National Emission Standards for Hazardous Air Pollutants: Mercury Emissions From Mercury Cell Chlor-Alkali Plants

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 63

[FRL-7236-6]

RIN 2060-AE85

National Emission Standards for Hazardous Air Pollutants: Mercury Emissions From Mercury Cell Chlor-Alkali Plants

AGENCY:

Environmental Protection Agency (EPA).

ACTION:

Proposed rule.

SUMMARY:

This action proposes national emission standards for hazardous air pollutants (NESHAP) for mercury cell chlor-alkali plants. The proposed standards would limit mercury air emissions from these plants. The proposed standards would implement section 112(d) of the Clean Air Act (CAA) which requires all categories and subcategories of major sources and area sources listed in section 112(c) to meet hazardous air pollutant emission standards reflecting the application of the maximum achievable control technology (MACT). The proposed standards would reduce nationwide mercury emissions from these sources by about 4,100 kilograms per year (kg/yr) (9,100 pounds per year (lb/yr)) from the levels allowed by the existing mercury NESHAP.

Mercury is a neurotoxin that accumulates, primarily in the especially potent form of methylmercury, in aquatic food chains. The highest levels are reached in predator fish species. Mercury emitted to the air from various types of sources (usually in the elemental or inorganic forms) transports through the atmosphere and eventually deposits onto land or water bodies. When mercury is deposited to surface waters, natural processes (bacterial) can transform some of the mercury into methylmercury that accumulates in fish. The health effect of greatest concern due to methylmercury is neurotoxicity, particularly with respect to fetuses and young children.

DATES:

Comments.

Submit comments on or before September 3, 2002.

Public Hearing.

If anyone contacts the EPA requesting to speak at a public hearing by July 23, 2002, a public hearing will be held on August 2, 2002.

ADDRESSES:

Docket.

Docket No. A-2000-32 contains supporting information used in developing the proposed standards for the mercury cell chlor-alkali plant source category. The docket is located at the U.S. EPA, 401 M Street, SW., Washington, DC 20460 in Room M-1500, Waterside Mall (ground floor), and may be inspected from 8:30 a.m. to 5:30 p.m., Monday through Friday, excluding legal holidays.

FOR FURTHER INFORMATION CONTACT:

Mr. Iliam Rosario, Metals Group, Emission Standards Division (C439-02), U.S. EPA, Research Triangle Park, North Carolina 27711, telephone number: (919) 541-5308, facsimile: (919) 541-5600, electronic mail address:

rosario.iliam@epa.gov.

SUPPLEMENTARY INFORMATION:

Comments.

Comments and data may be submitted by electronic mail (e-mail) to:

a-and-r-docket@epa.gov.

Electronic comments must be submitted as an ASCII file to avoid the use of special characters and encryption problems and will also be accepted on disks in WordPerfect® format. All comments and data submitted in electronic form must note the docket number: Docket No. A-2000-32. No confidential business information (CBI) should be submitted by e-mail. Electronic comments may be filed online at many Federal Depository Libraries.

Commenters wishing to submit proprietary information for consideration must clearly distinguish such information from other comments and clearly label it as CBI. Send submissions containing such proprietary information directly to the following address, and not to the public docket, to ensure that proprietary information is not inadvertently placed in the docket: OAQPS Document Control Office (C404-02) Attention: Iliam Rosario, Metals Group, Emission Standards Division, U.S. EPA, Research Triangle Park, NC 27711. The EPA will disclose information identified as CBI only to the extent allowed by the procedures set forth in 40 CFR part 2. If no claim of confidentiality accompanies a submission when it is received by the EPA, the information may be made available to the public without further notice to the commenter.

Public Hearing.

Persons interested in presenting oral testimony or inquiring as to whether a hearing is to be held should contact Cassie Posey, telephone number: (919) 541-0069. Persons interested in attending the public hearing must also call Cassie Posey to verify the time, date, and location of the hearing. The public hearing will provide interested parties the opportunity to present data, views, or arguments concerning the proposed emission standards.

Docket.

The docket is an organized and complete file of all the information considered by the EPA in rule development. The docket is a dynamic file because material is added throughout the rulemaking process. The docketing system is intended to allow members of the public and industries involved to readily identify and locate documents so that they can effectively participate in the rulemaking process. Along with the proposed and promulgated standards and their preambles, the contents of the docket will serve as the record in the case of judicial review. (

See

section 307(d)(7)(A) of the CAA.) The regulatory text and other materials related to this rulemaking are available for review in the docket or copies may be mailed on request from the Air Docket by calling (202) 260-7548. A reasonable fee may be charged for copying docket materials.

World Wide Web Information.

In addition to being available in the docket, an electronic copy of today's proposed rule will also be available through EPA's World Wide Web site. Following signature, a copy of the rule will be posted on our policy and guidance page for newly proposed or promulgated rules:

http://www.epa.gov/ttn/oarpg.

The web site provides information and technology exchange in various areas of air pollution control. If more information regarding the web site is needed, call our web site help line at (919) 541-5384.

Regulated entities.

Entities potentially affected by this action include plants engaged in the production of chlorine and caustic in mercury cells. Regulated categories and entities include those sources listed in the primary Standard Industrial Classification code 2812 or North American Information Classification System code 325181.

This description is not intended to be exhaustive, but rather provides a guide for readers regarding entities likely to be regulated by this action. To determine whether your facility, company, business, organization, etc., is regulated by this action, you should carefully examine § 63.8182 of the proposed rule. If you have questions regarding the applicability of this action to a particular entity, consult the person listed in the preceding

FOR FURTHER INFORMATION CONTACT

section.

Outline.

The information presented in this preamble is organized as follows:

I. Background

A. What is the source of authority for development of NESHAP?

B. What criteria are used in the development of NESHAP?

C. What is a mercury cell chlor-alkali plant?

D. What are the health effects associated with mercury?

E. How does this action relate to the part 61 Mercury NESHAP?

II. Summary of Proposed Standards

A. What is the source category?

B. What are the affected sources and emission points to be regulated?

C. What are the emission limitations?

D. What are the work practice standards?

E. What are the operation and maintenance requirements?

F. How are initial and continuous compliance with the emission limitations to be demonstrated?

G. How are initial and continuous compliance with the work practice standards to be demonstrated?

H. What are the notification and reporting requirements?

I. What are the recordkeeping requirements?

III. Rationale for Selecting the Proposed Standards

A. How did we select the source category?

B. How did we select the affected sources and emission points to be regulated?

C. How did we select the form of the standards?

D. How did we determine the basis and level of the proposed standards for existing sources?

E. How did we determine the basis and level of the proposed standards for new sources?

F. How did we select the testing and initial compliance requirements?

G. How did we select the continuous compliance requirements?

H. How did we select the notification, recordkeeping, and reporting requirements?

IV. Summary of Environmental, Energy, Cost, and Economic Impacts

A. What are the air emission impacts?

B. What are the non-air health, environmental, and energy impacts?

C. What are the cost and economic impacts?

V. Solicitation of Comments and Public Participation

VI. Administrative Requirements

A. Executive Order 12866, Regulatory Planning and Review

B. Executive Order 13132, Federalism

C. Executive Order 13175, Consultation and Coordination with Indian Tribal Governments

D. Executive Order 13045, Protection of Children from Environmental Health Risks and Safety Risks

E. Unfunded Mandates Reform Act of 1995

F. Regulatory Flexibility Act (RFA), as amended by the Small Business Regulatory Enforcement Fairness Act of 1996 (SBREFA)

G. Paperwork Reduction Act

H. National Technology Transfer and Advancement Act

I. Executive Order 13211, Actions Concerning Regulations that Significantly Affect Energy Supply, Distribution, or Use

I. Background

A. What Is the Source of Authority for Development of NESHAP?

Section 112 of the CAA contains our authorities for reducing emissions of hazardous air pollutants (HAP). Section 112(d) requires us to promulgate regulations establishing emission standards for each category or subcategory of major sources and area sources of HAP listed pursuant to section 112(c). Section 112(d)(2) specifies that emission standards promulgated under the section shall require the maximum degree of reductions in emissions of the HAP subject to section 112 that are deemed achievable considering cost and any non-air quality health and environmental impacts and energy requirements.

Each national emission standard for hazardous air pollutants (NESHAP) established reflects the maximum degree of reduction in emissions of HAP that is achievable. This level of control is commonly referred to as maximum achievable control technology (MACT).

Section 112(c)(6) requires us to list source categories and subcategories assuring that sources accounting for not less than 90 percent of the aggregate emissions of each of seven specific pollutants (including mercury) are subject to standards under section 112(d) of the CAA.

Mercury cell chlor-alkali plants are among the sources listed to achieve the 90 percent goal for mercury.

B. What Criteria Are Used in the Development of NESHAP?

Section 112(d)(2) specifies that NESHAP for new and existing sources must reflect the maximum degree of reduction in HAP emissions that is achievable, taking into consideration the cost of achieving the emissions reductions, any non-air quality health and environmental benefits, and energy requirements. This level of control is commonly referred to as MACT.

Section 112(d)(3) defines the minimum level of control or floor allowed for NESHAP. In essence, the MACT floor ensures that the standard is set at a level that assures that all affected sources achieve the level of control at least as stringent as that already achieved by the better-controlled and lower-emitting sources in each source category or subcategory. For new sources, the MACT floor cannot be less stringent than the emission control that is achieved in practice by the best-controlled similar source. The MACT standards for existing sources cannot be less stringent than the average emission limitation achieved by the best-performing 12 percent of existing sources in the category or subcategory (or the best-performing five sources for categories or subcategories with fewer than 30 sources).

In developing MACT, we also consider control options that are more stringent than the floor. We may establish standards more stringent than the floor based on the consideration of cost of achieving the emissions reductions, any non-air quality health and environmental impacts, and energy impacts.

C. What Is a Mercury Cell Chlor-alkali Plant?

1. Mercury Cell Chlor-Alkali Production Facilities

At a mercury cell chlor-alkali plant, mercury cell chlor-alkali production facilities are used to manufacture chlorine and caustic as co-products and hydrogen as a by-product through the electrolytic decomposition of brine in mercury cells. The central unit is the mercury cell which is a device comprised of an electrolyzer (electrolytic cell) and decomposer with one or more end boxes and other components linking them. While each mercury cell is an independent production unit, numerous cells are connected electrically in series to form a cell circuit. Cells are situated in a cell room and typically arranged in two rows separated by a center aisle. The cell room is generally a two-story structure in which mercury cells are housed on the upper floor. The lower floor houses various process and housekeeping functions. The number of mercury cells at a given plant ranges from 24 to 116 and averages 56. A mercury cell involves two distinct reactions which occur in separate vessels. The electrolyzer produces chlorine gas, and the decomposer produces hydrogen gas and caustic solution (sodium hydroxide or potassium hydroxide). The electrolyzer can be described as an elongated, shallow steel trough enclosed by side panels and a top cover. A typical electrolyzer measures about 15 meters (about 50 feet) in length and 1.5 meters (about 5 feet) in width and holds about 3,600 kilograms (around 8,000 pounds) of mercury. The decomposer is a 4-to-5 feet high cylindrical vessel located at the outlet end of the electrolyzer and is usually oriented vertically. The electrolyzer and the decomposer are typically linked by an inlet end box and an outlet end box.

A shallow stream of liquid mercury flows continuously between the

electrolyzer and the decomposer. The mercury enters the cell at the inlet end box and flows down a slight grade to the outlet end box, where it flows out of the cell into the decomposer. After being processed in the decomposer, the mercury is pumped back to the inlet end box of the cell.

Saturated brine (sodium chloride solution or potassium chloride solution) is fed to the electrolytic cell via the inlet end box and flows toward the outlet end box above the shallow layer of mercury. Both brine and mercury flow beneath dimensionally stable metal anodes, typically made of a titanium substrate with a metal catalyst that are suspended in the electrolyzer top. The flowing mercury serves as the cathode.

Electric current applied between the anodes and the mercury cathode causes a reaction that produces chlorine at the anode, while an alkali metal (sodium or potassium) binds with the mercury as an amalgam at the cathode. The chlorine gas is collected at the top of the cell and transported to an ancillary gas purification system followed in most cases by a liquefaction facility. The alkali metal/mercury amalgam exits via the outlet end box and enters the decomposer. The brine, whose salt content has been partially depleted in the reaction, also exits the cell via the outlet end box and is transferred to an ancillary brine preparation system.

The decomposer functions as a packed bed reactor in which the alkali metal/mercury amalgam contacts deionized water in the presence of a catalyst. The amalgam reacts with the water, liberating the mercury and yielding caustic soda (sodium hydroxide) or caustic potash (potassium hydroxide) and hydrogen. The caustic and mercury are separated in a trap at the end of the decomposer. The caustic and hydrogen are each transferred to ancillary treatment, and the mercury is pumped back to the inlet end of the cell.

As previously noted, end boxes serve as connections between the electrolyzer and decomposer in a mercury cell. The inlet end box collects and combines raw materials at the inlet end of the cell, and the outlet end box separates and directs various materials out of the cell. An end-box ventilation system, which is present at most but not all plants, evacuates the vapor spaces of the end boxes. The end-box ventilation system also commonly evacuates the vapor space of other vessels and process equipment, such as pump seals, wash water tanks, and caustic tanks and headers. In most cases, mercury contained in this equipment is covered with a layer of water or other aqueous liquid so the air being pulled into the end-box ventilation system is not in direct contact with mercury. However, due to the elevated temperatures in this equipment, particularly end boxes, mercury diffuses through the liquid and is present in the vapor spaces. The concentration of mercury in end-box ventilation systems before any steps are taken to remove mercury varies greatly depending on the vacated equipment. The collected gases are usually cooled and then treated in a mist eliminator and other control equipment prior to being discharged to the atmosphere. It is the mercury remaining in the treated stream that causes the end-box ventilation system vent to be a point source of mercury air emissions for plants that have these systems.

Important ancillary operations at a mercury cell chlor-alkali plant include chlorine purification and liquefaction, brine preparation, caustic purification, by-product hydrogen cleaning, and wastewater treatment.

Chlorine gas is collected under vacuum from each mercury cell and fed into a header system leading out of the cell room. The chlorine then undergoes cooling, mist elimination, and drying. Only trace amounts of mercury remain in the product chlorine gas, typically less than 0.03 parts per million (ppm). Thus, limited mercury emissions are associated with the chlorine purification operation, as this level is achieved without any steps for mercury removal and is consistent with final mercury concentrations for well-controlled gaseous by-product hydrogen streams. In most instances, further cooling, compression, and liquefaction are conducted to obtain liquid chlorine.

Brine flows in a continuous loop through the mercury cells and the brine preparation system which provides clean saturated brine for electrolysis. An important function of the brine system is the removal of impurities naturally associated with salt such as calcium, iron, and aluminum. The presence of these elements can adversely affect cell efficiency. These impurities are removed by the addition of caustic and sodium carbonate which react to form metal precipitates that are removed by filtration. Subsequently, the brine is acidified to remove excess caustic, subjected to heat exchange for temperature adjustment, and returned to the mercury cells as clean saturated brine. Mercury exists in the brine system in the form of dissolved mercuric chloride and on the order of 3 to 25 ppm. The low vapor pressure of mercuric chloride, which is approximately 30 times lower than that of elemental mercury at 35°C, limits the potential for emissions of mercury from the brine system.

Because the caustic solution produced directly from the decomposer is commercial grade, the only additional treatment needed is mercury removal. The concentration of mercury in the caustic stream leaving the decomposer ranges from about 3 to 15 ppm. Mercury is removed by cooling and filtration. Residual mercury contained in the caustic product is typically around 0.06 ppm.

Hydrogen gas exiting a decomposer contains mercury vapor. A mercury-saturated hydrogen gas stream typically leaves a decomposer at a temperature over 200°F. The mercury concentration of this stream can be as high as 3,500 milligrams per cubic meter (mg/m

3

). Accordingly, in most situations, each decomposer is equipped with an adjacent cooler through which the hydrogen gas stream is routed to condense mercury and return it to the mercury cell. After initial cooling, the hydrogen gas from each decomposer is collected into a common header. The combined gas is then treated for mercury with additional cooling and adsorption (or absorption) control equipment. The cleaned hydrogen gas is then either burned as fuel in a boiler, transferred to another process as a raw material, or vented directly to the atmosphere. Due to the mercury remaining in the treated stream, the by-product hydrogen stream is a point source of mercury air emissions.

Mercury cell chlor-alkali plants generate a variety of aqueous waste streams that contain mercury and are treated in a wastewater treatment system. These wastewaters originate from a variety of sources, ranging from wastewaters produced from cell room washdowns and cleanup activities to liquids or slurries produced from purged brine from the brine system and backwash water from the filtration equipment used for caustic purification.

Wastewater treatment applied at most mercury cell chlor-alkali plants entails three basic steps. First, sodium hydrosulfide is added to the wastewater (which contains both elemental mercury and mercury compounded as mercuric chloride) to form mercuric sulfide. This compound has a very low vapor pressure which practically eliminates the potential for mercury air emissions from wastewater treatment. Next, the mercuric sulfide is removed through precipitation and filtration which results in a liquid fraction and a mercuric sulfide filter cake. Any dissolved mercury contained in the liquid is removed by treatment in a carbon adsorber prior to being

discharged in accordance with a plant's discharge permit. The wastewater treatment sludges produced, which consist mainly of the mercuric sulfide filter cake, are classified as hazardous under Resource Conservation and Recovery Act (RCRA) regulations (40 CFR part 261, subpart D). This waste, designated as K106, must be treated for mercury removal prior to disposal or landfilling which generally means high temperature treatment.

2. Mercury Recovery Facilities

Nine mercury cell chlor-alkali plants have mercury recovery facilities on-site to recover elemental mercury from mercury-containing wastes. The wastes treated include those considered K106 wastes, as cited above, and debris and nondebris D009 wastes. The D009 wastes, as classified under RCRA regulations (40 CFR part 261, subpart D), are nonspecific mercury-containing wastes. Debris wastes include any contaminated material or item greater than 2

1/2

inches in any one dimension, such as hardware, protective gear, piping, and equipment. Nondebris wastes include graphite from decomposers, cell room sump sludges, spent carbon media from carbon adsorption control devices, and other small solids.

The most commonly used process is thermal recovery (retorting), where mercury-containing wastes are heated to volatilize the mercury which is then condensed and recovered. Six plants each operate a mercury thermal recovery unit. In such a unit, mercury in wastes is driven to the vapor phase at temperatures over 1,000°F inside one or more retorts. The retort off-gas, which is rich in mercury vapor, is routed through cooling equipment to condense the mercury for recovery. However, because it is not possible to condense all of the mercury, the off-gas is typically routed through polishing control equipment to further reduce mercury before the stream is discharged to the atmosphere. This causes the mercury thermal recovery unit vent to be a point source of mercury air emissions. Mercury that never vaporizes and subsequently is neither condensed nor emitted remains in the retort ash, whose mercury content is limited by RCRA land disposal restrictions (40 CFR part 268, subpart E).

Mercury thermal recovery units can be classified, based on the type of retort used, as oven type units and non-oven type units. Three plants have batch oven retorts, and three plants have non-oven retorts (rotary kiln or single hearth). There are differences between the two types related to operating temperature and residence time. Oven retorts have lower operating temperatures (around 1,000°F) and substantially longer residence times (24 to 54 hours) than do kilns which operate at around 1,375°F with residence times approaching 3 hours.

Noteworthy among all six thermal recovery units is the relatively small volume of exhaust gas generated. Volumetric flow rates range from around 50 standard cubic feet per minute (scfm) on one oven type unit to 1,200 scfm on one non-oven type unit. Non-oven type units have higher volumetric flow rates with an average flow rate of 1,000 scfm and a median of 1,075 scfm than oven type units with an average of 130 scfm and a median of 100 scfm.

Two of the nine plants use a chemical process in which mercuric sulfide and elemental mercury in wastes are chemically transformed to mercuric chloride from which elemental mercury is then precipitated. This process differs from mercury thermal recovery in that it is an entirely liquid-phase operation. Moreover, owing to the low vapor pressure of mercuric chloride, the potential for mercury air emissions from this process is limited. Mercury that is not converted and recovered remains in the processed waste materials whose mercury content is limited by RCRA land disposal restrictions for nonthermal mercury recovery processes (40 CFR part 268, subpart E).

The ninth plant uses a batch purification still for recovering elemental mercury only from end-box residues which are high in mercury content. The system involves heating small batches of end-box residues to volatilize the mercury contained followed by a condenser for mercury recovery. This contrasts with thermal recovery units that treat large volumes of low mercury content wastes. The still is operated under vacuum such that the gas stream after the condenser is routed through two carbon adsorption beds in series to limit mercury air emissions. The system is used only a few times per year for 1 to 2 days at a time. Due to the small volumetric flow rate and mercury concentration of the vented stream and limited operation of the still, mercury air emissions are very low from recovery in the batch purification still.

Fugitive mercury emissions can occur due to leaking equipment, liquid mercury spills, or accumulations in many locations throughout mercury cell chlor-alkali production facilities and mercury recovery facilities, including areas of maintenance activities, liquid mercury collection and handling, and storage for mercury-containing wastes. Most of these sources are associated with cell rooms. Liquid mercury exposed to the atmosphere evaporates at a rate depending on temperature, air flow, and other variables. Methods of controlling fugitive mercury emissions include the containment of liquid mercury leaks, clean up of liquid mercury spills and accumulations, repair of equipment leaking liquid mercury, and containment of mercury-containing wastes under an aqueous liquid or in closed containers. Since liquid mercury can be visually identified, routine visual inspections are an effective method to detect these problems. Mercury vapor leaks, by comparison, are much more difficult to detect and typically result in higher emissions. Vapor leaks occur mostly at the decomposer and in the hydrogen system.

D. What are the Health Effects Associated With Mercury?

Mercury is highly toxic, persistent, and bioaccumulates in the food chain. Most people have some exposure to mercury as a result of normal daily activities. People may be exposed to mercury through inhalation of ambient air; consumption of contaminated food, water, or soil; and/or dermal exposure to substances containing mercury. Also, exposures occur as the result of dental amalgams and from various other sources.

Mercury is a naturally occurring element that is found in air, water, and soil in various inorganic and organic forms. The three primary forms of interest are elemental mercury, inorganic mercury, and methylmercury. As mercury moves through environmental media, it undergoes complex transformations.

Mercury emitted to the air from various types of sources (usually in elemental or inorganic forms) transports through the atmosphere and eventually deposits onto land or water bodies. Once deposited, natural processes can transform some of the mercury into methylmercury which is a highly toxic, more bioavailable form that biomagnifies in the aquatic food chain (such as in fish). Generally, fish consumption dominates the pathway for human and wildlife exposure to mercury.

Inhalation is the primary direct exposure route of concern for elemental mercury because this form strongly partitions to air. Absorption of elemental mercury vapor occurs rapidly through the lungs. Once absorbed, elemental mercury is readily distributed throughout the body; it crosses both placental and blood-brain barriers. The

elemental mercury is oxidized to divalent mercury in most body tissues. Once elemental mercury crosses these barriers and is oxidized to divalent mercury, return to the general circulation is impeded, and mercury can be retained in brain tissue. Effects on the nervous system appear to be the most sensitive toxicological endpoint following exposure to elemental mercury. Exposures above the threshold level can result in tremors, nervousness, insomnia, neuromuscular changes (such as weakness, muscle atrophy, and muscle twitching), headaches, polyneuropathy, and memory loss.

Inhalation and ingestion exposure routes are of interest for inorganic mercury because this form is found in air and other media such as soils and water. There is some limited information suggesting that about 40 percent of the inhaled inorganic mercury is absorbed. Absorption of inorganic mercury through the gastrointestinal tract varies with the particular mercuric salt involved. The portion that is absorbed remains in the body for a considerable length of time. The reported half-life of inorganic mercury in blood is about 20 to 66 days. There is no evidence that inorganic mercury is methylated to form methylmercury in the human body. The inorganic mercury has a limited capacity for penetrating the blood-brain or placental barriers. Limited data suggest that inorganic mercury is a possible human carcinogen. The most sensitive general systemic adverse effect due to exposure to inorganic mercury is the formation of autoimmune glomerulonephritis (that is, inflammation of the kidney).

Ingestion is the primary exposure route of interest for methylmercury. Dietary methylmercury is almost completely absorbed into the blood and distributed to all tissues, including the brain. It also readily passes through the placenta to the fetus and fetal brain. Methylmercury has a relatively long half-life in the human body (about 70 to 80 days). Neurotoxicity is the health effect of greatest concern with methylmercury exposure. The developing fetus is considered most sensitive to the effects from methylmercury. Therefore, women of child-bearing age are the population of greatest concern. During several poisoning incidents in Minamata, Japan, in the 1950's and Iraq in the 1970's, children born of women who were exposed to high doses of methylmercury during pregnancy through ingestion of contaminated fish or grain suffered neurological harms. These harms included death, cerebral palsy, or delayed onset of walking and talking. Also, lower

in utero

exposures have resulted in delays and deficits in learning abilities.

E. How Does This Action Relate to the Part 61 Mercury NESHAP?

We promulgated the National Emission Standard for Mercury on April 6, 1973 (40 CFR part 61, subpart E).

1

Those standards (hereafter referred to as the Mercury NESHAP) limit mercury emissions from mercury cell chlor-alkali plants as well as mercury ore processing facilities and sludge incineration and drying plants. Specifically, the Mercury NESHAP limits mercury emissions from mercury cell chlor-alkali plants to 2,300 grams per day and requires that mercury emissions be measured (in a one-time test) from hydrogen streams, end-box ventilation systems, and the cell room ventilation system. As an alternative to measuring ventilation emissions from the cell room to demonstrate compliance, the Mercury NESHAP allows an owner or operator to assume a ventilation emission value of 1,300 grams per day of mercury providing the owner/operator adheres to a suite of approved design, maintenance and housekeeping practices. Every mercury cell chlor-alkali plant currently in operation in the United States complies with the cell room ventilation provisions by carrying out these practices rather than by measuring mercury emissions discharged from the cell room. Since every plant uses the 1,300 grams per day assumed value for its cell room ventilation emissions, subtracting the 1,300 grams per day cell room value from the 2,300 grams per day plantwide standard effectively creates an emission limit for the combined emissions from hydrogen streams and end-box ventilation systems of 1,000 grams per day.

1

This regulatory program was originally set forth at 38 FR 8826, April 6, 1973; and amended at 40 FR 48302, October 14, 1975; 47 FR 24704, June 8, 1982; 49 FR 35770, September 12, 1984; 50 FR 46294, November 7, 1985; 52 FR 8726, March 19, 1987; and 53 FR 36972, September 23, 1988.

The requirements in today's proposed standards are more stringent than the requirements in the Mercury NESHAP. Using the 1,000 grams per day value as the baseline, we estimate that the mercury emissions would be reduced to less than 60 grams per day (on average) by the proposed rule. This represents about 94 percent reduction from the Mercury NESHAP baseline for vents. In addition, the work practice standards in today's proposal represent the most explicit compilation of practices currently employed by the industry, along with detailed recordkeeping and reporting requirements and requirements that supplement existing RCRA provisions for the storage of mercury-containing wastes. While we cannot quantify the mercury emissions reductions that would be achieved by the proposed work practice standards, we are confident that their implementation would result in additional reductions in mercury emissions beyond that currently achieved by the existing Mercury NESHAP.

We believe that every aspect of the Mercury NESHAP that applies to mercury cell chlor-alkali plants is addressed in today's proposed 40 CFR part 63, subpart IIIII. In fact, as discussed above, the proposed requirements are more stringent than the respective requirements in the Mercury NESHAP. Consequently, we believe that when mercury cell chlor-alkali plants are required to comply with the proposed rule as the promulgated, the requirements of the Mercury NESHAP that apply to them will no longer be relevant or applicable. Therefore, upon the proposed compliance date as indicated in § 63.8186 of the proposed rule, mercury cell chlor-alkali plants will no longer have any obligation to comply with the Mercury NESHAP, nor will they be allowed to comply with the Mercury NESHAP instead of the applicable provisions in the proposed 40 CFR part 63, subpart IIIII. Specifically, we are proposing that affected sources subject to the proposed rule would no longer be subject to §§ 61.52(a), 61.53(b) and (c), and 61.55(b), (c) and (d) of 40 CFR part 61, subpart E, after the compliance date which is proposed to be 2 years following the promulgation of the final rule.

II. Summary of the Proposed Standards

A. What Is the Source Category?

The source category is Chlorine Production. However, this proposal only applies to one type of chlorine production process—the mercury cell chlor-alkali process. Today's proposal applies to all plants engaged in the manufacture of chlorine and caustic in mercury cells. Other chlor-alkali cell types used to produce chlorine and caustic, such as diaphragm cell and membrane cell technologies, would not be covered by this proposed rule because they do not emit mercury. Emissions of chlorine and HCL from all chlorine production facilities are addressed in a separate action elsewhere in today's

Federal Register

.

B. What Are the Affected Sources and Emission Points To Be Regulated?

The proposed rule defines two affected sources: mercury cell chlor-alkali production facilities and mercury recovery facilities. The former includes all cell rooms and ancillary operations used in the manufacture of chlorine, caustic, and by-product hydrogen at a plant site. The latter includes all processes and associated operations needed for mercury recovery from wastes.

Emission points addressed within mercury cell chlor-alkali production facilities include each mercury cell by-product hydrogen stream, each mercury cell end-box ventilation system vent, and fugitive emission sources throughout each cell room and various areas. Emission points addressed within mercury recovery facilities include each mercury thermal recovery unit vent and fugitive emission sources associated with storage areas for mercury-containing wastes.

C. What Are the Emission Limitations?

For new or reconstructed mercury cell chlor-alkali production facilities, the proposed rule would prohibit mercury emissions.

For existing mercury cell chlor-alkali production facilities with end-box ventilation systems, we are proposing that aggregate mercury emissions from all by-product hydrogen streams and end-box ventilation system vents not exceed 0.067 grams of total mercury emitted per megagram of chlorine produced (grams Hg/Mg Cl

2

), or 1.3 x 10

−4

pounds of total mercury per ton of chlorine produced (lb Hg/ton Cl

2

). For existing mercury cell chlor-alkali production facilities without end-box ventilation systems, we are proposing that mercury emissions from all by-product hydrogen streams not exceed 0.033 grams Hg/Mg Cl

2

, or 0.66 × 10

−4

lb Hg/ton Cl

2

. In addition, we are proposing that separate mercury concentration operating limits be established for each affected by-product hydrogen stream and each affected end-box ventilation system vent. The mercury concentration operating limits would be based only on elemental mercury, and each vent stream outlet would be continuously monitored for elemental mercury to show relative changes in mercury levels.

For new, reconstructed, or existing mercury recovery facilities with oven type mercury thermal recovery units, we are proposing that total mercury emissions not exceed 23 milligrams per dry standard cubic meter (mg/dscm) from each oven type unit vent. For new, reconstructed, or existing mercury recovery facilities with non-oven type mercury thermal recovery units, the proposed limit is 4 mg/dscm. Additionally, we are proposing that a mercury concentration operating limit (based on elemental mercury) be established concurrent with the initial performance test for each mercury thermal recovery unit vent.

D. What Are the Work Practice Standards?

We are proposing a set of work practice standards to address and mitigate fugitive mercury releases at mercury cell chlor-alkali plants. These provisions include specific equipment standards such as the requirement that end boxes either be closed (that is, equipped with fixed covers), or that end-box headspaces be routed to a ventilation system. Other examples include requirements that piping in liquid mercury service have smooth interiors, that cell room floors be free of cracks and spalling and coated with a material that resists mercury absorption, and that containers used to store liquid mercury have tight-fitting lids. The proposed work practice standards also include operational requirements. Examples of these include requirements to allow electrolyzers and decomposers to cool before opening, to keep liquid mercury in end boxes and mercury pumps covered by an aqueous liquid at a temperature below its boiling point at all times, to maintain end-box access port stoppers in good sealing condition, and to rinse all parts removed from the decomposer for maintenance prior to transport to another work area.

A cornerstone of the proposed work practice standards is the inspection program for equipment problems, leaking equipment, liquid mercury accumulations and spills, and cracks or spalling in floors and pillars and beams. Specifically, the proposed rule would require that visual inspections for equipment problems, such as end-box access port stoppers not securely in place, liquid mercury in open containers not covered by an aqueous liquid, or leaking vent hoses, be conducted twice each day (once every 12 hours). If a problem is found during an inspection, the owner or operator would need to take immediate action to correct the problem. Monthly inspections for cracking or spalling in cell room floors would also be required as well as semiannual inspections for cracks and spalling on pillars and beams. Any cracks or spalling found would need to be corrected within 1 month.

Visual inspections for liquid mercury spills or accumulations would be required twice per day. If a liquid mercury spill or accumulation is identified during an inspection, the owner or operator would need to initiate cleanup of the liquid mercury within 1 hour of its detection. Acceptable cleanup methods would include wet vacuum cleaning, washing to a trench or canal with an aqueous liquid cover, or a suitable alternative method approved upon petition.

In addition to cleanup, the proposed rule would require that an inspection of equipment in the area of the spill or accumulation be conducted to identify the source of the liquid mercury. If the source is found, the owner or operator would be required to repair the leaking equipment as discussed below. If the source is not found, the owner or operator would be required to reinspect the area every 6 hours until the source is identified or until no additional liquid mercury is found at that location.

Inspections of specific equipment for liquid mercury leaks would be required once per day. If leaking equipment is identified, the proposed rule would require that any dripping mercury be contained and covered by an aqueous liquid, and that a first attempt to repair leaking equipment be made within 1 hour of the time it is identified. The proposed rule would require that leaking equipment be repaired within 4 hours of the time it is identified, although there are provisions for delaying repair of leaking equipment for up to 48 hours.

Inspections for hydrogen gas leaks would be required twice per day (once each 12 hours). For a hydrogen leak at any location upstream of a hydrogen header, a first attempt at repair would be required within 1 hour of detection of the leaking equipment, and the leaking equipment would need to be repaired within 4 hours (with provisions for delay of repair if the leaking equipment is isolated). For a hydrogen leak downstream of the hydrogen header but upstream of final control, a first attempt at repair would be required within 4 hours, and complete repair would be required within 24 hours (with delay provisions if the header is isolated).

As a complement to the inspection program, the proposed rule also includes a requirement to institute a cell room monitoring program whereby owners and operators would continuously monitor mercury concentration in the upper portion of each cell room and take corrective actions as soon as practicable when elevated mercury vapor levels are

detected. The proposed rule does not include detailed requirements for this program. However, we do plan to develop specific criteria for such a program which would be issued either as guidance outside of the final rule or as an amendment to the final rule.

The program would not be a continuous monitoring system inasmuch as the results would be used only to determine relative changes in mercury vapor levels rather than compliance with a cell room emission or operating limit. Generally, the owner or operator would need to establish an action level for each cell room which would be based on preliminary monitoring to determine normal baseline conditions. The action level, or levels if appropriate, would then be established as a yet to be determined multiple of the baseline values. Once the action level(s) is established, continuous monitoring would need to be conducted. If an action level is exceeded, actions to correct the situation would need to be initiated as soon as possible. If the elevated mercury vapor level is due to a maintenance activity, the owner or operator would need to ensure that all work practices related to that maintenance activity are followed. If a maintenance activity is not the cause, inspections and other actions would be needed to identify and correct the cause of the elevated mercury vapor level.

For fugitive mercury emissions associated with storage areas for mercury-containing waste, the proposed rule would require that carbon media from decomposers and cell room sludges either be stored in closed containers or be stored in open containers under a layer of aqueous liquid that is replenished at least once per week. For all other mercury-containing wastes, the proposed rule would require that the wastes either be washed or chemically decontaminated to remove visible mercury or be stored in closed containers.

Finally, the proposed rule would establish the duty for owners and operators to routinely wash surfaces throughout the plant where liquid mercury could accumulate. Owners and operators would be required to prepare and follow a written washdown plan detailing how and how often specific areas specified in the proposed rule would be washed down to remove any accumulations of liquid mercury.

E. What Are the Operation and Maintenance Requirements?

We are proposing that each owner and operator would always operate and maintain affected source(s), including air pollution control and monitoring equipment, in a manner consistent with good air pollution control practices at least to the levels required by the proposed rule, as required under § 63.6(e)(1)(i) of the NESHAP General Provisions. The proposed rule would require each owner and operator to prepare and implement a written startup, shutdown, and malfunction plan according to the operation and maintenance requirements in § 63.6(e)(3) of the NESHAP General Provisions.

F. How Are Initial and Continuous Compliance With the Emission Limitations To Be Demonstrated?

The proposed rule would require compliance with emission limitations within 2 years from [DATE OF PUBLICATION OF THE FINAL RULE IN THE

Federal Register

].

To demonstrate initial compliance with the proposed emission limits for by-product hydrogen streams and end-box ventilation system vents, we are proposing that each owner and operator would conduct performance tests and perform specified calculations. A test would be needed for each by-product hydrogen stream using 40 CFR part 61, appendix A, Method 102. A test would also be required for each end-box ventilation system vent using 40 CFR part 61, appendix A, Method 101 or 101A. Each performance test would be conducted in accordance with a site-specific test plan prepared pursuant to the performance test quality assurance program requirements in § 63.7(c)(2) of the NESHAP General Provisions. Each performance test would be comprised of at least three runs, each lasting 2 hours at a minimum. Concurrent with each test run, the quantity of chlorine produced would need to be determined according to an equation contained in the proposed rule that calculates chlorine production based on cell line electric current load. Then, the mass of mercury emitted per unit mass of chlorine produced would be calculated for each test run, and the runs would be averaged for each tested vent. Initial compliance would be achieved if the sum of the average mass of mercury emitted per mass of chlorine produced of all by-product hydrogen streams and all end-box ventilation system vents is less than 0.067 gm Hg/Mg Cl

2

for plants with end-box ventilation systems, or if the sum of the average mass of mercury emitted per mass of chlorine produced of all by-product hydrogen streams is less than 0.033 gm Hg/Mg Cl

2

for plants without end-box ventilation systems.

To demonstrate initial compliance with the mercury thermal recovery unit emission limits, we are proposing that a performance test be conducted for each vent using Method 101 or 101A. Once again, the performance test would need to follow a site-specific test plan developed by the owner and operator according to § 63.7(c)(2) of the NESHAP General Provisions. The proposed rule would require that during the test, the type of waste resulting in the highest mercury concentration in the mercury thermal recovery unit vent be processed. Documentation of the mercury content of this type of waste and an explanation of why it results in the highest mercury concentration would be required as part of the site-specific test plan. Three test runs would need to be conducted at a point after the last control device for each vent. Initial compliance would be achieved if the average vent mercury concentration is less than 23 mg/dscm for each oven type vent or 4 mg/dscm for each non-oven type vent.

To continuously comply with the emission limit for each by-product hydrogen stream, end-box ventilation system vent, and mercury thermal recovery unit, we are proposing that each owner and operator would continuously monitor outlet elemental mercury concentration and compare the daily average results with a mercury concentration operating limit for the vent. This operating limit would be established during the required performance tests, as explained later in this section. Continuous compliance would be demonstrated by collecting outlet elemental mercury concentration data using a continuous mercury vapor monitor, calculating daily averages, and documenting that the calculated daily average values are no higher than established operating limits. Each daily average vent elemental mercury concentration greater than the established operating limit would be considered a deviation.

The proposed rule would require that each continuous mercury vapor monitor be installed, operated, and maintained in accordance with a site-specific monitoring plan. For each monitor, this plan would need to address installation and siting, monitor performance specifications, performance evaluation procedures and calibration criteria, ongoing operation and maintenance procedures, ongoing data assurance procedures, and ongoing recordkeeping and reporting procedures.

Owners or operators would establish a mercury concentration operating limit for each by-product hydrogen stream, end-box ventilation system vent, and mercury thermal recovery unit vent as part of the initial compliance demonstration. During each

performance test, the proposed rule would require that a continuous mercury vapor monitor be used to measure elemental mercury concentration in the vent stream at least once every 15 minutes for the entire duration of each performance test run. The average elemental mercury concentration measured during any valid test run conducted during the performance test in which mercury emissions did not exceed the applicable emission limit would then be established as the mercury concentration operating limit.

G. How Are Initial and Continuous Compliance With the Work Practice Standards To Be Demonstrated?

The proposed rule would require compliance with the work practice standards within 2 years from [DATE OF PUBLICATION OF THE FINAL RULE IN THE

Federal Register

]. The proposed work practice standards would primarily be requirements for ongoing operational activities. For these activities, there is no specific action called for to demonstrate initial compliance, other than a commitment by the owner or operator that the work practices standards will be met. Therefore, the major component of the initial compliance demonstration for the work practice standards would be a certification by the owner or operator that the work practice standards will be met. In addition, there are a few requirements that could cause an owner or operator to install new equipment or upgrade existing equipment. Documentation of any such actions would also be required in the initial compliance demonstration.

The proposed rule contains specific recordkeeping requirements related to the work practice standards. These include records of when inspections were conducted, problems identified, and actions taken to correct problems. Continuous compliance with work practice standards would be demonstrated by maintaining these required records.

Initial compliance with the washdown plan would be demonstrated by submission of the plan by the owner or operator and certification that they operate according to, or will operate according to, the plan. Continuous compliance with the plan would be demonstrated by maintaining related records. Records would also be required to demonstrate compliance with the cell room monitoring program.

H. What Are the Notification and Reporting Requirements?

The proposed rule would require that owners or operators submit the following notifications and reports:

• Initial Notification

• Notification of Intent to conduct a performance test

• Notification of Compliance Status (NOCS)

• Compliance reports.

For the Initial Notification, we are proposing that each owner or operator notify us that their plant is subject to the NESHAP for mercury cell chlor-alkali plants, and that they provide other basic information about the plant. For existing sources, this notification would need to be submitted no later than [DATE 120 CALENDAR DAYS AFTER PUBLICATION OF THE FINAL RULE IN THE

Federal Register

].

For the Notification of Intent report, we are proposing that each owner or operator notify us in writing of the intent to conduct a performance test at least 60 days before the performance test is scheduled to begin.

The Notification of Compliance Status for the work practice standards would be due [DATE 30 DAYS AFTER THE PUBLICATION OF THE FINAL RULE IN THE

Federal Register

] for existing sources. In this notification, the owner or operator would need to certify that the work practice standards are being or will be met. Furthermore, we are proposing that the washdown plan be submitted as part of this notification, and that the owner or operator certify that they operate or will operate according to the plan.

For the emission limits where a performance test is required to demonstrate initial compliance (that is, the emission limits for by-product hydrogen streams and end-box ventilation system vents and the mercury thermal recovery unit vent limits), the tests would have to be conducted within 180 days after the compliance date, and the Notification of Compliance Status would be due 60 days after the completion of the performance test. We are proposing that the site-specific plan addressing the use of continuous mercury vapor monitors for vents be submitted as part of this notification.

Reporting on continuous compliance would be required semiannually, with the first report due within the first 6 months after initial compliance.

I. What Are the Recordkeeping Requirements?

Records required by the proposed rule related to by-product hydrogen streams, end-box ventilation system vents, and mercury thermal recovery unit vents include the following: performance test results, records showing the establishment of the applicable mercury concentration operating limits (including records of the mercury concentration monitoring conducted during the performance tests), records of the continuous mercury concentration monitoring data, records of the daily average elemental mercury concentration values, and records associated with site-specific monitoring plans.

With regard to the work practice standards, the proposed rule would require that records be maintained to document when each required inspection was conducted and the results of each inspection. Records noting equipment problems (such as end-box cover stoppers not securely in place or mercury in an open container not covered by an aqueous liquid) identified during a required inspection and the corrective action taken would also be required. If equipment that is leaking mercury liquid or hydrogen/mercury vapor is identified during a required inspection or at any other time, the proposed rule would require records of when the leak was identified and when it was repaired. Similarly, if a mercury spill or accumulation is identified at any time, the proposed rule would require records of when the spill or accumulation was found and when it was cleaned up.

A copy of the current version of the washdown plan would need to be kept on-site and be available for inspection. Records of when washdowns were conducted would be required.

The proposed rule would require that copies of each notification and report that is submitted to comply with this subpart be kept and maintained for 5 years, the first 2 of which must be on-site.

III. Rationale for Selecting the Proposed Standards

A. How Did We Select the Source Category?

The mercury cell chlor-alkali production portion of the chlorine production source category was among the categories and subcategories of major and area sources listed for regulation under section 112(c)(6) of the CAA (63 FR 17838, April 10, 1998) to assure that sources accounting for not less than 90 percent of the aggregate mercury emissions nationwide are subject to standards under section 112(d). We estimate that mercury cell chlor-alkali production accounts for

over 5 percent of all stationary source emissions of mercury and over 25 percent of the emissions from stationary noncombustion sources. The Chlorine Production source category is comprised of 43 facilities engaged in the manufacture of chlorine and caustic in electrolytic cells. Cell types employed include the diaphragm cell, membrane cell, and mercury cell. Of these, only the mercury cell process has the potential to emit mercury. For the 1997 base year of the MACT analysis, twelve facilities employed mercury cells. We are aware that one of the twelve facilities ceased operations permanently in September 2000. Nonetheless, we considered it to be part of the source category for the development of MACT since it was in operation in 1997.

B. How Did We Select the Affected Sources and Emission Points To Be Regulated?

For the purposes of implementing NESHAP, an affected source is defined to mean the stationary source, the group of stationary sources, or the portion of a stationary source that is regulated by relevant standards or other requirements established under section 112 of the CAA. An affected source specifies the group of unit operations, equipment, and emission points that are subject to the standards. We can define an affected source as narrowly as a single piece of equipment or as broadly as all equipment at a plant site.

We decided to separate the unit operations and emission points related to the production of chlorine and caustic from the unit operations and emissions points related to mercury recovery. Mercury cell chlor-alkali production facilities include a number of integrated operations dedicated to the production, storage, and transfer of product chlorine, product caustic, and by-product hydrogen. In contrast, mercury recovery facilities are operations dedicated to the recovery of mercury from mercury-containing wastes. These operations are independent of the chlor-alkali process and are thus not integral to production. As a result, the proposed rule addresses emissions from two separate affected sources: mercury cell chlor-alkali production facilities and mercury recovery facilities.

Unit operations and emission points grouped within the mercury cell chlor-alkali production facilities affected source are by-product hydrogen streams, end-box ventilation system vents, and fugitive mercury emissions associated with cell rooms, hydrogen systems, caustic systems, and storage areas for mercury-containing wastes. As described previously, each is a potentially significant source of mercury emissions. Chlorine purification, brine preparation, and wastewater treatment operations are believed to have low mercury emissions to the air. Accordingly, today's proposal contains no requirements for these operations.

Unit operations and emission points grouped within the mercury recovery facilities affected source include all mercury thermal recovery unit vents and fugitive mercury emissions associated with mercury-containing waste storage areas. Chemical mercury recovery and recovery in a batch purification still are believed to have low mercury emissions to the air. Accordingly, today's proposal contains no requirements for these operations.

C. How Did We Select the Form of the Standards?

Section 112 of the CAA requires that standards be specified as numerical emission standards, whenever possible. However, if it is determined that it is not feasible to prescribe or enforce a numerical emission standard, section 112(h) indicates that a design, equipment, work practice, or operational standard may be specified.

With the exception of standards for fugitive emission sources, we are proposing numerical emission limits for all other mercury emission sources. Specifically, the proposed standards include numerical emission limits for by-product hydrogen streams, end-box ventilation system vents, and mercury thermal recovery unit vents.

Cell rooms bring together mercury, a large electrical load, and hot production equipment. Accordingly, most fugitive mercury emission sources at mercury cell chlor-alkali plants are associated with cell rooms. Reliable quantification of these cell room fugitive emissions would be costly, owing to the need to measure both mercury vapor concentration and air flow rate at ceiling apertures with sophisticated equipment. Some plants have many separate ceiling apertures, and plants in warm climates tend to be little enclosed on the sides. Moreover, levels of fugitive mercury vary with cell room operations, precluding the setting of a numerical limit.

Mercury cell chlor-alkali plant fugitive mercury emission sources are also associated with storage areas for mercury-containing wastes. The measurement of mercury emissions from mercury-containing waste storage areas is also impracticable as these are usually located in several places throughout a plant, many of which are open areas.

Not unexpectedly, emissions data on cell room and waste storage emissions are very limited as in the case of cell rooms, or nonexistent as in the case of waste storage areas. As such, we believe that it is not feasible to either prescribe or enforce numerical emission limit(s) for fugitive mercury emissions from cell rooms and waste storage areas. Consequently, today's proposed standards address fugitive emission sources at mercury cell chlor-alkali plants through the establishment of work practice standards.

D. How Did We Determine the Basis and Level of the Proposed Standards for Existing Sources?

Section 112 of the CAA establishes a minimum baseline or “floor” for MACT standards. For new sources, the standards for a source category or subcategory cannot be less stringent than the emission control that is achieved in practice by the best-controlled similar source. The standards for existing sources may be less stringent than standards for new sources, but they cannot be less stringent than the average emission limitation achieved by the best-performing 12 percent of existing sources for categories and subcategories with 30 or more sources, or the average emission limitation achieved by the best-performing five sources for categories or subcategories with fewer than 30 sources for which the Administrator has emissions information.

After the floor has been determined for a category or subcategory, the Administrator must set MACT standards that are technically achievable and no less stringent than the floor. Such standards must then be met by all sources within the category or subcategory. The regulatory alternatives selected for new and existing sources may be different because of different MACT floors, and separate emission limits may be established for new and existing sources.

The EPA generally determines the MACT floor and then considers beyond-the-floor control options. Here, EPA considers the achievable reductions in emissions of HAP (and possibly other pollutants that are co-controlled), cost and economic impacts, energy impacts, and other non-air environmental impacts. The objective is to achieve the maximum degree of HAP emission reduction without incurring unreasonable cost or other impacts.

1. By-Product Hydrogen Streams and End-Box Ventilation System Vents

The fundamental unit in the mercury cell chlor-alkali process is a mercury cell. The by-product hydrogen stream and the end-box ventilation system vent represent the mercury emission point sources that originate from a mercury cell. As discussed earlier, hydrogen gas is incidentally produced as a result of the catalyzed reaction of sodium/mercury amalgam and deionized water to produce caustic in a decomposer. The end-box ventilation stream is a collection of vapors from head spaces of end boxes and possibly other vessels, including pump tanks and seal legs, wash water tanks, and caustic tanks and headers. The mercury content of the by-product hydrogen stream and the end-box ventilation stream, prior to control, is a direct function of the design of the mercury cell. Ten different mercury cell models are used by the twelve mercury cell chlor-alkali plants. Given these differences in cell design and their effect on potential vent mercury emissions, we opted to develop a cell-wide standard for mercury emissions from both points.

Given the large variation among the plants in terms of production capacity (the largest plant is capable of producing over five times as much chlorine as the smallest) and mercury emissions potential, we concluded that any equitable assessment of MACT should account for this disparity. We selected the actual amount of chlorine produced by weight as the uniform parameter for our analysis for the following reasons: Chlorine is the primary product generated; chlorine production can be accurately determined; and chlorine and hydrogen are generated in the same stoichiometric quantities, that is one molecule of hydrogen is produced for each molecule of chlorine produced.

We then considered the fact that two plants do not have end-box ventilation systems. Both plants operate cells with closed end boxes. Consequently, there is no need for end-box ventilation and, therefore, no end-box ventilation system emission point. Next, we examined whether the mercury cells at the ten plants equipped with end-box ventilation systems could be reconfigured with closed end boxes. We concluded that the use of an end-box ventilation system is an inherent feature of the original design of a cell, and that it is not technically feasible to eliminate end-box ventilation systems at these plants. We have, therefore, decided to distinguish plants with end-box ventilation systems and plants without these systems for purposes of establishing MACT.

Accordingly, we are proposing, for plants with end-box ventilation systems, a single emission limit for mercury emissions from all by-product hydrogen streams and mercury emissions from all end-box ventilation system vents in units of mass of mercury emissions per mass of chlorine produced. For plants without end-box ventilation systems, we are proposing an emission limit for mercury emissions from all by-product hydrogen streams in units of mass of mercury emissions per mass of chlorine produced.

• Emission Limit for Plants With End-Box Ventilation Systems

In order to establish MACT for the combined mercury emissions from by-product hydrogen streams and end-box ventilation system vents, we relied on estimates of annual mercury emissions for each vent and information on annual chlorine production provided by the ten plants with end-box ventilation systems. A total of twenty mercury emission estimates were provided, one emission estimate for all by-product hydrogen streams and one emission estimate for all end-box ventilation system vents at each of the ten plants. Background information on these emission estimates is available in the docket to this rulemaking (No. A-2000-32).

Of the twenty emission estimates, fourteen (six for by-product hydrogen streams and eight for end-box ventilation system vents) are based on stack tests performed in accordance with established EPA reference methods specific to chlor-alkali plants. These include Method 101 for the determination of particulate and gaseous mercury from air streams (i.e., end-box ventilation system vents) and Method 102 for the determination of mercury in hydrogen streams. We obtained and reviewed copies of all available test reports and determined that the tests were conducted correctly. Six emission estimates (four for by-product hydrogen streams and two for end-box ventilation system vents) are based on periodic measurements of mercury concentration in the vent streams. The methods used for these periodic measurements are largely modifications of EPA reference test methods. As such, we believe that they provide reasonably accurate results consistent with what would otherwise be obtained with the EPA reference test methods. Our conclusion is that these data represent the best information available on mercury emissions from these vents, and that they are appropriate for use in establishing MACT.

The MACT floor was calculated as follows. For each plant, we divided the sum of the reported annual mercury emissions from all by-product hydrogen streams and end-box ventilation system vents by the annual chlorine production. The chlorine production values used are largely representative of actual annual chlorine production levels. We then ranked the plants from lowest to highest emitters for combined normalized mercury emissions. The normalized mercury emission values range from 0.067 grams Hg/Mg Cl

2

to 3.41 grams Hg/Mg Cl

2

. We should note that the lowest value, 0.067 grams Hg/Mg Cl

2

, is from the plant that closed permanently in September 2000. Nonetheless, we believe that it is appropriate to retain it in the pool of existing sources used to determine existing source MACT. Prior to closure, this plant was the lowest-emitting and best-performing source. The average (mean) of the best (lowest) five normalized values results in a floor value for existing sources of 0.14 grams Hg/Mg Cl

2

.

Of the ten plants with by-product hydrogen streams and end-box ventilation systems, we project that seven would need to install additional controls or upgrade existing controls to meet the 0.14 grams Hg/Mg Cl

2

floor level. We assume the following plant-specific actions: Two plants would need to install new carbon adsorbers on their by-product hydrogen streams (one plant would be replacing an existing adsorber with a new, larger adsorber); one plant would need to install a new packed scrubber on its end-box ventilation system vent; three plants would need to install new controls on both their by-product hydrogen streams and end-box ventilation system vents; and one plant would need to both upgrade carbon adsorber control on its by-product hydrogen stream by switching to impregnated carbon and replacing carbon more frequently as well as install a new packed scrubber on its end-box ventilation system vent.

We estimate that the total aggregate installed capital control costs needed to meet the existing source MACT floor for the seven affected plants to be about $660,000. We estimate total aggregate annual control costs, including costs for labor, materials, electricity, capital recovery, taxes, insurance, and administrative charges (excluding costs for monitoring, reporting, and recordkeeping) for the seven affected plants to be about $570,000 per year. Mercury emission reductions against actual emissions would total 556 kg/yr (1,225 lbs/yr) for the seven affected plants. Mercury emission reductions

against the potential-to-emit baseline, as represented by the allowable emissions under the Mercury NESHAP, would total over 3,400 kg/yr (over 7,500 lbs/yr) for the seven affected plants. The associated annual cost per unit of mercury emission reduction values would be approximately $465 per pound (actuals baseline) and under $80 per pound (potential-to-emit baseline), respectively.

Water pollution impacts due to the increased use of packed bed scrubbers involving aqueous hypochlorite scrubbing solution on end-box ventilation systems are estimated to total 1.2 million liters (320 thousand gallons) of additional wastewater. Impacts on solid waste due to increased use of carbon adsorption for by-product hydrogen streams are estimated to total 17 megagrams per year (Mg/yr), 19 tons per year (tpy), of mercury-containing spent carbon. Energy requirements are estimated to total an additional 878 thousand kilowatt-hours per year (kW-hr/yr). Estimated secondary air pollution impacts due to heightened energy consumption total 282 Mg/yr (311 tpy), with carbon dioxide emissions comprising 99 percent of the estimate.

We then examined beyond-the-floor MACT options. We selected the lowest normalized value among the ten plants, namely 0.067 grams Hg/Mg Cl

2

, as a beyond-the-floor option. As noted above, this 0.067 grams Hg/Mg Cl

2

value is from a plant that is now closed. Nonetheless, as stated previously, we believe it is appropriate to retain it in the pool of existing sources and to include it in the beyond-the-floor assessment.

The 0.067 grams Hg/Mg Cl

2

value corresponds to 0.05 grams Hg/Mg Cl

2

from the by-product hydrogen stream controlled by a condenser coupled with a molecular sieve adsorber, and 0.017 grams Hg/Mg Cl

2

from the end-box ventilation system vent, also controlled by a condenser coupled with a molecular sieve adsorber. It is our understanding that molecular sieve technology for mercury vapor emission control is no longer commercially available. We, thus, acknowledge some uncertainty associated with the achievability of this level of control. However, for the reasons set forth below, we believe that other technologies and operating practices exist that can achieve this level of emissions control.

Due to the very low volumetric flow rates associated with both by-product hydrogen streams and end-box ventilation system vents (typically less than 5,000 scfm and 4,500 scfm, respectively), we believe that the retrofit of control equipment to reduce mercury emissions is both practical and reasonable. We project that the nine plants with baseline emissions greater than 0.067 grams Hg/Mg Cl

2

would meet the 0.067 grams Hg/Mg Cl

2

beyond-the-floor option through the installation of new controls or the upgrading of existing controls. We assume the following plant-specific actions: two plants would need to install new carbon adsorbers on their by-product hydrogen streams (one plant would be replacing an existing adsorber with a new, larger adsorber); three plants would need to install a new packed scrubber on their end-box ventilation system vents; three plants would need to install new controls on both their by-product hydrogen streams and end-box ventilation system vents; and one plant would need to both upgrade existing carbon adsorber control on its by-product hydrogen stream by switching to impregnated carbon and replacing carbon more frequently as well as install a new packed scrubber on its end-box ventilation system vent. We project that the five new carbon adsorbers would need to accommodate a 25 percent higher carbon charge than assumed to meet the floor option. Upgrades to existing carbon adsorber control would involve more frequent carbon replacement than that assumed to meet the floor option. Five of the seven new packed scrubbers on end-box ventilation systems would need to be operated more efficiently than assumed to meet the floor option.

In evaluating regulatory options that are more stringent than the floor, we must consider the cost of achieving such emission reductions, and any non-air quality health and environmental impacts and energy requirements. The beyond-the-floor option would result in an additional 76 kg/yr (168 lb/yr) of total mercury emission reductions for the nine affected plants (a 48 percent incremental reduction from the floor option). For the nine affected plants, the incremental installed capital costs are estimated to total around $210,000, and the incremental annual costs are estimated to total around $150,000 per year. The incremental cost per unit of incremental mercury emission reduction is $900 per pound.

The incremental water pollution impacts are estimated to total 550 thousand liters (145 thousand gallons) of additional wastewater. The incremental solid waste impacts are estimated as 5.1 Mg/yr (5.6 tpy) of mercury-containing spent carbon in total. The incremental energy impacts are estimated as 110 thousand kW-hr/yr in total. The incremental secondary air pollution impacts are estimated to total 35 Mg/yr (39 tpy), with carbon dioxide emissions comprising 99 percent of the estimate.

We believe the additional emission reductions that would be achieved by the beyond-the-floor option are warranted. Further, we believe that the incremental costs of achieving such emission reductions, as well as incremental non-air environmental impacts and energy requirements, are reasonable for mercury. Therefore, we selected the 0.067 grams Hg/Mg Cl

2

beyond-the-floor option as MACT for plants with end-box ventilation systems.

If comments are received on this proposal that lead us to conclude that this level of control is unachievable, we retain the option of setting the standard at the next lowest normalized emission value. Accordingly, we have evaluated the impacts of an alternative 0.076 grams Hg/Mg Cl

2

mercury emission limit for plants with end-box ventilation systems.

We project that the eight plants with baseline emissions greater than 0.076 grams Hg/Mg Cl

2

would need to install new controls or upgrade existing controls to meet this level. This would result in an additional 65 kg/yr (143 lb/yr) of total mercury emission reductions for the eight affected plants (a 41 percent incremental reduction) from the floor option. We assume the same plant-specific actions as those assumed to meet the 0.067 grams Hg/Mg Cl

2

value, given the small difference in emission reductions at the two levels. For the eight affected plants, the incremental installed capital costs are estimated to total around $197,000, and the incremental annual costs are estimated to total around $125,000 per year. The incremental cost per unit of incremental mercury emission reduction is $875 per pound.

The incremental water pollution impacts are estimated to total 317 thousand liters (84 thousand gallons) of additional wastewater. The incremental solid waste impacts are estimated as 5.1 Mg/yr (5.6 tpy) of mercury-containing spent carbon in total. The incremental energy impacts are estimated as 105 thousand kW-hr/yr in total. The incremental secondary air pollution impacts are estimated to total 34 Mg/yr (37 tpy), with carbon dioxide emissions comprising 99 percent of the estimate.

• Emission Limit for Plants Without End-Box Ventilation Systems

In order to establish MACT for mercury emissions from by-product hydrogen streams for the two plants without end-box ventilation systems, we used estimates of annual mercury

emissions from by-product hydrogen streams and information on actual chlorine production provided by the two plants for 1997. Both emission estimates are based on periodic measurements of mercury concentration in the vent streams obtained using methods that are largely modifications of EPA reference test methods. Background information on these emission estimates is available in the docket to this rulemaking (No. A-2000-32).

For each plant, we divided the reported annual mercury emissions from by-product hydrogen streams by the annual chlorine production. The normalized values are 0.033 grams Hg/Mg Cl

2

and 0.17 grams Hg/Mg Cl

2

. Although there are fewer than five sources from which to constitute a MACT floor, we opted to take the average (mean) of the two normalized values, resulting in 0.10 grams Hg/Mg Cl

2

as the floor value for existing sources. We project that the higher emitting plant would need to upgrade existing controls to meet the 0.10 grams Hg/Mg Cl

2

floor level. Specifically, the carbon in its existing carbon adsorbers would need to be replaced more frequently. There would be no capital costs as more frequent carbon media replacement is only a recurring annual cost estimated at $13,000 per year. Mercury emission reductions against actual emissions would total 6 kg/yr (14 lbs/yr). Mercury emission reductions against the potential-to-emit baseline, as represented by the allowable emissions under the Mercury NESHAP, would total over 600 kg/yr (over 1,300 lbs/yr). The associated annual cost per unit of mercury emission reduction values would be approximately $940 per pound and less than $10 per pound, respectively. There are no associated secondary air pollution, water pollution, or energy impacts. Estimated solid waste impacts due to increased use of carbon adsorption total 1.0 Mg/yr (1.1 tpy).

We then examined beyond-the-floor MACT options. We selected the lowest normalized value among the two plants, namely 0.033 grams Hg/Mg Cl

2

, as a beyond-the-floor option. Controls applied to achieve this value include a condenser coupled with a carbon adsorber. For purposes of estimating impacts, we assumed that the higher-emitting plant would replace its existing carbon adsorber with a new, larger adsorber to meet the 0.033 grams Hg/Mg Cl

2

level.

In evaluating regulatory options that are more stringent than the floor, we must consider the cost of achieving such emission reduction, and any non-air quality health and environmental impacts and energy requirements. The beyond-the-floor option would result in an additional 6 kg/yr (14 lb/yr) of total mercury emission reductions (a 47 percent incremental reduction from the floor option). The incremental installed capital costs are estimated to total around $182,000. The incremental annual costs are estimated to total around $126,000 per year. The incremental cost per unit of incremental mercury emission reduction is approximately $9,000 per pound. There are no associated incremental water pollution impacts. The estimated incremental solid waste impacts total an additional 5.3 Mg/yr (5.8 tpy) of mercury-containing spent carbon. The incremental energy impacts are estimated as 252 thousand kW-hr/yr in total. The incremental secondary air pollution impacts are estimated to total 81 Mg/yr (89 tpy), with carbon dioxide emissions comprising 99 percent of the estimate.

We believe the additional emission reductions that would be achieved by the beyond-the-floor option are warranted. Further, we believe that the incremental costs of achieving such emission reductions as well as incremental non-air environmental impacts and energy requirements are reasonable for mercury. Therefore, we selected the 0.033 grams Hg/Mg Cl

2

level as MACT for plants without end-box ventilation systems, which is approximately half the level selected for plants with end-box ventilation systems.

2. Sources of Fugitive Mercury Emissions

As explained above, we have determined that work practice standards provide the most appropriate approach for addressing fugitive mercury emissions at mercury cell chlor-alkali plants. Every mercury cell chlor-alkali plant is currently subject to the Mercury NESHAP and implements the design, maintenance, and housekeeping practices referenced in the NESHAP to control fugitive cell room emissions. We believe that these existing requirements represent the MACT floor for existing mercury fugitive emission sources. Since these floor requirements are currently observed at each existing plant, a standard based on this floor level of control would not be expected to reduce mercury emissions from current levels or produce any associated cost, non-air environmental or energy impacts.

We then examined beyond-the-floor options. We noted that many of the existing work practice requirements are general in nature and nonspecific relative to the frequency and scope of inspections, as well as recordkeeping and reporting. We decided that clarification and elaboration on these general practices was warranted to make them more explicit and to improve assurance of compliance. Accordingly, we initiated a thorough examination of specific measures employed across the industry to limit fugitive mercury emissions.

In the summer of 1998, we conducted site visits to five mercury cell chlor-alkali plants to observe and document their design, operational, maintenance, housekeeping, and recordkeeping practices. The five plants were selected to provide a broad representation of ownership (the five plants are owned by five different companies) and different mercury cell models (mercury cells made by all three manufacturers and of varying sizes are represented). We also selected plants in different areas of the United States (U.S.) to account for geographical variations such as climate. In addition to the site visits, we obtained current standard operating procedures for mitigating sources of fugitive mercury emissions from all twelve plants. We used this knowledge and information to develop a detailed compilation of practices currently used across the industry to control fugitive mercury emissions.

We used this compilation to identify explicit practices for each individual plant area, equipment type, and inspection procedure and assembled them as beyond-the-floor work practice requirements. We feel that the resulting work practice standards represent the most stringent practices applied in the industry.

The types of enhancements from the MACT floor level requirements that are included in the beyond-the-floor option may be generally classified in three categories. First, the beyond-the-floor requirements add considerable specificity. The equipment and areas to be inspected are identified along with the required frequency of the inspections and the conditions that trigger corrective action. Response time intervals for when the corrective actions must occur are also included. Second, some types of inspections are required at more frequent intervals than required by the Mercury NESHAP (e.g., inspecting decomposers for hydrogen leaks once each 12 hours rather than once each day). Third, the beyond-the-floor option includes additional requirements not included in the floor level. The two most obvious examples are the detailed recordkeeping procedures and reporting provisions which are more fully developed than

those in the Mercury NESHAP and the requirements for storage of mercury-containing wastes.

Also included in the beyond-the-floor option is a requirement for owners and operators to develop and implement a plan for the routine washdown of accessible surfaces in the cell room and other areas. All plants currently wash down cell room surfaces regularly. However, due to plant-specific considerations, we are uncomfortable with issuing a specific set of requirements for washdowns that would apply at all plants. As a result, the beyond-the-floor option establishes the duty for owners or operators to prepare and implement a written plan for washdowns and identifies elements to be addressed in the plan. Although washdowns are an ongoing practice at all plants, we believe that including such a requirement in the beyond-the-floor option will elevate the importance of washdowns as part of an overall approach to reducing cell room fugitive emissions.

As a final element of the beyond-the-floor option, we considered the extent to which measurement of ambient mercury levels in the cell room air should be incorporated. Currently, all mercury cell chlor-alkali plants periodically monitor mercury vapor levels at the cell room floor plane, in keeping with Occupational Safety and Health Administration (OSHA) standards for worker exposure to mercury. Typically, on a daily basis, a plant operator measures and records the mercury vapor level in the cell room. Some plants use technologies that measure the mercury vapor level at a single point, such as portable mercury vapor analyzers based on ultraviolet light absorption or gold film amalgamation detection. Plant operators using these technologies take readings at specified locations in the cell room. Other plants utilize procedures that provide an aggregate reading, such as chemical absorption into potassium permanganate solution followed by separate cold vapor atomic absorption analysis in a laboratory setting. This composite sample is most often obtained by a plant operator walking through the cell room with a small sampling pump.

When a mercury vapor level above the OSHA personal exposure limit is measured, plant operators require the use of respirators in the area. They also take action to determine and eliminate the cause of the elevated mercury level.

Given the fact that all plants conduct cell room mercury vapor measurements, we determined that it was appropriate to include requirements to conduct cell room monitoring as a means to identify and correct situations resulting in elevated mercury levels (and obviously, increased mercury emissions) as part of the beyond-the-floor option for fugitive mercury emission sources. We considered basing such a program on periodic measurement, which would correspond to the programs currently in place at mercury cell chlor-alkali plants. We also considered basing such a program on the continuous measurement of mercury vapor levels in the upper portions of the cell room. We are aware of technologies, including extractive, cold vapor spectroscopy systems and open-path, differential optical absorption spectroscopy systems, designed for such continuous monitoring applications. In August of 2000, we studied cell room mercury vapor levels at a U.S. mercury cell chlor-alkali plant using both extractive and open-path technologies. In addition, we are aware of extractive systems currently in use in Europe for this purpose.

Upon consideration of the benefits of periodic versus continuous monitoring of the cell room mercury vapor levels, we selected continuous monitoring as part of the proposed cell room monitoring program for the following reasons. First, we believe that continuous monitoring would identify hydrogen leaks or other situations that result in elevated mercury levels in the cell room much more promptly than periodic monitoring. If periodic monitoring was conducted on a daily basis, hours could pass before such a leak was detected. We also believe that the continuous monitoring of mercury vapor levels during maintenance activities would provide information to help plant operators refine and improve such maintenance activities to reduce mercury emissions.

Finally, we believe that the monitoring on the cell room floor plane could fail to detect hydrogen leaks or other situations resulting in mercury vapor leaks that may occur at higher elevations. Continuous monitoring in the upper portion of the cell room would provide a representation of all areas of the cell room at all levels.

Therefore, we have included a program involving the continuous monitoring of mercury vapor levels in the cell room as part of the beyond-the-floor option. We envision the basic elements for this program to be as follows. Each owner or operator would be required to install a mercury monitoring system in each cell room and continuously monitor the elemental mercury concentration in the upper portion of the cell room. The type of technology, whether an extractive, cold vapor spectroscopy system or an open-path, differential optical absorption spectroscopy system, would be at the discretion of the owner or operator, provided that performance criteria, such as a minimum detection limit, were met. A sampling configuration would be specified to acquire a composite measurement representative of the entire cell room air. For example, the sampling configuration may involve sampling at least three points along the center aisle of the cell room and above the mercury cells at a height sufficient to ensure representative readings.

For each cell room, the owner or operator would need to establish an action level which would be based on preliminary monitoring to determine normal baseline conditions. The onset and duration of this preliminary monitoring would be specified as well as guidelines for setting the action level. Continuous monitoring would commence after a specified time period following establishment of the action level and its documentation in a notification to us. A minimum data acquisition requirement would be established, such as a requirement to collect and record data for at least a certain percent of the time in any 6-month period.

Actions to correct the situation as soon as possible would be required when measurements above the action level were obtained over a defined duration, such as a certain number of consecutive measurements or an average over a certain time period above the action level. If the elevated mercury vapor level was due to a maintenance activity, the owner or operator would need to keep records describing the activity and verifying that all work practices related to that maintenance activity are followed. If a maintenance activity was not the cause, then inspections and other actions would need to be conducted within specific time periods to identify and correct the cause of the elevated mercury vapor level.

In evaluating whether to establish the beyond-the-floor option as MACT, we looked at the incremental impacts on emissions, cost, energy, and other non-air effects. Relative to emissions, we firmly believe that although we are unable to actually quantify the reductions expected with the implementation of the beyond-the-floor option, substantial reductions would nonetheless occur. We know from experience and inference that the added scrutiny inherent in the suite of beyond-the-floor practices will of necessity result in fewer fugitive emissions. In considering the cost impacts of the

beyond-the-floor option, we attempted to estimate the cost associated with the equipment needed to carry out cell room monitoring as well as increased demand for labor and overhead needed to fully implement the proposed monitoring, inspection, recordkeeping, and reporting activities. We estimate the total installed capital costs needed to meet the beyond-the-floor option for fugitive mercury emissions to be around $663,000. We estimate the total annual costs to be around $840,000 per year, consisting of about $94,000 for annualized capital expenditure on mercury monitoring systems; about $736,000 per year for labor for monitoring, inspections, and recordkeeping, about $2,100 per year for mercury monitoring system utilities, and about $7,500 per year for mercury monitoring system replacement parts. We are unable to estimate increases in wastewater associated with washdown and cleanup activities for liquid mercury spills and accumulations as well as increases in solid waste since these would be highly plant-specific. Energy requirements for mercury monitoring systems are estimated to total an additional 53 thousand kW-hr/yr. Estimated secondary air pollution impacts due to heightened energy consumption total 17 Mg/yr (19 tpy), with carbon dioxide emissions comprising 99 percent of the estimate.

We believe the additional emission reductions that would be achieved by the beyond-the-floor option are warranted and that the estimated incremental costs to meet this level are reasonable. Therefore, we are selecting the beyond-the-floor work practice standards as MACT for fugitive mercury emission sources.

With regard to the cell room monitoring program, we acknowledge that there are uncertainties associated with the use of mercury monitoring systems for continuous monitoring that can only be addressed through actual field validation. We are specifically requesting comment on the feasibility of using such systems for continuous monitoring to prompt corrective actions for elevated mercury vapor levels in the cell room. We are also requesting comment on the detailed elements of the cell room monitoring program which we are unable to delineate in its entirety at this time.

Following proposal, we will involve the public in defining this program. Specifically, we will enter into a joint effort with industry, monitoring instrument suppliers, and other interested parties, to detail the elements and requirements of this program. We will take additional appropriate rulemaking steps as necessary to fully implement this program, including assuring opportunity for industry and the public to comment.

3. Mercury Thermal Recovery Unit Vents

As previously discussed, nine of the twelve mercury cell chlor-alkali plants have mercury recovery processes. Six of the nine plants operate a thermal recovery unit in which mercury-containing wastes are heated and the resulting mercury-laden off-gas is cooled and treated for mercury removal prior to being discharged to the atmosphere. Two plants recover mercury with a chemical process and one plant recovers mercury in a purification still; in both cases, mercury air emissions are believed to be low.

In establishing MACT for mercury thermal recovery units, we obtained information from all six plants with these units. Each plant provided descriptions of its thermal recovery operation, including the types of wastes processed and the control devices applied. Where available, plants also provided results of performance testing or periodic sampling and an estimate of their mercury emissions.

Each of the six plants operates one or more retorts (as part of its mercury thermal recovery unit) in which mercury-containing wastes are heated to a temperature sufficient to volatilize the mercury. The off-gas containing mercury vapor is then cooled in the mercury recovery/control system, causing the mercury to condense to liquid. The liquid mercury condensate is then collected from recovery devices for reuse in the mercury cells. The primary emission source is the mercury thermal recovery unit vent where off-gas that has passed through the recovery/control system is discharged to the atmosphere. Retorts used include three basic designs: batch oven (three plants), rotary kiln (two plants), and single hearth (one plant).

The batch ovens are D-tube retorts which are so named because each resembles an uppercase letter “D” on its side. Pans are filled with waste, typically around 10 cubic feet, and then placed into an oven. After inserting three or four pans, the oven door is closed and the retort is indirectly heated to about 1,000°F. The residence time varies from about 24 to 48 hours, depending on the type of waste being processed. While heating, the oven is kept under a vacuum and the mercury vapors are pulled into the mercury recovery/control system. After the cycle is completed, the unit is allowed to cool and the pans are then removed.

The rotary kilns are long, refractory-lined rotating steel cylinders in which the waste charge to be treated flows counter current to hot combustion gases used for heating. Wastes to be treated are conveyed into a ram feeder which inserts a waste charge into the kiln at regular intervals, typically about every 5 minutes. Each is directly fired with natural gas and is heated to over 1,300°F. The rotation of the kiln provides for mixing and transfer of the waste to the discharge end. The residence time is about 3 hours. The gas stream leaving the kiln passes through an afterburner where the temperature is increased to around 2,000°F to complete combustion reactions involving sulfur and carbon and then to a mercury recovery/control system.

The single hearth retort is comprised of a vertically mounted, refractory lined vessel with a single hearth and a rotating rabble. Waste is charged onto the hearth through a charge door by way of a conveyor. Once charged, the conveyor is withdrawn, the charge door is closed, and the heating or treatment cycle begins. The waste is stirred by the rabble rake, which turns continuously, and is heated to around 1,350°F. The residence time, which ranges according to waste type, is typically much longer than for rotary kilns. Similar to rotary kilns, the gas stream leaving the hearth retort passes through an afterburner where the temperature is increased to around 2,000°F to complete combustion reactions involving sulfur and carbon and then to a mercury recovery/control system.

As noted above, there are several important differences between the oven retorts and the non-oven (rotary kiln and single hearth) retorts related to operating temperature and residence time. There are also significant differences in the volumetric flow rates produced by the oven and the non-oven retorts. Oven retorts typically have volumetric flow rates around 100 scfm, which is an order of magnitude lower than flow rates for non-oven retorts which are around 1,000 scfm. Together, these differences can have a material impact on mercury concentration, mass flow rate of mercury, and other factors that influence mercury loadings to the recovery/control system. After evaluation of these technical and operational differences between oven retorts and non-oven retorts and their potential effect on emissions characteristics and control device applicability, we are proposing to distinguish between retort types for the purpose of establishing MACT.

With the exception of the plant with a single hearth retort that is controlled

with a scrubber as the final control device, the recovery/control system at each plant consists of condensation and carbon adsorption for final mercury control. The amount and type of carbon adsorbent used in the fixed bed, nonregenerative carbon adsorbers varies among the five plants. One plant uses activated carbon, one uses iodine-impregnated carbon, and three use sulfur-impregnated carbon. We believe that each type is effective in removing mercury provided the adsorbent is replaced at a frequency appropriate to prevent breakthrough.

In contrast, the plant with the single hearth retort utilizes a chlorinated brine packed-tower scrubber for final mercury control. In this scrubber, elemental mercury vapor is removed by chemically reacting with the chlorinated brine solution to form mercuric chloride, a nonvolatile mercury salt which is readily soluble in aqueous solutions. The resulting scrubber effluent is returned to the brine system causing the absorbed mercury to be recycled back to the mercury cells. Performance data for this brine scrubber system shows that the effectiveness is comparable to that of the condenser/carbon adsorber systems used at the other five plants.

While examining the performance capabilities of the condenser/carbon adsorber systems, we identified several factors that influence performance. We believe that a primary factor affecting mercury recovery and control is the temperature to which retort off-gas is cooled prior to entering the final control device. Because of the volatile nature of elemental mercury, temperature has a direct effect on the concentration of mercury vapor that can exist in a gas stream. For example, the concentration of mercury vapor that could exist in a gas stream at 50°F is 5 mg/m

3

, while the predicted concentration at 85°F is 30 mg/m

3

, a six-fold increase. At 100°F, the concentration could potentially be over 50 mg/m

3

.

A key factor relative to the performance of carbon adsorbers is contact time. As noted previously, we believe that generally each of the carbon adsorbents presently used in the industry can effectively collect mercury vapor. However, it is essential for optimum performance that the contact time between the gas stream to be treated and the carbon adsorbent be long enough to allow for maximum adsorption. Consequently, design and operational factors such as carbon bed depth, sorbent particle size, and gas velocity have an appreciable impact on collection efficiency. Another key consideration is the frequency at which the adsorbent is replaced since the adsorbing capacity of any sorbent decreases as saturation and breakthrough are approached.

In assessing potential formats for a numerical emission limit, we considered a limit on emissions in a specified time period, a limit normalized on the amount of wastes processed, and an outlet mercury concentration limit. The amounts and types of wastes processed at each plant and among plants vary considerably. We believe, generally, that mercury emissions from the thermal recovery unit vent are proportional to the amount of mercury-containing wastes processed and the amount of mercury contained in these wastes. Therefore, we concluded that limiting emissions over a specified time period would unfairly impact plants that process larger amounts of wastes and/or wastes that contain more mercury. A mercury emission limit normalized on the amount of wastes processed would eliminate this inequity. However, given the wide variation in the mercury content of different types of wastes and the varying mix of waste types processed at different plants, we concluded that setting and enforcing such an emissions limit is impractical.

Several factors influence the concentration of mercury in the thermal recovery unit vent exhaust. The most significant include the mercury content of the wastes being processed and the volumetric flow rate through the system. Volumetric flow rate is dependent on process rate, fuel usage, and the volume of combustion gas generated. The mercury concentration may also vary depending on the stage of the heating cycle. The mercury content of the exhaust stream leaving the condenser(s) or other type of cooling unit should remain relatively constant, provided that the outlet temperature is constant and the residence time is sufficient. Depending on the effectiveness of the carbon adsorber or brine scrubber, the mercury concentration would be further reduced. As a result, we conclude that concentration at the outlet of the final control device is the most meaningful and practical measure of the combined performance of each element of the mercury recovery/control system. Therefore, we have selected concentration for the format of the MACT standard for mercury thermal recovery units.

Finally, we evaluated how, or if, the proposed standards should address different waste types; that is, should different emission limits be set for different waste types or should one limit be set for the waste type shown to be the highest emitting. We analyzed all the available data but were unable to ascertain any relationship between the type of waste (K106, D009 debris, or D009 nondebris) being treated during testing or sampling and the outlet mercury concentration measured across all plants. As a result, we are proposing an outlet mercury concentration limit that is neutral to the type of waste being processed. The analysis also influenced our decision on the proposed requirements for performance testing. We are proposing that testing be conducted during conditions representative of the most extreme, relative to potential mercury concentration, expected to occur under normal operation. While we would have preferred that the proposed rule specify the type of waste to be processed during testing, our inability to discern a relationship between waste type and outlet mercury concentration across plants prevented us from doing so. Therefore, the proposed rule would obligate owners and operators to process mercury-containing wastes that result in the highest vent mercury concentration during performance testing.

In summary, our review and analysis of all the available information on mercury thermal recovery units leads us to the following conclusions:

• Separate MACT emission limits should be developed for oven type and non-oven (rotary kiln and single hearth) type mercury thermal recovery units.

• These emission limits should not distinguish among waste types processed.

• Concentration is the appropriate format for the numerical emission limits.

The following describes how we selected the proposed emission limits for oven type and non-oven type mercury thermal recovery units.

There are three plants that use oven retorts. All are owned and operated by the same company. One plant operates five ovens, another operates three ovens, and the third operates two ovens. Thermal recovery at all three plants is conducted between 6,000 to 7,000 hours per year. The amounts of waste processed and the amounts of mercury recovered range from 90 to almost 300 tpy and from 3 to 20 tpy, respectively. At all three plants, the mercury-laden off-gas leaving the retort is cooled and treated for particulates and acid gases in a wet scrubber with caustic solution, followed by further cooling in a condenser. The cooled gas is then routed through one or more fixed-bed, nonregenerative carbon adsorbers before being discharged to the atmosphere. We conducted an evaluation of the mercury

recovery/control systems at all three plants, considering the condenser outlet temperature and the amount of carbon in the beds.

The plant that ranked highest in this evaluation, which we consider to be the best-controlled plant, provided mercury emissions data (periodic sampling results) over 3 years. The other two plants were unable to provide emissions data. Therefore, data from this best-controlled plant were used to establish MACT. Since an emission limit based on the best-controlled plant would obviously be more stringent than the floor level, the selection of a level associated with the best-performing recovery/control system for this retort type clearly meets our statutory requirement regarding the minimum level allowed for NESHAP.

This best-controlled plant has five ovens and two separate but identical mercury recovery/control systems. One treats the exhaust gas from three ovens while the other services two ovens. Each system is comprised of a wet scrubber and condenser, which cool the exhaust gases to around 70°F, followed by a carbon adsorber with about 700 pounds of activated carbon. Available test data for this plant consist of bimonthly measurements for 1997, 1998, and 1999 on each stack. We reviewed the sampling method used to obtain these data which are largely based on EPA reference methods for mercury emissions from mercury cell chlor-alkali plants and concluded that it is capable of producing measurements of reasonable accuracy that are suitable for use as the basis for MACT. We removed six data points that we determined were statistical outliers and combined the data for both control systems into one data set comprised of 134 individual measurements.

We then evaluated options for how these data should be used to establish a numerical emission limit to represent MACT. While this limit must represent the performance of the controls in place at this best-controlled plant, it also must account for variability in outlet mercury concentration due to processing different mercury-containing waste types and normal variation in recovery/control equipment performance. As noted previously, we are proposing that performance tests for mercury thermal recovery units be conducted under the most challenging conditions, which we are defining as the processing of wastes that result in the highest recurring mercury concentration in the vent exhaust. Each performance test would consist of at least three runs, and the average concentration measured would be compared with the emission limit to determine compliance. Given our inability to establish a discernible correlation between waste type processed and emissions, our obligation to set standards that are achievable under the full range of normal acceptable operating conditions and the fact that initial performance is based on at least three separate test runs, we chose to set the standard based on the average of the three highest measured values in the data set of 134 measurements for the best-controlled plant. The three measured values are 20.4, 22.1, and 26.4 mg/m

3

. The average of the three is 23 mg/dscm, which we are proposing as the mercury concentration emission limit for oven type units.

Due to the very low volumetric flow rates associated with oven type mercury thermal recovery unit exhaust streams (typically less than 300 scfm), we believe that the retrofit of control equipment to reduce mercury emissions is both practical and reasonable. For purposes of estimating the impacts of the proposed emission limit, we assumed that the two plants with lower-performing control systems would need to install new, larger carbon adsorbers to meet the 23 mg/dscm level. The total installed capital control costs are estimated to be around $217,000 for all three plants, and the total annual control costs are estimated to be around $163,000 per year for all three plants. Estimated mercury emission reductions against actual baseline emissions would total 33 kg/yr (74 lbs/yr) for all three plants. The associated annual cost per unit of mercury emission reduction would be approximately $2,200 per pound.

Impacts on solid waste due to increased use of carbon adsorption are estimated total 5.2 Mg/yr (5.7 tpy) of mercury-containing spent carbon. Energy requirements are estimated to be an additional 473 thousand kW-hr/yr. Estimated secondary air pollution impacts due to heightened energy consumption are 152 Mg/yr (168 tpy), with carbon dioxide emissions comprising 99 percent of the estimate.

As noted previously, three plants operate retorts other than oven-type retorts. Thermal recovery at these three plants is conducted between 1,500 and 5,000 hours per year. The amounts of waste processed and the amounts of mercury recovered range from 50 to 500 tpy and from 3 to 12 tpy, respectively. The mercury recovery/control systems operated at the two plants with rotary kiln retorts consist of direct contact cooling, particulate and acid gas scrubbing, condensation, and carbon adsorption. The retort off-gas at both plants is cooled to a temperature of 55° F on average before being routed through two fixed-bed, nonregenerative adsorbers containing sulfur-impregnated carbon media. The mercury recovery/control system at the plant with a single hearth retort employs a direct contact water quench tower, a venturi scrubber, and a caustic packed-tower scrubber, which lower the retort off-gas temperature to an average of 80° F, and a chlorinated brine packed-tower scrubber as the final control device. The following summarizes the emissions data available and our approach to determining MACT for non-oven type units.

At one of the plants with a rotary kiln, the mercury concentration is determined daily at the outlet of the last carbon adsorber bed using a company-developed procedure derived from an OSHA method for determining worker exposures in the workplace. When submitting data obtained using this method, the company cautioned that although the routine sampling with the modified OSHA procedure produces credible information on relative changes in performance, it does not produce accurate information on actual mercury releases. Specifically, we believe the data obtained using this method are biased low. The average measured mercury concentration for this plant is an order of magnitude lower than averages for the other two plants (discussed below), and the minimum measured value is two orders of magnitude lower. It is our conclusion that data from this plant are unsuitable for standard setting, as they greatly understate emissions and thus overstate the performance of the mercury recovery/control system.

At the other plant with a rotary kiln, concentration measurements are made monthly using a method that is a modification of EPA Method 101 for determining mercury emissions from mercury cell chlor-alkali plants. Data were provided for each month in 1998. The measured mercury concentrations range from 1.4 mg/m

3

to 6.0 mg/m

3

, with a mean of 2.8 mg/m

3

.

Personnel at the plant with the single hearth retort conduct monthly measurements of the mercury concentration in the brine scrubber exhaust gas. The measurement method used is based on an EPA reference method and is very similar to the method used at the second rotary kiln plant discussed above. Data were provided for 1997, 1998, and 1999. The measured mercury concentrations range from 0.2 mg/m

3

to 10.8 mg/m

3

, with a mean and median value of 1.6 and 2.2 mg/m

3

, respectively.

In establishing the MACT floor and subsequently MACT, we focused on the two plants for which we have credible emissions data. We removed two points determined to be statistical outliers from the 3-year data set at the plant with the single hearth retort and determined there were no statistical outliers in the 1998 data set for the second plant with a rotary kiln. These data were used in the MACT determination for non-oven thermal recovery unit vents.

Although there are fewer than five sources from which to constitute a MACT floor, we opted to take the mean of the data from these two plants as the MACT floor option for existing sources. We averaged the three highest concentration data points for each plant and took the mean of the two plant averages (3.9 mg/dscm and 5.4 mg/dscm) rounded to one significant figure, 5 mg/dscm, as the floor value.

Of the three plants with non-oven type mercury thermal recovery unit vents, we project that only one plant would need to upgrade existing controls to meet the 5 mg/dscm floor level, and that this could be accomplished by replacing the carbon in its existing carbon adsorbers more frequently than current practice. There would be no capital costs as more frequent carbon media replacement is only a recurring annual cost estimated at $1,200 per year. Mercury emission reductions against actual baseline emissions would total about 2 kg/yr (5 lbs/yr) for the three plants. The associated annual cost per unit of mercury emission reduction would be approximately $240 per pound. With the assumption of more frequent carbon media replacement, there are no associated secondary air pollution, water pollution, or energy impacts. Estimated solid waste impacts due to increased use of carbon adsorption total 0.09 Mg/yr (0.1 tpy).

We then examined beyond-the-floor MACT options. A direct comparison of the data for the two plants providing credible data indicates that the emission levels recorded at one plant (with mean and median values of 1.2 and 0.7 mg/m

3

, respectively) are about half that recorded at the other plant (with mean and median values of 2.8 and 1.9 mg/m

3

, respectively). Further, the highest monthly values recorded were 4.3 mg/m

3

and 5.9 mg/m

3

, respectively. We used the data from the lower-emitting plant to establish a beyond-the-floor option. We averaged the three highest values for this plant (not including the values determined to be outliers) for a beyond-the-floor value of 4 mg/dscm.

Due to the very low volumetric flow rates associated with non-oven type mercury thermal recovery unit exhaust streams (typically less than about 2,000 scfm), we believe that the retrofit of control equipment to reduce mercury emissions is both practical and reasonable. For purposes of estimating impacts, we assumed that one plant would need to upgrade its controls, and that it would do this by further increasing its carbon replacement frequency to meet the 4 mg/dscm level. We assume that the remaining plant would not need to upgrade its existing controls to meet the beyond-the-floor level.

In evaluating regulatory options that are more stringent than the floor, we must consider the cost of achieving such emission reduction, and any non-air quality health and environmental impacts and energy requirements. The beyond-the-floor option would result in an additional 6 kg/yr (13 lbs/yr) of total mercury emission reductions for the three plants (a 10 percent incremental reduction from the floor option). The incremental annual costs are estimated to total around $5,800 per year. The incremental cost per unit of incremental mercury emission reduction is approximately $450 per pound. With the assumption of more frequent carbon media replacement, there are no associated incremental secondary air pollution, water pollution, or energy impacts. The estimated solid waste impacts total an additional 0.4 Mg/yr (0.5 tpy) of mercury-containing spent carbon.

We believe the additional emission reductions that would be achieved by the beyond-the-floor option are warranted. Further, we believe that the incremental costs of achieving such emission reductions, as well as incremental non-air environmental impacts and energy requirements, are reasonable for mercury. Therefore, we selected 4 mg/dscm as MACT for non-oven type mercury thermal recovery unit vents.

In summary, the proposed emission limits are 23 mg/dscm and 4 mg/dscm for oven type mercury thermal recovery unit vents and non-oven type mercury thermal recovery unit vents, respectively. We believe that both proposed limits are representative of the best-performing systems for each retort type based on available data and as such, each limit clearly meets our statutory safeguard regarding the minimum level of control allowed under the statute.

E. How did We Determine the Basis and Level of the Proposed Standards for New Sources?

Section 112(d)(3) of the CAA specifies that standards for new sources cannot be less stringent than the emission control that is achieved in practice by the best-controlled similar source, as determined by the Administrator.

In the case of mercury cell chlor-alkali production facilities, of the 43 chlor-alkali production facilities in operation in the U.S. at the time of this analysis, 32 use cell technologies other than mercury (23 use diaphragm cells and 9 use membrane cells). As explained further below, we consider these chlor-alkali facilities using non-mercury cell technology to be “similar sources,” and, as such, a suitable basis for the standard for new source MACT. Such a standard would effectively eliminate mercury emissions from new source chlor-alkali production facilities.

The impact of such a standard would be negligible given that in terms of cost, economic and air and non-air environmental impacts, we don't believe that a new mercury cell chlor-alkali plant would otherwise ever be constructed. No new mercury cell chlor-alkali plant has been constructed in the U.S. in over 30 years, and we have no indication of any plans for future construction. In addition, we believe that any future demand for new or replacement chlor-alkali production capacity would be met easily through the construction of new production facilities that do not use or emit mercury. Consequently, we believe it is appropriate to consider non-mercury cell facilities as similar sources and the prohibition of new mercury cell chlor-alkali production facilities achievable. Accordingly, we are proposing a complete prohibition on mercury emissions for new source MACT for mercury cell chlor-alkali production facilities. We are not proposing any initial and continuous compliance requirements related to this emission limit as we believe they are unnecessary since the emissions prohibition effectively precludes the new construction or reconstruction of a mercury cell chlor-alkali production facility.

As highlighted in the previous discussion on the selection of standards for existing sources, the emission levels achieved by the best-controlled sources were selected as the proposed existing source MACT levels for mercury recovery facilities. These best levels of control for point sources are 23 mg/dscm of exhaust from an oven type mercury thermal recovery unit vent, and 4 mg/dscm of exhaust from a non-oven type mercury thermal recovery unit vent. For fugitive emission sources, the best level of control identified is the work practice standard represented in

the beyond-the-floor option selected for proposal for existing sources.

In the case of mercury recovery facilities, we know of three plants that employ low emitting mercury recovery processes. These processes include chemical mercury recovery used at two plants and recovery in a batch purification still used at a third plant. Unlike thermal recovery units which are capable of treating a variety of waste types, the chemical recovery and the purification still processes have limited application. Both are suitable to treating only certain waste types, K106 wastes for the former and end-box residues for the latter. Plants using these nonthermal recovery processes transfer their remaining wastes off-site for treatment, which typically involves thermal recovery. Given this limitation, we do not believe that these nonthermal recovery processes qualify as a suitable basis for new source MACT. Consequently, for new source MACT for mercury recovery facilities, we are proposing numerical mercury emission limits consistent with that achieved by the best similar sources, 23 mg/dscm for oven type thermal recovery unit vent and 4 mg/dscm for non-oven type thermal recovery units.

F. How did We Select the Testing and Initial Compliance Requirements?

We selected the proposed testing and initial and continuous compliance requirements based on requirements specified in the NESHAP General Provisions (40 CFR part 63, subpart A). These requirements were adopted for mercury cell chlor-alkali plants to be consistent with other part 63 NESHAP. These requirements were chosen to ensure that we obtain or have access to sufficient information to determine whether an affected source is complying with the standards specified in the proposed rule.

The proposed rule would require initial and periodic compliance tests for determining compliance with the emission limits for by-product hydrogen streams and end-box ventilation system vents, and the emission limits for oven type and non-oven type mercury thermal recovery unit vents. The proposed rule would require the use of published EPA methods for measuring total mercury. Specifically, the proposed rule would allow the use of Method 101 or 101A (of appendix A of 40 CFR part 61) for end-box ventilation system vents and mercury thermal recovery unit vents and Method 102 for by-product hydrogen streams. Methods 101 and 102 were developed in the 1970's specifically for use at mercury cell chlor-alkali plants. Although Method 101A was developed to measure mercury emissions from sewage sludge incinerators, it is appropriate for use for end-box ventilation system vents and mercury thermal recovery unit vents.

The NESHAP General Provisions specify at § 63.7(e)(3) that each test consist of three separate test runs. The proposed rule would adopt this requirement. Further, the proposed rule would require that each test run be at least 2 hours long. This is the duration specified in Method 101 and referenced in Methods 101A and 102.

In the stack test data that were provided to us, there were numerous incidents where the results were reported as “less than” a certain level. We believe that this is primarily related to the sensitivity of the analytical instrument (that is, the absorption spectrophotometer) used to measure the amount of mercury in the collected sample. Method 101 states that the absorption spectrometer must be the “Perkin Elmer 303, or equivalent, containing a hollow-cathode mercury lamp and the optical cell * * * .” It is our understanding that this particular model is no longer commercially available, and that newer, more sensitive absorption spectrophotometers are available. We considered whether it was necessary to specify, either in the proposed rule or through a modification to the test method, that Perkin Elmer 303 did not have to be used. We concluded that the “or equivalent” language contained in Method 101 allows for the use of newer, more sensitive instruments and as a result, adding rule language or amending Method 101 was unnecessary.

Even with the 2-hour minimum test run period and the clarification that newer, more sensitive absorption spectrophotometers are allowed to be used, we remain concerned that quantifiable results of mercury emissions may not be obtained during performance tests. As a result, the proposed rule includes a requirement that the amount of mercury collected during each test run be at least 2 times the limit of detection for the analytical method used. This will assure that a reliably quantifiable amount of mercury is collected for each test run.

The emission limits for by-product hydrogen streams and end-box ventilation system vents are in the form of mass of mercury emissions per mass of chlorine produced. Therefore, criteria for the measurement of chlorine production during performance testing are also necessary. It is our understanding that instrumentation used to measure actual chlorine production, as well as the location and frequency of measurement, varies from plant to plant. Types of instruments used include rail car weigh scales, weigh cells on liquid storage tanks, and gas flow meters. Calibration procedures for these instruments are plant-specific and dependent on the involvement of third parties concerned with quantifying actual chlorine production for billing and other purposes. Moreover, at a given plant, an accurate value for actual chlorine production based on these measurements is generally obtained at the end of an operating month when mass balance calculations are performed to verify measurements.

For a compliance test run on the order of several hours, we, therefore, needed to rely on some other reasonable indicator of chlorine production. All mercury cell chlor-alkali plants measure the electric current through on-line mercury cells, also known as the cell line load or cell line current load, with a digital monitor that provides readings continuously. This cell line current load measurement can be used in conjunction with a theoretical chlorine production rate factor to obtain the instantaneous chlorine production rate. The theoretical factor is based on a statement of Faraday's Law that 96,487 Coulombs (Faraday's constant, where a Coulomb is a fundamental unit of electrical charge) are required to produce one gram equivalent weight of the electrochemical reaction product (chlorine). It is our understanding that chlorine production calculated in this manner would differ from the actual quantity produced at the plant by about 3 to 7 percent due to electrical conversion efficiency and reaction efficiency determined by equipment characteristics and operating conditions. We consider this degree of variability acceptable.

We, therefore, stipulate in the proposed rule that the cell line current load be continuously measured during a performance test run and that measurements be recorded at least every 15 minutes over the duration of the test run. We further specify equations for computing the average cell line current load and for calculating the quantity of chlorine produced over the test run.

In addition to the requirement to conduct performance tests to demonstrate compliance with the emission limits, owners or operators would be required to establish a mercury concentration operating limit for each vent as part of the initial compliance demonstration. Then, at least twice a permit term (at mid-term and renewal), they would conduct subsequent compliance demonstrations and at the same time reestablish

operating limit values. The proposed rule would require that these mercury concentration operating limits be determined directly from the concentration monitoring data collected concurrent with the initial performance test.

For the work practice standards, initial compliance is demonstrated by documenting and certifying that the standards are being met or will be met by submitting a washdown plan and by certifying that the plan is being followed or will be followed. This approach assures initial compliance by requiring the owner or operator to submit a certified statement in the Notification of Compliance Status report.

G. How Did We Select the Continuous Compliance Requirements?

For each of the proposed emission limits, which consist of the limits on mercury emissions from hydrogen streams, end-box ventilation systems, and thermal recovery units, we considered the feasibility and suitability of continuous emission monitors (CEM) as the means of demonstrating continuous compliance. While we were unable to identify any mercury cell chlor-alkali plant currently using a mercury CEM on any vent, we did determine that there are mercury CEM commercially available that may be suitable for use at mercury cell chlor-alkali plants. To date, most of the development work on mercury CEM has focused on the development of monitors for the continuous measurement of mercury air emissions from either coal-fired utility boilers or hazardous waste incinerators. Most mercury CEM are extractive monitors which extract a continuous or nearly continuous sample of gas, then transfer the gas to an instrument for spectroscopic analysis by way of either cold vapor atomic absorption or cold vapor atomic fluorescence.

These cold vapor techniques have similar limitations. Both detect mercury vapor only in its elemental form. To measure other forms of mercury vapor (

e.g.,

oxidized/inorganic/divalent mercury, such as mercuric chloride), the sampled gases must first pass through a converter which reduces any nonelemental mercury vapor present to the elemental form prior to analysis. None of the available monitors based on the cold vapor techniques are capable of measuring particulate or nonvapor phase mercury since the sample gas must be filtered to remove any particulate matter present prior to conversion and analysis. This would include elemental mercury condensed on particulate matter and any mercury compounds in particulate form. Monitors that are capable of measuring total vapor phase mercury range in price from $50,000 to $80,000. Simpler monitors that measure only elemental mercury vapor average about $10,000.

For the proposed emission limits for by-product hydrogen streams and end-box ventilation system vents, which are expressed in grams of mercury per megagram of chlorine produced, we evaluated two options: continuous compliance against the proposed gram per megagram standards, and continuous compliance against plant and vent specific operating limits expressed in terms of concentration. In addition to monitoring mercury concentration, the first option would require continuous monitoring of volumetric flow rate and a continuous, or at least periodic, measurement of chlorine production. The operating limits for the second option would be set at the time that initial compliance with the emission limit is demonstrated.

Since the predominant form of liquid mercury in mercury cells and other production facilities is elemental, we assumed that the mercury contained in the vent gas from either by-product hydrogen streams or end-box ventilation system vents is similarly largely in the elemental vapor form. Thus, the simpler, less expensive monitors for measuring elemental mercury vapor only should be suitable.

We concluded that monitoring only elemental mercury concentration provides a simpler, less expensive, and more reliable alternative to demonstrating continuous compliance than monitoring against the gram per megagram standards. As a result, we are proposing that continuous compliance for by-product hydrogen streams and end-box ventilation system vents be demonstrated through the continuous monitoring of elemental mercury concentration in the vent exhaust.

To the best of our knowledge, mercury contained in the exhaust gas of thermal recovery units, both oven and non-oven types, should exist as both vapor (elemental or nonelemental) and fine particulate matter. As highlighted above, none of the currently available monitors are capable of measuring particulate mercury. Consequently, continuous monitoring to demonstrate continuous compliance with the total mercury concentration limit would not be possible.

Similar to the by-product hydrogen streams and end-box ventilation system vents, we also considered the feasibility and usefulness of monitoring vapor phase mercury, specifically the elemental form. We concluded that the continuous monitoring of elemental mercury vapor as a surrogate to the total mercury emission limit using the simpler of the available monitors provides an acceptable and cost-effective means of tracking relative changes in emissions and control device performance. Therefore, as proposed for by-product hydrogen streams and end-box ventilation system vents, we are proposing for oven type and non-oven type mercury thermal recovery units that continuous compliance be demonstrated through continuous monitoring of elemental mercury concentration against an applicable concentration operating limit established as part of the initial compliance demonstration.

Another important aspect of continuous compliance is the time period over which continuous compliance is determined. One option would be an instantaneous period, where any measurement outside of the established range (that is, above the established concentration limit) would constitute a deviation. More commonly, the average of the monitoring data over a specified time period, for example an hour, is compared to the established limit.

While mercury cell chlor-alkali production facilities are generally operated continuously, there are process fluctuations that impact emissions. Mercury recovery facilities are operated intermittently, depending on the amount of mercury-containing waste to be treated and other factors. We believe that an emissions averaging period is necessary for both situations. We considered a daily averaging period and concluded that daily averaging would accommodate process variations while precluding avoidable periods of high emissions. Therefore, we are proposing a daily averaging period for demonstrating continuous compliance.

We also considered how to address monitoring data collected during startups, shutdowns, and malfunctions. We believe that it is important to continue to monitor the outlet mercury concentration during startups, shutdowns, and malfunctions to minimize emissions and to demonstrate that the plant's startup, shutdown, and malfunction plan is being followed. However, as provided for in the NESHAP General Provisions (40 CFR part 63, subpart A), we do not believe that the data collected during these periods should be used in calculating the daily average values. The emission limits were developed based on normal operation, and the performance tests will be conducted during representative operating conditions. Therefore, the

inclusion of monitoring data collected during startups, shutdowns, and malfunctions into the daily averages would be inconsistent with the data used to develop the emission limits and, subsequently, the mercury concentration operating limits.

While we did not identify situations in the mercury cell chlor-alkali industry where elemental mercury concentration is being continuously monitored, we believe that continuous elemental mercury concentration monitoring devices are available for use at mercury cell chlor-alkali plants. We recognize that the transfer of this monitoring technology to applications at mercury cell chlor-alkali plants will introduce uncertainties that can only be addressed through actual field demonstration. We are specifically requesting comment on the technical feasibility of using continuous elemental mercury concentration monitors for indicating relative changes in control system performance. We are also requesting comment on the proposed specifications for these devices.

Continuous compliance with the proposed work practice standards for the fugitive emission sources would be demonstrated by maintaining the required records documenting conformance with the standards and by maintaining the required records showing that the washdown plan was followed.

H. How Did We Select the Notification, Recordkeeping, and Reporting Requirements?

We selected the proposed notification, recordkeeping, and reporting requirements based on requirements specified in the NESHAP General Provisions (40 CFR part 63, subpart A). As with the proposed initial and continuous compliance requirements, these requirements were adapted for mercury cell chlor-alkali plants to be consistent with other part 63 national emission standards.

IV. Summary of Environmental, Energy, Cost, and Economic Impacts

A. What Are the Air Emission Impacts?

As discussed previously, the level of mercury emissions allowed by the Mercury NESHAP is 2,300 grams per day. If one assumes that all twelve plants in the source category emit mercury at this level and that each operates 365 days a year, total annual potential-to-emit baseline emissions would be 10,074 kg/yr (22,200 lb/yr). Annual potential-to-emit baseline emissions for fugitive emission sources would be 5,694 kg/yr (12,544 lb/yr), based on 1,300 grams per day assumed for each plant's cell room ventilation system when the eighteen design, maintenance, and housekeeping practices referenced in the Mercury NESHAP are followed. Annual potential-to-emit baseline emissions for by-product hydrogen streams, end-box ventilation system vents, and mercury thermal recovery unit vents would be 4,380 kg/yr (9,656 lb/yr), based on the remaining 1,000 grams per day allowed. We estimate that the proposed rule would reduce industrywide mercury emissions for by-product hydrogen streams, end-box ventilation system vents, and mercury thermal recovery unit vents from this annual potential-to-emit baseline to around 245 kg/yr (545 lb/yr), which is equivalent to about 94 percent reduction.

While the level of mercury emissions allowed by the Mercury NESHAP defines the potential-to-emit baseline, the sum of annual mercury emissions releases from by-product hydrogen streams, end-box ventilation system vents, and mercury thermal recovery vents, as estimated by mercury cell chlor-alkali plants, defines an annual actual baseline for vents of about 935 kg/yr (2,060 lb/yr). We estimate that the proposed rule would reduce industrywide mercury emissions for vents from this annual actual baseline to around 245 kg/yr (545 lb/yr), which is equivalent to about 74 percent reduction.

We estimate that secondary air pollution emissions would result from the production of electricity required to operate new control devices and new monitoring equipment assumed for plant vents. Assuming electricity production as based entirely on coal combustion for a worst-case scenario, we estimated plant-specific impacts for carbon dioxide, sulfur dioxide, nitrogen oxides, particulate matter, and carbon monoxide emissions. The total estimated secondary air impacts of the proposed requirements for point sources at the twelve mercury cell chlor-alkali plants is around 554 Mg/yr (611 tpy) for all pollutants combined, with carbon dioxide emissions comprising 99 percent of the estimate.

We are unable to quantify the primary air emission impacts associated with the proposed work practice standards, so no mercury emission reduction is assumed for fugitive emission sources. However, we believe strongly that the new and more explicit requirements contained in the proposed standards will in fact result in mercury emission reductions beyond baseline levels. Relative to secondary impacts, we expect that secondary air pollution emissions, principally carbon dioxide, would result from the production of electricity required to operate new monitoring equipment assumed for plant cell rooms. We estimate the secondary air impacts of the proposed rule for fugitive sources to be 17 Mg/yr (19 tpy).

B. What Are the Non-Air Health, Environmental, and Energy Impacts?

We do not expect that there will be any significant adverse non-air health impacts associated with the proposed standards for mercury-cell chlor-alkali plants.

We estimate that an increase in the amount of mercury-containing waters would result from the heightened use of packed tower scrubbing assumed for several plant vents. The total estimated water pollution impact of the proposed rule for point sources is about 1.8 million liters (466 thousand gallons) of additional wastewater per year. We estimate that an increase in the amount of mercury-containing solid wastes would result with the heightened use of carbon adsorption assumed for several plant vents. The total estimated solid waste impact of the proposed rule for point sources is about 34 Mg/yr (38 tpy) of additional mercury-containing spent carbon.

We are unable to quantify non-air environmental impacts associated with the proposed work practice standards, so no wastewater and solid waste impacts are assumed for fugitive emission sources.

We estimate that the proposed requirements for point sources would result in increased energy consumption, specifically additional fan power in conveying gas streams through new carbon adsorbers and new packed scrubbers assumed for certain plant vents and additional power consumed by new vent monitoring equipment. The total estimated energy impacts of the proposed requirements for point sources is about 1,724 thousand kW-hr/yr.

We estimate that the proposed requirements for fugitive sources would result in increased energy consumption required to operate new monitoring equipment assumed for plant cell rooms. The total estimated energy impacts of the proposed requirements for fugitive sources is about 53 thousand kW-hr/yr.

C. What Are the Cost and Economic Impacts?

For projecting cost impacts of the proposed rule on the mercury cell chlor-alkali industry, we estimate that all twelve plants would incur costs to meet the proposed work practice standards and the proposed monitoring,

recordkeeping, and reporting requirements. We estimate that ten plants would incur costs to meet the proposed emission limits for by-product hydrogen streams and end-box ventilation system vents, and three plants would incur costs to meet the proposed emission limits for mercury thermal recovery units. The total estimated capital cost of the proposed rule for the twelve mercury cell chlor-alkali plants is around $2.5 million, and the total estimated annual cost is about $2.2 million per year. Plant-specific annual costs in our estimate range from about $91,000 for the least-impacted plant to about $375,000 for the worst-impacted plant.

The purpose of the economic impact analysis is to estimate the market response of chlor-alkali production facilities to the proposed standards and to determine the economic effects that may result due to the proposed NESHAP. Chlor-alkali production jointly creates both chlorine and caustic, usually sodium hydroxide, in fixed proportions. Being joint commodities, the economic analysis considers the impacts of the proposed NESHAP on both the chlorine and sodium hydroxide markets.

The chlorine production source category contains 43 facilities, but only twelve facilities using mercury cells are directly affected by the proposed standards. These twelve facilities are located at twelve plants that are owned by eight companies. Although one of these twelve plants permanently closed due to reasons unrelated to this rulemaking, the following impacts are based on the twelve plants in operation at the time the analysis was conducted.

Chlor-alkali production in mercury cells leads to potential mercury emissions from hydrogen streams, end-box ventilation system vents, mercury thermal recovery units, and fugitive emission sources. The compliance costs for the proposed standards, therefore, relate to the purchase, installation, operation, and maintenance of pollution control equipment at the point sources, as well as the labor costs and overheads associated with observing work practices addressing fugitive emissions. The estimated total annual costs for the proposed NESHAP are $1.8 million. This cost estimate represents about 0.38 percent of the 1997 chlorine sales revenue for the twelve mercury cell chlor-alkali production facilities. Furthermore, the total annual costs represent only 0.01 percent of the revenues of owning the directly affected mercury cell chlor-alkali plants.

The economic analysis predicts minimal changes in industry outputs and the market prices of chlorine and sodium hydroxide as a result of the estimated control costs. The new market equilibrium quantities of chlorine and sodium hydroxide decrease by less than 0.1 percent. Equilibrium prices of chlorine and sodium hydroxide both rise by less than 0.1 percent due to the proposed standards. Based on these estimates, we conclude that the proposed standards are not likely to have a significant economic impact on the chlorine production industry as a whole or on secondary markets such as the labor market and foreign trade.

We perform an economic analysis to determine facility- and company-specific impacts. These economic impacts are measured by calculating the ratio of the estimated annualized compliance costs of emissions control for each entity to its revenues (

i.e.

, cost-to-sales ratio). After the cost-to-sales ratio is calculated for each entity, it is then multiplied by 100 to convert the ratio into percentages. Actual revenues at the facility level are not available, therefore, estimated facility revenues received from the sale of chlorine are used. Some of these facilities also produce caustic as potassium hydroxide, but the revenues from the sale of this product are not estimated. The twelve mercury cell chlor-alkali plants have positive cost-to-sales ratios. The ratio of costs to estimated chlorine sales revenue for these facilities range from a low of 0.16 percent to a high of 1.00 percent. The average cost-to-sales ratio for the twelve mercury process chlorine production facilities is 0.46 percent. More detailed economic analysis predicted minimal changes in chlorine production at each facility. Thus, overall, the economic impact of the proposed standards is minimal for the facilities producing chlorine.

The share of compliance costs to company sales are calculated to determine company level impacts. Since eight companies own the twelve affected facilities, all eight firms face positive compliance costs from the proposed NESHAP. The ratio of costs to estimated revenues range from a low of less than 0.01 percent to a high of 0.22 percent, and the average ratio of costs to company revenues is 0.06 percent. Again, more detailed economic analysis at the company level predicts little change in company output or revenues. So, at the company level, the proposed standards are not anticipated to have a significant economic impact on companies that own and operate the chlorine production facilities.

No facility or company is expected to close as a result of the proposed standards, and the economic impacts to consumers are anticipated to be minimal. The generally small scale of the impacts suggests that there will also be no significant impacts on markets for the products made using chlorine or sodium hydroxide. For more information, consult the economic impact analysis report entitled “Economic Impact Analysis for the Proposed Mercury Cell Chlor-Alkali Production NESHAP,” which is available in the docket to this rulemaking.

V. Solicitation of Comments and Public Participation

We seek full public participation in arriving at final decisions and encourage comments on all aspects of the proposed standards from all interested parties. You need to submit appropriate supporting data and analyses with your comments to allow us to make the best use of them. Be sure to direct your comments to the Air and Radiation Docket and Information Center, Docket No. A-2000-32 (see

ADDRESSES

).

VI. Administrative Requirements

A. Executive Order 12866, Regulatory Planning and Review

Under Executive Order 12866 (58 FR 51735, October 4, 1993), the EPA must determine whether the regulatory action is “significant” and, therefore, subject to review by the Office of Management and Budget (OMB) and the requirements of the Executive Order. The Executive Order defines “significant regulatory action” as one that OMB determines is likely to result in a rule that may:

(1) Have an annual effect on the economy of $100 million or more or adversely affect in a material way the economy, a sector of the economy, productivity, competition, jobs, the environment, public health or safety, or State, local, or Tribal governments or communities;

(2) Create a serious inconsistency or otherwise interfere with an action taken or planned by another agency;

(3) Materially alter the budgetary impact of entitlements, grants, user fees, or loan programs or the rights and obligations of recipients thereof; or

(4) Raise novel legal or policy issues arising out of legal mandates, the President's priorities, or the principles set forth in this Executive Order.

Pursuant to the terms of Executive Order 12866, it has been determined that the proposed rule is not a “significant regulatory action” because none of the listed criteria apply to this action. Consequently, this action was not submitted to OMB for review under Executive Order 12866.

B. Executive Order 13132 (Federalism)

Executive Order 13132, entitled “Federalism” (64 FR 43255, August 10, 1999), requires the EPA to develop an accountable process to ensure “meaningful and timely input by State and local officials in the development of regulatory policies that have federalism implications.” “Policies that have federalism implications” is defined in the Executive Order to include rules that have “substantial direct effects on the States, on the relationship between the national government and the States, or on the distribution of power and responsibilities among the various levels of government.” Under section 6 of Executive Order 13132, the EPA may not issue a rule that has federalism implications, that imposes substantial direct compliance costs, and that is not required by statute, unless the Federal government provides the funds necessary to pay the direct compliance costs incurred by State and local governments, or the EPA consults with State and local officials early in the process of developing the rule. The EPA also may not issue a rule that has federalism implications and that preempts State law unless the Agency consults with State and local officials early in the process of developing the rule.

If the EPA complies by consulting, Executive Order 13132 requires the EPA to provide to OMB, in a separately identified section of the preamble to the rule, a federalism summary impact statement (FSIS). The FSIS must include a description of the extent of the EPA's prior consultation with State and local officials, a summary of the nature of their concerns and the Agency's position supporting the need to issue the rule, and a statement of the extent to which the concerns of State and local officials have been met. Also, when the EPA transmits a draft final rule with federalism implications to OMB for review pursuant to Executive Order 12866, the EPA must include a certification from the Agency's Federalism Official stating that the EPA has met the requirements of Executive Order 13132 in a meaningful and timely manner.

The proposed rule will not have substantial direct effects on the States, on the relationship between the national government and the States, or on the distribution of power and responsibilities among the various levels of government, as specified in Executive Order 13132. The proposed rule is mandated by statute and does not impose requirements on States; however, States will be required to implement the rule by incorporating the rule into permits and enforcing the rule upon delegation. States will collect permit fees that will be used to offset the resource burden of implementing the rule. Thus, the requirements of section 6 of the Executive Order do not apply to the proposed rule. Although section 6 of Executive Order 13132 does not apply to the proposed rule, the EPA did consult with State and local officials in developing the proposed rule.

C. Executive Order 13175, Consultation and Coordination with Indian Tribal Governments

Executive Order 13175, entitled “Consultation and Coordination with Indian Tribal Governments” (65 FR 67249, November 6, 2000), requires the EPA to develop an accountable process to ensure “meaningful and timely input by tribal officials in the development of regulatory policies that have tribal implications.” “Policies that have tribal implications” is defined in the Executive Order to include regulations that have “substantial direct effects on one or more Indian tribes, on the relationship between the Federal government and the Indian tribes, or on the distribution of power and responsibilities between the Federal government and Indian tribes.”

The proposed rule does not have tribal implications. It will not have substantial direct effects on tribal governments, on the relationship between the Federal government and Indian tribes, or on the distribution of power and responsibilities between the Federal government and Indian tribes, as specified in Executive Order 13175. Thus, Executive Order 13175 does not apply to the proposed rule.

In the spirit of Executive Order 13175 and consistent with EPA policy to promote communications between EPA and tribal governments, EPA specifically solicits additional comment on the proposed rule from tribal officials.

D. Executive Order 13045, Protection of Children From Environmental Health Risks and Safety Risks

The Executive Order 13045 applies to any rule (1) that OMB determines is “economically significant,” as defined under Executive Order 12866, and (2) the EPA determines that the environmental health or safety risk addressed by the rule has a disproportionate effect on children. If the regulatory action meets both criteria, the EPA must evaluate the environmental, health, or safety aspects relevant to children and explain why the rule is preferable to other potentially effective and reasonably feasible alternatives considered by the EPA.

As with most rulemakings developed under section 112(d) of the CAA, today's proposal is based on MACT. Risks to public health and impacts on the environment are not typically considered in the development of emissions standards under section 112(d). Rather, these risks and impacts are considered later (within 8 years after promulgation of the MACT rule) under the residual risk program as required by section 112(f) of the CAA. While we do not believe the proposed rule to be “economically significant,” as defined under Executive Order 12866, we do beli

This text is long and has been trimmed here. Open the source document for the complete record.