flame Apparatus in Routine and Research
Tests of Air Filters," Filtration and Separation, 499-503 (Sept-
Oct 1969).
7. J. Dyment, "Use of a Goetz Aerosol Spectrometer for Measuring the
Penetration of Aerosols through Filters as a Function of Particle
Size," Aerosol Science, Vol. 1, 53-67 (1970).
8. R.N. Mitchell, et al., "Comparison of Respirator Filter Penetration
by Dioctyl Phthalate and Sodium Chloride," Am. Ind. Hyg. Assoc. J.,
357-364 (June 1971) .
9. W.J. Smith and N. Surprenant, "Properties of Various Filter Media
for Atmospheric Dust Sampling," ASTM Proc., Vol. 53, p. 1122 (1953).
10. B.I. Ferber, et al., Respirator Filter Penetration Using Sodium
Chloride Aerosol. U.S. Dept. Interior, Bureau of Mines, RI 7403
(June 1970).
11. A. Doyle, W.J. Smith, and N.M. Wiederhorn, Collection, Monitoring,
and Identification of Particles in Gas Distribution Systems, Am. Gas
Assn. (Cat. No. 46/OR), 1959.
12. E.A. Ramskill and W.L. Anderson, "The Inertial Mechanism in the
Mechanical Filtration of Aerosols," J. Colloid Science, ^, No. 5,
p. 416-28 (October 1951).
13. J. Dyment, "Assessment of Air Filters at Elevated Temperatures and
Pressures," Filtration and Separation, _7, No. 4, 441-445 (July-Aug
1970).
G-8
-------
14. M.W. Thring and W. Strauss, Trans. Inst. Chem. Engrs., 41, 248 (1963)
15. M.W. First, Performance of Absolute Filters at Temperatures from
Ambient to 1000°F, 12th AEG Air Cleaning Conference (Aug 1972).
16. H.L. Green and W.R. Lane, Particulate Clouds, 2nd ed. Ch. 6, Van
Nostrand (1964).
17. R.G. Dorman, "Filtration," Ch. 8 of Aerosol Science, edited by C.N.
Davies, Academic Press, London and New York (1966).
18. C.B. Moore, et al., "Airborne Filters for the Measurement of Atmos-
pheric Space Charge," J. Geophysical Research, Vol. 66, No. 10,
p. 3219 (October 1961).
19. R.B. Bent, "Testing of Apparatus for Ground Fair-Weather Space-
Charge Measurements," J. Atm. & Terrestrial Phys., Vol. 26, p. 313
(1964).
G-9
-------
APPENDIX H
CHARACTERIZATION OF PARTICULATES
DIRECTLY ON THE FILTER
Several sampling runs were conducted for very short times (5 seconds to
5 minutes) to obtain very lightly-loaded filters suitable for optical
examination. (Samples were from Runs No. NI-17A, NI-48, and NI-60.
Details of sampling conditions are given in Appendix D.) In this way,
it was hoped to obtain a typical particle size distribution and further
more, through x-ray analysis of isolated particles, to determine system-
atic variations in particle chemistry as a function of size. For these
studies, three filter media were employed, including glass fiber, organic
membrane, and silver membrane. Scanning electron micrographs of clean
filters of each type are shown in Figure H-l.
Characterization methods included optical examination, scanning electron
microscopy (SEM) and non-dispersive x-ray analysis of individual particles
in the SEM. In general, the particles were too small to obtain good
confirmation by optical microscopy. The excessive surface roughness
of the filters combined with the very limited depth of field of the
optical microscope made it very difficult to "focus-in" on particles,
and accurate descriptions of size and morphology were precluded, even
for the organic membrane filters which could be made transparent by a
suitable immersion oil and examined in transmission. Therefore, all
examinations were conducted with the SEM.
Because of its high electrical conductivity, the silver membrane filter
is very stable under electron bombardment in the SEM. The organic mem-
brane and glass fiber filters must be vapor-coated with carbon to avoid
building up of space charges and loss of image resolution. A non-
dispersive x-ray analyzer attachment to the SEM permits elemental analysis
of individual particles and is invaluable for characterizing particle
chemistry.
Several typical examples of filter catch particulate are presented in
Figures H-2 through H-6. Figure H-2, the glass fiber filter is seen
to trap large particles at the surface, whereas smaller particles
became embedded. Highly morphological particles are shown in Figure
H-3. The qualitative chemical analysis of one of these particles, as
determined by non-dispersive x-ray analysis, indicated a chemistry
consistent with a mixed potassium calcium sulfate.
In comparison, the collected particle exhibited in Figure H-4 exhibits
a morphology very similar to that of the particles shown in Figure H-3.
However, the chemistry is quite different, with enrichment in Cl, Al,
Si and Zn content and less Ca and S. These examples point out the
difficulty in identifying compounds by the methods. Individual particles
can have quite different chemistries dependant upon their combustion
H-l
-------
history. Also, the x-ray intensity of any given element is influenced
by the other elements present (matrix absorption and emission effects)
and also by the size and geometry of the particle being analyzed. Thus,
results are qualitative in nature, and the usefulness of the method
depends upon the objective of the analysis.
Other typical collected particles are shown in Figures H-5 and H-6. The
spherical glassy slag particle contains many smaller particles on the
surface which have a somewhat different chemistry. The fly ash particle
has an elemental chemistry very similar to what is observed in gross
chemical analysis of filter catches by x-ray fluorescence or emission
spectrographic analysis.
Examination of several dozen particles has indicated a similar chemistry
for sizes within the range of 2 to 5 u, with Si, Al, S and Cl being the
major species. Based upon this rather limited examination of filter
catch samples, we conclude that it is feasible to conduct particle size
distribution measurements and corresponding qualitative chemical analysis
on all particles down to approximately 1 ym.
H-2
-------
Organic Membrane (Millipore Cellulose Ester)
Silver Membrane (Selas Flotronics)
Glass Fiber (MSA)
FIGURE H-l Scanning Electron Micrographs of
Clean Filter Media
20ym
H-3
-------
'« ' -*"-^W*'«*T 7j«r . WPr
2f)um
5 ym
FIGURE H~2 Particulate Collected on Glass Fiber
Filter, Sample NI-17A
H-4
-------
Energy (KeV)
1.
2.
3.
4.
Na
S
K
Ca
2 \sm
Chemical Analysis
Strong - S
Medium - K, Ca
Weak - Na
FIGURE H-3 Particulate on Organic Membrane Filter Exhibiting
High Degree of Morphology and Corresponding X-ray
Analysis, Sample NI-48
H-5
-------
2ym
Chemical Analysis
Strong - S, Cl
Medium - Na, K
Weak - Al, Si, Zn
Energy (KeV)
1. Na 4. S
2. Al 5. ci
3- Si 6. K
7. Zn
FIGURE H-4 Particulate on Organic Membrane Filter,
Similar to Figure G-3, with Corresponding
X-ray Analysis, Sample NI-48
H-6
-------
Slag Particle
10 urn
Energy (KeV)
A. Small Particles
Strong - S,Cl,K,Ca
Medium - Al,Si
Weak - Na,P,Fe,Mn
1.
2.
3.
A.
5.
Na
Al
Si
P
S
11
121SEC 9396INT
VS: 509 HS : 50EV/CH
6.
7.
8.
9.
10.
Fe
Cl
K
Ca
V
Mn
Energy (KeV)
B. Large Sphere
Strong - S,Al,Si
Medium - Ca
Weak - Cl,Fe,V
FIGURE H-5 Slag Particles Collected on An Organic Membrane
Filter with Corresponding Qualitative Elemental
Analysis, Sample NI-60
H-7
-------
Flyash
10
Area 1
Strong - A1,C1
Medium - S
Weak - Na,Si,K,Fe,Cu
1. Na 6. K
2. Al 7. Ca
3. Si 8. Ti
4. S 9. Fe
5. Cl 10. Cu
11. Zn
Area 2
Strong - Cl
Medium - Na.Al,Si,S,K,Zn
Weak - Ca,Ti,FE
FIGURE H-6 Flyash Particle Collected on an Organic Membrane
Filter with Corresponding Qualitative Elemental
Analysis, Sample NI-48
H-8
-------
APPENDIX I. SULFUR OXIDE STUDIES
Table of Contents
Page
I. FIELD MEASUREMENTS OF SOX LEVELS 1-1
II. LABORATORY STUDIES 1-1
A. Reactions in the Impingers 1-1
1. Distilled Water 1-5
2. Impinger Catch 1-5
B. Oxidation of S02 by the Filter 1-7
C. Influence of Sample Handling on Oxidation
of S02 1-11
D. Combined Field and Laboratory Tests 1-14
E. Distribution of Sulfate in Sampling System 1-16
III. VAPOR PRESSURE DATA FOR S03 1-18
IV. DISCUSSION 1-18
1-1
-------
APPENDIX I
SULFUR OXIDE STUDIES
The role of sulfur oxides in the measurement of particulates is a major
issue which received attention during this program and continues to
deserve further study. As noted in Appendix F, not only were measurable
quantities of sulfate found in the filter catch, but also much of the
impinger catch was sulfuric acid. If the sulfate were simply the result
of scrubbing sulfur trioxide, then it could well fit into present approach
of inclusion as particulate. However, if the sulfate was due to the
oxidation of sulfur dioxide during passage through the sampling train or
in sample handling or treatments, the acid sulfate should not be classi-
fied as particulate.
I. FIELD MEASUREMENTS OF SO Levels
As given in Appendix A, literature values for sulfur oxide levels in the
emissions from municipal incinerators range from 30 to 60 ppm. In order
to aid us in understanding the fate of sulfur oxides in the sampling train,
direct measurements of sulfur oxide levels were made. To do this, 863
was collected using a controlled temperature condensation technique while
SC>2 was collected using an aqueous hydrogen peroxide scrubber. Analysis
for collected sulfate was made using a barium ion (thorin indicator)
titration procedure. The results of these measurements, made on three
separate days, are given in Table 1-1. It can be seen that the sulfur
levels in the incinerator emissions changed from August to November,
primarily as a result of increased S02« The latter measurements at
20-30 ppm SOX are more consistent with earlier published results.
II. LABORATORY STUDIES
The components of the sampling train and the places where reactions of
sulfur oxides might occur are described in Table 1-2. The purpose of
this laboratory study was to try and sort out which of these reactions
might be important.
A. Reactions in the Impingers
Initial experiments examining the reactivity of SCU in various water solu-
tions were carried out by conducting studies at room temperature
using the apparatus schematically depicted in Figure 1-1.
Air at 37 Jl/min was passed through a water saturater and mixed with a
stream of 1.5 £/min of 940 ppm of S02 in nitrogen so that a final con-
centration of 37 ppm S02 was bubbled through 200 ml of the solution under
1-2
-------
TABLE 1-1
SO Levels in Incinerator Emissions
— x
Date
8/71
Average
9/71
Average
12/8-71
Average
Levels of SOy (ppm)
SO,
2.7
20
SO-
1.3
0.5
2.5
SO
.34
3.2
22
Ratio
so3/so4
5
5
1
0.5
1
1
1
2
2
0.8
0.2
2
2
7
7
2
1.3
1.2
3
3
.25
.3
.5
.7
.2
.7
.7
2.7
6.8
0.7
0.5
0.6
0.5
0.3
0.4
3.3
7.3
1.0
0.9
.2
.1
.3
.4
,25
26
21
16
15
2
1.6
4
2
28
22.6
20
17
0.1
<0.1
0.2
0.1
0.1
1-3
-------
TABLE 1-2
Possible Fate of Sulfur Compounds
in the EPA Sampling Train
Sampling Train
Component
Probe/Cyclone/Filter
Impinger
Gaseous
Species
SO,
SO,
a)
Possible Reactions
250°F
S0
b) S02+particulate -
c) SO, Condensation
Fate of Sulfur
Compound
See SO, Reactions
j 3
Sulfites Trapped as Particulate
Trapped as Particulate
d) S03+H20 *H2S04
e) SO.+particulate *Sulfates
S) so2+o2
h) SO +cations
SO,
Condensation or
Absorption
Trapped as Particulate
Collected as Soluble SO-
then:
a) oxidized —*sulfate
b) volatilizes from
liquid
Trapped as Particulate
Sulfites Trapped as Particulate
50, Trapped as Particulate
Influence on
Particulate
Measurement
See S03
Increase Weight
Increase Weight
Increase Weight
Increase Weight
Increased Weight
No effect
Increase Weight
Increase Weight
Increase Weight
-------
Compressed Air
n-
940 ppm SO2/N2
H20 Saturater
(room temperature)
Sample Bottle
Flowmeters
FIGURE 1-1: APPARATUS FOR S02 SOLUTION REACTIVITY STUDIES
1-5
-------
study for 1 hour at a total flow rate (38.5 £/m, 1.36 cfm) . This flow
was comparable to that actually used in operating the EPA sampling train,
but the concentration of 862 was approximately 10 times higher than any
concentrations of SOX which we had observed at the incinerator and thus
represented an extreme test of the possibility that solution reactions
were contributing to the high sulfate values. The observed results
(listed in Table 1-3) were:
1. Distilled Water
The resulting solutions from the distilled water experiment were analyzed
for soluble S02 by 1 2 titration and total sulfate gravimetrically, after
oxidation of the SC>2 bromine. An average of 40 mg/Jl of S02 remained
dissolved in the water at the end of the experiment. This was about
equal to the total amount of sulfur (as sulfate) observed in the solu-
tion so that there was no evidence for oxidation of SC^ prior to analysis.
2. Impinger Catch
Several impinger catches from the reciprocating grate incinerator were
combined to give sufficient sample for study (see Table 1-3). After
dividing into three parts, two portions were treated in the same way as
the water sample above, and one portion was analyzed as a control (no
SOo bubbled through). On the average, the S02 treated solutions con-
tained slightly more sulfur oxides than the control. Since the results
were the same when measuring S02 and total sulfates, the differences can
be attributed entirely to the presence of dissolved S02. Therefore,
once again, a large amount of sulfate was not formed from the S02 in the
impinger. It is interesting to note the much lower amount of S02 dis-
solved in the impinger solution — undoubtedly due to the lower pH (2.5).
A final study was carried out with a different set of combined impinger
catches by treating them initially in the same manner as the previous set,
but then concentrating the solutions to about 5 ml and analyzing for
total sulfate (see Table 1-3). The purpose of this exercise was to see
if the additional concentration step and temperature effects encountered
in the EPA procedure would result in conversion to sulfate. The two S02
treated samples showed an average 5-6 mg/K, greater S0^~ value than the
control samples. This would only amount to 2-3 mg in a typical 300-
500 ml impinger catch which was equivalent to 10-15% of the impinger
sulfate catch.
In summary, it did not appear that the solution conversion of S02 to
can account for more than 10-20% of the observed impinger catch sulfate
values. Junge and Ryan (Quart. J. Roy. Meterol. Soc., 84_ 46 (1956))}
showed that S02 was not oxidized in water, but that conversion to sulfate
could proceed at pH's greater than 3 in the presence of transition metals
such as Mn, Cu, Fe, and Co (which effectiveness decreased in the order
listed) . At a pH of 3 the oxidation proceeds at half the rate that it
does at pH 4 and above, and it is completely stopped at a pH of 2. Since
the impinger solutions rapidly reach a pH of 2.5 due to the collected HC1,
it was not surprising that we did not observe S02 oxidation in impinger
solutions.
1-6
-------
TABLE 1-3
Sample
Distilled Water
Impinger Catch
Treatment
S02 treated
S02 treated
Control (no
added S02)
£.
A
Soluble AA
S02 (mg/£) Total SOV
As as S04 (mg/Ji)
Ave . SO, Ave .
33 54
40 60 61
48 68
10 48
9.5 14 50
9 52
6 6 9 47 47
Impinger Catch
Evaporated to 5ml
Evaporated to 5ml
Control, no added
S02
Control, no added
S02
12
10
*Via titration with iodine.
**0xidation and precipitation as
1-7
-------
B. Oxidation of SO^ by the Filter
In order to test the possibility that oxidation of S02 by the filter and
subsequent scrubbing of the resultant SOo by the impingers could account
for the sulfate in the impingers, two different experimental approaches
were employed. In the first, air was bubbled through water and then
joined with a sulfur dioxide in nitrogen (940 ppm) stream in the proper
proportions to give a final gas mixture containing 45 ppm SOp flowing
at 0.5 CFM. This gas mixture was passed through' a preheated stainless
coil (1/4 in. O.D. x 20 ft) and then through a metal filter holder con-
taining the filter to be studied, both contained in an oven maintained at
250°F. The exit gas stream was then bubbled through an impinger contain-
ing 200 ml of 80% isopropyl alcohol cooled in an ice water bath (see
Figure I-2a). Following the one-hour run, the impinger solution was
warmed to room temperature and air bubbled through it for ten minutes to
remove most of the dissolved S02- The results of these tests using
filters collected during the incinerator sampling, as well as a blank,
.are given in Table 1-4. It is clear from this data that not very much
SOo was collected in the isopropanol impinger, nor were measurable amounts
of sulfate being formed on the filter, based on the change in weight
data. Based on this data, for a standard 60 SCF sample, possibly 2-3 mg
sulfate would be collected by the impingers due to filter catalysis from
45 ppm SC>2 gas stream. We assume that even lesser amounts would result
from a 5 ppm stream. To ensure that 862 was in the gas stream at the
desired levels, one run was made with an impinger containing hydrogen
peroxide scrubbing solution. Roughly 95% of the calculated S02 for a
45 ppm stream was recovered as sulfate in the impinger, confirming that
the S02 was present as planned.
Because of the lack of evidence for 863 in the impinger from the above
experiment, we were concerned about the possibility that either the metal
filter holder may have reacted with the SOo, or the isopropanol impinger
was very inefficient and low SOg recoveries were obtained. To study the
possibility that the metal system and isopropanol impingers had an effect,
a second set of experiments were performed using the same gas mixture and
preheating system but then passing this mixture into the heated all-glass
filter holder on the RAG sampling train. Flows were maintained at 0.5
CFM for one hour for each run and scrubbing was accomplished via the use
of the normal three impingers arrangement in the RAC train (water at ice
bath temperatures). Schematic representation is shown in Figure I-2b.
The results of these experiments are given in Table 1-5. In this case,
there are two runs whereby significant quantities of sulfate were
found. The reason for not finding much sulfate in the other two runs
is not clear. Possibly, the system needed equilibration to overcome
initial buildup of SOo, or else some unknown variable was out of control.
1-8
-------
I
VO
TABLE 1-4
Results from FirstS02 Filter Oxidation Study
Code
(NI- )
—
—
—
74-1
28-1
13-1
80-1
Filter
Degree of
Loading (mg)a
Blank
Blank
Blank
293
231
246
108
Loss in
Weight (mg)
-2.5
-3.4
-4.1
-12.2
-4.6
-2.2
-6.4
Found (mg) in the Isopropanol
Solubleb Total S0xu
SO, as SO^
0.1 1.4
0.5 1.3
1.1
1.0
0.9
1.1
1.9
Impinger
Average
0.8
0.4
1.1
a Each filter lost weight during the run.
b Analysis: titration with iodine.
c Analysis: Oxidation and precipitation as
d Corrected for 0.5 mg sulfate blank due to soluble S02-
-------
TABLE 1-5
Results from Second S0n Oxidation Study
i
M
o
Filter
Code
24-1
46-1
78-1
14-1
Loading
Blank
Blank
870
436
153
237
so2
0.3
0.5
0.5
0.5
0.5
0.3
^
Found
Total SO,,
8
6
7
8
25
22
*
in Impingers (nig)
SO,, due to SO,"1"
5
3
4
5
22
19
Average
4
5
21
*Impinger solutions (210-230 ml) boiled and evaporated to less than 200 ml and then diluted
to 200 ml for analysis.
~*~Included rough correction for residual S02 and 2 mg SO^ due to oxidation.
-------
a)
Water
Saturater
Gas Mixing and
Heating Coil
Metal Filter
Holder
(254-274°F)
b)
EPA Glass
Filter Holder
@ 250°F
Vent
Isopropanol
Impinger
@ Ice Bath
Vent
EPA Impingers
@ Ice Bath
FIGURE 1-2 : SCHEMATIC OF APPARATUS EMPLOYED FOR SO2/SO3 OXIDATION STUDIES
-------
Because the results of these experiments were contradictory, further work
was performed. In order to avoid potential problems of 803 adsorption on
metal, this work was done with an RAC sampling train all glass system (from
the filter on) and, to decrease the solubility of S02 in the collection
system and more closely simulate the final impinger composition, we switched
to 0.01 N HC1 in the impingers. In actual field practice, the impinger
becomes acid very quickly so that a solution at pH of 2-3 is probably more
realistic than using distilled water with a pH of 5-7. The apparatus for
this work is shown schematically in Figure I-2b.
In the first series of runs, the 45 ppm S02 stream was passed through the
RAC train assembly for 1 hour at 0.5 CFM. At the completion of the run,
the water in the impingers was combined and evaporated via heating at about
90°C to less than 200 ml and then diluted back to 200 ml. Portions were
then analyzed for 802 via iodine titration and for total S02 by means of
precipitation with barium chloride after oxidation with bromine (see series
A, Table 1-6). For the second series of tests, Series B in Table 1-6, the
experimental runs were duplicated except that the impinger solutions
were analyzed directly rather than after an evaporation and a concentration
step. Conclusions from these runs were:
• Large quantities of sulfate were not collected in the impinger—
by any mechanism—when a 45 ppm S02 gas stream was passed over a
heated glass filter system at 250°F for as long as four hours.
Our experiments have shown that between 3 and 6 mg of sulfate
was collected; this would account for, at most, 10-30% of the
total sulfate found in the impinger solutions of a test run at
an incinerator. (The precision of these analyses is probably
such that the differences shown in Table 1-6 between series A
and B, should not be considered very significant.)
• Most of the sulfate may, in fact, have been soluble S02—not
803 collected by or produced in the impinger so that, it may be
that the 804 was produced by the sample handling procedures in
the analytical methodology.
• No differences were observed with or without a glass filter in
the system.
C. Influence of Sample Handling on Oxidation of SO?
To learn how the sample handling influences the levels of sulfate found,
another set of experiments were performed. In these, unheated S02 in air
at two levels (45 ppm and 10 ppm) was bubbled directly into the impingers
containing 0.01 N HC1 (at ice bath temperature) for one hour. The water
samples were analyzed immediately for S02 via iodine titration and then
evaporated to 5-10 ml and analyzed for total sulfate via barium precipita-
tion. The purpose of this test was to see how much 802 is lost during
evaporation and to examine the effect of S02 concentration in the gas
stream. The results of this study (see Table 1-7) confirm that:
1-12
-------
i
M
Co
Sample Conditions
Series
Filter Used
None
Blank
NI-24-1
NI-14-1
None
Blank
NI-24-1
NI-14-1
TABLE 1-6
Filter Oxidation Studies
Found in Impingers (mg)
as SO,,
0.4
0.5
0.3
0.3
4.6
3.8
3.6
2.9
so?c
SO, Equivalent
—
-------
TABLE 1-7
SO- Analysis Studies
S02 Level
in Gasa £>00 Sulfate Equiv. Evaporation0 Average
(ppm)
45
45
H 45
M
10
10
so_
2.1
3.3
3.6
5.1
1.6
0.7
1.1
1.1
— t.
Found in Impingers
S02 Levelsb
Sulfate Equiv.
3.2
4.9
5.4
7.6
2.4
1.1
1.7
1.7
(mg)
Sulfate after
Evaporation0
3.8
4.4
2.7
7.9
*
4.9
1.6
1.9
1.3
L 7
/c *» /• tr/_ *•» -T -r • /
1.6
j.,1 L.I ±.y
10
a - Estimated from flow rates.
b - SC>2 determined via iodine titration and sulfate equivalent calculated.
c - Impinger solutions evaporated to 5-10 ml and then total SO determined gravi-
metrically after oxidation with bromine.
* - Not included in average, if did the average would be 2.4 and conclusion essentially
the same.
-------
• Most of the S02 solubilized in the impingers at the end of the
sampling run remained available to be measured as sulfate even
after concentration of the impinger solution to a few ml. There-
fore, to explain the low loss of 862 during evaporation, most of
the S02 must have been oxidized fairly rapidly during the evapora-
tion step.
• The level of SC>2 in the gas stream had an influence on the S02
dissolved and thus the residual sulfate found in the impingers
(higher SC>2 leads to higher sulfate) . Thus the 45 ppm stream
gave the same results as in Table 1-6. However, with less SC>2
in the gas stream, the 10 ppm stream caused less to be absorbed by
the solution.
• The amount of sulfate generated in the impingers due to S0~ solu-
bilization will probably be low when the EPA train is used for
incinerator testing—because the S02 levels in the gas stream tend
to be low.
D. Combined Field and Laboratory Tests
To aid in our understanding of what happens to sulfur oxides in the sampl-
ing system, we performed a field test whereby the sulfur oxide levels
were measured at the same time as samples were taken using the EPA sampl-
ing train. Because it was operationally easier, we used two complete
sampling systems with the nozzles directly adjacent to each other in the
stack. Our assumption was that serious concentration differences within
a one-foot area in the gas stream, although possible, were not likely.
503 and 502 were collected separately via a controlled condensation col-
lector and a peroxide impinger respectively. (The resultant sulfate was
determined by the conventional barium titration procedure with thorin
indicator.) Sulfur dioxide and sulfate levels in the EPA train impinger
solutions were determined by iodine titration and barium precipitauion
respectively. The results of these experiments are given in Table 1-8
and they led to some interesting observations:
• The amount of sulfate in the impingers did not correlate with
the measured levels of SOo in the incinerator effluent.
• The collection of total SO in the impingers was about 35%
except for one case of 90%.
Developing a clear explanation of these results is not easy. The high
variability of the measurement of SO, at these low levels using the
controlled temperature condensation technique is one possibility. Thus,
maybe the SOo results are in error and the gas stream really had more
803—enough to account for what was found in the impingers. This
certainly could be the case for NI-86,87 and 88. However, it is an
unlikely explanation for NI-85. Either that run was a fluke or oxida-
tion of S02 must have occurred early in the sampling train and the
1-15
-------
TABLE 1-8
Determination of Sulfur Oxide Levels
in Incinerator Effluents and Impinger Solutions
Sample
No.
NI-85
NI-86
NI-87
NI-88
Concentration
in Stack Gas
(ppm)a
sp_2
26
21
16
15
S03
2
1.6
4
2
Measured
Sample
Volume
1.3
1.4
1.3
1.5
Calculated Available Sulfate
S02 and SO^
142
120
84
95
SO, as SO,
j H
11
9
21
13
b , V
(mg)
Total
153
129
105
108
Sulfur Oxides
in Impinger c
SO,,*
1
8
7
7
SO, e Total
H
139 140
38 46
28 35
32 39
Percent S0}
Found in
Impingers
90
33
33
35
a - Determined on separate gas sample collected at same time as EPA sample and close to same sampling
location.
o
b - Calculated based on S02 level and volume sampled (1 ppm SO,, represents 4.2 mg sulfate per m sampled)
c - Collected via standard EPA sampling procedure.
d - Determined by titration with iodine.
-------
resultant SO^ collected in the impingers. (We could not make this happen
in the laboratory but possibly field conditions were different.)
In an effort to help explain this dilemma, an experiment was performed
whereby an SC>2 containing gas stream was bubbled through impinger solu-
tion collected at an incinerator in another attempt to generate signi-
ficant amounts of sulfate in the sampling train in the laboratory. The
experimental procedure was similar to earlier work except that an effort
was made to utilize extreme conditions in the hope of generating some
sulfate. Thus, for this work, we used a 45 ppm SC^ in nitrogen gas
stream and the setup shown in Figure I-2b.
The gas stream, where temperature varied from 206-318°F, was passed at
a flow rate of 0.6 cfm for 4.7 hours over a filter previously used in
the field (Run No. NI-14) at 250°F, and through impinger solution from
Run No. NI-88. After determining the S02 and sulfate levels in the
original impinger solution, at the end of this run, the impinger solu-
tion was re-analyzed for sulfate and residual S02- The results were:
Sulfate
Amount Found (mg) Corrected
Sulfate Residual SO? for S00
-1 "* " ' —^—^— Z1 •
After 48 13 35
Before 32^ 11 21_
Difference 16 14
Looking at either the corrected or uncorrected (for residual SC^) sulfate
values, it would appear that a little sulfate (14-16 mg) did build up in
the impinger solution. However, when it is recognized that the total SO?
passed through this solution represented roughly 900 mg of sulfate, it
can be seen that the sulfate pickup in the impinger was not very signifi-
cant. Thus, we have been unable to duplicate conditions in the laboratory
whereby appreciable sulfate is collected in the impingers due to oxida-
tion of S02 by the filter, filter holder or impingers.
E. Distribution of Sulfate in Sampling System
A presentation of the distribution of sulfate between the various parts
of the sampling train for the four runs used for thorough quantitative
measurement is given in Table 1-9. As can be seen from this data, the
sulfate was fairly evenly distributed between the probe/cyclone, filter
and impingers—but on the average more in the latter. Whether all of
the measured sulfate was due to SO^ and S02 or includes solid sulfates
collected on the filter and probe in unknown.
1-17
-------
TABLE 1-9
Distribution of Sulfate in Sampling Train
Sample
(NI- )
77
H
^ 78
oo
79
80
Average
Distribution
Probe /Cyclone
28
16
28
28
25
of Sulfate
Filter
40
42
34
26
35
Impinger
32
42
38
46
45
Total Weight
Found (mg)
70
103
93
61
82
-------
III. VAPOR PRESSURE DATA FOR S(H
The dew point of small amounts of SOo in the presence of excess amounts
of water vapor has been predicted from theoretical thermodynamic cal-
culation by Muller^ and confirmed experimentally by Lisle and SensenbaugbA
From this work, the relationship between SO, concentration and temperature
is: J
Temp. (°F) S03 (ppm)
230 0.4
250 2
270 9
In the average sample taken for this program, one ppm SCL represents
6.5 mg sulfate.
IV. DISCUSSION
It has always been clear that SCU will be collected by the sampling train
in one way or another—the only question has been where. Our best
estimate at this time is that, due to vapor pressure limitation, no_
more than 10-15 mg of sulfate in our "standard catch" can be attributed
to SO-j originally in the gas stream. This assumes a 2 ppm vapor pres-
sure for 503 at a filter temperature of 250°F (the actual temperature
has varied from 210-270°F). During all of our experimental field
measurements, the measured 803 level has been above 2 ppm only once
(4 ppm), so that a 10-15 mg maximum probably is real. Because of its
high reactivity, most, if not all, of the 803 probably reacts with
materials on the filter or probe and never reaches the impingers. In
any case, this amount of sulfate added to the filter is a relatively
small percentage (5-10% of the total 150-250 mg) so that the question as
to whether the sulfate on the filter (or probe and cyclone) came from
803, S02> or particulate sulfate is not too important for incinerator
effluents.
Our laboratory results do not support the hypothesis that the filter at
250°F or the impingers themselves will convert significant quantities
of S02 to 503 or sulfate. Thus, even in the most extreme case, very
2. "A Contribution to the Problem of the Action of Sulfuric Acid
on the Dew Point Temperature of Flue Gases." P. Muller, Chem. Ing. Tech.,
28., 279, (1956).
3. "The Determination of Sulfur Trioxide and Acid Dew Point in
Flue Gases." E.S. Lisle and J.D. Sensenbaugh, Combustion, 36 (7),
January, 1965,
1-19
-------
little sulfate was formed during passage of SC>2 over a hot filter or
through an acidified solution. This was true even though we employed
metal or glass filter holders, clean glass filters or filters which con-
tained particulate catch from an incinerator. Even though the impingers
contain practically no mineral ash, it may be that the actual impinger
solution acts as a catalyst in a way that we have not duplicated in the
laboratory. Another possible alternative is that the probe, which is
generally hotter than 250°F acts as the catalyst for the oxidation.
This is the only portion of the sampling train that was not duplicated
in the laboratory.
1-20
-------
APPENDIX J. SOME IMPORTANT CONSIDERATIONS FOR SAMPLING OF PARTICULATES
Table of Contents
I. INTRODUCTION J-l
II. TESTING AN INCINERATOR J-2
A. Pretest Planning J-2
B. Source Factors J-2
1. Flue Gas Temperature J-3
2. Pressure J-3
3. Velocity J-3
4. Uniformity of Flow J-3
5. Flue Size J-4
6. Gas Composition and Water Content J-4
7. Process Parameters J-4
8. Control Devices J-4
9. Working Conditions J-4
C. The Sampling and Collection Train J-5
1. Nozzle J-5
2. Probes J-7
3. Collectors J-9
4. Flow Meter J-13
5. Pumps J-14
6. Supporting Element J-16
J-l
-------
APPENDIX J
SOME IMPORTANT CONSIDERATIONS FOR
SAMPLING OF PARTICULATES
I. INTRODUCTION
This chapter will focus on the needs and problems associated with source
factors and equipment used in characterizing particulate matter. Our
intent in doing this is to clearly delineate objectives and goals
associated with the determination of particulate loadings, the pitfalls
that one may encounter by the use of various pieces of sampling equipment,
and the tradeoffs that are available when deciding to use one device
rather than another. In addition, this discussion will serve as a back-
gound for the specific evaluations of individual methods which was given
in Appendix C.
Throughout this discussion it is important to remember that the basic
objective in measuring particulate concentrations in the emissions from
incinerators is to answer such questions as:
• Is the combustion process generating large quantities of pavti-
culate matter?
• How effective are the Air Pollution Control devices?
• What amount of particulate matter is being emitted to the
atmosphere?
A major problem in attempting to answer these questions is defining the
term particulate. Does it include only inorganic solids that are stable
at high temperature? hydrates? anything that will condense upon leav-
ing the stack? If condensibles are to be included, then how far from
the stack can condensation occur and what temperatures are relevant?
Such considerations serve to illustrate that the definition of particu-
late matter has a great influence on what methodology which can be con-
sidered. Traditionally, interest has been in the mass of dust emitted
which contributed to dust fall (solid particles in the gas stream
>l/ym). At the present time there is interest in many other effects
of air pollution and, as such, the size spectrum is widened. Cloud
formation in the atmosphere is related to nuclei population; atmospheric
visibility and haze are related to strong scattefers (0.1 to 1 ym);
toxicity is frequently interpreted in terms of a "respirable" fraction
(approximately 0.1 to 5 um); the mass of material emitted to the atmo-
sphere is in the larger sizes (> lym).
J-2
-------
Specific reference to sampling protocol or recommendation of specific
equipment has been deliberately omitted. Methods for testing incinerators
have been published in the Federal Register, December 23, 1971 and
provide complete detailed descriptions of an acceptable methodology.
The sections which follow will permit the reader to appreciate the many
considerations inherent in the specification and use of the Federal
Register methods.
II. TESTING AN INCINERATOR
A. Pretest Planning
Pretest planning considers the objectives of the test against the nature
of the operation under study. Such factors as steady-state vs. transient
operation, lengths of cycles, peak periods of emission, estimated gas
composition and pollutant concentrations, gas temperature and pressure,
duct size, gas velocity and humidity are examined to select an appropriate
collection method, method of analysis, sampling rate, sampling interval,
number of tests required for proper statistical analysis, as well as to
consider less obvious needs, such as special materials of construction
(temperature or corrosion effects, etc.). Such mundane factors as avail-
ability of ladders, scaffolding, sampling ports, lighting (for night work),
and coordination with operating personnel must be included in the test
plan.
Adequate planning increases the probability for representativeness of
the sample which is the real measure of success in the tests to follow.
The alternative to pretest planning is the too familiar trial-and-error
method which leads to extended experimental programs, overlooked key
data, and a generally more costly and less representative result.
B. Source Factors
Incinerator sources must be examined to evaluate those operational and
configurational parameters which influence the sampling process or are
needed to complete test computation. These factors include:
Flue gas temperature
Pressure
Flow velocity
Uniformity of flow, - gross geometry, bends, length
Flue size - cross sectional geometry
Flue gas composition and water content
Process parameters - e.g., for combustion, excess air, rate, type
of firing, firing rate, waste composition (heat content)
Control devices
Working conditions
The relationship of each factor or its influence on sampling configura-
tion or procedure is discussed below.
J-3
-------
1. Flue Gas Temperature
The temperature at the sampling point determines whether condensibles
are sampled as gases or particles. This obviously influences the choice
of sample train components and probe temperature (heated or cooled) and
the choice of materials of construction. Temperature also influences
gas density and must be considered in velocity and mass flow calculations.
2. Pressure
Pressure at the sampling point determines (with temperature and composition)
the density of flue gas and is reflected in the correction factors needed
to convert measured values to standard conditions.
3. Velocity
Velocity is, like temperature, a very important parameter. It is neces-
sary to know velocity to select a proper size probe and also to adjust
the sampling flow rate for isokinetic conditions (necessary to eliminate
sampling errors at the inlet). Calculations of emission rate are based
on velocity traverses and flue area.
The selection of an instrument to measure velocity will depend on the
magnitude. For the velocities characteristic of many large instal-
lations (velocity > 10 ft/sec), pitot tube measurements are most gen-
erally made. Pitot tubes with included monometers are not sufficiently
accurate for velocities much below this value and alternative methods
are used.
4. Uniformity of Flow
Closely related to the magnitude of velocity is its uniformity
• Across the sampling plane
• As a function of time at a given point.
The spatial uniformity depends primarily on the configuration of the flue,
duct, chamber or whatever conduit containing the effluent is being
sampled. There is perhaps more variability in this factor than in any
other. Every large installation is more or less unique. One is gen-
erally advised to sample several diameters downstream and a few diameters
upstream of any flew disturbing element. In practice it is rare to find
an ideal location.
Temporal uniformity depends upon the type of operation (batch or continu-
ous) and is very important in sampling.
J-4
-------
5. Flue Size
The dimensions of flue or duct determine the number of samples that must
be taken across the sampling plane to obtain a representative measure.
Small rectangular ducts require fewer measurements to obtain the same
degree of accuracy as large ducts. Accuracy for a given number of
points is not a function of size for circular ducts. Interference pro-
duced by the introduction of the sample probe may be a factor in small
units.
6, Gas Composition and Water Content
Gas composition must be known to completely characterize gas flow and
thus the total emission. In the case of combustion effluents, composi-
tion data are also needed to correct the measured values for particulate
loading to some normalized basis, e.g., 12% CC^, 50% excess air. Water
content (dew point) must also be considered to prevent condensation
before particle collection. Condensation can lead to plugging of the
collector and/or wetting of the sample.
7. Process Parameters - e.g., Incineration Processes -
Excess Air, Waste Composition, Heat Content, Type
of Firing, Load
Data on these factors is needed to provide a complete record of the con-
ditions present when the test is made. Type of waste, method of firing,
and load are major determinants in the type and magnitude of emissions.
During startup or load changes, excessive emissions can be produced.
8. Control Devices
The application of any control devices will change both the total load-
ing and particle size distribution in an effluent stream. Tests on units
with high efficiency collectors sometimes require different sampling
configurations or test procedures at the inlet and outlet due to the
large reduction in loading. Furthermore, many dry collectors
operate at elevated temperatures and would have little or no effect on
condensibles.
Wet scrubbers introduce a different type of problem. The effluent is
cooled, saturated, and often loaded with liquid water. Means to separate
the water and prevent filter plugging are required.
9. Working Conditions
Difficulties and hazards result from the necessity of working on poorly
accessible, often elevated, poorly lighted, cramped platforms in all
weather conditions. Also, size, weight, fragility, reliability of the
sampling system, long duration of sampling test program, and distance
J-5
-------
of sampling team from home base are practical factors that will be of
utmost importance. In spite of these hardships, it is essential that
capable and experienced personnel be directly involved in the field test
program.
C. The Sampling and Collection Train
In this section, the individual components normally associated with
sampling and collection of particulates will be discussed. The basic
reason for setting up and utilizing a sampling train is to obtain a
sample that is "representative" of that in the original environment.
As discussed in Section B above, meeting this objective is dependent on
a wide range of factors. Within the context of this program, we aimed for
all matter which would be a particle or condensible at 70°F. Therefore,
as the individual sampling train components are discussed, they must be
viewed in the light of this very specific definition.
The apparatus used for the sampling of all stationary combustion sources,
including incinerators, can be divided into modular components includ-
ing: sampling nozzle, probe (sometimes temperature controlled), collec-
tion device, water condenser and/or desiccant column, sample flow meter,
(gas temperature, pressure and water content at the meter are essential),
and air mover.
Each component is discussed below.
1. Nozzle
The first unit of a sampling system is the nozzle which is connected
directly either to a probe or a collecting device. The composition of
the materials used to construct nozzles is also important. Nozzles
are generally made of stainless steel, and less frequently of glass or
brass. Vycor glass is capable of withstanding the high temperatures but
the rigors of field work suggest non-glass nozzles. Hydrofluoric acid
is one of the few components in the gas stream which might affect glass
but it probably is not present in the effluent in high enough concentra-
tion to be a determining factor. Brass has been used (because of ease
in machining) but because of corrosion and possible reaction with sampled
gases, is not considered to be appropriate. Lining the nozzle with
porcelain or some other material might be considered but the material
would have to be able to withstand large changes in temperature and the
differential in coefficient of expansion would have to be small to pre-
vent the lining from spalling off. In general, the use of high nickel
alloy stainless steel offers the optimum in serviceability compromise.
It eliminated the breakage problem and since the nozzle is short, chem-
ical reactions with the stainless steel are minimized.
J-6
-------
Shape of nozzle is important to sampling accuracy both in respect to
providing passage for the solid particles into the sample and also the
creation of turbulence in the immediate area of collection. Whiteley,
and Reed^ conducted a series of experiments on the effects of the
shape of the sampling nozzle on the accuracy of sampling under isokinetic
conditions. Using tubing of 5/8" O.D. and 1/2" I.D., they constructed
several nozzles, each with a different angle of chamfer and tested them
against the collection accuracy of a nozzle with a 15° chamfer.
A nozzle with 120° angle of chamfer showed only 0.5 percent variation
in accuracy from that for the 15° angle; the 180° (flat) nozzle showed
a variation of +10% in collecting both large and small particles.
In the above series of tests, the upstream or right angle projection of
nozzle-probe combination was 3-5/16" long. From a subsequent series of
tests it was found that the projection of the nozzle at right angles
to the probe should be greater than 1-1/4" and probably should be as
long as 3-5/16".
The diameter of sampling nozzles is generally predicated on other factors
inherent in the sampling system. Generally the nozzle area is chosen
to achieve isokinetic conditions within the pumping capability of the
system for the particular stream velocity. Nozzles in a size range of
0.250 to 0.625 inches in diameter are used extensively as many sampling
systems are designed to sample at a rate of approximately 1 SCFM. These
sizes, coupled with the flexibility in the pumping source, are well
suited to achieve isokinetic sampling conditions in most situations.
Nozzles of larger size, 0.75 to +1.25 inches in diameter, have been
used on systems for sampling greater volumes of gases. There are no
upper limits on diameter of a nozzle other than the capability of a
system to pump and handle a large volume rate of sample and the dimen-
sions of the duct.
A minimum of 0.250 inches diameter for sample nozzles has been stipulated
in ASME Power Test Codes.
A second type of nozzle which in theory is self-adjusting has been used.
The so-called "null" type nozzle contains chambers surrounding the tip
in which static pressure just inside and just outside the nozzle is
measured. In theory when the sample flow is adjusted so that the two
are equal, isokinetic conditions are achieved. In practice, significant
errors are introduced by differences in flow, inside and out, surface
rouehness.. etc. Toynbee are Parkes^ ' have developed a null orobe
design for which they claim almost no deviations from isokinetic flow.
1. A.B. Whiteley and L.E. Reed, "The Effect of Probe Shape on the
Accuracy of Sampling Flue Gases for Dust Control," Journal of the
Inst. of Fuel, 32, pages 316-20, (1959).
2. P.A. Toynbee, W.J.S. Parkes, Intern. J. Air Water Pollution,
16, 13, (1962).
J-7
-------
However, an extremely sensitive micromanometer is required to measure
the balance of the static pressure tubes and considerable flexibility
as well as very fine adjustment is needed in the suction source. Any
blocking or deformity of the static holes introduces error. Null
nozzles must be calibrated at approximately the temperature and velo-
city (Reynolds Number) at which they are to be used. In general, the
null nozzle does not offer an advantage in precision or accuracy over
other nozzles, and, in fact, will introduce errors through use outside
of its calibrated range.
In addition, orientation of the nozzle relative to the flow lines of
the gas stream is important as shown in Figure J-l. The ratio of the
indicated concentration (c) and true concentration (Co) is plotted
against the angle between the nozzle and the flow lines. Results for
several particle sizes are shown.
Therefore, in summary, for optimum sampling, isokinetic conditions are
necessary and the nozzle should be constructed of a high nickel alloy
stainless steel, be between 0.25-0.6 inches in diameter, have between
15° and 120° chamfer tip, be 2-3" long and be pointed into and parallel
to the gas flow.
2. Probes
Probes are of two general classes, simple and temperature controlled.
Probes are frequently heated to prevent losses in sample lines through
condensation. Alternatively, one can use a simple probe, but one must
use very careful handling techniques and quantitatively wash condensate
and precipitate from the probe and sampling lines. This material is
then combined with the material in the collector. Cooled probes are
used where one is sampling at temperatures above 800°F. Generally,
the probe is introduced to a stack at right angles to the direction of
flow. This requires a 90° bend so that the nozzle will face the
nozzle into the flow. This should be a smooth 90° bend and free of burrs
or imperfections. The diameter of the probe is such as to maintain trans-
port velocity for particles and to minimize the time of contact with
corrosive gases.
Stainless steel probes are available commercially and are frequently
used. They have the advantages of durability, strength and ease of
attachment to other components. However, stainless has the disadvantage
of a high potential for reaction with sample gas—especially acids, and
because of the length of the probe (frequently up to fifteen feet)
minimizing corrosion is important. Reaction with the sampled gases
tends to cause pitting and deterioration of the probe with increasing
problems of cleaning and introduction of error.
Glass probes have been used on extremely corrosive gases and have the
advantage of being relatively inert except for reaction with fluorides.
Because these are not a common constituent in incinerator effluents,
this may not be a service problem. Glass probes have the disadvantage
of fragility and they are difficult to connect to other components. The
latter can be minimized by using special joints but repairs in the field
are made much more difficult.
J-8
-------
l.Q -
0.8 -
0.2 -
30 60 90 120 150 180 (degrees)
FIGURE J-l
Angle Between Nozzle and Wind Direction
J-9
-------
A glass-lined Vycor stainless steel probe combined the advantage of
strength of the one and corrosion resistance of the other. However,
breakage of the glass-lined steel probes due to thermal or mechanical
stresses coupled with its high initial cost imposes considerable hard-
ship on the users.
The length of the probe will vary and will be determined by the stack
diameter (six-to-thirty feet) and the needs of the sampling program.
The longer the probe, the more tendency for collection of particulates
in the probe due to deposition and agglomeration. Deposition of particles
in the probe can be lessened by using a boundary layer dilution probe.
The dilution system would increase the complexity of the sampling system
and none are known to be in commercial use today.
3. Collectors
The collector is the device that is used to separate the particles from
the gas sample and to retain the particles for subsequent analyses or
treatments.
a. Filters
Filtration is the method of broadest applicability and is the most used
for mineral particulates and soot (those particles that are already
formed and are stable at stack temperatures). In practice, a cyclone is
sometimes placed in front of a filter to minimize problems of plugging.
This practice generally limits the size of particles collected by the
filtering media to less than 3 to 5 um.
A wide variety of filter media are available permitting selection based
upon factors such as filtration efficiency, pressure drop per unit flow,
inertness, composition (glass, cellulose, alundum-refractory, asbestos,
polymeric materials) low background contamination (for special analytical
needs), color, solubility, thermal stability, price, etc. The effi-
ciency of a filter generally improves with the accumulation of collected
material but eventually the resistance can become so great that adequate
gas passage is impaired.
Filter media may be paper, glass fiber, alundum, cloth, membrane, etc.,
with the material selected being that best suited to sampling conditions,
collecting efficiency desired and analytical procedures to be followed.
Paper, glass fiber, and alundum (aluminum oxide) are used extensively
for thimbles. The alundum filter has the advantage of being able to
withstand temperatures of over 1000°F. Both the alundum and glass
fiber filters are relatively non-hydroscopic and require less time for
humidity conditioning prior to weighing than cellulose filter media.
Filter media can be used over a wide range of sampling rates and, if
properly selected, perform very satisfactorily. Membrane filters
collect the particulate matter on a flat surface which can then be
examined microscopically. They are used extensively for particle size
determination. They are restricted to relatively low sampling rate
J-10
-------
systems, e.g., liters per minute rather than cubic feet per minute.
Furthermore, they are limited to low temperature applications. The 47 mm
size is about as large as is commonly used in this application and the
cost of membrane filters is still considerably more than other media.
Cloth is used infrequently as a collecting medium as it is inferior to
other materials for small particle collection. However, it has some
application in large high-volume systems.
Filters can be placed either inside or outside of the stack. The former
introduces difficulties due to temperature and an increased potential
for reaction between the collected material and the gas stream. The
latter introduces errors due to sampling lines losses. In general, a
drop in temperature to 300°F to 250°F (via cooling of the gas) and
placing the filter outside will reduce the magnitude of chemical changes
in the sample during the collection step—especially if organics and con-
densibles are involved. On the other hand, this same step (lowering
the temperature) increases the likelihood of hydration and other forms
of chemical reaction occuring with the mineral particulates.
A further more detailed discussion of high efficiency glass fiber filters
is given in Appendix G.
b. Classifiers
A classifier is a device which separates particulates into one or more
size fractions. One type, the cyclone, is used before the main collection
device (filter or impinger) to minimize overloading and plugging, which
can cause loss of sample and changes in flow rate. Cascade impactors,
such as the Anderson sampler, also is a type of classifier. This is a
multijet, multistage impactor which collects weighable quantities of parti-
culate matter. High efficiency is reported for submicron particles and
above. The practical upper limit of size is < 50y. These impactors
generally must be operated at constant flow rates, have very low capacity,
and are much more expensive than a cyclone.
Cyclone collectors are used extenstively as a precollector or classifier
unit in a sampling system. Cyclones are efficient above a minimum particle
size; some commercial designs operate down to a 3 ym cutoff and research
sampling devices are designed to submicron cutoffs. They are relatively
inexpensive and have no moving parts to wear out, can be used both
inside and outside the stack, and they can be obtained in sizes and cap-
acities from a few liters per minute up to a number of cubic feet per
minute.
If the gas stream contains water droplets, those larger than the cutoff
size will be collected by the cyclone along with the particulates. This
can cause great difficulty if it is desired to use a cyclone for collec-
tion of particulates for characterization. Probably all one can do is
to collect the "mud", dry it and weigh the residue. Possibilities for
J-ll
-------
prior interactions with the water (chemical reactions, hydration, decomposi-
tion) are too high to make individual particle characterization meaningful.
The cyclone can be placed either inside or outside of the stack. When
placed inside right after the nozzle, it should minimize particle losses
in the probe (due to gravitational and turbulent deposition mechanisms)
by removing large particles immediately. Because the particulates fall
out of the main stream during collection in a cyclone, and thus are
collected primarily out of the flow of the gas, there should be little
interaction with the gas stream itself (if dry). Decomposition of the
sample after collection in a cyclone, due to longer residence in the stack
is a possibility, but may not be serious since the sample probably has
been exposed to much higher temperatures in the incinerator fire box
(1000-2200°F). The rate of chemical reactions of the gases after cooling
to stack temperature should be considerably lower. Combustion and pyrolysis
may still be a problem, however, for organic compounds and carbon parti-
cles especially where the stack gas is very hot. An extra advantage of
a cyclone inside the stack wall is that the pressure drop across the
cyclone can be employed as a flow measurement at stack conditions.
When placed outside, the chances for chemical decomposition (or reaction)
are reduced—especially if the stack gases are cooled first. However, the
advantages derived from the ability to measure flow at stack conditions
and the removal of large particulates before entering the probe will have
been lost.
A possible solution to the problems associated with handling a wet gas
stream in a cyclone is to dilute the flue gases ahead of the cyclone.
This would get the dew point of the gas stream to a point where liquid
would not collect in the cyclone. It would not solve the situation where
water droplets were already present before dilution.
As to materials of construction, the same problems would apply here as
with the nozzle and probe.
c. Impactors/Impingers
A number of collectors are based on impaction. When the particles are
collected on a target submerged in a liquid, the device is called an
impinger although it is essentially a single stage impactor. Impingers
have a possible advantage in that they collect gases too. Disadvantages,
of course, are caused by solubility, evaporation of the collecting
liquid and the nuisance factor introduced by the requirement to handle
liquids in the field (freezing, spilling, etc.)
The wet impingement device used is usually the Greenburg-Smith impinger
with the collecting liquid being initially distilled water. In the
Greenburg-Smith impinger gas is drawn through an orifice 2.3 mm in diameter
at a rate of approximately 1 cfm (0.028 m^/min) and impinged against a
J-12
-------
flat surface 5 mm distant. The particles are retained in liquid which
fills the irapinger flask to a depth of about 2 inches.
When an impinger contains a liquid, it can become a very efficient col-
lector for aerosol particles, very fine (submicron)particles, condensibles
(organic and inorganic) and soluble gases. Sometimes the impinger catch
is thought to be synonymous with condensibles but clearly that is not
necessarily the case since there is a high likelihood that gases dis-
solve as well as condense in the impinger solution. The impinger can
contain water (either purposely or as a result of condensation), a reactive
solution (such as sodium hydroxide), and organic solvent (such as an
alcohol or benzene), or a dry solid (such as silica, charcoal or alumina).
In some situations, a filter can be placed after the impinger to ensure
collection of aerosol or entrained particles.
The impinger is usually operated at ambient temperature or less within
the limits of the properties of the collecting liquid. The lower the
temperature, the more efficient the collection efficiency for the so-called
condensibles.
d. Other
Electrostatic precipitation is also used for particle sampling. This well
known method passes particles through a charging section and then preci-
pitates the charged particles in an electric field. High efficiency
performance is easily attained even for small particles. The pressure
drop across a precipitator is low since there is little to obstruct the
flow. A high voltage power supply is required and presents a safety
hazard. Furthermore, collection efficiency is dependent upon particle
properties (size, dielectric properties) which makes it independable in
variable situations such as incinerator sampling.
The principal of thermal precipitation has been exploited in a classical
air sampling instrument and is still used extensively in research applica-
tions but generally not in field work. The thermal precipitation forces
are very gentle and liquid droplets or fragile agglomerates are not
likely to shatter when they contact the collecting surface. Furthermore,
they are very efficient for submicron particles and have found use for
collecting low vapor pressure and fogs. The original instruments were
limited severely by a low sampling rate (10 cc/min). Other units have
been designed which employ high flow rates and time resolution of the
sample. Due to the high degree of control required and the potentially
low efficiency, they are not practical for incinerator sampling.
Particles can also be collected by gravitational sedimentation upon
horizontal surfaces. Devices have been constructed of stacks of closely
spaced horizontal plates to minimize the sedimentation distance and
increase the gas feed rate. Small units have been used as collectors
in atmospheric sampling. They are, however, extremely slow and bulky since
J-13
-------
the unit must be designed for viscous flow, and the residence time must
be sufficient to allow sedimentation. They are not considered appropriate
for incinerator sampling.
4. Flow Meter
In particulate stack sampling, two types of sample gas metering are
employed:
fi\ ri /dCvolume)\
(1) gas flow,^ -j£ )
(2) volume, (integrated flow)
A well-designed sampling system will have one of each type. The flow
meter is used in adjusting flow to isokinetic conditions ( a null nozzle
may be considered to satisfy the function of the flow meter), while the
gas volume meter is used for calculating particulate concentrations and
emissions. The accuracy of a concentration (hence, emission) determina-
tion depends as much upon the accuracy of sample gas volume measurements
as upon accuracy of particulate mass. This may be seen in equation (1)
below:
\r
(1)
where: C = particulate mass concentration
M = mass of collected particulates from a sample
V = volume of sample gas measured by the meter
Pm = absolute pressure of the sample gas at the meter
P fcj = standard pressure
Tm = absolute temperature of the sample gas at the meter
Tgt(j = standard temperature
It can also be seen from equation (1) the the accuracy of C also depends
in a first order manner, on the measurement of sample gas temperature
and pressure at the meter. It is easy to measure the gas temperature
within less than one percent. When the gas meter is located upstream of
the pump the meter must be leak free. The meter may be operated at a
few inches of mercury vacuum and the pressure must be measured with an
error of less than 0.3 inches of mercury for a concentration error of less
than one percent. Most reasonably priced mechanical vacuum gauges are
not this good when the vacuum is less than about 5 inches of mercury.
For this reason, any gas meter operated in this vacuum range should be
equipped with a mercury manometer or a high quality mechanical gauge.
Most gas meters are not rated to operate under more than about 1" Hg
J-14
-------
vacuum. When the meter is downstream of the pump (the pump must be leak-
free) , the meter pressure is virtually barometric and hence not likely
to introduce a significant error.
The required accuracy of a gas meter depends on its use. A flow meter
used in adjusting isokinetic flow need not be as accurate as a meter used
in measuring sample volume. Small discrepancies from isokinetic flow
do not affect the accuracy of results as significantly as errors in volume
measurement. Within accuracy constraints, the particular gas meter(s)
must be in harmony with the total sampling system and within the general
engineering requirements of:
(1) Portability
(2) Operational simplicity
(3) Reliability/stability
(4) Ruggedness
Types of gas meters are summarized in Table j-1.
The present state-of-the-art uses mainly the dry gas meter and calibrated
orifices with the calibrated cyclone used to a lesser extent. These are
time-tested techniques which will provide the necessary results,
although it seems that modern technology could be used to improve or
replace some of these techniques. An example would be to use electronic
analog devices with a calibrated cyclone to provide volume as well as
flow measurement. Also, a permanent record (strip chart) could be pro-
duced to reduce the possibility of human error of reading gauges.
5. Pumps
Suction sources may consist of a variety of motor-driven pumps or air
ejectors.
The rotary vacuum pump is a positive displacement unit utilizing vanes of
graphite or fiber providing pulse-free flow at a low initial cost and low
maintenance cost. These units are available in lubricated or oil-less
types with a wide range of volume and pressure characteristics from
several commercial sources.
The cycloid pump is also a positive displacement unit which provides
pulse-free flow. It has the disadvantage of being unable to pull as
great a vacuum as the rotary vane type pump due to slippage between the
cycloids. It is more expensive.
Diaphragm pumps, usually smaller than the previously described suction
sources, are used in sampling operations. These pumps do not operate
under high pressures of vacuum.
J-15
-------
TABLE J-l
Summary of Gas Meters
Meter
Dry gas
Meter
Calibrated
Orifice
Meter
Rotameter
Wet Test
Meter
Rotary Gas
Meter (Root
Type)
Turbine
Meter
Venturi
Meter
Mass Flow
Meter
Calibrated
Cyclone
Type
Volume
Flow
Flow
Volume
Volume
Volume or
Flow
Flow
Mass Flow
or Total
Mass
Flow
Use Typical
In Particulate Common Best Reliability Operational
Stack Sampling Capacity Accuracy Stability Simplicity
Very Common .1 to sev- ^1 Check cal. Good
eral cfm frequently
Very Common Any 1-2% Good Good
Occational *v
-------
Centrifugal pumps (compressors) are often used in sampling systems where
a large volume of flow is required at low head.
Ejectors may be used over a wide range of volumes and are capable of
operating at relatively large pressure drops. Compressed air or steam
requirements and noise are the two greatest problems in the use of
ejectors. Their use is minimal.
6. Supporting Element
Measurements to establish gas velocity and calculate gas volume are
usually made with a Pitot-static tube and an inclined manometer. These
measurements are used to determine isokinetic velocity and are an import-
ant factor in the accuracy of a sampling procedure.
A special "S" type "Staubscheibe" tube is frequently used on dirty or wet
gases. The advantage of this unit is that it does not plug as readily as
the conventional Pitot-static tube which has extremely small openings
(0.040 in. static ports), and thus lessens the chance fo.r large errors
due to plugging.
A correction factor must be applied to all velocity measurements made
with the "S" type tube. The factor can be determined by calibration
with a Pitot-static tube. Gas velocity as calculated from the "S"
type tube reading must be multiplied by the correction factor to obtain
the true gas velocity. Pre-calibrated units are available from manu-
facturers.
Velocity pressures are indicated on an inclined manometer which is
usually constructed to read up to two or more inches of water in incre-
ments of 0.01 inches.
Static pressure readings are also an important part of the supporting
sample data. Static pressure readings are usually obtained by using
the static pressure holes in the Pitot-static tube connected to the
draft gauge with the other tap of the gauge open to the atmosphere.
Alternatively, the static pressure is often measured with a water U-tube
manometer connected to tubing flush with the inside wall of the duct.
Instruments for temperature measurement are largely governed by the
temperature found and the frequency of measurement desired.
Dial thermometers and thermocouples with a potentiometer are frequently
used and provide suitable data.
Determination of the moisture content of the sample requires the conden-
sation of excess moisture at the sampling temperature and the immediate
measurement of the volume of gas at saturated conditions. Condensation
at ambient or a controlled temperature is achieved by condensing coils
or by passing the gas through a temperature controlled chamber where
the condensed water can be subsequently measured.
J-17
-------
The moisture content in the stack gas can be determined from wet and
dry bulb temperature readings by use of the following formula:
p = p1 - 0.01 (td -
where p = vapor pressure of gas in inches of mercury
p' = vapor pressure of saturated gas at tw in inches of mercury
t
PL = barometric pressure in inches of mercury
pg = stack pressure in inches of mercury
In many sampling operations, the density of the gas may be assumed to be
the same as air and no actual evaluation or correction made. Gas den-
sity can be measured by the Shilling or National Bureau of Standards Gas
Density Apparatus. As these instruments are not usually available, the
most commonly used method employs the Orsat gas analysis technique
to measure composition to permit calculation of molecular weight gas
density.
In the Orsat gas analysis, (X>2, CO, and 02 are determined by reaction
with appropriate reagents and the remaining volume of the original
sample is considered to be nitrogen. From the gas composition and known
molcular weights, the density of the sampled gas can be determined.
Data on C02 and 02 content are also required for statement of emissions
on a normalized basis.
The Orsat instrument is basically a laboratory device and does not
function well in the field, nor does it lend itself well to operation
by the inexperienced chemist or technician. Instruments are now on the
market to make volumetric (%) evaluations of C02 and ©2 rapidly and
directly from the stack or sampled gas much more easily and rapidly than
the Orsat and the equipment lends itself well to use under field condi-
tions •
J-18
-------
APPENDIX K
METHOD 5 - DETERMINATION OF PARTICULATE
EMISSIONS FOR STATIONARY SOURCES*
Definition of Particulate
Particulate matter is defined as "any finely divided liquid or solid mater-
ial, other than uncombined water, as measured by Method 5."
1. Principle and Applicability
1.1 Principle. Particulate matter is withdrawn isokinetically from
the source and its weight is determined gravimetrically after removal of
uncombined water.
1.2 Applicability. This method is applicable for the determination of
particulate emissions from stationary sources only when specified by the test
procedures for determining compliance with New Source Performance Standards.
2. Apparatus
2.1 Sampling Train. The design specifications of the particulate sampl-
ing train used by'EPA (Figure K-l) are described in APTD-0581. Commercial
models of this train are available.
2.1.1 Nozzle. Stainless steel (316)with sharp, tapered leading edge.
2.1.2 Probe. Pyrex (trade name) glass with a heating system capable
of maintaining a minimum gas temperature of 250°F at the exit end during
sampling to prevent condensation from occurring. When length limitations
(greater than about 8 ft) are encountered at temperatures less than 600°F
Incoloy 825 (trade name) or equivalent, may be used. Probe for sampling
gas streams at temperatures in excess of 600°F must have been approved by
the Administrator.
2.1.3 Pitot Tube. Type S or equivalent, attached to probe to monitor
stack gas velocity.
2.1.4 Filter Holder. Pyrex (trade name) glass with heating system
capable of maintaining minimum temperature of 225°F.
2.1.5 Impingers/Condenser. Four impingers connected in series with glass
ball joint fittings. The first, third, and fourth impingers are of the
Greenburg-Smith design, modified by replacing the tip with a 1/2-inch ID glass
tube extending to one-half inch from the bottom of the flask. The second
*Taken from "Standards of Performance for New Stationary Sources," Federal
Register, Vol. 36, No. 247, Thursday, December 23, 1971, p. 24888.
K-l
-------
Impinger Train Optional. May be Replaced
By An Equivalent Condenser
Heated Area
Filter Holder / Thermometer
Probe
Averse-Type
Pi tot Tube
Pi tot Manometer lingers Ice Bath
, , , , By-Pass Valve
Urifice ...••>
Thermometers a
Check
Valve
Vacuum
Line
Vacuum Gauge
Main Valve
Dry Test Meter
Air-Tight
Pump
FIGURE K-l
Particulate Sampling Train
K-2
-------
impinger is of the Greenburg-Smith design with the standard tip. A conden-
ser may be used in place of the impingers provided that the moisture content
of the stack gas can still be determined.
2.1.6 Metering System. Vacuum gauge, leak-free pump, thermometers
capable of measuring temperature to within 5°F, dry gas meter with 2%
accuracy, and related equipment, or equivalent, as required to maintain
an isokinetic sampling rate and to determine sample volume.
2.1.7 Barometer. To measure atmospheric pressure to ± 0.1 inches Hg.
2.2 Sample recovery.
2.2.1 Probe Brush. At least as long as probe.
2.2.2 Glass Wash Bottles. Two
2.2.3 Glass Sample Storage Containers.
2.2.4 Graduated Cylinder. 250 ml.
2.3 Analysis.
2.3.1 Glass Weighing Dishes.
2.3.2 Desiccator.
2.3.3 Analytical Balance. To measure to t 0.1 mg.
2.3.4 Trip Balance. 300 g capacity, to measure to - 0.05 g.
3. Reagents
3.1 Sampling.
3.1.1 Filters. Glass fiber, MSA 1106 BH (trade name) or equivalent,
numbered for identification and preweighed.
3.1.2 Silica Gel. Indicating type, 6-16 mesh, dried at 175°C
(350°F) for 2 hours.
3.1.3 Water.
3.1.4 Crushed Ice.
3.2 Sample Recovery.
3.2.1 Acetone. Reagent grade.
3.3 Analysis.
3.3.1 Water.
3.3.2 Dessicant. Drierite (trade name) indicating.
K-3
-------
4. Procedure
4.1 Sampling.
4.1.1 After selecting the sampling site and the minimum number of
sampling points, determine the stack pressure, temperature, moisture,
and range of velocity head.
4.1.2 Preparation of Collection Train. Weigh 'to the nearest gram
approximately 200 g of silica gel. Label a filter of proper diameter,
desiccate (dry using Drierite at 70°F - 10°F) for at least 24 hours and
weigh to the nearest 0.5 mg in a room where the relative humidity is
less than 50%. Place 100 ml of water in each of the first two impin-
gers, leave the third impinger empty, and place approximately 200 g of
preweighed silica gel in the fourth impinger. Set up the train without
the probe as in Figure K-l. Leak check the sampling train at the sampl-
ing site by plugging up the inlet to the filter holder and pulling a
15 in. Hg vacuum. A leakage rate not in excess of 0.02 c.f.m. at a
vacuum of 15 in. Hg is acceptable. Attach the probe and adjust the
heater to provide a gas temperature of about 250°F at the probe outlet.
Turn on the filter heating system. Place crushed ice around the impin-
ger s. Add more ice during the run to keep the temperature of the gases
leaving the last impinger as low as possible and preferably at 70°F
or less. Temperatures about 70°F may result in damage to the dry gas
meter from either moisture condensation or excessive heat.
4.1.3 Particulate Train Operation. For each run, record the data
required on the example sheet shown in Figure K-2. Take readings at
each sampling point, at least every 5 minutes, and when significant
changes in stack conditions necessitate additional adjustments in flow
rate. To begin sampling, position the nozzle at the first traverse
point with the tip pointing directly into the gas stream. Immediately
start the pump and adjust the flow to isokinetic conditions. Sample
for at least 5 minutes at each traverse point; sampling time must be
the same for each point. Maintain isokinetic sampling throughout the
sampling period. Nomographs are available which aid in the rapid adjust-
ment of the sampling rate without other computations. APTD-0576 details
the procedure for using these nomographs. Turn off the pump at the
conclusions of each run and record the final readings. Remove the probe
and nozzle from the stack and handle in accordance with the sample
recovery process described in Section 4.2.
4.2 Sample Recovery. Exercise care in moving the collection train
from the test site to the sample recovery area to minimize the loss of
collected sample or the gain of extraneous particulate matter. Set aside
a portion of the acetone used in the sample recovery as a blank for
analysis. Measure the volume of water from the first three impingers,
then discard. Place the samples in containers as follows:
Container No. 1. Remove the filter from its holder, place
in this container, and seal.
K-4
-------
Plant
Location
Operator
Date
Run No.
Sample Box No.
Meter Box No.
Meter AH
C Factor
SCHEMATIC OF STACK CROSS SECTION
Ambient Temperature
Barometric Pressure
Assumed Moisture,%
Heater Box Setting _
Probe Length,ra
Nozzle Diameter,in
Probe Heater Setting
Traverse Point
; Number
i
i _
t
i
I
Sampling
Time
(e) , min.
Static
Pressure
(Ps) in. Hg
-.
l
i
Stack
Temperature
(Ts) °F
•- - j ! •
i i
....
1- L
[
Velocity
Head
;i!,i
-------
Container No. 2. Place loose particulate matter and acetone
washings from all sample-exposed surfaces prior to the filter in this
container and seal. Use a razor blade, brush, or rubber policeman to
lose adhering particles.
Container No. 3. Transfer the silica gel from the fourth
impinger to the original container and seal. Use a rubber policeman
as an aid in removing silica gel from the impinger.
4.3 Analysis. Record the data required on the example sheet shown
in Figure K-3. Handle each sample container as follows:
Container No. 1. Transfer the filter and any loose parti-
culate matter from the sample container to a tared glass weighing dish,
desiccate, and dry to a constant weight. Report results to nearest
0.5 mg.
Container Mo. 2. Transfer the acetone washings to a tared
beaker and evaporate to dryness at ambient temperature and pressure.
Desiccate and dry to a constant weight. Report results to the nearest
0.5 mg.
Container No. 3. Weigh the spent silica gel and report to
the nearest gram.
5. Calibration
Use methods and equipment which have been approved by the Adminis-
trator to calibrate the orifice meter, pitot tube, dry gas meter, and
probe heater. Recalibrate after each test series.
6. Calculations
6.1 Average dry gas meter temperature and average orifice pressure
drop. See data sheet (Figure K-2).
6.2 Dry Gas Volume. Correct the sample volume measured by the dry
gas meter to standard conditions (70°F, 29.92 inches Hg) by using
Equation K-l.
std
equation K-l
K-6
-------
Plant
Date
Run No.
Container
Number
Weight of Particulate Collected, mg
Final Weight
Tare Weight
Weight Gain
Total
Final
Initial
Liquid Collected
Total Volume Collected
Volume of Liquid
Water Collected
Impinger
Volume ,
ml
Silica Gel
Weight,
g
g* ml
Convert Weight of Water to Volume by Dividing Total Weight Increase by
Density of Water (1 g ml) :
Increase, g ,r , „ ^ .
—T~, , ;XD = Volume Water ml
(1 g/ml)
FIGURE K-3. Analytical Data
K-7
-------
where:
Vm , . = Volume of gas sample through the dry gas meter (standard condi
tions), cu. ft.
Vm = Volume of gas sample through the dry gas meter (meter condi-
tions), cu. ft.
T .. = Absolute temperature at standard conditions, 530° R.
std
Tm = Average dry gas meter temperature, °R.
Pbar = Barometric pressure at the orifice meter, inches Hg.
AH = Average pressure drop across the orifice meter, inches H20.
13.6 = Specific gravity of mercury.
P .. = Absolute pressure at standard conditions, 29.92 inches Hg.
6.3 Volume of Water Vapor.
MH2o
equation K-2
where:
Vw , , = Volume of water vapor in the gas sample (standard conditions)
cu. ft.
Vlo = Total volume of liquid collected in impingers and silica gel
(see Figure K-3) ml.
pH~ = Density of water, 1 g/ml.
MhL = Molecular weight of water, 18 Ib/lb-mole.
R = Ideal gas constant, 21.83 inches Hg - cu. f t. /lb-mole-°R.
T . . = Absolute temperature at standard conditions, 530°R.
P . d = Absolute pressure at standard conditions, 29.92 inches Hg .
K-8
-------
6.4 Moisture Content.
wo " Vm . . + Vw . .
std std
equation 1C-3
where:
BWO = Proportion by volume of water vapor in the gas stream, dimen-
sionless.
Vw , , = Volume of water in the gas sample (standard conditions), cu. ft.
Vm , . = Volume of gas sample through the dry gas meter (standard
conditions), cu. ft.
6.5 Total Particulate Weight. Determine the total particulate
catch from the sum of the weights on the analysis data sheet (Figure K-3)
6.6 Concentration.
6.6.1 Concentration in gr./s.c.f.
equation K~4
wher e:
c = Concentration of particulate matter in stack gas, gr./s.c.f.,
dry basis.
Mn = Total amount of particulate mattar collected, mg.
Vm +ri = Volume of gas sample through dry gas meter (standard conditions) ,
cu. ft.
K-9
-------
6.6.2 Concentration in Ib./cu. ft.
I lb „ \
.600 mg. V _
453,600 mg. "n/ -6 Mn
c = ' ••""' =^-I 1^ = O Ofjc: y -in
5 Vm . . 2'205 X 10 Vm , ,
std std
equation K-5
where:
c = Concentration of particulate matter in stack gas, Ib./s.c.f.,
dry basis.
453,600 = Mg/lb.
M = Total amount of particulate matter collected, mg.
Vm ,, = Volume of gas sample through dry gas meter (standard conditions)
o I U f
cu. ft.
6.7 = isokinetic variation.
I =
eV P A
s s n
—/JX 100
(1.667=^) l"(o.00267in- »^'ft')Vlc + ^ (Pbar
V _ sec./ [A ml.- R / _ c Tm \
13. 6
6V P A
s s n
equation K-6
wher e:
I = Percent of isokinetic sampling.
VI = Total volume of liquid collected in impingers and silica gel
(see Fig. K-3) ml.
pH20 = Density of water, I g./ml.
R = Ideal gas constant, 21.83 inches Hg-cu. ft./lb. mole-°R.
K-10
-------
MH20 = Molecular weight of water, 18' Ib./lb.-mole.
Vm = Volume of gas sample through the dry gas meter (meter conditions),
cu. ft.
Tm = Absolute average dry gas meter temperature (see Figure K-2) °R.
Pbar = Barometric pressure at sampling site, inches Hg.
AH = Average pressure drop across the orifice (see Fig. K-2),
inches Hg.
TS = Absolute average stack gas temperature (see Fig. K-2), °R.
0 = Total sampling time, rain.
Vs = Stack gas velocity calculated by Method 2, Equation 2-2, ft/sec.
PS = Absolute stack gas pressure, inches Hg.
Ap = Cross-sectional area of nozzle, eq. ft.
6.8 Acceptable Results. The following range sets the limit on
acceptable isokinetic sampling results:
If 90% - I - 110%, the results are acceptable, otherwise, reject the
results and repeat the test.
K-ll
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-650/2-73-023
3. RECIPIENT'S ACCESSIOt*NO.
4. TITLE AND SUBTITLE
Manual Methods for Sampling and
Analysis of Particulate Emissions From Municipal
Incinerators
5. REPORT DATE
September, 1973
6. PERFORMING ORGANIZATION CODE
?.AUTHOR(S) John T_ Fun)chouser; E.T. Peters, P.L. Levins,
Arnold Doyle, Paul Giever & John McCoy
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Arthur D. Little, Inc
Cambridge, Massachusetts
10. PROGRAM ELEMENT NO.
1AA010
11. CONTRACT'GRANT NO.
EHSD 71-27
12. SPONSORING AGENCY NAME AND ADDRESS
Chemistry & Physics Laboratory
NERC
Research Triangle Pk., North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Municipal incinerators, and similar stationary sources, though contributing only
a few percent of the total national air pollution load, are important sources of pollu
near population centers. The particulate matter they emit has a significant adverse
effect on health, on materials of construction and on visibility; they are responsible
for many complaints. Therefore, the Federal Government, through the Environmental
Protection Agency, has promulgated standards that specify the permissible levels
of particulate matter emitted from newly constructed incinerators operating at or
above a charging rate of 50 tons per day. In many cases, however, the chemistry
of the particulate species present in these emissions is not well know. Consequently,
there is a need to define more thoroughly the chemical nature of particulate emissions
from incinerators and to gain a better understanding of how the sample collection
equipment used by the EPA influences the physical and chemical properties of the
particulate. The primary goal of the program has been to help develop the data base
and the technology which will permit representative measurements of source
particulate emissions to be obtained from waste incineration sources, and from the
particulate pollution control devices associated with such sources.
ion
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI I-'icld/Group
Incinerator Emissions
Particulate Emission Methods
Particulate Emission Sampling
IS. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
287
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
K-12
-------
ENVIRONMENTAL PROTECTION AGENCY
Technical Publications Branch
Office of Administration
ft*M«rch Triangle Park. N. C. 27711
OFFICIAL BUSINESS
AN EQUAL OPPORTUNITY EMPLOYER
POSTAGE AND FEES PAID
ENVIRONMENTAL PROTECTION AGENCY
EPA - 335
SPECIAL FOURTH-CLASS RATE
BOOK
Return this sheet if you do NOT wish to receive this material I 1,
or if change of address is needed f I. (Indicate change, including
ZIP code.)
------- |