National Emission Standards for Hazardous Air Pollutants; Proposed Standards for Hazardous Air Pollutants From Chemical Recovery Combustion Sources at Kraft, Soda, Sulfite, and Stand-Alone Semichemical Pulp Mills
Federal RegisterApr 15, 1998
Ask Donna
What actually matters in this document.
Text
SUMMARY: This action proposes national emission standards for hazardous
air pollutants (NESHAP) for the pulp and paper production source
category under section 112 of the Clean Air Act as amended (CAA). The
proposed standards focus on reducing hazardous air pollutants (HAP's)
from new and existing sources used in chemical recovery processes at
kraft, soda, sulfite, and stand-alone semichemical pulp mills. The
intent of the proposed standards is to protect the public health and
the environment by reducing HAP emissions to the level corresponding to
the maximum achievable control technology (MACT). The proposed
standards would reduce HAP emissions by about 2,600 megagrams per year
(Mg/yr) (2,800 tons per year [tons/yr]). In addition, emissions of
criteria pollutants such as particulate matter (PM) and volatile
organic compounds (VOC's) would be reduced by about 56,400 Mg/yr
(62,100 tons/yr).
DATES: Comments. The EPA will accept written comments on the proposed
rule until June 15, 1998.
Public Hearing. If requested, EPA will hold a public hearing
concerning the proposed rule beginning at 10 a.m. on May 15, 1998 at
the EPA Office of Administration Auditorium, Research Triangle Park,
North Carolina. Requests to present oral testimony must be made by May
6, 1998.
ADDRESSES: Requests to Speak at Hearing. Requests to present oral
testimony at the public hearing should be submitted to Ms. Cathy Coats,
Minerals and Inorganic Chemicals Group (MD-13), Emission Standards
Division, U.S. Environmental Protection Agency, Research Triangle Park,
NC 27711, telephone number (919) 541-5422. Persons interested in
attending the hearing should call Ms. Coats to verify that a hearing
will be held.
Comments. Interested parties may submit written comments (in
duplicate, if possible) to Public Docket No. A-94-67 at the following
address: U.S. Environmental Protection Agency, Air and Radiation Docket
and Information Center, 401 M Street, SW., Washington, DC 20460. The
EPA requests that a separate copy of the comments also be sent to the
contact person listed below in the FOR FURTHER INFORMATION CONTACT
section.
Comments may also be submitted electronically by sending electronic
mail (e-mail) to: [email protected]. Electronic comments
must be submitted as an ASCII file avoiding the use of special
characters and any form of encryption. Comments will also be accepted
on diskette in WordPerfect 5.1 or ASCII file format. All comments in
electronic form must be identified by the docket number (No. A-94-67).
No confidential business information should be submitted through e-
mail. Electronic comments may be filed online at many Federal
Depository Libraries.
FOR FURTHER INFORMATION CONTACT: Mr. Jeff Telander, Minerals and
Inorganic Chemicals Group, Emissions Standards Division (MD-13), U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina
27711, telephone number (919) 541-5427.
SUPPLEMENTARY INFORMATION:
Regulated Entities
Entities potentially regulated by this proposed rule are those
kraft, soda, sulfite, and stand-alone semichemical pulp mills with
chemical recovery processes that involve the combustion of spent
pulping liquor. Regulated categories and entities are listed below in
Table 1.
Table 1.--Regulated Categories and Entities
------------------------------------------------------------------------
Category Examples of regulated entities
------------------------------------------------------------------------
Industry.......................... Kraft pulp mills, soda pulp mills,
sulfite pulp mills, stand-alone
semichemical pulp mills.
------------------------------------------------------------------------
Table 1 is not intended to be exhaustive, but rather provides a
guide for readers regarding entities likely to be regulated by this
action. Table 1 lists the types of entities that EPA is now aware could
potentially be regulated by this action. Other types of entities not
listed in the table could also be regulated. To determine whether your
facility is regulated by this action, you should carefully examine the
applicability criteria in Sec. 63.860. 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.
Electronically Available Information
The preamble and the regulatory text for this proposed NESHAP for
chemical recovery combustion sources at kraft, soda, sulfite, and
stand-alone semichemical pulp mills are available on the Technology
Transfer Network (TTN), one of EPA's electronic bulletin boards. The
TTN provides a forum for technological and regulatory exchange in
various areas of air pollution control. The service is free, except for
the cost of a phone call. Dial (919) 541-5742 for up to a 14,400 baud
rate modem. If more information on the TTN is needed, call the TTN help
line at (919) 541-5384.
Docket
The docket (No. A-94-67) is available for public inspection and
copying from 8:30 a.m. to noon and from 1 to 3 p.m., Monday through
Friday at EPA's Air and Radiation Docket and Information Center,
Waterside Mall, Room M-1500 (ground floor), 401 M Street, SW.,
Washington, DC 20460.
The following documents and other supporting materials related to
this rulemaking are available for review in the docket center:
Technical Support Document: Chemical Recovery Combustion Sources at
Kraft and Soda Pulp Mills (docket entry No. II-A-31); Technical Support
Document: Chemical Recovery Combustion Sources at Sulfite Pulp Mills
(docket entry No. II-A-28); Profile of U.S. Stand-Alone Semichemical
Pulp Mills Memo (docket entry No. II-B-70); Nationwide Baseline HAP
Emissions for Combustion Sources at Stand-Alone Semichemical Pulp Mills
Memo (docket entry No. II-B-67); Nationwide Costs, Environmental
Impacts and Cost-Effectiveness of HAP Control Options for Combustion
Sources at Stand-Alone Semichemical Mills Memo (docket entry No. II-B-
69); the Nationwide Costs, Environmental Impacts, and Cost-
Effectiveness of Regulatory Alternatives for Kraft, Soda, Sulfite, and
Semichemical Combustion Sources Memo (docket entry No. II-B-63); the
Economic Analysis for the National Emission Standards for Hazardous Air
Pollutants for Source Category: Pulp and Paper Production; Effluent
Limitations Guidelines, Pretreatment Standards, and New Source
Performance Standards: Pulp, Paper, and Paperboard Category--Phase I
(docket entry No. II-A-32); the State of Washington PM Data for Kraft
Recovery Furnaces, Smelt Dissolving Tanks, and Lime Kilns Memo (docket
entry No. II-B-59); and the State of
[[Page 18755]]
Washington PM Data for Sulfite Combustion Units Memo (docket entry No.
II-B-40). Also, copies of this information may be obtained from the Air
Docket upon request by calling (202) 260-7548 or sending a FAX to (202)
260-4000. A reasonable fee may be charged for copies of docket
materials.
The information presented in the remainder of this preamble is
organized as follows:
I. Statutory Authority
II. Introduction
A. Background
B. NESHAP for Source Categories
C. Health Effects of Pollutants
D. Industry Profile
III. Summary of Proposed Standards
A. Applicability
B. Emission Limits and Requirements
1. PM HAP Standards for Kraft and Soda Pulp Mills
2. Total Gaseous Organic HAP Standards for Kraft and Soda Pulp
Mills
3. PM Standards for Sulfite Pulp Mills
4. Total Gaseous Organic HAP Standards for Stand-Alone
Semichemical Pulp Mills
C. Performance Test Requirements
D. Monitoring Requirements and Compliance Provisions
E. Recordkeeping and Reporting Requirements
IV. Rationale
A. Selection of Source Category
B. Selection of Emission Points
1. Emission Points--Kraft Pulp Mills
2. Emission Points--Soda Pulp Mills
3. Emission Points--Sulfite Pulp Mills
4. Emission Points--Stand-Alone Semichemical Pulp Mills
C. Selection of Definition of Affected Source
D. Selection of Pollutants
1. PM HAP's
2. Total Gaseous Organic HAP's
3. Hydrochloric Acid (HCl)
E. Determination of Subcategories and MACT Floors
1. MACT Floors--Kraft and Soda Pulp Mills
2. MACT Floors--Sulfite Pulp Mills
3. MACT Floors--Stand-Alone Semichemical Pulp Mills
F. Discussion of Regulatory Alternatives
1. Kraft and Soda Pulp Mills
2. Sulfite Pulp Mills
3. Stand-Alone Semichemical Pulp Mills
G. Selection of Proposed Standards for Existing and New Sources
1. Existing Sources
2. New Sources
H. Selection of Format of the Standards
1. PM HAP Standards for Kraft and Soda Pulp Mills
2. PM Standards for Sulfite Pulp Mills
3. Total Gaseous Organic HAP Standard for Kraft and Soda Pulp
Mills
4. Total Gaseous Organic HAP Standard for Stand-Alone
Semichemical Pulp Mills
I. Selection of Monitoring Requirements
J. Selection of Test Methods
K. Selection of Reporting and Recordkeeping Requirements
L. Relationship to Other Regulations
1. Noncombustion Source Rule and Chemical Recovery Combustion
Source Rule
2. NSPS (subpart BB of part 60) and Chemical Recovery Combustion
Source Rule
3. New Source Review/Prevention of Significant Deterioration
Applicability
M. Solicitation of Comments
V. Impacts of Proposed Standards
A. Number of Impacted Sources
B. Environmental Impacts
C. Energy Impacts
D. Cost Impacts
E. Economic Impacts
F. Benefits Analysis
VI. Administrative Requirements
A. Docket
B. Public Hearing
C. Executive Order 12866
D. Enhancing the Interdepartmental Partnership Under Executive
Order 12875
E. Unfunded Mandates Reform Act
F. Regulatory Flexibility
G. Paperwork Reduction Act
H. Clean Air Act
I. Statutory Authority
The statutory authority for this proposal is provided by sections
101, 112, 114, 116, and 301 of the Clean Air Act, as amended (42 U.S.C.
7401, 7412, 7414, 7416, and 7601).
II. Introduction
A. Background
On February 23, 1978, EPA promulgated new source performance
standards (NSPS) to limit emissions of PM and total reduced sulfur
(TRS) compounds from new, modified, and reconstructed kraft pulp mills
under authority of section 111 of the Act (43 FR 7568). In addition,
EPA issued retrofit guidelines in 1979 for control of TRS emissions at
existing kraft pulp mills not subject to the NSPS. The NSPS for kraft
pulp mills limit TRS emissions from recovery furnaces, smelt dissolving
tanks (SDT's), lime kilns, digesters, multiple effect evaporators,
black liquor oxidation (BLO) systems, brownstock washers, and
condensate strippers that were constructed, modified, or reconstructed
after September 24, 1976. The standards also limit PM emissions from
recovery furnaces, SDT's, and lime kilns that were constructed,
modified, or reconstructed after September 24, 1976. As required under
section 111(a) of the Act, these standards reflected the application of
the best technological system of continuous emission reduction that the
Administrator determined had been adequately demonstrated (taking into
consideration the cost of achieving such emission reduction, and any
nonair quality health and environmental impacts and energy
requirements).
Revisions to these standards were promulgated on May 20, 1986 (51
FR 18538). The revisions exempted BLO systems from the TRS standards;
revised the existing TRS limit and format of the standard for SDT's;
deleted the requirement to monitor the combustion temperature in lime
kilns, power boilers, and recovery furnaces; changed the frequency of
excess emission reports from quarterly to semiannually; and exempted
diffusion washers from the TRS standard for brownstock washers. The
revisions also required that monitored emissions be recorded and
specified the conditions [Sec. 60.284(e)] under which excess emissions
would not be deemed a violation of Sec. 60.11(d). Today's action does
not revise or change the TRS requirements of the NSPS. However, today's
standards do include PM emission limits, as a surrogate for measuring
PM HAP emissions, for combustion sources (existing and new) in the
chemical recovery area of the mill.
On December 17, 1993, EPA proposed (1) effluent limitations
guidelines and standards for the control of wastewater pollutants for
the pulp and paper industry and (2) NESHAP for noncombustion sources in
the pulp and paper industry (58 FR 66078), otherwise referred to as
``MACT I.'' The emission points covered in the proposed NESHAP for
noncombustion sources were limited to process units in the pulping and
bleaching processes (e.g., digesters, bleaching towers, and associated
tanks) and in the associated wastewater collection and treatment
systems at mills that chemically pulp wood fiber using kraft, sulfite,
soda, or semichemical methods. In March 1996, EPA proposed to include
for regulation additional noncombustion operations and mills not
covered under the December 17, 1993 proposal (e.g., mechanical pulping,
pulping of secondary fiber by nonchemical means, nonwood pulping, and
paper machines), otherwise referred to as ``MACT III'' (61 FR 9383).
The NESHAP for noncombustion sources and the effluent guidelines are
being promulgated as part of today's integrated rule, ``NESHAP for
Source Category: Pulp and Paper Production; Effluent Limitations
Guidelines, Pretreatment Standards, and New Source Performance
Standards: Pulp, Paper, and Paperboard Category.'' This proposed NESHAP
for chemical recovery combustion sources at kraft, soda, sulfite and
stand-alone semichemical pulp mills, otherwise referred to as ``MACT
II,'' does not revise or change the requirements of the
[[Page 18756]]
NESHAP for noncombustion sources that is being promulgated today.
B. NESHAP for Source Categories
Section 112 of the Act provides a list of 189 HAP's and directs EPA
to develop rules to control HAP emissions from both new and existing
major sources. The Act requires that the rules be established by
categories of emission sources considering all HAP's emitted, rather
than establishing rules based on the emission of a single pollutant
from a source category. The statute also requires that the standards
reflect the maximum degree of reduction in emissions of HAP's that is
achievable, taking into consideration the cost of achieving such
emission reduction and any nonair quality health and environmental
impacts and energy requirements. This level of control is commonly
referred to as MACT.
In addition, the Act sets out specific criteria to be considered
for establishing a minimum level of control and criteria (incremental
cost, energy impacts, etc.) for evaluating control options more
stringent than the minimum level of control. This minimum level of
control is commonly referred to as the MACT ``floor.'' The MACT floor
for new sources, as specified by the Act, is ``the emission control
that is achieved in practice by the best controlled similar source.''
The MACT floor for existing sources, as specified by the Act, is the
average emission limitation achieved by the best performing 12 percent
of existing sources in each category or subcategory of 30 or more
sources (CAA section 112(d)(3)). For smaller categories or
subcategories, the Act specifies that standards shall not be less
stringent than the average emission limitation achieved by the best
performing five sources in the category or subcategory. These floor
determinations are based on data available to the Administrator at the
time the standards are developed. The statutory provisions do not limit
how the standard is set, beyond requiring that it be applicable to all
sources in a category or subcategory and at least as stringent as the
MACT floor. The emission standards are to be reviewed and revised as
necessary no less often than every 8 years. Also, EPA may later
promulgate more stringent standards to address any unacceptable health
or environmental risk that remains after the imposition of controls
resulting from today's standards (CAA section 112(f)).
C. Health Effects of Pollutants
The Clean Air Act was created in part ``to protect and enhance the
quality of the nation's air resources so as to promote the public
health and welfare and the productive capacity of its population'' (CAA
section 101(b)(1)). Title III of the Act establishes a technology-based
control program to reduce stationary source emissions of HAP's. The
goal of section 112(d) is to apply such control technology to reduce
emissions and thereby reduce the hazard of HAP's emitted from
stationary sources.
This proposed rule is technology-based (i.e., based on MACT). The
MACT strategy avoids dependence on a risk-based approach as a pre-
requisite for regulating air toxics. Such risk assessments are limited
by incomplete information on what HAP's are emitted, what level of
emissions is occurring, what health and safety benchmarks are available
to assess risk, what health effects may be caused by certain
pollutants, and how best to model these effects, among other things.
Because of these issues, a quantitative risk assessment of potential
effects from all of the HAP's emitted from pulp and paper combustion
sources is not included in this rulemaking. However, as described in
section IV.D.3.d of this preamble, an exposure assessment was conducted
to determine if current emissions of hydrogen chloride (HCl) from pulp
and paper combustion sources result in exposures that provide an ample
margin of safety.
The EPA does recognize that the degree of adverse effects to health
can range from mild to severe. The extent and degree to which health
effects may be experienced is dependent upon (1) ambient concentrations
observed in the area, (2) duration of exposures, and (3)
characteristics of exposed individuals (e.g., genetics, age, pre-
existing health conditions, and lifestyle) which vary significantly
with the population. Some of these factors are also influenced by
source-specific characteristics (e.g., emission rates and local
meteorological conditions) as well as pollutant-specific
characteristics.
Available emission data, collected during development of this
proposed rule, show that metals, various organic compounds, and HCl are
the most significant HAP's emitted from pulp and paper combustion
sources. Following is a summary of the potential health and
environmental effects associated with exposures, at some level, to
these emitted pollutants.
Almost all metals appearing on the section 112(b) list are emitted
from pulp and paper combustion sources. These metals can cause a range
of effects, including mucous membrane effects (e.g., bronchitis,
decreased lung function), gastrointestinal effects, nervous system
disorders (from cognitive effects to coma or even death), skin
irritation, and reproductive and developmental disorders. Additionally,
several of the metals accumulate in the environment and in the human
body. Cadmium, for example, is a cumulative pollutant that can cause
kidney effects after cessation of exposure. Similarly, the onset of
effects from beryllium exposure may be delayed by months to years.
Further, some of the metal compounds have been classified by EPA as
known (e.g., arsenic and chromium (VI)) or probable (e.g., cadmium and
beryllium) human carcinogens.
All forms of mercury, a volatile metal, may be characterized as
quite toxic, with different health effects associated with different
forms of the pollutant. Methyl mercury is the most toxic form of
mercury to which humans and wildlife generally are exposed. Exposure to
methyl mercury occurs primarily through the aquatic food chain. The
target organ for methyl mercury toxicity in humans is the nervous
system. The range of neurotoxic effects can vary from subtle decrements
in motor skills and sensory ability to tremors, inability to walk,
convulsions, and death. Exposure to inorganic mercury is associated
with renal impairment. Some forms of mercury have also been classified
as possible human carcinogens. Exposure to mercury compounds can also
cause effects in plants, birds, and nonhuman mammals. Reproductive
effects are the primary concern for avian mercury poisoning.
Organic compounds emitted from pulp and paper combustion sources
include acetaldehyde, benzene, formaldehyde, methyl ethyl ketone,
methyl isobutyl ketone, methanol, phenol, styrene, toluene, and
xylenes. These organic compounds have a range of potential health
effects associated with exposure at some level. Some of the effects
associated with short-term inhalation exposure to these pollutants are
similar and include irritation of the eyes, skin, and respiratory tract
in humans; central nervous system effects (e.g., drowsiness, dizziness,
headaches, depression, nausea, irregular heartbeat); reproductive and
developmental effects; and neurological effects. Exposure to benzene
and methyl isobutyl ketone at extremely high concentrations may lead to
respiratory paralysis, coma, or death. Human health effects associated
with long-term inhalation exposure to the organic compounds listed
above may include mild symptoms such as nausea, headache, weakness,
insomnia, intestinal pain, and burning eyes; effects on the central
nervous system; disorders of the blood; toxicity to the immune
[[Page 18757]]
system; reproductive disorders in women (e.g., increased risk of
spontaneous abortion); developmental effects; gastrointestinal
irritation; liver injury; and muscular effects.
In addition to the noncancer effects described above, some of the
organic HAP's emitted from pulp and paper combustion sources have been
classified by EPA as either known (e.g., benzene) or probable (e.g.,
acetaldehyde and formaldehyde) human carcinogens.
Hydrogen chloride is an inorganic HAP which is highly corrosive to
the eyes, skin, and mucous membranes. Short-term inhalation of HCl by
humans may cause coughing, hoarseness, inflammation and ulceration of
the respiratory tract, as well as chest pain and pulmonary edema if
exposure exceeds threshold concentrations. Long-term occupational
exposure of humans to HCl has been reported to cause inflammation of
the stomach, skin, and lungs, and photosensitization.
The health and environmental effects associated with exposure to PM
and ozone are described in EPA's Criteria Documents, which support the
national ambient air quality standards (EPA 1996, ``Air Quality
Criteria for Ozone and Related Photochemical Oxidants,'' EPA-600/P-93-
004, RTP, NC; EPA 1996, ``Air Quality Criteria for Particulate
Matter,'' EPA-600/P-95-001, RTP, NC). Briefly, PM emissions have been
associated with aggravation of existing respiratory and cardiovascular
disease and increased risk of premature death. Volatile organic
compounds are precursors to the formation of ozone in the ambient air.
At ambient levels, human laboratory and community studies have shown
that ozone is responsible for the reduction of lung function,
respiratory symptoms (e.g., cough, chest pain, throat and nose
irritation), increased hospital admissions for respiratory causes, and
increased lung inflammation. Animal studies have shown increased
susceptibility to respiratory infection and lung structure changes.
Studies have shown that exposure to ozone can cause foliar injury
and disrupt carbohydrate production and distribution in plants. The
reduction in carbohydrate production and allocation can lead to reduced
root growth, reduced biomass or yield production, reduced plant vigor
(which can increase susceptibility to attack from insects and disease
and damage from cold), and diminished ability to successfully compete
with more tolerant species. These effects have been observed in native
vegetation in natural ecosystems and in a selected number of commercial
trees and agricultural crops.
D. Industry Profile
There are currently 122 kraft, 2 soda, 15 sulfite, and 14 stand-
alone semichemical pulp mills in the United States. The majority (52
percent) of kraft mills are located in the Southeastern United States.
The two soda pulp mills are located in Tennessee and Pennsylvania. The
majority of sulfite mills (67 percent) are located in Washington and
Wisconsin. Half of all stand-alone semichemical pulp mills are located
in the Midwestern United States.
The kraft process is the dominant pulping process in the United
States. The kraft and soda processes account for approximately 82
percent of all domestic pulp production; sulfite and stand-alone
semichemical processes account for approximately 2 and 6 percent of the
domestic pulp production, respectively.
Numerous HAP compounds are emitted from combustion sources in the
chemical recovery area at kraft, soda, sulfite, and stand-alone
semichemical pulp mills. The HAP compounds emitted in the largest
quantities are methanol and HCl. Methanol and HCl account for
approximately 70 percent of the total HAP's emitted from the chemical
recovery area.
All of the kraft, soda, sulfite, and stand-alone semichemical pulp
mills are believed to be major HAP emission sources (i.e., emissions
greater than or equal to 9.1 Mg/yr [10 tons/yr] for an individual HAP
or 23 Mg/yr [25 tons/yr] for total HAP's). In most cases, HAP emissions
from combustion sources in the chemical recovery area alone are
sufficient to characterize these mills as major sources.
III. Summary of Proposed Standards
A. Applicability
The proposed standards apply to all existing and new kraft, soda,
sulfite, and stand-alone semichemical pulp mills with chemical recovery
processes that involve the combustion of spent pulping liquor.
Specifically, the sources that are regulated by today's proposed
standards are--
(1) Nondirect contact evaporator (NDCE) recovery furnaces, direct
contact evaporator (DCE) recovery furnace systems, SDT's, and lime
kilns at kraft and soda pulp mills;
(2) Sulfite combustion units at sulfite pulp mills; and
(3) Semichemical combustion units at stand-alone semichemical pulp
mills.
All existing kraft and soda pulp mills have chemical recovery
processes that involve the combustion of spent pulping liquor. However,
several existing sulfite and stand-alone semichemical pulp mills do not
recover pulping chemicals by combusting spent liquor. Three of the 15
sulfite mills use a calcium-based sulfite process and do not have
chemical recovery combustion units and, thus, would not be impacted by
this proposed rule. One of the 14 stand-alone semichemical pulp mills
burns spent liquor in a power boiler and does not have chemical
recovery; therefore, that mill also would not be impacted by this
proposed rule.
B. Emission Limits and Requirements
Today's proposed standards would regulate PM HAP emissions and/or
total gaseous organic HAP emissions for chemical recovery combustion
sources in the pulp and paper source category. The proposed emission
standards are summarized in Table 2.
[[Page 18758]]
Table 2.--Summary of Proposed Emission Standards a
--------------------------------------------------------------------------------------------------------------------------------------------------------
PM HAP standard Alternate PM HAP standard Total gaseous organic HAP
------------------------------------ (``bubble'') standard
Subcategory Emission point --------------------------------------------------------------------
Existing New Existing New Existing New
--------------------------------------------------------------------------------------------------------------------------------------------------------
Kraft and soda............... Recovery PM PM Mill-specific PM No ``bubble'' No standard.... Total gaseous
furnaces (NDCE 0.10 gr/dscm 0.034 g/dscm or PM HAP alternate No standard b.. organic HAP
and DCE). (0.044 gr/dscf) (0.015 gr/dscf) emission limit standard for No standard b..
Smelt dissolving at 8% O2 OR PM at 8% O2. [kg/Mg (lb/ton) new sources. 0.012 kg/Mg
tanks. HAP PM BLS] based on (0.025 lb/ton)
Lime kilns...... 1.00E-03 kg/Mg 0.06 kg/Mg calculated BLS (as
(2.01E-03 lb/ (0.12 lb/ton) value of the measured by
ton) BLS. BLS. sum of the methanol).
PM PM individual No standard b.
0.10 kg/Mg 0.023 g/dscm emission limits No standard b.
(0.20 lb/ton) (0.010 gr/dscf) for recovery
BLS OR PM HAP at 10% O2. furnaces, smelt
dissolving
6.20E-05 kg/Mg tanks, and lime
(1.24E-04 lb/ kilns. See
ton) BLS. equations 1 and
PM 2 in section
0.15 g/dscm III.B.1.
(0.067 gr/dscf)
at 10% O2 OR PM
HAP
6.33E-03 kg/Mg
(1.27E-02 lb/
ton) CaO.
Sulfite...................... Sulfite PM PM Not applicable.. Not applicable. No standard b.. No standard b.
combustion 0.092 g/dscm 0.046 g/dscm
units. (0.040 gr/dscf) (0.020 gr/dscf)
at 8% O2. at 8% O2.
Stand-alone semi-chemical.... Semichemical No standard..... No standard..... Not applicable.. Not applicable. Total gaseous Total gaseous
combustion organic HAP organic HAP
units recovery
combustion 1.49 kg/Mg 1.49 kg/Mg
units. (2.97 lb/ton) (2.97 lb/ton)
BLS (as BLS (as
measured by measured by
THC) OR 90% THC) OR 90%
reduction. reduction.
--------------------------------------------------------------------------------------------------------------------------------------------------------
a BLS = black liquor solids; CaO = calcium oxide (lime); THC = total hydrocarbons; gr/dscf = grains per dry standard cubic foot; g/dscm = grams per dry
standard cubic meter; kg/Mg = kilograms per megagram; lb/ton = pounds per ton; O2 = oxygen.
b Emissions of total gaseous organic HAP's from these sources are regulated as part of the NESHAP for noncombustion sources at pulp and paper mills.
Hazardous air pollutants are proposed only for existing recovery
furnaces, SDT's, and lime kilns at kraft and soda pulp mills. Limits
for total gaseous organic HAP emissions are proposed for new kraft and
soda recovery furnaces and existing and new semichemical combustion
units. Either methanol or total hydrocarbons (THC), depending on the
subcategory, is used as a surrogate for total gaseous organic HAP
emissions. The emission standards for each subcategory are discussed in
the following sections by the pollutant regulated.
1. PM HAP Standards for Kraft and Soda Pulp Mills
Today's rule proposes PM HAP emission limits for existing recovery
furnaces, SDT's, and lime kilns at kraft and soda pulp mills. In
addition, PM emission limits are proposed as a surrogate for PM HAP
emission limits for both new and existing affected sources at kraft and
soda pulp mills. The EPA is using the term ``PM HAP'' in this preamble
to refer to the standards which can be measured either on a total PM
basis or on a HAP component of PM basis. For existing kraft and soda
recovery furnaces, SDT's, and gas-fired lime kilns, the proposed PM
emission limits are the same as the New Source Performance Standards
for Kraft Pulp Mills (43 FR 7568). Under today's proposed standards,
existing oil-fired lime kilns would be subject to a more stringent PM
standard than the NSPS requirements.
The proposed standards also would allow the use of a ``bubble
compliance alternative'' for determining compliance with the PM HAP
standard for existing sources at kraft and soda pulp mills. The bubble
compliance alternative would allow mills to set PM or PM HAP emission
limits for each existing affected source at the mill such that, if
these limits are met, the total emissions from all existing affected
sources would be less than or equal to a mill-specific bubble limit.
This mill-specific bubble limit is calculated based on the proposed
emission limits (referred to as reference concentrations or reference
emission rates) for each affected source and mill-specific gas flow
rates and process rates. Equation 1, below, would be used to calculate
the bubble limit based on PM emissions.
[[Page 18759]]
[GRAPHIC] [TIFF OMITTED] TP15AP98.039
Where:
ELPM=overall PM emission limit for all existing affected
sources at the kraft or soda pulp mill, kg/Mg (lb/ton) of black liquor
solids fired.
Cref,RF=reference concentration of 0.10 g/dscm (0.044 gr/
dscf) corrected to 8 percent oxygen for existing kraft or soda recovery
furnaces.
QRFtot=sum of the average gas flow rates measured
during the performance test from all existing recovery furnaces at the
kraft or soda pulp mill, dry standard cubic meters per minute (dscm/
min) (dry standard cubic feet per minute [dscf/min]).
Cref,LK=reference concentration of 0.15 g/dscm (0.067 gr/
dscf) corrected to 10 percent oxygen for existing kraft or soda lime
kilns.
QLKtot=sum of the average gas flow rates measured during the
performance test from all existing lime kilns at the kraft or soda pulp
mill, dscm/min (dscf/min).
F1=conversion factor, 1.44 minuteskilogram/daygram
(minkg/dg) (0.206 minutespound/daygrain
[minlb/dgr]).
BLStot=sum of the average black liquor solids firing rates
of all existing recovery furnaces at the kraft or soda pulp mill
measured during the performance test, megagrams per day (Mg/d) (tons
per day [tons/d]) of black liquor solids fired.
ER1ref,SDT=reference emission rate of 0.10 kg/Mg (0.20 lb/
ton) of black liquor solids fired for existing kraft or soda smelt
dissolving tanks.
Equation 2, below, would be used to calculate the total bubble
limit based on PM HAP emissions.
[GRAPHIC] [TIFF OMITTED] TP15AP98.040
Where:
ELPMHAP=overall PM HAP emission limit for all existing
affected sources at the kraft or soda pulp mill, kg/Mg (lb/ton) of
black liquor solids fired.
ERref,RF=reference emission rate of 1.00E-03 kg/Mg (2.01E-03
lb/ton) of black liquor solids fired for existing kraft or soda
recovery furnaces.
ERref,LK=reference emission rate of 6.33E-03 kg/Mg (1.27E-02
lb/ton) of CaO produced for existing kraft or soda lime kilns.
CaOtot=sum of the average lime production rates for all
existing lime kilns at the kraft or soda pulp mill measured as CaO
during the performance test, Mg CaO/d (ton CaO/d).
BLStot=sum of average black liquor solids firing rates of
all existing recovery furnaces at the kraft or soda pulp mill measured
during the performance test, Mg/d (ton/d) of black liquor solids fired.
ER2ref,SDT=reference emission rate of 6.20E-05 kg/Mg (1.24E-
04 lb/ton) of black liquor solids fired for existing kraft or soda
smelt dissolving tanks.
Owners or operators that choose to comply with the PM HAP standards
using the proposed bubble compliance alternative would be allowed to
meet either the PM bubble limit determined in Equation 1 or the PM HAP
bubble limit determined in Equation 2, but would not be required to
meet both bubble limits. The proposed bubble compliance alternative
would not be applicable to new sources. All new affected sources at
kraft and soda pulp mills would be required to meet the individual
emission limitations set for those sources. Also, owners or operators
of existing sources subject to the NSPS for kraft pulp mills would be
required to continue to meet the PM emission limits of that rule,
regardless of which option they choose for complying with today's PM
HAP standard.
Owners or operators that choose to comply with the PM HAP standards
using the proposed bubble compliance alternative would be required to
submit preliminary emission limits to the applicable permitting
authority for approval for each existing kraft or soda recovery
furnace, SDT, and lime kiln at the mill. Before the preliminary PM or
PM HAP emission limits would be approved, the owner or operator would
be required to submit documentation demonstrating that if the
preliminary emission limits for each emission source are met, the
entire group of affected sources would be in compliance with the mill-
wide allowable emission level. The allowable emission level would be
determined from the applicable bubble equation using the reference
concentrations and reference emission rates for each emission source
and source-specific factors for exhaust gas flow rates and process
rates. Once approved by the applicable permitting authority, the
emission limits would be incorporated in the operating permit for the
mill. Thereafter, the owner or operator of the kraft or soda pulp mill
would demonstrate compliance with the standards by demonstrating that
each recovery furnace, SDT, and lime kiln emitted less than or equal to
the approved emission limit for that source. In addition, the PM
emission limits for any existing recovery furnace, SDT, and lime kiln
subject to the NSPS for kraft pulp mills must be at least as stringent
as the PM emission limits established in the NSPS. An example of how
the bubble compliance alternative can be used to establish emission
limits for affected sources at an example mill is provided in the
docket (docket entry No. II-B-75).
With one exception, owners or operators that choose to comply with
the PM HAP standards using the proposed bubble compliance alternative
must include all existing sources in the bubble. Any existing affected
source that can be classified as a stand-by unit (i.e., a source that
operates for less than 6,300 hours during any calendar year) could not
be included as part of a bubble. Owners or operators of stand-by units
must accept either the proposed PM or proposed PM HAP emission limits
shown in Table 2 for those units. The EPA requests comments on the
proposal to exclude stand-by units from the proposed bubble compliance
alternative. Some have argued that stand-by units--especially units
operating less than 20 percent of the year--may be relatively expensive
to control. Thus, inclusion of stand-by units within a compliance
bubble may yield important cost savings by allowing a more stringent
control of other units to offset the relatively high cost emissions
from the stand-by unit. The EPA also requests comment on the proposed
definition of a stand-by unit as a unit operating less than 6,300 hours
in a calendar year.
2. Total Gaseous Organic HAP Standards for Kraft and Soda Pulp Mills
There are no standards under the proposed rule for total gaseous
organic
[[Page 18760]]
HAP's for existing NDCE recovery furnaces or DCE recovery furnace
systems. All new recovery furnaces at kraft and soda pulp mills would
be required to meet a total gaseous organic HAP limit, as measured by
methanol, of 0.012 kg/Mg (0.025 lb/ton) of black liquor solids fired.
3. PM Standards for Sulfite Pulp Mills
Existing sulfite combustion units would be required to meet a PM
emission limit of 0.092 g/dscm (0.040 gr/dscf) corrected to 8 percent
oxygen. New sulfite combustion units would be required to meet a PM
emission limit of 0.046 g/dscm (0.020 gr/dscf) corrected to 8 percent
oxygen.
4. Total Gaseous Organic HAP Standards for Stand-Alone Semichemical
Pulp Mills
All existing and new stand-alone semichemical pulp mills with
chemical recovery combustion units would be required to reduce total
gaseous organic HAP emissions (measured as THC) from these units by 90
percent, or meet a total gaseous organic HAP emission limit (measured
as THC) of 1.49 kg/Mg (2.97 lb/ton) of black liquor solids fired.
C. Performance Test Requirements
The following discussion identifies the test methods to be used for
compliance determinations.
Test Method 5, ``Determination of Particulate Emissions from
Stationary Sources'' (40 CFR part 60, appendix A)--in conjunction with
either the integrated sampling techniques of Test Method 3, ``Gas
Analysis for the Determination of Dry Molecular Weight'' (40 CFR part
60, appendix A) or Test Method 3A, ``Determination of Oxygen and Carbon
Dioxide Concentrations in Emissions from Stationary Sources'' (40 CFR
part 60, appendix A)--is the test method for determining compliance
with the PM emission standards for new and existing kraft and soda
recovery furnaces, SDT's, and lime kilns and for new and existing
sulfite combustion units. Test Method 17, ``Determination of
Particulate Emissions from Stationary Sources (In-Stack Filtration
Method)'' may be used as an alternative to Test Method 5 if a constant
value of 0.009 g/dscm (0.004 gr/dscf) is added to the results of Test
Method 17 and the stack temperature is no greater than 205 degrees
Centigrade ( deg.C) [400 degrees Fahrenheit ( deg.F)].
Test Method 29, ``Determination of Metals Emissions from Stationary
Sources'' (40 CFR part 60, appendix A) is the test method for
determining compliance with the PM HAP emission standards for existing
kraft and soda recovery furnaces, SDT's, and lime kilns. Test Method 29
also may be used as an alternative to Test Method 5 for measuring PM
emissions. The Agency also will allow operators or owners the option of
measuring all of the PM HAP's (except mercury) with Test Method 29 and
making a separate measurement of the mercury using Test Method 101A,
``Determination of Particulate and Gaseous Mercury Emissions from
Sewage Sludge Incinerators'' (40 CFR part 61, appendix A).
Test Method 308, ``Procedure for Determination of Methanol
Emissions from Stationary Sources'' is being promulgated today as part
of the final NESHAP for noncombustion sources at pulp and paper mills
and is the test method for determining compliance with the total
gaseous organic HAP emission limit for new kraft and soda NDCE recovery
furnaces that are not equipped with dry electrostatic precipitator
(ESP) systems and for DCE recovery furnace systems.
Test Method 25A, ``Determination of Total Gaseous Organic
Concentration using a Flame Ionization Analyzer'' (40 CFR part 60,
appendix A) is the test method for determining compliance with the
total gaseous organic HAP emission limit for new and existing
combustion sources at stand-alone semichemical pulp mills.
D. Monitoring Requirements and Compliance Provisions
Each owner or operator of an affected source would be required to
install, operate, calibrate, and maintain a continuous monitoring
system for each affected source. The owner or operator also would be
required to establish a range of values for each operating parameter
(associated with a process operation or with an emission control
device) to be monitored based upon values recorded during the initial
performance test or during qualifying previous performance tests using
the required test methods. If values from previous performance tests
are used to establish the operating parameter range, the owner or
operator would be required to certify that the control devices and
processes had not been modified subsequent to the testing upon which
the data used to establish the operating ranges were obtained. The
owner or operator could conduct multiple performance tests to establish
ranges of operating parameters. The owner or operator also could
establish expanded or replacement ranges during subsequent performance
tests. An exceedance of the operating parameters would occur when the
measured operating parameter levels, averaged over a specified time
period, are outside the established range for a predetermined duration.
However, with the exception of opacity exceedances, no more than one
exceedance would be attributed to an affected source during any given
24-hour period. The following paragraphs describe: (1) The operating
parameters to be monitored, (2) the averaging periods and frequency
with which these parameters should be monitored, (3) when corrective
action is required to return operating parameters to levels that are
within the established range, and (4) when operating parameter
exceedances constitute a violation of the standards.
Owners or operators of existing kraft or soda recovery furnaces
that are equipped with an ESP for PM or PM HAP control would be
required to install, calibrate, maintain, and operate continuous
opacity monitoring systems (COMS). The COMS would be required to
perform at least one cycle of sampling and analysis for each successive
10-second period and one cycle of data recording for each successive 6-
minute period. If 10 consecutive 6-minute average values of opacity
exceed 20 percent, the owner or operator would be required to initiate
the corrective actions contained in the mill's startup, shutdown, and
malfunction (SSM) plan. A violation would occur when 6 percent of the
6-minute average opacity values recorded during any 6-month reporting
period are greater than 35 percent.
Owners or operators of new kraft or soda recovery furnaces and new
or existing kraft or soda lime kilns that are equipped with ESP's for
PM or PM HAP control would also be required to install, calibrate,
maintain, and operate COMS. The COMS would be required to perform at
least one cycle of sampling and analysis for each successive 10-second
period and one cycle of data recording for each successive 6-minute
period. If 10 consecutive 6-minute average values of opacity are
greater than 20 percent, the owner or operator would be required to
initiate the corrective actions contained in the facility's SSM plan. A
violation would occur when 6 percent of the 6-minute average opacity
values within any 6-month reporting period are greater than 20 percent.
Owners or operators using wet scrubbers to meet the PM or PM HAP
emission limits for any kraft or soda recovery furnace, smelt
dissolving tank, or lime kiln or the PM limit for sulfite combustion
units would be required to install, calibrate, maintain, and operate a
continuous monitoring system capable of determining and permanently
recording the pressure drop and
[[Page 18761]]
scrubbing liquid flow rate at least once for each successive 15-minute
period. If any 3-hour average of the pressure drop or scrubbing liquid
flow rate falls outside the established range, the owner or operator
would be required to initiate the corrective actions included in the
facility's SSM plan. A violation would occur when six 3-hour average
values of either parameter are outside the established range during any
6-month reporting period.
Owners or operators using regenerative thermal oxidizers (RTO's) to
comply with the total gaseous organic HAP emission standard for
chemical recovery combustion units at stand-alone semichemical mills
would be required to establish a minimum RTO operating temperature that
indicates (1) at least a 90 percent reduction in HAP emissions
(measured as THC) or (2) outlet HAP emissions (measured as THC) of less
than or equal to 1.49 kg/Mg (2.97 lb/ton) of black liquor solids. To
ensure ongoing compliance, the owner or operator would be required to
install, calibrate, maintain, and operate a monitoring system to
measure and record the RTO operating temperature for each successive
15-minute period. If any 1-hour average of the operating temperature
falls below the minimum established temperature, the owner or operator
would be required to initiate the corrective actions contained in the
facility's SSM plan. A violation would occur when any 3-hour average of
the RTO operating temperature falls below the minimum established
temperature.
The owner or operator of an affected source that uses a wet
scrubber, ESP, or RTO to comply with today's standards may monitor
alternative operating parameters subject to prior written approval by
the applicable permitting authority.
The owner or operator of an affected source that is complying with
today's proposed standards through operational changes or by a control
device other than those described above would be required to submit a
plan proposing parameters to be monitored, parameter ranges, and
monitoring frequencies to be used to determine ongoing compliance,
subject to approval by the applicable permitting authority. If any 3-
hour average value of a monitored parameter falls outside the
established range, the owner or operator would be required to initiate
the corrective actions included in the facility's SSM plan. A violation
would occur when six 3-hour average values of a monitored parameter are
outside the established range during any 6-month reporting period.
Owners or operators complying with the total gaseous organic HAP
standard for new kraft and soda recovery furnaces through the use of an
NDCE recovery furnace equipped with a dry ESP system would not be
required to perform any continuous parameter monitoring for gaseous
organic HAP's; however, each owner or operator would be required to
maintain onsite a certification statement signed by a responsible mill
official that an NDCE recovery furnace equipped with a dry ESP system
is in use.
E. Recordkeeping and Reporting Requirements
In addition to all of the recordkeeping and reporting requirements
outlined in Sec. 63.10 of the General Provisions (subpart A of 40 CFR
part 63), owners or operators of kraft, soda, sulfite, and stand-alone
semichemical pulp mills would be required to maintain the following
records for each affected source: (1) Records of the black liquor
solids firing rates for all recovery furnaces at kraft and soda pulp
mills and spent liquor solids firing rates for all chemical recovery
combustion units at sulfite and stand-alone semichemical pulp mills;
(2) records of the lime production rates, calculated as CaO, for all
kraft and soda lime kilns; (3) records of all parameter monitoring
data; (4) records and documentation of supporting calculations for
compliance determinations; (5) records of the established monitoring
parameter ranges for each affected source; and (6) records of all
certifications made in order to determine compliance with the total
gaseous organic HAP standards. All records would have to be maintained
for a minimum of 5 years.
IV. Rationale
This section describes the rationale for the decisions made by the
Administrator in determining the proposed MACT floors for each source
category and in selecting the proposed standards.
A. Selection of Source Category
The list of source categories was published in the Federal Register
on July 16, 1992 and includes pulp and paper mills as major sources of
HAP's (57 FR 31576). Standards for the pulp and paper production source
category are being developed in phases. In December 1993, EPA proposed
the first set of emission standards for the source category (i.e., a
proposed NESHAP for noncombustion sources in the pulp and paper
industry, otherwise referred to as MACT I) as part of a ``cluster
rule'' that also included proposed effluent guidelines and standards
for the control of wastewater pollutants (58 FR 66078). In March 1996,
EPA proposed to include for regulation additional noncombustion
operations and mills not covered under the December 1993 proposal
(i.e., MACT III) (61 FR 9383). The NESHAP for noncombustion sources, as
well as the effluent guidelines and standards, are being promulgated as
part of today's cluster rule. An additional set of standards for the
source category is covered by today's proposed NESHAP for chemical
recovery combustion sources (i.e., MACT II). Today's proposed
``combustion sources'' NESHAP covers (1) combustion units in the
chemical recovery area at kraft, soda, sulfite, and stand-alone
semichemical pulp mills, (2) SDT's at kraft and soda pulp mills, and
(3) BLO systems at kraft pulp mills. Although kraft and soda SDT's and
kraft BLO systems are not combustion sources, these equipment are
included in today's proposed ``combustion sources'' NESHAP because they
are closely associated with the chemical recovery combustion equipment.
For the purposes of today's proposed standards, the combustion units,
SDT's, and BLO systems are collectively referred to as ``chemical
recovery combustion sources.'' Specifically, the chemical recovery
combustion sources are defined as (1) kraft and soda NDCE recovery
furnaces and DCE recovery furnace systems (which include BLO systems),
(2) kraft and soda SDT's, (3) kraft and soda lime kilns, (4) sulfite
combustion units, and (5) semichemical combustion units.
B. Selection of Emission Points
The following section identifies the HAP emission points for kraft,
soda, sulfite and stand-alone semichemical pulp mills that were
examined by the Agency for control under the proposed rule. General
descriptions of the chemical recovery process and equipment also are
included in this section. More detailed information on the emission
points and chemical recovery process can be found in the technical
support documents listed under the ADDRESSES section.
1. Emission Points--Kraft Pulp Mills
Emission points at kraft pulp mills that were examined by the
Agency for control under the proposed standards are NDCE recovery
furnaces and DCE recovery furnace systems, SDT's, and lime kilns. These
emission points are integral parts of the kraft chemical recovery
process, in which cooking liquor chemicals (i.e., sodium hydroxide
[NaOH] and sodium sulfide [Na2S]) are recovered from spent
cooking liquor. Cooking liquor, which is used in the pulping process,
is commonly referred
[[Page 18762]]
to as white liquor; spent cooking liquor is commonly referred to as
black liquor.
a. NDCE Recovery Furnaces and DCE Recovery Furnace Systems. There
are an estimated 209 recovery furnaces operating at U.S. kraft pulp
mills. The kraft recovery furnace is essentially a chemical recovery
unit and steam generator that uses black liquor as its fuel. More
specifically, the kraft recovery furnace (1) recovers inorganic pulping
chemicals from black liquor as smelt by reducing sodium sulfate
(Na2SO4) to Na2S and (2) combusts
organic compounds in black liquor to produce steam for mill processes.
Kraft recovery furnaces can be classified based on the type of
final-stage evaporator used to increase the solids content of black
liquor prior to firing in the furnace. The final-stage evaporator,
which follows the multiple-effect evaporator (MEE), may be either an
NDCE or DCE. Direct contact evaporators use flue gases from the
recovery furnace to concentrate the black liquor. In the 1970's, as
energy costs increased and Federal and State regulations were passed
that limited TRS emissions from kraft pulp mills, the use of NDCE's (or
concentrators) became more prevalent. By using an NDCE, the heat that
was formerly used to concentrate black liquor in the DCE can be used to
produce steam by extending the economizer section of the furnace, and
the TRS emissions (associated with the DCE) will be decreased. For
newer recovery furnaces, all of which use NDCE's, the NDCE is often
considered an integral part of the MEE. Approximately 61 percent of
kraft recovery furnaces are NDCE recovery furnaces, and 39 percent are
DCE recovery furnace systems. For the purposes of today's proposed
rule, an ``NDCE recovery furnace'' is defined as a recovery furnace
that is equipped with an NDCE that concentrates black liquor by
indirect contact with steam. A ``DCE recovery furnace system'' is
defined to include a DCE recovery furnace and any BLO system, if
present, at the pulp mill; a ``DCE recovery furnace'' is defined as a
recovery furnace that is equipped with a DCE that concentrates strong
black liquor by direct contact between the hot recovery furnace exhaust
gases and the strong black liquor.
All kraft recovery furnaces have a PM control device, typically an
ESP. The PM collected in the ESP, which is predominantly
Na2SO4, is returned to the concentrated black
liquor that is fired in the recovery furnace. The mechanism for
returning the PM to the black liquor may be a dry system or may use
either black liquor or process water.
In DCE recovery furnace systems, black liquor is oxidized prior to
evaporation in the DCE. Black liquor oxidation reduces emissions of TRS
compounds, which are stripped from black liquor in the DCE when the
black liquor contacts hot flue gases from the recovery furnace. Black
liquor can be oxidized using either air or pure (molecular) oxygen.
Air-sparging units operate by bubbling air through the black liquor
using multiple diffuser nozzles. Air-sparging units have from one to
three tanks (or stages) that operate in series and a corresponding
number of emission points. At two mills, vent gases from air-sparging
BLO units are routed to a power boiler to reduce TRS emissions via
incineration. Molecular oxygen BLO systems resemble pipeline reactors
and require relatively short residence times (i.e., 30 seconds to 5
minutes compared to 1 or more hours for air-sparging units). Because
all of the oxygen is consumed in the reaction, no system vent is
required with molecular oxygen BLO in-line reactors, and therefore, no
emission point is associated with these systems. There are an estimated
46 BLO systems operating at kraft pulp mills. Mills with multiple DCE
recovery furnaces have one BLO system. At present, only four mills
(with seven DCE recovery furnaces) use a molecular oxygen BLO system.
The emission potential for DCE recovery furnace systems is higher
than that for NDCE recovery furnaces because of the increased
opportunity to strip HAP compounds from the black liquor in the process
equipment. In the DCE recovery furnace system, gaseous organic HAP
compounds can be stripped from the black liquor in the air-sparging BLO
system and in the DCE. Similarly, the emission potential for NDCE
recovery furnaces with ESP's that use black liquor or HAP-contaminated
process water in the ESP bottom or PM return system is higher than that
for NDCE recovery furnaces that have dry ESP systems (i.e., dry-bottom
ESP's and dry PM return systems). As with the air-sparging BLO systems
and DCE's, stripping of gaseous organic HAP compounds can occur if
black liquor or HAP-contaminated process water is used in the bottom of
the ESP or in the PM return system.
In addition to the criteria pollutants (i.e., PM, NOX,
SO2, CO, and VOC [ozone precursor]) and TRS, the compounds
emitted in the largest quantities from NDCE recovery furnaces and DCE
recovery furnace systems are methanol and HCl. For a given process
emission rate, the total gaseous organic HAP emissions from DCE
recovery furnace systems are, on average, approximately 14 times higher
than NDCE recovery furnaces with dry ESP systems. Also, for a given
process emission rate, the total gaseous organic HAP emissions from
NDCE recovery furnaces with wet ESP systems (i.e., ESP's that use black
liquor or HAP-contaminated process water in the ESP bottom or PM return
system) are, on average, approximately 3.5 times higher than NDCE
recovery furnaces with dry ESP systems. Of the total gaseous organic
HAP's emitted, methanol emissions account for approximately 67 percent
of emissions from DCE recovery furnace systems and 13 percent of
emissions from NDCE recovery furnaces with dry ESP systems.
For a given process emission rate, HCl emissions are approximately
equivalent for both NDCE recovery furnaces and DCE recovery furnace
systems. Hydrogen chloride emissions account for approximately 19
percent of the total gaseous HAP emissions from DCE recovery furnace
systems and 76 percent of the total gaseous HAP emissions from NDCE
recovery furnaces with dry ESP systems.
Particulate matter HAP's account for approximately 0.2 percent of
the PM emissions and 0.3 percent of the total HAP emissions from
recovery furnaces. Although the PM inlet loadings to the PM control
devices for NDCE recovery furnaces are higher than for DCE recovery
furnaces due to removal of 20 to 40 percent of the PM in the DCE unit,
equivalent outlet PM emissions can be achieved with the use of add-on
controls.
b. Smelt Dissolving Tanks. There are an estimated 227 SDT's at U.S.
kraft pulp mills. This estimate is higher than the estimated number of
recovery furnaces because some furnaces have two SDT's. The SDT is a
large, covered vessel located below the recovery furnace and is the
discharge point for molten smelt, which is the main product from the
combustion of black liquor. Smelt, which is predominantly sodium
carbonate (Na2CO3) and Na2S, filters
through the char bed at the bottom of the recovery furnace and is
continuously discharged through water-cooled spouts into the SDT. As
the smelt exits the water-cooled spouts, the smelt stream is shattered
with medium-pressure steam so that it can be safely dissolved in the
SDT. In the SDT, smelt is dissolved in weak wash water from the
recausticizing area to form unclarified green liquor, an aqueous
solution of Na2CO3 and Na2S.
Large volumes of steam are generated when the smelt is quenched in
the SDT. Residual water vapor and PM generated
[[Page 18763]]
during quenching are drawn off the tank through a venturi scrubber or
other PM control device using an induced-draft fan. Particulate matter
HAP's account for approximately 0.06 percent of the PM emissions from
SDT's. The water used in the scrubber, which is typically weak wash,
drains directly into the SDT. Gaseous organic HAP compounds (primarily
methanol) also are emitted from SDT's as a result of the use of weak
wash in the SDT and PM control device. Because of the elevated
operating temperature of the SDT, gaseous organic HAP compounds present
in the weak wash can volatilize and subsequently be released to the
atmosphere.
c. Lime Kilns. An estimated 190 lime kilns operate at U.S. kraft
pulp mills. The lime kiln is part of the recausticizing process in
which green liquor from the SDT is converted to white liquor.
Specifically, Na2CO3 in the green liquor is
converted to NaOH, a main constituent of white liquor, by adding
reburned lime (CaO) from the lime kiln. The resulting white liquor
solution contains NaOH, Na2S, and calcium carbonate
(Ca2CO3) precipitate (referred to as ``lime
mud''). Lime mud is removed from this solution in a white liquor
clarifier. The lime mud is then washed, dewatered, and calcined in a
lime kiln to produce reburned lime, which is recycled back to the green
liquor.
Most kilns in use at kraft pulp mills are large rotary kilns (98
percent); a few fluidized-bed calciners are also used. Natural gas or
fuel oil typically provides the energy for the calcining process. The
majority of lime kilns at kraft pulp mills also burn noncondensible gas
streams (NCG's) from various process vents, such as digester and
evaporator vents.
Lime kiln exhaust gases consist of combustion products, carbon
dioxide released during calcination, water vapor evaporated from the
mud, and entrained lime dust. Particulate in the exhaust gases is
mainly CaO, Ca2CO3, and sodium salts.
Approximately 1.4 percent of the PM emissions from lime kilns is PM
HAP's. Exhaust gases are routed through a PM control device prior to
being discharged to the atmosphere. Venturi scrubbers and ESP's are the
two most common types of PM control devices used to control PM
emissions from lime kilns.
As with SDT's, gaseous organic HAP compounds (primarily methanol)
also are emitted from lime kilns due primarily to the use of weak wash
as the scrubbing liquor in the PM control device and lime mud washer.
Because of the elevated gas stream temperature, gaseous organic HAP
compounds present in the weak wash can volatilize and subsequently be
released to the atmosphere.
2. Emission Points--Soda Pulp Mills
Emission points at soda pulp mills that were examined by the Agency
for control under today's proposed standards are recovery furnaces,
SDT's, and lime kilns. The processes and equipment used in the chemical
recovery areas of soda and kraft pulp mills are similar, except that
the soda process, because it is a nonsulfur process, does not require
black liquor oxidation. With the exception of sulfur-containing
compounds, the types and quantities of compounds emitted from soda pulp
mills are comparable to the types and quantities of compounds emitted
from kraft pulp mills. There are only two soda pulp mills in the United
States, and no new soda mills are expected to be constructed. There are
a total of two recovery furnaces (one NDCE and one DCE), two SDT's, and
two lime kilns at the soda mills.
3. Emission Points--Sulfite Pulp Mills
The emission point at sulfite pulp mills that was examined by the
Agency for control under the proposed standard is the chemical recovery
combustion unit. The chemical recovery combustion unit is an integral
part of the chemical recovery process, which recovers cooking liquor
chemicals from spent cooking liquor (also called red liquor). The types
of chemical recovery combustion units used at sulfite mills are
recovery furnaces, fluidized-bed reactors, and combustors. There are 18
recovery furnaces, 2 fluidized-bed reactors, and 1 combustor operating
at sulfite pulp mills. For the purposes of today's proposed rule, these
various combustion units are collectively referred to as ``sulfite
combustion units.''
The process and equipment used to recover sulfite cooking liquor
chemicals depend on the chemical base of the cooking liquor. Sulfite
cooking liquors use one of four chemical bases--magnesium (Mg), ammonia
(NH3), calcium (Ca), or sodium (Na). Cooking liquor
chemicals can be recovered for the Mg-, NH3-, and Na-based
sulfite processes. Recovery of cooking liquor chemicals is not
practical for the Ca-based sulfite process, and, therefore, no sulfite
combustion units are used at the existing Ca-based sulfite mills.
Additionally, there are currently no operating Na-based sulfite mills.
There are currently six Mg-based sulfite mills and six NH3-
based sulfite mills. Information on the sulfite combustion units at Mg-
and NH3-based sulfite pulp mills follows.
At the six Mg-based sulfite mills, red liquor is fired in a
recovery furnace or fluidized-bed reactor. There are nine recovery
furnaces and two fluidized-bed reactors. Multiple-effect evaporators,
which may be followed by a DCE or NDCE, are used to increase the solids
content of the red liquor prior to firing in the combustion unit.
Magnesium-based sulfite combustion units differ from kraft recovery
furnaces in that there are no smelt beds. Combustion of the spent
liquor produces both heat for steam generation and exhaust gases that
contain magnesium oxide (MgO) particulate and SO2 gas. When
a recovery furnace is used, the major portion of the MgO is recovered
as a fine white powder from the exhaust gases using multiple cyclones.
When a fluidized-bed reactor is used, MgO from the exhaust gases is
collected in a cyclone and from the bed of the reactor as pulverized
bed material. The MgO from the recovery furnace or fluidized-bed
reactor is then slaked with water to form magnesium hydroxide
(Mg(OH)2), which is used as circulating liquid in a series
of absorption towers and/or venturi scrubbers designed to recover
SO2 from combustion gases. In the absorption towers/venturi
scrubbers, SO2 is recovered by reaction with
Mg(OH)2 to form a magnesium bisulfite solution. The
magnesium bisulfite solution is then fortified with makeup
SO2 and subsequently used as cooking liquor. Some mills have
installed air pollution control devices, such as a fiber-bed demister
system or an educted venturi scrubber, downstream of the SO2
absorption equipment, to further reduce PM and/or SO2
emissions.
At the six NH3-based sulfite pulp mills, red liquor is
fired in a recovery furnace or combustor. There are nine recovery
furnaces and one combustor. The solids content of the red liquor is
increased using MEE's, which may be followed by a DCE or NDCE.
Combustion of the spent liquor produces both heat for steam generation
and combustion gases that contain recoverable SO2. The
ammonia base is consumed during combustion, forming nitrogen and water.
A small amount of ash is produced and periodically removed from the
furnace bottom. (There are no smelt beds.) Sulfur dioxide is recovered
from cooled flue gas in an acid-gas absorption tower to form an
ammonium bisulfite solution. Fresh aqueous NH3 is used as
the circulating liquor in the absorption system. The ammonium bisulfite
solution is fortified with makeup SO2 and used as cooking
liquor. Exit gases from the absorption system are typically
[[Page 18764]]
routed to a fiber-bed demister system for PM removal and mist
elimination prior to being discharged to the atmosphere. Some mills
have installed a scrubber or mesh-pad mist eliminator upstream of the
fiber-bed demister system for additional PM and SO2 emission
control and to improve the efficiency and operation of the fiber-bed
demister system.
4. Emission Points--Stand-Alone Semichemical Pulp Mills
The emission point at stand-alone semichemical pulp mills that was
examined for control under today's proposed standards is the chemical
recovery combustion unit. The combustion unit is used in the chemical
recovery process to recover the inorganic cooking chemicals, produce
steam, and remove the organic compounds in the black liquor by
combustion. Cooking liquor chemicals are recovered as either smelt or
ash, which is dissolved in water and mixed with make-up cooking
chemicals to form white liquor.
There are 14 chemical recovery combustion units currently operating
at stand-alone semichemical pulp mills. Five different types of
chemical recovery combustion units are in operation: fluidized-bed
reactors, recovery furnaces, smelters, rotary liquor kilns, and
pyrolysis reactors. For the purposes of today's standards, these
various combustion units are collectively referred to as ``semichemical
combustion units.''
a. Fluidized-Bed Reactors. Seven fluidized-bed reactors are
currently in use at seven stand-alone semichemical pulp mills.
Fluidized-bed reactors are used extensively because the recovered
chemicals are in the form of solid pellets, which can be stored in
silos until the chemicals are needed to make fresh cooking liquor. This
practice requires less storage space than when recovered chemicals are
routed directly to a dissolving tank and stored in solution.
In the fluidized-bed reactor, concentrated black liquor is fired
from a single spray gun located at the top of the reactor. As the
liquor falls towards the bed, evaporation and some combustion occurs,
causing the liquor to pelletize. Fluidizing gas rises through the bed
of solid pellets, setting the bed in fluid motion. The soda ash
(Na2CO3) pellets are recovered from the reactor
and stored in silos.
b. Recovery Furnaces. Two NDCE recovery furnaces are currently in
use at two stand-alone semichemical pulp mills. Semichemical recovery
furnaces, like kraft recovery furnaces, are used to recover cooking
liquor chemicals by burning concentrated black liquor and to produce
process steam with the heat of combustion. Semichemical and kraft
recovery furnaces are similar in design.
c. Smelters. Two smelters are currently in use at a nonsulfur-
based, stand-alone semichemical pulp mill. Smelters operate in a manner
similar to recovery furnaces, except that smelters do not produce
excess steam for mill processes and are actually net users of heat. The
units currently in use are actually converted small kraft recovery
furnaces.
d. Rotary Liquor Kilns. Two rotary liquor kilns are currently in
use at two nonsulfur-based, stand-alone semichemical pulp mills. Unlike
lime kilns used in the kraft chemical recovery process, rotary liquor
kilns are used for the combustion of black liquor at semichemical pulp
mills. In the kiln, fuel oil is burned in the lower end. An induced-
draft fan at the upper end draws combustion air into the lower end and
draws combustion gases through the kiln. Approximately halfway between
the lower and upper ends, black liquor is fired into the kiln. Sodium
carbonate ash created from contact between black liquor and combustion
gases falls to the lower end of the kiln, then is routed to an ash
dissolving tank. The combustion gases are routed to a waste heat boiler
to produce steam.
e. Pyrolysis Reactor. One pyrolysis reactor is currently in use at
a stand-alone semichemical pulp mill. ``Pyrolysis'' means chemical
change caused by heat, not by combustion. In the pyrolysis reactor,
fuel oil or propane is burned to provide the heat for pyrolysis. Black
liquor is injected under high pressure in a finely atomized spray
through several nozzles arranged around the wall of the pyrolysis
chamber. The hot combustion gases travel downward at high velocity and
contact the liquor sprays at high turbulence and rapid mixing.
Pyrolysis reactions occur, converting the sodium in the liquor into
a solid ash powder composed mainly of soda ash
(Na2CO3), and the other constituents into a
gaseous mixture of hydrogen sulfide (H2S) mixed with CO,
carbon dioxide (CO2), hydrogen (H2), methane
(CH4), nitrogen (N2), and water vapor.
f. HAP Emissions from Semichemical Combustion Sources. Test data
indicate that chemical recovery combustion units at stand-alone
semichemical pulp mills are significant sources of gaseous organic HAP
emissions. The major HAP compounds emitted from chemical recovery
combustion units are methanol, benzene, methyl ethyl ketone,
formaldehyde, and toluene. The fluidized-bed reactors emit the highest
quantities of HAP's, while emissions from other semichemical combustion
unit types (e.g., recovery furnaces and rotary liquor kilns) are much
lower. For example, based on available HAP emissions data, the
fluidized-bed reactors have total HAP emissions approximately 20 to 75
times higher per ton of black liquor solids fired than the other
semichemical combustion unit types. Some of the other semichemical
combustion unit types (e.g., recovery furnaces and rotary liquor kilns)
are inherently lower-emitting because they achieve more complete
combustion of organic compounds. (No HAP emission data were available
for the pyrolysis unit; however, that unit is scheduled to be
decommissioned by 1998 due to operational difficulties, and no more
pyrolysis units are expected to be installed at stand-alone
semichemical pulp mills.) Unlike kraft recovery furnaces, most of the
HAP's emitted from fluidized-bed reactors at stand-alone semichemical
pulp mills are formed in the reactor due to incomplete combustion, not
from contact of the exhaust stream with black liquor or HAP-
contaminated water in the DCE or wet ESP systems. Carbon monoxide
emissions, an indicator of combustion efficiency, have been measured
from fluidized-bed reactors at levels as high as 50,000 parts per
million by volume (ppmv); by contrast, kraft recovery
furnaces typically emit less than 1,000 ppmv of CO. No add-
on control devices are currently being used to control total gaseous
organic HAP emissions from combustion sources at stand-alone
semichemical pulp mills; however, at least one RTO will be installed to
control emissions from a fluidized-bed reactor at a semichemical mill
by the end of 1997.
C. Selection of Definition of Affected Source
Most industrial plants consist of numerous pieces or groups of
equipment that emit HAP and that may be viewed as emission ``sources.''
The Agency, therefore, uses the term ``affected source'' to designate
the equipment within a particular kind of plant that is chosen as the
``source'' covered by a given standard. For today's rulemaking, EPA is
proposing to define the affected source as each individual process unit
within the chemical recovery area at kraft, soda, sulfite, and stand-
alone semichemical pulp mills. For kraft and soda pulp mills, each
recovery furnace and its associated SDT('s) are considered together as
an affected source. The Agency decided to
[[Page 18765]]
consider these emission points as one source because recovery furnaces
and SDT's are generally sold as one unit, although the emissions from
the recovery furnace and the SDT are treated separately in nearly all
cases. In today's proposed rulemaking, five process units are examined:
(1) Kraft and soda NDCE recovery furnaces (and associated SDT's), (2)
kraft and soda DCE recovery furnace systems (and associated SDT's), (3)
kraft and soda lime kilns, (4) sulfite combustion units, and (5)
semichemical combustion units.
D. Selection of Pollutants
For purposes of this rule, the HAP's emitted from combustion
sources at pulp mills have been divided into three categories: (1) PM
HAP's, (2) total gaseous organic HAP's, and (3) HCl. The EPA proposes
to regulate emissions of PM HAP's and gaseous organic HAP's.
1. PM HAP's
Available emission data indicate that PM HAP's are emitted from
kraft and soda recovery furnaces, SDT's, and lime kilns and sulfite
combustion units. Particulate matter HAP's represent approximately 0.2
percent of the PM emitted from these combustion sources. Particulate
matter was selected as a surrogate for HAP metals emitted in the form
of particulate. Available data on PM control device performance
indicate that control systems that control PM also control the HAP
portion of the PM. (See Technical Support Document: Chemical Recovery
Combustion Sources at Kraft and Soda Pulp Mills, Chapter 3; docket
entry No. II-A-31.) However, as a means of maximizing compliance
flexibility, the proposed rule also includes a PM HAP emission limit
for existing affected sources at kraft and soda mills that choose to
measure PM HAP's directly, as opposed to measuring PM.
2. Total Gaseous Organic HAP's
Available emission data indicate that the following gaseous organic
HAP's are emitted from kraft and soda NDCE recovery furnaces and DCE
recovery furnace systems and semichemical combustion units:
acetaldehyde, benzene, formaldehyde, methyl ethyl ketone, methyl
isobutyl ketone, methanol, phenol, styrene, toluene, and xylenes.
Methanol is the predominant gaseous organic HAP emitted from kraft and
soda NDCE recovery furnaces and DCE recovery furnace systems.
Methanol was selected as a surrogate for gaseous organic HAP
compounds for demonstrating compliance with the total gaseous organic
HAP limits for new kraft and soda NDCE recovery furnaces and DCE
recovery furnace systems because methanol is the predominant HAP
emitted from these sources, and controls in place for methanol also
would result in the control of other gaseous organic HAP compounds.
(See Technical Support Document: Chemical Recovery Combustion Sources
at Kraft and Soda Pulp Mills, Chapter 2; docket entry No. II-A-31.) For
example, the major emission mechanism for the release of gaseous
organic HAP compounds is the stripping of the compounds from the black
liquor in the BLO unit, the DCE, and some ESP systems. Reducing contact
between the gas streams and the black liquor in these units reduces not
only methanol emissions but also emissions of other gaseous organic
HAP's. In addition, performance tests are more expensive when a range
of organic compounds must be measured. The measurement of methanol as a
surrogate for gaseous organic HAP's reduces compliance costs.
Therefore, the Agency selected methanol as a surrogate for total
gaseous organic HAP emissions for new kraft and soda NDCE recovery
furnaces and DCE recovery furnace systems.
For new and existing semichemical combustion units, THC emissions
were selected as a surrogate for total gaseous organic HAP emissions.
Emissions from semichemical combustion units are primarily the result
of incomplete combustion, and THC emissions were found to correlate
with HAP emissions. (See Correlation of THC Emissions with HAP
Emissions Memo; docket entry No. II-B-71.)
3. Hydrochloric Acid (HCl)
The Agency proposes not to regulate HCl emissions from recovery
furnaces. Under the authority of section 112(d)(4), the Agency has
determined that no further control is necessary because HCl is a
``health threshold pollutant,'' and HCl levels emitted from recovery
furnaces are below the threshold value within an ample margin of
safety. The following discussion provides the basis for the Agency's
decision not to regulate HCl emissions from recovery furnaces.
Specifically, this section discusses (1) the statutory authority for
considering the health threshold when establishing standards, (2) the
determination of HCl as a threshold pollutant, (3) the exposure
assessment modeling of HCl emissions from recovery furnaces, (4) an
ecological assessment of HCl, and (5) the Agency's conclusions.
a. Statutory Authority. The Act includes certain exceptions to the
general statutory requirement to establish emission standards based on
the performance of MACT. Of relevance here, section 112(d)(4) provides
EPA with authority, at its discretion, to develop risk-based standards
for HAP's ``for which a health threshold has been established'',
provided that the standard achieves an ``ample margin of safety.'' (The
full text of the section 112(d)(4): ``[w]ith respect to pollutants for
which a health threshold has been established, the Administrator may
consider such threshold level, within an ample margin of safety, when
establishing emission standards under this subsection.'')
The EPA presumptively applies section 112(d)(4) only to HAP's that
are not carcinogens because Congress clearly intended that carcinogens
be considered nonthreshold pollutants. (Staff of the Senate Committee
on Environment and Public Works, A Legislative History of the Clean Air
Act Amendments of 1990, Vol. 1 at 876, statement of Senator Durenberger
during Senate Debate of October 27, 1990: ``With respect to the
pollutants for which a safe threshold can be set, the authority to set
a standard less stringent than maximum achievable control technology is
contained in subsection (d)(4). With respect to carcinogens and other
non-threshold pollutants, no such authority exists in subsection (d) or
in any other provision of the Act.'') The legislative history further
indicates that if EPA invokes this provision, it must assure that any
emission standard results in ambient concentrations less than the
health threshold, with an ample margin of safety, and that the
standards must also be sufficient to protect against adverse
environmental effects (S. Rep. No. 228, 101st Cong. at 171). Costs are
not to be considered in establishing a standard pursuant to section
112(d)(4) (Ibid.).
Therefore, EPA believes it has the discretion under section
112(d)(4) to develop risk-based standards for some categories emitting
threshold pollutants, which may be less stringent than the
corresponding ``floor''-based MACT standard would be. If EPA decided to
develop standards under this provision, it would seek to assure that
emissions from every source in the category or subcategory are less
than the threshold level to an individual exposed at the upper end of
the exposure distribution. The upper end of the exposure distribution
is calculated using the ``high end exposure estimate,'' defined as ``a
plausible estimate of individual exposure for those persons at the
upper end of the exposure distribution, conceptually above the 90th
percentile, but not higher than the individual in the population who
has the highest exposure'' (EPA Exposure Assessment Guidelines, 57 FR
22888, May 29, 1992).
[[Page 18766]]
The EPA believes that assuring protection to persons at the upper end
of the exposure distribution is consistent with the ``ample margin of
safety'' requirement in section 112(d)(4).
The EPA emphasizes that use of section 112(d)(4) authority is
wholly discretionary. As the legislative history described above
indicates, cases may arise in which other considerations dictate that
the Agency should not invoke this authority to establish less stringent
standards, despite the existence of a health effects threshold that is
not jeopardized. For instance, EPA does not anticipate that it would
set less stringent standards where evidence indicates a threat of
significant or widespread environmental effects, although it may be
shown that emissions from a particular source category do not approach
or exceed a level requisite to protect public health with an ample
margin of safety. The EPA may also elect not to set less stringent
standards where the estimated health threshold for a contaminant is
subject to large uncertainty. Thus, in considering appropriate uses of
its discretionary authority under section 112(d)(4), EPA intends to
consider other factors in addition to health thresholds, including
uncertainty and potential ``adverse environmental effects,'' as that
phrase is defined in section 112(a)(7).
b. Health Effects Assessment. Several factors are considered in the
Agency's decision of whether a pollutant should be categorized as a
health threshold pollutant for the purposes of section 112(d)(4). These
factors include evidence and classification of carcinogenic risk and
evidence of noncarcinogenic effects. The following discussion focuses
on these factors.
Consideration is given to any evidence of human carcinogenic risk
associated with the pollutant. Based on Congress's intent, for the
purposes of section 112(d)(4), the Administrator presumptively
concludes that HAP's classified as either Group A (known carcinogen),
Group B (probable carcinogen), or Group C (possible carcinogen) (as
defined under the EPA's 1986 Carcinogen Risk Assessment Guidelines (51
FR 33992; September 24, 1986)) should not be categorized as threshold
pollutants (as per section 112(f)(2)(A) of the Act, which requires EPA
to consider residual risk standards for pollutants classified as
``known, probable, or possible human carcinogens''). The EPA recognizes
that advances in risk assessment science and policy, as incorporated in
future EPA risk assessment guidelines, may affect the way EPA
differentiates between threshold and non-threshold HAP's. The EPA's
draft Guidelines for Carcinogen Risk Assessment (public review draft,
April, 1996) suggest that carcinogens be assigned non-linear dose-
response relationships where data warrant. It is possible that dose-
response curves for some substances may reach zero risk at a dose
greater than zero, creating a threshold for carcinogenic effects. The
EPA will consider both the state of the science and legislative intent
in future rulemaking under section 112(d)(4). Under EPA's current
guidelines, the Agency considers the data on carcinogenicity in humans
and/or animals for pollutants with A, B, or C classifications adequate
support for consideration of a HAP as a nonthreshold pollutant.
By definition, the Agency does not have enough evidence available
to conclude whether HAP's with the weight of evidence classification of
Group D (as defined under the EPA's 1986 Carcinogen Risk Assessment
Guidelines [51 FR 33992; September 24, 1986]) pose a human cancer risk.
Thus, the Agency will determine, on a case-by-case basis, whether the
available evidence is sufficient to conclude whether a ``safety
threshold for exposure'' exists for each HAP that is classified as a
Group D pollutant. For the purposes of this action, the Agency believes
it is reasonable to classify HCl as a Group D pollutant (see Health
Assessment Document for Chlorine and Hydrogen Chloride, Review Draft;
EPA-600/8-87/041A, August 1994). This classification is based on only
one animal study, and no human data are available for review. In the
animal study, no carcinogenic response was observed in rats exposed via
inhalation. Based on the limited negative carcinogenicity data, and on
EPA's knowledge of how HCl reacts in the body and its likely mechanism
of action (discussed further below), the Agency presumptively considers
HCl to be a threshold pollutant.
Under current EPA science policy, HAP's classified as Group E
pollutants (evidence of noncarcinogenicity for humans) are
presumptively considered by the Agency, for the purposes of section
112(d)(4), to have a ``safety threshold of exposure.'' Therefore, Group
E pollutants are considered threshold pollutants, unless there is
adequate evidence to the contrary. The EPA has developed new risk
assessment guidelines for reproductive effects (see http://www.epa.gov/
ORD/WebPubs/repro), and is in the process of developing others (e.g.,
developmental effects and neurotoxicity) that may influence
determinations of thresholds for specific pollutants.
For pollutants such as HCl that are considered to have a
``threshold of safety'' below which adverse effects are not expected,
the information on noncarcinogenic effects must be evaluated to
determine the potential hazards associated with exposure to the
pollutant. One approach for determining potential hazards of a
pollutant is to use its Inhalation Reference Concentration (RfC). The
RfC is defined as an estimate (with uncertainty spanning perhaps an
order of magnitude) of a daily inhalation exposure that, over a
lifetime, would not likely result in the occurrence of noncancer health
effects in humans. A health benchmark such as the RfC can be
established by applying uncertainty factors to the critical toxic
effect derived from the lowest or no-adverse-effect level of a
pollutant (see EPA-600/8-90-066F, October 1994, Methods for Derivation
of Inhalation Reference Concentrations and Applications of Inhalation
Dosimetry). The confidence in the RfC (which is given a qualitative
ranking of either high, medium, or low) is based on the number of
studies available and the quality of the data base, among other things.
The RfC for HCl is based on a single animal study, which used only
one dose and had limited toxicological measurements. In that study,
laboratory rats exposed to 15,000 g/m3 HCl for 6
hours per day, 5 days per week for life, developed an increased
incidence of hyperplasia of the larynx and trachea, compared to
controls (Health Assessment Document for Chlorine and Hydrogen
Chloride, Review Draft; EPA-600/8-87/041A, August 1994). Effects on
laboratory animals exposed to even higher concentrations of HCl for 90
days included damage to the organs of the respiratory system, but not
to more distant organs. Chronic exposure studies involving lower
concentrations (less than 15,000 g/m3) have not
been done, nor have comprehensive epidemiological studies of humans
(Health Assessment Document for Chlorine and Hydrogen Chloride, Review
Draft; EPA-600/8-87/041A, August 1994).
The RfC for HCl is 20 g/m3 (EPA, 1995,
Integrated Risk Information System (IRIS), Reference Concentration
(RfC) for Inhalation Exposure for Hydrogen Chloride. National Center
for Environmental Assessment, Cincinnati, OH. On-Line). This
concentration is a low confidence RfC with an uncertainty factor of 300
applied to the lowest adverse effect level noted in animals (Ibid).
Generally, information on developmental and reproductive effects
would provide additional confidence in
[[Page 18767]]
the adequacy of the health benchmark for characterizing health risk. No
information is available on the developmental or reproductive effects
associated with HCl exposure in humans or animals. However, no
additional uncertainty is applied for the lack of these studies because
HCl that deposits in the lung is not expected to have any effects at
sites distant from the lung. Hydrogen chloride, in solution, quickly
dissociates to H+ (which, in small doses, is buffered in the
tissue or blood) and Cl- (which is ubiquitous in the body).
Therefore, HCl is expected to have only local effects at the site of
initial deposition. Furthermore, HCl is not thought to be directly
genotoxic (Health Assessment Document for Chlorine and Hydrogen
Chloride, Review Draft; EPA-600/8-87/041A, August 1994).
Based on the information presented above, the Administrator has
determined that HCl is a health threshold pollutant for the purpose of
section 112(d)(4) of the Act. The Administrator also concludes that, in
this case, the RfC is an appropriate threshold value for assessing risk
to humans associated with exposure to this pollutant through
inhalation.
c. Exposure Assessment. Based on emission tests of 14 kraft
recovery furnaces, uncontrolled HCl emissions from DCE and NDCE
recovery furnaces range from 0 to 923 Mg/yr (0 to 1,016 tons/yr);
however, the concentrations of HCl in recovery furnace exhaust gases
(0.3 to 95.6 ppmv) are relatively low due to the high volume
of the exhaust gases. Chlorides enter the liquor cycle primarily
through the wood used for pulping and the caustic used as makeup
chemical during white liquor preparation, although mill process water
can also be a significant contributor. A small portion of the chlorides
in the black liquor fed to the recovery furnace can be emitted from the
furnace as HCl gas. The remaining chlorides in the black liquor exit
the recovery furnace as inorganic alkali salts, either as particulate
in the exhaust gases or as a constituent of the smelt.
For sulfite combustion units, HCl emissions are negligible because
acid-gas absorption systems are an integral part of the sulfite
chemical recovery process. Hydrochloric acid emissions data are
available for only one sulfite combustion unit; HCl emissions from this
unit were approximately 1 ppmv following the acid-gas
absorption system. No data are available on HCl emissions prior to the
acid-gas absorption systems. No HCl emission data are available for
semichemical combustion units. However, neither process nor technical
considerations indicate that HCl emissions would be significant.
Inputs for the exposure assessment model were developed for kraft
and soda recovery furnaces, which have the higher HCl emissions. The
inputs were developed using available test data and mill-specific
process data. Estimated HCl emission rates were based on the highest
available HCl emission factors (in units of kilograms [kg] of HCl per
kg of black liquor solids fired) for both NDCE and DCE recovery
furnaces. Because the HCl emission rates were based on mill-specific
process data (e.g., black liquor solids firing rate), each recovery
furnace type at each mill had a unique set of emissions estimates.
Stack parameters (i.e., height, diameter, temperature and velocity)
were based on information obtained from the AIRS data base; average
values from AIRS were assigned to those sources for which AIRS data
were not available. For mills with multiple recovery furnaces (e.g.,
two NDCE recovery furnaces), HCl emissions from the furnaces were
summed, and the stack parameters for those recovery furnaces were
averaged.
This exposure assessment was conducted following the principles
described in the Agency's Exposure Assessment Guidelines (57 FR 22888,
May 29, 1992). There is no expectation that the population will be
exposed to higher long-term levels of HCl than those predicted by the
model. In this case, a screening analysis was used to determine if
emissions of HCl could result in exposures above Agency-established
health threshold concentrations. The assessment was conducted for 106
mills. The applied approach incorporates into the analysis ranges of
values for those variables meeting the following criteria: where
mathematical distributions are available; where the variables are
independent; and, most importantly, where the variables are believed to
significantly influence the results of the analysis. This probabilistic
procedure uses Monte Carlo simulation to produce distributions with
associated probability estimations (e.g., there is a 95 percent
probability that the estimated exposure to the most exposed population
group (census block) is less than the RfC for HCl).
The distributions used in the Monte Carlo analysis were taken
primarily from EPA sources (such as the Exposure Factors Handbook; EPA/
600/8-89/043, July 1989) and the literature. Best judgments were used
in selecting the distributions and, in some cases, in using only
portions of the distributions that are provided in the Handbook. Use of
other distributions may result in different final outcomes for the
Monte Carlo analysis.
The results of this analysis show that, at the 95 percent
confidence interval, the maximum concentration predicted to which
people are estimated to be exposed is 0.3 g/m \3\, 60 times
less than the inhalation reference concentration.
In addition, terrain (e.g., hills and valleys) is known to affect
concentration estimates predicted near facilities with elevated
pollutant releases (e.g., stacks). The effect of terrain on estimated
HCl concentrations was investigated by including terrain in the
modeling of the ten recovery furnaces that produced the highest
estimated HCl concentrations at census blocks in the exposure
assessment described above. The terrain analysis and a Monte Carlo
assessment similar to that described above resulted, at the 95 percent
confidence interval, in a maximum concentration to which people are
expected to be exposed of 2 g/m \3\, which is 10 times less
than the inhalation reference concentration.
d. Ecological Assessment. The standards for emissions must also
protect against significant and widespread adverse environmental
effects to wildlife, aquatic life, and other natural resources.
Approaches to ecological risk assessments are being developed and
applied by EPA for several areas of concern regarding the effects of
pollutants. For HCl emitted by these source categories, a formal
ecological risk assessment as such has not been made. However,
publications in the literature have been reviewed to determine if there
would be reasonable expectation for serious or widespread adverse
effects to natural resources.
Aspects of pollutant exposure and effects that should be considered
are: toxicity effects from acute and chronic exposures to expected
concentrations around the source (as measured or modeled), persistence
in the environment, local and long-range transport, and tendency for
bio-magnification with toxic effects manifest at higher trophic levels.
No research has been identified for effects on terrestrial animal
species beyond that cited in the development of the RfC. The evidence
available to date, discussed in section IV.D.3.b of this preamble,
indicates that HCl is a threshold pollutant for the purposes of section
112(d)(4) of the Act. Modeling calculations indicate that there is
little likelihood of chronic or widespread exposure to HCl at
concentrations above the threshold around pulp and paper mills. Based
on these considerations, EPA believes that the RfC can reasonably be
expected to protect
[[Page 18768]]
against widespread adverse effects in other animal species as well.
Plants also respond to airborne HCl levels. Chronic exposure to
about 600 g/m \3\, can be expected to result in discernible
effects, depending on the plant species. Plants respond differently to
HCl as an anhydrous gas than to HCl aerosols. Relative humidity is
important in plant response; there appears to be a threshold of
relative humidity above which plants will incur twice as much damage at
a given dose (Medical and Biological Effects of Environmental
Pollutants: Chlorine and Hydrogen Chloride, National Academy of
Sciences, 1976). Effects include leaf injury and decrease in
chlorophyll levels in various species given acute, 20-minute exposures
of 6,500 to 27,000 g/m \3\ (Health Assessment Document for
Chlorine and Hydrogen Chloride, Review Draft; EPA-600/8-87/041A, August
1994). A field study reports different sensitivity to damage of foliage
in 50 species growing in the vicinity of an anhydrous aluminum chloride
manufacturer. American elm, bur oak, eastern white pine, basswood, red
ash and several bean species were observed to be most sensitive.
Concentrations of HCl in the air were not reported. Chloride ion in
whole leaves was 0.2 to 0.5 percent of dry weight; sensitive species
showed damage at the lower value, but tolerant species displayed no
injury at the higher value. Injury declined with distance from the
source with no effects observed beyond 300 meters (Harper and Jones,
1982, ``The Relative Sensitivity of Fifty Plant Species to Chronic
Doses of Hydrogen Chloride,'' Phytopathology 72: 261-262).
Prevailing meteorology strongly determines the fate of HCl in the
atmosphere (Health Assessment Document for Chlorine and Hydrogen
Chloride, Review Draft; EPA-600/8-87/041A, August 1994). However, HCl
is not considered a strongly persistent pollutant, or one where long
range transport is important in predicting its ecological effects. In
the atmosphere, HCl can be expected to be absorbed into aqueous
aerosols, due to its great affinity for water, and removed from the
troposphere by rainfall. In addition, HCl will react with hydroxy ions
to yield water plus chloride ions. However, the concentration of
hydroxy ions in the troposphere is low, so HCl may have a relatively
long residence time in areas of low humidity. No studies are reported
of HCl levels in ponds or other small water bodies or soils near major
sources of HCl emissions. Toxic effects of HCl to aquatic organisms
would likely be due to the hydronium ion, or acidity. Aquatic organisms
in their natural environments often exhibit a broad range of pH
tolerance. Effects of HCl deposition to small water bodies and to soils
will primarily depend on the extent of neutralizing by carbonates or
other buffering compounds (Health Assessment Document for Chlorine and
Hydrogen Chloride, Review Draft; EPA-600/8-87/041A, August 1994).
Chloride ions are essentially ubiquitous in natural waters and soils,
so minor increases due to deposition of dissolved HCl will have much
less effect than the deposited hydronium ions. Deleterious effects of
HCl on ponds and soils, where such effects might be found near a major
source emitting to the atmosphere, likely will be local rather than
widespread, as observed in plant foliage.
Effects of HCl on tissues are generally restricted to those
immediately impacted and are essentially acidic effects. The rapid
solubility of HCl in aqueous media releases hydronium ions, which can
be corrosive to tissue when above a threshold concentration. The
chloride ions may be concentrated in some plant tissues, but may be
distributed throughout the organism, as most organisms have chloride
ions in their fluids. Leaves or other tissues exposed to HCl may show
some concentration above that of their immediate environment; that is,
some degree of bioconcentration can occur. However, long-term storage
in specific organs and biomagnification of concentrations of HCl in
trophic levels of a food chain would not be expected. Thus, the
chemical nature of HCl results in deleterious effects, that when
present, are local rather than widespread.
e. Conclusions. The results of the exposure assessment modelling
showed exposure levels to HCl emissions from kraft and soda recovery
furnaces below the health threshold value. Furthermore, the threshold
value, for which the RfC was determined to be an appropriate value, was
not exceeded when taking into account an ample margin of safety.
Finally, no significant or widespread adverse environmental effects
from HCl are anticipated. Therefore, the Agency, under authority of
section 112(d)(4), has determined that further control of HCl emissions
from kraft and soda recovery furnaces and sulfite and semichemical
combustion units is not necessary.
E. Determination of Subcategories and MACT Floors
The first step in establishing MACT floors is to determine whether
the source category warrants subcategorization. In evaluating the
chemical recovery process for subcategorization, the Agency took into
consideration the type of equipment used in the process, the emission
potential of each emission point, and any variations in the process due
to pulp type. The Agency determined that the chemical recovery areas at
kraft and soda pulp mills do not warrant subcategorization because the
recovery areas are comparable in processes, equipment, and HAP
emissions. The Agency determined that separate subcategories are
warranted for sulfite and stand-alone semichemical pulp mills because
the recovery processes used at sulfite and stand-alone semichemical
pulp mills are specifically different from each other and from those
used at kraft and soda pulp mills.
The proposed MACT floors for each category were established on an
emission point basis. For existing sources at kraft and soda pulp
mills, the MACT floor was established by examining the emission level
achievable by the control technology used by the source at the 94th
percentile (i.e., the median emission limitation achieved by the top 12
percent of sources). Because there are fewer than 30 sulfite combustion
units nationwide, the proposed MACT floor for existing sources at
sulfite pulp mills was established by examining the emission level
achieved by the control technology used by the best-performing five
existing sources at sulfite pulp mills. The MACT floor approach used
for existing sources at sulfite pulp mills was also used for existing
sources at stand-alone semichemical pulp mills because there are fewer
than 30 semichemical combustion sources. The MACT floor technologies
for new sources at kraft, soda, sulfite, and stand-alone semichemical
pulp mills are based on the best-performing similar source for each
subcategory. The control technologies and corresponding emission levels
that represent the proposed MACT floors were determined based on
technology and emission data that were available to the Administrator.
1. MACT Floors--Kraft and Soda Pulp Mills
This section provides a brief description of the MACT floor
determinations for kraft and soda NDCE recovery furnaces, DCE recovery
furnace systems, lime kilns, and SDT's.
a. NDCE Recovery Furnaces. An estimated 128 NDCE recovery furnaces
operate at 96 U.S. kraft and soda pulp mills. Information regarding the
furnace type, size, and add-on control devices is available for
approximately 88 percent of these recovery furnaces. Ninety-seven
percent of NDCE recovery furnaces are
[[Page 18769]]
equipped with an ESP, 2 percent are equipped with an ESP followed by a
wet scrubber, and the remaining 1 percent are equipped with two wet
scrubbers in series. The add-on control devices were installed
primarily for control of PM emissions.
The following paragraphs describe the proposed MACT floor control
technologies for new and existing kraft and soda NDCE recovery furnaces
for both PM/PM HAP and total gaseous organic HAP control and the
emission levels achievable with each proposed MACT floor technology.
(1) PM and PM HAP MACT Floors. Properly designed and operated ESP's
used on kraft recovery furnaces routinely achieve PM removal
efficiencies of 99 percent or greater. Although emission test data from
recovery furnace ESP's on PM HAP performance are limited, available
data on ESP performance indicate that those systems that achieve the
greatest PM removal show the best performance for the HAP portion of
the PM. (See Technical Support Document: Chemical Recovery Combustion
Sources at Kraft and Soda Pulp Mills, Chapter 3; docket entry No. II-A-
31.) Therefore, PM can be used as a surrogate for PM HAP's.
The NSPS for kraft pulp mills requires that PM emissions from
recovery furnaces constructed, reconstructed, or modified after
September 24, 1976 be less than or equal to 0.10 g/dscm (0.044 gr/dscf)
of flue gas corrected to 8 percent oxygen. Approximately 39 percent of
NDCE recovery furnaces are subject to the NSPS, and even more (80
percent) reportedly achieve the NSPS limit.
Long-term (monthly) PM emission data are available for eight NDCE
recovery furnaces. Particulate matter emissions from each of these
eight NDCE recovery furnaces varied significantly from month to month;
however, PM emissions from seven of the eight NDCE recovery furnaces
consistently met the NSPS limit of 0.10 g/dscm (0.044 gr/dscf)
corrected to 8 percent oxygen over a 4- to 6-year period. Collectively,
emissions from these seven NDCE recovery furnaces ranged from 0.002 to
0.10 g/dscm (0.001 to 0.044 gr/dscf), corrected to 8 percent oxygen.
(See State of Washington Data Memo, docket entry No. II-B-59.) Thus,
the long-term data demonstrate that NDCE recovery furnaces equipped
with ESP's can meet the NSPS level of 0.10 g/dscm (0.044 gr/dscf)
corrected to 8 percent oxygen on a long-term basis. Because greater
than 6 percent of NDCE recovery furnaces are capable of meeting the
NSPS limit on a long-term basis with ESP's, the proposed MACT floor PM
control technology for existing kraft and soda NDCE recovery furnaces
is an ESP capable of meeting the NSPS, which typically has a specific
collecting area (SCA) of 100 m2/(m3/sec) (530
ft2/1,000 acfm). The application of the proposed MACT floor
PM control technology is represented by a PM emission level of 0.10 g/
dscm (0.044 gr/dscf) corrected to 8 percent oxygen.
The proposed MACT floor control technology for PM HAP is the same
as the proposed MACT floor control technology for PM and is represented
by a PM HAP emission level of 1.00E-03 kg/Mg (2.01E-03 lb/ton) of black
liquor solids fired. The proposed MACT floor PM HAP emission level is
based on available test data and is equivalent to the average PM HAP
emission factor for recovery furnaces with PM emissions that achieve
the NSPS level of 0.10 g/dscm (0.044 gr/dscf) corrected to 8 percent
oxygen.
With respect to MACT for new sources, the best-performing PM
control system of the eight NDCE recovery furnaces for which long-term
PM emission data are available is an ESP with an operating SCA between
110 and 130 m2/(m3/sec) (570 and 670
ft2/1,000 acfm) followed by a cross-flow, packed-bed
scrubber. Monthly PM emissions data from the NDCE recovery furnace with
this control system varied from 0.002 to 0.025 g/dscm (0.001 to 0.011
gr/dscf) corrected to 8 percent oxygen over a 6-year period. Taking the
variability of the data into consideration, a PM emission level of
0.034 g/dscm (0.015 gr/dscf) was selected to represent the MACT floor
PM emission level for new NDCE recovery furnaces. Therefore, the
proposed MACT floor PM control technology for new kraft and soda NDCE
recovery furnaces is an ESP capable of achieving a PM emission level of
0.034 g/dscm (0.015 gr/dscf) corrected to 8 percent oxygen (i.e., an
ESP with a typical SCA between 110 and 130 m2/
[m3/sec] [570 and 670 ft2/1,000 acfm]) followed
by a packed-bed scrubber.
Although the proposed MACT floor PM control technology for new NDCE
recovery furnaces includes both the ESP and the cross-flow, packed-bed
scrubber, the scrubber was installed as a heat recovery device and for
SO2 control and is not expected to provide much, if any,
additional PM control. Because of the high PM removal efficiencies
achievable with newer ESP's, the proposed MACT floor PM emission level
of 0.034 g/dscm (0.015 gr/dscf) corrected to 8 percent oxygen for new
NDCE recovery furnaces could be achieved with the application of the
ESP alone.
A PM HAP emission level was not established for new NDCE recovery
furnaces because insufficient PM HAP data are available from NDCE
recovery furnaces representing MACT for new sources.
(2) Total Gaseous Organic HAP MACT Floors. The ESP systems applied
to existing NDCE recovery furnaces conform to one of two designs: wet
ESP systems or dry ESP systems. A wet ESP system uses unoxidized black
liquor or water in the ESP bottom or in the PM return system. A dry ESP
system includes both a dry-bottom ESP and a dry PM return system. Wet
ESP systems that use black liquor or HAP-contaminated water emit higher
levels of gaseous organic HAP's than dry ESP systems due to the
stripping of gaseous organic HAP's from the black liquor or HAP-
contaminated water in the ESP bottom or PM return system. Based on the
available emission data, NDCE recovery furnaces with dry ESP systems
emit, on average, approximately 72 percent less total gaseous organic
HAP's than NDCE recovery furnaces with wet ESP systems.
Although information is available to classify almost all (99
percent) of NDCE recovery furnace ESP's as wet- or dry-bottom, little
information is available regarding the use of black liquor or HAP-
contaminated water in the recovery furnace ESP PM return systems. Based
on the limited available information on ESP return systems,
approximately 5 percent of NDCE recovery furnaces are estimated to be
equipped with dry ESP systems. Because the estimated percentage of NDCE
recovery furnaces equipped with dry ESP systems is less than 6 percent,
the proposed MACT floor control technology for total gaseous organic
HAP emissions from existing kraft and soda NDCE recovery furnaces is a
wet ESP system, and, thus, no control of total gaseous organic HAP's is
achieved at the floor. However, because NDCE recovery furnaces equipped
with dry ESP systems represent the best-controlled source for total
gaseous organic HAP emissions, the proposed MACT floor total gaseous
organic HAP control technology for new kraft and soda NDCE recovery
furnaces is a dry ESP system. Emission data from three NDCE recovery
furnaces equipped with dry ESP systems indicate that a total gaseous
organic HAP emission level, as measured by methanol, of 0.012 kg/Mg
(0.025 lb/ton) of black liquor solids fired or less is achievable. The
methanol emission level corresponds to the highest three-run average
obtained for a dry ESP system on an NDCE recovery furnace plus an
additional amount to
[[Page 18770]]
account for the variability in the dry ESP system data set and the lack
of long-term data. Therefore, the total gaseous organic HAP emission
level, as measured by methanol, associated with the proposed MACT floor
control technology (i.e., a dry ESP system) is 0.012 kg/Mg (0.025 lb/
ton) of black liquor solids fired.
b. DCE Recovery Furnace Systems. The DCE recovery furnace system
includes the recovery furnace, DCE, and the BLO system. An estimated 83
DCE recovery furnaces are in operation at 48 U.S. kraft and soda pulp
mills. An estimated 46 BLO systems are in operation at these 48 pulp
mills. Of the two mills without BLO systems, one is a soda pulp mill,
and the other is a kraft pulp mill. Information regarding the furnace
type, size, and add-on control devices and the associated BLO systems
is available for approximately 93 percent of DCE recovery furnace
systems.
Like NDCE recovery furnaces, all DCE recovery furnaces are equipped
with some type of add-on control device to reduce PM emissions from the
furnace. In the case of DCE units, 90 percent are controlled with an
ESP, 8 percent are controlled with an ESP followed by a wet scrubber,
and the remaining 2 percent are controlled with two ESP's in series. As
with NDCE recovery furnaces, MACT floor control technologies for DCE
recovery furnace systems were selected for both PM/PM HAP and total
gaseous organic HAP emissions. The following paragraphs describe the
proposed MACT floor control technologies for new and existing kraft and
soda DCE recovery furnace systems and the emission levels achievable
with each proposed MACT floor technology.
(1) PM and PM HAP MACT Floors. As discussed above for NDCE recovery
furnaces, properly designed and operated ESP's used on kraft recovery
furnaces routinely achieve PM removal efficiencies of 99 percent or
greater. Using installation dates to determine NSPS applicability,
three DCE recovery furnaces (i.e., 4 percent of the DCE recovery
furnace population) are subject to the NSPS emission limit of 0.10 g/
dscm (0.044 gr/dscf) corrected to 8 percent oxygen for kraft recovery
furnaces. Long-term (monthly) PM emission data are available for an
additional four DCE recovery furnaces that are not subject to the NSPS
but have consistently met the NSPS emission level of 0.10 g/dscm (0.044
gr/dscf) corrected to 8 percent oxygen over a 3- to 6-year period, even
though PM emissions from each of these four DCE recovery furnaces
varied significantly from month to month. Collectively, the PM
emissions from these four DCE recovery furnaces varied from 0.011 to
0.10 g/dscm (0.005 to 0.044 gr/dscf) corrected to 8 percent oxygen over
the 3- to 6-year period. (See State of Washington Data Memo; docket
entry No. II-B-59.) The combination of those DCE recovery furnaces
subject to the NSPS and those for which data show an ability to achieve
the NSPS level on a long-term basis represent a total of seven DCE
recovery furnaces, or 9 percent of the DCE recovery furnace population.
Because greater than 6 percent of DCE recovery furnaces are capable
of meeting the NSPS PM limit on a long-term basis with ESP's, the
proposed MACT floor PM control technology for existing kraft and soda
DCE recovery furnace systems is an ESP capable of meeting the NSPS,
which typically has an SCA of 90 m2/(m3/sec) (430
ft 2/1,000 acfm). The application of the proposed MACT floor
PM control technology is represented by a PM emission level of 0.10 g/
dscm (0.044 gr/dscf) corrected to 8 percent oxygen.
The proposed MACT floor control technology for PM HAP is the same
as the proposed MACT floor control technology for PM and is represented
by a PM HAP emission level of 1.00E-03 kg/Mg (2.01E-03 lb/ton) of black
liquor solids fired. As with existing NDCE recovery furnaces, the
proposed MACT floor PM HAP emission level is equivalent to the average
PM HAP emission factor for kraft and soda recovery furnaces with PM
emissions that achieve the NSPS level of 0.10 g/dscm (0.044 gr/dscf)
corrected to 8 percent oxygen.
The best-performing PM control system for both NDCE and DCE
recovery furnaces is an ESP with an operating SCA between 110 and 130
m2/(m3/sec) (570 and 670 ft 2/1,000
acfm) followed by a cross-flow, packed-bed scrubber. Monthly PM
emissions data from the recovery furnace with this control system
varied from 0.002 to 0.025 g/dscm (0.001 to 0.011 gr/dscf) corrected to
8 percent oxygen over a 6-year period. Taking the variability of the
data into consideration, a PM emission level of 0.034 g/dscm (0.015 gr/
dscf) was selected to represent the MACT floor PM emission level for
new DCE recovery furnaces. Therefore, the proposed MACT floor PM
control technology for all new kraft and soda DCE recovery furnaces is
an ESP capable of achieving a PM emission level of 0.034 g/dscm (0.015
gr/dscf) corrected to 8 percent oxygen (i.e., an ESP with a typical SCA
between 110 and 130 m2/[m3/sec] [570 and 670 ft
2/1,000 acfm]) followed by a packed-bed scrubber.
Although the proposed MACT floor PM control technology for new
kraft and soda DCE recovery furnaces includes both the ESP and the
cross-flow, packed-bed scrubber, the scrubber was installed as a heat
recovery device and for SO2 control and is not expected to
provide much, if any, additional PM control. Because of the high PM
removal efficiencies achievable with newer ESP's, the proposed MACT
floor PM emission level of 0.034 g/dscm (0.015 gr/dscf) corrected to 8
percent oxygen for new DCE recovery furnaces could be achieved with the
application of the ESP alone.
The EPA is not proposing a MACT floor PM HAP emission level for new
kraft and soda DCE recovery furnaces for the same reason stated above
for new NDCE recovery furnaces.
(2) Total Gaseous Organic HAP MACT Floors. Four of the estimated 46
BLO systems in operation are pipeline molecular oxygen-based systems,
which have no emission points. No emission data are available from DCE
recovery furnaces with molecular oxygen BLO systems for comparison with
DCE recovery furnaces with air-based BLO systems. Therefore, the effect
of molecular oxygen BLO systems on total emissions from the DCE
recovery furnace system is uncertain. With air-based BLO systems,
gaseous organic HAP's are stripped from the black liquor and emitted to
the atmosphere as the air bubbles and black liquor make contact. Unlike
air-based systems, molecular oxygen systems use pure oxygen, and, thus,
no diluents are introduced that could strip organic compounds from the
black liquor; consequently, organic compounds not released from the
black liquor during the oxidation process could be subsequently
stripped, in theory, from the oxidized black liquor when the black
liquor enters the direct contact evaporator. For this reason, molecular
oxygen BLO systems are not viewed by the Agency as a control option for
DCE recovery furnace systems.
The gaseous organic HAP emissions from 2 of the estimated 42 air-
based BLO systems are controlled via incineration in power boilers; the
remainder are uncontrolled. However, the two air-based BLO units with
controlled emissions represent less than 6 percent of DCE recovery
furnace systems. Therefore, the proposed MACT floor for total gaseous
organic HAP control for existing kraft and soda DCE recovery furnace
systems is no control.
The DCE recovery furnace systems emit more gaseous organic HAP's
than
[[Page 18771]]
NDCE recovery furnaces because more opportunities exist for gaseous
organic HAP compounds to be stripped from the black liquor. In DCE
systems, gaseous organic HAP compounds can be stripped from the black
liquor in the BLO system, the DCE, and the ESP system. Based on the
available emission data, NDCE recovery furnaces with dry ESP systems
emit approximately 93 percent less total gaseous organic HAP's than DCE
recovery furnace systems.
The NDCE recovery furnaces with dry ESP systems also have lower TRS
emissions compared to DCE recovery furnace systems. The need for TRS
emission reductions and the need for additional recovery furnace
capacity have resulted in mills converting older and smaller DCE units
into larger NDCE units. Approximately 24 percent of the existing NDCE
recovery furnaces are converted DCE recovery furnaces. For these
reasons, and also because NDCE recovery furnaces are more energy
efficient than DCE recovery furnaces, all new recovery furnace
installations are of the NDCE design. Because of its lower HAP emission
potential, an NDCE recovery furnace equipped with a dry ESP system was
selected as the MACT floor total gaseous organic HAP control technology
for all new kraft and soda NDCE recovery furnaces and DCE recovery
furnace systems. This proposed MACT floor control technology is capable
of achieving a total gaseous organic HAP emission level, as measured by
methanol, of 0.012 kg/Mg (0.025 lb/ton) of black liquor solids fired.
c. Lime Kilns. An estimated 192 lime kilns operate at 124 U.S.
kraft and soda pulp mills. Information regarding the lime kiln type,
size, and add-on control devices is available for approximately 85
percent of these lime kilns. All of the add-on control systems in place
on lime kilns are for the control of PM or TRS emissions. No add-on
controls designed to remove gaseous organic HAP's are applied to lime
kilns.
Gaseous organic HAP emissions from lime kilns are primarily
attributable to the use of HAP-contaminated process waters in the lime
mud washers and lime kiln scrubbers. Therefore, gaseous organic HAP
emissions from lime kilns can be minimized by reducing the HAP content
of process waters used in the lime mud washers and scrubbers. These
process waters are being regulated as part of the final NESHAP for
noncombustion sources at pulp and paper mills. Therefore, no MACT floor
has been established for total gaseous organic HAP's for new and
existing kraft and soda lime kilns as part of this proposed NESHAP. The
following paragraphs describe the proposed MACT floor PM/PM HAP control
technologies and the associated emission levels for existing and new
kraft and soda lime kilns.
Particulate matter emissions from most (90 percent) of the lime
kilns are controlled by wet scrubbers. Venturi scrubbers are the most
common type of wet scrubber in use on lime kilns. Particulate matter
emissions from the remaining 10 percent of lime kilns are controlled by
ESP's (9 percent) or the combination of an ESP and wet scrubber (1
percent). Properly designed and operated venturi scrubbers and ESP's
used on kraft lime kilns are capable of reducing PM emissions by
greater than 99 percent.
The NSPS for kraft pulp mills requires that PM emissions from gas-
fired lime kilns constructed, reconstructed, or modified after
September 24, 1976 be less than or equal to 0.15 g/dscm (0.067 gr/dscf)
of flue gas corrected to 10 percent oxygen. Approximately 19 percent of
lime kilns are subject to the NSPS limit for gas-fired lime kilns, and
even more (i.e., 64 percent of all lime kilns, including oil-fired lime
kilns) have reported average PM emissions less than the gas-fired NSPS
limit.
Long-term (monthly) PM emission data are available for four gas-
fired lime kilns that are subject to the NSPS PM limit for gas-fired
lime kilns. No long-term data are available for oil-fired lime kilns.
Two of the four lime kilns for which long-term PM emission data are
available are equipped with venturi scrubbers, and two are equipped
with ESP's. Particulate matter emissions from the four lime kilns
varied from 0.002 to 0.15 g/dscm (0.001 to 0.067 gr/dscf) corrected to
10 percent oxygen over a 4-to 7-year period. The long-term data
demonstrate that existing lime kilns equipped with either venturi
scrubbers or ESP's can meet an emission level of 0.15 g/dscm (0.067 gr/
dscf) corrected to 10 percent oxygen on a long-term basis. Because
greater than 6 percent of lime kilns are capable of meeting the gas-
fired NSPS limit on a long-term basis with venturi scrubbers or ESP's,
the proposed MACT floor control technology for existing kraft and soda
lime kilns is either a venturi scrubber or an ESP. The application of
these proposed MACT floor PM control technologies is represented by a
PM emission level of 0.15 g/dscm (0.067 gr/dscf) corrected to 10
percent oxygen. The proposed MACT floor control technology for PM HAP
is the same as the proposed MACT floor control technology for PM and is
represented by a PM HAP emission level of 6.33E-03 kg/Mg (1.27E-02 lb/
ton) of CaO produced. The proposed MACT floor PM HAP emission level is
equivalent to the average PM HAP emission factor for lime kilns with
outlet PM emissions that achieve the NSPS level of 0.15 g/dscm (0.067
gr/dscf) corrected to 10 percent oxygen.
Of the four lime kilns for which long-term PM emission data are
available, the best-performing PM control system is an ESP with an
operating SCA of 220 m2/(m3/sec) (1,120
ft2/1,000 acfm), which is substantially higher than the
typical SCA for an ESP designed to meet the NSPS (i.e., 90
m2/[m3/sec] [460 ft2/1,000 acfm]). The
monthly PM emissions from the best-performing lime kiln varied from
0.002 to 0.018 g/dscm (0.001 to 0.008 gr/dscf) corrected to 10 percent
oxygen over a 7-year period. To account for the variability in the
data, a PM emission level of 0.023 g/dscm (0.010 gr/dscf) was selected
to represent the MACT floor PM emission level for new lime kilns.
Therefore, the proposed MACT floor PM HAP control technology for new
kraft and soda lime kilns is an ESP capable of achieving a PM emission
level of 0.023 g/dscm (0.010 gr/dscf) corrected to 10 percent oxygen
(i.e., an ESP with a typical SCA of 220 m2/[m3/
sec] [1,120 ft\2\/1,000 acfm]).
A MACT floor PM HAP emission level was not established for new lime
kilns for the same reasons stated above for new NDCE recovery furnaces.
d. Smelt Dissolving Tanks. An estimated 227 SDT's operate at 124
U.S. kraft and soda pulp mills. Information regarding the SDT size and
add-on control devices is available for approximately 83 percent of the
SDT's. The add-on control systems in place on SDT's are for control of
PM emissions. No add-on controls designed to remove gaseous organic
HAP's are applied to SDT's.
As discussed above for lime kilns, gaseous organic HAP emissions
from SDT's are primarily the result of the use of HAP-contaminated
process waters. The HAP-contaminated process waters are typically used
in the SDT scrubbers as makeup water to the SDT. Therefore, gaseous
organic HAP emissions from SDT's can be minimized by reducing the HAP
content of process waters used in the SDT and SDT scrubber. However, as
stated above for lime kilns, the control of HAP emissions from process
waters is being regulated as part of the final NESHAP for noncombustion
sources at pulp and paper mills. Therefore, no MACT floor has been
established for total gaseous organic HAP emissions for new and
existing kraft and soda SDT's as part of this proposed NESHAP.
Particulate matter emissions from most (87 percent) of the SDT's
are
[[Page 18772]]
controlled by wet scrubbers. Particulate matter emissions from the
majority of the remaining SDT's are controlled by mist eliminators.
Based on the available performance data for wet scrubbers and mist
eliminators installed on SDT's, wet scrubbers are more effective at
controlling PM emissions from SDT's than mist eliminators. (See
Technical Support Document: Chemical Recovery Combustion Sources at
Kraft and Soda Pulp Mills, Chapter 3; docket entry No. II-A-31.)
Properly designed wet scrubbers used on kraft SDT's are capable of
reducing PM emissions by greater than 99 percent.
The NSPS for kraft pulp mills require that PM emissions from SDT's
that are constructed, modified, or reconstructed after September 24,
1976 be less than 0.10 kg/Mg (0.20 lb/ton) of black liquor solids
fired. Approximately 29 percent of SDT's are subject to the NSPS PM
limit, and even more (75 percent) have reported average PM emissions
less than the NSPS PM limit. Although no long-term PM emission data are
available for SDT's equipped with wet scrubbers that are subject to the
NSPS limit of 0.10 kg/Mg (0.20 lb/ton) of black liquor solids fired,
the prevalence of wet scrubbers on SDT's and the high PM removal
efficiencies achieved with this technology are sufficient to establish
wet scrubbers as the proposed MACT floor PM control technology for
existing kraft and soda SDT's. The application of this control
technology is represented by a PM emission level of 0.10 kg/Mg (0.20
lb/ton) of black liquor solids fired. The proposed MACT floor control
technology for PM HAP is the same as the proposed MACT floor control
technology for PM and is represented by a PM HAP emission level of
6.20E-05 kg/Mg (1.24E-04 lb/ton) of black liquor solids fired. The
proposed MACT floor PM HAP emission level is equivalent to the average
PM HAP emission factor for SDT's with outlet PM emissions that achieve
the NSPS PM level of 0.10 kg/Mg (0.20 lb/ton) of black liquor solids
fired.
Long-term (monthly) PM emission data are available for three SDT's
equipped with wet scrubbers designed to meet a PM permit limit (0.06
kg/Mg [0.12 lb/ton] of black liquor solids fired) that is more
stringent than the NSPS. The high-efficiency wet scrubbers installed on
these three SDT's represent the best-performing PM control systems
installed on kraft and soda SDT's. Collectively, monthly PM emissions
from these three SDT's varied from 0.0045 to 0.055 kg/Mg (0.009 to 0.11
lb/ton) of black liquor solids fired over a 2- to 6-year period. (See
State of Washington Data Memo, docket entry No. II--B-59.) The long-
term data demonstrate that SDT's equipped with high-efficiency wet
scrubbers can achieve a maximum outlet PM level of 0.06 kg/Mg (0.12 lb/
ton) of black liquor solids fired on a long-term basis. Therefore, the
proposed MACT floor PM HAP control technology for new kraft and soda
SDT's is a high-efficiency wet scrubber capable of achieving a PM
emission level of 0.06 kg/Mg (0.12 lb/ton) of black liquor solids
fired.
2. MACT Floors--Sulfite Pulp Mills
An estimated 21 combustion units operate at sulfite pulp mills.
Information regarding the chemical recovery equipment and add-on
control devices is available for approximately 95 percent of these
combustion units. Because there are less than 30 sulfite combustion
units, the MACT floor for existing sources is based on the 5 best-
performing sources. Thirteen of the 21 sulfite combustion units (62
percent) are equipped with fiber-bed demister systems. The remainder of
the combustion units are equipped with venturi scrubbers or packed-bed
scrubbers. These add-on control devices were installed on sulfite
combustion units for PM control and additional SO2 control.
All sulfite combustion units are equipped with absorption towers prior
to the PM control device to recover SO2 for reuse in the
pulping process.
Long-term PM emission data are available for two sulfite combustion
units equipped with fiber-bed demister systems. Based on these long-
term data and additional long-term data for sulfite combustion units
equipped with wet scrubbers, fiber-bed demister systems are more
effective than wet scrubbers at controlling PM emissions from sulfite
combustion units. Monthly PM emission data from the two sulfite
combustion units equipped with fiber-bed demister systems ranged from
0.005 to 0.088 g/dscm (0.002 to 0.038 gr/dscf) corrected to 8 percent
oxygen over a 6-to 7-year period. Because the fiber-bed demister system
represents the best-performing control technology and at least five
sources are equipped with fiber-bed demister systems, this technology
was selected to represent the proposed MACT floor control technology
for existing sulfite combustion units. To account for variability in
the data, a PM emission level of 0.092 g/dscm (0.040 gr/dscf) corrected
to 8 percent oxygen was selected to represent the MACT floor PM
emission level for existing sulfite combustion units.
Monthly PM emission data from the best-performing sulfite
combustion unit equipped with a fiber-bed demister system ranged from
0.009 to 0.039 g/dscm (0.004 to 0.017 gr/dscf) corrected to 8 percent
oxygen over a 6-year period. This sulfite combustion unit also is
equipped with a wet scrubber between the SO2 absorption
towers and the fiber-bed demister system. The scrubber was added to the
system for additional PM and SO2 control. Because the best-
performing source is equipped with a wet scrubber and fiber-bed
demister system, the combination of these technologies was selected to
represent the proposed MACT floor control technology for new sulfite
combustion units. To account for the variability in the data, a PM
emission level of 0.046 g/dscm (0.020 gr/dscf) corrected to 8 percent
oxygen was selected to represent the MACT floor PM emission level for
new sulfite combustion units.
3. MACT Floors--Stand-Alone Semichemical Pulp Mills
An estimated 14 chemical recovery combustion units operate at 13
U.S. stand-alone semichemical pulp mills. Information regarding the
design and operation of chemical recovery combustion units is available
for all of these units. Although chemical recovery combustion units at
stand-alone semichemical pulp mills are equipped with a variety of PM
control devices, insufficient PM data and no PM HAP data are available
to establish MACT floors for PM or PM HAP. In addition, none of the
existing semichemical mills are currently controlling gaseous organic
HAP emissions from semichemical combustion sources. Therefore, no
control of total gaseous organic HAP emissions is achieved at the MACT
floor for existing or new sources.
However, the Agency has selected a beyond-the-floor option to
represent MACT for gaseous organic HAP control for existing and new
semichemical combustion sources. The beyond-the-floor option is based
on the use of an RTO preceded by a wet ESP. (A wet ESP or other PM
control device is necessary because the RTO requires a high degree of
PM control for proper operation.) Pilot study results at a stand-alone
semichemical mill indicate that an RTO is well-suited to reducing
gaseous organic HAP emissions from fluidized-bed reactors, which emit
the highest known quantities of HAP's of the combustion technologies
currently in use at semichemical pulp mills. The semichemical mill that
conducted the pilot study is currently installing a full-scale RTO
based on the results of the pilot study.
[[Page 18773]]
During the pilot study, the RTO reduced THC emissions from the
mill's fluidized-bed reactor by an average of 97 percent. However,
because the RTO has not yet been demonstrated full-scale at a
semichemical mill, EPA estimated the total gaseous organic HAP emission
level that corresponds to MACT using the average THC emission reduction
(90 percent) achieved during the pilot study test run with the lowest
level of control. The estimated 90 percent THC emission reduction was
applied to the average uncontrolled THC emissions (measured as carbon)
from a fluidized-bed reactor. Based on the results of the calculation,
the application of an RTO preceded by a wet ESP is estimated to be
representative of either a total gaseous organic HAP emission level of
1.49 kg/Mg (2.97 lb/ton) of black liquor solids fired, or a 90 percent
reduction in total gaseous organic HAP emissions. (Total gaseous
organic HAP's are measured as THC, as carbon, in both cases.)
F. Discussion of Regulatory Alternatives
The proposed standards were selected based on a review of the
regulatory alternatives developed for the affected sources. Table 3
presents the regulatory alternatives examined for existing affected
sources at kraft and soda pulp mills; Tables 4 and 5 present the
regulatory alternatives for existing affected sources at sulfite and
stand-alone semichemical pulp mills, respectively. For existing
affected sources, regulatory alternative I (RA I) represents the
proposed MACT floor, and additional regulatory alternatives represent
beyond-the-MACT-floor options. The regulatory alternatives are
increasingly more stringent in terms of total HAP emission reduction
requirements. The most stringent regulatory alternative examined for
existing sources is representative of MACT for new sources. A
discussion of the regulatory alternatives is provided below.
Table 3.--Regulatory Alternatives for Existing Affected Sources at Kraft and Soda Pulp Mills
----------------------------------------------------------------------------------------------------------------
Basis of alternative
-----------------------------------------------------------------------------------------------------------------
Recovery furnace systems
Regulatory alternatives (RA) ---------------------------------------------- Smelt dissolving Lime kilns
NDCE DCE tanks
----------------------------------------------------------------------------------------------------------------
RAI (MACT floor for existing NDCE recovery furnace DCE recovery furnace Wet scrubber 1.. ESP 1 or wet
sources). with ESP 1. with ESP 1. scrubber 1
RA II....................... NDCE recovery furnace DCE recovery furnace Wet scrubber 1.. ESP 1 or wet
with ESP 1. with ESP 1 plus BLO scrubber 1
vent controlled by
incineration.
RA III........................ NDCE recovery furnace NDCE recovery furnace Wet scrubber 1.. ESP 1 or wet
with dry ESP 1 with dry ESP 1 scrubber 1
system. system.
RA IV (MACT floor for new NDCE recovery furnace NDCE recovery furnace Wet scrubber 2.. ESP 2
sources)a. with dry ESP2 system with dry ESP 2
and packed-bed system and packed-
scrubber. bed scrubber.
----------------------------------------------------------------------------------------------------------------
a Tighter PM control is achieved for new sources through the use of a more efficient ESP design (ESP 2) or
scrubber design (wet scrubber 2) than that used under regulatory alternatives I through III (ESP 1 or wet
scrubber 1) for existing sources.
Table 4.--Regulatory Alternatives for Existing Affected Sources at
Sulfite Pulp Mills
------------------------------------------------------------------------
Regulatory alternatives (RA) Basis of alternative
------------------------------------------------------------------------
RA I (MACT floor for existing sources). Fiber-bed demister system.
RA II (MACT floor for new sources)..... Wet scrubber followed by fiber-
bed demister system.
------------------------------------------------------------------------
Table 5.--Regulatory Alternatives for Existing Affected Sources at
Semichemical Pulp Mills
------------------------------------------------------------------------
Regulatory alternatives (RA) Basis of alternative
------------------------------------------------------------------------
RA I (MACT floor for existing and new No control.
sources).
RA II (Beyond-the-MACT floor for Wet ESP followed by
existing and new sources). regenerative thermal oxidizer.
------------------------------------------------------------------------
1. Kraft and Soda Pulp Mills
As shown in Table 5, four regulatory alternatives were considered
for MACT selection for affected sources at kraft and soda pulp mills.
The first regulatory alternative (RA I) represents the proposed MACT
floor for existing affected sources, and the other three alternatives
(RA II, RA III, and RA IV) represent beyond-the-MACT-floor options.
Each of these regulatory alternatives is discussed below by emission
point.
a. NDCE Recovery Furnaces. For NDCE recovery furnaces, the
regulatory alternatives are based on two levels of PM HAP control and
two levels of total gaseous organic HAP control, as measured by
methanol. Under RA I (proposed MACT floor for existing sources), PM HAP
emissions would be controlled through the application of an ESP with a
typical operating SCA of 100 m\2\/(m\3\/sec) (530 ft\2\/1,000 acfm);
the ESP would reduce PM HAP emissions by greater than 99 percent.
The regulatory alternatives RA II and RA III are based on the same
PM HAP control equipment specifications for the NDCE recovery furnace
as RA I (the proposed MACT floor); therefore, no further reduction in
PM HAP emissions would be achieved under RA II and RA III than that
achieved at the floor. However, under RA III, total gaseous organic HAP
emissions would be controlled to levels beyond the proposed MACT floor
through the application of a dry ESP system (i.e., a dry-bottom ESP
with a dry PM return system). The use of a dry ESP system would result
in a reduction in total gaseous organic HAP emissions from those mills
currently using wet ESP systems (i.e., wet-bottom ESP's or dry-bottom
ESP's with wet PM return systems). Wet ESP systems emit greater
quantities of gaseous organic HAP's because these compounds are
stripped from the black liquor in the bottom of the ESP and in the PM
return system.
[[Page 18774]]
The most stringent beyond-the-floor regulatory alternative (RA IV)
combines the conversion of the ESP system with more stringent PM HAP
control requirements for the furnace. The more stringent PM HAP control
would be obtained through the application of an ESP followed by a
packed-bed scrubber; the typical operating SCA of the ESP would be
between 110 and 130 m\2\/(m\3\/sec) (570 and 670 ft\2\/1,000 acfm).
Although the packed-bed scrubber is capable of reducing HCl emissions
from the NDCE recovery furnace by as much as 99 percent, as stated in
section IV.E.1.a of this preamble, the ESP could be used alone to meet
the PM emission limit for new NDCE recovery furnaces because the
scrubber removes little, if any, of the PM remaining in the gas stream
exiting the ESP. Because the PM HAP control costs for RA IV are based
on an ESP followed by a packed-bed scrubber, those costs are
overstated. Regulatory alternative IV is representative of the best-
controlled similar source for NDCE recovery furnaces.
b. DCE Recovery Furnace Systems. For DCE recovery furnace systems,
the regulatory alternatives are based on two levels of PM HAP control
and three levels of total gaseous organic HAP control, as measured by
methanol. Under the proposed MACT floor regulatory alternative RA I, PM
HAP emissions would be reduced through the application of an ESP with a
typical operating SCA of 90 m\2\/(m\3\/sec) (430 ft\2\/1,000 acfm).
The beyond-the-floor regulatory alternative RA II is based on the
same PM HAP control equipment specifications for the DCE recovery
furnace as RA I; however, total gaseous organic HAP emissions also
would be reduced by controlling the vent gases from air-based BLO
systems to a beyond-the-floor level via incineration. The use of an
incineration device such as a power boiler or thermal oxidizer could
achieve total gaseous organic HAP emission reductions of 98 percent or
greater from air-based BLO systems, which would translate to a 38
percent reduction of total gaseous organic HAP emissions from the
entire DCE recovery furnace system.
The beyond-the-floor regulatory alternative RA III is based on the
conversion of the DCE recovery furnace to an NDCE recovery furnace
equipped with a dry ESP system with a typical operating SCA of 100
m\2\/(m\3\/sec) (530 ft\2\/1,000 acfm). The conversion of the DCE
recovery furnace would reduce total gaseous organic HAP emissions from
the DCE recovery furnace system by approximately 93 percent. No further
reduction in PM HAP emissions would be achieved under RA III than that
achieved at the floor (RA I) for DCE recovery furnaces.
The most stringent beyond-the-floor regulatory alternative (RA IV)
combines the conversion of the DCE recovery furnace with more stringent
PM HAP control requirements for the furnace. The more stringent PM HAP
control requirements are based on an ESP with a typical operating SCA
between 110 and 130 m\2\/(m\3\/sec) (570 and 670 ft\2\/1,000 acfm)
followed by a packed-bed scrubber. Although the packed-bed scrubber is
capable of reducing HCl emissions from the DCE recovery furnace by as
much as 99 percent, as stated in section IV.E.1.a of this preamble, the
ESP could be used alone to meet the PM emission limit for new recovery
furnaces because the scrubber removes little, if any, of the PM
remaining in the gas stream exiting the ESP. Because the PM HAP control
costs for RA IV are based on an ESP followed by a packed-bed scrubber,
those costs are overstated. Regulatory alternative IV is representative
of the best-controlled similar source for DCE recovery furnace systems.
c. Smelt Dissolving Tanks. For SDT's, the regulatory alternatives
are based on two levels of PM HAP control. Regulatory alternatives I
through III are based on the use of a wet scrubber designed to meet the
NSPS PM emission level. The beyond-the-floor regulatory alternative RA
IV is based on the use of a high-efficiency wet scrubber designed to
reduce PM emissions from SDT's. Based on current information, no
controls more stringent than the use of high-efficiency wet scrubbers
are being applied to SDT's.
d. Lime Kilns. Two PM HAP control levels were considered for lime
kilns. Under regulatory alternatives I through III, the PM control
level is based on the level achievable with a wet scrubber or an ESP
designed to meet the NSPS. Under the beyond-the-floor regulatory
alternative RA IV, increased PM control is obtained through the
application of an ESP with a typical operating SCA of 220 m\2\/(m\3\/
sec) (1,120 ft\2\/1,000 acfm).
2. Sulfite Pulp Mills
As shown in Table 4, two regulatory alternatives were considered
for sulfite combustion units. Both of these alternatives would reduce
PM HAP emissions from the sulfite combustion unit. Regulatory
alternative I represents the proposed MACT floor for existing sulfite
combustion units and is based on the use of a fiber-bed demister
system. Regulatory alternative II is more stringent than the proposed
MACT floor option and is based on the use of a wet scrubber followed by
a fiber-bed demister system.
3. Stand-Alone Semichemical Pulp Mills
As shown in Table 5, two regulatory alternatives for total gaseous
organic HAP's were considered for combustion sources at stand-alone
semichemical pulp mills. Regulatory alternative I represents the MACT
floor for existing sources, which is no control. Regulatory alternative
II is more stringent than the MACT floor option and is based on the use
of a wet ESP followed by an RTO to reduce HAP emissions from the
semichemical combustion units.
G. Selection of Proposed Standards for Existing and New Sources
1. Existing Sources
The proposed standards for each emission point are based on the
emission level achievable when MACT is applied to that source. For
existing sources, MACT was determined by evaluating the regulatory
alternatives presented in Tables 3 through 5. The Agency selected RA I,
or the MACT floor alternative, as MACT for existing sources at kraft,
soda, and sulfite pulp mills. The decision to select RA I was based on
a comparison of the costs and benefits of the regulatory alternatives
for existing sources at kraft, soda, and sulfite pulp mills. The Agency
concluded that the benefits of additional controls beyond the MACT
floor for kraft, soda, and sulfite pulp mills do not outweigh the high
capital costs (shown in Tables 6 and 7).
Table 6.--Nationwide Costs Associated With Regulatory Alternatives for Kraft and Soda Affected Sources
----------------------------------------------------------------------------------------------------------------
Total capital
Regulatory alternatives (RA) investment, Total annual
dollar cost, dollar/yr
----------------------------------------------------------------------------------------------------------------
RA I (MACT floor for existing sources).................................... 219,000,000 23,000,000
[[Page 18775]]
RA II (Beyond the floor for existing sources)............................. 343,000,000 57,000,000
RA III (Beyond the floor for existing sources)............................ 1,450,000,000 64,400,000
RA IV (Beyond the floor for existing sources; MACT floor for new sources). 2,080,000,000 152,000,000
----------------------------------------------------------------------------------------------------------------
Table 7.--Nationwide Costs Associated With Regulatory Alternatives for Sulfite Affected Sources
----------------------------------------------------------------------------------------------------------------
Total capital
Regulatory alternatives (RA) investment, Total annual
dollar cost, dollar/yr
----------------------------------------------------------------------------------------------------------------
RA I (MACT floor for existing sources).................................... 11,400,000 5,120,000
RA II (Beyond the floor for existing sources; MACT floor for new sources). 19,600,000 8,770,000
----------------------------------------------------------------------------------------------------------------
Table 8.--Nationwide Costs Associated With Regulatory Alternatives for Semichemical Affected Sources
----------------------------------------------------------------------------------------------------------------
Total capital
Regulatory alternatives (RA) investment, Total annual
dollar cost, dollar/yr
----------------------------------------------------------------------------------------------------------------
RA I (MACT floor for existing and new sources)............................ 0 0
RA II (Beyond the floor for existing and new sources)..................... 28,100,000 6,860,000
----------------------------------------------------------------------------------------------------------------
The Agency selected RA II, or the beyond-the-floor alternative, as
MACT for existing sources at stand-alone semichemical pulp mills. The
decision to select RA II was based on (1) the suitability of RTO
technology for use with fluidized-bed reactors, which emit the highest
quantities of gaseous organic HAP's of the chemical recovery combustion
technologies currently in use at stand-alone semichemical pulp mills;
(2) the plans of one semichemical mill to install a full-scale RTO
system (preceded by a wet ESP) following a successful RTO pilot study;
and (3) the low cost-effectiveness value associated with a combination
wet ESP and RTO. (The cost-effectiveness value is less than $2,800/Mg
HAP's [$2,500/ton HAP's] based on conservative cost estimates.) Table 8
presents the costs associated with the regulatory alternatives for
existing sources at stand-alone semichemical pulp mills.
Information on the costs and environmental impacts of each
alternative can be found in the memorandum entitled ``Nationwide Costs,
Environmental Impacts, and Cost-Effectiveness of Regulatory
Alternatives for Kraft, Soda, Sulfite, and Semichemical Combustion
Sources'' (docket entry No. II-B-63). The economic impacts of each
alternative are discussed in ``Economic Analysis for the National
Emission Standards for Hazardous Air Pollutants for Source Category:
Pulp and Paper Production; Effluent Limitations Guidelines,
Pretreatment Standards, and New Source Performance Standards: Pulp,
Paper, and Paperboard Category--Phase I'' (docket entry No. II-A-32),
hereafter referred to as the ``Economic Analysis Document.''
2. New Sources
The most stringent regulatory alternatives examined for existing
sources (RA IV for kraft and soda pulp mills; RA II for sulfite pulp
mills; and RA II for stand-alone semichemical pulp mills) are
representative of MACT for new sources. The proposed standards are
equivalent to the emission level achieved by the application of MACT.
The proposed new source MACT for kraft and soda pulp mills is
represented by (1) an NDCE recovery furnace equipped with a dry ESP
system with an SCA between 110 and 130 m2/(m3/
sec) (570 and 670 ft2/1,000 acfm) followed by a packed-bed
scrubber for both NDCE and DCE recovery furnaces, (2) a wet scrubber
designed to meet a PM emission limit of 0.06 kg/Mg (0.12 lb/ton) of
black liquor solids fired for SDT's, and (3) an ESP with an SCA of 220
m2/(m3/sec) (1,120 ft2/1,000 acfm) for
lime kilns. The proposed new source MACT for sulfite combustion units
is represented by a wet scrubber followed by a fiber-bed demister
system. The proposed new source MACT for semichemical combustion units
is represented by a wet ESP followed by an RTO.
H. Selection of Format of the Standards
1. PM HAP Standards for Kraft and Soda Pulp Mills
In selecting the type and format of the proposed PM HAP standard
for kraft and soda pulp mills, the Agency took into consideration the
fact that the HAP fraction of the PM emitted was small (approximately
0.25 percent). Consequently today's proposed standards provide owners
and operators of existing affected sources at kraft and soda pulp mills
several alternatives for meeting the proposed PM HAP standards. Owners
or operators of existing affected sources would be allowed to comply
with either the PM or the PM HAP emission limit set for each source. In
addition, as an alternative to meeting either the PM or PM HAP emission
limits for each existing affected source, the proposed rule would allow
owners or operators to comply with the PM HAP standards by using a
bubble compliance alternative that groups PM or PM HAP emissions from
all existing sources together. Under the proposed bubble compliance
alternative, owners or operators could control PM or PM HAP emissions
more than required at one emission point, where control costs are
relatively low, in return for a comparable relaxation of controls at a
second emission point where control costs are higher. This approach
allows the owner or operator the maximum degree of flexibility in
developing the PM or PM HAP control strategy for existing sources in
the chemical recovery area while reducing HAP emissions to the same
levels that would be achieved through the application of MACT for each
affected source.
The proposed bubble compliance alternative only applies to existing
sources at kraft and soda pulp mills.
[[Page 18776]]
New sources must meet the applicable PM emission limits proposed for
new sources.
This text is long and has been trimmed here. Open the source document for the complete record.
This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.