Rule: Air Pollutants, Hazardous National Emission Standards: Portland Cement Manufacturing Industry

Wednesday,
December 20, 2006
Part IV
Environmental
Protection Agency
40 CFR Part 63
National Emission Standards for
Hazardous Air Pollutants From the
Portland Cement Manufacturing Industry;
Final Rule and Proposed Rule
pwalker on PRODPC60 with RULES3
VerDate Aug<31>2005 18:52 Dec 19, 2006 Jkt 211001 PO 00000 Frm 00001 Fmt 4717 Sfmt 4717 E:\FR\FM\20DER3.SGM 20DER3
76518 Federal Register / Vol. 71, No. 244 / Wednesday, December 20, 2006 / Rules and Regulations
ENVIRONMENTAL PROTECTION the maximum achievable control Publicly available docket materials are
AGENCY technology (MACT) floor standards for available either electronically through
hydrogen chloride (HCl), mercury, and www.regulations.gov or in hard copy at
40 CFR Part 63 total hydrocarbons (THC), and metal EPA Docket, EPA/DC, EPA West, Room
[EPA–HQ–OAR–2002–0051; FRL–8256–4] hazardous air pollutants (HAP). 3334, 1301 Constitution Ave., NW.,
EPA published a proposed response Washington, DC. The Public Reading
RIN 2060–AJ78 to the court’s remand on December 2, Room is open from 8:30 a.m. to 4:30
2005. We received over 1700 comments p.m., Monday through Friday, excluding
National Emission Standards for
on the proposed response. This action legal holidays. The telephone number
promulgates EPA’s final rule for the Public Reading Room is (202)
Portland Cement Manufacturing
amendments in response to the court’s 566–1744, and the telephone number for
Industry
remand and the comments received on EPA Docket Center is (202) 566–1742.
AGENCY: Environmental Protection the proposed amendments. FOR FURTHER INFORMATION CONTACT: Mr.
Agency (EPA). DATES: This final rule is effective on Keith Barnett, EPA, Office of Air Quality
ACTION: Final rule. December 20, 2006. Planning and Standards, Sector Policies
SUMMARY: On June 14, 1999, under the ADDRESSES: EPA has established a
and Programs Division, Metals and
authority of section 112 of the Clean Air docket for this action under Docket ID Minerals Group (D243–02), Research
Act (CAA), EPA promulgated national No. EPA–HQ–OAR–2002–0051. All Triangle Park, NC 27711; telephone
emission standards for hazardous air documents in the docket are listed on number (919) 541–5605; facsimile
pollutants (NESHAP) for new and the www.regulations.gov Web site. number (919) 541–3207; e-mail address
existing sources in the Portland cement Although listed in the index, some barnett.keith@epa.gov.
manufacturing industry. On December information is not publicly available, SUPPLEMENTARY INFORMATION:
15, 2000, the United States Court of e.g., confidential business information
I. General Information
Appeals for the District of Columbia (CBI) or other information whose
Circuit (D.C. Circuit) remanded parts of disclosure is restricted by statute. A. Does this action apply to me?
the NESHAP for the Portland cement Certain other material, such as Entities potentially affected by this
manufacturing industry to EPA to copyrighted material, is not placed on action are those that manufacture
consider, among other things, setting the Internet and will be publicly Portland cement. Regulated categories
standards based on the performance of available only in hard copy form. and entities include:
TABLE 1.—REGULATED ENTITIES TABLE

Category NAICS 1 Examples of regulated entities
Industry ...................................................... 32731 ........... Owners or operators of Portland cement manufacturing plants.
State .......................................................... None ............ None.
Tribal associations ..................................... None ............ None.
Federal agencies ....................................... None ............ None.
1 North American Industry Classification System.
This table is not intended to be procedure raised with reasonable VI. Responses to Major Comments
exhaustive, but rather provides a guide specificity during the period for public VII. Summary of Environmental, Energy, and
for readers regarding entities likely to be comment can be raised during judicial Economic Impacts
regulated by this action. This table lists review. Moreover, under section A. What facilities are affected by the final
amendments?
the types of entities that may potentially 307(b)(2) of the CAA, the requirements
B. What are the air quality impacts?
be regulated by this action. To established by the final NESHAP may C. What are the water quality impacts?
determine whether your facility is not be challenged separately in any civil D. What are the solid waste impacts?
regulated by this action, you should or criminal proceeding brought to E. What are the energy impacts?
carefully examine the applicability enforce these requirements. F. What are the cost impacts?
criteria in 40 CFR 63.1340 of the rule. C. How is this Document Organized? G. What are the economic impacts?
If you have questions regarding the VIII. Statutory and Executive Order Reviews
The information presented in this
applicability of this action to a A. Executive Order 12866, Regulatory
preamble is organized as follows: Planning and Review
particular entity, consult the person
listed in the preceding FOR FURTHER I. General Information B. Paperwork Reduction Act
II. Background C. Regulatory Flexibility Analysis
INFORMATION CONTACT section.
III. Summary of the National Lime D. Unfunded Mandates Reform Act
B. Judicial Review. The NESHAP for Association v. EPA Litigation E. Executive Order 13132, Federalism
the Portland Cement Manufacturing IV. EPA’s Final Action in Response to the F. Executive Order 13175, Consultation
Industry were proposed in December 2, Remand and Coordination With Indian Tribal
2005 (70 FR 72330). This action A. Determination of MACT for Mercury Governments
announces EPA’s final decisions on the Emissions G. Executive Order 13045, Protection of
NESHAP. Under section 307(b)(1) of the B. Determination of MACT for HCl Children From Environmental Health
CAA, judicial review of the final Emissions Risks and Safety Risks
C. Determination of MACT for THC H. Executive Order 13211, Actions That
NESHAP is available only by filing a Emissions Significantly Affect Energy Supply,

petition for review in the U.S. Court of D. Evaluation of a Beyond-the-Floor Distribution, or Use
Appeals for the D.C. Circuit by February Control Option for Non-Volatile HAP I. National Technology Transfer and
20, 2007. Under section 307(d)(7)(B) of Metal Emissions Advancement Act
the CAA, only an objection to a rule or V. Other Rule Changes J. Congressional Review Act
Federal Register / Vol. 71, No. 244 / Wednesday, December 20, 2006 / Rules and Regulations 76519
II. Background emission limitation in the proposed standards for HCl, mercury, and THC.
Section 112(d) of the CAA requires amendments. This final action reflects (Id. At 641.)
EPA to set emissions standards for our consideration of these comments. The Sierra Club also challenged EPA’s
major stationary sources based on We have previously amended the decision not to set beyond-the-floor
performance of the MACT. The MACT Portland Cement NESHAP. Consistent emission limits for mercury, THC, and
standards for existing sources must be at with the terms of a settlement agreement non-volatile HAP metals (for which PM
least as stringent as the average between the American Portland Cement is a surrogate). The court only addressed
emissions limitation achieved by the Alliance and EPA, EPA adopted final the absence of beyond-the-floor
best performing 12 percent of existing amendments and certain interpretative emission limits for non-volatile HAP
sources in the category or subcategory or clarifications to the rule on April 5, metals since EPA was already being
the best performing five sources for 2002 (76 FR 16614), July 5, 2002 (67 FR required to reconsider MACT floor
source categories with less than 30 44766), and December 6, 2002 (67 FR emission standards for mercury, THC,
sources (CAA section 112(d)(3)(A) and 72580). These amendments generally and HCl, and thus, by necessity, also
(B)). This level is called the MACT floor. relate to the rule’s applicability, and to must consider whether to adopt beyond-
For new sources, MACT standards must the performance testing, and monitoring the-floor standards for these HAP. The
be at least as stringent as the control provisions of the rule. In this action, we Sierra Club argued, and the court
level achieved in practice by the best are also amending the rule to re-insert agreed, that in considering beyond-the-
controlled similar source (CAA section two paragraphs relating to the floor standards for non-volatile HAP
112(d)(3)). EPA also must consider more applicability of the Portland cement metals, EPA considered cost and energy
stringent ‘‘beyond-the-floor’’ control new source performance standards that requirements but did not consider non-
options. When considering beyond-the- were deleted in error in a previous air quality health and environmental
floor options, EPA must consider not amendment. impacts as required by the CAA (Id. at
only the maximum degree of reduction It should be noted that the rule text 634–35). The court also found EPA’s
in emissions of HAP, but must take into presented in this notice includes parts analysis of beyond-the-floor standards
account costs, energy, and non-air of the rule that are not being amended. deficient in its assertion that there were
quality health environmental impacts This is done because, in some cases, no data to support fuel switching
when doing so. adding additional rule text reduces the (switching to natural gas) as a viable
On June 14, 1999 (64 FR 31898), in possibility of errors in updating the option of reducing emissions of non-
accordance with these provisions, EPA Code of Federal Regulations. volatile HAP metals (Id. at 635).
published the final rule entitled IV. EPA’s Final Action in Response to
‘‘National Emission Standards for III. Summary of the National Lime
Association v. EPA Litigation the Remand
Portland Cement Manufacturing Following promulgation of the A. Determination of MACT for Mercury
Industry’’ (40 CFR part 63, subpart NESHAP for Portland cement Emissions
LLL).1 manufacturing, the National Lime 1. Floor Determinations
The legacy public docket for the final Association and the Sierra Club filed
rule is Docket No. A–92–53. The final In developing the proposed
petitions for review of the standards in amendments we systematically
rule provides protection to the public by the D.C. Circuit. The American Portland
requiring Portland cement evaluated all possible means of
Cement Alliance, although not a party to developing a quantified floor standard
manufacturing plants to meet emission the litigation, filed a brief with the court
standards reflecting the performance of for mercury emissions from these
as amicus curiae. The court denied sources, including both back end
the MACT. Specifically, the 1999 final essentially all of the petition of the
rule established MACT-based emission technology-based pollution control
National Lime Association, but granted devices and front end feed and fuel
limitations for particulate matter (as a part of the Sierra Club petition.
surrogate for non-volatile HAP metals), control. See National Lime, 233 F. 3d at
In National Lime Association v. EPA, 634 (finding that EPA had erred in
dioxins/furans, and for greenfield 2 new 233 F. 3d 625 (D.C. Cir. 2000), the court
sources, THC (as a surrogate for organic examining only technological (i.e., back-
upheld EPA’s determination of MACT end) controls in considering a level for
HAP). We considered, but did not floors for particulate matter (PM) (as a
establish limits for, THC for existing a mercury floor). We also were unable
surrogate for non-volatile HAP metals) to devise any type of work practice
sources and HCl or mercury for new or and for dioxin/furan. However, the
existing sources. In response to the standard that would result in mercury
court rejected EPA’s determination that emissions reductions (70 FR 72332—
mandate of the D.C. Circuit arising from it need not determine MACT floors for
litigation summarized below in this 72335, December 2, 2005).3
the remaining HAP emitted by these In response to comments on the
preamble, on December 2, 2005, we sources, namely, mercury, other organic
proposed amendments addressing proposed standards, we have performed
HAP (for which THC are a surrogate), additional evaluations of potential
standards for these pollutants. We and HCl (233 F. 3d at 633). The court
received over 1700 comments on the floors for mercury emissions (and also
specifically rejected the argument that performed additional evaluations of
proposed amendments. Most of these EPA was excused from establishing
comments were from the general public floor levels because no ‘‘technology- 3 Indeed, most of the options EPA considered are
and addressed the lack of a mercury based pollution control devices’’ exist to really beyond-the-floor alternatives, because they
control the HAP in question (Id. at 634). reflect practices that differ from those now in use
1 Cement kilns which burn hazardous waste are
by any existing source (including the lowest
in a separate class of source, since their emissions The court noted that EPA is also emitters). (Coal switching, switching to natural gas,
differ from Portland cement kilns as a result of the specifically obligated to consider other and raw material switching are examples.) In EPA’s
hazardous waste inputs. Rules for hazardous waste- pollution-reducing measures including view, a purported floor standard which forces every
burning cement kilns are found at subpart EEE of process changes, substitutions of source in a category to change its practices is a
part 63. beyond-the-floor standard. Such a standard may not
2 A new greenfield kiln is a kiln constructed after materials inputs, or other modifications be adopted unless EPA takes into account costs,
March 24, 1998 at a site where there are no existing (Id.). The court remanded the rule to energy, and non-air health and environmental
kilns. EPA to set MACT floor emission impacts. 70 FR 72335.
beyond-the-floor options for mercury were facilities with multiple stack tests certain that the variability reflected in
control). We obtained additional for mercury. We do have multiple test these results fails to cabin the total raw
mercury emissions test data during and results for one of the lowest mercury material and emissions variability
after the two comment periods on the emitters in the data base. During the experienced by the plants in the source
proposed amendments and once again first test with the raw mill on 7 the category, since we have only a handful
evaluated setting a floor based on the facility was one of the lower emitting of results. These data confirm our
median of the 12 percent of the kilns facilities in the source category tentative conclusion at proposal that
demonstrating the lowest mercury demonstrating emissions of 7.8 constantly changing concentrations of
emissions in stack tests. We discuss micrograms per dry standard cubic mercury in kiln inputs leave no reliable
each of these possibilities in turn below. meter (µg/dscm) (all test values are way to quantify that variability. 70 FR
a. Control of Mercury in Primary 4 Raw corrected to seven percent oxygen). 72333.
Materials and Fossil Fuels. i. Mercury During a second test 8 years later In the proposed amendments we also
Emission Levels Reflecting Raw Material (reflecting raw materials from the same evaluated requiring facilities to switch
and Fossil Fuel Contributions are quarry) mercury emissions with the raw from coal to natural gas as a method to
Inherently Site-Specific. mill on were 60 µg/dscm, a variability reduce mercury emissions, or requiring
As stated at proposal, mercury factor of roughly 8 times. We could use of so-called clean coal (70 FR
emissions come from the predominant identify no facility operational changes 72333–34). We tentatively concluded
input to a cement kiln by volume: The between the times of the two tests that that this was not feasible on a national
limestone which is the chief raw would account for this large difference basis due to insufficient supply and lack
material for the kiln.5 Small amounts of in mercury emissions. of infrastructure, and reiterate that
mercury also are found in other raw conclusion here. One commenter noted
We also obtained data from a facility
material inputs to the process.6 Fossil that petroleum coke was another fuel
that was retested for mercury in July
fuel, almost always coal, is the other that is lower in mercury and is currently
2005, within 3 months of an initial test.
source of mercury emissions. Mercury used as a cement kiln fuel. However, a
With the raw mill on, mercury
levels in limestone vary enormously, mercury standard based on requiring
emissions averaged 0.00138 pounds per
both within a single quarry and between fuel switching to petroleum coke suffers
hour in the April test and 0.00901
quarries, the result being that a single from the same defects as requiring
pounds per hour in the July test, a
source may be unable to replicate its facilities to switch to natural gas. This
own performance in different tests, and variability factor of 7. With the raw mill
fuel may not be available in all areas of
could not duplicate a second source’s off, emissions averaged 0.00823 pounds
the country and there may not be
performance since a kiln lacks access to per hour in the April test and 0.0189
sufficient availability of the fuel to
any other kiln’s limestone. Mercury pounds per hour in the July test. We
replace a significant percentage of the
levels in coal likewise vary also noted that during the April test
coal burned in cement kilns. Petroleum
significantly, although mercury mercury emissions with the raw mill off
coke is a byproduct of petroleum
emissions due to coal are normally were below mercury emissions with the
refining, therefore the supply is limited
swamped by the emissions attributable raw mill on in the July test. Because it by the demand for refined petroleum
to limestone (70 FR 72333–34). is known that when the raw mill is on fuels. Petroleum coke has a low volatile
In an attempt to quantify the potential the raw meal adsorbs mercury, thereby matter content which can lead to
variability, we looked to see if there reducing measured mercury emissions ignition problems if burned without a
in the short term, we can only assume supplemental fuel. It also typically has
4 We discuss in section IV.A.1.c below floor that the uncontrolled variation in the a higher sulfur content than coal. This
determinations for cement kilns using secondary mercury levels in the raw materials—all can adversely affect kiln refractory life
materials (utility fly ash) as raw materials, in place of which come from the same quarry—
of primary materials. and increase internal corrosion of the
5 Limestone makes up approximately 75 percent
was so great between the two tests that kiln shell. As previously noted, each
of the mass input to the kiln. Typically the way a it negated the effect of the operating individual facility has specific
cement plant is sited is that a limestone quarry condition of the raw mill. requirements for raw material additives
suitable for cement production and that is expected We also assessed potential variability
to provide many years of limestone is identified and based on the chemical composition of
the plant is built next to the quarry. There are cases by examining daily variations in cement its limestone. The minerals present in
where a cement plant may purchase small amounts kilns’ raw materials and fuel mercury the coal ash fulfill part of those
of limestone to blend with the limestone from its contents. We obtained data from an requirements. Therefore, replacing part
quarry. However, this close proximity of the quarry operating facility that analyzed samples
and cement plant is an inherent part of the cement or all of the coal currently used at a
manufacturing process and, therefore, a cement of raw material and fuel each day over facility with petroleum coke, which has
plant does not have the flexibility to obtain the bulk a 30 day period. We calculated average almost no ash, may force the facility to
of its limestone from any other source. See 70 FR daily emissions assuming all the
72333. incorporate additional raw material
6 Post-proposal review of available data on other
mercury in the raw materials and fuel additives containing mercury to
mercury raw materials indicates that other feed was emitted. The average daily compensate for the loss of the coal ash.
materials also contribute some mercury, though, in emissions would vary from a low of 0.09 Thus, we adhere to the tentative
most cases, less than limestone. Other raw materials lb to a maximum of 16.44 lb, or a factor conclusion reached at proposal: front
include (but are not limited to): shale or clay to of 183 (See Summary of Mercury Test
provide alumina; iron ore to provide iron; and sand end feed and fuel control of cement
to provide silica. These raw materials are used in data in Docket 2002–0051). kilns is inherently site specific, and
lesser amounts than limestone, and a cement plant These are enormous swings in basing limits on kiln performance in
may have some flexibility in the sources of other variability.8 Moreover, it is virtually
raw materials. As noted in the preamble to the
individual performance tests which
proposed amendments, there are cases where a reflect only those inputs will result in
7 See section c. below discussing operation of the
facility made changes to their raw materials (other limitations that kilns can neither
then limestone) to reduce mercury emissions. in-line raw mill and its implication for mercury
control.
duplicate (another kiln’s performance)
However, this type of control is site specific based
on the available materials and the chemical 8 Variability of emissions based on the operation nor replicate (its own).
composition of the limestone. The site specific of air pollution controls are typically lower that
factors preclude using this as a basis for a national those shown above because air pollution controls reduction or outlet emissions levels and to account
rule (70 FR 72334). are typically designed to meet certain percent for variations in inlet conditions.
ii. Implications of Permit Limits for specific limit based on the same air quality health and environmental
Mercury. There are currently 19 cement procedures the facilities with permits impacts. See 70 FR 72335.
kilns (out of 70 cement kilns for which used: determining in the course of the We are aware that in the case of the
we reviewed permit requirements) that permitting process what its maximum NESHAP for Industrial, Commercial,
have permit limits for mercury. At first conceivable mercury emissions are and Institutional Boilers and Process
blush, it might be argued that these likely to be based on the facility’s raw Heaters (Boiler NESHAP), we used short
permit limits demonstrate that material and fuel inputs, and tacking on term emissions data and applied a
variability of mercury emissions can be an additional variability factor. variability factor to determine a floor for
controlled, since sources must comply However, this would require that we set mercury emissions (69 FR 55236,
with the limitations. It might further be a separate limit for each facility, with September 13, 2004). We do not believe
argued that these permit limits are each facility being its own subcategory that approach is applicable to the
‘‘emission limitations achieved,’’ the (i.e. a different type of facility) based on Portland cement source category. First,
statutory basis for establishing floors for its site specific raw materials and fuels. in the case of the Boiler NESHAP the
existing sources under section 112(d)(3). See 70 FR 72334, alluding to this floor was based on performance of a
Likewise, for new sources, the lowest possibility. EPA has great discretion in control technology, fabric filters, which
permit limit is arguably a measure of deciding whether or not to means that facilities were exercising
performance of the ‘‘best controlled subcategorize within a source category. some control over mercury emissions
similar source’’ (the permit itself being We do not believe a decision to and variability could be realistically
the means of control). We have individually subcategorize is warranted cabined and quantified, so that an
determined, however, that for most considering the fact that the result will emission limit could be replicable and
facilities, the permit limit was be no discernable environmental benefit duplicable. Though the majority of
established based on an estimate because conduct will be unaltered. cement kilns also use fabric filters, the
provided by the facility of the annual Chemical Mfr’s Ass’n v. EPA, 217 F. 3d collected particulate in this source
amounts of mercury that would enter 861, 866–67 (D.C. Cir. 2000) (arbitrary category consists of product and, to
the kiln with the raw materials and and capricious for EPA to impose costly some extent, unprocessed raw materials.
fuels. One facility had a mercury limit regulatory obligations without some As a result most of the collected
based on its estimated annual emission showing that the requirement furthers particulate is recycled back to the
from an emissions test, and one facility the CAA’s environmental goals). process, largely negating any impact of
had a limit based on a State law, Therefore, we have determined that the particulate control technology on
although in neither case did the even though these permit limits exist, mercury emissions.9 Second, the
resulting permit cause a cement kiln they have not resulted in a quantifiable variabilities seen as a result of fuel
source to alter or otherwise modify its reduction of mercury emissions. Any inputs in the Boiler NESHAP are much
existing practices to meet the limit. option to develop a MACT floor for lower than the variabilities indicated in
Thus, we find no cases where a facility mercury with these limits would either the Portland cement industry where the
actually has had to take any steps, either result in an unnecessarily complex rule mercury fuel variability is a distant
through the imposition of process with no environmental benefit, or a rule second to the enormous variability of
changes or add-on controls, to reduce its mercury in the raw materials. We do not
which improperly imposes a de facto
mercury emissions as a result of any of believe the data exist to accurately
beyond-the-floor standard without the
these permit limits. See ‘‘Summary of quantify this variability.
required consideration of costs, energy
Cement Kiln Permit Data for Mercury’’ Another option we considered was
and non-air quality impacts.
in the docket. using long term data to set a floor.
We considered the option of setting iii. Why not Average the Performance
However, since, to our knowledge,
an emissions limit, either on a pounds Test Data? Some commenters stated that
continuous emission monitors for
per year (lb/yr) or a pound per ton of EPA must simply average the results of
mercury have not been demonstrated on
clinker basis, based on the median of the 12 per cent lowest mercury
cement kilns, and none currently exist
the top 12 percent of the 17 kilns with performance test data to establish the
on cement kilns, there is no long term
permit limitations. However, we repeat floor for existing sources, and establish
stack performance data on mercury
that none of the facilities with permit the new source performance floor at the
emissions from cement kilns that we
limits were required to actually take level of the lowest test result. We
could use to set a numerical emissions
action to reduce mercury emissions. rejected this approach at proposal, and
limit. The only available long term data
Their limits were all based on site do so here, because it fails to account for
of which we are aware is from several
specific factors (expected maximum the variability of mercury levels in raw
facilities which have a requirement to
conceivable levels of mercury materials and fuels and hence
perform monthly analyses of
emissions), and were set at a level that variability in performance. See 70 FR
composited daily samples of fuels and
did not require the imposition of add- 72335; see also 70 FR 59436 (Oct. 12,
raw materials to calculate a 12 month
on controls, feed or fuel substitution, or 2006). We must, of course, account for
mercury emissions total. However, all
any other constraint. Any limit we set sources’ variability in establishing a
these kilns are located in one state
based on these permits would require MACT floor. Mossville Environmental
(Florida) with unrepresentatively low
that at least some facilities apply Action Now v. EPA, 370 F. 3d 1232,
levels of mercury in limestone (so far as
beyond-the-floor control technology to 1241–42 (D.C. Cir. 2004). The only way
we can determine). We do not believe
meet the limit since feed and fuel all kilns, including the kilns with the
these data would be representative of
control via substitution is not possible. lowest emission levels in individual
Such a standard would impermissibly tests, could meet this type of standard 9 As explained in the following section of the
apply beyond-the-floor emission control continuously, as required, would be to preamble, however, EPA has determined that the
without consideration of costs and other install backend technology-based floor for both existing and new sources involves the
non-air health and environmental control equipment. However, this would removal from the kiln system of collected
particulate under designated circumstances. In
impacts. be a de facto beyond-the-floor standard, addition, the floor for new sources reflects
We also considered a limit where adopted impermissibly because of reductions in mercury based on performance of a
each facility would set their own site failure to assess cost, energy, and non- wet scrubber.
the source category as a whole. More ash used relative to other raw materials. its use as a feedstock, as well as the
basically, basing a standard on one set These data apparently do not exist, with associated air emissions.
of kilns’ raw material inputs still suffers one exception discussed in the next It should also be noted that there are
from the defect that no facility has paragraph. We do know that the two at least two new facilities whose permits
access to another’s raw materials. highest mercury emitting facilities (in specifically require use of fly ash as
b. Floors for Facilities Using Utility individual performance tests) do not use their alumina source, as they have no
Fly Ash as Raw Material. Some cement fly ash. Without data on the actual source for shale or clay, the primary
kilns use utility fly ash as an alternative mercury contributions of all materials, material alternatives for alumina.
raw material to replace shale or clay.10 we do not believe we can draw any Finally, a facility that currently uses fly
These kilns replace a natural material, valid general conclusions on the impact ash may not be able to return to using
shale or clay, with a secondary material of the use of fly ash on mercury the natural (i.e. primary) raw materials
(i.e. a recycled air pollution control emissions. it replaced. For example, if the replaced
residue), fly ash. Approximately 34 We do have detailed data from one raw materials were shale, the shale
cement manufacturing facilities are facility that used fly ash where 50 quarry may now be closed and the
currently using utility boiler fly ash as percent of the total mercury input to the facility may not have access to a suitable
a feedstock. We reviewed the available kiln is in the fly ash. However, even for shale supply.
data and have come to the conclusion this facility, we cannot accurately Given the lack of any data to
that cement kilns using fly ash are a quantify the impact on mercury positively state the impact of fly ash on
different type of kiln, within the emissions of the decision to replace the mercury emissions for the source
meaning of section 112 (d) (1) of CAA, shale used at this facility with fly ash category in general, as well as the
and that for cement kilns currently because we have been unable to obtain positive environmental effects of using
using fly ash, the current use would be data on the mercury content of the shale fly ash, there is no basis for a floor
considered the MACT floor. Our the fly ash replaced. We also have no standard based on substituting other
reasoning is as follows. mercury analysis data from the time potential raw materials (such as shale or
Use of fly ash can have an effect on clay) for fly ash. At the same time, we
period when the facility used shale.
mercury emissions since fly ash do not see any means of identifying a
contains mercury in varying amounts. There are other factors to consider
floor for existing fly ash users based on
As discussed below, mercury emissions when we evaluate the environmental
substituting different fly ash types
may be higher or lower depending on effects—generally quite positive—of
reflecting different mercury content.
the amounts of mercury involved vis-à- substituting fly ash for shale or clay.
The recycled fly ash is not fungible.
vis the raw materials that would First, fly ash in general has a lower
Cement kilns must carefully select only
otherwise be used (if available). But as organic material content than shale or
fly ash with needed properties within a
also explained more fully below, some clay. At the facility just mentioned,
relatively small tolerance. Cement kilns
cement kilns using fly ash do not have replacing the shale with fly ash reduced
also usually are limited to fly ash
an alternative raw material source. emissions of THC from around 80 parts
available from boilers which are
Given that these kilns use a different per million by volume (ppmv) to 3
reasonably close to the kiln (typically
raw material, not always replaceable, ppmv. Because fly ash can reduce kiln
within a few hundred miles) or shipping
and that the material affects mercury fuel consumption, it reduces emissions
expense becomes prohibitive. The fly
emissions, we believe that these kilns of sulfur dioxide (SO2), oxides of
ash selection process is involved; it has
are a separate kiln type, and hence a nitrogen (NOX), and carbon monoxide
taken years for kilns to identify a
separate subcategory, for purposes of (CO2). Using fly ash as a kiln feed
suitable fly ash source. Accordingly, we
mercury emissions. For a similar reduces the landfill requirements for
evaluate fly ash like the other raw
conclusion see 64 FR at 52871 (Sept. 30, disposal of utility fly ash. Use of fly ash
material inputs into cement kilns, and
1999) (cement kilns that choose to burn reduces cement plant power
do not believe that a floor that is based
hazardous waste in place of fossil fuels consumption because it is usually fine
on substitution of either raw materials
are a separate source category for MACT enough that it can be added directly to
or other fly ash is justified because the
purposes). the kiln rather then being ground in a
input is variable and uncontrollable. We
We attempted to determine if, in mill. Use of fly ash also reduces fuel
discuss in section IV.A.2 below the one
general, facilities that use fly ash have consumption because compared to the
exception to this conclusion for fly ash
higher emissions of mercury than those raw materials it typically replaces it is
where the mercury content has been
that do not. An analysis of data for already highly calcined; it does not have
artificially increased by sorbent
EPA’s toxic release inventory and the the same types of large crystals as the
injection.
National Emissions Inventory did not raw materials it replaces (this improves c. Control of Collected Particulate
show differences significant enough that burnability); some fly ashes have lower (Cement Kiln Dust). There are two
we could draw any definitive metal alkali content, thus avoiding hard operation factors that impact measured
conclusions. We considered reviewing burning to drive off alkali metals and mercury emissions at the kiln stack.
the available mercury emissions test reducing the need to operate the alkali These are the use of in-line raw mills
data to determine if we could discern a bypass; it is drier than quarried and the recycling of cement kiln dust
trend. However, as previously materials, thus saving fuel used to dry (CKD).
discussed, we do not believe these data materials. Many domestic cement plants Many (but not all) kiln systems have
are representative of long term mercury have high pyrites in their quarry, in-line raw mills. In these systems the
emissions. We also attempted to obtain especially in the shale or clay. In most kiln exhaust gas is routed through the
data on the important issue of the cases, this pyrite is the main source of raw mill to dry the raw materials. This
amounts and mercury contents of fly SO2 emissions from the kiln. Using fly process results in mercury contained in
ash can significantly reduce the SO2 the flue gas being adsorbed by the raw
10 Though these are also raw materials inputs, the
emissions that result from pyrite in the meal.11 This results in an apparent
mass of clay or shale is typically less than 15 raw materials. It also reduces the energy
percent of the mass input to the kiln. Limestone
makes up approximately 80 percent of the mass required for the quarring, milling, and 11 More specifically, when the mill is on-line, the
input. transporting of the shale or clay prior to kiln gas containing volatilized mercury is used to
reduction if mercury emissions are both existing and new sources is the However, for new sources mercury
being measured at the kiln stack. work practice that cement kiln dust be emissions would not be uncontrolled—
However, the captured mercury is removed from the kiln system at the solely dependent on raw material
reintroduced into the kiln which creates point that recirculation causes adverse mercury content—but rather would
a recycle loop of mercury until the effect on product. reflect performance of ‘‘the best
captured mercury eventually escapes d. Standards Based on Performance of controlled similar source’’ (section 112
and is emitted to the atmosphere. Also, Wet Scrubbers. There are at least five (d)(3)). A kiln so-equipped would thus
raw mills do not run continuously. cement kilns that have limestone (wet) have the best performance over time,
When the raw mill is turned off, this scrubbers installed for control of SO2. since variability in mercury attributable
effect of raw meal adsorption of mercury Commenters noted that based on to raw material and fuel inputs would
is negated and mercury emissions experience with utility boilers, and be controlled in part.13
appear to increase. However, the other similar combustion devices, there We believe there is a reasonable basis
increase is actually mercury that would is reason to expect that the scrubbers that wet scrubbers remove oxidized
have previously been emitted but was installed on cement kilns also remove mercury from cement kiln emissions.
captured by the raw meal and returned oxidized mercury. First, wet scrubbers are known to
to the kiln. The net effect is that an in- To our knowledge, we obtained all the remove oxidized mercury in most
line raw mill does not increase or available data on wet scrubber combustion applications though
reduce mercury emissions over the long controlled facilities after the comment removal rates vary. We have speciated
term; it simply alters the time at which period on the proposed amendments. mercury test data on two kilns that
the mercury is released. This consists of data from 2004 and indicate that there is a significant
Mercury is also adsorbed on the CKD 2005 tests at two facilities measured amount of oxidized mercury in at least
collected in the particulate control exclusively at the scrubber outlet. These some cement kilns. See mercury
device, typically a fabric filter or an data range from 0.42 to 30 µg/dscm. emission test data for Holcim, Dundee,
electrostatic precipitators (ESP). Variability of mercury emissions at the MI and Lafarge, Alpena, MI, in docket
Because the collected CKD mainly scrubber-equipped kilns for which we EPA–HQ–OAR–2002–0051. Second, the
consists of product, and sometimes have multiple test data differs by orders limited data we have from cement kilns
small amounts of raw materials, the of magnitude. These data fall within the equipped with wet scrubbers is among
collected CKD is recycled back to the range of test data from all kilns (those the lowest end-of-stack mercury data in
kiln to the extent possible. The portion with wet scrubbers and those without our data base (although not the lowest),
that cannot be recycled to the kiln is wet scrubbers). We have no test data for which could indicate that some removal
either sent to a landfill, or used in some mercury measured at the scrubber inlet. mechanism is involved. An important
other manner (i.e. some type of As a result, we cannot, on the basis of caveat, however, is that these data are
beneficial use). Most facilities require the current data, determine with exclusively end-of-stack, without paired
that a portion of the CKD be removed absolute certainty (though we believe it inlet concentrations. These data thus do
from the kiln system rather than is reasonably certain) if the outlet not with absolute certainty demonstrate
returned to the kiln. This is done to mercury emissions from the wet that mercury removal is occurring or
bleed the kiln system of alkali materials scrubber equipped kilns are a result of how much.
that build up as they circulate which mercury removal by the scrubber, or We estimated the performance of the
would otherwise contaminate product simply reflect the amounts of mercury best performing scrubber, and hence the
and damage the kiln lining. This in the raw materials. We now discuss new source MACT floor, to be 41 µg/
practice necessarily reduces the overall the implications of this information for dscm (corrected to 7 percent oxygen)
volume of mercury emitted by cement purposes of existing and new source using the following rationale. First, we
kilns, as noted by several commenters, floors. Note that the following limited the analysis to data from wet
since the entrained mercury in the CKD discussion assumes the scrubbers scrubber equipped kilns because, as just
is no longer available for release from remove oxidized mercury for reasons
the kiln. The amount of reduction is discussed below. the mercury emission levels of the kilns equipped
kiln-specific, based on the level of alkali with wet scrubbers with other kilns whose mercury
First, there are an insufficient number levels reflect raw material and fuel mercury levels
materials in the kiln’s raw materials and of wet-scrubber equipped kilns on at the time of the performance test, the resulting
required product specifications, and which to base an existing source floor. limit would not be achievable over time by any
therefore not quantifiable on a national The scrubber-equipped kilns would source other than one with a wet scrubber.
basis. Nor would kiln-by-kiln site- Ostensible best performers would consequently
represent the best performing sources have to retrofit with back end control, since
specific emission standards be since data from other kilns simply otherwise they could not consistently achieve the
warranted, for the same reasons that reflect the mercury levels in kiln inputs results of their own performance tests.
site-specific limits based on mercury on the day of the test. There are 158 13 That is, variability would no longer be purely
levels on raw material and fuel inputs operating kilns, and the information a function of the happenstance of the amount of
are not justified. EPA is instead mercury in raw materials (and fossil fuels) used in
available to us indicates that only five the test condition. As explained more fully below,
determining that a floor standard for of them are equipped with wet performance of wet scrubbers, however, is variable,
scrubbers. The median kiln of the top 12 based not only on operation of the device but on
sweep the mill of the finely ground raw feed percent would, therefore, not be a mercury levels in input materials. Wet scrubbers on
particles. Since the mill temperature is only about utility boilers, for example, are documented to
90 to 120 °C during this operational mode, the fine scrubber equipped kiln.12 remove between 0 to 72 percent of incoming
PM can adsorb the mercury in the gas stream, and mercury. See Control of Mercury Emissions from
the particles containing condensed mercury are 12 Choosing the median source for assessing an Coal-Fired Electric Utility Boilers: Interim Report
stored in the raw feed silos. This stored raw mix existing source floor here is a reasonable manner of Including Errata available at http://www.epa.gov/
then is fed to the kiln. The captured mercury is determining ‘‘the average emission limitation nrmrl/pubs/600r01109/600r01109.htm. We should
again volatilized and returned in the gas stream to achieved by the best performing 12 percent of note, however, that because utility boilers do not
the raw mill, only to be captured again in the raw existing sources’’ (section 112 (d)(3)). Not only can have the significant levels of alkaline materials that
mill, as described above. This process continues as the statutory term ‘‘average’’ be reasonably are present in cement kilns, which alkaline
long as the raw mill is on-line, and the raw feed interpreted to mean median, but it is appropriate to materials would impede mercury oxidation and
continues to adsorb additional mercury through this do so here in order not to adopt a de facto beyond scrubber efficacy, we do not view utility boilers as
process. the floor standard. If one were simply to combine a ‘‘similar source’’ for purposes of section 112(d)(1).
discussed, the wet scrubber equipped work practice, and in addition, to meet and, for the subcategory of kilns that
kilns represent the best performing a standard of either 41 µg/dscm or a site- currently use fly ash as a raw material,
sources, regardless of their actual outlet specific limit based on performance of replacing the fly ash with a lower
emissions levels in individual a properly designed and operated wet mercury raw material. Substituting raw
performance tests. Second, we ranked scrubber. materials or fossil fuels with lower-
all the wet scrubber mercury emissions As just explained, the mercury data mercury inputs could in theory reduce
with the raw mill off. We believe this is on which the new source floor is based mercury emissions, but this alternative
appropriate because the condition of are not only limited, but fails to is infeasible for the reasons explained at
raw mill off represents a normal definitively answer the critical question 70 FR 72333–72334.
operating mode for a cement kiln (albeit of whether wet scrubbers are removing Generally, once mercury enters a kiln
the operating mode when mercury oxidized mercury, and, if so, to what system, it has five potential fates: it may
emissions would be highest, as extent. We are taking immediate steps to remain unchanged and become part of
discussed above in section a.i). We then address this issue and augment the data the final product; it may react with raw
took the mean raw mill off value for base. In an action published elsewhere materials and exit the kiln in the
mercury emissions from a cement kiln in this Federal Register, we are granting clinker; it may vaporize in the high
in our (limited) data base, and added to reconsideration of the new source temperature of the kiln and/or
it a variability factor to account for standard adopted in this rule, both due preheater; it may condense or react with
normal variation in emissions. This to substantive issues relating to the cement kiln dust and be removed
variability factor is the standard performance of wet scrubbers and from the system; or it may exit the kiln
deviation of the data multiplied by because information about their system in vapor form or be adsorbed to
2.326 (the z statistic) to produce the performance in this industry has not a dust particle through the stack. In
99th confidence interval. We looked to been available for public comment. We general, mercury in the fuel becomes
all of the data, rather than to the data also have initiated actions to obtain volatilized near the kiln’s combustion
from the single lowest emitting kiln, inlet and outlet test data for cement zone and is carried toward the feed end
because there are too few data points kilns equipped with wet scrubbers in of the system along with combustion
from that kiln (or from any one kiln) to order to determine if these controls gases. Some of the mercury compounds
estimate that kiln’s variability. Given remove mercury, and to what extent. In pass through the entire system and exit
that variability is known to occur, we addition, we are committing to in vapor phase through a stack.
believe that this is the best completing this reconsideration process However, as the flue gas cools, some
approximation of variability of the best within one year from December 20, mercury may adsorb/condense onto
performing kiln presently available. 2006. dust particles in the cooler regions of
The result of this analysis is a new the kiln system. Much of this dust
2. Beyond-the-Floor Determinations
source floor of 41 µg/dscm that must be containing condensed mercury would
met continuously (raw mill on and raw During development of the original then be captured by the PM control
mill off) (see further discussion in NESHAP for Portland cement device and for most kiln systems,
section A.3 below). This is an emissions manufacturing, we conducted MACT returned to the kiln.
limit that we believe will not be floor and beyond-the-floor analyses for We evaluated, requiring a facility to
exceeded 99 percent of the time by the kiln and in-line kiln/raw mill mercury further reduce the recycling of CKD
best performing kiln whose performance emissions (63 FR 14182, March 24, 1998 beyond the wastage already needed to
is used to set the standard. and 64 FR 31898, June 14, 1999). We protect product quality, the floor for
Because of the limited performance also conducted a beyond-the-floor both existing and new sources. For a
data characterizing performance of the analysis for mercury, based on the 600,000 tpy (tpy) kiln the estimated
lowest-emitting scrubber-equipped kiln, performance of activated carbon total annual cost would be $3.7 million
the rule also contains an alternative new injection with an additional PM control just for replacement of CKD (which is
source mercury floor. The best device. Costs for the system would actually product) and disposal of
performing kiln is equipped with a wet include the cost of the carbon injection additional solid waste. This cost does
scrubber, although there could be system and an additional fabric filter not account for the increased raw
questions about its performance over (FF) to collect the carbon separately materials costs and energy costs
time. Therefore, if a new source installs from the CKD. Based on the low levels associated with reducing the recycling
a properly designed and operated wet of mercury emissions from individual of the CKD. The mercury emissions
scrubber, and is unable to achieve the Portland cement kilns, as well as the reduction would range from 0.012 to
41 µg/dscm standard, then whatever high cost per ton of mercury removed by 0.055 tpy based on assumed CKD
emission level the source achieves (over the carbon injection/FF system, we mercury concentrations of 0.33 and 1.53
time, considering all normal sources of determined that this beyond-the-floor parts per million (ppm) respectively.
variability) would become the floor for option was not justified (63 FR 14202, The cost per ton of mercury reduction
that source. Based on the design of the March 24, 1998). would range from $67 million to $308
wet scrubber that is the basis of the new At proposal, EPA again concluded million. See Costs and Impacts of
source floor, this would be a packed bed tentatively that a beyond the floor Wasting Cement Kiln Dust or Replacing
or spray tower wet scrubber with a standard based on performance of Fly Ash to Reduce Mercury Emissions
minimum liquid-to-gas ratio of 30 activated carbon is not justified (70 FR in docket EPA–HQ–OAR–2002–0051.
gallons per thousand cubic feet of 72335). We have since reevaluated We note that the median value for the
exhaust gas. beyond-the-floor control options for mercury content of recycled CKD for
In sum, we conclude that floors for mercury emissions. This evaluation one study was only 0.053 ppm. See the
mercury for all existing cement kilns included both process changes and add- report Mercury and Lead Content in
should be to remove accumulated on control technology. Raw Materials in docket EPA–HQ–

mercury-containing cement kiln dust There are two potential feasible OAR–2002–0051. This would indicate
from the system at the point product process changes that have the potential that for the majority of the facilities the
quality is adversely affected. The floor to affect mercury emissions. These are costs per ton would be even higher that
for new sources is to utilize this same removing CKD from the kiln system those presented above. In addition, we
estimate that wasting 50 percent of the Impacts of Wasting Cement Kiln Dust or mercury content due to sorbent
recycled CKD would reduce the energy Replacing Fly Ash to Reduce Mercury injection at the utility site to increase
efficiency of the process by six percent Emissions in docket EPA–HQ–OAR– mercury emissions from cement kilns,
due to the need to process and calcine 2002–0051. Also, this option would and for the increase to be much more
additional feed to replace the wasted mean that all the fly ash currently being significant than emissions attributable
CKD. It is possible that in some cases used as a cement kiln feed would now to the current fly ash being used.
the wasted cement kiln dust could be potentially have to be landfilled. This We do not see a ban on the use of this
mixed with the cement product rather would generate an additional 3 million type of fly ash as significantly affecting
than landfilled, or that some other tpy of solid waste, with potential the overall current beneficial uses of fly
beneficial use could be found. This adverse health and environmental ash. First, we do not anticipate the
would reduce the costs and non-air impacts associated with management of widespread use of activated carbon
adverse impacts of this option. these wastes. There would also be injection ACI in the utility industry
However, there are currently barriers to adverse environmental air and non-air until 2010 or later. Therefore, both the
directly mixing CKD with clinker due to quality health and environmental cement industry and the utility industry
product quality and product impacts associated with the mining of will have a significant amount of time
specification issues. We do not have additional raw materials that would to adjust to this requirement. Second, a
data available to evaluate the potential have to be utilized. In addition, the utility boiler that decides to apply ACI
for beneficial use of the CKD. Based on overall kiln efficiencies (i.e. the amount for mercury control has the option of
these costs, the adverse energy impacts, of fuel required per ton of clinker collecting the fly ash from sorbent
and the increased adverse waste produced) at the facilities using fly ash injection systems separately from the
disposal impacts (see 64 FR 45632, would be expected to decrease if the fly rest of the facility’s fly ash (e.g., EPRI’S
45635–36 (Aug. 20, 1999) for examples ash were replaced with shale or clay. TOXECON system). Therefore, the
of potential hazards to human health This decrease may be as large as 10 utility boiler could continue to supply
and the environment posed by disposal percent (See Site Visit to Lafarge non-sorbent fly ash to a cement kiln
of cement kiln dust), we do not believe Cement in Alpena Michigan in the even after the application of ACI for
this beyond-the-floor option is justified docket). mercury control. Finally, technology is
and therefore are not selecting it. Based on the cost, energy, and adverse being developed that would allow the
As previously noted, for the non-air quality health and mineral-rich portion of fly ash to be
subcategory of facilities that use utility environmental impacts, we believe that separated from the high carbon/high
boiler fly ash as a kiln feed we banning the current use of utility boiler mercury portion.
determined that the current use fly ash is not justified. Based on these factors, we are
represented the MACT floor. We We also separately evaluated the use banning the use of utility boiler fly ash
considered two beyond-the-floor of fly ash from a utility boiler where in cement kilns where the fly ash
options for this subcategory. One option activated carbon, or some other type of mercury content has been increased
was to ban the use of any fly ash if it sorbent injection, has been used to through the use of activated carbon or
resulted in a mercury emissions collect mercury. This practice does not any other sorbent unless the facility can
increase over a raw material baseline, currently occur. See 70 FR 72344 demonstrate that the use of that fly ash
and the second was to only ban the use (voicing concern about potential for will not result in an increase in mercury
of fly ash whose mercury content had increased mercury emissions from
emissions over baseline emissions (i.e.,
been artificially increased through the cement kilns were such fly ash to be
emissions not using the mercury
use of a sorbent to capture mercury in used). The mercury concentration in
increased fly ash). The facility has the
the utility boiler flue gas. this type of fly ash will vary widely.
burden of proving there has been no
If we were to ban the use of utility However, full scale testing of fly ash
emissions increase over baseline. This
boiler fly ash for any case where it has from utility boilers using various
requirement, adopted as a beyond-the-
been shown to increase mercury sorbent injection processes has
floor control, applies to both existing
emissions from the kiln over a raw indicated there is a potential for sorbent
and new sources.
material baseline, facilities would have injection to significantly increase fly ash
We also reevaluated our analysis of
to revert to using their previous raw mercury content (Characterization of
potential control options based on add-
materials, or to find alternative raw Mercury-Enriched Coal Combustion
on control technology. These were
materials that provide the same Residues from Electric Utilities Using
chemical constituents as the fly ash. As Enhanced Sorbents for Mercury Control control options based on the use of a
previously noted, if a facility replaces in the docket EPA–HQ–OAR–2002– limestone scrubber, and ACI.
their shale or clay with fly ash, the 0051). Testing to date has shown As previously noted there are at least
quarry for that material may now be increases by a factor of 2 to 10, and in five cement kilns that have limestone
closed and it may not be possible to one case of a very low mercury fly ash (wet) scrubbers. As discussed in section
cost-effectively obtain the previously the increase was by a factor of 70. IV.A.1.d above, there is a reasonable
used raw materials. And for at least two Data from 16 cement facilities basis for believing that the wet
new facilities, the original raw materials currently using fly ash not reflecting scrubbers remove the oxidized mercury.
used at startup will include fly ash, so sorbed mercury showed mercury There are no data available to allow us
there is no previously used material concentrations in the fly ash from 0.002 to definitively estimate the percent
with which to compare the mercury ppm to 0.685 ppm with a median of reduction expected. We performed a
content of the fly ash. Due to the site 0.136 ppm. Data on the fly ash mercury cost analysis based on an assumed
specific costs associated with raw content of currently operating utility mercury removal efficiency of 42
materials, we don’t have any data to boilers testing sorbent injection showed percent, which is transferred (solely for
purposes of analysis) 14 from
calculate the costs of the beyond-the- levels ranging from 0.071 ppm up to
floor option for the industry as a whole. 1.529 ppm with a median level of 1.156 14 As explained in section 1.d, there are no data
In one example, we estimated the cost ppm, significantly higher than the fly to definitively state that the percent reduction
as approximately $136 million per ton ash currently in use. Therefore, we see achieved by wet scrubbers in the utility industry are
of mercury reduction. See Costs and a potential for fly ash with enhanced Continued
performance of wet scrubbers in the the wet scrubber will achieve cobenefits existing MACT dioxin standard for
utility boiler category and represents the of reducing SO2 and dioxins (although Portland cement kilns). The results of
greatest degree of removal one could dioxin removal would be relatively that analysis for an existing model large
expect to be consistently achieved for modest since any removal would be kiln are as follows:
Portland cement kilns. We also note that incremental to that required by the
TABLE 1.—PACKED BED SCRUBBER—COSTS AND EMISSION REDUCTIONS

Emissions reduction
Clinker production in tpy Total annualized cost
(tpy) ($/yr) D/F Hg
SO2 (tpy) (g/yr) (lb/yr)
600,000 1,542,000 297 0.11 16.8–147
Based on this analysis the cost per ton from cement kiln exhaust gas, we have would be approximately 30 tpy or less.
of mercury removed ranges from $21 no data on the actual expected removal This THC emissions reduction is based
million per ton to $184 million per ton, efficiency. Data are available for one on an assumed control efficiency of 50
a result that is not at all cost effective. emissions test on a cement kiln burning percent. We do not see these small THC
In addition, a wet scrubber for a large hazardous waste. In this test the emission reductions (of which organic
kiln will generate approximately 45,500 mercury removal efficiency averaged 89 HAP are a small subset) to be a reason
tpy of solid waste and require percent removal. However, the inlet to alter our tentative decision at
approximately 980,000 kilowatt hour mercury concentration during the test proposal that a standard based on
per year (kWhr/year) of electricity. varied from 65 to 267 µg/dscm. A performance of ACI is not justified as a
Based on the significant cost impacts review of the data for the individual test beyond-the-floor control option.
per ton of emission reduction, and the runs implies that the percent reduction Finally, for greenfield new sources
adverse energy and solid waste impacts, decreases as the inlet concentration (sources being newly built at a site
and the uncertainty of the actual decreases. Almost all the non-hazardous without other cement kilns), we
mercury emission reductions, we do not waste cement kilns tested had mercury considered the option of requiring such
consider this control option to be concentrations well below 65 µg/dscm. a kiln to be sited at a low-mercury
reasonable for existing sources. Therefore, the long term performance of quarry. This concept has intuitive
At proposal, EPA discussed and ACI on mercury emissions from cement appeal: such a new kiln is not tied to an
rejected a beyond-the-floor option based kilns is very uncertain. We also note existing source of limestone, and so can
on the use of activated carbon injection. that the application of ACI to a cement choose where to be sited. The difficulty
See 70 FR 72335. Commenters noted kiln (either alone or in combination is in quantifying this type of standard.
that our costs for ACI had not been with a wet scrubber) will generate We cannot presently quantify what
updated from the costs calculated in approximately 1,600 tpy of solid waste ‘high mercury quarry’ or ‘low mercury
development of the original NESHAP. In for a new or existing large kiln. quarry’ means, and cannot responsibly
response, we have now updated our ACI Recycling of the waste would be select an arbitrary number that might
costs based on more recent information. unlikely due to the toxics content. make it impossible to build a new
The total annualized costs for a large cement kiln in major parts of the
new or existing kiln ranges from For existing sources we rejected a country.
$510,000 to $676,000 per year. control option based on the performance
Assuming an 80 percent reduction in of ACI due to the significant cost per ton 3. Conclusion
mercury emissions, the cost per ton of of mercury removed, increased energy In sum, we conclude that the
mercury removal ranged from $4 use, and the adverse non-air quality standards for mercury for all existing
million to $42 million per ton for health and environmental impacts (in cement kilns are to remove accumulated
existing kilns. The wide range in cost the form of additional mercury and mercury-containing cement kiln dust
per ton of removal is mainly influenced organic-laden waste generated). For new from the system at the point product
by the baseline mercury emissions. sources we rejected the option based on quality is adversely affected. The
Based on the wide variation we have the performance of ACI combined with standard for new sources is to utilize
seen in actual mercury emissions in this a wet scrubber for essentially the same this same work practice, and in
source category, the actual cost per ton reasons: significant cost per ton of addition, to meet a standard of either 41
would also vary widely from site to site mercury removed, increased energy use µg/dscm or a site-specific limit based on
as shown above. and adverse non-air quality health and performance of a properly designed and
We also evaluated a beyond-the-floor environmental impacts. For both new operated wet scrubber.
option for new kilns based on and existing sources we also rejected In addition, we are banning the use of
combining ACI and a wet scrubber. The this control option due to the utility boiler fly ash in cement kilns
incremental cost of ACI in this uncertainty of the actual performance where the fly ash mercury content has
application is $9 to $89 million per ton levels achieved, which leads to been increased through the use of
of mercury removed, which we regard uncertainty of the actual cost per ton of activated carbon or any other sorbent
as a very high cost. mercury emissions reduction. We also unless the facility can demonstrate that
Our cost estimates assumed 80 note that the application of ACI the use of that fly ash will not result in
percent emissions reduction for potentially could result in a THC an increase in mercury emissions over
mercury. Though we are reasonably emission reduction of up to 117 tpy per baseline emissions (i.e., emissions not
certain that ACI will remove mercury kiln, though in most cases the reduction using the mercury increased fly ash).
representatie of the percent reduction in the figure in beyond-the-floor analyses as an upper

Portland cement source category. We used this bound best case for potential emission reductions.
Because the final standard is more were operated only intermittently, their decided to adopt this approach. Under
stringent than the standard EPA performance could not be used as a the second approach, which we are not
proposed, the compliance date for basis for the MACT floor for new adopting, site specific risk analyses
sources which commenced construction sources. Alkaline scrubbers were then would be used to establish standards on
after December 2, 2005, and before considered for beyond-the-floor a case-by case basis.
promulgation of this final rule is three controls. Using engineering assessments After careful consideration of the
years from December 20, 2006. See from similar technology operated on comments on the proposed
section 112(i)(2). New sources that municipal waste combustors and amendments, we are not requiring
commence construction after the date of medical waste incinerators, we control of HCl emissions from cement
promulgation of today’s action must estimated costs and emissions kilns under section 112(d). Under the
comply with the final rule upon start- reductions. Based on the costs of control authority of section 112(d)(4) of the
up. However, as we are reconsidering and emissions reductions that would be CAA, we have determined that no
the new source mercury standard and achieved, we determined that beyond- further control is necessary because HCl
plan to take final action on that the-floor controls were not warranted is a ‘‘health threshold pollutant,’’ and
reconsideration in no more than a year (63 FR 14203, March 24, 1998). human health is protected with an
and as construction of a new kiln In the proposed amendments, we ample margin of safety at current HCl
generally takes at least 20–24 months, it reexamined establishing a floor for emission levels. The following explains
is unlikely that any new source will be control of HCl emissions from new the statutory basis for considering
subject to the standard before Portland cement sources. Since health thresholds when establishing
completion of reconsideration. promulgation of the NESHAP, wet standards and the basis for today’s
We are also requiring that new scrubbers have been installed and are decision, including a discussion of the
sources demonstrate compliance by operating at a minimum of five Portland risk assessment conducted to support
doing mercury emission testing with the cement plants. See section IV.A.1.d the decision.
raw mill off and with the raw mill on. above. For the reasons described above, Section 112 of the CAA includes
The reason to test under both conditions this is an insufficient number of exceptions to the general statutory
is that (as explained in section A.1.c scrubbers on which to base an existing requirement to establish emission
above) one other operation factor source floor for this category (id.). We standards based on MACT. Of relevance
besides wet scrubber performance did, however, propose to base the floor here, section 112(d)(4) allows us to
affecting emissions is the recycling of for new sources on the performance of develop risk-based standards for HAP
CKD. A facility could cut off CKD continuously operated alkaline ‘‘for which a health threshold has been
recycling for purposes of meeting the scrubbers, and proposed emissions established’’ provided that the standards
emission limit during testing with raw levels of 15 ppmv at the control device achieve an ‘‘ample margin of safety.’’
mill off, and then start recycling after outlet, or a 90 percent HCl emissions Therefore, we believe we have the
the test which could result in the reduction measured across the scrubber, discretion under section 112(d)(4) to
emissions limit being exceeded. We as the new source floor. develop standards which may be less
could simply limit CKD recycling to the We also reexamined the MACT floor stringent than the corresponding
level during the raw mill off test, but we for existing sources. The only potential technology-based MACT standards for
believe this would potentially and controls identified as a floor option was threshold hazardous air pollutants
needlessly restrict the ability of a the operation of the kiln and PM control emitted by some source categories. See
facility to recycle CKD during raw mill device themselves. Because the kiln and 67 FR 78054, December 20, 2002 and 63
on operation. During the test under each PM control system contain large FR 18765, April 15, 1998.
condition, the facility must record the amounts of alkaline CKD, the kilns In evaluating potential standards for
amount of CKD recycle. The amount of themselves remove a significant amount HCl for this source category, we seek to
CKD recycle becomes an operating limit of HCl (which reacts with the CKD and assure that emissions from every source
not to be exceeded. is captured as particulate). See 70 FR in the category result in exposures not
The limit for new sources adopted 72337 and 69 FR 21259 (April 20, 2004). causing adverse effects, with an ample
here also applies to both area and major We proposed as a floor the operation of margin of safety, even for an individual
new sources. We have applied this limit the kiln and PM control as a work exposed at the upper end of the
to area sources consistent with section practice standard. exposure distribution. The upper end of
112(c)(6). We also evaluated requiring the use of the exposure distribution is calculated
For facilities that elect to meet an alkaline scrubber as a beyond-the- using the ‘‘high end exposure estimate,’’
mercury emissions limits using ACI, we floor control option for existing sources. defined as a plausible estimate of
are incorporating the operating and We found that the costs and non-air individual exposure for those persons at
monitoring requirements for ACI that quality health and environmental the upper end of the exposure
are applicable when ACI is used for impacts were not reasonable for the distribution, conceptually above the
dioxin control. emissions reductions achieved. 90th percentile, but not higher than the
We also solicited comment on individual in the population who has
B. Determination of MACT for HCl adopting alternative risk-based emission the highest exposure. We believe that
Emissions standards for HCl pursuant to section assuring protection to persons at the
In developing the 1999 Portland 112(d)(4) of the CAA (70 FR 72337). We upper end of the exposure distribution
Cement NESHAP we concluded that no suggested two possible approaches for is consistent with the ‘‘ample margin of
add-on air pollution controls were being establishing such standards. Under the safety’’ requirement in section 112(d)(4).
used whose performance could be used first approach an alternative risk-based Our decision not to develop standards
as a basis for the MACT floor for standard would be based on national for HCl from cement kilns is based on
existing Portland cement plants. For exposure standards determined by EPA the following. First, we consider HCl to
new source MACT, we identified two to ensure protection of public health be a threshold pollutant. See 63 FR
kilns that were using alkaline scrubbers with an ample margin of safety, and to 18767, 67 FR 78054, and 70 FR 59407,
for the control of SO2 emissions. But we be protective of the environment. For October 12, 2005. Second, we have
concluded that because these devices reasons discussed below we have defined threshold values for HCl in the
form of an Inhalation Reference cement companies and conservative exposure is not expected. Impacts at
Concentration (RfC) and acute exposure assumptions regarding (among other potential residential locations would be
guideline level (AEGL). Third, HCl is factors) HCl stack concentrations, expected to be significantly below those
emitted from cement kilns in quantities operating conditions, receptor locations, presented in the analysis.
that result in human exposure in the and dispersion characteristics. The kilns The PCA study generated estimates of
ambient air at levels well below these for which data were provided cover a chronic (annual average) concentrations
threshold values with an ample margin full range of kiln types, operating for comparison to the relevant health
of safety. Finally, there are no adverse conditions, and stack parameters. The reference values or threshold levels.
environmental effects associated with three-tiered modeling approach consists Chronic exposures were compared to
HCl emissions from cement kilns. The of: the RfC of 20 µg/m3 for long-term
bases and supporting rationale for these • Tier 1—Lookup tables. continuous exposure.
conclusions are as follows. • Tier 2—Screening dispersion Noncancer risk assessments typically
For the purposes of section 112(d)(4), modeling. use a metric called the Hazard Quotient
several factors are considered in our • Tier 3—Detailed dispersion (HQ) to assess risks of exposures to
decision on whether a pollutant should modeling. noncarcinogens. The HQ is the ratio of
be categorized as a health threshold The concentration estimates from exposure (or modeled concentration) to
pollutant. These factors include each modeling tier should be more the health reference value or threshold
evidence and classification of accurate and less conservative than the level (i.e., RfC or REL). HQ values less
carcinogenic risk and evidence of previous one. As a result, the level of than 1 indicate that exposures are below
noncarcinogenic effects. For a detailed complexity of the modeling and data the health reference value or threshold
discussion of factors that we consider in input information required for each tier level and are likely to be without
deciding whether a pollutant should be is greater than for the previous tier. If a appreciable risk of adverse effects in the
categorized as a health threshold plant showed emissions below the exposed population. HQ values above
pollutant, please see the April 15, 1998, threshold concentration in any tier, that 1.0 do not necessarily imply that
Federal Register document (63 FR plant was not included in the next tier adverse effects will occur, but that the
18766). In the April 15, 1998, action of modeling. potential for risk of such effects
cited above, we determined that HCl, a In order to evaluate potential health increases as HQ values exceed 1.0.
Group D pollutant, is a health threshold impacts it is necessary to establish long For the PCA assessment, if the HQ
pollutant for the purpose of section term concentration thresholds. The RfC was found to be less than one for any
112(d)(4) of the CAA (63 FR 18753). is a long-term threshold, defined as an of the tiers using conservative defaults
The Portland Cement Association estimate of a daily inhalation exposure and modeling assumptions, the analysis
(PCA) conducted a risk assessment to that, over a lifetime, would not likely concluded with that tier. On the other
determine whether the emissions of HCl result in the occurrence of significant hand, if the HQ exceeded one, analysis
from cement kilns at the current noncancer health effects in humans. We proceeded to subsequent tiers.
baseline levels resulted in exposures have determined that the RfC for HCl of The Tier 1 modeling resulted in an
below the threshold values for HCl. We 20 micrograms per cubic meter (µg/m3) HQ above 1 for most facilities.
reviewed the risk assessment report is an appropriate threshold value for Therefore, a Tier 2 analysis was
prepared by the PCA and believe that it assessing risk to humans associated required. In the Tier 2 analysis, all
uses a reasonable and conservative with exposure to HCl through inhalation facilities except for five showed an HQ
methodology, is consistent with EPA (63 FR 18766, April 15, 1998). below 1.
methodology and practice, and reaches Therefore, the PCA used this RfC as the For the five facilities with an HQ
a reasonable conclusion that current threshold value in their exposure above 1, additional data were obtained
levels of HCl emissions from cement assessment for HCl emitted from cement on the actual HCl and stack moisture
kilns would be well under the threshold kilns. concentrations at these facilities and the
level of concern even for assumed The general approach was that actual Tier 2 modeling analysis was rerun. The
worst-case human receptors. release characteristics were used for refined Tier 2 analysis resulted in HQ
The PCA analysis evaluated long-term stack height, stack diameter, exit values of 0.30 or less for all five
and short-term ambient air temperature, and exit velocity, based on facilities.
concentrations resulting from emissions information provided by the individual Thus, we have evaluated and are
of HCl from Portland cement kilns in facilities modeled by the PCA. The comfortable with PCA’s calculations
order to quantify potential non-cancer analyses performed under each tier and feel confident that exposures to HCl
risks associated with such emissions, as assumed worst case operating scenarios, emissions from the facilities in question
well as to characterize potential such as maximum production rate and are unlikely to ever exceed an HQ of 1.0.
ecological effects of those emissions. 24 hours per day, 365 days per year Therefore, we believe that the predicted
The approach is based on the USEPA operation, and that all kilns were exposures from these facilities should
guidance document entitled ‘‘A Tiered located 10 meters from the property still be protective of human health with
Modeling Approach for Assessing the boundary line. HC1 emission rates were an ample margin of safety. Put another
Risks Due to Sources of Hazardous Air assumed to be 130 ppmv for all kiln way, total exposures for nearby
Pollutants’’ (USEPA 1992) (Tiered types. This is an extremely conservative residents would not exceed the short-
Modeling Approach) and is consistent number. Hydrogen chloride emission term or long-term health based
with EPA risk characterization guidance rates are below 10 ppmv at most threshold levels or health reference
‘‘Air Toxics Risk Assessment Reference facilities, and the highest value for values. Similarly, based on the PCA
Library—Volume 2—Facility-Specific which we have data is below 45 ppmv. analysis we believe that the acute
Assessment’’ (USEPA, 2004). The PCA In the Tier 2 analyses, worse case exposure to HC1 for these facilities
conducted dispersion modeling for 67 metrological conditions were assumed. would not exceed the short-term,
cement plants and 112 cement kilns, Further, it is important to note that health-based threshold level.
representing about two-thirds of all these predicted impacts are located The standards for emissions must also
operating cement plants in the U.S., adjacent to facility property lines, many protect against significant and
using stack parameter data provided by times in locations where chronic widespread adverse environmental
effects to wildlife, aquatic life, and other HCl will have much less effect than the account the adverse energy and
natural resources. The PCA did not deposited hydronium ions. environmental impacts, we determined
conduct a formal ec0ological risk In conclusion, acute and chronic that the precalciner/no preheater design
assessment. However, we have reviewed exposures to expected HCl did not represent MACT (63 FR 14202,
publications in the literature to concentrations around cement kilns are March 24, 1998). We also considered
determine if there would be reasonable not expected to result in adverse feed material selection for existing
expectation for serious or widespread environmental toxicity effects. HC1 is sources as a MACT floor technology and
adverse effects to natural resources. not persistent in the environment. concluded that this option is not
We consider the following aspects of Effects of HCl on ponds and soils are available to existing kilns, or to new
pollutant exposure and effects: toxicity likely to be local rather than kilns located at existing plants because
effects from acute and chronic widespread. Finally, HCl is not believed these facilities generally rely on existing
exposures to expected concentrations to result in biomagnification or raw material sources located close to the
around the source (as measured or bioaccumulation in the environment. source due to the cost of transporting
modeled), persistence in the Therefore, we do not anticipate any the required large quantities of feed
environment, local and long-range adverse ecological effects from HCl. materials. However, for new greenfield
transport, and tendency for The results of the exposure kilns, feed material selection as
biomagnification with toxic effects assessment showed that exposure levels achieved through appropriate site
manifest at higher trophic levels. to baseline HCl emissions from cement selection and feed material blending is
No research has been identified for production facilities are well below the demonstrated and is the basis for new
effects on terrestrial animal species health threshold value. Additionally, source MACT (63 FR 14202, March 24,
beyond that cited in the development of the threshold values, for which the RfC 1998).
the HCl RfC. Modeling calculations and AEGL values were determined to be In our proposed amendments we
indicate that there is little likelihood of appropriate values, were not exceeded reexamined MACT for THC for both
chronic or widespread exposure to HCl when considering conservative new and existing facilities. We proposed
at concentrations above the threshold estimates of exposure resulting from to adopt the same standards for Portland
around cement manufacturing plants. cement kiln emissions as well as cement kilns as are applicable to kilns
Based on these considerations, we considering background exposures to that fire hazardous waste (40 CFR
believe that the RfC can reasonably be HCl and therefore, represent an ample 63.1220(a)(5)). Those standards are
expected to protect against widespread margin of safety. Furthermore, no based on using good combustion
significant or widespread adverse conditions to destroy hazardous air
adverse effects in other animal species
environmental effects from HCl are pollutants in fuels. Our rationale for
as well.
anticipated. Therefore, under authority proposing to adopt these standards was
Plants also respond to airborne HCl
of section 112(d)(4), we have that the THC and carbon monoxide (CO)
levels. Chronic exposure to about 600
determined that further control of HCl standards guarantee that the kiln will
µg/m3 can be expected to result in operate under good combustion
discernible effects, depending on the emissions from new or existing cement
manufacturing plants under section conditions and will minimize formation
plant species. Further, in various (and hence, emissions) of non-dioxin
species given acute, 20 minute 112(d) is not necessary.
organic HAP from fuel combustion. We
exposures of 6,500 µg/m3, field studies C. Determination of MACT for THC believed that the control of THC
report different sensitivity to damage of Emissions emissions from cement kilns which do
foliage. The maximum modeled long- not fire hazardous waste should be no
term HCl concentration (less than 100 1. Floor Determinations
more difficult to control than emissions
µg/m3) is well below the 600 µg/m3 THC serve as a surrogate for non- for kilns that do fire hazardous waste
chronic threshold, and the maximum dioxin organic HAP emissions for this because GCP are maintainable by either
short-term HCl concentration (less than source category. During the type of kiln, and the hazardous waste
1600 µg/m3) is well below the 6,500 µg/ development of the 1999 Portland cement kilns would be the more
m3 acute exposure threshold. Therefore, Cement NESHAP, EPA identified no challenged in that regard. Because we
no adverse exposure effects on plant add-on air pollution control technology had no data upon which to set a
species are anticipated. being used in the Portland cement different standard, and because we
HCl is not considered to be a strongly industry whose performance could be believed these levels were indicative of
persistent pollutant or one where long used as a basis for establishing a MACT good combustion in any case, the
range transport is important in floor for controlling THC emissions adoption of the standards for cement
predicting its ecological effects. In the from existing sources. EPA did identify kilns firing hazardous waste was
atmosphere, HCl can be expected to be two kilns using a system consisting of deemed appropriate.
absorbed into aqueous aerosols, due to a precalciner (with no preheater), which We continue to believe that good fuel
its great affinity for water, and removed essentially acts as an afterburner to combustion conditions are indicative of
from the troposphere by rainfall. Toxic combust organic material in the feed. the performance of the median of the
effects of HCl to aquatic organisms The precalciner/no preheater system best performing 12 percent of existing
would likely be due to the hydronium was considered a possible basis for a sources for controlling non-dioxin
ion, or acidity. Aquatic organisms in beyond-the-floor standard for existing organic HAP. However, based on
their natural environments often exhibit kilns and as a possible basis for a MACT comments received on the proposed
a broad range of pH tolerance. Effects of floor for new kilns. However, this amendments, and additional emission
HCl deposition to small water bodies system was found to increase fuel data analysis, we believe our proposed
and to soils will primarily depend on consumption relative to a preheater/ quantified method of monitoring good
the extent of neutralizing by carbonates precalciner design, to emit six times as fuel combustion, i.e. setting specific
or other buffering compounds. Chloride much SO2, two and one half times as THC or CO levels, was flawed.
ions are essentially ubiquitous in much NOX, and 1.2 times as much CO2 Industry commenters had noted that
natural waters and soils so minor as a preheater/precalciner kiln of the majority of the THC emissions from
increases due to deposition of dissolved equivalent clinker capacity. Taking into a cement kiln main stack result from the
introduction of feed materials into the demonstrating that good fuel term performance capabilities of this
cold end of the kiln. These emissions combustion conditions exist. control device considering variable
are essentially a function of the organic In the final amendments, we are organic levels in raw materials and
content of the raw materials, and cannot requiring that existing kilns and in-line other process variabilities.
be controlled using GCP, which is the kilns/raw mills must implement GCP We are retaining the proposed THC
basis of our MACT floor. At proposal we designed to minimize THC from fuel emission limit of 20 ppmv measured at
agreed with this assessment (and combustion. GCP include training all the main kiln stack as the MACT floor
continue to agree with it), but believed operators and supervisors to operate and for all new or reconstructed kilns and
that the fact that cement kilns that burn maintain the kiln, calciner, and inline raw mill/kilns. An alternative to
hazardous waste can meet these pollution control systems in accordance the 20 ppmv floor level is that a facility
standards indicated that the proposed with good engineering practices. The may demonstrate a 98 percent reduction
level could be met by all cement kilns training shall include operating the kiln, in THC emissions from uncontrolled
under good combustion conditions, calciner, and pollution control system levels—the level of emission reduction
even considering the fact that good in a manner to minimize excess required by permit for the best
combustion cannot control THC or CO emissions. performing source in the category. We
emissions emanating from organic We have also reexamined the have determined in other rules that a 20
materials in the feed. We also believed proposed MACT floor for new sources. ppmv outlet emissions level or 98
that by allowing a facility to monitor CO There are currently two cement kilns percent destruction efficiency represent
as a surrogate for THC, we had provided with add-on controls which reduce the long term performance of an RTO
sufficient flexibility to account for emissions of THC. At one facility, under the varying conditions typically
variations in feed material organic activated carbon is injected into the flue encountered in industrial applications.
content. gas and collected in the PM control See Thermal Incinerators and Flares in
We have reevaluated these device. The carbon adsorbs some of the Docket EPA–HQ–OAR–2002–0051. As
assumptions. First, we obtained THC. The collected carbon is then noted above, the one cement facility
additional THC emission data from reinjected into the kiln in a location that with an RTO operating full-time has
several facilities. These data ensures destruction of the collected actual and permitted emission levels
demonstrate that there are certain THC. However, the THC emissions from which are below 20 ppmv. However, the
cement facilities where THC emissions, this facility are the highest for any performance guarantee at this facility is
with no indication of poor fuel facility for which we have data. based on the combined emissions of CO
combustion practices, exceed 20 ppmv. Therefore, we do not consider this to and THC. Therefore, all new facilities
The data also indicate that achieving the represent the best controlled source. could meet the permitted emission
100 ppmv CO level, even for cement This same facility also has an alternative levels of the one facility that has an RTO
kilns with low organic content feed and control scheme for THC of a limestone only if they all have the same levels of
good fuel combustion conditions, is not scrubber followed by a regenerative CO in the exhaust gas. We have no data
possible without use of a control device. thermal oxidizer (RTO). However, these to support that all new kilns will have
See Lehigh CO and THC data in docket control devices have not operated sufficient CO in the exhaust streams to
EPA-HQ-OAR–2002–0051. Moreover, continuously due to significant guarantee that they can meet the same
the analogy with hazardous waste- operation problems caused by the site level of performance as the one facility
burning cement kilns breaks down. If a specific constituents in the flue gas. (See noted above, or, conversely that this one
cement kiln that fires hazardous waste e-mail from Michael D. Maillard, facility would continue to meet the
cannot meet the THC or CO limits in the Michigan Department of Environmental same THC levels if CO levels in its
Hazardous Waste Combustor (HWC) Quality in docket EPA–HQ–OAQ–2002– exhaust gas differed. We thus believe
NESHAP due to organic materials in 0051.) Because these controls have not long term performance for THC alone is
their feed, they can (and have) simply been demonstrated to have the ability to better characterized based on the well-
stopped firing hazardous waste. This operate continuously, we cannot established data documenting
can either be done permanently, or consider them as the basis for a new performance of RTO for THC. Moreover,
temporarily anytime the kiln operator source MACT floor (or an emission the percent reduction achievable by an
notes that THC or CO emissions are standard, for that matter). RTO is dependent on the inlet
approaching the emission limits. This A second facility also has a limestone concentration of organics. See Thermal
option is not available to cement kilns scrubber followed by an RTO. The Incinerators and Flares in Docket EPA–
that do not fire hazardous waste; they scrubber is necessary to prevent fouling, HQ–OAR–2002–0051. Thus, we believe
cannot stop making cement without plugging, and corrosion of the RTO. In that a limit based on the demonstrated
ceasing business altogether. This would this case the scrubber/RTO operates performance of RTO under a variety of
mean that facilities with higher levels of continuously and efficiently. This circumstances is the best measure of the
organic materials in the raw materials facility has been tested and showed long term performance of this device
would be forced to adopt some type of VOC (essentially the same as THC) under the circumstances likely to be
add-on control to meet the emissions emission levels of 4 ppmv (at 7 percent encountered by new cement kilns,
limits. As we have previously stated, we oxygen), and currently has a permit especially varying levels of organics in
believe this would result in the limit for VOC of approximately 9 ppmv. the feed.
imposition of a beyond-the-floor The RTO has a guaranteed destruction
efficiency of 98 percent of the combined 2. Beyond-the-Floor Determinations
standard without the mandated
consideration of costs and other emissions of CO and THC. Based on this In the December 2005, proposed
impacts. See 70 FR 72335. information we believe this facility is amendments we considered beyond-the-
As a result, although we adhere to our the best controlled source, and that the floor options for existing sources of
approach at proposal that the MACT performance of a limestone scrubber substituting raw materials with lower
floor for control of non-dioxin organic followed by an RTO is the basis for new organic contents, but we determined
HAP at existing sources is operating source MACT floor for non-dioxin this beyond-the-floor option was not
under good combustion conditions, we organic HAP, measured as THC. We feasible (70 FR 72340). We also
are adopting a different means of explain below how we assess the long- considered a beyond-the-floor THC
standard of 20 ppmv based on the use date for this standard is one year from we see no reason that the floor level of
of the scrubber/RTO control system. December 20, 2006. Because all facilities control should not apply in the case
Based on the available data, we estimate already have some type of training where there is a separate raw material
that approximately 75 percent of program, we believe one year is dryer. We note that in the original
existing kilns could meet a 20 ppmv sufficient to comply with this NESHAP, the 50 ppmv standard also
standard without the addition of requirement. See section 112(1)(3) applied to raw material dryers.
controls. For an existing preheater/ (compliance dates for MACT standards We are adopting our proposed
precalciner kiln that could not meet a 20 ‘‘shall provide for compliance as requirement that compliance for a THC
ppmv standard without controls, the expeditiously as practicable’’). standard will be demonstrated using a
capital cost would be approximately The standard for new sources is to CEM and a 1-hour averaging period. See
$10.7 million and the total annualized meet a THC standard of either 20 ppmv 70 FR 72340. The previous 50 ppmv
cost would be approximately or a 98 percent reduction in THC standard for new greenfield sources was
$3.9 million. The cost per ton of THC emissions from uncontrolled levels. based on a monthly average. We believe
reduction would be in the area of However, as explained above, a monthly average was appropriate for
$20,000, assuming an inlet performance of a back-end control that standard (and are retaining monthly
concentration of about 63 ppmv. We device (i.e. the RTO, preceded by an averaging for kilns subject to that
estimate that approximately 5 percent of enabling scrubber) was not the basis of standard) because the standard’s basis is
the THC is actually organic HAP. the proposed new source standard. selection of raw materials. There can be
Therefore, the cost of organic HAP Information that one kiln utilizes an significant short term variations in raw
reduction would be $398,000 per ton. In RTO, as well as information regarding materials, even if a facility can meet the
addition, the energy use for one large the technical capabilities of RTO, standard in the long term. In the case of
kiln to operate an RTO would be emerged following the public comment these final amendments the required
approximately 99.7 billion British period and therefore has not previously level of performance is based on an
thermal units per year, a very high been available for public comment. To emissions control technology.
energy consumption rate. The wet afford opportunity for comment, EPA is Therefore, we do not anticipate the
scrubber required upstream of the RTO itself immediately granting same type of short term variability that
would also result in 40 million gallons reconsideration of the new source existed with the previous 50 ppmv
per year of additional water usage and standard for THC in a notice published standard.
create 45,500 tpy of solid waste (from elsewhere in today’s Federal Register. Because the final standard is more
dewatered scrubber sludge). Based on The original Portland Cement stringent than the standard EPA
the costs, significant adverse energy NESHAP contains a 50 ppmv THC proposed, the compliance date for
impacts, and adverse non-air quality emissions limit for new greenfield kilns, sources which commenced construction
health and environmental impacts, we kilns/inline raw mills, and raw after December 2, 2005, and before
do not believe a beyond-the-floor materials dryers. There are no situations promulgation of this final rule is 3 years
standard is justified. we can identify where a 50 ppmv limit from December 20, 2006. See section
We also examined a beyond-the-floor would be more stringent than a 98 112(i)(2). We consider the final standard
regulatory option based on the use of percent reduction limit. Since this 50 to be more stringent than the proposed
ACI for THC control. The total annual ppm limit is less stringent than the new standard because it is based on the
cost for this option would be $470,000 source standard we are adopting in this performance of a control device
to $600,000 for an existing preheater/ rule reflecting performance of an RTO, (notwithstanding that the numeric limit
preclaciner kiln. The cost per ton of it is obviously not appropriate to retain is the same as proposed), and now
THC reduction would be in the area of it. We are thus finding that the floor for controls both THC emissions from fuel
$5,000, assuming an inlet concentration greenfield new sources (and all other combustion and THC emissions
of about 63 ppmv. We estimate that new sources under this rule) is 20 ppm/ resulting from the organic materials in
approximately 5 percent of the THC is 98 percent THC, with one exception. the kiln feed, and is more likely to result
actually organic HAP. Therefore, the This new source limit will, at least for in significant costs and changes in
cost of organic HAP reduction would be some new facilities, require the operation than the proposed standard.
$100,000 per ton. In addition, this application of a back end control. For For new sources that elect to meet
control option would generate this reason, we do not believe this limit THC emissions limits using ACI, we are
approximately 850 tpy of solid waste. should be applied retroactively to incorporating the operating and
Based on the high costs, energy impacts, sources constructed prior to December monitoring requirements for ACI that
and adverse non-air quality health and 2, 2005, the date of proposal for the are applicable when ACI is used for
environmental impacts, we do not amendment. See the response to dioxin control.
believe a beyond-the-floor standard is comment concerning new sources in
section VI for our rationale for this D. Evaluation of a Beyond-the-Floor
justified. Control Option for Non-Volatile HAP
We did not examine a beyond-the- decision. So for sources constructed
prior to December 2, 2005, we are not Metal Emissions
floor regulatory option for new sources
because there are no controls that amending the 50 ppmv THC limit. In our MACT determination for PM
would, on average, generate a greater Consistent with section 112(c)(6) we (the surrogate for non-volatile HAP
THC reduction than a combination of a are applying the 20 ppmv/98 percent metals), we concluded that well-
wet scrubber/RTO. Thus, the floor level reduction limit to both major and area designed and properly operated FF or
is also new source MACT. new sources. We are also applying the ESP designed to meet the new source
limit to raw materials dryers. We performance standards (NSPS) for
3. Conclusion anticipate that all new kilns will be Portland cement plants represent the
In sum, we conclude that the preheater/precalciner kilns with an MACT floor technology for control of
standards for THC for all existing inline raw mill (i.e. there will be no PM from kilns and in-line kiln/raw
cement kilns is implementing GCP separate dryer exhaust). This is the mills. Because no technologies were
designed to minimize THC emissions design of the kilns that form the basis identified for existing or new kilns that
from fuel combustion. The compliance of new source MACT for THC. However, would consistently achieve lower
emissions than the NSPS, EPA We proposed two resolutions to this Response: Although we disagree with
concluded that there was no beyond- issue. They were: the premise of this comment, the
the-floor technology for PM emissions (1) Changing the wording of 40 CFR comment is moot because we are setting
(63 FR 14199, March 24, 1998). 63.1340(c) to make it clear that all raw standards for all HAP which was
In National Lime Association v. EPA, materials storage and handling is addressed by the court’s mandate. We
the court held that EPA had failed to covered by the NESHAP, but that agree that the court stated the absence
adequately document that substituting crushers (regardless of their location) of technology-based pollution control
natural gas for coal was an infeasible are not. devices for HCl, mercury, and THC did
control option, and also that EPA had (2) Including crushers as an affected not excuse EPA from setting emission
not assessed non-air quality health and source in the Portland Cement NESHAP standards for those pollutants. In
environmental impacts when and incorporating the current response to the court’s opinion, we have
considering beyond-the-floor standards requirements applicable to crushers evaluated all possibilities of setting
for HAP metals (233 F. 3d at 634–35). contained in 40 CFR part 60, subpart standards, including technology based
As a result, the court remanded the OOO (and correspondingly, exempting control, fuel and raw materials changes,
beyond-the-floor determination for HAP crushers covered by the Portland and process modifications. We believe
metals for further consideration by EPA. Cement NESHAP from 40 CFR part 60, this evaluation is what the court
We presented our reexamination of a subpart OOO). intended. See 70 FR 72335.
beyond-the-floor MACT control Comment: Regarding EPA’s rejection
We received several comments from
standard for HAP metals in the of beyond-the-floor standards for each
State and local agencies supporting our
preamble to the proposed amendments, HAP, one commenter states that EPA’s
contention that the intent of the rule
addressing the remand by showing that reasoning is both unrelated to the
language at issue was to exclude
substitution of fuel or feed materials are relevant statutory mandate and arbitrary
crushers, and that our interpretation of
either technically infeasible or cost and capricious, as well as completely
the rule language was correct. We
prohibitive and therefore that a beyond- ignoring currently available control
the-floor standard for HAP metals is not considered simply deleting the
measures of which EPA is aware and
reasonable. (See 70 FR 72340–72341). (potentially) confusing language and
which would result in reductions of
We also indicated that non-air health adding clarifying language that a
emissions of mercury, HCl, THC and
and environmental impacts would be crusher located after raw materials
other HAP.
minimal, as would energy use storage would be covered by this Response: Where we have rejected
implications (id. at 72341). We received subpart. However, we have not been beyond-the-floor standards we have
no data in the comments on the able to identify any facilities where the evaluated all available control methods
proposed amendments that have altered crusher is located after raw materials that have been demonstrated for this
our previous analysis. Therefore, we are storage. In addition, we do not have data source category. We also evaluated
not including a beyond-the-floor PM to determine the impacts of adding control technologies that have not been
standard in these final amendments. coverage of this piece of equipment to demonstrated, but that we have reason
this subpart. For that reason, we are to believe may be effective (such as
V. Other Rule Changes modifying the language in § 63.1340(c) ACI). With one exception, which is
On April 5, 2002, we amended the to state that crushers are not covered by banning the use of fly ash with elevated
introductory text of 40 CFR 63.1353(a) this subpart regardless of their location. mercury contents that result from
to make it more clear that affected There are currently no regulations that sorbent injection where such a practice
sources under the Portland Cement regulate existing crushers in this would increase mercury emissions, in
NESHAP were not subject to 40 CFR application. New crushers would no case did we find that a beyond-the-
part 60, subpart F (67 FR 16615, April potentially be subject to the floor standard was justified
20, 2002). In making this change, we requirements of 40 CFR 60, subpart (‘‘achievable’’ in the language of section
inadvertently deleted paragraphs (a)(1) OOO. 112(d)(2)) taking into consideration
and (2) of 40 CFR 63.1353. The language VI. Responses to Major Comments costs, energy, and non-air quality health
in these paragraphs is still necessary for and environmental impacts.
determining the applicability of 40 CFR This section presents a summary of Comment: According to one
part 60, subpart F. We proposed to responses to major comments. A commenter, EPA’s refusal to set mercury
reinstate these paragraphs as originally summary of the comments received and standards demonstrates contempt of
written in the final rule. We received no our responses to those comments may court. The commenter states that EPA’s
comments on this issue and are be found in Docket ID No. EPA–HQ– reconsideration of MACT for mercury
therefore reinstating the two paragraphs OAR–2002–0051. did not satisfy the court’s directive to
as proposed. Comment: According to several establish emissions standards and not
In the proposed amendments we commenters, EPA’s proposal did not just reconsider the issue.
requested comment on amending satisfy the mandate issued by the DC Citing the CAA’s requirements to set
language published on April 5, 2002, Circuit Court of Appeals. On EPA’s emission standards for each HAP listed
whose purpose was to clarify that analysis of MACT for mercury, HCl, and in 112(b) and, as directed in 112, for
crushers were not subject to this THC; EPA’s beyond-the-floor analysis; each category of sources for the HAP
NESHAP. The PCA believed that there and the risk-based exemptions from HCl applying the maximum achievable
had been misinterpretation of the standards, one commenter states they degree of reduction, the commenter
amended rule text. However, we are unlawful, arbitrary, capricious, and states that EPA’s decision to not set
explained in the proposed amendments irrelevant. These commenters state that mercury emission standards is
that we believe the PCA interpretation the court was clear in its directive to unlawful.
is not reasonable when reading the EPA that the absence of technology- Response: EPA strongly disagrees
entire final NESHAP. However, we based pollution control devices for HCl, with the commenter’s characterization
agreed that the rule language as written mercury, and THC did not excuse EPA of the proposed standards in the
is conceivably open to more than one from setting emission standards for proposed rule. EPA issued the proposed
interpretation. See 70 FR 72341. those pollutants. rule consistent with the court’s
instructions in the remand. In response EPA’s decision not to set mercury be doing so deliberately to reduce
to comments received, however, EPA standards; and fourth the commenter mercury emissions. According to the
has modified the proposal and adopted claims EPA mistakenly cites the Copper commenter, whether the sources are
specific standards for each HAP covered Smelters (Sierra Club) and PVC MACT achieving low mercury emissions levels
by the court’s mandate. Thus, this cases (Mossville) as justification for its through deliberate measures or
comment is moot, even accepting the approach. According to the commenter, coincidentally are statutorily irrelevant.
commenter’s premise (which EPA does these cases pertain to beyond-the-floor Response: We disagree with all the
not), since EPA is establishing standards standards and do not apply to floor points raised in the comment above and
(in the sense the commenter uses the standards, which require EPA to set preceding comments that EPA’s
term) for each HAP covered by the floors at emission levels that the best arguments for not setting mercury
court’s mandate. Moreover, as explained sources achieved, regardless of what standards are without merit. As noted
in other parts of this preamble, EPA has EPA thinks is achievable. above, we believe we have met the
carefully analyzed many different Response: The commenter’s reading court’s directive by evaluating all
possibilities for setting standards for the of Mossville is not correct. The case available methods of mercury control,
HAP covered by the remand, examining involved a floor standard. See 370 F. 3d including changes to fuels, raw
not only technology-based back end at 1240–42. We explained at proposal materials, and process controls. We do
controls but control of inputs to cement why we believe the discussion of raw not agree that the court directed us to
kilns as well. We believe that our action materials in Sierra Club is also set standards regardless of the facts, nor
fully satisfies both the letter and spirit applicable to a floor determination. See do we agree that section 112(d)(3) of the
of the court’s mandate. 70 FR at 72335 n. 4. CAA requires us to set floor standards
Comment: The commenter above Comment: The commenter further that cannot be met without requiring
states that EPA’s arguments for not states that EPA’s argument that its even the best performing facilities to
setting mercury standards are without emissions data do not reflect apply beyond-the-floor controls—
merit and provide several justifications performance over time, merely relates to controls not used by any sources in the
for its view. First the commenter states the sufficiency of EPA’s data. The source category, even those which are
that EPA’s arguments for not setting commenter states that EPA is required ostensibly the best performing (i.e. the
mercury standard are irrelevant because to develop an approach to setting a floor lowest emitters in individual
EPA has a clear statutory obligation to standard, including collecting more performance tests).
set mercury standard and any reason for emissions data if needed. The commenter correctly noted that
not doing so must be invalid. Response: Floor standards are to we are required to set standards based
Response: This comment is now reflect the performance of sources ‘‘for on facilities for which the administrator
moot, as just explained. which the Administrator has emissions has emissions information. However, as
Comment: According to the same information’’ (section 112(d)(3)), which explained previously in the notice, the
commenter, EPA’s view as to what is provision does not create an obligation emissions levels in the data available to
achievable cannot replace the CAA to gather a specified amount of the administrator are mainly influenced
requirement to set MACT floors information. Moreover, not only must by factors that are beyond the control of
reflecting what the best performing MACT standards, including standards the facilities tested, and the test results
sources are achieving. The commenter reflecting the MACT floor, reflect can neither be replicated by the
states that the CAA mandates a floor performance variability but EPA may individual facilities nor duplicated by
reflecting the average emission reasonably estimate what that variability other facilities. In addition, these are
limitation achieved by the best can be, and is not limited to stack short term data that we believe are not
performing 12 percent of the existing emissions measured in single indicative of the sources’ long term
sources (for which the Administrator performance tests as the commenter emissions. The commenter states that
has emissions information) and not apparently believes. See Mossville, 370 we should get better data. However,
what EPA believes would be achievable. F. 3d at 1242 (setting standard at a level they do not indicate how we would be
The commenter states that the court slightly higher than the highest data able to perform this task given the fact
expressly required EPA to set emission point experienced by a best performing that there are no long term data
standards based on what the best source ‘‘reasonably estimates the available for mercury emissions from
performers are actually achieving and performance of the best * * * cement kilns: We know of no case
not what EPA thinks is achievable. performing sources’’). Most basically, where any cement facility has applied
Response: As Mossville and earlier because MACT standards must be met at mercury continuous emission
cases make clear, because MACT all times, a standard must reflect monitoring (CEM) technology, or
standards (based on floors or otherwise) performance variability that occurs at all gathered any long term emissions data
must be met at all times, the standards times, and this variability is simply not we could use to set a national standard.
must reflect maximum possible accounted for in single stack test results (We do note, however, that we are
variability (assuming proper design and for mercury from a cement kiln. ourselves granting reconsideration of
operation of the various control Comment: The same commenter the new source standard for mercury, in
mechanisms). See discussion at 70 FR disagrees with EPA’s position that part to initiate field testing of cement
72335 and 70 FR 59436. setting the floor at emission levels kilns equipped with wet scrubbers.)
Comment: The same commenter achieved by the relevant best sources The commenter further states that
disagrees with EPA’s argument that the would require kilns to install back-end docket records for Portland cement, the
governing case law (National Lime controls, thus bypassing beyond-the- hazardous waste standards, and electric
Ass’n and CKRC) did not involve facts floor requirements of achievability, utilities demonstrate that various
where the levels of performance tests considering cost and other statutory pollution controls have the ability to
are dependent entirely on composition factors. Contrary to EPA’s position, the reduce mercury emissions. We agree
of raw materials and fuel and cannot be commenter argues that sources are using with this comment in part. We believe
replicated or duplicated. The low mercury fuel and feed and some both ACI and wet scrubbers will reduce
commenter states that the governing kilns are using controls that reduce mercury from cement kilns (and the
case law addresses that exact issue: mercury emissions, albeit they may not floor for mercury for new sources is
based on performance of a wet a particular HAP, the facility is taking considering the ‘‘cost of achieving such
scrubber). We did evaluate these specific actions that results in a emission reduction’’ and other
controls as beyond-the-floor control reduction of the pollutant. For example, enumerated statutory factors. According
options and determined, based on what a facility that installs a thermal oxidizer to the commenter, the only relevant
we consider reasonable assumptions of to reduce total hydrocarbons also factors regarding the cost measures are:
their performance, that requiring reduces organic HAP, even though the (1) Whether it is too costly to be
facilities to apply these controls was not thermal oxidizer may not have been ‘‘achievable’’; and (2) whether it would
achievable, within the meaning of installed for purposes of HAP reduction. yield additional reductions, i.e., without
section 112(d)(2) of the CAA, after However, the facility is still taking a the measure, the standard would not
considering costs, energy impacts, and specific action that reduces HAP reflect the ‘‘maximum’’ achievable
non-air quality health and emissions. Also, another facility can degree of reduction. The commenter
environmental impacts. install a similar control device and states that EPA does not claim that the
We also agree that fabric filters and expect to achieve the same result. use of ACI would not be achievable,
ESPs can reduce mercury emissions Results thus can be duplicated from site only that ACI is not ‘‘justified.’’ This
because there is some mercury retained to site. position, according to the commenter,
in the collected CKD. As explained In the case of cement kilns, the contravenes the CAA and exceeds EPA’s
earlier, we agree that this forms the ‘‘actions’’ being taken that in some cases authority and would allow EPA to avoid
basis of a MACT floor (and standard), may reduce mercury emissions are the properly determining the maximum
although the degree of mercury result of site specific factors that cannot degree of reduction achievable
reduction is site-specific based on the necessarily be duplicated elsewhere. For considering cost and the other
rate of recycling per kiln. Because the example, facility A may achieve lower enumerated factors.
amount of emission reduction mercury emissions than facility B Response: We disagree with the
associated with the practice is site simply because the limestone quarry commenter’s interpretation.
specific and not directly measurable, we used by facility A has a lower mercury The statute requires that EPA consider
are expressing the standard as a work content (at least on the day of the ‘‘the cost of achieving such emission
practice. We also explained why respective performance tests). Facility A reduction’’ (section 112 (d)(2)) in
requiring further reductions based on is not achieving lower mercury determining the maximum emission
more CKD wastage is not justified as a emissions deliberately, but it is still reduction achievable. This language
beyond-the-floor standard based on achieving a lower level. However, does not mandate a specific method of
considerations of cost and adverse non- because facility B does not have access taking costs into account, as the
air quality health and environmental to facility A’s quarry, it would have to commenter would have it, but rather
impacts (increased waste generation and use some other control technique to leaves EPA with significant discretion
disposal), as well as increased energy match facility A’s mercury emissions. as to how costs are to be considered. See
use. The commenter never disputes that Husqvarna AB v. EPA, 254 F. 3d 195,
In no case did we find that any of the requiring facility B (and quite possibly 200 (D.C. Cir. 2001). In that case, the
control options discussed by the A) to match the performance will court interpreted the requirement in
commenter could be considered as the require installation of a control device section 213 (a) (3) of the CAA (which
basis for a MACT floor for new or not used in the industry. As explained mirrors the language in section
existing sources (with the two at proposal and earlier in the preamble, 112(d)(2))that nonroad engines ‘‘achieve
this amounts to an impermissible de the greatest degree of emission
exceptions just noted) for reasons
facto beyond-the-floor standard. reduction achievable through the
previously discussed.
We also note that the HWC NESHAP The commenter also states that the application of [available] technology
does have mercury limits. However, best performing kilns are achieving * * * giving appropriate consideration
these limits are achieved by controlling lower mercury emission using a variety to the cost of applying such
of methods, but does not offer any data technology’’, and held that this language
the mercury input of the hazardous
or analysis as to what these methods ‘‘does not mandate a specific method of
waste feed (through source separation,
are, or how other facilities could cost analysis’’. The court therefore
blending, or other means). Therefore,
duplicate the performance of the lower ‘‘f[ound] reasonable EPA’s choice to
any comparison of the mercury limits
emitting facilities without adding some consider costs on the per ton of
for cement kilns that burn hazardous
type of back end controls. In addition, emissions removed basis’’.
waste with cement kilns that do not is Moreover, where Congress intended
due to the wide variation in emissions
misplaced.15 that economic achievability be the
level due to variations in raw materials,
The commenter notes that cement means of assessing the reasonableness of
we have no data to show conclusively
kilns are achieving superior mercury costs of technology-based
that even if back end controls were
emissions through a variety of different environmental standards, it says so
applied that kilns with higher mercury
means, and further states that whether explicitly. See Clean Water Act section
emissions due to higher mercury
they are doing this intentionally is contents in their limestone could 301 (b) (2) (A) (direct dischargers of
legally irrelevant. The comment is achieve the same emissions levels as toxic pollutants to navigable waters
correct that the reason for application of facilities with naturally occurring low must meet standards reflecting ‘‘best
a particular control technique is mercury limestone used in the (one- available technology economically
irrelevant. National Lime, 233 F. 3d at time, snapshot) performance test. achievable’’ (emphasis added). There is
640. But the commenter fails to consider Comment: Regarding EPA’s rejection no such explicit directive in section 112
that even in the case where a facility of a beyond-the-floor mercury standard (d)(2). EPA accordingly does not accept
applies some type of control scheme, on the basis of low levels of mercury the commenter’s interpretation.
and that scheme happens to also reduce
emissions and high costs of reducing Comment: Several comments support

15 Indeed, the entire reason that hazardous waste
emissions, one commenter states that EPA’s decision not to develop either an
burning cement kilns are a different source category
the CAA requires that EPA’s standards existing or new source floor for
is the impact and potential controllability of the must reflect the ‘‘maximum degree of mercury. The commenters state that an
hazardous waste inputs. See 64 FR at 52871. reduction that is achievable’’ achievable floor cannot be developed
because wide variation in mercury EPA should collect data on mercury which results in predictable and
concentrations in raw materials and emissions and then determine mercury consistent differences in mercury
fuels used by cement kilns would make limits based on data. Recommendations emissions. While the statement that
compliance impossible. One commenter for collecting additional data included only some facilities use fly ash is
also agrees with EPA’s statement that a soliciting test data from State and local correct, there are no data to indicate that
national conversion of cement kilns to agencies. Several commenters state that the use of fly ash results in consistent
natural gas is not possible due to serious EPA should conduct a new MACT floor and predictable differences in mercury
supply problems and the lack of an and beyond-the-floor evaluation based emissions. All the raw materials and
adequate natural gas infrastructure. on current and complete data— fuels that enter the kiln affect mercury
Response: We agree with these including data from state and local emissions. The decision to use fly ash
comments that the Agency cannot agencies where cement plants are may or may not affect mercury
establish a floor based on raw material located—on mercury emissions from emissions based on the mercury content
or fuel inputs. Portland cement plants. According to of the raw materials the fly ash replaces.
Comment: One commenter restates its one commenter, EPA explained that its The only way to predict the impact on
original position that EPA’s arguments decision not to set mercury standards mercury emissions of fly ash for the
regarding its inability to establish floors was due to a lack of emissions data plant currently using this material
are irrelevant, unlawful and arbitrary. while in reality it chose not to gather would be to obtain long term detailed
The commenter states that evidence data under an incorrect statutory raw materials and fuel analyses for
made available since the original interpretation that it did not have to set every plant, including analyses of the
comment period closed confirms that: standards if it believed there was no replaced materials. However, in many
(1) Some kilns perform better than control technology available. The cases the replaced materials may no
others; (2) consistent and predictable commenter states that now EPA has longer be available. Neither are the data
differences in emission levels can be access to more mercury emissions data available for the current materials being
attributed to differences in the raw than it initially claimed including: (1) used. In no way does the use of fly ash
materials, fuel, kiln design and control Toxic release inventory (TRI) data based make the mercury emissions any more
technology; and (3) additional measures on mercury stack monitoring by 35 consistent than for facilities not using
for controlling mercury emissions are plants and, (2) as indicated by EPA, data fly ash, or vice versa. All kilns are still
available to kilns. The commenter states on mercury content of coal fly ash, subject to uncontrollable variations in
that there is evidence that: (a) Some shale, and clay that is either already raw materials and fuels, of which fly ash
kilns use raw materials that are available or can be easily obtained from is only a small part. In fact, the two
consistently higher or lower in mercury existing sources—the commenter notes facilities with the highest measured
than other kilns as evidenced by a that Florida DEP reports that kilns mercury emissions do not use fly ash,
cement kiln in Tehachapi, California collect several samples of the mercury and one of these facilities, which
that uses limestone from a quarry levels in their raw materials on a daily happens to have 30 days of feed
adjacent to an abandoned mercury mine basis. materials analyses for mercury, shows
and consistently reports high (2000 lb/ Response: We disagree that our significant variations in mercury
yr) mercury emissions—other kilns have arguments regarding the inability to emissions. There are no data to support
consistently lower mercury levels establish floors are irrelevant, unlawful any contention that using fly ash will
because they use raw materials with low and arbitrary. We agree that some kilns inevitably result in a mercury emissions
mercury levels; (b) there are many emit less mercury than others in increase at any specific site.
measures by which mercury emissions individual performance tests. The The commenter also stated that kiln
can be reduced as exemplified by argument that these kilns consistently design—wet versus dry—affects
Holcim’s statement that mercury perform better over time than other mercury emissions. There are no data to
emissions can be controlled by careful kilns is not correct, however, as shown support that statement, nor are we
input control and EPA’s in section IV.A.1.a above, where we aware of any reason a wet or dry kiln
acknowledgement that mercury showed that one of the lowest emitting would perform differently with respect
emissions are affected by the use of kilns in a single test was one of the to mercury emissions. The information
mercury-contaminated fly ash—as only highest emitting in a later test due to referred to by the commenter is from the
39 of 112 plants choose to use fly ash, raw material mercury variability. We TRI. These data do not differentiate
the commenter states that a plant’s thus do not believe it is appropriate to between kilns that burn hazardous
deliberate choice about using fly ash (as use the term ‘‘perform better then waste, which are a different class of kiln
well as the choice by some to burn tires, others’’ because this implies that the subject to different regulations, and
or choosing to burn a rank of coal lower emission levels achieved are the result those that do not. Cement kilns that
in mercury, and use of by products from of some controllable action or otherwise burn hazardous waste tend to be wet
steel mills and foundries and flue gas will perform over time at some kilns and also tend to have higher
dryer sludge) results in consistent and predictable level. A facility cannot mercury emission than kilns that do not
predictable differences in their mercury achieve a performance level similar to burn hazardous waste, because of higher
emissions; (c) wet kilns emit more another facility by varying its inputs mercury levels in the hazardous waste
mercury than dry kilns (twice as much because, as previously discussed, one fuels burned by these kilns. Therefore,
according to EPA), showing that the kiln facility does not have access to another’s the data cited by the commenter do not
design results in a consistent and raw materials (or fuels), and therefore support their conclusion.
predictable difference in mercury cannot be expected to necessarily Several commenters also suggested
emissions; and (d) additional emissions achieve the same mercury emissions that EPA collect additional emission test
data confirm that some kilns are levels based on input control. The data from State and local agencies. We
achieving consistently better emission commenter acknowledges that facilities collected additional data, and have
levels than others. Several comments have significant variations in raw begun the process of gathering more.
were received regarding the adequacy of materials mercury content. See section IV.A.1.b above, and the
the emissions data used in EPA’s The commenter also notes that only separate notice in today’s Federal
analyses. Several commenters state that some facilities choose to use fly ash Register announcing reconsideration of
the new source standard for mercury. lignite coke, sulfur containing be considered the basis of a MACT floor.
We believe data in the record chemicals, or combinations of these Dry scrubbers and wet sorbent injection
conclusively show that because of the compounds. Sorbent injection systems systems have been applied at one
variations in raw materials mercury are demonstrated at the Holcim Dundee location each, but these systems do not
content show that any mercury limit plant which is limited by its permit to operate continuously and would
based on these data would not be 115 lb/yr mercury, most of which is therefore not be considered as a floor
achievable on a continuous basis, even assumed to be from coal. Mercury limits technology. We evaluated the carbon
by the kilns that form the basis of the are also in place under the hazardous injection system mentioned by the
floor, without the requirement of waste combustor rule (70 FR 59402): commenter. However, the configuration
applying beyond-the-floor back end 120 µg/dscm for new or existing cement of this system is different from the
control technology. The TRI monitoring kilns; 130 µg/dscm for hazardous waste configuration required to achieve a
data referenced by one commenter is incinerators; 80 µg/dscm for large mercury reduction. The fact that the
actually short term tests. To our municipal waste combustors. The facility meets a specific mercury limit is
knowledge, there are no cement kilns commenter states that these limits set a not attributable to the sorbent injection
using mercury continuous monitors. precedent for establishing more system, which is configured for control
The data the commenter referenced stringent mercury emission limits and of total hydrocarbons. (See section IV.C.
from Florida are daily samples, but they that there are abatement technologies on why this facility does not represent
are only analyzed on a monthly basis. In available to exceed requirements. The new source MACT for THC emissions.)
any case, any emission limit based on commenter provided emissions data for We also are aware that wet scrubber
these data would not solve the problem several U.S. cement kilns as well as technology has been applied to at least
that other facilities do not have access emissions data from cement kilns five cement kilns, and therefore we did
to the same raw materials. operating in Europe. The commenter evaluate wet scrubbers as a floor
Comment: In commenting on the states that sorbent injection control technology for both new and existing
adequacy of EPA analysis of the MACT technology is proven for mercury sources and as a beyond-the-floor
floor for existing and new sources, control and states that this technology technology for existing sources. Our
several comments were received has been demonstrated on full-scale analysis and conclusions are set out in
recommending that EPA give further demonstrations in the electric sections IV.A.1.d and IV.A.2 above.
consideration to requiring the use of generating sector. According to the We did not evaluate control
emission control technology for commenter, activated carbon is also technologies other than wet scrubbers
reducing mercury emissions. used to remove SO2, organic and ACI as a potential beyond-the-floor
Several commenters state that EPA’s compounds, ammonia, ammonium, HCl, technology. We have no data to indicate
analysis should have considered wet hydrogen fluoride, and residual dust that these controls are any more
scrubbers, dry scrubbers, wet absorbent after an ESP or FF and that the spent or efficient or cost effective than the
injection, dry absorbent injection, and used sorbent can be used as a fuel in the controls we did evaluate. In addition the
fly ash retorting with mercury controls. kiln and the particles are trapped in the performance of these controls is less
One commenter states that in evaluating clinker. The commenter notes that a certain than either wet scrubbers or ACI.
the MACT floor, EPA should establish a cement manufacturer in Switzerland, The commenter also notes that
link between mercury emissions and fueled with renewable sludge waste, mercury limits have been applied to
existing controls for sulfur and used activated carbon to achieve up to other source categories and to cement
particulate matter and examine 95 percent reduction in SO2 which kilns that burn hazardous waste. The
potential co-benefit reductions. correlates to an emission rate of less application of an emission limit to
According to the commenter, this would than 50 µg/m3. another source category or class of
be similar to the approach used by EPA One commenter states that EPA cement kiln does not, in and of itself,
in establishing the initial mercury caps should also consider pre-combustion indicate that a mercury emissions limit
in the Clean Air Mercury Rule (CAMR). technology for coal that has been is required or appropriate here. With
The commenter believes that specific demonstrated in the utility sector. One respect to the mercury standards for
control equipment will result in a such technology, pre-combustion coal cement kilns that burn hazardous waste,
percent reduction of mercury whether beneficiation, transforms relatively low as noted earlier, these standards are
the mercury is from feedstock or from cost, low rank western coal (lignite or based exclusively on control of mercury
fuel. Standards could be expressed as a subbituminous) into a cleaner more levels in the hazardous waste fuel
desired percent control achieved using efficient energy source (k-FuelTM). This inputs, and hence are not applicable to
a specific control technology technology applies heat and pressure to the Portland cement kiln category. See
combination for sulfur and particulate reduce moisture and can increase heat 70 FR 59648. In addition, we note that
matter as was done in the coal-fired value by 30–55 percent for low rank the limits mentioned are well above the
electric steam generating unit coals. The result is higher output per emission test data for all but two cement
determinations. The commenter states ton of coal while lowering emissions kilns that do not burn hazardous waste.
that such an approach is necessary to including reduction in mercury content Cement kilns that burn hazardous waste
determine a new source standard for by up to 70 percent or more and typically have stack gas concentrations
Portland cement kilns. The commenter reduced emissions of SO2 and NOX. of 43 to 196 µg/dscm resulting from the
included the tables that were developed Response: We have reevaluated the hazardous waste alone (69 FR 21251,
for the percent reduction determination available emission control technology April 20, 2004). These levels, which
for electric utilities. One commenter for reducing mercury emissions. The reflect only the mercury emissions
states that more than 60 U.S. and 120 commenters mentioned numerous attributable to the hazardous waste, are
international waste-to-energy plants control technologies including wet themselves higher then the majority of
fueled with municipal or industrial scrubbers, dry scrubbers, wet sorbent the emission levels from cement kilns
waste or sewage sludge use sorbent injection, dry sorbent injection, and fly that do not burn hazardous waste, the
injection ahead of fabric filters to ash retorting. Dry sorbent injection and majority of which are below 43 to 196
remove mercury from flue gases. The fly ash retorting have not been applied µg/dscm. See ‘‘Summary of Mercury
sorbents used include activated carbon, to cement kilns. Therefore, they cannot Test Data’’ in Docket EPA–HQ–OAR–
2002–0051. Therefore, we believe it is publication of the PCA documenting Response: We agree that reducing the
reasonable to assume that cement kilns this. The two commenters state that one recycling of CKD has, in some cases,
that do not fire hazardous waste are opportunity to avoid the recycling of been shown to reduce mercury
much lower emitters of mercury than CKD is by mixing it with clinker to emissions and that this practice creates
the hazardous waste-firing cement kilns. make masonry and other types of a floor for both existing and new
The commenter also mentioned pre- cement. One commenter states that CKD sources. See section IV.A.1.c above. The
combustion technology for mercury has numerous beneficial uses and can amount of CKD recycled versus the CKD
control, including k-Fuel. Coal cleaning be sold as a byproduct by cement plants. wasted at any facility is based on the
is another option for removing mercury The commenter addresses some of the concentration of alkali metals in the raw
from the fuel prior to combustion. In barriers to the practice of mixing materials. Also, the effect of this
some states, certain kinds of coal are materials with clinker to make materials practice on mercury emissions will be
commonly cleaned to increase its for sale. In response to comments that highly variable because the amount of
quality and heating value. the industry apply various non-ACI mercury present in the cement kiln dust
Approximately 77 percent of the eastern controls or work practices to reduce varies from facility to facility. Thus, we
and midwestern bituminous coal mercury emissions, one commenter have adopted a work practice standard
shipments are cleaned in order to meet states that none of these practices have which will reflect these site-specific
customer specifications for heating been demonstrated to be effective in practices. We also have evaluated a
value, ash content and sulfur content. controlling mercury emissions from beyond-the-floor control option based
See Mercury Study Report to Congress: cement kilns. on further reducing the recycling of
Volume VIII: An Evaluation of Mercury One commenter states that EPA could CKD back to the cement kiln and
Control Technologies and Costs, consider prohibiting or limiting CKD determined it was not achievable
December 1997. Given the fact that most recycling in cement kilns while (within the meaning of section 112
coal is already cleaned, we believe that requiring ACI in conjunction with (d)(2)) after considering costs, energy
any benefits of mercury reduction from existing particulate matter control impacts, and non-air quality health and
coal cleaning are already being realized. devices. According to the commenter, environmental impacts. This would also
There is only one k-Fuel production this approach would avoid the expense be the case if one combined ACI and
plant of which we are aware, so this fuel of an additional control device and its reduced or eliminated the recycling of
is not available in sufficient quantities associated waste stream. The CKD.
to be considered as a potential commenter recognizes that there is a One commenter also suggested the
alternative fuel. We are not aware of any possibility that the mercury and carbon use of lower mercury fuels, specifically
widely available coals that have been level in the CKD may cause it to be petroleum coke, and setting a limit for
subjected to more advanced coal considered a hazardous waste. mercury emission based on the upper
cleaning techniques. We also note that Two commenters support the use of bounds of the limits of mercury in the
advanced coal cleaning techniques have alternative feed and fuel materials as feed and fuel. The comment on
an estimated cost of approximately $140 techniques for reducing mercury petroleum coke is addressed above in
million per ton of mercury reduction. emissions. One commenter states that section IV.A.1.a.i. We rejected this later
These costs per ton of removal are EPA’s evaluation of low-mercury fuels option because it would set a limit that
higher than costs of other potential should have included petroleum coke. has no environmental benefit because it
beyond-the-floor technologies such as According to the commenter, testing at achieves no emissions reduction. See
ACI and wet scrubbers. one kiln has shown that petroleum coke section I.A.1.b above. Another
Comment: Several comments were contained significantly less mercury commenter mentioned the problems
received regarding the need for EPA to than the coal previously used to fuel the with setting a limit based on changes to
include in its analysis of the MACT kiln. The commenter also suggested fuels, namely that limited supply would
floor the use of work practices alone or evaluating the increasing use of tire- preclude any MACT floor based on fuel
in combination with control derived fuel and its impact on mercury switching, and would likewise preclude
technologies to reduce mercury emissions. One commenter states that any beyond-the-floor option. We agree
emissions. Two commenters state that data are available that indicate that with those comments. See 70 FR 72334.
the work practice of wasting a portion mercury content of fuel and feed used Comment: Several comments support
of the control device catch, that is by kilns is not so variable that an upper EPA’s decision not to set ‘‘beyond-the-
disposing of a portion of the catch rather limit for mercury in coal and feed could floor’’ mercury standards for the
than recycling it back to the kiln, can not be set by EPA. One commenter following reasons: (1) Any possible
reduce total mercury emissions. One states that EPA should collect sufficient activated carbon injection ‘‘back-end’’
commenter cites a European report data on the variability of mercury in control technology would be
showing that lowering the gas feed and fuel materials to actually prohibitively expensive; (2) the cost per
temperature upstream of the baghouse determine what the variability is. mass of mercury emissions reduced
accompanied by disposing of part of the One commenter responded to would be astronomical; and (3) the
catch is an effective measure in comments recommending that kilns application of such possible activated
reducing mercury emissions. According switch from coal to petroleum coke, fuel carbon injection would generate
to the commenter, material removal is oil, and tire-derived fuel because these additional solid waste and increase
already practiced at many kilns in the have lower mercury concentrations. The energy use.
U.S. for other reasons than mercury commenter states that limited supply, Response: We agree with these
removal. This occurs for example when long distances, and permitting issues comments for the reasons previously
CKD is wasted or when a bypass is used make it impossible to replace a discussed.
at kilns with preheaters to relieve significant percentage of the coal burned Comment: A commenter states that in
buildups of volatile components, e.g., with alternative fuels. The commenter the beyond-the-floor evaluation, EPA
chlorides or sulfates. The commenter states, however, that the industry could failed to consider other control
states that such kilns emit less mercury utilize a much larger amount of these measures that reduce mercury
through the stack than kilns that do not fuels if permitting barriers were emissions. The commenter cited coal
waste CKD. The commenter cites a lowered. cleaning, mercury-specific coal
treatments, optimization of existing control technologies for this source The commenter states that EPA’s
control (the commenter supplied a list category. In the case of any coal estimate of the cost of ACI and the
of optimizing technologies), as well as cleaning technology, we did not amount of mercury that would be
currently available control technologies specifically evaluate these technologies. reduced are arbitrary and capricious
such as enhanced wet scrubbing, We know of no case where these and, therefore, so is EPA’s reliance on
Powerspan-ECO, Advanced Hybrid technologies have been used in the cost per ton estimates as a basis for
Filter, Airborne Process, LoTox process, cement industry, or any other industry, rejecting ACI as a beyond-the-floor
and MerCAP. According to the as the basis for control of mercury technology.
commenter, mercury reductions for emissions, therefore they cannot be Two commenters state that, given
these technologies range from 20 considered a floor technology. We also mercury’s toxicity and the significant
percent to over 90 percent. According to do not consider these technologies to be mercury emissions from Portland
the commenter, EPA’s failure to demonstrated to the point where we cement plants, they strongly disagree
evaluate any of these measures is would consider them as the basis of a with EPA’s conclusion that standards to
arbitrary and capricious and beyond-the-floor standard. As noted limit mercury emissions are ‘‘not
contravenes CAA 112(d)(2) which above, most coals are already cleaned. justified.’’
requires the agency to set standards Coals that have been cleaned using Response: The commenters did not
reflecting the maximum degree of advanced cleaning techniques are not provide data to support their claims that
reduction achievable through the full generally available. In addition, data mercury emissions from this source
range of potential reduction measures. from an evaluation of advanced coal category are significantly
In a later comment, the same cleaning indicated that the costs were underestimated. We are aware that
commenter states that EPA failed to approximately $140 million per ton of recent tests at several facilities have
satisfy the CAA by not considering end- mercury reduction. See Mercury Study indicated that they had significantly
of-stack controls. As an example of a Report to Congress: Volume VIII: An underestimated their mercury
controlled source, the commenter states Evaluation of Mercury Control emissions. In some cases the mercury
that Holcim’s Zurich plant successfully Technologies and Costs, December emissions were significantly higher. We
uses the Polvitec system, a carbon filter 1997. are also aware of recent tests where the
system that controls mercury as well as Comment: Citing the information used measured mercury emissions were low,
organic pollutants. to estimate costs and mercury and in at least one case was actually
One commenter objects to EPA’s reductions associated with ACI as below previous estimates. We do not
refusal to set beyond-the-floor mercury outdated, unsupported and agree that these few cases indicate that
standards as unlawful and arbitrary. The unexplained, one commenter states that our current estimates of mercury
commenter states that EPA failed to EPA’s estimates are inadequate and, emissions are significantly in error.
consider eliminating the use of fly ash furthermore, ignores the more recent Comments: Several commenters state
as a beyond-the-floor standard even ACI data used in EPA’s power plant that EPA has ignored or undervalued
though it is possible for kilns not to use rulemaking. non-air impacts. Commenters state that
fly ash—a majority of kilns do not use Response: We have updated our ACI EPA should consider non-air
any fly ash—and not using fly ash costs based on more recent information. environmental, economic, and societal
would reduce mercury emissions. For As explained above in discussions of impacts resulting from contamination of
example, the commenter states that potential beyond-the-floor options based water bodies and their lost recreational
more than half the mercury emissions on performance of ACI, we still do not and commercial fishing uses negatively
from an Alpena, MI kiln are from fly find such standards to be achievable affecting tourism and jobs; and
ash. According to the commenter, kilns within the meaning of section 112 neurological effects on children caused
could also reduce mercury emissions by (d)(2). by mercury exposures among females of
using cleaner fuel (e.g., natural gas), Comment: One commenter states that child-bearing age. According to
using coal with lower mercury content, recent tests for mercury emission from commenters, local advisories against
refraining from the use of other mercury Portland cement plants in New York eating fish due to mercury tissue levels
containing by-products from power and Michigan show that EPA does not undercut efforts to encourage fish
plants, steel mills, and foundries, and have an accurate picture of mercury consumption as a way to reduce risk of
refraining from the use of flue gas dryer emissions from this industry. The heart disease. One commenter states
sludge. One commenter recommends commenter states that the lack of that in failing to set maximum degree of
that EPA conduct a new beyond-the- accurate information affected EPA’s reduction standards that are achievable,
floor evaluation based on up-to-date and analysis of ACI as a beyond the floor EPA did not consider the costs of not
complete data. control. The commenter recommends setting mercury standards, including the
Response: We have conducted that EPA conduct additional stack public health costs of increased
additional beyond-the-floor analyses for testing to collect accurate emissions exposure to mercury in children as well
all demonstrated control techniques for data. as the societal costs of contaminated
cement kilns. This included banning One commenter also states that EPA water bodies, fish, and other wildlife.
use of utility boiler fly ash as feed to does not provide information on the Response: The purpose of 112(d)
cement kilns, reducing the recycling of amount of mercury that would be standards is to apply maximum
CKD, use of wet scrubbers, and use of reduced by ACI. The commenter states achievable control technology. The
ACI. The statement that not using fly that self-reported mercury emission data consideration of impacts such as those
ash would reduce mercury emissions is provided by industry in EPA’s TRI, discussed above is performed during the
not supported by existing data, as appear to grossly underestimate actual section 112(f) residual risk phase. See
explained in section IV.A.1.b above. kiln mercury emissions and provides Sierra Club v. EPA, 353 F. 3d 976, 989–
These are discussed in section I.A.2 examples of such under-reporting. 90 (D.C. Cir. 2004) (rejecting the
above. The commenters mentioned Based on the limited emissions test commenter’s argument). We have begun
other additional control techniques data, the commenter states that actual this analysis for this source category.
including both add-on controls and coal mercury emissions data could be ten The results of this analysis will be
cleaning. These are not demonstrated times greater than the TRI estimates. included in a separate rulemaking.
Comment: Several commenters raised companies having trouble finding ways economic impacts of lost revenue from
concerns related to the local impacts of to manage fly ash that would not the sale of fly ash, landfill disposal fees,
industrial mercury emissions. increase impacts on land use and other and the potential rate increases for
According to one commenter, the high ecosystem values. These commenters electricity consumers; and the
temperature of cement kilns results in state that further study of such trade-offs environmental impacts of relying on
mercury emissions that fall out and are is necessary. virgin feedstock—which contains
deposited much closer to the source Another commenter notes that mercury as well as organic
than was previously thought. One approximately 2.5 million tons of fly compounds—including increased
commenter cites research that confirms ash is used annually in cement kilns, energy use, additional air emissions,
that mercury disproportionately affects thus reducing the need for an equivalent and impacts on natural resources.
nearby residents and that shows that amount of natural materials that would One commenter states that there are
nearly 70 percent of the mercury in an come from virgin sources. Another many advantages (a list of the
area’s rainwater comes from nearby commenter notes that some environmental and energy benefits is
coal-burning industrial plants. One configurations of coal-fired electric included as part of the comment)
commenter states that EPA did not generating unit control equipment can associated with the use of fly ash as an
consider impacts of mercury hot spots, reduce the level of ash-bound mercury, alternative for some naturally occurring
citing Florida and EPA research and that research is being conducted on raw materials. The commenter states
showing a reduction in local and methods that capture and stabilize that they also understand the impacts
regional fish mercury levels when mercury, producing a secondary waste that the use of fly ash may have on
MACT standards for medical and product separate from the ash stream. mercury emissions and are looking at
municipal incineration were One commenter adds that the costs of approaches that may be used to
implemented. The commenter provided replacing fly ash with other materials minimize mercury emissions from use
documentation of impacts on local could be in excess of $10 million per of fly ash. They state that they will
environments of lowering local or ton of mercury removed. This provide additional information on a
regional mercury emissions. One commenter also states that the use of preferred approach should one be
commenter states that they are some alternate materials could result in identified.
concerned over the documented levels emissions of HAP, including mercury, One commenter opposes a blanket
of mercury in fish in their county and and increased emissions of criteria ban on use of fly ash without regard to
the fact that three recently permitted pollutants either directly or as the result its source or the use of analysis to
Portland cement plants in their county of increased fuel usage per ton of clinker determine mercury content. The
are permitted to emit over 400 lb/yr of produced. One commenter agrees with commenter agrees that setting mercury
mercury in addition to a coal fired EPA that fly ash from electric utility emission limits is inappropriate given
electrical generating plant that emits boilers may progressively contain more the variability in concentration in raw
over 70 lbs of mercury annually. mercury as the electric utility industry materials and that it would be contrary
Response: These factors will be reduces its mercury emissions. to case law under CAA section 112. The
considered in the section 112(f) residual According to the commenter, some commenter lists the manufacturing and
risk analysis discussed above. It is boiler fly ash is of a quality that allows environmental benefits of using fly ash
impermissible to consider these risk- it to be added directly as a raw material as a substitute for other raw materials:
based factors in setting the technology- for concrete where most of the mercury reduced fuel consumption in kiln;
based standards at issue here. is permanently bound; lower quality fly reduced power consumption for
Comment: EPA solicited comments on ash is unusable in concrete and instead grinding; reduce emissions of organics
a potential ban of the use of mercury- is added as a raw material additive to (THC) and combustion emissions (NOX,
containing fly ash from utility boilers as the cement kiln. This commenter, SO2, and CO); reduce need to dispose of
an additive to cement kiln feed. however, recommends that EPA fly ash; and reduced SO2 emissions from
Numerous commenters state that a ban consider work practices, monitoring, reduced use of raw materials containing
is premature for several reasons, with and mercury controls rather than a ban pyrites. The commenter states that in
their objections falling into one of on fly ash. some regions, fly ash is the only source
several groupings: anti-Resource Two commenters state that data from of aluminum for some cement plants.
Conservation and Recovery Act (RCRA) TRI showing that 64 percent of kilns not Also, they state that like other raw
policy to encourage recycling that is using fly ash account for 60 percent of materials, the mercury content of fly ash
protective of human health and the mercury emissions, while the 36 percent can vary widely. The commenter
environment, CAMR in litigation, that do use fly ash account for about 40 recommends an approach that allows
mercury removal technology not yet percent of mercury emissions, do not the use of fly ash if companies can
developed, substitutes may be more justify a conclusion that fly ash demonstrate that mercury emissions
harmful, and cost of a ban has not been feedstock from utility boilers that will not be significantly impacted. Such
considered. Due to these concerns about control mercury is a culprit in mercury an approach is being developed by the
the completeness of data they believe emissions from cement kilns. commenter and will be submitted to
are relevant to banning the use of fly ash Two commenters, citing EPA’s EPA as a supplement to their comments.
as a cement plant raw material, the positing that wet kilns may emit more Response: We have considered the
commenters suggest the fly ash ban be mercury than dry kilns, suggest that the comments above and have come to the
postponed and studied further for now. driver for mercury emissions from kilns conclusion that a ban on the current use
Two commenters add that banning fly may be the type of kiln rather than the of utility boiler fly ash is not warranted.
ash use, thereby requiring cement feedstock. See section I.A.1.b above.
manufacturers to use substitutes for raw Two commenters note that EPA Comment: Several commenters are
materials, cannot be used as the basis of acknowledges that the proposed ban opposed to allowing the use of fly ash
a national rule due to the variability of fails to consider the solid waste and if it means increased mercury
mercury content of fly ash. These economic impacts of diverting 2–3 emissions. One commenter cited a study
commenters also state that banning the million tons/yr from beneficial use to showing that fly ash mercury content
use of fly ash could result in power disposal in landfills, including the can vary from 0.005 to 120 micrograms
per cubic gram of ash as evidence that kilns using fly ash. We also note unavailability of low-chlorine feed or
EPA needs to limit the use of fly ash in numerous positive environmental fuel justifies a decision not to set HCl
cement and should also evaluate other effects of using fly ash in lieu of shale standards for existing kilns is irrelevant,
additives, including cement kiln dust, and clay, including increases in overall because EPA has an unambiguous legal
for their mercury emissions potential. kiln energy efficiency, and a potential obligation to set floors reflecting the HCl
One commenter states that if the reduction in THC emissions. Given the emission levels achieved by the relevant
mercury in fly ash will cause the fly ash lack of data that the use of fly ash best performing kilns.
to be classified as a hazardous waste, its adversely affects mercury emissions (i.e. One commenter states that in setting
use should be banned until the fate of causes an increase in emissions over work practice standards for HCl, EPA
mercury in the cement manufacturing raw materials that would be used in did not satisfy the CAA criteria that
process is better understood. place of the fly ash) other then at one apply when it is ‘‘not feasible to
One commenter states that EPA facility, and the other positive prescribe or enforce an emission
should take into consideration future environmental benefits, we do not standard.’’ The commenter states that a
increases in the mercury content of coal believe any action is warranted on fly work practice standard is unlawful
combustion products (CCP) resulting ash use as currently practiced in the because EPA did not and could not
from the Clean Air Interstate Rule and industry. claim that: (1) HCl cannot be emitted
the CAMR. They state that the higher The commenters also expressed through a conveyance designed and
mercury content of CCP used in concern that as utility boilers apply ACI constructed to emit or capture such
producing Portland cement as well as or other sorbents to reduce their pollutant or that such conveyance
the recycling of cement kiln dust could mercury emissions, utility boiler fly ash would be inconsistent with any existing
cause mercury emissions to increase. will have significantly increased law; or (2) the application of
Several commenters understand that mercury concentrations, likely well in measurement methodology is not
fly ash is a necessary component in the excess of levels in clay and shale that practicable due to technological and
manufacturing process, but believe would be used in its place. We agree economic limitations.
measures should be implemented to with this concern. As previously noted Response: The comment is moot. EPA
avoid increased mercury emissions. One the available data indicate that ACI (or is not requiring section 112(d) control of
commenter recommends the use of fly other sorbent) can significantly increase HCl emissions since emissions of this
ash as long as control requirements are fly ash mercury content. For this reason, HAP from cement kilns will remain
included in the rule, e.g., work practice we have added a provision in the final protective of human health with an
standards and other strategies to prevent rule to ban the use as a cement kiln feed ample margin of safety and will not
an increase in mercury emissions from utility boiler fly ash whose mercury result in adverse effects on the
the fly ash. One commenter states that content has been artificially increased environment, even under highly
EPA should require either: (1) Carbon through the use of sorbent injection, conservative worst case assumptions as
injection with fabric filtration without unless it can be shown that the use of to potential exposure. See section IV.B
insufflation; or (2) treatment of the ash this fly ash will not increase mercury above, and CAA section 112(d)(4). The
to remove and capture the mercury. emissions over a cement kiln’s raw court’s opinion does not address the
According to the commenter, if these do material baseline. possibility of using the section 112(d)(4)
not adequately reduce mercury Comment: Regarding EPA’s decision authority on considering technology-
emissions, the fly ash should not be to not set HCl standards for existing based standards for HCl and EPA’s use
used. Another commenter states that kilns, a commenter states that EPA’s of that authority violates nothing in
EPA should include provisions for action is unlawful, contemptuous of either the letter or spirit of the court’s
pollution prevention plans, in which court, and arbitrary for all of the reasons mandate.
monitoring and testing of mercury cited above by the commenter in their Comment: Two commenters took
sources are conducted and appropriate comment on EPA’s action on the issue with EPA’s proposed definition of
work practices or other measures are mercury rule. In addition, the ‘‘new’’ sources as it applies to the
evaluated and implemented to control commenter also finds EPA’s proposal proposed HCl limits for new kilns.
mercury emissions. The commenter regarding HCl unlawful and arbitrary for Regarding EPA’s new source standards
states that the facility can then the following reasons. for HCl (15 ppmv or 90 percent HCl
determine the least cost approach for The commenter states that EPA reduction), one commenter states that
achieving mercury reductions. asserts that it ‘‘reexamined’’ the MACT EPA has created a compliance loophole
One commenter states that EPA needs floor for existing sources whereas the for kilns that commenced construction
to further investigate the practice of court directed EPA to ‘‘set’’ HCl before December 2, 2005 and is
adding fly ash to understand the standards. Thus, according to the unlawful. According to the commenter,
concentration of mercury being added commenter, EPA’s stated reason for not the CAA defines new source where
and subsequent emissions of mercury. setting HCl standards for existing kilns construction or reconstruction
The commenter states that if alternatives (the number of kilns equipped with commenced after the Administrator
are available, EPA should consider scrubbers is insufficient to constitute 12 ‘‘first’’ proposes regulations. The
banning the use of fly ash. percent of the kilns) is irrelevant. commenter states that EPA first
Response: We received comments According to the commenter, the proposed standards on March 24, 1998,
both for and against the use of utility approach EPA is required to take is to and that any kiln at which construction
boiler fly ash. As previously noted in average the emission levels with those or reconstruction was commenced after
this notice, we performed our own of the other best performing sources to March 24, 1998, is a new source and
evaluation of the practice based on the set the floor. The commenter states that must meet new source standards. The
available data. The result of our analysis such a level would not reflect the commenter states that EPA ignores that
was that even though we are aware of performance of scrubbers, rather it its violation of a clear statutory duty,
one facility where the use of fly ash would reflect the level achieved by the (i.e., its failure to promulgate HCl
contributes to approximately half of the best performing sources as required by standards in the 1998 rulemaking), is
facility’s mercury emissions, we cannot the CAA. The commenter states also the reason that sources built after March
state that this occurs at other cement that EPA’s reasoning that the 24, 1998, have not already installed
pollution controls necessary to meet context of a court action which in the NESHAP (section 63.6(e)). The
new source HCl standards. essentially required EPA to reexamine same commenter later submitted
Response: We disagree with these the entire issue, and re-determine what another comment restating their
comments. First, the comment is moot the standard should be. Under such position on HCl that standards for
with respect to an HCl new source circumstances, the only reasonable date existing and new kilns are not necessary
standard because, based on the for determining new source and do not represent the MACT floor.
authority of section 112(d)(4), EPA has applicability for a resulting standard Response: This comment is also moot
determined that no such standard is would be the date EPA proposes it. given EPA’s decision not to set a section
required because emissions will be at Moreover, even under the commenter’s 112(d) standard for HCl based on the
levels which are protective of human (strained) reading, EPA did not propose authority of section 112(d)(4) of the
health with an ample margin of safety, standards for mercury, hydrocarbons, or CAA.
and will not have an adverse effect on HCl for these sources in the 1998 Comment: One commenter states that
the environment. However, the same proposal until December 2, 2005; this is EPA has not demonstrated that it has
issue of the applicability date for new why the rule was remanded by the D.C. examined the costs associated with
sources is presented for mercury and Circuit.17 Hence, for the HAP covered alkaline scrubbers in establishing a
THC, so we are responding to the by this rule, the new source trigger date MACT floor for new sources. The
comment. would be December 2, 2005, even under commenter states that EPA’s scrubber
The whole premise of new source the commenter’s reading. However, we costs are not representative of a wet
standards being potentially more strict repeat that we disagree with the scrubber that can meet limits of up to 90
than for existing sources, and requiring commenter’s interpretation because it percent control of SO2. According to the
new sources to comply immediately results in situations antithetical to the commenter, EPA’s cost are for dry or
with those requirements (see section underlying premise of a new source wet lime spray systems incapable of 90
112(d)(3) (new source floor criteria are standard: namely that amendments to percent reduction on preheater/
more stringent than those for existing new source standards will result in precalciner kilns. The commenter
sources) and 112(i)(1)), is that these existing sources having to comply provides capital and annualized costs
sources are being newly constructed and immediately with both new source for a 1 million tpy kiln of $18 to $25
hence can immediately install the best standards and immediate compliance million and $4.5 to $7 million,
pollution controls without incurring the dates. This would be both unfair and respectively. The commenter states that
time or the expense of retrofitting. Put impossible. Congress simply cannot using EPA’s range of 12 to 200 tpy of
another way, new sources know from have intended this result. HCl removal, this translates to a cost of
the beginning of the construction effort Comments: Regarding the proposed between $35,000 and $375,000 per ton
what controls will be required, and do work practice standards for existing of HCl removed. The commenter states
not have to incur the higher costs and kilns (operate at normal operating that this range is higher than the range
the time-consuming disruptions conditions and operate a particulate EPA considered unreasonable for
normally associated with control control device), one commenter states existing kiln beyond-the-floor controls
retrofits. If we were to require ‘‘new that there is not enough information to ($8,500 to $28,000 per ton removed).
sources’’ that commenced construction require ‘‘normal operating conditions’’ The commenter concludes that wet
prior to December 2, 2005, to for kilns and air pollution control scrubbers are not a reasonable option.
The commenter adds that dry or wet
retroactively install controls because we device. According to the commenter,
lime spray systems can remove SO2
have changed rule requirements, then ‘‘normal’’ kiln conditions may not be
prior to the raw mill but essentially
these particular sources would have to best for HCl removal. This commenter
perform the same function as the raw
bear retrofit costs that we do not believe also states that existing operating &
mill, and therefore achieve an
were intended by the CAA. Immediate maintenance (O&M) and start up, shut incremental removal efficiency far
compliance would also be an down, and malfunction (SSM) plans below 90 percent. The commenter states
impossibility.16 already ensure normal operation. Other that this would be less cost effective
The commenter states that the statute commenters state that this proposed than EPA described for existing kiln
mandates this result because a new work practice is arbitrary as there is no beyond-the-floor technology.
source is defined as a source ‘‘normal operating condition’’ for all Response: This comment is also moot
constructed or reconstructed after the kilns in the U.S. The commenters state in relation to HCl given EPA’s decision
Administrator ‘‘first proposes’’ that a multitude of factors—combustion not to set a section 112(d) standard for
regulations ‘‘establishing an emission parameters, kiln design, raw material HCl based on the authority of section
standard’’ applicable to the source. The inputs, fuel characteristics, etc—make 112(d)(4) of the CAA. However, it now
commenter thus concludes that the new this requirement unworkable. has relevance in regards to the costs of
source trigger date must be March 24, One commenter notes that 40 CFR controlling mercury emissions because
1998, the proposal date of the 1999 rule. 63.6(e) already requires plants to we evaluated wet scrubbers for mercury
This reading makes no sense in the minimize emissions during an SSM control from existing sources as a
event to the extent consistent with good beyond-the-floor option and new
16 As it happens, under this rule, the compliance
air pollution practices and with safety sources as a floor option. We did further
date for sources which [0] commenced construction considerations. The commenter states
after December 2, 2005, and before promulgation of investigation of the potential costs of
this final rule is 3 years because the standards EPA should clarify that the proposed alkaline (wet) scrubbers and revised our
adopted are more stringent than those proposed on requirement to continuously operate cost estimates after proposal based on
December 2, 2005. See CAA section 112(i)(2). kilns under normal conditions and data developed as part of the Industrial
However, the same issue will arise should EPA operate a particulate control device is
adopt revised standards as a result of the periodic Boiler NESHAP. The scrubber costs are
subject to the SSM provisions elsewhere
review mandated by section 112(d)(6). There is no based on alkaline scrubbers specifically

indication that Congress intended the draconian designed to remove HCl and/or SO2
result of sources constructed at the time of the 17 Greenfield cement kilns, for which EPA
initial MACT rule (which could be decades in the adopted a new source standard for THC in 1999, are
from a coal-fired boiler and we have
past for a section 112 (d)(6) revised standard) to be a separate type of new source for purposes of this made the required adjustments in cost
considered new sources. analysis. to account for differences in the flue gas
characteristics of a cement kiln versus a RfC is an estimate (with uncertainty the above information submitted by
coal-fired boiler. spanning perhaps an order of another commenter that risks to health
Comment: One commenter states that magnitude) of a daily inhalation from HCl are well below levels
EPA’s proposed risk-based exemptions exposure of the human population acceptable for both chronic and acute
from HCl standards are unlawful, (including sensitive subgroups) that is impacts.
arbitrary and capricious. On the likely to be without an appreciable risk Response: As discussed in section
proposal to develop a single national of deleterious effects during a lifetime. IV.B above, we have reviewed the risk
risk-based HCl standard based on the The existence of a threshold for analysis provided by the commenter
RfC for HCl the commenter states no noncancer effects of HCl is established and agree that additional control of HCl
national risk-based HCl standard exists by general toxicological principles, i.e., is not required.
making it impossible to comment that organisms are able to repair some Comment: Regarding emission
effectively on any provisions in the amount of corrosive tissue damage of standards for THC, one commenter
cement rule that might rely on a the type caused by HCl. If the damage states that although EPA has proposed
hypothetical future rulemaking. The does not exceed an organisms’ ability to limits, they have not set standards for
commenter continues stating that any repair it, then no adverse effects will the main kiln stack at existing sources
attempt to set risk-based standards on a occur. Although the underlying data for and new sources at existing plants. The
national rule that does not exist and is HCl did not identify subthreshold commenter states that EPA’s position on
not currently available for review, exposures for chronic effects, this was THC standards is unlawful,
would contravene the CAA notice and due to experimental design issues rather contemptuous of court, and arbitrary for
comment requirements. The commenter than the absence of a threshold. EPA is the same reasons given by the
states further that 112(d)(4) allows EPA unaware of any studies, theory, or commenter above regarding EPA
to set health-based emission standards experts that suggest HCl does not have position on mercury standards (see
only for those pollutants for which a a threshold for adverse effects. above). The same commenter in a later
health threshold has been established, Comment: Two commenters submission, states that the preamble to
and that no cancer threshold has been submitted comments on the need for the proposed rule appears to indicate
set for HCl (nor is there any HCl standards. According to the that EPA did not set emission standards
classification of HCl with respect to commenters, based on a risk analysis for THC emissions from the kiln’s main
carcinogenicity and none exists). Also, using 14 preheat/precalciner kilns at 13 stack, although the regulatory text does
the commenter states that no non-cancer cement plants using a range of in-stack specify emission limits for the kiln’s
threshold has been set for HCl and that HCl concentrations as well as a main stack.
the integrated risk information system sensitivity analysis using higher Response: Since EPA is setting
(IRIS) RfC, on which EPA attempts to hazardous waste kiln HCl standards for THC (as a surrogate for
rely, does not purport to be an concentrations, risks are well below the non-dioxin organic HAP), and also
established threshold. According to the short-term and long-term thresholds. proposed to do so, this comment is not
commenter, disclaimers in IRIS negate Based on this minimal risk, the factually accurate (and, as noted in
any notion that it provides an commenters state that there is no need earlier responses, mischaracterizes the
established threshold for HCl. for an HCl standard for new kilns or the court’s mandate in any case). In
Response: We largely disagree with proposed operational standard for addition, as previously discussed, we do
these comments. Section 112 of the existing kilns. The commenters state not agree with the commenter that the
CAA includes exceptions to the general that additional data will be submitted to court’s mandate required us to set
statutory requirement to establish demonstrate that there is minimal risk standards regardless of the facts. The
emission standards based on MACT. Of and no need for HCl standards. court noted that we had inappropriately
relevance here, section 112(d)(4) As stated in its comments on the limited our analysis to add-on back end
effectively allows us to consider risk- original proposal, one commenter states control technologies. As is the case with
based standards for HAP ‘‘for which a that a standard for HCl is not warranted mercury and HCl, setting some type of
health threshold has been established’’ for either existing or new sources. Since emission limits based on test data
provided emissions of the HAP are at the close of the previous comment would mean that many facilities would
levels that provide an ‘‘ample margin of period, the commenter conducted a have to apply a beyond-the-floor add-on
safety.’’ Therefore, we believe we have study to evaluate the long term and control technology to meet the floor
the discretion under section 112(d)(4) to short term health risks of HCl emissions level of control without consideration of
develop standards which may be less from 112 kilns at 67 plants. According the costs, energy, and non-air health and
stringent than the corresponding to the commenter, risks were assessed environmental impacts.
technology-based MACT standards for using EPA modeling guidance and Comments: One commenter states that
some categories emitting threshold conservative modeling assumptions. EPA has improperly borrowed standards
pollutants, or not to set a standard if it The commenter states that based on from its 1999 regulations for hazardous
is apparent that emissions from the their analysis, both chronic and acute waste combustors, which were found
source category (i.e. from any source in risks are below acceptable levels and unlawful and vacated 18 rather than
the category, or any potential new that none of the kilns studied have the setting standards that reflect the THC or
source) would remain protective of potential to generate HCl emissions that CO emission levels actually achievable
human health and the environment with result in air concentrations exceeding by the best performing sources (12
an ample margin of safety and EPA’s RfC threshold for chronic health percent of cement kilns for existing and
protective of the environment. effects or Cal EPA’s reference exposure best performing cement kiln for new).
The data are inadequate to make a level threshold for acute effects. Based The commenter states further that
determination as to whether HCl is on these results, the commenter states although maintaining good combustion
carcinogenic in either humans or that there is no justification for an HCl

18 This is incorrect; the THC rules for hazardous
animals, so EPA has not developed an standard for new or existing cement
waste incinerators/cement kilns/lightweight
assessment for carcinogenicity of HCl. kilns. The commenter included a copy aggregate kilns were not challenged and were
The IRIS noncancer assessment for of the health risk analysis with their therefore not vacated by the D.C. Circuit. See CKRC,
HCl provides a RfC for inhalation. An comments. Another commenter refers to 255 F.3d at 872.
conditions affects THC emissions, it is For new greenfield facilities we commenters, at issue is the control of
not the only factor that does so and cites established in the 1999 rule that a products of incomplete combustion
the plants’ selection of raw materials as facility would have the option to site the (PIC) vs. control of hydrocarbons from
affecting THC emissions. The quarry at a location with low enough feed materials. They state that HWC
commenter states that EPA’s new organic content that they could meet a have the option ceasing to burn
greenfield source standard reflects that 50 ppmv THC emissions limit. We hazardous waste when exceeding the
use of low organic feed materials affects determined that this was feasible limit (and can do so easily using
THC emissions and also cites statements because two facilities had already done automatic waste feed cutoff systems)
by Florida DEP and Holcim that so at the time we promulgated the and that the HWC THC standard only
selection of feed materials can affect original NESHAP. This limit was not applies when hazardous waste is being
THC emissions. The commenter states remanded by the court and is currently burned.
that EPA admits that add-on controls, in effect. Three commenters state that the HWC
e.g., ACI and scrubber/RTO (in use on As we have previously discussed, we MACT standards were based on EPA’s
two kilns), as well as precalciner/no do not agree that the court decision RCRA Boiler and Industrial Furnace
preheater technology reduce THC compels us to set a THC standard that rules, which in turn were based on the
emissions. According to the commenter, will require some sources to install a need to safely manage hazardous waste,
because these other factors can affect beyond-the-floor control technology a need that is irrelevant to the facilities
THC emissions, EPA has incorrectly set under the guise of a floor standard. covered under the current proposal.
the floor based on good combustion These facts have not changed from the Response: We agree with this
control only. The commenter states that original NESHAP. comment and have removed the
EPA concedes that cement kilns may be However, at proposal we noted that proposed quantified limits for existing
able to achieve better THC emission facilities could control THC resulting sources. We have not removed the limit
levels than through the use of good from combustion of fuel.20 We for new sources because the basis of the
combustion alone when it discusses in explained that the basis of the MACT new source floor (and standard) is
the proposed rule that nonhazardous floor for cement kilns firing hazardous performance of a RTO (preceded by a
waste cement kilns should be ‘‘less waste was also good combustion, and scrubber to enable the RTO to function).
challenged’’ than hazardous waste kilns these kilns had established limits for Application of an RTO (in series with a
in meeting the proposed limits and that THC as a quantitative measure of good scrubber) would allow new cement
the ‘‘lack of any hazardous waste feed combustion conditions. Given the fact kilns to meet a 20 ppmv standard, or to
for a non-hazardous waste (NHW) that both classes of kilns were using the remove 98 percent of incoming organic
cement kiln should make it easier to same method of control, we proposed to HAP measured as THC.
control the combustion process.’’ The apply the same limits to kilns that did Comment: Three commenters state
commenter states that EPA did not not burn hazardous waste. We have no that EPA has no empirical data
account for the fact that nonhazardous data, and none were supplied by the demonstrating that any NHW kiln can
waste burning kilns can control their commenter, to make any judgments achieve the proposed limits on a
combustion conditions and thus THC about whether or not kilns that do not continuous basis. One commenter states
emission more easily than hazardous burn hazardous waste could actually that bench scale studies estimated that
waste burning kilns, instead just meet a more stringent standard. Because for varying organic levels, 47 percent of
borrowing the standard for hazardous the standards are based on complete samples would have resulted in
waste burning kilns without attempting combustion of the fuel, and because of emissions that exceed the 20 ppmv
to show that the proposed limits reflect the extremely high temperatures in the limit.
end of the kiln where the fuels are Response: We agree with this
what is actually achievable by the
introduced (both those that burn comment and have removed the
relevant best performers. According to
hazardous waste and those that do not), proposed limits for existing sources. We
the commenter, EPA’s arguments that it
we believe that both types of kilns have not removed the limit of new
does not have to consider factors other
should achieve comparable complete sources because the basis for the new
than good combustion were rejected by
destruction of organic materials present source floor is now the performance of
the court as irrelevant and EPA must set
in the fuels under normal operating a RTO. Application of an RTO would
the THC limits reflecting the average
conditions reflecting good combustion. allow the facilities noted in the
emission level that the best sources
Simply because we state that controlling comment to meet a 20 ppmv standard.
actually achieve. Comment: Three commenters state
Response: In the original NESHAP, THC emissions from kilns that do not
burn hazardous waste should be less that the contribution to THC/CO from
we noted that THC emissions were
difficult than controlling emissions from raw materials outweighs the measure of
primarily a function of the organic
kilns that do burn hazardous waste does THC/CO for good combustion of
materials in the kiln feed. As we have
not imply that one type of kiln can hazardous waste fuels. Thus, THC and
previously discussed, a facility has a
achieve a measurably lower THC CO are not useful indicators of good
starkly limited ability to change their
emission level than another. combustion. One commenter notes that
raw materials to reduce their organic
Comments: Several commenters state available information shows that it is
content. The fact that individual
that it is inappropriate to set THC floor difficult to correlate HC and HAP
facilities have successfully reduced
limits based on a different source emissions. The commenter further states
organic contents of their feed materials
category, i.e., HWC. According to the that several studies show that neither
to reduce THC emissions does not
THC nor CO is a reliable surrogate for
indicate that this option is available to
being a paper exercise not producing environmental good combustion or PIC or HAP
all facilities. Therefore, we cannot use
benefit. emissions. According to the commenter,
this option as the basis of a national
20 Fuel organics can be controlled because they HC emissions are a function of: (1) Raw
standard for existing facilities.19 are fed into the hot end of the kiln. Feed materials material organic content; (2) source of
are fed into the other end of the kiln and therefore
19 EPA could subcategorize each source based on have the opportunity to vaporize and leave with the
fuel and firing location; (3) temperature
its raw material organic content (each source being exhaust gas before they reach the portions of the profile; (4) oxygen concentration; and
a different ‘‘type’’), but rejects this alternative as kiln which are hot enough to combust them. (5) type of manufacturing process. One
commenter states that the high uncontrollable variability in organic the-floor analysis for THC contravenes
temperatures required for the formation HAP levels in raw materials. CAA 112(d)(2) which requires that
of cement clinker (>2700F) ensure as Comment: A commenter states that EPA’s final standards reflect the
complete combustion of fuels as is EPA failed to consider the reduction in maximum degree of reduction
possible. THC as part of the beyond-the-floor achievable through any and all
Response: We agree with the analysis of ACI. According to the reduction measures, and any claim that
comment that because organic commenter, organic HAP potentially EPA’s THC standard reflects the
contributions from processing raw controlled by ACI include maximum achievable degree of
materials is the chief contributor to polychlorinated biphenyls, polycyclic reduction would be arbitrary and
measured THC levels (since such organic matter, and polyaromatic capricious in light of EPA’s failure to
emissions are not combusted and hence hydrocarbons. According to the consider these technologies or explain
are not largely destroyed), having a commenter, to determine the maximum its decision not to base beyond-the-floor
quantified limit for THC as a measure of degree of reduction in THC emissions standards on any or all of them.
good combustion is not appropriate for that is achievable for cement kilns, the Response: We have no actual test data
existing cement kilns that do not burn CAA requires that EPA evaluate the to establish the impact of ACI on THC
hazardous waste. We disagree with the reductions achievable through the use of emissions, but are using a figure of 50
more general statements regarding the ACI. percent, which reflects the best
appropriateness of a THC indicator for One commenter states that: (1) EPA estimates of the one facility using ACI
organic HAP, and indeed are continuing did not determine, as required by the for organics control. As explained in
to utilize THC as an indicator for new CAA for beyond-the-floor standards, the section IV.C above, the facility in
sources. As noted in the proposal of the maximum degree of reduction in THC question is extremely unusual in that
original NESHAP, the organic HAP emissions achievable through GCP; (2) the uncontrolled THC emission levels
component of THC emissions varies EPA did not show that its standards are much higher than any other facility
widely (63 FR 14196). However, THC reflect the maximum degree of in the source category, so the 50 percent
emissions do contain organic HAP. reduction achievable through reduction figure is probably more
Applying MACT to THC emissions will combustion controls in light of its efficient than would be achieved
also control organic HAP, but will be findings that NHW burning kilns should industry-wide. As explained in section
less costly than attempting to set be able to achieve the THC standards IV.A.2 above, however, even assuming
individual limits for each individual more easily than hazardous waste this degree of reduction, we did not find
organic HAP (64 FR 31918). burning kilns; (3) EPA did not a beyond-the-floor option based on
We also agree with the comment that determine the maximum degree of performance of ACI to be achievable
combustion conditions in the hot end of reduction achievable through the within the meaning of section 112(d)(2).
the kiln where fuels are fired should judicious selection of raw materials The commenter also stated that we
assure destruction of organics although they acknowledge that such did not assess the maximum degree of
(including organic HAP) in the fuel. For methods will control THC emissions THC reduction achievable by optimized
this reason, we adhere to our position at and that kilns are already using it and combustion practices. There are no data
proposal that good combustion can control THC emissions through the available to perform this type of analysis
conditions in the cement kiln should use of other materials such as fly ash and none were provided by the
assure destruction of organic HAP in and kilns can and do import raw commenter. Moreover, THC levels
fuel and represents the measure of best materials from sources that are not co- significantly below those associated
performance for reducing emissions of located or immediately nearby; (4) EPA with good combustion conditions are
organic HAP from existing cement kilns. did not determine the degree of not necessarily indicative of further
As explained in section I.C above, we reduction achievable through the use of organic HAP reductions. See discussion
have chosen a different means of end-of-stack controls already in use in at 70 FR 59462–59463 (October 12,
expressing good combustion conditions the cement industry, including ACI, 2005).
than the quantified THC limit which we which EPA only considered for mercury We also did not evaluate the degree to
proposed. and dioxin control and which would which ‘‘judicious selection’’ of raw
Comment: Three commenters state reduce THC emissions significantly and materials can be used to reduce THC
that it is inappropriate to apply limits also reduce mercury and dioxin emissions, except that we have
for non-dioxin organic HAP when feed emissions; 21 (5) EPA failed to determine previously established that a greenfield
materials have varying levels of the maximum degree of reduction facility can limit THC emission to 50
organics, which EPA acknowledges by achievable through the use of limestone ppmv by selection of limestone with
setting THC limits only for new scrubber/RTO even though the agency is sufficiently low organic materials
greenfield sources (EPA also applied aware that such devices can contents. We are aware that cement
variability of feed/fuel materials in significantly reduce emissions of THC production facilities can import some
justifying rules or lack of rules for (as well as HCl) and are already in use raw materials from sources other than
mercury, HCl and non-mercury metals). in the industry and does not contend those nearby. However, the fact that in
Two commenters add that a Reaction that they are too expensive; and (6) EPA some cases materials can be imported
Engineering study shows that organics failed to consider or determine the from a farther distance does not change
emitted from kiln feed is extremely maximum degree of reduction the fact that each individual cement
variable across the country with levels achievable through the use of a carbon facility has specific raw materials needs
varying by over four orders of coke filter system such as the Polvitec based on their particular limestone and
magnitude. system in use at Holcim’s Zurich plant. other raw materials. We do not have
Response: We agree with these data, nor are data available, to develop
For the reasons (1–7) listed above, the
comments and have made appropriate a national rule that would cover every
commenter states that EPA’s beyond-
changes in the final rule to the proposed possible raw material substitution to
floor for existing cement kilns’ non- 21 Since the rule already contains a standard for reduce THC emissions.
dioxin organic HAP emissions to dioxin, incremental reductions attributable to use of The commenter also stated we did not
account for the essentially ACI are quite small; see section IV.a.2 above. assess the maximum degree of emission
reduction achievable through the use of We also considered requiring periodic that might potentially be covered. We
end-of-stack controls. However, as stack tests rather then THC CEM. considered using the current
previously discussed, there are no data However, the current rule already Nonmetallic Mineral NSPS, which
available for us to perform this analysis requires kilns at greenfield area sources established standards of performance for
for any controls other than an RTO. In to install a THC CEM. We could see no new crushers. But we have no data to
the case of an RTO, we have evaluated justification for allowing a more lenient determine if the NSPS for this source
its performance as a beyond-the-floor THC monitoring option for new kilns at category would be an appropriate
control for existing sources. In that case, non-greenfield facilities. MACT floor. Finally, we believe we can
we determined requiring a facility to Comment: One commenter states that resolve the issue by simply stating that
apply an RTO as a beyond-the-floor the requirement for THC CEM will crushers are not covered by this
option was not achievable, within the impose additional cost for no benefit. regulation. It was never our intent that
meaning of section 112(d)(2), due to the The commenter recommends that EPA this rule regulate equipment typically
high costs and adverse energy eliminate numerical limits or require associated with another source category.
utilization impacts. The new source less costly monitoring options, e.g., Comment: One commenter states that
standard for THC is based on periodic stack testing. The commenter all of the raw material handling and
performance of an RTO (in tandem with recommends that if EPA does require storage, except crushing, should be
a scrubber), as discussed previously. We CEM, extend the compliance date to at
covered by the Portland cement
do not believe any further control is least 2 years because the State
NESHAP. They state that the only non-
technically feasible. certification process requires more than
metallic mining activities subject to the
The commenter also stated we had 1 year.
Response: We have not adopted a NSPS subpart OOO are at the quarry
not considered the use of a carbon coke and at the crusher. The commenter
system. The source for this comment requirement that existing sources install
a THC monitor. For new sources, the states that under the alternative
notes that there was one facility in interpretation offered by EPA, several
Europe. We note the plant in question compliance date is ordinarily the
effective date of the rule or startup, steps characteristic of cement
was designed to burn pelletized sewage manufacturing would not be included in
sludge. The source of the comment does whichever is later. See section 112(i)(1).
However, in this case, because the new subpart LLL, for example the ‘‘on-line’’
not indicate the performance or costs of measurement devices such as cross-belt
source standard is more stringent than
this system. We assume it would neutron analyzers that are used in the
proposed (see discussion in section
perform similarly to a carbon adsorption preblending and proportioning steps.
IV.C.3 above), sources which
system, which achieves emission The commenter states further that the
commenced construction or
reductions similar to those of an RTO. raw mix fed to the raw mill is the
reconstruction after December 2, 2005,
We believe that the wet scrubber/RTO product of the very careful
but before December 20, 2006, will have
system, which is demonstrated on a instrumentally-aided proportioning and
until December 21, 2009 to comply. See
cement kiln in the United States, is a blending operation that is one of the
section 112(i)(2).
viable beyond-the-floor option. Given Comment: Two commenters favor most important series of steps in the
the lack of demonstration of a carbon including all crushers in the Portland cement manufacturing process.
coke filter in this country, the fact that cement NESHAP and establishing Response: We agree with this
we have a viable alternative as a emission limits for crushers based on comment.
beyond-the-floor option (an RTO), and the requirements in 40 CFR, subpart
the fact that the carbon coke filter is OOO, if they satisfy the requirements of VII. Summary of Environmental,
unlikely to perform any better than an the CAA. One commenter cites State Energy, and Economic Impacts
RTO, we do not believe consideration of requirements for primary crushers of 10 A. What facilities are affected by the
a carbon coke filter is warranted. percent opacity, work practices, and a final amendments?
Comment: Several commenters baghouse with outlet concentration of
oppose EPA’s proposed regulation of 0.01 grams per dry standard cubic feet; We estimate that there are
area sources for THC. Three commenters secondary crushers are subject to a 20 approximately 94 cement plants
state that there is no legal basis for percent opacity limit. The commenter currently in operation. These 94 plants
regulating area sources. The provided a copy of their State have a total of 158 NHW cement kilns.
commenters note that there is no requirements for crushers at cement We estimate that 20 new kilns with a
‘‘statement of basis and purpose’’ as manufacturing facilities. capacity of 20,900,000 tpy of clinker
required by CAA 307(d)(3). One commenter states that capacity will be subject to the final
One commenter recommends that applicability based on location relevant amendments by the end of the fifth year
EPA exempt area sources, which would to other sources is confusing and after promulgation of the amendments.
experience the same cost as major recommended that EPA put all Note that national impacts are based on
sources with fewer benefits; or consider appropriate requirements for the sources the estimated capacity increase, not on
less stringent options, e.g., periodic in one requirement and remove a specific number of model kilns.
stack test rather than CEM. 63.1340(c) altogether. B. What are the air quality impacts?
Response: As previously noted, in the Response: We agree that applicability
original 1999 NESHAP for this source based on location relevant to other For existing kilns, we estimate that
category we regulated THC emissions sources is confusing. However, in our the impacts of the amendments will
from area sources because the THC final determination on this issue we essentially be zero because we believe
emissions from a cement kiln are likely decided that crushers should not be that all existing kilns are already
to contain polycyclic organic matter. covered under this NESHAP. The performing the work practices
This pollutant is listed in section reasons are first, we have no definitive prescribed in the amendments. For the
112(c)(6) of the CAA as a pollutant. The information that there are any facilities 20 new kilns the variation in mercury
commenter provided no data that would that currently have crushers after raw and hydrocarbon emissions from kilns
lead us to change this determination (63 materials storage. Second, we have no makes it difficult to quantify impacts on
FR 14193–94). data to set a floor for existing crushers a national basis with any accuracy.
For mercury emissions we estimate a In addition to the direct air emissions new 650,000 tpy kiln to install a THC
new kiln with a capacity of 650,000 typ impacts, there will be secondary air monitor is $140,000, to install a wet
of clinker will have an emission impacts that result in the increased scrubber is $2.7 million, and to install
reduction ranging from zero to 280 lb/ electrical demand generated by new a wet scrubber/RTO is $10.7 million.
yr. We estimate the national mercury sources’ control equipment. These For kilns where the uncontrolled THC
emissions reduction to be 1300 to 3000 emissions will be an increase in emissions are below 40 ppmv, we are
lb/yr in the fifth year after emissions of pollutants from utility assuming they will opt for a lower cost
promulgation. boilers that supply electricity to the THC control, such as ACI. The
Reported hydrocarbon emission test Portland cement facilities. Assuming estimated capital cost for ACI applied to
results range from less than 1 ppmv dry two new kilns will install a scrubber a new 650,000 tpy kiln is $1.0 to $1.6
basis (at 7 percent oxygen) to over 140 followed by an RTO, three will install million. The total estimated national
ppmv dry basis (Docket A–92–53) an ACI system, and five will install wet capital cost at the end of the fifth year
measured at the main kiln stack. For 52 scrubbers, we estimate these increases after promulgation is $64 to $67 million.
kilns tested for hydrocarbon emissions to be 105 tpy of NOX, 47 tpy of CO, 157 The estimated annualized cost per
(Docket A–92–53), approximately 25 tpy of SO2, and 5 tpy of PM at the end new 650,000 tpy kiln is an estimated as
percent had emissions of hydrocarbons of the fifth year after promulgation. $34,000 to $37,000 for kilns a THC
that exceeded the 20 ppmv THC limit at monitor, $470,000 to $597,000 for ACI,
C. What are the water quality impacts? $1.4 to $1.5 million for a wet scrubber,
the main stack. The average
hydrocarbon emissions for the kilns There should be no water quality and $3.6 to $3.9 million for a wet
exceeding 20 ppmv was 62.5 ppmv. impacts for the proposed amendments. scrubber/RTO. National annualized
Assuming that most new kilns will be The requirement for new sources to use costs by the end of the fifth year will be
sited at existing locations this would alkaline scrubbers upstream of the RTO an estimated $26 to $28 million.
imply that 15 out of 20 new kilns will will produce a scrubber slurry liquid G. What are the economic impacts?
have no THC emissions reduction as a waste stream. However, we are
assuming the scrubber slurry produced EPA conducted an economic analysis
result of the THC Standard. For a new of the amendments to the NESHAP
kiln that, in the absence of the standard, will be dewatered and disposed of as
solid waste. Water from the dewatering which have cost implications. For
would emit near the average existing sources the only requirement
hydrocarbon level of 62.5 ppmv, the process will be recycled back to the sc
in the form of aqueous discharges, with any cost implication is the
application of new source MACT requirement to keep records of CKD
consisting of an RTO would result in a addition of a scrubber will increase
water usage by about 41 million gallons wastage. These costs are very small. We
reduction of about 196 tpy for a 650,000 assessed earlier Portland cement
tpy kiln. We also estimate that for 15 per year (gyps) for each new 650,000 tpy
kiln that installs a scrubber, or a regulations with greater per source
percent of the new kiln capacity will costs, and those costs did not have a
have uncontrolled emissions that national total of 460 million gyps.
significant effect on the cost of goods
exceed the 20 ppmv limit, but will use D. What are the solid waste impacts? produced. Since the conditions that
alternatives to application of an RTO produced those conclusions still exist
The solid waste impact will be the
(such as ACI) to meet the THC today, EPA believes these new
generation of scrubber slurry that is
emissions limit. These kilns will regulations will not have a discernible
assumed to be dewatered and disposed
achieve an emissions reduction of impact on the Portland cement market
of as solid waste, and solid waste from
approximately 103 tpy for a new for existing sources.
the ACI systems. The amount of solid
650,000 tpy new kiln. The total national For new sources, both the magnitude
waste produced is estimated as 519,300
reduction will be 1100 tpy in the fifth of control costs needed to comply with
tpy in the fifth year after promulgation
year after promulgation of the standard. the final amendments and the
of the amendments.
The THC and mercury standards for distribution of these costs among
new sources will also result in E. What are the energy impacts? affected facilities have a role in
concurrent control of SO2 emissions. Requiring new kilns to install and determining how the market will
For kilns that elect to use an RTO to operate alkaline scrubbers and RTO will change. The final amendments will
comply with the THC emissions limit it result in increased energy use due to the require all new kilns constructed on or
is necessary to install an alkaline electrical requirements for the scrubber after December 2, 2005, to install THC
scrubber upstream of the RTO to control and increased fan pressure drops, and monitors. As with existing sources, the
acid gas and to provide additional natural gas to fuel the RTO. We estimate cost on a THC monitor is not significant
control of PM. We estimate that the additional electrical demand to be compared to the costs assessed in the
approximately 25 percent of the 41 million kWhr per year and the earlier regulations. However, the cost for
additional capacity built in the next five natural gas use to be 271 billion cubic ACI or for the wet scrubbers/RTO
years will have to install wet scrubbers feet by the end of the fifth year. systems are significant. We estimate that
for mercury control, and 10 percent will 3 of the 20 new kilns will have to install
install a wet scrubber/RTO system for F. What are the cost impacts? ACI, 2 of 20 new kilns will be required
THC control. The SO2 emissions The final rule amendments should to install a wet scrubber/RTO system to
reductions for a new 650,000 tpy kiln impose minimal costs on existing meet the limits for THC, and five kilns
will be approximately 320 tpy, and is sources. These costs will be will install a wet scrubber to meet the
estimated as 3640 nationally. recordkeeping costs to document CKD new source mercury limits.
Note that we have determined that wastage. The costs for new sources Because of the high cost of
reducing SO2 emissions also results in include the THC monitor and transportation compared to the value of
a reduction in secondary formation of recordkeeping costs for CKD wastage on Portland cement, the market for
fine PM because some SO2 is converted all new kilns, a wet scrubber for Portland cement is localized and
to sulfates in the atmosphere. Therefore, mercury control on five new kilns, and characterized by imperfect competition.
the THC standards will also result in a a wet scrubber/RTO on two of the new The possible outcomes of the final
reduction in emissions of fine PM. kilns. The estimated capital cost for a amendments are either a deferral in
bringing the new kiln into production or increase by 1.0 to 7.2 percent, to values submitted for approval to the OMB
a price increase in the immediate region of $86 to $91, respectively. under the Paperwork Reduction Act, 44
around the two new kilns that face With the increasing demand for U.S.C. 3501 et seq. The information
control costs. For perfect competition, Portland cement and the high capacity collection requirements are not
control costs at a new facility will be utilization of existing plants and the enforceable until OMB approves them.
completely passed on in the long run to nature of the regional markets, it is
unlikely that the new kilns would be The information requirements are
the purchaser of the good. With
imperfect competition the outcome is delayed. Because of the imperfect based on notification, recordkeeping,
harder to predict. Less than full cost competition, it is likely in the regions and reporting requirements in the
pass through is a likely possibility. around the two new kilns facing control, NESHAP General Provisions (40 CFR
The model new kilns used in this the price of the Portland cement would part 63, subpart A), which are
analysis have a clinker capacity of increase but by less than the 1.0 to 7.2 mandatory for all operators subject to
650,000 tons/yr. The annual control cost percent that would be required to fully national emission standards. These
would be up to $597,000 for kilns that cover the control costs. recordkeeping and reporting
apply ACI, $1.5 million for a kiln that requirements are specifically authorized
VIII. Statutory and Executive Order
applies a wet scrubber, and $3.9 million by section 114 of the CAA (42 U.S.C.
Reviews
for a kiln that applies an scrubber/RTO, 7414). All information submitted to EPA
in 2002 dollars. Clinker is an A. Executive Order 12866, Regulatory pursuant to the recordkeeping and
intermediate good in the production of Planning and Review reporting requirements for which a
Portland cement and corresponds to a Under Executive Order (EO) 12866 claim of confidentiality is made is
Portland cement capacity of 720,000 (58 FR 51735, October 4, 1993), this safeguarded according to Agency
tons/yr. To compare the costs to the action is a ‘‘significant regulatory policies set forth in 40 CFR part 2,
value of the Portland cement in 2004 of action’’ because it raised novel legal and subpart B.
$85 for a national average mill value we policy issues. Accordingly, EPA
use the Chemical Engineering Plant Cost These requirements include records of
submitted this action to the Office of
Index for 2004 and 2002 to get a 2004 CKD removal from the kiln system at all
Management and Budget (OMB) for
annual cost of $640,000 for kilns that existing and new sources, and
review under EO 12866 and any
require ACI, $1.7 million for kiln that changes made in response to OMB requirements for new kilns constructed
apply wet scrubbers, and $4.4 million recommendations have been after December 2, 2005, to install and
for those that apply an scrubber/RTO. documented in the docket for this test a continuous monitor to measure
The value of the Portland cement action. THC. We expect these additional
produced in a year at the $85 price requirements to affect 94 facilities over
would be $61 million. If the cost were B. Paperwork Reduction Act the first 3 years. The estimated annual
to be fully passed on to the purchaser The information collection average burden is outlined below.
in a higher price the price would requirements in this final rule have been
Total annual
Affected entity Total hours Labor costs Total costs
O&M costs
Industry ............................................................................................................ 4,159 $679,105 $161,672 $840,777

Implementing Agency ...................................................................................... 213 16,100 NA 16,100
Burden means the total time, effort, or CFR are listed in 40 CFR part 9. When on small entities, small entity is defined
financial resources expended by persons this information collection request is as: (1) A small business as defined by
to generate, maintain, retain, or disclose approved by OMB, the Agency will the Small Business Administration’s
or provide information to or for a publish a technical amendment to 40 (SBA) regulations at 13 CFR 121.201; (2)
Federal agency. This includes the time CFR part 9 in the Federal Register to a small governmental jurisdiction that is
needed to review instructions; develop, display the OMB control number for the a government of a city, county, town,
acquire, install, and utilize technology approved information collection school district or special district with a
and systems for the purposes of requirements contained in this final population of less than 50,000; and (3)
collecting, validating, and verifying rule. a small organization that is any not-for-
information, processing and profit enterprise which is independently
maintaining information, and disclosing C. Regulatory Flexibility Act owned and operated and is not
and providing information; adjust the The Regulatory Flexibility Act (RFA) dominant in its field.
existing ways to comply with any generally requires an agency to prepare After considering the economic
previously applicable instructions and a regulatory flexibility analysis of any impacts of this final rule amendments
requirements; train personnel to be able rule subject to notice and comment on small entities, I certify that this
to respond to a collection of rulemaking requirements under the action will not have a significant
information; search data sources; Administrative Procedure Act or any economic impact on a substantial
complete and review the collection of other statute unless the agency certifies number of small entities. The small
information; and transmit or otherwise that the rule will not have a significant entities directly regulated by the final
disclose the information. economic impact on a substantial rule amendments are small businesses.
An agency may not conduct or number of small entities. Small entities We determined there are 6 or 7 small
sponsor, and a person is not required to include small businesses, small businesses in this industry out of a total
respond to a collection of information organizations, and small governmental of 44. Each small business operates a
unless it displays a currently valid OMB jurisdictions. single plant with one or more kilns. The
control number. The OMB control For purposes of assessing the impact total annualized cost of the standards in
numbers for EPA’s regulations in 40 of today’s proposed rule amendments the amendments for an existing kiln is
nominal. The revenue for the entire was not adopted. Before EPA establishes F. Executive Order 13175, Consultation
small business sector is estimated to be any regulatory requirements that may and Coordination With Indian Tribal
around $260 million (2003 dollars). significantly or uniquely affect small Governments
New sources, will incur higher costs governments, including tribal Executive Order 13175 entitled
because new kilns must install a THC governments, it must have developed ‘‘Consultation and Coordination with
monitor, and approximately three of the under section 203 of the UMRA a small Indian Tribal Governments’’ (65 FR
20 new kilns will have to install ACI, government agency plan. The plan must 67249, November 9, 2000), requires EPA
two will have to install wet scrubbers, provide for notifying potentially to develop an accountable process to
and two will have to install a wet affected small governments, enabling ensure ‘‘meaningful and timely input by
scrubber/RTO system for THC control.
officials of affected small governments tribal officials in the development of
For new sources that must install
to have meaningful and timely input in regulatory policies that have tribal
controls, the cost of control is estimated
the development of EPA regulatory implications.’’ The final rule
to be one to seven percent of the
proposals with significant Federal amendments do not have tribal
expected revenue from a new kiln. We
intergovernmental mandates, and implications, as specified in Executive
currently do not have any information
on plans for small businesses to build informing, educating, and advising Order 13175, because tribal
new kilns. small governments on compliance with governments do not own or operate any
Although the final rule amendments the regulatory requirements. sources subject to today’s action. Thus,
will not have a significant economic Executive Order 13175 does not apply
EPA has determined that the final rule to the proposed rule amendments.
impact on a substantial number of small amendments do not contain a Federal
entities, EPA nonetheless has tried to mandate that may result in expenditures G. Executive Order 13045, Protection of
reduce the impact of the final of $100 million or more for State, local, Children From Environmental Health
amendments on small entities. The and tribal governments, in the aggregate, Risks and Safety Risks
emission standards are representative of
or the private sector in any 1 year, nor Executive Order 13045 (62 FR 19885,
the floor level of emissions control,
do the amendments significantly or April 23, 1997) applies to any rule that:
which is the minimum level of control
allowed under CAA. Further, the costs uniquely impact small governments, (1) Is determined to be ‘‘economically
of required performance testing and because they contain no requirements significant’’ as defined under Executive
monitoring for non-dioxin organic HAP that apply to such governments or Order 12866, and (2) concerns an
emissions from new sources have been impose obligations upon them. Thus, environmental health or safety risk that
minimized by specifying emissions these final rule amendments are not EPA has reason to believe may have a
limits and monitoring parameters in subject to the requirements of sections disproportionate effect on children. If
terms a surrogate for organic HAP 202 and 205 of the UMRA. the regulatory action meets both criteria,
emissions, which surrogate (THC) is less the Agency must evaluate the
E. Executive Order 13132, Federalism environmental health or safety effects of
costly to measure.
the planned rule on children, and
D. Unfunded Mandates Reform Act Executive Order 13132 (64 FR 43255,
explain why the planned regulation is
August 10, 1999), requires EPA to
Title II of the Unfunded Mandates preferable to other potentially effective
develop an accountable process to and reasonably feasible alternatives
Reform Act of 1995 (UMRA), Public
ensure ‘‘meaningful and timely input by considered by the Agency.
Law 104–4, establishes requirements for
Federal agencies to assess the effects of State and local officials in the EPA interprets Executive Order 13045
their regulatory actions on State, local, development of regulatory policies that as applying only to those regulatory
and tribal governments and the private have federalism implications.’’ ‘‘Policies actions that are based on health or safety
sector. Under section 202 of the UMRA, that have federalism implications’’ is risks, such that the analysis required
EPA generally must prepare a written defined in the Executive Order to under section 5–501 of the Executive
statement, including a cost-benefit include regulations that have Order has the potential to influence the
analysis, for proposed and final rules ‘‘substantial direct effects on the States, rule. The final rule amendments are not
with ‘‘Federal mandates’’ that may on the relationship between the national subject to Executive Order 13045
result in expenditures to State, local, government and the States, or on the because they are based on technology
and tribal governments, in the aggregate, distribution of power and performance and not on health or safety
or to the private sector, of $100 million responsibilities among the various risks.
or more in any 1 year. Before levels of government.’’
promulgating a rule for which a written H. Executive Order 13211, Actions That
The final rule amendments do not Significantly Affect Energy, Supply,
statement is needed, section 205 of the
have federalism implications. The final Distribution, or Use
UMRA generally requires EPA to
identify and consider a reasonable rule amendments will not have This rule is not a ‘‘significant energy
number of regulatory alternatives and substantial direct effects on the States, action’’ as defined in Executive Order
adopt the least costly, most cost- on the relationship between the national 13211, ‘‘Actions Concerning Regulations
effective, or least burdensome government and the States, or on the That Significantly Affect Energy Supply,
alternative that achieves the objectives distribution of power and Distribution, or Use’’ (66 FR 28355 (May
of the rule. The provisions of section responsibilities among the various 22, 2001)) because it is not likely to
205 do not apply when they are levels of government, as specified in have a significant adverse effect on the
inconsistent with applicable law. Executive Order 13132, because State supply, distribution, or use of energy.
Moreover, section 205 allows EPA to and local governments do not own or These rule requirements will have
adopt an alternative other than the least operate any sources that would be energy effects due to the energy
costly, most cost-effective, or least subject to the proposed rule requirements for the control devices
burdensome alternative if the amendments. Thus, Executive Order required for new sources. We estimate
Administrator publishes with the final 13132 does not apply to the final rule the additional electrical demand to be
rule an explanation why that alternative amendments. 15 million kWhr per year and the
natural gas use to be 270 billion cubic use with impingers, as in EPA method, agency promulgating the rule must
feet by the end of the fifth year. We do although the standard describes submit a rule report, which includes a
not consider these energy impacts to be procedures for the use of impingers. copy of the rule, to each House of the
significant. This European standard is validated for Congress and to the Comptroller General
the use of fritted bubblers only and of the United States. EPA will submit a
I. National Technology Transfer and
requires the use of a side (split) stream report containing this rule and other
Advancement Act
arrangement for isokinetic sampling required information to the U.S. Senate,
Section 12(d) of the National because of the low sampling rate of the the U.S. House of Representatives, and
Technology Transfer and Advancement bubblers (up to 3 liters per minute, the Comptroller General of the United
Act (NTTAA) of 1995 (Pub. L. 104–113, maximum). Also, only two bubblers (or States prior to publication of the rule in
Section 12(d), 15 U.S.C. 272 note) impingers) are required by EN 13211, the Federal Register. A major rule
directs EPA to use voluntary consensus whereas EPA method requires the use of cannot take effect until 60 days after it
standards (VCS) in its regulatory six impingers. In addition, EN 13211 is published in the Federal Register.
activities, unless to do so would be does not include many of the quality This action is not a ‘‘major rule’’ as
inconsistent with applicable law or control procedures of EPA methods, defined by 5 U.S.C. 804(2). This rule
otherwise impractical. The VCS are especially for the use and calibration of will be effective on December 20, 2006.
technical standards (e.g., materials temperature sensors and controllers,
specifications, test methods, sampling sampling train assembly and List of Subjects in 40 CFR Part 63
procedures, and business practices) that disassembly, and filter weighing. Environmental protection,
are developed or adopted by VCS The voluntary consensus standard Administrative practice and procedure,
bodies. The NTTAA directs EPA to CAN/CSA Z223.26–M1987, Air pollution control, Hazardous
provide Congress, through OMB, ‘‘Measurement of Total Mercury in Air substances, and Reporting and
explanations when the Agency does not Cold Vapour Atomic Absorption recordkeeping requirements.
use available and applicable VCS. Spectrophotometeric Method,’’ is not
This final rule involves technical Dated: December 8, 2006.
acceptable as an alternative to EPA
standards. EPA cites EPA Method 29 of Stephen L. Johnson,
Method 29 (for mercury). This standard
40 CFR part 60 for measurement of is not acceptable because of the lack of Administrator.
mercury emissions in stack gases for detail in quality control. Specifically, ■ For the reasons stated in the preamble,
new cement kilns. CAN/CSA Z223.26 does not include title 40, chapter I, part 63 of the Code
Consistent with the NTTAA, EPA specifications for the number of of Federal Regulations is amended as
conducted searches to identify calibration samples to be analyzed, follows:
voluntary consensus standards in procedures to prevent carryover from
addition to these EPA methods. The one sample to the next, and procedures PART 63—[Amended]
search and review results are in the for subtraction of the instrument ■ 1. The authority citation for part 63
docket for this rule. response to calibration blank as in EPA
One voluntary consensus standard continues to read as follows:
method. Also, CAN/CSA Z223.26 does
was identified as an acceptable not require that the calibration curve be Authority: 42 U.S.C. 7401, et seq.
alternative to an EPA test method for the forced through or close to zero (or a
purposes of the final rule. The voluntary Subpart LLL—[Amended]
point no further than ±2 percent of the
consensus standard ASTM D6784–02, recorder full scale) as in EPA method. ■ 2. § 63.1342 is revised to read as
‘‘Standard Test Method for Elemental, Also, CAN/CSA Z223.26 does not follows:
Oxidized, Particle-Bound and Total include a procedure to assure that two
Mercury Gas Generated from Coal-Fired consecutive peak heights agree within 3 § 63.1342 Standards: General.
Stationary Sources (Ontario Hydro percent of their average value and that Table 1 to this subpart provides cross
Method),’’ is an acceptable alternative to the peak maximum is greater than 10 references to the 40 CFR part 63, subpart
EPA Method 29 (portion for mercury percent of the recorder full scale, as in A, general provisions, indicating the
only) as a method for measuring EPA methods. CAN/CSA Z223.26 does applicability of the general provisions
mercury. not include instructions for a blank and requirements to subpart LLL.
The search for emissions a standard to be run at least every five ■ 3. Section 63.1343 is revised to read
measurement procedures identified two samples, and specifications for the peak as follows:
other voluntary consensus standards. height of the blank and the standard as
EPA determined that these two in EPA method. § 63.1343 Standards for kilns and in-line
standards identified for measuring Section 63.1349 to subpart LLL of this kiln/raw mills.
emissions of the HAP or surrogates rule lists the testing methods included (a) General. The provisions in this
subject to emission standards in this in the regulation. Under § 63.7(f) and section apply to each kiln, each in-line
rule were impractical alternatives to § 63.8(f) of Subpart A of the General kiln/raw mill, and any alkali bypass
EPA test methods for the purposes of Provisions, a source may apply to EPA associated with that kiln or in-line kiln/
this rule. Therefore, EPA does not for permission to use alternative test raw mill. All gaseous, mercury and
intend to adopt these standards for this methods or alternative monitoring D/F emission limits are on a dry basis,
purpose. The reasons for the requirements in place of any required corrected to 7 percent oxygen. All total
determinations for the two methods are testing methods, performance hydrocarbon (THC) emission limits are
discussed below. specifications, or procedures. measured as propane. The block
The voluntary consensus standard EN averaging periods to demonstrate
13211:2001, ‘‘Air Quality—Stationary J. Congressional Review Act compliance are hourly for 20 ppmv total
Source Emissions—Determination of the The Congressional Review Act, 5 hydrocarbon (THC) limits and monthly
Concentration of Total Mercury,’’ is not U.S.C. 801 et seq., as added by the Small for the 50 ppmv THC limit.
acceptable as an alternative to the Business Regulatory Enforcement (b) Existing kilns located at major
mercury portion of EPA Method 29 Fairness Act of 1996, generally provides sources. No owner or operator of an
primarily because it is not validated for that before a rule may take effect, the existing kiln or an existing kiln/raw mill
located at a facility that is a major (5) Contain mercury from the main bypass in excess of 41 µg/dscm if the
source subject to the provisions of this exhaust of the kiln, or main exhaust of source is a new or reconstructed source
subpart shall cause to be discharged into the in-line kiln/raw mill, or the alkali that commenced construction after
the atmosphere from these affected bypass in excess of 41µg/dscm if the December 2, 2005. As an alternative to
sources, any gases which: source is a new or reconstructed source meeting the 41 µg/dscm standard you
(1) Contain particulate matter (PM) in that commenced construction after may route the emissions through a
excess of 0.15 kg per Mg (0.30 lb per December 2, 2005. As an alternative to packed bed or spray tower wet scrubber
ton) of feed (dry basis) to the kiln. When meeting the 41 µg/dscm standard you with a liquid-to-gas (l/g) ratio of 30
there is an alkali bypass associated with may route the emissions through a gallons per 1000 actual cubic feet per
a kiln or in-line kiln/raw mill, the packed bed or spray tower wet scrubber minute (acfm) or more and meet a site-
combined particulate matter emissions with a liquid-to-gas (l/g) ratio of 30 specific emissions limit based on the
from the kiln or in-line kiln/raw mill gallons per 1000 actual cubic feet per measured performance of the wet
and the alkali bypass are subject to this minute (acfm) or more and meet a site- scrubber.
emission limit. specific emissions limit based on the ■ 4. Section 63.1344 is amended as
(2) Exhibit opacity greater than 20 measured performance of the wet follows:
percent. scrubber. ■ a. Revising paragraphs (c) through (e);
(3) Contain D/F in excess of: (d) Existing kilns located at area ■ b. Adding paragraphs (f) through (i).
(i) 0.20 ng per dscm (8.7 × 10–11 gr per sources. No owner or operator of an
dscf) (TEQ); or existing kiln or an existing in-line kiln/ § 63.1344 Operating limits for kilns and in-
(ii) 0.40 ng per dscm (1.7 × 10–10 gr raw mill located at a facility that is an line kiln/raw mills.
per dscf) (TEQ) when the average of the area source subject to the provisions of * * * * *
performance test run average this subpart shall cause to be discharged (c) The owner or operator of an
temperatures at the inlet to the into the atmosphere from these affected affected source subject to a mercury,
particulate matter control device is sources any gases which: THC or D/F emission limitation under
204 °C (400 °F) or less. (1) Contain D/F in excess of 0.20 ng § 63.1343 that employs carbon injection
(c) Reconstructed or new kilns located per dscm (8.7 × 10–11 gr per dscf) (TEQ); as an emission control technique must
at major sources. No owner or operator or operate the carbon injection system in
of a reconstructed or new kiln or (2) Contain D/F in excess of 0.40 ng accordance with paragraphs (c)(1) and
reconstructed or new inline kiln/raw per dscm (1.7 × 10–10 gr per dscf) (TEQ) (c)(2) of this section.
mill located at a facility which is a when the average of the performance (1) The three-hour rolling average
major source subject to the provisions of test run average temperatures at the activated carbon injection rate shall be
this subpart shall cause to be discharged inlet to the particulate matter control equal to or greater than the activated
into the atmosphere from these affected device is 204 °C (400 °F) or less. carbon injection rate determined in
sources any gases which: (e) New or reconstructed kilns located accordance with § 63.1349(b)(3)(vi).
(1) Contain particulate matter in at area sources. No owner or operator of (2) The owner or operator shall either:
excess of 0.15 kg per Mg (0.30 lb per a new or reconstructed kiln or new or (i) Maintain the minimum activated
ton) of feed (dry basis) to the kiln. When reconstructed in-line kiln/raw mill carbon injection carrier gas flow rate, as
there is an alkali bypass associated with located at a facility that is an area source a three-hour rolling average, based on
a kiln or in-line kiln/raw mill, the subject to the provisions of this subpart the manufacturer’s specifications. These
combined particulate matter emissions shall cause to be discharged into the specifications must be documented in
from the kiln or in-line kiln/raw mill atmosphere from these affected sources the test plan developed in accordance
and the bypass stack are subject to this any gases which: with § 63.7(c), or
emission limit. (1) Contain D/F in excess of: (ii) Maintain the minimum activated
(2) Exhibit opacity greater than 20 (i) 0.20 ng per dscm (8.7 × 10–11 gr per carbon injection carrier gas pressure
percent. dscf) (TEQ; or drop, as a three-hour rolling average,
(3) Contain D/F in excess of: (ii) 0.40 ng per dscm (1.7 × 10–10 gr based on the manufacturer’s
(i) 0.20 ng per dscm (8.7 × 10–11 gr per per dscf) (TEQ) when the average of the specifications. These specifications
dscf) (TEQ); or performance test run average must be documented in the test plan
(ii) 0.40 ng per dscm (1.7 × 10–10 gr temperatures at the inlet to the developed in accordance with § 63.7(c).
per dscf) (TEQ) when the average of the particulate matter control device is (d) Except as provided in paragraph
performance test run average 204 °C (400 °F) or less. (e) of this section, the owner or operator
temperatures at the inlet to the (2) Contain total hydrocarbons (THC), of an affected source subject to a
particulate matter control device is from the main exhaust of the kiln, or mercury, THC or D/F emission
204 °C (400 °F) or less. main exhaust of the in-line kiln/raw limitation under § 63.1343 that employs
(4) Contain total hydrocarbons (THC), mill, in excess of 20 ppmv if the source carbon injection as an emission control
from the main exhaust of the kiln, or is a new or reconstructed source that technique must specify and use the
main exhaust of the in-line kiln/raw commenced construction after brand and type of activated carbon used
mill, in excess of 20 ppmv if the source December 2, 2005. As an alternative to during the performance test until a
is a new or reconstructed source that meeting the 20 ppmv standard you may subsequent performance test is
commenced construction after demonstrate a 98 percent reduction of conducted, unless the site-specific
December 2, 2005. As an alternative to THC emissions from the exit of the kiln performance test plan contains
meeting the 20 ppmv standard you may to discharge to the atmosphere. If the documentation of key parameters that
demonstrate a 98 percent reduction of source is a greenfield kiln that affect adsorption and the owner or
THC emissions from the exit of the kiln commenced construction on or prior to operator establishes limits based on
to discharge to the atmosphere. If the December 2, 2005, then the THC limit those parameters, and the limits on
source is a greenfield kiln that is 50 ppmv. these parameters are maintained.
commenced construction on or prior to (3) Contain mercury from the main (e) The owner or operator of an
December 2, 2005, then the THC limit exhaust of the kiln, or main exhaust of affected source subject to a D/F, THC, or
is 50 ppmv. the in-line kiln/raw mill, or the alkali mercury emission limitation under
§ 63.1343 that employs carbon injection atmosphere. If the source is a greenfield percent reduction in THC across the
as an emission control technique may dryer constructed on or prior to control device using the performance
substitute, at any time, a different brand December 2, 2005, then the THC limit test requirements in 40 CFR part 63,
or type of activated carbon provided is 50 ppmv, on a dry basis corrected to subpart SS.
that the replacement has equivalent or 7 percent oxygen. (5) The owner or operator of a kiln or
improved properties compared to the (b) New or reconstructed raw in-line kiln/raw mill subject to the
activated carbon specified in the site- materials dryers located at a facility that 41 µg/dscm mercury standard shall
specific performance test plan and used is an area source cannot discharge to the demonstrate compliance using EPA
in the performance test. The owner or atmosphere any gases which contain Method 29 of 40 CFR part 60. ASTM
operator must maintain documentation THC in excess of 20 ppmv, on a dry D6784–02, Standard Test Method for
that the substitute activated carbon will basis as propane corrected to 7 percent Elemental, Oxidized, Particle-Bound
provide the same or better level of oxygen if the source commenced and Total Mercury Gas Generated from
control as the original activated carbon. construction after December 2, 2005. As Coal-Fired Stationary Sources (Ontario
(f) Existing kilns and in-line kilns/raw an alternative to the 20 ppmv standard, Hydro Method), is an acceptable
mills must implement good combustion you may demonstrate a 98 percent alternative to EPA Method 29 (portion
practices (GCP) designed to minimize reduction in THC emissions from the for mercury only). If the kiln has an in-
THC from fuel combustion. GCP include exit of the raw materials dryer to line raw mill, you must demonstrate
training all operators and supervisors to discharge to the atmosphere. If the compliance with both raw mill off and
operate and maintain the kiln and source is a greenfield dryer constructed raw mill on. You must record the hourly
calciner, and the pollution control on or prior to December 2, 2005, then recycle rate of CKD during both test
systems in accordance with good the THC limit is 50 ppmv, on a dry basis conditions and calculate an average
engineering practices. The training shall corrected to 7 percent oxygen. hourly rate for the three test runs for
include methods for minimizing excess ■ 6. Section 63.1349 is amended as each test condition.
emissions. follows: * * * * *
(g) No kiln and in-line kiln/raw mill ■ a. By revising paragraph (b)(4);
■ 7. Section 63.1350 is amended as
may use as a raw material or fuel any ■ b. By adding paragraph (b)(5);
■ c. By removing paragraph (f).
follows:
fly ash where the mercury content of the
■ a. Revising paragraphs (g), (h) and (n);
fly ash has been increased through the
§ 63.1349 Performance Testing and
use of activated carbon, or any other Requirements. ■ b. Adding paragraphs (o) and (p).
sorbent unless the facility can
demonstrate that the use of that fly ash * * * * *
(b) * * * § 63.1350 Monitoring requirements.
will not result in an increase in mercury (4)(i) The owner or operator of an * * * * *
emissions over baseline emissions (i.e. affected source subject to limitations on (g) The owner or operator of an
emissions not using the fly ash). The emissions of THC shall demonstrate affected source subject to an emissions
facility has the burden of proving there initial compliance with the THC limit limitation on D/F, THC or mercury
has been no emissions increase over by operating a continuous emission emissions that employs carbon injection
baseline. monitor in accordance with as an emission control technique shall
(h) All kilns and in-line kilns/raw comply with the monitoring
Performance Specification 8A of
mills must remove (i.e. not recycle to requirements of paragraphs (f)(1)
appendix B to part 60 of this chapter.
the kiln) from the kiln system sufficient through (f)(6) and (g)(1) through (g)(6) of
The duration of the performance test
cement kiln dust to maintain the desired this section to demonstrate continuous
shall be three hours, and the average
product quality. compliance with the D/F, THC or
THC concentration (as calculated from
(i) New and reconstructed kilns and
the one-minute averages) during the mercury emissions standard.
in-line kilns/raw mills must not exceed
three-hour performance test shall be (1) Install, operate, calibrate and
the average hourly CKD recycle rate
calculated. The owner or operator of an maintain a continuous monitor to record
measured during mercury performance
in-line kiln/raw mill shall demonstrate the rate of activated carbon injection.
testing. Any exceedance of this average
initial compliance by conducting The accuracy of the rate measurement
hourly rate is considered a violation of
separate performance tests while the device must be ±1 percent of the rate
the standard.
raw mill of the in-line kiln/raw mill is being measured.
■ 5. Section 63.1346 is revised to read
under normal operating conditions and (2) Verify the calibration of the device
as follows:
while the raw mill of the in-line kiln/ at least once every three months.
§ 63.1346 Standards for new or raw mill is not operating. (3) The three-hour rolling average
reconstructed raw material dryers. (ii) The owner or operator of an activated carbon injection rate shall be
(a) New or reconstructed raw material affected source subject to limitations on calculated as the average of 180
dryers located at facilities that are major emissions of THC who elects to successive one-minute average activated
sources can not discharge to the demonstrate compliance with the carbon injection rates.
atmosphere any gases which: alternative THC emission limit of 98 (4) Periods of time when one-minute
(1) Exhibit opacity greater than ten percent weight reduction must averages are not available shall be
percent, or demonstrate compliance by also ignored when calculating three-hour
(2) Contain THC in excess of operating a continuous emission rolling averages. When one-minute
20 ppmv, on a dry basis as propane monitor in accordance with averages become available, the first one-
corrected to 7 percent oxygen if the Performance Specification 8A of minute average is added to the previous
source commenced construction after appendix B to part 60 at the inlet to the 179 values to calculate the three-hour
December 2, 2005. As an alternative to THC control device of the kiln, inline rolling average.
the 20 ppmv standard, you may kiln raw mill, or raw materials dryer in (5) When the operating status of the
demonstrate a 98 percent reduction in the same manner as prescribed in raw mill of the in-line kiln/raw mill is
THC emissions from the exit of the raw paragraph (i) above. Alternately, you changed from off to on, or from on to
materials dryer to discharge to the may elect to demonstrate a 98 weight off, the calculation of the three-hour
rolling average activated carbon (4) For new facilities complying with June 14, 1999, or upon startup of
injection rate must begin anew, without the 20 ppmv THC emissions limit, any operations, whichever is later.
considering previous recordings. hourly average THC concentration in (c) The compliance date for an
(6) The owner or operator must any gas discharged from a raw material existing source to meet the requirements
install, operate, calibrate and maintain a dryer, the main exhaust of a greenfield of GCP for THC is December 20, 2007.
continuous monitor to record the kiln, or the main exhaust of a kiln or in- (d) The compliance date for a new
activated carbon injection system carrier line kiln/raw mill, exceeding 20 ppmvd, source which commenced construction
gas parameter (either the carrier gas flow reported as propane, corrected to seven after December 2, 2005, and before
rate or the carrier gas pressure drop) percent oxygen, is a violation of the December 20, 2006 to meet the THC
established during the mercury, THC or standard. emission limit of 20 ppmv/98 percent
D/F performance test in accordance * * * * * reduction or the mercury standard of
with paragraphs (g)(6)(i) through 41 µg/dscm or a site-specific standard
(n) Any kiln or kiln/in-line raw mill
(g)(6)(iii) of this section. based on application of a wet scrubber
(i) The owner or operator shall install, using a control device (other then ACI)
to comply with a mercury emissions will be December 21, 2009.
calibrate, operate and maintain a device
to continuously monitor and record the limit or equipment standard will ■ 9. Section 63.1355 is amended by
parameter value. monitor the control device parameters adding paragraphs (d), (e) and (f) to read
(ii) The owner or operator must as specified in 40 CFR part 63 subpart as follows:
calculate and record three-hour rolling SS.
§ 63.1355 Recordkeeping requirements.
averages of the parameter value. (o) For kilns and in-line kilns/raw
(iii) Periods of time when one-minute mills complying with the requirements * * * * *
averages are not available shall be in Section 63.1344(g), each owner or (d) You must keep annual records of
ignored when calculating three-hour operator must obtain a certification from the amount of CKD which is removed
rolling averages. When one-minute the supplier for each shipment of fly ash from the kiln system and either
averages become available, the first one- received to demonstrate that the fly ash disposed of as solid waste or otherwise
minute average shall be added to the was not derived from a source in which recycled for a beneficial use outside of
previous 179 values to calculate the the use of activated carbon, or any other the kiln system.
three-hour rolling average. sorbent, is used as a method of mercury (e) You must keep records of the
(h) The owner or operator of an emissions control. The certification amount of CKD recycled on an hourly
affected source subject to a limitation on shall include the name of the supplier basis.
THC emissions under this subpart shall and a signed statement from the (f) You must keep records of all fly
comply with the monitoring supplier confirming that the fly ash was ash supplier certifications as required
requirements of paragraphs (h)(1) not derived from a source in which the by § 63.1350(o).
through (h)(3) of this section to use of activated carbon, or any other ■ 10. Section 63.1356 is amended by
demonstrate continuous compliance sorbent, is used as a method of emission revising paragraph (a) to read as follows:
with the THC emission standard: control.
(1) The owner or operator shall (p) If the facility opts to use a fly ash § 63.1356 Exemption from new source
install, operate and maintain a THC performance standards.
derived from a source in which the use
continuous emission monitoring system of activated carbon, or any other (a) Except as provided in paragraphs
in accordance with Performance sorbent, is used as a method of mercury (a)(1) and (2) of this section, any
Specification 8A, of appendix B to part emissions control and demonstrate that affected source subject to the provisions
60 of this chapter and comply with all the use of this fly ash does not increase of this subpart is exempt from any
of the requirements for continuous mercury emissions, they must obtain otherwise applicable new source
monitoring systems found in the general daily fly ash samples, composites performance standard contained in
provisions, subpart A of this part. monthly, and analyze the samples for subpart F or subpart OOO of part 60 of
(2) The owner or operator is not this chapter.
mercury.
required to calculate hourly rolling (1) Kilns and in-line kiln/raw mills, as
averages in accordance with section 4.9 ■ 8. Section 63.1351 is revised to read
as follows: applicable, under 40 CFR 60.60(b),
of Performance Specification 8A if they located at area sources are subject to PM
are only complying with the 50 ppmv § 63.1351 Compliance dates. and opacity limits and associated
THC emissions limit. reporting and recordkeeping, under 40
(3) For facilities complying with the (a) Except as noted in paragraph (c)
below, the compliance date for an CFR part 60, subpart F.
50 ppmv THC emissions limit, any
thirty-day block average THC owner or operator of an existing affected (2) Greenfield raw material dryers, as
concentration in any gas discharged source subject to the provisions of this applicable under 40 CFR 60.60(b),
from a greenfield raw material dryer, the subpart is June 14, 2002. located at area sources, are subject to
main exhaust of a greenfield kiln, or the (b) Except as noted in paragraph (d) opacity limits and associated reporting
main exhaust of a greenfield in-line below, the compliance date for an and recordkeeping under 40 CFR part
kiln/raw mill, exceeding 50 ppmvd, owner or operator of an affected source 60, subpart F.
reported as propane, corrected to seven subject to the provisions of this subpart * * * * *
percent oxygen, is a violation of the that commences new construction or [FR Doc. E6–21405 Filed 12–19–06; 8:45 am]
standard. reconstruction after March 24, 1998, is BILLING CODE 6560–50–P

Rule: Air Pollutants, Hazardous National Emission Standards: Portland Cement Manufacturing Industry

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Rule: Air Pollutants, Hazardous National Emission Standards: Portland Cement Manufacturing Industry

Hochgeladen von

Copyright:

Verfügbare Formate

Wednesday,

December 20, 2006

TABLE 1.—REGULATED ENTITIES TABLE

NESHAP is available only by filing a Emissions Significantly Affect Energy Supply,

should be to remove accumulated on control technology. Raw Materials in docket EPA–HQ–

TABLE 1.—PACKED BED SCRUBBER—COSTS AND EMISSION REDUCTIONS

600,000 1,542,000 297 0.11 16.8–147

representatie of the percent reduction in the figure in beyond-the-floor analyses as an upper

emissions and high costs of reducing Comment: Several comments support

review mandated by section 112(d)(6). There is no based on alkaline scrubbers specifically

carcinogenic in either humans or that there is no justification for an HCl

Industry ............................................................................................................ 4,159 $679,105 $161,672 $840,777

Das könnte Ihnen auch gefallen