Mercury Usage and Alternatives in the Electrical and Electronics IndustriesMERCURY USAGE AND ALTERNATIVES IN THE ELECTRICAL AND ELECTRONICS INDUSTRIES
Bruce M. Sass, Mona A. Salem, and Lawrence A. Smith Battelle
Columbus, Ohio 43201
Project Officer
Risk Reduction Engineering Laboratory Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY CINCINNATI, OHIO 45268
NOTICE
This material has been funded wholly or in part by the U.S. Environmental Protection Agency (EPA) under Contract No. 68-C0-0003 to Battelle. It has been subjected to the Agency’s peer and administrative review and approved for publication as an EPA document. Approval does not signify that the contents necessarily reflect the views and policies of the EPA or Battelle; nor does mention of trade names or commercial products constitute endorsement or recommendation for use. This document is intended as advisory guidance only to the electrical and electronics industries in developing approaches to pollution prevention. Compliance with environmental and occupational safety and health laws is the responsibility of each individual business and is not the focus of this document.
ii
FOREWORD
Today’s rapidly developing and changing technologies and industrial products and practices frequently carry with them the increased generation of materials that, if improperly dealt with, can threaten both public health and the environment. 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. These laws direct EPA to perform research to define our environmental problems, measure the impacts, and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for planning, implementing, and managing research, development, and demonstration programs to provide an authoritative, defensible engineering basis in support of the policies, programs, and regulations of EPA with respect to drinking water, wastewater, pesticides, toxic substances, solid and hazardous wastes. Superfund-related activities, and pollution prevention. This publication is one of the products of that research and provides a vital communication link between the researcher and the user community.
Passage of the Pollution Prevention Act of 1990 marked a significant change in the U.S. policies concerning the generation of hazardous and nonhazardous wastes. This bill implements the national objective of pollution prevention by establishing a source reduction program at the EPA and by assisting States in providing information and technical assistance regarding source reduction. In support of the emphasis on pollution prevention, projects have been designed, with the coordina- tion and cooperation of the Office of Solid Waste (OSW), to identify and evaluate source reduction and recycling options for selected RCRA wastestreams. This report describes the current usage of mercury as well as alternative technologies to reduce mercury use and disposal in the electronics industry.
E. Timothy Oppelt, Director Risk Reduction Engineering Laboratory
. . .III
ABSTRACT
Many industries have already found alternatives for mercury or have greatly decreased mercury use. However, the unique electromechanical and photoelectric properties of mercury and mercury compounds have made replacement of mercury difficult in some applications. This study was initiated to identify source reduction and recycling options for mercury in the electrical and electronics industries (SIC 36) and measurement and control instrument manufacture (SIC 382). The project identified trends in pollution prevention for mercury use throughout the U.S. economy by a review of the sources and use of mercury in the economy. Regulatory trends encouraging mercury pollution prevention were examined, and current practices in the electrical and electronics industries were reviewed in detail to identify potential source reduction and reuse options for mercury. Industrial and economic data suggest that the quantity of mercury used in electrical and electronic control and switching devices is significant. Opportunities have been identified to replace mercury-containing devices. For applications where mercury cannot be avoided, recycling, mainly by vacuum retorting, is commercially available.
This report was submitted in partial fulfillment of Contract Number 68-C0-0003, Work Assignment 3-36 under the sponsorship of the U.S. Environmental Protection Agency. This report covers a period from August, 1992, to December, 1993, and the study was completed as of January 31, 1994.
iv
CONTENTS
SECTION 1: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
SECTION 2: Technical Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Literature Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Technical Associations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Academia.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Battelle Staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Conferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Site Visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
SECTION 3: Mercury Economic Data and Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Historical Uses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Recent Mercury Usage Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Mercury Cell Chloralkali Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Batteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Switching Devices and Control Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Electrical Lamps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Mercury-Cadmium-Telluride Semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Paints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Catalysts.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
State and Federal Regulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Mercury Treatment Standards Under RCRA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 State Regulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
SECTION 4: Source Reduction Alternatives for Mercury in the Electrical and Electronics Industries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Electrical Lighting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Batteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Switching Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Mercury Electronic Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Silent Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Reed Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Proximity Sensors and Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Control Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Thermostats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Mercury Switch Thermostats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Thermostat Market Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Non-Mercury Switch Thermostats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Mercury-Cadmium-TellurideSemiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 MCT Alternative Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
SECTION 5: Recycling Alternatives for Mercury in the Electronics Industry . . . . . . . . . . . . . . 33 Industry Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Recycling Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
SECTION 6: Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
SECTION 7: References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
10 11 12
Mercury consumption in the United States, by use . . . . . . . . . . . . . . . . . . . . . . . . . 7 U.S. mercury consumption in electronic products (metric tons), 1980-1992 . . . . . . . . 9 Discards of mercury in products in the municipal solid wastestream, 1970 to 2000 (in short tons) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Typical composition range for K106 nonwastewater . . . . . . . . . . . . . . . . . . . . . . . . 10 Examples of dry cell-type batteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Comparison between the mercury switch and its alternatives . . . . . . . . . . . . . . . . . . 21 Market drivers governing thermostat sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Comparison between the mercury switch thermostat and its alternatives . . . . . . . . . . 27
FIGURES
U.S. mercury consumption in electronic products, 1980-1992 . . . . . . . . . . . . . . . . 8 Typical packaged high-speed optical switch has electrical input ports and output fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Typical bimetal shapes: (a) strip, (b) coil, (c) U-shape, (d) spiral . . . . . . . . . . . . . . . . 23 Typical diaphragm forms used for temperature sensing . . . . . . . . . . . . . . . . . . . . . . 23 Mercury tilt switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Honeywell T87 thermostat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Mechanical bimetallic snap switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Diagram of magnetic snap switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Hot-wire vacuum switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Typical fluorescent lamp recovery processing flowchart . . . . . . . . . . . . . . . . . . . . . . 34 Vacuum retort for mercury recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Reverse distribution scenarios for recycling of thermostats . . . . . . . . . . . . . . . . . . . . 38
vi
ACKNOWLEDGMENTS
This report was prepared under the direction and coordination of Paul M. Randall of the U.S. Environmental Protection Agency (EPA), Office of Research and Development, Risk Reduction Engineering Laboratory, Pollution Prevention Branch, in Cincinnati, Ohio. Technical review was provided by
Dr. David Allen Professor, UCLA Los Angeles, California
Mr. Rodney Everett Marketing Manager, HVAC General Electric Company Morrison, Illinois
Mr. S. Garry Howell U.S. Environmental Protection Agency Risk Reduction Research Engineering Laboratory Pollution Prevention Research Branch Cincinnati, Ohio
Mr. Steve Keefe Director, State Government Affairs Honeywell, Inc. Minneapolis, Minnesota
Mr. Richard Robinson National Electrical Manufacturers Association Washington, D.C.
Mr. Ronald J. Turner U.S. Environmental Protection Agency Risk Reduction Engineering Laboratory Physical/Chemical Separations Branch Cincinnati, Ohio
vii
INTRODUCTION
This study was conducted as part of the U.S. Environmental Protection Agency’s (EPA) effort to develop pollution prevention options for RCRA wastestreams that have been difficult or expensive to treat. Pollution prevention is the use of materials, processes, or practices that reduce or eliminate the creation of pollutants or wastes. Pollution prevention should be considered the first step in a hierarchy of options for reducing the generation of pollution. The next step in the hierarchy is responsible recycling of any wastes that cannot be reduced or eliminated at the source. Wastes that cannot be recycled should be treated in accordance with environmental standards. Finally, any wastes that remain after treatment should be disposed of safely.
The objective of the study and this resulting report has been to identify source reduction and recycling options for mercury in the electronics industry. To accomplish this objective, the sources and use of mercury in the U.S. economy were reviewed and regulatory trends encouraging mercury pollution prevention were examined to provide a background for a detailed review of the electronics industry. Current practices in the electrical and electronics industries (SIC 36) and measurement instrument and control instrument and control instrument manufacture (SIC 382) were reviewed in detail to identify potential source reduction and reuse options for mercury. Industrial and economic data suggest that the quantity of mercury used in electronic control and switching devices is significant. Some opportunities were identified to replace mercury-containing devices. It was found that recycling of mercury, mainly by vacuum retorting, is becoming commercially available for some electronic components.
The steady decline in mercury consumption in the United States is well documented in previous studies on mercury usage in batteries and fluorescent lamps; however, details of how the electronics industry had reduced its need for mercury and what new technologies were involved were not well known. The data collected identify possible approaches to reduce mercury use and increase recycling in the subject industries.
Although mercury was known to be toxic for many centuries, the level of health hazard has come to light only since the 1970s. Metallic mercury, its vapor, and many of its compounds are protoplasmic poisons, which are toxic to all forms of life. Ingesting sufficient quantities, by mouth, through the skin, or by inhalation, can cause severe neurological damage and fatality in humans (Budavari, 1989). The alkyl organic compounds are the most toxic forms of mercury. Alkyl mercury compounds, such as dimethylmercury, are used as intermediates in some chemical processes. It is now known that some marine organisms can biologically methylate inorganic mercury and concentrate it up to 3,000 times.
Mercury-containing RCRA wastes are difficult to treat reliably by conventional techniques such as solidification/stabilization. This project was undertaken, with the coordination and cooper- ation of the Office of Solid Waste, to help define pollution prevention technologies for mercury- containing RCRA problem wastes.
Reduction of mercury disposal is also promoted by the 33/50 Program. The 33/50 Program is EPA’s voluntary pollution prevention initiative to reduce national pollution releases and off-site transfers of 17 toxic chemicals by 33% by the end of 1992 and by 50% by the end of 1995. EPA is asking companies to examine their own industrial processes to identify and
1
Implement cost-effective pollution prevention practices for these chemicals. Company participation in the 33/50 Program is completely voluntary. The Program aims, through voluntary pollution prevention activities, to reduce releases and off-site transfers of a targeted set of 17 chemicals from a national total of 1.4 billion pounds in 1988 to 700 million pounds by 1995, a 50% overall reduction. The Toxics Release Inventory (TRI) (established by federal law, the Emergency Planning and Community Right-to-Know Act of 1986) will be used to track these reductions using 1988 data as a baseline. As required by the Pollution Prevention Act of 1990, TRI industrial reporting requirements were to be expanded, beginning in calendar year 1991, to include information on pollution prevention.
2
SECTION 2
TECHNICAL APPROACH
The study supplemented literature data sources with industry and academic sources to give insight into current uses for mercury and to identify and evaluate practical alternatives to reduce mercury use.
LITERATURE SEARCH
Battelle performed an extensive literature search for information on the use of mercury and its alternatives in the electronics industry. Technical journals were utilized to obtain the necessary information. A literature search was performed at the Electronic Industries Association. The U.S. Bureau of Mines database was searched for information quantifying the production and consumption of mercury. The Electric Power Research Institute (EPRI) database, the Alternative Treatment Technology Information Center (ATTIC), and the Pollution Prevention Information Exchange System (PIES) also were employed.
The ATTIC network is maintained by the Technical Support Branch of EPA’s Risk Reduction Engineering Laboratory (RREL). This network has four online databases that can be searched by external users.
ATTIC Database. Contains abstracts and bibliographic citations to technical reports, bulletins, and other publications produced by EPA, other federal and state agencies, and industry dealing with technologies for treatment of hazardous wastes. Performance and cost data, quality assurance information, and a contact name and phone number are given for the technologies.
Risk Reduction Engineering Laboratory (RRELI Treatability Database. Provides information about contaminants - physicochemical properties, environmental data, treatment technologies, contaminant concentration, media or matrix, performance, and quality assurance.
Technical Assistance Directory. Lists experts from government, universities, and consulting firms who can provide guidance on technical issues or policy questions.
Calendar of Events. Extensive list of conferences, seminars, and workshops on treatment of hazardous wastes. International as well as U.S. events are covered.
Two other databases are available through a system operator. The Robert S. Kerr Environmental Research Laboratory Soil Transport and Fate Database deals with the movement and fate of contaminants in soil matrices. The Hazardous Waste Collection Database is a collection of reports,
3
commercially published books, and directives and legislation on hazardous waste. There is no charge for the ATTIC service. It is available via modem over standard telephone lines. The phone number for the ATTIC modem contact is (903) 908-2138 (1200 or 2400 baud), and the modem settings are no parity, 8 data bits, 1 stop bit, and full duplex. The system operator for ATTIC can be reached at (908) 321-6677 or by fax at (908) 906-6990. The user’s manual (U.S. EPA, ND) also is available.
PIES is a bulletin board system that links to several databases and provides messaging capabilities and forums on various topics related to pollution prevention. Through its link to the United Nation’s International Cleaner Production Information Clearinghouse, it provides a communi- cation link with international users. PIES is part of the Pollution Prevention Information Center (PPIC), which is supported by EPA’s Office of Environmental Engineering and Technology Demon- stration and Office of Pollution Prevention and Toxics. PIES contains information about current events and recent publications relating to pollution prevention. Summaries of federal, state, and corporate pollution prevention programs are provided. The two sections of the database cover case studies and general publications and can be searched by keywords related to specific contaminants, pollution prevention technologies, or industries. The phone number for dial-up access is (703) 506-1025; qualified state and local officials can obtain a toll-free number by calling PPIC at (703) 821-4800. Modem settings are 2400 baud, no parity, 8 data bits, 1 stop bit, and full duplex. The system operator for PIES can be reached at (703) 821-4800.
TECHNICAL ASSOCIATIONS
Battelle contacted the following organizations to obtain data on electrical and electronics industry practices:
American Electronics Institute (AEl) Chemical Manufacturers Association (CMA) American Institute of Pollution Prevention (AIPP) Electronic Industries Association (EIA) Electric Power Research Institute (EPRI) Instrument Society of America (ISA) National Electrical Manufacturers Association (NEMA) Institute of Electrical and Electronics Engineers, Inc. (IEEE) American Electronics Association (AEA).
ACADEMIA
Battelle contacted a number of universities to gather information on current research on alternatives to mercury use. The Electrical Engineering Departments at the University of Illinois and the University of Missouri were particularly responsive to the information request.
INDUSTRY
Battelle solicited many industry members through phone calls and questionnaires, shown in the Appendix. The companies contacted were General Electric Company, AT&T, Digital Equip- ment Corporation, Honeywell, Microswitch, CP Clare, Motorola, Thomson CSF, Lutron, Alph
4
International, Leviton, Philips Lighting Company, Hamlen, and Sylvania. Battelle also contacted waste exchanges in an effort to identify reuse options for mercury-bearing wastes.
BATTELLE STAFF
The Product Design and Engineering and the Electronic Systems Operations staff at Battelle provided input on the operational characteristics of current mercury-containing electronic devices and identified alternative technologies to mercury use.
CONFERENCES
Battelle staff attended the National Conference on Minimization and Recycling of Industrial and Hazardous Waste 92, held in Arlington, Virginia, September 22-24, 1992, and the First IEEE International Symposium on Electronics and the Environment, held in Arlington, Virginia, May 10-12, 1993, to review current research into mercury-free or reduced-mercury-content electrical, electronic, instrument, or control options.
SITE VISITS
Battelle staff visited Honeywell’s thermostat manufacturing facilities in Minneapolis, Minnesota. Honeywell, other thermostat manufacturers, and the National Electrical Manufacturers Association have worked with the Minnesota Pollution Control Agency to initiate a recycling program for all brands of mercury-switch thermostats. Battelle staff also visited Honeywell’s recycling pilot facility, also located in Minneapolis.
5
MERCURY ECONOMIC DATA AND REGULATION
Mercury is a shiny metal and is the only common element that is liquid at room tempera- ture. The chemical symbol, Hg, is derived from the Greek word hydrargyrum, meaning “liquid sil- ver.” Mercury’s atomic number 80, and its atomic weight is 200.59. Mercury has a high density (13.6 g/cm3).
Mercury is an unreactive, corrosion-resistant metal which melts at -38.87OC (-37.97OF) and boils at 356.58OC (673.84OF). When heated to near its boiling point, mercury oxidizes in air to form mercuric oxide (HgO). Mercuric oxide decomposes at 500°C (930°F), releasing oxygen and forming mercury metal.
Mercury is estimated to occur in concentrations of 0.010 to 0.3 mg/kg in typical soils (Swartzburg et al., 1992 and 1993). By contrast, ore-grade materials average about 0.5% mercury content.
The primary source of mercury is the sulfide ore, cinnabar. In a few cases, mercury occurs as the principal ore product. Mercury is more commonly obtained as the by-product of pro- cessing complex ores that contain mixed sulfides, oxides, and chloride minerals, which are usually associated with base and precious metals, particularly gold. Native, or metallic, mercury, is found in very small quantities in some ore sites.
Mercury can be recovered from its ores by relatively simple methods. Some early methods for purification included leaching the ores in sodium sulfide and sodium hydroxide solu- tions or in a sodium hypochlorite solution. Today, mercury is recovered from the sulfide ores or secondary sources by a high-temperature retorting process. The ore is ground and heated to about 580°C (1080°F) in the presence of oxygen. The sulfide ore decomposes to form mercury vapor and sulfur dioxide. The mercury vapor is condensed and washed with nitric acid. The cleaned mercury is further purified by single or triple distillation, depending on the grade required.
HISTORICAL USES
Mercury was among the first metals to be identified and used as native metal. Archaeol- ogists found mercury in an Egyptian tomb dating from 1500 BC. Mercury compounds also have been in use since early times. The Egyptians and the Chinese are believed to have used cinnabar as a red pigment. The Greeks used mercury as a medicine.
RECENT MERCURY USAGE PATTERNS
Mercury for domestic use in 1990 came from domestic mines, sales of surplus from government stocks, imports, and waste recovery. Mercury was produced as the main product of the McDermitt Mine and as a by-product of eight gold mines in Nevada, California, and Utah. The McDermitt Mine has since been closed (U.S. Bureau of Mines, 1993). Market expectations indicate
6
a continuing decline in mercury use and increased reliance on recycled mercury (Greenberg et al., 1993).
Smaller amounts of mercury are produced when secondary sources are reprocessed. In 1992, commercial secondary mercury reprocessors produced 176 metric tons of mercury (U.S. Bureau of Mines, 1993). Common secondary mercury sources include spent batteries, mercury vapor and fluorescent lamps, switches, dental amalgams, measuring devices, control instruments, and laboratory and electrolytic refining wastes. The secondary processors typically use high- temperature retorting to recover mercury from compounds and distillation to purify the contam- inated liquid mercury metal.
The main use areas for mercury are chemical production, particularly chlorine/caustic manufacture; electrical and electronic components; and instruments and related products. Recent mercury use patterns are indicated by Table 1. As shown in the table, both the supply and the demand for mercury have declined in response to regulatory pressures particularly in paints and chemicals. More detail on the annual use in electrical and electronics applications is shown in Figure 1 and in Table 2. Note that in Figure 1, data for Wiring Devices and Switches, Measuring and Control Instruments, and Other Electrical and Electronic and Other Instruments (after 1987) have been combined in bar chart format. This was done because there is some ambiguity regarding how specific devices may have been placed in any one of these three Standard Industrial Classification (SIC) categories. The combined data roughly represent mercury consumption in all electrical, electronic, and instrument applications exclusive of electrical lighting and batteries. Overall, the data in Table 2 and Figure 1 suggest that mercury usage has declined over the past decade, but aside from batteries, usage in electrical and electronic devices has remained fairly constant.
TABLE 1. MERCURY CONSUMPTION IN THE UNITED STATES, BY USE
Use Use in 1989 Use in 1992
(MT)“’ (MT)‘“’
Mercury cell chloralkali process Laboratory uses Paint Other chemical related uses
Electrical and Electronics
Instruments and Related Products
Other
Total
87 52 39 37
32 148
1,212 621
(a) MT = metric ton (1 MT is equivalent to 1000 kg, 2,205 lb, 1,102 short tons, and 29 flasks).
Source: U.S. Bureau of Mines (1993).
7
TABLE 2. U.S. MERCURY CONSUMPTION IN ELECTRONIC PRODUCTS (METRIC TONS), 1980-1992
Year SIC
Measuring and Other Electrical and Electrical Wiring Devices Control Electronic, Other Lighting and Switches Batteries Instruments Instruments”’
3641 3643 3692 382 Other 1980 36 106 960
1981 36 91 1016
1982 28 69 858
1983 44 80 806
1984 51 94 1025
1985 40 95 953
1986 41 103 751
1987 45 131 533
1988 31 176 448
1989 31 141 250
1990 33 70 106
1991 29 25 18
1992 55 69 16
(a) Data revised from earlier reports in U.S. Bureau of Mines 1993 report. (b) Revised data not available before 1988. Source: U.S. Bureau of Mines (1993).
105 (b)
196 (b)
106 (b)
85 (b)
99 (b)
79 (b)
63 ( b )
59 (b)
77 55
87 32
108 25
70 165
52 148
Table 3 shows the estimated disposal of mercury in products in the municipal solid wastestream. Household batteries are the major source of mercury in the municipal solid waste- stream. Electrical lighting components comprise the second largest source of mercury in municipal solid waste. These numbers correspond with the mercury consumption trend.
Mercury Cell Chloralkali Process
Production of chlorine gas and caustic soda accounts for most mercury used in the United States. The manufacturing process also is responsible for the largest loss of mercury into the environment. One process uses mercury and a cathode in an electrolytic cell into which sodium chloride brine is introduced. A current is applied to electrolytically oxidize chloride anions to form chlorine gas (Cl,), which is collected at the anode, and an alkali metal amalgam is formed with the mercury cathode. The amalgam is then decomposed with water to form caustic soda (sodium hydroxide), hydrogen, and relatively pure mercury metal. Although mercury metal is recycled back to the cell, large losses occur in brine purification muds and in wastewater treatment sludges. The brine sludge contains small amounts of mercuric ions, mostly as the tetrachloro complex, HgCl4
2
9
TABLE 3. DISCARDS’“’ OF MERCURY IN PRODUCTS IN THE MUNICIPAL SOLID WASTESTREAM, 1970 TO 2000 (IN SHORT TONS[b])
Products 1970 1980 1989 2000
Household batteries
0.4 1.9
0.0 0.0
709.0 172.7
(a) Discards before recovery. (b) Weights in this report are converted to short tons of 2,000 pounds. Source: U.S. EPA, 1992a, EPA/530-R-92-013.
A mercury-bearing sludge results from treatment of effluents from electrolytic processing to generate chlorine gas and sodium hydroxide. This sludge is a Resource Conservation and Recovery Act (RCRA)-listed waste with the waste code K106. A typical composition range for K106 nonwastewater is shown in Table 4 (Dungan, 1992).
TABLE 4. TYPICAL COMPOSITION RANGE FOR K106 NONWASTEWATER
Component Composition
Inorganic salts, mainly 3 to 15% (dry basis) sodium chloride
Water 24 to 57%
Bat te r ies
Batteries make up a significant but decreasing use of mercury. Mercury historically has been used to coat the zinc anode (negative electrode) in nonrechargeable household batteries. A few examples of these dry cell-type batteries are listed in Table 5.
Mercury is used to prevent the evolution of hydrogen gas from the battery, which results from internal chemical reactions. Hydrogen may pressurize the cell and cause internal leaking or explosion; blowout valves are now installed at the tips to minimize this possibility. In the alkaline- manganese battery, zinc anodic material is added as a powder. In the past, 1 to 3% mercury was mixed with the powdered zinc to form a mercury-zinc amalgam that inhibits zinc oxidation caused by chemical reactions with other components in the battery. The proportion of mercury in the amalgam decreases the rate of oxidation. Because alternatives to mercury have been identified by the battery manufacturing industry, mercury use in this industry is declining.
Switching Devices and Control Instruments
Mercury is used in both high-voltage and low-voltage mercury-arc rectifiers, oscillators, power control switches for motors, phanatrons, thyratrons, ignitrons, reed switches, silent switch- es, thermostats, and cathode tubes in radios, radar, and telecommunications equipment. Current rectifiers use electron tubes that consist of a metal, ceramic, or glass shell containing electrodes that maintain and control current flow. Electron tubes are used to generate, rectify, amplify, or convert electrical signals. Electron tubes are classified as vacuum and gas-filled tubes. In practice, the distinction is not absolute as the degree of vacuum and the amount and type of gas may vary widely. In general, gas-filled tubes permit higher currents than do vacuum tubes due to ionization of mercury vapor in the tube.
Mercury also is used in many medical and industrial instruments to control or measure reactions and equipment functions. This list includes mostly metallic mercury equipment, such as thermometers, manometers, barometers, and other pressure-sensing devices, gauges, valves, seals, and controls. The calomel (mercurous chloride) electrode commonly is used in conjunction with glass electrodes to measure hydrogen ion (pH) and other ion activities.
Electrical Lamps
Mercury vapor is used in both low-pressure “fluorescent” lamps and high-pressure mercu- ry lamps. Fluorescent lamps commonly are used for indoor lighting, whereas high-pressure mercury
TABLE 5. EXAMPLES OF DRY CELL-TYPE BATTERIES
Battery Type Size/Configuration
Alkaline-manganese dioxide AAA, AA, C, D, 9V
Mercuric oxide Button ceils (hearing aids)
Zinc-silver oxide Button cells
11
lamps are used for street lighting, industrial work areas, aircraft hangers, and floodlighting. Other mercury-vapor lamps are used for photographic purposes, including motion picture projection, and for heat therapy.
Mercury-Cadmium-Telluride Semiconductors
Mercury-cadmium-telluride (MCT) is an important semiconducting material used for infrared (IR) detection. Its photoconductive and photovoltaic properties rival those of more mature Ill-V semiconductors such as GaAs, Gap, and InSb. Other potential uses for MCT are in IR lasers (Mahavadi et al., 1990) and y-ray detectors (Mullin et al., 1985). The bulk composition of the MCT alloy is Hg1-xCdxTe (0</- x </-1). The compositional parameter x can be varied indefinitely to optimize detector characteristics, which is done particularly in the 8 to 14 um wavelength range (Irvine and Mullin, 1981) and in the 3 to 5 um range (Korenstein et al., 1990).
Pain ts
Mercury compounds were used until recently in water-based latex paints as a biocide and preservative. Four mercury compounds were formerly registered for this use:
phenylmercuric acetate 3-(chloromethoxy)propylmercuric acetate di(phenylmercury) dodecenylsuccinate phenylmercuric oleate.
The mercury compounds helped to control bacterial and fungal growth and to prevent mildew attack after application when added to exterior paints.
Since the cancellation of registrations for mercury biocides and preservatives by May 1991, no such mercury compounds are allowed in future manufacturing of paints (U.S. EPA, 1992a, EPA 530-R-92-01 3). Paint used, stored, or manufactured at a Comprehensive Environmen- tal Response, Compensation, and Liability Act (CERCLAI site is likely to predate the cancellation of registration of mercury compounds. Therefore, mercury compounds may be present in latex paint or paint wastes at the site.
Catalysts
Mercury chloride (HgCI2) is used as a catalyst primarily for the production of vinyl chloride monomers (Ulrich, 1988, p. 73) and urethane foams (Oertel, 1985, p. 114). HgCI2 also is used to produce anthraquinone derivatives and other products.
STATE AND FEDERAL REGULATIONS
Solid wastes containing leachable mercury above the Toxicity Leaching Characteristic Procedure (TCLP) limit (0.2 mg/L) and certain source-specific wastestreams are regulated at the federal level under RCRA (40 CFR 261.10). Mercury air emissions are regulated at the federal level under the National Emissions Standard for Hazardous Air Pollutants (NESHAP, 40 CFR 60.50). States are beginning to enact legislation to limit the quantities of mercury in non-RCRA-listed wastes entering municipal waste disposal facilities.
12
Mercury Treatment Standards Under RCRA
In the mid-l 1980s to early 1990, the EPA collected and evaluated process performance data to identify Best Demonstrated Available Technologies (BDATs) for the treatment of RCRA- listed wastes. These studies collected performance data for industrial applications of recycling for a wide range of metals-contaminated wastes including mercury-bearing wastes. The EPA BDAT process considered recycling as a treatment alternative for many nonwastewater streams and identified recycling as the BDAT for some nonwastewater subcategories.
Recycling of mercury received increased momentum from the development of land ban restrictions on mercury-containing wastes. Like other metals, mercury cannot be destroyed. Fur- ther, EPA review of treatment data for the development of BDAT indicated that mercury is difficult to reliably stabilize when present either at high concentrations or in elemental form. The analysis of treatability data did, however, indicate that low concentrations of elemental mercury could be stabilized to meet the leachability levels acceptable for land disposal. Applicable technologies for the low-concentration mercury wastes were stabilization, amalgamation, or acid leaching followed by sulfide precipitation.
Due to the concerns about the ability to stabilize wastes containing high levels of mercury, the EPA examined a range of extraction and concentration techniques for recovering mercury for reuse. The classical technologies for recovery of mercury from sludges are the thermal processes of roasting and retorting. These processes sublimate mercury from metal-bearing wastes and capture the mercury, which require further refining prior to reuse.
Aqueous-based mercury recovery methods also were considered. These included acid leaching to form a solution that further concentrated by precipitation, amalgamation, ion exchange, electrodialysis, or electrowinning. Mercury concentrated by precipitation, amalgamation, or ion exchange will require further treatment such as roasting followed by triple vacuum distillation to produce a refined product.
BDAT treatment standards for organomercury wastes require pretreatment to remove or destroy the organic material(s). The organic constituents may interfere with the recovery or treatment of mercury-bearing wastes.
Due to a lack of data on mercury waste treatment by acid leaching followed by solution processing, the EPA established roasting and retorting as the BDAT for all mercury nonwastewaters having total mercury concentrations above 260 mg/kg, except for radioactive mixed wastes. The affected RCRA wastes are D009 (mercury characteristic), P065 (mercury fulminate), P092 (phenyl mercury acetate), U151 (mercury), and K106 (wastewater treatment sludge from the mercury cell process in chlorine production). The EPA also established incineration as a pretreatment step for P065, P092, and D009 (organics) prior to retorting in its June 1, 1990 rule (June 1, 1990, 55 FR 22572 and 22626).
The regulated community has expressed concern over lack of capacity, particularly for incineration of pretreatment of organomercury wastes. At least 17,260 metric tons of nonwaste- water forms of D009, K106, P065, P092, and U151 were generated in 1988 (Labiosa, 1992). By 1994, the estimated capacity for commercial processing of mercury-containing wastes is about 1,140 metric tons per year. Also, the operating commercial retorting facilities for RCRA wastes are permitted only for D009 wastes except for one facility that is permitted to retort K151 wastes.
New capacity is planned for retorting of K106 and D009 wastes. However, these facilities will be located at chloralkali plants and will be used to treat only plant waste on site. Operators of existing hazardous waste incinerators are reluctant to accept mercury-containing wastes. There is concern that existing provisions of 40 CFR 268.42(a) will cause the incinerator ashes and wastewater treatment sludges to be regulated as high-content mercury wastes that will require retorting.
13
Despite the short-term capacity shortage, the EPA concluded that maintaining retorting and roasting as the BDATs for mercury waste is sound and consistent with the EPA methods for establishing BDATs (Labiosa, 1992). Roasting and retorting are demonstrated for a variety of mercury species such as sulfides in ore concentrates and mixed materials in sludge and for solid wastes such as batteries and wastewater treatment sludge. Although the commercial capacity for mercury roasting and retorting is limited, equipment is available for purchase. In addition, new technologies for mercury recovery are being developed and source reduction efforts are reducing mercury waste production (Dungan, 1992).
The BDAT technology code RMERC is defined as retorting or roasting in a thermal processing unit capable of volatilizing mercury and subsequently condensing the volatilized mercury for recovery. The retorting or roasting unit (or facility) must be subject to one or more of the following:
a National Emissions Standard for Hazardous Air Pollutants (NESHAP) for mercury
a Best Available Control Technology (BACT) or a Lowest Achievable Emission Rate (LAER) standard for mercury imposed pursuant to a Prevention of Significant Deterioration (PSD) permit
a state permit that establishes emission limitations (within the meaning of Section 302 of the Clean Air Act) for mercury.
All wastewater and nonwastewater residues derived from the RMERC process must comply with the corresponding treatment standards for the applicable waste code, including consideration of any applicable subcategories (e.g., high or low mercury subcategories).
State Regulations
Several states have enacted or are considering legislation to prohibit mercury disposal in municipal waste, discourage or prohibit mercury use, or encourage mercury recycling.
Under California’s hazardous waste regulations (Title 22, Hazardous Waste Control Law, Health and Safety Code, Division 20), used fluorescent bulbs are considered hazardous because their mercury content exceeds the state’s Total Threshold Limit Concentration (TTLC) for mercury (20 mg/kg). Although California does not have a “conditionally exempt small quantity generator” classification, California’s Department of Health Services (DHS) has instituted a policy (not a regulation), that limits the disposal of fluorescent bulbs by a generator to no more than 25 used tubes and/or mercury vapor lamps “at any one time in one day. . ." (about 7 kg/day, or 210 kg/month). This policy was developed to allow small-quantity generators (e.g., individuals, small businesses) to dispose of their bulbs in the Subtitle D wastestream, while still regulating large-quantity generators (e.g., large companies, relamping companies) under the state’s hazardous waste program.
However, some larger generators apparently have misinterpreted the policy. Anecdotal reports indicate that some companies are disposing of the first 25 bulbs a day in municipal solid waste, with the rest going to recycling facilities, or are storing used bulbs and gradually disposing of them 25 bulbs at a time in their municipal solid waste. Since such practices are not within the intent of California’s regulatory policy, the DHS currently is reviewing the policy to eliminate such abuses (U.S. EPA, 1992b).
14
Connecticut
A recent bill passed by the legislature in Connecticut, PA91-377, requires the Public Works Commissioner to establish a pilot program for collecting and recycling fluorescent bulbs at a state facility. The bill also requires that a report be prepared, by January 1993, for the legislature’s Environment Committee that will address the feasibility and costs associated with such programs. To date, the Public Works Department has initiated neither a pilot program nor a formal study (U.S. EPA, 1992b).
Florida
Florida state law 93-207 section 55, “environmentally sound management of mercury- containing devices and lamps,” has been approved by the governor. The law prohibits incineration or disposal to a landfill of mercury-containing devices after January 1, 1996. The prohibition may be applied as early as July 1, 1994 on a local basis if recycling capacity is available. A mercury- containing device is any electrical product, other than batteries or lamps, that is determined by the Florida Department of Environmental Protection as proven to release mercury into the environment.
Incineration or landfill disposal of mercury-containing lamps is prohibited after July 1, 1994. Recycling rather than disposal in a permitted facility may be required as of July 1, 1994 on a local basis if recycling capacity is available.
Michigan
Proposed legislation in Michigan (Senate Bill 583), if passed, would prohibit the sale of any product that contains lead, mercury, or cadmium that was intentionally introduced into the product unless the product or the product’s packaging is labeled with the statement, “This product contains heavy metals that may be hazardous to human health and the environment.”
Minnesota
The Minnesota legislature has recently enacted a law (1992 Minnesota Law Chapter 560) regulating the disposal of mercury-containing products. The new law limits sales and use of mercury, and bans placing mercury and mercury-containing articles in solid waste or wastewater streams. Sellers of mercury must obtain the buyer’s written agreement to use the mercury only for certain purposes and to abide by the disposal regulations. Buyers must certify that the mercury will be used only for medical, dental, instructional, research, or manufacturing purposes.
Mercury-containing products (i.e., thermostats, thermometers, electric switches, appliances, and medical or scientific instruments) must be clearly labeled as to their mercury content and to the fact that their disposal is now regulated. When such items are removed from use, the mercury must be reused, recycled, or otherwise managed so that it does not enter the municipal solid waste stream or wastewater disposal system. Thermostat manufacturers must provide incentives and sufficient information to ensure that thermostats being removed from service are so reused, recycled, or otherwise managed. The act also includes limits on distribution of thermometers and bans games and toys containing mercury.
In addition, the Minnesota Pollution Control Agency (MPCA) is charged with conducting a study to propose waste management rules that address the disposal of mercury-containing light bulbs (U.S. EPA, 1992b). See Section 5 for a discussion of MPCA’s participation in a thermostat recycling program.
New Jersey
Proposed legislation in New Jersey (Assembly Bill 2046), if passed, would limit develop- ment of new solid waste incinerators and prohibit incineration or disposal in sanitary landfill of
15
metal containers, chlorinated plastics, scrap iron, glass, plastic beverage containers, batteries, used tires, scrapped corrugated cardboard, yard waste, vegetative waste, food waste, newsprint, office paper, and any other material deemed reusable, compostable, or recyclable. The list of proscribed items may be expanded to include any other material in the solid wastestream that is a source of cadmium, lead, dioxin, mercury, chlorine, or halogens for which removal would reduce the heavy metal content of residual ash from combustion of solid waste.
New York
New York State’s Energy, Resources, and Development Agency is working with Mercury Refining (a processor of general mercury wastes near Albany) to increase their capacity for handling and treating various mercury-containing wastes. The state is concerned in general with reducing mercury wastes in municipal wastestreams; fluorescent bulbs are only one of several products of concern.
Mercury Refining is one of a few companies in the United States that reclaims mercury from various wastes. Their retorting process will be expanded to handle batteries and other household waste products. The State of New York and Mercury Refining are investigating various bulb-handling systems for the process (U.S. EPA, 1992b).
Vermont
In Vermont, a bill recently has been introduced to bar the disposal of fluorescent lamps, motor oil, antifreeze, and organic solvents. The bill would require manufacturers to ensure that a collection system is available and to publicize its existence (U.S. EPA, 1992b).
16
SOURCE REDUCTION ALTERNATIVES FOR MERCURY IN THE ELECTRICAL AND ELECTRONICS INDUSTRIES
The industry sectors covered by this report are electrical and electronic device manufac- ture (SIC 36) and measuring and control instrument manufacture (SIC 382). Source reduction alternatives to mercury use are being sought and have been used in the electrical lighting, battery, switching device, instrument, and thermostat manufacturing areas. These alternatives are discussed in the following sections.
ELECTRICAL LIGHTING
In 1992, approximately 55 metric tons of mercury were consumed by the electrical lighting industry. Mercury-containing lamps include fluorescent lamps and high-intensity discharge (HID) lamps. Examples of HID lamps include mercury vapor, metal halide, and high-pressure sodi- um lamps. Today, fluorescent lamps and HID fluorescent lamps are the second largest source of mercury in municipal solid waste, as shown in Table 3 (U.S. EPA 1992a). By the year 2000, mer- cury contamination resulting from the disposal of fluorescent lamps to municipal solid waste is pro- jected to increase to 40.9 short tons. Although manufacturers are working to reduce the mercury content of each lamp, increased fluorescent lamp usage is expected due to their energy efficiency. The average life of an electrical fluorescent lamp is 4 years, whereas that of a HID lamp is less than 1 year. More than 550 million fluorescent lamps were used in 1992 (information obtained from NEMA). The input of mercury to municipal solid waste from the primary source - household batteries - is projected to decline from 621.2 short tons in 1989 to 98.5 short tons in 2000.
All fluorescent lamps contain mercury. Mercury acts as a multiphoton source in fluor- escent lamps. The mercury content typically ranges from 20 to 50 mg per tube depending on the size. Ultraviolet (UV) light is produced by mercury when it is bombarded by electrons produced by current flowing through the tube. Phosphor powders coated on the inside glass tube convert the UV light to visible light.
Major lighting companies such as General Electric, Sylvania, Philips, and Siemens have expended serious efforts to identify alternatives to mercury as a photon source in lighting. The U.S. Department of Energy (DOE) funded a multimillion dollar contract with Sylvania to find a multiphoton phosphor to provide an alternative to mercury use as a photon source. The project terminated last year without success.
Most of the alternatives tested have failed the performance tests. Cesium and cadmium have good discharges but are not economical. They also are hazardous and thus offer no pollution prevention advantage. The use of alternative phosphors would require changes to the design of the lamp power supply. The new lamp design would require changing the ballast and fixture struc- ture, which in turn would require the disposal of millions of ballasts and fixtures, thus contributing to the solid and hazardous wastestreams.
The research to date shows that there is no economically feasible alternative to mercury in fluorescent lighting. However, research is being done to find a way to reduce the amount of
17
mercury used in electrical lighting (Meyer, 1992). Light bulbs produced today contain 60% less mercury than those manufactured 10 years ago. Today a standard fluorescent lamp contains 0.05 mg/m3 mercury, approximately 0.02% of the total weight of the bulb.
A low-energy mercury vapor light bulb was developed by two companies in California. This bulb has an operating life of 20,000 hours, or 20 times that of an incandescent bulb. The mercury vapor is excited using a high-frequency radio wave to give off UV light, which then strikes the phosphor coating on the inside of bulb walls to produce visible light.
Although reduction of mercury in lamps has reached practical limits, there is a growing market for recycling the mercury, glass, and aluminum from fluorescent and mercury vapor lamps. A standard fluorescent lamp contains up to 80% glass by weight. The end assemblies constitute about 15% of the total tube weight. These end pieces are aluminum with a coated tungsten fila- ment held in a glass mount. The filling gas is argon, or in the case of energy-conserving models, argon and krypton. The inside surface of the tube is coated with a phosphor powder to produce light. The powder is mostly calcium phosphate plus trace quantities of activators such as manga- nese, antimony, chloride, fluoride, tin, yttrium, or titanium. The activators control the color of the light.
Prior to 1988, cadmium was used to increase light output efficiency. Because cadmium- containing lamps still exist in the inventories of suppliers, some mercury waste processors will not accept any fluorescent lamps.
Fluorescent lamps can be processed to recover several valuable resources. The recovery process typically involves crushing the tube and separating the metal end pieces from the glass. Metal components such as the end caps often are sent to other recyclers for recovery. The tube components are then roasted and retorted to recover mercury. The glass, phosphor, and mercury may be treated together, or the glass may be separated and only the phosphor treated. The resulting glass often is recycled. Mercury recovered by retorting is purified by distillation for reuse.
Processing typically costs 10 cents per foot of tube for standard tubes. The cost covers shipping to the processing facility. High-pressure mercury vapor lamps and U-tube fluorescent lamps or other lamps with ceramic bases require some hand disassembly. It typically costs about 50 cents per lamp to process these items (Watson, 1992).
Some states such as California and Minnesota have passed legislation restricting the disposal of fluorescent lamps containing 40 to 50 mg of mercury per tube, depending on the size. In California, more than 25 types of fluorescent lamps have been classified as hazardous waste. In Minnesota, all lamps discarded from commercial sources are considered hazardous. Such legisla- tion has caused several companies to start recovering mercury from spent lamps.
Other states are suggesting similar regulations. Currently, there are no federal regulations for the disposal of fluorescent lamps. Used fluorescent lamps that show a toxic leachable charac- teristic by the TCLP test are considered RCRA hazardous waste and are subject to RCRA Subtitle C regulations.
BATTERIES
In 1992, approximately 16 metric tons of mercury were consumed in the United States by the battery manufacture industry. In the past, mercury was added to alkaline-manganese and zinc-carbon batteries to control gassing. U.S. manufacturers were successful in reducing the mer- cury content to below 250 ppm (Balfour, 1992). In 1992 U.S. manufacturers started producing mercury-free alkaline-manganese batteries. Most zinc-carbon batteries manufactured in the United States no longer contain any mercury.
Batteries represent the largest current source of mercury in municipal solid waste, as shown in Table 3 (U.S. EPA, 1992a). In 1989, household batteries accounted for 621.2 short tons
18
of the mercury discarded in municipal solid waste. It is estimated that by the year 2000, household batteries will be responsible for only 98.5 short tons of the mercury discarded in municipal solid waste.
Beginning in 1992, several battery manufacturers began selling mercury-free alkaline batteries. Other metals such as indium, gallium, and magnesium are used as substitutes for mercury (U.S. Bureau of Mines, 1993). In addition, the use of mercuric oxide batteries, primarily for hearing aids and pagers, is being replaced by zinc-air batteries. However, mercuric oxide batteries will continue to be used for medical and military applications because, currently, there are no acceptable substitutes (U.S. Bureau of Mines, 1993).
Many states, such as Minnesota, New Jersey, and Connecticut, have passed legislation targeting the recovery of household batteries. As a result, battery collection and recycling pro- grams have been implemented. Mercury recovery rates from household batteries are improving. Currently, nearly 6% of the mercury in batteries is recovered.
SWITCHING DEVICES
Industrial and economic data suggest that the quantity of mercury used in electronic control and switching devices is significant. Research shows that mercury is still used in the devices described below.
Mercury Electronic Switches
Ignitrons, thyratrons, and trigger-tubes containing mercury are applied as an electronic switch via grid control. The thyratron is adapted to control a moderate amount of power in an on/off switching operation. This type of switch is used in communications and has been replaced largely by solid-state alternatives. The communications industry currently is performing much research in the fiber optic switch field. Optical switching technology is especially suited for appli- cation in the communications industry. Much research is being done in this area, and new appli- cations are still developing (Korotky, 1989; Hinton, 1992). Figure 2 illustrates an example of a high-speed optical switch.
The transistor, developed in 1947, has replaced most vacuum tubes and some gas-filled tubes. Some tubes, however, have not been replaced by solid-state devices, although inroads are
Figure 2. Typical packaged high-speed optical switch has electrical input ports and output fibers (after Korotky, 1989).
19
being made. Special-purpose transistors now are available to amplify signals in the microwave region of frequencies. Applications requiring the ability to amplify high-power signals, however, still require gas-filled tubes. Examples of such applications include microwave ovens, radar instal- lations, X-ray machines, or mercury-arc rectifiers.
The few remaining applications of gas-filled tubes use mercury vapor tubes. Gas-filled mercury tubes use a pool of mercury as the cold cathode pool. In the excitron type, a small arc is continually maintained between the cathode and an auxiliary excitation anode. When the main anode is positive, current is carried through the tube by transport of ionized mercury vapor. The ignitron-type tube is similar, except that a spark is created during each positive cycle by a flash of current passed between an ignitor electrode and the pool of mercury at the cathode. The mercury vapor becomes ionized because of electron emissions caused by the spark, and an arc is estab- lished between the cathode and anode.
Silent Switches
Silent switches using mercury are small tubes with electrical contacts at one end of the tube. As the tube tilts, the mercury collects at the lower end, providing a conductive path to complete the circuit. Mercury switches are available in voltage ratings up to 250 volts and current ratings up to a maximum of 45 amps. This type of switch is used in numerous applications. For example, in the electrical light switch, when the switch is tilted it makes the circuit and when it is tilted back it breaks it. These switches also are used in such diverse applications as sump pump float controls, automobile trunk lamps, and washing machine lift covers. Silent switches are referred to as such because they prevent electrical noise from occurring. When the contact is made the electrical flow is smooth.
There are several alternatives to mercury switches. One alternative is the micro switch; this is a quick-acting snap switch that is actuated by a small travel distance of 1/16 inch or less. It is used to shut off the power that drives a traveling mechanism when the traveling unit reaches a predetermined point. It is operated either manually or mechanically. This switch is a good replace- ment in case of safety switch application or when sudden power interruption is desired, but is not suitable for all mercury switch applications. The primary disadvantage of all hard contact switches is that they may fail due to contamination or corrosion in the contacts. Table 6 compares switches that use mercury with alternative switch types.
Reed Switches
Reed switches are small circuit controls that are used in electronic devices. Their electrical contacts are wetted with mercury to provide an instantaneous circuit when the switch is closed and to permit instantaneous current interruption when the circuit is broken. Reed switches eliminate the static produced in ordinary hard-contact-type switches. Reed switches are used in applications where static would impair the operation of the electronic device.
Alternatives to reed switches are being found in solid-state and electro-optical switches (Table 6). Reed switches are less expensive than solid-state alternatives and therefore still hold a significant place in the market. However, the trend is for solid-state relays to steadily replace mercury-wetted switches.
Proximity Sensors and Switches
An area of growth in the solid-state switch market is in proximity sensors and switches Engineering Materials and Design, 1989). One design uses an inductive coil to sense motion and is used to detect prop shaft rotation and movement of conveyors. Sensing distances are as yet
20
TABLE 6. COMPARISON BETWEEN THE MERCURY SWITCH AND ITS ALTERNATIVES
Type Properties Application Hazardous Content”’
Mercury switch
Hard-contact switch
Solid-state switch
Electra-optical switch
Inductive sensor
Capacitive sensor
Photoelectric sensor
Ultrasonic sensor
Metal-to-metal contact, may be open or sealed, versatile, inexpensive
More sophisticated design features, versatile
Higher speed, expensive, multiple user
Senses metal targets, 10 to 20 mm detection
Senses mass
Senses nontransparent, nonreflective materials, up to 50 m away; high speed
Senses all objects, range of about 0.5 m; high speed
On/off relay, thermostats, circuit control
On/off relay, general circuit con- trols, high or low voltage
Mercury
None
Communications
Conveyors None
(a) Indicates hazardous materials other than lead which may be used in solder.
fairly short, in the neighborhood of 10 to 20 mm. Work is in progress by major companies to increase the sensing distance without significantly increasing switch size. Other problems are that the coil detects only metal targets, temperature affects performance, and hysteresis may not allow opening and closing at the same switch position.
An alternative design uses capacitive sensors to detect mass, so the target may be metallic or nonmetallic targets. However, they may be affected adversely by electromagnetic and radiofrequency interference, as well as by moisture and dust.
Photoelectric sensors also are undergoing expanding development. Typically, they use either a consolidated light beam or a diffuse light source. Distances range from up to 20 to 50 m by beam methods and 2 m by diffuse methods. Fiber optics can be added to detect objects as small as 0.1 mm. Disadvantages are that target materials cannot be transparent or reflective. Other problems are dust, moisture, and ambient light.
Ultrasonic sensors are gaining new ground because they overcome many problems endemic to other kinds of sensors. Ultrasonic sensing does not depend on color, optical reflectivity, shape, or material. Its sensitivity is not diminished by dust or moisture. However, audible noise from machinery may cause misreads. Initially, ultrasonic sensors operated in the 20 to 30 kHz band, but some now operate above 200 kHz. Their sensing range typically is around 0.5 m.
Still more specialized applications such as in telecommunications systems, may employ electro-optical switches. Research in electro-optical switches, or photonics, in proceeding along two
21
paths. Guided-wave photonics is the more highly developed of the two. It combines a large number of signals into a single physical channel within optical fibers and other structures. Free-space pho- tonics, the newer technology, processes signals in parallel using structures such as lenses, mirrors, holograms, and arrays of optical logic gates or electro-optical integrated circuits (Hinton, 1992).
CONTROL INSTRUMENTS
Mercury is used in many instrumentation devices such as thermometers and mercury manometers. Mercury manometers are considered reliable absolute-pressure gages, and they provide the accuracy needed for a system analysis. A common application is in the steam jet air ejectors used in process plants that have a supply of available steam. However, some mercury-free units such as electronic vacuum gages are accurate, portable pressure-measuring instruments. Formerly, gas regulators used mercury in a safety device that was designed to divert gas flow outside of a building if the gas line pressure became too high. This device consisted of a U-shape tube with mercury at the base of the tube. If the pressure were to exceed a safe value, a weighed amount of mercury would be ejected through an outside vent, subsequently relieving gas pressure. Modern gas regulators use a mechanical spring mechanism instead of mercury. Older homes may still have gas regulators that contain mercury.
THERMOSTATS
Thermostats are temperature control devices that usually consist of a temperature- sensing element, an electrical switch that activates heating and cooling equipment, and a mecha- nism for adjusting nominal temperature. Thermostats are used to control temperature in large building spaces, individual rooms, and appliances. Some types of thermostats use mercury in the switch mechanism. Historically, mercury switches have proved to be quite reliable, accurate, long- lived, and cost efficient. These are important qualities because thermostats control the dispensa- tion of large amounts of electrical power and their operational efficiency has a large impact on fuel consumption. Unoptimized thermostatic control can lead to many times more energy consumption than necessary. Poor performance may be caused by one of several reasons. ‘The main reason is due to hysteresis in the temperature-sensing component, the electrical switch, or both. Hysteresis may lead to large differentials, or swings, in room temperature.
Most residential and appliance thermostats are two-wire, or on/off, type electromechan- ical devices. They contain a temperature-sensing element that mechanically moves an electrical switch into a position where it can be energized. The temperature-sensing device most commonly used in the United States is a bimetal element, which operates on the principle of differential expansion of materials (Haines, 1961). It is composed of two thin layers of dissimilar metals, which have different coefficients of thermal expansion, either welded or brazed together. Bimetals can have many different shapes, such as a strip, coil, U-shape, and spiral, as shown in Figure 3. When heated, the bimetal bends toward the metal which has the lower of the two rates of thermal expansion, and when cooled it bends the opposite direction. Another type of temperature-sensing element is a gas-filled diaphragm. Diaphragm sensors employ a gas- or liquid-filled form that expands when heated and contracts when cooled (Figure 4). Normally, a refrigerant gas is used. Replacement gases that are non-ozone-depleting are being investigated. These types of sensors are used in some air-conditioning equipment, but less in heating equipment. Diaphragm sensors are more popular in Canada and Europe than in the United States. Their tolerances are not as great as, for example, the mercury tilt switch, and they are relatively expensive because they require more engineering design and calibration prior to sale.
22
(d)
Figure 3. Typical bimetal shapes: (a) strip, (b) coil, (c) U-shape, (d) spiral.
Thermostat switches are mounted on a temperature-sensing element such that they can be energized or de-energized at certain temperatures above a nominal setting, known as the control point. The maximum difference between the minimum and maximum operating temperatures is the temperature differential, which normally should be within 2OF of the control point for comfort heating. Larger temperature differentials would be too noticeable, whereas smaller differentials would cause a heating or cooling system to run more frequently than necessary and would be uneconomical. Another problem that affects system performance is control hunt. If the heating or cooling equipment is mismatched with the room size, temperature may vary widely about the control point. Then, the problem of undershooting or overshooting the control point temperature will result. Control hunt is associated with the heating or cooling system, whereas temperature differential is particular to the thermostat itself.
To achieve better control of temperature differentials, manufacturers fit most all types of thermostats with a component to achieve heat leveling (Haines, 1961). Heat leveling prevents room overheating and is accomplished in either of two ways. One method, called heat anticipation, uses a small resistance element to heat a bimetal sensor at the beginning of the call-for-heat cycle. This method normally is applied to two-wire circuits, that are typical of single-stage switches. The second method, called heat acceleration, is used in series with the common wire in three-wire circuits. A heat accelerator does not cause the full effect of the artificial heat to be felt by the thermostat until the room temperature has risen enough to break the low-temperature contact. The effect of the heat accelerator is to accelerate shutdown of a burner after it has been running. In either case, the resistance heat element is located below and close to the bimetal sensor, so that artificial heat is quickly detected by the heat-sensing element.
In modern residential heating, electromechanical thermostats operate at 24 VAC. However, line ‘voltage may be used in older equipment, in electric strip heating, and in inductive circuits such as the kind used to activate ventilation fans in attics and other semiclosed spaces.
Figure 4. Typical diaphragm forms used for temperature sensing (after Miles, 1966).
23
Conversion from line voltage (nominally 115 VAC) to 24 VAC usually is accomplished by rewiring the thermostat to a transformer supply. Newer, fully electronic thermostats operate at 5 VDC internally. Conversion from 24 VAC to 5 VDC usually is done within the thermostat assembly itself. Commercial heating and cooling equipment often uses fully electronic or pneumatic controls, neither of which utilize mercury.
Mercury Switch Thermostats
The mercury tilt switch is a type of electrical switch that is commonly used in thermo- stats. Mercury’s unique natural properties make it extremely effective in mercury switch thermo- stats, and it has been used in thermostats for more than 40 years. At room temperature mercury has excellent conductivity and its high surface tension enables the mercury to move freely in a cohesive mass within the switch assembly (Figure 5). Each bulb contains approximately 3 grams of mercury. Normally, a mercury tilt switch is mounted to a piece of bimetal. The switch follows the motion of the bimetal as it responds to changes in room temperature by rotating one way or another. The switch thus controls a circuit by being moved to an opened or closed position. A drop of mercury within a sealed glass or metallic tube moves under the force of gravity, where it either makes or breaks an electrical circuit. Mercury’s physical properties, particularly high density and surface tension, are such that the mercury tilt switch performs exceptionally well. As the mercury drop flows down the tube, its weight shifts past the center of the tube to accelerate the tilting motion. Temperature differentials normally may be within 1 to 1.5OF for mercury switch thermostats, which is optimal for most heating or cooling system. Thus, thermostats containing mercury provide accurate and reliable temperature control. A schematic showing the components of the popular Honeywell T87 thermostat is shown in Figure 6.
Mercury switch thermostats operate quietly and efficiently, do not require a power source, and require little or no maintenance. The typical service life of a mercury thermostat is 20 to 40 years. They are sufficiently accurate for residential heating/cooling systems and, by reducing temperature differentials, they provide highly economic temperature control. Mercury tilt switches are sold by suppliers at $0.75 to $1.45 each. The primary markets for mercury thermo- stats are single-stage heating/cooling systems and multistage systems such as heat pumps for resi- dential applications. Single-stage systems, where a thermostat controls either heating or cooling, normally require one mercury tilt switch per thermostat. Multistage heating and cooling systems, such as are used in residential heat pumps, commonly require 2 to 6 mercury tilt switches per
Figure 5. Mercury tilt switch.
24
COVER
THERMOMETER
Figure 6. Honeywell T67 thermostat. (Courtesy of Honeywell, Inc., Minneapolis, Minnesota.)
25
thermostat, depending on the type of system controlled. Residential heat pumps commonly use 2 mercury tilt switches for 3-stage operation (2 heating stages, 1 cooling stage).
Thermostat Market Assessment
Analysis of thermostat markets indicates that approximately 10 to 15 metric tons of mercury are consumed annually in the United States for the production of thermostats, primarily for home heating and cooling applications. Today, about 70 million thermostats are in residential use in the United States’. It is estimated that 90% use mercury. Thermostat manufacturers estimate that 2 to 3 million thermostats are brought out of service each year. Most of these thermostats are replaced by the homeowner or contractor.
There are three main market drivers governing U.S. purchases of thermostats, which are summarized in Table 7. One force driving the market is that existing equipment may require ser- vice. This could occur for several reasons. If the heating system in a home is changed, and/or if the thermostat is old, the homeowner may elect to replace the thermostat. Thermostats usually are installed by the HVAC dealer who purchases the equipment through a wholesaler. A second driver is that the homeowner may choose to modernize existing thermostats that are still in working order. This decision may be motivated by interest in achieving fuel cost savings through purchase of a more efficient thermostat, or for convenience features such as time-temperature programmabil- ity. The third market driver is the purchase of new equipment for a new house or for remodeling or additions in an existing house. Table 7 also shows estimated annual U.S. consumption of thermostats.
Non-Mercury Switch Thermostats
Alternative devices to replace mercury tilt switches would have to address the issues of cost, performance, fuel management, and environmental concerns. For example, if conventional mercury switch thermostats were no longer available, they might be replaced with switches of similar cost. Market research by thermostat manufacturers shows that consumers are driven pri- marily by price and that the majority of consumers will select a replacement thermostat of equiva- lent cost to the original, even if it offers substantially greater temperature differential swings. This would increase the amount of unnecessary energy used to heat buildings. The net environmental impact of energy consumption would have to be compared with the environmental impact of using mercury. For example, burning fossil fuels releases air pollutants, including mercury. Borrowing from studies on fluorescent lamps, which have shown that if they were replaced by incandescent lamps (which would require more energy to produce the same amount of light), the increased mercury entering the environment from burning coal would exceed the amount of mercury con- tained in fluorescent lamps. Although no similar studies have been performed on mercury switch thermostats, thermostats control a much larger amount of energy and have a much longer life than fluorescent lamps, so the results could be even more pronounced. Most important, mercury switch thermostats are relatively easy to recycle, so that the mercury never need enter the environment. Section 5 of this report discusses current recycling efforts by Honeywell, the largest U.S. supplier of mercury switch thermostats.
Several alternatives to mercury switches are available in the market. With the exception of the fully electronic type, the switch technologies discussed below are mature, and each is used today in particular niche applications. All thermostat switches, including the mercury tilt switch
1 Estimate based on consultations with thermostat manufacturers and on housing data from U.S. Bureau of the Census, 1993.
26
Market Driver
Service call
1 to 2 M(b)
Dealer New construction Home remodeling
2 to 3 M(c)
(a) First week of sustained cold or hot weather prompts purchases of new thermostats. (b) Consumption rate based on survey of thermostat manufacturers. (c) Based on new housing starts, which ranged from 1.2 to 1.8 million/yr in the period 1980 to
1992, with corrections for multiple thermostats per housing unit and replacements during remodeling. (Housing data from U.S. Bureau of the Census, 1993.)
(Figure 5), have the basic function of transmitting movement of a thermal sensor to a control component, which then will be regulated with respect to further changes in the thermal sensing unit. The control component is a switch in an electrical circuit, an amplifier in an electronic circuit, and a pressure-actuated valve in a pneumatic system. The electrical and electronic types are summarized in Table 8 and discussed below.
TABLE 6. COMPARISON BETWEEN THE MERCURY SWITCH THERMOSTAT AND ITS ALTERNATIVES
Switch Type Performance Applications Thermostat
Price(b)
Inexpensive, less reliable
Premium residential heating/cooling
$40-80
$10-30
$30-50
$60-100
$70-14O(c)
(a) Primarily used on line-voltage equipment. (b) Manufacturer’s list price; includes thermostat unit, without clock or other options available in product line. (c) Includes programmable features.
27
Open-Contact Mechanical Snap Switch
The mechanical snap switch is perhaps the simplest example of how movement of the bimetal can be transmitted to activate an electrical switch. The example shown in Figure 7 uses the Otter principle (Miles, 1965). Thin bimetal of large surface area is held under tension due to a crimp placed in the metal. This enables the center leg of the device to snap downward as the bimetal expands due to increasing temperature and to snap upward as the temperature declines. An electric circuit is completed by means of a stationary contact above the moving contact. Mechanical snap switches press their contacts together and open them instantaneously (hence the term “snap”), which ensures positive electrical contact with minimum contact wear, and eliminates the need for a separate switch. The primary use of the open-contact snap switch for heating purposes is in electric strip heating, which operates on line voltage rather than on 24 VAC. Fewer than 5% of homes in the United States use electric strip heating, which is mainly confined to areas in the Pacific Northwest and New England. In addition, three-season porches sometimes are heated by electric strip units.
In addition to use in thermostats for home heating, mechanical snap switches can be used as temperature regulators for electric irons, flame-failure devices, and overtemperature controls for electric motors. Advantages of the mechanical snap switch over other types of thermostats include low cost, light weight, design simplicity, and orientation independence. Disad- vantages over other types include lower accuracy, higher temperature differential, shorter life under continuous use, and failure if contacts become dirty.
Open-Contact Magnetic Snap Switch
The magnetic snap switch uses a thermal sensing element, such as a bimetal or gas-filled diaphragm, to act against an armature which is poised to bring a movable contact into position with a fixed contact, as shown in Figure 8. As this occurs, the armature responds to the force of a magnet and suddenly is drawn toward it. The purpose of the magnet is to prevent the contacts from arcing and chattering as they are drawn closer together. The mechanism is designed such that with the contacts engaged, the magnetic armature still is slightly separated from the magnet. Then, as the temperature increases, a spring exerts a restoring force away from the magnet to the point where the armature’s attraction to the magnet is overcome, and the moving blade snaps away from the fixed contact. Advantages of the open-contact magnetic snap switch over other types of thermostats are low to moderate cost, good precision, and orientation independence.
Figure 7. Mechanical bimetallic snap switch.
28
Figure 8. Diagram of magnetic snap switch.
It is an alternative that already is used in many low-end, single-stage thermostats. A primary disadvantage is that this type of switch is easily contaminated and subject to failure due to common household items such as dust that prevent the contacts from closing properly. This type of switch is not acceptable in premium equipment such as heat pumps.
Sealed Magnetic Snap Switches
The sealed magnetic snap switch is similar to the open-contact magnetic switch de- scribed above, except that the contacts and armature assembly are sealed in a canister to prevent contamination by dust. This type has been used in single-stage heating/cooling systems for more than 20 years. It is considered a higher cost alternative to other types of switches. However, recent cost increases in manufacturing all types of switches have brought these two types into closer alignment. The sealed magnetic-snap switch offers superior performance when used with a U-shaped bimetal rather than a spiral bimetal. It has the advantages of not being position-sensitive and of being resistant to chattering due to wall vibration. One problem is that the spacing required to eliminate magnetic interference between switches is difficult to overcome in multistage thermo- stats. The major problem in staging of snap switches is being able to stage the switches without cascading through the small interstage differentials required.
Electronic Control Systems
Fully electronic thermostats use all solid-state electronic components for sensing tempera- ture and for switching the heating or cooling circuit. The primary temperature sensing element is the thermistor. The cost of employing electronic control systems in low-end and single-stage applica- tions is high. However, electronic switch costs are declining and the mercury switch cost is steady, causing the gap to narrow. Unlike mercury-containing thermostats, electronic thermostats require a power source. Power can be provided directly from a 24-VAC source or from self-contained
29
batteries. The primary advantage of electronic thermostats is programmability and the fuel cost savings that can result by lowering room temperature automatically at preset times. Disadvantages are higher cost and possible shorter lifetimes than electromechanical thermostats. However, it remains to be seen if electronic thermostats produced today will operate with as few needs for repairs as have been documented for electromechanical types. One disadvantage of the battery- powered units is that the batteries require periodic changing and may create a different environmental concern. Older electronic units had a tendency to be larger than any of the electro- mechanical units. Because the thermostat market is price-sensitive at the different performance levels, the penetration of electronics into the middle- and low-end portions of the markets has been prevented. Current manufacturer’s estimates for the single-stage electronic thermostats indicate a retail price increase of 35% to 100% over the electromechanical units, with the cost differential for multistage units in the 20% to 50% range.
Hot- Wire Vacuum Switch
The hot-wire vacuum switch is a secondary relay that operates on the principle that no arc is formed in a vacuum when an electric circuit is made or broken. This allows even highly inductive loads, such as motors, to be switched. Electrical loads up to 25 amps at line voltage can be accommodated with appropriate surge suppression, with temperature differentials of only about 0.1OC (Miles, 1965). A schematic of this switch is shown in Figure 9. The distance between the
Figure 9. Hot-wire vacuum switch.
30
contacts may be very small, typically on the order of 0.001 inch. A resistance wire is wound around an insulating bobbin and is rigidly fixed to its terminals. Tension on the resistance wire forces the level to rotate about a fulcrum and compress a spring. If a small current is passed through the wire it is heated and expands, which allows the spring to force the level, thus closing the gap and completing a circuit between the main terminals. The hot-wire vacuum switch is ideal for applications where fine temperature control is needed. One unusual feature of this relay is that after the primary control circuit is made, some time is taken before the resistance wire expands and the load circuit is made. The time delay may be made up to 20 seconds. The primary control device may be a hard contact bimetal switch. The hot-wire vacuum switch is used for temperature control in some luxury automobiles.
MERCURY-CADMIUM-TELLURIDE SEMICONDUCTORS
The potential health and environmental hazards of preparing mercury-cadmium-telluride (MCT) materials is well recognized in the electronics industry. Although method improvements have led to a decrease in the amount of mercury used, difficulties involved in preparing MCT are such that efficiency has not been a primary consideration. Production of MCT requires very con- trolled conditions. Although MCT can be made in bulk, it is more successfully prepared using epitaxial growth techniques (Liu et al., 1991). Bulk methods require high Hg partial pressure at the maximum melting point, which is approximately 35 atm for x=0.2 compositions, where x is a com- position variable in the formula Hg1-xCdxTe. Bulk methods of preparation also tend to yield crystals that are nonuniform in x, due to segregation effects in the melt (Irvine and Mullin, 1981). For