United States
Environmental Protection
Agency
AIR
Office of Air Quality
Planning And Standards
Research Triangle Park, NC 27711
EPA-454/R-97-012
December 1997
LOCATING AND ESTIMATING AIR
EMISSIONS FROM SOURCES OF
MERCURY AND MERCURY COMPOUNDS
-------
EPA-454/R-97-012
Locating And Estimating Air Emissions
From Sources of Mercury and Mercury Compounds
Office of Air Quality Planning and Standards
Office of Air and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
December 1997
-------
This report has been reviewed by the Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, and has been approved for publication. Mention of trade
names and commercial products does not constitute endorsement or recommendation for use.
EPA-454/R-97-012
-------
TABLE OF CONTENTS
Section Page
EXECUTIVE SUMMARY xi
1.0 PURPOSE OF DOCUMENT 1-1
2.0 OVERVIEW OF DOCUMENT CONTENTS 2-1
3.0 BACKGROUND 3-1
3.1 NATURE OF THE POLLUTANT 3-1
3.2 OVERVIEW OF PRODUCTION, USE, AND EMISSIONS 3-1
3.2.1 Production 3-1
3.2.2 End-Use 3-3
3.2.3 Emissions 3-6
4.0 EMISSIONS FROM MERCURY PRODUCTION 4-1
4.1 PRIMARY MERCURY PRODUCTION 4-1
4.1.1 Process Description 4-2
4.1.2 Emission Control Measures 4-2
4.1.3 Emissions 4-4
4.2 SECONDARY MERCURY PRODUCTION 4-4
4.2.1 Process Description 4-4
4.2.2 Emission Control Measures 4-6
4.2.3 Emissions 4-6
4.3 MERCURY COMPOUNDS PRODUCTION 4-6
4.3.1 Process Description 4-6
4.3.2 Emission Control Measures 4-7
4.3.3 Emissions 4-7
5.0 EMISSIONS FROM MAJOR USES OF MERCURY 5-1
5.1 CHLORINE PRODUCTION USING THE MERCURY CELL PROCES S 5-1
5.1.1 Process Description 5-1
5.1.2 Emission Control Measures 5-4
5.1.3 Emissions 5-5
5.2 BATTERY MANUFACTURING 5-5
5.2.1 Process Description 5-6
5.2.2 Emission Control Measures 5-8
5.2.3 Emissions 5-9
5.3 ELECTRICAL USES 5-9
5.3.1 Electric Switches 5-9
5.3.2 Thermal Sensing Elements 5-15
5.3.3 Tungsten Bar Sintering 5-15
5.3.4 Copper Foil Production 5-16
5.3.5 Fluorescent Lamp Manufacture and Recycling 5-16
5.4 INSTRUMENT MANUFACTURING AND USE (THERMOMETERS) 5-19
5.4.1 Process Description 5-19
5.4.2 Emission Control Measures 5-19
5.4.3 Emissions 5-20
in
-------
TABLE OF CONTENTS (continued)
Section Page
6.0 EMISSIONS FROM COMBUSTION SOURCES 6-1
6.1 COAL COMBUSTION 6-3
6.1.1 Coal Characteristics 6-3
6.1.2 Process Description 6-6
6.1.3 Emission Control Measures 6-8
6.1.4 Emissions 6-9
6.2 FUEL OIL COMBUSTION 6-13
6.2.1 Fuel Oil Characteristics 6-13
6.2.2 Process Description 6-17
6.2.3 Emission Control Measures 6-18
6.2.4 Emissions 6-18
6.3 WOOD COMBUSTION 6-22
6.3.1 Process Description 6-22
6.3.2 Emission Control Measures 6-24
6.3.3 Emissions 6-25
6.4 MUNICIPAL WASTE COMBUSTION 6-25
6.4.1 Municipal Solid Waste Characteristics 6-26
6.4.2 Process Description 6-26
6.4.3 Emission Control Measures 6-29
6.4.4 Emissions 6-30
6.5 SEWAGE SLUDGE INCINERATORS 6-31
6.5.1 Process Description 6-32
6.5.2 Emission Control Measures 6-33
6.5.3 Emissions 6-34
6.6 HAZARDOUS WASTE COMBUSTION 6-35
6.6.1 Process Description 6-35
6.6.2 Emission Control Measures 6-36
6.6.3 Emissions 6-37
6.7 MEDICAL WASTE INCINERATION 6-37
6.7.1 Process Description 6-38
6.7.2 Emission Control Measures 6-41
6.7.3 Emissions 6-41
7.0 EMISSIONS FROM MISCELLANEOUS SOURCES 7-1
7.1 PORTLAND CEMENT MANUFACTURING 7-1
7.1.1 Process Description 7-2
7.1.2 Emission Control Measures 7-4
7.1.3 Emissions 7-4
7.2 LIME MANUFACTURING 7-5
7.2.1 Process Description 7-5
7.2.2 Emission Control Measures 7-8
7.2.3 Emissions 7-8
7.3 CARBON BLACK PRODUCTION 7-9
7.3.1 Process Description 7-9
7.3.2 Emission Control Measures 7-9
7.3.3 Emissions 7-9
7.4 BYPRODUCT COKE PRODUCTION 7-9
7.4.1 Process Description 7-12
iv
-------
TABLE OF CONTENTS (continued)
Section Page
7.4.2 Emission Control Measures 7-12
7.4.3 Emissions 7-16
7.5 PRIMARY LEAD SMELTING 7-16
7.5.1 Process Description 7-17
7.5.2 Emission Control Measures 7-17
7.5.3 Emissions 7-20
7.6 PRIMARY COPPER SMELTING 7-20
7.6.1 Process Description 7-20
7.6.2 Emission Control Measures 7-24
7.6.3 Emissions 7-24
7.7 PETROLEUM REFINING 7-24
7.7.1 Process Description 7-25
7.7.2 Emission Control Measures 7-26
7.7.3 Emissions 7-26
7.8 MUNICIPAL SOLID WASTE LANDFILLS 7-26
7.8.1 Process Description 7-27
7.8.2 Emission Control Measures 7-27
7.8.3 Emissions 7-27
7.9 GEOTHERMAL POWER PLANTS 7-28
7.9.1 Emission Control Measures 7-28
7.9.2 Emissions 7-28
7.10 PULP AND PAPER PRODUCTION 7-29
7.10.1 Process Description 7-29
7.10.2 Emission Control Measures 7-33
7.10.3 Emissions 7-35
8.0 EMISSIONS FROM MISCELLANEOUS FUGITIVE AND AREA SOURCES 8-1
8.1 MERCURY CATALYSTS 8-1
8.1.1 Process Description 8-1
8.1.2 Emission Control Measures 8-1
8.1.3 Emissions 8-1
8.2 DENTAL ALLOYS 8-2
8.2.1 Process Description 8-2
8.2.2 Emission Control Measures 8-2
8.2.3 Emissions 8-2
8.3 MOBILE SOURCES 8-2
8.4 CREMATORIES 8-3
8.5 PAINT USE 8-3
8.6 SOIL DUST 8-3
8.7 NATURAL SOURCES OF MERCURY EMISSIONS 8-7
9.0 SOURCE TEST PROCEDURES 9-1
9.1 INTRODUCTION 9-1
9.2 DEDICATED MERCURY SAMPLING METHODS 9-1
9.2.1 EPA Method 101-Determination of Particulate and Gaseous 9-1
9.2.2 EPA Method lOlA-Determination of Particulate and Gaseous
Mercury Emissions from Stationary Sources 9-4
-------
TABLE OF CONTENTS (continued)
Section Page
9.2.3 EPA Method 102-Determination of Participate and Gaseous Mercury
Emissions from Chlor-Alkali Plants-Hydrogen Streams
(40 CFR, Part 61, 1992) 9-4
9.3 MULTIPLE METALS SAMPLING TRAINS 9-4
9.3.1 Method 0012-Methodology for the Determination of Metals Emissions
in Exhaust Gases from Hazardous Waste Incineration and Similar
Combustion Sources 9-4
9.3.2 CARB Method 436-Determination of Multiple Metals Emissions
from Stationary Sources 9-6
9.4 ANALYTICAL METHODS FOR DETERMINATION OF MERCURY 9-6
9.5 SUMMARY 9-7
10.0 REFERENCES 10-1
APPENDIX A. NATIONWIDE EMISSION ESTIMATES A-l
APPENDIX B. SUMMARY OF COMBUSTION SOURCE MERCURY
EMISSION DATA B-l
APPENDIX C. SELECTED INFORMATION FOR CEMENT KILNS
AND LIME PLANTS C-l
APPENDIX D. CRUDE OIL DISTILLATION CAPACITY D-l
APPENDIX E. PULP AND PAPER MILLS IN THE UNITED STATES
IN 1994 E-l
APPENDIX F. EMISSION FACTORS BY SOURCE CLASSIFICATION CODE .... F-l
VI
-------
LIST OF FIGURES
Figure Page
3-1. End-use pattern of mercury 3-5
4-1. Major components of mercury recovery from gold ores 4-3
4-2. Process flow diagram for secondary recovery at a battery plant 4-5
4-3. Mercuric/mercurous chloride production 4-8
4-4. Mercuric oxide production via mercuric chloride and mercuric nitrate intermediates .... 4-9
5-1. Basic flow diagram for a mercury-cell chlor-alkali operation 5-3
5-2. General flow diagram for mercuric oxide battery (button cell) manufacture 5-7
5-3. Manufacture of mercury buttons for wall switches 5-12
5-4. Thermostat switch manufacture 5-13
6-1. Process flow diagram for sludge incineration 6-33
6-2. Major components of an incineration system 6-39
7-1. Process flow diagram of portland cement manufacturing process 7-3
7-2. Process flow diagram for lime manufacturing process 7-7
7-3. Process flow diagram for carbon black manufacturing process 7-11
7-4. Schematic of byproduct coke oven battery 7-14
7-5. Types of air pollution emissions from coke oven batteries 7-15
7-6. Typical primary lead processing scheme 7-19
7-7. Typical primary copper smelter process 7-22
7-8. Relationship of the chemical recovery cycle to the pulping and product forming
processes 7-30
7-9. Kraft process-chemical recovery area (including direct contact evaporator recovery
furnace) 7-31
9-1. Typical dedicated mercury sampling train 9-3
9-2. Typical multiple metals sampling train 9-5
vn
-------
LIST OF TABLES
Table Page
ES-1 ESTIMATED NATIONWIDE EMISSIONS xi
3-1. PHYSICAL AND CHEMICAL PROPERTIES OF MERCURY 3-2
3-2. U.S. SUPPLY AND DEMAND FOR MERCURY, 1991 TO 1995 3-4
3-3. END-USE PATTERN OF MERCURY FOR INDUSTRIAL CONSUMPTION 3-4
3-4. ESTIMATED 1994-1995 NATIONWIDE MERCURY EMISSIONS FOR
SELECTED SOURCE CATEGORIES 3-7
4-1. BYPRODUCT MERCURY-PRODUCING GOLD MINES IN THE UNITED
STATES IN 1995 4-1
4-2. MERCURY COMPOUND PRODUCERS 4-6
5-1. 1996 MERCURY CELL CHLOR-ALKALI PRODUCTION FACILITIES 5-2
5-2. MERCURIC OXIDE, ALKALINE MANGANESE, OR ZINC-CARBON
BATTERY MANUFACTURERS IN 1996 5-6
5-3. METHODS FOR REDUCING WORKER EXPOSURE TO MERCURY
EMISSIONS IN BATTERY MANUFACTURING 5-8
5-4. EMISSION SOURCE PARAMETERS FOR AN INTEGRATED MERCURY
BUTTON CELL MANUFACTURING FACILITY 5-10
5-5. MEASURES TO REDUCE WORKPLACE EXPOSURE TO MERCURY
VAPOR EMISSIONS IN THE ELECTRIC SWITCH INDUSTRY 5-11
5-6. MANUFACTURERS OF ELECTRIC SWITCHES AND ELECTRONIC
COMPONENTS REPORTING IN THE 1994 TOXIC RELEASE INVENTORY 5-14
5-7. U.S. FLUORESCENT LAMP MANUFACTURERS' HEADQUARTERS 5-16
6-1. 1994 DISTRIBUTION OF FOSSIL FUEL CONSUMPTION IN THE UNITED
STATES 6-2
6-2. COAL HEATING VALUES 6-4
6-3. EXAMPLES OF COAL HEAT CONTENT VARIABILITY 6-5
6-4. MERCURY CONCENTRATION IN COAL BY COAL TYPE 6-6
6-5. MERCURY CONCENTRATION IN COAL BY REGION 6-7
6-6. CALCULATED UNCONTROLLED MERCURY EMISSION FACTORS FOR
COAL COMBUSTION 6-11
6-7. MEASURED MERCURY EMISSION FACTORS FOR COAL COMBUSTION 6-11
6-8. BEST TYPICAL MERCURY EMISSION FACTORS FOR COMMERCIAL/
INDUSTRIAL/ RESIDENTIAL COAL-FIRED BOILERS 6-14
6-9. TYPICAL HEATING VALUES OF FUEL OILS 6-15
6-10. TYPICAL FUEL OIL HEATING VALUES FOR SPECIFIC REGIONS 6-16
6-11. MERCURY CONCENTRATION IN OIL BY OIL TYPE 6-17
6-12. CALCULATED UNCONTROLLED MERCURY EMISSION FACTORS FOR
FUEL OIL COMBUSTION 6-19
6-13. MERCURY CONCENTRATIONS IN RESIDUAL OIL AND MERCURY
EMISSION FACTORS FROM RESIDUAL COMBUSTION 6-20
6-14. MERCURY EMISSION FACTORS FOR FUEL OIL COMBUSTION GENERATED
FROM CALIFORNIA "HOT SPOTS" TESTS 6-21
6-15. BEST TYPICAL MERCURY EMISSION FACTORS FOR FUEL OIL
COMBUSTION 6-22
6-16. SUMMARY OF GEOGRAPHICAL DISTRIBUTION OF MWC FACILITIES
LARGER THAN 35 Mg/d 6-27
6-17. COMPOSITION OF DISPOSED RESIDENTIAL AND COMMERCIAL WASTE
(WEIGHT PERCENT) 6-28
6-18. AVERAGE EMISSION FACTORS FOR MUNICIPAL WASTE COMBUSTORS ... 6-31
viii
-------
LIST OF TABLES (continued)
Table Page
6-19. SUMMARY OF MERCURY EMISSION FACTORS FOR SEWAGE SLUDGE
INCINERATORS 6-34
6-20. MERCURY EMISSION FACTORS FOR MWI'S 6-42
7-1. LIME PRODUCERS IN THE UNITED STATES IN 1994 7-6
7-2. CARBON BLACK PRODUCTION FACILITIES 7-10
7-3. BYPRODUCT COKE PRODUCERS IN THE UNITED STATES IN 1991 7-13
7-4. DOMESTIC PRIMARY LEAD SMELTERS AND REFINERIES 7-16
7-5. U.S. PRIMARY COPPER SMELTERS 7-17
7-6. GEOTHERMAL POWER PLANTS OPERATING IN THE UNITED STATES
IN 1992 7-21
7-7. MERCURY EMISSION FACTORS FOR GEOTHERMAL POWER PLANTS 7-28
7-8. MERCURY EMISSION FACTORS FOR COMBUSTION SOURCES AT PULP
AND PAPER MILLS 7-29
8-1 NUMBER OF CREMATORIES AND CREMATIONS BY STATE 8-4
9-1 MERCURY SAMPLING METHODS 9-2
IX
-------
EXECUTIVE SUMMARY
The emissions of mercury and mercury compounds into the atmosphere are of special significance
because of the Clean Air Act Amendments of 1990 (CAAA). Sections of the CAAA that may require
information on mercury emissions include 112(n)(l)(A, B, C), 112(c)(6), 112(m), 112(o)(l), 112(k), and
129. This document is designed to assist groups interested in inventorying air emissions of mercury by
providing a compilation of available information on sources and emissions of these substances.
In the U.S., mercury is produced primarily as a byproduct of gold mining and as a result of
secondary production (i.e., recycling or mercury recovery from products or by-products); the last mercury
mine was closed in 1990. In 1995, the total U.S. supply of mercury was 911 Mg (1,002 tons), of which
approximately 41 percent resulted from imports. The demand for mercury in the U.S. has decreased sharply
(64 percent) since 1989. In 1995, the U.S. demand was 436 Mg (480 tons) or 48 percent of the supply.
In 1995, seven source categories accounted for the U.S. demand for mercury; the chlor-alkali
industry was the major user. Other major users of mercury were for wiring devices and switches and
production of measurement and control instruments. These three source categories accounted for about
65 percent of the total U.S. demand for mercury; the other four source categories accounted for the remaining
35 percent.
Nationwide mercury emissions were estimated for several source types for the years 1994/1995.
These were the latest years for which adequate information was available for almost all source types. The
total nationwide mercury emissions estimate was 140 Mg (154 tons) from five major source types.
Table ES-1 shows the estimated nationwide emissions by major source types and the percent contribution of
each type to the total emissions. The three specific sources emitting the largest quantities of mercury were
coal combustion, municipal waste combustion, and medical waste combustion.
TABLE ES-1. ESTIMATED NATIONWIDE EMISSIONS
Major source type
Mercury and mercury compound
production
Major uses of mercury
Combustion sources
Miscellaneous manufacturing
processes
Other miscellaneous sources
TOTAL
Estimated nationwide emissions,
Mg (tons)
0.13(0.14)
7.3 (8.0)
123.0(135.6)
8.1(8.9)
1.3(1.5)
140(154)
Percent of total emissions
<0.1
5.2
88.0
5.8
0.9
100
XI
-------
1.0 PURPOSE OF DOCUMENT
The U. S. Environmental Protection Agency (EPA), State, and local air pollution control agencies are
becoming increasingly aware of the presence of substances in the environment that may be toxic at certain
concentrations. This awareness, in turn, has led to attempts to identify source/receptor relationships for these
substances and to develop control programs to regulate emissions. Typically, however, little information
exists on the magnitude of the emissions of these substances or about the sources that may be emitting them
to the atmosphere.
To assist groups interested in inventorying air emissions of various hazardous chemicals and metals,
EPA is preparing a series of documents such as this that compiles available information on sources and
emissions of these substances. Prior documents in the series are listed below:
Substance
Acrylonitrile
Carbon Tetrachloride
Chloroform
Ethylene Bichloride
Formaldehyde
Nickel
Chromium
Manganese
Phosgene
Epichlorohydrin
Vinylidene Chloride
Ethylene Oxide
Chlorobenzene
Polychlorinated Biphenyls (PCB's)
Polycyclic Organic Matter (POM)/
Polycyclic Aromatic Hydrocarbons (PAH)
Benzene
Perchloroethylene and Trichloroethylene
Municipal Waste Combustion
Coal and Oil Combustion Sources
1,3-Butadiene
Chromium (Supplement)
Sewage Sludge
Styrene
Cadmium and Cadmium Compounds
Cyanide Compounds
Methylene Chloride
Medical Waste Incinerators
TCDD/TCDF
Toluene
Xylenes
EPA Publication No.
EPA-450/4-84-007a
EPA-450/4-84-007b
EPA-450/4-84-007c
EPA-450/4-84-007d
EPA-45 0/4-91-012
EPA-450/4-84-007f
EPA-450/4-84-007g
EPA-450/4-84-007h
EPA-450/4-84-007i
EPA-450/4-84-007J
EPA-450/4-84-007k
EPA-450/4-84-0071
EPA-450/4-84-007m
EPA-450/4-84-007n
EPA-450/4-84-007p
EPA-45 0/4-
EPA-450/2-
EPA-450/2-
EPA-450/2-
EPA-450/2-
EPA-450/2-
EPA-450/2-
EPA-454/R.
EPA-454/R.
EPA-454/R.
EPA-454/R.
EPA-454/R.
Draft
EPA-454/R.
EPA-454/R.
84-007q
89-013
89-006
89-001
89-021
89-002
90-009
93-011
93-040
93-041
93/006
93-053
93-047
93-048
1-1
-------
Methyl Ethyl Ketone EPA-454/R-93-046
Methyl Chloroform EPA-454/R-93-045
Chlorobenzene (Update) EPA-454/R-93-044
Benzene Update Draft
Polycyclic Organic matter (POM) Update Draft
1,3-Butadiene Update EPA-454/R-96-008
Lead Draft
Arsenic Draft
This document deals specifically with an update of the previous document on emissions of mercury
and mercury compounds (EPA-454/R-93-023); however, the majority of the information contained in this
document concerns elemental mercury emissions.
In addition to the information presented in this document, another potential source of emissions data
for mercury and mercury compounds is the Toxic Chemical Release Inventory (TRI) form required by Section
313 of Title III of the 1986 Superfund Amendments and Reauthorization Act (SARA 313). SARA 313
requires owners and operators of facilities in certain Standard Industrial Classification Codes that
manufacture, import, process or otherwise use toxic chemicals (as listed in Section 313) to report annually
their releases of these chemicals to all environmental media. As part of SARA 313, EPA provides public
access to the annual emissions data. The TRI data include general facility information, chemical information,
and emissions data. Air emissions data are reported as total facility release estimates for fugitive emissions
and point source emissions. No individual process or stack data are provided to EPA under the program.
The TRI requires sources to use stack monitoring data for reporting, if available, but the rule does not require
stack monitoring or other measurement of emissions if data from these activities are unavailable. If
monitoring data are unavailable, emissions are to be quantified based on best estimates of releases to the
environment.
The reader is cautioned that the TRI will not likely provide facility, emissions, and chemical release
data sufficient for conducting detailed exposure modeling and risk assessment studies. In many cases, the
TRI data are based on annual estimates of emissions (i.e., on emission factors, material balance calculations,
and engineering judgment). We recommend the use of TRI data in conjunction with the information provided
in this document to locate potential emitters of mercury and to make preliminary estimates of air emissions
from these facilities.
Mercury is of particular importance as a result of the Clean Air Act Amendments of 1990 (CAAA).
Mercury and mercury compounds are included in the Title III list of hazardous air pollutants (HAPs) and will
be subject to standards established under Section 112, including maximum achievable control technology
(MACT). Also, Section 112(c)(6) of the 1990 CAAA mandate that mercury (among others) be subject to
standards that allow for the maximum degree of reduction of emissions. These standards are to be
promulgated no later than 10 years following the date of enactment. In addition to Section 112(c)(b), other
sections of the CAAA that may require data on mercury emissions include the electric utility steam-
generating units, Section 112(n)(l)(A); the National Institute of Environmental Health Sciences (NIEHS)
health effects study, Section 112(n)(l)(B); the mercury report to Congress, Section 112(n)(l)(C); the Great
Waters Program, Section 112(m); the National Academy of Sciences (NAS) risk assessment methodology
study, Section 112(o)(l); the area source program, Section 112(k); and the solid waste combustion program,
Section 129.
The data on mercury emissions are based, whenever possible, on the results of actual test procedures.
Data presented in this document are total mercury emissions and do not differentiate the chemical forms of
the mercury. The sampling and analysis procedures employed for the determination of the mercury
concentrations from various sources are presented in Section 9, Source Test Procedures. These procedures do
not provide data on the speciation of the mercury in the emissions.
1-2
-------
2.0 OVERVIEW OF DOCUMENT CONTENTS
As noted in Section 1, the purpose of this document is to assist Federal, State, and local air pollution
agencies and others who are interested in locating potential air emitters of mercury and mercury compounds
and estimating air emissions from these sources. The information summarized in this document should not
be assumed to represent the source configuration or emissions of any particular facility.
This section provides an overview of the contents of this document. It briefly outlines the nature,
extent, and format of the material presented in the remaining sections of this document. As stated in Section
1, this document represents a revision and update of the locating and estimating document on mercury and
mercury compounds published in 1993. In addition to an update of the emission estimates, some sources
were deleted and new sources were added. Previous sections on natural gas combustion and oil shale
retorting were deleted from this document. Mercury emissions estimates from natural gas combustion were
based on a single test report and the accuracy of the data in that report have been questioned. Oil shale
retorting was deleted because it is not conducted in the United States. New sections have been added for
hazardous waste incineration, pulp and paper production, and municipal waste landfills.
Section 3 of this document provides a brief summary of the physical and chemical characteristics of
mercury and mercury compounds and an overview of their production and uses. A chemical use tree
summarizes the quantities of mercury produced by various techniques as well as the relative amounts
consumed by various end uses. To the extent possible, the emissions data are presented for the 1994/1995
time period. This background section may be useful to someone who wants to develop a general perspective
on the nature of the substance and where it is manufactured and used.
Sections 4 to 7 of this document focus on the major industrial source types that emit mercury.
Section 4 discusses the production of mercury and mercury compounds. Section 5 discusses the different
uses of mercury as an industrial feedstock. Section 6 discusses emissions from combustion sources. Section
7 discusses emissions from miscellaneous manufacturing processes, and Section 8 discusses emissions from
miscellaneous fugitive and area sources. For each major industrial source category described, process
descriptions and flow diagrams are given wherever possible, potential emission points are identified, and
available emission factor estimates are presented that show the potential for mercury emissions before and
after controls are employed by industry. Individual companies are named that are reported to be involved
with the production and/or use of mercury based on industry contacts, reference materials, the Toxic Release
Inventory (TRI), and available trade publications.
Section 9 of this document summarizes available procedures for source sampling and analysis of
mercury. Details are not provided nor is any EPA endorsement given or implied for any of these sampling
and analysis procedures. Section 10 provides references. Appendix A presents calculations used to derive
the estimated 1994/1995 nationwide mercury emissions. Appendix B presents a summary of the combustion
source test data. Appendix C lists U.S. Portland cement manufacturers. Appendix D presents U.S. crude oil
distillation capacity. Appendix E presents 1994 U.S. pulp and paper mills.
This document does not contain any discussion of human health or environmental impacts of
mercury, nor does it include any discussion of ambient air levels or ambient air monitoring techniques.
2-1
-------
Comments on the content or usefulness of this document are welcome, as is any information on
process descriptions, operating practices, control measures, and emissions that would enable EPA to improve
the document. All comments should be sent to:
Leader, Emission Factor and Inventory Group (MD-14)
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
2-2
-------
3.0 BACKGROUND
3.1 NATURE OF THE POLLUTANT
Mercury, also called quicksilver, is a heavy, silver-white metal that exists as a liquid at ambient
temperatures. Its chemical symbol, Hg, comes from the Latin word, hydrargyrum, meaning liquid silver.
Mercury and its major ore, cinnabar (HgS), have been known and used for thousands of years. Table 3-1
summarizes the major chemical and physical properties of mercury.1
Mercury metal is widely distributed in nature at very low concentrations. In uncontaminated soil,
mercury concentrations range from 30 to 500 parts per billion (ppb) with an average of about 100 ppb. For
most rocks, the mercury content ranges from 10 to 20,000 ppb. Except where special geologic conditions
prevail or where anthropogenic sources lead to increases, surface fresh waters generally contain less than
0.1 ppb total mercury, and seawater averages 0.1 to 1.2 ppb of mercury.
Metallic mercury can be found in small quantities in some ore deposits; however, it usually occurs as
a sulfide. It occurs sometimes as the chloride or the oxide, typically in conjunction with base and precious
metals. Although HgS is by far the predominant mercury mineral in ore deposits, other common
mercury-containing minerals include corderoite (HggS^CL,), livingstonite (HgSb4S7), montroydite (HgO),
terlinguaite (Hg2OCl), calomel (HgCl), and metacinnabar, a black form of cinnabar.
Because metallic mercury has a uniform volume expansion over its entire liquid range and a high
surface tension, it is used in barometers, manometers, thermometers, and other measuring devices. It also is
used extensively in electrical applications, including batteries, electrical lamps, and wiring and switching
devices. Its low electrical resistivity makes it one of the best electrical conductors among the metals.
In the ionic form, mercury exists in one of two oxidation states (or valences): Hg(I), or the
mercurous ion, and Hg(II), or the mercuric ion. Of the two states, the higher oxidation state, Hg(II), is the
more stable.
Mercury has a tendency to form alloys or amalgams with almost all metals except iron, although at
higher temperatures it will even form alloys with iron. Mercury forms amalgams with vanadium, iron,
niobium, molybdenum, cesium, tantalum, or tungsten to produce metals with good to excellent corrosion
resistance. A mercury-silver amalgam traditionally has been used for teeth fillings.
Mercury is stable at ambient temperatures. It does not react with air, ammonia, carbon dioxide,
nitrous oxide, or oxygen but readily combines with the halogens and sulfur. Mercury will react with any
hydrogen sulfide present in the air and should be kept in covered containers. It is not affected by hydrochloric
acid but is attacked by concentrated sulfuric acid. Mercury can be dissolved in either dilute or concentrated
nitric acid, resulting in the formation of either mercurous [Hg(I)] salts (if the mercury is in excess or no heat
is applied) or mercuric [Hg(II)] salts (if excess acid or heat is used).
3.2 OVERVIEW OF PRODUCTION, USE, AND EMISSIONS
3.2.1 Production
Primary production of mercury occurs principally as a byproduct of gold mining. Mercury was
previously mined from mercury ores in Nevada, but that mine closed in 1990. It is still produced in relatively
small quantities as a byproduct from gold ores in Nevada, California, and Utah.2
Secondary production (recycling) of mercury includes the processing of scrapped mercury-containing
products, and industrial waste and scrap. Sales of scrap mercury from U.S. Government
3-1
-------
TABLE 3-1. PHYSICAL AND CHEMICAL PROPERTIES OF MERCURY
Property
Value
Atomic weight
Crystal system
CAS registry number
Atomic number
Valences
Outer electron configuration
lonization potentials, normal, eV
1 st electron
2nd electron
3rd electron
Melting point, °C
Boiling point, ° C
Latent heat effusion, J/g (cal/g)
Latent heat of vaporization, J/g (cal/g)
Specific heat, J/g (cal/g)
Solid
-75.6°C
-40 °C
-263.3°C
Liquid
-36.7°C
210°C
Electrical resistivity, Q-cm, at 20 °C
Density, g/cm3
at20°C
at melting point
at-38.8°C (solid)
atO°C
Thermal conductivity, W/(cm2-K)
Vapor pressure, 25 ° C
Solubility in water, 25 °C
200.59
Rhombohedral
7439-97-6
80
1,2
5d106s2
10.43
18.75
34.20
-38.87
356.9
11.80(2.8)
271.96(65.0)
0.1335(0.0319)
0.141 (0.0337)
0.0231 (0.00552)
0.1418(0.0339)
0.1335(0.0319)
95.8xlO'6
13.546
14.43
14.193
13.595
0.092
2xlO-3mmHg
20-30 //g/L
Source: Reference 1.
3-2
-------
stockpiles were a major secondary source of mercury until July 1994 when Congress suspended sales.2
Major sources of recycled mercury are dental amalgams, scrap mercury from instrument and electrical
manufacturers (including fluorescent lamps), wastes and sludges from research laboratories and electrolytic
refining plants, and mercury batteries.1
Table 3-2 presents the 1991 to 1995 supply-and-demand figures for mercury. The information
contained in Table 3-2 was obtained from the U.S. Geological Survey.2 Values for secondary production,
industry stocks, and industrial consumption are based on voluntary response to USGS questionnaires. The
values presented are based on limited questionnaire response and USGS estimates. As shown in Table 3-2,
the total U.S. supply of mercury in 1995 was 911 Mg (1,002 tons). An estimated 59 percent of the total
supply resulted from primary and secondary mercury production processes. Table 3-2 also shows that of the
total 1995 U.S. mercury supply, approximately 48 percent (436 Mg [480 tons]) was used to meet domestic
demands, while 20 percent met export demands.
The supply-and-demand figures presented in Table 3-2 illustrate a dramatic change in the overall
structure of the industrial demand for mercury in the U.S. Since 1992, U.S. industrial demand for mercury
has steadily declined from 621 Mg (683 tons) to 436 Mg (480 tons), a decrease of 30 percent. U.S. exports
of mercury have undergone greater decline, falling from 977 Mg (1,075 tons) to 179 Mg (197 tons), a
reduction of over 80 percent. Conversely, imports of mercury have risen from 56 Mg (62 tons) in 1991 to
377 Mg (415 tons) in 1995, an increase of 673 percent. The decline of mercury exports and the sharp
increase in mercury imports are due in large part to the suspension by Congress of sales of mercury from U.S.
Government stockpiles.
3.2.2 End-Use
Table 3-3 summarizes the end-use pattern for industrial consumption of mercury in the U.S. in 1991,
1994, and 1995.2 The percentage of the total 1995 mercury supply for industrial consumption that was
consumed by each end-use category is shown in Figure 3-1. The chlor-alkali industry, at 35.3 percent,
accounts for the largest percentage consumption of mercury. Wiring devices and switches manufacture and
measuring and control instruments manufacture represent the second and third largest consumers of mercury
at 19.3 percent and 9.9 percent, respectively. The remaining source categories, as outlined in Table 3-3,
account for approximately 35 percent of total industrial mercury consumption in 1995.2
During the period from 1991 to 1995, the demand picture for mercury has continued to undergo
significant change in the overall demand among industries. The magnitude of these overall changes and the
dramatic change in mercury demand for specific industries is shown in Table 3-3. The most dramatic change
occurred in the battery manufacturing industry where demand dropped from 78 Mg (86 tons) in 1991, to less
than 0.5 Mg (0.6 tons) in 1995. Other industries showing significant decreases in demand from 1991 levels
were measuring and control instrument manufacture and chlorine production.2
Three industries showed an increase in mercury consumption from 1991 to 1995. The most
significant increase occurred in the wiring devices and switches industry, where demand rose from 25 Mg
(27.5 tons) in 1991 to 84 Mg (92.4 tons) in 1995. The dental equipment and supplies industry also
underwent a significant increase in mercury demand, rising from 27 Mg (29.7 tons) in 1991 to 32 Mg
(35.2 tons) in 1995. The only other industry exhibiting an increase in mercury demand is the electric lighting
industry with a slight increase from 29 Mg (31.9 tons) in 1991 to 30 Mg (33 tons) in 1995. Despite the
increases in these three industries, the net change in total U.S. demand for mercury from 1991 to 1995 is a
decrease of 118 Mg (130 tons) or 21 percent from the 1991 level.
The demand decreases in end-use areas will affect the magnitude of mercury emissions in the U.S.
and will lead to secondary impacts. One secondary impact on emissions will be in the area of waste disposal,
particularly in municipal and medical waste combustion. In medical waste, used batteries constitute a major
source of mercury emissions during incineration. Mercury use in battery production decreased by over
99 percent from 1991 to 1995. This decrease should be evident in mercury emissions from both medical
waste and municipal waste incineration. In addition, the significant decrease in demand for the measuring and
control instruments industry may also be felt in emissions from municipal waste incineration. This impact
would occur further in the future than the impact from batteries because of the longer equipment life
expectancy.
3-3
-------
TABLE 3-2. U.S. SUPPLY AND DEMAND FOR MERCURY, 1991 TO 1995
(metric tons, Mg)a
1991
1992
1993
1994
1995
Supply:
No. of producing mines
Mine production, byproduct
Secondary production:
Industrial
Government0
Shipments from NDSd
Imports for consumption
Total supply6
8
58
165
215
103
56
597
9
64
176
103
267
92
702
9
Wb
350
543
40
933
7
W
466
86
129
681
8
W
534
0
377
911
Demand:
Industrial consumption
Exports
Total demand6
554
786
1,340
621
977
1,598
558
389
947
483
316
799
436
179
615
Source: Reference 2.
fFor values in U. S. shprt tons, multiply metric tons (Mg) by 1.1.
W = Withheld to avoid disclosing company proprietary data.
^Secondary mercury shipped from U.S. Department of Energy stocks.
Primary mercury shipped from the National Defense Stockpile.
6 From the table it is obvious that the supply and demand figures do not agree. In discussions of this discrepancy
with J. Plachy (U.S.G.S), he indicated confidence in all figures in this table except industrial consumption. The
individual consumption figures are based in large part on U.S.G.S. estimates and constitute the greatest area of
uncertainty.
TABLE 3-3. END-USE PATTERN OF MERCURY FOR INDUSTRIAL CONSUMPTION
Industry
Chlorine production
Wiring devices and switches
Measuring and control instruments
Dental equipment and supplies
Electric lighting
Other chemical and allied products'3
Laboratory uses
Batteries
Paint
Other uses6
Total demand
Mercury demand, Mga
1991
184
25
70
27
29
18
10
78
6
107
554
1994
135
79
53
24
27
25
24
6
d
110
483
1995
154
84
43
32
30
c
c
<0.5
d
93
436
Source: Reference 2.
aFor values in U. S. short tons, multiply metric tons (Mg) by 1.1.
Includes pharmaceutical uses and miscellaneous catalysts.
GWithheld to avoid disclosing company proprietary data; included in "Other uses."
Not reported separately.
Includes other electrical and electronic uses, other instruments and related products, and unclassified uses.
For 1995, it also includes "Laboratory uses" and "Other chemical and allied products."
3-4
-------
Wiring Devices &
Switches (19.3%)
Other Uses (21.2%)
Dental Equipment
& Supplies (7.3%)
Batteries (0.1%)
Electric Lighting (6.9%)
Measuring & Control
Instruments (9.9%)
Chlorine production (35.3%)
(chlor-alkali industry)
Figure 3-1. End-use pattern of mercury
,2
-------
3.2.3 Emissions
The source of emissions information used to determine a portion of the source categories is the 1994
Toxic Chemicals Release Inventory System (TRI) form required by Section 313 of Title III of the 1986
Superfund Amendments and Reauthorization Act (SARA 313).3 This section requires owners and operators
of Federal facilities and facilities in Standard Industrial Classification (SIC) codes 20-39 that manufacture,
import, process, or otherwise use toxic chemicals to report their annual air releases of these chemicals. All
facilities in these SIC's are not required to report; there are thresholds concerning the number of full-time
equivalent employees and quantity of the compound used, below which facilities are not required to report
releases. The emissions are to be based on source tests (if available); otherwise, emissions may be based on
emission factors, mass balances, or other approaches. Certain source categories (e.g., combustion sources)
that account for substantial mercury emissions, but which are not reported in TRI, were included in the
estimates presented.
It should be noted that, in selected cases, facilities reported to TRI under multiple SIC codes. As a
result, it was difficult to assign emissions to a specific SIC code. In this case, efforts were made to determine
the appropriate SIC codes associated with the emissions. However, if that was not possible, the data were not
used in the analysis. Other reference sources provided additional potential emission source categories that
may not have been included in TRI.4
Another source of emissions information used to determine annual emissions from several of the
source categories is information collection requests authorized under Section 114 of the Clean Air Act
Amendments of 1990 (CAAA). These requests for information are distributed primarily for the purpose of
developing or assisting in the development of implementation plans under Section 110, standards of
performance under Section 111, or emission standards under Section 112 of the CAAA. These requests are
typically in the form of a questionnaire and often request detailed information on air emissions, control
technologies, and related process parameters.
Table 3-4 provides a summary of the estimated 1995 nationwide mercury emissions for those source
types where adequate information was available (i.e., emission factors and production data). Appendix A
presents the data used for each of these estimates, assumptions, and the emission calculations for each
category of these source types. The estimated emissions were based on emission factors provided in this
document or calculated from source test data and appropriate process information, if available.
The total 1995 nationwide mercury emissions estimate was 140 Mg (154 tons) for those source types
identified in Table 3-4. The three specific categories emitting the largest quantitites of mercury were coal
combustion (67.8 Mg [74.6 tons]), municipal waste combustion (26 Mg [29 tons]), and medical waste
combustion (14.5 Mg [16.0 tons]). These three specific categories combined accounted for approximately
78 percent of the total mercury emissions listed in Table 3-4.
Of the five major source types, mercury emissions resulting from combustion categories accounted
for a total of 123.0 Mg (135.6 tons), or approximately 88 percent of the total estimated emissions. Within
the combustion group, the major contributor to mercury emissions was from the combustion of coal, followed
by municipal waste, and medical waste. Coal combustion accounted for 55 percent of the total emissions
from combustion sources and 48 percent of the total emissions from all source types. The other six
combustion areas, wood, municipal waste, medical waste, hazardous waste, sewage sludge, and oil,
collectively accounted for 45 percent of the total emissions from combustion groups and 39 percent of the
total emissions from all source types.
3-6
-------
TABLE 3-4. ESTIMATED 1994-1995 NATIONWIDE MERCURY EMISSIONS
FOR SELECTED SOURCE CATEGORIES
Source type
Mercury emissions
Mg
Mercury and mercury compound production
Primary mercury production
Secondary mercury production
Mercury compound production
0.13
| Tons
Basisa
NA
0.14
NA
No longer mined
Emission factor
No emission factors
Major uses of mercury
Chlorine production
Battery manufacture
Electrical uses
Measurement/control instruments
6.5
5E-04
0.4
0.4
7.1
6E-04
0.5
0.4
1994 TRI report
Emission factor
Emission factor
Emission factor
Combustion sources
Coal combustion
Oil combustion
Municipal waste combustion
Sewage sludge combustion
Hazardous waste combustion
Medical waste combustion
Wood combustion0
67.8
7.6
26
0.9
6.3
14.5
0.1
74.6
8.4
29
0.9
6.9
16.0
0.1
Emission factor/EMF factor
Emission factor
Capacity data/F-factors
Emission factors
EPA/OSW estimates
Capacity data/F-factors
Emission factor
Miscellaneous manufacturing processes
Portland cement production
Lime manufacturing
Carbon black production
Byproduct coke production
Primary lead smelting
Primary copper smelting
Petroleum refining
Municipal solid waste landfills
Geothermal power plants6
Pulp and paper production
4.0
0.1
0.3
0.6
0.1
0.06
0.07
1.3
1.6
4.4
0.1
0.3
0.7
0.1
0.06
NA
0.08
1.4
1.8
Emission factor
Emission factor
Emission factor
Emission factor
Raw materials
Plant data
No emission factor
Test data
Emission factor
Emission factor
Other miscellaneous sources
Mercury catalysts
Dental alloys
Mobile sources
Crematories
Paint
TOTAL
0.6
0.7
140
NA
0.7
NA
0.8
NA
154
No production data
Emission factor
No emission factor
Emission factor
No emission factor
NA = Not applicable.
aSee Appendix A for details of the estimation procedure.
bEmissions summary year not provided.
cEmissions based on 1980 wood-fired boiler capacity.
Emissions based on 1991 production capacity.
eEmissions based on 1993 capacity.
3-7
-------
4.0 EMISSIONS FROM MERCURY PRODUCTION
In 1995, the total supply of metallic mercury (Hg) in the United States was estimated to be 1,045 Mg
(1,152 tons)2. Of this total, approximately 51 percent resulted from secondary production processes
(industrial reclamation); 36 percent was due to imports; about 2 percent was from shipments from the
National Defense Stockpile; and 11 percent was from industry stocks (see Section 3, Figure 3-1). There were
16 facilities in the United States that produced mercury. Of these facilities, eight produced mercury as a
byproduct from gold ore, and eight were secondary mercury production facilities that reclaim mercury.
Mercury emissions occur primarily during the metal production process and during mercury reclamation
processes. In this section, mercury emissions were estimated only for mercury reclamation; no data were
available for the other source types. For mercury reclamation, the mercury emissions for 1994 were
estimated to be 0.13 Mg (0.14 tons).
This section presents information on the identification of the producers and descriptions of typical
production processes. Process flow diagrams are given as appropriate, and known emission control practices
are presented. Estimates of mercury emissions are provided in the form of emission factors wherever data
were available.
4.1 PRIMARY MERCURY PRODUCTION
Mercury is currently produced in the United States only as a byproduct from the mining of gold ores.
Production from mercury ore had occurred at the McDermitt Mine in McDermitt, Nevada, but the mine
ceased operation in 1990. In 1995, eight U.S. gold mines produced metallic mercury as a byproduct;
Table 4-1 presents a list of these mines. As shown in the table, six of the mines are in Nevada, one is in
California, and one is in Utah. None of the operating gold mines in Alaska produce byproduct mercury. In
1995, the total quantity of mercury recovered at these mines was withheld to avoid disclosing company
proprietary data.2
TABLE 4-1. BYPRODUCT MERCURY-PRODUCING GOLD MINES IN
THE UNITED STATES IN 1995
Mine
County and State
Operator
Getchell
Carlin Mines Complex
Alligator Ridge
Enfield Bell
McLaughlin
Mercura
Paradise Peak
Pinson Mine
Humboldt, NV
Eureka, NV
White Pine, NV
Elko, NV
Napa, CA
Tooele, UT
Gabbs, NV
Humboldt, NV
FMC Gold Co.
Newmont Gold Co.
Placer Dome U.S.
Independence Mining Co., Inc.
Homestake Mining Co.
Barrick Mercur Gold Mines, Inc.
FMC Gold Co.
Pinson Mining Co.
Source: Reference 2.
aMine closed in 1997.
In 1994, 86 Mg (95 tons) of primary mercury were shipped from the National Defense Stockpile).2
Because of a suspension of sales in 1994, there were no sales from the stockpile in 1995.
4-1
-------
4.1.1 Process Description
4.1.1.1 Production from Mercury Ores. No process description of the McDermitt Mine operation
will be presented because the existing equipment has been removed from the site, thereby negating any
possibility that the facility could reopen at a future date using the same process and equipment.
4.1.1.2 Byproduct from Gold Ores. Recovery of mercury as a byproduct from gold ores is the only
remaining ore-based production process; all other processes for mercury production are either reclamation or
government surplus stock. A simplified flow diagram depicting mercury recovery from a gold cyanidation
process is shown in Figure 4-1. The flow diagram and process description for mercury recovery from gold
mining is not intended to reflect any specific gold mine operation but to summarize the types of processes and
controls that could be employed. Actual processes will vary from mine to mine.
The incoming gold ore is crushed using a series of jaw crushers, cone crushers, and ball mills. If the
incoming ore is an oxide-based ore, no pretreatment is required, and the crushed ore is mixed with water and
sent to the classifier. If the ore is a sulfide-based ore, it must be pretreated using either a fluidized-bed or
multiple hearth pretreatment furnace (roaster) to convert metallic sulfides to metallic oxides.5 The exhaust
gas from either of these units is sent through wet electrostatic precipitators (ESP's) and, if necessary, through
carbon condensers. The exhaust gas then passes through a scrubber in which SO2 is removed by lime prior to
discharging to the atmosphere. If the treated sulfide ore is high in mercury content, the primary mercury
recovery process occurs from the wet ESP's. If the concentration is sufficiently low, no attempt is made to
recover the mercury for sale. The pretreated ore is mixed with water and sent to the classifier, where the ore
is separated (classified) according to size. Ore pieces too large to continue in the process are returned to the
crusher operation.
From the classifier, the slurry passes through a concentrator to reduce the water content and then to a
series of agitators containing the cyanide leach solution. From the agitators, the slurry is filtered, the filter
cake is sent to disposal, and the filtrate containing the gold and mercury is transferred to the electrowinning
process. If the carbon-in-pulp (CIP) process is used, the cyanide pulp in the agitators is treated with activated
carbon to adsorb the gold and mercury. The carbon is filtered from the agitator tanks and treated with an
alkaline cyanide-alcohol solution to desorb the metals. This liquid then is transferred to the electrowinning
tanks. In the electrowinning process, the gold and mercury are electrodeposited onto a stainless steel wool
cathode, which is sent to a retort to remove mercury and other volatile impurities. The stainless steel wool
containing the gold is transferred from the retort to a separate smelting furnace where the gold is melted and
recovered as crude bullion.
The exhaust gas from the retort, containing mercury, SO2, participate, water vapor, and other volatile
components, passes through condenser tubes where the mercury condenses as a liquid and is collected under
water in the launders. From the launders, the mercury is purified and sent to storage. After passing through
the condenser tubes, the exhaust gas goes through a venturi and impinger tower to remove particulate and
water droplets and then moves through the SO2 scrubber prior to discharging to the atmosphere.
Gold ores in open heaps and dumps also can be treated by cyanide leaching. In this process, the gold
ore is placed on a leaching pad and sprayed with the cyanide solution. The solution permeates down through
the ore to a collection system on the pad, and the resulting pregnant solution is sent to a solution pond. From
this pond, the leachate liquors, which contain gold and mercury, are transferred to the gold recovery area
where the liquor is filtered and sent to the electrowinning process.
4.1.2 Emission Control Measures
Potential sources of mercury emissions from gold processing facilities are at locations where
furnaces, retorts, or other high temperature sources are used in the process and where the mercury is removed
from the launders. The treated gas discharged to the atmosphere is also a source of mercury emissions.
These sources are denoted in Figure 4-1 with a solid circle.
When pretreatment roasting is required, the exhaust gases from the furnace pass through a cyclone to
remove particulate and then move through wet ESP's to remove arsenic, mercury, and some of the SO2. If the
mercury concentration in the gold ore is high, the ESP's will not remove all of the mercury, and an activated
carbon adsorber bed may be required for additional mercury removal. The gas passes through a
4-2
-------
ATMOSPHERE
CYCLONE
1
•
^
•«
PRETREATMENT
ROASTER
to
»
V
AQUEOUS STREAM
~. MERCURY fc
•"" RECOVERY —
V
\
•»
^. LIME S02
SCRUBBER
STACK
ORE
CRUSHED
ORE
CLASSIFIER
T OVERFLOW
F H-n -M
LIME S02
SCRUBBER
«-
IMPINGER
TOWER
Hg
POTENTIAL MERCURY EMISSION SOURCES
Figure 4-1. Major components of mercury recovery from gold ores.
-------
lime scrubber to remove SO2; if the SO, concentration is low, a caustic scrubber may be used.5 From the
scrubber, the gas is discharged through the stack to the atmosphere. Essentially the same emission control
measures are used from the exhaust gas from the retort. After the gas passes through the condenser tubes to
remove the mercury, a venturi and a cyclone are used to remove particulate and water droplets. These
controls are followed by the lime scrubber to remove the SO2 prior to discharging the clean gas to the
atmosphere.
4.1.3 Emissions
The major sources of mercury emissions for gold processing facilities are the pretreatment roaster (if
required) and the retort. Other sources of emissions are from the purification process after removal of
mercury from the launders and the stack emissions to the atmosphere. No emissions data have been
published for facilities producing mercury as a byproduct from gold ore. Limited data were published for
emission sources at facilities that produced mercury from the primary ore.6'7 While treatment techniques to
recover the mercury, after the mercury has been vaporized in a retort or furnace, and the emission sources are
very similar to production from primary ore, the overall production process is considerably different. The
emission factors for production from primary ore should not be used to estimate emissions from gold mining
operations.
4.2 SECONDARY MERCURY PRODUCTION
There are two basic categories of secondary mercury production: recovery of liquid mercury from
dismantled equipment and mercury recovery from scrap products using extractive processes. On an annual
basis, the total quantity of mercury recovered as liquid mercury is much greater than that recovered by
extractive processes. Three areas that have contributed to a large proportion of the liquid mercury recovery
category are: (1) dismantling of chlorine and caustic soda manufacturing facilities; (2) recovery from
mercury orifice meters used in natural gas pipelines; and (3) recovery from mercury rectifiers and
manometers. In each of these processes, the liquid mercury is drained from the dismantled equipment into
containers and sold on the secondary mercury market. The second category involves the processing of
scrapped mercury-containing products and industrial wastes and sludges using thermal or chemical extractive
processes because the mercury cannot be decanted or poured from the material. One mercury recycler
(Bethlehem Apparatus Company) estimated that this second category accounted for 15 to 20 percent of the
total quantity of mercury reported as recycled from industrial scrap in 1995.
In 1995, an estimated 534 Mg (588 tons) of mercury was recycled from industrial scrap.2 These
totals do not include in-house mercury reclamation at industrial plants using mercury. According to the
USGS, eight major companies were reported to be involved in secondary mercury production using purchased
scrap material (mercury recyclers) in 1995.2 The three dominate companies in this market are Bethlehem
Apparatus Company in Hellertown, Pennsylvania; D. F. Goldsmith in Evanston, Illinois; and Mercury
Refining Company in Albany, New York.
4.2.1 Process Description
The predominant method to recover metallic mercury for recycling from scrap products is thermal
treatment.l Figure 4-2 provides a general process diagram for secondary mercury recovery at a battery
plant.8 This process is generally representative of the recovery of mercury by thermal treatment of scrap.
Generally, the mercury-containing scrap is reduced in size and is heated in retorts or furnaces at about
538 °C (1000°F) to vaporize the mercury. The mercury vapors are condensed by water-cooled condensers
and collected under water.8
Vapors from the condenser, which may contain particulate, organic compounds, and possibly other
volatile materials from the scrap, are combined with vapors from the mercury collector line. This combined
vapor stream is passed through an aqueous scrubber to remove particulate and acid gases (e.g., HC1, SO2).
From the aqueous scrubber, the vapor stream passes through a charcoal filter to remove organic components
prior to discharging into the atmosphere.
The collected mercury is further purified by distillation, collected, and then transferred to the filling
area. In the filling area, special filling devices are used to bottle small quantities, usually 0.464 kg (1 Ib) or
2.3 kg (5 Ib) of distilled mercury. With these filling devices, the mercury flows by gravity through tubing
from a holding tank into the flask until the flask overflows into an overflow bottle.
4-4
-------
Hg
SCRAP
CONDENSER
RETORT
CHAMBER
Hg
H20
Hg COLLECTOR
DENOTES POTENTIAL MERCURY EMISSION SOURCE
H2O
i
SCRUBBER
DISTILLATION
ATMOSPHERE
CHARCOAL
FILTERS
FILLING AREA
Figure 4-2. Process flow diagram for secondary recovery at a battery plant.
-------
The desired amount of mercury is dispensed into the shipping bottle by opening a valve at the bottom of the
flask. The shipping bottle is then immediately capped after the filling and sent to the storage area.8
4.2.2 Emission Control Measures
Information on specific emission control measures is very limited and site specific. If a scrubber is
used, as shown in Figure 4-2, mercury vapor or droplets in the exhaust gas may be removed by condensation
in the spray. There is no information to indicate that chemical filters would be effective in removing mercury
vapors. No information was found for other control measures that are used in secondary mercury production
processes. Concentrations in the workroom air due to mercury vapor emissions from the hot retort may be
reduced by the following methods: containment, local exhaust ventilation, dilution ventilation, isolation,
and/or personal protective equipment. No information was provided to indicate that these systems are
followed by any type of emission control device.8 Vapor emissions due to mercury transfer during the
distillation or filling stages may be reduced by containment, ventilation (local exhaust or ventilation), or
temperature control.
4.2.3 Emissions
During production of mercury from waste materials using an extractive process, emissions may vary
considerably from one type of process to another. Emissions may potentially occur from the following
sources: retort or furnace operations, distillation, and discharge to the atmosphere from the charcoal filters.
The major mercury emission sources are due to condenser exhaust and vapor emissions that occur during
unloading of the retort chamber. These sources are indicated in Figure 4-2 by a solid circle. Mercury
emissions also can occur in the filling area when the flask overflows and during the bottling process. '9
Mercury Refining Company reported results from two emission test studies conducted in 1994 and
1995 that showed average mercury emissions of 0.85 kg/Mg (1.7 Ib/ton) of mercury recovered.10 In 1973,
emission factors were estimated to be 20 kg (40 Ib) per megagram (ton) of mercury processed due to
uncontrolled emissions over the entire process.6
Mercury emission data were reported in the 1994 TRI only for Mercury Refining Company, Inc., in
Albany, New York, and Bethlehem Apparatus Company in Hellertown, Pennsylvania.3 Mercury Refining
reported plant emissions to the atmosphere of 116 kg (255 Ib) for 1994, and Bethlehem Apparatus reported
plant emissions to the atmosphere of 9 kg (20 Ib) for 1994. The other major recycler, D. F. Goldsmith, does
not use extractive processes; their recycling is primarily from purchases of mercury decanted from old
equipment. Mercury emission data were not available for the other five facilities.
The total mercury emissions were estimated to be 0.13 Mg (0.14 tons) for 1994; see Appendix A for
calculations.
4.3 MERCURY COMPOUNDS PRODUCTION
The production of mercury compounds presents a potential source of release of mercury into the
atmosphere. Table 4-2 lists several producers of inorganic mercury compounds. No U.S. producers of
phenylmercury acetate (PMA) or thimerosal (merthiolate) were identified.11 No facility reported mercury
emissions in the 1994 TRI.3
TABLE 4-2. MERCURY COMPOUND PRODUCERS
Producer
Elf Atochem North America, Inc., Chemical
Specialties Division
GFS Chemicals, Inc.
Johnson Matthey, Inc.
R.S.A Corporation
Location
Tulsa, OK
Columbus, OH
Ward Hill, MA
Danbury, CT
Compound(s)
HgF2,Hg2F2
HgBr2,HgI2,Hg(N03)2,
HgS04
Hg2(N03)2
Hg(SCN)2
Source: Reference 11.
4.3.1 Process Description
4-6
-------
Numerous inorganic mercury compounds are produced annually in the United States using metallic
mercury as the starting material. The production processes for mercuric chloride and mercuric oxide were
selected to serve as typical examples. The production processes for each compound have been studied at
Troy Chemical Corporation.12 A synopsis of these two production processes is provided below; additional
information can be found in Reference 8.
4.3.1.1 Mercuric Chloride and Mercurous Chloride. The production of these two compounds occurs
by the direct reaction of mercury with chlorine gas according to the following equations:
2Hg° + C12 - Hg2Cl2
Hg° + C12 - HgCl2
Figure 4-3 presents a process diagram for the production of mercuric chloride. Elemental mercury
(Stream A) is pumped from a holding tank into a reactor where it reacts with excess chlorine gas (Stream B).
The reaction products (Stream C) are ducted to a precipitation unit where the dry product (HgCl2) settles and
is raked out. Mercuric chloride (Stream D) is packaged and sealed in drums for shipping. >iy The exhaust
from the reactor (Stream E) is sent to a caustic scrubber where unreacted mercury is recovered and is then
recycled back (Stream F) to the reactor. A similar process is used to produce mercurous chloride.
4.3.1.2 Mercuric Oxide. Two different processes have been used for mercuric oxide production:
(1) production via mercuric chloride and (2) production via mercuric nitrate intermediates. Both processes
are shown in Figure 4-4.
In production via mercuric chloride, mercury (Stream A) and chlorine in brine solution (Stream B)
are mixed in a reactor where mercuric chloride is produced in solution by oxidation of the liquid mercury.
The mercuric chloride (Stream C) is then transferred to a second reactor and an aqueous caustic (NaOH)
solution is added, resulting in the formation of mercuric oxide. The mercuric oxide precipitate (Stream D) is
then washed, dried, screened, and packaged9.
In the process using the mercuric nitrate intermediate, (also shown in Figure 4-4), mercury
(Stream A) and nitric acid (Stream B) are combined in a reactor, resulting in the formation of mercuric nitrate
(Hg(NO3)2). The mercuric nitrate (Stream C) is then transferred to a second reactor where mercuric oxide is
precipitated by adding an aqueous caustic solution (NaOH). The mercuric oxide (Stream D) is washed, dried,
ground, and packaged.8
4.3.2 Emission Control Measures
No information was found on specific emission control devices to remove or treat the mercury
emissions. Only methods designed to reduce the workplace concentrations without subsequent treatment
were presented Methods suitable for reducing workroom air concentrations of mercury during the
production of mercury compounds are similar to those described for primary and secondary mercury
processing. Particulate concentrations in the workplace resulting from several process operations (e.g.,
addition of dry chemicals to reactors, filtration, drying, grinding, and packaging) may be reduced by
containment, exhaust ventilation, dilution ventilation, and personal protective equipment. Mercury vapor
concentrations from mercury transfer to reactors and from the reactors may be reduced by containment.
During mercuric oxide production, grinding and packaging operations are done in an enclosed system
under vacuum, including material transfers. A cyclone dust collector separates fine dust from product-sized
HgO particles, which are channeled to the packaging station. The fine dust is collected and transferred
periodically to fiber drums. The vacuum pump discharge also goes through a cyclone dust separator before it
exhausts to the roof. Collected dust is recycled through the grinder.12
4.3.3 Emissions
During the production of these compounds, emissions of mercury vapor and particulate mercury
compounds may occur at the following sources: reactors, driers, filters, grinders, and transfer operations.
These emission sources are indicated in Figures 4-3 and 4-4 by a solid circle.
4-7
-------
4-8
-------
VENT TO AIR 0.1%
Hg
NaOH
t
SCRUBBER
t
Hg
CARBON
FILTER
i
^
1 ^-
A
EXCESS ^.
CU B
EXHAUST HOOP • "SZZXS"
REACTOR
(COMBUSTION)
°'2 D ,
c
PRECIPITATION
UNIT
D
DRUMMING
(1 .9% OF ORIGINAL
AMOUNT OF HG
RECYCLED)
*-HgC,2
I
HEAT
Figure 4-3. Mercuric/mercurous chloride production.
-------
Hg
Cl, /Brine
4
i
•
Reaction 1:
Oxidation
to HgCI2
C
r
Reaction 2:
Precipitation
ofHgO
D
Washing
I
•
Drying
Screening
Packaging
NaOH
Hg
HN03 .
1
1
•
Reaction 1 :
Formation of
Hg(N03)2
C
p
Reaction 2:
Precipitation
ofHgO
Filtration of
Cryst. Red Oxide
or Powd.
Yellow Oxide
NaOH '
D
Filtrate
Oven
Drying
Enclosed
Dumping
Station
Grinder
Vacuum
t
Packaging
in50-lb
Drums
Feeding
Figure 4-4. Mercuric oxide production via mercuric chloride and mercuric nitrate intermediates.
-------
Emission factors are not available for production of mercury compounds. No test data for mercury
emissions were found that would permit the calculation of emission factors.
4-11
-------
5.0 EMISSIONS FROM MAJOR USES OF MERCURY
Emissions from industrial processes that use mercury are discussed in this section. The four
commercial uses discussed in this section are (1) chlorine production using the mercury cell process,
(2) primary battery production, (3) production of electrical lighting, wiring devices, and electrical switches,
and (4) production of measuring and control instruments. A summary of the estimated mercury emissions
from each of these industries is as follows:
Industry Emissions, Mg (tons)
Chlorine production 6.5(7.1)
Primary battery production 5 E-04 (6 E-04)
Electrical equipment production 0.4 (0.5)
Measurement/control instruments 0.4 (0.5)
This section is divided into four subsections, one devoted to each of the four commercial uses listed
above. Each of the subsections presents a general discussion of the production process and where mercury is
used in the process, descriptions of existing mercury emission control measures, and estimates of mercury
emission factors. The level of detail varies according to the availability of information, particularly for
emissions where data may be incomplete or absent.
5.1 CHLORINE PRODUCTION USING THE MERCURY CELL PROCESS
In 1996, the mercury cell process, which is the only chlor-alkali process using mercury, accounted for
12.1 percent of all U.S. chlorine production.13 Although most chlor-alkali plants use diaphragm cells, the
mercury cell is still used at 14 facilities. The chlor-alkali industry, however, is gradually moving away from
mercury cell production and toward a membrane cell process because the membrane cell process does not use
mercury, is 12 to 14 percent more energy efficient, and produces mercury free products. Table 5-1 presents
the location and capacity of mercury cell chlor-alkali production facilities operating in the U.S. in 1996.11
5.1.1 Process Description
The mercury cell process consists of two electrochemical cells, the electrolyzer and the decomposer.
A basic flow diagram for a mercury cell chlor-alkali production operation is shown in Figure 5-1.
Saturated (25.5 weight percent) purified sodium or potassium brine (Stream A) flows from the main
brine saturation section, through the inlet end box, and into the electrolyzer cell. The cell is an elongated
trough that is inclined approximately 1 ° to 2.5 ° with sides that are typically lined with rubber. The brine
flows between stationary activated titanium anodes suspended from above into the brine; mercury, which is
the cathode, flows concurrently with the brine over a steel base.
The electrochemical reaction that occurs at the titanium anodes is shown in equation (1); the reaction
at the mercury cathode is shown in equation (2); and the overall reaction is shown in equation (3).
2Cl--Cl2t+2e (1)
Hg + 2Na+ + 2e - Na-Hg amalgam (2)
Hg + 2Na+ + 2C1- - C121 + Na-Hg amalgam (3)
5-1
-------
TABLE 5-1. 1996 MERCURY CELL CHLOR-ALKALI PRODUCTION FACILITIES
Facility3
Ashta Chemicals, Inc.
Georgia-Pacific West, Inc.
The BFGoodrich Company, BFGoodrich
Specialty Chemicals
Holtrachem Manufacturing Company
Occidental Chemical Corporation, Basic
Chemicals Group, Electrochemicals
Olin Corporation
Pioneer Chlor-Alkali Company, Inc.
PPG Industries, Inc., Chemicals Group
Vulcan Materials Company, Vulcan
Chemicals Division
Location3
Ashtabula, OH
Bellingham, WA
Calvert City, KY
Reigelwood, NC
Orrmgton, ME
Deer Park, TX
Delaware City, DE
Muscle Shoals, AL
Augusta, GA
Charleston, TN
St. Gabriel, LA
Lake Charles, LA
New Martinsville, WV
Port Edwards, WI
TOTAL
Capacity
103 Mg/yr
36
82
109
48
76
347
126
132
102
230
160
233
70
65
1,816
103 tons/yr
40
90
120
53
80
383
139
146
112
254
176
256
77
72
1,998
1991 emissions,
lb/yrc
N/A
200
1,206
528
735
1,290
532
184
1,540
1,892
1,240
1,440
1,085
1,030
12,902
1994 emissions,
lb/yr3
1,660
1,290
842
1,095
582
1,040
510
233
1,317
1,509
N/A
1,230
1,130
N/A
12,438
aReference
b
SRI figures adjusted based on questionnaire responses. References 11, 15-27.
°Emissions data based on responses to Section 114 information collection requests from the following: References 15-27.
TRI emissions data. Reference 3.
-------
BASIC TREATMENT CHEMICALS
(SODA ASH, CAUSTIC LIME,
ACID, CaCL2, ETC.)
SOLID
NaCL
FEED
PRODUCT
CHLORINE
CHLORINE
OTHER
BRINE
DECHLORINATOR
SPENT BRINE
MAIN BRINE
SATURATION,
PURIFICATION, AND
FILTRATION
TREATED BRINE (A)
INLET END-BOX -
END-BOX
VENTILATION
SYSTEM
END-BOX
VENTILATION SYSTEM
(B)
COOLING, DRYING,
COMPRESSION, AND
LIQUEFACTION
(C)
ELECTROLYZER
-OUTLET END-BOX
• END-BOX
-*• VENTILATION
SYSTEM
WATER COLLECTION
SYSTEM
END-BOX
VENTILATION SYSTEM
(D)
Hg PUMP
DECOMPOSER
(DENUDER)
AMALGAM
I
(G)
CAUSTIC SODA
SOLUTION
HYDROGEN
GAS
BYPRODUCT
(F)
DENOTES POTENTIAL MERCURY EMISSIONS
COOLING,
FILTRATION,
AND MERCURY
RECOVERY
PRODUCT
• CAUSTIC
SODA
Figure 5-1. Basic flow diagram for a mercury-cell chlor-alkali operation9.
5-3
-------
Chlorine gas (Stream B), formed at the electrolyzer anode, is collected for further treatment. The
spent brine (Stream C) contains 21-22 weight percent NaCl and is recycled from the electrolyzer to the main
brine saturation section through a dechlorination stage. Sodium forms an amalgam, containing from 0.25 to
0.5 percent sodium, at the electrolyzer cathode. The resulting amalgam flows into the outlet end box at the
end of the electrolyzer. In the outlet end box, the amalgam is constantly covered with an aqueous layer to
reduce mercury emissions. The outlet end box also allows removal of a thick mercury "butter" that is formed
by impurities. The sodium amalgam (Stream D) flows from the outlet end box into the second cell, the
decomposer.
The decomposer is a short-circuited electrical cell in which the sodium amalgam acts as the anode
and graphite as the cathode in sodium hydroxide solution. Fresh water is added to the decomposer where it
reacts with the sodium amalgam to produce elemental mercury (Stream E), sodium hydroxide (Stream F), and
byproduct hydrogen gas (Stream G). Stream E is then stripped of sodium and the mercury (Stream H) is
recirculated back to the electrolyzer through the inlet end box. The inlet end box provides a convenient
receptacle on the inlet end of the electrolyzer to receive the recycled mercury from the decomposer and keep it
covered with an aqueous layer to reduce mercury emissions.
The caustic soda solution (Stream F) leaving the decomposer at a typical concentration of 50 weight
percent is cooled and filtered. The byproduct hydrogen gas (Stream G) may be vented to the atmosphere,
burned as a fuel, or used as a feed material for other processes.9'14
5.1.2 Emission Control Measures
Several control techniques are employed to reduce the level of mercury in the hydrogen streams and
in the ventilation stream from the end boxes. The most commonly used techniques are (1) gas stream
cooling, (2) mist eliminators, (3) scrubbers, and (4) adsorption on activated carbon or molecular sieves.
Mercury vapor concentrations in the cell room air are not subject to specific emission control measures but
rather are maintained at acceptable worker exposure levels using good housekeeping practices and equipment
maintenance procedures.
Gas stream cooling may be used as the primary mercury control technique or as a preliminary
removal step to be followed by a more efficient control device. The hydrogen gas stream from the
decomposer exits the decomposer at 93° to 127°C (200° to 260°F) and passes into a primary cooler. In this
indirect cooler, a shell-and-tube heat exchanger, ambient temperature water is used to cool the gas stream to
32° to 43 °C (90° to 110°F). A knockout container following the cooler is used to collect the mercury. If
additional mercury removal is desired, the gas stream may be passed through a more efficient cooler or
another device. Direct or indirect coolers using chilled water or brine provide for more efficient mercury
removal by decreasing the temperature of the gas stream to 3° to 13 °C (37° to 55 °F). If the gas stream is
passed directly through a chilled water or brine solution, the mercury condenses and is collected under water
or brine in lined containers. Mercury in the ventilation air from the end boxes can be removed using either
direct or indirect cooling methods. In situations where the ventilation air from the exit box contains mercuric
chloride particulates, the direct method may be preferred. The direct cooling method not only cools the gas
stream, but also removes the particulate from the stream. Regardless of the gas stream treated, the water or
brine from direct contact coolers requires water treatment prior to reuse or discharge because of the dissolved
mercury in the liquid.
Mist eliminators can be used to remove mercury droplets, water droplets, or particulate from the
cooled gas streams. The most common type of eliminator used is a fiber pad enclosed by screens. With the
fiber pad eliminator, trapped particles are removed by periodic spray washing of the pad and collection and
treatment of the spray solution.
Scrubbers are used to chemically absorb the mercury from both the hydrogen stream and the end box
ventilation streams. The scrubbing solution is either depleted brine from the mercury cell or a sodium
hypochlorite (NaOCl) solution. These solutions are used in either sieve plate scrubbing towers or packed-bed
scrubbers. Mercury vapor and mist react with the sodium chloride or hypochlorite scrubbing solutions to
form water-soluble mercury complexes. If depleted brine is used, the brine solution is transferred from the
scrubber to the mercury cell where it is mixed with fresh brine and the mercury is recovered by electrolysis in
the cell.
Sulfur- and iodine-impregnated carbon adsorption systems are commonly used to reduce mercury
levels in the hydrogen gas and end box streams. This method requires pretreatment of the gas stream by
5-4
-------
primary or secondary cooling followed by mist eliminators to remove about 90 percent of the mercury content
of the gas stream. As the gas stream passes through the carbon adsorber, the mercury vapor is initially
adsorbed by the carbon and then reacts with the sulfur or iodine to form the corresponding mercury sulfides
or iodides. Depending upon the purity requirements and final use for the hydrogen gas, several adsorber beds
may be connected in series to reduce the mercury levels to the very low ppb range.
A proprietary molecular sieve adsorbant was used by five facilities to remove mercury from the
hydrogen gas stream until 1984 when the supply of the adsorbant was discontinued by the manufacturer. The
technique used dual adsorption beds in parallel such that while one bed was being used for adsorption, the
other was being regenerated. A portion of the purified hydrogen gas from one adsorption bed was diverted,
heated, and used to regenerate the second adsorption bed.9
In addition to the control measures described above, the conversion of mercury cell chlor-alkali
plants to the membrane cell process would eliminate all mercury emissions from this industry. As mentioned
earlier, the chlor-alkali industry is gradually moving away from mercury cell production and toward the
membrane cell process.
5.1.3 Emissions
The three primary sources of mercury emissions to the air are (1) the byproduct hydrogen stream, (2)
end box ventilation air, and (3) cell room ventilation air. Emission sources (1) and (2) are indicated on
Figure 5-1 by solid circles.
The byproduct hydrogen stream from the decomposer is saturated with mercury vapor and may also
contain fine droplets of liquid mercury. The quantity of mercury emitted in the end box ventilation air
depends on the degree of mercury saturation and the volumetric flow rate of the air. The amount of mercury
in the cell room ventilation air is variable and comes from many sources, including end box sampling,
removal of mercury butter from end boxes, maintenance operations, mercury spills, equipment leaks, cell
failure, and other unusual circumstances.9
Mercury emissions data for end box ventilation systems and hydrogen gas streams from 21 chlor-
alkali production facilities are included in a 1984 EPA report.9 The dates of the emission tests included in the
report range from 1973 to 1983. These data should not be applied to current mercury cell operations in part
because of the variability in the emission data reported. No evaluation of the variability in the data was
presented in the EPA report. In addition, control techniques at current facilities differ from the techniques
employed during these tests. Even if the general technique (e.g., scrubbing, carbon adsorption) is the same,
improvements in control efficiency have likely been made since these tests were conducted.
The most recent AP-42 section on the chlor-alkali process presents emission factors for emissions of
mercury from mercury cell hydrogen vents and from end boxes.28 These emission factors are based on two
1972 emission test reports. The emission factors were not used to estimate emissions from the chlor-alkali
industry because process operations and control techniques have likely changed considerably since these tests
were conducted. If available, recent test data and information on control system design and efficiency should
be used to estimate emissions for site-specific mercury cell operations.
Total 1994 mercury emissions for this industry are estimated to be 6.5 Mg (7.1 tons); see Appendix
A for details.
5.2 BATTERY MANUFACTURING
Three main types of primary batteries have historically used mercury: (1) mercuric oxide (also
known as mercury-zinc); (2) alkaline; and (3) zinc-carbon (or Leclanche). The mercury served two principal
functions: (1) in the cathode of mercuric oxide batteries and (2) as an inhibitor for corrosion and side
reactions in zinc-carbon and alkaline batteries. Zinc air, silver oxide, and alkaline manganese button cell
batteries also use very small amounts of mercury to control gassing. Prior to the late 1980's, most primary
batteries and some storage batteries contained mercury in the form of mercuric oxide (HgO), zinc amalgam
(Zn-Hg), mercuric chloride (HgCl2), or mercurous chloride (Hg2Cl2). Since 1989, the use of mercury in
primary batteries has decreased from 250 Mg (275 tons) in 1989 to less than 0.5 Mg (<0.6 tons) in 1995 (see
Table 3-2). The two major reasons for this decrease were reduction in the production of mercuric oxide
batteries and the discontinued use of mercury as a corrosion inhibitor in alkaline and zinc carbon batteries.
This decrease occurred as a result of the enactment on May 13, 1996 of the "Mercury-Containing and
5-5
-------
Rechargeable Battery Management Act" (Public Law 104-142). Upon enactment, this law prohibited the sale
of mercuric oxide button cells and alkaline batteries containing mercury as well as the use of mercury as a
corrosion inhibitor in zinc carbon batteries. Under the law, it also became illegal to sell larger mercuric oxide
batteries unless the manufacturer or importer provides purchasers with information on licensed recycling or
disposal facilities. The sale of mercury oxide button cells was discontinued as early as 1993 and use of
mercury as a corrosion inhibitor in alkaline batteries ceased in 1992-1993.29 Since the only type of battery
that uses mercury to any measurable degree is the mercuric oxide, it is the only battery discussed in this
section.
Table 5-2 presents the U.S. manufacturers and production sites for mercuric oxide, alkaline
manganese, or zinc-carbon batteries in 1996. The only facilities that produce mercuric oxide batteries are
AMC, Inc. and Eveready in Bennington, Vermont.
TABLE 5-2. MERCURIC OXIDE, ALKALINE MANGANESE, OR ZINC-CARBON
BATTERY MANUFACTURERS IN 1996
Manufacturer
Production site
Alexander Manufacturing Company (AMC, Inc.)
Duracell, USA
Eagle-Picher Industries, Inc.
Eveready Battery Company, Inc.
Muteca
Rayovac Corp.
Mason City, IA
Cleveland, TN
LaGrange, GA
Lancaster, SC
Lexington, NC
Colorado Springs, CO
Maryville, MO
Fremont, OH (to be closed)
Bennington, VT
Asheboro, NC (2 plants)
Columbus, GA (Corporate offices)
Madison, WI
Fennimore, WI
Portage, WI
Source: References 29 and 33.
aMutec is a joint venture between Eastman Kodak and Panasonic.
Mercuric oxide batteries were produced in two sizes: button cells and larger sizes. Button cells are
small, circular, relatively flat batteries that were used in transistorized equipment, walkie-talkie's,
photoelectric exposure devices, hearing aids, electronic watches, cardiac pacemakers, and other items
requiring small batteries. Larger mercuric oxide batteries are produced for a variety of medical, military,
industrial, and other nonhousehold equipment.
5.2.1 Process Description
The basic flow diagram for the manufacture of mercuric oxide batteries is shown in Figure 5-2. The
mercuric oxide-zinc cells use mercuric oxide (mixed with graphite and manganese dioxide) as the cathode.
The anode is a zinc-mercury amalgam. According to the NEMA, the basic flow diagram in Figure 5-2 was
based on a Rayovac mercuric oxide battery production facility in Portage, Wisconsin, that discontinued
production of this battery type in 1986.30
In the production of the cathodes, mercuric oxide (Stream A), manganese dioxide (Stream B), and
graphite (Stream C) are manually metered through a hopper to the blending area.9 The resulting mixture
(Stream D) is sent to a processing unit where it is compacted into tablets by "slugging" (compression in a
rotary pressing device to a specified density). These tablets are then granulated into uniformly sized
5-6
-------
ANODE
(MERCURIC OXIDE-ZINC CELLS)
Uncontrolled
.DENOTES POTENTIAL MERCURY EMISSION SOURCE
Emissions
Monitoring*
HgO.
Graphite-
MnO2 .
Amalgamating
Emissions
Mixing and
Blending
Processing,
Blending, Anode
Gel**
Emissions
Processing
(Pelleting,
Granulation,
Pelleting,
Consolidation)
Baghouse
Cell
Manufacturing
Monitoring*
Batteries (Button Cells)
CATHODE
(MERCURIC OXIDE-ZINC 6 MERCURIC OXIDE-CADMIUM CELLS)
*Mercuiy emissions monitored to ensure compliance with state limits.
''Process operations controlled to maintain compliance with state emission limits.
Figure 5-2. General flow diagram for mercuric oxide battery (button cell) manufacture.
-------
particles, and then pelletized in a rotary press. The pellets are consolidated into small metal cans less than
1.3 cm (0.5 in.) in diameter.8
For the production of the anodes, elemental mercury (Stream E) and zinc powder (Stream F) are
metered from hoppers or hold tanks into an enclosed blender to produce a zinc-mercury amalgam. The
amalgam (Stream G) is sent to a processing area where it is blended and the anode gel formed.8 Highly
controlled process operations are enforced to maintain mercury vapor emissions to levels within compliance
to State limits.
The completed anodes and cathodes then are sent to the cell manufacturing area. Separators,
electrolyte, and other components are assembled with the anode and cathode to produce the HgO-Zn cell.
Assembly may be automatic or semiautomatic. The assembled cathode, anode, electrolyte, and cover are
sealed with a crimper. Depending on the design, other components may be added. Those additional
components may include an insulator, an absorber, and a barrier.
An integrated mercuric oxide battery plant may also produce HgO and recycled mercury onsite.
Mercuric oxide production is discussed in Section 4 under mercury compound production. Secondary
recovery of mercury at the battery plant is discussed in Section 4 under secondary mercury production.
5.2.2 Emission Control Measures
Baghouses are used to control particulate emissions from the mixing/blending and processing steps
in the production of cathodes. Mercury vapor emissions from the anode processing and cell manufacturing
areas are generally discharged to the atmosphere uncontrolled. Ventilation air in the assembly room is
recirculated through particulate filters. One plant reported an average of 73 percent mercury vapor removal
efficiency in the cell assembly room when an air handler system, consisting of a particulate prefilter and a
charcoal filter, was operated using 75 percent recirculating air and 25 percent fresh air.8
In addition to the emission control measures, other methods can be used to reduce potential worker
exposure in the workplace.8 Table 5-3 summarizes the types of methods used in the workplace to reduce
worker exposure to mercury vapor and particulate during battery manufacturing.
TABLE 5-3. METHODS FOR REDUCING WORKER EXPOSURE TO MERCURY
EMISSIONS IN BATTERY MANUFACTURING
Control methods
Process modification and substitution
Containment
Ventilated enclosure
Local exhaust ventilation
Temperature control
Dilution ventilation
Isolation
Mercury removal from air stream
Personal protective equipment
Particulate
Xa
Xa
xb,c
xa,b,c
xa,b,c
Xa,c
Xa,b,c
Xa,b
Vapor
xd,e
xd,e
xd,e
xd,e
xd,e
xd,e
Source: Reference 8.
fP articulate emissions during loading of mixers and blenders in cathode preparation.
Particulate emissions from grinding, slugging, and pelletizing in cathode production.
^Particulate emissions from drying, screening, and pelletizing in anode production.
Vapor emissions from blending, drying, and pelletizing during anode production.
eVapor emission from product components.
Reject materials such as anodes, cathodes, chemical mixes, and cells can be stored under water to
suppress mercury vaporization.
5-S
-------
Machinery for grinding, mixing, screening, pelletizing, and/or consolidating can be enclosed with
little or no need for worker access. Two mercuric oxide button cell manufacturers in 1983 were using such
enclosures and glove boxes to reduce worker exposure. Iris ports allowed access to the enclosed equipment.
Exhaust airstreams are generally ducted to a baghouse. These facilities also used ventilated enclosures to
store completed anodes and cathodes on the cell assembly lines; the exhaust air takeoffs from these
enclosures led to a baghouse.
5.2.3 Emissions
During the manufacture of mercuric oxide batteries, mercury may potentially be emitted from several
processes as particulate and as vapor emissions. These release points are indicated in Figure 5-2 by a solid
circle. The processes include grinding, mixing, sieving, pelleting, and/or consolidating.
The only reported emission factor for a mercuric oxide production facility was for one plant in
Wisconsin.31 This facility used a combination of a baghouse and charcoal filter to treat the exhaust
ventilation air. Annual use of mercury was 36.17 Mg (39.8 tons) and annual emissions were reported as
36.3 kg (80 Ib) of mercury as HgO particles. For this specific facility, the mercury emission factor would be
1.0 kg/Mg (2.0 Ib/ton) of mercury used. This facility discontinued production of mercuric oxide batteries in
1986:30
This emission factor should be used with extreme caution for several reasons. The facility ceased
production of mercuric oxide batteries and the emission controls cited in Reference 31 are probably not
applicable to facilities currently producing this type of battery. Although it is not specifically stated in
Reference 31, it is also presumed that the mercury emission quantity was an estimate by the manufacturer
because no reference is made to any emissions testing performed at the facility. Moreover, this factor is for
1 year at one specific site so that extrapolation of this factor to current mercuric oxide battery manufacturing
facilities can lead to erroneous results.
Based on another study, the emission source rates from an integrated mercury button cell plant are
summarized in Table 5-4.9 Major emission points were the pelleting and consolidating operations (up to 42
g/d; 0.094 Ib/d) and cell assembly (29 g/d; 0.063 Ib/d). Emission controls were not in place for mercury
vapor emissions from the main plant.
Total 1995 mercury emissions for this industry are estimated to be 5 x 10"4 Mg (6 x 10"4 tons); see
Appendix A for details.
5.3 ELECTRICAL USES
Mercury is one of the best electrical conductors among the metals and is used in five areas of
electrical apparatus manufacturing: electric switches, thermal sensing elements, tungsten bar sintering,
copper foil production, and fluorescent light manufacture.
5.3.1 Electric Switches
The primary use of elemental mercury in electrical apparatus manufacturing is in the production of
silent electric wall switches and electric switches for thermostats. The mercury "buttons" used in wall
switches consist of mercury, metal electrodes (contacts), and an insulator. The thermostat switches are
constructed of a short glass tube with wire contacts sealed in one end of the tube. An outside mechanical
force or gravity activates the switch by causing the mercury to flow from one end of the tube to the other, thus
providing a conduit for electrical flow.
The National Electrical Manufacturers Association (NEMA) was contacted in 1993 to identify
manufacturers of electric switches that may use mercury in their devices.32 Of the 15 companies identified by
NEMA in 1993, 10 currently use no mercury at their production facilities. General Electric Corporation
stated that thermostats, both with and without mercury, were produced at their Morrison, Illinois, facility.
Honeywell, Inc. produces microswitches that contain mercury at their Freeport, Illinois, facility. The only use
of mercury by Emerson Electric is by its White Rodgers Company that manufactures mercury bulb switches
at a plant in Afton, Missouri and mercury bulb switches, used for thermostats, at a plant in Puerto Rico.29
No information is available for the two companies shown below.
5-9
-------
TABLE 5-4. EMISSION SOURCE PARAMETERS FOR AN INTEGRATED
MERCURY BUTTON CELL MANUFACTURING FACILITY
Building/source No. description*1
Emission rateb
g/d
Ib/d
Exit temp., °K, and control
device
Main Plant
Control Room
1. Blending, slugging, compacting,
granulating
2. Slugging, granulating
3. Pelleting, consolidating
4. Pelleting, consolidating
4a. Pelleting, consolidating
5. Blending, compacting,
granulating, pelleting,
consolidating
6.12
1.22
1.63C
42.46
6.53
1.36C
0.0135
0.0027
0.0036C
0.0936
0.0144
0.003C
297; Baghouse
297; Baghouse
295; Baghouse
297; Baghouse
297; Baghouse
297; Baghouse
Anode room
6. Amalgam, dewatering
6a. Vacuum dryer
6b. Blending
7. Pelleting, zinc amalgam
1.82C
0.46C
0.91C
4.08C
0.004C
0.001C
0.002C
0.009C
297; Uncontrolled
297; Uncontrolled
297; Uncontrolled
295; Baghouse
Cell assembly area
8. Assembling cells
28.58
0.0630
295; Baghouse for particulate.
Vapor by recirculating air
through prefilters and charcoal
filters
Source: Reference 9.
^Source numbers are the same code used by facility.
Emission rates were measured by facility except where noted.
°Estimated emission rate by facility.
5-10
-------
Company Corporate Headquarters
Ranco, Inc. Plain City, OH
United Technologies Huntingdon, IN
In 1995, 84 Mg (92 tons) of mercury were used in the production of wiring devices and switches.2
5.3.1.1 Process Description.
5.3.1.1.1 Mercury buttons for wall switches. A process flow diagram for the manufacture of
mercury buttons for wall switches is shown in Figure 5-3. A metal ring, glass preform, ceramic center, and
center contact are assembled on a semiautomatic loader (Step 1) and fused together in a sealing furnace (Step
2). Each subassembly is then transferred to a rotating multistation welding machine, located in an isolation
room, where it is filled with about 3 g (0.11 ounces) of mercury (Step 3). The mercury used to fill the
subassembly is stored in an external container. During the subassembly filling step, the mercury container is
pressurized with helium; this pressurization transfers the mercury from the large storage container to a
smaller holding tank. Mercury is released in a controlled manner from the holding tank by using a rotating
slide gate that is synchronized to the welding machine speed. The filled subassembly is placed in the can,
evacuated, and welded shut to form the button (Step 4). The assembled buttons then leave the isolation room
and are cleaned (Step 5), zinc plated (Step 6), and assembled with other components (Step 7) to form the
completed wall switches.8
5.3.1.1.2 Thermostat switches. The production process for thermostat switches used for household
heating/air conditioning control and other applications is shown in Figure 5-4. First, metal electrodes
(contacts) are inserted into one end of a glass tube 0.89 to 1.5 cm (0.35 to 0.59 in.) in diameter (Step 1). This
end of the tube is then heated, crimped around the electrodes, and sealed. The apparatus is then cleaned,
transferred to the isolation fill room, and loaded onto the filling machine where the tubes are evacuated (Step
2). At the filling machine (Step 3), the vacuum in the glass tube is released and mercury is drawn into the
tube. The open end of the mercury-filled tube is then heated, constricted, and sealed (Step 4). Filling of
switch tubes produced in low volume is performed manually using the same sequence of steps. Excess glass
at the seal is discarded into a bucket of water (Step 5). The filled tube leaves the isolation room and falls into
a transport container (Step 6). Attachment of wire leads to the electrode contacts completes the switch
assembly (Step 7).
5.3.1.2 Emission Control Measures. Table 5-5 shows typical emission control methods used in the
mercury switch industry to reduce worker exposure to mercury vapor. The use of isolation rooms and
automated systems for fill operations in the manufacture of mercury buttons has considerably reduced the
manual handling of elemental mercury. For example, a refiner can supply mercury in 363 kg (800 Ib)
stainless steel storage containers that are individually mounted in steel frames to permit lifting and transport
by forklift. This eliminates the need to manually transfer the mercury from 35-kg (76-lb) iron flasks to the
holding tank.
The use of effective gaskets and seals allows containment of mercury in the process streams. Reject
and broken switches are discarded under water to suppress mercury vaporization.
Exhaust ventilation, which is custom designed to fit specific equipment, is often used to reduce
worker exposure to mercury vapor, mercury particulate, or both. For example, a specially designed circular
slot hood may be used to cover the filling and welding machine. Plastic strip curtains may be suspended from
the hood to help prevent airflow from the hood into the work room.
Temperature control is widely practiced as one of the most effective measures to reduce mercury
emissions. Reducing the fill room temperature to between 18° and 20 °C (64° and 68 °F) can be effective in
lowering mercury emissions. Some industry operations shut down and require personnel evacuation from the
room when temperatures rise above 21 °C (70°F).
Dilution ventilation of fill room air, without apparent control, has been practiced at mercury switch
plants. The negative pressure in the fill room prevents escape of mercury vapor into adjacent assembly areas.
5-11
-------
HELIUM
PREESURIZATIOM
OF MERCURY IN
CONTAINER
PRESSURIZED
HELIUM
1
ASSEMBLY OF
METAL RING, GLASS
PREFORM, CERAMIC
CENTER AND
CENTER CONTACT
ISOLATION ROOM
DENOTES POTENTIAL MERCURY EMISSION SOURCE
Figure 5-3. Manufacture of mercury buttons for wall switches8
-------
1
ELECTRODES
(CONTACTS)
INTO GLASS
TUBES
HB AH
2 3 •
^ TURF ^ MERCURY
^ EVACUATION ^ FILLING
(AUTOMATIC
OR MANUAL)
a
Hg A
4 •
WELDING
FILLED TUBES
1 '
5
EXCESS GLASS
DISCHARGE
INTO WATER
6
FINISHED
THERMOSTAT
SWITCH
ISOLATION FILL ROOM
DENOTES POTENTIAL MERCURY EMISSION SOURCE
Figure 5-4. Thermostat switch manufacture8
-------
TABLE 5-5. MEASURES TO REDUCE WORKPLACE EXPOSURE TO MERCURY
VAPOR EMISSIONS IN THE ELECTRIC SWITCH INDUSTRY
Control method
Process modification and
substitution
Containment
Ventilated enclosure
Local exhaust ventilation
Temperature control
Dilution ventilation
Isolation
Sources
Hg purification and
transfer
X
X
X
X
X
X
Hg filling
X
X
X
Product
testing
X
X
X
X
Spills, breakage,
rejects
X
X
X
X
Source: Reference 8.
Examples of technologies for removing mercury from exhaust streams were not found. However,
controls used at other manufacturers of electrical and electronic items may be effective at mercury switch
plants. These controls are discussed in subsequent subsections.8
In 1994, a major manufacturer of thermostats announced a pilot project to recycle mercury
thermostats. Homeowners and contractors can send unneeded thermostats back to the manufacturer so the
mercury can be removed and recycled. In addition, in 1995, the U.S. EPA announced a "Universal Waste
Rule" (which includes thermostats) that effectively allows for the transportation of small quantities of
mercury from specific products. This ruling should encourage recycling.33 In late 1996, the three major
thermostat manufacturrers, Honeywell, White Rodgers (a subsidiary of Emerson Electric), and General
Electric, agreed to form the Thermostat Recycling Corporation (TRC) to initiate a nationwide mercury switch
wholesaler take-back program utilizing the universal waste rule. The TRC plans to commence operations in
most of the Great Lake states and Florida in late 1997 or 1998. The TRC will request participation by all
contractors and wholesalers in the target states. Under the plan, HVAC dealers bring used thermostats to
participating wholesalers and place the mercury-containing switch in recycling containers. When the
container is full, the wholesaler ships the container to a consolidation facility where the mercury bulbs are
removed from the thermostat. The mercury bulbs will be shipped to a mercury recycling facility for mercury
reclamation.29
5.3.1.3 Emissions. During the manufacture of electric switches (wall and thermostat), mercury may
be emitted during welding or filling, as a result of spills or breakage, during product testing, and as a result of
material transfer. The mercury emission sources are indicated in Figures 5-3 and 5-4 by a solid circle.
Table 5-6 lists the three manufacturers of electric switches that reported mercury air emissions in the
1994 Toxic Release Inventory (TRI). Total reported emissions from these manufacturers was 6.4 kg (14
pounds).3
TABLE 5-6. MANUFACTURERS OF ELECTRIC SWITCHES AND ELECTRONIC
COMPONENTS REPORTING IN THE 1994 TOXIC RELEASE INVENTORY
Facility
Durakool, Inc.
Hermaseal Co.
Micro Switch
Honeywell Div.
Location
Elkhart, IN
Elkhart, IN
Freeport, IL
Comments
Hg used as an article component
Hg used as an article component
Hg used as an article component
Total annual air
emissions, Ib
5
5
4
Source: Reference 3.
5-14
-------
No mercury emission data have been published for other manufacturers of electrical switches. In the
production of either mercury buttons for wall switches or thermostat switches, the principal sources of
mercury emissions occur during filling processes that are conducted in isolated rooms. The isolation rooms
are vented to maintain the room at a slight negative pressure and prevent mercury contamination of adjacent
work areas. No emission data or results of tests are available to develop an estimate of mercury emissions
from the two processes. One 1973 EPA report, however, presents an emission factor for overall electrical
apparatus manufacture of 4 kg of mercury emitted for each megagram of mercury used (8 lb/ton).6 This
factor pertains only to emissions generated at the point of manufacture. This emission factor should be used
with caution, however, as it was based on engineering judgment and not on actual test data. In addition,
electrical switch production and the mercury control methods used in the industry have likely changed
considerably since 1973. The emission factor could, therefore, substantially overestimate mercury emissions
from this industry and should not be used to estimate current mercury emissions.
Total 1995 mercury emissions for this industry are estimated to be 0.4 Mg (0.5 tons); see
Appendix A for details.
5.3.2 Thermal Sensing Elements
In certain temperature-sensing instruments, a bulb and capillary temperature-sensing device is an
integral part of the instrument. These devices use the expansion force of mercury as it is heated to activate
the external controls and indicators of the instrument.
5.3.2.1 Process Description. A thermal sensing instrument consists of a temperature-sensing bulb, a
capillary tube, a mercury reservoir, and a spring-loaded piston. The bulb is made by cutting metal tubing to
the correct size, welding a plug to one end of the tube, and attaching a coupling piece to the other end. The
capillary tube is cut to a specified length and welded to the coupling at the open end of the bulb. The other
end of the capillary is welded to a "head" that houses the mechanical section of the sensor.
The bulb and capillary assembly are filled with mercury by a multistation mercury filling machine
that is housed in a ventilated enclosure. After filling, the sensor is transferred to a final assembly station
where a return spring and plunger are set into a temporary housing on the head of the sensor. To complete
the temperature instrument, the sensor is then attached to a controller and/or indicating device.8
5.3.2.2. Emission Control Measures. No information was found on specific emission control
devices or measures to control mercury emissions during the filling process. Although the filling machine is
typically in a ventilated enclosure, no information is available concerning any subsequent treatment of the
exhaust gas prior to discharge to the atmosphere.
5.3.2.2.1 Emissions No emission factors for mercury emissions from thermal sensing element
manufacturing were found in the literature, and no emission test data were available to calculate emission
factors.
5.3.3 Tungsten Bar Sintering
5.3.3.1 Process Description. Tungsten is used as a raw material in the manufacture of incandescent
lamp filaments. The manufacturing process starts with tungsten powder pressed into long, thin bars of a
specified weight. These bars are pretreated and then sintered using a high-amperage electrical current.
During the tungsten bar sintering process, mercury is used as a continuous electrical contact. The mercury
contact is contained in pools (mercury cups) located inside the sintering unit.
After the sintering process is completed, the bars are cooled to ambient temperature and the density
of the tungsten bars is determined. Metallic mercury is normally used in these measurements because of its
high specific gravity. To calculate the density of the tungsten bars, the bars are dipped into a pool of
mercury, and the weight of the displaced mercury is determined. When the bars are removed from the
mercury pool, the mercury is brushed off into a tray of water that is placed in front of the pool.8
5.3.3.2 Emission Control Measures. No specific information on emission control measures for
sintering tungsten bars was found in the literature.
5.3.3.3 Emissions. Mercury is used only during the actual sintering and the final density
measurements. For this reason, it is assumed that these two operations account for all the mercury emitted
5-15
-------
from the process. No specific data for mercury emissions from the tungsten sintering process were found in
the literature, and no emission test data were available to calculate mercury emission factors.
5.3.4 Copper Foil Production
High purity copper foil, used as a laminate in printed circuit boards, is produced by an
electrodeposition process using mercury as the electrical contact.
5.3.4.1 Process Description. The initial step in the foil production process is the dissolution of scrap
copper in sulfuric acid to form copper sulfate. The solution is then fed to the plating operation where the
copper ions are electrodeposited on rotating drums as copper metal. Each plating drum is composed of a
concrete cell containing the copper sulfate solution, an anode (lead), a rotating titanium drum (cathode), and a
winding roll. During the electrodeposition process, a current passes between the lead anode and the rotating
drum cathode. As the drum rotates, the copper metal is electrodeposited on the drum surface in the form of a
continuous thin foil sheet.
The plated foil is peeled from the drum and wound on a roll. When the roll reaches a specified size,
it is removed from the plating drum unit and transferred to the treating room where it is specially treated,
annealed, slit, wrapped, and prepared for shipping.8
Elemental mercury is used as the continuous electrical contact between the rotating shaft of the drum
and the electrical connections. The liquid mercury is contained in a well located at one end of the rotating
drum shaft.8
5.3.4.2 Emission Control Measures. Manufacturing processes that require mercury as an electrical
contact generally use ventilated enclosures for controlling vapor emissions from mercury pools. In copper
foil production, the mercury wells are located in ventilated enclosures, and exhaust gases are directed to a
mercury vapor filter. Another method of controlling emissions from mercury wells is to reduce the
temperature of mercury in the well. Generally, mercury wells operate at 82°C (180°F); at this temperature,
mercury has a vapor pressure of 0.10 mmHg. A temperature reduction to 21 ° C (70 ° F) decreases the mercury
vapor pressure to 0.0013 mmHg.
5.3.4.3 Emissions. Mercury can be emitted from the drum room and treating room of the copper
plating process. No information was available on mercury release rates to the atmosphere through ventilation
systems. No specific data for mercury emissions from the production of copper foil were found in the
literature, and no emission test data were available for calculating emission factors.
5.3.5 Fluorescent Lamp Manufacture and Recycling
All fluorescent lamps contain elemental mercury as mercury vapor inside the glass tube. Mercury
has a unique combination of properties that make it the most efficient material for use in fluorescent lamps.
Of the 680 million mercury-containing lamps sold in the U.S. annually, approximately 96 percent are
fluorescent lamps.34 The names and division headquarters of the four fluorescent lamp manufacturers in the
U.S. in 1995 are shown in Table 5-7.
TABLE 5-7. U.S. FLUORESCENT LAMP MANUFACTURERS' HEADQUARTERS
Company
DURO-LITE Corp.
General Electric
OSRAM Sylvania, Inc.
Philips Lighting Company
Division Headquarters
North Bergen, NJ
Cleveland, OH
Danvers, MA
Somerset, NJ
Source: References 29 and 33..
In 1995, 30 Mg (33 tons) of mercury were purchased for the manufacture of electric lighting,
including fluorescent, mercury vapor, metal halide, and high-pressure sodium lamps.2 Lamps do not contain
all of the mercury purchased for the manufacture; mercury not retained in the lamps is returned to mercury
recyclers for purification and reuse. In 1994, 15.7 Mg (17.3 tons) of the 27 Mg (30 tons) of mercury were
actually contained in the lamps.34
5-16
-------
There are presently few mercury recycling facilities in the country. Data from a 1994 EPA report
indicate that approximately 600 million fluorescent lamps are disposed each year, with only 2 percent of that
number being recycled.35 That translates into approximately 12 million fluorescent lamps recycled annually.
The number of fluorescent lamps recycled has been increasing so the 2 percent figure in the 1994 report may
underestimate the current recycling efforts.
5.3.5.1 Fluorescent Lamp Manufacture.
5.3.5.1.1 Process description. Fluorescent lamp production begins with the preparation of the lamp
tube. Precut glass tubes are washed to remove impurities, dried with hot air, and coated with a liquid
phosphor emulsion that deposits a film on the inside of the lamp tube. Mount assemblies, consisting of a
short length of glass exhaust tube, lead wires, and a cathode wire, are fused to each end of the glass lamp
tube. The glass lamp tube, with attached mount assemblies, is then transferred to the exhaust machine.
On the exhaust machine, the entire glass tube system is exhausted and a small amount (15 to 100 mg
[3.3 x 10"5 to 2.2 x 10"4 lb]) of mercury is added. A few high wattage HID lamps may contain up to 250 mg
of mercury. Over the life of the lamp, some of the mercury combines with the glass, internal metals, and the
emulsion coating on the interior of the lamp tube. Following the addition of mercury, a vacuum is drawn
through the glass lamp tube system to remove the air and small quantities of excess mercury. The glass tube
system is then filled with inert gas and sealed. After the lamp tubes are sealed, metal bases are attached to the
ends of the lamp tube and are cemented in place by heating.8
5.3.5.1.2 Emission control measures. No add-on emission control measures were identified for
exhaust or ventilation gases. The only methods identified were those used to reduce worker exposure.
Mercury air concentrations due to handling are usually reduced by containment, local exhaust ventilation,
temperature control, isolation, and/or mercury removal from the air stream. Mercury air levels during the
lamp production steps are reduced by process modifications, containment, ventilated enclosures, local exhaust
ventilation, and temperature control.
The use of mercury-containing fluorescent and other high-efficiency lighting systems is increasing
because of the energy efficiency of these systems. However, the mercury content of fluorescent lamps has
decreased by 53 percent between 1989 and 1995 to an average of 22.8 mg of mercury per lamp. Continued
product design changes that further reduce mercury use by the industry could also further reduce mercury
emissions from the industry.
5.3.5.1.3 Emissions. Mercury emissions from fluorescent lamp manufacturing may occur during
mercury handling operations and during lamp production. Handling operations that may result in mercury
vapor emissions include mercury purification, mercury transfer, and parts repair. During lamp production,
mercury may be emitted from the mercury injection operation and from broken lamps, spills, and waste
material.
One 1973 EPA report presents an emission factor for overall electrical apparatus manufacture of
4 kg of mercury emitted for each megagram of mercury used (8 lb/ton).6 This factor pertains only to
emissions generated at the point of manufacture. This emission factor should be used with extreme caution,
however, as it was based on engineering judgment and not on actual test data. In addition, electric light
production and the mercury control methods used in the industry have likely changed considerably since
1973. The emission factor may, therefore, substantially overestimate mercury emissions from this industry.
A 1984 emission rate of 10.2 g/d (0.02 Ib/d) was found in the National Air Toxics Information
Clearinghouse (NATICH) for a GTE lamp manufacturing facility in Kentucky.36 However, no information
was available on the quantity of mercury used at the facility, the number of units produced, or other data that
would permit a comparison of this emission rate with other facilities. In addition, no data were presented to
allow calculation of an annual quantity.
Only one lamp manufacturing facility (General Electric Company Bucyrus Lamp Plant) reported
mercury emissions in the 1994 TRI; their annual emissions were 0.21 Mg/yr (0.23 tons/yr).3
5-17
-------
5.3.5.2 Fluorescent Lamp Recycling.
5.3.5.2.1 Process description. The crushing of fluorescent lamps to separate the glass from the
phosphor powder in the lamp is commonly the first step in recycling of mercury; although some companies
use other methods, such as removal of the phosphor powder by air vortex or by flushing with hydrochloric
acid.35 The simplest crushers are essentially single units with a crusher mounted on top of a barrel, usually a
55-gallon drum. This system is used in many industrial facilities to crush their fluorescent lamps as a means
to reduce the solid waste volume before disposing the material in a landfill. In this version, lamps are hand-
fed to a feeder chute of variable length and diameter. The lamps pass to the crushing unit, typically
consisting of motor-driven blades, which implode and crush the lamps. From here, the crushed powder drops
into the barrel below the crusher. Some systems include a vacuum system which collects air from beneath the
crusher, preventing mercury laden air from exiting through the feed chute. Material collected in the vacuum
system first passes through a cyclone separator. This removes glass particles, which drop into the drum. Air
from the cyclone separator contains phosphor powder and some mercury vapor. These are removed by
further control.
After crushing of the lamps, mercury recovery is often the next step in the recycling process. Most
commonly, lamps that are not landfilled undergo retorting or roasting which recovers mercury by distillation.
Different versions exist, but in each, the material is heated to vaporize the mercury and recover it as a liquid.
This can be accomplished in closed vessels (retorts) or in open-hearth furnaces, ovens, or rotary kilns
(roasting). Recovery of the vaporized mercury can be done with condensers and separators or with a venturi
scrubber and decanter, followed by an air pollution control system.
Retorting generally gives higher recovery rates than does roasting, and is also well-suited to wastes
containing volatile forms of mercury. Thus retorting is generally the recovery method of choice for
fluorescent lamps. Typically, the mercury-containing wastes are placed in a retort, and heated for 4 to
20 hours to a temperature above the boiling point of mercury (357°C [675 °F]) but below 550°C (1022°F).
Vaporized material from this process is condensed in the scrubber or condenser, and then recovered in a
collector or decanter. This recovered mercury may require additional treatment, such as nitric acid bubbling,
to remove impurities.
5.3.5.2.2 Emission control measures. The simplest fluorescent lamp crushers have no air pollution
control devices. More sophisticated versions of the barrel-mounted crusher utilize a negative air exhaust
system to draw the crushed debris and prevent it from reemerging through the feeder tube. The drawn air is
then passed through a high efficiency particulate air (HEPA) filter to remove particulate matter from the
exhausted flow. Other control techniques include gasketing around the connection between the crusher and
drum, total enclosure, and disposable collection barrels.
One crushing system utilizes a vacuum system which collects air and tube materials from beneath the
crusher, which then passes through a cyclone separator to remove glass particles. From the cyclone, the air
passes through a baghouse, several particulate matter filters and HEPA filters to ensure that all lamp particles
have been removed. The exhaust then passes through activated carbon beds, which trap the mercury vapor.
The air is then passed through more particulate filters which trap any carbon that may have been carried away
from the activated carbon bed. The air from the containment room (in which the crusher and filters are
located) is blended with the cleaned crusher exhaust air and sent through another series of particulate filters
and more activated carbon.35 No efficiencies of this control system are available.
Another crusher uses a system similar to the one mentioned above. The entire system operates under
negative pressure and the crushed debris is collected in a cyclone. The exhaust continues through a reverse jet
baghouse, a HEPA filter, and then through a potassium iodide-impregnated carbon filter. This removes the
mercury by precipitating it in the form of mercuric iodide (no removal efficiencies were provided). The air in
the building that houses the crusher is also under negative pressure and is drawn through the entire filter
system as well.35
No information was found describing control devices for mercury recovery systems beyond the
condensers, separators, and venturi scrubbers designed for product recovery.
5.3.5.1.1 Emissions. Mercury emissions from fluorescent lamp recycling may occur from crusher
feed chutes, connections between crushers and receiving barrels, collection barrels themselves, control system
outlets for crushers or retorts, and scrubber system wastewater.
5-18
-------
In many cases, actual emission estimates have not been determined for lamp recycling processes;
rather, occupational exposure estimates have been derived from ambient air measurements taken in the
workplace. Approximations of mercury emissions are available for two fluorescent lamp crushers based on
reported production rates, air flow rates, and typical exhaust characteristics for a carbon adsorber controlling
mercury vapor emissions.35 The emission rates for these two crushers range from 0.14 to 10 mg/min (3.1 x
10'7 to 2.2 x 10'5 Ib/min) and 0.002 to 0.16 mg/lamp (4.4 x 10'9 to 3.5 x 10'7 Ib/lamp). The average
emission factor for the two crushers is 0.071 mg/lamp (1.6 x 10"7 Ib/lamp). This emission factor should be
used with caution, however, as it was based on engineering judgment and not on actual test data.
Mercury emission test data from a 1994 test are available for one fluorescent bulb crusher. The unit
is an enclosed system vented to a HEPA fabric filter and a carbon adsorber. The average mercury emission
rate for the three test runs was 0.003 g/hr (0.000007 Ib/hr). Using the reported tube processing rate of
3,414 bulbs/hr, a mercury emission factor of 0.00088 mg/lamp (1.9 x 10"9 Ib/lamp) can be estimated which is
about two orders of magnitude lower than the average emission factor estimated in the previous paragraph.37
No mercury emission data were available from which to calculate emission factors for recovery
processes.
5.4 INSTRUMENT MANUFACTURING AND USE (THERMOMETERS)
Mercury is used in many medical and industrial instruments for measurement and control functions
including thermometers; manometers, barometers, and other pressure-sensing devices; gauges; valves; seals;
and navigational devices. Because mercury has a uniform volume expansion over its entire liquid range and a
high surface tension, it is extremely useful in the manufacture of a wide range of instruments. It is beyond the
scope of this report to discuss all instruments that use mercury in some measuring or controlling function.
Although there is potential for mercury emissions from all instruments containing mercury, this section
focuses only on the production of thermometers because they represent the most significant use, and more
information is available on thermometer manufacture than on the manufacture of other instruments.
There are generally two types of clinical thermometers: 95 percent are oral/rectal/baby
thermometers, and 5 percent are basal (ambient air) temperature thermometers. An oral/rectal/baby
thermometer contains approximately 0.61 g (0.022 oz.) of mercury and a basal thermometer contains
approximately 2.25 g (0.079 oz.) of mercury.38
In 1995, 43 Mg (47 tons) of mercury were used in all measuring and control instrument
manufacture.2
5.4.1 Process Description
The manufacture of temperature measurement instruments varies according to the type of bulb or
probe. In addition, the mercury filling procedure varies among different instrument manufacturers. The
production of glass thermometers begins with the cutting of glass tubes (with the appropriate bore size) into
required lengths. Next, either a glass or metal bulb, used to contain the mercury, is attached to one end of the
tube.
The tubes are filled with mercury in an isolated room. A typical mercury filling process is conducted
inside a bell jar. Each batch of tubes is set with open ends down into a pan and the pan set under the bell jar,
which is lowered and sealed. The tubes are heated to approximately 200°C (390°F), and a vacuum is drawn
inside the bell jar. Mercury is allowed to flow into the pan from either an enclosed mercury addition system
or a manually filled reservoir. When the vacuum in the jar is released, the resultant air pressure forces the
mercury into the bulbs and capillaries. After filling, the pan of tubes is manually removed from the bell jar.
Excess mercury in the bottom of the pan is purified and transferred back to the mercury addition system or
filling reservoir.
Excess mercury in the tube stems is forced out the open ends by heating the bulb ends of the tubes in
a hot water or oil bath. The mercury column is shortened to a specific height by flame-heating the open ends
(burning-off process). The tubes are cut to a finished length just above the mercury column, and the ends of
the tubes are sealed. All of these operations are performed manually at various work stations. A temperature
scale is etched onto the tube, completing the assembly.8'9
5-19
-------
5.4.2 Emission Control Measures
Vapor emissions from mercury purification and transfer are typically controlled by containment
procedures, local exhaust ventilation, temperature reduction to reduce the vapor pressure, dilution ventilation,
or isolation of the operation from other work areas. The tube bore size also can be modified to reduce the use
of mercury.
The major source of mercury emissions in the production of thermometers may be in the mercury
filling step. Several emission control measures have been identified for production processes that require, in
part, filling an apparatus with metallic mercury. In the previous discussion of the electric switch industry,
Table 5-5 presented several control methods that are used by that industry to reduce workplace exposure to
mercury vapor emissions. These controls or combinations of controls are generally applicable to the
production of thermometers.
One of the latter steps in the production of thermometers involves heating the mercury in a high
temperature bath and the subsequent heating of the open ends with a flame (burning-off process). A possible
control scenario for these operations would include an isolation room with local exhaust ventilation and
dilution ventilation, to create a slight negative pressure in the room. This arrangement would prevent escape
of mercury vapor into adjacent assembly or work areas.
Additionally, product substitutions in the marketplace may reduce mercury emissions from
instrument manufacturing and use. One notable example of such a substitution is the replacement of mercury
thermometers with digital devices.
5.4.3 Emissions
Mercury emissions can occur from several sources during the production of thermometers. Many of
the procedures used in thermometer production are performed manually, and as a result, emissions from these
procedures are more difficult to control. The most significant potential sources of emissions are mercury
purification and transfer, mercury filling, and the heating out (burning-off) process. Vapor emissions due to
mercury spills, broken thermometers, and other accidents may contribute to the level of mercury emissions.
No specific data for mercury emissions from manufacturing thermometers or any other instrument
containing mercury were found in the literature, and no emission test data were available from which to
calculate emission factors. One 1973 EPA report, however, presents an emission factor for overall
instrument manufacture of 9 kg of mercury emitted for each megagram of mercury used (18 lb/ton).6 This
emission factor should be used with extreme caution, however, as it was based on survey responses gathered
in the 1960's and not on actual test data. In addition, instrument production and the mercury control methods
used in instrument production have likely changed considerably since the time of the surveys.
Total 1995 mercury emissions for this industry are estimated to be 0.4 Mg (0.5 tons); see
Appendix A for details.
5-20
-------
6.0 EMISSIONS FROM COMBUSTION SOURCES
Mercury is often found as a trace contaminant in fossil fuels or waste materials. When these
materials are combusted, the combination of the elevated temperature of the process and the volatility of
mercury and mercury compounds results in mercury being emitted in the combustion gas exhaust stream.
This section addresses mercury and mercury compound emissions from seven stationary source combustion
processes:
- Coal combustion
- Oil combustion
- Wood combustion
- Municipal waste combustion
- Sewage sludge incineration
- Hazardous waste combustion
- Medical waste incineration
These seven processes fall into two general categories. The first three involve fuel combustion for energy,
steam, and heat generation, while the last four are primarily waste disposal processes, although some energy
may be recovered from these processes. A summary of the estimated emissions from each of the above
categories is as follows:
Category Emissions, Mg (tons)
Coal combustion 67.8 (74.6)
Oil combustion 7.4(8.1)
Wood combustion 0.1 (0.1)
Municipal waste combustion 26 (29)
Sewage sludge incineration 0.86 (0.94)
Hazardous waste combustion 6.3 (6.9)
Medical waste incineration 14.5 (16)
The paragraphs below provide a general introduction to the two combustion categories. As part of
this introduction, a summary of nationwide fuel usage is presented in detail. This information was used to
develop nationwide emissions of mercury for different sectors and fuels.
In 1994, the total annual nationwide energy consumption in the United States was
93.584 x 1012 megajoules (MJ) (88.789 x 1015 British thermal units [Btu]).39 Of this total, about
54.889 x 1012 MJ (52.077 x 10 Btu) or 59 percent involved consumption of coal, petroleum products, and
natural gas in nontransportation combustion processes. (No data were available on energy consumption for
wood combustion from the U.S. Department of Energy.) Table 6-1 summarizes the 1994 U.S. distribution of
fossil fuel combustion as a function of fuel type in the utility, industrial, commercial, and residential sectors.
The paragraphs below provide brief summaries of fuel use patterns; additional details on fuel consumption by
sector for each State can be found in "State Energy Data Report, Consumption Estimates, 1994"39.
As shown in Table 6-1, at 22.129 x 1012 MJ (20.995 x 1015 Btu) per year, the industrial sector is the
largest consumer of fossil fuels. This sector uses a mixture of natural gas (46 percent), fuel oil (7 percent),
other petroleum fuels (35 percent), and coal (12 percent). The other petroleum fuels that are
6-1
-------
TABLE 6-1. 1994 DISTRIBUTION OF FOSSIL FUEL CONSUMPTION IN THE UNITED STATES
Fuel
Bituminous/lignite coal
Anthracite coal
Distillate oil
Residual oil
Other petroleum fuels
Natural gas
Total
Annual energy consumption, 1012 MJ (1015 Btu)
Utilities
17.760
(16.850)
0.018
(0.017)
0.100
(0.095)
0.893
(0.847)
0.027
(0.026)
3.222
(3.057)
22.020
(20.892)
Industrial
2.642
(2.507)
(-)
1.169
(1.109)
0.448
(0.425)
7.710
(7.315)
10.160
(9.639)
22.129
(20.995)
Commercial
0.076
(0.072)
0.012
(0.011)
0.489
(0.464)
0.184
(0.175)
0.121
(0.115)
3.139
(2.978)
4.021
(3.815)
Residential
0.041
(0.039)
0.017
(0.016)
0.928
(0.880)
(-)
0.485
(0.460)
5.249
(4.980)
6.719
(6.375)
Total
20.519
(19.468)
0.047
(0.044)
2.686
(2.548)
1.525
(1.447)
8.343
(7.916)
21.770
(20.654)
54.889
(52.077)
Source: Reference 39.
used include primarily liquified petroleum gas, asphalt and road oil, and other nonclassified fuels. Again, the
distribution among the three fuel types varies substantially from State to State, with each of the three
contributing significant fractions in most States. Notable exceptions are Hawaii, which relies almost
exclusively on petroleum fuels; Alaska, which relies primarily on natural gas; and the northeastern States of
Connecticut, New Hampshire, Rhode Island, and Vermont, which use almost no coal.
The utility sector is the second largest fossil fuel energy consumer at the rate of 22.020 x 1012 MJ
(20.892 x 1015 Btu) per year. About 81 percent of this energy was generated from coal combustion, with
bituminous and lignite coal contributing substantially greater quantities than anthracite coal. In fact,
Pennsylvania is the only State in which anthracite coal is used for electric power generation. Although most
States rely primarily on coal for power generation, the distribution among fossil fuels varies from State to
State, and several States rely heavily on natural gas and fuel oil for power generation. In California, natural
gas provides about 97 percent of the fossil-fuel based electricity production, and no coal is used. In Hawaii,
fuel oil is used exclusively, while in Oklahoma and Texas, a mixture of coal and natural gas are used. In
Florida, Louisiana, Massachusetts, and New York, coal, fuel oil, and natural gas each represent a substantial
fraction of the power generation. The States of Idaho, Maine, Rhode Island, and Vermont have no coal fired
utilities. Idaho relies exclusively on hydroelectric power, while the New England States use a mixture of fuel
oil, natural gas, nuclear, and hydroelectric power.
As shown in Table 6-1, substantially smaller quantities of fossil fuel are used in the commercial and
residential sectors than are used in the utility and industrial sectors. The fuels used are primarily natural gas,
fuel oil, and liquified petroleum gas (the "other petroleum fuels" in the residential category). Almost all
States use a mixture of the fuels, but the distributions vary substantially, with some States like California and
Louisiana using primarily natural gas and others like New Hampshire and Vermont using a much greater
fraction of fuel oil. One unique case is Pennsylvania where anthracite coal is used in both the residential and
commercial sectors.
In the individual sections below, additional information will be presented on the mercury content of
the different fuels and on the relationship between fuel type and emissions. However, for any geographic
area, the contribution of energy generation sources to mercury emissions will be a function of the distribution
of fuels used in the different sectors within the area.
The sources within the second combustion category are engaged primarily in waste disposal.
Mercury emissions from these processes are related to the mercury levels in the waste. The different waste
types are generally characterized with distinct source categories.
Furthermore, these waste disposal practices are not strongly related. Consequently, each of these
categories will be characterized individually within the sections below rather than in a general discussion here.
6-2
-------
The seven sections below have a consistent organization. First, the characteristics of the fuel or waste are
described and, in the case of the waste combustion processes, the general source category is also described.
Second, process descriptions are presented and emission points are identified. Third, available emission
control measures are identified and described. Finally, emission factors are presented. A discussion of the
sampling and analytical methods used to determine the mercury emission levels from combustion sources is
presented in Section 9.
6.1 COAL COMBUSTION
As presented in Table 6-1, most coal combustion in the United States occurs in the utility and
industrial sectors, with about 87 percent being bituminous and lignite combustion within the utility sector and
about 13 percent being bituminous and lignite combustion in the industrial sector. Consequently, the focus of
the discussion below will be on bituminous and lignite coal combustion in utility and industrial boilers.
However, limited information on anthracite coal combustion will also be presented.
6.1.1 Coal Characteristics
The coal characteristics of greatest interest in evaluating mercury emissions from coal combustion
are coal heating values and coal mercury content. Mercury emissions are a direct function of the mercury
content, while heating values are used to convert emission factors between mass input-based and heat
input-based activity levels. This section briefly summarizes the information about coal heating levels and
mercury content.4'41'42 More complete summaries can be found in Reference 40 and detailed analyses of
coal mercury content as a function of coal type and geographic region can be found in Reference 41 and
Reference 42.
Coal is a complex combination of organic matter and inorganic ash formed in geologic formations
from successive layers of fallen vegetation and other organic matter. Coal types are broadly classified as
anthracite, bituminous, subbituminous, or lignite, and classification is made by heating values and amounts of
fixed carbon, volatile matter, ash, sulfur, and moisture.43 Formulas for differentiating coals based on these
properties are available.44 These four coal types are further subdivided into 13 component groups. Table 6-2
summarizes information about the heating values for these component groups.
The heating value of coal varies among coal regions, among mines within a region, among seams
within a mine, and within a seam. The variability is minimal compared to that found with trace metal levels
described below, but it may be important when fuel heat content is used as the activity level measure for
source emission calculations. Data presented in Table 6-3 illustrate the regional variability of coal heat
content. Heat content among coals from several different mines within a region appears to exhibit greater
variability than either variability within a mine or within a seam. For the sample points presented in
Table 6-3, intermine variability averaged 15 percent, intramine variability 7 percent, and intraseam variability
3 percent. Because few combustion sources burn coal from just one seam or one mine, coal heat content
variability may significantly affect emission estimates that are being calculated using emission factors, coal
use data, and coal heat content data, even if the source gets all its coal from the same area of the country.40
To an even greater extent than the heating value, the mercury content of coal varies substantially
among coal types, at different locations in the same mine, and across geographic regions. The most
comprehensive source of information on coal composition is the United States Geological Survey (USGS)
National Coal Resources Data System (NCRDS). Geochemical and trace element data are stored within the
USCHEM file of NCRDS. As of October 1982, the file contained information on 7,533 coal samples
representing all U.S. coal provinces. Trace element analysis for about 4,400 coal samples were included in
the data base. This computerized data system was not accessed during the current study due to time and
budgetary constraints and information from USGS that indicated that few data had been added to the system
since 1972; however, a summary of the data presented in Reference 40 was reviewed. The most extensive
source of published trace element data was produced in Reference 42. This report contains data for 799 coal
samples taken from 150 producing mines and includes the most important U.S. coal seams. Data from
Reference 42 was the initial input into the USCHEM file of NCRDS. The information presented here
summarizes the review presented in Reference 40 of the results published in References 41 and 42. Note that
those results are consistent with unpublished analyses conducted by USGS on the data
6-3
-------
TABLE 6-2. COAL HEATING VALUES
Coal class
Anthracite
Bituminous
Subbituminous
Lignite
Component
group
Al
A2
A3
Bl
B2
B3
B4
B5
SI
S2
S3
LI
L2
Definition
Meta- anthracite
Anthracite
Semianthracite
Low volatile
bituminous
Medium volatile
bituminous
High volatile
A bituminous
High volatile
B bituminous
High volatile
C bituminous
Subbituminous A
Subbituminous B
Subbituminous C
Lignite A
Lignite B
Sourcea
PA,RI
CO,PA,NM
AR,PA,VA
AR,MD,OK,PA,
WV
AL,PA,VA
AL,CO,KS,KY,
MO,NM,PA,
TN,TX,UT,VA,
WV
IL,KY,MO,OH,
UT,WY
IL,IN,IA,MI
MT,WA
WY
CO,WY
ND,TX
NA
Heating value, kJ/kg (Btu/lb)
Rangea
21,580-29,530
(9,310-12,740)
27,700-31,800
(11,950-13,720)
27,460-31,750
(11,850-13,700)
30,640-34,140
(13,220-14,730)
31,360-33,170
(13,530-14,310)
28,340-35,710
(12,230-14,510)
26,190-30-480
(11,300-13,150)
24,450-27,490
(10,550-11,860)
23,940-25,820
(10,330-11,140)
21,650-22,270
(9,340-9,610)
19,280-19,890
(8,320-8,580)
16,130-17,030
(6,960-7,350)
NA
Meana
25,560
(11,030)
30,270
(13,000)
29,800
(12,860)
32,400
(13,980)
32,170
(13,880)
31,170
(13,450)
28,480
(12,290)
26,030
(11,230)
24,890
(10,740)
21,970
(9,480)
19,580
(8,450)
16,660
(7,190)
NA
Source: Reference 40.
aNA = not available.
6-4
-------
TABLE 6-3. EXAMPLES OF COAL HEAT CONTENT VARIABILITY
Variability
[ntermine variability
Intramine variability
[ntraseam variability
Coal source
Eastern U.S.
Central U.S.
Western U.S.
Eastern U.S.
Central U.S.
Western U.S.
Eastern U.S.
Central U.S.
Western U.S.
Coal heat content, Btu/lb
Mean
12,320
10,772
11,227
12,950
10,008
12,000
12,480
10,975
10,351
12,230
10,709
11,540
Rangea
10,750- 13,891
9,147- 12,397
9,317- 13,134
NA
9,182- 10,834
11,335- 12,665
NA
9,667 - 12,284
9,791- 10,911
NA
10,304- 11,113
NA
Percent variation
about the mean
12.7
15
17
4.8b
8.0
5.5
5.7C
12.0
5.4
3.0d
3.7
2.5e
Source: Reference 40.
aNA = not available.
bBased on a standard deviation of 624.
cBased on a standard deviation of 708.
dBased on a standard deviation of 371.
eBased on a standard deviation of 291.
6-5
-------
contained in NCRDS as of 1989.45 More information on the sampling and analysis of mercury in coal is
presented in Section 9.
Table 6-4 presents information on the mean concentration of mercury in coal and on the distributions
of mercury concentrations by coal type. Bituminous and anthracite coals have the highest mean mercury
concentrations, 0.21 parts per million by weight (ppmwt) and 0.23 ppmwt, respectively. The standard
deviation of each mean either approaches or exceeds the mean, indicating strong variation in the data.
According to Reference 40 subbituminous coals have the greatest reported range of mercury concentrations
(0.01 to 8.0 ppm). Based on conversations with USGS personnel, the means reported in Table 6-4 are
regarded as typical values for in-ground mercury concentration in coals in the United States.45
TABLE 6-4. MERCURY CONCENTRATION IN COAL BY COAL TYPE
Coal type
Bituminous
Subbituminous
Anthracite
Lignite
No. of samples
3,527
640
52
183
Mercury concentration, ppmwt
Range
<0.01to3.3
0.01 to 8.0
0.16 to 0.30
0.03 to 1.0
Arithmetic mean
0.21
0.10
0.23
0.15
Standard
deviation
0.42
0.11
0.27
0.14
Source: Reference 40.
Other estimates of mercury concentration in coal have been developed. The U. S. EPA, in the Study
of Hazardous Air Pollutant Emissions from Electric Utility Steam Generating Units, used a USGS data base
containing analyses of 3,331 core and channel samples of coal from the top 50 (1990 and later) economically
feasible coal streams in the U.S.33'46 Industry reviewed the USGS data set and, under a separate effort,
screened the data to remove about 600 entries representing coal seams that could not be mined
economically.47 Because the average mercury concentration of the screened data set was virtually the same as
the mercury concentration when the full USGS data set was used, EPA elected to use the USGS data set in its
entirety.33 Other data sets showing concentrations about 50 percent lower than the USGS data set average
are based on significantly lower numbers of samples.4
47
The concentration of mercury in coal also varies by geographic region from which the coal is mined.
Based on the "best typical" values for each region, which are footnoted in Table 6-5, coals from the
Appalachian and Gulf Provinces have the highest mean mercury concentration, 0.24 ppmwt for both regions.
Also, based on the best available data, the lowest mean concentration is found in coals from the Alaska region
(0.08 ppmwt). However, note that another study showed substantially higher levels (4.4 ppmwt). That study
also showed that the greatest range of concentration is found in coals from the Alaska region with a reported
range of 0.02 to 63 ppmwt.40 The means reported in Table 6-5 may be regarded as typical in-ground
concentrations of mercury in coals from each geographic region.
6.1.2 Process Description
As shown in Table 6-1, almost all coal combustion occurs in utility and industrial boilers. Almost all
of the coal burned is bituminous and subbituminous (95 percent) and lignite (4 percent).40 However, the
processes used for the different coals are comparable. The paragraphs below first describe the boilers used
for bituminous coal combustion. Then, lignite and anthracite combustion are described briefly.
References 48 and 43 offer additional details on these processes.
The two major coal combustion techniques used to fire bituminous and subbituminous coals are
suspension firing and grate firing. Suspension firing is the primary combustion mechanism in pulverized coal
and cyclone systems. Grate firing is the primary mechanism in underfeed and overfeed stokers. Both
mechanisms are employed in spreader stokers.
Pulverized coal furnaces are used primarily in utility and large industrial boilers. In these systems,
the coal is pulverized in a mill to the consistency of talcum power (i.e., at least 70 percent of the particles
6-6
-------
TABLE 6-5. MERCURY CONCENTRATION IN COAL BY REGION
Region
Appalachian
Interior
Illinois Basin0
Gulf Province
Northern Plains
Rocky Mountains
Alaska
No. of
samples
2,749
331
592
155
82
38
34
371
490
184
124
107
18
Mercury concentration, ppmwt
Range
<0.01-3.3
0.01-0.83
0.01-1.5
0.03-1.6
0.16-1.91
0.03-1.0
0.01-3.8
0.01-1.48
0.01-8.0
0.02-63
Arithmetic mean
0.24*
0.24b
0.14*
0.14b
0.15
0.21
0.24*
0.1 8b
O.lla
0.11
0.09*
0.06b
0.11
0.08a
4.413
Standard deviation
0.47
0.14
0.22
0.19
0.10
0.12
0.07
Source: Reference 40.
a Values from Reference 41 are based on the most comprehensive data set currently available (the NCRDS)
and may be used as typical values for mercury in coal from these regions.
Values from Reference 42 are included in the NCRDS. Arithmetic means from the entire NCRDS are more
representative than means from this study, since the NCRDS contains many more coal samples. The
Reference 42 data are included to give an idea of the range of values for mercury content in individual coal
samples from each region.
0 Eastern section of Interior Province.
will pass through a 200-mesh sieve). The pulverized coal is generally entrained in primary air and
suspension-fired through the burners to the combustion chamber. Pulverized coal furnaces are classified as
either dry or wet bottom, depending on the ash removal technique. Dry bottom furnaces fire coals with high
ash fusion temperatures, and dry ash removal techniques are used. In wet bottom (slag tap) furnaces, coals
with low ash fusion temperatures are used, and molten ash is drained from the bottom of the furnace.
Cyclone furnaces burn low ash fusion temperature coal crushed to a 4-mesh size. The coal is fed
tangentially, with primary air, to a horizontal cylindrical combustion chamber. Small coal particles are
burned in suspension, while the larger particles are forced against the outer wall. Because of the high
temperatures developed in the relatively small furnace volume, and because of the low fusion temperature of
the coal ash, much of the ash forms a liquid slag that is drained from the bottom of the furnace through a slag
tap opening. Cyclone furnaces are used mostly in utility and large industrial applications.
In spreader stokers, a flipping mechanism throws the coal into the furnace and onto a moving grate.
Combustion occurs partially in suspension and partially on the grate. Because the entrained particles in the
furnace exhaust have substantial carbon, fly ash reinjection from mechanical collectors is commonly used to
improve boiler efficiency. Ash residue in the fuel bed is deposited in a receiving pit at the end of the grate.
In overfeed stokers, coal is fed onto a traveling or vibrating grate and burns on the fuel bed as it
progresses through the furnace. Ash particles fall into an ash pit at the rear of the stoker. "Overfeed" applies
because the coal is fed onto the moving grate under an adjustable gate. Conversely, in "underfeed" stokers,
coal is fed upward into the firing zone by mechanical rams of screw conveyers. The coal moves in a channel,
known as a retort, from which it is forced upward, spilling over the top of each side to feed the fuel bed.
Combustion is completed by the time the bed reaches the side dump grates from which the ash is discharged
to shallow pits.
The next most common coal used in the U.S. is lignite. Lignite is a relatively young coal with
properties intermediate to those of bituminous coal and peat. Because lignite has a high moisture content (35
6-7
-------
to 40 weight percent) and a low wet basis heating value (16,660 kJ/kg [7,190 Btu/lb]), it generally is used as
a fuel only in areas in which it is mined. Lignite is used mainly for steam/electric production in power plants
and typically is fired in larger pulverized coal-fired or cyclone-fired boilers.
Anthracite coal is a high-rank coal with more fixed carbon and less volatile matter than either
bituminous coal or lignite. Because of its low volatile matter content and slight clinkering, anthracite is most
commonly fired in medium-sized traveling grate stokers and small hand-fired units. Some anthracite
(occasionally with petroleum coke) is used in pulverized coal-fired boilers, and it may be blended with
bituminous coal. Because of its low sulfur content (typically less than 0.8 weight percent) and minimal
smoking tendencies, anthracite is considered a desirable fuel in areas where it is readily available. In the
United States, anthracite is mined primarily in northeastern Pennsylvania and consumed mostly in
Pennsylvania and surrounding States. The largest use of anthracite is for space heating. Lesser amounts are
employed for steam/electric production, typically in underfeed stokers and pulverized coal dry-bottom boilers.
Although small quantities of mercury may be emitted as fugitive particulate matter from coal storage
and handling operations, the primary source of mercury and mercury compound emissions from coal
combustion is the combustion stack. Because the combustion zone in boilers operates at temperatures in
excess of 1100°C (2000°F), the mercury in the coal is vaporized and exits the combustion zone as a gas. As
the combustion gases pass through the boiler and the air pollution control system, they cool, and some of the
mercury and mercury compounds may condense on the surface of fine particles. The relative fractions of
vapor- and particle-phase mercury in the exhaust stack depend primarily on the temperature of the air
pollution control system, and the amount of residual carbon in the coal fly ash (some of the vaporous mercury
and mercury compounds may adsorb onto carbon at temperatures present in some air pollution control
devices). To date, little information has been obtained on these distributions.
6.1.3 Emission Control Measures
Emission control measures for coal-fired boilers include controls based on combustor design and
operating practices that are directed primarily at nitrogen oxides (NOX) and particulate matter (PM) control
and add-on air pollution control devices that are designed for acid gas and PM control.43 Those measures
that are most likely to affect mercury control are add-on control systems designed for both PM and acid gas
control. The primary types of PM control devices used for coal combustion include multiple cyclones,
electrostatic precipitators, fabric filters (baghouses), and wet scrubbers, while both wet and dry flue gas
desulfurization (FGD) systems are used for sulfur dioxide (SO,). Some measure of PM control is also
obtained from ash settling in boiler/air heater/economizer dust hoppers, large breeches and chimney bases,
but these mechanisms will not reduce mercury emissions.
Electrostatic precipitators (ESP) are the most common high efficiency control devices used on
pulverized coal and cyclone units. These devices are also being used increasingly on stokers. Generally, PM
collection efficiencies are a function of the specific collection area (i.e., the ratio of the collection plate area
per volumetric flow rate of flue gas through the device). Particulate matter efficiencies of 99.9 weight percent
have been measured with ESP's. Fabric filters have recently seen increased use in both utility and industrial
applications both as a PM control measure and as the collection mechanism in dry FGD systems, generally
achieving about 99.8 percent PM control. Wet scrubbers are also used to control PM emissions, although
their primary use is to control emissions of sulfur oxides. Because, unlike the other PM control devices, wet
scrubbers reduce the gas stream temperature, they may be more effective than the other controls in removing
condensible PM, such as mercury. The other PM control devices would require some type of acid gas control,
such as a spray dryer.
Mechanical collectors, generally multiple cyclones, are the primary means of control on many stokers
and are sometimes installed upstream of high efficiency control devices in order to reduce the ash collection
burden. Depending on application and design, multiple cyclone PM efficiencies can vary tremendously.
However, these systems are relatively inefficient for fine particles and are not likely to provide measurable
control of mercury emissions, which are primarily in the vapor and fine particle fractions of the exhaust.
The section on emissions below presents the available data on emission control system performance.
However, in evaluating the potential emissions from a facility or group of facilities, any assumptions about
control system performance, including those based on the data presented herein, should be examined carefully
to assure that they are supported by reliable test data obtained via methods comparable to those described in
Section 9. Also, performance estimates must be consistent with the physical and chemical properties of the
compounds being emitted and with the operating characteristics of the systems being evaluated.
6-8
-------
6.1.4 Emissions
This revision of the coal-fired boiler section of the previous mercury L&E document presents
separate sections for utility boilers and commercial/industrial/residential coal-fired boilers.32 Since the
previous mercury L&E document was published in 1993, EPA conducted a comprehensive study to estimate
hazardous air pollutant (HAP) emissions, including mercury, from utility boilers.46 The results of this study
were published in a report covering trace metal, organic HAP, and radionuclide emissions; control techniques
from utility boilers; and a comprehensive risk assessment.46 Additionally, a brief description and
presentation of the results of the study specifically with respect to mercury emissions and controls was
published in the Mercury Study Report to Congress.33 These EPA reports quantified the impact on mercury
emissions from coal-fired utility boilers of both coal cleaning and existing combinations of boilers and
control devices. The reports included data from multiple emission test programs and represent the most
comprehensive mercury emission estimates available for coal-fired utility boilers. Therefore, the approach
described in these documents for coal-fired utility boilers has been adopted for this document. For
commercial/industrial/residential coal-fired boilers, the approach adopted in the previous mercury document
(EPA, 1993) is relatively unchanged.32
In providing comments on the draft of this L&E document, EPRI suggested that EPA use the results
of the EPRI report Mercury and Other Trace Metals in Coal, to develop mercury emission estimates from
coal combustion.4 This report presented the results of the analysis of 154 coal samples from full-scale
power plants. These results were also available in Reference 47. For bituminous coal, an average mercury
concentration of 0.087 ppm is reported, a level more than 50 percent lower than the 0.21 ppm average
concentration for the USGS data set. EPRI considers the data presented in Reference 49 to be of better
quality than the USGS data set because of the use of more accurate sampling and analytical techniques.50
Additionally, EPRI asserts that the 154 samples are "coal-as-burned" samples versus those in the USGS data
set that include samples from coal seams containing "significant levels of noncombustibles and uneconomic
samples."50
For the purposes of this L&E document, it was important that the mercury emission estimates be
consistent with mercury emission estimates developed by other groups within EPA. Therefore, for agency
consistency, and, as decribed in this section of the L&E document, the mercury emission estimates presented
reflect those developed in the Utility HAP study. While EPA does not dispute the validity of the mercury in
coal data in Reference 51, these data were not included in the development of the mercury emission estimates
presented in the Utility HAP study and, therefore, are not included in the mercury emission estimates
presented in this section. However, these data may be included in the revised Utility HAP study that is
expected to be released in early 1998. For now, the estimates presented in section 6.1.4 reflect EPA's
position on mercury emission estimates from coal combustion.
6.1.4.1 Utility Boilers. The approach used to develop mercury emission estimates in the Utility
HAP study comprised a two-step process.32'33 First, the mercury concentration in the coal was estimated.
Then, using the boiler-specific data in the Utility Data Institute (UDI)/Edison Electric Institute (EEI) Power
Statistics data base (1991 edition), the estimated mercury concentration in the fuel was multiplied by the fuel
feed rate to obtain the total amount of mercury entering each boiler listed in the data base. Second, "emission
modification factors" (EMF's) were developed based on test data that represent the level of mercury control
that could be expected across various boiler configurations and control devices. The EMF's developed from
the tested units were applied to all other similar units to give mercury emission estimates on a per-unit
basis32'33.
The estimates of mercury concentrations in coal were developed by using a USGS data base of trace
element concentrations in coal by State of coal origin for 3,331 core and channel samples of coal. These
samples came from 50 coal beds having the highest coal production in the United States. The average
mercury content of each of these beds was calculated and the location of each bed was matched with a State.
Using the UDI/EEI data base and records of actual coal receipts, the State from which each utility purchased
the majority of its coal was identified. Then, the mercury content of the coal fired by each utility was
assigned based on the average concentration of mercury calculated for each coal bed.32'33
To account for the impact of coal cleaning on mercury concentration in coal, a 21 percent reduction
in mercury concentration was attributed to coal cleaning for those boilers purchasing bituminous coal from
States where coal cleaning is common practice.32'33 While approximately 77 percent of the eastern and
midwestern bituminous coals are cleaned, the 21 percent reduction was assumed for all boilers burning
6-9
-------
bituminous coal east of the Mississippi River.51 No coal cleaning reductions were applied to lignite or
subbituminous coals, or bituminous coal when the State of coal origin was west of the Mississippi River.32'33
The mercury input to each boiler in the data base was calculated by multiplying the boiler feed rate
by the mercury content in the assigned coal and assessing the 21 percent reduction attributed to coal cleaning,
as appropriate.32'33
Emissions data were available from 51 emission tests conducted by the U.S. EPA, the Electric Power
Research Institute (EPRI), the Department of Energy (DOE), and individual utilities. The EMF's were
calculated from the emission test data by dividing the amount of mercury exiting either the boiler or control
device by the amount of mercury entering the boiler. Boiler-specific emission estimates were then calculated
by multiplying the calculated inlet mercury input by the appropriate EMF for each boiler configuration and
control device. The utility emission test data are listed in Section 10 of the Utility HAP study and in
Appendix B of the U.S. EPA Mercury Report to Congress.33 The EMF's for the various boiler
configurations and control devices are shown in Appendix C of the Utility HAP study and in Appendix C of
the U.S. EPA Mercury Report to Congress.33'46
To calculate the mercury emissions from a specific boiler, the following equation was used:
. \ / Mercury x . . . . Coal \ f \
I Mercury \ / •*_ \ / Boiler \ / , . \ /™m\
I . •1 = 1 content 1 x I _ , ,_ 1 x I cleaning Ixl EMF 1
I emissions I I . ,1 I feed rate I I , I I I
\ / \ in coal / \ / \ factor / \ /
For boilers burning bituminous coal when the State of coal origin was east of the Mississippi River, a coal
cleaning factor of 0.79 (reflecting a mercury reduction of 21 percent) was applied to the above equation. For
all other boilers in the data base, no coal cleaning reductions were applied, i.e., in the above equation, the coal
cleaning factor for these boilers was equated to one. The results of applying this operation to the boilers in
the data base indicate that the total nationwide mercury emissions from coal-fired utility boilers are
approximately 51 ton/yr or 46.3 Mg/yr.33
6.1.4.2 Commercial/Industrial/Residential Boilers. For commercial/industrial/residential boilers, the
data presented above on mercury concentrations in coal and coal heating values were used to develop mass
balance-based emission factors.
The information presented in the literature indicates that virtually 100 percent of the mercury
contained in the coal is emitted from the furnace as either a vapor or fine PM. Consequently, the coal heating
values presented in Table 6-2 and the coal mercury concentrations presented in Table 6-4 can be used to
develop uncontrolled emission factors for major coal types under the conservative assumption that all
mercury in the coal is emitted. Furthermore, note that the coal composition data in Table 6-4 are based on
in-ground mercury concentrations and that calculated emission factors shown in Table 6-6 are based on the
conservative assumption that as-fired coal contains equivalent concentrations. The emission factors do not
account for coal washing. To account for coal washing, a mercury emission reduction of 21 percent can be
applied to the factors in Table 6-6.
The uncontrolled emission factors listed in Table 6-6 were calculated using the coal heating values
from Table 6-2 and the coal mercury concentrations in Table 6-4. These calculated emission factors were
compared with the latest emission factors for coal combustion published in AP-42.52'53 In AP-42, separate
emission factors were developed for bituminous/subbituminous and for anthracite coal combustion based on
available emission test data. For bituminous/subbituminous coal, the AP-42 uncontrolled emission factor is
16 lb/1012 Btu and has an E rating. This factor is identical to the calculated uncontrolled emission factor for
bituminous coal presented in Table 6-6. For anthracite coal combustion, the AP-42 uncontrolled
6-10
-------
TABLE 6-6. CALCULATED UNCONTROLLED MERCURY EMISSION FACTORS
FOR COAL COMBUSTION
Coal type
Bituminousa
Subbituminousb
Anthracite0
Lignited
Calculated mercury emission factors
kg/1015J
7.0
4.5
7.6
9.0
lb/1012Btu
16
10
18
21
g/Mg coal
0.21
0.10
0.23
0.15
10-3lb/toncoal
0.42
0.20
0.46
0.30
aBased on arithmetic average of the five average heating values in Table 6-2.
bBased on arithmetic average of the three average heating values in Table 6-2.
°Based on average heating value for coal category A2 in Table 6-2.
dBased on average heating value for coal category LI in Table 6-2.
emission factor is 0.13 x 10"3 Ib/ton of coal and also has an E rating. This factor, while smaller than the
calculated value for anthracite coal (0.46 x 10"3 Ib/ton of coal) presented in Table 6-6, is of the same order of
magnitude as the calculated value. The AP-42 did not present a separate emission factor for lignite coal
combustion. The emission factors presented in Table 6-6 are considered to be better factors to use in
developing nationwide mercury emission estimates than the AP-42 factors for the following reasons. The two
AP-42 emission factors were developed using limited data while the calculated uncontrolled emission factors
represent a significant volume of mercury-concentration-in-coal data. Calculated uncontrolled factors were
developed for each coal type while the AP-42 emission factors were developed only for
bituminous/subbituminous and anthracite coals.
A comprehensive summary of the test data generated prior to 1989 for coal-fired boilers and furnaces
is presented in Reference 40. The data from individual tests that are presented in that report are compiled in
Table B-l in Appendix B of Reference 54. Table 6-7 summarizes these data as a function of coal type and
control status. Note the wide range of emission factors for each coal type. In addition to the variability in
coal heat content and the uncertainty in mercury sampling and analysis, this range reflects the substantial
variation in coal mercury content and highlights the need to obtain coal-specific mercury data to calculate
emission estimates whenever possible. Also note that the data are combined across industry sector and boiler
type because these parameters are not expected to have a substantial effect on emission factors.
The test data summarized in Table 6-7 from Reference 40, although limited, indicate that essentially
no control of mercury in flue gas is achieved by multiclones, up to 50 percent control is achieved by ESP's,
and limited scrubber data show mercury efficiencies of 50 and 90 percent. Long-term scrubber performance
will depend on the blowdown rate for the scrubber, with efficiency falling if the system approaches
equilibrium. However, according to literature references discussed in Reference 40, these control efficiencies
may be biased high because they are based on data collected using older test methods, which tended to collect
mercury vapor inefficiently. Consequently, these estimates represent upper bounds of efficiencies. More
information on the methods for sampling and analysis of mercury in flue gas is presented in Section 9.
The test data reported in the Utility HAP study comprises data that was collected using more up to
date test methods. This study reported the following mercury control efficiencies for individual control
devices controlling emissions from coal-fired utility boilers: 0 to 59 percent for FGD systems (6 tests);
0 percent control for hot-side ESP's (ESP's located upstream of an FGD unit) (2 tests); zero to 82 percent
control for cold-side ESP's (17 tests); zero to 73 percent control for fabric filters (5 tests); and zero to
55 percent control for spray dryer absorber/fabric filter (SDA/FF) systems (4 tests).
Based on review of the available data, the best estimates for uncontrolled emission factors for typical
coal combustion facilities are those obtained from a mass balance using coal composition data. This
approach was selected because the available uncontrolled test data for commercial/industrial/
6-11
-------
TABLE 6-7. MEASURED MERCURY EMISSION FACTORS FOR COAL COMBUSTION
Coal
typea
Bd
Bd
Bd
Bd
Bd
Bd
SBe
SBe
SBe
Lf
Lf
AS
Control
status"
UN
MPorMC
ESP or
MP/ESP
ESP-2
stage
WSor
MC/WS
FF
UN
ESP or
MP/ESP
WS
MC
ESP
UN
No. of
boilers
17
9
29
1
5
1
3
o
3
2
4
o
3
3
No. of
data
points
34
15
59
5
5
1
5
5
2
4
3
3
Measured mercury emission factors
kg/1015 J
Mean
3.8
12.9
3.4
0.086
7.9
2.0
13.0
1.2
3.4
4.1
0.18
2.3
Range0
0.005-133
0.60-77
0.18-9.6
0.005-0.25
b.d.-37
-
0.28-35
0.16-1.8
2.1-4.7
1.9-9.5
0.099-0.23
1.5-3.0
lb/1012Btu
Mean
8.8
29.9
8.0
0.20
18.4
4.6
30.2
2.7
8.0
9.6
0.41
5.3
Range0
0.011-308
1.4-180
0.41-22.3
0.011-0.56
b.d.-86
-
0.64-81
0.37-4.1
4.9-11
4.4-22
0.23-0.53
3.5-7.0
g/Mg coal
Mean
0.11
0.39
0.10
0.0026
0.24
0.060
0.29
0.027
0.075
0.068
0.0030
0.070
Range0
0.00015-4.0
0.018-2.3
0.0055-0.29
0.00015-0.0075
b.d.-l.l
-
0.0062-0.78
0.0035-0.040
0.047-0.10
0.032-0.16
0.0016-0.0038
0.045-0.091
10-3lb/toncoale
Mean
0.23
0.78
0.21
0.0052
0.48
0.12
0.58
0.052
0.15
0.14
0.0059
0.14
Range0
0.00029-8.0
0.036-4.7
0.011-0.58
0.00029-0.015
b.d.-2.2
-
0.012-1.5
0.0071-0.078
0.094-0.21
0.063-0.32
0.0033-0.0076
0.091-0.18
ON
K^
to
Source: Reference 40.
?B = bituminous, SB = subbituminous L = lignite, A =anthracite.
UN = uncontrolled, MP = mechanical precipitation system, MC =multiclone, ESP = electrostatic precipitator, WS =wet scrubber.
°b.d. = below detection limits.
Based on arithmetic average of the five average heating values in Table 6-2.
JBased on arithmetic average of the three average heating values in Table 6-2.
Based on average heating value for coal category LI in Table 6-2.
gBased on average heating value for coal category A2 in Table 6-2.
-------
residential boilers are of uncertain quality, and the coal concentration data are representative of a much larger
industry segment. Utilizing the available data from Reference 40, and the Utility HAP's study, controlled
emission factors were obtained by applying the following percent removal efficiencies to the uncontrolled
emission factors in Table 6-6. Zero percent efficiency for mechanical collectors, 0 to 82 percent control for
ESP's, 0 to 60 percent control for wet scrubbers and FGD systems, 0 to 73 percent for fabric filters, and 0 to
55 percent for SDA/FF systems. The resultant best typical emission factors are shown in Table 6-8.
The mercury emission factors presented in Table 6-8 should be viewed as the most realistic
nationwide estimates possible, based on the little data that are available. It should be recognized that there is
considerable uncertainty in these estimates. The uncertainty in the estimates is due to the wide variability in
mercury concentrations in coal, the variability in coal heat content, and the uncertainty in sampling and
analytical methodologies for detecting mercury. Therefore, these estimates should not be used to determine
emissions from specific coal combustion facilities.
Estimates of the total 1994 nationwide mercury emissions from coal-fired commercial/industrial/
residential boilers are 21.5 Mg (23.6 tons); for additional details, see Appendix A. The total 1994
nationwide mercury emission estimates for coal combustion (utility plus commercial/industrial/residential)
are 67.8 Mg (74.6 tons).
6.2 FUEL OIL COMBUSTION
As shown in Table 6-1, based on energy consumption estimates by the U.S. Department of Energy,
fuel oil use spans the four sectors of energy users. Distillate fuel oil is used in all sectors with the largest use
in the residential (35 percent) and the industrial (43 percent) sectors, but also with amounts used in both the
commercial (18 percent) and utility (4 percent) sectors. Residual oil is used primarily in the industrial
(29 percent) and utility (59 percent) sectors. Because the oil combustion process is not complex, and control
systems are not widely applied to oil-fired units, the discussion below will focus on fuel characteristics and on
emissions from oil-fired units.39
6.2.1 Fuel Oil Characteristics
The fuel oil characteristics of greatest importance for characterizing mercury emissions from fuel oil
combustion are the heating value and the mercury content of the oil. The heating value is used for converting
from emission factors with mass- or volume-based activity levels to those with activity levels based on heat
input.
The term fuel oil covers a variety of petroleum products, including crude petroleum, lighter
petroleum fractions such as kerosene, and heavier residual fractions left after distillation.40 To provide
standardization and means for comparison, specifications have been established that separate fuel oils into
various grades. Fuel oils are graded according to specific gravity and viscosity, with No. 1 Grade being the
lightest and No. 6 the heaviest. The heating value of fuel oils is expressed in terms of kJ/L (Btu/gal) of oil at
16 °C (60 °F) or kJ/kg (Btu/lb) of oil. The heating value per gallon increases with specific gravity because
there is more weight per gallon. The heating value per mass of oil varies inversely with specific gravity
because lighter oil contains more hydrogen. For an uncracked distillate or residual oil, heating value can be
approximated by the following equation:
Btu/lb = 17,660 + (69 x API gravity)
For a cracked distillate, the relationship becomes:
Btu/lb = 17,780 + (54 x API gravity)
Table 6-9 provides an overall summary of the heating values of typical fuel oils used in the U.S., and
Table 6-10 shows the variability in fuel oil heating values used in various regions of the country.
Appendix B of Reference 40 provides additional details.
The data base for mercury content in fuel oils is much more limited than the coal mercury content
data base. A number of petroleum industry associations were contacted, but none who responded have
6-13
-------
TABLE 6-8. BEST TYPICAL MERCURY EMISSION FACTORS FOR
COMMERCIAL/INDUSTRIAL/RESIDENTIAL COAL-FIRED BOILERS
Coal
typea
B
B
B
B
B
B
Control statusb
Uncontrolled
Mechanical collector
ESP
WS/FGD
FF
SDA/FF
Typical mercury emission factors
kg/1015 J
7.0
7.0
1.3-7.0
0.7-7.0
1.9-7.0
3.2-7.0
lb/1012 Btu
16
16
2.9-16
1.6-16
4.3-16
7.2-16
g/Mg coal
0.21
0.21
0.038-0.21
0.021-0.21
0.012-0.21
0.095-0.21
10-3lb/toncoal
0.42
0.42
0.08-0.42
0.042-0.42
0.11-0.42
0.19-0.42
SB
SB
SB
SB
SB
SB
Uncontrolled
Mechanical collector
ESP
WS/FGD
FF
SDA/FF
4.5
4.5
0.81-4.5
0.4-4.5
1.2-4.5
2.0-4.5
10
10
0.18-10
1-10
2.7-10
4.5-10
0.10
0.10
0.018-0.10
0.010-0.10
0.027-0.10
0.045-0.10
0.20
0.20
0.036-0.020
0.02-0.20
0.05-0.20
0.09-0.20
A
A
A
A
A
A
Uncontrolled
Mechanical collector
ESP
WS/FGD
FF
SDA/FF
7.6
7.6
1.4-7.6
0.7-7.6
2.1-7.6
3.4-7.6
18
18
3.2-18
1.8-18
4.9-18
8.1-18
0.23
0.23
0.04-0.23
0.023-0.23
0.06-0.23
0.10-0.23
0.46
0.46
0.08-0.46
0.046-0.46
0.12-0.46
0.21-0.46
L
L
L
L
L
L
Uncontrolled
Mechanical collector
ESP
WS/FGD
FF
SDA/FF
9.0
9.0
1.6-9.0
0.9-9.0
2.4-9.0
4.1-9.0
21
21
3.8-21
2.1-21
5.7-21
9.5-21
0.15
0.15
0.03-0.15
0.015-0.15
0.04-0.15
0.07-0.15
0.30
0.30
0.05-0.30
0.030-0.30
0.08-0.30
0.14-0.30
Source: Reference 32.
aB = bituminous, SB = subbituminous, A = anthracite, L = lignite.
bESP = electrostatic precipitator, WS/FGD = wet scrubber or flue gas desulfurization system,
FF = filter fabric, and SDA/FF = spray dryer absorber/fabric filter system.
6-14
-------
TABLE 6-9. TYPICAL HEATING VALUES OF FUEL OILS
Type
Color
Heating valuea
kJ/L
(Btu/gal)
kJ/kg
(Btu/lb)
FUEL OIL GRADES
No. 1
Distillate
Light
38,200
(137,000)
45,590-46,030
(19,670-19,860)
No. 2
Distillate
Amber
40,900
(141,000)
44,430-45,770
(19,170-19,750)
No. 4
Very light
residual
Black
40,700
(146,000)
42,370-44,960
(18,280-19,400)
No. 5
Light residual
Black
41,200
(148,000)
41,950-44,080
(18,100-19,020)
No. 6
Residual
Black
41,800
(150,000)
40,350-43,800
(17,410-18,900)
Crude
40,000-42,300
(144,000- 152,000)
40,700-43,300
(17,500-18,600)
ON
Source: Reference 40; and Reference 54.
a The distillate samples, as well as the residual samples, analyzed for Btu/gal and Btu/lb heating values are different; therefore, the heating values presented do not
directly correspond to one another.
b These crude oil values are based on a limited number of samples from West Coast field sites presented in Reference 55 and may not be representative of the
distribution of crude oils processed in the United States.
-------
TABLE 6-10. TYPICAL FUEL OIL HEATING VALUES FOR SPECIFIC REGIONS
Region
Eastern
Southern
Central
Rocky
Mountain
Western
No. 1 fuel oil
Heating value, kJ/L (Btu/gal)
No. of
samples
33
13
27
14
16
Range
36,900-37,800
(132,500-135,700)
37,000-37,700
(132,900-135,400)
36,900-37,800
(132,500-135,700)
37,100-37,600
(133,100-135,100)
36,700-37,900
(131,700-136,200)
Average
37,400
(134,200)
37,400
(134,300)
37,300
(134,000)
37,400
(134,200)
37,500
(134,600)
No. 2 fuel oil
No. of
samples
56
19
35
17
18
Range
37,100-40,800
(133,100-146,600)
38,000-39,400
(136,400-141,500)
37,800-40,800
(135,900-146,600)
37,900-39,100
(136,100-140,400)
37,900-39,100
(136,100-140,500)
Average
38,800
(139,500)
38,800
(139,400)
38,800
(139,200)
38,700
(139,000)
38,700
(139,000)
No. 4 fuel oil
No. of
samples
1
0
2
2
0
Range
...
...
40,700-41,800
(146,000-
150,100)
41,800-41,900
(150,100-
150,500)
—
Average
40,700
(146,000)
...
41,200
(148,000)
41,900
(150,300)
—
Region
Eastern
Southern
Central
Rocky
Mountain
Western
No. 5 fuel oil (light)
Heating value, kJ/L (Btu/gal)
No. of
samples
1
0
4
2
0
Range
...
...
41,300-42,200
(148,400-151,500)
42,900-43,600
(153,900-156,500)
...
Average
41,300
(148,400)
...
41,700
(149,900)
43,200
(155,200)
...
No. 6 fuel oil
No. of
samples
17
14
10
7
12
Range
40,900-43,900
(147,000-157,600)
41,900-43,600
(150,500-156,500)
41,900-44,200
(150,600-158,900)
42,300-44,300
(151,900-159,200)
41,700-45,500
149,900-163,500)
Average
43,300
(151,900)
42,600
(152,900)
42,600
(152,900)
43,100
(154,600)
43,000
(154,400)
ON
I
ON
Source: Reference 40.
-------
done any research on metals content in fuel oils. No single centralized data base is available, and the
information presented below is based on limited data from individual studies.
Concentrations of mercury in fuel oil depend upon the type of oil used. No comprehensive oil
characterization studies have been done, but data in the literature report mercury concentrations in crude oil
ranging from 0.023 to 30 ppmwt, while the range of concentrations in residual oil is 0.007 to 0.17 ppmwt.
Because only a single mean value was found in the literature for mercury concentration in distillate oil, no
conclusions can be drawn about the range of mercury in distillate oil. Table 6-11 lists typical values for
mercury in oils, which were obtained by taking the average of the mean values found in the literature. The
value for distillate oil is the single data point found in the literature and may not be as representative as the
values for residual and crude oils.
TABLE 6-11. MERCURY CONCENTRATION IN OIL BY OIL TYPE
Fuel oil type
Residual No. 6
Distillate No. 2
Crude
No. of samples
46
Mercury concentration, ppmwt
Range
0.002-0.006
0.007-30
Typical value
0.004a
<0.12b
3.5C
Source: References 40, 50, 56.
aMidpoint of the range of values.
bAverage of data from three sites.
cAverage of 46 data points was 6.86; if the single point value of 23.1 is eliminated, average based on 45
remaining data points is 1.75. However, the largest study with 43 data points had an average of
3.2 ppmwt. A compromise value of 3.5 ppmwt was selected as the best typical value.
6.2.2 Process Description
Fuel oils are broadly classified into two major types: distillate and residual. Distillate oils (fuel oil
grade Nos. 1 and 2) are more volatile and less viscous than residual oils, having negligible ash and nitrogen
contents and usually containing less than 0.1 weight percent sulfur. No. 4 residual oil is sometimes classified
as a distillate; No. 6 is sometimes referred to as Bunker C. Being more viscous and less volatile than
distillate oils, the heavier residual oils (Nos. 5 and 6) must be heated to facilitate handling and proper
atomization. Because residual oils are produced from the residue after lighter fractions (gasoline and
distillate oils) have been removed from the crude oil, they contain significant quantities of ash, nitrogen, and
sulfur. Small amounts of crude oil are sometimes burned for steam generation for enhanced oil recovery or
for refinery operations.43'48
Oil-fired boilers and furnaces are simpler and have much less variation in design than the coal-fired
systems described earlier. The primary components of the system are the burner, which atomizes the fuel and
introduces it along with the combustion air into the flame, and the furnace, which provides the residence time
and mixing needed to complete combustion of the fuel. The primary difference in systems that fire distillate
oil and residual oil is that the residual oil systems must have an oil preheater to reduce the viscosity of the oil
so that it can be atomized properly in the burner. Systems that fire distillate oil and residual oil also have
different atomization methods.
The only source of mercury emissions from oil-fired boilers and furnaces is the combustion stack.
Because the entire fuel supply is exposed to high flame temperatures, essentially all of the mercury and
mercury compounds contained in the fuel oil will be volatilized and exit the furnace with the combustion
gases. Unless these combustion gases are exposed to low-temperature air pollution control systems and high-
efficiency PM control systems, which typically are not found on oil-fired units, the mercury and mercury
compounds will be exhausted in vapor phase through the combustion stack.
6-17
-------
6.2.3 Emission Control Measures
The three types of control measures applied to oil-fired boilers and furnaces are boiler modifications,
fuel substitution, and flue gas cleaning systems.40'48 Only fuel substitution and flue gas cleaning systems
may affect mercury emissions. Fuel substitution is used primarily to reduce SO2 and NOX emissions.
However, if the substituted fuels have lower mercury concentrations, the substitution will also reduce mercury
emissions. Because PM emissions from oil-fired units are generally much lower than those from coal-fired
units, high-efficiency PM control systems are generally not employed on oil-fired systems. However, the flue
gas cleaning systems that are used on oil-fired units are described briefly below.
Flue gas cleaning equipment generally is employed only on larger oil-fired boilers. Mechanical
collectors, a prevalent type of control device, are primarily useful in controlling PM generated during soot
blowing, during upset conditions, or when a very dirty heavy oil is fired. During these situations, high
efficiency cyclonic collectors can achieve up to 85 percent control of PM, but negligible control of mercury is
expected with mechanical collectors.
Electrostatic precipitators are used on approximately one-third of the oil-fired power plants. Older
ESP's may remove 40 to 60 percent of the PM, but negligible mercury control is expected. Newer ESP's may
be more efficient, but no data are available for oil-fired power plants. Recent test data indicate mercury
control efficiencies for ESP's controlling emissions from oil-fired utility boilers of 42 and 83 percent.4
Scrubbing systems have been installed on oil-fired boilers to control both sulfur oxides and PM. Similar to
systems applied to coal combustion (presented in Reference 40), these systems can achieve PM control
efficiencies of 50 to 90 percent. Because they provide gas cooling, some mercury control may be obtained,
but little data are available on their performance.
6.2.4 Emissions
The only substantive source of mercury emissions from fuel oil combustion operations is the
combustion gas exhaust stack. Three types of information were used to develop emission factors for oil
combustion. First, the data described above on fuel oil heating value and mercury content of fuel oils were
used to develop emission factors by mass balance, assuming conservatively that all mercury fired with the
fuel oil is emitted through the stack. Second, the emission factors developed in AP-42 for residual and
distillate oil combustion and in Reference 47 for residual oil combustion were evaluated. Third, rated
emission test data were evaluated and summarized. The paragraphs below first present the results generated
from each of the three sources. Then, the relative merits of the emission factors generated via each of the
procedures are discussed, and the best "typical" emission factors are identified.
The literature on fuel oil combustion suggests that essentially all mercury in the fuel oil is vaporized
in the combustion zone and exhausted as a vapor in the combustion gas stream. Using the assumption that
100 percent of the mercury in fuel oil leaves the boiler or furnace in the exhaust gases, the data in Tables 6-9
and 6-11 were used to calculate uncontrolled emission factors for No. 2 distillate and No. 6 residual oil. Data
presented in Reference 52, which show average crude oil heating values of 42,500 kJ/kg (18,300 Btu/lb) and
41,300 kJ/L (148,000 Btu/gal), can be combined with the mercury content data in Table 6-11 to calculate
uncontrolled emission factors for crude oil combustion. The results of these calculations are presented in
Table 6-12.
The calculated emission factors in Table 6-12 were compared to the available emission factors for
fuel oil combustion from AP-42. The AP-42 presents emission factors for No. 2 and No. 6 fuel oils; no
emission factors are developed for crude oil in AP-42.53 The AP-42 emission factor for residual oil (No. 6)
combustion is based on emission tests from 15 sites conducted from April 1990 through April 1994. The
average emission factor reported for mercury emissions is 1.13 E-04 lb/103 gallons(0.73 lb/1012 Btu). This
emission factor is rated C. The comparable calculated emission factor for residual oil in Table 6-12 based on
the mercury content in the oil is 3.3 E-05 lb/103 gallons (0.21 lb/1012 Btu).
The AP-42 emission factor for distillate oil (No. 2) combustion (3.0 lb/1012 Btu) is actually based on
the average concentration of mercury in residual oil. It is not based on any emission test data and is rated E.
Additionally, the residual oil mercury concentration data used to develop this estimate are somewhat dated.
The comparable calculated emission factor for distillate oil in Table 6-12 is 6.2 lb/1012
6-18
-------
TABLE 6-12. CALCULATED UNCONTROLLED MERCURY EMISSION FACTORS
FOR FUEL OIL COMBUSTION
Fuel oil type
Residual No. 6a
Distillate No. 2a
Crudeb
Calculated mercury emission factors
kg/1015 J
0.092
2.7
82
lb/1012 Btu
0.21
6.2
190
g/Mg
fuel oil
0.004
0.12
3.5
ID'3 Ib/ton
fuel oil
0.008
0.24
7.0
g/103L
fuel oil
0.0039
0.10
3.4
lb/106 gal
fuel oil
0.033
0.86
28
aBased on typical heating values in Table 6-9 and mercury concentrations in Table 6-11.
bBased on average crude oil heating values in Reference 54 and mercury concentrations in Table 6-11.
Btu and is based on the average of the mercury concentration measured in distillate oil samples at three sites
as part of the California AB2588 study.50
Reference 40 contains some mercury emission test data for the combustion of residual oil, distillate
oil, and a 1:1 mixture of residual/crude oil. All of these data were developed from 1979 through 1981 and
were presented in the previous mercury L&E. In an effort to eliminate mercury emission test data collected
using older, less reliable emission test methods, EPA elected to utilize only post-1990 emission test data.
This approach is consistent with the approach utilized in EPA's Utility HAP Study. Therefore, the emission
test data from Reference 40 are not utilized here; instead, more recent test data are presented.
Table 6-13 presents the results of a series of emission tests for the combustion of residual oil
reported in Reference 47. As part of this test program, residual oil mercury concentrations were also
measured; these data are also presented in Table 6-13. The data show that the mercury emissions from
residual oil combustion are highly variable and that in most cases, the measured stack emissions are higher
than the inlet fuel levels. Because these data are not normally distributed and appear to be log normal, a
geometric mean was calculated to better represent the range of the data (References 47 and 56). The
geometric mean for these data is 0.46 lb/1012 Btu. Data are not available for distillate or crude oil
combustion in Reference 47.
In summary, three mercury emission factors are presented for residual oil combustion: the
0.73 lb/1012 Btu factor from AP-42, 0.46 lb/1012 Btu from EPRI, and 0.21 lb/1012 Btu from the EPRI
residual oil analyses. Because the 0.46 lb/1012 Btu emission factor is essentially the midpoint of the range of
the three values, this factor was selected as the best "typical" emission factor for residual oil combustion.
Because there are no emission test data for distillate oil combustion, the mass balance approach was used to
estimate the best "typical" emission factor for distillate oil combustion.
As a part of the previous L&E study, two test reports prepared as a part of the California "Hot
Spots" program were reviewed.54'57 The emission factors generated from these three reports are summarized
in Table 6-14. Each of the reports contained the data on fuel oil characteristics needed to calculate mercury
input rates, so Table 6-14 contains both calculated emission factors based on mercury input levels and
measured emission factors based on stack tests. Because mercury levels in all of the fuel oils tested were
below detection limits, all calculated emission factors are reported as "less than" values. Note that only one
of the two tests showed mercury levels above the detection limit in the stack. That test showed measured
emissions to be substantially greater than mercury input to the process, making the results suspect. These
discrepancies may be a function of the analytical problems that have been reported for mercury methods
applied to combustion sources. These problems are discussed in more detail in Section 9. On balance, these
data provide little information for emission factor development.
The available information on uncontrolled mercury emissions from crude oil combustion is
ambiguous. The limited test data presented in Table 6-14 show measured factors that range from less than
0.05 to 15 kg/1015 J (<0.12 to 34 lb/1012 Btu), a range of almost three orders of magnitude. The calculated
emission factor of 84 kg/1015 J (190 lb/1012 Btu), which is based on limited fuel composition
6-19
-------
TABLE 6-13. MERCURY CONCENTRATIONS IN RESIDUAL
OIL AND MERCURY EMISSION FACTORS FROM RESIDUAL COMBUSTION
Unit name
117
118
112
13
103
106
107
104
105
108
109
13
118
112
13
117
Residual oil mercury
concentration, ppmw
0.0023
0.0040
0.0060
<0.040
<0.090
<0.10
<0.10
<0.10
<0.10
<0.10
<0.90
<0.030
0.0040
0.0060
<0.040
0.0023
Mean mercury emission
factor, lb/10r2 Btu
0.60
0.98
1.3
0.23
<3.6
<5.0
<37
12
<4.7
<32
1.8
0.16
0.50
0.24
<0.066
0.49
Source: Reference 47.
6-20
-------
TABLE 6-14. MERCURY EMISSION FACTORS FOR CRUDE OIL COMBUSTION
GENERATED FROM CALIFORNIA "HOT SPOTS" TESTS
Process type
Pipeline/
process heater^
Generated
Fuel oil
type
Crude
Crude
Calculated mercury emission factoi%
kg/1015J
<2.4
<2.4
lb/1012
Btu
<5.6
<5.6
g/Mg fuel
oil
<0.10
<0.10
10'3
Ib/ton
fuel oil
<0.20
<0.21
g/lO'Lfuel
oil
<0.097
<0.10
lb/106 gal
fuel oil
<0.81
<0.83
Measured mercury emission factor^
kg/10 J
<0.052
14.7
'fb/lO^Btu
<0.12
34.1
g/Mg fuel
oil
<0.0022
0.62
10'3lb/ton
fuel oil
<0.0044
1.2
g/lO'Lfuel
oil
<0.0021
0.61
lb/106 gal
fuel oil
<0.018
5.1
Source: Reference 54; Reference 57.
aEmission factors were based on assumed crude oil heating value of 42,500 kJ/kg (18,300 Btu/lb) and density of 0.97 kg/L (8.1 Ib/gal).
bMercury detection limit is 0.1 mg/kg.
cMercury detection limit is 0.1 mg/L.
Os
NJ
-------
and heating value data, expands the range even further. Because these data are quite sparse and the relative
quality of the data is uncertain, the midpoint of the range was selected as the best "typical" emission factor.
The uncontrolled emission factors for distillate, residual, and crude oil are presented in Table 6-15.
Data are insufficient to develop controlled emission factors for fuel oil combustion. There is considerable
uncertainty in these emission factor estimates due to the variability of mercury concentrations in fuel oil, the
incomplete data base on distillate oil, and the uncertainty in sampling and analysis for detecting mercury.
Therefore, these estimates should not be used to determine emissions from specific oil-fired units.
TABLE 6-15. BEST TYPICAL MERCURY EMISSION FACTORS FOR FUEL
OIL COMBUSTION
Fuel oil type
Residual No. 6
Distillate No. 2
Crude
Typical mercury emission factors
kg/1015 J
0.20
2.7
41
lb/1012 Btu
0.46
6.2
95
g/Mg fuel
oil
0.009
0.12
1.7
10-3lb/tonfuel
oil
0.017
0.24
3.5
g/103 L
fuel oil
0.0085
0.10
1.7
lb/106 gal fuel
oil
0.071
0.86
14
Total 1994 mercury emissions from oil combustion (utility, industrial, and commercial/residential)
are estimated to be 7.6 Mg (8.4 tons); see Appendix A for details.
6.3 WOOD COMBUSTION
Wood and wood wastes are used as fuel in both the industrial and residential sectors. In the
industrial sector, wood waste is fired in industrial boilers to provide process heat, while wood is burned in
fireplaces and wood stoves in the residential sector. Studies have shown that wood and wood wastes may
contain mercury; however, insufficient data are available to estimate the typical mercury content in wood and
wood wastes. The information below includes process descriptions for the three combustion processes
(boilers, fireplaces, and wood stoves), descriptions of the control measures used for wood-fired processes,
and emission factors.
6.3.1 Process Description
6.3.1.1 Industrial Boilers. Wood waste combustion in boilers is confined primarily to those
industries in which wood waste is available as a byproduct. These boilers are used to generate heat energy
and to alleviate potential solid waste disposal problems. In boilers, wood waste is normally burned in the
form of hogged wood, bark, sawdust, shavings, chips, mill rejects, sanderdust, or wood trim. Heating values
for this waste range from about 9,300 to 12,000 kJ/kg (4,000 to 5,000 Btu/lb) of fuel on a wet, as-fired basis.
The moisture content of as-fired wood is typically near 50 weight percent, but may vary from 5 to 75
weight percent, depending on the waste type and storage operations. Generally, bark is the major type of
waste burned in pulp mills; either a mixture of wood and bark waste or wood waste alone is burned most
frequently in the lumber, furniture, and plywood industries.58 One National Council of the Paper Industry for
Air and Stream Improvement (NCASI) study found the mercury content of bark waste to range from <0.08 to
0.84ppmwt.59
As of 1980, there were about 1,600 wood-fired boilers operating in the U.S., with a total capacity of
approximately 30.5 gigawatts (GW) (1.04 x 1011 Btu/hr).60 No specific data on the distribution of these
boilers were identified, but most are likely to be located where pulp and paper mills or other wood product
plants are located (i.e., in the Southeast, Pacific Northwest, Wisconsin, Michigan, and Maine).
Various boiler firing configurations are used for burning wood waste. One common type of boiler
used in smaller operations is the Dutch oven. This unit is widely used because it can burn fuels with very
high moisture content. Fuel is fed into the oven through an opening in the top of a refractory-lined furnace.
The fuel accumulates in a cone-shaped pile on a flat or sloping grate. Combustion is accomplished in two
stages: (1) drying and gasification and (2) combustion of gaseous products. The first stage takes place in the
primary furnace, which is separated from the secondary furnace chamber by a bridge wall. Combustion is
6-22
-------
completed in the secondary chamber before gases enter the boiler section. The large mass of refractory helps
to stabilize combustion rates but also causes a slow response to fluctuating steam demand.58
In another boiler type, the fuel cell oven, fuel is dropped onto suspended fixed grates and is fired in a
pile. Unlike the Dutch oven, the refractory-lined fuel cell also uses combustion air preheating and positioning
of secondary and tertiary air injection ports to improve boiler efficiency. Because of their overall design and
operating similarities, however, fuel cell and Dutch oven boilers have comparable emission characteristics.58
The most common firing method employed for wood-fired boilers with a steam generation rate
greater than 45,000 kg/hr (100,000 Ib/hr) is the spreader stoker. With this boiler, wood enters the furnace
through a fuel chute and is spread either pneumatically or mechanically across the furnace, where small pieces
of the fuel burn while in suspension. Simultaneously, larger pieces of fuel are spread in a thin, even bed on a
stationary or moving grate. The burning is accomplished in three stages in a single chamber: (1) moisture
evaporation; (2) distillation and burning of volatile matter; and (3) burning of fixed carbon. This type of
boiler has a fast response to load changes, has improved combustion control, and can be operated with
multiple fuels. Natural gas, oil, and/or coal, are often fired in spreader stoker boilers as auxiliary fuels. The
fossil fuels are fired to maintain a constant steam supply when the wood waste moisture content or mass rate
fluctuates and/or to provide more steam than can be generated from the wood waste supply alone. Although
spreader stokers are the most common stokers among larger wood-fired boilers, overfeed and underfeed
stokers are also utilized for smaller units.58
Another boiler type sometimes used for wood combustion is the suspension-fired boiler. This boiler
differs from a spreader stoker in that small-sized fuel (normally less than 2 mm [0.08 in.]) is blown into the
boiler and combusted by supporting it in air rather than on fixed grates. Rapid changes in combustion rate
and, therefore, steam generation rate are possible because the finely divided fuel particles burn quickly.58
A recent innovation in wood firing is the fluidized bed combustion (FBC) boiler. A fluidized bed
consists of inert particles through which air is blown so that the bed behaves as a fluid. Wood waste enters in
the space above the bed and burns both in suspension and in the bed. Because of the large thermal mass
represented by the hot inert bed particles, fluidized beds can handle fuels with moisture contents up to near
70 percent (wet basis). Fluidized beds also can handle dirty fuels (up to 30 percent inert material). Wood
fuel is pyrolyzed faster in a fluidized bed than on a grate due to its immediate contact with hot bed material.
As a result, combustion is rapid and results in nearly complete combustion of the organic matter, thereby
minimizing emissions of unburned organic compounds.58
6.3.1.1 Residential Wood Stoves. Wood stoves are enclosed wood heaters that control burning or
burn time by restricting the amount of air that can be used for combustion. They are commonly used in
residences as space heaters, both as the primary source of residential heat and as a supplement to
conventional heating systems. Based on known variations in construction, combustion, and emission
characteristics, there are five different categories of residential wood burning devices: (1) the conventional
wood stove; (2) the noncatalytic wood stove; (3) the catalytic wood stove; (4) the pellet stove; and (5) the
masonry heater.61
The conventional stove category comprises all stoves without catalytic combustors not included in
the other noncatalytic categories (i.e., noncatalytic and pellet). Conventional stoves do not have any
emissions reduction technology or design features and, in most cases, were manufactured before July 1, 1986.
Stoves of many different airflow designs may be in this category, such as updraft, downdraft, crossdraft, and
S-flow.61
Noncatalytic wood stoves are those units that do not employ catalysts but do have emission-reducing
technology or features. Typical noncatalytic design includes baffles and secondary combustion chambers.61
Catalytic stoves are equipped with a ceramic or metal honeycomb device, called a combustor or
converter, that is coated with a noble metal such as platinum or palladium. The catalyst material reduces the
ignition temperature of the unburned volatile organic compounds (VOC's) and carbon monoxide (CO) in the
exhaust gases, thus augmenting their ignition and combustion at normal stove operating temperatures. As
these components of the gases burn, the temperature inside the catalyst increases to a point at which the
ignition of the gases is essentially self-sustaining.61
6-23
-------
Pellet stoves are those fueled with pellets of sawdust, wood products, and other biomass materials
pressed into manageable shapes and sizes. These stoves have active air flow systems and unique grate design
to accommodate this type of fuel.61
Masonry heaters are large, enclosed chambers made of masonry products or a combination of
masonry products and ceramic materials. Masonry heaters are gaining popularity as a cleaner-burning, heat-
efficient form of primary and supplemental heat, relative to some other types of wood heaters. In a masonry
heater, a complete charge of wood is burned in a relatively short period of time. The use of masonry
materials promotes heat transfer. Thus, radiant heat from the heater warms the surrounding area for many
hours after the fire has burned out.61
6.3.1.2 Residential Fireplaces. Fireplaces are used primarily for aesthetic effects and secondarily as
a supplemental heating source in homes and other dwellings. Wood is most commonly used as fuel, but coal
and densified wood "logs" also may be burned. The user intermittently adds fuel to the fire by hand.62
Fireplaces can be divided into two broad categories: (1) masonry (generally brick and/or stone,
assembled on site, and integral to a structure) and (2) prefabricated (usually metal, installed on site as a
package with appropriate duct work).62
Masonry fireplaces typically have large fixed openings to the fire bed and have dampers above the
combustion area in the chimney to limit room air and heat losses when the fireplace is not being used. Some
masonry fireplaces are designed or retrofitted with doors and louvers to reduce the intake of combustion air
during use.6
Prefabricated fireplaces are commonly equipped with louvers and glass doors to reduce the intake of
combustion air, and some are surrounded by ducts through which floor level air is drawn by natural
convection, heated, and returned to the room.62
All of the systems described above operate at temperatures that are above the boiling point of
mercury and mercury compounds. Consequently, any mercury contained in the wood fuel will be emitted with
the combustion gases via the exhaust stack.
6.3.2 Emission Control Measures
Although some wood stoves use emission control measures such as catalysts and secondary
combustion chambers to reduce VOC and CO emissions, these techniques are not expected to affect mercury
emissions. However, wood-fired boilers employ PM control equipment which may provide some reduction.
These systems are described briefly below.
Currently, the four most common control devices used to reduce PM emissions from wood-fired
boilers are mechanical collectors, fabric filters, wet scrubbers, and electrostatic precipitators (ESP's).58 Of
these controls, only the last two have the potential for significant mercury reduction.
The most widely used wet scrubbers for wood-fired boilers are venturi scrubbers. With gas-side
pressure drops exceeding 4 kilopascals (kPa) (15 inches of water), PM collection efficiencies of 90 percent or
greater have been reported for venturi scrubbers operating on wood-fired boilers.58 No data were located on
the performance of these systems relative to mercury emissions. However, some control is expected
(probably in the range of 50 to 90 percent) based on results achieved for coal combustion sources.
Fabric filters (i.e., baghouses) and ESP's are employed when PM collection efficiencies above
95 percent are required. Collection efficiencies of 93 to 99.8 percent for PM have been observed for ESP's
operating on wood-fired boilers, but mercury efficiencies are likely to be substantially lower (probably
50 percent or less) based on the performance of ESP's in controlling mercury from coal combustion
sources.58 The performance of ESP's in controlling mercury emissions depends on operating temperature and
the amount of carbon in the fly ash.
Fabric filters have had limited applications to wood-fired boilers. The principal drawback to fabric
filtration, as perceived by potential users, is a fire danger arising from the collection of combustible
carbonaceous fly ash. Despite potential complications, fabric filters are generally preferred for boilers firing
salt-laden wood. This fuel produces fine PM with a high salt content for which fabric filters can achieve high
collection efficiencies. In two tests of fabric filters operating on salt-laden wood-fired boilers, PM collection
6-24
-------
efficiencies were above 98 percent.58 No data are available on mercury emission reduction for fabric filters,
but results for other combustion sources suggest that efficiencies will be very low.
6.3.3 Emissions
The primary source of mercury emissions from wood combustion processes is the combustion gas
exhaust stack. Very small quantities of mercury also may be emitted with the fugitive PM emissions from
bottom and fly ash handling operations.
The data on mercury emissions from wood combustion are limited. A recent National Council of the
Paper Industry for Air and Stream Improvement (NCASI) report provided a range and average emission
factor for boilers without electrostatic precipitators (ESP's) and for boilers with ESP's.63 The boilers without
ESP's included a variety of control devices including cyclones, multiclones, and various wet scrubbers. The
average emission factor reported for boilers without ESP's was 3.5 x 10~6 kg/Mg (6.9 x 10~6 Ib/ton) of dry
wood burned. The average emission factor reported for boilers with ESP's was 1.3 x 10~6 kg/Mg
(2.6 x 10~6 Ib/ton) of dry wood burned.
The most recent AP-42 section on wood waste combustion in boilers provided an average
uncontrolled emission factor for mercury emissions based on four emission test reports.58 The AP-42
uncontrolled emission factor for mercury emissions from wood waste combustion is 2.6 x 10~6 kg/Mg
(5.2 x 10~6 Ib/ton) of wet, as-fired wood burned.
The NCASI average emission factor reported for wood-fired boilers with ESP's of 1.3 x 106 kg/Mg
(2.6 x 10~6 Ib/ton) of dry wood burned is recommended for estimating mercury emissions from wood waste
combustion in boilers.
For residential wood combustion, only one emission factor was found in the literature.64 This
emission factor is based on one test burning one type of wood (pine) at a single location. In 1987, the
Department of Energy estimated that 22.5 million households burned approximately 42.6 million cords of
wood.65 Given that the density of wood varies greatly by wood species and moisture content, and that the
above emission factor is from a single test, nationwide emissions of mercury from residential wood
combustion were not estimated.
Total 1994 mercury emissions from wood combustion are estimated to be 0.1 Mg (0.1 tons); see
Appendix A for details.
6.4 MUNICIPAL WASTE COMBUSTION
Refuse or municipal solid waste (MSW) consists primarily of household garbage and other
nonhazardous commercial, institutional, and nonmanufacturing industrial solid waste. Municipal waste
combustors (MWC's) are used to reduce the mass and volume of MSW that ultimately must be landfilled. In
fact, MWC's reduce the volume of MSW by about 90 percent.
In the previous mercury L&E, it was estimated that there were over 160 MWC plants in operation in
the United States with capacities greater than 36 megagrams per day (Mg/d) [40 tons per day (ton/d)] and a
total capacity of approximately 100,000 Mg/d (110,000 ton/d) of MSW.32 A number of MWC plants have
closed since 1991. At the beginning of 1995, over 130 MWC plants with aggregate capacities of greater than
36 Mg/d (40 ton/d) of MSW were operating in the United States. The number of combustion units per
facility ranges from one to six, with the average being two. Total facility capacity ranges from 36 to 2,700
Mg/d (40 to 3,000 ton/d). Together these plants have a total capacity of approximately 90,000 Mg/d
(99,000 ton/d).66
In addition to the MWC's discussed above, a number of smaller MWC's in the United States have
plant capacities of less than 36 Mg/d (40 ton/d). This population of smaller MWC's comprises a very small
fraction of the nation's total MWC capacity.
Table 6-16 shows the geographic distribution of MWC units and capacities by States for MWC
plants larger than 35 Mg/d. This distribution reflects the MWC's that were operational in 1995.67
6-25
-------
6.4.1 Municipal Solid Waste Characteristics
Municipal solid waste is a heterogeneous mixture of the various materials found in household,
commercial, institutional, and nonmanufacturing industrial wastes. Major constituents in typical municipal
waste are listed in Table 6-17. In 1994, a total of 145.24 million Mg (159.76 million tons) of municipal solid
waste was discarded in the municipal waste stream. Of this total, 74.8 percent was due to materials in
discarded products and 25.2 percent other waste, such as food wastes and yard trimmings.68 Known sources
of mercury in MSW are batteries, discarded electrical equipment and wiring, fluorescent bulbs, paint residues,
and plastics. As of 1989, 644 Mg (709 tons) of mercury were reported to be discarded in the municipal solid
waste stream, and the concentration of mercury in solid waste is reported to be in the range of less than 1 to
6 ppm by weight with a typical value of 4 ppm by weight. However, because of changes in mercury
consumption, the quantity of mercury discarded in the municipal solid waste stream has decreased
dramatically since 1989 and is expected to decrease in the future.69'70
The most recent report on mercury discarded in solid municipal waste was a 1992 EPA report based
on 1989 data with projections to the year 2000. One of the most common sources of mercury in this waste
was from the discard of batteries; in 1989, it was estimated that about 88 percent of the total discard of
mercury was from batteries. Of the 88 percent, about 28 percent was from mercuric oxide batteries and the
remainder from alkaline and other batteries.38 According to the Bureau of Mines (now part of USGS)
estimates, 250 Mg (275 tons) of mercury were used in battery production in 1989; current USGS estimates
for 1995 are 6 Mg (6.6 tons) and for 1996, less than 0.5 Mg (0.55 tons).2 As discussed in Section 5.2, only
mercuric oxide button cells and the larger mercuric oxide batteries use mercury to any extent. The proportion
of mercury usage between the button cells and the larger batteries is not available but essentially all of the
larger batteries are used in hospital and military applications and, therefore, would generally not be contained
in the municipal solid waste stream. Battery discards from hospital and military applications would be either
recycled or disposed at the facility.30'71 Hospital battery discards incinerated at the facility would be a
component of the medical waste combustion estimates.
6.4.2 Process Description
The three principal MWC classes are mass burn, refuse-derived fuel (RDF), and modular
combustors. The paragraphs below briefly describe some of the key design and operating characteristics of
these different combustor types.72'73
In mass burn units, the MSW is combusted without any preprocessing other than removal of items
too large to go through the feed system. In a typical mass burn combustor, unprocessed waste (after removal
of bulky, noncombustible items) is delivered by an overhead crane to a feed hopper. From the feed hopper,
refuse is fed into the combustion chamber on a moving grate. Combustion air in excess of stoichiometric
amounts is supplied below (underfire air) and above (overfire air) the grate. Mass burn combustors are
usually erected at the site (as opposed to being prefabricated at another location) and range in size from 46 to
900 Mg/day (50 to 1,000 tons/d) of MSW throughput per unit. The mass burn combustor category can be
divided into mass burn refractory wall (MB/REF), mass burn/waterwall (MB/WW), and mass burn/rotary
waterwall (MB/RC) designs. The two most common, MB/WW and MB/REF, are described below.
The MB/WW design represents the predominant technology in the existing population of large
MWC's, and it is expected that the majority of new units will be MB/WW designs. In MB/WW units, the
combustor walls are constructed of metal tubes that contain pressurized water and recover radiant energy
from the combustion chamber. Trucks deliver MSW to a large pit, where the waste is mixed and bulky items
are removed. After removal of large, bulky items and noncombustibles, unprocessed waste is delivered by an
overhead crane to a feed hopper that conveys the waste into the combustion chamber. Nearly all modern
MB/WW facilities utilize reciprocating grates or roller grates to move the waste through the combustion
chamber. The grates typically include two or three separate sections where designated
6-26
-------
TABLE 6-16. SUMMARY OF GEOGRAPHICAL DISTRIBUTION
OF MWC FACILITIES LARGER THAN 35 Mg/d
State
AK
AL
AR
CA
CT
FL
GA
HI
ID
IL
IN
MA
MD
ME
MI
MN
MS
MT
NC
NH
NJ
NY
OH
OK
OR
PA
SC
TN
TX
UT
VA
WA
WI
Total
No. of MWC
facilities
2
1
4
3
6
13
1
1
1
1
1
9
4
4
5
12
1
1
3
3
6
12
2
2
2
7
2
2
3
1
6
4
4
129a
State MWC capacity
Mg/d (ton/a)
109(120)
627 (690)
257 (283)
2,324 (2,560)
5,489 (6,045)
15,480(17,048)
454 (500)
1,961 (2,160)
45 (50)
1,453(1,600)
2,145 (2,362)
9,770 (10,760)
4,821 (5,310)
1,816(2,000)
4,744 (5,225)
4,633 (5,102)
136(150)
65 (72)
657 (724)
755 (832)
5,286 (5,822)
9,584(10,555)
545 (600)
1,117(1,230)
613(675)
7,901 (8,702)
790 (870)
1,135(1,250)
177(195)
363 (400)
5,743 (6,325)
1,251 (1,378)
755(831)
93,000(102,400)
Percentage of total MWC
capacity in the United
States
<0.5
0.67
0.28
2.5
5.9
17
0.49
2.1
0.05
1.6
2.3
11
5.2
2.0
5.1
5.0
<0.5
<0.5
0.71
0.81
5.7
10
0.59
1.2
0.66
8.5
0.85
1.2
0.19
0.39
6.2
1.4
0.81
100
"There are a total of 129 MWC facilities which operate approximately
305 units.
Source: Reference 67.
6-27
-------
TABLE 6-17. COMPOSITION OF DISPOSED
RESIDENTIAL AND COMMERCIAL
WASTE (WEIGHT PERCENT)
Component Year, 1994
Paper and Paperboard
Yard Wastes
Food Wastes
Glass
Metals
Plastics
Wood
Textiles
Rubber and Leather
Miscellaneous
Totals
32.9
14.8
8.5
6.4
6.3
11.8
8.2
3.6
3.7
3.8
100.0
Source: Reference 68.
stages in the combustion process occur. On the initial grate section, referred to as the drying grate, the
moisture content of the waste is reduced prior to ignition. In the second grate section, the burning grate, the
majority of active burning takes place. The third grate section, referred to as the burnout or finishing grate, is
where remaining combustibles in the waste are burned. Bottom ash is discharged from the finishing grate
into a water-filled ash quench pit or ram discharger. From there, the moist ash is discharged to a conveyor
system and transported to an ash loading area or storage area prior to disposal. Because the waste bed is
exposed to fairly uniform high combustion temperatures, mercury and mercury compounds will be exhausted
as vapors with the combustion gases.
The MB/REF combustors are older facilities that comprise several designs. This type of combustor
is continuously fed and operates in an excess air mode with both underfire and overfire air provided. The
waste is moved on a traveling grate and is not mixed as it advances through the combustor. As a result, waste
burnout or complete combustion is inhibited by fuel bed thickness, and there is considerable potential for
unburned waste to be discharged into the bottom ash pit. Rocking and reciprocating grate systems mix and
aerate the waste bed as it advances through the combustion chamber, thereby improving contact between the
waste and combustion air and increasing the burnout of combustibles. The system generally discharges the
ash at the end of the grates to a water quench pit for collection and disposal in a landfill. The MB/REF
combustors have a refractory-lined combustion chamber and operate at relatively high excess air rates to
prevent excessive temperatures, which can result in refractory damage, slagging, fouling, and corrosion
problems.
Because of their operating characteristics, the tracking grate systems may have cool ash pockets in
which mercury and mercury compounds are not exposed to high temperature and are thereby retained in the
ash, rather than being exhausted with the combustion gas stream. Consequently, mercury and mercury
compounds may be emitted as fugitive emissions from ash handling. However, the combustion stack is the
primary source of mercury emissions. In the rocking and reciprocating grate systems, essentially all mercury
will be exhausted with the combustion gas.
Refuse-derived fuel combustors burn MSW that has been processed to varying degrees, from simple
removal of bulky and noncombustible items accompanied by shredding, to extensive processing to produce a
finely divided fuel suitable for co-firing in pulverized coal-fired boilers. Processing MSW to RDF generally
raises the heating value of the waste because many of the noncombustible items are removed.
A set of standards for classifying RDF types has been established by the American Society for
Testing and Materials (ASTM). The type of RDF used is dependent on the boiler design. Boilers that are
6-28
-------
designed to burn RDF as the primary fuel usually utilize spreader stokers and fire fluff RDF in a semi-
suspension mode. This mode of feeding is accomplished by using an air swept distributor, which allows a
portion of the feed to burn in suspension and the remainder to be burned out after falling on a horizontal
traveling grate. The number of RDF distributors in a single unit varies directly with unit capacity. The
distributors are normally adjustable so that the trajectory of the waste feed can be varied. Because the
traveling grate moves from the rear to the front of the furnace, distributor settings are adjusted so that most of
the waste lands on the rear two-thirds of the grate to allow more time for combustion to be completed on the
grate. Bottom ash drops into a water-filled quench chamber. Underfire air is normally preheated and
introduced beneath the grate by a single plenum. Overfire air is injected through rows of high pressure
nozzles, providing a zone for mixing and completion of the combustion process. Because essentially all of
the waste is exposed to high combustion temperatures on the grate, most of the mercury in the RDF will be
discharged with the combustion gas exhaust.
In a fluidized-bed combustor (FBC), fluff or pelletized RDF is combusted in a turbulent bed of
noncombustible material, such as limestone, sand, or silica. In its simplest form, the FBC consists of a
combustor vessel equipped with a gas distribution plate and an underfire air windbox at the bottom. The
combustion bed overlies the gas distribution plate. The RDF may be injected into or above the bed through
ports in the combustor wall. The combustor bed is suspended or "fluidized" through the introduction of
underfire air at a high pressure and flow rate. Overfire air is used to complete the combustion process.
Good mixing is inherent in the FBC design. Fluidized-bed combustors have uniform gas
temperatures and mass compositions in both the bed and in the upper region of the combustor. This
uniformity allows the FBC's to operate at lower excess air and temperature levels than conventional
combustion systems. Waste-fired FBC's typically operate at excess air levels between 30 and 100 percent
and at bed temperatures around 815°C (1500°F). At this temperature, most mercury and mercury
compounds will be volatilized and exhausted with the combustion gas stream as a vapor.
In terms of number of facilities, modular starved-(or controlled-) air (MOD/SA) combustors
represent a noteable segment of the existing MWC population. However, because of their small sizes, they
account for only a small percentage of the total capacity. The basic design of a MOD/SA combustor consists
of two separate combustion chambers, referred to as the "primary" and "secondary" chambers. Waste is
batch-fed intermittently to the primary chamber by a hydraulically activated ram. The charging bin is filled
by a front-end loader or by other mechanical systems. Waste is fed automatically on a set frequency, with
generally 6 to 10 minutes between charges.
Waste is moved through the primary combustion chamber by either hydraulic transfer rams or
reciprocating grates. Combustors using transfer rams have individual hearths upon which combustion takes
place. Grate systems generally include two separate grate sections. In either case, waste retention times in
the primary chamber are lengthy, lasting up to 12 hours. Bottom ash is usually discharged to a wet quench
pit.
The quantity of air introduced in the primary chamber defines the rate at which waste burns.
Combustion air is introduced in the primary chamber at substoichiometric levels, resulting in a flue gas rich in
unburned hydrocarbons. The combustion air flow rate to the primary chamber is controlled to maintain an
exhaust gas temperature set point [generally 650° to 980°C (1200° to 1800°F)], which corresponds to about
40 to 60 percent theoretical air. As the hot, fuel-rich flue gases flow to the secondary chamber, they are
mixed with excess air to complete the burning process. Because the temperature of the exhaust gases from
the primary chamber is above the autoignition point, completing combustion is simply a matter of introducing
air to the fuel-rich gases. The amount of air added to the secondary chamber is controlled to maintain a
desired flue gas exit temperature, typically 980° to 1200° (1800° to 2200°F). At these primary chamber and
secondary chamber temperatures, essentially all of the mercury contained in the waste is expected to be
emitted as a vapor from the secondary chamber with the combustion gas stream.
6.4.3 Emission Control Measures
Mercury emissions from MWC units are controlled to a limited extent by adsorbing the mercury
vapors from the combustion chamber onto the acid gas sorbent material and then removing the particle-phase
mercury with a high-efficiency PM control device. The PM control devices most frequently used in the
United States are ESP's and fabric filters. To achieve this mercury control, reducing flue gas temperature at
the inlet to the control device to 175°C (350°F) or less is beneficial.74 Typically, newer MWC systems use a
6-29
-------
combination of gas cooling and duct sorbent injection (DSI) or spray dryer (SD) systems upstream of the PM
device to reduce temperatures and provide a mechanism for acid gas control.
The information contained in Reference 74 suggests that these combined acid gas/PM systems can
achieve improved mercury control by injecting activated carbon or modified activated carbon into the flue gas
upstream from the DSI or SD unit. With activated carbon injection, mercury control is increased to
90 percent. The paragraphs below briefly describe the DSI and SD processes. Because the ESP's and FF's
used on MWC's are comparable to those used on other combustion systems, they are not described.
Spray drying in combination with either fabric filtration or an ESP is the most frequently used acid
gas control technology for MWC's in the United States. Spray dryer/fabric filter systems are more common
than SD/ESP systems and are used most on new, large MWC's. In the spray drying process, lime is slurried
and then injected into the SD through either rotary atomizer or dual-fluid nozzles. The key design and
operating parameters that significantly affect SD acid gas performance are the SD's outlet temperature and
lime-to-acid gas stoichiometric ratio. The SD outlet temperature, which affects mercury removal, is
controlled by the amount of water in the lime slurry.72
With DSI, powdered sorbent is pneumatically injected into either a separate reaction vessel or a
section of flue gas duct located downstream of the combustor economizer. Alkali in the sorbent (generally
calcium) reacts with HC1 and SO2 to form alkali salts (e.g., calcium chloride [CaCl2] and calcium sulfite
[CaSO3]). Some units also use humidification or other temperature control measures upstream from the
collection device. Reaction products, fly ash, and unreacted sorbent are collected with either an ESP or fabric
filter.72
Recent test programs using activated carbon injection have been conducted in the United States.
Recent test results have shown mercury removal efficiencies from 90 percent to over 95 percent with
activated carbon injection.67 Other test results show mercury reductions ranging from 50 to over 95 percent,
depending on the carbon feed rate, with typical outlet Hg concentrations of less than 50 //g/dscm.67'7 >74 As
a result of the emission standards developed for municipal waste combustors under section 129 of the Clean
Air Act Amendments, new (subpart Eb) and existing (subpart Cb), MWC's will typically operate with spray
dryer/fabric filter systems with activated carbon injection.
6.4.4 Emissions
The primary source of mercury emissions from municipal waste combustors is the combustion gas
exhaust stack. However, small amounts of mercury may be emitted as part of the fugitive PM emissions from
fly ash handling, particularly if highly efficient dry control systems are used.
A recent EPA report documenting 1995 estimates of the mercury emissions from municipal waste
combustors indicates that mercury emissions from MWC's decreased by 48 percent between 1990 and
1995.67 Estimated 1990 mercury emissions were 49 Mg (54 tons) and for 1995, emissions are estimated to
be 26 Mg (29 tons). This decrease in mercury emissions is attributed to retrofits of air pollution controls on
some MWC's, retirement of several existing MWC's, and significant reductions in uncontrolled mercury
emissions due to decreased levels of mercury in consumer products such as batteries. The inventory of
MWC's used to develop the 1995 estimates of mercury emissions is presented in Appendix B. Relative to the
1990 nationwide emissions of mercury from MWC's, a 92 percent reduction in mercury levels (to 4.0 Mg or
4.4 tons) is projected by about 2000 as a result of the section 129 emission standards (subpart Eb) and
guidelines (subpart Cb) for MWC's.67
A recent study conducted to update the municipal waste combustor section of AP-42 provided a
comprehensive review of the available MWC mercury emission data. The study found that most of the test
reports contained insufficient process data to generate emission factors. The authors of the municipal waste
combustion section of AP-42 concluded that the development of emission factors for MWC's, using only the
test reports which estimated feed rates, would eliminate data from so many facilities that the values derived
were not likely to be representative of the entire MWC population. In addition, the subjective nature of the
refuse feed rates called into question the validity of the limited data. Consequently, emission factors were
developed using the F-factor, which is the ratio of the gas volume of the products of combustion (e.g., flue
gas volume) to the heating value of the fuel. This approach, presented in EPA Method 19, requires an
F-factor and an estimate of the fuel heating value. For MWC's, the F-factor is 0.257 dscm/MJ (9,570
dscf/106 Btu) (at 0 percent Cv) of MSW fired. For all combustor types, except RDF combustors, a heating
value of 10,500 kJ/kg (4,500 Btu/lb) refuse was assumed. For RDF combustor units, the processed refuse
6-30
-------
has a higher heating value, and a heating value of 12,800 kJ/kg (5,500 Btu/lb) was assumed. Overall, these
data are representative of average values for MWC's.72 While this procedure does provide good average
emission factors that represent an industry cross section, the assumed F-factor and waste heating values
above may not be appropriate for specific facilities.
As mentioned earlier, the concentration of mercury in consumer products has declined since 1989.
As a result, the concentration of mercury in municipal solid has declined. The same methodology used to
develop the AP-42 emission factors was applied to the average mercury concentrations presented in
Reference 67. These average mercury concentrations and the resultant average emission factors are presented
in Table 6-18. While the procedure used to develop the emission factors presented in Table 6-18 does
provide good average emission factors that represent the industry cross section, the assumed F-factors and
waste heating values above may not be appropriate for specific facilities.
TABLE 6-18. AVERAGE EMISSION FACTORS FOR MUNICIPAL WASTE COMBUSTORS
Combustor type
Non-RDF without AG control
Non-RDF with AG control
Non-RDF with AG control and carbon
RDF without AG control
RDF with AG control
Mercury
concentration
ug/dscm @ 7% O2
340
205
19
260
35
Average emission factors
g/Mg waste
1.4
0.83
0.077
1.3
0.17
10'3lb/ton waste
2.8
1.7
0.15
2.6
0.34
AG = acid gas control (includes SD, DSI/FF, SD/ESP, DSI/ESP, SD/FF, and SD/ESP
configurations)
Non-RDF = Combustors that burn MSW (e.g, MB/WW, MB/RW, MOD/EA, MOD/SA)
RDF = Combustors that burn refused derived fuel
Total 1995 mercury emissions from municipal waste combustion are estimated to be 26 Mg (29 tons); see
Appendix A for details.
6.5 SEWAGE SLUDGE INCINERATORS
Currently, there are 166 active sewage sludge incinerators (SSI's) in the United States using one of
three technologies: multiple hearth, fluidized-bed, and electric infrared. Over 80 percent of the identified,
operating SSI's are multiple hearth units. About 15 percent of the SSI's are fluidized-bed combustors;
3 percent are electric infrared; and the remainder cofire sewage sludge with municipal solid waste.75
Most sewage sludge incineration facilities are located in the eastern United States, but a substantial
number are also located on the West Coast. New York has the largest number of SSI facilities with 33,
followed by Pennsylvania and Michigan with 21 and 19, respectively. About 785,000 Mg (865,000 tons) of
sewage sludge on a dry basis are estimated to be incinerated annually.76
The most recent data on the mercury content of sewage sludge obtained from the 1988 National
Sewage Sludge Survey show a mean mercury concentration of 5.2 ppmwt (parts per million by weight).77
Earlier data obtained in the mid 1970's indicate that mercury concentrations in municipal sewage sludge range
from 0.1 to 89 ppmwt with a mean value of 7 ppmwt and a median value of 4 ppmwt Other early data
collected by EPA from 42 municipal sewage treatment plants in the early 1970 s showed a range of 0.6 to
43 ppmwt, with a mean value of 4.9 ppmwt on a dry solids basis.78 The potential for the formation of
volatile organomercury compounds during the waste treatment process was considered. According to two
sources, no test data are available for emissions of organomercury compounds from this source.79^0 These
sources expect any level of formation would be very low.
The sections below provide SSI process descriptions, a discussion of control measures, and a
summary of mercury emission factors.
6-31
-------
6.5.1 Process Description
Figure 6-1 presents a simplified diagram of the sewage sludge incineration process, which involves
two primary steps. The first step in the process of sewage sludge incineration is the dewatering of the sludge.
Sludge is generally dewatered until it is about 20 to 35 percent solids. Systems using Thermal Conditioning
Processes regularly obtain dewatered sludge that contains in excess of 40 percent solids. When it is more
than 25 percent solids, the sludge will usually burn without auxiliary fuel. After dewatering, the sludge is
sent to the incinerator, and thermal oxidation occurs. The unburned residual ash is removed from the
incinerator, usually on a continuous basis, and is disposed in a landfill or reused (i.e., bricks, concrete,
asphalt, etc.). A portion of the noncombustible waste, as well as unburned volatile organic compounds, is
carried out of the combustor through entrainment in the exhaust gas stream. Air pollution control devices,
primarily wet scrubbers, are used to remove the entrained pollutants from the exhaust gas stream. The gas
stream is then exhausted, and the collected pollutants are sent back to the head of the wastewater treatment
plant in the scrubber effluent. As shown in Figure 6-1, the primary source of mercury emissions from the SSI
process is the combustion stack. Some fugitive emissions may be generated from ash handling, but the
quantities are expected to be small. Because mercury and mercury compounds are relatively volatile, most
mercury will leave the combustion chamber in the exhaust gas; concentrations in the ash residue are expected
to be negligible.
The paragraphs below briefly describe the three primary SSI processes used in the United States.75
The basic multiple hearth furnace is cylindrical in shape and is oriented vertically. The outer shell is
constructed of steel, lined with refractory, and surrounds a series of horizontal refractory hearths. A hollow
cast iron rotating shaft runs through the center of the hearths. Attached to the central shaft are the rabble
arms with teeth shaped to rake the sludge in a spiral motion, alternating in direction from the outside in, then
inside out, between hearths. Typically, the upper and lower hearths are fitted with four rabble arms, and the
middle hearths are fitted with two. Cooling air for the center shaft and rabble arms is introduced into the
shaft by a fan located at its base. Burners that provide auxiliary heat are located in the sidewalls of the
hearths.
In the majority of multiple hearth incinerators, dewatered sludge is fed directly onto the top hearth.
For a number of incinerators, the sludge is fed directly to a lower hearth. Typically, the rabble arms move the
sludge through the incinerator as the motion of the rabble arms rakes the sludge toward the center shaft,
where it drops through holes located at the center of the hearth. This process is repeated in all of the
subsequent hearths, with the sludge moving in opposite directions in adjacent hearths. The effect of the
rabble motion is to break up solid material to allow better surface contact with heat and oxygen.
Ambient air is first ducted through the central shaft and its associated rabble arms. The center shaft
cooling air exhaust is either sent back to a lower hearth or it is piped to the incinerator's exhaust stack for
"plume suppression". The combustion air flows upward through the drop holes in the hearths, countercurrent
to the flow of the sludge, before being exhausted from the top hearth.
Multiple hearth furnaces can be divided into three zones. The upper hearths comprise the drying
zone where most of the moisture in the sludge is evaporated. The temperature in the drying zone is typically
between 425° and 760°C (800° and 1400°F). Sludge combustion occurs in the middle hearths (second
zone) as the temperature is increased to a maximum of 925 °C (1700°F). When exposed to the temperatures
in both upper zones, most mercury will be volatilized and discharged as vapor in the exhaust gas. The third
zone, made up of the lowermost hearth(s), is the cooling zone. In this zone, the ash is cooled as its heat is
transferred to the incoming combustion air.
Fluidized-bed combustors (FBC's) are cylindrically shaped and vertically oriented. The outer shell is
constructed of steel and lined with refractory. Tuyeres (nozzles designed to deliver blasts of air) are located
at the base of the furnace within a refractory-lined grid. A bed of sand rests upon the grid. Dewatered sludge
is fed into the bed of the furnace. Air injected through the tuyeres, at pressures from
6-32
-------
GAS EXHAUST
SLUDGE
• FUGITIVE EMISSIONS
VENTURI WATER
• POTENTIAL SOURCES OF MERCURY EMISSIONS
Figure 6-1. Process flow diagram for sludge incineration.
20 to 35 kPa (3 to 5 psig), simultaneously fluidizes the bed of hot sand and the incoming sludge. Normally a
temperature of 677°(1250 °F), which is sufficient to vaporize most mercury contained in the sludge, is
maintained in the bed. As the sludge burns, fine ash particles and mercury vapor are carried out the top of the
furnace with the exhaust gas.
An electric infrared incinerator consists of a horizontally oriented, insulated furnace. A woven wire
belt conveyor extends the length of the furnace, and infrared heating elements are located in the roof above
the conveyor belt. Combustion air is preheated by the flue gases and injected into the discharge end of the
furnace. Electric infrared incinerators consist of a number of prefabricated modules that are linked together
to provide the required furnace length. The dewatered sludge cake is conveyed into one end of the incinerator.
An internal roller mechanism levels the sludge into a continuous layer approximately 2.5 centimeters (cm)
[1 inch (in.)] thick across the width of the belt. The sludge is sequentially dried and then burned as it moves
beneath the infrared heating elements. Ash is discharged into a hopper at the opposite end of the furnace.
The preheated combustion air enters the furnace above the ash hopper and is further heated by the outgoing
ash. The direction of air flow is countercurrent to the movement of the sludge along the conveyor.
In addition to the three technologies discussed above, other technologies have been used for
incineration of sewage sludge. Three of these processes are cyclonic reactors, rotary kilns, and wet oxidation
reactors; none of these processes find widespread usage in the United States.
6.5.2 Emission Control Measures
Most SSI's are equipped with some type of wet scrubbing system for PM control. Because these
systems provide gas cooling as well as PM removal, they can potentially provide some mercury control.
Limited data obtained on mercury removal efficiencies are presented in the emission factor discussion. The
paragraphs below briefly describe the wet scrubbing systems typically used on existing SSI's.75
6-33
-------
Wet scrubber controls on SSI's range from low pressure drop spray towers and wet cyclones to
higher pressure drop venturi scrubbers and venturi/impingement tray scrubber combinations. The most
widely used control device applied to a multiple hearth incinerator is the impingement tray scrubber. Older
units use the tray scrubber alone, while combination venturi/impingement tray scrubbers are widely applied to
newer multiple hearth incinerators and to fluidized-bed incinerators. Most electric incinerators and some
fluidized-bed incinerators use venturi scrubbers only.
In a typical combination venturi/impingement tray scrubber, hot gas exits the incinerator and enters
the precooling or quench section of the scrubber. Spray nozzles in the quench section cool the incoming gas,
and the quenched gas then enters the venturi section of the control device. Venturi water is usually pumped
into an inlet weir above the quencher. The venturi water enters the scrubber above the throat and floods the
throat completely. Most venturi sections come equipped with variable throats to allow the pressure drop to
be increased, thereby increasing PM efficiency. At the base of the flooded elbow, the gas stream passes
through a connecting duct to the base of the impingement tray tower. Gas velocity is further reduced upon
entry to the tower as the gas stream passes upward through the perforated impingement trays. Water usually
enters the trays from inlet ports on opposite sides and flows across the tray. As gas passes through each
perforation in the tray, it creates a jet that bubbles up the water and further entrains solid particles. At the top
of the tower is a mist eliminator to reduce the carryover of water droplets in the stack effluent gas.
6.5.3 Emissions
The primary source of mercury emissions from sewage sludge incineration is the combustion gas
exhaust stack. However, small quantities of mercury also may be emitted with the fugitive PM emissions
generated from bottom and fly ash handling operations.
As a part of the recent update of AP-42, data have been developed on mercury emissions from SSI's.
These data are summarized in Table 6-19.
TABLE 6-19. SUMMARY OF MERCURY EMISSION FACTORS
FOR SEWAGE SLUDGE INCINERATORS
Incinerator typea
MH
FB
Control statusb
CY
CY/VS
IS
VS/IS
VS/IS
Mercury emission factors
g/Mg dry sludge
2.3
1.6
0.97
0.005
0.03
10'3lb/ton dry sludge
4.6
3.2
1.9
0.01
0.06
Source: Reference 75.
aMH = multiple hearth; FB = fluidized bed.
CY = cyclone; VS = venturi scrubbers; IS = impingement scrubber.
The emission factors in Table 6-19 should be used cautiously in that available data suggest that both
mercury concentrations in sludge and control efficiencies vary widely. Mercury emissions from SSI's are
limited by a NESHAP to 3,200 grams per 24 hours for an entire facility. All SSI's are required to conduct
more frequent monitoring/testing if the facility emits 1,600 or more grams per 24 hours.
Total 1994 mercury emissions from sewage sludge incineration are estimated to be 0.86 Mg
(0.94 tons); see Appendix A for details.
6-34
-------
6.6 HAZARDOUS WASTE COMBUSTION
Based on a 1994 listing of hazardous waste incinerators, the EPA Office of Solid Waste estimates
that there are 190 permitted or interim status incinerators, 41 cement kilns, and 11 light-weight aggregate
kilns that burn hazardous waste in the United States. Of these facilities, the commercial operations burn
about 635,200 Mg (700,000 tons) of hazardous waste per year. The remaining facilities are onsite or captive
units and burn about 726,000 Mg (800,000 tons) per year.81'82 The incinerators generally utilize one of five
basic technologies depending upon the types of waste to be treated: liquid injection, gas or fume, fixed or
multiple hearth, rotary kiln, and fluidized bed. Of these, the liquid injection and rotary kiln are probably the
two most prevalent types of incinerators currently in use.81
Lightweight aggregate kilns process a wide variety of raw materials (such as clay, shale, or slate)
which, after thermal processing, can be combined with cement to form concrete products. Lightweight
aggregate concrete is produced either for structural purposes or for thermal insulation purposes. A
lightweight aggregate plant is typically composed of a quarry, a raw material preparation area, a kiln, a
cooler, and a product storage area. The material is taken from the quarry to the raw material preparation area
and from there is fed into the rotary kiln.
The sections below provide a description of the hazardous waste combustion process and types of
incinerators and light-weight aggregate kilns, a discussion of control measures, and a summary of mercury
emissions and factors. A discussion of the production of Portland cement, cement kiln control measures, and
mercury emission sources is presented in Section 7.1. The mercury emission estimates discussed in
Section 7.1 are for the use of nonhazardous waste fuel.
6.6.1 Process Description
6.6.1.1 Incinerators. In most processes, the waste to be treated is transported from a storage area to
the incinerator where thermal oxidation occurs. Solid wastes are typically transported in drums or similar
containers, and liquids or gases are piped from the storage area. Depending upon the type of incinerator and
the wastes to be treated, either solid or liquid wastes or a combination may be fed into the incinerator along
with an auxiliary (supplemental) fuel and combustion air. Unburned residual ash is removed from the
incinerator, usually on a continuous basis, and is disposed. A portion of the noncombustible waste, as well as
small amounts of unburned volatile organic material, are carried out of the primary incinerator chamber
through entrainment in the exhaust gas stream. For some units (e.g., rotary kilns), the exhaust gas passes
through a secondary combustion chamber (afterburner) before going to the air pollution control devices. Air
pollution control devices, typically wet scrubbers, fabric filters, or electrostatic precipitators, are used to
remove the entrained pollutants from the exhaust gas stream. The gas stream exits to the atmosphere through
a stack, and the pollutants collected by the control devices are disposed. Scrubber effluents from the control
devices are sent to wastewater treatment and solids from fabric filters typically are landfilled. Because of the
high temperature in the combustion chambers, the primary source of mercury emissions from hazardous
waste incineration is the stack; concentrations in the ash residue are expected to be very small. Some fugitive
emissions may be generated during ash handling but the quantities also are expected to be very small.
The five basic incinerator types used in the United States are discussed in the following paragraphs.
Two of the types, fixed (or multiple hearth) and fluidized bed, were described in the previous section on
sewage sludge incinerators (Section 6.5.1) and will not be repeated here. The only major difference between
their use for dewatered sewage sludge and hazardous waste may be differences in combustion temperatures;
otherwise the units are essentially the same. The three designs discussed below are liquid injection, gas or
fume, and rotary kilns.83
The liquid injection is one of the most common designs. In this design, a pumpable and atomizable
waste is delivered to the incinerator and passes through burners into the combustion chamber. Burners
consist of an atomizing nozzle and a turbulent mixing section where the waste is mixed with primary air. The
incineration chamber is rectangular or cylindrical in shape, lined with refractory, and oriented vertically or
horizontally. Vertically aligned chambers may be fired from either the top or bottom. Atomized waste is
combusted at temperatures ranging from 870° to 1200°C (1300° to 3000°F) and residence times from 0.5 to
3 seconds. If the heat content of the liquid waste is insufficient to maintain the required combustion
temperature, an auxiliary (supplemental) fuel is used. Following combustion, the exhaust gases pass through
air pollution control devices and exit a stack.83
6-35
-------
Gas or fume incinerators are very similar to liquid injection types except the treated waste is a gas or
volatilized material (fume) instead of an atomized liquid. These incinerators are the simplest type to design
and operate. Waste storage and pumping systems are designed for particular gas temperature/pressure
considerations. The incinerator combustion chamber, combustion temperatures, and exhaust gas control
devices are comparable to liquid injection systems.83
Rotary kiln incinerators generally are considered to be the most versatile and durable of the common
types of incinerators. Using a mix to maintain necessary heat content, rotary kilns can simultaneously treat
solid wastes, liquid organic wastes, and aqueous wastes. A rotary kiln is a refractory lined cylindrical steel
shell tilted on the horizontal axis. The shell is usually supported on two or more steel tracks (trundles), which
band the shell, and ride on rollers to allow the kiln to rotate around its horizontal axis. Waste material is
tumbled through the kiln by gravity as the kiln rotates at a rate of 1 to 2 revolutions per minute. The rate of
rotation and angle of tilt determine the solids residence time in the kiln. Rotary kiln diameters range from 1.2
to 4.9 meters (4 to 16 feet), and length-to-diameter ratios are typically 5:1. The kilns typically operate at
temperatures of 870° to 980°C (1600° to 1800°F).83
In rotary kilns, solid waste is fed through the nonrotating upper end of the kiln using an auger screw
or ram feeder. Pumpable wastes (e.g., sludges) can be fed through a water-cooled tube (wand) and liquid
organic wastes, aqueous wastes, and/or auxiliary fuel are injected through burner nozzles. Waste continues to
heat and burn as it travels down the inclined kilns. Combustion air is provided through ports on the face of
the kiln; the kiln usually operates at 50 to 200 percent excess air. At the end of the kiln, the residual ash
drops into an ash pit, is cooled, and removed for disposal. The exhaust gases, containing unburned
components, are routed to an afterburner (secondary combustion chamber) operating at about 1100° to
1400°C (2000° to 2500°F) and 100 to 200 percent excess air. Auxiliary fuel and/or pumpable liquid wastes
usually are used to maintain the afterburner temperature. The flue gases leave the afterburner, pass through
air pollution control devices, and exit to the atmosphere through a stack.83
6.6.1.2 Lightweight Aggregate Kilns. A rotary kiln consists of a long steel cylinder, lined internally
with refractory bricks, which is capable of rotating about its axis and is inclined at an angle of about
5 degrees to the horizontal. The length of the kiln depends in part upon the composition of the raw material
to be processed but is usually 30 to 60 meters (98 to 197 feet). The prepared raw material is fed into the kiln
at the higher end, while firing takes place at the lower end. The dry raw material fed into the kiln is initially
preheated by hot combustion gases. Once the material is preheated, it passes into a second furnace zone
where it melts to a semiplastic state and begins to generate gases which serve as the bloating or expanding
agent. In this zone, specific compounds begin to decompose and form gases such as SO2, CO,, SO3, and O2
that eventually trigger the desired bloating action within the material. As temperatures reach their maximum
(approximately 1150°C [2100°F]), the semiplastic raw material becomes viscous and entraps the expanding
gases. This bloating action produces small, unconnected gas cells, which remain in the material after it cools
and solidifies. The product exits the kiln and enters a section of the process where it is cooled with cold air
and then conveyed to the discharge.
Kiln operating parameters such as flame temperature, excess air, feed size, material flow, and speed
of rotation vary from plant to plant and are determined by the characteristics of the raw material. Maximum
temperature in the rotary kiln varies from about 1120°C to 1260°C (2050°F to 2300°F), depending on the
type of raw material being processed and its moisture content. Typical exit temperatures may range from
about 427° to 650°C (800° to 1200°F), again depending on the raw material and on the kiln's internal
design. Approximately 50 to 200 percent excess air is forced into the kiln to aid in expanding the raw
material.
6.6.2 Emission Control Measures
Incinerators are equipped with a wide variety of air pollution control devices (APCDs) that range in
complexity from no control to complex, state-of-the-art systems that provide control for several pollutants.
Units with no controls are limited to devices burning low ash and low chlorine content wastes. The hot flue
gases from the incinerators are cooled and purged of air pollutants before exiting through the stack to the
atmosphere. Cooling is done primarily by water quenching; water is atomized and sprayed directly into the
hot flue gases. The cooled gases then pass through various APCDs to control particular matter (PM), acid
gases, metals (including mercury), and organic components. Common APCDs for gaseous pollutant control
include packed towers, spray dryers, and dry scrubbers; of these, packed towers are the most common. For
PM control, venturi scrubbers, wet or dry electrostatic precipitators (ESPs), or fabric filters are common
controls. Activated carbon injection is being used at one facility for control of dioxins and mercury.83
6-36
-------
Lightweight aggregate kilns may use one or a combination of air pollution control devices, including
fabric filters, venturi scrubbers, cyclones and dry scrubbers. All of the facilities utilize fabric filters as the
main type of emissions control, although a spray dryer, venturi scrubber and dry scrubber may be used in
addition to a fabric filter.
One of the major factors in control of mercury emissions is reduction of the flue gas temperature.
Because wet scrubbing systems provide gas cooling as well as PM control, they can potentially provide some
degree of mercury removal. Wet APCD devices, such as packed towers, wet ESPs, and high pressure drop
venturi scrubbers, would be expected to show some degree of mercury control. Fabric filters would not be
expected to show significant mercury reduction because of the high flue gas temperature.
6.6.3 Emissions
The principal source of mercury emissions from hazardous waste incinerators and lightweight
aggregate kilns is the flue gas (combustion gas) exhaust stack. Small quantities of mercury compounds also
may be emitted with fugitive PM emissions generated from incinerator ash handling operations.
As a part of the EPA proposed revised standards for hazardous waste combustors, baseline national
emissions estimates were made for hazardous air pollutant (HAP) emissions, including mercury, from
hazardous waste incinerators.81 The baseline estimate entailed estimation of mercury emissions from the
78 hazardous waste incinerators in the EPA Office of Solid Waste (OSW) data base and then determination
of the number of facilities not represented by the OSW data base. For facilities contained in the OSW data
base, mercury average hourly emissions and stack gas flow rates (Ib/hr) were calculated for each incinerator
with emission measurements. Similar but untested units were assumed to have the same emission rate as
tested units. The total number of units not represented in the OSW data base was determined and multiplied
by the average mercury emission rate to obtain a total hourly mercury emission rate. Based on these data, an
average mercury baseline emission rate was calculated for incinerators. Using similar calculations, an
average mercury baseline emission rate for cement kilns and light-weight aggregate kilns was also calculated.
Details on the methodologies used to estimate the mercury emissions from hazardous waste incinerators,
cement kilns, and lightweight aggregate kilns may be obtained from docket materials prepared by the EPA
Office of Solid Waste for the proposed hazardous waste combustion MACT standards.
Total 1996 mercury emissions from hazardous waste combustion are estimated to be 6.3 Mg
(6.9 tons); see Appendix A for details.
6.7 MEDICAL WASTE INCINERATION
Medical waste includes infectious and noninfectious wastes generated by a variety of facilities
engaged in medical care, veterinary care, or research activities such as hospitals, clinics, doctors' and dentists'
offices, nursing homes, veterinary clinics and hospitals, medical laboratories, and medical and veterinary
schools and research units. Medical waste is defined by the EPA as "any solid waste which is generated in
the diagnosis, treatment, or immunization of human beings or animals, in research pertaining thereto, or in the
production or testing of biologicals." A medical waste incinerator (MWI) is any device that burns such
medical waste.84 Based on comments received following proposal of the new source performance standards
(NSPS) and emission guidelines (EG) for MWI's, EPA may elect to modify this definition by making it more
specific.85
In the 1993 mercury document, estimates developed by EPA suggested that about 3.06 million Mg
(3.36 million tons) of medical waste were produced annually in the United States.32 The EPA estimated that
approximately 5,000 MWI's located at hospitals, veterinary facilities nursing homes, laboratories, and other
miscellaneous facilities across the U.S. were used to treat this waste. Following proposal of the NSPS and
EG for new and existing MWI's, the EPA received new information regarding the number of MWI's operating
throughout the United States. More recent estimates developed for the MWI EG indicate that there are
approximately 2,400 MWI's operating in the United States. These 2,400 MWI's are used to treat
approximately 767 thousand Mg (846 thousand tons) of medical waste per year. The lower estimate of the
number of existing MWI's operating in the U.S. has led to a lower estimate of the mercury emissions
produced by MWI's. The EPA currently estimates that MWI's emit approximately 14.5 Mg (16.0 tons) of
mercury per year. However, the upcoming EG are expected to reduce mercury emissions by more than
90 percent.85
6-37
-------
Available information indicates that MWI systems can be significant sources of mercury emissions.
Mercury emissions result from mercury-bearing materials contained in the waste. Although concentrations of
specific metals in the waste have not been fully characterized, known mercury sources include batteries;
fluorescent lamps; high-intensity discharge lamps (mercury vapor, metal halide, and high-pressure sodium);
thermometers; special paper and film coatings; and pigments. Prior to 1991, batteries, primarily alkaline and
mercury-zinc batteries, were a major mercury source. Prior to 1991, the concentration of mercury in alkaline
batteries was about 1 percent and that of mercury-zinc batteries ranged from 35 to 50 percent mercury. In
1991, several battery manufacturers reduced the mercury content in alkaline batteries to less than
0.025 percent. Additionally, the use of zinc-air batteries as a replacement for the mercury-zinc batteries
became more prevalent. Alkaline batteries are general purpose batteries that are used in a variety of
equipment. Mercury-zinc batteries previously were used in transistorized equipment, hearing aids, watches,
calculators, computers, smoke detectors, tape recorders, regulated power supplies, radiation detection meters,
scientific equipment, pagers, oxygen and metal monitors, and portable electrocardiogram monitors.
Cadmium-mercury pigments are primarily used in plastics but also are used in paints, enamels, printing inks,
rubber, paper, and painted textiles.69'87 Hospital laboratory facilities use polyvinyl alcohol (PVA) fixatives
to preserve and examine stool specimens for internal parasites; these diagnostic tools contain mercuric
chloride and may be disposed in the MWI waste stream. All of these materials can be routed to an MWI,
thereby contributing to mercury emissions from this source category.
6.7.1 Process Description
Although the ultimate destination of almost all medical waste produced in the United States is a solid
waste landfill, the waste generally must be treated before it can be landfilled. The primary functions of MWI
facilities are to render the waste biologically innocuous and to reduce the volume and mass of solids that must
be landfilled by combusting the organic material contained in the waste. Over the years, a wide variety of
MWI system designs and operating practices have been used to accomplish these functions. To account for
these system differences, a number of MWI classification schemes have been used in past studies, including
classification by waste type (pathological, mixed medical waste, red bag waste, etc.), by operating mode
(continuous, intermittent, batch), and by combustor design (retort, fixed-hearth, pulsed-hearth, rotary kiln,
etc.). Some insight into MWI processes, emissions, and emissions control is provided by each of these
schemes. However, because the available evidence suggests that mercury emissions are affected primarily by
waste characteristics, the characterization and control of mercury emissions from MWI's can be discussed
without considering other MWI design and operating practices in detail. The paragraphs below provide a
generic MWI process description and identify potential sources of mercury emissions.
A schematic of a generic MWI system that identifies the major components of the system is shown in
Figure 6-2. As indicated in the schematic, most MWI's are multiple-chamber combustion systems that
comprise primary, secondary, and possibly tertiary chambers. The primary components of the MWI process
are the waste-charging system, the primary chamber, the ash handling system, the secondary chamber, and the
air pollution control system, which are discussed briefly below.
Medical waste is introduced to the primary chamber via the waste-charging system. The waste can
be charged either manually or mechanically. With manual charging, which is used only on batch and smaller
(generally older) intermittent units, the operator opens a charge door on the side of the primary chamber and
tosses bags or boxes of waste into the unit. When mechanical feed systems are employed, some type of
mechanical device is used to charge the waste to the incinerator. The most common mechanical feed system
is the hopper/ram assembly. In a mechanical hopper/ram feed system, the following steps occur: (1) waste is
placed into a charging hopper manually, and the hopper cover is closed; (2) a fire door isolating the hopper
from the incinerator opens; (3) the ram moves forward to push the waste into the incinerator; (4) the ram
reverses to a location behind the fire door; and (5) after the fire door closes, a water spray cools the ram, and
the ram retracts to the starting position. The system now is ready to accept another charge. The entire
hopper/ram charging sequence normally functions as a controlled, automatically-timed sequence to eliminate
overcharging. The sequence can be activated by the
6-38
-------
CONTROL AND
MONITORING
ON
WASTE
1
WASTE
CHARGE
SYSTEM
PRIMARY CHAMBER
TO
ATMOSPHERE
STACK
TO
ATMOSPHERE
STACK
ASH
REMOVAL
SYSTEM
SECONDARY
CHAMBER
WASTE
HEAT
BOILER
I (
AIR »
POLLUTION'
CONTROL *
SYSTEM
T
Figure 6-2. Major components of an incineration system.
-------
operator or for larger, fully automated incinerators, it may be activated at preset intervals by an automatic
timer.88'89
The potential for mercury emissions from the waste-charging systems is low. Mechanical systems
are generally operated with a double-door system to minimize fugitive emissions. Small quantities of fugitive
emissions may be generated while the chamber door is open during manual charging, but no data are available
on the magnitude of these emissions.
The primary chamber (sometimes called the "ignition" chamber) accepts the waste and begins the
combustion process. Most modern MWI's operate this chamber in a "controlled-air" mode to maintain
combustion air levels at or below stoichiometric requirements. The objectives of this controlled-air operation
are to provide a more uniform release of volatile organic materials to the secondary chamber and to minimize
entrainment of solids in these off-gases. Three processes occur in the primary chamber. First, the moisture in
the waste is volatilized. Second, the volatile fraction of the waste is vaporized, and the volatile gases are
directed to the secondary chamber. Third, the fixed carbon remaining in the waste is combusted.
The primary chamber generates two exhaust streams~the combustion gases that pass to the
secondary chamber and the solid ash stream that is discharged. Any metal compounds in the waste, including
mercury, are partitioned to these two streams in one of three ways. The metals may be retained in the primary
chamber bottom ash and discharged as solid waste; they may be entrained as PM in the combustion gases; or
they may be volatilized and discharged as a vapor with the combustion gases. Because mercury and mercury
compounds are generally quite volatile and because the primary chamber typically operates in the range of
650° to 820°C (1200° to 1500°F), most of the mercury in the waste stream will be exhausted as a vapor to
the secondary chamber.
The primary chamber bottom ash, which may contain small amounts of mercury or mercury
compounds, is discharged via an ash removal system and transported to a landfill for disposal. The ash
removal system may be either manual or mechanical. Typically, batch units and smaller intermittent units
employ manual ash removal. After the system has shut down and the ash has cooled, the operator uses a rake
or shovel to remove the ash and place it in a drum or dumpster. Some intermittent-duty MWI's and all
continuously operated MWI's use a mechanical ash removal system. The mechanical system includes three
major components: (1) a means of moving the ash to the end of the incinerator hearth—usually an ash transfer
ram or series of transfer rams, (2) a collection device or container for the ash as it is discharged from the
hearth, and (3) a transfer system to move the ash from the collection point. Generally, these automatic
systems are designed to minimize fugitive emissions. For example, one type of collection system uses an ash
bin sealed directly to the discharge chute or positioned within an air-sealed chamber below the hearth. A door
or gate that seals the chute is opened at regular intervals to allow the ash to drop into the collection bin.
When the bin is filled, the seal-gate is closed, and the bin is removed and replaced with an empty bin. In
another system, the ash is discharged into a water pit. The ash discharge chute is extended into the water pit
so that an air seal is maintained. The water bath quenches the ash as the ash is collected. A mechanical
device, either a rake or drag conveyor system, is used to intermittently or continuously remove the ash from
the quench pit. The excess water is allowed to drain from the ash as it is removed from the pit, and the wetted
ash is discharged into a collection container.
The potential for mercury emissions from both mechanical and manual ash discharge systems is
minimal. As described above, most mechanical systems have seals and provide ash wetting as described
above to minimize fugitive PM emissions. While manual systems can generate substantial fugitive PM, the
concentrations of mercury have generally been shown to be quite low.9 Consequently, fugitive mercury
emissions are negligible.
The primary function of the secondary chamber is to complete the combustion of the volatile organic
compounds that was initiated in the primary chamber. Because the temperatures in the secondary chamber
are typically 980°C (1800°F) or greater, essentially all of the mercury that enters the secondary chamber will
be exhausted as a vapor. The hot exhaust gases from the secondary chamber may pass through an energy
recovery device (waste heat boiler or air-to-air heat exchanger) and an air pollution control system before they
are discharged to the atmosphere through the combustion stack. This combustion stack is the major route of
mercury emissions from MWI's.
6-40
-------
6.7.2 Emission Control Measures
A number of air pollution control system configurations have been used to control PM and gaseous
emissions from the MWI combustion stacks. Most of these configurations fall within the general classes of
wet systems and dry systems. Wet systems typically comprise a wet scrubber designed for PM control
(venturi scrubber or rotary atomizing scrubber) in series with a packed-bed scrubber for acid gas removal and
a high-efficiency mist elimination system. Most dry systems use a fabric filter for PM removal, but ESP's
have been installed on some larger MWI's. These dry systems may use sorbent injection via either dry
injection or spray dryers upstream from the PM device to enhance acid gas control. Additionally, some
systems incorporate a combination dry/wet system that comprises a dry sorbent injection/fabric filter system
followed by a venturi scrubber. Because the systems described above are designed primarily for PM and acid
gas control, they have limitations relative to mercury control. However, recent EPA studies indicate that
sorbent injection/fabric filtration systems can achieve improved mercury control by adding activated carbon
to the sorbent material. The emission data presented in the section below provide information on the
performance of some of the more common systems.
6.7.3 Emissions
The primary source of emissions from medical waste incineration is the combustion gas exhaust
stack. However, small quantities of mercury may be contained in the fugitive PM emissions from ash
handling operations, particularly if the fly ash is collected in a dry air pollution control system with high
mercury removal efficiencies.
Over the past 8 years, mercury emissions have been measured at several MWI's through the EPA's
regulatory development program, MWI emission characterization studies conducted by the State of
California, and compliance tests conducted in response to State air toxic requirements. In the 1993 mercury
L&E document, mercury emission data were available from 20 MWI's. Of these, data from 14 of the
facilities were considered to be adequate for emission factor development.32 Since publication of the
previous document, an additional 27 emission test reports were obtained by EPA to be used in the
reassessment of the performance of add-on air pollution control devices (APCD's). These test reports were
reviewed by EPA's Emission Measurements Center (EMC) for completeness to determine if the test data was
suitable for use in the development of the maximum achievable control technology (MACT) standards and
guidelines for MWI's. The results of the EMC review are documented in a memorandum describing the
general selection rules for MWI and APCD emission test data.91
Emission factors for MWI's with combustion controls, wet scrubbers, fabric filter/packed bed
systems, and dry scrubbers (with and without activated carbon injection) were developed for the MWI
standards and guidelines. The MWI emission factors were developed by (1) developing exhaust gas flow
rate-to-waste burned factors (dscf/lb factors), (2) developing pollutant concentrations for each control
technology, and (3) calculating emission factors by multiplying together the results of the first two steps.
Approximately 26 emissions test reports, including those from EPA's emissions test program and test reports
reviewed by EMC with sufficient process data, were used to develop the dscf/lb factors. The average dscf/lb
factor for intermittent and continuous MWI's was determined to be 2.67 dscf/lb (at 7 percent O2). The
mercury emissions data from 19 emissions test reports (8 from EPA's emissions test program and 11
additional reports qualified by EMC) was used to determine the achievable emissions concentrations for
MWI's with combustion controls, wet scrubbers, and dry scrubbers.93'94'95 The emission factors for each
control technology were calculated by multiplying the average dscf/lb factor by the achievable mercury
concentration for each control technology.
Table 6-20 presents the MWI emission factors for each control technology developed by EPA for the
MWI NSPS and EG. The emission factors presented in Table 6-20 are average emission factors that
represent emissions from continuous and intermittent MWI's that burn a mixture of noninfectious waste and
infectious (red bag) waste. While the procedure used to calculate the MWI emission factors provides average
emission factors that represent the industry cross section, it should not be used to determine emission factors
for individual facilities. The dscf/lb factor presented above may not be appropriate for specific facilities due
to variations in auxiliary fuel usage and excess air ratios.
Total 1996 mercury emissions from medical waste incineration are estimated to be 14.5 Mg
(16 tons); see Appendix A for details.
6-41
-------
TABLE 6-20. MERCURY EMISSION FACTORS FOR MWI'S
Air pollution control
Combustion control
Wet scrubber
Dry scrubber w/o carbon
Dry scrubber w/ carbon
Fabric filter/packed bed
Mercury emission factor
g/Mg waste
37
1.3
37
1.7
1.3
10'3lb/ton waste
74
2.6
74
3.3
2.6
6-42
-------
7.0 EMISSIONS FROM MISCELLANEOUS SOURCES
Mercury has been found to be emitted from various miscellaneous sources including the following:
1. Portland cement manufacturing,
2. Lime manufacturing,
3. Carbon black production,
4. Byproduct coke production,
5. Primary lead smelting,
6. Primary copper smelting,
7. Petroleum refining,
8. Municipal solid waste landfills,
9. Geothermal power plants, and
10. Pulp and paper production.
Raw materials processed at the facilities listed above include minerals, ores, and crudes extracted
from the earth. Many of these raw materials contain mercury. At various stages of processing, the raw
materials are heated. Therefore, each of the manufacturing processes listed above may emit mercury during
various steps of raw materials processing. A summary of the estimated mercury emissions from each of the
above industries is as follows:
Emissions, Mg
Industry (tons)
Portland cement manufacture 4.0 (4.4)
Lime manufacture 0.1 (0.1)
Carbon black production 0.3(0.3)
Byproduct coke production 0.6 (0.7)
Primary lead smelting 0.1 (0.1)
Primary copper smelting 0.06 (0.07)
Municipal solid waste landfills 0.07 (0.08)
Geothermal power plants 1.3(1.4)
Pulp and paper production 1.6(1.8)
No emission estimate was developed for petroleum refining because the only emission factors were for
auxiliary processes not specifically associated with petroleum refining.
This section presents process information, air pollution control measures, and estimates of mercury
emissions for these sources.
7.1 PORTLAND CEMENT MANUFACTURING
More than 30 raw materials are used to manufacture portland cement. These materials can be
classified into four basic classes of raw materials: calcarious, siliceous, argillaceous, and ferriferous. Two
processes, the wet and dry processes, can be used to manufacture portland cement. In 1995, there was a total
of208U.S. cement kilns with a combined total clinker capacity of 76.3 x 106Mg(83.9x 106tons). Of this
total, six kilns with a combined capacity of 1.7 x 106 Me (1.9 x 106 tons) were inactive. The total number of
active kilns was 202 with a clinker capacity of 74.7 x 10 Mg (82.2 x 106 tons).96 The name, location, and
clinker capacity (in metric tons) of each kiln is presented in Appendix C. Based on 1995 U.S. cement kiln
capacity data, an estimated 72 percent of the portland cement is manufactured using the dry process, and the
remaining 28 percent based on the wet process. A description of the processes used to manufacture portland
cement and the emissions resulting from the various operations is presented below.97
7-1
-------
7.1.1 Process Description
Figure 7-1 presents a basic flow diagram of the portland cement manufacturing process. The process
can be divided into four major steps: raw material acquisition and handling, kiln feed preparation,
pyroprocessing, and finished cement grinding.97
The initial step in the production of portland cement manufacturing is raw materials acquisition.
Calcium, which is the element of highest concentration in portland cement, is obtained from a variety of
calcareous raw materials, including limestone, chalk, marl, sea shells, aragonite, and an impure limestone
known as "natural cement rock." The other raw materials—silicon, aluminum, and iron-are obtained from
ores and minerals, such as sand, shale, clay, and iron ore. Mercury is expected to be present in the ores and
minerals extracted from the earth. The only potential source of mercury emissions from raw material
acquisition would be due to wind blown mercury- containing particulate from the quarry operations. Mercury
emissions are expected to be negligible from these initial steps in portland cement production.
The second step involves preparation of the raw materials for pyroprocessing. Raw material
preparation includes a variety of blending and sizing operations designed to provide a feed with appropriate
chemical and physical properties. The raw material processing differs somewhat for wet- and dry-processes.
At facilities where the dry process is used, the moisture content in the raw material, which can range from less
than 1 percent to greater than 50 percent, is reduced to less than 1 percent. Mercury emissions can occur
during this drying process but are anticipated to be very low because the drying temperature is generally well
below the boiling point of mercury. However, some dryers do attain a temperature above the boiling point of
mercury, which would react in emissions. At some facilities, heat for drying is provided by the exhaust gases
from the pyroprocessor. At facilities where the wet process is used, water is added to the raw material during
the grinding step, thereby producing a pumpable slurry containing approximately 65 percent solids.
Pyroprocessing (thermal treatment) of the raw material is carried out in the kiln, which is the heart of
the portland cement manufacturing process. During pyroprocessing, the raw material is transformed into
clinkers, which are gray, glass-hard, spherically-shaped nodules that range from 0.32 to 5.1 cm (0.125 to 2.0
in.) in diameter. The chemical reactions and physical processes that take place during pyroprocessing are
quite complex. The sequence of events can be divided into four stages:
1. Evaporation of uncombined water from raw materials as material temperature increases to 100°C
(212°F),
2. Dehydration as the material temperature increases from 100°C to approximately 430°C (800°F)
to form the oxides of silicon, aluminum, and iron,
3. Calcination, during which carbon dioxide (CO2) is evolved, between 900°C (1650°F) and 982°C
(1800°F) to form calcium oxide,
4. Reaction of the oxides in the burning zone of the rotary kiln to form cement clinker at
temperatures about 1510°C (2750°F).
The rotary kiln is a long, cylindrical, slightly inclined, refractory-lined furnace. The raw material mix is
introduced into the kiln at the elevated end, and the combustion fuels are usually introduced into the kiln at
the lower end, in a countercurrent manner. The rotary motion of the kiln transports the raw material from the
elevated end to the lower end. Fuel such as coal or natural gas, or occasionally oil, is used to provide energy
for calcination. Mercury is present in coal and oil. Tables 6-4 and 6-11 presented data pertaining to mercury
content in coal and oil, respectively. Use of other fuels, such as chipped rubber, petroleum coke, and waste
solvents, is becoming increasingly popular. Combustion of fuel during the pyroprocessing step contributes to
potential mercury emissions. Mercury may also be present in the waste-derived fuel mentioned above.
Because mercury evaporates at approximately 350°C (660°F), most of the mercury present in the raw
materials can be expected to be volatilized during the pyroprocessing step.
7-2
-------
DRY PROCESS
WET PROCESS
Figure 7-1. Process flow diagram of portland cement manufacturing process5.7
-------
Pyroprocessing can be carried out using one of five different processes: wet process, semi-dry, dry
process, dry process with a preheater, and dry process with a preheater/precalciner. These processes
essentially accomplish the same physical and chemical steps described above. The last step in the pyro-
processing is the cooling of the clinker. This process step recoups up to 30 percent of the heat input to the
kiln system, locks in desirable product qualities by freezing mineralogy, and makes it possible to handle the
cooled clinker with conventional conveying equipment. Finally, after the cement clinker is cooled, a sequence
of blending and grinding operations is carried out to transform the clinker into finished portland cement.
7.1.2 Emission Control Measures97
The primary pollutants resulting from the manufacture of portland cement are PM and PM-10, NOX,
SO2, CO, and CO2. Emissions of metal compounds occur from the portland cement kilns and can be grouped
into three general classes: volatile metals (including mercury), semivolatile metals, and refractory or
nonvolatile metals. Although partitioning of the metals is affected by kiln operating conditions, the refractory
metals tend to concentrate in the clinker, the semivolatile metals tend to be discharged through the bypass
stack, and the volatile metals through the primary exhaust stack. The largest emission source of volatile
metals within the cement plant is the pyroprocessing system that includes the kiln and clinker cooler exhaust
stacks.
Process fugitive emission sources include materials handling and transfer, raw milling operations in
dry process facilities, and finish milling operations. Potential mercury emission sources are indicated in
Figure 7-1 by solid circles. Typically, particulate emissions from these processes are captured by a
ventilation system with a fabric filter. Because the dust from these units is returned to the process, they are
considered to be process units as well as air pollution control devices. The industry uses shaker, reverse air,
and pulse jet filters, as well as some cartridge units, but most newer facilities use pulse jet filters. For process
fugitive operations, the different systems are reported to achieve typical outlet PM loadings of 45 milligrams
per cubic meter mg/m3 (0.02 grains per actual cubic foot [gr/acfj). Because some fraction of the mercury is
in particle form, the performance of these systems relative to particulate mercury control is expected to be
equivalent to this overall particulate performance.
In the pyroprocessing units, PM emissions are controlled by fabric filters (reverse air, pulse jet, or
pulse plenum) and ESP's. Clinker cooler systems are controlled most frequently with pulse jet or pulse
plenum fabric filters. A few gravel bed (GB) filters have been used on clinker coolers.
The dust collected by the various fabric filters at the cement manufacturing facility is called cement
kiln dust (CKD). This dust is typically recycled into the process as a feed ingredient and substantially passes
through the cement kiln again, where a fraction of the residual mercury in the dust is volatilized. As dust is
continually recycled, essentially all of the mercury input to the process will eventually leave the system as a
vapor from the kiln stack. If the CKD is disposed, however, the particulate mercury remaining in the CKD
also goes to disposal, and only the mercury volatilized during the single pass through the cement kiln escapes
to the atmosphere as vapor from the kiln stack.
7.1.3 Emissions
The mercury emissions discussed in this section for the manufacture of portland cement are only for
the use of fossil fuels and nonhazardous waste auxiliary fuels. Mercury emissions from the use of hazardous
waste fuels were discussed in Section 6.6, Hazardous Waste Combustion.
The principal sources of mercury emissions are expected to be from the kiln and preheating/
precalcining steps. Negligible quantities of emissions would be expected in the raw material processing and
mixing steps because the only source of mercury would be fugitive dust containing naturally occurring
quantities of mercury compounds from the raw materials. Processing steps that occur after the calcining
process in the kiln would be expected to be a much smaller source of emissions than the kiln. Potential
mercury emission sources are denoted by solid circles in Figure 7-1. Emissions resulting from all processing
steps include particulate matter. Additionally, emissions from the pyroprocessing step include other products
of fuel combustion such as SO2, NOX, CO2, and CO. Carbon dioxide from the calcination of limestone will
also be present in the flue gas.
7-4
-------
Cement kiln test reports have been reviewed by EPA (and its contractor) in its development of the
Portland cement industry NESHAP, and by a private company. Test reports for Certification of Compliance
(COC) emissions tests (required of all kilns burning hazardous waste derived fuel) (WDF) and test reports
for facilities not burning hazardous waste were reviewed.98'99 The results from the Gossman study showed
an average emission factor of 0.65 x 10~4 kg/Mg of clinker (1.3 x 10~4 Ib/ton of clinker) for nonhazardous
waste fuels. The RTI study evaluated tests based on both nonhazardous waste fuel and hazardous waste fuel.
For the hazardous waste tests, the mercury emissions data were corrected to reflect only the mercury
emissions originating from the fossil fuel and raw material. The emissions data for nonhazardous waste and
the corrected hazardous waste were combined and showed an average mercury emission factor of 0.65 x 10"
4 kg/Mg of clinker (1.29 x 10'4 Ib/ton of clinker).
Total 1994 mercury emissions from this industry are estimated to be 4.0 Mg (4.4 tons); see
Appendix A for details.
7.2 LIME MANUFACTURING
Lime is produced in various forms, with the bulk of production yielding either hydrated lime or
quicklime. In 1994, producers sold or used 17.4 x 106 Mg (19.2 x 106 tons) of lime produced at 109 plants
in 33 States and Puerto Rico. The 1994 production represented a 3.6 percent increase over 1993 production.
In 1990, there were 113 lime production operations in the U.S. with a annual production of 15.8 x 106 Mg
(17.4 x 106 tons). The leading domestic uses for lime include steelmaking, pulp and paper manufacturing,
and treatment of water, sewage, and smokestack emissions.100
Appendix C provides a list of the active lime plants in the United States in 1991. The list includes
company headquarters' locations, plant locations by State, and the type of lime produced at each plant. The
geographical locations, by State, of the lime operations and quantities of lime produced are shown in
Table 7-1.
7.2.1 Process Description
Lime is produced by calcining (removal of CO2) limestone at high temperature. Limestone is
commonly found in most states but only a small portion can be used for lime production. To be classified as
limestone, the rock must contain 50 percent or more calcium carbonate. If the rock contains 30 to 45 percent
magnesium carbonate, it is called dolomite. The product of the calcining operation is quicklime; this material
can be hydrated with water to produce hydrated lime or slaked lime (Ca(OH)2). The product of calcining
dolomite is dolomitic quicklime; it also can be hydrated. Figure 7-2 presents a flow diagram for the lime
manufacturing process. Lime manufacturing is carried out in five major steps. These are:
1. Quarrying raw limestone,
2. Preparing the limestone for calcination,
3. Calcining the limestone,
4. Processing the lime by hydrating, and
5. Miscellaneous transfer, storage, and handling processes.
The manufacturing steps in lime production are very similar to that of the dry portland cement
process, which was discussed in the previous section. The most important process step with respect to
emissions of mercury and other air pollutants is the calcination. During calcination, kiln temperature may
reach 1820°C (3300°F). Approximately 90 percent of the lime produced in the United States is
manufactured by calcining limestone in a rotary kiln. Other types of lime kilns include the vertical or shaft
kiln, rotary hearth, and fluidized bed kilns. Fuel, such as coal, oil, petroleum coke, or natural gas, may be
used to provide energy for calcination. Petroleum coke is usually used in combination with coal; oil is rarely
used as a fuel source. Approximately one-third of the U.S. lime kilns are fired with natural gas. Auxiliary
fuels such as chipped rubber and waste solvents may potentially be used; at the present time, however, no
lime kilns use these auxiliary fuels.101
Mercury is expected to be present in very small quantities in the limestone and in coal and oil used as
fuel. Data pertaining to the mercury content in coal and oil are presented in Sections 6.1 and 6.2,
respectively. As with the production of portland cement, any mercury present in the raw materials can be
7-5
-------
TABLE 7-1. LIME PRODUCERS IN THE UNITED STATES IN 1994
State
Alabama
Arizona, Nevada, Utah
California
Colorado, Montana, Wyoming
Idaho, Oregon, Washington
Illinois, Indiana, Missouri
Iowa, Nebraska, South Dakota
Kentucky, Tennessee, West Virginia
Michigan
North Dakota
Ohio
Pennsylvania
Puerto Rico
Texas
Virginia
Wisconsin
Otherd
Total
No. of
plants
4
8
7
10
8
8
5
5
9
3
9
8
1
6
5
4
9
109
Lime production x 103 Mg (xlQ-3 tons)
Hydrateda
184 (203)
243 (268)
26 (29)
(--)
25 (28)
464 (511)
W (W)
132 (145)
26 (29)
(--)
W (W)
263 (290)
23 (25)
471 (519)
121 (133)
124 (137)
213 (235)
2,310 (2,546)
Quicklime?1
1,470 (1,620)
1,570 (1,730)
178 (196)
335 (369)
597 (658)
2,910 (3,207)
W (W)
1,800 (1,984)
611 (673)
108 (119)
W (W)
1,330 (1,466)
<0.5 (<0.6)
740 (815)
621 (684)
383 (422)
2,430 (2,678)
15,100 (16,640)
Total*
1,660 (1,829)
1,810 (1,995)
203 (224)
335 (369)
622 (685)
3,380 (3,725)
(242) (267)te
1,930 (2,127)
637 (702)
108 (119)
(1,850) (2,039)c
1,590 (1,752)
23 (25)
1,210 (1,333)
742 (818)
507 (559)
2,640 (2,909)
17,400 (19,175)
ON
Source: Reference 100.
fMetric ton data rounded by the U.S.G.S. to three significant digits; may not add to totals shown.
Withheld to avpid disclosing company proprietary data; included in "Other" category.
°Total included in total for "Other category.
Includes Arkansas, Louisiana, Massachusetts, Minnesota, Oklahoma, and data indicated by "W".
-------
HIGH CALCIUM AND
DOLOMITIC LIMESTONE
I I 6
§ § jg
§ 8 1
i i i
LIMESTONE PRODUCTS
QUARRY AND NINE OPERATIONS
(DRILLING, BLASTING, AND
CONVEYING BROKEN LIIVESTONE)
i
RAW MATERIAL STORAGE
PRIMARY CRUSHER
SCREENING AND CLASSIFICATION
0.64-6.4 cm LIMESTONE
FOR VERTICAL KILNS
SECONDARY CRUSHING
LI i vies iunc
FOR ROTARY—^ CALCINATION
PULVERIZING
SCREENING AND CLASSIFICATION
_ PULVERIZED
STONE
MAX. SIZE 0.64 -1.3 cm
GROUND AND PULVERIZED QUICKLIME
\
HIGH CALCIUM AND DOLOMITIC
NORMAL HYDRATED LIME
INSPECTION
1
QUICKLIME
i
|- AND^LOMmCONLY 1 CRUSHING AN° PULVERIZING |
WATER
I
i
PEBBLE AND LUMP QUICKLIME
DOLOMITIC
QUICKLIME
ONLY
i
WATER
AND/OR STEAIU
j
PRESSURE HYDRATOR
I
DOLOMITIC PRESSURE
HYDRA TED LIME
Figure 7-2. Process flow diagram for lime manufacturing process.101
7-7
-------
expected to be emitted in the lime kiln. Combustion of fuel in the lime kiln is the major contributor to
mercury emissions.
7.2.2 Emission Control Measures
With the exception of the lime kiln, the emission sources in the lime manufacturing industry can be
classified as either process emissions or fugitive emissions. The primary pollutants resulting from these
fugitive sources are PM. No specific control measures for the lime industry are reported in the literature for
the fugitive sources. The reduction measures used for fugitive dust sources at portland cement manufacturing
facilities may also be applicable at lime manufacturing industries.
Air pollution control devices for lime kilns are primarily used to recover product or control fugitive
dust and PM emissions. Calcination kiln exhaust is typically routed to a cyclone for product recovery, and
then routed through a fabric filter or ESP's to collect fine particulate emissions. Other emission controls
found at lime kilns include wet scrubbers (typically venturi scrubbers). How well these various air pollution
control devices perform, relative to vapor phase mercury emissions in lime production, is not well
documented. The control efficiencies are expected to be similar to those observed in the production of
portland cement because of the similarities in the process and control devices.
7.2.3 Emissions
Mercury emissions from fuel combustion will occur from the lime kiln (calcination) as shown in
Figure 7-2. Mercury present in the limestone will also be emitted from the kiln. All other potential emission
sources in the process are expected to be very minor contributors to overall mercury emissions. Emissions
resulting from all five processing steps include particulate matter. Additionally, emissions from the lime kiln
include other products of fuel combustion such as SO2, NOX, CO, and CO2.
Lime kiln test reports are available for two facilities in the United States and one in Canada. The
source test reports for the U. S. facilities have been reviewed by EPA ( and its contractor) in its development
of the lime manufacturing industry NESHAP.102 The test report for the Canadian facility was provided by
the National Lime Association and reviewed as a part of this document.103 At the Canadian facility, two
different kilns were tested; one was a coal/coke-fired rotary kiln and the other was a natural gas-fired vertical
kiln. For the coal/coke-fired rotary kiln, the results from the tests showed an average mercury emission factor
of 9.0 x 10~6 kg/Mg of lime produced (1.8 x 10~5 Ib/ton of lime produced); the emission factors ranged from
0.8 x 10'5 to 1.0 x 10'5 kg/Mg of lime produced (1.6 x 10'5 to 2.0 x 10'5 Ib/ton of lime produced) over the
four test runs. For the natural gas-fired vertical kiln, the results showed an average mercury emission factor
of 1.5 x 10"6 kg/Mg of lime produced (3.0 x 10"6 Ib/ton of lime produced); the emission factors ranged from
1.45 x 10'6 to 1.6 x 10'6 kg/Mg of lime produced (2.9 x 10'6 to 3.2 x 10'6 Ib/ton of lime produced) over the
four test runs. Process data from the tests at the Canadian facility were used to calculate the quantity of
limestone fed required to produce 0.91 Mg (1.0 ton) of lime. Based on process data for the rotary kiln, the
average ratio of limestone feed to lime produced was 0.50 (i.e., 2 tons of limestone are required to produce 1
ton of lime). The average ratio for the vertical kiln was calculated to be 0.51.
The test results from the two U. S. facilities were evaluated for EPA by its contractor, TRI. Both of
the facilities, APG Lime Company and Eastern Ridge Lime company, employed coal-fired rotary kilns. The
results of the tests at APG showed an average mercury emission factor of 1.9 x 10"6 kg/Mg of limestone feed
(3.8 x 10"6 Ib/ton of limestone feed). Based on the 2:1 limestone feed to lime produced ratio, this
corresponds to an emission factor of 3.8 x 10"6 kg/Mg of lime produced (7.6 x 10"6 Ib/ton of lime produced).
At Eastern Ridge, the results showed an average mercury emission factor of 4.7 x 10"6 kg/Mg of limestone
feed (9.4 x 10 Ib/ton of limestone feed). Using the 2:1 conversion ratio, this corresponds to a mercury
emission factor of 9.4 x 10"6 kg/Mg of lime produced (1.9 x 10"5 Ib/ton of lime produced). The average
mercury emission factors for the coal-fired rotary kilns from the one Canadian facility and the two U. S.
facilities were combined and showed an overall average mercury emission factor of 7.4 x 10"6 kg/Mg of lime
produced (1.5 x 10 Ib/ton of lime produced).
The total 1994 mercury emissions from this industry are estimated to be 0.10 Mg (0.10 tons); see
Appendix A for details.
7-8
-------
7.3 CARBON BLACK PRODUCTION
Carbon black is produced by pyrolizing petrochemical oil feedstock. A compilation of facilities,
locations, types of processes, and annual capacities is presented in Table 7-2. A description of the process
used to manufacture carbon black and the emissions resulting from the various operations is presented below.
7.3.1 Process Description
Carbon black is produced by partial combustion of hydrocarbons. The most common production
process (which accounts for more than 98 percent of carbon black produced) is based on a feedstock
consisting of a highly aromatic petrochemical or carbo chemical heavy oil. Mercury can be expected to be
present in the feedstock. Although the mercury content in the feedstock used to manufacture carbon black is
not known, mercury content in petroleum crude is reported to range between 0.023 and 30 ppm by weight.105
Figure 7-3 contains a flow diagram of this process.
Three primary raw materials are used in this process: preheated feedstock (either the petrochemical
oil or carbochemical oil), which is preheated to a temperature between 150 and 250°C (302 and 482°F);
preheated air; and an auxiliary fuel such as natural gas. The preheated oil and air are introduced into a
furnace, or reactor, that is fired with the auxiliary fuel. A turbulent, high-temperature zone is created in the
reactor by combusting the auxiliary fuel, and the preheated oil feedstock is introduced in this zone as an
atomized spray. In this zone of the reactor, most of the oxygen is used to burn the auxiliary fuel, resulting in
insufficient oxygen to combust the oil feedstock. Thus, pyrolysis (partial combustion) of the feedstock is
achieved, and carbon black is produced. Most of the mercury in the feedstock is emitted in the hot exhaust
gas from the reactor.
The product stream from the reactor is quenched with water, and any residual heat in the product
stream is used to preheat the oil feedstock and combustion air before recovering the carbon in a fabric filter.
Carbon recovered in the fabric filter is in a fluffy form. The fluffy carbon black may be ground in a grinder,
if desired. Depending on the end use, carbon black may be shipped in a fluffy form or in the form of pellets.
Pelletizing is done by a wet process in which carbon black is mixed with water along with a binder and fed
into a pelletizer. The pellets are subsequently dried and bagged prior to shipping.
7.3.2 Emission Control Measures
High-performance fabric filters are used to control PM emissions from main process streams during
the manufacture of carbon black.104 It is reported that the fabric filters can reduce PM emissions to levels as
low as 6 mg/m3 (normal m3). Mercury emissions from the reactor are primarily in the vapor phase. These
emissions will proceed through the main process streams to the fabric filters. If the mercury remains in the
vapor phase, the mercury control efficiency of the fabric filters is expected to be low. If the product gas
stream is cooled to below 170°C (325 °F), the fabric filter may capture a significant fraction of the condensed
mercury, thus providing a high degree of emission control.
7.3.3 Emissions
The processing unit with the greatest potential to emit mercury is the reactor. Mercury emission
sources are indicated in Figure 7-3 by solid circles. Mercury, which is present in the oil feedstock, can
potentially be emitted during the pyrolysis step. However, no data are available on the performance of the
fabric filter control systems for mercury emissions. The only available data are for emissions from the
oil-furnace process. These data show mercury emissions of 0.15 g/Mg (3 x 10"4 Ib/ton) from the main
process vent.106 The source of these data could not be obtained to verify the validity of the emission factors.
Because the factors are not verified, they should be used with extreme caution.
Total 1995 mercury emissions from this industry are estimated to be 0.25 Mg (0.28 tons); see
Appendix A for details.
7.4 BYPRODUCT COKE PRODUCTION
Byproduct coke, also called metallurgical coke, is a primary feedstock for the integrated iron and
steel industry. Byproduct coke is so named because it is produced as a byproduct when coal is heated in an
7-9
-------
TABLE 7-2. CARBON BLACK PRODUCTION FACILITIES
Company
Cabot Corporation
North American Carbon Black
Division
Chevron Chemical Company
Olevins and Derivatives Division
Columbian Chemicals Company
Continental Carbon Company
Degussa Corporation
Pigment Group
Ebonex Corporation
Engineered Carbons, Inc.
General Carbon Company
Hoover Color Corporation
Sir Richardson Carbon Company
Location
Franklin, Louisiana
Pampa, Texas
Villa Platte, Louisiana
Waverly, West Virginia
Cedar Bayou, Texas
El Dorado, Arkansas
Moundsville, West
Virginia
North Bend, Louisiana
Ulysses, Kansas
Phenix City, Alabama
Ponca City, Oklahoma
Sunray, Texas
Aransas Pass, Texas
Belpre, Ohio
New Iberia, Louisiana
Melvindale, Michigan
Baytown, Texas
Borger, Texas
Orange, Texas
Los Angeles, California
Hiwassee, Virginia
Addis, Louisiana
Big Spring, Texas
Borger, Texas
Type of
process3
F
F
F
F
A
F
F
F
F
F
F
F
F
F
F
C
F
FandT
F
C
C
F
F
F
TOTAL
Annual capacityb
103
161
29
100
91
9
57
88
100
36
36
120
59
54
54
109
4
86
102
61
0.5
0.5
120
54
129
1,660
106lb
355
65
220
200
20
125
195
220
80
80
265
130
120
120
240
8
190
225
135
1
1
265
120
285
3,665
Source: Reference 11.
aA = acetylene decomposition
C = combustion
F = furnace
T = thermal
bCapacities are variable and based on SRI estimates as of January 1, 1996.
7-10
-------
Main Unit
Bag Filter
T
_\
-
_
-
-
Burner
Fuel
Vent to
Vacuum -
Cleanup
Water
ATMOSPHERIC EMISSIONS
Tall Gas
-Vent
-Flare
-Steam/Power Generation
-Dryer Fuel
-Feedstock Heating
•=Mercury
emission
source
Pneumatic Conveying Gas
Pneumatic System Vent
(Optional Closed-Loop Recycle)
Dryer Purge Gas
*• Indirect Heating
Exhaust Gas
•Dust Bag Filter
To Storage
Tall Gas
or Fuel
Figure 7-3. Process flow diagram for carbon black manufacturing process1.04
04
-------
oxygen-free atmosphere (coked) to remove the volatile components of the coal. The material remaining is a
carbon mass called coke. The volatile components are refined to produce clean coke-oven gas, tar, sulfur,
ammonium sulfate, and light oil.107 Table 7-3 contains a list of byproduct coke oven facilities reported to be
in operation in 1991.108 A description of the process used to manufacture byproduct coke and the emissions
resulting from the various operations is presented below.
7.4.1 Process Description
Coke is currently produced in two types of coke oven batteries: the slot oven byproduct battery and
the nonrecovery battery. The slot oven byproduct type is the most commonly used battery; over 99 percent of
coke produced in 1990 was produced in this type of battery. The nonrecovery battery, as the name suggests,
is one where the products of distillation are not recovered and are immediately combusted to provide energy
within the plant. The nonrecovery battery is currently used at only one location; however, it is expected to be
a more popular choice when existing plants are reconstructed. Figures 7-4 and 7-5 present the general layout
and the emission points of a typical byproduct coke oven battery.
The byproduct coke oven battery consists of 20 to 100 adjacent ovens with common side walls that
are made of high quality silica and other types of refractory brick. Typically, the individual slot ovens are 11
to 16.8 m (36 to 55 ft) long, 0.35 to 0.5 m (1.1 to 1.6 ft) wide, and 3.0 to 6.7 m (9.8 to 22 ft) high. The wall
separating adjacent ovens, as well as each end wall, is made of a series of heating flues.107 Depending on the
dimensions, the production capacity may range from 6.8 to 35 Mg (7.5 to 39 tons) of coke per batch.
Pulverized coal, which is the feedstock, is fed through ports located on the top of each oven by a car
(referred to as a larry car) that travels on tracks along the top of each battery. After the oven is charged with
coal, the ports are sealed ("luted") with a wet clay mixture, and gaseous fuel (usually clean coke oven gas) is
combusted in the flues located between the ovens to provide the energy for the pyrolysis).107
The operation of each oven is cyclic, but the battery contains a sufficiently large number of ovens to
produce an essentially continuous flow of raw coke oven gas. The individual ovens are charged and emptied
at approximately equal time intervals during the coking cycle. The coking process takes between 15 and 30
hours, at the end of which almost all the volatile matter from the coal is driven off, thus forming coke. The
coking time is determined by the coal mixture, moisture content, rate of underfiring, and the desired
properties of the coke. When demand for coke is low, coking times can be extended to 24 to 48 hours.
Coking temperatures generally range from 900° to 1,000°C (1,650° to 2,000°F). The gases that evolve
during the thermal distillation are removed through the offtake system and sent to the byproduct plant for
recovery.107
At the end of the coking cycle, doors on both ends of the oven are removed and the incandescent coke
is pushed from the oven by a ram that is extended from the pusher machine. The coke is pushed through a
coke guide into a special railroad car called a quench car. The quench car carries the coke to a quench tower
where it is deluged with water to prevent the coke from burning after exposure to air. The quenched coke is
discharged onto an inclined coke wharf to allow the excess water to drain and to cool the coke to a reasonable
handling temperature. The coke is then crushed and screened to the proper size for the blast furnace
operation.107
The mercury content in coal was presented in Section 6.1. Table 6-4 presented data pertaining to
mercury levels in various types of U.S. coals. Depending on the type of coal used, the mercury content can be
as high as 8 ppm by weight; however, values of about 1 ppm are more typical. Consequently, the gases that
evolve from the coking operation are likely to contain mercury.110
7.4.2 Emission Control Measures
Emissions from charging coal into the ovens are controlled by stage charging in which coal is
discharged from the larry car hoppers in an ordered sequence that maintains an open tunnel head at the top of
the oven to provide an exit space for the gas until the last hopper is emptied. An important aspect of stage
charging is adequate aspiration, which is used to pull the gas generated during charging from the ovens into
the regular gas handling equipment.107
7-12
-------
TABLE 7-3. BYPRODUCT COKE PRODUCERS IN THE UNITED STATES IN 1991
Facility
Acme Steel, Chicago, IL
Armco, Inc., Ashland, KY
Armco, Inc., Middleton, OH
Bethlehem Steel, Bethlehem, PA
Bethlehem Steel, Burns Harbor, IN
Bethlehem Steel, Lackawanna, NY
Bethlehem Steel, Sparrows Point, MD
Geneva Steel, Orem, UT
Gulf States Steel, Gadsden, AL
Inland Steel, East Chicago, IN
LTV Steel, Pittsburgh, PA
LTV Steel, Chicago, IL
LTV Steel, Cleveland, OH
LTV Steel, Warren, OH
National Steel, Granite City, IL
National Steel, Ecorse, MI
USS, Div. of USX Corp., Clairton, PM
USS, Div. of USX Corp., Gary, IN
Wheeling-Pittsburgh Steel, East Steubenville,
WV
Total
No. of
batteries
2
2
3
3
2
2
3
1
2
6
5
1
2
1
2
1
12
6
4
58
Total No. of
ovens
100
146
203
284
164
152
210
208
130
446
315
60
126
85
90
78
816
422
224
4,259
Total capacity,
tons/d
1,600
2,700
4,535
3,944
4,380
1,872
4,069
2,250
2,800
5,775
5,404
1,600
3,200
1,500
1,520
925
12,640
7,135
3,800
71,649
Source: Reference 108.
7-13
-------
Door Emissions
(SCC 3-03-003-08)
4 i
Charge Lid Emissions
(SCC 3-03-003-14) Combustion
Quenching V Stack Emissions
Emission \ (SCC 3-03-003-06)
(SCC 3-03-003-04) \ i
A '
Quench Tower
Combustion
(Underfire) Stack
Charging
Emissions
(SCC 3-03-003-02)
Collector Main
(SCC 3-03-003-15)
Sandpipe Caps
(SCC 3-03-003-14)
Pushing Emissions
(SCC 3-03-003-03)
4
Offtake
' j Emissions
.''(SCC 3-03-003-14)
Coal Conveyor
.(SCC 3-03-003-09)
Quench Car
Emission stream
Figure 7-4. Schematic of byproduct coke oven battery.
,107
-------
TYPES OF AIR POLLUTION EMISSIONS
FROM COKE OVEN BATTERIES
© Pushing emissions
(2) Charging emissions
(3) Door emissions
(?) Topside emissions
(§) Battery underfire emissions
W//////7///////////////////////////S////'.
Figure 7-5. Types of air pollution emissions from coke oven batteries.109
7-15
-------
During the coking cycle, pollutants are emitted from leaks on the battery, including leaks from doors,
from lids that cover the charging ports, and from the offtake system. Because the oven is maintained under a
positive pressure, these leaks occur from small openings, such as gaps where metal seals mate against some
other part of the oven. Small gaps seal by the condensation of tar. Door leaks on most batteries are
controlled by repairing and maintaining doors, door seals, and jambs to prevent large gaps between the metal
seal and the jamb. The manual application of a supplemental sealant such as sodium silicate is used at some
plants to further reduce door leaks. A few batteries control door leaks by the external application of a luting
material to provide a seal (called hand-luted doors). Lid leaks and offtake leaks are controlled by applying
luting material around sealing edges to stop leaks and reluting when leaks are observed. The control of leaks
requires a diligent work practice program that includes locating leaks and then identifying and correcting their
cause.107
Pushing coke into the quench car is a significant source of PM emissions. Most facilities control
pushing emissions by using mobile scrubber cars with hoods, shed enclosures evacuated to a gas cleaning
device, or traveling hoods with a fixed duct leading to a stationary gas cleaner. Emission control devices used
to control emissions from quenching include ESP's, fabric filters, and wet scrubbers. These control devices
are effective primarily for PM control. No data are available on the performance of these systems for control
of mercury emissions. However, because these devices typically operate at elevated temperatures (>170°C
[>325 °F]), mercury removal is anticipated to be limited.
Fugitive particulate matter emissions are generated from material handling operations such as
unloading, storing, and grinding of coal; as well as screening, crushing, storing, and loading of coke. These
coal and coke handling PM emissions may be controlled by the use of cyclones.107
7.4.3 Emissions
Mercury emissions can be generated in small quantities during coal preparation and handling as
fugitive particulate matter because mercury is present as a trace contaminant in coal. Mercury also may be
volatilized and released during charging and pushing operations. During the coking cycle, mercury may be
volatilized and released to the atmosphere through poorly sealed doors, charge lids, and offtake caps, and
through cracks which may develop in oven brickwork, the offtakes, and collector mains.
There are no mercury emission data for byproduct coke ovens in the U.S. However, emission factors
used in Germany for coke production range from 0.01 to 0.03 g/Mg (2 x 10 to 6 x 10 Ib/ton) of coke
produced.33 It is important to note that U.S. coke producers use a high quality cleaned coal while their
European counterparts do not. If it is assumed that the coal cleaning process results in a 20 percent reduction
in mercury emissions (see Section 6.1.4.11 then the resultant U.S. mercury emission factor for coke
production would be 0.025 g/Mg (5 x 10"5 Ib/ton).
Total 1991 mercury emissions from this industry are estimated to be 0.59 Mg (0.65 tons); see
Appendix A for details.
7.5 PRIMARY LEAD SMELTING
Lead is recovered from a sulfide ore, primarily galena (lead sulfide-PbS), which also contains small
amounts of copper, iron, zinc, and other trace elements such as mercury. In 1994, the production of refined
primary lead from domestic ores and base bullion was 328,000 Mg (361,500 tons), which represents an
increase of about 6 percent over 1993 production.111 A list of primary lead smelters currently in operation
within the United States is given in Table 7-4.1 *1 A description of the process used to manufacture lead and
the emissions resulting from the various operations are presented below.
7-16
-------
TABLE 7-4. DOMESTIC PRIMARY LEAD SMELTERS AND REFINERIES
Smelter
ASARCO, East Helena, MT
ASARCO, Glover, MO
Doe Run, Herculaneum, MO
Refinery
ASARCO, Omaha, NEa
Same site
Same site
Estimated refinery capacity,
Mg (tons)
60,000(66,100)
125,000 (137,800)
200,000 (220,400)
Source: Reference 111.
aClosed permanently for lead refining as of May 31, 1996. There is limited refinery capacity at East
Helena, MT.
7.5.1 Process Description
Lead ores are concentrated at or near the mine and shipped to the smelter as the ore concentrate.
Figure 7-6 contains a process flow diagram of primary lead smelting. The recovery of lead from the lead ore
consists of three main steps: sintering, reduction, and refining.112
Sintering occurs in a sintering machine, which is essentially a large oven containing a continuous
steel pallet conveyor belt. Each pallet consists of perforated grates, beneath which are wind boxes connected
to fans to provide excess air through the moving sinter charge. The sintering reactions take place at about
1000°C (1832°F) during which lead sulfide is oxidized to lead oxide. The gas from the front end of the
sintering machine, containing 2.5 to 5 percent SO2, is vented to gas cleaning equipment before being sent to a
sulfuric acid plant. Gases from the rear of the sinter machine are recirculated through the moving grate and
then, typically, vented to a baghouse. The desulfurized sinter roast is crushed and then transported to the
blast furnace in charge cars. Since mercury and its compounds vaporize below the sintering temperature,
most of the mercury present in the ore can be expected to be emitted during sintering either as elemental
mercury or as mercuric oxide.
Reduction of the sintered lead is carried out in a blast furnace at a temperature of 1600°C (2920°F).
The furnace is charged with a mixture of sinter (80 to 90 percent of charge), metallurgical coke (8 to
14 percent of charge), and other materials, such as limestone, silica, litharge, and other constituents, which
are balanced to form a fluid slag. In the blast furnace, the charge descends through the furnace shaft into the
smelting zone, where it becomes molten, and then into a series of settlers that allow the slag to separate from
the lead. The slag is cooled and sent to storage; the molten lead, about 85 percent pure, is transported in pots
to the dross area for refining. Any residual mercury remaining in the roast from the sintering operation is
expected to be released during this reduction process.
The dressing area consists of a variety of interconnected kettles heated by natural gas. The lead pots
from the blast furnace are poured into receiving kettles and cooled. The copper dross rises to the top and can
be skimmed off for processing. The remaining lead dross is transferred to a finishing kettle where other
materials are added to facilitate further separation of impurities. In the finishing kettles, the lead dross
bullion settles to the bottom, is removed, and sent to the refinery. The matte and speiss rise to the top, are
removed, and sent to copper smelters.
Further refining of the lead bullion is carried out in cast iron kettles. Refined lead, which is 99.99 to
99.999 percent pure, is cast into pigs for shipment.
7.5.2 Emission Control Measures
Emission controls on lead smelter operations are employed for controlling PM and SO, emissions
resulting from the blast furnace and sintering machines. Centrifugal collectors (cyclones) may be used in
conjunction with fabric filters or ESP's for PM control. The blast furnace and the sintering machine operate
at very high temperatures (in excess of 1000°C [1832°F]); as a result, mercury is emitted from these sources
in vapor form. Therefore, particulate control devices would have little effect on mercury
7-17
-------
Concentrated ore
Limestone
Silica
Sinter recycle
Flue dust
Coke
C
Aso2
Dust and Fume
A SO2 A Dust and Fume AFume
1 Dust and Fume T T
Sinter Sinter ^
Machine A ^
Limest
Silica
Sinter
Lead c
Coke
E
Fu
Slag
one
recycle
xide
,,ast Bullion Dross.na crossing p Refined
rnace ^ — Area Bu||ion — ' Lead —
I
Ammonium chloride3
Soda ash
Sulfur . .
Flue dust T
Coke Matte and Speiss
t
Slag fuming
oal furnacea.b
Dump
Denotes potential
mercury emission source
Zinc oxidea'b
aNot at ASARCO, East Helena, Montana.
bNot at ASARCO, Glover, Missouri.
Figure 7-6. Typical primary lead processing scheme!12
-------
emissions from the sintering machine and blast furnace. However, no collection efficiency data are available
for mercury using these systems.
Control of SO2 emissions is achieved by absorption to form sulfuric acid in the sulfuric acid plants,
which are commonly part of lead smelting plants.
7.5.3 Emissions
Mercury, which may be present in the ore, may be emitted during the sintering and blast furnace
steps and, to a lesser extent, in the dressing area because these processes take place at high temperatures.
None of the primary lead smelters crush the ore at the smelter; all ore crushing and concentration is
performed at or near the mine. The smelters receive the ore concentrate for processing. Mercury emission
sources are indicated on Figure 7-6 by solid circles.
The available emission factor data for mercury emissions from primary lead smelting are for a
custom smelter operated by ASARCO in El Paso, Texas; this facility ceased operating in 1985. No recent
mercury emission factors are available for the three current primary lead smelters. The custom smelter in El
Paso obtained lead ore from several sources both within and outside the United States. These ores had a
variable mercury content depending upon the source of the ore. Two of the three current smelters are not
custom smelters; they typically process ore from the vicinity of the smelter. The two smelters in Missouri use
ore only from southeast Missouri; these ores have a very low mercury content. These two smelters combined
have over 84 percent of the total U.S. smelter capacity. The ASARCO-East Helena plant, although a custom
smelter, generally also processes low mercury concentrates. None of the three primary lead smelters reported
mercury emission data in the 1994 TRI.
Because the El Paso facility data were based on ores with a variable mercury content, and the current
major sources of lead ore have a very low mercury content, use of those emission factors will lead to an
overestimation of current emissions. A better estimating method would be to use the actual mercury content
of the ore and estimate emissions based on those data. The major domestic source of lead ore concentrate is
from the southeast Missouri area near the Glover and Herculaneum smelters. Data on mercury content in lead
concentrates from this area indicate the mercury concentration to be less than 0.2 parts per million (ppm).113
Using a mercury concentration of 0.2 ppm and particulate matter (PM) emission factors from the EPA AP-42
section on primary lead smelting, the upper limit for mercury emissions is estimated to be 0.10 Mg
(0.11 tons); see Appendix A for estimation procedures.112
7.6 PRIMARY COPPER SMELTING
Copper is recovered from a sulfide ore principally by pyrometallurgical smelting methods. Copper
ores contain small quantities of arsenic, cadmium, lead, antimony, and other heavy metals including mercury.
Data pertaining to mercury content in the ore are not available.
A list of primary copper smelters currently in operation within the U.S. is given in Table 7-5. In
1995, the total U.S. capacity for the eight primary copper smelters was 1,413,000 Mg (1,557,000 tons). The
total capacity in 1996 decreased to 1,354,000 Mg (1,492,000 tons) due to the closure of the Copper Range
smelter.114 In 1996, there were 19 refineries for primary copper processing; five used an electrolytic process
and 14 used an electrowinning process. In addition, there were seven refineries for secondary copper
processing.114 Since the mercury levels in the copper from the smelter would be very low, the mercury
emissions from the refining process are expected to be very small. Therefore, no tabulation of the individual
refineries is presented.
A description of the process used to manufacture copper and the emissions resulting from the various
operations is presented below.115
7.6.1 Process Description
The pyrometallurgical copper smelting process is illustrated in Figure 7-7.115 The traditionally used
process includes roasting of ore concentrates to produce calcine, smelting of roasted (calcine feed) or
unroasted (green feed) ore concentrates to produce matte, and converting of the matte to yield blister
7-19
-------
TABLE 7-5. U.S. PRIMARY COPPER SMELTERS
Company
ASARCO, Inc.
Copper Range Companya
Cypruss Climax Metals
Company
Kennecott
Magma Copper Company
Phelps Dodge Corp.
Location
El Paso, TX
Hayden, AZ
White Pine, MI
Globe, AZ
Garfield, UT
San Manuel, AZ
Hildalgo,NM
Hurley, NM
Process
ConTop flash furnace
Inco flash
Reverberatory
Isasmelt/Electric
Outokumpu
Outokumpu flash
Outokumpu flash
Inco flash
Total
Capacity, 103 Mg (103 tons)
1995
91 (100)
172(190)
68 (75)
163(180)
256 (282)
309 (340)
200 (220)
154(170)
1,413(1,557)
1996
100(110)
172(190)
0(0)
163(180)
256 (282)
309 (340)
200 (220)
154(170)
1,354(1,492)
Source: Reference 114.
aClosed in February 1995.
7-20
-------
Ore Concentrates with Silica Fluxes
0
O)
JO
CO
O
•e
0)
O
O
Fuel
Air
ROASTING3
OR DRYING b
OFF GAS
i
Fuel
Air
FLASH
SMELTING
to Dump
5% CA\\
MATTE (-
OFF GAS
Air
CONVERTING
OFF GAS
Natural or Reformulated Gas
Green Poles or Logs
Fuel
Air
Slag to Converter
Blister Copper (98.5% Cu)
i
FIRE REFINING
OFF GAS
Anode Copper (99.5% Cu)
To Electrolytic Refinery
• Denotes potential mercury emission source
aFirst step in the traditionally used copper-smelting process.
"First step in the currently used copper-smelting process.
Figure 7-7. Typical primary copper smelter process.115
7-21
-------
copper product (about 99 percent pure). Typically, the blister copper is fire refined in an anode furnace, cast
into "anodes" and sent to an electrolytic refinery for further impurity elimination. Copper smelters currently
process ore concentrates by drying them in fluidized bed dryers and then converting and refining the dried
product in the same manner as the traditionally used process.115
In roasting, charge material of copper concentrate mixed with a siliceous flux (often a low grade ore)
is heated in air to about 650°C (1200°F), eliminating 20 to 50 percent of the sulfur as SO,. Portions of such
impurities as antimony, arsenic, and lead are driven off, and some iron is converted to oxide. The roasted
product, calcine, serves as a dried and heated charge for the smelting furnace. Either multiple hearth or
fluidized bed roasters are used for roasting copper concentrate. Multiple hearth roasters accept moist
concentrate, whereas fluid bed roasters are fed finely ground material (60 percent minus 200 mesh). With
both of these types, the roasting is autogenous. Because there is less air dilution, higher SO2 concentrations
are present in fluidized bed roaster gases than in multiple hearth roaster gases. Because mercury has a boiling
point of 350°C (660°F), most of the mercury in the ore may be emitted during roasting.
In the smelting process, either hot calcines from the roaster or raw unroasted or dried concentrate is
melted with siliceous flux in a flash smelting furnace to produce copper matte, a molten mixture of cuprous
sulfide (Cu2S), ferrous sulfide (FeS), and some heavy metals. The required heat comes from partial oxidation
of the sulfide charge and from burning external fuel. Most of the iron and some of the impurities in the
charge oxidize with the fluxes to form a slag atop the molten bath, which is periodically removed and
discarded. Copper matte remains in the furnace until tapped. Mattes produced by the domestic industry
range from 35 to 65 percent copper, with 45 percent the most common. The copper content percentage is
referred to as the matte grade. Currently, four smelting furnace technologies are used in the United States:
electric, ConTop (flash) Outokumpu (flash), and Inco (flash). There are no reverberatory furnaces currently
in operation in the U.S. Flash furnaces may operate at temperatures as high as 1200 to 1300°C (2190 to
2370°F). Even though the exact temperatures at which the other furnace technology (electric) operates is not
known, it is probable that it operates at temperatures higher than the boiling point of mercury. Therefore, any
residual mercury that remains in the calcine may be emitted during the smelting step.
For smelting in electric arc furnaces, heat is generated by the flow of an electric current in carbon
electrodes lowered through the furnace roof and submerged in the slag layer of the molten bath. The feed
generally consists of dried concentrates or calcines, and charging wet concentrates is avoided. The matte is
periodically tapped and the slag is skimmed at frequent intervals. Electric furnaces do not produce fuel
combustion gases, so effluent gas flow rates are low and SO2 concentrations are high.
Flash furnace smelting combines the operations of roasting and smelting to produce a high grade
copper matte from concentrates and flux. In flash smelting, dried ore concentrates and finely ground fluxes
are injected, together with oxygen, preheated air, or a mixture of both, into a furnace of special design, where
temperature is maintained at approximately 1200 to 1300°C (2190 to 2370°F). Most flash furnaces, in
contrast to reverberatory and electric furnaces, use the heat generated from partial oxidation of their sulfide
charge to provide much or all of the energy (heat) required for smelting. They also produce offgas streams
containing high concentrations of SO^. Other flash furnaces, such as ConTop cyclone reactors, use oxyfuel
combustion to generate the heat required for oxidation.
Slag produced by flash furnace operations typically contains higher amounts of copper than does that
from electric furnace operations. As a result, the flash furnace and converter slags are treated in a slag
cleaning furnace to recover the copper (not conducted at the ASARCO, Hayden facility). Slag cleaning
furnaces usually are small electric furnaces. The flash furnace and converter slags are charged to a slag
cleaning furnace and are allowed to settle under reducing conditions, with the addition of coke or iron sulfide.
The copper, which is in oxide form in the slag, is converted to copper sulfide, is subsequently removed from
the furnace and is charged to a converter with regular matte. If the slag's copper content is low, the slag is
discarded.
The final step in the production of blister copper is converting, with the purposes of eliminating the
remaining iron and sulfur present in the matte and leaving molten "blister" copper. All but one U. S. smelter
uses Fierce-Smith converters, which are refractory lined cylindrical steel shells mounted on trunnions at either
end, and rotated about the major axis for charging and pouring. An opening in the center of the converter
functions as a mouth through which molten matte, siliceous flux, and scrap copper are charged and gaseous
products are vented. Air or oxygen-rich air is blown through the molten matte. Iron sulfide (FeS) is oxidized
to iron oxide (FeO) and SO,, and the FeO blowing and slag skimming are repeated until an adequate amount
of relatively pure Cu2S, called "white metal", accumulates in the bottom of the converter. A renewed air blast
7-22
-------
oxidizes the copper sulfide to SO2, leaving blister copper in the converter. The blister copper is subsequently
removed and transferred to refining facilities. This segment of converter operation is termed the finish blow.
The SO2 produced throughout the operation is vented to pollution control devices.
One domestic smelter uses Hoboken converters. The Hoboken converter is essentially like a
conventional Fierce-Smith converter, except that this vessel is fitted with a side flue at one end shaped as an
inverted U. This flue arrangement permits siphoning of gases from the interior of the converter directly to the
offgas collection system, leaving the converter mouth under a slight vacuum. The Hoboken converters are
also equipped with secondary hoods to further control emissions.
Blister copper usually contains from 98.5 to 99.5 percent pure copper. Impurities may include gold,
silver, antimony, arsenic, bismuth, iron, lead, nickel, selenium, sulfur, tellurium, and zinc. To purify blister
copper further, fire refining and electrolytic refining are used. In fire refining, blister copper is placed in an
anode furnace, a flux is usually added, and air is blown through the molten mixture to oxidize remaining
impurities, which are removed as a slag. The remaining metal bath is subjected to a reducing atmosphere to
reconvert cuprous oxide to copper. Temperature in the furnace is around 1100°C (2010°F). The fire-refined
copper is cast into anodes. Further refining separates the copper from impurities by electrolysis in a solution
containing copper sulfate and sulfuric acid. Metallic impurities precipitate from the solution and form a
sludge that is removed and treated to recover precious metals. Copper is dissolved from the anode and
deposited at the cathode. Cathode copper is remelted and cast into bars, rods, ingots, or slabs for marketing
purposes. The copper produced is 99.95 to 99.97 percent pure. Any mercury emissions during the refining
step will only be minimal.
7.6.2 Emission Control Measures
Emission controls on copper smelters are employed for controlling PM and SO2 emissions resulting
from roasters, smelting furnaces, and converters. Electrostatic precipitators are the most common PM control
devices employed at copper smelting facilities. Control of SO2 emissions is achieved by absorption to
sulfuric acid in the sulfuric acid plants, which are commonly part of all copper smelting plants.
7.6.3 Emissions
The main source of mercury will be during the roasting step and in the smelting furnace. Converters
and refining furnaces may emit any residual mercury left in the calcine. These sources are denoted by solid
circles in Figure 7-7.
In 1993, the Emission Standards Division of the U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, issued an information collection request to all eight of the primary
copper smelters operating at that time for data on mercury emissions. The reported values represented
annualized mercury emissions, in pounds per year (Ib/yr), from both stack and fugitive emission points at
each smelter. With the exclusion of Copper Range, which is closed, the self-reported values for mercury
emissions in 1993 ranged from 0 to 35 Ib/yr and the total for all smelters was 55.0 kg/yr (121.2 Ib/yr) or
0.055 Mg (0.061 tons/yr).116 In 1994, smelter production from domestic and foreign ores increased about
3.15 percent over 1993 production.114
Total 1994 mercury emissions from this industry are estimated to be 0.057 Mg (0.063 tons); see
Appendix A for details.
7.7 PETROLEUM REFINING
[This section is a condensation of the petroleum refining section in the 1993 Mercury L&E document;
except for plant location information, no new data have been added.]
Petroleum refining involves the conversion of crude petroleum oil into refined products, including
liquified petroleum gas, gasoline, kerosene, aviation fuel, diesel fuel, fuel oils, lubricating oils, and feedstocks
for the petroleum industry.
As of January 1995, there were 34 oil companies in the United States with operable atmospheric
crude oil distillation capacities in excess of 100,000 barrels per calendar day. These oil companies operated
refineries at a total of 107 different locations. In addition, there were 53 companies with distillation
7-23
-------
capacities of less than 100,000 barrels per calendar day. A listing of all companies, specific refinery
locations, and distillation capacities is presented in Appendix D.
Mercury is reported to be present in petroleum crude, and its content in petroleum crude is reported
to range between 0.023 and 30 ppm by weight.105 A description of the processes used in petroleum refining
and emissions resulting from the various operations is presented below.
7.7.1 Process Description
Petroleum refining is a very complex and highly integrated process. This process description
represents a general petroleum refining operation and highlights only the common process components.
Actual processes may vary among refineries depending upon the specific products produced. The operations
at petroleum refineries are classified into five general categories, as listed below:1 ;119
1. Separation processes,
2. Petroleum conversion processes,
3. Petroleum treating processes,
4. Feedstock and product handling, and
5. Auxiliary facilities.
7.7.1.1 Separation Processes. Constituents of crude oil include paraffmic, naphthenic, and aromatic
hydrocarbon compounds; impurities include sulfur, nitrogen, and metals. Three separation processes used to
separate these constituents include: atmospheric distillation, vacuum distillation, and recovery of light ends
(gas processing).
Atmospheric distillation results in the formation of bottoms consisting of high-boiling-point
hydrocarbons. Topped crude withdrawn from the bottoms of atmospheric distillation can be separated further
by vacuum distillation.
In vacuum distillation, the topped crude is heated in a process heater to temperatures ranging from
370° to 425 °C (700° to 800 °F) and subsequently flashed in a multi-tray vacuum distillation column,
operating at vacuums ranging from 350 to 1,400 kg/m2 (0.5 to 2.0 psia). Standard petroleum fractions
withdrawn from the vacuum distillation include lube distillates, vacuum oil, asphalt stocks, and residual oils.
Distillation is carried out at temperatures higher than the boiling point of mercury and can be expected to be
the primary source of mercury emissions.
7.7.1.2 Conversion Processes. Conversion processes include (1) cracking, coking, and visbreaking,
which break large molecules into smaller molecules; (2) isomerization and reforming processes to rearrange
the structures of molecules; and (3) polymerization and alkylation to combine small molecules into larger
ones. Residual mercury from the separation processes is probably emitted during the conversion processes.
Catalytic cracking uses heat, pressure, and catalysts to convert heavy oils into lighter products.
Feedstocks are usually gas oils from atmospheric distillation, vacuum distillation, coking, and deasphalting
processes, with a boiling range of 340° to 540 °C (650° to 1000°F). Two types of cracking units, the
fluidized catalytic cracking (FCC) unit and the moving-bed catalytic cracking unit, are used in the refineries.
Visbreaking is a thermal cracking process used to reduce the viscosity of the topped crude or vacuum
distillation residues. The feedstock is heated and thermally cracked at a temperature ranging between 455 °
and 480°C (850° and 900°F) and pressure ranging between 3.5 and 17.6 kg/cm2 (50 and 250 psia). The
cracked products are quenched with gas oil and flashed into a fractionator. The vapor overhead from the
fractionator is separated into light distillate products. A heavy distillate is recovered from the fractionator
liquid.
Coking is also a thermal cracking process used to convert low value residual fuel oil to higher value
gas oil and petroleum coke. It is carried out at high temperature and low pressure, and the resulting products
include petroleum coke, gas oils, and lighter petroleum stocks.
The conversion steps, cracking, coking, and visbreaking, described above can be expected to be
secondary sources of mercury emissions.
7-24
-------
Equipment commonly used during conversion includes process heaters and reformers. Process
heaters are used to raise the temperature of petroleum feedstocks to a maximum of 510°C (950°F). Fuels
burned include refinery gas, natural gas, residual fuel oils, or combinations. Reformers are reactors where the
heat for the reaction is supplied by burning fuel.
7.7.1.3 Treatment Processes. Petroleum treatment processes include hydrodesulfurization,
hydrotreating, chemical sweetening, acid gas removal, and deasphalting. These treatment methods are used to
stabilize and upgrade petroleum products. Removal of undesirable elements, such as sulfur, nitrogen, and
oxygen, is accomplished by hydrodesulfurization, hydrotreating, chemical sweetening, and acid gas removal.
Deasphalting is carried out to separate asphaltic and resinous materials from petroleum products.
Hydrotreating is a process in which the oil feed is treated by mixing with hydrogen in a fixed-bed catalyst
reactor. Removal of acid gas involves controlling emissions of SO2. Elemental sulfur is recovered as a
byproduct.
Any residual mercury left over in the feedstock after the separation and conversion steps can be
expected to be emitted during the treatment step.
7.7.1.4 Feedstock and Product Handling. This includes storage, blending, loading, and unloading of
petroleum crude and products. No mercury emissions are expected during these steps.
7.7.1.5 Auxiliary Facilities. Auxiliary facilities include boilers, gas turbines, wastewater treatment
facilities, hydrogen plants, cooling towers, and sulfur recovery units. Boilers and gas turbines cogeneration
units within petroleum refineries may burn refinery gas.
7.7.2 Emission Control Measures
Control of VOC (and in some instances, CO) emissions from distillation, catalytic cracking, coking,
blowdown system, sweetening, and asphalt blowing is achieved by flares. In some instances, the VOC-laden
gas stream is also used as fuel in process heaters.1
Control of PM emissions from catalytic cracking is achieved by using cyclones in conjunction with
ESP's.
7.7.3 Emissions
Emissions of mercury can be expected during the process steps where petroleum crude is processed
at high temperatures, such as the distillation, cracking, visbreaking, and other conversion steps. Other
emissions from petroleum refining operations include mainly PM, VOC, and products of fuel combustion.118
An emission factor for uncontrolled emissions from the fluid coking unit in the conversion step was cited in
SPECIATE but the source of these data could not be obtained in order to verify the validity of the emission
factors so the factor is not cited. The only other available data pertain to emissions from process heaters and
reformers. Based on a series of emission tests carried out in California, emission estimates for mercury were
presented for refinery gas-fired process heaters, boilers, gas turbine cogeneration units, and asphalt fume
incinerators. 12° These sources are auxiliary equipment that use fuels from a combination of sources. These
data are not representative of mercury emissions from the refining process.
7.8 MUNICIPAL SOLID WASTE LANDFILLS
A municipal solid waste (MSW) landfill is a discrete area of land or an excavation that receives
household waste and is not a land application unit, surface impoundment, injection well, or waste pile. A
MSW landfill may also receive other types of wastes, such as commercial solid waste, nonhazardous sludge,
and industrial solid waste.121
Municipal solid waste management in the United States is dominated by disposal in landfills. In
1994, approximately 60 percent of municipal solid waste was landfilled, 16 percent was incinerated, and
24 percent was recycled or composted.68 There were an estimated 3,600 active MSW landfills in the United
States in 1996.122 In 1994, active landfills received an estimated 115 million megagrams (Mg) (127 million
tons) of MSW.68
7-25
-------
7.8.1 Process Description
The three major design methods for MSW landfills are the area method, the trench method, and the
ramp method. These methods use a three-step process that consists of spreading the waste, compacting the
waste, and covering the waste with soil. The trench and ramp methods are not commonly used, particularly
when liners and leachate collection systems are used. The area fill method involves placing waste on the
ground surface or landfill liner, spreading it in layers, and compacting with heavy equipment. The trench
method entails excavating trenches designed to receive a day's worth of waste. The soil from the excavation
is often used for cover material and wind breaks. The ramp method is typically employed on sloping land,
where waste is spread and compacted in a manner similar to the area method; however, the cover material is
generally obtained from the front of the working face of the filling operation.
Modern landfill design often incorporates liners constructed of soil (e.g., recompacted clay) or
synthetics (e.g., high density polyethylene), or both to provide an impermeable barrier to leachate (i.e., water
that has passed through the landfill), and gas migration from the landfill.
7.8.2 Emission Control Measures
Landfill gas collection systems are either active or passive systems. Active collection systems
provide a pressure gradient in order to extract landfill gas by use of mechanical blowers or compressors.
Passive systems allow the natural pressure gradient created by the increased pressure within the landfill from
landfill gas generation to mobilize the gas for collection.
Landfill gas control and treatment options include (1) combustion of the landfill gas and
(2) purification of the landfill gas. Combustion techniques include techniques that do not recover energy
(e.g., flares and thermal incinerators), and techniques that recover energy (i.e., gas turbines and internal
combustion engines) and generate electricity from the combustion of landfill gas. Boilers can also be used to
recover energy from landfill gas in the form of steam. These combustion techniques are not expected to
provide any control of mercury emissions.
Purification techniques can be used to process raw landfill gas to pipeline quality natural gas by
using adsorption, absorption, and membranes. Mercury emissions may be reduced by adsorption, but no data
are available to determine the extent of control of mercury emissions (if any).
7.8.3 Emissions
Landfill gas, composed of approximately 50 percent methane and 50 percent CO,, is produced by
anaerobic decomposition of MSW in landfills.12 In 1994, MSW landfills were estimated to release 10.2
million Mg (11.2 million tons) of methane.121 Landfill gas also contains trace constituents, including
mercury. Mercury comes from the breakage of waste materials that contain mercury, such as certain types of
batteries, fluorescent light bulbs, and light switches. Data from nine landfills show landfill gas mercury
concentrations that range from 7.0 x 10"7 ppm to 8.8 x 10"4 ppm and average 1.4 x 10"4 ppm.123 Data
provided by EPA's Emission Factor and Inventory Group (EFIG) for the Freshkills Landfill in New York
indicate landfill gas mercury concentrations between 2.5 x 10"3 ppm and 7.1 x 10"4 ppm.124 The midpoint of
the data from Freshkills, 1.6 x 10"3 ppm, was averaged with the other nine data points to calculate an average
landfill gas mercury concentration of 2.9 x 10"4 ppm. Future releases of the EPA document "Compilation of
Air Pollutant Emission Factors" (AP-42) will incorporate additional data (that is not yet available) for
mercury emissions from MSW landfills in Section 2.4, Landfills.
Total 1994 mercury emissions from this source category are estimated to be 0.074 Mg (0.081 tons);
see Appendix A for details.
7.9 GEOTHERMAL POWER PLANTS
Geothermal power plants are either dry-steam or water-dominated.125 For dry-steam plants, steam is
pumped from geothermal reservoirs to turbines at a temperature of about 180°C (360°F) and a pressure of
7.9 bars absolute. For water-dominated plants, water exists in the producing strata at a temperature of
approximately 270 °C (520 °F) and at a pressure slightly higher than hydrostatic. As the water flows towards
the surface, pressure decreases and steam is formed, which is used to operate the turbines. In 1992,
18 geothermal power plants were operating in the United States, and one new plant began operating in
1993.126'127 Table 7-6 lists the names, locations, plant types, and capacities of these facilities.
7-26
-------
TABLE 7-6. GEOTHERMAL POWER PLANTS OPERATING
IN THE UNITED STATES IN 1992a
Facility
The Geysers CA
Salton Sea, CA
Heber, CA
East Mesa, CA
Coso, CA
Casa Diablo, CA
Amedee, CA
Wendel, CA
Puna, HI
Dixie Valley, NV
Steamboat Hot Springs, NV
Beowawe Hot Springs, NV
Desert Peak, NV
Wabuska Hot Springs, NV
Soda Lake, NV
Stillwater, NV
Empire and San Emidio, NV
Roosevelt Hot Springs, UT
Cove Fort, UT
Type
Dry-steam
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Not specified
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Water-dominated
Net capacity (MW)
1,805.7
218.3
47.0
106.0
247.5
34.0
2.0
0.7
25.0
57.0
19.3
16.7
9.0
1.7
15.7
12.5
3.2
20.0
12.1
Total 2,653
Source: References 126 and 127.
aPuna, Hawaii data from Reference 127. Puna facility began operating in 1993. All other data
taken from Reference 126.
7.9.1 Emission Control Measures
No information is available pertaining to air pollution control systems used in geothermal power
plants.
7.9.2 Emissions
Mercury emissions at geothermal power plants are documented to result from two sources: off-gas
ejectors, and cooling towers. Table 7-7 contains the mercury emission factors for these two sources. These
data are based on measurements taken in 1977.125 No process data are given in the documentation
containing the test results and the primary source of these data could not be obtained to verify the validity of
the emission factors. If significant process modifications or changes in control strategies have been
incorporated since 1977, the emission factors reported in Table 7-7 may no longer be valid.
Total 1993 mercury emissions from this source category are estimated to be 1.3 Mg (1.4 tons); see
Appendix A for details.
7-27
-------
TABLE 7-7. MERCURY EMISSION FACTORS FOR GEOTHERMAL POWER PLANTS
Source
Off-gas ejectors
Cooling tower exhaust
Emission factor range,
g/MWe/hr
0.00075 - 0.02
0.026 - 0.072
Average emission factor
g/MWe/hr
0.00725
0.05
Ib/MWe/hr
0.00002
0.0001
Source: Reference 125.
7.10 PULP AND PAPER PRODUCTION
In the pulp and paper industry, wood pulp is produced from raw wood via chemical or mechanical
means or a combination of both. When chemical pulping methods are used to produce pulp, the chemicals
used in the process are recycled for reuse in the process. Combustion sources located in the chemical
recovery area of pulp and paper mills represent potential sources of mercury emissions. Power boilers
located at pulp mills are another potential source of mercury emissions; mercury emissions from power
boilers are discussed in Sections 6.1 and 6.2. A list of the 153 U. S. pulp mills currently in operation that
have chemical recovery combustion sources is provided in Appendix E. A description of the pulping
processes and chemical recovery combustion sources at these pulp mills and the estimated mercury emissions
from the chemical recovery combustion sources are discussed below.
7.10.1 Process Description128'129'130
The wood pulping process may involve chemical or mechanical treatment of the wood or a
combination of both. Four principal chemical wood pulping processes currently in use are (1) kraft, (2) soda,
(3) sulfite, and (4) semichemical. (The semichemical process requires both chemical and mechanical
treatment of the wood.) The kraft process is the dominant pulping process in the United States, accounting
for approximately 80 percent of domestic pulp production. Currently, there are estimated to be 122 kraft,
2 soda, 15 sulfite, and 14 stand-alone semichemical pulp mills in the United States with chemical recovery
combustion sources.131'132'133
In the kraft pulping process, wood chips are "cooked" under pressure in a digester in an aqueous
solution of sodium hydroxide (NaOH) and sodium sulfide (Na2S), referred to as "cooking liquor," or "white
liquor." Cooking the wood chips in white liquor results in the extraction of cellulose from the wood by
dissolving the lignin that binds the cellulose fibers together. The contents of the digester are then discharged
to a blow tank, where the softened chips are disintegrated into fibers or "pulp."
The pulp and spent cooking liquor are subsequently separated in a series of brown stock washers.
Spent cooking liquor, referred to as "weak black liquor," from the brown stock washers is routed to the
chemical recovery area. Weak black liquor is a dilute solution of lignins, organic materials, sodium sulfate
(Na2SO4), sodium carbonate (Na2COo), and white liquor. The purpose of the chemical recovery area is to
recover the cooking liquor chemicals from the spent cooking liquor. After the brown stock washers, the
washed pulp may be subjected to a bleaching sequence, before being pressed and dried to yield the finished
product.
Some of the mercury that is present in the wood chips will also be present in the finished product,
and the rest will be present in the spent cooking liquor. The levels of mercury in the product and in the liquor
are expected to be very low because the levels of mercury in the wood chips are not expected to be higher
than the background levels of mercury in the environment. However, no data are currently available to
confirm this assumption. The amount of mercury that is present in the wood chips is expected to vary
somewhat from mill to mill based on the origin of the wood that the mills process.
Emissions of PM (including metals such as mercury) are associated with combustion units located in
the chemical recovery area. The chemical recovery area at a kraft pulp mill includes chemical recovery
furnaces, smelt dissolving tanks (SDT's), and lime kilns. Figure 7-8 shows the relationship of the chemical
recovery cycle to the pulping and product forming process areas. Figure 7-9 shows a process flow diagram of
the chemical recovery area at kraft pulp mills and identifies the mercury emission points.
7-28
-------
WOOD
PULPING
PULP
PRODUCT
FORMING
PAPER OR
PULP PRODUCT
WHITE LIQUOR
BLACK LIQUOR
CHEMICAL RECOVERY
Figure 7-8. Relationship of the chemical recovery cycle to the pulping
and product forming processes.
7-29
-------
PULPING
WOOD CHIPS
OJ
o
,-'' CHEMICAL RECOVERY
A
1 ^ ~
FLUE GAS
MERCURY EMISSION SOURCES
Figure 7-9. Kraft process—chemical recovery area (including direct contact evaporator recovery furnace^8
-------
In the chemical recovery cycle, weak black liquor is first directed through a series of multiple-effect
evaporators (MEE's) to increase the solids content of the black liquor. The "strong" black liquor from the
MEE's is then either (1) oxidized in the black liquor oxidation (BLO) system if it is further concentrated in a
direct contact evaporator (DCE) or (2) routed directly to a concentrator (i.e., nondirect contact evaporator
[NDCE]).
Concentrated black liquor is then sprayed into the recovery furnace, where organic compounds are
combusted, and the Na9SO4 is reduced to Na2S. The combustion process typically occurs at temperatures
around 982 °C (1800°F) or higher, which would be high enough to volatilize any mercury present in the black
liquor. The black liquor burned in the recovery furnace has a high energy content, which is recovered as
steam for process requirements at the pulp mill. The heat is recovered through the heat exchanger section of
the furnace (i.e., superheater, boiler bank, and economizer). The design economizer exit gas temperature
ranges from 177 to 190°C (350 to 375 °F) for recovery furnaces with NDCE's (i.e., NDCE recovery
furnaces). For recovery furnaces with DCE's (i.e., DCE recovery furnaces), the heat from the recovery
furnace is used to evaporate the black liquor. As a result, the required economizer exit gas temperature for
DCE recovery furnaces, 371 to 427°C (700 to 800°F), is much higher than that for NDCE recovery furnaces.
Based on the exit gas temperatures, there is no substantial difference in the potential for mercury to
remain vaporized when it exits a DCE recovery furnace than an NDCE recovery furnace. However, most
recovery furnaces are of the NDCE design. According to available data, approximately 61 percent of kraft
and soda recovery furnaces are NDCE recovery furnaces, while 39 percent are DCE recovery furnaces. The
NDCE recovery furnace design is a more recent design, with greater capacity, greater energy efficiency, and
lower odorous total reduced sulfur (TRS) emissions. Most recovery furnaces installed since the 1970's have
been NDCE recovery furnaces.
In addition to the steam energy provided by combustion of black liquor in the recovery furnace,
energy for pulping processes at the mill can also be provided by a power boiler. Power boilers located at pulp
and paper mills are usually wood-fired boilers, although coal-, oil-, gas-, and combination fuel-fired boilers
are also used. Process and emissions information for these boilers are provided in Section 6.0.
After the black liquor has been combusted in the recovery furnace, molten inorganic salts, referred to
as "smelt," collect in a char bed at the bottom of the furnace. Smelt, at approximately 1040 to 1150°C (1900
to 2100°F), is drawn off from the furnace and dissolved in weak wash water in the SDT to form a solution of
carbonate salts called "green liquor," which is primarily Na2S and Na2CO3. The green liquor formed in the
SDT also contains insoluble unburned carbon and inorganic impurities, called dregs, which are removed in a
series of clarification tanks.
Although the high temperature of the smelt discharged from the furnace is sufficient to volatilize any
mercury present in the smelt, the smelt is cooled as it enters the SDT when it is shattered by high-pressure
steam or shatter sprays of recirculated green liquor. Large volumes of steam are generated when the molten
smelt is dissolved in the weak wash water, which releases more heat. The vapor space above the liquid level
provides an opportunity for water vapor and PM resulting from the quenching of the smelt to settle out of
suspension into the green liquor. An induced draft fan constantly draws the vapor and entrained PM through
a PM control device, generally a wet scrubber. Because of the cooling of the smelt that occurs in the SDT,
there is some opportunity for mercury to remain in the liquor upon exiting the SDT.
Decanted green liquor from the SDT is transferred to the causticizing area of the mill, where the
Na2CO3 is converted to NaOH by the addition of lime. The green liquor is first transferred to a slaker tank,
where lime from the lime kiln reacts with water to form calcium hydroxide (Ca[OH]2). From the slaker,
liquor flows through a series of agitated tanks, referred to as causticizers, that allow the causticizing reaction
to go to completion (i.e., Ca[OH]2 reacts with Na2CO3 to form NaOH and CaCO3). The causticizing product
is then routed to the white liquor clarifier, which removes CaCO3 precipitate, referred to as "lime mud." The
lime mud, along with dregs from the green liquor clarifier, is washed in the mud washer to remove the last
traces of sodium. The filtrate from the mud washer, known as "weak wash," is used in the SDT to dissolve
recovery furnace smelt. The white liquor (NaOH and Na2S) from the clarifier is recycled to the digesters in
the pulping area of the mill.
The mud from the mud washer is dried and calcined in the lime kiln to produce "reburned" lime,
which is reintroduced to the slaker. The calcining reaction requires a minimum temperature of 815 ° C
(1500°F), which is sufficient to volatilize any mercury present in the lime mud. The combustion gases exit
7-31
-------
the lime mud feed end of the kiln at temperatures of approximately 150° to 200°C (300° to 400°F), which
will also result in volatilization of mercury.
The other pulping processes are similar to the kraft pulping processes but with some significant
differences. The soda pulping process is essentially the same as the kraft process, except that soda pulping is
a nonsulfur process (Na2CO3 only or a mixture of Na2CO3 and NaOH), and, therefore, does not require black
liquor oxidation to reduce the odorous TRS emissions.
The sulfite pulping process is also carried out in a manner similar to the kraft process, except that an
acid cooking liquor is used to cook the wood chips. The sulfite chemical pulping processes currently used at
U.S. mills can be classified as either acid sulfite or bisulfite; these processes use magnesium, ammonia, or
calcium bases to buffer the sulfite cooking liquor. Chemical recovery is only practiced at those sulfite mills
that use the magnesium or ammonia-based sulfite process. The system used to recover cooking chemicals is
specific to the base. Similar to kraft pulp mills, the spent liquor is recovered at sulfite pulp mills by being
burned in a type of combustion unit. Combustion units used at sulfite pulp mills include recovery furnaces
and fluidized-bed reactors. Typical combustion temperatures for sulfite combustion units are about 278 °C
(500°F) lower relative to kraft recovery furnaces, ranging from 704° to 760°C (1300° to 1400°F). These
temperatures are sufficiently high to volatilize any mercury present.
The semichemical pulping process is used to produce corrugating medium, which is the inside layer
of corrugated containers. The semichemical pulping process uses a combination of chemical and mechanical
pulping methods. Wood chips first are partially softened in a digester with chemicals, steam, and heat; once
chips are softened, mechanical methods complete the pulping process. Three types of chemical pulping
methods are currently in use at semichemical mills—neutral sulfite semichemical (NSSC) (sodium-based
sulfite process), kraft green liquor, and nonsulfur (Na2CO3 only or a mixture of Na2CO3 and NaOH).
Semichemical and kraft pulping processes are co-located at 13 mills. At those mills, the spent liquor
from the semichemical pulping process is burned in the kraft recovery furnace. Fourteen mills use the
semichemical pulping process only. Those mills, referred to as "stand-alone semichemical pulp mills," use a
variety of chemical recovery equipment for combusting the spent liquor, but the predominant type (50
percent) appears to be the fluidized-bed reactor. Other types of chemical recovery equipment used at stand-
alone semichemical pulp mills include recovery furnaces, smelters, rotary liquor kilns, and pyrolysis units.
Typical combustion temperatures in the recovery furnaces and smelters are similar to those for kraft and soda,
while typical combustion temperatures in the fluidized-bed reactors and rotary liquor kilns are about 278 °C
(500°F) lower, around 704° to 760°C (1300° to 1400°F). These temperatures are sufficiently high to
volatilize any mercury present.
Similar to the kraft process, cooking liquor chemicals at semichemical mills are recovered from the
chemical recovery combustion equipment as ash or smelt, which is mixed with water in a dissolving tank to
form green liquor. The green liquor is then combined with makeup chemicals to form fresh cooking liquor.
A typical temperature at the dissolving tank vent would be 180°F, which is well below the volatilization
temperature for mercury. Therefore, mercury is expected to be in particulate form at the dissolving tank vent.
7.10.2 Emission Control Measures128'129'134
Due to State and Federal regulations regarding PM emissions, almost all chemical recovery
combustion units at kraft pulp mills (i.e., recovery furnaces, SDT's, and lime kilns) are equipped with add-on
PM control devices.
The PM emitted from kraft recovery furnaces is mainly Na7SO4 (about 80 percent), with smaller
amounts of K2SO4, Na2CO3, and NaCl. There are economic benefits from recycling the predominantly
Na2SO4 PM catch from the recovery furnace flue gases because the recovery of chemicals reduces the costs
of using "makeup" chemicals. The PM emissions from approximately 95 percent of kraft recovery furnaces
are controlled with an ESP alone; the PM emissions from the remaining furnaces are controlled with an ESP
followed by a wet scrubber (4 percent) or with a wet scrubber alone (1 percent). Properly designed and
operated ESP's used on kraft recovery furnaces routinely achieve PM removal efficiencies of 99 percent or
greater. Direct-contact evaporators used to concentrate black liquor also serve to control PM emissions,
removing between 20 and 50 percent of the particulate load prior to the ESP.
To obtain optimal control of PM on a continuous basis, the ESP should be operated at temperatures
between 150° and 260°C (300° and 500°F). Below 150°C (300°F), corrosion is accelerated due to
7-32
-------
concentration of acid gases; above 260°C (500°F), the ESP's PM collection efficiency starts to decline.
Typical exit gas temperatures for ESP's installed on NDCE and DCE recovery furnaces would be 199 and
160°C (390 and 320°F), respectively. Recovery furnaces controlled with wet scrubbers would have even
lower exit gas temperatures (e.g., 82°C [180°F]).
Finely divided smelt (Na2CO3 and Na2S) entrained in water vapor accounts for most of the PM
emissions from SDT's. The PM emissions from approximately 87 percent of kraft SDT's are controlled with
a wet scrubber. Venturi scrubbers are the most commonly used type of wet scrubber installed on SDT's,
comprising about 43 percent of SDT wet scrubbers. Reported PM removal efficiencies for venturi scrubbers
installed on SDT's range from 97 to greater than 99 percent. Typical inlet and outlet temperatures for SDT
wet scrubbers would be 93 and 77 °C (200 and 170°F), respectively. These temperatures are below the
volatilization temperature for mercury. Therefore, it is expected that most of the mercury present will be in
particulate form and should be collected by the wet scrubber.
Particulate matter emissions from most of the remaining SDT's (11 percent) are controlled with a
mist eliminator alone. Mist eliminators are generally less effective than wet scrubbers at controlling PM
emissions. Inlet and outlet temperatures for SDT mist eliminators are similar to those for SDT wet scrubbers,
so any difference in mercury control relative to wet scrubbers would be a result of the lower PM control
efficiency of mist eliminators relative to wet scrubbers.
Lime kiln PM emissions are mainly sodium salts, CaCO3, and CaO, with uncontrolled emissions
comprising mainly calcium compounds and controlled emissions comprising mainly sodium salts. Sodium
salts result from the residual Na2S in the lime mud after washing. There are economic advantages to
recovering the PM emissions from lime kilns because after the PM is recovered, it can be returned to the
system for calcining. Particulate matter emissions from approximately 90 percent of kraft lime kilns are
controlled with a wet scrubber. Two percent of these scrubbers are operated in series with a second scrubber.
Venturi scrubbers are the most commonly used type of wet scrubber installed on lime kilns, comprising about
89 percent of lime kiln wet scrubbers. Particulate matter collection efficiencies for venturi scrubbers installed
on kraft lime kilns average 99 percent. Typical inlet and outlet temperatures for lime kiln venturi scrubbers
would be 249 and 71 °C (480 and 160°F), respectively. These temperatures are below the volatilization
temperature for mercury. Therefore, it is expected that most of the mercury present will be in particulate
form and should be collected by the venturi scrubber.
Particulate matter emissions from the remaining 10 percent of kraft lime kilns are controlled by
ESP's (9 percent) or the combination of an ESP and wet scrubber (1 percent). Installing ESP's to control PM
emissions from lime kilns has been more widespread in recent years; about half of the APCD installations on
lime kilns since 1990 have been ESP's. Properly designed and operated ESP's used on kraft lime kilns
routinely achieve PM removal efficiencies of 99 percent or greater. Typical inlet and outlet temperatures for
the lime kiln ESP are expected to be similar, about 249°C (480°F). Although the outlet temperature for the
ESP is higher than that for the lime kiln venturi scrubber, it is still below the volatilization temperature for
mercury. Therefore, it is expected that most of the mercury will be in particulate form and should be
collected by the ESP.
Some of the equipment operated downstream of the chemical recovery combustion units at sulfite
pulp mills serve a dual role as process equipment and emission control equipment (e.g., absorption towers
used to recover SO2 for reuse in the process and to reduce emissions to the atmosphere). Other equipment
have been installed primarily to reduce emissions. Control devices installed to reduce PM emissions at sulfite
pulp mills include fiber-bed mist eliminators and wet scrubbers. A typical exit gas temperature for these
devices would be 49°C (120°F), which is below the volatilization temperature for mercury. Therefore, most
of the mercury present should be in particulate form and should be recovered using the PM control devices.
Similar to sulfite pulp mills, some of the equipment installed downstream of the chemical recovery
combustion unit at stand-alone semichemical pulp mills are used as both process equipment and emission
control equipment. For example, in addition to controlling PM emissions, venturi scrubbers at some mills
also serve as direct contact evaporators to increase the solids content of the black liquor. Other control
devices installed to reduce PM emissions include cyclones and wet and dry ESP's. Typical exit gas
temperatures for venturi scrubbers and ESP's installed on combustion units at semichemical pulp mills would
be similar to those for comparable control devices installed on combustion units at kraft pulp mills.
Fugitive emissions from sources in a pulp mill include coal piles, paved and unpaved roads, bulk
materials handling (lime, limestone, starch, etc.), and wood handling. Control strategies include wetting; the
7-33
-------
use of chemical agents, building enclosures, and windscreens; paving or wetting roads; and modifying
handling equipment. No information is available on the amount of mercury emitted as fugitive emissions, but
the amount is expected to be very small, around background levels.
There are only limited emission test data from pulp and paper combustion sources on the
performance of add-on controls for metals such as mercury. However, data collected from other combustion
sources on the relative performance of add-on control devices for metals indicate that systems that achieve the
greatest PM removal also provide the best performance for metals. Therefore, particulate mercury may also
be controlled to the same extent as PM. Although no data are available for confirmation, some of the
mercury may be emitted from the control devices in vapor form, especially from the ESP's, which have higher
outlet temperatures compared to the wet scrubbers.
7.10.3 Emissions
Mercury can be introduced into the pulping process through the wood which is being pulped, in the
process water used in the pulping process, and as a contaminant in makeup chemicals added to the process. If
the mercury is not purged from the process in the wastewater or as dregs, it can accumulate in the chemical
recovery area and subsequently be emitted from the chemical recovery combustion sources. The amount of
mercury emitted may depend on how tightly closed the pulping process is (i.e., the degree to which process
waters are recycled and reused). Mercury emission points in the chemical recovery area are shown in
Figure 7-9.
Mercury emissions data are only available from combustion units at kraft pulp mills. Detectable
mercury emissions data are available for eight recovery furnaces, one SDT, and three lime kilns, located at 11
kraft pulp mills. The mercury emissions data for these kraft combustion units were summarized in a
memorandum based on the following sources: (1) test data presented in Technical Bulletin No. 650 from the
National Council of the Pulp and Paper Industry for Air and Stream Improvement (NCASI), (2) test data
provided in a response to a survey sent to pulp and paper mills, and (3) test data in emission test reports
provided by pulp and paper mills.135
Average mercury emission factors were estimated for recovery furnaces, SDT's, and lime kilns based
on the available mercury emissions data. The average mercury emission factors for the recovery furnaces,
SDT, and lime kilns are presented in Table 7-8. Where necessary, test data were blank corrected to be
consistent with EPA Method 29 procedures (which can be used to measure mercury emissions).136 Only
those emission tests that had detectable emissions in at least one test run were included in the mercury
average. For those test runs with nondetect data, the values of the nondetects were calculated as one-half the
detection limit. Data sets for which all three runs were below method detection limits were not included in the
average.
Nationwide 1994 mercury emissions were estimated from these emission factors for kraft and soda
recovery furnaces, SDT's, and lime kilns. The total mercury emissions were estimated to be 1.6 Mg
(1.8 tons); see Appendix A for details. As shown in Appendix A, the single largest source of mercury
emissions in the chemical recovery area is the recovery furnace. Nationwide, mercury emissions account for
only 0.003 percent of PM emissions from kraft and soda recovery furnaces, SDT's, and lime kilns.137
7-34
-------
TABLE 7-8. MERCURY EMISSION FACTORS FOR COMBUSTION SOURCES AT PULP AND
PAPER MILLS
Kraft combustion
source
Recovery furnace
Smelt dissolving
tank
Lime kiln
Emission factor
kg/Mg
1.95xlO-5a
2.61xlO-8a
1.46xlO-6b
Ib/ton
3.90xlO-5a
5.23xlO-8a
2.91xlO-6b
Number of units tested/control device
Eight recovery furnaces, each controlled
with an ESP
One SDT, controlled with a mist eliminator
Three lime kilns, each controlled with a wet
scrubber
Source: Reference 135.
aPer Mg (or ton) of black liquor solids fired in the recovery furnace.
bPer Mg (or ton) of lime produced in the lime kiln.
7-35
-------
8.0 EMISSIONS FROM MISCELLANEOUS FUGITIVE AND AREA SOURCES
Mercury has been found to be emitted from various miscellaneous fugitive and area sources including
the following:
1. Mercury catalysts;
2. Dental alloys;
3. Mobile sources;
4. Crematories;
5. Paint use;
6. Soil dust; and
7. Natural sources
Nationwide mercury emission estimates were developed only for the dental alloy and crematories source
categories. For the remaining categories, either mercury use has been discontinued or no emission factors
could be identified. Mercury emissions from dental alloys were estimated to be 0.64 Mg (0.7 tons) and
emissions from crematories were estimated to be 0.73 Mg (0.80 tons).
8.1 MERCURY CATALYSTS
Mercury catalysts have been used in the production of polyurethane and vinyl chloride. According to
1995 data, U.S. consumption of refined mercury for "other chemical and allied products" includes
Pharmaceuticals and miscellaneous catalysts. This category is no longer reported as a separate category but
is included in the "other uses" category. No data are available for any quantities of mercury used for catalytic
purposes.2
8.1.1 Process Description
Catalysts involved in the production of polyurethane have been composed of the phenylmercuric
compounds (C6H5Hg+), but few facilities currently use this catalyst and phenylmercuric compounds are no
longer produced in the United States.11 The locations of facilities using these compounds are unknown.
Two processes can be used to manufacture vinyl chloride: one process based on acetylene uses
mercuric chloride on carbon pellets as a catalyst, and the other is based on the oxychlorination of ethylene.
Vinyl chloride is produced by oxychlorination at all facilities except at Borden Chemical and Plastics
Corporation in Geismar, Louisiana. Borden Chemical and Plastics produces vinyl chloride using mercuric
chloride as a catalyst with acetylene.11
8.1.2 Emission Control Measures
No information was found in the literature concerning specific control measures for mercury
emissions from the production of vinyl chloride.
8.1.3 Emissions
No emission factors were found in the literature, and no test data that could be used to calculate
emission factors was found. In the 1994 TRI inventory, Borden Chemical and Plastics reported no mercury
emissions at the Louisiana production facility.3
8-1
-------
8.2 DENTAL ALLOYS
Dental amalgams used to fill cavities in teeth include an appreciable quantity of mercury. The
amalgamation process is fairly generic industrywide, although some dental facilities use ready-made dental
capsules to reduce worker exposure to elemental mercury. Dental fillings contain mixtures of metals, usually
silver (67 to 70 percent), tin (25 to 28 percent), copper (0 to 5 percent), and zinc (0 to 2 percent), which are
blended with mercury in a 5:8 proportion to form an amalgam *
8.2.1 Process Description
The dental alloy and mercury are placed inside a two-part plastic capsule that contains a pestle.
Mercury is added with a dispenser that delivers a drop (or "spill") when a button is pressed. Usually, only
one or two drops are necessary to mix the amalgam. The plastic capsule then is closed and placed in an
agitator where the contents are mixed for approximately 15 seconds. Once mixing is completed, the capsule
is opened to remove the amalgam, which then is placed in a container for immediate application in the
cavity.8
8.2.2 Emission Control Measures
No emission controls are noted for handling mercury used in amalgam production. One work
practice is the use of ready-made dental capsules that already contain a pestle and premeasured amounts of
mercury and alloy.8 This practice eliminates any unnecessary handing and accidental spilling of mercury.
8.2.3 Emissions
The total amount of mercury used in the dental industry in 1995 was 32 Mg (35 tons); this accounts
for about 7 percent of the industrial consumption of mercury.2 A 1981 report estimates that 2 percent of the
mercury used in dental applications is emitted to the atmosphere.138 Using the 2 percent figure, 1995
mercury emissions are estimated to be 0.64 Mg (0.7 tons); see Appendix A for estimation procedure.
8.3 MOBILE SOURCES
For the purposes of this document, mobile sources are defined as diesel- and gasoline-powered, on-
road vehicles. The potential for emissions from other types of mobile sources such as ships, motorcycles,
snowmobiles, and other nonhighway mobile sources are not included in this section due to absence of data.
A 1983 study indicated an estimated mercury emission factor of 1.3 x 10"3 milligram (mg) per
kilometer (km) (4.6 x 10"9 Ib/mile) for motor vehicles without resolution of emission rates into vehicle
types.139 The population of vehicles studied was 81.9 percent gasoline-powered passenger cars, 2.4 percent
gasoline-powered trucks, and 15.7 percent diesel trucks. This emission factor was based on a 1977 ambient
sampling study, which was before the widespread use of catalytic converters and unleaded gasoline, and
before State-regulated inspection and maintenance programs were widely mandated. Additionally, both
gasoline and diesel vehicles are now subject to much more stringent tailpipe emission standards than they
were in 1977. Thus, any emissions of mercury from highway motor vehicles are likely to be substantially
reduced from 1977 levels. A 1979 study characterized regulated and unregulated exhaust emissions from
catalyst and non-catalyst equipped light-duty gasoline operated automobiles operating under malfunction
conditions. 14° An analysis for mercury was included in the study but no mercury was detected; the analytical
minimum detection limit was not stated. A 1989 study measured the exhaust emission rates of selected toxic
substances for two late model gasoline-powered passenger cars.141 The two vehicles were operated over the
Federal Test Procedure (FTP), the Highway Fuel Economy Test (HFET), and the New York City Cycle
(NYCC). Mercury was among the group of metals analyzed but was not present in detectable quantities. The
analytical minimum detection limits for mercury in the three test procedures were: FTP--0.025 mg/km (8.9 x
10'8 Ib/mile) HFET-0.019 mg/km (6.7 x 10'8 Ib/mi), and NYCC--0.15 mg/km (53.2 x 10'8 lb/mi).142 These
minimum detection limits are over ten times higher than the estimated emission factor presented in the 1983
study. Because of the large differences between the mercury emission factors and the overall lack of test data,
no average mercury emission factor is recommended for this source.
8-2
-------
8.4 CREMATORIES
Mercury resulting from the thermal instability of mercury alloys of amalgam tooth fillings during
cremation of human bodies may potentially be a source of mercury air emissions. In 1995, approximately
488,224 cremations were performed at the 1,155 crematories located throughout the United States.143
Table 8-1 lists the number of crematories located in each State and the estimated number of cremations
performed in each State for 1995 and projections of these totals for the years 1996, 2000, and 2010.
Only one set of data are available for the average quantity of mercury emitted for a cremation in the
United States. Tests were conducted for a propane-fired incinerator at a crematorium in California. Results
of the testing for uncontrolled mercury emissions ranged from 6.26 E-03 to 2.26 E-03 kg/body burned
(1.38 E-04 to 4.9 E-03 Ib/body); the average mercury emission factor was 1.50 E-03 kg/body burned (3.3 E-
03 Ib/body). The test results were obtained from tests conducted by the California Air Resources Board.144
Total 1995 mercury emissions from this category are estimated to be 0.73 Mg (0.80 tons); see
Appendix A for details.
8.5 PAINT USE
Four mercury compounds—phenylmercuric acetate, 3-(chloromethoxy) propylmercuric acetate,
di(phenylmercury) dodecenylsuccinate, and phenylmercuric oleate-were registered as biocides for interior
and exterior paint but in May 1991, all registrations for mercury-based biocides in paints were voluntarily
cancelled by the registrants. According to the 1996 EPA Report to Congress, the demand for mercury to be
used in paints was eliminated in 1992.
Mercury compounds were added to paints to preserve the paint in the can by controlling microbial
growth and to preserve the paint film from mildew attack after it is applied to a surface. During and after
application of paint, these mercury compounds can be emitted into the atmosphere. One source estimates that
66 percent of the mercury used in paints is emitted into the atmosphere; however, this emission rate, which
was derived using engineering judgement, is based on a 1975 study performed when the demand for mercury
in paint was high.145 The age of the data and the method by which the emission factor was calculated limit
the reliability of the factor, making emission estimates generated from it quite uncertain. Furthermore, no
conclusive information is available regarding the time frame over which mercury in paint is emitted into the
atmosphere after it is applied to a surface. However, limited information suggests that emissions could occur
for as long as 7 years after initial application, although the distribution of emissions over this time period is
unknown Based on the voluntary cancellation of mercury-based biocide registrations in May 1991 and
rapidly declining usage in 1990, it is assumed that current mercury emissions from this source are very small
or zero.
8.6 SOIL DUST
[This section is the same as it appeared in the 1993 document.]
Mercury levels in soil dust have been measured at a few locations in the western United States.146
The mercury level in soil dust near a phosphate fertilizer operation in Pocatello, Idaho was found to be
0.002 (20 ppm) weight percent and levels in dust from an unpaved road near the same facility were at
0.001 weight percent. This reference also cited mercury levels to be about 0.001 weight percent in soil dust
near a courthouse in Medford, Oregon; at a school in Bend, Oregon; near the downtown area of Grant's Pass,
Oregon; and near Key Back in Eugene, Oregon. Samples taken near a silicone manufacturing plant in
Springfield, Oregon, showed mercury levels at 0.004 weight percent in the soil dust. Tests at LaGrande dock
in LaGrande, Oregon, showed mercury in the soil dust at levels of 0.003 weight percent.
The validity of these levels cannot be verified because the original references could not be located to
evaluate the test methods and procedures used in these studies. In addition, the mercury levels found in the
soils of these areas probably are not indicative of soil levels in other areas of the country. The soils in the
Idaho and Oregon areas are primarily volcanic in geologic origin and have higher soil mercury levels than
other areas of the U.S.
8-3
-------
TABLE 8-1. NUMBER OF CREMATORIES AND CREMATIONS BY STATE
United States
New England
Connecticut!
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
Middle Atlantic
New Jersey
New York
Pennsylvania
East North Central
Illinois
Indiana
Michigan
Ohio
Wisconsin
West North Central
Iowa
Kansas
Minnesota
Missouri
Nebraska
North Dakota
South Dakota
1995
Crematories
1,155
47
11
4
13
8
6
5
110
17
43
50
197
48
28
40
49
32
83
16
9
22
24
7
2
3
1995
Cremations
488,224
30,268
6,336
4,079
11,979
2,945
2,759
2,170
64,570
18,385
27,629
18,556
70,707
20,579
5,964
17,529
18,083
8,552
21,229
3,448
1,918
8,501
5,356
2,006
NA
NA
Projected
1996
Crematories
1,177
46
109
202
86
1996
Cremations
514,100
32,500
66,900
76,000
24,600
2000
Crematories
1,321
49
117
227
97
2000
Cremations
606,200
41,200
78,400
93,900
29,800
2010
Crematories
1,678
57
138
288
123
2010 Cremations
836,500
62,800
107,300
138,800
42,800
-------
TABLE 8-1. (continued)
South Atlantic
Delaware
B.C.
Florida
Georgia
Maryland
N. Carolina
S. Carolina
Virginia
W. Virginia
East South Central
Alabama
Kentucky
Mississippi
Tennessee
West South Central
Arkansas
Louisiana
Oklahoma
Texas
Mountain
Arizona
Colorado
Idaho
Montana
Nevada
New Mexico
1995
Crematories
247
4
0
124
21
17
31
14
29
7
30
9
5
4
12
70
12
7
9
42
120
30
28
14
15
12
12
1995
Cremations
95,894
1,245
NA
61,070
6,086
6,797
8,074
3,127
8,396
1,099
7,493
1,661
1,768
1,077
2,987
23,800
3,039
2,923
2,193
15,645
41,554
13,479
10,408
2,895
3,402
6,557
3,264
Projected
1996
Crematories
247
30
74
126
1996
Cremations
100,500
8,400
26,000
44,200
2000
Crematories
294
37
83
144
2000
Cremations
119,200
11,400
33,400
51,800
2010
Crematories
413
53
105
187
2010 Cremations
165,900
18,900
51,900
70,900
-------
TABLE 8-1. (continued)
Utah
Wyoming
Pacific
Alaska
California
Hawaii
Oregon
Washington
1995
Crematories
6
3
251
6
147
8
37
53
1995
Cremations
1,549
NA
131,405
1,328
92,646
4,214
11,736
21,481
Projected
1996
Crematories
257
1996
Cremations
135,000
2000
Crematories
273
2000
Cremations
147,100
2010
Crematories
314
2010 Cremations
177,200
Source: Reference 143.
-------
8.7 NATURAL SOURCES OF MERCURY EMISSIONS
[This section is the same as it appeared in the 1993 document.]
Mercury is emitted from natural sources (rock, soils, water and biota) primarily as elemental mercury
vapor and to a lesser degree as particulate and vaporous oxides, sulfides and halides of mercury.
Organomercuric compounds (methylmercury vapors) are also a significant component of natural emissions
(some evidence of dimethyl-mercury emissions also exists).147 However, few direct measurements of
mercury flux and speciation from natural sources are available in the literature. There is general agreement
that the principal natural sources of mercury emissions include, in order of probable importance,
volatilization in marine and other aquatic environments, volatilization from vegetation, degassing of geologic
materials, PM and vapor emissions during volcanic and geothermal activity, wind-blown dust, and PM and
vapor emissions during forest and brush fires. Recent studies strongly emphasize the importance of the air-
water exchange of mercury as well as biologically mediated volatilization in both marine and terrestrial
environments.14?-150 These sources represent a relatively constant flux to the atmosphere and may comprise
30 to 50 percent of total natural emissions.150 In contrast, volcanic, geothermal, and burning biomass
activities are widely variable temporally and spatially. Volcanic eruptions, in particular, can cause massive
perturbations in atmospheric trace metal cycles. Volcanic activity alone may comprise 40 to 50 percent of
total natural mercury emissions at times.
Published estimates of total global emissions of mercury from natural sources range widely from 100
to 30,000 megagrams (Mg) (110 to 33,000 tons) per year. However, the more recent estimates cluster in the
2,000 to 3,000 Mg per year range.147 Reference 147, citing work done in 1988, estimates natural emissions
to be 3,000 Mg (3,300 tons) per year or approximately 40 percent of total global emissions from all sources.
The supporting data for individual source categories are limited for each of these estimates, and it is clear that
any quantitative understanding of natural mercury flux is lacking.
As a result of reemission, current levels of mercury emitted to the atmosphere by natural processes
are elevated relative to preindustrial levels. More than two thirds of world mercury production has occurred
since 1900, and mercury emissions have been widely dispersed and recycled. In other words, present day
emissions from natural sources are comprised of yesterday's anthropogenic emissions, in part. It is not
possible to quantify the contribution of recycled mercury to the natural emissions estimates and, therefore, the
estimates cited above must be viewed with even greater uncertainty.
8-7
-------
9.0 SOURCE TEST PROCEDURES
9.1 INTRODUCTION
A number of methods exist to determine mercury (Hg) emissions from stationary sources. Several
EPA offices and some State agencies have developed source specific or dedicated sampling methods for Hg.
Other industry sampling methods exist, including continuous emission monitors (CEMs), but these methods
have not been validated and are not discussed in this section.
Subsequent parts of this section discuss EPA reference or equivalent sampling methods for Hg.
Sampling methods fall into one of two categories: (1) dedicated Hg methods for specific sources or
(2) multiple metals sampling trains that include Hg for multiple sources. Each category of methods is
described, differences among the methods are discussed, and a citation is provided for more detailed
information about the methods. A summary of methods is presented in Table 9-1.
Sampling methods included in this section were selected from EPA reference methods and State
methods. To be a reference method, a sampling method must undergo a validation process and be published.
To qualify as an equivalent method, a sampling method must be demonstrated to the EPA Administrator,
under specific conditions, as an acceptable alternative to the normally used reference methods.
9.2 DEDICATED MERCURY SAMPLING METHODS
9.2.1 EPA Method 101-Determination of Particulate and Gaseous
Mercury Emissions from Chlor-Alkali Plants (40 CFR, Part 61, 1992)
This method can be used to determine particulate and gaseous Hg emissions from chlor-alkali plants
and other sources (as specified in the regulations) where the carrier-gas stream in the duct or stack is
principally air.151 Particulate and gaseous Hg emissions are withdrawn isokinetically from the source and
collected in an acidic iodine monochloride (IC1) solution. The Hg collected (in the mercuric form) is reduced
to elemental Hg and then aerated and precipitated from the solution into an optical cell and measured by
atomic absorption spectrophotometry (AAS). A diagram of a sampling train typical of dedicated Hg
sampling trains is presented in Figure 9-1.
After initial dilution, the range of this method is 0.5 to 120 micrograms of Hg per milliliter
(,ug Hg/ml). The upper limit can be extended by further dilution of the sample. The sensitivity of this
method depends on the selected recorder/spectrophotometer combination.
Analytical interferences include SO2, which reduces IC1 and causes premature depletion of the IC1
solution. Also, concentrations of IC1 greater than 10"4 molar inhibit the reduction of the Hg(II) ion in the
aeration cell. Condensation of water vapor on the optical cell windows of the AAS causes a positive
interference.
Estimates of precision and accuracy were based on collaborative tests, wherein 13 laboratories
performed duplicate analyses on two Hg-containing samples from a chlor-alkali plant and on one laboratory-
prepared sample of known Hg concentration. The estimated within-laboratory and between-laboratory
standard deviations are 1.6 and 1.8 /^g Hg/ml, respectively.
9-1
-------
TABLE 9-1. MERCURY SAMPLING METHODS
Method
EPA 101
EPA 101 A
EPA 102
EPA 29
SW-846 0012
OSW-BIF
CARB 436
Filter
None
Glass fiber
(optional)
None
Quartz or glass fiber
Quartz or glass fiber
Quartz or glass fiber
Quartz or glass fiber
Impinger
3XIC1
1 X silica gel
1 X KMnO4
2 X KMnO4
1 X silica gel
3XIC1
1 X silica gel
1 X empty (optional)
2 X HNO3/H2O2
1 X empty
2 X KMnO4/H2SO4
1 X silica gel
1 X empty (optional)
2 X HNO3/H2O2
1 X empty
2 X KMnO4/H2SO4
1 X silica gel
1 X empty
2 X HNO3/H2O2
1 X empty
2 X KMnO4/H2SO4
1 X silica gel
1 X empty
2 X HNOJH2O2
2 X KMn04/H2SO4
1 X silica gel
Range
0.5 to 1 20 ug Hg/ml
20-800 ng Hg/ml
0.5 to 1 20 ug Hg/ml
ngHg/ml to //g Hg/ml
ngHg/ml to jug Hg/ml
ngHg/ml to ug Hg/ml
ngHg/ml to ug Hg/ml
Chemical interference
SO2
Oxidizable organic matter,
Water vapor on optical
window
SO2
None
None
None
None
Detection limit
Not listed
Not listed
Not listed
0.2 ng Hg/ml
0.2 ng Hg/ml
0.2 ng Hg/ml
0.2 ng Hg/ml
•o
-------
STACK WALL
•O
OJ
PROBE
THERMOMETER
CHECK VALVE
REVERSE-TYPE
PITOTTUBE
TEMPERATURE SENSOR
I- o III '.' Ill
I III III
ILjl [Ljl [Ljl [Ljl
PITOT MANOMETER
THERMOMETERS
IMPINGERS
BY-PASS VALVE
AIN VALVE
ORIFICE
MANOMETER
VACUUM LINE
VACUUM GAUGE
TIGHT PUMP
DRY TEST METER
Figure 9-1. Typical dedicated mercury sampling train.
-------
9.2.2 EPA Method IQlA-Determination of Particulate and Gaseous Mercury Emissions from Stationary
Sources (40 CFR. Part 61. 1996)
This method is similar to Method 101, except acidic potassium permanganate (KMnO4) solution is
used for collection instead of acidic IC1.152 This method is used to determine particulate and gaseous Hg
emissions from stationary sources. This method is a revised version of EPA Method 101 as published in
40 CFR, Part 61, 1992, which was entitled "Determination of Particulate and Gaseous Mercury Emissions
from Sewage Sludge Incinerators."
Particulate and gaseous Hg emissions are withdrawn isokinetically from the source and collected in
acidic KMnO, solution. The Hg collected (in the mercuric form) is reduced to elemental Hg, which is then
aerated from the solution into an optical cell and measured by AAS or by any atomic absorption unit with an
open sample presentation area in which to mount the optical cell.
After initial dilution, the range of this method is 20 to 800 nanograms of Hg per milliliter
(ng Hg/ml). The upper limit can be extended by further dilution of the sample. The sensitivity of the method
depends on the selected recorder/spectrophotometer combination.
Analytical interferences include excessive oxidizable organic matter in the stack gas, which
prematurely depletes the KMnO^ solution and thereby prevents further collection of Hg. Condensation of
water vapor on the optical cell windows of the AAS causes a positive interference.
Based on eight paired-train tests, the within-laboratory standard deviation was estimated to be 4.8 /j-g
Hg/ml in the concentration range of 50 to 130 micrograms of Hg per cubic meter (/j,g Hg/m3).
9.2.3 EPA Method 102-Determination of Particulate and Gaseous Mercury Emissions from
Chlor-Alkali Plants-Hydrogen Streams f40 CFR. Part 61. 1992)
Although similar to Method 101, Method 102 requires changes to accommodate extracting the
sample from a hydrogen stream.153 Sampling is conducted according to Method 101, except for the
following procedures:
1. Operate only the vacuum pump during the test. The other electrical equipment, e.g., heaters, fans,
and timers, normally are not essential to the success of a hydrogen stream test.
2. Calibrate the orifice meter at flow conditions that simulate the conditions at the source as
described in APTD-0576 (see Citation 9 in Section 10 of Method 101). Calibration should either be done
with hydrogen or some other gas having a similar Reynolds Number so that there is a similarity between the
Reynolds Numbers during calibration and sampling.
9.3 MULTIPLE METALS SAMPLING TRAINS
9.3.1 Method 0012-Methodology for the Determination of Metals Emissions in Exhaust Gases from
Hazardous Waste Incineration and Similar Combustion Sources
Method 0012 was developed for the determination of 16 metals, including Hg, from stack emissions
of hazardous waste incinerators and similar combustion processes.154 While Method 0012 can be used to
determine particulate emissions from these sources, the filter heating/desiccation modifications to the sample
recovery and analysis procedures for determining particulate emissions may potentially impact the front-half
Hg determination. A diagram of a sampling train typical of a multiple metals sampling train is presented in
Figure 9-2.
The stack sample is withdrawn isokinetically from the source. Particulate emissions are collected in
the probe and on a heated filter; gaseous emissions are collected in a series of moisture knockout traps,
chilled impingers, and silica gel traps. Of the four solution charged impingers, two contain an aqueous
solution of dilute nitric acid (HNOo) combined with dilute hydrogen peroxide (H2O2) and two contain acidic
potassium permanganate (KMnOJ solution. Materials collected in the sampling train are digested with acid
solutions using conventional Parr® Bomb, or microwave digestion techniques to dissolve
9-4
-------
THERMOCOUPLE
GLASS
PROBE•
TIP
REVERSE-TYPE
PITOT TUBE
THERMOMETER
ALL GLASS SAMPLE EXPOSED SURFACE TO HERE.
(EXCEPT WHEN TEFLON FILTER SUPPORT IS USED)
GLASS
FILTER
HOLDER
THERMOCOUPLE
GLASS PROBE
LINE
IMPINGERSWITH
ABSORBING SOLUTIONS
HEATED
AREA
PITOT
MANOMETER
EMPTY (OPTIONAL MOISTURE KNOCKOUT)
5%HN03/10%H2O2
EMPTY
4%KMnO4/10%H2SO4
THERMOCOUPLES
BY-PASS VALVE
VACUUM GAUGE
VACUUM LINE
AIR-TIGHT PUMP
DRY GAS METER
Figure 9-2. Typical multiple metals sampling train.
154
-------
inorganics and to remove organic constituents that may create analytical interferences. As many as six
separate samples can be recovered from the sampling train. The HNO3/H2O2 impinger solution, the acidic
KMnO4 impinger solution, the hydrochloric acid (HC1) rinse solution, the acid probe rinse, the acetone probe
rinse, and digested filter solutions can be analyzed for Hg by cold vapor atomic absorption spectroscopy
(CVAAS). As few as three sample fractions can be analyzed for Hg: the combined probe rinse and filter, the
combined HNOo/H^Oj impinger solutions, and the combined KMnO4 impinger and rinse solutions. The
detection limit for Hg by CVAAS is approximately 0.2 ng Hg/ml.
The corresponding in-stack method detection limit can be calculated by using (1) the procedures
described in this method, (2) the analytical detection limits described in the previous paragraph, (3) a volume
of 300 ml for the front-half and 150 ml for the back-half samples, and (4) a stack gas sample volume of
1.25m3:
C
where: A = analytical detection limit, //g Hg/ml
B = volume of sample prior to aliquot for analysis, ml
C = sample volume, dry standard cubic meter (dscm)
D = in-stack detection limit, /j,g Hg/m3
The in-stack method detection limit for Hg using CVAAS based on this equation is 0.07 /j,g Hg/m3
for the total sampling train. A similar determination using AAS is 5.6 /^g Hg/m3.
Two other multiple metals sampling methods developed by EPA can be used to collect Hg. These
methods are the Methodology for the Determination of Metals Emissions in Exhaust Gases from Hazardous
Waste Incineration and Similar Combustion Sources and EPA Method 29-Methodology for the
Determination of Metals Emissions from Stationary Sources.133'156 Both methods are virtually identical to
Method 0012 in sampling approach and analytical requirements.
9.3.2 CARB Method 436-Determination of Multiple Metals Emissions from Stationary Sources
This method can be used to determine the emissions of metals, including Hg, from stationary
sources. 157 This method is similar to SW-846 Method 0012 in sampling approach and analytical
requirements. Method 436 suggests that the concentrations of target metals in the analytical solutions be at
least 10 times the analytical detection limits. This method may be used in lieu of Air Resource Board
Methods 12, 101, 104, 423, 424, and 433.
9.4 ANALYTICAL METHODS FOR DETERMINATION OF MERCURY158'159
This section contains brief descriptions of two analytical techniques generally used for Hg
determinations.
The two Hg analysis methods are Method 7470 and 7471, from SW-846.158'159 Both methods are
cold-vapor atomic absorption methods, based on the absorption of radiation at the 253.7-nm wavelength by
mercury vapor. Mercury in the sample is reduced to the elemental state and aerated from solution in a closed
system. The Hg vapor passes through a cell positioned in the light path of an atomic absorption
spectrophotometer. Absorbance (peak height) is measured as a function of mercury concentration. Cold-
Vapor AA (CVAA) uses a chemical reduction to selectively reduce Hg. The procedure is extremely sensitive
but is subject to interferences from some volatile organics, chlorine, and sulfur compounds. The typical
detection limit for these methods is 0.0002 mg/L.
The two methods differ in that Method 7470 is approved for analysis of Hg in mobility -procedure
extracts, aqueous wastes, and ground waters.158 Method 7471 is approved for analysis of Hg in soils,
sediments, bottom deposits, and sludge-type materials.159 Analysis of samples containing high amounts of
organics presents special problems: (1) the tendency to foam during the reduction step, which blocks the flow
of sample to the absorption cell and (2) the reduction of Hg(II) to Hg before addition of stannous chloride
(SnCl2).
9-6
-------
Two analytical considerations are common to both methods. First, stannous chloride should be
added immediately prior to analysis to ensure the reduction of Hg(II) to Hg occurs in the vaporization cell
only. Second, moisture in the absorption cell can reduce the reliability of the method and should be
eliminated or minimized. Finally, a closed-loop system may provide a more reliable system than an open-
loop system for introducing the sample to the reaction flask.
9.5 SUMMARY
All of the above source sampling methods collect a sample for analysis of multiple metals, including
Hg, or a sample for Hg analysis alone. Significant criteria and characteristics of each method are presented in
Table 9-1. This table is a summary of information presented in various methods. The major differences
among the methods involve (1) the type of impinger solutions, (2) the amount or concentration of impinger
solutions, (3) the sequence and types of sample train recovery solutions, and (4) the use and/or type of
particulate filter.
In assessing Hg emissions from test reports, the age or revision number of the method indicates the
level of precision and accuracy of the method. Older methods are sometimes less precise or accurate than
those that have undergone more extensive validation. Currently, EPA Method 301 from 40 CFR Part 63,
Appendix A, can be used to validate or prove the equivalency of new methods.
9-7
-------
10.0 REFERENCES
1. S. C. DeVito, 1995. Mercury. (In) Kirk-Othmer Encyclopedia of Chemical Technology, Volume 16,
4th ed., J. Kroschivitz, exec, ed., John Wiley and Sons, New York. pp. 212-228.
2. J. Plachy, 1996. Mercury. (In) Minerals Yearbook, Volume l--Metals and Minerals, U.S. Geological
Survey, U.S. Department of Interior, Washington, D.C.
3. U. S. Environmental Protection Agency, 1996. Toxics Release Inventory, 1994 TRI data, Office of
Pollution Prevention and Toxics (OPPT), Washington, D.C.
4. XATEF, 1991. Crosswalk/Air Toxic Emission Factor Data Base, Version 1.2 for October 1991
Update. Office of Air Quality Planning and Standards, U. S. Environmental Protection Agency,
Research Triangle Park, NC.
5. R.B. Coleman, 1990. Roasting of Refractory Gold Ore and Concentrates. (In) Proceedings of the
Gold '90 Symposium--Gold'90, Salt Lake City, UT. February 26-March 1, 1990. Society of Mining
Engineers of AIME, Littleton, CO.
6. D. Anderson, 1973. Emission Factors for Trace Substances. EPA-450/2-72-001. U.S.
Environmental Protection Agency, Research Triangle Park, NC.
7. P.T. Bartlett and T. L. Muldoon, 1982. Propose and Evaluate Methods of Controlling Mercury Vapor
Emissions in a Processing Mill Furnace Room. BUMINES-OFR-44-83 (NTIS PB 83-171751).
National Technical Information Service, Springfield, VA.
8. R.P. Reisdorf and D.C. D'Orlando, 1984. Survey of Health Hazard Control Systems for Mercury Use
and Processing. NTIS PB85-107241. National Technical Information Service, Springfield, VA.
9. U. S. Environmental Protection Agency, 1984. Review of National Emission Standards for Mercury.
EPA-450/3-84-004. Emission Standards and Engineering Division, Research Triangle Park, NC.
10. Mercury Refining Company, 1997. Excerpts from emission source test reports conducted by General
Testing Corporation in September 1994 and Galson Corporation in June 1995, Submitted to Midwest
Research Institute , Cary, NC, September 3, 1997.
11. SRI International, 1996. Directory of Chemical Producers: United States of America. SRI
International, Menlo Park, CA.
12. Dynamac Corporation, 1982. Mercury Control Technology Assessment Study: Troy Chemical
Corporation, Newark, NJ. Preliminary Survey Report for the Site Visit of October 15, 1981.
ECTB-109-32A (NTIS PB89-13026). National Technical Information Service, Springfield, VA.
13. Chlorine Institute, 1997. Comment letter from the Chlorine Institute to D. Beauregard, U.S.
Environmental Protection Agency; Emissions, Monitoring, and Analysis Division, Research Triangle
Park, NC, June 20, 1997.
14. F. Rauh, 1991. Alkali and Chlorine Products: Mercury Cell Process. (In) Kirk-Othmer Encyclopedia
of Chemical Technology, Volume 1, 4th ed., J.I. Kroschivitz, exec, ed., John Wiley and Sons, New
York.
15. BF Goodrich Company of Calvert City, Kentucky, 1992. Response to Section 114 Questionnaire
Chlor-Alkali Information Request. O'Brien, W.C. to Jordan, B.C., U.S. EPA/ESD.
10-1
-------
16. Georgia-Pacific Corporation of Bellingham, Washington, 1993. Response to Section 114
Questionnaire—Chlor-Alkali Information Request. Dahlgre, E. to Jordan, B.C., U.S. EPA/ESD.
17. LCP Chemicals of Riegelwood, North Carolina, 1993. Response to Section 114 Questionnaire--
Chlor-Alkali Information Request. H. L. Croom to B. C. Jordan, U.S. EPA/ESD. February 1993.
18. LCP Chemicals of Orrington, Maine, 1993. Response to Section 114 Questionnaire—Chlor-Alkali
Information Request. D. R. Tonini to B. C. Jordan, U.S. EPA/ESD. February 1993.
19. Occidental Petroleum Corporation, 1993. Response to Chlorine Production Information Collection
Requests submitted to U.S. EPA for the Deer Park Plant.
20. Occidental Petroleum Corporation, 1993. Response to Chlorine Production Information Collection
Requests submitted to U.S. EPA for the Delaware City Plant.
21. Occidental Petroleum Corporation, 1993. Response to Chlorine Production Information Collection
Requests submitted to U.S. EPA for the Muscle Shoals Plant.
22. Olin Chemicals of Augusta, GA, 1993. Response to Section 114 Questionnaire—Chlor-Alkali
Information Request. J.C. Rytlewski to B.C. Jordan, U.S. EPA/ESD. February 1993.
23. Olin Chemicals of Charleston, TN, 1993. Response to Section 114 Questionnaire—Chlor-Alkali
Information Request. J.P. Newman to B.C. Jordan, U.S. EPA/ESD. February 1993.
24. Pioneer Chlor Alkali Company, Inc. of St. Gabriel, LA, 1993. Response to Section 114
Questionnaire—Chlor-Alkali Information Request. B.L. Bennett to B.C. Jordan, U.S. EPA/ESD.
March 1993.
25. PPG Industries, 1993. Response to Chlorine Production Information Collection Requests submitted to
U.S. EPA for PPG Industries' Natrium, WV, and Lake Charles, LA, chlorine production facilities.
26. PPG Industries of Lake Charles, LA, 1993. Response to Section 114 Questionnaire—Chlor-Alkali
Information Request. A.P. Plauche to B.C. Jordan, U.S. EPA/ESD. January 1993.
27. Vulcan Materials Co., 1993. Response to Chlorine Production Information Collection Requests
submitted to U.S. EPA for Vulcan Chemicals' Wichita, KS, Geismar, LA, and Port Edwards, WI,
chlorine production facilities.
28. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section
8.11, Chlor-Alkali, Research Triangle Park, NC.
29. R. Erdheim, 1997. Comment letter from National Electrical Manufacturers Association (NEMA) to
D. Beauregard, U.S. Environmental Protection Agency; Emissions, Monitoring, and Analysis
Division, Research Triangle Park, NC, June 19, 1997.
30. M.M. Dierlich, 1994. Comment letter from National Electrical Manufacturers Association (NEMA)
to M. H. Keating, U. S. Environmental Protection Agency, Emission Standards Division (MD-13),
Research Triangle Park, NC, February 4, 1994.
31. Wisconsin Bureau of Air Management (WBAM), 1986. Mercury Emissions to the Atmosphere in
Wisconsin. Publication No. PUBL-AM-014. Wisconsin Department of Natural Resources, Madison,
WI.
32. U. S. Environmental Protection Agency, 1993. Locating and Estimating Air Emissions from Sources
of Mercury and Mercury Compounds, EPA-454/R-93-023. U. S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park, NC, September 1993.
33. U. S. Environmental Protection Agency, 1996. Mercury Study—Report to Congress, Volume II: An
Inventory of Anthropogenic Mercury Emissions in the United States (SAB Review Draft), EPA-
452/R-96-001b, Office of Air Quality Planning and Standards and Office of Research and
Development, Research Triangle Park, NC, June 1996.
10-2
-------
34. National Electrial Manufacturers Association (NEMA), 1996. Fluorescent Lamps and the
Environment. Rosslyn, VA. July, 1996.
35. W. Battye, U. McGeough, and C. Overcash. (EC/R, Inc.), 1994. Evaluation of Mercury Emissions
from Fluorescent Lamp Crushing. EPA-453/R-94-018. U. S. Environmental Protection Agency,
Research Triangle Park, NC.
36. National Air Toxics Information Clearinghouse Data Base (NATICH), 1992. Emissions Standards
Division, Office of Air Quality Planning and Standards, U. S. Environmental Protection Agency,
Research Triangle Park, NC.
37. Environmental Enterprises, Inc., 1994. Emissions Testing for Particulate Matter and Mercury.
Prepared for USA Lights of Ohio, Cincinnati, OH.
38. U. S. Environmental Protection Agency, 1992. Characterization of Products Containing Mercury in
Municipal Solid Waste in the United States, 1970 to 2000. Office of Solid Waste, Washington, DC.
39. U. S. Department of Energy (DOE), 1996. State Energy Data Report, Consumption Estimates, 1994.
Report No. DOE/EIA-0214(94), Energy Information Administration, U.S. Department of Energy,
Washington, DC, October 1996.
40. G. Brooks, 1989. Estimating Air Toxic Emissions from Coal and Oil Combustion Sources.
EPA-450/2-89-001. Prepared by Radian Corporation for U. S. Environmental Protection Agency,
Research Triangle Park, NC. April 1989.
41. D.M. White, L.O. Edwards, A.G. Eklund, D.A. DuBose, and F.D. Skinner, 1984. Correlation of Coal
Properties with Environmental Control Technology Needs for Sulfur and Trace Elements.
EPA-600/7-84-066. U. S. Environmental Protection Agency, Research Triangle Park, NC.
42. V.E. Swanson, J.H. Medlin, J.R Hatch, S.L. Coleman, G.H. Wood, S.D. Woodruff, and
R.T. Hildebrand, 1976. Collection, Chemical Analysis, and Evaluation of Coal Samples in 1975.
USGS Report No. 76-468. U.S. Department of the Interior, Geological Survey. Washington, DC.
43. U. S. Environmental Protection Agency, 1988. Compilation of Air Pollutant Emission Factors,
AP-42, Fourth Edition, Supplement B. U. S. Environmental Protection Agency, Research Triangle
Park,NC. September 1988.
44. Babcock and Wilcox Company. Steam: Its Generation and Use. Babcock and Wilcox, New York,
NY. 1978.
45. R.B. Finkelman, U.S. Department of the Interior, 1993. Metal Concentrations in Coal. Letter to D. D.
Wallace, Midwest Research Institute, February 8, 1993.
46. U. S. Environmental Protection Agency, 1996. Study of Hazardous Air Pollutant Emissions from
Electric Utility Steam Generating Units—Interim Final report, EPA-453/R-96-013a. Office of Air
Quality Planning and Standards, Research Triangle Park, NC, October 1996.
47. Electric Power Research Institute (EPRI), 1994. Electric Utility Trace Substances Synthesis Report
and Appendices. EPRI TR-104614. November 1994.
48. A.J. Buonicore and W.T. Davis, eds. 1992. Air Pollution Engineering Manual. VanNostrand
Reinhold, New York, NY.
49. Electric Power Research Institute (EPRI), 1997. Mercury and Other Trace Metals in Coal.
EPRI TR-106950. January 1997.
50. Review comments from Dr. L. Levin, Electric Power Research Institute (EPRI), Palo Alto, CA, to
D. Beauregard, U. S. Environmental Protection Agency, Research Triangle Park, NC, June 20, 1997.
51. D.J. Akers, et al., 1993. The Effect of Coal Cleaning on Trace Elements. Draft Report to the Electric
Power Research Institute, Palo Alto, CA. CQ Inc., Homer City, PA, December 1993.
10-3
-------
52. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42,
Section 1.1, Bituminous and Subbituminous Coal Combustion, Research Triangle Park, NC.
53. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42,
Section 1.2, Anthracite Coal Combustion, Research Triangle Park, NC.
54. W.E. Osborne and M.D. McDannel, 1990. Emissions of Air Toxic Species: Test Conducted Under
AB2588 for the Western States Petroleum Association. Report No. CA 72600-2601. CARNOT,
Tustin, CA. May 1990.
55. D.J. Boron, Controlling Toxic Emissions. Coal. June 1990.
56. P. Chu and D. B. Porcella, "Mercury Stack Emissions From U. S. Electric Utility Power Plants",
Proceedings of the Third International Conference on Mercury as a Global Pollutant, Whistler,
Britich Columia, Canada, July 10-14, 1994.
57. Pape & Steiner Environmental Services, 1990. AB2588 Testing at Texaco Trading and
Transportation, Inc. Panoche Station-Heater, June 18 through June 29, 1990. Report PS-90-2187.
Pape & Steiner Environmental Services Bakersfield, CA. September 1990.
58. U. S. Environmental Protection Agency, 1996. Emission Factor Documentation for AP-42, Section
1.6, Wood Waste Combustion in Boilers, Research Triangle Park, NC.
59. National Council of the Paper Industry for Air and Stream Improvement, Inc. (NCASI), 1991. Current
Status of Nonrecyclable Paper Burning in the Forest Products Industry. NCASI Technical Bulletin
No. 615. September 1991.
60. U. S. Environmental Protection Agency, 1982. Background Information for New Source Performance
Standards: Nonfossil Fuel Fired Industrial Boilers. Draft EIS. EPA-450/3-82-007. Office of Air
Quality Planning and Standards, Research Triangle Park, NC.
61. U. S. Environmental Protection Agency, 1996. Emission Factor Documentation for AP-42, Section
1.10, Residential Wood Stoves, Research Triangle Park, NC.
62. U. S. Environmental Protection Agency, 1996. Emission Factor Documentation for AP-42, Section
1.9, Residential Fireplaces, Research Triangle Park, NC.
63. National Council of the Paper Industry for Air and Stream Improvement, Inc. (NCASI), 1995.
Compilation of Air Toxic and Total Hydrocarbon Emissions Data For Sources at Chemical Wood
Pulp Mills. NCASI Technical Bulletin No. 701. October 1995.
64. D.G. DeAngelis, D.S. Ruffin, J.A. Peters, and RB. Reznik, 1980. Source Assessment: Residential
Wood Combustion. U. S. Environmental Protection Agency, Research Triangle Park, NC.
65. B. Phillips, 1993. Residential Wood Consumption. U.S. EPA/OAQPS, facsimile to T. Campbell,
Midwest Research Institute, November 12, 1993.
66. D. Fenn and K. Nebel, Radian Corporation, 1992. Memorandum to W. Stevenson, U. S.
Environmental Production Agency, March 9, 1992.
67. U. S. Environmental Protection Agency, 1996. National Emissions for Municipal Waste Combustors,
Office of Air Quality Planning and Standards, Research Triangle Park, NC, October 1996.
68. U. S. Environmental Protection Agency, 1996. Characterization of Municipal Solid Waste in the
United States: 1995 Update. Washington, D.C., March 1996.
69. Franklin Associates, Ltd., 1989. Characterization of Products Containing Mercury in Municipal Solid
Waste in the United States, 1970 to 2000. EPA-530/SW-89-015A. U. S. Environmental Protection
Agency, Washington, D.C. January 1989.
10-4
-------
70. Solid Waste Association of North America (SWANA), 1993. Mercury Emissions from Municipal
Solid Waste Combustors: An Assessment of the Current Situation in the United States and Forecast
of Future Emissions. National Renewable Energy Laboratory, Golden, CO. May 1993.
71. B. Strong, (Midwest Research Institute), 1997. Pollution Prevention. Memorandum to R Copland,
EPA/ESD, January 2, 1997.
72. U. S. Environmental Protection Agency, 1993. Emission Factor Documentation for AP-42, Section
2.1, Refuse Combustion, Research Triangle Park, NC.
73. Radian Corporation, 1989. Locating and Estimating Air Toxics Emissions from Municipal Waste
Combustors. EPA-450/2-89-006. U. S. Environmental Protection Agency, Research Triangle Park,
NC. April 1989.
74. K.L. Nebel and D.M. White, 1991. A Summary of Mercury Emissions and Applicable Control
Technologies for Municipal Waste Combustors, U. S. Environmental Protection Agency, Research
Triangle Park, NC. September 1991.
75. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section
2.2, Sewage Sludge Incineration. Research Triangle Park, NC.
76. B. Southworth, EPA, Office of Water Programs, 1996. Number of current incinerators and quantity of
sludge processed. Personal communication to Midwest Research Institute, November 1996.
77. U. S. Environmental Protection Agency, 1990. National Sewage Sludge Survey. Federal Register.
Vol. 55, 47239, November 9, 1990.
78. U. S. Environmental Protection Agency, 1974. Background Information on National Emission
Standards for Hazardous Air Pollutants-Proposed Amendments to Standards for Asbestos and
Mercury. EPA-450/2-74-009A. Research Triangle Park, NC, October, 1974.
79. D. Mallory, Califorinia Air Resources Board, 1997. Personal communication to Midwest Research
Institute, January 1997. Potential for formation of organomercury compounds in activated sewage
sludge.
80. R. Caballero, Los Angeles Sanitation System, 1997. Personal communication to Midwest Research
Institute, January 1997.
81. U. S. Environmental Protection Agency, 1996. Proposed Revised Technical Standards for Hazardous
Waste Combustion Facilities. Federal Register. Vol. 61, 17357, April 19, 1996. Draft Technical
Support Document, February 1996.
82. F. Behan, EPA Office of Solid Waste, 1997. Personal communication to Midwest Research Institute,
April 1997. National emissions estimates for mercury from hazardous waste combustion and number
of facility data.
83. American Society of Mechanical Engineers (ASME), 1988. Hazardous Waste Incineration: A
Resource Document. The ASME Research Committee on Industrial and Municipal Wastes, New
York, NY, January 1988.
84. Midwest Research Institute (MRI), 1992. Medical Waste Incinerators-Background Information for
Proposed Standards and Guidelines: Industry Profile Report for New and Existing Facilities, Draft
Report. Cary,NC. April 1992.
85. U. S. Environmental Protection Agency, 1996. Standards of Performance for New Stationary Sources
and Emission Guidelines for Existing Sources: Medical Waste Incinerators, Proposed Rule, Federal
Register. June 20, 1996.
10-5
-------
86. U. S. Environmental Protection Agency, 1992. Medical Waste Incinerators-Background Paper for
New and Existing Facilities, Draft. Research Triangle Park, NC. June 1992.
87. K. Amos, 1991. Getting Ready for the Mercury Challenge at Municipal Waste Incinerators. Solid
Waste and Power. April 1991.
88. R. Neulicht, M. Turner, L. Chaput, D. Wallace, and S. Smith, 1989. Operation and Maintenance of
Hospital Waste Incinerators, EPA-450/3-89-002. U. S. Environmental Protection Agency, Research
Triangle Park, NC. March 1989.
89. R. Neulicht, L. Chaput, D. Wallace, M. Turner, and S. Smith, 1989. Hospital Incinerator Operator
Training Course: Volume I Student Handbook, EPA-450/3-89-003. U. S. Environmental Protection
Agency, Research Triangle Park, NC. March 1989.
90. Midwest Research Institute (MRI), 1992. Medical Waste Incinerators-Background Information for
Proposed Standards and Guidelines: Control Technology Performance Report for New and Existing
Facilities, Draft Report. Cary, NC. July 1992.
91. M. Turner, (Midwest Research Institute), 1995. Description of General Selection Rules for Medical
Waste Incinerator/Air Pollution Control Device Emission Test Data. Memorandum to R. Copland,
EPA/ESD, September 15, 1995.
92. D. Randall and B. Hardee (Midwest Research Institute), 1996. Emission Factors for Medical Waste
Incinerators. Memorandum to R. Copland, EPA/ESD, April 8, 1996.
93. B. Strong, (Midwest Research Institute), 1996. Acid Gases and Metals Typical Performance and
Achievable Emission Levels for Medical Waste Incinerators with Combustion Controls.
Memorandum to R. Copland, EPA/ESD, May 20, 1996.
94. M. Turner, (Midwest Research Institute), 1996. Wet Scrubber Performance. Memorandum to R.
Copland, EPA/ESD, May 20, 1996.
95. M. Turner, (Midwest Research Institute), 1996. Dry Scrubber Performance. Memorandum to R.
Copland, EPA/ESD, May 20, 1996.
96. Portland Cement Association, 1996. U.S. and Canadian Portland Cement Industry: Plant Information
Summary. Washington, DC. November 1996.
97. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section
11.6, Portland Cement Manufacturing, Research Triangle Park, NC.
98. Research Triangle Institute (RTI), 1996. Mercury Emission from Cement Kilns, Memorandum from
Elizabeth Heath to Mr. Joseph Wood, Emission Standards Division, U. S. Environmental Protection
Agency, Research Triangle Park, NC, March 20, 1996.
99. Gossman Consulting, Inc., 1996. Report to the Portland Cement Associaiton on Mercury Emissions
Data Quality from Testing at Cement Kilns, Draft Report, submitted to Portland Cement Association,
February 1996.
100. M. Miller, 1996. Lime. (In) Minerals Yearbook, Volume 1-Metals and Minerals, U.S. Geological
Survey, U.S. Department of Interior, Washington, D.C.
101. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section
11.17, Lime Manufacturing, Research Triangle Park, NC.
102. J. Wood, 1997. Written communication from Joseph Wood, U.S. Environmental Protection Agency,
Research Triangle Park, NC to Tom Lapp, Midwest Research Institute, Cary, NC, July 28, 1997.
103. National Lime Association (NLA), 1997. Testing of Hazardous Air Pollutants at Two Lime Kilns,
Final Test Report, National Lime Association, Arlington, VA, February, 1997.
10-6
-------
104. B.R. Taylor, 1992. Section 12. Carbon Black. Air Pollution Engineering Manual. Air and Waste
Management Association, Pittsburg, PA. 1992.
105. T.F. Yen, (1975). The Role of Trace Metals in Petroleum, Ann Arbor Science Publishers, Ann
Arbor, MI.
106. R.W. Serth and T.W. Hughes, 1980. Polycyclic Organic Matter (POM) and Trace Element Contents
of Carbon Black Vent Gas. Environmental Science & Technology, Volume 14 (3), pp. 298-301.
107. U. S. Environmental Protection Agency, 1996. Emission Factor Documentation for AP-42, Section
12.2, Coke Production (DRAFT), Research Triangle Park, NC.
108. W.D. Huskonen, 1991. Adding the Final Touches. 33 Metal Producing, Volume 29, pp. 26-28, May
1991.
109. U. S. Environmental Protection Agency, 1986, Compilation of Air Pollution Emission Factors, AP-42,
Fourth Edition. Research Triangle Park, NC.
110. T.W. Easterly, P. E. Stefan, P. Shoup, and D. P. Kaegi, 1992. Section 15. Metallurgical Coke. Air
Pollution Engineering Manual. Air and Waste Management Association, Pittsburg, PA.
111. G.R. Smith, 1996. Lead. (In) Minerals Yearbook, Volume l~Metals and Minerals, U.S. Geological
Survey, U.S. Department of Interior, Washington, D.C.
112. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section
12.6, Primary Lead Smelting, Research Triangle Park, NC.
113. J. Richardson, 1993. Primary lead smelting process information and mercury emission factors.
ASARCO, Inc., Salt Lake City, Utah, facsimile to Midwest Research Institute. August 24, 1993.
114. D.L. Edelstein, 1996. Copper. (In) Minerals Yearbook, Volume 1-Metals and Minerals, U.S.
Geological Survey, U.S. Department of Interior, Washington, D.C.
115. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section
12.3, Primary Copper Smelting, Research Triangle Park, NC.
116. E.P. Grumpier, 1995. Mercury Emissions from Primary Copper Smelters, Memorandum to
A. E. Vervaert, EPA/ESD, September 18, 1995.
117. National Petroleum Refiners Association (NPRA), 1995. United States Refining Capacity.
Washington, DC, June 1995.
118. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section
5.1, Petroleum Refining, Research Triangle Park, NC.
119. J.E. Rucker and R P. Streiter, 1992. Section 17. The Petroleum Industry. Air Pollution Engineering
Manual. Air and Waste Management Association, Pittsburgh, PA.
120. Almega, 1990. AB2588 Pooled Source Emission Test Program, The Almega Corporation Project
16551, The Almega Corporation Report 16551-4. Volume I. Prepared for Western States Petroleum
Association, Glendale, Ca. July 1990.
121. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42,
Section 2.4, Municipal Solid Waste Landfills, Research Triangle Park, NC.
122. A. Geswein, (EPA/ Office of Solid Waste), 1996. Telephone Communication with B. Shrager,
Midwest Research Institute, November 12, 1996.
123. ESCOR, Inc., 1982. Landfill Methane Recovery Part II: Gas Characterization. Final Report.
December 1981 to December 1982. for Gas Research Institute in cooperation with Argonne National
Laboratory, Northfield, IL, December 1982.
10-7
-------
124. R. Myers, 1996. Telephone communication between Ron Myers, U.S. Environmental Protection
Agency, Research Triangle Park, NC, and Brian Shrager, Midwest Research Institute, Gary, NC.
November 1996.
125. D.E. Robertson, E.A. Crecelius, J.S. Fruchter, and J.D. Ludwick, 1977. Mercury Emissions from
Geothermal Power Plants, Science, Volume 196(4294), pp. 1094-1097.
126. M. Reed, U.S. Department of Energy, Geothermal Division, 1993. Location and Capacity Information
on U.S. Geothermal Power Plants. Facsimile to T. Campbell, Midwest Research Institute, February
1993.
127. International Geothermal Association (IGA), 1995. Data on Proposed and Existing Geothermal
Power Plants in the United States. Internet Web Page, http//www.demon.co.uk/geosci/wrusa.html.
128. U. S. Environmental Protection Agency, 1996. Technical Support Document: Chemical Recovery
Combustion Sources at Kraft and Soda Pulp Mills, EPA-453/R-96-012. Office of Air Quality
Planning and Standards, Research Triangle Park, NC, October 1996.
129. A. Someshwar and J. Pinkerton, 1992. Wood Processing Industry. (In) Air Pollution Engineering
Manual, Air and Waste Management Association, NY, 1992, pp. 835-849.
130. U. S. Environmental Protection Agency, 1995. Combustion Sources at Sulfite Pulp Mills—Technical
Support Document for Proposed Standards, Draft Report. June 1995.
131. R. Nicholson, (Midwest Research Institute), 1996. Addendum to Summary of Responses to the 1992
NCASI "MACT" Survey, to J. Telander, EPA/MICG, June 13, 1996.
132. V. Soltis, (Midwest Research Institute), 1995. U.S. Population of Sulfite and Stand-Alone
Semichemical Pulp Mills, April 24, 1995.
133. S. McManus, (Midwest Research Institute), 1996. Industry Profile of Combustion Sources at Stand-
Alone Semichemical Pulp Mills.
134. R. Mcllvane, 1993. Removal of Heavy Metals and Other Utility Air Toxics. Presented at the EPRI
Hazardous Air Pollutant Conference, 1993.
135. T. Hollow ay, (Midwest Research Institute), 1996. Summary of PM and HAP Metals Data. Submitted
to the project files. June 14, 1996.
136. V. Harris, (Midwest Research Institute), 1996. Review Findings of Blank Metals Data. Memorandum
to R. Nicholson, Midwest Research Institute. May 16, 1996.
137. T. Holloway and R. Nicholson (Midwest Research Institute), 1996. Nationwide Costs, Environmental
Impacts, and Cost-Effectiveness of Regulatory Alternatives for Kraft, Soda, Sulfite, and Semichemical
Combustion Sources. Memorandum to J. Telander, EPA/MICG.
138. J. Perwak, et al. (A. D. Little, Inc.), 1981. Exposure and Risk Assessment for Mercury, EPA-440/4-
85-011. Office of Water and Waste Management. U. S. Environmental Protection Agency,
Washington, D.C.
139. W.R. Pierson and W. W. Brachaczek, 1983. Particulate Matter Associated with Vehicles on the Road.
II. Aerosol Science and Technology 2:1-40.
140. C.M. Urban and R.J. Garbe, 1979. Regulated and Unregulated Exhaust Emissions from
Malfunctioning Automobiles. Presented at the Society of Automotive Engineers (SAE) Passenger Car
Meeting, Dearborn, Michigan.
141. M.A. Warner-Selph and J. DeVita, 1989. Measurements of Toxic Exhaust Emissions from Gasoline-
Powered Light-Duty Vehicles. Presented at the Society of Automotive Engineers (SAE) International
Fuels and Lubricants Meeting and Exposition, Baltimore, Maryland.
10-8
-------
142. M.A. Warner-Selph, U. S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, 1993. Personal communication with T. Lapp, Midwest Research Institute. Analytical
detection limits for mercury in the 1989 study. April 1993.
143. Cremation Association of North America (CANA), 1996. 1996 Projections to the Year 2010.
Chicago, IL.
144. California Air Reseouces Board (CARS), 1992. Evaluation Test on Two Propane Fired Crematories
at Camellia Memorial Lawn Cemetary. Test Report No. C-90-004. October 29, 1992.
145. W. Van Horn, 1975. Materials Balance and Technology Assessment of Mercury and Its Compounds
on National and Regional Bases. EPA-560/3-75-007 (NTIS PB-247 000/3). Office of Toxic
Substance, U. S. Environmental Protection Agency, Washington, DC.
146. SPECIATE. Volatile Organic Compound (VOC)/Particulate Matter (PM) Speciation Data System,
Version 1.4. Office of Air and Radiation, Office of Air Quality Planning and Standards, U. S.
Environmental Protection Agency, Research Triangle Park, NC. October 1991.
147. O. Lindqvist, K. Johansson, M. Aatrup, A. Andersson, L. Bringmark, G. Hovsenius, L. Hakanson, A.
Iverfeldt, M. Meili, and B. Timm, 1991. Mercury in the Swedish Environment: Recent Research on
Causes, Consequences and Corrective Methods. Water Air Soil Pollut. 55(1-2): 26-30, 38-39, 65-
70.
148. World Health Organization (WHO), 1976. Environmental Health Criteria 1. Mercury. Geneva.
149. D.H. Klein, 1972. Some Estimates of Natural Levels of Mercury in the Environment. In:
Environmental Mercury Contamination, R Hartung and B. D. Dinman, eds. Ann Arbor Science
Publishers, Inc. Ann Arbor, Michigan.
150. J.O. Nriagu, 1989. A Global Assessment of Natural Sources of Atmospheric Trace Metals. Nature.
Vol.338. March 2, 1989.
151. U. S. Environmental Protection Agency, 1992. EPA Method 101, Determination of Particulate and
Gaseous Mercury Emissions from Chlor-Alkali Plants. 40 Code of Federal Regulations, Part 61,
Appendix B. Washington, D.C.
152. U. S. Environmental Protection Agency, 1996. EPA Method 101A, Determination of Particulate and
Gaseous Mercury Emissions from Stationary Sources. 40 Code of Federal Regulations, Part 61,
Appendix B. Washington, D.C.
153. U. S. Environmental Protection Agency, 1992. EPA Method 102, Determination of Particulate and
Gaseous Mercury Emissions from Chlor-Alkali Plants - Hydrogen Streams. 40 Code of Federal
Regulations, Part 61, Appendix B. Washington, D.C.
154. U. S. Environmental Protection Agency, 1988. EPA Method 0012, Methodology for the
Determination of Metals Emissions in Exhaust Gases from Hazardous Waste Incineration and Similar
Combustion Sources, Test Methods for Evaluating Solid Waste: Physical/Chemical Methods.
SW-846, Third Edition. Office of Solid Waste and Emergency Response. Washington, D.C.,
September 1988.
155. U. S. Environmental Protection Agency, 1990. Methodology for the Determination of Metals
Emissions in Exhaust Gases from Hazardous Waste Incineration and Similar Combustion Sources,
Methods Manual for Compliance with the BIF Regulations Burning Hazardous Waste in Boilers and
Industrial Furnaces. E.P.A./530-SW-91-010. Office of Solid Waste and Emergency Response.
Washington, D.C, December 1990.
156. U. S. Environmental Protection Agency, 1996. EPA Method 29, Methodology for the Determination
of Metals Emissions in Exhaust Gases from Stationary Sources. 40 Code of Federal Regulations, Part
60, Appendix A. Washington, D.C.
10-9
-------
157. CARB Method 436, Undated. Determination of Multiple Metals Emissions from Stationary Sources.
State of California Air Resources Board, Sacramento, CA.
158. U. S. Environmental Protection Agency, 1988. EPA Method 7470, Mercury in Solid or Semisolid
Waste (Manual Cold-Vapor Technique), Test Methods for Evaluating Solid Waste:
Physical/Chemical Methods. SW-846, Third Edition. Office of Solid Waste and Emergency
Response. Washington, D.C. September 1988.
159. U. S. Environmental Protection Agency, 1988. EPA Method 7471, Mercury in Solid or Semisolid
Waste (Manual Cold-Vapor Technique), Test Methods for Evaluating Solid Waste: Physical/
Chemical Methods. SW-846, Third Edition. Office of Solid Waste and Emergency Response.
Washington, D.C. September 1988.
10-10
-------
APPENDIX A
NATIONWIDE EMISSION ESTIMATES
-------
SECTION 4
EMISSIONS FROM MERCURY PRODUCTION
Primary Mercury Production ~
Mercury is no longer mined as a primary product in the United States.
Secondary Mercury Production -
Basis of Input Data
1. In the 1994 TRI summary, mercury emissions were reported for 2 of the 3 major secondary
mercury producers. Mercury Refining Company reported emissions of 116 kg (255 Ib) and
Bethlehem Apparatus Company reported emissions of 9 kg (20 Ib). The third major
company, D.F. Goldsmith, does not reclaim mercury from scrap materials using extractive
processes.
2. Emissions from secondary mercury production are uncontrolled.
Calculation
Total 1994 emissions = 116 kg + 9 kg = 125 kg = 0.125 Mg = 0.13 Mg = 0.14 tons
Mercury Compounds Production -
No emission factors are available for mercury emissions from this process.
A-l
-------
SECTION 5
EMISSIONS FROM MAJOR USES OF MERCURY
Chlorine Production ~
Basis of Input Data
1. Table 5-1 presents two sets of mercury emissions data for mercury-cell chlor-alkali facilities.
The 1991 data are based on Section 114 information collection requests. The 1994 data are
based on voluntary reporting in TRI. Because the totals for the two data sets are essentially
the same (12,902 Ib vs. 12,438 Ib, a difference of less than 4 percent), the TRI data set was
used to calculate emissions because these data represent more recent emission estimates.
2. In the 1994 TRI summary, mercury emissions were reported for 12 of the 14 U.S. mercury
cell facilities.3 Mercury emissions for those 12 facilities totaled 12,438 Ib.
3. Mercury-cell capacity of the 12 facilities reporting mercury emissions totaled 1,750,000 tons
of chlorine.
4. The total number of U.S. chlor-alkali facilities is 14.
5. Total mercury-cell capacity of all 14 U.S. chlor-alkali facilities is 1,998,000 tons of
chlorine.4 SRI figures were adjusted based on Section 114 information collection request
responses.
6. Emission data were prorated for the remaining two facilities.
Calculation
Total 1994 emissions for all 14 chlor-alkali facilities =
= 12,438 Ib x 1>998>00°
1,750,000
= 12,438 Ib x 1.14
= 14,201 Ib
= 7.1 tons or 6.5 Mg
Battery Manufacture ~
Basis of Input Data
1. The 1995 consumption of mercury in the production of primary batteries was less than 0.5
Mg (<0.6 tons).2
2. A mercury emission factor of 1.0 kg/Mg used (2.0 Ib/ton) was obtained from a Wisconsin
study of a mercury oxide battery plant, which is the only type of battery using mercury.5
3. The plant used to develop this emission factor discontinued production of this type of battery
in 1986. This emission factor may be representative of an outdated production process.
A-2
-------
Calculation
Total 1995 emissions = 1.0 kg/Mg x 0.5 Mg = 0.5 kg
0.5 kg = 5 x 10'4 Mg = 6 x 10'4 tons
Electrical Uses
Electric Switches -
Basis of Input Data
1. The 1995 consumption of mercury was 84 Mg (92 tons).2
2. A mercury emission factor of 4 kg/Mg (8 Ib/ton) of mercury consumed for overall electrical
apparatus manufacture was obtained from a 1973 EPA report.1 This factor pertains only to
emissions generated at the point of manufacture.
3. This factor should be used with caution as it is based on engineering judgment and not on
actual test data. In addition, fluorescent lamp production and the mercury control methods
used in the industry have likely changed considerably since 1973. The emission factor may,
therefore, substantially overestimate mercury emissions from this industry.
Calculation
Total 1995 emissions = 92 tons x = 736 Ib or 0.4 tons
ton
= 84 Mg x = 336 kg or 0.3 Mg
Mg
Thermal Sensing Elements ~
No emission factors are available for mercury emissions from this process.
Tungsten Bar Sintering ~
No emission factors are available for mercury emissions from this process.
Copper Foil Production ~
No emission factors are available for mercury emissions from this process.
Fluorescent Lamp Manufacture ~
Basis of Input Data
1. The 1995 consumption of mercury was 30 Mg (33 tons).2
2. A mercury emission factor of 4 kg/Mg (8 Ib/ton) of mercury consumed for overall electrical
apparatus manufacture was obtained from a 1973 EPA report.1 This factor pertains only to
emissions generated at the point of manufacture.
3. This factor should be used with caution as it is based on engineering judgment and not on
actual test data. In addition, fluorescent lamp production and the mercury control methods
A-3
-------
used in the industry have likely changed considerably since 1973. The emission factor may,
therefore, substantially overestimate mercury emissions from this industry.
Calculation
Total 1995 emissions = 33 tons x - = 264 Ib = 0.1 tons
ton
= 30 Mg x = 120 kg = 0.1 Mg
Mg
Fluorescent Lamp Recycling ~
Basis of Input Data
1. Data from a 1994 EPA report indicate that approximately 600 million fluorescent lamps are
disposed each year, with only 2 percent of that number (or 12 million lamps) being recycled
annually.6
2. A mercury emission factor of 0.00088 mg/lamp (or 1.9 x 10"9 Ib/lamp) was obtained from a
1994 test report for one fluorescent lamp crusher.7
3. A large degree of uncertainty is associated with this emission estimate because of the limited
data from which the emission factor was developed.
Calculation
T + i mn/i • • 12 x 106 lamps 8.8 x 10"4 mg 3
Total 1994 emissions = - — x - - = 10.56 x 10 mg
yr lamp
= 10.56 g
= 0.011 kg
= 1.1 x 10"5 Mg
or 1.2 x 10"5 tons
Measurement and Control Instrument Manufacturing
Basis of Input Data
1. In 1995, 43 Mg (47 tons) of mercury were used in all measuring and control instrument
manufacture.2
2. A 1973 EPA report presents an emission factor for overall instrument manufacture of 9
kg/Mg (18 Ib/ton) of mercury consumed.1
3. This emission factor should be used with caution as it is based on survey responses gathered
in the 1960's and not on actual test data. In addition, instrument production and the mercury
control methods used in the industry have likely changed considerably since the time of the
surveys.
A-4
-------
Calculation
1 & 1V\
Total 1995 emission = 47 tons x = 846 Ib = 0.4 tons
ton
43 Mg x --- = 387 kg = 0.4 Mg
Mg
A-5
-------
SECTION 6
EMISSIONS FROM COMBUSTION SOURCES
Coal Combustion
Coal-Fired Utility Boilers ~
Basis of Input Data
1. Develop average mercury emission concentrations for the major coal seams in the USGS
data base and identify these seams with States.
2. Using the UDI/EEI data base of specific boiler configurations, calculate the mercury input to
each boiler by matching coal from States in (1) and multiplying the mercury content of the
coal by the boiler annual coal consumption rate.
3. Adjust the mercury input in (2) for those boilers burning bituminous coal located east of the
Mississippi River as a result of coal cleaning by multiplying the input in (2) by 0.79 (a 21
percent reduction in mercury content).
4. Multiply the resulting mercury input from (2) or (3) by the EMF factor that applies to the
particular boiler. The EMF factors are found in Table B-l, Appendix B.
5. Sum the estimated mercury emissions for each boiler.
6. The total nationwide mercury emission estimate from utility coal-fired boilers is 46.3 Mg/yr
(51 tons/yr).
Coal-Fired Industrial Boilers ~
Basis of Input Data
1. From Table 6-8, emission factor for bituminous coal combustion = 7.0 x 10"15 kg/J and for
anthracite coal combustion = 7.6 x 10"15 kg/J.
2. No control of emissions from industrial boilers was assumed.
3. Energy from coal combustion in industrial sector from Table 6-1.
Calculations
Total 1994 emissions = 7.0 x 10'15 kg/J * 2.892 x 1018 J/yr
+ 7.6 x 10'15 kg/J * 0.009 x 1018 J/yr
= 20.3 Mg = 22.3 tons
Coal-Fired Commercial and Residential Boilers ~
Basis of Input Data
1. From Table 6-8, emission factor for bituminous coal combustion = 7.0 x 10"15 kg/J and for
anthracite coal combustion = 7.6 x 10"15 kg/J.
2. No control of emissions from commercial/residential boilers was assumed.
3. Energy from coal combustion in commercial/residential sectors from Table 6-1.
A-6
-------
Calculations
Total 1994 emissions = 7.0 x 10'15 kg/J * 0.130 x 1018 J/yr
+ 7.6 x 10'15 kg/J * 0.032 x 1018 J/yr
= 1.2 Mg= 1.3 tons
Oil Combustion
Oil-Fired Utility Boilers -
Basis of Input Data
1. From Table 6-15, emission factor for distillate oil combustion = 2.7 x 10"15 kg/J and for
residual oil combustion = 0.2 x 10"15 kg/J.
2. Air pollution control measures assumed to provide no mercury emission reduction.
3. Energy consumption from fuel oil combustion from Table 6-1.
Calculations
Total 1994 emissions = 2.7 x 10'15 kg/J * 0.100 x 1018 J/yr
+ 0.2 x 10'15 kg/J * 0.893 x 1018 J/yr
= 0.45 Mg = 0.49 tons
Oil-Fired Industrial Boilers ~
Basis of Input Data
1. From Table 6-15, emission factor for distillate oil combustion = 2.7 x 10"15 kg/J and for
residual oil combustion = 0.2 x 10"15 kg/J.
2. Air pollution control measures assumed to provide no mercury emission reduction.
3. Energy consumption from fuel oil combustion from Table 6-1.
Calculations
Total 1994 emissions = 2.7 x 10'15 kg/J * 1.169 x 1018 J/yr
+ 0.02 x 10'15 kg/J * 0.448 x 1018 J/yr
= 3.2Mg = 3.6tons
Oil-Fired Commercial/Residential Boilers ~
Basis of Input Data
1. From Table 6-15, emission factor for distillate oil combustion = 2.7 x 10"15 kg/J and for
residual oil combustion = 0.2 x 10"15 kg/J.
2. Air pollution control measures assumed to provide no mercury emission reduction.
3. Energy consumption from fuel oil combustion from Table 6-1.
Calculations
Total 1994 emissions = 2.7 x 10'15 kg/J * 1.417 x 1018 J/yr
+ 0.2 x 10'15 kg/J * 0.184 x 1018 J/yr
= 3.9Mg = 4.3tons
A-7
-------
Wood Combustion
Industrial Boilers -
Basis of Input Data
1. NCASI Technical Bulletin 701 gives an average emission factor for mercury emissions from
wood-fired boilers with ESP's of 1.3 x 10'6 kg/Mg (2.6 x 10'6 Ib/ton) dry wood fuel.
2. Total U.S. wood-fired boiler capacity is assumed to be 1.04 x 1011 Btu/hr, which is the same
rate as 1980.8
3. Heating value of dry wood fuel is 18 x 106 Btu/ton.
4. The U.S. wood consumption rate:
1.04 x 1011 Btu/hr
18 x 106 Btu/ton
= 5,778 tons (dry)/hr
Assuming operation at capacity for 8,760 hours/year, total annual wood consumption =
5,778 tons/year x 8,760 hr/yr = 50,615,280 tons/yr
Calculation
Total 1994 emissions = 50.62 x 106 tons/yr x 2.6 x 10'6 Ib Hg/ton
= 1321bHg/hr
= O.ltonsorO.lMg
Residential Wood Stoves ~
No emission factors are available for mercury emissions from this process.
Residential Fireplaces ~
No emission factors are available for mercury emissions from this process.
Municipal Waste Combustion ~
Basis of Input Data
1. The following average concentrations presented in "National Emissions for Municipal Waste
Combustors" were applied to the inventory of municipal waste combustors (provided in
Appendix B) to determine the nationwide emissions for refused derived fuel (RDF) and non-
RDF combustors:9
A-8
-------
Calculations
1.
Combustor type
Non-RDF without acid gas control
Non-RDF with acid gas control
Non-RDF with acid gas control and carbon
RDF without acid gas control
RDF with acid gas control
Average mercury concentration,
ug/dscm @ 7% O2
340
205
19
260
35
2. The F-factor used for municipal waste combustors was 9,570 dscf/MMBtu at 0 percent
oxygen. Higher heating values were given as 4,500 Btu/lb for unprocessed MSW, and
5,500 Btu/lb for RDF.y
3. Average capacity factors, which represent the percentage of operational time a plant would
operate during a year at 100 percent capacity were presented in the EPA report on mercury
emissions from municipal waste combustors. For all units, except modular/starved-air
combustors, the annual capacity factor was 91 percent (0.91). For modular/starved-air
combustors, the annual capacity factor was 74 percent (0.74).
The F-factor and higher heating values were used to develop volumetric flow factors for non-
RDF and RDF units as follows:
Volumetric flow factor (non-RDF) = (9,750 dscf @ 0%O2/MMBtu) * (4,500 Btu/lb) *
(2,000 Ib/ton) * (20.9/(20.9-7))/(35.31 dscf/dscm)/(106 Btu/MMBtu) = 3,670 dscm @ 7%
O2/ton MSW
Volumetric flow factor (RDF) = (9,750 dscf @ 0%O2/MMBtu) * (5,500
Btu/lb) * (2,000 Ib/ton) * (20.9/(20.9-7))/(35.31 dscf/dscm)/(106 Btu/MMBtu) =
4,457 dscm @ 7% O2/ton RDF
2. The following equation was used to convert the mercury stack concentrations to megagrams
per year for each unit in the municipal waste combustor inventory:
E = CxVxTxCF/10
12
where:
E = annual mercury emissions (Mg/yr)
C = flue gas mercury concentration (ug/dscm @ 7% O2)
V = volumetric flow factor (dscm @ 7% O2/ton waste)
T = MWC unit capacity (ton/year), and
CF = capacity factor (unitless).
The annual mercury emissions from each MWC in the inventory were summed to determine the
nationwide mercury emissions from municipal waste combustors. The total nationwide emissions of
mercury from municipal waste combustors are 26 Mg/yr (29 ton/yr).
A-9
-------
Sewage Sludge Incinerators -
Basis for Input Data
1. Total sludge processed in 1995 was 785,000 Mg (864,000 tons).10
2. From the Draft AP-42, Section 2.2, Sewage Sludge Incineration, an average emission factor
for units with a venturi control device was 0.018 g/Mg (3.5x10 Ib/ton). For other control
devices, the average emission factor was 1.6 g/Mg (3.2 x 10"3 Ib/ton).11
3. In the U.S., there are 166 active sewage sludge incinerators; of this population, 47 use
venturi control devices, 97 use other control devices, and no information was available for
22 units. Of the 144 units for which data are available, 47/144 or 33 percent use venturi
controls and 97/144 or 67 percent use other controls. This percentage distribution is
assumed to be representative for all 166 units.10'11
Calculation
Total 1995 emissions = 785,000 Mg/yr x 0.33 x 0.018 g/Mg + 785,000 x 0.67 x 1.6 g/Mg
= 0.86Mg
= 0.94 tons
Hazardous Waste Combustion ~
Basis of Input Data
1. Mercury national emissions estimate data were obtained from the EPA Office of Solid
Waste Studies for the proposed hazardous waste combustion MACT standards. Details on
the methodologies used to estimate the mercury emissions from hazardous waste
incinerators, cement kilns, and lightweight aggregate kilns may be obtained from docket
materials prepared by the EPA Office of Solid Waste for the proposed hazardous waste
combustion MACT standards.12
2. For 1996, emissions from cement kilns permitted to burn hazardous waste were derived by
EPA for the 41 hazardous waste burning cement kilns in the United States. The national
mercury emissions estimate for cement kilns is 5,860 Ib/yr. This corresponds to 2.66 Mg/yr
(2.93 tons/yr).
3. For 1996, emissions from hazardous waste incinerators were derived by EPA for 190 units
in operation.13 The national mercury emissions estimate for incinerators is 7,700 Ib/yr.
This corresponds to 3.5 Mg/yr (3.95 tons/yr).
4. For 1996, emissions from lightweight aggregate kilns were derived by EPA based on
11 kilns. The national mercury emissions estimate for lightweight aggregate kilns is
156 Ib/yr. This corresponds to 0.07 Mg/yr (0.08 tons/yr).
Calculation
Total annual emissions = 2.7 Mg + 3.5 Mg + 0.07 Mg = 6.27 Mg
= 6.3 Mg = 6.9 tons
Medical Waste Incineration ~
Basis of Input Data
1. The annual emission estimates are based on the calculation procedure employed in
developing the environmental impacts of the emission guidelines for medical waste
incinerators (MWI's). An inventory of existing MWI's was the basis of the emission
calculations for the emission guidelines.
A-10
-------
2. The waste incineration capacity of each MWI was included in the inventory. Waste was
assumed to be charged at two-thirds of the design capacity because average hourly waste
charging rates measured during emissions testing are about two-thirds of the design rate
specified by MWI manufacturers.
3. The type of emissions control at each facility was estimated based on applicable State permit
limits.
4. The annual hours of operation for each MWI was based on the hours of operation for model
plants.
Calculation
1. The annual emissions for each MWI in the inventory was calculated with the following
formula:
Emission (Ib/yr) = CxHxRxF
where, C is the MWI design capacity (Ib/hr), H is the annual charging hours (hr/yr), R is the
ratio of the actual charging rate to the design capacity (2/3), and F is the emission factor for
the appropriate level of control (from Table 6-20).
2. The total emissions from all MWI's in the inventory were calculated by summing the
emissions for each individual unit as shown below14
2, 400
Annual emissions = 2_/ emissions for each MWI i
1=1
= 32,000 Ib/yr = 16.0 tons = 14.5 Mg
A-ll
-------
SECTION 7
EMISSIONS FROM MISCELLANEOUS SOURCES
Portland Cement Manufacturing ~
Basis of Input Data
1. The estimated 1995 total production of clinker from nonhazardous waste fueled kilns was
62.3 x 106 Mg (68.7 x 10 tons). These clinker production levels were estimated using the
same percentage of total clinker production from nonhazardous waste fueled kilns as cited
byRTI.15
2. The average emission factor is 6.5 x 10"5 kg/Mg (1.3 x 10"4 Ib/ton) of clinker produced.15
Calculation
Total 1994 emissions = 62.3 x 106 Mg x 6.5 x 10'5 kg/Mg = 4.0 Mg = 4.4 tons
This mercury emission estimate is for the use of nonhazardous waste as a fuel; emission estimates for
cement kilns burning hazardous waste are presented in Section 6, Hazardous Waste Combustion.
Lime Manufacturing ~
Basis of Input Data
1. The estimated 1994 total production of lime was 17.4 x 106 Mg (19.2 x 106 tons).16
2. An emission factor of 7.4 x 10"6 kg/Mg of lime produced (1.5 x 10"5 Ib/ton) is used for coal-
fired rotary kilns and 1.5 x 10"6 kg/Mg of lime produced (3.0 x 10"6 Ib/ton) for natural gas-
fired vertical kilns.17'18. Natural gas is used to fire 33 percent of the lime kilns.
Calculation
Total 1994 emissions = 17.4 x 106 Mgx 7.4 x 10'6 kg/Mg x 0.67 +17.4 x 106 Mgx 1.5 x 10'6
kg/Mg x 0.33 = 86 kg + 8.6 kg = 95 kg
95 kg = 0.095 Mg = 0.10 tons
Carbon Black Manufacturing ~
Basis of Input Data
1. The mercury emission factor for the main process vent is 0.15 g/Mg (3x10"4 Ib/ton).19
2. The 1995 total annual production capacity of carbon black is 1,660,000 Mg (1,832,500
tons)4
Calculation
The total 1995 emission estimate of mercury from carbon black manufacturing is:
0.15 g/Mg x 1,660,000 Mg/yr = 249,000 g = 0.25 Mg
or
0.00030 Ib/ton x 1,832,500 ton/yr = 550 Ib = 0.28 ton
A-12
-------
By-Product Coke Production ~
Basis of Input Data
1. No mercury emission data are available for U.S. byproduct coke ovens.
2. An emission factor is available for German coke ovens of 6 x 10"5 Ib/ton coke product.20
3. Assume that the U.S. coal cleaning process results in a 20% reduction in mercury emissions
from U.S. byproduct coke ovens (see Section 6.1.4.1). This results in a mercury emission
factor for U.S. coke ovens of 5 x 10"5 Ib/ton coke produced.
4. 1991 total U.S. coke production capacity was 71,649 tons/day.21
5. Assuming operation 365 days/year, 1991 total annual U.S. coke production capacity was
26.15xl06tons.
Calculation
Total 1991 emissions = 26.15 x 106 tons coke * 5 x 10"5
ton coke
= 1,308 Ib
= 0.65 tons or 0.59 Mg
Primary Lead Smelting ~
Basis of Input Data
1. Based on background information in the NSPS for lead smelters, 100 units of ore yields 10
units of ore concentrate, 9 units of sinter, and 4.5 units of refined lead.22
2. The estimated 1994 lead in ore concentrate quantity was 3.7 x 105 Mg (4.07 x 105 tons).23
3. Recent data from lead ore mines indicates that the mercury content of lead ore concentrate is
less than 0.2 ppm.24 It is assumed that the particulate emissions from the process have the
same mercury concentration as the lead ore concentrate (i.e., no concentrating of the mercury
occurs). A mercury concentration of 0.2 ppm is used as an upper limit value. Based on this
concentration, the mercury content is estimated to be 0.4 x 10"3 Ib Hg per ton of ore
concentrate.
4. The mercury emission factors from AP-42 for three emission sources in the process are:
a. sinter machine (weak gas): 0.051 kg/Mg (0.10 Ib/ton) of sinter produced
b. sinter building fugitives: 0.118 kg/Mg (0.24 Ib/ton) of sinter produced
c. blast furnace = 0.21 kg/Mg (0.43 Ib/ton) of bullion
Calculation
Emissions from sinter machine (weak gas):
0.1 Ib/ton * 4.07 x 105 tons * 1/0.9 * 0.4 x 10'3 = 18.1 Ib Hg = 8.23 kg
A-13
-------
Emissions from sinter building fugitives:
0.24 Ib/ton * 4.07 x 105 tons * 1/0.9 * 0.4 x 10'3 = 43.4 Ib Hg = 19.73 kg
Emissions from blast furnace:
0.43 Ib/ton * 4.07 x 105 tons * 1/0.45 * 0.4 x 10'3 = 155.6 Ib Hg = 70.73 kg
Total 1994 emissions:
18.1 Ib + 43.4 Ib + 155.6 Ib = 217. lib = 0.11 tons = 0.10 Mg
Primary Copper Smelting -
Basis of Input Data
1. In 1993, the Emission Standards Division requested all eight of the primary copper smelters
in operation for data on mercury emissions.
2. With the exclusion of Copper Range, which is closed, the total of the self-reported values for
mercury emissions in 1993 was 0.055 Mg (0.06 tons).25
3. In 1994, smelter production from domestic and foreign ores increased 3.15 percent over
1993 production26
Calculation
Total 1994 emissions = 0.055 Mg x 1.0315 = 0.057 Mg = 0.063 tons
Petroleum Refining ~
No reliable emission factors are available for mercury emissions.
Municipal Solid Waste Landfills ~
Basis of Input Data
1. The average mercury concentration in landfill gas is 2.9 x 10"4 ppmV.27'28
2. Methane emissions from landfills in 1994 totaled 10.2 x 106 Mg (11.2 x 106 tons) 29
3. The methane gas volume was:
Volume = 2.24 x 1010 Ib/yr x 1/16 (Ib mole/lb methane) x 385.3 dscf/lb mole = 5.394 x 1011 dscf/yr
4. The total landfill gas volume is twice the methane volume, or 1.079 x 1012 dscf/yr.29
Calculation
Total 1994 emissions = 2.9 x 10'4 ppm x 10'6 x 1.079 x 1012 dscf/yr x 200.59 Ib Hg/lb mole x
lib mole/3 85.3 dscf
= 162.9 Ib Hg = 0.081 tons = 0.074 Mg
A-14
-------
Geothermal Power Plants ~
Basis of Input Data
1. The mercury emission factors for geothermal power plants are:30
Off-gas ejectors: 0.00725 g/MWe/hr (0.00002 Ib/MWe/hr)
and
Cooling tower exhaust: 0.05 g/MWe/hr (0.0001 Ib/MWe/hr)
2. The total annual capacity (MW) of U. S. geothermal power plants in 1993 was 2,653
MW.31'32
3. Assumption: All plants operate at capacity, 24 hrs per day, 365 days per year.
4. Based on the above assumption, annual capacity in MW hr is:
2,653 MW x 24 hr/day x 365 days/yr = 2.32 x 107 MW hr
Calculation
The total 1993 emission estimate of mercury from geothermal power plants based on the above
capacity data and assumption is:
2.32 x 107 MW hr x (0.00725 + 0.05) g/MWe/hr x 10'6 Mg/g= 1.3 Mg = 1.4 tons
Pulp and Paper Production ~
Basis of Input Data
1. The nationwide daily black liquor solids firing rate for kraft and soda recovery furnaces is
2.36 X 105 Mg/d (2.60 x 105 tons/d).33 The same firing rate also applies to kraft and soda
SDT's, which are associated with the recovery furnaces. The nationwide daily lime
production rate for kraft and soda lime kilns is 3.76 x 104 Mg/d (4.15 x 104 tons/d).34
2. Kraft and soda combustion sources nationwide are assumed to operate 24 hr/d for 351 d/yr.
This operating time accounts for 14 days of scheduled shutdown annually for maintenance
and repair.
3. The chemical recovery areas at kraft and soda pulp mills are considered sufficiently similar
to justify applying the mercury emission factors for the kraft combustion sources to the soda
combustion sources. No information is available on mercury emission factors for sulfite or
stand-alone semichemical pulp mills, and the two processes are sufficiently different from
the kraft process that the mercury emission factors for the kraft combustion sources were not
applied to the sulfite and semichemical combustion sources. Therefore, mercury emissions
for the sulfite and semichemical combustion sources will not be included in the nationwide
mercury emission estimate.
4. The average mercury emission factor for kraft and soda recovery furnaces is
1.95 x 10 kg/Mg (3.90 x 10"5 Ib/ton) of black liquor solids fired. The average mercury
emission factor for kraft and soda SDT's is 2.61 x 10'8 kg/Mg (5.23 x 10'8 Ib/ton) of black
liquor solids fired. The average mercury emission factor for kraft and soda lime kilns is
1.46 x 10'6 kg/mg (2.91 x 10* Ib/ton) of lime produced.35
Calculation
Emissions from kraft and soda recovery furnaces = 1.95 x 10"5 kg/Mg * 2.36 x 105 Mg/d * 351 d/yr
= 1.62 x 103 kg/yr = 1.62 Mg/yr
A-15
-------
Emissions from kraft and soda SDT's = 2.61 x 10'8 kg/Mg * 2.36 x 105 Mg/d * 351 d/yr =
2.16kg/yr = 0.00216Mg/yr
Emissions from kraft and soda lime kilns = 1.46 x 10'6 kg/Mg * 3.76 x 104 Mg/d * 351 d/yr:
19.3 kg/yr = 0.0193 Mg/yr
Total 1994 emissions from kraft and soda combustion sources = 1.62 Mg + 0.00216 Mg +
0.0193 Mg = 1.64 Mg = 1.81 ton
A-16
-------
SECTION 8
EMISSIONS FROM MISCELLANEOUS FUGITIVE AND AREA SOURCES
Mercury Catalysts ~
No data are available for any quantities of mercury used for catalytic purposes. Zero emissions have
been assumed.
Dental Alloys ~
Basis of Input Data
1. In 1995, the total use of mercury in dental equipment and supplies was 32 Mg (35 tons).2
2. It has been estimated that 2.0 percent of the mercury used in dental applications is emitted
to the atmosphere. This would correspond to an emission factor of 20 kg/Mg (40 Ib/ton)
of mercury used.
Calculation
Total 1995 emissions = 32 Mg x 20 kg/Mg = 0.64 Mg = 0.70 tons
Mobile Sources -
No reliable emission factors are available for mercury emissions from mobile sources.
Crematories ~
Basis for Input Data
1. In 1995, there were 488,224 cremations in the U.S.37
2. Only one set of data are available for the average quantity of mercury emitted for a
cremation in the U.S. The estimated average emission factor is 1.5 x 10 kg
(3.3 x 10"3 Ib) per cremation. This emission factor will be used for estimations for the
U.S.
Calculation
Total 1995 emissions = — x 488,224 cremations
cremation
= 0.732 kg = 0.73 Mg = 0.80 tons
Paint Use ~
All registrations for mercury-based biocides in paints were voluntarily canceled by the registrants in
May 1991. Based on the voluntary cancellation, it is assumed that mercury emissions from this
source are very small or zero.
Soil Dust --
There are no emission factors for mercury emissions from soil dust.
Natural Sources of Mercury Emissions ~
There are no emission factors for mercury emissions from natural sources.
A-17
-------
REFERENCES FOR APPENDIX A
1. Anderson, D., 1973. Emission Factors for Trace Substances. EPA-450/2-72-001. U. S. Environmental
Protection Agency, Research Triangle Park, NC.
2. Plachy, J., 1996. Mercury. (In) Minerals Yearbook, Volume l~Metals and Minerals, U.S. Geological
Survey, U.S. Department of Interior, Washington, D.C.
3. U. S. Environmental Protection Agency, 1996. Toxics Release Inventory, 1994 TRI data, Office of
Pollution Prevention and Toxics (OPPT), Washington, D.C.
4. SRI International, 1996. Directory of Chemical Producers: United States of America. SRI
International, Menlo Park, CA.
5. Wisconsin Bureau of Air Management (WBAM), 1986. Mercury Emissions to the Atmosphere in
Wisconsin. Publication No. PUBL-AM-014. Wisconsin Department of Natural Resources, Madison,
WI.
6. Battye, W., U. McGeough, and C. Overcash. (EC/R, Inc.), 1994. Evaluation of Mercury Emissions
from Fluorescent Lamp Crushing. EPA-453/R-94-018. U. S. Environmental Protection Agency,
Research Triangle Park, NC.
7. Environmental Enterprises, Inc., 1994. Emissions Testing for Particulate Matter and Mercury. Prepared
for USA Lights of Ohio, Cincinnati, OH.
8. U. S. Environmental Protection Agency, 1982. Background Information for New Source Performance
Standards: Nonfossil Fuel Fired Industrial Boilers. Draft EIS. EPA-450/3-82-007. Office of Air
Quality Planning and Standards, Research Triangle Park, NC.
9. U. S. Environmental Protection Agency, 1996. National Emissions for Municipal Waste Combustors,
Office of Air Quality Planning and Standards, Research Triangle Park, NC, October. 1996.
10. Southworth, B., EPA, Office of Water Programs, 1996. Number of current incinerators and quantity of
sludge processed. Personal communication to Midwest Research Institute, November 1996.
11. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section 2.2,
Sewage Sludge Incineration. Research Triangle Park, NC.
12. Behan, F., EPA Office of Solid Waste, 1997. Personal communication to Midwest Research Institute,
April 1997. National emission estimates for mercury from hazardous waste combustion and number of
facility data.
13. U. S. Environmental Protection Agency, 1996. Proposed Revised Technical Standards for Hazardous
Waste Combustion Facilities. Federal Register. Vol. 61, 17357, April 19, 1996. Draft Technical
Support Document, February 1996.
14. Hardee, B., and Hanks, K., (MRI) 1996. Memorandum to Copland, R., EPA/ESD. May 20, 1996. Air
Emission Impacts of the Regulatory Options for Medical Waste Incinerators (MWI's).
15. Research Triangle Institute (RTI), 1996. Mercury Emission from Cement Kilns, Memorandum from
Elizabeth Heath to Mr. Joseph Wood, Emission Standards Division, U. S. Environmental Protection
Agency, Research Triangle Park, NC, March 20, 1996.
16. Miller, M. M., 1996. Lime. (In) Minerals Yearbook, Volume 1-Metals and Minerals, U.S. Geological
Survey, U.S. Department of Interior, Washington, D.C.
17. Wood, J., 1997. Written communication from Joseph Wood, U.S. Environmental Protection Agency,
Research Triangle Park, NC to Tom Lapp, Midwest Research Institute, Gary, NC, July 28, 1997.
18. National Lime Association (NLA), 1997. Testing of Hazardous Air Pollutants at Two Lime Kilns, Final
Test Report, National Lime Association, Arlington, VA, February 1997.
A-18
-------
19. Serth, R.W., and T.W. Hughes, 1980. Polycyclic Organic Matter (POM) and Trace Element Contents of
Carbon Black Vent Gas. Environmental Science & Technology, Volume 14 (3), pp. 298-301.
20. Jockel, W. and J. Hartje, 1991. Batenerhebung Uber die Emissionen Unweltgefahrdenden
Schwermetalle Forschungsbericht 91-104-02 588, TUV, Rheinland eV. Koln, Germany.
21. Huskonen, W. D., 1991. Adding the Final Touches. 33 Metal Producing, Volume 29, pp. 26-28, May
1991.
22. U. S. Environmental Protection Agency, 1974. Background Information on National Emission
Standards for Hazardous Air Pollutants-Proposed Amendments to Standards for Asbestos and
Mercury. EPA-450/2-74-009A. Research Triangle Park, NC, October, 1974.
23. Smith, G.R., 1996. Lead. (In) Minerals Yearbook, Volume 1--Metals and Minerals, U.S. Geological
Survey, U.S. Department of Interior, Washington, D.C.
24. Richardson,!., 1993. Primary lead smelting process information and mercury emission factors.
ASARCO, Inc., Salt Lake City, Utah, facsimile to Midwest Research Institute. August 24, 1993.
25. Grumpier, E. P., 1995. Mercury Emissions from Primary Copper Smelters, Memorandum to
A. E. Vervaert, EPA/ESD, September 18, 1995.
26. Edelstein, D. L., 1996. Copper. (In) Minerals Yearbook, Volume 1-Metals and Minerals, U.S.
Geological Survey, U.S. Department of Interior, Washington, D.C.
27. ESCOR, Inc., 1982. Landfill Methane Recovery Part II: Gas Characterization. Final Report. December
1981 to December 1982. for Gas Research Institute in cooperation with Argonne National Laboratory,
Northfield, IL, December 1982.
28. Myers, R., 1996. Telephone communication between Ron Myers, U.S. Environmental Protection
Agency, Research Triangle Park, NC, and Brian Shrager, Midwest Research Institute, Gary, NC.
November 1996.
29. U. S. Environmental Protection Agency, 1995. Emission Factor Documentation for AP-42, Section 2.4,
Municipal Solid Waste Landfills, Research Triangle Park, NC.
30. Robertson, D.E., E.A. Crecelius, J.S. Fruchter, and J.D. Ludwick, 1977. Mercury Emissions from
Geothermal Power Plants, Science. Volume 196(4294), pp. 1094-1097.
31. Reed, M. (Department of Energy, Geothermal Division), 1993. Location and Capacity Information on
U.S. Geothermal Power Plants. Facsimile to T. Campbell, Midwest Research Institute, February 1993.
32. International Geothermal Association (IGA), 1995. Data on Proposed and Existing Geothermal Power
Plants in the United States. Internet Web Page, http//ww.demon.co.uk/geosci/wrusa.html.
33. Randall, D., Jones, R, and Nicholson, R, (MRI) 1995. Memorandum to Telander, J., EPA/MICG.
April 25, 1995. Nationwide Baseline Emissions for Combustion Sources at Kraft and Soda Pulp Mills.
34. Holloway, T., (MRI) 1996. Memorandum to the project files. November 7, 1996. Nationwide Lime
Production Rates for Kraft and Soda Lime Kilns.
35. Holloway, T., (MRI) 1996. Memorandum to the project files. June 14, 1996. Summary of PM and
HAP Metals Data.
36. Perwak, J., et al. (A.D. Little, Inc.), 1981. Exposure and Risk Assessment for Mercury, EPA-440/4-85-
011. Office of Water and Waste Management, U. S. Environmental Protection Agency, Washington,
D.C.
A-19
-------
37. Cremation Association of North America (CANA), 1996. 1996 Projections to the Year 2010. Chicago,
IL.
38. California Air Resources Board (CARB), 1992. Evaluation Test on Two Propane-Fired Crematories at
Camellia Memorial Lawn Cemetery. Test Report No. C-90-004. October 29, 1992.
A-20
-------
APPENDIX B
SUMMARY OF COMBUSTION SOURCE MERCURY EMISSION DATA
-------
TABLE B-l. EMISSION MODIFICATION FACTORS FOR UTILITY
BOILER EMISSION ESTIMATESa
Type of APCD or boiler
EMF factor
Fabric filter
Spray dryer adsorber (includes a fabric filter)
Electrostatic precipitator (cold-side)
Electrostatic precipitator (hot-side)
Electrostatic precipitator (oil-fired unit)
Particulate matter scrubber
Fluidized gas desulfurization scrubber
Circulating fluidized bed scrubber
Cyclone-fired boiler without NOX control (wet bottom, coal-fired)
Front-fired boiler without NOX control (dry bottom, coal-fired)
Front-fired boiler without NOX control (dry bottom, gas-fired)
Tangential-fired boiler without NOX control (before a hot-side ESP, coal-fired)
Tangential-fired boiler with NOX control (before a hot-side ESP, coal-fired)
Front-fired boiler without NOX control (dry bottom, oil-fired)
Front-fired boiler with NOX control (dry bottom, oil-fired)
Opposed-fired boiler without NOX control (dry bottom oil-fired)
Tangentially-fired boiler without NOX control (dry bottom, oil-fired)
Tangentially-fired boiler with NOX control (dry bottom, oil-fired)
Opposed-fired boiler with NOX control (dry bottom, coal-fired)
Front-fired boiler without NOX control (wet bottom, coal-fired)
Tangentially-fired boiler without NOX control (dry bottom, coal-fired)
Tangentially-fired boiler with NOX control (dry bottom, coal-fired)
Vertically-fired boiler with NOX control (dry bottom, coal-fired)
0.626
0.701
0.684
1.000
0.315
0.957
0.715
1,000
0.856
0.706
1.000
1.000
0.748
1.000
1.000
0.040
1.000
1.000
0.812
0.918
1.000
0.625
0.785
aTo calculate mercury control efficiency for a specific boiler/control device configuration, the EMF is
subtracted from 1.
Source: Mercury Study Report to Congress, Volume II: An Inventory of Anthropogenic Mercury Emissions
in the United States. EPA-452/8-96-001b. June 1996.
B-l
-------
TABLE B-2. SUMMARY OF MUNICIPAL WASTE COMBUSTOR DATA
Unit name
Juneau RRF
Sitka WTE Plant
Huntsville Refuse-Fired Steam
Fac.
Tuscaloosa Solid Waste Fac.
Batesville
Blytheville Incinerator
North Little Rock RRF
Osceola
Stuttgart Incinerator
Commerce Refuse-to-Energy Fac.
Lassen Community College
Long Beach (SERRF)
Modesto
Bridgeport RESCO
Bristol RRF
Lisbon RRF
Mid-Connnecticut Project
Southeastern Connecticut RRF
Stamford I
Stamford II Incinerator
Town of New Canaan Volume
Reduction Plant
Wallingford RRF
Windham RRF
Solid Waste Reduction Center
No.l
Kent
Pigeon Point
Sussex
Bay Resource Mgt. Center
Broward Co. RRF North
Broward Co. RRF South
Dade Co. RRF
Dade Co. RRF Expansion
Hillsborough Co. RRF
Lake Co. RR
Lee Co. RRF
Mayport NAS
McKay Bay REF
Miami International Airport
North Co. Region RR Project
Pasco Co. Solid Waste RRF
Southernmost WTE
Wheelabrator Pinellas RRF
Location
Juneau
Sitka
Huntsville
Tuscaloosa
Batesville
Blytheville
North Little Rock
Osceola
Stuttgart
Commerce
Susanville
Long Beach
Crows Landing
Bridgeport
Bristol
Lisbon
Hartford
Preston
Stamford
Stamford
New Canaan
Wallingford
Windham
Washington
Wilmington
Panama City
Pompano Beach
Pompano Beach
Miami
Miami
Tampa
Okahumpka
Fort Myers
Mayport NAS
Tampa
Miami
West Palm Beach
Hudson
Key West
St. Petersburg
State
AK
AK
AL
AL
AR
AR
AR
AR
AR
CA
CA
CA
CA
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
DC
DE
DE
DE
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
Project
status
OP
OP
OP
IA
OP
OP
IA
OP
OP
OP
IA
OP
OP
OP
OP
uc
OP
OP
IA
IA(1994)
OP
OP
IA
IA
On Hold
IA
On Hold
OP
OP
OP
OP
On Hold
OP
OP
UC
OP
OP
OP
OP
OP
OP
OP
Total plant
capacity,
tons/d
70
50
690
300
100
70
100
50
63
380
100
1,380
800
2,250
650
500
2,000
600
150
360
125
420
108
1,000
1,800
600
600
510
2,250
2,250
3,000
750
1,200
528
1,200
50
1,000
60
2,000
1,050
150
3,000
No. of
units
2
1
2
4
2
2
4
2
5
1
3
2
3
2
3
2
1
1
1
3
3
4
5
2
3
3
4
3
2
2
1
4
1
2
3
2
3
Combustor type
MOD/SA
MOD/EA
MB/WW
MOD/SA
MOD/SA
MOD/SA
MOD/SA
MOD/SA
MOD/SA
MB/WW
MOD
MB/WW
MB/WW
MB/WW
MB/WW
MB/WW
RDF
MB/WW
MB/REF
MB/REF
MB/REF
MOD/EA
MOD/SA
MB/REF
MB
MOD
MB/RC
MB/WW
MB/WW
RDF
MB/WW
MB/WW
MB/WW
MOD/EA
MB/REF
MOD/SA
RDF
MB/WW
MB/WW
MB/WW
Air pollution
control devices
ESP
ESP DSI
FFSD
ESP
None
None
None
None
None
FF SD SNCR
FFDSI
FF SD SNCR
FF SD SNCR
FFSD
FFSD
FF SD SNCR
FFSD
FFSD
ESP
ESP
WS
FFSD
FFSD
ESP
None
ESP
None
ESP
FFSD
FFSD
ESP
FF SD SNCR CI
ESP
FFSD
FF SD SNCR CI
Cyc
ESP
None
ESPSD
FFSD
ESP
ESP
B-2
-------
TABLE B-2. (continued)
Unit name
Savannah RRF
Honolulu Resource Recovery
Venture
Waipahu Incinerator
Burley
Beardstown
Havana WTE Fac.
Northwest WTE
Robbins RRF
West Suburban Recycling and
Energy Center
Bloomington
Indianapolis RRF
Kentucky Energy Assoc.
Louisville Energy Generating
Fac.
Louisville Incinerator
Fall River Incinerator
Framingham
Haverhill Lawrence RDF
Haverhill RRF
Montachusetts RRF
North Andover RESCO
Pittsfield RRF
Saugus RESCO
SEMASS RRF
Units 1 & 2
SEMASS RRF
Unit3
Springfield RRF
Wheelabrator Millbury
Harford Co. WTE Fac.
Montgomery Co. North RRF
Unit #2
Montgomery Co. North RRF
Unit #3
Montgomery Co. RRF
Montgomery Co. South RRF
Unit #2
Montgomery Co. South RRF
Unit #3
Pulaski
Southwest RRF (RESCO)
Frenchville
Greater Portland Region RRF
Maine Energy Recovery
Location
Savannah
Honolulu
Honolulu
Burley
Beardstown
Havana
Chicago
Robbins
Summit
Bloomington
Indianapolis
Corbin
Louisville
Louisville
Fall River
Framingham
Lawrence
Haverhill
Shirley
North Andover
Pittsfield
Saugus
Rochester
Agawan
Millbury
Aberdeen Proving
Grounds
Dickerson
Baltimore
Baltimore
Frenchville
Portland
Biddeford - Saco
State
GA
HI
HI
ID
IL
IL
IL
IL
IL
IN
IN
KY
KY
KY
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MD
MD
MD
MD
MD
MD
MD
MD
ME
ME
ME
Project
status
OP
OP
IA
OP
P
P
OP
P
P
On Hold
OP
P
On Hold
IA
OP
IA
OP
OP
UC
OP
OP
OP
OP
OP
OP
OP
OP
OP
OP
uc
OP
OP
OP
OP
IA
OP
OP
Total plant
capacity,
tons/d
500
2,160
600
50
1,800
1800
1,600
1,600
1,800
300
2,362
500
250
100
600
500
710
1,650
243
1,500
240
1,500
1,800
900
360
1,500
360
300
300
1,800
300
300
1,500
2,250
50
500
600
No. of
units
2
2
2
1
4
2
3
4
2
2
1
2
2
2
2
2
1
3
2
4
1
1
1
1
5
3
1
2
2
Combustor type
MB/WW
RDF
MB/REF
MOD/SA
RDF
RDF
MB/WW
RDF/FB
RDF/WW
MB
MB/WW
MB
RDF/FB
Unknown
MB/REF
MB/REF
RDF
MB/WW
MB/WW
MB/WW
MOD/EA
MB/WW
RDF
MOD
MB/WW
MOD/SA
MB/RC/REF
MB/RC/REF
MB/WW
MB/RC/REF
MB/RC/REF
MB/REF
MB/WW
Unknown
MB/WW
RDF
Air pollution
control devices
ESP FF(r) SD(r)
ESPSD
ESP
None
FF SD SNCR
FF SD SNCR
ESP
FF SD SNCR
FF SD SNCR
FF SD SNCR
FFSD
Cyc FF SNCR
WS
WS
FFSD
ESP FSI(r)
ESPSD
FF SD SNCR CI
ESP FSI(r)
ESPWS
FF(r) SD(r)
ESPSD
FF SD SNCR
FFDSI
ESPSD
ESP
ESP FSI
ESP FSI
FF SD SNCR CI
ESP FSI
ESP FSI
ESP
ESP
None
ESPSD
FFSD
B-3
-------
TABLE B-2. (continued)
Unit name
Mid Maine Waste Action Corp.
Penobscot Energy Recovery
Comp.
Central Wayne Co. Sanitation
Auth
Clinton Township
Greater Detroit RRF Unit #1
Greater Detroit RRF Unit #2
Greater Detroit RRF Unit #3
Jackson Co. RRF
Kent Co. WTE Fac.
Oakland Co. WTE Fac.
Elk River FFR
Fergus Falls
Hennepin Energy Recovery
Facility
Olmstead WTE Facility
Perham Renewable RF
Polk Co. Solid Waste Resource
Recovery
Pope-Douglas Solid Waste
Ramsey- Washington
Red Wing Solid Waste Boiler
Facility
Richards Asphalt Co. Facility
Western Lake Superior Sanitary
District
Wilmarth Plant
Ft Leonard Wood RRF
St Louis WTE
Pascagoula Energy Recovery
Facility
Livingston/Park County MWC
Carolina Energy Corp
Fayetteville RRF
New Hanover Co. WTE Unit 1 &
2
New Hanover Co. WTE Unit 3
NIEHS
University City RRF
Wrightsville Beach Incinerator
Lamprey Regional SW Coop.
Pittsfield Incinerator
SES Claremont RRF
Wheelabrator Concord
Camden RRF
Essex Co. RRF
Location
Auburn
Orrington
Dearborn Heights
Clinton Township
Detroit
Jackson
Grand Rapids
Auburn Hills
Anoka
Fergus Falls
Minneapolis
Rochester
Perham
Fosston
Alexandria
Red Wing
Red Wing
Scott
Duluth
Mankato
Ft Leonard Wood
St Louis
Moss Point
Park County
Kinston
Fayetteville
Wilmington
RTF
Charlotte
Wrightsville
Beach
Durham
Pittsfield
Claremont
Concord
Camden
Newark
State
ME
ME
MI
MI
MI
MI
MI
MI
MI
MI
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MO
MO
MS
MT
NC
NC
NC
NC
NC
NC
NC
NH
NH
NH
NH
NJ
NJ
Project
status
OP
OP
OP
OP
OP
OP
OP
OP
OP
On Hold
OP
OP
OP
OP
OP
OP
OP
OP
OP
OP
OP
OP
IA
P
OP
OP
P
uc
OP
OP
OP
OP
IA
OP
IA
OP
OP
OP
OP
Total plant
capacity,
tons/d
200
700
500
600
1,100
1,100
1,100
200
625
2,000
1,500
94
1,200
200
114
80
72
720
72
70
260
720
78
1,200
150
72
600
600
200
249
40
235
50
132
48
200
500
1,050
2,277
No. of
units
2
2
2
2
1
1
1
2
2
3
2
2
2
2
2
2
2
2
1
2
2
3
2
2
1
2
2
1
2
2
2
3
2
2
2
3
3
Combustor type
MB
RDF
RDF
MB/REF
RDF
RDF
RDF
MB/WW
MB/WW
MB
RDF
MOD/SA
MB/WW
MB/WW
MOD/SA
MOD/SA
MOD/EA
RDF
MOD/EA
MOD
RDF
RDF
MOD/SA
MOD/EA
MOD/SA
RDF
RDF/FB
MB/WW
MB/WW
MOD/SA
MB/WW
MOD/SA
MOD/EA
MOD/SA
MB/WW
MB/WW
MB/WW
MB/WW
Air pollution
control devices
FFSD
FFSD
ESP
ESP
FF(r) SD(r)
FF(r) SD(r)
FF(r) SD(r)
FFSD
FFSD
FF SD CI
FFDSI
WS
FF SD SNCR
CI(r)
ESP
ESP
ESP
ESP
ESP
ESP
ESP
vs
FF(r) SD(r)
None
FF SD SNCR
ESP
None
FF DSI SNCR CI
DSI SNCR CI
ESP SD(r)
FF SD SNCR
None
ESP
None
Cyc
None
FFDSI
FFDSI
ESP SD CI
ESP SD CI
B-4
-------
TABLE B-2. (continued)
Unit name
Fort Dix RRF
Gloucester County
Union Co. RRF
Warren Energy RF
Adirondack RRF
Albany Steam Plant
Babylon RRF
Belts Ave. Incinerator
Cattaraugus Co. WTE Plant
Dutchess Co. RRF
Glen Cove
Green Island WTE Plant
Green Point Incinerator
Hempstead
Henry St. Incinerator
Huntington RRF
Kodak RRF
Long Beach RRF
MacArthur WTE
MER Expansion
Monroe Co. RRF
Niagara Falls RDF WTE
Oceanside RRF
Oneida Co. ERF
Onondaga Co. RRF
Oswego Co. WTE
Port of Albany WTE Fac.
South West Brooklyn Incinerator
Westchester RESCO
Akron Recycle Energy System
City of Columbus SW Reduction
Fac.
Euclid
Mad River RRF
Montgomery Co. North RRF
Unit#l
Montgomery Co. South RRF
Unit#l
Miami RRF
Walter B. Hall RRF
Coos Bay Incinerator
Location
Wrightstown
Westville
Railway
Oxford Township
Hudson Falls
Albany
Babylon
Queens
Cuba
Poughkeepsie
Glen Cove
Green Island
Green Point
Westbury
Brooklyn
Huntington
Rochester
Long Beach
Islip/Ronkonkoma
Islip/Ronkonkoma
Rochester
Niagara Falls
Oceanside
Rome
Jamesville
Fulton
Port of Albany
Brooklyn Bay 4 1st
St.
Peekskill
Akron
Columbus
Euclid
Springfield
Dayton
Dayton
Miami
Tulsa
Coquille
State
NJ
NJ
NJ
NJ
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
OH
OH
OH
OH
OH
OH
OK
OK
OR
Project
status
OP
OP
OP
OP
OP
IA
OP
IA
IA
OP
IA
P
IA
OP
IA
OP
OP
OP
OP
On Hold
IA
OP
IA
OP
uc
OP
P
IA
OP
IA
IA
IA
IP
OP
OP
OP
OP
OP
Total plant
capacity,
tons/d
80
575
1,440
400
432
600
750
1,000
112
400
250
1,500
100
2,505
750
150
200
518
350
2,000
2,200
750
200
990
200
1,300
960
2,250
1,000
2,000
200
1,750
300
300
105
1,125
125
No. of
units
4
2
3
2
2
2
2
4
3
2
2
3
3
1
1
2
2
4
3
4
4
3
3
6
2
1
1
3
3
3
Combustor type
MOD/SA
MB/WW
MB/WW
MB/WW
MB/WW
RDF
MB/WW
MB/REF
MOD/SA
MB/RC
MB/WW
MB
Unknown
MB/WW
Unknown
MB
RDF
MB/WW
MB/RC
MB
RDF
RDF
MB/WW
MOD/SA
MB/WW
MOD/SA
MB
MB/REF
MB/WW
RDF
RDF
MB/REF
MB/WW
MB/RC/REF
MB/RC/REF
MOD/SA
MB/WW
MOD/SA
Air pollution
control devices
FF WS CI
FF SD CI
FF SD SNCR CI
FFSD
ESPSD
ESP
FFSD
ESP
None
FFDSI
FF(r) DSI
FF SD SNCR
ESP
FFSD
ESP
FF SD SNCR
ESP
ESP
FFDSI
FF
None
ESP
ESP
ESP
FF SD SNCR CI
ESP
FF SD SNCR CI
FF(r) DSI(r)
SD(r) SNCR(r)
CI(r)
ESP FSI (r)
ESP
ESP FF(r) SD(r)
ESP
FF SD SNCR CI
ESP FSI
ESP FSI
None
ESP
None
B-5
-------
TABLE B-2. (continued)
Unit name
Marion Co. WTE
Delaware Co. RRF
Glendon RR Project
Harrisburg WTE
Lancaster Co. RRF
Montgomery Co. RRF
Philadelphia EC
Philadelphia NW
Potter Co. RR
Westmoreland WTE Fac.
Wheelabrator Falls RRF
York Co. RR Center
San Juan
Central Falls RRF
Johnston RRF
North Kingston Solid Waste Fac.
Quonset Point RRF
Chamber Medical Tech. of SC
Foster Wheeler Charleston RR
Dyersburg RRF
Lewisburg RRF
Nashville Thermal Transfer Corp
Resource Authority in Sumner
Co.
Center RRF
City of Clebume
Panola Co. WTE
Waxahachie Solid Waste RR
Davis Co. WTE
Alexandria/ Arlington RRF
Arlington - Pentagon
Galax City SW Steam Recovery
Unit
Harrisonburg RRF
Henrico Co. RRF
1-95 Energy RRF
NASA Refuse-fired Steam
Generator
Norfolk Naval Station
Norfolk Navy Yard
Prince William and London
Counties
Salem Waste Disposal Energy
Recovery
Location
Brooks
Chester
Glendon
Harrisburg
Bainbridge
Conshohoken
Philadelphia EC
Philadelphia NW
Greensburg
Falls Township
Manchester
Township
San Juan
Central Falls
Johnston
North Kingston
Quonset Point
Hampton
Charleston
Dyersburg
Lewisburg
Nashville
Gallatin
Center
Clebume
Carthage
Waxahachie
Layton
Alexandria
Arlington -
Pentagon
Galax
Harrisonburg
Richmond
Lorton
Hampton
Norfolk Naval
Station
Norfolk
Manassass
Salem
State
OR
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PR
RI
RI
RI
RI
SC
SC
TN
TN
TN
TN
TX
TX
TX
TX
UT
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
Project
status
OP
OP
P
OP
OP
OP
IA
IA
P
OP
OP
OP
P
P
P
P
P
OP
OP
IA
IA
OP
OP
OP
OP
OP
IA
OP
OP
OP
IA
OP
IA
OP
OP
IA
OP
P
IA
Total plant
capacity,
tons/d
550
2,688
500
720
1,200
1,200
750
750
48
50
1,500
1,344
1,200
750
750
750
710
270
600
100
60
1,050
200
40
115
40
50
400
975
50
56
100
250
3,000
200
360
2,000
1,700
100
No. of
units
2
6
2
3
2
2
2
2
2
3
3
3
2
2
1
3
2
1
3
1
2
2
3
1
1
2
4
2
2
4
4
Combustor type
MB/WW
MB/RC/WW
MB/WW
MB/WW
MB/WW
MB/WW
MB/WW
MW/WW
MOD
MOD/SA
MB/WW
MB/RC/WW
MB/WW
MB
MB/WW
MB
MB/WW
MOD/SA
MB/WW
MOD/SA
MOD
MB/WW
MB/RC
MOD/SA
MOD/SA
MOD/SA
MOD/SA
MB/REF
MB/WW
MOD/SA
MB/RC/WW
MB/WW
RDF/FB
MB/WW
MB/WW
MB/WW
RDF
MB/WW
MOD/SA
Air pollution
control devices
FFSD
FFSD
FF SD SNCR CI
ESP
FFSD
FFSD
ESP
ESP
FFSD
ESP
FF SD SNCR CI
FFSD
FF SD SNCR CI
None
FF SD SNCR CI
None
FF SD SNCR CI
ESP DSI SD
ESPSD
None
WS
ESP
ESP
WS
ESP
WS
None
ESP DSI
ESP DSI CI(r)
None
FF
ESP
None
FFSD
ESP
ESP SD(r)
ESP FF(r) SD(r)
FF SD SNCR CI
None
B-6
-------
TABLE B-2. (continued)
Unit name
Rutland RR Center
Fort Lewis RRF
Recomp Bellingham RRF
Skagit Co. RRF
Spokane Regional Disposal Fac.
Tacoma
Barron Co. WTE Fac.
LaCrosse Co.
Madison Power Plant
Muscoda RRF
Sheboygan
St. Croix Co. WTE Fac.
Waukesha RRF
Winnebago
Location
Rutland
Fort Lewis
Bellingham
Mt. Vernon
Spokane
Tacoma
Almena
French Island
Madison
Muscoda
Sheboygan
New Richmond
Waukesha
Winnebago
State
VT
WA
WA
WA
WA
WA
WI
WI
WI
WI
WI
WI
WI
WI
Project
status
IA
UC
OP
OP
OP
OP
OP
OP
IA
IA
OP
OP
IA
P
Total plant
capacity,
tons/d
240
120
100
178
800
300
100
400
120
120
216
115
175
500-1,000
No. of
units
2
3
2
2
2
2
2
2
2
2
1
3
2
Combustor type
MB/MOD
MB/WW
MOD/SA
MB/WW
MB/WW
Cofired RDF/FB
MOD/SA
RDF/FB
Cofired RDF
MOD/SA
MB/REF
MOD/SA
MB/REF
Air pollution
control devices
ESPWS
FF SD SNCR
FFWS
FFSD
FF SD SNCR
FFDSI
ESP
DSI EGB
ESP
FFDSI
WS
FFDSI
ESP
None
OP = operating; IA = inactive (temporarily or permanently shutdown); UC - under construction; On hold = construction plans on
hold; and P = planned.
B-7
-------
APPENDIX C.
SELECTED INFORMATION FOR CEMENT KILNS AND LIME PLANTS
C. 1 - UNITED STATES PORTLAND CEMENT KILN CAPACITIES--1995
C.2 - LIME PLANTS IN THE UNITED STATES IN 1991
-------
TABLE C-l. PORTLAND CEMENT PRODUCTION FACILITIES--1995
Company and location
Alamo Cement Co.
San Antonio, TX
Allentown Cement Co., Inc.
Blandon, PA
Armstrong Cement & Sup. Co.
Cabot, PA
Ash Grove Cement Co.
Nephi, UT
Louisville, NE
Durkee, OR
Foreman, AR
Montana City, MT
Chanute, KS
Inkom, ID
Seattle, WA
Blue Circle Inc.
Ravena, NY
Atlanta, GA
Tulsa, OK
Calera, AL
Harleyville, SC
Calaveras Cement Co.
Redding, CA
Tehachapi, CA
California Portland Cement
Mojave, CA
Colton, CA
Rillito, AZ
Capitol Cement Corporation
Martinsburg, WV
Capitol Aggregates, Inc.
San Antonio, TX
Centex
Laramie, WY
La Salle, IL
Fernley, NV
Continental Cement Co., Inc.
Hannibal, MO
Dacotah Cement
Rapid City, SD
Dixon-Marquette
Dixon, IL
Dragon Products Company
Thomaston, ME
No. /type of kiln
1 -Dry
2 -Dry
2 -Wet
1-Dry
2 -Dry
1-Dry
3 -Wet
1-Wet
2 -Wet
2 -Wet
1 -Dry
2 -Wet
2 -Dry
2 -Dry
2 -Dry
1 -Dry
1 -Dry
1 -Dry
1 -Dry
2 -Dry
4 -Dry
3 -Wet
1 -Dry/1 -Wet
2 -Dry
1 -Dry
2 -Dry
1 -Wet
1 - Dry/2 - Wet
4 -Dry
1-Wet
Clinker capacity,11
103Mg/year
740
844
294
570
885
422
910
289
478
205
681
1,596
546
544
578
644
590
818
1,126
680
1,171
868
456/319
606
498
418
544
526/286
474
392
C-l
-------
TABLE C-l. (continued)
Company and location
Essroc Materials
Nazareth, PA
Nazareth, PA
Speed, IN
Bessemer, PA
Frederick, MD
Logansport, IN
Florida Crushed Stone
Brooksville, FL
Giant Cement Holding, Inc.
Harleyville, SC
Bath, PA
Glens Falls Cement Co.
Glens Falls, NY
Hawaiian Cement Company
Ewa Beach, HI
Holnam, Inc.
Midlothian, TX
Theodore, AL
Clarksville, MO
Holly Hill, SC
Mason City, IA
Florence, CO
Fort Collins, CO
Dundee, MI
Artesia, MS
Seattle, WA
Three Forks, MT
Ada, OK
Morgan, UT
Independent Cement Corp.
Catskill, NY
Hagerstown, MD
Kaiser Cement Corp.
Permanente, CA
Kosmos Cement Co.
Kosmosdale, KY
Pittsburgh, PA
LaFarge Corporation
Buffalo, IA
Grand Chain, IL
Alpena, MI
Whitehall, PA
Sugar Creek, MO
Pauldmg, OH
Fredonia, KS
No. /type of kiln
4 -Dry
1-Dry
2 -Dry
2 -Wet
2 -Wet
2 -Wet
1 -Dry
4 -Wet
2 -Wet
1 -Dry
1 -Dry
1 -Dry
1 -Dry
1 -Wet
2 -Wet
2 -Dry
3 -Wet
1 -Dry
2 -Wet
1 -Wet
1 -Wet
1 -Wet
2 -Wet
2 -Wet
1-Wet
1 -Dry
1 -Dry
1 -Dry
1-Wet
1 -Dry
2 -Dry
5 -Dry
3 -Dry
2 -Dry
2 -Wet
2 -Wet
Clinker capacity,21
103Mg/year
530
1,067
921
518
338
412
537
788
546
463
227
953
1,362
1,179
967
835
761
422
956
463
404
327
562
288
544
463
1,451
707
349
843
1,050
2,094
791
478
432
349
C-2
-------
TABLE C-l. (continued)
Company and location
Lehigh Portland Cement
Mason City, IA
Leeds, AL
Union Bridge, MD
Mitchell, IN
York, PA
Waco, TX
Lone Star Industries
Cape Girardeau, MO
Greencastle, IN
Oglesby, IL
Pryor, OK
Sweetwater, TX
Medusa Cement Co.
Charlevoix, MI
Clmchfield, GA
Wampum, PA
Demopolis, AL
Mitsubishi Cement Corp.
Lucerne Valley, CA
Monarch Cement Company
Humboldt, KS
National Cement Company of Alabama
Ragland, AL
Natl. Cement Co. of Califorinia
Lebec, CA
North Texas Cement
Midlothian, TX
Pennsuco Cement Co.
Medley, FL
Phoenix Cement Company
Clarkdale, AZ
RC Cement Company, Inc.
Independence, KS
Stockertown, PA
Festus, MD
Chattanooga, TN
Rinker Portland Cement Corp.
Miami, FL
Rio Grande Cement Corp.
Tijeras, NM
Riverside Cement Co.
Oro Grande, CA
Riverside, CA
RMC Lonestar
Davenport, CA
No. /type of kiln
1-Dry
1-Dry
4 -Dry
3 -Dry
1-Wet
1-Wet
1 -Dry
1 -Wet
1 -Dry
3 -Dry
3 -Dry
1 -Dry
1 -Dry/1 -Wet
3 -Dry
1 -Dry
1 -Dry
3 -Dry
1 -Dry
1 -Dry
3 -Wet
3 -Wet
3 -Dry
4 -Dry
2 -Dry
2 -Dry
2 -Wet
2 -Wet
2 -Dry
7 -Dry
2 -Dry
1 -Dry
Clinker capacity,21
103Mg/year
731
644
900
661
91
78
1,032
616
522
631
435
1,237
542/189
638
735
1,547
611
811
590
768
881
639
292
828
1,102
398
500
432
1,070
100
726
C-3
-------
TABLE C-l. (continued)
Company and location
Roanoke Cement Company
Cloverdale, VA
Royal Cement Co.
Logendale, NV
Southdown, Inc.
Victorville, CA
Brooksville, FL
Knoxville, TN
Fairborn, OH
Lyons, CO
Odessa, TX
St. Mary's Peerless Cement Co.
Detroit, MI
Sunbelt Cement Corp.
New Braunfels, TX
Texas Industries
New Braunfels, TX
Midlothian, TX
Texas-Lehigh Cement Co.
Buda, TX
Total capacity reported
No. /type of kiln
5 -Dry
1 -Dry
2 -Dry
2 -Dry
1-Dry
1-Dry
1-Dry
2 -Dry
1 -Wet
1 -Dry
1 -Dry
4 -Wet
1 -Dry
136 -Dry 772 -Wet
Clinker capacity,21
103Mg/year
899
177
1,461
1,102
580
544
380
478
590
880
760
1,144
988
76,335
Source: U.S. and Canadian Portland Cement Industry: Plant Information Summary. December 31, 1995.
Portland Cement Association, Skokie, Illinois. November, 1996.
aNote: All Kilns, including inactive Kilns.
Kilns reported as inactive in 1995
California Portland Cement
Centrex
Lafarge Corporation
Medusa Cement Company
Pennsuco Corporation
St. Mary's Peerless Cement Corp.
Total active capacity
Colton, CA 1 kiln
Laramie, WY 1 kiln
Whitehall, PA 1 kiln
Clinchfield, GA 1 kiln
Medley, FL 1 kiln
Detroit, MI 1 kiln
Clinker capacity
103 Mg/yr
340
211
177
189
156
590
74,672
C-4
-------
TABLE C-2. LIME PLANTS ACTIVE IN THE UNITED STATES IN 199 la
(Source: National Lime Association)
Company /headquarters location
Alabama
Allied Lime Company (HQ),
Birmingham, AL
Blue Circle, Inc.
Calera, AL
Cheney Lime & Cement Company
Allgood, AL
Dravo Lime Company
Saginaw, AL
Arizona
Chemstar Lime, Inc. (HQ)
Phoenix, AZ
Magma Cooper Company (C)
San Manuel, AZ
Arkansas
Arkansas Lime Company
Batesville, AR
California
Spreckles Sugar Company, Inc. (C)
Woodland, CA
Chemstar Lime, Inc. (HQ)
Phoenix, AZ
Delta Sugar Corp. (C)
Clarksburg, CA
Holly Sugar Corp. (C)
Colorado Springs, CO
Marine Magnesium Company (C)
S. San Francisco, CA
National Refractories & Minerals Corp.
Moss Landing, CA
Union Sugar Division of Holly Sugar Corp. (C)
Santa Maria, CA
Colorado
Calco, Inc.
Salida, CO
Western Sugar Company
Fort Morgan, CO
Greeley, CO
Idaho
The Amalgamated Sugar Company (C)
Nampa, ID
Paul, ID
Twin Falls, ID
Phoenix, AZ
Plant location/name
Alabaster
Montevallo
Roberta
Landmark
Allgoodb
Longview Div.
Douglas
Nelson
San Manuel
Batesville
Woodland
City of Industry13
Stocktonb
Clarksburg
Hamilton City
Brawley
Tracy
Sonora
Natividad
Betteravia
Salida
Fort Morgan
Greeley
Nampa
Mini-Cassia
Twin Falls
Ten Milec
Type of lime produced
Q
Q,H
Q,H
Q,H
H
Q,H
Q
Q,H
H
Q,H
Q
H
H
H
Q
Q
Q
Q
DL
Q
Q
Q
Q
Q
Q
Q
Q
C-5
-------
TABLE C-2. (continued)
Company /headquarters location
Illinois
Marblehead Lime Company (HQ)
Chicago, IL
Vulcan Materials Company
Countryside, IL
Inland Steel Company (C)
E. Chicago, IN
Iowa
Linwood Mining & Minerals Corp.
Davenport, IA
Kentucky
Dravo Lime Company (HQ)
Pittsburgh, PA
Louisiana
Dravo Lime Company (HQ)
Pittsburgh, PA
USG Corp. (HQ)
Chicago, IL
Massachusetts
Lee Lime Corp.
Lee, MA
Pfizer, Inc.
Adams, MA
Michigan
Detroit Lime Company
Detroit, MI
The Dow Chemical Company (C)
Ludington, MI
Marblehead Lime Company (HQ)
Chicago, IL
Michigan Sugar Company (C)
Saginaw, MI
Monitor Sugar Company (C)
Bay City, MI
Minnesota
American Crystal Sugar Company (C)
Moorhead, MN
Southern Minn. Sugar Corp. (C)
Renville, MN
Plant location/name
South Chicago
Thornton
Buffington
McCook
Indiana Harbor
Linwood (UG)
Black River Div. (UG)
Maysville Div. (HG)
Pehcanb
New Orleans
Lee
Adams
River Rouge
Ludington
River Rouge
Brennan
Sebawaing
Carollton
Crosswell
Caro
Bay City
Moorhead
Crookston
East Grand Forks
Renville
Type of lime produced
Q,H
DL, DH, DB
Q
DL
Q
Q,H
Q,H
Q
H
Q,H
DL,DH
Q
Q
DL
Q
Q,H
Q
Q
Q
Q
Q
Q
Q
Q
Q
C-6
-------
TABLE C-2. (continued)
Company /headquarters location
Missouri
Ash Grove Cement Company
Springfield, MO
Mississippi Lime Company (HQ)
Alton, IL
Resco Products of Missouri, Inc. (HQ)
Clearfield, PA
Continental Lime, Inc.
Townsend, MT
Holly Sugar Corp. (C)
Colorado Springs, CO
Western Sugar Company
Billings, MT
Nebraska
Western Sugar Company (C)
Bayard, NE
Mitchell, NE
Scottsbluff, NE
Nevada
Chemstar Lime, Inc. (HQ)
Phoenix, AZ
Continental Lime, Inc.
Wendover, NV
North Dakota
American Crystal Sugar Company (C)
Drayton, ND
Hillsboro, ND
Minn-Dak Farmers Corp. (C)
Wahpeton, ND
Ohio
Elkem Metals Company (C)
Astabula, OH
GenLime Group LP
Genoa, OH
The Great Lakes Sugar Company (C)
Fremont, OH
Huron Lime Company
Huron, OH
LTV Steel (C&S)
Grand River, OH
Martin Marietta (C&S)
Woodville, OH
National Lime & Stone Company
Findlay, OH
Ohio Lime Company
Woodville, OH
Plant location/name
Springfield
Ste. Genevieve (UG)
Bonne Terre
Indian Creek
Sidney
Billings
Bayard
Mitchell
Scottsbluff
Apex
Henderson
Pilot Peak
Drayton
Hillsboro
Minn-Dak
Ashtabula
Genoa
Fremont
Huron
Grand River
Woodville
Carey
Woodville
Millersville
Type of lime produced
Q,H
Q,H
DL, Q, DB
Q
Q
Q
Q
Q
Q
Q,H
DL,DH
Q
Q
Q
Q
Q
DL,DH
Q
Q
Q
DL,DB
DL,DH
DL
DL
C-7
-------
TABLE C-2. (continued)
Company /headquarters location
Oklahoma
St. Clair Lime Company
Oklahoma City, OK
Oregon
The Amalgamated Sugar Company (C)
Nyssa, OR
Ash Grove Cement Company
Portland, OR
Pennsylvania
I.E. Baker Company (C&S)
York, PA
Bellefonte Lime Company
Belief onte, PA
Centre Lime & Stone Company
Pleasant Gap, PA
Con Lime Company
Bellefonte, PA
Corson Lime Company
Plymouth Meeting, PA
Mercer Lime & Stone Company
Pittsburgh, PA
Warner Company
Devault, PA
Wimpey Minerals PA, Inc.
Annville, PA
Puerto Rico
Puerto Rican Cement Company, Inc.
Ponce, PR
South Dakota
Pete Lien & Sons, Inc.
Rapid City, SD
Tennessee
Bowater Southern Paper Corp. (C)
Calhoun, TN
Term Luttrell Company
LuttrelL TN
Plant location/name
Marble City (UG)
Nyssa
Portland
York
Bellefonte
Pleasant Gap
Bellefonte (UG)
Plymouth Meeting
Branchton
Cedar Hollow
Hanover
Annville
Ponce
Rapid City
Calhoun
Luttrell (UG)
Type of lime produced
Q,H
Q
Q,H
DB
Q,H
Q,H
Q,H
DL,DH
Q,H
DL,DH
DL, Q
Q,H
Q,H
Q,H
Q
0,H
C-8
-------
TABLE C-2. (continued)
Company /headquarters location
Texas
APG Lime Corp.
New Braunfels, TX
Austin White Lime Company
Austin, TX
Chemical Lime, Inc.
Clifton, TX
Holly Sugar Corp. (C)
Colorado Springs, CO
Redland Stone Products Company
San Antonio, TX
Texas Lime Company
Cleburne, TX
Utah
Chemstar Lime, Inc. (HQ)
Phoenix, AZ
Continental Lime, Inc.
Delta, UT
M.E.R.R. Corp.
Grantsville, UT
Virginia
APG Lime Corp
Ripplemead, VA
Chemstone Corp.
Strasburg, VA
W.S. Frey Company, Inc.
York, PA
Riverton Corp. (C)
Riverton, VA
Shenvalley Lime Corp.
Stephens City, VA
Virginia Lime Company
Ripplemead, VA
Washington
Northwest Alloys, Inc. (C)
Addy, WA
Continental Lime, Inc.
Tacoma, WA
West Virginia
Germany Valley Limestone Company
Riverton, WV
Wisconsin
CLM Corp. (HQ)
Duluth, MN
Rockwell Lime Company
Manitowoc, WI
Western Lime & Cement Company
West Bend, WI
Plant location/name
New Braunfels
McNeil
Clifton
Marble Falls
Hereford
San Antonio
No. 1
Round Rockd
Dolomite
Cricket Mountain
Marblehead Mt.e
Kimballton (UG)
Dominion
Clearbrook
Riverton
Stephens City
Kimballton (UG)
Addy
Tacoma
Riverton
Superior
Manitowoc
Green Bay
Eden
Type of lime produced
Q, H, DL, DH
Q,H
Q,H
DL
Q
Q,H
Q,H
Q,H
DL,DH
Q
DL
Q,H
Q,H
Q
H
H
Q,H
DL
Q,H
Q,H
Q,H
DL,DH
Q,H
DL,DH
C-9
-------
TABLE C-2. (continued)
Company /headquarters location
Wyoming
Holly Sugar Company (C)
Colorado Springs, CO
The Western Sugar Company (C)
Lovell, WY
Plant location/name
Torrington
Worland
Lowell
Type of lime produced
Q
Q
Q
KEY:
C = Lime plant is operated predominantly for captive consumption.
C&S = Captive and sales—captive consumption with significant commercial sales.
DB = Refractory, dead-burned dolomite.
DH = Dolomitic hydrate.
DL = Dolomitic quicklime.
H = Hydrated lime.
HQ = Headquarters address.
Q = Quicklime.
UG = Underground mine.
aExcludes regenerated lime.
bHydrating plant only.
°New plant, scheduled to come on-line August 1992.
dPlant did not operate in 1991; it has been mothballed.
eClosed December 1991, last shipments made May 1992.
C-10
-------
APPENDIX D.
CRUDE OIL DISTILLATION CAPACITY
-------
Table 5. Refiners' Operable Atmospheric Crude Oil Distillation Capacity as of January 1,1995
Refiner
Companies with Capacity
Over 100,000 bbl/cd
Chevron USA Inc
Pascagoula, Mississippi
El Segundo California
Richmond California
Port Arthur Texas
E| Pas/i Texas ,----,,,-
Perth Amboy, New Jersey
Honolulu Hawaii
Salt Lake City Utah
Amoco Oil Co
Texas City Texas
Whiting Indiana
Mandan North Dakota
Yorktown, Virginia
Salt Lake City Utah
Exxon Co U.S.A
Baton Rouge, Louisiana
Baytown, Texas
Benicia, California
Billings, Montana
Mobil Oil Corp
Beaumont, Texas
Joliet, Illinois
Chalmette, Louisiana
Torrance, California
Paulsboro New Jersey . . .
Shell Oil Co
Wood River, Illinois
Norco, Louisiana
Martinez, California
Anacortes, Washington
Odessa, Texas
BP America Inc
BP Oil Corp.
Belle Chasse (Alliance) Louisiana
Marcus Hook, Pennsylvania
Lima, Ohio
Toledo, Ohio
Sun Co Inc
Marcus Hook, Pennsylvania
Toledo, Ohio
Tulsa, Oklahoma
Sun Refining & Marketing
Philadelphia, Pennsylvania
Star Enterprise
Port Arthur/Neches, Texas
Convent, Louisiana
Delaware City, Delaware
USX Corp
Marathon Oil Co.
Garyville, Louisiana
Barrels per
Calendar Day
1 206 000
295,000
230 000
230000
185000
87000
80,000
54 000
45000
998,000
433 000
41 0 000
58 000
53,000
44000
992 000
424000
396 000
128000
44000
929 000
315,000
188000
170000
130000
1 26 000
761,000
268 000
215000
148,900
100500
28600
700 500
231 500
172,000
161 000
136000
700000
175,000
125,000
85,000
315,000
600000
235,000
225000
140,000
570,000
255,000
Refiner
Robinson, Illinois
Detroit, Michigan
Texas City Texas
Petroleos De Venezuela
Citgo Petroleum Corp
Lake Charles Louisiana
Citgo Refining & Chemical Inc
Corpus Christ! Texas
Citgo Asphalt Refining Co
Paulsboro New Jersey
Savannah Georgia
Koch Industries Inc
Koch Refining Co
Corpus Christ! Texas
St Paul (Pine Bend) Minnesota
Tosco Corp
Bayway Refining Co.
Bayway New Jersey
Tosco Refining Co
Martinez (Avon) California
Tosco Northwest Co
Ferndale Washington
Atlantic Richfield Co
Arco Products Co
Los Angeles California.
Ferndale (Cherry Point) Washington
Arco Alaska Inc
Prudhoe Bay Alaska
Kuparuk Alaska
E I Du Pont De Nemours & Co
Conoco Inc
Westlake Louisiana
Ponca City Oklahoma..
Commerce City Colorado
Billings, Montana '.
Texaco Refining & Marketing Inc
Anacortes (Puget Sound) Washington
El Dorado Kansas . .
Wilmington (Los Angeles) California
Bakersfield California
Ashland Oil Inc
Cat lettsbur g , Kentucky
St. Paul Minnesota
Canton Ohio
Phillips Petroleum Co
Phillips 66 Co.
Sweeny Texas
Borger Texas
Woods Cross Utah .
Lyondell Petrochemical Co.
Lyondell Citgo Refining Co Ltd
Houston, Texas
Barrels per
Calendar Day
.. .. 175000
70000
70000
503000
305 000
1 30 000
40 000
28 000
485 000
255 000
230 000
470 000
215 000
1 60 000
95 000
453 000
237 000
189000
15 000
12 000
438 000
191 000
140 000
57 500
49500
350 600
136 000
94600
64 000
56000
346 500
213400
67 100
66 000
320000
185 000
110000
25000
265 000
See footnotes at end ol table.
30
D-l
-------
Table 5. Refiners' Operable Atmospheric Crude Oil Distillation Capacity as of January 1,1995
(Continued)
Refiner
Solomon Inc
Phibro Energy U.S.A. Inc.
Texas City Texas
Houston Texas
Krotz Springs Louisiana
Coastal Corp., The
Coastal Eagle Point Oil Co
Westville, New Jersey
Coastal Refining & Marketing Inc
Corpus Christi, Texas
Coastal Mobile Refining Co
Chickasaw, Alabama
Fina Oil & Chemical Co
Port Arthur Texas
Big Spring, Texas ...
Unocal Corp
Wilmington (Los Angeles) California
Rodeo (San Francisco) California
Arroyo Grande (Santa Maria), California
Mapco Petroleum Inc
North Pole, Alaska
Memphis, Tennessee
Shell Oil/PMI Holdings North America
Deer Park Refg Ltd Partnership
Deer Park, Texas . ... .. . ..
Diamond Shamrock Refining & Marketing Co
Sunray (McKee) Texas
Three Rivers Texas
Total Petroleum Inc ...
Ardmore, Oklahoma
Arkansas City Kansas
Alma, Michigan ...
Colorado Refining Co.
Commerce City, Colorado
Crown Central Petroleum Corp
Pasadena, Texas
La Gloria Oil & Gas Co
Tyler, Texas . .
Kerr-McGee Corp
Southwestern Refining Co. Inc.
Corpus Christi, Texas
Kerr-McGee Refining Corp
Wynnewood Oklahoma
Cotton Valley, Louisiana
Uno-Ven Co.
Lemont (Chicago), Illinois
Horsham Corp
Clark Refining & Marketing
Blue Island, Illinois
Hartford, Illinois
Barrels per
Calendar Day
254,500
123,500
71 000
60,000
236,500
125,000
95,000
16,500
230,000
175000
55,000
220,700
105600
73,100
42,000
217,200
. . 1 28,200
89,000
215,900
207,000
132000
75,000
197,600
68,000
56,000
45,600
28,000
155,000
100.000
55,000
154,800
104,000
43000
7,800
147,000
143,015
80,515
62 500
Refiner
Murphy Oil U.S.A. Inc
Meraux, Louisiana
Superior, Wisconsin
Sinclair Oil Corp
Tulsa, Oklahoma
Sinclair, Wyoming
Little America Refining Co.
Evansville (Casper), Wyoming
Castle Energy Corp .
Indian Refining
Lawrenceville, Illinois
Powerine Oil Co.
Santa Fe Springs, California
Cenex
National Cooperative Refinery Assoc.
McPherson • Kansas .
Cenex
Laurel, Montana
Companies with Capacity
30,001 to 100,000 bbl/cd
BHP Petroleum Americas Refining Inc.
Ewa Beach Hawaii . .
Tesoro Petroleum Corp
Kenai Alaska
LL&E Petroleum Marketing, Inc.
Saraland (Mobile) Alabama
Farmland Industries Inc.
Coffeyville Kansas
American Ultramar Ltd
Ultramar Refg
Wilmington, California
Holly Corp
Navajo Refining Co.
Artesia New Mexico . ...
Montana Refining Co.
Great Falls, Montana
Pennzoil Co Inc. . ..
Pennzoil Producing Co
Shreveport, Louisiana
Rouseville Pennsylvania
United Refining Co.
Warren, Pennsylvania
Lion Oil Co.
El Dorado Arkansas
Barrels per
Calendar Day
133,200
100,000
33.200
132500
54,000
54,000
24,500
127,250
80,750
46 500
117,050
75600
41,450
.
93500
72 000
71 000
68600
68,000
64,000
57000
7,000
61 ,900
46,200
15,700
60,000
51 ,000
See footnotes at end of table.
D-2
31
-------
Table 5. Refiners' Operable Atmospheric Crude Oil Distillation Capacity as of January 1,1995
(Continued)
Refiner
The Coastal Corp/Sinochem
Pacific Refining Co.
Hercules, California
Placid Refining Co.
Port Allen, Louisiana
Paramount Acquisition Corp.
Paramount Petroleum Corp.
Paramount, California
Pride Refining Inc.
Abilene Texas .... ..
Enjet
St. Rose Refining Inc.
St. Rose, Louisiana
Frontier Refining Co.
Cheyenne, Wyoming
Petro Star Inc. ...
Valdez, Alaska
North Pole, Alaska
Hunt Consolidated Inc.
Hunt Refining Co.
Tuscaloosa, Alabama
Time Oil Co.
U.S. Oil & Refining Co.
Tacoma, Washington
Companies with Capacity
10,001 to 30,000 bbl/cd
Valero Refining Co.
Corpus Christi, Texas
Gold Line Refining Ltd.
Lake Charles, Louisiana
Neste Trrfinery Petro Serve a
Corpus Christi, Texas
Crysen Corp ....
Crysen Refining Inc.
Woods Cross, Utah
Sound Refining Inc.
Tacoma, Washington
San Joaquin Refining Co. Inc.
Bakersfield, California . .
Flying J Petroleum Inc.
Big West Oil Co.
North Salt Lake, Utah
Ergon Inc.
Vicksburg, Mississippi ...
Barrels per
Calendar Day
50,000
48,500
46,500
42750
40,000
38670
36 300
26,300
10 000
33,500
32,400
•T" "7978,620 •
29,900
27 600
27 000
24400
12,500
11,900
24300
24,000
23000
Refiner
Countrymark Cooperative Inc.
Mount Vernon, Indiana
Kern Oil & Refining Co.
Bakersfield, California
Giant Industries Inc.
Giant Refining Co.
Gallup, New Mexico
World Oil Co
Sunland Refining Corp.
Bakersfield California
Lunday Thagard
South Gate, California
Barrett Refining Corp
Thomas (Custer), Oklahoma
Vicksburg, Mississippi
VGSCorp
Southland Oil Co
Sandersville, Mississippi
Lumberton Mississippi
Gary Williams Co.
Bloomfield Refining Co.
Bloomfield, New Mexico '.
Huntway Refining Co
Benicia California
Wilmington, California
Wyoming Refining Co.
Newcastle, Wyoming
Transworld Oil U.S.A. Inc.
Calcasieu Refining Co.
Lake Charles, Louisiana
Quaker State Corp.
Newell (Congo), West Virginia
Asphalt Materials
Laketon Refining Corp.
Laketon, Indiana
Apex Oil Co Inc
Petroleum Fuel & Terminal
Long Beach, California
total. ""'..: "Lz " '~~~ ' ~~." .~ ""
Companies with Capacity
10,000 bbl/cd or Less
Witco Corp.
Bradford, Pennsylvania
Anchor Gasoline Corp.
Canal Refining Co.
Church Point Louisiana
Barrels per
Calendar Day
22 000
21,400
20800
. . 20 100
12 000
8 100
18 500
10,500
8,000
1 6 800
11 000
5 800
16 800
14 100
8 600
5 500
12,555
12 500
11 500
11 100
10800
^•*rV' 389,155 <
10,000
9500
See footnotes at end of table.
32
D-3
-------
Table 5. Refiners' Operable Atmospheric Crude Oil Distillation Capacity as of January 1,1995
(Continued)
Refiner
Calumet Lubricants Co. L.P.
Princeton Louisiana
Cayman Resources
Cyril Petrochemical Corp
Cyril, Oklahoma .
Arcadia Refining b
Lisbon, Louisiana
Bechtel Investment Inc.
Petro Source Refining Partners
Eagle Springs, Nevada .
Martin Gas Sales Inc.
Berry Petroleum Co.
Stephens Arkansas
Cross Oil & Refining Co. Inc.
Smackover, Arkansas ....
Age Refining & Marketing
San Antonio Texas
Barrels per
Calendar Day
8200
7,500
7 350
7000
6700
6,200
6,000
Refiner
Young Refining Corp.
Douglasville, Georgia .
Somerset Refinery Inc.
Somerset, Kentucky
Oil Holdings Inc.
Tenby Inc.
Oxnard, California
Unico, Inc.
Intermountain Refining Co., Inc.
Fredonia, Arizona
Howell Corp.
Howell Hydrocarbons & Chemical Inc.
Channelview, Texas
Petrolite Corp.
Kilgore, Texas
Total ™~ "' . „»:. "<^":- "*~~^
U.S. Total ..~_..........,,,......~..........™ ....._
Barrels per
Calendar Day
5,540
5,500
4 000
3800
1 400
1,000
„.;.„,. '"„„„."' 89,630 "
—..I,-™.....™/ 15,434,280;
Source: United States Refining Capacity, January 1, 1995
National Petroleum Refiners Association, Washington, DC
aFormerly Petroserve Ltd. (Trifinery)
Formerly Dubach Gas Co.
bbl/cd = Barrels per Calendar Day.
Source: Energy Information Administration (EIA), Form EIA-820, 'Annual Refinery Report.'
D-4
33
-------
APPENDIX E.
PULP AND PAPER MILLS IN THE UNITED STATES IN 1994
-------
TABLE E-l. PULP AND PAPER MILLS IN THE UNITED STATES IN 1994
Mill name
Alabama Pine Pulp
Alabama River Pulp
Appleton Papers, Inc.
Arkansas Kraft
Badger Paper Mills, Inc.
Boise Cascade Corp.
Boise Cascade Corp.
Boise Cascade Corp.
Boise Cascade Corp.
Boise Cascade Corp.
Boise Cascade Corp.
Bowater Inc. Carolina Division
Bowaters
Champion International
Champion International
Champion International
Champion International
Champion International
Champion International
Champion International
Chesapeake Paper Products Co.
Consolidated Packaging Corp.
Consolidated Papers
Container Corp. of America
Cross-Pointe Paper Co.
Federal Paper Board Co.
Federal Paper Board, Inc.
Finch, Pruyn, & Co., Inc.
Gaylord Container Corp.
Gaylord Container Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Georgia-Pacific Corp.— Nekoosa Paper Co.
Georgia-Pacific Corp.
Location
Perdue Hill, AL
Perdue Hill, AL
Roaring Springs, PA
Oppelo, LA
Peshtigo, WI
Deridder, LA
International Falls, MN
Jackson, AL
Rumford, ME
St. Helens, OR
Wallula, WA
Catawba, SC
Calhoun, TN
Canton, NC
Courtland, AL
Lufkm, TX
Quinnesec, MI
Roanoke Rapids, NC
Sheldon, TX
Cantonment, FL
West Point, VA
Fort Madison, IA
Wisconsin Rapids, WI
Femandina Beach, FL
Park Falls, WI
Augusta, GA
Riegelwood, NC
Glens Falls, NY
Bogalusa, LA
Pine Bluff, AR
Ashdown, AR
Bellingham, WA
Big Island, VA
Brunswick, GA
Cedar Springs, GA
Crossett, AR
Monti cello, MS
Nekoosa, WI
New Augusta, MS
Palatka, FL
Port Edwards, WI
Toledo, OR
Type of
pulping process
Kraft
Kraft
Kraft
Kraft
Sulfite
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Semichemical
Kraft
Kraft
Sulfite
Kraft
Kraft
Sulfite
Kraft
Kraft
Kraft
Sulfite
Semichemical
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Sulfite
Kraft
E-l
-------
TABLE E-l. (continued)
Mill name
Georgia-Pacific Corp.
Georgia-Pacific Corp.
Oilman Paper Co.
Great Northern Paper Co.
Groveton Paper
Gulf States Paper Corp.
ITT Rayonier, Inc.
ITT-Rayonier, Inc.
ITT Rayonier, Inc.
Inland Container Corp.
Inland-Orange, Inc.
Inland-Rome, Inc.
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
International Paper
Interstate Paper
JSC/Container
JSC/Container
James River Corp.
James River Corp.
James River Corp.
James River Corp.
James River Corp.
James River Corp.
James River Paper Co.
Jefferson Smurfit
Ketchikan Pulp Co.
Kimberly-Clark Corp.
Location
Woodland, ME
Zachary, LA
St. Mary's, GA
Millinocket, ME
Groveton, NH
Demopolis, AL
Femandina Beach, FL
Jesup, GA
Port Angeles, WA
New Johnsonville, TN
Orange, TX
Rome, GA
Bastrop, LA
Camden, AR
Ene, PA
Gardiner, OR
Georgetown, SC
Jay, ME
Mansfield, LA
Mobile, AL
Moss Point, MS
Natchez, MS
Pine Bluff, AR
Pineville, LA
Selma, AL
Texarkana, TX
Ticonderoga, NY
Vicksburg, MS
Riceboro, GA
Brewton, AL
Jacksonville, FL
Berlin, NH
Camas, WA
Camas, WA
Clatskanie, OR
Pennington, AL
St. Francisville, LA
Old Town, ME
Circleville, OH
Ketchikan, AK
Coosa Pines, AL
Type of
pulping process
Kraft
Kraft
Kraft
Sulfite
Semichemical
Kraft
Sulfite
Kraft
Sulfite
Semichemical
Kraft
Kraft
Kraft
Kraft
Soda
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Sulfite
Kraft
Kraft
Kraft
Kraft
Semichemical
Sulfite
Kraft
E-2
-------
TABLE E-l. (continued)
Mill name
Lincoln Pulp & Paper
Longview Fibre Co.
Louisiana Pacific
MacMillan Bloedel, Inc.
Mead Coated Board
Mead Corp.
Mead Corp.
Mead Paper
Mead Paper/Chillicothe Division
Menasha Corp.
Mosinee Paper
P.H. Glatfelter
Packaging Corp. of America
Packaging Corp. of America
Packaging Corp. of America
Packaging Corp. of America
Pope & Talbot
Port Townsend Paper Corp.
Potlatch Corp.
Potlatch Corp.
Potlatch Corp.
Procter & Gamble
Procter & Gamble Cellulose
Procter & Gamble Cellulose
Riverwood International Georgia
Riverwood International
S.D. Warren Co.
S.D. Warren Co.
Scott Paper Co.
Scott Paper Co.
Scott Paper Co.
Simpson Paper
Simpson Paper
Sonoco Products
St. Joe Forest Products
Stone Container Corp.
Stone Container Corp.
Stone Container Corp.
Stone Container Corp.
Stone Container Corp.
Stone Container Corp.
Location
Lincoln, ME
Longview, WA
Samoa, CA
Pine Hill, AL
Phemx City, AL
Kingsport, TN
Stevenson, AL
Escanaba, MI
Chillicothe, OH
Otsego, MI
Mosinee, WI
Spring Grove, PA
Counce, TN
Filer City, MI
Tomahawk, WI
Valdosta, GA
Halsey, OR
Port Townsend, WA
Cloquet, MS
Lewiston, ID
McGehee, AR
Mehoopany, PA
Ogelthorpe, GA
Perry, FL
Macon, GA
West Monroe, LA
Muskegon, MI
Westbrook, ME
Everett, WA
Mobile, AL
Skohegan, ME
Pasadena, TX
Tacoma, WA
Hartsville, SC
Port St. Joe, FL
Coshocton, OH
Florence, SC
Hodge, LA
Missoula, MT
Ontonagon, MI
Panama City, FL
Type of
pulping process
Kraft
Kraft
Kraft
Kraft
Kraft
Soda
Semichemical
Kraft
Kraft
Semichemical
Kraft
Kraft
Kraft
Semichemical
Semichemical
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Sulfite
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Sulfite
Kraft
Kraft
Kraft
Kraft
Semichemical
Kraft
Semichemical
Kraft
Kraft
Kraft
Semichemical
Kraft
E-3
-------
TABLE E-l. (continued)
Mill name
Stone Container Corp.
Stone Hopewell, Inc.
Stone Savannah River
Temple-Inland Forest Products
Thilmany International
Union Camp Corp.
Union Camp Corp.
Union Camp Corp.
Union Camp Corp.
Virginia Fibre Corp.
Wausau Paper Mills Co.
Weston Paper & Manufacturing Corp.
Westvaco Corp.
Westvaco Corp.
Westvaco Corp.
Westvaco Corp.
Weyerhaeuser Paper Co.
Weyerhaeuser Paper Co.
Weyerhaeuser Paper Co.
Weyerhaeuser Paper Co.
Weyerhaeuser Paper Co.
Weyerhaeuser Paper Co.
Weyerhaeuser Paper Co.
Weyerhaeuser Paper Co.
Williamette Industries, Inc.
Williamette Industries, Inc.
Williamette Industries, Inc.
Williamette Industries, Inc.
Williamette Industries, Inc.
Location
Snowflake, AZ
Hopewell, VA
Port Wentworth, GA
Evadale, TX
Kaukauna, WI
Eastover, SC
Franklin, VA
Prattville, AL
Savannah, GA
Amherst/Riverville, VA
Brokaw, WI
Terra Haute, IN
Covington, VA
Luke, MD
N. Charleston, SC
Wickhffe, KY
Columbus, MS
Cosmopolis, WA
Longview, WA
New Bern, NC
Plymouth, NC
Rothschild, WI
Springfield, OR
Valhant, OK
Albany, OR
Bennettsville, SC
Campti, LA
Hawesville, KY
Johnsonburg, PA
Type of
pulping process
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Semichemical
Sulfite
Semichemical
Kraft
Kraft
Kraft
Kraft
Kraft
Sulfite
Kraft
Kraft
Kraft
Sulfite
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Kraft
Sources: Midwest Research Institute (MRI), 1996. Memorandum from Nicholson, R., MRI, to Telander, J..
EPA/MICG. June 13, 1996. Addendum to Summary of Responses to the 1992 NCASI "MACT"
Survey.
Midwest Research Institute (MRI), 1995. Memorandum from Soltis, V., MRI to the project file.
April 24,1995. U.S. Population of Sulfite and Stand-Alone Semichemical Pulp Mills.
E-4
-------
APPENDIX F.
EMISSION FACTORS BY SOURCE CLASSIFICATION CODE
-------
TABLE F-l. SUMMARY OF EMISSION FACTORS BY SOURCE CLASSIFICATION CODE
sec/
Description
3-13-020-00
Mercury Oxide
Battery Manufacture
3-13-005-00
Electrical Switch
Manufacture
3-13-011-00
Fluorescent Lamp
Manufacture
3-13-012-00
Fluorescent Lamp
Recycling
3-15-027-00
Thermometer
Manufacture
1-01-002, 1-02-002,
1-03-002
Bituminous and
Subbituminous Coal
Combustion
1-01-001, 1-02-001,
1-03-001
Anthracite Coal
Combustion
1-01-004
No. 6 Oil Fired
Emissions Source
Overall Process
Overall Process
Overall Process
Lamp Crusher
Overall Process
Industrial Boilers;
Commercial and
Residential Boilers
Industrial Boilers;
Commercial and
Residential Boilers
Utility Boilers;
Industrial Boilers;
Commercial and
Residential Boilers
Emission Factora
Control Device(s) Range Average
Uncontrolled — 2b
Uncontrolled — 8b
Uncontrolled — 8b
Fabric Filter and — 1.9E-09C
Carbon Adsorber
Uncontrolled — 18b
Uncontrolled — 16d
Uncontrolled — 18d
Uncontrolled — 0.46d
~ Factor
Rating
U
U
U
E
U
E
E
E
-------
TABLE F-l. (continued)
sec/
Description
1-01-005
Distillate Oil Fired
1-01-009, 1-02-009,
1-03-009
Wood Waste Combustion
5-01-005-15
Sewage Sludge Incinerators
3-05-006, 3-05-007
Portland Cement
Manufacture
3-05-016-18
Lime Manufacture
3-05-016-19
Lime Manufacture
3-01-005-04
Carbon Black
Manufacturing
3-03-003
Coke Production
1-01-015-01
1-01-015-02
Geothermal Power Plants
Emissions Source
Utility Boilers;
Industrial Boilers;
Commercial and
Residential Boilers
Industrial Boilers
Multiple Hearth
Incinerators
Kiln Stack
Coal-Fired Rotary
Kiln Stack
Gas-Fired Vertical
Kiln Stack
Main Process Vent
Overall Process
Off-Gas Ejectors
Cooling Tower
Exhaust
Control Device(s)
Uncontrolled
Uncontrolled
Venturi Scrubber
Cyclone
Fabric Filter; ESP;
Venturi Scrubber
Cyclone and Fabric
Filter
Fabric Filter
Fabric Filter
Fabric Filter; ESP
Uncontrolled
Uncontrolled
Emission Factora
Range Average
6.2d
2.6 E-06e
3.5 E-05f
3.2 E-03f
1.3 E-04S
7.6E-06- 1.5E-0511
1.8E-05
3.0 E-06h
3 E-04J
6 E-05k
2 E-05m
1 E-04m
~ Factor
Rating
E
E
E
E
E
E
E
U
U
U
U
-------
TABLE F-l. (continued)
sec/
Description
3-07-001-04
3-07-001-10
Chemical Wood Pulping
3-07-001-05
3-07-001-06
3-15-025-00
Dental Alloy (Mercury
Amalgam) Production
3-15-021-01
Crematoriums
Emissions Source
Kraft/Soda
Recovery Furnace
Kraft/Soda SDTs
Kraft/Soda Lime
Kiln
Overall Process
Crematory Stack
Control Device(s)
ESP; Wet Scrubber
Venturi Scrubber;
Wet Scrubber
Wet Scrubber; ESP
Uncontrolled
Uncontrolled
Emission Factora
Range Average
3.9 E-0511
5.23 E-0811
2.91 E-06h
40b
3.3 E-03P
Factor
Rating
U
U
U
U
E
aTo convert from Ib/ton to kg/Mg, multiply by 0.5.
blb/ton of mercury used.
clb/lamp crushed.
dlb/1012 Btu.
elb/ton of dry wood fuel.
flb/ton of sludge processed.
glb/ton of clinker produced.
hlb/ton of lime produced.
Jlb/ton of carbon black produced.
klb/ton of coke produced.
mlb/MWe/hr.
nlb/ton of black liquor solid fuel.
plb/body burned.
------- |