EPA/600/R-02/104
December 2002
Use and Release of
Mercury in the United States
by
Barry R. Leopold
Science Applications International Corporation
Reston, Virginia 20190
Contract Nos. 68-C-0027 and 68-C7-0011
Project Officer
Kenneth Stone
Sustainable Technology Division
National Risk Management Research Laboratory
Cincinnati, Ohio 45268
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
Notice
The information in this document has been funded by the United States Environmental Protection Agency under
Contract Nos. 68-C6-0027 and 68-C7-0011 to Science Applications International Corporation (SAIC). It has been
subj ected to the Agency' s peer and administrative review and has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the Nation's land, air, and
water resources. Under a mandate of national environmental laws, the Agency strives to formulate and implement
actions leading to a compatible balance between human activities and the ability of natural systems to support and
nurture life. To meet this mandate, EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our ecological resources
wisely, understand how pollutants affect our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for investigation of
technological and management approaches for preventing and reducing risks from pollution that threaten human
health and the environment. The focus of the Laboratory's research program is on methods and their cost-effectiveness
for prevention and control of pollution to air, land, water, and subsurface resources; protection of water quality in
public water systems; remediation of contaminated sites, sediments and ground water; prevention and control of
indoor air pollution; and restoration of ecosystems. NRMRL collaborates with both public and private sector partners
to foster technologies that reduce the cost of compliance and to anticipate emerging problems. NRMRL's research
provides solutions to environmental problems by: developing and promoting technologies that protect and improve
the environment; advancing scientific and engineering information to support regulatory and policy decisions; and
providing the technical support and information transfer to ensure implementation of environmental regulations and
strategies at the national, state, and community levels.
This publication has been produced as part of the Laboratory's strategic long-term research plan. It is published and
made available by EPA's Office of Research and Development to assist the user community and to link researchers
with their clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
111
-------
Abstract
Although mercury use is decreasing in the United States, mercury continues to pose a serious risk to human health
and the environment. This report presents the results of a two-year effort sponsored by EPA's Office of Research and
Development to quantify and map the flows of mercury throughout the U.S. economy and released into the
environment. Data contained in the report are intended to help prioritize research and development efforts.
Using a materials flow analysis (MFA), this report quantifies cradle-to-grave mercury use, reuse, release and disposal
associated with products and processes that use mercury. Among the 100 data sources examined were industry
estimates, government statistics, literature sources, and USEPA data such as the Toxics Release Inventory and
emission factors. Specific industries and sectors within the following major divisions were evaluated: mercury
supply, mercury use in manufacturing processes, incidental mercury use associated with coal combustion, incidental
mercury use associated with non-coal sources, and other sources of mercury resulting from previous use. For each
sector, mercury use is described, mercury-containing raw materials and products catalogued, and reported mercury
releases into media (air, water, and solid waste) quantified. For sectors in which data are available concerning mercury
speciation and geographic distributions, this information is presented.
IV
-------
Table of Contents
Abstract iv
List of Exhibits vi
List of Acronyms and Abbreviations viii
Acknowledgments ix
Chapter 1 Introduction 1
1.1 Background 1
1.2 Purpose of Report 1
1.3 Methodology Used in Report 2
1.4 Scope of Report 2
1.5 Summary of Results 3
1.6 Data Limitations and Uncertainty 3
Chapter 2 Supply of Mercury 16
2.1 Secondary Mercury Production 16
2.2 Imports and Exports 19
2.3 U.S. Government Stockpiles 19
2.4 Miscellaneous U.S. Government Uses 19
Chapter 3 Manufacturing Processes Involving Mercury 20
3.1 Chlor-Alkali Manufacturing 20
3.2 Lamp Manufacturing, Use, and Disposal 23
3.3 Thermometers and Other Instruments 26
3.4 Thermostats 29
3.5 Switches and Relays 31
3.6 Organic Chemical Production 34
3.7 Dental Preparations 34
3.8 Pharmaceutical Use 36
3.9 Laboratory Use 37
3.10 Batteries 38
3.11 Miscellaneous 38
Chapter 4 Incidental Mercury Use Associated With Coal Combustion or Coal Use 39
4.1 Coal Combustion by Utilities 39
4.2 Lime Manufacturing 43
4.3 Residential, Commercial, and Industrial Coal Combustion 45
4.4 Byproduct Coke Production 47
4.5 Portland Cement Manufacturing 48
4.6 Coal Combustion Waste Products 50
-------
Chapter 5 Incidental Mercury Use Associated With Non-Coal Sources 52
5.1 Oil Combustion 52
5.2 Carbon Black Production 54
5.3 Gold Mining 56
5.4 Primary Lead and Zinc Mining and Smelting 57
5.5 Primary Copper Mining and Smelting 58
5.6 Pulp and Paper Manufacturing 59
5.7 Oil Refining 60
5.8 Rubber and Plastic Products 60
5.9 Geothermal Power 61
5.10 Wood-Fired Boilers 61
5.11 Utility Natural Gas Combustion 62
Chapter 6 Additional Sources of Mercury Resulting from Disposal or Other Final Disposition 63
6.1 Hazardous Waste Combustion 63
6.2 Crematories 64
6.3 Sewage Treatment and Sludge Incineration 66
6.4 Municipal Waste Combustion 68
6.5 Landfills 68
6.6 Medical Waste Incineration 69
Chapter 7 Geographic Distribution of Mercury 70
7.1 Purpose 70
7.2 Data Sources and Limitations 70
7.3 Findings 72
7.4 Speciation 83
Chapter 8 Conclusions 87
8.1 Conclusions 87
References 94
VI
-------
List of Exhibits
Exhibit 1-1. Summary of the U.S. Mercury Life Cycle 5
Exhibit 1-2. Summary of Mercury Mass Balance and Data Quality 6
Exhibit 1-3. Annual Flow of Mercury in Supply and Manufacturing Sectors 8
Exhibit 1-4. Annual Flow of Mercury Associated with Incidental Mercury Use 9
Exhibit 1-5. Summary of Data Quality for Selected Sectors 10
Exhibit 1-6. Review of Data Quality for Selected Sectors 11
Exhibit 2-1. Mercury in Secondary Mercury Production 17
Exhibit 3-1. Mercury in Chlor-Alkali Manufacturing 21
Exhibit 3-2. Chlor-Alkali Mercury Cell Process Mercury Used , Emitted, Recycled, and Disposed 21
Exhibit 3-3. Mercury in Electrical Lighting 24
Exhibit 3-4. Lighting Industry Mercury Consumed 24
Exhibit 3-5. Mercury in Thermometers 27
Exhibit 3-6. Mercury Used to Manufacture Thermometers in the U.S. in 1997 27
Exhibit 3-7. Mercury Consumption by SIC Code 382 - Measuring and Control Instruments (tons) 28
Exhibit 3-8. Mercury in Thermostats 30
Exhibit 3-9. Mercury in Switches and Relays Manufacturing 32
Exhibit 3-10. Mercury Consumption by SIC Code 3643 - Wiring Device and Switches 33
Exhibit 3-11. Mercury Content of Various Mercury Switches and Relays 33
Exhibit 3-12. Mercury in Dental Preparations 35
Exhibit 3-13. Mercury in Pharmaceuticals 37
Exhibit 4-1. Mercury in Utility Coal Combustion 40
Exhibit 4-2. Efficiencies of Various Control Devices in Removing Mercury from Flue Gas 41
Exhibit 4-3. Mercury in Lime Production 44
Exhibit 4-4. Mercury in Residential, Commercial, and Industrial Coal Combustion 46
Exhibit 4-5. Mercury in Coke Production 48
Exhibit 4-6. Mercury in Cement Manufacturing 49
Exhibit 5-1. Mercury in Utility, Non-utility, and Residential Oil Combustion 53
Exhibit 5-2. Mercury in Carbon Black Production 55
Exhibit 5-3. Mercury in Gold Mining 57
Exhibit 5-4. Mercury in Primary Lead and Zinc Production 58
Exhibit 5-5. Mercury in Primary Copper Production 59
Exhibit 5-6. Mercury Content of Crude Oil and Petroleum Products 60
Exhibit 5-7. Mercury in Oil Refining 61
Exhibit 6-1. Mercury in Hazardous Waste Combustion 64
Exhibit 6-2. Mercury in Crematories 65
Exhibit 6-3. Mercury Flow in Sewage Treatment 67
Exhibit 7-1. List of Data Sources for Geographic Distribution of Mercury 71
Exhibit 7-2. Total Mercury Releases by States (1999 TRI) 73
Exhibit 7-3. Total Mercury Air Releases by States (1999 TRI) 73
Exhibit 7-4. Total Mercury Air Release by State (NTI 1996) 74
Exhibit 7-5. Mercury Emission Density by State (NTI 1996) 74
Exhibit 7-6. Number of Mercury Spills Reported per State (NRC 2000) 75
Exhibit 7-7. Mercury Fish Advisories by State 76
Exhibit 7-8. Coal-fired Utility Boilers Release of Mercury by State (ICR 2000) 77
Exhibit 7-9. NTI County Density Map for Mercury Compounds 79
vii
-------
Exhibit 7-10. 25 Counties Reporting Highest Mercury Air Emissions from NTI 80
Exhibit 7-11. 25 Counties Reporting Highest Mercury Air Emissions from Coal-fired Boilers 81
Exhibit 7-12. 25 Counties Reporting Highest Multimedia TRI Releases 82
Exhibit 7-13. 25 Counties Reporting Highest Releases from 3 Data Sources 83
Exhibit 7-14. Speciation Emission for Coal-fired Utility Boiler by State 84
Exhibit 8-1. Summary of Sectors with 100+Tons of Mercury in a Life Cycle Stage 90
vin
-------
List of Acronyms and Abbreviations
Btu British Thermal Unit
CaO Calcium Oxide
CGLI Council of Great Lakes Industries
CIBO Council of Industrial Boiler Owners
USDHHS United States Department of Health and Human Services
USDOE United States Department of Energy
EEI Edison Electric Institute
USEPA United States Environmental Protection Agency
EPRI Electric Power Research Institute
ESP Electrostatic Precipitator
FETC Federal Energy Technology Center
FGD Flue Gas Desulfurization
GLNPO Great Lakes National Program Office
GLNWF Great Lakes National Wildlife Federation
HAP Hazardous Air Pollutant
HC1 Hydrogen Chloride
Hg Mercury
LCA Life Cycle Assessment
LCI Life Cycle Inventory
MFA Materials Flow Analysis
MSW Municipal Solid Waste
NaCl Sodium Chloride
NAICS North American Industry Classification System
NaOH Sodium Hydroxide
NEMA National Electrical Manufacturers Association
NPDES National Pollutant Discharge Elimination System
PBT Persistent, Bioaccumulative and Toxic
POTW Publicly Owned Treatment Works
ppm parts per million
ppmwt parts per million (weight)
RCRA Resource Conservation and Recovery Act
SBIR Small Business Innovative Research Program
SIC Standard Industrial Classification
TCLP Toxicity Characteristic Leaching Procedure
TSCA Toxic Substances Control Act
UCR University Coal Research Program
UDI Utility Data Institute
USBM United States Bureau of Mines
USGS United States Geological Survey
WLSSD Western Lake Superior Sanitary District
IX
-------
Acknowledgments
This report was prepared under the direction and coordination of Kenneth R. Stone of the U.S. Environmental
Protection Agency, Systems Analysis Branch, Sustainable Technology Division, National Risk Management Research
Laboratory, Cincinnati, Ohio. This report was prepared by Science Applications International Corporation (SAIC)
in Reston, Virginia, under Contract nos. 68-C6-0027 and 68-C7-0011.
-------
Chapter 1
Introduction
1.1 Background
The United States Environmental Protection Agency
(USEPA) is charged by Congress with protecting the
Nation's land, air, and water resources. Under a
mandate of national environmental laws, the Agency
strives to formulate and carry out actions leading to a
compatible balance between human activities and the
ability of natural systems to support and nurture life.
These laws direct the USEPA to define our
environmental problems, measure the impacts, and
search for solutions. The National Risk Management
Research Laboratory is responsible for planning,
implementing, and managing research, development, and
demonstration programs. These provide an authoritative
defensible engineering basis in support of the policies,
programs, and regulations of the USEPA with respect to
drinking water, wastewater, pesticides, toxic substances,
solid and hazardous wastes, and Superfund-related
activities.
USEPA has identified mercury as one of 12 persistent,
bioaccumulative, and toxic (PBT) substances (63 Federal
Register 63926, November 17, 1998). Although
mercury use is decreasing in the United States, mercury
continues to pose a serious risk to public health and the
environment. For several years USEPA has been
coordinating efforts with other North American
countries, as well as its own national and regional
programs, as part of the PBT Program. The goal of
USEPA's PBT Program is to reduce the risk and future
exposure to PBTs using a cross-goal and cross-media
approach.
Recent USEPA activities have focused on studying and
reducing the impacts of mercury and other PBT
chemicals on the Great Lakes, and developing initiatives
to promote the recycling of mercury-containing products
to reduce the quantity of mercury in landfilled wastes.
These efforts have focused attention on mercury use
disposal and have led to the development of some
valuable use and release data. However, a thorough
inventory of the mercury life cycle has not yet been
developed.
Previous reports published by USEPA, States, and
foreign countries have identified sources and uses of
mercury. These reports include USEPA's \991Mercury
Study Report to Congress, which presented nationwide
air release estimates for individual sectors where
mercury is used intentionally or is present as a
contaminant in raw materials. Additional data
describing mercury use have been available from the
U.S. Geological Survey (USGS) and the U.S.
Department of Commerce's Bureau of Census. These
sources assist in identifying significant industrial and
consumer sectors where mercury is used, released to the
environment, or both. While the quality of the data in
the Report to Congress is necessarily variable due to the
diversity of sectors, it serves as a useful starting point for
information on most sectors where mercury can enter the
environment. Other valuable resources include
USEPA's 1997 Locating and Estimating Air Emissions
from Sources of Mercury and Mercury Compounds and
several other sources developed through the EPA's Great
Lakes National Program Office (GLNPO). The 1997
Mercury Study Report to Congress, for example,
identifies over 40 types of sources releasing mercury to
air.
As part of the Agency-wide PBT Program, the Office of
Research and Development (ORD) is identifying
economic and environmental effects of implementing
possible pollution prevention opportunities. The first
step of this effort is to quantify rates of mercury
production, use, recycling, and environmental releases.
The data in this report will be used in future USEPA and
ORD efforts that identify industry-specific pollution
prevention opportunities and evaluate the economic and
environmental effects of their implementation.
1.2 Purpose of Report
The purpose of this report is to quantify the use and
release of mercury in the United States. Using a
Materials Flow Analysis (MFA), this report quantifies
-------
cradle-to-grave mercury use, reuse, release, and disposal
associated with products and processes that use mercury.
An MFA is largely based on the Life Cycle Assessment
(LCA) concept, which is a process to evaluate the
complete "life cycle" environmental impact of a product,
process, or activity. Rather than take a broad multi-
chemical approach, typical of an LCA, this report
focuses on the flow of mercury throughout each stage,
from raw materials acquisition through ultimate
disposition.
1.3 Methodology Used in Report
This study examines mercury use in the United States
from supply through disposal. To assess the life cycle
implications of mercury use in the United States, this
report used the USEPA's 1997Mercury Study Report to
Congress as a starting point for identifying domestic
sectors where mercury is present. This report identifies
the following five categories of sectors:
• Mercury supply. This accounts for facilities who
sell purified (usually elemental) mercury for
industrial use.
• Mercury use in manufacturing processes. Facilities
may use mercury as part of a production process
(e.g., in chlor-alkali production, mercury is used on-
site but is not intended to be in products) or in a
mercury-containing product (e.g., fluorescent
lamps).
• Incidental mercury use associated with coal
combustion. Mercury is a coal contaminant. Coal is
used not only for large-scale power generation, but
in other areas for combustion or as a raw material.
• Incidental mercury use associated with non-coal
sources. Mercury is present as a contaminant in
non-coal fossil fuels and in mined minerals.
• Other sources of mercury resulting from previous
use. In some cases, the life cycle of mercury cannot
be easily traced. For example, mercury may be
found in wastewater which is then treated as sewage.
While the initial sources of the mercury discussed in
this category may be covered under another sector,
such as the manufacture of a mercury bearing
product, mercury from previous use is addressed in
keeping with the life cycle approach to this report.
For each sector, the following data elements on a
national level and annual basis are identified:
(1) A short description of how mercury is used in
processes or products.
(2) Raw materials containing mercury, and the quantity
of mercury in these raw materials.
(3) Products containing mercury, and the quantity of
mercury in these products.
(4) Potential release points of mercury, and quantitative
emissions of mercury to air, water, and land.
(5) A discussion of the quality and consistency of the
above elements.
For each sector, these estimates are presented in a flow
diagram format. The basis for the estimates (e.g., the
data source and/or calculation method), the quality of the
underlying data, and their uncertainty are also presented.
This format allows easy identification of sectors in
which information regarding raw materials, product
content, and mercury release is incomplete, uncertain, or
contradictory. When such instances occur, a speculation
is made as to whether these mass imbalances are
reflective of unreported emissions, data quality
limitations, or other reasons. The methodology for
obtaining these estimates is presented in the individual
chapters of this report.
As a final step in the methodology, maps were created to
show which regions of country release the largest
amounts of mercury. The emissions from several data
sources were mapped showing the quantity and density
of mercury emissions. In addition, available speciation
data were added into the maps.
1.4 Scope of Report
Using an MFA, this study profiles mercury in raw
materials acquisition, product manufacturing, use,
recycling, and final disposition. Additionally, it
provides information from data sources that track
mercury releases to different media. The report
identifies uses of mercury for which data are available,
and identifies uses of certain raw materials (such as coal)
where mercury is present as a contaminant. However, all
such raw materials are not identified, and all products
containing mercury (for widespread downstream use) are
not included.
In addition, this report attempts to incorporate speciation
data into the analysis. Data sources were examined and
mapped to provide an overview of which regions of the
United States emit the largest amounts of mercury.
Mercury speciation data were then collected and used to
quantify each species of mercury released for a particular
sector. Speciation data were available only for the
combustion of utility coal and municipal waste.
-------
The results of this study can be used in conjunction with
other data to better identify mercury use and release
patterns. For most of the sectors identified in this report,
pollution prevention opportunities are discussed in the
report titled Identifying Research and Development
Priorities to Reduce Mercury Use and Environmental
Releases in the United States. This report containing
pollution prevention opportunities and the MFA are
intended to be used as companion documents. The
results of these two efforts could be used to help target
and reduce the mercury use in specific sectors.
1.5 Summary of Results
7.5.7 Description of Exhibits
Exhibits 1-1 through 1-4 summarize the findings of this
report for mercury use sectors and their categories. A
"sector" is a single industry or class of similar products
or processes where mercury is present. Subsequently,
sector is a very flexible term because it can mean, for
example, the industry of chlor-alkali manufacturing, the
production and use of mercury containing-lights, or the
process of oil combustion. As discussed above, there are
five categories of sectors presented in this report: (1)
mercury supply; (2) manufacturing and use of
manufactured products; (3) incidental usage of mercury
in coal; (4) incidental usage of mercury in non-coal
materials; and (5) other sources resulting from previous
use. Each exhibit identifies these major categories,
while Exhibits 1-2 through 1-4 identify the individual
sectors comprising the categories.
These exhibits present the flow of mercury on a national
aggregate annual basis for mercury sectors and
categories of sectors. The actual flow changes from year
to year as a result of changes in demand, industry
initiatives to reduce or measure mercury, and
implementation of regulatory efforts by local, state, and
federal governments. Additionally, the same quality of
data are not available on a consistent basis from year to
year for all sectors or even within individual sectors,
therefore the data presented in these exhibits are
intended to represent mercury flow in the latest year
where data are available, rather than for any specific
year.
Exhibit 1-1 provides an overview of the available data.
The data are presented according to major categories.
The purpose of Exhibit 1-1 is to identify principal flows
of mercury through the U.S. Economy. Mercury sources
are listed at the top of Exhibit 1-1. Each of these
constitute 'inputs' to the flow of mercury in the United
States. Mercury then flows to subsequent categories,
including manufacturing and use, or straight to final
disposition (represented as wastes, exports, and
recycling).
Exhibit 1-2 quantifies the flow of mercury for the
individual sectors evaluated in this report. The first
column accounts for mercury inputs for each sector; this
includes the use of elemental mercury and the presence
of mercury as a contaminant in raw materials. The next
three columns indicate the quantity of releases of
mercury into the environment for air, water, and solids.
The last column presents the quantity of mercury that
has accumulated as a result of historical use (reservoir).
Other notations are used in Exhibit 1-2 to acknowledge
that specific sectors represent a source of mercury but
that an estimate is not provided. Reasons include an
overall lack of data or unreliable data, or that all relevant
quantities are small relative to the other sources
presented in this exhibit (less than one ton). The basis
for the estimates, or a discussion of why no estimate is
presented, is provided in the subsequent chapters of this
report.
Exhibits 1-3 and 1-4 present further sector-specific detail
for each of the sectors evaluated in this report. Exhibit
1-3 illustrates the quantities of mercury in supply,
manufacturing processes, and subsequent use of
manufactured products. These are discussed in Chapters
2 and 3 of this report. Exhibit 1-4 illustrates the
quantities of mercury in cases where mercury is
incidentally used and in final disposition. These are
discussed in Chapters 4 through 6 of this report. Data
are presented for the quantity consumed (present in raw
materials) which was previously shown in Exhibit 1-2.
In these exhibits, however, the outputs are detailed
according to the quantities in product, exports,
multimedia releases (sum of air, water, and land), and
recycling.
1.6 Data Limitations and Uncertainty
The quantitative data in this report are based on a
combination of government data, industry data, and
estimates. The quality of these data often differ by
sector due to industry-specific initiatives undertaken in
response to regulatory or voluntary efforts, such as
reporting requirements. The use and management of
mercury has been changing in the last several years in
response to these initiatives, which makes even recent
mercury use data unreliable. Where this is the case,
-------
trend information has been presented, where available,
to give an indication of the changes in process. Finally,
it is unusual to obtain consistent information on raw
material use, releases, recycling, and product content.
The result is that the data presented in this report are of
varying quality and is subject to future change. It is not
possible to obtain data for a single consistent year
throughout the report. For example, some data are
applicable for 1995, others for 1999, and other estimates
are calculated using data for a variety of years. This was
a function of two factors:
• Use of data which was readily available; for
example, the Toxics Release Inventory (TRI) data.
The most recent data available, from the 1999 TRI,
were used throughout this report.
• Data were only available for a certain year; for
example, Department of Commerce (Bureau of
Census) data. Until 1994, the Department of
Commerce collected data for lamp production. In
this report, such data were used because it was the
most recent. However, based on later indirect
indications as discussed in the report, such data may
be adequate for approximating current production.
USEPA has finalized revisions to the Toxics Release
Inventory (TRI), which reduce the reporting threshold of
mercury to 10 pounds (64 Federal Register 58666,
October 29, 1999). These changes took effect for the
year 2000 reporting year, and the data were not available
to the public until mid-2002. These changes mean that
a facility that uses as little as 10 pounds of mercury in a
year will be required to report its release and recycling
activities (even if its releases are zero). Under previous
reporting years, a facility was not required to report its
releases unless it used 10,000 pounds of mercury in a
year. The result of this change is that many more
facilities that use and release mercury will be reporting
this information, which will improve the quality of
several of the numerical estimates presented in this
report.
Additional caution should be used for Exhibits 1-2
through 1-4. Estimates are sometimes presented as
single numbers, and sometimes as ranges. Ranges are
used to express the range of uncertainty in the available
data. However, there is uncertainty associated with
much of the data expressed as discrete values, which is
discussed in detail in the individual chapters of this
report.
Integrating and interpreting data from various
information sources in orderto characterize processes, in
this case the flow of mercury through key industries,
inherently results in some degree of uncertainty. In this
study, some sources of information were equivocal due
to factors such as low sample sizes, conflicting reports,
or a of lack of data. The degree of uncertainty in data
collected for this report varies among sectors and by
estimate (releases, consumption, and reservoirs).
Despite associated uncertainties, estimations are
expected to provide a relative scale ranging from
industries that are heavily involved with mercury to
those that are associated with the element to a much
lesser degree.
In this report, sources of uncertainty are addressed in
detail for each sector. The purpose of this section is to
provide a summary of the uncertainty associated with the
mercury usage estimates of several key sectors. Exhibit
1-5 lists the mercury quantities associated with these key
sectors (use, release, and reservoir) and assigns each
relevant estimate a data quality score in order to aid
interpretations of this report's findings. Data scores use
an A, B, or C system designed to render the sector
estimates into 3 tiers of data confidence. Exhibit 1-6
discusses the supporting materials within each scored
industry and is intended to elucidate the basis of the data
quality scores. Justifications of data quality scores for
remaining sectors are not explicitly identified, rather, a
discussion of data quality for each sector is provided in
their respective chapters.
-------
The U.S. Mercury Life Cycle
Raw
Materials
Acquisition
Product
Manufacture
Domestic sources of
mercury
Product Use
Final
Disposition
Imported sources
of mercury
Mercury used in
products and
processes
Combustion sources of
mercury
Mercury found in
commercial and
professional products
Recycled
mercury
wastes
Mercury
releases to air,
water, and land
Exported
mercury
Exhibit 1-1. Summary of the U.S. Mercury Life Cycle
-------
Exhibit 1-2. Summary of Mercury Mass Balance Data
Mercury Sector
Mercury Supply (discussed in Chapter 2)
Secondary Mercury Production
Imports and Exports
U.S. Government Stockpiles
Total for Mercury Supply
Mercury Use in Manufacturing Processes (discussed in Chapter 3)
Chlor-alkali Manufacturing
Electrical Lighting: Manufacturing
Electrical Lighting: Use and Disposal
Thermometers: Manufacturing
Thermometers: Use and Disposal
Thermostats: Manufacturing
Thermostats: Use and Disposal
Switches and Relays: Manufacturing
Switches and Relays: Use and Disposal
Organic Chemical Production
Dental Preparations Manufacturing
Dental Office: Use
Mercury Compounds
Batteries
Total for mercury use in manufacturing
Mercury Used in Raw
Materials (ton/yr)
430
-83
0
347
79
16
17*
9- 17
9-17*
15-21
13-20*
36-63
36-63*
—
34-54
34 - 54 *
—
negligible
189 - 250
Waste Releases (ton/yr)
Air
0.4
0
0
0.4
6.3
0.3
3
<0.2
2-3
0
1-2
0
7-13
—
0
0.8
—
negligible
20.4 - 28.5
Water
0
0
0
0
0.1
—
0
0
0
0
0
0
0
—
0
7.4
—
negligible
7.5
Solid
0.1
0
0
0.1
21.5
0.2
11
0
7-14
0
6-8
0
29-50
—
0
0
—
negligible
74.7-104.7
Mercury Reservoir
(ton)
—
—
4,850
4,850
2,000
0
65-75
0
45-85
0
230
0
630
—
0
1,200
—
—
4,170-4,220
Incidental Mercury Use Associated with Coal Combustion (discussed in Chapter 4)
Utility Coal Combustion
Lime Manufacturing
Residential/Comm/Industrial Coal Combustion
Byproduct Coke Production
Portland Cement Manufacturing
Coal Combustion Wastes
Total for Coal Combustion
Incidental Mercury Use Associated with Non-Coal Sources (discussed
105
2.7-5.0
7.6-21.2
3.2
3.4-5.7
3
124.9-143.1
in Chapter 5)
48
0.1
21.2-23.6
0.7
4.2
0
74.2 - 76.6
1
0
0
0
0
0
7
33
0.1
0.6
1.5
0.6
3
38.8
0
0
0
0
0
—
0
-------
Waste Releases (ton/yr)
Mercury Sector
Oil Combustion: U.S. Utility
Oil Combustion: Other
Carbon Black Production
Gold Mining
Primary Lead and Zinc Mining and Smelting
Primary Copper Mining and Smelting
Mercury Used in Kaw
Materials (ton/yr)
0.06
0.08
0.11
1,370
0.2
6.4
Air
0.2
7.8-10.9
0.3
6.2
0.2
0.1
Water
0
0
0
0
0
0
Solid
<0.6
<0.1
0
1,342
7.6
16.4
Mercury Reservoir
(ton)
0
0
0
—
0
0
Pulp and Paper Manufacturing —
Oil Refining 2.2-11.5
Rubber and Plastic Products —
Geothermal Power —
Wood-fired Boilers —
Utility Natural Gas Combustion negligible
Total for Incidental Non-Coal Sources 1,379 - 1,389
Other Sources of Mercury Resulting from Previous Use (discussed in Chapter 6)
Hazardous Waste Combustion 1.0
Crematories 1.4
Sewage Treatment and Sludge Incineration 12
Municipal Waste Combustion —
Landfills —
Medical Waste Incineration —
Historical Mining Activities —
Total for Previous Use Sources 14
negligible
14.8-17.9
7.1
<0.1
O.9
negligible
0
0
0
5.5
negligible
1,365
0
0
5.5
5.5
5.5
— No data available or estimate was not made.
* Not included in subtotal to avoid 'double counting.'
-------
Exhibit 1-3. Annual Flow of Mercury in Supply and Manufacturing Sectors
Mercury Sector
Manufacturing (tons/year) Use (tons/year) Disposal (tons/year)
Consumption Product Exports Releases Recycling Domestic Imports Releases Recycling
Source
Supply of Mercury (Discussed in Chapter 2)
Recycling
Net Imports
U.S. Government Stockpiles
Subtotal
Manufacturing Processes Involving Mercury
Chor-Alkali
Electrical Lighting
Thermometers
Thermostats
Switches and Relays
Organic Chemical Production
Dental Preparations
Mercury Compounds
Batteries
Subtotal
430
-83
0
347
(Discussed in Chapter 3)
79
16
9-17
15-21
36-63
—
34-54
—
—
189-250
429 0 1.2 o — - -
-83—0 o — - -
0 ______
346 0 1.2 0 — - -
<0.5 0 27.8 13.04 — — — —
16 2 0.5 1.1 14 3 15 2
9-17 — <0.2 — 9-17 — 9-17 —
11-17 2-3 0 4.1 9-14 4-6 7-10 —
36-63 — 0 0 36-63 — 36-63 —
— — — — — — — —
34-54 0 0 2.5 34-54 — 8.2 —
— — — — — — — —
— — — — — — — —
106-167 4-5 28.26 18.24 102-162 7-9 75-113 2
Estimates are on an annual basis of mercury (short tons per year).
-------
Exhibit 1-4. Annual Flow of Mercury Associated with Incidental Mercury Use
Mercury Sector
Incidental Mercury Use Associated with Coal Combustion
Coal Combustion by Utilities
Lime Manufacturing
Residential/Commercial/Industrial Coal Combustion
Coke Production
Portland Cement Manufacturing
Coal Combustion Wastes
Subtotal
Incidental Mercury Use Associated with Non-coal Sources
Oil Combustion
Carbon Black Production
Gold Mining
Primary Lead and Zinc Mining and Smelting
Primary Copper Mining and Smelting
Pulp and Paper Manufacturing
Oil Refining
Rubber and Plastic Products
Geothermal Power
Wood-fired Boilers
Utility Natural Gas Combustion
Subtotal
Consumption
(tons/yr)
Product Releases
(tons/yr) (tons/yr)
Recycling
(tons/yr)
or Coal Use (Discussed in Chapter 4)
105
2.7-5.0
7.6-21.2
3.2
3.4-5.7
3
125-143
(Discussed in Chapter 5)
0.14
0.11
1,370
0.2
6.4
—
2.2-11.5
—
—
—
—
1,379-1,389
3 88
— 0.2
0 21.8-24.2
0 1.8
0 4.8
0 3
3 120-122
0 8.7-11.8
— 0.28
0 1,348
0 7.8
0 16.4
— —
0.93 —
— —
— —
— —
— —
0 1,380-1,383
0
0
0
0
0
0
0
0
0
21.6
0
0
—
—
—
—
—
—
22
Additional Sources of Mercury Resulting from Disposal or Final Disposition (Discussed in Chapter 6)
Hazardous Waste Combustion
Crematories
Sewage Treatment and Sludge Incineration
Municipal Waste Combustion
Landfills
Medical Waste Incineration
Historical Mining Activities
Subtotal
1
1.4
12
—
—
—
—
14
0 7.1
0 <0.1
0 12
— —
— —
— —
— —
0 19
0
0
0
—
—
—
—
0
Estimates are on an annual basis of mercury (short tons per year).
-------
Exhibit 1-5. Summary of Data Quality for Selected Sectors
Sectors Ranked by Total Releases
Sector Total
(tons/year) Data Quality
Gold Mining
Utility Coal
Combustion
Switches and Relays:
Use and Disposal
Chlor-alkali
Manufacturing
Thermometers: Use and
Disposal
Thermostats: Use and
Disposal
Dental Preparations
Secondary Mercury
Production
Oil Refining
Imports and Exports
U.S. Government
Stockpiles
1,348
88
36-63
27.8
9-17
7-10
8.2
0.5
—
—
—
B
A (air)
B (others)
B (releases)
C (recycling)
B
B (releases)
C (recycling)
B
B
C
C
—
—
Sectors Ranked by Consumption
Sector Total Data
(tons/year) Quality
Gold Mining
Secondary Mercury
Production
Utility Coal
Combustion
Imports and Exports
Chlor-alkali
Manufacturing
Switches and
Relays: Use and
Disposal
Dental Preparations
Thermostats: Use
and Disposal
Thermometers: Use
and Disposal
Oil Refining
U.S. Government
Stockpiles
1,370
430
105
-83
79
36-63
34-54
13-20
9- 17
2.2- 11.5
—
B
B
A
A
A
B
B
B
B
B
—
Sectors Ranked by Reservoir
Sector Total Data
(tons) Quality
U.S. Government
Stockpiles
Chlor-alkali
Manufacturing
Dental Preparations
Switches and Relays:
Use and Disposal
Thermostats: Use and
Disposal
Thermometers: Use and
Disposal
Secondary Mercury
Production
Utility Coal Combustion
Imports and Exports
Gold Mining
Oil Refining
4,850
2,000
1,200
630
230
45-85
—
—
—
—
—
A
B
C
B
B
B
—
—
—
—
—
Data Quality Legend: A: Expected to be well documented, B: Data available but uncertain, C: Very little data available
10
-------
Exhibit 1-6. Review of Data Quality for Selected Sectors
Estimate
Data Source(s)
Review of supporting materials
Data Quality
Secondary Mercury Production
Consumption
Releases
USGS data from
1997
US EPA 1997
Mercury Report to
Congress
1999 TRI
USGS estimate of 430 tons is from latest year available. This value is uncertain and higher than a
conflicting industry estimate. Reassess estimations in 2000 TRI reports when more mercury
recyclers may report.
Air releases estimate from Mercury Report to Congress extrapolation based on 1994 TRI data.
Uncertainty results from changes in facilities and the industry since 1994. Water and solid releases
are uncertain due to low sample size (n = 2). Reassess all estimations in 2000 TRI reports when
more mercury recyclers may report.
B
C
Imports and Exports
Consumption
USGS data for
2000
Estimate from Census Bureau is expected to be accurate. Principle uncertainty is that no data are
available on trade of mercury -containing scrap or waste. Figure based on available trade data.
A
U.S. Government Stockpiles
Reservoirs
US Defense
Logistics Agency
Estimate expected to be accurate, from agency in charge of stockpile management.
A
Chlor-alkali Manufacturing
Consumption
Releases
Reservoir
Chlorine Institute
data for 2000
1999 TRI
US EPA 1997
Mercury Report To
Congress and
Chlorine Institute
(2000 data)
Data are based on a survey of all eleven industry facilities over a 4 year period. Principle
uncertainty is that usage data are variable from year to year, possibly reflecting intermittent use.
Data are based on the response of 13 mercury cell plants representing 96% of total production.
There may be some error associated with plant's ability to measure releases.
Broad estimation predicated on large storage of mercury in plants, mercury cell capacity, and
contamination in pipes, equipment, etc. based on number cells in operation (accurately known) and
quantity of mercury per cell (not accurately known or uniform).
A
B
B
11
-------
Estimate
Data Source(s)
Review of supporting materials
Data Quality
Thermometers: Use (Rather than Manufacturing)
Consumption
1997 Bureau of
Census Data
US EPA report
from 1992
regarding mercury
in municipal waste
High-end estimate. The number of thermometers produced, as recorded by the Bureau of Census,
does not delineate data for liquid-in-glass thermometers sold by liquid type. Estimations of mercury
content per thermometer from a 1992 EPA report are expected to be accurate.
Key assumptions: No import or export data available. All liquid-in-glass thermometers contain
mercury.
B
Releases
Same as
Consumption
references
Release projections are based on consumption values, which consist of some uncertainties.
Estimation of landfill waste releases (7 -14 tons) similar to prior EPA estimations (16.3 tons in
1989). No estimate for recycling.
Key assumptions'. 80% of thermometer mercury is landfilled and 20% combusted, consistent with
overall municipal solid waste management.
C (releases)
C (recycling)
Reservoir
USEPA 1992b
The validity of this estimate is unclear due to lack of data (e.g. number of mercury thermometers
still in use).
Key assumptions'. Thermometer life-span of 5 years. Mercury consumption is 9 - 17 tons per year.
B
Thermostats: Use (Rather than Manufacturing)
Consumption
1997 Bureau of
Census Data
US EPA report
from 1992
regarding mercury
in municipal waste
The ultimate calculation of this value integrates figures from 3 separate studies (average amount of
mercury per thermostat, number of thermostats produced annually, and percentage of thermostats
exported, imported, and sold domestically). Potentially, the sum of any errors from each study may
be significant, although the largest error is likely to be in the thermostat production data.
Key assumptions'. Even distribution of mercury and non-mercury devices among total thermostats
produced.
B
Releases
US EPA from
1994 report
The estimate is primarily based on the number of thermostats brought out of service in 1994. While
this number has certainly changed 7 years later, the largest uncertainty for a present-day estimate is
the quantity recycled, which is unknown.
Key assumptions'. 80% of solid wastes are landfilled and 20% incinerated, consistent with overall
municipal solid waste management.
C (releases)
C (recycling)
12
-------
Estimate
Data Source(s)
Review of supporting materials
Data Quality
Reservoir
US EPA reports
from 1992 and
1994
The estimate is on the low-end since it is based on an estimation of thermostats in use among U.S.
residences only and not commercial or government sites.
B
Switches and Relays: Use (Rather than Manufacturing)
Consumption
USGS 1997 data
US 1997 Bureau of
Census data
The USGS estimate of 63 tons excludes mercury reed relays, which typically have a high mercury
content but have unknown production. The 36 tons/year estimate is based on US DOC data may be
highly variable because it includes non-mercury switches and also excludes mercury reed relays.
Switches and relays, in general, are difficult to track due to the high number of categories used to
describe them.
B
Releases
Same as
Consumption
references
The total releases were estimated by assuming that outflow equals inflow (disposal = consumption).
No data were available for recycling.
Key assumptions: The amount used in switch and relay manufacturing must eventually be disposed,
with 80% landfilled and 20% incinerated, consistent with overall municipal solid waste
management.
C (releases)
C (recycling)
Reservoir
USGS 1990 -1997
data
The estimate is but probably low. USGS estimated the consumption of mercury containing wiring
devices and switches over the period of 1990 to 1997. The 630 ton reservoir figure is the sum of
consumption over these 8 years. Switch life is typically greater than 8 years. Moreover, the data do
not include all types of mercury switches and relays or imports.
B
Dental Preparation
Consumption
USGS 1997 data
US Dept. of Health
and Human
Services 1993
study
1994 literature
paper
The range provided (34 - 54 tons/year) represents 2 different estimates. The lower estimate of 34
tons / year was provided by USGS (1997), the most recent year available. The higher estimate is
based on the number of fillings and the mercury content of fillings.
Key assumptions: Use of amalgams continued to decrease through time.
B
13
-------
Estimate
Releases
(From Dental
Offices)
Reservoir
(Population)
Data Source(s)
US EPA 1997
Mercury Report to
Congress
1996 research by
Arenholt
1999 DAMS report
Same as
Consumption
references
Review of supporting materials
Estimated releases are probably conservative. Low-end estimates for average mercury in dental
wastewater and the number of dental offices were used. Air releases based on assumption; waste
water releases from single study and conservative estimate of the number of dental offices.
Key assumptions: Two percent of the total amount of mercury used is emitted from spills and scrap.
Very broad estimate based consumption data.
Key assumptions: Twenty to forty year life-span for fillings.
Data Quality
B (releases)
C (recycling)
C
Utility Coal Combustion
Consumption
Releases
1999 EPA ICR and
US EPA 1997
Report to Congress
US EPA Fossil
Fuel Waste
Reports To
Congress in 1998
and 1999
The estimate of mercury into utility coal combustion is based on consumption and sampling data at
over 450 coal-fired utilities during 1999.
Key assumptions: 15% of the mercury present in coal nationwide is removed prior to introduction to
the boiler. This is based on average coal cleaning efficiency from the 1997 Report to Congress.
The approximation of air releases (40 tons/year) is expected to be reliable since it was based on a
large sample of measurements during 1999. Measurements for solid waste are not as extensive
although waste generation quantities are based on relatively recent (1997) facility survey data.
Water releases (7 tons per year) may be understated since generation and / or sampling data are
unavailable for several wastewater sources including: Pile runoff, boiler blowdown, gas-side wastes,
and FGD liquor.
A
A (air)
B (others)
Gold mining
Consumption
and
Releases
1999 TRI
Total releases from gold mining is based on 1999 TRI data submitted by 8 Nevada-based facilities,
which were probably the largest sites but not all of the sites. As smaller facilities report year 2000
releases, the estimate may rise slightly. Reported air emissions are suspected to be based on
estimates and not measurements. Other media releases are of unknown quality.
Consumption estimate is largely based on same uncertainties as the releases estimate.
Key assumptions: Mercury input from trace impurities in gold ore to the gold mining process is
assumed to be equal to the amount released.
B
14
-------
Estimate
Data Source(s)
Review of supporting materials
Data Quality
Oil Refining
Consumption
and
Products
Releases
1999 sampling
data from
Minnesota and US
DOE Petroleum
Supply Annual for
2000
1999 TRI
The throughput data tabulated by US DOE is highly reliable. The mercury content of crude oil,
however, is variable by nature and the mercury concentration published by Minnesota is based on a
sample size of only two refineries.
Waste releases according to 1999 TRI data appeared to be minimal but were largely unavailable for
most refineries. Only six oil refineries and bulk fuel terminals reported out of approximately 150
facilities.
B
C
15
-------
Chapter 2
Supply of Mercury
Mercury in the United States is supplied by three
sources: secondary mercury production (recycling),
imports, and a government stockpile. Presently, the
government stockpile does not actually supply mercury,
but rather represents a source of mercury which must be
managed at some point in the future. Mercury is also
obtained as a byproduct of gold mining, as discussed in
Section 5.3 of this report.
2.1 Secondary Mercury Production
Facilities conducting secondary mercury production may
be classified under one of several different business
classifications:
SIC Code 2819: Industrial Inorganic Chemicals, Not
Otherwise Specified (including redistilled mercury)
NAICS Code 325188: All Other Basic Inorganic
Chemical Manufacturing
SIC Code 3341: Secondary Smelting and Refining
of Nonferrous Metals (except copper and aluminum)
NAICS Code 331492: Secondary Smelting,
Refining, and Alloying of Nonferrous Metals
(except copper and aluminum)
SIC Code 4953: Refuse Systems
NAICS Code 562211: Hazardous Waste Treatment
and Disposal
Secondary mercury production is the production of
mercury through processing scrapped mercury-
containing materials. Mercury-containing materials that
may be recycled include dental amalgam, spent batteries,
electrical switches, control instruments, thermometers,
spent catalysts from chlorine and caustic soda production
and laboratory and electrolytic refining wastes. These
waste products are sent to mercury recycling facilities,
which then process the waste to produce mercury for
resale.
The production of secondary mercury from scrap began
rising in 1990, as industrial consumption has been
falling. In 1997, the most recently reported year,
secondary mercury production was 430 tons and
domestic consumption was 381 tons (USGS 2002a).
2.1.1 Mercury Recovery Process
Mercury may be recovered using two methods:
extractive processes to recover mercury from scrap, and
removal of liquid mercury from dismantled equipment.
The extractive processes may involve either thermal or
chemical treatment; thermal treatment is the most
common. Extractive processes are used to recover
mercury from scrapped products as well as industrial
solid and liquid waste when liquid mercury cannot be
drained. Because these methods involve chemical and
thermal manipulation of the mercury, extractive
processes are more likely to result in higher mercury
emissions and waste.
Thermal Extractive Process
In thermal extraction processes, mercury-bearing scrap
is heated to about 538°C (1000°F) to vaporize the
mercury. The mercury vapors are condensed and the
mercury is collected under water. Vapors from the
condenser are combined with vapors from the mercury
collector line, then purified with an aqueous scrubber to
remove particulate matter (PM) and acid gases such as
HC1 and SO2. Organic matter is removed by passing the
vapor through a charcoal filter, then the vapor is
discharged to the atmosphere.
Chemical Extractive Process
There are several chemical methods for extracting
mercury from aqueous mercury-bearing waste streams.
Metallic mercury may be precipitated by treating the
waste stream with sodium borohydride or passing the
waste stream through a zinc-dust bed. Mercuric sulfide
may be precipitated using a water-soluble sulfide. Ionic
mercury may be recovered using ion-exchange systems.
Mercuric ions may be trapped with a chemically
modified cellulose.
16
-------
Liquid Removal Process
Liquid mercury may be removed from waste mercury-
containing equipment by dismantling the equipment and
draining the liquid mercury. Mercury recovered through
this process is distilled to purify the product before
resale.
The liquid removal process is easier and less expensive
than extractive processes, and may be used wherever
liquid mercury can be effectively drained from
equipment. Because there is no heating or chemical
processing of the mercury under these circumstances,
very little mercury emissions or other mercury waste is
expected to be generated by liquid removal activities.
2.1.2 Materials Flow
Exhibit 2-1 illustrates the consumption, release, and
product content of mercury in secondary mercury
recovery.
Input
While there are no quantities available for the amount of
mercury contained in the equipment, scrap, and waste
sent to the recycling facilities, it was assumed that this
amount is equal to the mercury emissions plus the
mercury recovered, presented in Exhibit 2-1.
Consumption:
430tons/yr
Secondary
Mercury
Production
Product:
430 tons/yr
Releases: 0.5 tons/yr
^-Air: 0.4 tons/yr
- Water: 0 tons/yr
- Disposal: 0.135 tons/yr
Sources/a Mercury Release and Recycling: Air releases from EPA (1997a). Disposal releases
from 1999 TRI data.
Mercury in Product: USGS (2002a).
Mercury Consumption: Assumed equal to mercury in product and mercury recycled.
Exhibit 2-1. Mercury in Secondary Mercury Production
Output
In 1997, an estimated 389 metric tons (430 tons) of
secondary mercury were produced (USGS 2002a). An
industry source estimates that a much lower quantity of
mercury is produced: 75 to 150 tons (Lawrence 2000).
Another data source is TRI which requires all facilities
to estimate and report the quantity of chemicals recycled
onsite, if any. Such data for mercury represents the
quantities actually recovered. Their most recent data
show that the total quantity recycled onsite is 52.5 tons
(two provided data for 1999). However, facilities are not
required to report TRI releases if they use less than
10,000 pounds per year, therefore, some releases may
not be included in these data.
Due to the uncertainty in the TRI data, the USGS data
are considered more reliable and complete; these data are
presented in Exhibit 2-1.
Air Releases
Two techniques used for estimating air emissions are
presented here: the use of facility-aggregated reported
releases and the use of an emission factor. Each
technique generates different results, with advantages
and disadvantages to each.
The facility-aggregated approach totals reported air
releases from all facilities in the industry. Two mercury
recovery facilities reported air emissions from their 1999
TRI data, totaling 0.003 tons. This estimate is
incomplete because many additional mercury recovery
17
-------
facilities did not submit TRI reports.
As a more appropriate alternative, emission factors can
be used to estimate industry-wide emissions, based on
data from a small number of facilities and extrapolated
to the industry as a whole. This technique was used in
the Mercury Study Report to Congress, where an
estimate of 0.4 tons was developed by extrapolating
available air emissions data (using the 1994 TRI as a
source) to the industry as a whole (USEPA 1997a). This
estimate, however, also leads to uncertainty because
some secondary mercury producers claim to not emit
mercury to the atmosphere when they recycle, and it is
difficult to account for this variability among reporters
using available information. The air release estimates
from EPA (1997a), generated using this approach, are
shown in Exhibit 2-1.
Water Releases
Wastewater is generated during the vapor-condensing
phase of thermal extraction. These liquid wastes are
filtered to remove impurities (such as mercury). An
industry-wide estimate for mercury releases to water is
not available. TRI data are available from 1999 for two
mercury recovery facilities. No releases were reported
for either facility. Therefore, this release is shown as
zero in Exhibit 2-1.
Reservoir
Mercury recycling facilities are expected to have State or
federally permitted storage requirements for incoming
wastes. No information is available regarding the
quantities of mercury awaiting recycling by consumers
or other users.
Solid Waste Releases
Solid waste is generated as a byproduct after mercury is
removed from scrap and equipment. Most of this solid
waste is disposed of in landfills. A relatively small
amount of solid waste is sent for further treatment or
recycling to recover metals other than mercury.
Two techniques used for estimating releases to land
disposal are presented here: the use of reported facility-
aggregated releases and the use of a release factor. Each
technique generates different results and offers
advantages and disadvantages.
Release factors can be used to estimate industry-wide
releases, based on data from one or two facilities and
applied to the industry as a whole. Based on the
sampling of 'before and after' wastes from mercury
recovery operations of the late 1980s, approximately 98
percent of mercury is recovered as product with the
remaining present in the residue (USEPA 1998a).
Therefore, using this 2 percent loss rate applied to the
1997 production level of 430 tons results in a loss of 8.6
tons. This estimate may be high due to improved
process efficiencies in recent years.
Using the 1999 TRI data, two mercury recovery facilities
reported releases to land and to other sites likely to treat
waste prior to land disposal (e.g., commercial water
treatment, waste brokers). These releases totaled 0.135
tons and are expected to be low because the estimate
omits other recycling facilities. This quantity is used in
Exhibit 2-1, however, because of the age of the above
recovery data.
2.1.3 Discussion
There is some uncertainty in the estimated quantity of
mercury recovered in the United States. The quantity
presented in Exhibit 2-1 is based on USGS data.
However, one mercury recovery company has indicated
that the USGS data may be a high estimate because of
the practice of extrapolating results for non-reporting
facilities (Lawrence 2000).
There is some uncertainty in the amount of mercury in
air and solid waste releases emissions generated from
secondary mercury production. The 0.4 tons per year
cited earlier for air releases is an extrapolated quantity
based on EPA (1997a). However, each of the different
methods of secondary mercury production (thermal
extraction, chemical extraction, and liquid drainage)
produce widely varied amounts of air emissions. The
quantity of mercury remaining in the waste residue is
also dependent on the form of mercury and the extent to
which it can be removed from the spent material. The
plants not reporting emissions data may be using a
different combination of recovery processes than the
plants for which data are available. New mercury
recovery facilities open each year, so all aggregate
estimates may be low. Without detailed knowledge of
how much mercury-bearing scrap, sludge, and
equipment is processed using each method, a truly
representative emissions figure for this sector cannot be
developed. Belter estimates are expected when 2000
TRI data become available in summer 2002 for most
facilities in this industry.
18
-------
2.2 Imports and Exports
Mercury-containing materials are imported to and
exported from the United States. Such materials include
elemental mercury, mercury-containing scrap or waste
for disposal or recovery, and mercury-containing
products such as fluorescent lamps. No data are
available for the trade of scrap or waste, while the trade
of mercury-containing products is discussed in
subsequent chapters for each particular product. Data
for the trade of elemental mercury are presented here.
USGS reported that 113 tons of mercury were imported
and 196 tons of elemental mercury were exported in
2000 (USGS 2002a), resulting in a net export of 83 tons.
Principal trading partners include Australia and Germany
(sources of imported mercury) and India and the
Netherlands (destinations for exported mercury) (USGS
2000a).
2.3 U.S. Government Stockpiles
The U.S. Government has previously purchased mercury
for the National Defense Stockpile to satisfy contingency
requirements for national emergencies. A total of 4,850
tons are presently being stored in five locations in the
eastern United States. Note that this cannot be directly
compared to the other supply sources because this is a
'one-time' quantity while the others are annual
quantities.
There has been no need for stockpiled mercury as a
national security requirement, and sales of this material
have been suspended since 1994. Therefore, mercury is
not presently being added to or removed from the
stockpile. The Defense Logistics Agency, which is
responsible for maintaining the stockpile, is in the
process of preparing an environmental impact statement
to examine alternatives for stockpile management. At
present, these alternatives include no action, and the
consolidation of supply, sales, and disposal (DLA
2002). Identification and implementation of an
alternative will likely take a minimum of several years.
U.S. government stockpiled mercury is being stored
with no transport or processing presently being
conducted. The only potential releases are from the
storage. Such releases are expected to be minimal
because no handling occurs.
2.4 Miscellaneous Government Uses
Another U.S. government mercury reservoir is in the
Spallation Neutron Source research center; this DOE
facility uses elemental mercury as the target for the
neutron (SNS 2002). Because this is a new projected
use, reduction opportunities are not considered. The
facility is expected to begin operating in 2006.
19
-------
Chapter 3
Manufacturing Processes Involving Mercury
3.1 Chlor-Alkali Manufacturing
3.1.1 Introduction
Facilities producing chlor-alkali are classified under the
following business classifications:
SIC Code 2812: Alkalies and Chlorine
NAICS Code 325181: Alkalies and Chlorine
Manufacturing
More mercury is used in chlorine and caustic soda
manufacturing than in any other industrial sector in the
United States. The SIC Code 2812 and NAICS Code
325181 describe all industries primarily engaged in
manufacturing alkalies (e.g., NaOH) and chlorine (C12).
Chlorine and alkali manufacturing are linked because of
a shared production process. Electrolysis separates the
sodium and chlorine in salt brine (NaCl), producing 1.1
tons of caustic soda for every ton of chlorine. Since
chlorine cannot be economically stored or moved over
long distances, chlor-alkali facilities are often located
near industries that require chlorine. The two largest
industries for chlorine are vinyl chloride monomer
manufacturing and pulp and paper manufacturing (Kirk-
Othmer 1991).
There are three electrolytic methods used in chlor-alkali
production: diaphragm cell, mercury cell, and membrane
cell production. Although all new chlor-alkali facilities
being built use either membrane cell or diaphragm cell
technologies - processes that do not use mercury -
several chlor-alkali facilities still use the mercury cell
process. In the United States, mercury cell plants
account for 10% of the chlorine production capacity;
production takes place at 11 facilities using mercury
cells (Chlorine Institute 2001 and ChemExpo 2000).
One advantage of the mercury cell process is that it
produces a low-salt caustic soda and it is much easier to
scale production levels of chlorine and caustic soda
based upon demand (Genna 1998).
Unlike the diaphragm cell and membrane cell processes,
which are one-step processes, the mercury cell process is
a two-step process: an electrolyzing stage produces the
chlorine gas and a decomposing stage produces the
caustic soda (USEPA 1997a). Flowing at the bottom of
the cell, a few millimeters below the suspended metal
anode, the mercury acts as the cathode in the electrolytic
process. Each cell may contain three tons of mercury
(USEPA 1997a), and through most of the 1990s there
were a total of 762 mercury cells (Chlorine Institute
2001). An aqueous salt brine solution (NaCl) flows
between the anode and the cathode, releasing chlorine
gas at the anode. The remaining sodium and mercury
amalgam flows from the electrolyzer cell to the
decomposing cell, which separates the mercury from the
sodium that produces sodium hydroxide (NaOH) and
recycles the mercury back to the electrolyzer cell (Kirk-
Othmer 1991).
3.1.2 Materials Flow
Exhibit 3-1 illustrates the consumption, release, and
product content of mercury in chlor-alkali production.
The environmental release estimates in Exhibit 3-1 are
based on 1999 Toxic Release Inventory (TRI) mercury
release data for 13 of the 14 plants operating in that year.
Mercury use data are based on data from the Chlorine
Institute (2001); mercury in products is estimated as
described below.
The calculation does not match the amount consumed
with the total amount released. It is unknown why there
is a discrepancy between consumption and release.
Studies are being conducted by EPA to find out where
the missing mercury goes. One explanation could be that
the consumption is from the Chlorine Institute and the
release data is from TRI.
20
-------
Mercury Consumption
Exhibit 3-2 shows estimates of mercury consumption
for several years along with mercury release data. The
Chlorine Institute's (2001) estimate of mercury
consumption from the domestic chlor-alkali facilities
decreased from 222 tons in 1990 to 79 tons in 2000.
The Chlorine Institute also provided estimates of
mercury purchases by the chlor-alkali facilities. On a
year-to-year basis, mercury purchases did not
necessarily equal mercury use.
However, over a multi-year period, mercury purchases
were roughly equivalent to mercury use. Chlor-alkali
facilities may purchase more mercury than they
anticipate using, storing the excess mercury for later
use (Dungan 1999). The 2000 consumption estimate
is used in Exhibit 3-1.
Air Releases
Air releases of mercury from chlor-alkali production
result from elevated process temperatures. The heat
generated by the electrolysis process used to
Consumption:
79tons/yr
Chlor-Alkali
Production
Product:
<0.5 tons/yr
(caustic soda)
Releases: 27.8 tons/yr
- Air: 6.3 tons/yr
--Water: 0.073 tons/yr
- Disposal: 21.5 tons/yr
Recycling: 13.04 tons/yr (offsite)
Sources: Mercury Consumption: Chlorine Institute (2001).
Mercury Release and Recycling: 1999 TRI data.
Mercury in Product: Estimated from production capacity (Chlorine Institute 2001) and product
concentration (WLSSD 1997).
Exhibit 3-1. Mercury in Chlor-Alkali Manufacturing
Exhibit 3-2. Chlor-Alkali Mercury Cell Process Mercury Used, Emitted, Recycled, and Disposed
Quantity (tons)
Total Mercury Used1
Total Mercury Air Emissions 2
Total Mercury Water Emissions 2
Total Onsite Mercury Recycling 2
Total Offsite Mercury Recycling 2
Total Mercury Disposal2
Total Mercury in Caustic Soda3
1990 1991 1992 1993 1994 1995 1996 1997 1998
222 175 148 104 146 165 137 118 104
7.3
0.1
469
8.2
7.2
0.6
1999
88
6.3
0.07
374
13
21.5
0.6
2000
79
-
-
-
-
-
-
1 Source: Chlorine Institute (2001 ).
2 Source: 1999 TRI data for 13 of 14 chlor-alkali facilities using the mercury cell process. These thirteen facilities represented 96 percent of the total
production capacity, signifying that the reported releases are an excellent estimate of industry-wide releases.
3 Source: Estimated; represents high estimate of mercury likely to be in product. See text.
This table does not summarize environmental release data from the TRI prior to 1998.
21
-------
separate the chlorine from the salt brine contributes to
mercury volatilization (Johnson 1999). There are three
primary sources of mercury air emissions at a mercury
cell chlor-alkali facility: (1) byproduct hydrogen steam,
(2) end box ventilation air, and (3) cell room ventilation
air (USEPA 1997a). Ventilation systems and scrubbers
reduce the amount of mercury emitted to the atmosphere.
The mercury is transferred to water or to solid waste,
where it may be recycled or disposed. As shown in
Exhibit 3-1, industry-wide air releases are estimated at
6.3 tons based on 1999 TRI data.
Water Releases
Releases of mercury-containing water result from the
large quantities of water used in the electrolysis process.
Mercury is also found in the wastewater and brine of the
mercury cell process. Some mercury is found in the
water collected from the periodic wash-down of floors
and equipment. As shown in Exhibit 3-1, industry-wide
water releases are estimated as 0.07 tons based on 1999
TRI data. The estimate represents the mercury content
in waters discharged to a surface water or to a publicly
owned treatment works (POTW).
Solid Waste Releases and Recycling
Wastewater treatment sludges from chlor-alkali facilities
were routinely landfilled until 1992, when USEPA
banned the landfilling of certain mercury-containing
sludges (USEPA 1988b). Because of the restriction,
many mercury cell facilities now use retort and
hydrometallurgical processes to remove the mercury
from their wastes prior to landfilling and recycle the
recovered mercury back into the mercury cell process
(USEPA 1998a). As shown in Exhibit 3-2, a large
quantity of mercury is recycled onsite. (However, it is
not possible to identify how each facility reported this
quantity, and a portion of the quantity may represent
mercury that is continuously re-inserted back into the
process when using mercury as a catalyst.) This onsite
recycling is not accounted for as recycled quantities in
the summary of Exhibit 3-1 because it is internal to the
industry, rather than being sent to a commercial
recycling facility such as those discussed in Section 2.
As shown in Exhibit 3-1, industry-wide land disposal
releases are estimated as 21.5 tons based on 1999 TRI.
Product
Because mercury has a high vapor pressure at normal
operating conditions, mercury is found in trace amounts
in the reaction products (chlorine and caustic soda). No
estimates of mercury content in chlorine gas were found.
In 1987, a Wisconsin wastewatertreatment district found
that caustic soda (sodium hydroxide) can contain
mercury ranging from 10 to 300 parts per billion
(WLSSD 1997). The Chlorine Institute identified an
average level of 100 parts per billion, based on 1995
survey data (Chlorine Institute 2000).
Using conservative assumptions, the industry-wide
mercury content of caustic soda is estimated as no more
than 0.5 tons per year. These assumptions include using
the upper end of the mercury concentration range of 300
parts per billion, and estimating annual sodium
hydroxide production of 1.7 million tons per year (which
is equivalent to the capacities for mercury cell facilities
reported in Chlorine Institute (2001)). This estimate
assumes no mercury contributions from other processes.
Such contributions are possible at facilities where the
mercury cell process was replaced but where residual
mercury may still be present.
Reservoir
A considerable quantity of mercury is present inside a
chlor-alkali facility. This is partly due to the function of
mercury as a catalyst; as discussed above, each cell may
contain three tons of mercury (USEPA 1997a) and there
were 762 cells operating for most of the 1990s prior to
the most recent closures (Chlorine Institute 2001). An
industry source estimates that a single plant holds
between 75 and 750 tons of mercury, which would be
available to the secondary market upon dismantling of
the plant (Lawrence 2000). Additional mercury is also
expected to be present within pipes, equipment, etc., as
an amalgam, which may not be easily recoverable.
Based on these data, this report estimates that at least
2,000 tons of mercury is present at operating and
recently closed chlor-alkali production facilities.
3.1.3 Discussion
There is an apparent discrepancy between the mercury
consumed by the chlor-alkali industry and the mercury
emitted. Mercury consumed by the chlor-alkali industry
is used to replenish production losses. However,
mercury consumption is much larger than the reported
mercury emissions (Johnson 1999), and the mercury
contained in the product is not a significant fraction.
Therefore, approximately 50 tons of mercury appear to
be "missing" based on the 1999/2000 data. There is
increasing concern among state and federal regulators
regarding this "missing mercury" (Johnson 1999). Olin
Corporation, a major chlor-alkali producer, is working
with USEPA to eliminate mercury discharges from its
22
-------
two mercury cell chlor-alkali facilities (Johnson 1999).
The head of Olin Corporation is urging the chlor-alkali
industry to develop better methods to measure and
control fugitive mercury emissions (Johnson 1999).
Data for 2000 mercury consumption were provided by
the Chlorine Institute. USGS has not reported mercury
usage statistics since 1997. Historically, USGS and
Chlorine Institute data have differed. For example, the
quantity of mercury consumption in 1995 as provided by
the Chlorine Institute, 165 tons, is slightly lower than the
quantity of mercury consumption in 1995 provided by
USGS, 170 tons. This discrepancy is even more
apparent when comparing 1997 data (118 tons Chlorine
Institute vs. 176 tons USGS). The USGS data may
include extrapolations for non-respondents. The
Chlorine Institute is actively tracking mercury
consumption at the plants using the mercury cell process.
3.2 Lamp Manufacturing, Use, and Disposal
3.2.1 Introduction
Facilities manufacturing lamps and lighting equipment
may be classified under the following business
classification:
SIC Code 3641:
Tubes
Electric Lamp Bulbs and
NAICS Code 33511: Electric Lamp Bulb and
Part Manufacturing
SIC Code 3641 and NAICS Code 33511 are comprised
of establishments primarily engaged in manufacturing
electric bulbs, tubes, and related light sources. Mercury
is a key component of fluorescent lamps and high
intensity discharge (HID) lamps (including mercury
vapor, metal halide, and high pressure sodium lamps).
Fluorescent lamps are widely used for indoor lighting in
businesses and increasingly in residences, while HID
lamps are used for heat lamps, film projectors, dental
exams, photochemistry, water purification, and street
lighting. When an electrical current passes through
mercury vapor, it emits ultraviolet light. In a fluorescent
lamp, this ultraviolet light is converted into visible light
when it excites the phosphorus coating inside the tube,
causing it to fluoresce.
The mercury content in fluorescent bulbs in the United
States has steadily decreased during the past two
decades. In 1989, the average mercury content in a
fluorescent bulb was 48.2 mg (USEPA 1999a),
decreasing to 11.6 mg in 1999 for a typical four-foot
lamp (NEMA 2000). In 1995, Philips Lighting
introduced a low-mercury fluorescent lamp containing
only 4.4 mg of mercury (USEPA 1999a). OSRAM
Sylvania introduced a mercury-free high intensity
discharge (HID) lamp in 1998 (Sylvania 1998).
3.2.2 Materials Flow
Exhibit 3-3 illustrates the consumption, release, and
product content of mercury in electrical lighting,
spanning manufacturing, use, and final disposal.
3.2.3 Manufacture
Mercury use in lamps depends on the quantity of lamps
manufactured and the mercury content of the bulbs.
Philips Lighting estimates that low-mercury lamps
constitute 85% of its current lamp production and that
they have reduced their mercury use by 13 tons per year
(USEPA 1999a). Similar production information from
other manufacturers was not available. OSRAM
Sylvania estimates that introduction of their mercury-
free HID lamp should reduce mercury consumption by
0.17 tons per year (Sylvania 1998).
Mercury Consumption
As shown in Exhibit 3-4, mercury consumption by
domestic lighting manufacturers has declined from a
peak of 61 tons per year in 1992 to about 32 tons per
year in 1997, based on data from USGS. While these
data are useful for identifying trends, the USGS estimate
is not reflected in Exhibit 3-3. Instead, a lower estimate
of 16 tons based on data from the Bureau of Census and
the National Electrical Manufacturers Association
(NEMA) was used. The NEMA estimate was used
because it is based on more recent lamp composition
data and, due to uncertainties with the USGS data
identified in Section 3.1, the USGS data may
overestimate actual use.
The U.S. Department of Commerce's Bureau of Census
(USDOC 1995) estimates that 599 million fluorescent
lamps and 28.5 million HID lamps were produced in the
United States in 1994. Assuming an average mercury
content of 11.6 mg of mercury per fluorescent lamp
(NEMA 2000) and 25 mg per HID lamp (USEPA
1992b), lamp manufacturing consumed 16 tons of
mercury in 1994. The quantity was used as an estimate
for present day usage.
In 1994, the Census Bureau stopped collecting data on
lamp production. Based on National Electrical
23
-------
Manufacturers Association (NEMA) data, lighting
system sales increased from $7.8 billion to $8.4 billion
in 1997 (NEMA 1999), an increase of 8 percent.
Therefore, the 1994 Bureau of Census estimate of 599
million fluorescent lamps manufactured in the United
States appears to be a reasonable estimate for 1997.
Releases
Mercury can be released during transfer and parts
repair, mercury handling, mercury injection into the
lamps, accidents, and spills (USEPA 1997a). Two
lamp manufacturing companies submitted TRI reports
for 1999, reporting the release of 0.26 tons of mercury
to the
Consumption:
16tons/yr
Lighting
Manufacturing
Product:
16tons/yr
Exports:
2 tons/yr
Domestic Use:
14 tons/yr
Releases: 0.46 tons/yr
- Air: 0.26 tons/yr
• - Water: Not Available
- Disposal: 0.2 tons/yr
Recycling: 3.7 tons/yr (off-site)
Domestic:
1 4 tons/yr
^-
^-
Imports:
3 tons/vr
Use
(66-75 tons
in use)
17 tons/yr
(4 year lag)
Disposal
Releases: 14 tons/yr
- Air: 3 tons/yr
"- Disposal: 11 tons/yr
Recycling: 3 tons/yr
Sources:n Mercury Consumption: Extrapolated from the Bureau of Cencsus (DOC 1995), NEMA (2000), and EPA(1992b).
Mercury in Product: Extrapolated from lamps sold and exported (DOC 1995), lamps imported (EPA 1999b), and mercury content (NEMA 2000).
Mercury Release and Recycling: 1999 TRI data for manufacturing.
Exports and Imports: Bureau of Census (DOC 1995).
Emissions for Use: Recycling rate from NEMA (2000). Air and land disposal extrapolated using EPA(1997c).
Exhibit 3-3. Mercury in Electrical Lighting
Exhibit 3-4. Lighting Industry Mercury Consumed
1990 1991 1992 1993 1994 1995
1996 1997
Total Mercury Consumed1 (tons)
36
43
61
42
30
33
32
32
'Source: United States Geological Survey, Mineral Industry Surveys 1990-97
24
-------
air and 3.7 tons recycled. Releases from these facilities
are larger than the industry-wide estimate of 0.06 tons
in the Mercury Report to Congress (USEPA 1997a).
TRI data for these manufacturers are used in Exhibit 3-3.
This estimate may be low due to the small number of
facilities, but extrapolating to a larger population is
difficult due to a lack of facility-specific information.
Exports
An estimated 68 million fluorescent lamps and 4 million
HID lamps were exported in 1994 (USDOC 1995). This
is approximately 1.8 tons of mercury (using the mercury
content assumptions above). This number is used in
Exhibit 3-3.
3.2.4 Use
Mercury Consumption
Of the nearly 600 million fluorescent lamps
manufactured in the United States in 1994, 517 million
lamps were sold domestically; the remainder were
exported or stayed in inventory (USDOC 1995). An
additional 100 million fluorescent lamps containing an
estimated 2.5 tons of mercury were imported in 1995
(USEPA 1999b). Therefore, approximately 620 million
fluorescent lights were sold in the United States
containing 16 tons of mercury.
Of the 29 million HID lamps manufactured in 1994, 25
million were sold domestically; the remainder were
exported or remained in inventory (USDOC 1995). An
additional 3.5 million HID lamps containing an
estimated 0.1 tons of mercury were imported in 1995
(USEPA 1999b). Therefore, approximately 29 million
HID lamps (0.8 tons of mercury) were sold in the United
States in 1994. The total quantity of mercury consumed
from lighting (17 tons in Exhibit 3.3) reflects the
combination of fluorescent lamps (16 tons) and HID
lamps (1 ton).
Reservoir
Assuming a 20,000-hour lifespan for fluorescent lamps,
these lamps should last about four years. Assuming the
620 million lamps sold each year are replacing one-
fourth of the lamps in use, there were between 2.5 and 3
billion fluorescent bulbs in use in 1997, constituting 65
to 75 tons of mercury throughout the United States
(assuming 11.6 mg of mercury per lamp).
Because HID lamps typically have a usable life of
10,000 hours and most are used 24 hours per day,
USEPA (1992b) assumed that HID lamps are replaced
annually. Therefore, all 29 million lamps are
replacement lamps and they contained 0.8 tons of
mercury (assuming 25 mg of mercury per lamp).
3.2.5 Disposal
Since fluorescent lamps have a lifespan of about four
years, the quantity of mercury used in lamps today does
not reflect the quantity of mercury being disposed.
Instead, there is a four year lag from initial use to
disposal. The estimated 620 million fluorescent lights
purchased in 1994 probably entered the waste stream in
1997 - 1998. The 29 million HID lamps sold that year
probably entered in 1995. Together, they equal about 17
tons of mercury removed from service in 1997.
Until 1995, most fluorescent lights were disposed of as
municipal solid waste (MSW). USEPA (1992a)
estimated in 1992 that 82 percent of mercury-containing
lamps were landfilled, 16 percent were incinerated, and
2 percent were recycled. The number of companies
collecting lamps for recycling has increased since the
early 1990s to more than 60 companies. More recent
estimates by the Association of Lighting and Mercury
Recyclers state that the recycling rate in the late 1990s
was 15 percent (NEMA 2000). Assuming rates of 15
percent recycled, 67 percent landfilled, and 18 percent
incinerated (consistent with USEPA 1997c percentages
of wastes that are landfilled and incinerated), this results
in 11 tons that entered landfills, 3 tons incinerated, and
3 tons recycled.
The mercury lamp recycling rate is expected to continue
to increase due to changes in USEPA's universal waste
rule in June 1999. In this rule, USEPA streamlined
recycling requirements for mercury-containing
fluorescent, mercury vapor, sodium halide, and metal
halide lamps that exceed mercury concentrations set by
USEPA's Toxicity Characteristic Leaching Procedure
(TCLP) test. A goal of this rule is to encourage
recycling by making it easier for generators to collect,
store, and transport bulbs destined for recycling (USEPA
1999c).
3.2.6 Discussion
The quantity of mercury consumed for production was
assumed to equal the quantity estimated to be present in
domestically manufactured products (16 tons). This
estimate was used instead of the much greater Bureau of
Mines (USGS 1997) estimate for mercury consumption
of 32 tons in 1996. Therefore, this represents a source of
uncertainty because additional methods to verify either
25
-------
of these estimates are not available.
The quantity of mercury in lamps is expected to
decrease, but based on current research, elimination in
fluorescent lamps is not expected. As a result, future
releases of mercury will decrease slightly.
A second source of uncertainty is the extent to which
mercury in post-consumer lamps is currently recycled.
The recycling rates are expected to increase due to
regulatory changes such as the 1999 regulatory changes
by USEPA. Therefore, the quantities ultimately recycled
and disposed by commercial, industrial, and consumer
users are uncertain.
3.3 Thermometers and Other Instruments
3.3.1 Introduction
Facilities manufacturing thermometers and other
instruments may be classified under the following
business classifications:
SIC Code 38295: Commercial, geophysical,
meteorological, and general purpose
instruments. Applicable SIC (Product) Codes
are as follows :
Barometers:
20 - Barometers
Liquid in glass thermometers:
22 - Scientific thermometers
23 - Industrial thermometers (food, air
conditioning, and refrigeration)
24 - Household and commercial thermometer
34 - Medical thermometer
NAICS Code 339112: Surgical and medical
instrument manufacture.
NAICS Code 334519: Other measuring and
controlling device manufacturing.
Mercury is often used in medical and scientific
instruments because it is non-reactive, metallic, and
liquid over a relatively wide range of temperatures. The
most common use of mercury as a medical and scientific
instrument is in the liquid-in-glass thermometer.
Mercury is also used in instruments such as barometers
and other pressure-sensing devices. Liquid-in-glass
thermometers are commonly used for household,
industrial, clinical, and scientific purposes. The U.S.
Census Bureau provided estimates for each of these
classes of thermometers bought and sold in the United
States in 1997 (USDOC 1998). The Census Bureau did
not distinguish between mercury-filled thermometers and
those filled with other liquids, nor did they provide an
estimate for thermometer imports and exports.
Therefore, estimates for mercury use based on these
Census quantities are likely to overestimate actual
quantities of mercury consumed. USEPA (1997a)
expects mercury use and emissions from thermometers
to remain steady, with decreases resulting from digital
thermometers to be offset by increased demand for
thermometers by a growing population.
3.3.2 Materials Flow
Exhibit 3-5 illustrates the consumption, release, and
product content of mercury in thermometers and similar
instruments in manufacturing, use, and final disposal.
3.3.3 Manufacturing
Mercury Consumption
USEPA (1992b) identified that oral/rectal/baby
thermometers contained 0.61 grams mercury, and basal
thermometers contained 2.25 grams mercury. They also
estimated that 95 percent of clinical thermometers are
oral/rectal/baby thermometers and basal thermometers
comprised the remaining five percent. USEPA (1992b)
did not provide an estimate of mercury content for
scientific and industrial thermometers; therefore, the
mercury content of these instruments were assumed to be
equal to the quantity present in basal thermometers.
The U.S. Bureau of Census estimates that approximately
8.5 million medical and household thermometers (valued
at $12.2 million, or $1.44 each) and 0.58 million
industrial thermometers (valued at $10.2 million, or
$17.60 each) were bought and sold in the United States
in 1997 (USDOC 1998). The Bureau of Census did not
provide an estimate for scientific thermometer
production, but did provide an estimated value of $5.8
million. Assuming that each scientific thermometer
costs between $1.44 and $17.60 (derived from the other
thermometer types), an estimated 0.33 to 4.0 million
scientific thermometers were bought and sold in the
United States in 1997. The Bureau of Census also did
not specify whether these thermometers were
manufactured domestically or imported, although
USEPA (1992b) states that thermometer imports have
been increasing and assumes that exports are minimal.
26
-------
Consumption:
9-17 tons/yr
Thermometer
Manufacturing
Product:
9-17 tons/yr
Releases: <0.2 tons/yr
- Air: <0.2 tons/yr
- - Water: 0 tons/yr
- Disposal: 0 tons/yr
Recycling: 0 tons/yr
Domestic:
9-17 tons/yr
Imports:
Not Estimated
Use
(45-85 tons
in use)
9-17 tons/yr
(20 year lag)
Exports:
Not Estimated
Domestic Use:
9-17 tons/yr "
Disposal
Releases: 9-17 tons/yr
_-Air: 2-3 tons/yr
- Disposal: 7-14 tons/yr
Recycling: Not Available
Sources:n Mercury Consumption: Calculated from USDOC (1998) and USEPA (1992b).
Mercury Releases: Air releases during manufacturing from USEPA (1997a). Other releases from general waste management data from
USEPA (1997c).
Exhibit 3-5. Mercury in Thermometers
Exhibit 3-6. Mercury Used to Manufacture Thermometers in the U.S. in 1997
Thermometer Type
Quantity Manufactured1
Mercury Content per Total Mercury
Thermometer (tons)
(grams)2
Medical and household thermometer -
Basal
425,000
2.25
1.05
Medical and household thermometer -
Oral/rectal/baby
Industrial thermometers
Scientific thermometers
Total
8,100,000
583,000
330,000 to 4,000,000
8,300,000 to 11,900,000
0.61
2.25
2.25
-
5.45
1.45
0.74 to 9.0
8.7 to 17.0
1 U.S. Census (USDOC 1998), estimate for scientific thermometers is extrapolated from dollar value (see text).
2USEPA (1992b), mercury content for household thermometers is assumed to be same for oral/rectal/baby thermometers, mercury content for industrial and
basal thermometer is assumed to be same as basal thermometers.
27
-------
Product
For an upper-end estimate using Bureau of Census
(USDOC 1998) data of liquid-in-glass thermometers
bought and sold in the U.S. and USEPA (1992b) data for
mercury content, Exhibit 3-6 shows that about 9 to 17
tons of mercury were contained in thermometers
produced in the United States in 1997 (assuming no
imports and that all liquid-in-glass thermometers are
mercury-filled). This quantity is high because it assumes
that all liquid-filled thermometers contain mercury.
Releases
Mercury thermometers are produced by creating a
vacuum in the capillary glass tube to draw mercury into
the bulb and glass tube. USEPA (1997a) cites a 1973
USEPA estimate of 18 pounds of mercury emitted for
every ton of mercury used in instrument manufacture.
However, USEPA (1997a) warns that this estimate is
based on a survey of manufacturers during the 1960s and
may be an overestimate of actual emissions. Using the 9
to 17 ton consumption estimate above, approximately
160 to 300 pounds (0.08 to 0.15 tons) of mercury are
emitted to the air as a result of mercury thermometer
manufacturing. Thermometer manufacturers reported no
mercury releases to any media in the 1999 TRI.
3.3.4 Use
Using an estimated lifespan of 5 years (USEPA 1992b)
and an annual production rate of 9 to 17 tons per year (as
described in the manufacturing section), it is estimated
that 45 to 85 tons of mercury are currently in use in
thermometers in the United States. Because the mercury
is completely contained in the thermometer, release and
exposure to the mercury are unlikely under normal
operating conditions.
3.3.5 Disposal
USEPA (1997a) reports that there is little data regarding
mercury disposal. Most thermometers are discarded
when they are cracked or broken and enter the waste
stream from residential and clinical settings (USEPA
1997a). USEPA (1992b) estimates that five
percent of the glass thermometers are broken each year.
USEPA (1997a) cites a 1989 study that estimated that
16.3 tons of mercury were discarded in landfills from
thermometers. It can be assumed that the quantity of
mercury used in thermometer production (9 to 17 tons
for 1997) requires eventual recycling or disposal.
Assuming 80 percent is landfilled and 20 percent is
combusted (based on typical municipal waste
combustion rates), 7 to 14 tons are expected to be
disposed to land and 2-3 tons are expected to be emitted
to the air via combustion.
Increasing awareness regarding recycling of mercury
thermometers has lead to programs such as Fisher
Scientific's mercury thermometer trade-in program that
offers to reclaim a mercury thermometer for every non-
mercury thermometer ordered (Fisher Scientific 1999).
Because of these recycling programs, disposal estimates
may be high; there is no estimate available for the
amount of mercury recycled from thermometers.
3.3.6 Discussion
The quantity of mercury in thermometers was estimated
at 9 to 17 tons for 1997, based on Department of
Commerce data addressing domestic sales of
thermometers. Because all of the thermometers were
assumed to be mercury-filled, this was intended to
represent a high estimate for mercury consumption. The
only other estimate is USGS data. As shown in Exhibit
3-7, 26 tons of mercury were used in 1997 for
'measuring and control instruments,' which is intended
to include both mercury thermometers and thermostats
(Reese 1999).
No estimates for other values could be found, so the
remaining quantities on Exhibit 3-5 were calculated from
this consumption quantity. Also, mercury recycling
facilities are known to accept thermometers for
recycling, but quantities are not available. Therefore, the
quantities presented in Exhibit 3-5 as ultimately recycled
and disposed are uncertain.
Exhibit 3-7. Mercury Consumption by SIC Code 382 - Measuring and Control Instruments (tons)
1990 1991 1992 1993 1994 1995 1996 1997
Total Mercury Used1
119
99
72
58
47
45
26
'Source: USGS (1990-7)
28
-------
3.4 Thermostats
3.4.1 Introduction
Facilities manufacturing thermostats may be classified
under the following business classification:
SIC Code 3822: Controls for Monitoring Residential
and Commercial Environments and Appliance
Regulating Controls.
NAICS Code 334512: Automatic Environmental
Control Manufacturing for Residential, Commercial,
and Appliance Use.
A thermostat is a type of switch that turns on or off
depending on the temperature. Thermostats are used to
control the temperature in individual rooms, building
spaces, appliances, and refrigerators. Mercury switch
thermostats have been commonly used to control room
temperatures in commercial and residential spaces for
more than 50 years (USEPA 1994), although mercury-
free alternatives are available. Typically, a mercury
switch is mounted on a piece of bimetal. Bimetal is
composed of a strip or coil of two thin layers of
dissimilar metals that bend at different rates when heated
or cooled. As the bimetal bends with the temperature
change, a drop of mercury in a tube within the mercury
switch moves under force of gravity to either complete
or break an electrical circuit. Mercury thermostats have
proven to be an accurate, reliable, and inexpensive
means to control temperature (USEPA 1994).
3.4.2 Materials Flow
Exhibit 3-8 illustrates the consumption, release, and
product content of mercury in thermostats during
manufacturing, use, and final disposal.
3.4.3 Manufacturing
Mercury Consumption
Manufacturing of mercury switch thermostats consists of
filling a short glass tube with a bead of mercury and
sealing one end with wire contacts. There is little data
available on mercury consumption in the manufacturing
of thermostats. Using U.S. Bureau of Census data and
consultations with thermostat manufacturers, USEPA
(1994) estimates that 3 to 5 million mercury switch
thermostats were manufactured in 1994. USEPA
(1992b) estimates that each thermostat contains about 3
grams of mercury, therefore 11 to 17 tons of mercury are
used to produce thermostats annually.
The U.S. Census Bureau (USDOC 1998) estimates that
45 million thermostats were manufactured in the United
States in 1997, but some of these units may not contain
mercury. Exhibit 3-7 shows the USGS estimates for
domestic industrial consumption of mercury for SIC
Code 382, which includes thermostat and thermometer
manufacturing. Export data on mercury switch
thermostats were not available.
Review of the 1999 TRI data shows four facilities
involved in electronic component manufacturing (SIC
Code 3679) reporting mercury releases; it is not known
for certain whether these releases are a result of
thermostat manufacturing (as opposed to switches or
other products produced by the facility). Mercury may
be emitted during the manufacturing process from spills
and breakages, product testing, and product transfer
(USEPA 1997a). Total emissions from these three
facilities show negligible releases to air (0.002 tons), no
releases to water, and 3.9 tons of mercury recycled off-
site in 1999. This recycling quantity may be the result of
off-spec product, spill collection, etc.
These quantities may not reflect other companies
involved in thermostat production, and may be
overestimates by including releases resulting from
unrelated facility activities. Therefore, in Exhibit 3-8, it
is assumed that the quantity present in products (11-17
tons), plus the quantity recycled (4 tons) equaled
consumption (15-21 tons).
3.4.4 Use
Because mercury is contained in a sealed glass tube
within the mercury switch thermostat, release and
exposure to the mercury is unlikely under normal
operating conditions. USEPA (1994) estimates that 70
million mercury switch thermostats were used in U.S.
residences in 1994, which is associated with 230 tons of
mercury (assuming 3 grams per thermostat as explained
above). Since a mercury switch thermostat is a
mechanical device with few moving parts, its lifespan is
typically between 20 and 40 years, often exceeding that
of the room or building within which it is housed
(USEPA 1994). USEPA (1997a) cites a 1995 National
Electrical Manufacturing Association finding that
upgrading, remodeling, and building demolition are the
principal causes of mercury switch thermostat removal.
29
-------
Consumption:
15-21 tons/yr
Thermostat
Manufacturing
Product:
11—17 tons/yr
Exports:
2-3 tons/yr
Domestic Use:
9-14 tons/yr
Releases: 0 tons/yr
Recycling: 3.9 tons/yr (off-site)
Domestic:
9-1 4 tons/yr
Imports:
4—6 tons/vr
Use
(230 tons
in use)
7-10 tons/yr
(20 year lag)
Disposal
Releases: 7—10 tons/yr
_-Air: 1-2 tons/yr
" — Disposal: 6—8 tons/yr
Recycling: Not Available
Sources:^ Mercury Consumption: Sum of mercury in product and releases.
Mercury Release and Recycling: 1999 TRI for 3 facilities
Mercury in Product: Estimated from average mercury content (EPA 1992b) and number of thermostats produced (DOC
Exhibit 3-8. Mercury in Thermostats
Imports and exports may also affect the flow of mercury
in thermostats. Bureau of Census data (USDOC 1998)
indicate that the total value of thermostats produced was
$718 million in 1997, the quantity imported was $259
million (36 percent of domestic production) and exports
were $121 million (17 percent of domestic production).
These data include mercury and non-mercury devices.
Assuming an even distribution of mercury and non-
mercury devices and a constant annual production rate,
this indicates 11 to 17 tons of mercury are present in
domestically produced devices, 4 to 6 tons of mercury
are in imported products, and 2 to 3 tons are in exported
products. The net result is that 13 to 20 tons of mercury
annually enter the domestic consumer market in
thermostats.
3.4.5 Disposal
USEPA (1994) estimates that 2 to 3 million thermostats
were brought out of service in 1994. Assuming that all
of the disposed thermostats contained mercury at 3
grams per unit, this corresponds to 7 to 10 tons of
mercury per year. In the past, most thermostats have
been disposed of as municipal solid waste. Assuming
that 80 percent of solid wastes are landfilled and the
remaining is sent to municipal waste combustors, 80
percent of the mercury (6 to 8 tons) is landfilled and the
remainder emitted to the air.
Efforts to recycle mercury switch thermostats are
increasing; however, it is unknown what proportion of
the thermostat wastestream is being recycled. USEPA
(1999b) cites Thermostat Recycling Corporation as
recycling 120 pounds (0.06 tons) of mercury in the Great
Lakes region in 1998.
3.4.6 Discussion
Since mercury switch thermostats have such long lives,
they are expected to enter the waste stream for at least
30
-------
the next 30 to 40 years. USEPA (1992b) projects
programmable (non-mercury) thermostats to steadily
replace mercury switch thermostats, gaining an
additional one percent of the market share annually.
The quantity of mercury in thermostats estimated to
enter the consumer market (13 to 20 tons) is greater than
the quantity of mercury estimated to be in thermostats
removed from service (7 to 10 tons). This may represent
an inaccuracy in one or both of these estimates.
Alternatively, and perhaps more likely, it may be
indicative of the large lag time between generation and
disposal. For example, it may be the case that a larger
number of thermostats are being sold today than 20 to 40
years ago (i.e., the thermostats just now being removed
from service), or that new construction (rather than
replacement) comprises a significant percentage of the
new thermostat market and the number of thermostats in
buildings in the United States increases every year.
The USGS consumption data were not used for
consumption estimates because the data are combined
with other product categories (i.e., thermometers). The
estimate used here, however, is consistent with the
USGS estimate from Exhibit 3-7.
Estimates are not available addressing the quantity of
mercury in used thermostats sent for recycling.
Potentially, this is a significant data gap, because
numerous programs are in place to recycle mercury
containing thermostats. Identifying an accurate estimate
is difficult due to the varied methods by which the
thermostats may enter the recycling market, not all of
which are accountable. Therefore, the quantities
ultimately recycled and disposed by commercial,
industrial, and consumer use are uncertain.
3.5 Switches and Relays
3.5.1 Introduction
Facilities manufacturing switches and relays may be
classified under the following business classifications:
SIC Code 36251 66: Relays and Industrial Controls,
General Purpose and Other Relays, Reed Relays;
Mercury Wet Reed
SIC Code 36433 69: Wiring Devices and Supplies,
Current Carrying Wires, Switches for Electrical
Circuitry, All Other Switches: Appliance and
Fixture, Including Surface Mounted, Mercury, etc.
Mercury switches and relays are used in many household
and automotive applications. Mercury switches are
typically used to detect motion. A mercury switch
consists of a glass or ceramic tube with electrical
contacts at one end. When the tube is tilted or jolted, a
bead of mercury flows over the electrical contact and
completes the circuit. A mercury switch is often called
a "silent switch" because electrical contact is established
instantaneously due to the surface tension of the
mercury. In a hard contact switch, the microscopic
"bounce" that occurs as contact is established may cause
electrical noise (USEPA 1994).
Tilt switches are mercury switches that are used to sense
tilt. Mercury tilt switch applications include level
controls, security alarm systems, vending machine
alarms, washing machine covers, and automobile trunk
light switches. Mercury tilt switches are also used as
motion and vibration sensors in anti-theft devices,
"smart appliances" that turn off when not in use, and
automobile anti-lock brakes.
A relay is an electromechanical switch where the
variation of current in one electrical circuit controls the
current in another circuit. A relay consists of an
electromagnet that is connected to a moveable contact.
When the electromagnet is energized, the contact is
moved to either complete or break a circuit. In a
mercury reed relay, the electrical contacts are wetted
with mercury to provide an instantaneous circuit
(USEPA 1994).
3.5.2 Materials Flow
Exhibit 3-9 illustrates the consumption, release, and
product content of mercury in switches and relays, in
manufacturing, use, and final disposal.
3.5.3 Manufacturing
Mercury Consumption
USGS data in Exhibit 3-10 show that the total amount of
mercury used to produce wiring devices and switches
peaked in 1995 at 92 tons and dropped to 63 tons in
1997. The USGS estimate does not include mercury
reed relays because relays are classified under SIC Code
3625. The mercury content of various switches and
relays is shown in Exhibit 3-11.
Data from both the Department of Commerce's Bureau
of Census and USGS (1997) were used for estimating
mercury flow in this sector. The data are expressed as a
range: 36-63 tons per year.
31
-------
Consumption:
36-63 tons/yr
Switch and Relay
Manufacturing
Product:
36-63 tons/yr
Releases: 0 tons/yr
Recycling: 0 tons/yr
Domestic:
36-63 tons/yr
^
Imports:
Not Estimated
Use
(630 tons
in use)
36-63 tons/yr
"(10-15 year lag)"
Exports:
Not Estimated
Domestic Use:
36-63 tons/yr'
Disposal
Releases: 36-63 tons/yr
__- Air: 7-13 tons/yr
"-Disposal: 29-50 tons/yr
Recycling: Not Available
Sources: Mercury Consumption: Low end range from USDOC (1998) and M2P2 (1996). High end range from USGS (1997).
Mercury Releases: General waste management data from USEPA (1997c).
Exhibit 3-9. Mercury in Switches and Relays Manufacturing
The number of mercury switches manufactured in the
United States is uncertain. Mercury switches could be
included under various product codes within SIC code
36433 (Switches for Electrical Circuitry). Mercury
switches are specified in product sub-code 69 (All other
general use switches, including mercury). However, the
quantities and values in the 1998 Current Industrial
Report (USDOC 1998) combines product code 69 with
other AC-DC switches (product code 51) to protect
proprietary information. Moreover, mercury switches
may be found within other product codes such as
automotive switches and other special type of switches.
Mercury reed relays are classified under SIC code 3625,
however, specific production data were withheld in the
1997 Manufacturing Profiles report (USDOC 1998).
Assuming that all 16.5 million general use switches (SIC
Code 36433-69) bought and sold in the United States
(USDOC 1998) are mercury switches and each contains
2 grams of mercury results in approximately 36 tons of
mercury. This estimate could be high because SIC Code
36433-69 includes non-mercury switches, but it could
also be low because it does not include mercury reed
relays and may not include automotive and other
switches. In 1996, 11.2 tons of mercury was used in
U.S.-made vehicles, primarily as lighting switches (GLU
2001).
Releases
Mercury may be released during the manufacturing
process from spills and breakages, product testing, and
product transfer (USEPA 1997a). The wastes associated
32
-------
with mercury switch manufacturing are uncertain.
Mercury switch manufacturing consists of filling a glass
or ceramic tube with 0.5 to 3 grams of mercury and
sealing the end with electrical contacts. Although four
facilities within SIC Code 3679 (Electric component
manufacturing) reported mercury waste emissions to the
Toxics Release Inventory in 1999, it is uncertain whether
those releases result from manufacturing mercury
switches or relays. Total emissions from these three
facilities show negligible releases to air (0.0025 tons), no
releases to water, and 3.9 tons of mercury recycled off-
site in 1999. These estimates were previously accounted
for in thermostat manufacturing, and are not repeated
here. Applied to the industry as a whole, these quantities
are not necessarily representative of other switch and
relay manufacturers.
3.5.4 Use
Mercury switches are very reliable, and certain types of
mercury switches can last up to 50 years (USEPA
1992b). Because the mercury is contained in a sealed
glass or ceramic tube within the mercury switch, it is
unlikely that it will be released under normal conditions.
Because mercury switches are used in various
applications, from lighting switches to anti-lock brakes,
the number of switches currently in use is not easy to
determine. Using the USGS mercury consumption data
since 1990 (see Exhibit 3-10), and assuming that the
mercury contained in those switches is still in use, there
are at least 630 tons of mercury contained in switches in
the United States. This estimate is probably low because
of the long life span of mercury switches; most switches
manufactured in the 1970s and 1980s are probably still
in use. The amount of mercury imported into the U.S.
contained in imported mercury switches is also
unknown.
3.5.5 Disposal
USEPA (1992b) estimates that 1.9 tons of mercury are
discarded from mercury electric light switches each year,
assuming that 10 percent of the switches are disposed
after 10 years of production, 40 percent discarded after
30 years of production, and the remaining 50 percent
after 50 years. However, that estimate does not include
other mercury switches such as those found in household
appliances, automobiles, and mercury reed relays.
Exhibit 3-10 assumes the amount used in switch and
relay manufacturing (36-63 tons/year) must eventually
be disposed, with 80 percent landfilled and 20 percent
incinerated.
3.5.6 Discussion
Because mercury switches have such long life spans,
they are expected to steadily enter the waste stream for
at least the next 30 to 40 years. The automobile industry
is working to reduce mercury consumption (CGLI
1999). Mercury reed relays are gradually being replaced
by solid state relays (USEPA 1994).
Exhibit 3-10. Mercury Consumption by SIC Code 3643 - Wiring Device and Switches
1990
1991
1992
1993
1994
1995
1996
1997
Total Mercury Used1 (tons)
77
78
90
91
87
92
54
63
'Source: USGS (1990-7)
Description
Exhibit 3-11. Mercury Content of Various Mercury Switches and Relays
Mercury Content (mg)
Automobile trunk and hood light switch
500-1,000
Freezer light
2,000
Silent Switches
2,600
Mercury Reed Relay
140-3,000
Source: M2P2 1996
33
-------
The wide variety of mercury switches and their
applications in consumer and industrial products makes
accounting extremely difficult. This variety results in
different classifications of switches with some portion of
each containing mercury. The most significant example
is the classification of thermostats, which contain a
mercury switch but is categorized separately (and
discussed elsewhere in this report). Nevertheless, there
may be difficulties with data interpretation, especially in
cases where a manufacturer produces a wide variety of
mercury-containing products that contain switches, but
classifies its business activities according to a more
limited set of SIC codes. Such a problem was apparent
when interpreting TRI data for manufacturers of various
electrical devices and attributing the data to different
products (such as thermostats and switches).
The quantity of mercury used for switches is uncertain.
The quantity provided by USGS (1997) is 63 tons.
Bureau of Census data (36 tons) were also used to
account for switches that are likely to contain mercury.
Both values are included in Exhibit 3-10, as a range, to
account for this uncertainty.
3.6 Organic Chemical Production
3.6.1 Introduction
Facilities producing organic chemicals may be classified
under the following business classifications:
SIC Code 2869: Industrial Organic Chemicals
NAICS Code 325: Chemical Manufacturing
Mercury is used as a catalyst in the organic chemicals
industry. One known use is in the production of vinyl
chloride monomer using acetylene as a raw material. In
this process, acetylene (C2H2) is combined with
hydrogen chloride (HC1) and flows through a fixed bed
of solid mercuric chloride catalyst. The product is vinyl
chloride (C2H3C1), which is subsequently purified. This
process is used by a single facility, Borden Chemicals
and Plastics in Geismar, Louisiana. In 1996, this facility
had a capacity to produce 950 million pounds of vinyl
chloride per year, but by 1998 was expected to increase
this capacity by 250 million pounds per year (USEPA
2000a). As a result, the quantity of mercury used and
subsequently released is expected to increase. In 1999,
a total of three facilities (the Geismar facility was not
one of them) reported releases of mercury; however, the
releases were negligible (0.0005 tons).
3.6.2 Materials Flow
No estimates of mercury consumption data were
available for this industry, therefore neither consumption
nor release data can be presented due to insufficient data.
3.7 Dental Preparations
3.7.1 Introduction
Facilities manufacturing or using dental equipment may
be classified underthe following business classifications:
SIC Code 3843: Dental Equipment and Supplies
NAICS Code 339114: Dental Equipment and
Supplies Manufacturing
SIC Code 8021: Offices and Clinics of Dentists
NAICS Codes 6212 and 62121: Offices of Dentists
This section focuses on use of mercury by the dental
profession. Amalgam fillings, used to fill cavities in
teeth, contain about 50 percent mercury. Not all of the
mercury used by the dental profession ends up in the
fillings. Some is lost as air emissions, some is
discharged in wastewater, and some is disposed as
hazardous waste or is recycled.
3.7.2 Material Flows
Exhibit 3-12 illustrates the flow of mercury in the dental
profession.
Mercury Consumption
Mercury consumption is assumed equal to the amount of
mercury used in amalgam fillings.
Mercury is a major component (50 percent) in amalgam
fillings. Using data from USGS (1997), USEPA (1997a)
assumed that 34 tons of mercury were used in the dental
industry during 1996, including amounts found in
equipment and supplies. However, another approach
presented below results in a slightly higher estimate of
54 tons per year. To account for this uncertainty both
estimates are given in Exhibit 3-12. In 1990, about 96
million of the more than 200 million restorative
procedures that were performed used amalgam
(USDHHS 1993). Amalgam use decreased by 12.5
percent among dentists from 1990 to 1995, and since the
beginning of 1993 the trend has been steady (USDHHS
1997). Assuming that amalgam use continued to
steadily decline results in 81.6 million amalgam fillings
in 1996. According to a study by Yoshida (1994), an
34
-------
Consumption:
34-54 tons/yr
Dental
Preparations
Manufacturing
Releases: 0 tons/yr
- Air: 0 tons/yr
• -Water: 0tons/yr
-Disposal: 0tons/yr
Recycling: 2.5 tons/yr (off-site)
Use:
34-54 tons/yr
- Disposal: No estimate
Recycling: No estimate
Dental
Offices
k.
Releases: 8.1 tons/yr
- Air: 0.7 tons/yr
Population
(1200 tons
in use)
k.
Releases: 0.1 tons/yr
-Air: 0.1 tons/yr
- Disposal: 0 tons/yr
Recycling: Not
Applicable
Sources: Consumption and Use: USGS (1997) for low end, and Yoshida (1994) and USDHHS(1993, 1997) for high end.
Manufacturing Releases: 1999 TRI.
Dental Office and Popluation Releases: Air releases from EPA (1997a), water and population releases from DAMS (1999).
Exhibit 3-12. Mercury in Dental Preparations
amalgam filling contains 0.6 grams of mercury.
Therefore, almost 49,000 kilograms (54 tons) of mercury
were used in fillings during 1996.
Air Releases
Mercury in fillings can be released in various ways,
including emissions from spills and scrap, air discharged
by the dental office's vacuum pump system (Rubin
1996), and constant emissions from the fillings in
people's mouths over time. USEPA (1997a) assumed
that two percent (0.7 ton out of 34 tons) of the total
amount of mercury used is emitted from spills and scrap,
but admits that number is likely an underestimate of the
total emissions. This estimate is reflected in Exhibit 3-
12.
Studies have been conducted to determine the amount of
mercury that is released from fillings once they are
placed in people's mouths. As presented in USDHHS
(1993), a study by Mackert found that, on average, a
person's intake of mercury from fillings is 1.24
micrograms each day; results from other studies ranged
from 1.7 to 27.0 micrograms per day. Using the U.S.
Census Bureau estimate of 281 million people in the
U.S. in 2000 (USDOC 2001) and 1.24 micrograms of
daily release per person results in 0.35 kilograms (1
pound) of mercury released from fillings per year. This
estimate is not presented in Exhibit 3-12 because it is not
directed towards media releases to the environment but
rather direct exposure.
Water Releases
Wastewater from a dental office may contain, on
average, 270 milligrams per day (range 65 to 842)
(based on data from Arenholt [1996] in DAMS [1999]).
Using the mean daily level of 270 milligrams per day per
office times 250 working days per year times 100,000
dental offices (conservative estimate, DAMS 1999)
35
-------
results in 6,750 kilograms or 7.4 tons of mercury
entering wastewater each year.
Amalgam separators can reduce the mean mercury
content in a dental office's wastewater from 270 to 35
milligrams per day (based on data from Arenholt [1996]
in DAMS [1999]). However, few dental offices in the
United States have amalgam separators (DAMS 1999).
Solid Waste Releases
No data are available to estimate quantities of mercury
recycled or disposed.
Recycling
The quantity of mercury recycled by dental offices is
unknown. 1999 TRI data showed one dental equipment
manufacturing company reporting mercury emissions.
Offsite recycling was reported as 2.5 tons, while releases
to all other media were reported as zero.
Reservoir
The total quantity of mercury in the population is
estimated based on the annual use rate of 34 to 54 tons.
No data were available to estimate the lifetime of a
filling. Assuming a 20 to 40 year span, this results in an
estimated quantity of 1200 tons in use.
3.7.3 Discussion
The estimate used for the amount of mercury per filling
was based upon a Japanese study. The mercury content
of fillings in Japan may be higher than in the United
States, which may help account for the discrepancy
between this estimate of 54 tons of total mercury used
and USEPA's estimate of 34 tons. Additionally, this is
a single average value, where in reality the quantity used
is a function of many factors such as the patient's needs
and the technique of the dentist. Furthermore, non-
amalgam fillings are being used for certain applications.
3.8 Pharmaceutical Use
3.8.1 Introduction
Mercury finds its way into a variety of pharmaceutical
products, including opthalmics, vaccines, and topical
products. Although use of mercury in these products has
been scaled back in recent years both from voluntary
actions by manufacturers due to increasing concerns over
mercury toxicity and as the result of Food and Drug
Administration (FDA) regulations, mercury is still found
in many products.
To assess the presence of mercury in food and drugs, the
FDA issued a request for data to identify food and drug
products that contain intentionally introduced mercury
compounds (63 FR 68775, December 14, 1998). FDA's
analysis of these responses (USFDA 1999) indicate that
three mercury compounds are intentionally introduced as
apreservative into both prescription and overthe counter
(OTC) nonhomeopathic products such as nasal spray.
These preservatives are thimerosal (TM),
phenylmercuric acetate (PMA) and phenylmercuric
nitrate (PMN). The responses also showed that more
than twenty other mercury compounds are used in
homeopathic drug products, usually as therapeutic
ingredients.
3.8.2 Materials Flow
Exhibit 3-13 shows the consumption and release of
mercury pharmaceutical product manufacturing._
Mercury Consumption
USFDA (1999) calculated that approximately 75,000
grams (0.08 tons) of mercury compounds are used per
year. The FDA calculated this amount by tallying the
responses received from the request for data;
categorizing the responses by compound used and
product type; searching its databases for additional
products that fall into these product type categories that
were not reported in the responses to the request for
data; applying the same average amounts of mercury
compounds reported for that product type and category
to the additional products found in the databases; then
totaling the amounts of mercury compounds from each
category to reach an estimated total amount of mercury
compounds used in pharmaceutical products.
The 75,000 grams of mercury compounds estimated
comes exclusively from the three common preservatives:
TM, PMA, and PMN. While many homeopathic product
uses were reported, the FDA concluded that the dilutions
of mercury compounds in products were so low as to be
negligible in comparison to pharmaceutical use.
Thimerosal in products accounts for approximately 99%
of the mercury compounds included in the FDA's
estimate.
Releases
Releases may result from the manufacture or formulation
of the mercury compounds themselves. No
pharmaceutical manufacturers reported mercury releases
in 1999 to the TRI.
36
-------
Consumption:
0.08 tons/yr
Pharmaceuticals
Product:
0.08 tons/yr
Releases: unknown
Recycling/Reuse: unknown
Sou/ve: Mercury Consumption: FDA (1999).
Mercury Release and Product: Assumed equal to consumption.
Exhibit 3-13. Mercury in Pharmaceuticals
The potential release mechanisms for mercury in
pharmaceutical products include excretion, exhalation,
volatilization, spillage during administration, and the
destruction or disposal of unused products. Because
these products may be administered in any location,
especially in the case of OTC products, there is no way
to quantify the amounts that are spilled or discarded.
Many studies have been conducted examining the output
of mercury from the human body, but these are largely
dependent on dose, method of exposure, and the specific
mercury compound. Because mercury in
Pharmaceuticals can be introduced orally, nasally,
dermally, ocularly, or through injection, it is also
impossible to quantify the output of these compounds
once introduced to the human body.
3.8.3 Discussion
While mercury preservatives in pharmaceuticals were
reported to the FDA in a large array of products, their
use is dwindling due to consumer and regulatory
pressure. The few uses remaining are likely to be
discontinued due to the requirements of the New Drug
Approval process, which requires demonstration that a
product is safe and effective. The estimated total amount
of mercury compounds used annually, 0.08 tons,
indicates that pharmaceutical use is negligible in
comparison to other sources and uses of mercury.
3.9 Laboratory Use
3.9.1 Introduction
This section focuses on the use of mercury and mercury
compounds in laboratory chemicals. Mercury
compounds are used in laboratories in two ways: as
chemical reagents in experiments and processes and in
chemical products used for laboratory work. Mercury is
also found in many laboratory instruments, such as
thermometers and manometers, as discussed in other
sections of this report; this section focuses specifically
on non-equipment use.
Histology, the processing of body tissues for
examination, comprises several types of steps. These
steps include fixation and staining, both of which
frequently use mercury-bearing compounds. It is
important to note that these chemicals often contain
mercury in concentrations less than 1 percent, so the
mercury compound may not be listed on the product
Material Safety Data Sheet (MSDS). A certification of
analysis from the manufacturer will reveal the small
amounts of mercury in these products.
3.9.2 Materials Flow
Mercury Consumption
Because there are a wide variety of mercury compounds
used in laboratories, and these chemicals are made by
many different manufacturers, it is not possible to
determine the quantity of mercury annually being used
in laboratory settings. One source notes a decline from
35 tons of mercury compounds used in 1990 to 11 tons
of these compounds used in 1991 (NC DEHNR 1996).
It can be assumed that the current total usage of the
chemicals has continued to decline in the past nine years,
in light of the recent revisions of standard analytical
methods and growing concern over environmental
hazards.
Releases
Releases may result from the manufacture or formulation
37
-------
of the mercury compounds themselves. No
manufacturers of laboratory chemicals reported mercury
releases in 1999 to the TRI.
Mercury can to be released in two additional ways: as
unused product (e.g., expired or otherwise discarded
reagent), and as a result of use (e.g., in samples at dilute
concentrations). In general, laboratories prepare their
own guidelines regarding handling procedures for these
waste materials. Releases as solid waste and as water
discharges are expected to be most prevalent. Solid
wastes are expected to be the result of unused reagent
that is sent offsite for recycling or disposal as a
hazardous waste. Water releases would result from the
disposal of analyzed samples which contain small
concentrations of the reagent, which is rinsed down the
sink. Water releases may also result from the disposal of
unused reagent down the sink. The presence or absence
of local regulations or permitting requirements regarding
sewer discharges is expected to influence the laboratory
practices used.
3.9.3 Discussion
Because data estimating the use of mercury-containing
laboratory chemicals and equipment are not avaiable, it
is impossible to determine the contribution of this sector
to domestic mercury use and release. The only available
estimate of use, 11 tons of mercury-bearing chemicals
used in 1991 (NC DEHNR 1996), does not estimate the
amount of mercury in these chemicals; because most of
these chemicals contain only trace amounts of mercury
(less than one percent), it can be assumed that the
amount of mercury used and released from laboratory
chemicals is negligible in comparison to other sources
and uses of mercury.
3.10 Batteries
3.10.1 Introduction
Facilities manufacturing or storing batteries may be
classified under the following business classification
code:
SIC Code 3691: Storage battery manufacturing
SIC Code 3692: Primary battery manufacturing
NAICS Code 33591: Battery manufacturing
The use of mercury in electrical batteries has decreased
significantly from more than 1,000 tons annually in the
early 1980s to less than 1 ton in 1996 (USGS 2000e).
The use of mercury in battery production was sharply
reduced in the early 1990s. Mercury is presently used in
two types of batteries: button cell batteries and mercuric
oxide batteries. Button cell batteries are used in watches
and other consumer electronics. Mercuric oxide
batteries are larger cylindrical batteries used mostly for
non-consumer use items such as medical or military
applications (USEPA 1997a). The Mercury-Containing
and Rechargeable Battery Management Act of 1996, in
part, phased out the use of alkaline-manganese and zinc-
carbon batteries containing intentionally added mercury
and button cell mercuric-oxide batteries (USGS 2000e).
3.10.2 Materials Flow
At present, most batteries are expected to last no more
than a few years either as a result of use or slow
discharge over time. Therefore, little to no mercury is
expected to be present as part of consumer use of
batteries from applications prior to 1992. Furthermore,
such a quantity from past use cannot be estimated.
Releases of mercury to air from battery manufacturing
were estimated by USEPA (1997a). This estimate
showed negligible mercury emissions (<0.001 tons) in
1995. Examination of the 1999 TRI data showed one
battery manufacturer reporting mercury releases of
0.0125 tons. This facility corresponded to the only
domestic mercuric oxide battery manufacturer (Maine
2000).
3.10.3 Discussion
In conclusion, mercury is consumed in very small
amounts for battery production, relative to other sources.
Quantities of mercury used and subsequently released
are correspondingly small. For this reason, no exhibit
illustrating mercury flow is presented.
3.11 Miscellaneous
TRI data for 1999 identified several facilities reporting
releases of mercury that do not appear to be engaged in
the manufacturing processes described above. The
industries include electroplating (three facilities),
explosives (one facility), food (one facility),
transportation (one facility), j ewelry and precious metals
(one facility). The combined releases of all the industries
is 0.21 tons of mercury.
38
-------
Chapter 4
Incidental Mercury Use Associated With Coal Combustion or Coal Use
4.1 Coal Combustion by Utilities
4.1.1 Introduction
This section focuses on utilities that burn coal for
electric power generation. This sector is of concern
because electric utilities are the largest source of
anthropogenic air emissions of mercury in the United
States (USEPA 1997a). Facilities may fall under several
SIC and NAICS Codes.
SIC codes:
491: Electric Services
4911: Electric Services
493: Combination Electric and Gas, and Other
Utility
4931: Electric and Other Services Combined
4939: Combination Utilities, Not Elsewhere
Classified
NAICS codes:
22: Utilities
221: Utilities
2211: Electric Power Generation, Transmission and
Distribution
22111: Electric Power Generation
221112: Fossil Fuel Electric Power Generation
Utility boilers that generate electricity can be fired by
coal, oil, natural gas, or some combination of these fuels.
This section focuses on utilities that use coal. Coal is
burned in a boiler to heat water and produce steam. The
steam is used to generate electricity and, in some cases,
heat. There are approximately 440 coal-fired utility
plants in the United States, and they use about 1250
boilers (USEPA 1999d). Coal-fired utilities tend to be
concentrated in the Northeast and Midwest.
In 1994, 81 percent of the energy generated from utility
boilers came from the burning of coal (USEPA 1997a).
Coal accounts for over three-quarters of electricity
generation in some areas including the Great Lakes area
(USEPA 1988a). For example, Ohio generates roughly
90 percent of its electricity from coal (GLNWF 1997).
Coal consumption is expected to increase 26 percent
between 1997 and 2020 as utilities use more of their
generation capacity, costs of natural gas and oil rise and
nuclear plants close (USEPA 1999d).
Mercury is present in the mined coal. After mining, the
coal may be cleaned to remove sulfur and improve
burning characteristics. It is then transported by rail to
end users such as utilities. Coal is stored in storage piles
or silos at the plant. From storage, the coal is subjected
to mechanical sizing operations and is charged to the
boiler. There are three basic types of boilers: pulverized,
cyclone, or stoker systems. Most (92%) coal-fired
boilers are pulverized coal systems where the coal is
pulverized before combustion. Cyclone systems, named
because of the cyclone-like vortex created by the coal
particles in the furnace during combustion, make up 8
percent of utility boilers. A third, less common type
(<1%) is the "stoker," which is used for smaller
capacities (e.g., 20-30 megawatts) and burns coal in a
variety of sizes (USEPA 1999d; USEPA 1988a).
In the removal of sulfur during coal cleaning, some
portion of the mercury is coincidentally removed as well.
The most widely used methods of coal cleaning use
specific gravity, relying on the principle that heavier
particles (i.e., impurities) separate from lighter ones (i.e.,
coal) when settling in liquid. A common method for
cleaning coarse pieces is to pulse currents of water
through abed of coal in a jig so impurities like shale and
pyrite sink. A mixture of water and ground magnetite is
used to clean coarse and medium-sized pieces. A
concentrating table, an inclined vibrating platform with
diagonal grooves that trap the impurities, is also used to
clean intermediate sized pieces. Fine coal particles are
often cleaned with froth flotation. The coal pieces are
coated with oil and then agitated in a controlled mixture
of air, water, and reagents until froth is formed on the
surface. Bubbles tend to attach to the coal, keeping it
buoyant, while heavier particles remain dispersed in the
water (USEPA 1988a). On average, coal cleaning
removes about 21 percent of the mercury contained in
39
-------
coal (USEPA 1998c). Seventy percent of coal used by
electric utilities is cleaned to some extent (USEPA
1999d). Putting these two estimates together, 15 percent
of the mercury present in coal nationwide is removed
prior to introduction to the boiler.
Mercury is atrace element (i.e., contaminant) in coal and
is a highly volatile metal that vaporizes at the
temperatures reached when coal is burned.
Consequently, mercury is emitted in the gas stream
during combustion. The concentration of mercury in the
coal varies considerably depending on the coal type,
where it was mined, and how it is processed before
combustion (Massachusetts 1996). There are four types
of coal: anthracite, bituminous, subbituminous, and
lignite. Ninety-one percent of the coal burned by coal-
fired utilities in the United States in 1997 was
bituminous and subbituminous, 9 percent was lignite,
and less than one percent was anthracite (USEPA
1999d). The different types of coal have varying
mercury content.
To estimate the quantity of mercury present in coal used
by coal-fired utility boilers, data from the EPA's 1999
Information Collection Request (ICR) were used, which
measured coal samples from every U.S. coal-fired power
plant as well as coal usage data. Using these data, a
consumption 89 tons mercury was calculated (after
processing and cleaning) from 925 million tons of coal.
In the 1999 ICR, EPA identified the mercury content of
coal at all 450 coal-fired utilities used in that year, and
also conducted air sampling at a subset of these units.
Based on analysis of the data, EPA estimated that 48
tons of mercury was released to the air (USEPA 2001).
(The remainder is assumed to be collected in air
pollution control residues and handled as solid waste.)
The combustion of coal results in the vaporization of
much of the contained mercury and its release to the
atmosphere, from which it is ultimately deposited in soil
or into bodies of water (USBM 1994). The part of the
mercury in coal that is not emitted to the atmosphere
during combustion is trapped in wastes such as bottom
ash and recoverable fly ash. Landfills are often the
ultimate repositories for these wastes (USBM 1994).
The ash can also be used in products such as concrete.
4.1.2 Materials Flow
Exhibit 4-1 demonstrates the flow of mercury during the
coal combustion process.
Mercury Consumption
Based on evaluation of EPA's 1999 Information
Collection Request, an estimated 89 tons of mercury
entered boilers in cleaned coal. As noted above, about
15% of mercury is removed from coal during cleaning
nationwide. Therefore, the quantity of mercury leaving
the mine is estimated as 105 tons, which is reflected in
Exhibit 4-1.
Consumption
105tons/yr
Coal
Cleaning
Consumption
89 tons/vr
Utility Coal
Combustion
Releases: 16 tons/yr
- Air: 0 tons/yr
•-Water: 0 tons/yr
- Disposal: 16 tons/yr
Recycling/Reuse: 0 tons/yr
Releases: 72 tons/yr
- Air: 48 tons/yr
--Water: 7 tons/yr
- Disposal: 17 tons/yr
Recycling/Reuse: 0 tons/yr
Souice: Mercury Consumption: USEPA (2001) to estimate boiler input.
Mercury Release: Air releases from EPA (2001). Water releases calculated using EPA (1988a). Disposal releases from coal cleaning
based on cleaning efficiency from EPA(1997a). Disposal and product releases from combustion: Based on EPA(1999d) and EPA (1988a).
Exhibit 4-1. Mercury in Utility Coal Combustion
40
-------
Air Releases
In 1999, the nationwide rate of mercury emissions from
coal-fired utility boilers was estimated to be 48 tons per
year (USEPA 2001). This estimate is reflected in
Exhibit 4-1.
Although small quantities of mercury may be emitted as
fugitive particulate matter from coal storage and
handling, the primary source of mercury from both coal
and combustion in utility boilers is the combustion stack.
Because the combustion zone in boilers operates at
temperatures above 1100 °C (2000 °F), mercury in the
coal is vaporized and exhausted as a gas. Some of the
gas may cool and condense as it passes through the
boiler and air pollution control
device. Mercury is released in both elemental and
compound form (believed to be mercuric chloride), with
significant variation due to site-specific factors.
Additional discussion regarding speciation is presented
in Section 7 of this report.
The primary types of control devices include
electrostatic precipitators (ESPs), wet scrubbers, fabric
filters or baghouses, and mechanical collectors. ESPs
are the most widely used control device by the electric
utility industry. None are specifically designed to
remove mercury, but all have some effect. Extensive
efficiency data were collected by EPA in 1999 (USEPA
2001). These efficiency data are presented in Exhibit 4-
2.
Exhibit 4-2. Efficiencies of Various Control Devices in Removing Mercury from Coal-fired Boiler Flue Gas
Control Device
Fabric filter
Cold-side Electrostatic precipitator (ESP)
Hot-side ESP
Fabric filter followed by wet desulfurization
(scrubber)
Cold-side ESP followed by wet desulfurization
(scrubber)
Hot-side ESP followed by wet desulfurization
(scrubber)
Median Mercury Removal Efficiency (%)
Bituminous Coals
89
29
11
97
78
39
Subbituminous Coals
73
3
0
No data
16
8
Source: USEPA 2001
Water Releases
Releases of mercury to water come from three main
sources: runoff from coal piles, wastewater from coal
cleaning, and maintenance and cleaning wastes (e.g.,
boiler blowdown, cooling tower blowdown,
demineralizer reagents, boiler cleaning wastes, and
liquors from flue gas desulfurization (FGD)). Each of
these sources may contain mercury. Waters such as
these are typically sent to settling basins prior to
discharge, where mercury may either be present in the
settled solids or be discharged with the effluent. As
shown below, only one source, cooling tower blowdown,
contained significant quantities of mercury or had
sufficient data to assess its contribution.
Cooling tower blowdown is waste removed periodically
from recirculating cooling tower systems to maintain
water quality. The average production is 2.6 billion
gallons per year per plant, and from the limited data
available, the concentration of mercury was measured as
1.5 micrograms per liter (USEPA 1988a). This results in
7 tons of mercury contained in cooling tower blowdown
each year industry-wide (i.e., generated by 440 plants).
This estimate is reflected in Exhibit 4-1.
Demineralizer regenerants are wastes resulting from the
periodic cleaning and regeneration of ion exchange beds
used to remove mineral salts from boiler makeup water.
The average plant production is 5 million gallons per
year, with a mercury concentration of 0.05 micrograms
per liter (USEPA 1998c). This results in less than 1
pound of mercury contained in these wastes each year.
41
-------
"Water-side" boiler cleaning wastes result from the
periodic cleaning of the boiler tubes, the superheater, and
the condenser. The average plant production is 180,000
gallons per year for water-side boiler cleaning. No
mercury data were available for water-side alkaline
cleaning wastes, but the mercury concentration for
water-side hydrochloric acid cleaning wastes ranges
from 0.0 to 0.002 milligrams per liter (USEPA 1998c).
Assuming acid is used 100 percent of the time and using
the high limit of mercury results in less than 2 pounds of
mercury in this waste each year.
Coal pile runoff is produced by precipitation falling on
coal storage areas. A typical coal storage pile is 25-40
feet high and can cover an area up to 75 acres. Most
utilities keep a supply on hand of at least 90 days' worth
of coal, which equals about 600-1800 cubic meters per
megawatt of generating capacity (USEPA 1999d). The
mercury content in runoff ranges from 0.0002 - 0.007
mg/L, and the average runoff is 20 inches per year per
plant (USEPA 1988a). No estimate for the quantity of
mercury from this source can be made.
Boiler blowdown is waste continuously or intermittently
removed from boilers that recirculate water to maintain
water quality. The average plant production is 11
million gallons per year (USEPA 1998c). No mercury
concentration data were available.
"Gas-side" or "fire-side" wastes are produced during
maintenance of the gas-side of the boiler, which includes
the air preheater, economizer, superheater, stack and
ancillary equipment. The residues are normally removed
with water only. The average plant production is
700,000 gallons per year for gas-side boiler cleaning, but
no data on mercury content in those wastes were
available (USEPA 1988a).
FGD sludge is the waste produced during the process of
removing sulfur oxide gases from the flue gas and is
discussed in more detail later. Wet systems use aqueous
solutions to remove the sulfur oxides from the flue gas.
A portion of FGD waste is wet FGD sludge liquors. In
this waste stream, mercury ranges from 0.00006 to 0.1
mg/L, with a median concentration of 0.005 mg/L
(USEPA 1988a). No data on generation quantities were
available.
Solid Waste Releases
Wastes from the coal combustion process go primarily to
landfills and surface impoundments. There are two main
solid waste streams from coal combustion: ash and flue
gas desulfurization (FGD) sludge.
In addition, coal cleaning may generate solid and/or
aqueous wastes containing mercury. No hard data on
amounts of wastewater or solid wastes generated by coal
cleaning facilities were available, but assuming that 15
percent of mercury in coal is removed in this process
(discussed above) results in 16 tons of mercury in coal
cleaning wastes. As identified above, mercury is likely
associated with the solids, however, an unknown portion
of the mercury may be present in water as suspended
solids or dissolved mercury. This estimate is reflected in
Exhibit 4-1.
Ash is the noncombusted waste material that remains
after coal is burned. Ash may be collected from the flue
gas (fly ash and FGD sludge), or remain in the boiler
(bottom ash and boiler slag).
Fly ash is small, uncombusted material carried out of the
boiler with the flue gases. In mechanical hopper fly ash,
mercury content ranges from 0.008 to 3.00 mg/kg of
coal, with a median of 0.073 mg/kg, and in fine fly ash
mercury content ranges from 0.005 to 2.50 mg/kg with
a median concentration of 0.10 mg/kg (USEPA 1988a).
There were 60.26 million tons of fly ash produced in
1997 (USEPA 1999d). Using the median concentration
of 0.10 mg/kg, an estimated 6.0 tons/year of mercury is
present in fly ash.
FGD sludge is the waste produced from the removal of
sulfur oxide gases from the flue gas. Wet systems use
aqueous solutions to remove the sulfur oxides from the
flue gas. Dry FGD systems use no water for sulfur oxide
removal, although dry FGD wastes may be mixed with
water before disposal (USEPA 1988a). Fly ash is the
primary source of most of the trace elements found in
scrubber sludge (as shown in Exhibit 4-2, scrubbers do
not have particularly high mercury removal efficiencie s).
In wet scrubbers that also serve as fly ash collection
devices, more than 50 percent of the sludge solids may
be ash (USEPA 1988a). In wet FGD sludge solids,
mercury ranges from 0.01 to 6.0 mg/kg, with a median
concentration of 0.4 mg/kg (USEPA 1988a) (dry FGD
sludge solids are assumed to have similar
concentrations). Utility boilers produced 25.16 million
tons of FGD wastes in 1997 and are expected to produce
50 million tons of sludge in 2000 (USEPA 1999d).
Using the median concentration of 0.4 mg/kg and the
1997 waste generation quantity, an estimated 10.1
42
-------
tons/year of mercury is present in fly ash.
Bottom ash and boiler slag are uncombusted material
that does not completely melt and settles on the bottom
of the boiler. In both materials, mercury content ranges
from 0.005 to 4.2 mg/kg of coal, with a median of 0.023
mg/kg(USEPA 1988a). There were 16.9 million tons of
bottom ash and 2.7 million tons of boiler slag produced
in 1997 (USEPA 1999d). Using the median
concentration of 0.023 mg/kg, an estimated 0.45
tons/year of mercury is present in bottom ash and boiler
slag.
A sum of the above sources (fly ash, FGD sludge,
bottom ash, and boiler slag) results in an estimated 16.6
tons per year of mercury. This estimate is shown in
Exhibit 4-1.
Product
In 1997,26.8 percent of all waste generated at coal-fired
electric utility power plants was reused (e.g., as cement
additives, high volume road construction material,
wallboard, flowable fill, and blasting grit). The reused
quantity included 19.3 million tons of fly ash (31.5% of
fly ash generated), 2.18 million tons (7.9%) of FGD
wastes, 5.10 million tons of bottom ash (27.7%), and
2.58 million tons of boiler slag (92.9%). Using the
mercury concentrations discussed previously, an
estimated 3.0 tons/year of mercury were contained in the
reused material. This quantity is already included in
Exhibit 4-1 as solid waste. No quantity is listed in
Exhibit 4-1 for recycled, because mercury is not
recovered from any coal combustion waste.
4.1.3 Discussion
The total annual quantity of mercury in coal processed in
utility boilers is estimated to be 89 tons. The annual
quantity of mercury released or in products is estimated
to be 48 tons to air, 17 tons to solid waste and products
and 7 tons to water; therefore, 89 tons per year are
assumed to enter the process and 72 tons per year leaves
the process. This discrepancy is due to the different
sources used in compiling these estimates. In these
calculations, air emissions do not assume all the mercury
present in the fuel is emitted in stack gas.
The following list mentions some of the actions that
have been taken to address the problem of mercury in
coal and resulting environmental release:
• Coal cleaning reduces the amount of ash produced,
thereby reducing the amount of mercury released.
On average, coal cleaning removes about 21 percent
of the mercury contained in coal (USEPA 1997a).
Seventy percent of the coal used by electric utilities
is cleaned to some extent (USEPA 1999d).
Control devices have reduced mercury air emissions.
The effectiveness of current control devices at
removing mercury was discussed earlier in this
section.
• Many collaborative efforts on trace element research
focused on mercury have been conducted by
Department Of Energy/Federal Energy Technology
Center (USDOE/FETC) and USEPA, the U.S.
Geological Survey (USGS), the Electric Power
Research Institute (EPRI), the utility industry, other
governmental agencies at both the federal and state
levels, and other U.S. and foreign research
organizations. Other groups such as the Small
Business Innovative Research Program (SBIR),
University Coal Research Program (UCR), and the
Jointly Sponsored Research Program at the
University of North Dakota Energy and
Environmental Research Center have focus areas in
research and development for the coal-fired utility
industry.
EPA's Information Collection Request data
represent a comprehensive analysis of mercury both
entering and being emitted from boilers.
4.2 Lime Manufacturing
4.2.1 Introduction
Lime is produced from the calcination of limestone.
Limestone is present throughout the United States and
comprises primarily calcium and magnesium in a
carbonate form. The limestone is fed to a rotary kiln
where it is heated and rotated slowly to ensure mixing.
This drives off carbon dioxide (and other volatile species
such as water). The product, quicklime, is discharged
from the opposite end of the kiln from which the
limestone is introduced. Most domestic kilns use coal as
a heat source, although the kiln can be adapted to oil or
natural gas. The coal is combusted separately and the
offgases travel through the kiln, to avoid mixing coal ash
with the lime product (Kirk-Othmer 1995). Because
both the limestone and the coal are heated to
temperatures well above the volatilization point of
mercury, it is expected that any mercury initially in the
raw materials is discharged to the air.
43
-------
Consumption:
2.7-5.0 tons/yr
Lime Production
Product:
Estimate not available
Releases: 0.2 tons/yr
-Air: 0.1 tons/yr
- - Water: 0 tons/yr
- Disposal: 0.1 tons/yr
Recycling: 0 tons/yr
Source:n Mercury Consumption: Estimated from coal and limestone use and concentration data from Kirk-Othmer
(1995), USGS(2002b), USGS (1998), and CIBO (1997).
Mercury Release: Air releases from EPA Mercury Study Report to Congress (1997a). Disposal releases estimated
using utility coal combustion information.
Mercury in Product: Estimate not available.
Exhibit 4-3. Mercury in Lime Production
4.2.2 Materials Flow
Exhibit 4-3 illustrates the consumption, release, and
product content of mercury in lime production. No lime
manufacturing facilities reported mercury releases to the
1999 TRI.
Mercury Consumption
The concentration of mercury in limestone is estimated
as 0.04 mg/kg (Council of Industrial Boiler Owners
1997). The quantity of limestone used as raw material is
estimated to be about 2 tons per ton produced (Kirk-
Othmer 1995), and the quantity of lime produced in
2000 is 21.6 million tons (USGS 2002b). This
corresponds to approximately 1.7 tons of mercury in the
limestone feed. The concentration of mercury is
expected to be variable, but no data demonstrating this
variability are available.
The other source of mercury in the feed results from the
coal fuel. The energy consumption of lime production
is variable depending on the efficiencies of the kiln.
Energy consumption is estimated as 5.5 to 8 million Btu
per ton lime (Kirk-Othmer 1995), or approximately 124
to 181 trillion Btu in 1999. The concentration of
mercury in coal is estimated as 4.8 to 36.4 pounds per
trillion Btus for 14 different coal types (USGS 1998).
Accounting for the two orThe quantity is assumed to be
zero.
Solid Waste Releases
Mercury is potentially present in coal combustion wastes
generated from the burning of fossil fuels in an onsite
boiler. Estimates of coal combustion wastes generated
from lime production are not available. However, a very
rough estimate can be obtained by using the results of
the assessment of utility coal combustion (even though
characteristics regarding particulate control and burner
technology may be different). In Section 4.1.3, air
releases of 48 tons and solid waste releases of 17 tons
were estimated. Applying this proportion to the 0.1 tons
of mercury released to air, no more than 0.1 ton of
mercury is estimated to be in the waste residues. This
estimate is reflected in Exhibit 4-3.
Product
Approximately 21.6 million tons of lime were produced
in 2000 (USGS 2002b). No data regarding the mercury
content in lime are available.
4.2.3 Discussion
The quantity of mercury entering the lime production
process results from limestone and coal. The total is
estimated to be 2.8 to 5.1 tons per year. However, the
quantity of mercury leaving the process is only estimated
to be 0.1 ton from air and 1 ton from solids. This
discrepancy can be due to the following factors:
• Poor estimates of mercury input. The quantity of
mercury present in the feed limestone is based on a
single concentration value of limestone, from a
single location. The mercury content of limestone is
expected to vary by location throughout the United
States. Additionally, the energy use in kilns is
obtained from a single source (Kirk-Othmer 1995),
and although this is useful for a 'ballpark' estimate,
a second data source would be required to help
ensure representativeness.
44
-------
Poor estimates of mercury output. The only estimate
of mercury releases or mercury in product is an air
estimate from theMercury Study Report to Congress
(USEPA 1997a), and an estimate of mercury present
in coal combustion ash. Other estimates of mercury
in water releases and in the product are not
available, although the quantity of mercury in water
is likely to be zero. However, even small
concentrations of mercury in the product would
re suit in sizable accounting (e.g., if the concentration
of mercury in the product equaled the concentration
of mercury in the limestone feed, this would account
for an additional 0.9 tons). In addition, quantities of
mercury in coal combustion ash as a solid waste
were only roughly estimated.
4.3 Residential, Commercial, and Industrial
Coal Combustion
4.3.1 Introduction
This section focuses on residential, commercial and
industrial boilers that burn coal to produce steam. This
sector is of concern because mercury in the coal is
vaporized during combustion and appears as a trace
contaminant in the gas exhaust stream. Facilities will
fall under many SIC or NAICS Codes. There is no
category specifically for industries that use coal-fired
boilers, and residential sources do not fall under the
purview of SIC or NAICS.
While boilers can be fired by coal, oil, natural gas, or a
combination, this section focuses on the use of coal.
Coal is burned in a boiler to heat water and produce
steam. The steam is used to generate heat or electricity
or as a production process input.
In 2000, residential, commercial, and industrial facilities
(excluding power producing utilities/non-utilities and
coke production) consumed approximately 69 million
tons of coal (USDOE 2002).
There are a wide range of boiler sizes and types used in
the commercial and industrial sector. Larger boilers use
a suspension-fired system similar to the systems in place
at coal-fired utilities. Moderate and small boilers use
grate-fired systems (USEPA 1997a). Residential boilers
tend to be small, stoker systems.
Mercury is a contaminant in coal, the raw material used
for combustion. The mercury content in coal can range
from 4.8 to 36.4 pounds per trillion Btus (USGS 1998).
In USEPA's Mercury Study Report to Congress, the
emission factor is determined by coal type. For
bituminous coal it is assumed that 16 pounds of mercury
per trillion Btus is emitted; for anthracite coal, 18 pounds
per trillion Btus (USEPA 1997a); estimates which are in
the range of the mercury content of coal from the USGS
data showing that much of the mercury was assumed to
be emitted to the air.
4.3.2 Materials Flow
Exhibit 4-4 demonstrates the flow of mercury during the
coal combustion process for industrial, commercial, and
residential boilers.
Mercury Consumption
Mercury is a contaminant in coal, the raw material used
for combustion. As identified above, approximately 69
million tons of coal for commercial, industrial, and
residential applications are used annually. In Section
4.1, it was estimated that 925 million tons of coal
contained 105 tons of mercury (about 0.11 tons mercury
per million tons coal). Assuming that the mercury
content of coal burned in each industry is similar,
approximately 7.6 tons of mercury is contained in
incoming coal for residential, commercial, and industrial
boilers.
An alternative calculation was presented in USEPA
(1997a). The same energy consumption of 2.8
quadrillion Btus was used but a different mercury
concentration in coal was assumed: for bituminous coal
it was assumed that 16 pounds of mercury per trillion
Btus are present, and 18 pounds per trillion Btus for
anthracite coal. This resulted in an estimate of 21.2 tons
of mercury per year. Exhibit 4-4 contains both numbers
as a range.
Air Releases
In USEPA (1997a), it was assumed that all mercury
present in the raw material would be released to the air.
Mercury control practices, including the purchase of
washed coal and the control of emissions, were not
assumed to occur. As a result, this source estimated that
mercury releases totaled 21.2 tons annually,
corresponding to 20.7 tons per year for
commercial/industrial boilers and 0.5 tons per year for
residential boilers (USEPA 1997a; USEPA 1993a). It is
similar to an estimate of 23.6 tons in USEPA (1997b);
calculated using the same emission factors for
bituminous and anthracite coal, but slightly different
energy consumption data. The air release estimate
shown in Exhibit 4-4 presents both estimates.
45
-------
Consumption:
7.6-21.2 tons/yr
Residential, Commercial,
and Industrial Coal
Combustion
Product:
0 tons/yr
Releases: 21.8-24.2 tons/yr
-Air: 21.2-23.6 tons/yr
->• -Water: 0 tons/yr
-Disposal: 0.6 tons/yr
Recycling/Reuse: 0 tons/yr
Source:D Mercury Consumption: Estimated using USEPA (2001) and USDOE (2002) for low
end and USEPA (1997a) for high end.
Mercury Release: Air releases estimated using USEPA (1997a) and USEPA
(1997b). Disposal releases estimated using USEPA (1999d) and USEPA (1988a).
Exhibit 4-4. Mercury in Residential, Commercial, and Industrial Coal
Combustion
Water Releases
No water release estimate is available.
Solid Waste Releases
Mercury may be present in ash that is generated from the
combustion process and is subsequently landfilled. The
ash may also be used as a product; the estimate presented
here includes ash managed using both methods. Ash
generation for these boilers includes bottom ash and fly
ash. The quantity and composition of each are a
function of the boiler technology as well as the specific
coal used (e.g., coal with high ash content generates
larger quantities of ash). Furthermore, the degree to
which air pollution control devices are used is extremely
variable.
Control devices used include mechanical (e.g., cyclone),
fabric filter, and electrostatic precipitators. Facilities
may not use any control devices at all or may use
devices with low collection efficiency (USEPA 1999d).
The type of air pollution control equipment in place
affects whether fly ash will be generated at all, as well as
its characteristics.
Through the consideration of these factors on a plant-
specific basis, an annual ash generation rate from
commercial and industrial non-utilities was presented in
USEPA (1999d) using data from the 1990 National
Interim Emission Inventory Database (USEPA 1990).
This estimate is 5.8 x 106 tons ash per year, for all types
of ash. As shown above, much less coal is used in
residential applications and its contribution to ash
generation is ignored. Data regarding the mercury
content of ash from non-utilities are not available.
Therefore, data from utility coal combustion wastes were
used: median concentrations of mercury in fly ash are
about 0.10 mg/kg and 0.02 mg/kg in bottom ash
(USEPA 1988a). Using the fly ash concentration as a
conservative value, about 0.6 tons of mercury are present
in land disposed wastes.
Product
Mercury content of byproducts (e.g., recycled ash) are
included in the "disposal" quantities.
4.3.3 Discussion
The total annual quantity of mercury in the raw material
is estimated to be 9.3 tons. The annual quantity of
mercury released or in products is estimated to be 21.2
tons to air, 0.6 tons to solid waste and products, and 0
tons to water. Therefore, about 9 tons per year are
assumed to enter the process and 22 tons per year are
assumed to leave the process. This discrepancy is due to
the different sources used in compiling these estimates.
There is a wide range in the quantity of mercury present
in the raw material. USEPA (1997a) used median values
in estimating this quantity, resulting in an estimate
within the range used here. This estimate was
subsequently used as the basis for air emissions. An
independent source for air emissions data is not
46
-------
available; therefore, it is very difficult to better identify
air releases. The above estimates show that almost all
the mercury present in the coal burned is emitted in the
stack gas.
Several other industrial sectors discussed in this report
use coal combustion as an onsite source for energy or
steam (this differs from coke production where coal is a
raw material). As discussed in USEPA (1999d),
industries accounting for a significant portion of coal-
fired non-utility generating capacity include pulp and
paper, primary metals, food products, and chemical
production. Therefore, there is potential for 'double
counting' in cases where coal combustion industries are
specifically discussed in this report.
4.4 Byproduct Coke Production
4.4.1 Introduction
Facilities producing coke byproducts may be classified
under the following business classifications:
SIC Code 3312: Steel Works, Blast Furnaces
(including Coke Ovens), and Rolling Mills (except
coke ovens not integrated with steel mills).
NAICS Code 324199: Coke oven products (e.g.,
coke, gases, tars) made in coke oven establishments
not integrated with steel mills.
NAICS Code 333111: Coke oven products made in
steel mills.
Coke has been used in iron and steel production for over
100 years. As a byproduct of certain types of coal
combustion, coke is composed of nearly pure carbon.
Coke revolutionized iron and steel production in the
1870s because it burned at much higher temperatures
than coal (Keller 1997). Coke is also used in other
metallurgical applications, for ferrous and nonferrous
metal production, forming, and recycling activities.
Coke may be produced at large integrated steel mills that
use coke for blast furnace operation. It may also be
produced by independent facilities who subsequently sell
the product in a wide variety of markets. Coke is
produced by burning coal in an oxygen-poor
environment at temperatures in excess of 2,200°F (Buss
1999), releasing the noncombustible contents within the
coal as gases (Keller 1997). In 1991, there were 19
byproduct coke producers in the United States
(Huskannen 1991 cited in USEPA 1997a). Most coke in
the United States is produced in slot oven byproduct
batteries (USEPA 1997a). The slot oven coke battery
consists of a series of narrow ovens with heating flues
between each oven pair. Pulverized coal is fed into each
oven and combusted for 12 to 20 hours, burning off
nearly all volatile matter and forming coke. The coke is
then unloaded into a rail car where it is cooled by a water
rinse (USEPA 1997a).
4.4.2 Materials Flow
Exhibit 4-5 illustrates the flow of mercury in coke
production.
Mercury Consumption
Kirk-Othmer (1993) estimates that 27 million tons of
coke were produced in the United States for the steel
industry in 1990. The quantity of coal used as raw
material for 2000 is estimated as 29 million tons
(USDOE 2002). In Section 4.3.2, the mercury content
of coal was estimated as about 0.11 tons of mercury per
million tons of coal (using USEPA 2001). This
nationwide weighted average is most appropriate for fuel
coal and may not be applicable to coal used for coke
production. Nevertheless, using these data results in an
estimate of 3.2 tons mercury in the incoming coal per
year, which is reflected in Exhibit 4-5.
Product
There were no data regarding the mercury content in
product coke. Additionally, there were no data
regarding mercury emissions in the iron and steel
manufacturing process where coke is used. It is assumed
that nearly all of the mercury is volatilized from the coke
during the coke production process.
Air Release
There are no reported mercury emissions from byproduct
coke plants in the Toxics Release Inventory. However,
mercury is probably present in the volatilized gases
released during the coking operation (USEPA 1997a).
Mercury may also be emitted through door leaks and
from the stacks. Sang and Lourie (1995) report that 306
kg (0.7 tons) of mercury are released from coke-making
operations in the Great Lakes Basin. Using emissions
factors from European coke plants, USEPA (1997a) also
estimates that potential emissions from domestic coke
plants are about 0.7 tons per year. This estimate is used
in Exhibit 4-5.
47
-------
Consumption:
3.2 tons/yr
Coke Production
Product:
Estimate not available
Releases: 2.2 tons/yr
- Air: 0.7 tons/yr
• - Water: 0 tons/yr
- Disposal: 1.53 tons/yr
Recycling/Reuse: 0 tons/yr
Source: Mercury Consumption: Estimated from coal use (USDOE 2002) and concentration data (USEPA 2001).
Mercury Release and Recycling: Air releases estimated using EPA Mercury Study Report to Congress (1997a).
Disposal releases from 1999 TRI.
Mercury in Product: Estimate not available.
Exhibit 4-5. Mercury in Coke Production
Water Release
Mercury may be found in the rinse water. However no
estimate of this quantity is available.
Solid Waste
Evaluation of 1999 TRI data showed three steel
production companies reporting mercury emissions.
These releases were assumed to be the result of coke
production. A total of 1.53 tons were disposed in
landfills; zero releases were reported to other media.
This estimate is shown in Exhibit 4-5.
4.4.3 Discussion
Due to the severity of the process conditions, the coking
process should volatilize nearly all of the mercury within
the coal. There is, however, a large apparent
discrepancy between the estimate of mercury within the
coal entering the coking facility (4.9 tons), and the
estimate of mercury leaving the facility as emissions (1.8
tons). This discrepancy may be the result of differences
in the data sources used.
4.5 Portland Cement Manufacturing
4.5.1 Introduction
Facilities manufacturing portland cement may be
classified under the following business classifications:
SIC Code 3241: Hydraulic Cement Manufacture
NAICS Code 32731: Cement Manufacturing
USEPA's Report to Congress on Cement Kiln Dust
(1993b) states that United States clinker production in
1990 was 65.1 million tons from 115 plants, representing
a production capacity of 76 million tons per year. This
is somewhat lower than the data presented in the
Mercury Study Report to Congress (81 million tons of
capacity at 212 plants for 1990).
Mercury emissions are a byproduct of Portland cement
manufacturing because the raw materials and fuel
contain small amounts of mercury. Portland cement is
manufactured using a mixture of gypsum, limestone, and
silica. After the rock is quarried, a series of crushers
reduce it to an appropriate size to be used as cement kiln
feed. The final rock size is approximately three inches
or smaller. The raw material is then processed through
either the "wet" or the "dry" process, depending on the
cement manufacturing facility. In the "wet" process, the
raw material is mixed with water during the grinding
step to form a slurry, and is then fed to the kiln as a
liquid. In the "dry" process, rather than mixing the raw
materials with water, the raw materials are dried to
reduce the moisture content, then fed to the kiln. The
remainder of the cement production process is essentially
the same.
The wet or dry material is pyroprocessed at abouthi,700"
F in a rotary kiln fed with powdered coal, oil or gas.
Some gases are released during this process, and the raw
material is transformed into clinker, hard gray nodules
about the size and shape of marbles. The clinker is then
48
-------
cooled, and the heated air from the coolers is returned to
kilns to save fuels and increase burning efficiency
(Portland Cement Association 1999).
4.5.2 Materials Flow
Exhibit 4-6 presents the flow of mercury through the
cement manufacturing process.
Mercury Consumption
Approximately 74.6 million tons of limestone were
consumed in cement production in 1990 (USEPA
1993b). This represents the most significant raw
material by weight, accounting for 85 percent of kiln
inputs (coal or other fuels were not included). The
mercury content of limestone is between 0.02 - 2.3 ppm
(USBM 1994), with a second source (CIBO 1997)
estimating a value of 0.04 ppm. Using the CIBO value
gives an estimated mercury input of 3 tons of mercury
nationwide. The use of the USBM data would result in
an extremely wide range of mercury.
Additional mercury is present in fuels combusted onsite
for heating the kilns. These fuels include coal, oil,
natural gas, and hazardous wastes (e.g., organic
solvents). However, in 1990 coal was the dominant fuel
used, accounting for 71 percent of the total heating value
(147xl012 Btu) (USEPA 1993b). The mercury content
in coal ranges from 4.8 to 36.4 pounds per 1012 Btu
(USGS 1998), accounting for about 0.35 to 2.7 tons of
mercury per year for this process.
Adding the contributions from coal and limestone, the
two most significant contributors, gives an estimated
mercury loading of 3.4 to 5.7 tons per year.
Consumption:
3.4-5.7tons/yr
Portland
Cement
Production
Product:
Estimate not available
Releases: 4.8 tons/yr
-Air: 4.2 tons/yr
- - Water: 0 tons/yr
- Disposal: 0.6 tons/yr
Recycling: 0 tons/yr
Mercury Consumption: Estimated from coal use and concentration data (USGS 1998; USEPA 1993b).
Mercury Release and Recycling: Air releases estimated using EPA Mercury Study Report to Congress (1997a). Disposal
releases estimated using Cement Kiln Dust Report to Congress (USEPA 1993b).
Mercury in Product: Estimate not available.
Exhibit 4-6. Mercury in Cement Manufacturing
Product
No data regarding the mercury content of the final
cement product is available. Because mercury
evaporates at approximately 660*F while the kilns
operate at 2700«F, most of the mercury present in the
raw materials probably volatilizes during production.
Air Release
Most of the mercury emitted during cement production
comes from the kiln and preheating/precalcining steps.
Minor sources of mercury emissions may include
particulate matter (PM) from raw material processing
and emissions from fuel combustion. The mercury
emission rate for the entire cement production process
was estimated to be 1.3xlO"4 pounds of mercury per ton
of clinker (USEPA 1997a). With 65.1 million tons of
clinker produced in 1990, this results in approximately
4.2 tons mercury emitted to the air each year.
Solid Waste Release
Particulate emissions are controlled during the
pyroprocessing steps by fabric filters and ESPs. The
resultant material from dust collection is cement kiln
dust, a material that can be reused onsite (i.e., in the
cement production process) or disposed. A total of 14.2
million tons of this material was generated in 1990, of
which 5 million tons was not recycled to the system.
Based on the analysis of 17 samples collected by
USEPA, a mercury concentration range of 0.005 to 14.4
mg/kg (median of 0.11 mg/kg) was determined. Using
49
-------
the median value and the quantity of 5 million tons
cement kiln dust that was not recycled, a mercury
loading of 0.6 tons of mercury is estimated (range 0.03
to 72 tons). There are no additional data on the specific
ability of these systems to capture mercury emissions
from cement kilns, so the quality of this estimate cannot
be compared with other data sources.
4.5.3 Discussion
There are no mercury TRI data for Portland cement
manufacturing facilities for 1999.
The estimates presented here are based on data from
1990. An uncertainty is the trends of the domestic
production of cement, as well as trends in raw material
use. As shown above, coal contributes a significant
amount of mercury to the raw material input of kilns.
Most of the estimates vary widely, mostly due to the
large quantities of materials involved. The result is that
the quantity of a given raw material used or waste
generated is not likely to have as much error as the
composition of that material with regard to mercury.
Limited data with wide ranges were used in
characterizing the composition of these materials. A
more accurate approach would be to consider a plant-by-
plant analysis, considering the composition of mercury
in the raw materials and emissions together with the site-
specific generation rate. However, such data (or
immediate plans to obtain it) are not available.
4.6 Coal Combustion Waste Products
4.6.1 Introduction
Large quantities of coal combustion wastes are used or
sold for applications other than disposal. Wastes (such
as ash) generated from utilities, industries, and
commercial applications can be re-used. In all cases, any
mercury in the ash is present as a contaminant which
may potentially be released to the environment during or
following use.
In 1997, the following quantities of utility coal
combustion wastes were reused: 19.3 million tons of fly
ash (31.5% of all fly ash generated), 5.1 million tons of
bottom ash (27.7% of all bottom ash generated), 2.6
million tons of boiler slag (92.9% of all boiler slag
generated), and 2.2 million tons of flue gas
desulfurization (FGD) wastes (7.9% of all FGD wastes
generated). Although similar quantitative data are not
available for non-utility sources of ash, available
information indicates that the wastes are used in similar
or identical applications. The largest uses of utility coal
combustion wastes are the following (in decreasing order
of quantity used in 1997): cement and concrete;
structural fill; waste stabilization; road base; blasting
grit; mining applications; wallboard; snow and ice
control; mineral filler; flowable fill; and agriculture
(USEPA 1999d).
4.6.2 Materials Flow
Using the compositional data presented in Section 4.1,
the quantity of mercury in product uses is estimated to be
3.0 tons per year (median), with a range up to 103 tons
per year (the upper end of the range assumes that all
wastes exhibit their highest concentrations).
By reviewing the above list of product applications,
many of the uses involve direct placement on the land
where the material is not expected to be moved. For
example, in agricultural uses the waste is applied directly
to soil, and in structural fill or road base applications the
material is used as a base for further construction that is
expected to last for many years. In waste stabilization,
the ash becomes part of the solid waste matrix which is
subsequently landfilled.
Other uses, including use as wallboard, blasting grit, and
mineral filler, do not include immediate placement on
the land. Instead, any mercury in the coal combustion
waste would be incorporated into the commercial or
consumer product and then eventually landfilled. Using
the compositional data presented in Section 4.1, the
quantity of mercury in these three uses is estimated to be
0.73 tons (median), with a range up to 22 tons (the upper
end of the range assumes that all wastes exhibit their
highest concentrations). Therefore, this quantity of
mercury is used in commerce, then probably disposed in
a landfill.
Very little information is available discussing the fate of
contaminants, including mercury, in product
applications. For example, it is not known if the
mercury migrates from its land-based applications to air,
stormwater runoff, or other media.
4.6.3 Discussion
Section 4.1 of this report identified that solid coal
combustion wastes, such as fly ash and bottom ash, are
either disposed of or are used as products. The estimates
from Section 4.1 are intended to present all management
methods for coal combustion wastes, and therefore the
estimates presented in this section necessarily duplicate
50
-------
those identified in Section 4.1. This section is intended
to highlight specific management methods of coal
combustion wastes. The same uncertainties regarding
the quality of the estimates in Section 4.1 are applicable
here. Chief among these concerns is the variability of
waste composition and the small amount of data that
exists regarding mercury in wastes. Methods being
considered for reducing mercury in stack air emissions
include capturing the mercury in fly ash. The use of
such controls are expected to increase in the future, with
a corresponding increase in the quantity of mercury
present in the generated solid wastes. However, the
magnitude of such changes cannot be predicted.
As discussed in this section of the report, many of these
applications involve placement of the material on the
land so that, regardless of whether the waste is disposed
or used as a product, the mercury present in the waste is
placed on the land. Once on the land, however, there is
no information regarding its environmental fate as
staying in the ash matrix, entering the air, or entering the
water.
51
-------
Chapter 5
Incidental Mercury Use Associated With Non-Coal Sources
5.1 Oil Combustion
5.1.1 Introduction
Oil combustion is used by utilities to generate electricity
and is used by industrial and commercial (non-utility)
facilities to generate steam, electricity, or heat for
miscellaneous industrial applications (USEPA 1999d).
Additionally, there are residential applications of oil
combustion as a heating fuel. Fuel usage and
technologies differ between each of these three sectors,
and therefore the use and releases of mercury differ in
each as well.
In the utility sector, the total amount of electricity
generated from oil combustion is small relative to the
total generation of electricity by coal and other
technologies. Oil is used at a much larger number of
industrial and non-utility facilities, and accounts for a
larger capacity as well: oil-fired utilities have a capacity
of 43,000 MW, and oil-fired non-utilities have a capacity
of 54,000 MW, based on 1994 data (USEPA 1999d).
Utilities predominantly use residual (No. 6) fuel oil,
while lighter and more expensive distillate (No. 2) fuel
oil is used for auxiliary or start-up purposes. Residual
oil has a higher ash content than distillate oil, leading to
increased levels of combustion bottom ash and air
pollution control fly ash. However, the majority of oil-
fired utility power plants do not use air pollution or
particulate control equipment, and therefore do not
collect fly ash. This is because the ash content of oil
(even residual oil) is much lower than coal and their
emissions characteristics may not require the addition of
air pollution control equipment (USEPA 1999d). The
predominant fuel used in non-utility and residential
applications is distillate oil. These units are smallerthan
combustion units in the utility sector. Therefore, the
quantity of ash generation in these sectors are expected
to be even less than in the utility sector. Nonetheless,
mercury releases still occur.
5.1.2 Materials Flow
Exhibit 5-1 illustrates the flow of mercury in oil
combustion in utility, non-utility, and residential
combustion units.
Mercury Consumption
Data are available quantifying the amount of fuel oil
used in utility, non-utility, and residential applications.
For the utility and non-utility sectors, the quantity of fuel
oil used in 1996 is as follows (USEPA 1999d):
• Utility, residual oil: 3,900 million gallons
• Utility, distillate oil: 684 million gallons
• Non-utility, residual oil: 3,100 million gallons
• Non-utility distillate oil: 5,500 million gallons
For the residential sector, the quantity of oil is not
directly available. However, the heating content of oil
used in the non-utility and residential sectors is reported
as 2,180 and 880 trillion Btu, respectively, in 1994.
Assuming that the heating value of oil used in these two
sectors is similar, and that only distillate oil is used for
residential applications, the quantity of oil estimated to
be used in residential applications is 3,500 million
gallons of distillate oil.
The mercury content of these fuels is expected to be
variable. Minnesota (1999) provides estimates of the
mercury content of product oils, reporting that residual
oil has 0.004 ppm mercury and distillate oil has 0.001
ppm mercury. Using the concentration data together
with the above volume data provides the following
estimates for the mercury content of raw materials in
1996:
• Utility: 0.06 tons of mercury (corresponding to 0.06
tons from residual oil and 0.003 tons from distillate
oil)
• Non-utility: 0.07 tons of mercury (corresponding to
0.05 tons from residual oil and 0.02 tons from
distillate oil)
52
-------
Consumption:
0.06 tons/yr
U.S.
Utility Oil
Combustion
Product:
0 tons/yr
Releases: <0.75 tons/yr
- Air: 0.2 tons/yr
• - Water: 0 tons/yr
- Disposal: <0.55 tons/yr
Recycling: 0 tons/yr (offsite)
Consumption:
0.07 tons/yr
U.S.
Non-Utility
Oil Combustion
Product:
0 tons/yr
Releases: 5.1-7.8 tons/yr
-Air: 5.0-7.7tons/yr
- - Water: 0 tons/yr
- Disposal: <0.13 tons/yr
Recycling: 0 tons/yr (offsite)
Consumption:
0.01 tons/yr
U.S.
Residential
Oil Combustion
Product:
0 tons/yr
Releases: 2.8-3.2 tons/yr
-Air: 2.8-3.2 tons/yr
- - Water: 0 tons/yr
- Disposal: 0 tons/yr
Recycling: 0 tons/yr (offsite)
Source/a Mercury Consumption: Composition data from Minnesota (1999). Use data from USEPA (1999d).
Mercury Release and Recycling: Disposal releases estimated from USEPA (1999d). Air releases
estimated from USEPA (1997a) and USEPA (1997b).
Exhibit 5-1. Mercury in Utility, Non-Utility, and Residential Oil Combustion
53
-------
Residential: 0.01 tons of mercury (corresponding to
0.01 tons from distillate oil).
Product
There are no products. In cases where ash is recycled,
these estimates are incorporated as solid waste releases.
Air Releases
The estimates of mercury releases from utility, non-
utility, and residential combustion of oil vary depending
on the mercury emission factors that were used in the
calculations. These estimates are as follows:
Utility: 0.2 tons/year
Non-utility: 5.0-7.7 tons
Residential: 2.8 - 3.2 tons
For utility boilers, mercury emissions were estimated
using emissions data available from 58 emission tests
conducted by USEPA, the Electric Power Research
Institute (EPRI), the Department of Energy (USDOE),
and individual utilities. Boiler-specific emission
estimates were then calculated by multiplying the
calculated inlet mercury concentration by the appropriate
emission factor for each boiler configuration and control
device.
For non-utility and residential boilers, the ranges account
for the different mercury emission factors for oil that
were used in the Mercury Study Report to Congress
(USEPA 1997a) and in Locating and Estimating Air
Emissions from Sources of Mercury and Mercury
Compounds (USEPA 1997b). The following factors
were used in the Report to Congress:
Residual Oil (No. 6): 2.9 kg/1015 J
Distillate Oil (No. 2): 3.0 kg/1015 J
The following emission factors were used in the
Locating and Estimating Air Emissions document:
Residual Oil (No. 6): 2.7 kg/1015 J
Distillate Oil (No. 2): 0.02 kg/1015 J
The mercury emission factors for residual oil and
distillate oil were multiplied by oil consumption
estimates in order to estimate the amount of mercury
released to air.
Solid Waste Releases
Air pollution control devices are most frequently used in
the utility sector. An estimated quantity of 23,000 tons
of oil combustion waste were collected in 1995. Air
pollution control equipment is less frequently used in the
non-utility sector, and therefore fewer oil combustion
solid wastes are produced. An estimated quantity of
5,500 tons of oil combustion waste are collected
annually (USEPA 1999d). Solid waste releases in the
residential market is assumed to be negligible. This
sector operates smaller boilers than the utility sector, and
is more likely to use lower ash distillate oil. As a result
of these factors, this sector is the least likely to employ
air pollution control devices which generate solid waste.
The mercury content in these wastes is variable,
depending on their type and other facility-specific
factors. The overall range for mercury is approximately
0.06 ppm to 24 ppm (USEPA 1999d). Using the upper
end of this range in conjunction with the waste quantities
listed above, the mercury loadings from these solid
wastes are as follows:
• Utility: <0.55 ton
• Non-utility: <0.13 ton
• Residential: Negligible
Water Releases
No water releases are expected. Water is used for air
pollution control, but solids in the water which may
contain mercury are expected to settle prior to discharge
or other release to the environment.
5.1.3 Discussion
In all cases the quantity of mercury assumed to be
released is less than the quantity of mercury assumed to
be present in the raw material. This discrepancy is
probably due to the variability of mercury in the raw
material. The data reported in USBM (1994) are
probably high.
The estimates for air releases are based on the use of
emission factors. Emissions are expected to vary from
facility to facility based on the mercury content of the
raw material and the type of control technology in place.
5.2 Carbon Black Production
5.2.1 Introduction
Carbon black consists of fine particles of carbon usually
formed by incomplete combustion of hydrocarbons in
the oil-furnace method. This substance is frequently
used as a filler in rubber manufacturing to add both
toughness and abrasion resistance to the final product.
54
-------
The mercury in carbon black production comes from the
hydrocarbon feedstock, similar to Number 6 (residual)
fuel oil, which has an average mercury content of 0.06
ppm (USBM 1994).
The three primary raw materials used in the production
of carbon black are feedstock (either a petrochemical oil
or a carbochemical oil), air, and an auxiliary fuel such as
natural gas. The feedstock is preheated to a temperature
of between 150 and 250°C, and the air is also preheated.
A turbulent, high-temperature zone is created in the
reactor by combusting the auxiliary fuel and the
preheated oil feedstock. The feedstock is introduced into
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 of the feedstock is achieved,
and carbon black is produced (USEPA 1997a). The air
stream containing the product is cooled and the product
is collected in a fabric filter. It is shipped in dry form,
primarily for use in the rubber industry.
Mercury Consumption
Mercury may be present in the residual oil feed.
Assuming that the concentration of mercury in the feed
is 0.06 ppm (USBM 1994), and the quantity of oil
consumed is equal to the production capacity of the
carbon black (1,830,000 tons/year from USEPA 1997a),
the quantity of mercury in the feed is estimated as 0.11
tons/year.
Product
The quantity of mercury in the product is not known. It
is reasonable to expect that some of the mercury would
be collected in the fabric filter.
Air Releases
In 1995, mercury emissions from carbon black
production were estimated to be 0.28 tons (USEPA
1997a). This estimate is expected to be an overestimate
because it is based on production capacity (rather than
actual production, which is not known), and the use of a
furnace emission factor developed from 1980.
5.2.2 Materials Flow
Exhibit 5-2 illustrates the consumption, release, and
product content of mercury in carbon black production.
Water and Solid Waste Releases
No release points for these media are identified.
Releases are estimated as zero.
Consumption:
0.11 tons/yr
U.S.
Carbon Black
Production
Product:
Not available
Releases: 0.28 tons/yr
-Air: 0.28tons/yr
• - Water: 0 tons/yr
- Disposal: 0 tons/yr
Recycling: 0 tons/yr
Source: Mercury Consumption: USBM, 1994.
Mercury Release and Recycling: EPA, 1997a
Mercury in Product: No estimate available.
Exhibit 5-2. Mercury in Carbon Black Production
55
-------
5.2.3 Discussion
The estimates for the mercury in raw material and
mercury in air releases are both uncertain because of the
limited information each is based upon. Additionally,
the quantity estimated to be released is greater than the
quantity estimated to be fed to the process, indicating
additional difficulties with the data quality. Finally, an
estimate for the quantity of mercury in the product could
not be developed although this may not necessarily be a
zero quantity.
5.3 Gold Mining
5.3.1 Introduction
While the major source of mercury is supplied by
secondary sources, additional quantities of mercury are
obtained as a byproduct of gold mining. Mercury is
present in gold ore; one source estimates a concentration
of 9 ppm although the concentration is expected to vary
based on location (USBM 1994). As of 2001, less than
10 gold mines recovered mercury from the ore (USGS
2002a). The remaining gold mines did not recover
mercury, although some mercury is expected to be
present in the gold ore and in waste materials.
In gold mining, gold-containing ore is crushed and then,
if necessary, roasted to remove sulfur. Mercury that is
present in the ore is vaporized and collected in air
pollution control devices; mercury in these wastes may
be either disposed or recovered in an onsite retort
furnace. Following roasting, the ore is mixed with water
and reacted with a cyanide leach solution where gold and
mercury are dissolved and solids removed via filtration.
The purified solution is sent to an electrowinning
process, where the gold is deposited on a steel cathode.
If necessary, the cathode is sent to a retort furnace for
mercury removal, then to a smelting furnace to volatilize
and purify the gold. In a retort furnace, the mercury is
collected by a condenser for subsequent sale (USEPA
1997a). Therefore, mercury present in gold ore may be
released to the land (e.g., in disposed air pollution
control wastes and spent ore tailings), to the air (e.g., not
removed by air pollution control devices), or in the gold
product (i.e., as an impurity).
Mercury was used through the early 1900's throughout
the western United States during gold mining (i.e., gold
amalgam process). Mercury was added to the ores to aid
in recovery, which resulted in widespread contamination
of mine waters and sediments (USGS 2000b). In the
United States, mercury is no longer used in gold
recovery operations although such use continues in other
parts of the world.
5.3.2 Materials Flow
Exhibit 5-3 illustrates the consumption, release, and
product content of mercury in gold mining.
Mercury Consumption
The mercury input from trace impurities in gold ore to
the gold mining process is assumed to be equal to the
amount released, as estimated below.
Releases
Mercury that is present in the ore can remain in the
waste rock or can be vaporized; the volatile mercury can
be released to the atmosphere or be collected in air
pollution control devices for mercury recovery or
disposal. Specific sources of mercury during gold ore
processing have been estimated by one company. Of
1,500 pounds of mercury estimated to be released to air
during processing, 23 percent were from milling (e.g.,
crushing), 29 percent from autoclaves, 31 percent from
electrowinning and retort, and 15 percent from furnace
stack. Other facilities reporting significant mercury
releases also operate autoclaves or roasters (Elko Daily
2000b). An estimated 10 to 50 percent of mercury
contained in the rock is removed for recovery (or
release) later in the process (Menne 1998).
There is contradictory information regarding the quantity
of mercury recovered during gold mining. USGS (1997)
indicates that this quantity is insignificant in comparison
to the quantity of mercury produced from mercury
recyclers (420 tons in 1997), while a newspaper
indicates that Nevada mines alone supply 110 to 150
tons per year (Elko Daily 2000a). An industry source
estimates that total worldwide byproduct production
(from gold, copper, etc.) is 400 tons per year (Lawrence
2000). Finally, 1999 TRI data shows that four Nevada
facilities reported onsite recycling (recovery) of 21.6
tons of mercury. Onsite recycling data from TRI were
not available for any other domestic gold or silver
mining facility. The TRI estimate is shown in Exhibit 5-
3 as a recovered product.
Both industry and academic sources suspect that most of
the 1999 TRI air emission estimates for mercury in this
industry are based on estimated, rather than measured,
mercury data (Elko Daily 2000b).
56
-------
Consumption:
1,370 tons/yr
Gold Mining
Product:
0 tons/yr
Releases: 1,348 tons/yr
-Air: 6.2 tons/yr
-Water: 0tons/yr
'-Disposal: 1,342 tons/yr
Recycling: 21.6 tons/yr
(produced on-site)
Mercury Consumption: Equal to releases plus recycling.
Mercury Release and Recycling: 1999 TRI Data.
Exhibit 5-3. Mercury in Gold Mining
Eight facilities, all from Nevada, reported releases of
mercury (or mercury compounds) in the 1999 TRI. Not
all of these facilities produce mercury as a product.
Total air releases were 6.2 tons. Total water releases
were 4 pounds (0.002 tons). Total (offsite) mercury
recycling was 0.06 tons. Total onsite land releases were
2,700,000 pounds (1,342 tons). The high disposal
quantity consists primarily of tailings in waste rock.
These estimates are shown in Exhibit 5-3.
5.3.3 Discussion
The available estimates of mercury releases from gold
mining vary widely, from 400 tons per year to 1,350 tons
per year, indicating a need for better data.
Beginning with the 1998 reporting year, mining
operations (including gold mining) have been required
to complete toxic release inventory (TRI) reports. These
data will show reported releases of mercury (and other
TRI pollutants), and may subsequently serve as
incentives to better monitor or control these emissions so
that companies can report decreases for these emissions.
Lowering of the reporting threshold for mercury in the
2000 reporting year for TRI will result in many smaller
facilities (including gold mining operations that
presently do not recover mercury) being required to
report multimedia emissions, which will improve the
quality of data for this sector.
5.4 Primary Lead and Zinc Mining and Smelting
5.4.1 Introduction
Mercury is potentially present in lead ores. Lead is
primarily mined in Missouri and Alaska for smelting
(USGS 2000c). The variability of mercury in lead ore is
expected to be less than the variability from other mined
materials that are recovered from a wider area of the
U.S.
Zinc ore is primarily mined in Alaska, with smaller
quantities obtained from Tennessee, New York, and
Missouri. Ore is processed in one of three domestic U.S.
smelters (USGS 2000a).
The ores are mined and then concentrated, generating
tailings as a waste. The concentrate is fed to a sintering
process, where sulfides are driven off using heat
(mercury is likely volatilized in this step). The sintered
material is fed to a blast furnace with coke and slag
forming constituents. Crude metal (elemental lead or
zinc) is removed as molten material and then refined.
5.4.2 Materials Flow
Mercury Consumption
The concentration of mercury in lead ore concentrate
(representative of the ore presently mined) is less than
0.2 ppm (USEPA 1997a). This results in approximately
0.18 tons/year; this quantity is reflected in Exhibit 5-4.
This is based on primary (mined) lead production of
400,000 tons in 1994, use of emission factors for air
57
-------
pollution control equipment following the sintering and
furnace operations, and the assumption that 10 tons of
ore concentrate produces 4.5 tons of lead. No data are
available for zinc ores.
Air Releases
Based on the lead ore concentration and the use of an
emission factor, USEPA (1997a) estimated annual air
emissions of mercury from lead smelting as 0.11 ton
(USEPA 1997a). Evaluation of 1999 TRI data showed
that one zinc smelting company reported mercury
releases. Mercury released into the air was 0.07 ton.
The TRI air release is shown in Exhibit 5-4.
Water Releases
No water releases were identified.
Consumption:
0.18tons/yr
U.S.
Primary Lead
and Zinc
Production
Product:
0 tons/yr
Releases: 7.8 tons/yr
-Air: 0.18tons/yr
• - Water: 0 tons/yr
- Disposal: 7.6 tons/yr
Recycling: 0 tons/yr
Source'. Mercury Consumption: Based on USEPA (1997a) for lead ore.
Mercury Release and Recycling: Air releases estimated using USEPA (1997a) and 1999
TRI data. Disposal releases estimated using 1999 TRI data.
Exhibit 5-4. Mercury in Primary Lead and Zinc Production
Solid Waste Releases
It is assumed that mercury present in the lead ore is
either released to the atmosphere or collected in air
pollution control waste and disposed. The quantity of
mercury present in such solid wastes and disposed is
estimated as 0.07 tons, which is the difference between
the mercury in the ore and mercury released to air.
Evaluation of 1999 TRI data showed one company in
this sector reported mercury release. This facility is a
zinc smelter that accepts both ores and waste materials
(electric arc furnace dust from iron production) (USGS
2000a); mercury is a contaminant in these raw materials.
Mercury releases were 7.6 tons as solid waste. The sum
of these two releases are reflected in Exhibit 5-4.
5.4.3 Discussion
Exhibit 5-4 shows a large discrepancy between
consumption and release. This is potentially due to the
absence of information regarding the mercury content of
raw materials accepted by zinc processing facilities; only
the mercury content of raw materials accepted by lead
processing facilities are identified in Exhibit 5-4.
5.5 Primary Copper Mining and Smelting
5.5.1 Introduction
Mercury is potentially present in copper ores. As with
other ores, copper ore is mined and then concentrated,
generating tailings as a waste. The copper concentrate
is fed to a smelting furnace with coke and slag forming
constituents. Crude elemental copper is removed as
molten material and then is further processed, using heat
to remove iron and other impurities. Mercury may be
driven off in the furnace or subsequent melting of the
copper (USEPA 1997a).
5.5.2 Materials Flow
Copper was recovered at 27 mines in the U.S. (USGS
2000d) and loadings of mercury in the ore were reported
for five facilities. Air releases for seven plants were
provided, with air releases much less than the reported
ore loadings on a plant-by-plant basis. The cumulative,
industry-wide total for mercury in ore concentrate is 6.4
tons, and the industry-wide total for mercury in air
releases is 0.06 tons (USEPA 1997a). The remaining
mercury is assumed to be present in the solid air
pollution control wastes. Exhibit 5-5 illustrates the
58
-------
consumption, release, and product content of mercury in
primary copper mining and smelting.
Evaluation of 1999 TRI data showed two facilities
reported mercury. One facility was an ore concentrator
operation and the second facility was a smelting
operation. Mercury releases were 16.35 tons as solid
waste, 0 tons to surface water, and 0.036 tons released to
the air. The air emissions from the facility were added
to the industry-wide total for air releases, as shown in
Exhibit 5-5.
5.5.3 Discussion
The discrepancy between consumption and release in
Exhibit 5-5 may be the result of variation in the ores.
Consumption:
6.4 tons/yr
U.S.
Primary Copper
Production
Product:
0 tons/yr
Releases: 16.4 tons/yr
-Air: 0.1 tons/yr
-Water: 0tons/yr
- Disposal: 16.35 tons/yr
Recycling: 0 tons/yr
Mercury Consumption: USEPA1997a.
Mercury Release: USEPA (1997a) for air releases and 1999 TRI data for other releases.
Exhibit 5-5. Mercury in Primary Copper Production
5.6 Pulp and Paper Manufacturing
5.6.1 Introduction
Mercury can be present in pulp and paper facilities as a
raw material impurity. It may also be present at facilities
that operate chlor-alkali mercury cells as part of the pulp
and paper manufacturing process; however, the latter
operations are discussed elsewhere in this report.
Pulp and paper plants use a variety of raw materials that
potentially contain mercury. These include the wood,
purchased chemicals containing mercury as a
contaminant, and coal used in onsite boilers for steam
generation. Key process steps of pulp manufacturing
with regard to potential mercury use include: (1)
debarking and chipping of the logs; (2) chemical pulping
using sodium hydroxide and sodium sulfide as typical
raw materials; (3) bleaching using chlorine, chlorinated
compounds, and sodium hydroxide as raw materials; and
(4) combustion or recovery processes (Kirk-Othmer
1996). There are approximately 150 pulp mills in the
U.S. (USEPA 1997a).
5.6.2 Materials Flow
Based on the above process description, and on available
mercury content information for several raw materials,
mercury may be present in the following raw materials:
coal (used in onsite boilers), sodium hydroxide (mercury
may be present if generated from chlor-alkali process),
bark (mercury may be present at levels from 0.08 ppm to
0.84 ppm [USEPA 1999d]), and sulfuric acid (mercury
was found to be present in sulfuric acid purchased from
a lead smelter [USEPA 1997b]). Any mercury present
in these raw materials is likely released to the
environment through air, water, or land disposal.
The quantities of mercury entering a pulp and paper
process are not known. In cases where concentration
data are available, the quantity of the raw material is
typically not available. Data for releases are incomplete.
An estimate of 1.9 tons of mercury per year, presented in
USEPA's Report to Congress (1997a), is principally
based on the combustion of coal and/or waste products
such as bark. Quantities of mercury in water and land
disposal are unavailable; no pulp and paper facilities
59
-------
reported releases to the 1999 TRI.
5.6.3 Discussion
Pulp mills are one of many industries that combust coal
for steam or electricity production. The combustion of
these industrial sources of coal are presented elsewhere
in this report, and would therefore account for at least
some of the use and release of mercury from this sector.
5.7 Oil Refining
5.7.1 Introduction
Mercury is present in crude oil in varying amounts
depending on its source. U.S. refineries process, or
refine, crude oil from domestic and imported sources.
The mercury present in the crude oil subsequently is
transferred to the products or is released to air, water, or
other media.
5.7.2 Materials Flow
Quantitatively, there are several sources of data available
to estimate the mercury content of crude oil processed in
the U.S. Recent data from Minnesota (1999) identifies
a range of 2.5 to 13 ppb. USEPA's (1997a) Mercury
Study Report to Congress reports the mercury content of
crude oil as 0.023 to 30 ppm weight, and USBM (1994)
gives a 'typical' mercury content of 3.5 ppm. Like other
properties of crude oil, it is likely that the mercury
content is extremely variable. The 2000 U.S. refinery
throughput was 5,514 million barrels (USDOE 2001), or
about 865 million tons per year. Using the concentration
range of 2.5 to 13.3 ppb from Minnesota (1999) results
in a range of 2.2 to 11.5 tons per year. These data are
summarized in Exhibit 5-6.
Exhibit 5-6. Mercury Content of Crude Oil and Petroleum Products
Material
Crude Oil
Gasoline
Distillate Oil
Residual Oil
Jet fuel/ kerosene
Other Products
Mercury Cone., ppb
2.5 to 13.3
1
1
4
1
—
Production, million bbl
5,514
2,910
1,310
255
612
1,224
Total Mercury Throughput, tons
2.2 to 11.5
0.46
0.21
0.16
0.10
E =0.93
—
Source: Mercury content of materials from Minnesota (1999). Nationwide throughput data for 2000 from U.S. DOE (2001).
Mercury release data for six oil refineries and bulk fuel
terminals are available from the 1999 TRI. Total
releases from these six facilities are 5 pounds to water
and 10 pounds to land. Other estimates are not available,
although additional refineries and bulk terminals are
expected to report mercury releases to the 2000 TRI due
to a change in reporting requirements for mercury (data
expected to be released in Summer 2002).
Refineries produce many products. The 2000 production
volume of these products and their mercury content are
presented in Exhibit 5-6. As shown, the apparent
mercury content of crude oil (2.2 to 11.5 tons) is greater
than the mercury content of the products (0.93 tons),
indicating that 'missing mercury' is unaccounted for.
Exhibit 5-7 summarizes the mercury flow in petroleum
refining.
5.7.3 Discussion
The quantity of mercury in crude oil can be extremely
variable. The release quantity in Exhibit 5-7 is
underestimated because data are available for only six of
more than 100 US refineries. Additionally, mercury
releases to certain media, especially air, are not routinely
measured by refineries and emissions of volatile metals
are difficult to estimate using conventional approaches.
Therefore, obtaining accurate accounting of mercury
from petroleum refining activities is a particular research
need.
5.8 Rubber and Plastic Products
In rubber manufacturing, carbon black is used as a raw
material. Carbon black is commonly produced from
petroleum products, which may contain mercury as an
impurity.
60
-------
Consumption:
2.2 to 11.5 tons
Oil Refining
Products:
0.93 ton
Releases:
< 0.01 ton
Source: Mercury content of materials from Minnesota (1999).
Nationwide throughput data for 2000 from U.S. DOE (2001).
Releases: 1998TRI.
Exhibit 5-1. Mercury from Oil Refining
There are no data regarding the mercury content of the
carbon black product. A lack of data regarding the
presence of mercury in this industry prevents any
estimates of mercury use and release.
5.9 Geothermal Power
5.9.1 Introduction
Geothermal power accounted for 2,650 megawatts of
power production capacity in 1992(USEPA 1997a). For
comparison, coal combustion (the principal source of
energy in the U.S.) accounted for approximately 300,000
megawatts of power production in 1996 (USEPA
1999d).
Geothermal plants operate in the western United States,
specifically in California, Hawaii, Nevada, and Utah.
Turbines in the plants are powered by steam that is
naturally present in the form of hot, high pressure water
or steam below the earth's surface (USEPA 1997a).
5.9.2 Materials Flow
Sources of mercury in geothermal plants are expected to
result from off-gas ejectors and cooling towers (USEPA
1997a). Quantitative estimates of air emissions from
these sources are presented in USEPA (1997a) based on
estimates developed from 1977 data. An estimate of 1.4
tons of mercury released to the air was developed.
Estimates to other media, and estimates of the quantity
of mercury present in the raw materials (geothermal
steam or water) are unavailable.
Mercury may be present in water or solid wastes.
Hydrogen sulfide (present in the raw material) requires
removal prior to venting of the gas; the sulfur is
collected in a solid form for disposal where mercury may
simultaneously be collected. Condensed water is also
collected, where mercury may also be present (Kirk-
Othmer 1994). No estimates for these quantities are
available.
5.9.3 Discussion
The estimate for air releases of mercury has uncertainty.
The data were developed from 1977, when operations
and air pollution control configurations may have
differed from today; which would affect the partitioning
of mercury to other media. Additionally, mercury
compounds were used in cooling towers as abiocide and
it is unknown if the factor developed in 1977 was
developed from a site where mercury was present in this
fashion. If so, present day mercury releases would be
overestimated because mercury is no longer used in
cooling towers.
5.10 Wood-Fired Boilers
5.10.1 Introduction
No mass balance estimates are available for wood-fired
boilers. These boilers are used in both residences and
industries. On a residential scale, wood is burned as logs
in a small stove. On an industrial scale, wastes or
byproducts from onsite processing of wood are burned
in an onsite boiler; these wastes could include sawdust
and wood chips (USEPA 1997a).
61
-------
5.10.2 Materials Flow comparison to quantities in other industries as to be
The mercury content of bark is reported to range from negligible.
0.08 to 0.8 ppm mercury (USEPA 1997a). The mercury
content of other wood products is not available. The
total quantity of wood burned in industrial boilers is
estimated as 100 million tons per year (USEPA 1997a).
Due to the uncertainty in mercury composition of the
feed, an estimate is not presented.
Air releases of mercury from industrial boilers are
presented as 0.26 tons per year (USEPA 1997a).
Another USEPA document estimates that the total 1994
mercury emission from wood combustion are 0.1 tons
(USEPA 1997b), but notes that the data are suspect.
Water releases are likely to be negligible, and estimates
for solid releases (e.g., combusted wood) are not
available. USEPA (1997a) presented air release factors
for industrial and residential boilers. The release factor
for residential boilers was not used because it was based
on a single data point (e.g. one wood type and one
burner). The value of this release factor was
approximately 4 orders of magnitude higher than the
value of the release factor used for industrial boilers.
Use of this release factor would have resulted in an
unreasonable estimate of mercury releases and therefore
was not used.
5.11 Utility Natural Gas Combustion
5.11.1 Introduction
Natural gas is used as a fuel at electric generating utility
power plants, and it is the second most significant fossil
fuel behind coal. In the production of electricity from
natural gas combustion, the gas is fed to a furnace with
excess air (USEPA 1999d). Generated heat is used to
transform water to steam, which drives a turbine to
generate electricity.
5.11.2 Materials Flow
An estimate for mercury air releases from natural gas
combustion was provided as 0.002 tons per year
(USEPA 1997a). Estimates for the quantity of mercury
in the incoming fuel were not available. Additionally,
no solid or aqueous wastes are expected from the
combustion of natural gas (USEPA 1999d). Therefore,
any mercury present in the fuel is probably released to
the air.
5.11.3 Discussion
No estimates for the mass balance of mercury in natural
gas combustion are presented. Only one number is
available (for air releases), and the value is so low in
62
-------
Chapter 6
Additional Sources of Mercury Resulting from Disposal or Other Final Disposition
Information in this chapter may overlap in part with
solid waste disposal figures from other sectors and may
also capture data from numerous unreported sources.
Waste management facilities were required to report to
the TRI for the first time in 1998. In 1999, seven
hazardous waste management facilities reported releases
of mercury or mercury compounds. These are facilities
that conduct a variety of services such as landfilling,
stabilization, incineration, consolidation, etc. Therefore
the data can not be neatly presented in any one section of
this report.
These seven facilities reported the following releases of
mercury and mercury compounds in 1999. A total of
272 tons were landfilled, 42 tons were sent offsite for
mercury recovery, and 0.03 tons were released to air.
This is a total of 314 tons.
6.1 Hazardous Waste Combustion
6.1.1 Introduction
USEPA's hazardous waste regulations have been in
effect for approximately 20 years. Under these
regulations, both commercial and captive (on-site)
facilities must have a permit in order to combust
hazardous waste. The types of facilities that combust
hazardous waste include incinerators (which almost
exclusively combust hazardous waste) and industrial
furnaces (which have the dual purpose of destroying
hazardous waste and deriving energy for use in other
industrial processes (USEPA 2000b)). An example of an
industrial furnace is a cement kiln.
Almost all hazardous wastes must be treated prior to
land disposal, and combustion is a common method to
remove organic constituents from wastes. Inorganic
constituents commonly remain in the ash or waste
residue, or are collected by air pollution control devices.
Such wastes would be subsequently treated or disposed.
Hazardous wastes are extremely variable in physical
form and composition, and include spent solvents, tank
bottoms, and electroplating sludge. However, not all of
these wastes are amenable to hazardous waste
combustion, or in fact undergo combustion as treatment.
For example, hazardous wastes high in mercury are sent
to mercury recovery and recycling facilities. Mercury
may be present in other hazardous wastes in small
amounts, and contribute to the mercury loading of a
hazardous waste combustion facility.
6.1.2 Materials Flow
Exhibit 6-1 illustrates the flow of mercury in hazardous
waste combustion.
Mercury Consumption
The quantity of hazardous waste combusted, both onsite
and offsite, is tracked biannually by USEPA. However,
the composition of this waste is not reported; therefore,
assumptions must be made regarding which of these
wastes are expected to contain mercury and the
concentration of mercury in the wastes.
Approximately 1,800,000 tons per year of hazardous
waste are combusted in commercial combustion units,
based on data from 1993 (USEPA 2000b); additional
waste is combusted in onsite (captive and
noncommercial) units. Extremely rough estimates can
be made regarding the quantity of mercury present in
these wastes. Specifically, by accounting only for three
hazardous waste types that are known to contain
mercury, and ignoring the mercury content of other
waste types, the total quantity of these wastes combusted
is 58,000 tons in 1995 (USEPA 1998b). These wastes
were probably combusted because they contained
organic constituents in addition to mercury. The
concentration of mercury in these types of wastes can
hypothetically range from less than 1 part per million to
100 percent, although such high mercury wastes are
typically not incinerated.
63
-------
Consumption:
1.0 tons/yr
Hazardous
Waste
Combustion
Product:
0 tons/yr
Releases: 7.1 tons/yr
-Air: 7.1 tons/yr
• - Water: 0 tons/yr
- Disposal: 0 tons/yr
Recycling/Reuse: 0 tons/yr
Source: Mercury Consumption: Estimated using USEPA (1998b) and USEPA (1998d).
Mercury Release and Recycling: Estimated using EPA Mercury Study Report to Congress (1997a).
Exhibit 6-1. Mercury in Hazardous Waste Combustion
Based on data from a separate and less comprehensive
survey of hazardous waste treatment operations, an
estimate for the quantity of mercury in mercury-
containing wastes can be made. Over one hundred waste
treaters (not all of them conducting combustion)
provided data regarding the average composition of their
wastes. Mercury composition data for wastes being
combusted was extracted from this data set. The
concentration of mercury ranged from not detected (or
not reported) to 18 mg/kg, with the median concentration
below 1 mg/kg (USEPA 1998d). Applying this high
concentration of 18 mg/kg to the 58,000 tons of
mercury-containing waste results in approximately 1 ton
of mercury in the raw material feed. This estimate has
significant uncertainty. It may be biased low because
other wastes that may contain mercury in low
concentrations are not accounted for. It may also be
biased high because the highest reported concentration
is used as representative of all wastes. Finally, the
inherent variability of mercury composition in hazardous
wastes prevents an accurate accounting.
Releases
Air releases from hazardous waste combustion are
estimated as 7.1 tons of mercury in 1995 (USEPA
1997a). Mercury releases to other media are not known.
It can be assumed that all of the mercury in the waste is
vaporized during combustion. Potential releases include
the collection of mercury in air pollution control devices
for subsequent disposal.
6.1.3 Discussion
While the quantities of hazardous waste combusted are
carefully tracked by USEPA, the quantity of mercury in
these wastes is largely unknown. The imbalance in the
input and output of mercury from hazardous waste
combustion is the result of using two different sources of
estimates.
6.2 Crematories
6.2.1 Introduction
This section focuses on crematories, i.e. establishments
that cremate human corpses. Facilities may fall under
the following SIC orNAICS codes.
SIC Code 7261: Funeral Service And Crematories
NAICS Codes 81222: Cemeteries and Crematories
Cremation is the process of reducing a body to ash and
bone fragments through the process of high heat.
Mercury associated with crematories comes from the
volatilization of amalgam tooth fillings that contain
approximately 50 percent mercury. The combustion of
fillings results in the vaporization of much of the
contained mercury and its release to the atmosphere,
from which it is ultimately deposited in soil or into
bodies of water.
64
-------
Consumption:
1.4 tons/yr
Crematories
Product:
Not Applicable
Releases: <0.1 tons/yr
-Air: <0.1 tons/yr
• - Water: 0 tons/yr
- Disposal: 0 tons/yr
Recycling/Reuse: 0 tons/yr
Mercury Consumption: Extrapolated from USDHHS (1993) and Yoshida (1994).
Mercury Relase and Recycling: Estimated using EPA Mercury Study Report to Congress (1997a).
Exhibit 6-2. Mercury in Crematories.
6.2.2 Materials Flow
Exhibit 6-2 demonstrates the flow of mercury during the
cremation process.
Mercury Consumption
Mercury is a component in tooth fillings used to prevent
further tooth decay in humans. Amalgam, or "silver
filling," is made from fifty percent elemental liquid
mercury, thirty-five percent silver and fifteen percenttin,
or tin mixed with copper, and sometimes small amounts
of zinc, palladium, or indium (Kennedy 1996). The
filling is placed in the cavity of a tooth after a dentist
drills out decay. A Japanese study showed that the
mercury content per amalgam filling is 0.6 grams
(Yoshida 1994). From 1971-74, U.S. adults 18 to 74
years old had an estimated average of 6.9 filled teeth, but
since 1979, amalgam use has decreased 38 percent
(USDHHS 1993). Assuming the average number of
cavities per person has similarly decreased, the 4.3
fillings per person results in 1.4 tons (1260 kg) of
mercury in the teeth of the 488,224 people cremated in
the U.S. in 1995.
Air Releases
There were 488,224 cremations in the United States
during 1995. Mercury emissions from a body during
cremation range from 3.84 x 10"8to 1.46x 10"6kilograms
(8.45 x 10'8 to 3.21 x 10'6 pounds). The average
emission is 0.94 x 10"6 kilograms per body (2.06 x 10"6
pounds per body), resulting in 0.46 kilograms (5.1 x 10"4
tons, or 1 pound) of mercury emissions from cremation
in 1995 (USEPA 1997a).
Solid Waste Releases
Cremated remains are the noncombustible bone
fragments. No data were available for mercury
concentration in remains.
6.2.3 Discussion
Differences between the input and output could be
accounted for in several ways . Only one set of data were
used to determine the average quantity of mercury
emitted during a cremation (USEPA 1997a). The data
are inconsistent with previous literature. For instance,
previous USEPA research indicated that, on average, 1
gram of mercury is emitted during a cremation, but that
estimate was based on European data that may not
accurately reflect U.S. dental practices and thus is
somewhat uncertain (Massachusetts 1996). Using 1
gram per body results in 1,076 pounds per year (or 0.5
tons per year). In addition, an estimated 40 pounds of
mercury were released via cremation in Michigan in
1994 (M2P2 1996). Using cremation emissions in
Michigan as a per capita average emission rate for the
total U.S. population (USDOC 2001) results in 1,100
pounds or 0.6 tons per year of mercury as an air release
across the country. This supports the hypothesis that
temperatures in a crematory (1400-1800 °F) are high
enough to combust all the mercury, which boils at 674
Mercury vapors are constantly emitted from fillings.
However, since the average daily intake for a person
with fillings is 1.24 micrograms of mercury (USDHHS
1993), the amount of mercury "lost" before cremation is
65
-------
minuscule.
POTWs.
No information regarding pollution control or other
actions that have been taken to address mercury
emissions from crematoria. While control devices are
present on stacks, no information regarding mercury
capture is available.
6.3 Sewage Treatment and Sludge Incineration
POTWs are likely to be included under the following
SIC and NAICS codes:
SIC Code 4952: Sewerage systems
NAICS Code 221320: Sewage Treatment Facilities
6.3.1 Introduction
Publicly owned treatment works (POTWs) accept and
treat wastewaters from domestic and industrial
operations. Mercury may enter these wastewaters from
the sources listed in this report. Additionally, mercury
emitted to the air may ultimately deposit and be present
in stormwater discharged to the treatment plant.
The sewage treatment process includes the following
steps: (1) collection of wastewater for centralized
treatment; (2) treatment of the wastewater through
processes including filtration, biological treatment, and
clarification; are (3) discharge of treated wastewater to
a surface water. The sludge generated may be managed
by land treatment (e.g., in designated treatment areas or
as use on public or private lands for soil enrichment
purposes) or by incineration.
6.3.2 Materials Flow
Exhibit 6-3 illustrates the flow of mercury in sewage
treatment, accounting for management of sludge by both
incineration and land disposal or land application.
Mercury Consumption
Due to the varied sources of mercury potentially present
in waters influent to a POTW, it is impossible to identify
the quantity of mercury entering such a facility without
monitoring data. Monitoring data for industrial facilities
discharging to a POTW are sometimes available for
mercury; however, the lack of automation for these data
in a national framework makes the data extremely
cumbersome to use for the many industrial sources. Due
to these data limitations, an estimate of the quantity of
mercury entering POTWs on a national level was
developed from the quantity of mercury released from
Water Releases
The Permit Compliance System (PCS) identifies
monitoring data for facilities with NPDES permits (i.e.,
those that discharge to a surface water). The PCS was
searched for monitoring data relevant to mercury which,
depending on the facility-specific permit, may be present
in the database in a number of forms such as total,
dissolved, etc. Furthermore, the search was limited to
those facilities identified in SIC code 4952 (as identified
above, such facilities conduct sewage treatment). Using
the calculation procedure below, it is estimated that 5.5
tons of mercury were released to water.
Data were available from PCS for approximately 700
facilities that reported monitoring data for mercury in
1997. While there are several different forms of mercury
presented in the database, two forms were predominant:
"total mercury" and "total recoverable mercury." In
compiling the data, the most complete, recent year's data
were used for each facility. In most cases this was 2000
or 2001, but in some cases 1999 represented the most
complete monitoring for mercury. For each facility, a
calculation was done to develop a single average value.
PCS presents data as concentrations (e.g., ppm) as well
as quantity loadings (e.g., Ibs. per day). Quantity
loadings were used preferentially; when not available,
the average concentration was multiplied by the facilities
permitted flow rate (which overestimates releases).
In measuring mercury, many facilities report non-detect
values. These are handled in two different ways: the
concentration is assumed to be zero when the facility did
not detect mercury at anytime during the year. If at least
one measurement was detected, then non-detect values
were assumed to be one-half of the detection limit.
The following uncertainties apply to the above estimate:
• Data are only available for facilities actually
monitoring for mercury. There may be instances
where additional facilities are discharging mercury
but are not recorded in PCS.
Facilities do not conduct continuous monitoring for
mercury. These loadings may be calculated from as
little as a single data point collected during the year.
Therefore, the collected data may not necessarily be
representative of the actual discharges.
66
-------
Consumption:
12tons/yr
Sewage
Treatment and
Sludge
Incineration
Product:
No Product
Releases: 12 tons/yr
- Air: <0.9 tons/yr
- - Water: 5.5 tons/yr
- Disposal: 5.5 tons/yr
Recycling/Reuse: 0 tons/yr
Source/a Mercury Consumption: Estimated from releases.
Mercury Release and Recycling: Water and disposal releases estimated from 1999-2001 PCS data, and EPA
(1985) sludge partitioning data. Air releases from USEPA (1997a).
Exhibit 6-3. Mercury Flow in Sewage Treatment
Data in PCS may, in some cases, be incorrect due to
entry errors. In developing the U.S. estimate,
particular attention was given to double checking
facilities with significant impacts on the results.
Air Releases
Potential sources of air releases are from the wastewater
treatment itself, and the incineration of the sludge
generated from treatment. No estimates are available for
fugitive air emissions from the sewage treatment plant;
the release of mercury is likely to be much less than the
release of mercury from incineration. The results from
USEPA (1985) indicate that a very small percentage of
mercury volatilizes during treatment (this source is
explained in greater detail below). Estimates are
available from the incineration of the treatment sludge,
however. A total of 0.94 tons of mercury are estimated
to be released to air from incineration (USEPA 1997a).
This estimate is calculated from the estimated quantity of
sludge incinerated in a year, average emission factor for
various types of combustion and air pollution control
units, and distributions of the type of combustion units
and air pollution control units.
Solid Waste Releases
Potential sources of solid waste releases (releases to
land) result from the disposal of the generated sludge and
disposal of any ash or air pollution control wastes
generated from combustion of the sludge. Estimates
regarding mercury in air pollution control residues are
not available. The data in USEPA (1997a) did not
indicate the efficiency associated with various control
devices for sewage sludge incineration.
The quantity of mercury in sewage sludge can be
roughly estimated based on the results of an USEPA
study from the 1980s (USEPA 1985). As part of this
study, removal efficiencies of various contaminants were
estimated using data from 40 POTWs as well as from
other sources (e.g., USEPA research projects). Mercury
was estimated to partition 50 percent to sludge and 50
percent to released effluent, with negligible air
emissions. Applying this percentage to the present-day
data, it can be estimated that 5.5 tons of mercury are
present in sludge. This quantity is assumed to be land
disposed directly or remain in air pollution control
residues for incineration.
The use of the 50 percent figure likely represents an
average value of various systems applicable at the time
of the study. There is probably variation in mercury
partitioning on a facility-specific basis.
6.3.3 Discussion
As discussed above, estimates for mercury loading to
water were available using PCS. Using these data,
loadings to solid waste disposal were subsequently
estimated. Releases to air were available from a second
source. No data are available regarding the loading of
mercury to POTWs. This necessitated the 'back
calculation' of this quantity using the release data, and
could not serve as a check for the accuracy of the release
data.
67
-------
To better support the estimate, additional research could
be conducted regarding the PCS data used for the water
release estimate. This could include determining the
portion of POTWs represented by the data set,
verification of the loadings calculations, and better
assessing the influence of non-detect values on the
calculations.
6.4 Municipal Waste Combustion
6.4.1 Introduction
USEPA has estimated that in 1995 approximately 80
percent of municipal solid waste is landfilled and 20
percent is combusted (this excludes material removed
from the wastestream for recycling) (USEPA 1997c). In
previous sections of this report, this partitioning was
used to estimate the quantity of a particular product (e.g.,
fluorescent lamps) that was eventually released as a solid
waste (i.e., sent to a landfill) and that which is emitted as
an air release (i.e., sent to a combustion unit).
6.4.2 Materials Flow
Municipal waste combustors were estimated to emit 30
tons of mercury in 1995 (USEPA 1997a). Mercury is
present in the municipal solid wastes that are burned in
these units, such as from consumer products.
Subsequently, mercury may also be present in the bottom
ash (noncombustible) residue, in air pollution control
wastes such as collected fly ash, or in the stack gas
emitted to the atmosphere.
The quantity of mercury emitted to the air during
combustion is estimated to be decreasing. This is due to
the decrease of mercury-containing products in
municipal solid waste (due to both source reduction and
recycling) and the implementation of state and federal
laws to control air emissions from municipal waste
combustion units. For example, final emission standards
have been promulgated for "large" municipal waste
combustors in 1995 (60 Federal Register 65387;
December 19, 1995) and proposed for "small" units in
1999 (64 Federal Register 47233; August 30, 1999), as
discussed in
http://www.epa. gov/ttn/atw/129/mwc/rimwc2 .html. In
both cases, mercury is one of the contaminants addressed
in the regulations.
The quantity of mercury estimated to be present in
incoming wastes to a municipal solid waste combustion
unit can be developed from examination of the data
presented in Chapter 3 of this report. Specifically, atotal
of 13 to 21 tons of mercury in the products were
assumed to enter a combustion unit as part of municipal
solid waste. Finding alternative methods of estimating
the quantities of mercury present in incoming wastes is
difficult due to its heterogenous nature. Furthermore,
while this report estimates that 13 to 21 tons per year
may enter municipal waste combustion units nationwide,
no estimate is presented regarding the quantity that is
eventually released to the air, to the land (as pollution
control waste), or sent for mercury recovery (if any).
6.4.3 Discussion
Data regarding the mercury content of ash are available
from characterization studies of the late 1980s and early
1990s. However, such data may not be representative of
present-day ash. This is due to changes in the mercury
composition of municipal solid waste, as well as revised
control technologies in place since the 1995 air rules.
Various control technologies are available for
controlling mercury emissions. In response to the new
air regulations, such controls are likely to be added or
optimized. Such controls include removal of mercury
entrained on ash in particulate collection devices
(USEPA 1997a), and the control of vapor through
activated carbon (Krishnan 1994).
6.5 Landfills
6.5.1 Introduction
As mentioned throughout this report, a variety of
mercury-containing wastes are disposed on the land.
This includes industrial wastes (e .§., from manufacturing
processes where mercury is used), air pollution control
wastes (e.g., where mercury is present in the influent
fossil fuel or ore), and municipal solid wastes (e.g.,
where mercury is present in consumer products. These
materials can be disposed in industrial waste landfills,
municipal solid waste landfills, and hazardous waste
landfills. Much of the mercury containing waste
described in this report may be managed with general
household trash, which may be combusted or landfilled.
6.5.2 Materials Flow
Mercury may be re-released from these landfills in the
form of air emissions, runoff, and leachate. An
estimated 0.08 tons of mercury is emitted in air releases
from municipal solid waste landfills (USEPA 1997a).
This is a small quantity in comparison to the total
quantity of mercury disposed to land. It is also small in
comparison to other air releases.
68
-------
6.5.3 Discussion
All instances of landfill disposal were discussed
previously in this report as part of individual products or
processes involving mercury. These estimates are not
presented here in order to avoid "double counting" of
such sources.
6.6 Medical Waste Incineration
6.6.1 Materials Flow
There are an estimated 7,000 hospitals in the U.S. with
approximately 34% operating their own incinerator (62
FR 48347; September 15, 1997). Sources vary with
respect to the amount of waste hospitals generate
annually. High-end estimates indicate 2 million tons of
waste is generated (Citizens for Environmental Health
2002) while another projection suggests the level may be
as low as 600,000 tons (Valenti 2000). USEPA (1997a)
estimated that the quantity of mercury emitted to the air
from this source is 16.0 tons per year in the 1994 to 1995
time-frame. Since that estimate was made, USEPA
published a final rule relevant to the control of mercury
and other emissions from this source category (62 FR
48347; September 15, 1997).
6.6.2 Discussion
Several pollution prevention activities are underway for
the reduction of mercury in the medical field, which
should reduce the amount of mercury fed into medical
waste incineration. For example, a Memorandum of
Understanding was developed between the American
Hospital Association and USEPA on June 25, 1998.
This memorandum discusses the elimination of mercury
in hospital wastes (AHA 1998).
69
-------
Chapter 7
Geographic Distribution of Mercury
7.1 Purpose
In previous chapters, data were presented to show how
product and industrial sectors use and release mercury in
the U.S. on a national basis. That profile, or Materials
Flow Analysis (MFA), shows the flow of mercury
through the U.S. economy and released into the
environment. The data obtained from the MFA can be
used to target sectors that emit the largest quantities of
mercury throughout their product lifecycle.
In conjunction with quantifying mercury use and release
on an industry-specific basis, this chapter presents data
geographically. Specifically, the releases from all
industries are identified for each State and for selected
local regions. This analysis allows for the targeting of
specific areas of the country in which the largest
quantities of mercury are released. The profile is useful
in identifying local, regional, and national importance of
mercury activity.
7.2 Data Sources and Limitations
Exhibit 7-1 lists the data sources presented in this
chapter that allow for regional distinctions. Exhibit 7-1
includes many of the same data sources used for
estimating the materials flow of mercury on a national
level. However, not all of the data sources discussed
earlier in this report are presented in Exhibit 7-1. In
most cases, this is because insufficient information is
available to allow for adequate treatment below the
national level. For example, the estimation of releases of
mercury from lighting use is dependent on data
regarding lamp sales at the national level. Since state-
level sales data are not available for these and other
products, it is not possible to estimate the geographic
distribution of mercury use and release from this
particular source.
An important limitation with a geographic presentation
of data is that a single complete geographic distribution
cannot be presented. This is because the analysis uses
multiple data sources which vary in scope and objective.
For example, TRI data present multimedia releases of
mercury but omit a significant number of facilities
known to release mercury. Alternatively, the 1996
National Toxics Inventory (NTI), while somewhat dated,
presents a more complete geographic description of
mercury releases to air but it is not possible to identify
the degree of overlap with other data sources because not
all of the data are facility-specific. It is very difficult to
combine these different data sets into a single
presentation to identify a 'single' release estimate for a
particular locality.
A second limitation of presenting data by geographic
region is that the release estimates cannot be 'rolled up'
to identify national estimates of flow and release. Using
TRI as an example, the data provide differing levels of
coverage depending on the industry and therefore
estimates are made using different data sources from one
industry to the next. In addition, as mentioned
previously, estimates used for national estimates such as
lighting sales data do not have a local component.
Each source identified in Exhibit 7-1 is described in
detail below, with descriptions of the availability,
completeness, and quality of the data source. Data from
each source are presented later in this chapter.
7.2.7 Toxics Release Inventory
TRI provides facility-specific environmental data for
mercury, mercury compounds (this category does not
distinguish between the type of compound), and other
chemicals. The data include releases to air, water, land,
and solid waste. TRI does not require all facilities that
emit mercury to report emissions; a facility only reports
emissions if it meets thresholds set by the TRI program.
Data are provided annually (the most recent data are for
1999) through Envirofacts
(http://www.epa.gov/enviro/html/tris/). Limitations of
the data are that, as discussed above, not all facilities
70
-------
Exhibit 7-1. List of Data Sources For Geographic Distribution of Mercury
Data Source
Toxics Release Inventory
(TRI)
National Toxics Inventory
(NTI)
National Response Center
Database (NRC)
U.S. EPA Utility Information
Collection Request
National Fish and Wildlife
Fish Consumption Advisories
Media
Multimedia
Air
Spills
Air
Environment
Description
U.S. EPA data from 1999 for approximately 80
facilities where relatively large quantities of mercury
are present during a year.
U.S. EPA data from 1996 identifying mercury
releases for virtually all U.S. counties.
Data for accidental releases of about 100 materials
identified as containing mercury as a primary
component in 2000.
Data collected by EPA in 1999 and extrapolated to
each of the approximately 450 U.S. coal-fired power
generation facilities.
Identification of all existing and news wildlife
advisories in 2000 where mercury is identified as the
reason for the advisory.
are required to report mercury or are required to identify
releases from all mercury-related activities. In addition,
the data reported by the facilities themselves are of
varying quality, being either estimates or measured
results.
7.2.2 National Toxics Inventory (NTI) Data
The NTI database provides air release information for
mercury as well as for other chemicals. The inventory
contains estimates of emissions from major, area, and
mobile source categories. The database is different from
TRI in that no specific reporting threshold is included in
the inventory. Larger sources are identified by facility
and smaller sources (e.g., gasoline stations) are grouped
as area sources and categorized both by industry and
location. Major and area sources both are stationary
sources differing in their potential to release air toxics
(as well as differing in their regulatory requirements and
the availability of data). Mobile sources include
highway traffic, aircraft, etc. The data are updated every
three years and the latest data available are from the
1996 inventory. The 1996 inventory incorporates
information collected from states, TRI data, other EPA
information, and estimation procedures. The data are
available from EPA's web site
(http://www.epa.gov/ttn/chief/net/index.html).
7.2.3 National Response Center Database
The National Response Center (NRC) manages the
reporting of all chemical and fuel spills (or other
accidental releases) in the U.S., consolidating and
simplifying reporting required by many legislative
statutes. Facilities are required to report spills or
accidental releases if they are subject to certain statutes,
such as the Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA). These
requirements generally require the reporting of spills
greater than certain quantities although in reviewing the
data for 2000 there is a wide range in the quantities
reported. The data are continuously updated with the
most recent whole-year data available for 2000. The
data are available as data files from the National
Response Center's Internet site
(http ://www.nrc .uscg.mil/index.htm).
These data were reviewed for spills involving mercury.
Identifying quantities from this data source is
problematic for a number of reasons. First, the data
collected by NRC are based on preliminary information,
and therefore quantities reported may be subj ect to error.
In other instances the quantity of mercury spilled is not
known. For this reason, data regarding the number of
incidents are expected to provide a better indication of
accidental mercury releases than the quantities reported.
7.2.4 National Listing of Fish and Wildlife
Consumption Advisories
EPA collects advisory information from states regarding
the consumption of fish by general and sensitive
populations.
71
-------
This data source is unique among the others identified in
this chapter because it identifies mercury in the
environment, rather than from the source. In this
instance, it is impossible to know the source of the
mercury. Data on mercury advisories are available from
fact sheets from EPA's web site
(http://www.epa.gov/waterscience/fish/').
A principal limitation of using the fish advisory data is
in reaching conclusions. Because there are many
species, waterways, and potential contaminants, a state
is limited by the extent of sampling it can conduct.
Therefore, differences in the number of fish advisories
from one year to the next, or from one state to another,
may simply be the result of differences in the number of
assessments conducted or the sampling and analysis
techniques employed.
7.2.5 U.S. EPA Information Collection Request
EPA has collected extensive data regarding the mercury
emissions of coal-fired power plants and the mercury
content of coals used at coal-fired power plants. These
data were collected in 1999 and 2000 for a one-time data
collection effort. These data are extremely useful
because they can characterize the use and emissions of
mercury from all coal-fired boilers in the 1999
timeframe. Data from this effort are available from EPA
(http: //www. epa. gov/ttnatwO 1 /combust/utiltox/utoxpg.
html).
Data regarding the mercury content of coals, coal usage,
and boiler characteristics were collected from all coal-
fired power plants. More comprehensive air sampling
data were collected from 10 to 20 percent of the
facilities, which EPA statistically extrapolated to the
population as a whole. The air sampling activity
included quantifying total mercury emissions as well as
the species emitted.
7.3 Findings
The data gathered from the above sources were used to
map the releases of mercury in the United States. Each
data source provides estimates of the overall quantity of
mercury emitted. The resulting data maps can be used to
identify local and regional hotspots of mercury
emissions. Examination of the maps allow researchers
to identify which local and regional areas of the country
emit the largest quantities of mercury. In addition, the
map will assist in determining which areas of the country
produce mercury of global concern.
7.3.1 State-Level Maps
Each of the five data sources discussed above were used
to generate data for mercury at the State level. Data for
each parameter identified in Exhibit 7-1 were aggregated
for each state and plotted on a map of the U.S. using a
simple spreadsheet. The results are presented in Exhibits
7-2 through 7-8.
Exhibits 7-2 and 7-3 present 1999 TRI data for
multimedia releases and air releases, respectively, for
facilities reporting for both 'mercury' and 'mercury
compounds.'
Exhibit 7-2 sums all quantities of mercury reported to be
released. The predominance of releases are from the
Eastern states. Many States have no facilities reporting
mercury releases. The only Western states that had
significant emissions according to TRI was Nevada and
Arizona. The Nevada emissions are mostly due to gold
and copper mining in the state, and represent by far the
highest emissions.
In the context of TRI, release refers to virtually any
quantity of mercury entering the environment in any
form, regardless of the potential risk posed or the media
impacted. Therefore, for a more comparable analysis of
TRI data and to facilitate comparisons with other data
sources, Exhibit 7-3 presents data for air releases only.
Again, Nevada leads the States for air releases of
mercury.
Exhibits 7-4 and 7-5 present NTI data. Exhibit 7-4
presents total mercury air releases for each State while
Exhibit 7-5 presents emission density (i.e., emissions per
square mile) for each state. While related, there are
some differences. For example, small mid-Atlantic
States with moderate emissions have extremely high
emission densities, while some larger Western States
with low emissions have even lower emission densities.
The NTI map shows the highest emissions and
concentration of mercury occur in the Midwest and
Eastern United States. The density of mercury are
highest in the following states: Delaware, Maryland,
Massachusetts, New Jersey, and Rhone Island (Appendix D
B). However, the overall top five state mercury emitters
are Texas, Florida, New York, Pennsylvania, and
Indiana.
72
-------
Exhibit 7-2. Total Mercury Releases by State (1999 TRI)
Total Mercury Releases
(tons)
220to 1,350 (1)
40 to 220 (2)
20to 40 (1)
10to 20 (2)
D Oto 10 (25)
Exhibit 7-3. Total Mercury Air Releases by State (1999 TRI)
Total Mercury Air Releases
(tons)
0.3 to 6.22 (10)
n 0.05to0.3 (5)
n 0.03to0.05 (2)
DO.01toO.03 (3)
n 0 toO.01 (11)
73
-------
Total Mercury Air Release by State (NTI1996)
Mercury Air Releases
(short tons)
4.1 to 23 .6 (10)
3.2 to 4.1 (8)
1.4 to 3.2 (12)
0.8 to 1 .4 (9)
0 to 0.8 (12)
Exhibit 7-4. Total Mercury Air Releases by State (NTI 1996)
Mercury Emission Density by State (NTI 1996)
Mercury Emission Density
• 0.000213 toO.000421 (6)
• 0.000103 toO.000213 (8)
D 7e-005 toO.000103 (7)
D 4e-005 to7e-005 (6)
D 1.7e-005 to4e-005 (8)
D 1.4e-005 tol.7e-005 (5)
D 3e-006 tol.4e-005 (8)
Exhibit 7-5. Mercury Emission Density By State (NTI 1996)
74
-------
Number of Mercury Spills Reported per State (NRC 2000
Number of Mercury Spills
• 8to14 (5)
H 7 to 8 (1)
• 5 to 7 (2)
D4to 5 (1)
• 3to 4 (7)
• 2to 3 (6)
D 1to 2 (12)
Exhibit 7-6. Number of Mercury Spills Reported per State (NRC, 2000)
75
-------
Mercury Fish Advisories by State
New Hampshire 5
Massachusetts 89
Rhode Island 2
Connecticut 6
New Jersey 30
Delaware 5
Exhibit 7-7. Mercury Fish Advisories By State
76
-------
Coal-fired Utility Boilers Release of Mercury by State (ICR 2000;
y
Mercury Releases by State
(tons)
• 1.74105.03 (7)
B 1.12101.74 (7)
n 0.86101.12 (7)
n 0.53100.86 (5)
n 0.34100.53 (5)
n O.OStoO.34 (9)
n 0 to 0.08 (7)
Exhibit 7-8. Coal-fired Utility Boilers Releases of Mercury per State
77
-------
The NTI results are consistent with the TRI data in that
mercury releases are concentrated in the Eastern States,
whether measured by total emissions or emission
density. However, two important deviations exist. Most
significantly Nevada is identified from the NTI data as
having virtually no releases, while it was a nationwide
leader for TRI air emissions. This is likely due to the
differences in the age of the data. NTI data are from
1996 when estimates for releasing mercury from mining
activity were less available. A second observation with
the NTI data is that estimates are provided for all States
in the continental U.S., versus much less complete
coverage from the TRI.
Exhibit 7-6 presents data for spills from the NRC over
the course of a year (2000). Even in the states with the
highest number of spills, the rate is just over once per
month in the entire State, which is not an abnormally
high occurrence. As stated previously, data for spill
quantities are not expected to provide useful information
and therefore were not analyzed. Therefore, on the
whole, data from the NRC do not suggest that spills of
mercury containing materials are particularly
problematic for any State.
Exhibit 7-7 presents the number of fish advisories for
each state in 2000. Advisories are highly clustered
around the Great Lakes and in general are not well
correlated with releases in a State. For example, of the
ten States identified from Exhibit 7-4 as having the
highest airborne mercury releases from NTI, only three
(Indiana, Ohio, and Florida) are identified as having the
highest number offish advisories. This information can
be interpreted in several ways. Some states may be more
diligent in their monitoring and advisory efforts than
others, or it may reflect the fact that mercury emitted
from one location can be transported to another.
Exhibit 7-8 presents air emissions data from coal-fired
utilities. With a few exceptions, these data provide an
excellent correlation with air release data from TRI.
Releases in the Appalachian and Great Lakes states and
Texas are similar for both the NTI data and the coal-fired
utility data. Significant exceptions include California
and the Northeast.
7.3.2 County-Level Data
Only three of the data sources identified above were
evaluated below the aggregate State level: TRI, NTI, and
coal-fired utility emissions (ICR).
For this analysis, county level data were used as a
reasonable aggregate of a local area: data below this
level (e.g., ZIP codes) would be expected to be too
'noisy,' while data above this level (e.g., Congressional
District or Regional area) would be too difficult to
generate. NRC and Fish advisory data were not
evaluated below the State level. The small number of
spills in each State would not be expected to be
informative at the County level. For fish advisories, data
would be too cumbersome to analyze when dealing with
bodies of water that transverse multiple counties.
County level data for NTI are presented in Exhibit 7-9.
NTI was expected to be the most informative for
evaluation at the county level because of the large
number of data points. Exhibit 7-9 presents emission
densities at the county level. As expected, variations are
seen within each state. Sometimes higher emission
densities are consistent with higher population densities,
as seen near cities. In other cases, high emission
densities outside of population centers are the result of a
small number of significant point source emissions.
Exhibits 7-10 through 7-12 present tabular data of the 25
counties from each data source where mercury emissions
are highest. In the case of the NTI and ICR data sources,
the top 25 counties represent about one-third of all air
emissions from all counties. For TRI however, the top
25 counties represent greater than 99 percent of the
nationwide emissions. This is a further result of the
relatively low number of data points available from the
TRI for 1999. Additionally, Table 7-12 shows that two
Nevada counties account for a disproportionate amount
of releases, due to mining activities.
7.3.3 Integration of County-Level Data
County-level data for TRI, NTI, and coal-fired boilers
were aggregated in an attempt to better draw conclusions
from all of the data sources. The maps in Exhibits 7-2
through 7-8 allow for comparisons at the State level.
However, such a visual comparison is impossible for the
thousands of counties in the U.S. A quick comparison of
Exhibits 7-10 through 7-12 ('top 25' counties for each
data source) shows very few counties repeating from one
data source to another. For this reason, an analysis was
conducted which attempts to combine TRI, NTI, and
ICR air release data for each county.
78
-------
1996 County Emission Densities
Mercury Compounds — United States Counties
Distribution of U.S. Emission Densities
Highest in U.S. , , 0.055
*5
90
Percent! le 75
50
25
Lowest in U.S.
Pollutant Emission Density by County
- - - - -' * *
O.O0044
0.000 17
o™3 C tons / yea r/sq. mile)'
O.OOOOO1 0
0
Source: U.S. EPA / QAQPS
NATA Nationa I—Sea le Air Toxics Assessment
Exhibit 7-9. NTI County Density Map for Mercury Compounds
79
-------
Table 7-10. 25 Counties Reporting Highest Mercury Air Emissions from NTI
County
Jasper
Dade
Pinellas
Hillsborough
Westchester
Greene
Noble
Frederick
Whatcom
Broward
Los Angeles
Harris
Cook
Essex
Marshall
Delaware
Grant
Titus
Marion
Fairfax
Rusk
Wood
Tulsa
New York
Calhoun
State
TX
FL
FL
FL
NY
NY
IN
MD
WA
FL
CA
TX
IL
MA
KY
PA
WV
TX
IN
VA
TX
WI
OK
NY
TX
All Others (3, 191 total)
Total
NTI emissions Ib
24,473
5,846
5,683
5,373
4,472
4,384
3,916
3,596
3,536
3,436
3,296
3,240
3,108
2,939
2,768
2,734
2,460
2,452
2,359
2,283
2,277
2,164
2,151
2,109
2,098
222,873
295.707
% of Total NTI Releases
7.51%
1.79%
1.74%
1.65%
1.37%
1.34%
1.20%
1.10%
1.08%
1.05%
1.01%
0.99%
0.95%
0.90%
0.85%
0.84%
0.75%
0.75%
0.72%
0.70%
0.70%
0.66%
0.66%
0.65%
0.64%
68.36%
100%
80
-------
Table 7-11. 25 Counties Reporting Highest Mercury Air Emissions from Coal-fired Boilers
County
Walker
Armstrong
Titus
San Juan
Indiana
Will
Rusk
Montour
Monroe
Jefferson
Tazewell
Kanawha
Mercer
Beaver
Person
Spencer
Gallia
Grant
Leon
Mason
Monroe
Clearfield
Coshocton
Rosebud
Shelby
State
AL
PA
TX
NM
PA
IL
TX
PA
GA
OH
IL
WV
ND
PA
NC
IN
OH
WV
TX
WV
MI
PA
OH
MT
AL
All Others (368 total)
Total
ICR Hg total Ib
2,490
2,154
2,093
2,089
1,848
1,600
1,363
1,216
1,201
1,179
1,125
1,093
1,057
1,036
1,024
1,018
1,011
974
964
963
936
926
897
891
877
63.747
95.772
% of Total ICR Releases
2.60%
2.25%
2.19%
2.18%
.93%
.67%
.42%
.27%
.25%
1.23%
1.17%
.14%
.10%
.08%
.07%
.06%
.06%
.02%
.01%
.01%
0.98%
0.97%
0.94%
0.93%
0.92%
66.56%
100%
81
-------
Table 7-12. 25 Counties Reporting Highest Multimedia TRI Releases
County
Humboldt
Elko
Tooele
Eureka
St. Clair
Pershing
Whatcom
Salt Lake
Sumter
Sumter
Beaver
Lorain
King
Harris
Porter
Bradley
New Castle
Ashtabula
Platte
Penobscot
Marshall
Iberville
Colbert
Marshall
Richmond
State
NV
NV
UT
NV
IL
NV
WA
UT
SC
AL
PA
OH
WA
TX
IN
TN
DE
OH
WY
ME
KY
LA
AL
WV
GA
All others (49 total)
Total
Multimedia TRI
Ib
1,231,260
1,190,814
404,140
220,359
81,599
43,008
35,807
32,802
28,325
24,841
15,230
14,943
4,458
3,667
2,800
2,640
2,172
,895
,824
,734
,662
,512
,499
,316
,268
4,996
3.356.571
% of Total
TRI Releases
36.68%
35.48%
12.04%
6.57%
2.43%
1.28%
1.07%
0.98%
0.84%
0.74%
0.45%
0.45%
0.13%
0.11%
0.08%
0.08%
0.06%
0.06%
0.05%
0.05%
0.05%
0.05%
0.04%
0.04%
0.04%
0.15%
100%
82
-------
Exhibit 7-13. 25 Counties Re
County
Humboldt
Elko
Tooele
Jasper
Eureka
Titus
Walker
Armstrong
Indiana
St. Clair
San Juan
Whatcom
Hillsborough
Rusk
Will
Dade
Grant
Pinellas
Beaver
Monroe
Montour
Jefferson
Mercer
Gallia
Person
State
NV
NV
UT
TX
NV
TX
AL
PA
PA
IL
NM
WA
FL
TX
IL
FL
WV
FL
PA
GA
PA
OH
ND
OH
NC
NTI Rank
0.03%
0.00%
0.36%
7.51%
0.00%
0.75%
0.22%
0.45%
0.62%
0.05%
0.25%
1.08%
1.65%
0.70%
0.36%
1.79%
0.75%
1.74%
0.16%
0.34%
0.32%
0.31%
0.37%
0.37%
0.34%
porting Highest Releases from 3 Data Sources
ICR Rank
0.01%
0.00%
0.00%
0.00%
0.00%
2.19%
2.60%
2.25%
1.93%
0.00%
2.18%
0.00%
0.48%
1.42%
1.67%
0.00%
1.02%
0.00%
1.08%
1.25%
1.27%
1.23%
1.10%
1.06%
1.07%
Multimedia TRI Rank
36.68%
35.48%
12.04%
0.00%
6.57%
0.00%
0.00%
0.00%
0.00%
2.43%
0.00%
1.07%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.45%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
Cumulative Rank
36.72%
35.48%
12.40%
7.51%
6.57%
2.94%
2.82%
2.70%
2.55%
2.48%
2.43%
2.15%
2.13%
2.12%
2.03%
.79%
.77%
.74%
.70%
.59%
.59%
.55%
.47%
.43%
.41%
83
-------
Speciation Emission for Coal-Fired Utility Boilers by State
Speciation Emissions
Participate
Ionized
Elemental
Exhibit 7-14. Speciation Emissions by State for Coal-Fired Utility Boilers
84
-------
Simply adding the releases reported would not be
appropriate, because each data source identifies a different
total quantity of mercury across the U.S. In addition, TRI
provides multimedia release information while the NTI and
ICR data represent air emissions only.
Instead, the far right hand columns in Exhibits 7-10 through
7-12 were used. These columns represent the proportion of
releases from the data source originating from that county,
and serves as a 'score' for the data source. The percentages
from each data source are summed to create a basis for a
final ranking. Exhibit 7-13 presents these results for the top
25 counties. Exhibit 7-13 shows that, in most instances, a
county's score is driven by a single data source. For
example, the top two counties reported extremely high
releases to TRI but low releases from the other two data
sources. In addition, in only two cases does a county
identified in Exhibit 7-13 report a non-zero ICR release and
a non-zero TRI release, indicating that in most instances
these releases occur in different localities. (NTI data are
available for essentially every county.)
This analysis can be repeated for additional data sources as
new data become available. For example, annual updates to
TRI can be integrated into the analysis, while additional
mercury release data sources not discussed in this chapter
can be added and integrated into the 'scoring' system.
However, Exhibit 7-13 shows that data limitations (in this
case, the low number of counties reporting TRI data) can
greatly influence the results.
7.4 Speciation
Methyl mercury is, from an environmental perspective, the
form of mercury which is of most concern. However, methyl
mercury is not known to be emitted from any anthropogenic
source in significant quantities. Instead, methylmercury is
formed within the environment through a complex series of
transformations. Nevertheless, the species of mercury
emitted from a given source is important for assessing
geographical impacts.
This section is limited to assessing speciation in air releases.
When released to water, mercury has an obvious local
impact. Landfill and similar releases are more complex
because migration of the mercury through the environment
is dependent on leaching and volatilzation, but nevertheless
is also of local concern. Air releases, however, do not
necessarily have immediate local impacts. Oxidized forms
of mercury (mercury compounds such as mercuric chloride)
readily deposit in a localized area once emitted. In contrast,
emissions of elemental mercury can remain airborne for long
periods of time and be transported across the country, or the
world, prior to deposition (Hanisch 1998).
There are limited data available for mercury speciation.
Speciation data are only available for emissions from two
categories: utility coal combustion and municipal waste
combustion. However, these two sources are both significant
in terms of mercury air releases. Data from the Mercury
Study Report to Congress (EPA 1997a) show that these two
sources comprise about 50 percent of air emissions
quantified in that report.
The most comprehensive data available is the EPA's ICR
from the coal-fired utility boilers. The results show that,
nationwide, the average mercury speciation breakdown is 54
percent elemental, 43 percent oxidized, and 3 percent
particulate. Therefore, the 48 tons of mercury emitted by
coal-fired utility boilers (see Chapter 4) breaks down to 26
tons elemental, 20 tons oxidized, and 1.5 ton particulate.
The studies that estimated mercury speciation had
limitations. Although sampling was conducted at a large
number of boilers and facilities, it is difficult to apply data to
individual plants since it was only a snapshot in time and
may be affected by future changes in coal supply, plant
operations, etc. Additionally, as facilities install control
devices which affect mercury capture, the resulting
speciation profile will change.
Exhibit 7-14 presents speciation data for each state from the
ICR data analysis. Some interesting trends are apparent from
this map. First, many states in the midwest and west emit
elemental mercury as the predominant species, while the
ionic form predominates in most states in the Eastern U.S.
One reason is due to the type of coal burned in each area.
Data from the ICR showed that western coals, in general,
emit a higher proportion of elemental mercury than do
Appalachian coals.
Secondly, there is wide variation in speciation results from
one state to the next. This further demonstrates that
generalities concerning a national distribution may not
necessarily apply to a local condition.
Further data collection for mercury speciation can result in
similar maps created for emissions. For example, by
combining speciation and emissions data for multiple
industries, a map can be created showing areas of the country
emitting a particular species of mercury. Such a map was not
prepared for this report because it would essentially only
reflect coal combustion emissions. Sufficiently robust
85
-------
speciation and emissions data are unavailable for other
industries.
Limited speciation data for municipal waste combustion are
also available from EPA's Office of Air (phone interview
with Jim Kilgore, EPA). EPA estimated that air emissions
from MSW combustion consisted of 85 percent ionic
mercury, <5 percent particulate, and 10 percent was
elemental. Due to MACT regulations, many facilities are in
the process of installing mercury control technologies which
will affect the speciation of stack gas. Additionally,
variation is expected between facilities due to site-specific
factors such as waste composition, boiler configuration, etc.
Therefore, the estimation has its limitations.
Nevertheless, these data show that municipal waste
combustion emissions are significantly different than coal
combustion emissions, with a much higher proportion of
oxidized mercury. One potential reason for this is the
increased chloride loadings in MSW feed, resulting in higher
rates of transformations to mercuric chloride in MSW stack
emissions.
The predominance of oxidized mercury in MWC combustor
stack gas is confirmed by a European study (Paur 1999),
although quantitative estimates are not provided.
Additionally, this source identified that sewage sludge
incineration resulted in higher levels of elemental mercury.
Data for mercury speciation in mining are available, however
this reflects data only for solid tailings material. The study
(Kim 2001) estimated the make-up of mercury from twelve
U.S. mercury and gold mine tailings using X-ray absorption
spectroscopy (XAS). The study found that most tailings
consist of cinnabar (HgS, hex) and metacinnabar (HgS, cub).
Other species of mercury found included: montroydite
(HgO), schuetteite (Hg3O2SO4), corderoite (Hg3S2C12),
and various chlorides which may be more mobile than the
cinnabar.
86
-------
Chapter 8
Conclusions
8.1 Conclusions
Even if all mercury use ended today, mercury would be
present well into the future due to the quantities of the
element in reservoirs and stockpiles. For example, some
items in commerce, such as thermostats, will not be
disposed or removed from service for a long period of
time. Given that the use of mercury in the economy and
its associated health implications will not end for many
years, mercury management will continue to be a
prominent responsibility.
The 3 7 sectors discussed in this report present a thorough
analysis of the life cycle of mercury throughout the U.S.
economy. Ideally, these data will be utilized in some
way to assist mercury management decisions, such as
policy formulation or the prioritization of research and
development efforts. This section summarizes the
estimates calculated in the report and categorizes
findings by life cycle stages and releases in key sectors.
Additionally, a section devoted to high-profile sectors is
included. Finally, observations concerning the
application of these findings to prioritize future research
needs are provided.
8.1.1 Life Cycle Stages
Acquisition
Mercury can be supplied by secondary production
facilities (430 tons per year), imports (variable), or
government stockpiles, and is often present within raw
materials as a contaminant. The majority of mercury
supply for use in product manufacturing can be
accounted for by secondary production, and secondly
from imports. Government stockpiles generally don't
supply mercury but act as a reservoir. In fact, the U.S.
Government stockpiles represent just over 50% of the
total domestic mercury reservoir (9,050 tons), indicating
that at least half of the US reservoir can be tightly
managed.
According to the data in this report, the supply of
mercury to product manufacturing is greater than the
demand. If this observation indicates a trend, then the
main domestic supply of mercury may be compromised
since the industry may become at risk of financial non-
viability in such a business environment. It seems likely
that a decrease in mercury demand is inevitable as
product manufacturers gradually use less mercury or
substitute materials, as seen in electrical lighting
manufacturing. A consistent fall in mercury demand
will probably result in decreased prices for mercury and
lower profits for secondary production facilities. In such
a scenario, recycling facilities may close, resulting in
increased use of landfills and incinerators for mercury-
containing wastes.
Product Manufacture
Mercury use among industries can be characterized as
either intentional (i.e., from secondary mercury
production or government stockpiles) or unintentional
(i.e., constituent in raw material). Mercury is
intentionally used as a raw material in product
manufacturing, these mercury-containing products are
subsequently used commercially as well as by the
general populace. Alternatively, mercury is used
unintentionally when present in trace concentrations
within raw materials. The amount of mercury used for
product manufacturing is comparable to the quantity of
mercury present from incidental uses where mercury is
a contaminant. Therefore, both uses (intentional and
unintentional) contribute significant quantities of
mercury to the total amount consumed.
Approximately 90% of the 1,700 tons per year of
mercury consumption in the U.S. can be attributed to
three sectors: gold mining (80%), chlor-alkali (5%), and
utility coal combustion (5%). The use of mercury by
gold mining and utility coal combustion is incidental
since the element is a constituent in the raw materials
used. These results limit the ability to rely on
innovations in product design as a solution for the
industrial handling of mercury. Industries that use
mercury intentionally have more technological options
87
-------
for reducing mercury consumption compared to
industries that use mercury unintentionally. For
example, a lamp manufacturer has a broader
technological horizon for manufacturing a product
using/releasing reduced or zero mercury. Conversely, a
gold mining or utility coal combustion facility may find
it difficult, if not impossible, to choose raw materials
without mercury. Consequently, most industries that use
mercury incidentally are limited to managing the releases
of mercury consumed. While eliminating mercury from
production processes is critical, the discussion above
demonstrates the importance of continued innovation in
mercury control technologies.
Where sufficient data are available to estimate the
quantities of both mercury used (consumed) and mercury
leaving the system or process, in most cases the quantity
reported entering is greater than the quantity reported
leaving. This indicates that mercury is either
accumulating within the system, or an error exists for at
least one of the estimates. This mercury 'imbalance' is
most evident for chlor-alkali manufacturing. The
estimated consumption is 79.1 tons per year, while the
quantity of mercury leaving the process in waste or
product is 27.9, a difference of 51.2 tons per year. This
may be partly accounted for by chlor-alkali's large
mercury reservoir (22% of total domestic mercury
reservoir). However, it is unlikely that these
manufacturers are storing an additional 51.2 tons of
mercury per year, rather, mercury leaving the system
may be under-reported or there is a substantial
unaccounted for sink within the system.
Product Use
Mercury is found in various commercial and professional
products that tend to be long-lived such as thermometers,
electrical lighting, and thermostats. Dental office
preparations were found to have the highest quantity of
mercury in use accounting for 13% of the total domestic
mercury reservoir. With the exception of chlor-alkali
manufacturing and dental amalgams, mercury can be
contained or recovered following product use.
Compared to the manufacturing stage, use of mercury-
containing products is characterized by considerably
higher releases of mercury into the environment.
Despite numerous efforts to collect fluorescent lights,
thermometers, thermostats, switches, and relays a
majority of the mercury in use is disposed of as solid
waste. As mentioned above, there are several industry
efforts to decrease the mercury content in certain
products and to develop mercury-free products. This
does not imply, however, that managing the disposal of
these products has a limited utility since past and present
mercury-containing products are expected to remain in
use for several years.
Final Disposition
The disposition of mercury occurs through recycling,
exports, and releases into the environment. Much of
present mercury consumption is met by existing recycled
material. However, available data do not provide a full
inventory of the total quantity of mercury recycled.
Exhibit 1-2 indicates that 430 tons per year of mercury
is estimated to be produced from U.S. mercury
recycling. However, based on examination and
accounting of mercury-containing materials, only 17
tons of mercury is accounted for as scrap materials sent
to recycling facilities. Because most data on recycling
rates are based on older or unpublished sources, these
estimates are assumed to be low. This highlights an
information gap that occurs throughout the report in
estimating the quantities of mercury in scrap and wastes
being recycled. Data detailing sector-specific quantities
of mercury recycling rates were not available.
Generally, the export of mercury is in the form of
elemental mercury (220 tons per year) as opposed to
mercury-containing products or scrap. Accordingly,
exports do not currently act as a significant mode of
spent product disposition. It is conceivable, however,
that if the secondary mercury production industry
diminishes due to reasons discussed above, industries
still generating mercury-containing scrap may prefer to
export the waste to overseas recycling facilities or waste
sites. This situation is a possibility if the total cost of
exporting the waste is less than domestic disposal.
Otherwise, scrap would be disposed of by landfilling and
incineration.
8.1.2 Releases
Segregated by media type, releases to solid waste are
greatest (1,500 tons/year), followed by air (125
tons/year) and water (20 tons/year). Gold mining alone
accounts for 90% of solid waste releases. Without gold
mining, annual releases to solid waste falls to 158 tons.
After gold mining, switches and relays disposal (29-50
tons/year) and utility coal combustion (33 tons/year)
contribute the greatest amount to solid waste releases.
Management of mercury-containing solid waste is
complicated because the mercury is embodied in various
forms such as a switches, flyash, or ore tailings. Solid
wastes are generally stored in landfills where the
-------
possibility of mercury leaching into the environment is
monitored. It is expected that mercury-containing solid
waste will decrease as product manufacturers develop
alternative, mercury-free processes. However, this is not
expected to be as noticeable for industries handling
mercury incidentally such as gold mining and utility coal
combustion, particularly as these industries grow and as
mercury removed from air continues to be disposed as
solid waste.
Mercury releases to air are clearly a significant issue
because of their transport over long and short distances.
The most significant input of mercury into air is utility
coal combustion. This industry has already been
identified as a significant input and has received
substantive attention over the past decade. Second to
this industry is coal combustion by residential,
commercial, and industrial sources with 21.2 to 23.6 tons
per year released into air. Based on this report's mercury
use and release estimates, utility coal combustion
releases approximately 45% of mercury input into the
air, whereas non-utility coal combustion may release 39 -
100% of the mercury consumed into air. A lower
effectiveness of mercury capture by non-utility coal
combustion renders this sector as a major source of
mercury releases, particularly if industries utilizing coal
combustion (non-utility) continue to grow.
Similar to air, mercury releases to water can result in
serious health issues. According to this study's data,
sewage treatment and sludge incineration (5.5 tons per
year) as well as dental offices (7.4 tons per year) release
some of the highest quantities of mercury into water.
Compared to mercury releases into air and solid waste,
water releases were considerably lower. However, data
on mercury releases to water were only found for four of
the thirty-seven industries included in this report.
Because data are not available for many industries,
mercury releases to water may be a larger problem than
depicted in this study. The limited data may reflect a lack
of industry efforts to monitor mercury releases to water.
For example, laboratories commonly dump chemicals
into sinks in low concentrations but potentially high
volumes. Ultimately, the burden is passed to POTWs,
which probably explains why this sector has the highest
releases of mercury to water. The problem may not be
the inability of POTWs to handle mercury-containing
waste, rather, it is the disposal of mercury in places such
as sinks and drains without monitoring.
Data on the geographic distributions of mercury as a
pollutant demonstrate that total releases are most
abundant in the eastern United States and Nevada.
Furthermore, available data revealed differences in
mercury speciation in air emissions between utility coal
combustion and municipal waste combustion sites.
Specifically, the average mercury speciation breakdown
for coal-fired utility boilers is 54% elemental, 43%
oxidized, and 3% particulate. When broken down by
state, the eastern boilers release mostly oxidized forms
while the west is characterized by higher elemental
releases. This contrast is most likely due to differences
in coal. Average municipal waste combustion releases
are estimated to be 85% oxidized, 5% particulate, and
10% elemental. Given that different mercury species
have varied regional or global impacts, this finding is
significant relative to air pollution control strategies.
These data have much potential for illustrating the large-
scale patterns and potential effects of mercury releases.
While the data in this report provide a reasonably
accurate depiction, more consistent data would minimize
uncertainties.
8.1.3 Key Sectors
To help identify research priorities, the sectors
corresponding to the highest quantities in mercury use,
release, and reservoirs are identified. Sectors included
are those estimated as using or releasing at least 100 tons
of mercury annually, or representing a mercury reservoir
of at least 100 tons. The selection of 100 tons is
somewhat arbitrary, but allows for a narrowing of
sectors from the 37 evaluated to a more manageable
number. A summary of this review can be found in
Exhibit 8-1. Ten sectors are listed and discussed below.
Chlor-alkali manufacturing
Chlor-alkali production using the mercury cell process
(the only process that employs mercury) is conducted at
11 U.S. locations accounting for 12% of total US
chlorine production capacity. Approximately 79 tons of
mercury are used in chlorine production annually, with
an additional 2,000 tons present in the U.S. operating
plants. These data suggest that the presence of mercury
in the chlor-alkali industry is a concern both in terms of
currently used quantities as well as the large quantities
contained in the plant that might be released into the
economy or the environment at a later date. The latter
concern is particularly relevant given the 50 ton gap
between mercury consumption and releases discussed in
section 8.1.1. Mercury releases from chlor-alkali
production is currently regulated by USEPA air, water,
and solid waste policies.
89
-------
Exhibit 8-1. Summary of Sectors with 100+ Tons of Mercury in a Life Cycle Stage
Mercury Sector
Mercury Present in
Raw Materials
> 100 Tons per Year
Mercury Released to All
Media > 100 Tons per
Year
Mercury Present as Reservoir
> 100 Tons Total
Chlor-alkali
Manufacturing
Dental
Preparations
Gold Mining
Landfills
Secondary
Mercury
Production
Thermometers
Thermostats
Switches and
Relays
Government
Stockpiles
Utility Coal
Combustion
2,000
1,200
1,370
430
1,348
230
630
4,850
105
* - There were no available data on mercury quantities in landfills. Presumably, landfills have large mercury reservoirs based on the
amount of industrial solid waste delivered to the sites.
** - Thermometers have an estimated reservoir of 45-85 tons; the sector was included based on the widespread use of these
instruments and, consequently, its relatively high reservoir.
Ongoing research needs include more accurately
quantifying the emissions and destination of mercury
used as a raw material. Based on consumption data and
release estimates, much greater quantities of mercury are
consumed per year than are estimated to be released.
More accurate accounting would serve to identify where
pollution prevention and control activities could be
targeted, or, if previously unknown, the identification of
mercury 'sinks' within the plants could potentially be
addressed.
Dental Preparations
Up to 1,200 tons of mercury are present in the U.S.
population as part of use in dental preparations (i.e.,
amalgam fillings). Among all sectors, the use of
mercury in fillings is the most intimate and direct with
respect to the manner in which it is consumed (i.e., in a
person's mouth). Unfortunately, very little opportunity
exists to address this mercury. Instead, pollution
prevention and control opportunities focus on activities
in dental offices across the United States, including
alternatives to mercury fillings and better management
of mercury wastes (including old fillings). Many
technologies are in use but are not universally adopted.
Non-mercury fillings have been successfully used, but
costs are reportedly higher. For facilities that continue
to use mercury, technologies and practices can be
employed to reduce the releases of mercury to the
environment. One technology reduces the quantity of
mercury in washwater, consisting of an amalgam
separator to recover the mercury from the water prior to
sewer discharge. The development of less expensive
non-mercury alternatives could be one key to increased
usage.
90
-------
Gold mining
Mercury is present as a contaminant in the ore and
requires removal during processing, typically as mine
tailings (solid waste) rather than as air or water releases.
Among all industries included in this report, gold mining
was distinguished by having the highest quantity of
mercury released into solid waste. Gold is mined at
about 100 locations in the United States, mostly located
in Nevada and Alaska. Presumably, solid waste from
gold mining presents limited risk to the general
population since mines are typically located in
population-sparse areas. Ranking states by population
size places Nevada at 35th (although it is the fastest
growing population in the U.S.) and Alaska at 48th
(USDOC 2001). However, ecological risk from the
relatively high quantities of mercury in mine tailings
may be an issue.
Several pollution prevention opportunities have been
implemented by a portion of the gold mining industry.
One such opportunity is mercury recovery. As a result
of processing, mercury becomes concentrated in certain
wastes which can be managed onsite in retorters.
However, not all facilities have retorting processes.
Therefore, this may be a candidate technology for
technology transfer emphasis. Other pollution
prevention and control opportunities involve removing
mercury early in the process. In gold processing, the
valuable elements (e.g., gold and silver) are leached from
the ore and concentrated prior to being recovered in solid
form. Other impurities such as mercury can be
simultaneously leached from the rock as well, and
similarly concentrated and separated.
An additional need is to better quantify the mercury
actually present and released in order to develop a more
accurate and consistent measure of both mercury in
incoming ore and mercury released to air, water, and
solids. Additionally, waiting for more data as relatively
small gold mines are being required to report may help
clarify uncertainties. Such programs will assist in
identifying whether this particular sector should be a
priority for additional research needs, and may help to
identify facility-specific pollution prevention and control
needs. Similarly, information on the extent to which
mercury leaches from tailing piles into the environment
can help gauge the size of the potential problem.
Landfills
Unknown quantities of mercury are present in thousands
of surface landfills used for disposal of municipal and
industrial waste, however, mercury quantities are
expected be relatively high as a result of previous and
current disposal of products containing mercury. Few
options are available to address the mercury already in
the landfill, however, opportunities exist to decrease the
quantities of mercury entering landfills. As long as the
mercury remains in the landfill, its effects are much less
severe than if it migrates to air or groundwater.
Monitoring of mercury in the vented gas and of ground
water from down-gradient wells will identify any site-
specific mercury concerns, and will assist in identifying
necessary remedial actions. Such monitoring technology
already exists, although application of monitoring and
analysis requirements is on a site-specific basis.
Secondary Mercury Production
Over 400 tons per year of mercury is supplied from
secondary mercury sources to satisfy existing demand.
Opportunities to impact the quantity of mercury
produced depend entirely on mercury demand, which is
better addressed from other sectors. Recovering mercury
from scrap likely results in overall reduced releases,
since without this sector the same scrap would be
landfilled or incinerated. Current estimates of releases
from secondary mercury production are relatively low,
however, aggregate sums may increase as the number of
recovery facilities continues to increase. A more up-to-
date estimate will be available in summer 2002 when
2000 TRI data are released. However, the discrepancy
between the quantity reported in secondary mercury
facility TRIs and the aggregate recycling rate reported
among various industries is likely to remain.
Thermometers. Thermostats. Switches, and Relays
Up to 860 tons of mercury is associated with
thermostats, switches, and relays in commerce. An
additional 45-85 tons of mercury is estimated to be
associated with thermometers in commerce. These
relatively high quantities are due to a wide array of
manufactured products which utilize mercury, the
uneven application of existing mercury recycling
programs for such products, and the long life of these
products (in the case of switches, the device typically
lasts longer than the product containing the switch).
Government Stockpiles
For government stockpiles, management opportunities
are currently being studied and assessed as part of a
government-wide strategy led by the Defense Logistics
Agency. Almost 5,000 tons of elemental mercury are
stored in locations across the country as part of the U.S.
91
-------
Government Stockpiles. It is likely that stabilization,
treatment, and disposal alternatives will be identified as
potential management options. These options may
require additional research to assess their feasibility.
Utility Coal Combustion
Utility coal combustion is practiced at over 400 utility
plants nationwide and has been identified as a significant
source of mercury air emissions. USEPA (1997d)
estimates that this industry accounts for 33% of all
mercury air releases. An estimated 105 per year is
present in coal that is burned in boilers and 88 tons per
year enters the environment as air, water, or solid waste
releases (the discrepancy is probably due to different
calculation methods for consumption and release).
Mercury air releases from utility coal combustion are
approximately 54%elemental (Hg°). When compared to
other mercury species, elemental mercury generally has
the largest global impact since it can be carried over long
distances. The USEPA has published its intent to
regulate mercury-containing air releases from utility coal
combustion (65 FR 79825; December 20, 2000). To
date, mercury associated with utility coal combustion has
received the most attention compared to other sectors in
the report. Reactions have included strong public
concern, policy actions, and significant R&D efforts.
Energy conservation measures represent the most
obvious method to decrease electricity demand and
therefore decrease associated mercury emissions.
Additionally, a migration toward existing and emerging
alternative energy applications represents substantial
potential for reduction of coal use; such applications
include an array of products and design alternatives,
including passive and active solar building design, and
a variety of distributed generation technologies such as
geothermal heat pumps and fuel cells. Changes to the
coal combustion process itself may also reduce mercury
releases. These process changes could include enhanced
coal pretreatment to precipitate and capture mercury
from coal prior to combustion.
8.1.4 Prioritizing Research Needs
The estimates presented in this report act as a balance
sheet or "snap shot" of mercury in the U.S. economy.
While the data are useful for understanding relative
magnitudes, they provide little information on the
temporal trends of mercury use. Changes in the use,
release, and disposal of mercury that occur over time
from industry to industry are important when
establishing new and ongoing priorities for research and
development efforts. A valuable area for further work
includes generating updates to the baseline data
presented in this report so that trends can be identified.
Such data could help reveal important changes that can
have an influence on future mercury priorities.
Appearing from this "snap shot" is a divergence between
the quantity of mercury supplied and the quantity
demanded for use in manufacturing processes. This
imbalance may have significant ramifications with
respect to the future of secondary mercury production.
Specifically, if the recycled mercury supply consistently
and increasingly outpaces demand, the industry will
likely deteriorate as prices fall. Consequently, a
potential research priority is to analyze current and
future financial solidity of mercury recycling markets in
order to determine if intervention is necessary to
maintain viability. The secondary mercury market plays
an important role throughout the use of mercury in the
economy. Most apparent is that secondary mercury
production reduces the quantities of mercury-containing
products from being incinerated (releasing mercury into
air) and / or reaching landfills. If the industry were to
collapse, there would be significant implications for
mercury reservoir and disposal management.
Clearly, these data alone can be helpful in providing a
foundation for any prioritization of research and
development expenditures. For example, relative sector
rankings of mercury quantities may be sufficient for
high-level prioritization. Alternatively, additional
information can be overlaid on this report's data for a
more refined analysis. Relevant information could
include the presence (or absence) of regulatory drivers or
the existing level of support by EPA and other entities
going toward new technological developments in
specific sectors. Whether the data are used alone or
juxtaposed with new information, they can serve two
purposes. First, they can act as away of calling attention
to sectors deserving of research and development
prioritization. Secondly, the data can serve as a baseline
from which to project and measure the quantifiable
impact of existing and new technological and policy
developments.
While data quality is important, it should not be given
overriding emphasis in a prioritization scheme. The
reasonable accuracy of these estimates may be sufficient
for the purpose of differentiating between sectors
associated with negligible quantities of mercury and
92
-------
those associated with substantial quantities of mercury.
Accordingly, efforts to improve data should address
significant gaps or inconsistencies rather than trying to
obtain exact estimates in industries that are already
relatively well-documented. Examples of significant
data gaps include estimates of recycled mercury
available compared to the reported quantities sent to
recycling facilities or the apparent accumulation of
mercury in chlor-alkali facilities.
Lastly, these data can form the start of a foundation of
mercury use, release, monitoring and exposure data to
begin to draw connections between mercury use, release,
transport, fate, exposure and risk. This would involve
juxtaposing facility and location-specific data used in
this report with geographically-specific monitoring and
exposure data to detect patterns and relationships
between points of release and points of exposure.
93
-------
References
Note: Due to the dynamic nature of the Internet, the location and content of World Wide Web sites given in this document may change over I
time. If you find a broken link to an EPA document, use the search engine at http://www.epa.gov/ to find the document. Links to web sites I
outside the U.S. EPA web site are provided for the convenience of the user, and the U.S. EPA does not exercise any editorial control over the [
information you may find at these external web sites.
American Hospital Association (AHA). 1998.
Memorandum of Understanding with U.S.
Environmental Protection Agency. June 25,
http: //www .h2e -online .org/about/mou .htm.
Arenholt-Bindslew, D. and Larsen, A.H.. 1996.
Mercury Levels and Discharge in Waste Water from
Dental Clinics. Water Air Soil Pollution. Volume
86(1-4), pp. 93-9.
Buss, W. E., M. Merrhof, H Piduch, R. Schumacher,
and K. Ulrich. 1999. Iron and Steel Engineer,
http://www.aise.org.
ChemExpo. 2000. Chlorine Chemical Profile.
September, http://www.chemexpo .com.
Chlorine Institute. 2001. Fourth Annual Report of the
Chlorine Institute to the United States Environmental
Protection Agency, http://www.cl2.com.
—. 2000. Guidelines for Technologies to Reduce
Mercury in Sodium Hydroxide. April.
Citizens for Environmental Health. 2002. Health Care
Without Harm.
http: //family .knick .net/cectoxic/hcwh .htm.
Council of Great Lakes Industries (CGLI). 1999.
Great Lakes Binational Toxics Strategy Success
Stories, http://www.cgli.org/success.html.
Council of Industrial Boiler Owners (CIBO). 1997.
Report to the U.S. Environmental Protection Agency
on Fossil Fuel Combustion Byproducts from Fluidized
Bed Boilers.
Dental Amalgam Mercury Syndrome Inc. and
Consumers for Dental Choice (DAMS). 1999.
Mercury Free and Healthy: The Dental Amalgam
Issue. Prepared for the National Institute of Science,
Law, and Public Policy Washington D.C., August,
http://www.amalgam.org.
Defense Logistics Agency (DLA). 2002. Mercury
Management Environmental Impact Statement.
www.mercuryeis.com.
Dungan, A. 1999. Phone conversation. Vice
President - Safety, Health, and Environment,
Chlorine Institute, November 2.
Elko Daily Free Press. 2000a. Report: Mines Big
Source of Mercury Pollution. Elko Daily Free Press
(Nevada), March 2.
—. "Clarification Sought on Mercury Emissions."
2000b. Elko Daily Free Press (Nevada), March 7.
Fisher Scientific. 1999. Fisher and ERTCO Mercury
Exchange Program.
http://www.fishersci.com.
Genna, A. 1998. Will the Buyers Market Continue?
Purchasing Online. Dec. 10,
http://www.manufacturing.net/magzinearchives/1998/
purl210.98/121chem2.htm.
Great Lakes National Wildlife Federation. 1997. Press
Release. Ohio's Coal-Burning Power Plants Poison
Fish, Rivers. December 11,
http: //www .nwf. org/greatlake s/.
Great Lakes United (GLU). 2001. Mercury in Cars
Project. January 2001,
http: //www. glu. org/cptf/mainpage .htm.
94
-------
Hanisch, C. 1998. "Where is Mercury Deposition
Coming From?" Environmental Science and
Technology. April 1, pp. 176A - 179A.
Johnson, J. 1999. "Olin to End Hg releases from
Chlor-Alkali Plants." C&EN. Apr. 8.
Keller, V. 1997. An 1870s Iron Making Primer.
http://www.usaor.net/users/vagelk/primer.htm
Kennedy, R., M.D. 1996. "The Thinking Person's
Guide to Perfect Health: Dental Amalgam Mercury
Poisoning."
http: //www. sonic .net/~nexus/amalgam .html
Kirk-Othmer Encyclopedia of Chemical Technology.
1996. "Pulp." Volume 20, Fourth edition.
—. 1995. "Lime and Limestone." Volume 15, Fourth
edition.
—. 1994. "Geothermal Energy." Volume 12, Fourth
edition.
—. 1993. "Coal Conversion Processes
(Carbonization)." Volume 6, Fourth edition.
—. 1991. "Alkali and Chlorine Products." Volume 1,
Fourth edition.
Kim 2001. Kim, Christopher S.; Lowry, Greg V.;
Shaw, Samuel; Rytuba, James J.; Brown Jr., Gordon E.
Speciation of Mercury-Bearing Mine Wastes Using X-
ray Absorption Fine Structure (XAFS) Spectroscopy:
Physical and Geochemical Effects. Eleventh Annual
V.M. Goldschmidt Conference. Hot Springs Virginia.
May, http://www.lpi.usra.edu/meetings/gold2001.
Krishnan, S. V., Gullett, B. K., Jozewicz, W. 1994.
Sorption of Elemental Mercury by Activated Carbons.
Environ. Sci. Technol. 28(8), pp. 1506-12.
Lawrence, B. 2000. The Mercury Marketplace:
Sources, Demand, Price, and the Impacts of
Environmental Regulation. Bethlehem Apparatus Co.
Presentation at Workshop on Mercury in Products,
Processes, Waste and the Environment. March 22-23,
2000 Baltimore Maryland.
Maine 2000. Mercury Management in the Health Care
Environment. Department of Environmental
Protection. October.
http: //www. state .me .us/dep/rwm/mercurymedical .htm
Massachusetts Department of Environmental
Protection. 1996. Mercury in Massachusetts: An
Evaluation of Sources, Emissions, Impacts, and
Controls. June.
Menne, D. 1998. Mercury in Gold Metallurgy. June
14, Perth Australia.
Michigan Mercury Pollution Task Force (M2P2).
1996. Mercury Pollution Prevention in Michigan,
April.
Minnesota. 1999. Mercury in Petroleum Refining;
Crude Oil and Refined Products. Final Report to
Legislative Commission on Minnesota Resources,
August 20.
National Electrical Manufacturers Association
(NEMA). 2000. Environmental Impact Analysis:
Spent Mercury-Containing Lamps, (fourth edition).
January, http://www.nema.org.
—. 1999. "Domestic Shipments."
http: //www .nema.org
North Carolina Department of Environment, Health,
and Natural Resources (NC DEHNR). 1996. Waste
Reduction and Proper Waste Management of Products
Containing Mercury.
http://www.p2pavs.0rg/ref/01/00127.htm
Paur, H.; Buchele, H.; Schrader, C.; Bolin, P.;
Winkler, W. Removal and Measurement of Mercury
in the Thermal Treatment of Wastes. 1999. Joint
Power Generation Conference, Volume 1, 443- 8.
Portland Cement Association (1999). Homepage.
http: //www .portcement.org.
Reese, R. 1999. Personal communication.
Geological Survey. November 9.
U.S.
Rubin, P.O., and M.H. Yu. 1996. Mercury Vapor in
Amalgam Waste Discharged from Dental Office
Vacuum Units. Archives of Environmental Health.
July-August; Volume 51 (4), 335-7.
95
-------
Sang, S., and B. Lourie. 1995. Mercury in Ontario:
An Inventory of Sources, Uses, and Releases."
Proceedings of 1995 Canadian Mercury Network
Workshop.
Spallation Neutron Source (SNS). 2002.
http: //www .sns.gov.
Sylvania. 1998. OSTREUM Sylvania Announces
Two Major Product Initiatives Under ECOLOGIC
Banner at Lightfair 1998. Press Release, May 21.
United States Bureau of Mines (USBM). 1994. The
Materials Flow of Mercury in the United States.
Information Circular 9412.
United States Department of Commerce, Bureau of
Census (USDOC). 2001. States Ranked by
Population: 2000. http://www.census.gov.
—. 1998. Current Industrial Report: Selected
Instruments and Related Products.
http: //www .census. gov/mp/www/pub/cir/mscir 17a.htm
1.
—. 1995. Current Industrial Reports: MQ36B
Electrical Lamps, 1994.
United States Department of Energy (USDOE). 2002.
U.S. Coal Consumption by End-Use Sector, 1995 to
2001. Energy Information Administration.
http://www.eia.doe.gov.
—. 2001. Petroleum Supply Annual 2000. DOE/EIA-
0340(00)/1. June 2001. Energy Information
Administration.
United States Department of Health and Human
Services, Public Health Service (USDHHS). 1997.
Dental Amalgam and Alternative Restorative
Materials. October.
—. 1993. Appendix III: Evaluation of Risks
Associated with Mercury Vapor from Dental
Amalgam. Dental Amalgam: A Scientific Review and
Recommended Public Health Service Strategy for
Research, Education and Regulation: Final Report of
the Subcommittee on Risk Management of the
Committee to Coordinate Environmental Health and
Related Programs. January,
http://web.health.gov/environment/amalgaml/ct.htm.
United States Environmental Protection Agency
(USEPA). 2001. U.S. Environmental Protection
Agency. Results of Part III of Information Collection
Request. June,
http: //www. epa. gov/ttnatwO 1 /combust/utiltox/utoxpg .h
tml.
—. 2000a. Listing Background Document for the
Chlorinated Aliphatics Listing Determination. June.
—. 2000b. Background Document for Capacity
Analysis for Land Disposal Restrictions: Newly
Identified Chlorinated Aliphatics Production Wastes
(final rule). September.
—. 1999a. Waste Minimization: Reducing Mercury
Use, Philips Lighting. OSWER EPA530-F-99-041.
June.
—. 1999b. Nationwide Generation of Mercury
Bearing Wastes: Pollution Prevention Analysis and
Technical Report. Office of Solid Waste, September.
—. 1999c. Environmental Fact Sheet: Some Used
Lamps are Universal Wastes. Office of Solid Waste,
July, EPA 530-F-99-024.
—. 1999d. Report to Congress: Wastes from the
Combustion of Fossil Fuels, Volume 2 - Methods,
Findings, and Recommendations, EPA 530-R-99-010,
Office of Solid Waste and Emergency Response,
Washington, DC, March.
—. 1998a. Waste Specific Evaluation of RMERC
Treatment Standard. July 6.
—. 1998b. Analysis of Alternatives to Incineration
for Organic-Mercury Wastes. July.
—. 1998c. Study of Hazardous Air Pollutant
Emissions from Electric Utility Steam Generating
Units. Final Report to Congress. Office of Air
Quality Planning and Standards. EPA-453/R-98-004.
—. 1998d. National Hazardous Waste Constituent
Survey. October.
96
-------
—. 1997a. Mercury Study Report to Congress.
Volume II: An Inventory of Anthropogenic Mercury
Emissions in the United States, EPA 452-R-97-004.
Office of Air Quality Planning and Standards and
Office of Research and Development. December.
—. 1997b. Locating and Estimating Air Emissions
from Sources of Mercury and Mercury Compounds,
EPA 454/R-97-012. Office of Air Quality Planning
and Standards, December.
—. 1994. Mercury Usage and Alternatives in the
Electrical and Electronics Industries.
Office of Research and Development, EPA/600/R-
94/047, January.
—. 1993a. Locating and Estimating Air Emissions
from Sources of Mercury and Mercury Compounds,
EPA/454/R-93-023. Research Triangle Park, NC.
—. 1993b.
December.
Report to Congress on Cement Kiln Dust.
—. 1992a. Management of Used Fluorescent Lamps:
Preliminary Risk Assessment. Submitted by Robert
Truesdale, Stephen Beaulieu, and Terrence Pierson.
Research Triangle Institute, Office of Solid Waste,
October.
—. 1992b. Characterization of Products Containing
Mercury in Municipal Solid Waste in the United
States, 1970 to 2000. Submitted by Franklin
Associates. Office of Solid Waste, January.
—. 1990. U.S. EPA 1990 National Interim Emissions
Inventory. Electronic database.
—. 1988a. Report to Congress: Wastes from the
Combustion of Coal by Electric Utility Power Plants.
EPA/530-SW-88-002, February.
—. 1988b. Case Studies of Mercury-Bearing Wastes
at a Mercury Cell Chlor-Alkali Plant. EPA/600/2-
88/011.
—. 1985. Domestic Sewage Study. Draft Final
Report Volume 1: Technical Report. December.
United States Food and Drug Administration
(USFDA). 1999. Mercury Compounds in Drugs and
Food. November,
http://www.fda.gov/cder/index.html.
United States Geological Survey (USGS). 2002a.
Mercury - Mineral Commodity Summaries. January.
—. 2002b. Mineral Commodity Summaries - Lime.
January.
—. 2000a. Minerals Yearbook- Zinc - 2000.
—. 2000b. Mercury Contamination from Historic
Gold Mining in California. USGS Fact Sheet FS-061-
00, May.
—. 2000c. Minerals Yearbook- Lead - 2000.
—. 2000d. Minerals Yearbook - Copper. 2000.
—. 2000e. The Materials Flow of Mercury in the
Economies of the United States and the World.
USGS Circular 1197. June.
—. 1998. Mercury in US Coal, USGS Open-File
Report 98-0772. August.
—. Mineral Industry Surveys: Mercury. 1990-1997.
Valenti, M. 2000. Rx for Medical Waste. Mechanical
Engineering. September, http://www.memagazine.org.
Western Lake Superior Sanitary District (WLSSD).
1997. Blueprint for Mercury Elimination, updated
May 12,
http://www.wlssd.duluth.mn.us/blueprint%20for%20m
ercury/HG 1 .htm.
Yoshida, M.; Kishimoto, T.; Yamamura, Y.; Tabuse.
M.; Akama, Y.; and H. Satoh. 1994. "Amount of
mercury from dental amalgam filling released into the
atmosphere by cremation." Nippon Koshu Eisei
Zasshi. July, Volume 41(7), 618 - 24, ISSN:
0546-1766.
97
------- |