National Emission Standards for Hazardous Air Pollutants
(NESHAP) FOR Lime Manufacturing Background Information
Document - Vol II
PUBLIC COMMENTS AND RESPONSES
August 2003
Emission Standards Division
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
-------
TABLE OF CONTENTS
Section Page
1.0 Summary 1-1
1.1 Table 1-1 List of Commenters 1-2
2.0 Rule Changes Since Proposal 2-1
2.1 Emission Limits 2-1
2.2 Monitoring Requirements 2-1
2.3 Emission Testing and Reports 2-2
2.4 Applicability 2-2
2.5 Averaging Periods for Monitoring 2-2
2.6 Averaging Kiln Emissions 2-3
3.0 Comment Summaries and EPA Responses 3-1
3.1 Subcategory for Scrubber-Equipment Kilns 3-1
3.2 PM Emission Limit for Scrubber-Equipment Kilns 3-17
3.3 PM Emission Limit for New Kilns 3-18
3.4 PM Emission Limit - General 3-19
3.5 Monitoring Requirements for Scrubber-Equipped Kilns 3-22
3.6 Monitoring Requirements for Kilns w/ Dry Pollution Control Devices 3-23
3.7 Operating Parameters 3-27
3.8 Opacity Limit 3-29
3.9 Stone Handling Operations 3-31
3.10 Test Methods 3-32
3.11 Economic Impact Analysis 3-36
3.12 Risk Analysis 3-42
3.13 Applicability 3-43
3.14 General Comments 3-44
l
-------
1.0 SUMMARY
On December 20, 2002, the U.S. Environmental Protection Agency (EPA) proposed national
emission standards for hazardous air pollutant (HAP) emissions from lime manufacturing plants
located at major source facilities (67 FR 78046). These proposed standards implemented
section 112(d) of the Clean Air Act as amended in 1990 (CAA). There were 24 comment letters
on the proposal (see Table 1-1), and the commenters consisted of trade associations,
manufacturers, and engineering firms, and the general public. Summaries of the comments, and
the EPA's responses, are presented in this background information document (BID Volume II).
This summary of comments and responses served as the basis for the revisions made to the rule
between proposal and promulgation. Besides summarizing the comments and responses, this
document also presents a summary of the rule revisions. This document supplements the
information used to develop the proposed rule, which is the docket for this rulemaking.
1-1
-------
Table 1-1 List of Commenters
(See Docket No. OAR-2002-0052 and Docket A-95-41 -National Emission Standards for
Hazardous Air Pollutants for Lime Manufacturing Plants for the comments)
Document
Number
Date Rcvd in
Docket
Cementer, Addressee. Title or Description,
etc.
Date of
Document
IV-D-01
02-13-03
John S. Morawetz, Center Director, ICWUC
Center for Worker Health and Safety
Education, Cincinnati, OH.
02-14-03
IV-D-02
02-18-03
A. J. Paris, President, Huron Lime Inc.
02-14-03
IV-D-03
02-18-03
William M. Brant, Director, Miami-Dade
Water and Sewer Department, Miami, FL.
02-18-03
IV-D-04
02-18-03
Arline M. Seeger, Executive Director, National
Lime Association, Arlington, VA.
02-18-03
IV-D-05
02-18-03
Tony C. Panchyshyn, Senior Environmental
Engineer, Graymont Western US Inc., Salt
Lake City, UT.
02-14-03
IV-D-06
02-18-03
J. Oscar Robinson, General Partner, Austin
White Lime Co., Austin, TX.
02-16-03
IV-D-07
02-18-03
Dana Stone, Vice President, Operations, Cutler-
Magner Company, Duluth, MN.
02-13-03
IV-D-08
02-18-03
Harold W. Robbins, President, Bison
Engineering Inc., Helena, MT.
02-14-03
IV-D-09
02-18-03
James R. Carson, Senior Staff Engineer,
Environmental Affairs, Ispat Inland Inc., East
Chicago, IN.
02-17-03
IV-D-10
02-18-03
Jeffrey P. LaCosse, President and Principal
Scientist, Spectral Insights LLC, Durham, NC.
02-17-03
IV-D-11
02-18-03
Steve Raffensperger, Vice President-General
Manager, MidAmerica Division, Martin
Marietta Aggregates, Mason, OH.
02-17-03
IV-D-12
02-18-03
Lucie Macalister, Environmental Engineer,
Linwood Mining and Minerals Corp.,
Davenport, IA.
02-14-03
1-2
-------
Table 1-1 List of Commenters
(See Docket No. OAR-2002-0052 and Docket A-95-41 -National Emission Standards for
Hazardous Air Pollutants for Lime Manufacturing Plants for the comments)
Document
Number
Date Rcvd in
Docket
Cementer, Addressee. Title or Description,
etc.
Date of
Document
IV-D-13
02-18-03
Robert J. Bartosh, Senior Vice President &
COO, Dakota Coal Company, Bismarck, ND.
02-13-03
IV-D-14
02-18-03
Louis Wolfe, Private Citizen, Newark, DE.
02-15-03
IV-D-15
02-18-03
Stephen A. Loeschner, Private Citizen, Fort
Wayne, IN.
02-18-03
IV-D-16
02-18-03
Ron Downey, Director, Environmental
Compliance, LWB Refractories, York, PA.
02-18-03
IV-D-17
02-18-03
Darin Klewsaat, Assistant Operations Director,
Southern Lime Company, Calera, AL.
02-21-03
IV-D-18
02-11-03
Lecia Craft, Vice President of Environment,
Pete Lien & Sons Inc., Rapid City, SD.
02-11-03
IV-D-19
02-10-03
Craig Clapsaddle, Mechanical Systems Inc.
(MSI), Raleigh, NC.
02-10-03
IV-D-20
02-17-03
Fred Nast, Chief Executive Officer, Western
Lime, West Bend, WI.
02-17-03
IV-D-21
02-18-03
Spencer C. Stinson, General Manager, Global
Stone Chemstone Corp., Strasburg, VA.
02-18-03
IV-D-22
02-17-03
Johnney G. Bowers, Vice President,
Manufacturing, United States Lime & Minerals
Inc., Dallas, TX.
02-12-03
IV-D-23
02-15-03
Comment entitled " CLEAN AIR IS
ESSENTIAL." submitted by
JeanPublic@yahoo.com.
02-15-03
IV-G-01
02-24-03
J. Steven Castleberry, Environmental Affairs
Manager, Mississippi Lime, Alton, IL. (This
document contains Confidential Business
Information. Please contact the project
officer if you need to review.)
02-18-03
1-3
-------
Table 1-1 List of Commenters
(See Docket No. OAR-2002-0052 and Docket A-95-41 -National Emission Standards for
Hazardous Air Pollutants for Lime Manufacturing Plants for the comments)
1-4
-------
2.0 Rule Changes Since Proposal
2.1 Emission Limits
We proposed a particulate matter (PM) standard (as a surrogate for non-mercury
hazardous air pollutant [HAP] metals) of 0.12 pounds PM per ton stone feed (lb/tsf) reflecting
the performance of dry pollution control systems (baghouses). We also solicited comment on
having a separate PM standard of 0.60 lb/tsf for kilns controlled with wet scrubbers. In the final
rule, we have decided to adopt these two different standards for PM emissions from existing lime
kilns. We are also indicating that existing kilns subject to the 0.60 lb/tsf PM emission limit are
not to be included in any averaging scheme for demonstrating compliance with a PM standard.
2.2 Monitoring Requirements
In the proposed NESHAP, we required facilities using wet scrubbers to monitor scrubber
pressure drop and liquid flow rate. We have written the final NESHAP to explicitly state that
alternative monitoring procedures are allowed under the procedures described in 40 CFR 63.8(f).
However, we do not delegate that authority.
The proposed NESHAP stated that you must install, operate, and maintain a continuous
opacity monitoring system (COMS) as required by 40 CFR part 63, subpart A, General
Provisions, and according to PS-1 in Appendix B to 40 CFR part 60. We have stated in the rule
that COMS installed, relocated, or substantially refurbished after February 6, 2001, must meet
the requirements of PS-1 as revised on August 10, 2000. Any COMS installed on or before
February 6, 2001, should continue to meet the requirements in effect at the time of installation
unless specifically required by the local regulatory agency to re-certify the COMS in question.
In the proposed NESHAP, we required you to monitor the performance of fabric filters
(FF) with either a COMS or a PM detector. In the final NESHAP, we are allowing existing
facilities to monitor FF performance using daily EPA Method 9, in Appendix A to 40 CFR part
60, visible emission readings if the facility has a positive pressure FF with multiple stacks, or if
it is infeasible to install a COMS in accordance with PS-1 in Appendix B to 40 CFR part 60.
In the proposed NESHAP, we allowed three alternatives for monitoring electrostatic
precipitator (ESP) performance. These were a COMS, a PM detector, or monitoring ESP voltage
and current. In the final NESHAP, we are allowing only two alternatives, a COMS or a PM
detector. There are no requirements to establish ESP voltage and current operating limits.
2-1
-------
In the proposed NESHAP, we specified that EPA Method 9 in Appendix A to 40 CFR
part 60 should be used to determine opacity from fugitive emissions. We have retained this
requirement in the final NESHAP, but we have added additional requirements on how EPA
method 9 in Appendix A to 40 CFR part 60 should be implemented to determine fugitive visible
emissions. This language was taken directly from 40 CFR 60.675(c)(1).
In the proposed NESHAP, §63.7120(b) could be interpreted to imply that processed stone
handling (PSH) operations must be continuously monitored. In the final NESHAP, PSH
operations are subject to monthly (not continuous) visible emission testing.
2.3 Emission Testing and Reports
In the proposed NESHAP, we required that lime kiln emission testing be conducted at the
highest production level reasonably expected to occur. In the final NESHAP, we require that
lime kilns be tested under representative operating conditions.
In the proposed NESHAP, we required reporting of deviations from operating, visible
emissions, and opacity limits, including those deviations that occur during periods of startup,
shutdown, or malfunction. In the final NESHAP, we require that reports are to be made in
accordance with 40 CFR 63.10(d).
In the proposed NESHAP, we required testing of all kilns in order to claim area source
status. In the final NESHAP, we have included a provision that allows the permitting authority
to determine if idled kilns must be tested, and also to determine whether all kilns that use
identical feed materials, fuels, and emission controls must still all be tested.
2.4 Applicability
In the proposed NESHAP, the raw material storage bin was the first emission unit in the
sequence of lime manufacturing that was part of the affected source. Materials processing
operations between the storage bin and the kiln were also covered. In the final NESHAP,
material stockpiles prior to the processed stone storage bin are not covered, open processed stone
piles are not covered, storage bins are defined as manmade enclosures, and use the term
processed stone handling operations instead of materials processing operations.
In the proposed NESHAP, we included as an affected source lime kilns that produced
2-2
-------
lime product from any calcareous substance. In the final NESHAP, we have excluded lime kilns
that produce lime from water softening sludge that contain calcium carbonate.
In the proposed NESHAP, we excluded materials handling operations associated with
lime product. In the final NESHAP, we have specifically stated that nuisance dust collectors are
part of lime product handling systems and, therefore, are not part of the affected source.
In the proposed NESHAP, we defined the affected source as the collection of all of the
lime kilns, lime coolers and materials processing operations. We noted that this language could
be misinterpreted to imply that a new lime kiln erected at an existing lime manufacturing plant
would be considered existing, not new. In the final NESHAP, we have written the language in
40 CFR
63.7082 to make our intent clear. New lime kilns, whether or not they are built at an existing
lime manufacturing plant, must meet the PM emission limits for new sources.
2.5 Averaging Periods for Monitoring
In the proposed NESHAP, we required that facilities use rolling 3-hour averages to show
compliance with wet scrubber operating limits. We noted that in the proposed rule, we did not
clearly state how to calculate the rolling average. Based on compliance requirements of other
NESHAP, we determined that a rolling average was not necessary to ensure compliance, but did
increase the complexity of the average calculation and record keeping process. Therefore, in the
final NESHAP, we require block 3-hour averages instead of rolling 3-hour averages, which is
consistent with the requirement to use block averaging required for ESP that choose to monitor
using COMS.
2.6 Averaging Kiln Emissions
In the proposed NESHAP, we allowed averaging among all lime kilns and coolers at
existing sources, and all new lime kilns and coolers at new sources, but did not allow averaging
of existing and new lime kilns and coolers together. In addition, the averaging provisions and
equations applied whether or not the facility desired to average. We have written the final
NESHAP to state that each individual new lime kiln and its associated cooler must meet a 0.10
lb/tsf PM emission limit, and each individual existing lime kilns and its associated cooler must
meet a 0.12 lb/tsf PM emission limit. Averaging is optional, so that if each individual kiln meets
its emission limit, averaging is not required. The exception to this is for existing kilns which are
subject to the 0.60 lb/tsf PM emission limit. These kilns are not eligible for averaging.
If the lime manufacturing plant has multiple kilns and wants to average kilns together to
2-3
-------
meet the PM emission limit, this is allowed (with one limitation discussed below, and the
exception for kilns subject to the 0.60 lb/tsf PM emission limit noted above) and the averaging
equations in the final rule must be used. However, in no case may a new kiln exceed a 0.10
lb/tsf emission limit. Where there are both new and existing lime kilns at a facility, then the PM
emission limit will be an average of the existing and new kiln PM emissions limits, weighted by
the annual actual production rates of the individual kilns. We believe that allowing averaging is
appropriate here because of the identity of the units (kilns and coolers in all cases), and the
emissions (same HAP in same type of emissions, since all emissions result from kilns and
coolers). Averaged emissions under these circumstances would, thus, still reflect MACT for the
affected source. The averaging provisions are included in the final NESHAP as a result of the
recommendations of the Small Business Advocacy Panel convened as required by section 609(b)
of the Regulatory Flexibility Act (RFA) and improves the compliance flexibility options for
small businesses, which is the intent of the RFA.
The only limitations we are requiring on averaging are: 1) any new kiln, when
considered alone, must meet the 0.10 lb/tsf emission limit; 2) kilns equipped with wet scrubbers
for PM emissions control are not eligible for averaging; 3) we are not allowing averaging for
other emission sources, or for opacity.
2-4
-------
3.0 Comment Summaries and EPA Responses
3.1 Subcategory for Scrubber-Equipped Kilns
Comment
In the preamble to the proposed rule, EPA requested comment on establishing a
subcategory for existing kilns equipped with wet scrubbers, if it could be demonstrated factually
that there would be significant environmentally counterproductive effects due to increased
emissions of acid gases, increased energy use, or increased water use. (See 67 FR at 78058.)
Several commenters (IV-D-04, IV-D-06, IV-D-02, IV-D-16, and IV-D-22) asked that a
subcategory for scrubber-equipped kilns be established. One commenter (IV-D-04) also asked
that kilns vented through a mine chamber be defined as scrubber-equipped kilns. The basis for
this request, in the commenters' view, is that wet scrubbers cannot meet the proposed particulate
matter (PM) standard of 0.12 pounds of PM per ton solid feed (lb/tsf), and therefore kilns with
scrubbers would have to replace them with baghouses. (They also assert that in most cases, wet
scrubbers have higher annualized costs compared to baghouses. Therefore, even if a wet
scrubber could meet a 0.12 lb/tsf emission level, facilities will opt to use baghouses due to cost
considerations. See comments on costs which follow.) This will result in an increase in
emissions of HC1 [a hazardous air pollutant (HAP)] and S02 (a non-HAP criteria pollutant), for a
nominal decrease in HAP metal emissions.
In support of this request one commenter provided estimates that not establishing the
requested wet scrubber subcategory would result in a metal HAP emissions decrease of 3 tons
per year (tpy), but would result in an increased emissions of 1823 tpy for HC1 and 2894 tpy for
S02.
One commenter (IV-D-12) stated that it had previously provided information on the
effect of replacing wet scrubbers with baghouses. They claimed that replacing one plant's wet
PM control system with a baghouse will increase HC1 by over 20 tpy, but only decrease metal
HAP by 481 pounds per year (Docket Item No. I-B-l 16, Appendix C).
Another commenter (IV-D-02) estimated that HAP currently being discharged from its
three existing kilns are 29 tpy of HC1 and 0.38 tpy of HAP metals. The emissions from the same
kilns if dry air pollution control equipment were used would be 652 tpy of HC1 and 0.075 tpy of
HAP metals. They also claim that the higher operating temperatures of dry systems cause metals
to vaporize and pass through a particulate collector, resulting in a lower metal concentration in
the captured particulate. As a result, they claim that even though dry control equipment may
reduce HAP metals emissions, the reduction will be minimal, while the release of HC1 will
increase significantly. In addition, the commenter notes that approximately 39 percent of the
HAP metal emissions attributed to the lime industry is manganese, which is not classified as a
known or suspected carcinogen. The commenter provides data which it claims show the only
3-1
-------
conventional pollutant that will be reduced with the installation of a dry control system will be
particulate emissions, and, "fugitive dust emissions from a dry system could more than offset the
improved particulate collection on the kiln exhausts."
Response
Standards implementing section 112 (d) of the Act must, of course, be of a minimum
level of stringency, usually referred to as the MACT floor. For existing sources, this floor level
of control cannot be less stringent than "the average emission limitation achieved by the best
performing 12 percent of the existing sources (for which the Administrator has emissions
information)." In this rule, EPA is establishing section 112 (d) standards to control emissions of
HAP metals, for which PM is a surrogate. None of the commenters challenge that the level of
PM emissions reflecting the average of the 12 per cent of the best performing sources is 0.12
lb/tsf. Notwithstanding, the commenters contend that EPA should subcategorize on the basis of
the type of air pollution control device used, and then separately determine the floor for each
subcategory.
Although the Act contemplates that EPA may establish subcategories when promulgating
MACT standards,1 subcategorization typically reflects "differences in manufacturing process,
emission characteristics, or technical feasibility". See 67 FR at 78058; see also 63 FR at 18768
(April 15, 1998) determining whether or not to subcategorize based on "the type of equipment
used in the process, the emission potential of each emission point, and any variations in the
process due to pulp type". A classic example, provided in the legislative history to section 112
(d), is of a different process leading to different emissions and different types of control
strategies - the specific example being Soderberg and prebaked anode primary aluminum
processes. rSee A Legislative History of the Clean Air Act Amendments of 1990 at 1138-39
(floor debates on Conference Report)].
Normally, it would be legally impermissible to subcategorize based on the type of air
pollution control device. The problem with subcategorizing on the basis of pollution control
device, quite simply, is that it leads to situations where floors are established based on something
other than the "best performing" sources. For example, suppose a source category consists of
100 sources using the same process and having the same emission characteristics, but that 50
sources use control device A to control HAP emissions, and 50 use control device B which is
Section 112 (d) (1) provides that "[t]he Administrator may distinguish among classes, types, and
sizes of sources within a category or subcategory in establishing such standards ". In addition,
section 112 (c) (1) encourages EPA to adhere to the subcategorization scheme already used in
establishing NSPS: "To the extent practicable, the categories and subcategories listed under this
subsection shall be consistent with the list of source categories established pursuant to section
111 [requiring EPA to establish technology-based New Source Performance Standards] and Part
C" (although the NSPS for the lime industry does not contain subcategories, nor was it ever
suggested when establishing NSPS that subcategories would be appropriate).
3-2
-------
two orders of magnitude less efficient. If one subcategorized based on the type of pollution
control device, the MACT floor for the 50 sources with control device B would reflect worst,
rather than "best" performance. Although the disparity in levels of emission control between the
best performing sources here, and the best performing sources using wet scrubbers is not this
dramatic - roughly a factor of 5, not orders of magnitude - the difference is nonetheless ample
and evident.
The question of whether it is permissible to subcategorize on the basis of pollution
control device has been litigated in the context of the effluent guideline provisions of the Clean
Water Act. These provisions, which were the prototype of the MACT provisions in the Clean
Air Act [see 5 Legislative History at 8473-74 (Senate Report)], require control of toxics
reflecting performance of best technology, although the Clean Water Act gives EPA more
flexibility to subcategorize than the Clean Air Act, since there is no floor provision, and EPA is
required to consider manufacturing process, age, engineering aspects, costs, and types of control
equipment (among other factors) in setting the standards. Clean Water Act section 304 (b).
Nonetheless, the Fifth Circuit rejected the argument that EPA erred in not subcategorizing based
on different types of pollution control equipment:
[Petitioners' plants do not differ materially from other [chemical] plants with respect to
manufacturing operations or wastewater characteristics.... Their plants differ only with
respect to the type of treatment petitioners voluntarily chose to employ. Petitioners did
not choose to employ the type of... treatment system that is most commonly used by
good performers in the [chemical] industry .... Instead, petitioners chose to employ ...
treatment systems which generate additional pollutants .... The EPA asserts that it
therefore reasonably required petitioners' plants to meet the limitation demonstrated to be
achievable by the 'best' plants in the industry, rather than providing less stringent
limitations for these plants simply because they chose to employ less effective treatment
technology....
Petitioners cannot now be excused from meeting the BPT [Best Practicable Technology]
limits on the grounds that their present method of treatment is less effective than the
average of the best.... Such a construction of section 304 (b) ( \) fB) would defeat the very
purpose of establishing BPT limits. The EPA is not required to base BPT upon a
technology that is less effective than the best practicable technology demonstrated in the
industry....We therefore hold that the EPA was entirely reasonable in declining to create a
subcategory for plants employing [the less effective] treatment systems.
Chemical Manufacturers Association v. EPA. 870 F. 2d 177, 218-19 (5th Cir. 1989) (emphasis
supplied), modf d on different grounds on rehearing, 885 F. 2d 253 (5th Cir. 1989).
This remains EPA's concern in the final rule. Subcategorizing purely on the basis of
pollution control device type "would defeat the very purpose of establishing [MACT]" because
3-3
-------
such a construction would mean the MACT floor would no longer reflect the performance of
best performing sources, as required.
It has been suggested here, however, that the best performing sources with respect to
HAP metal emissions should not be considered "best performing" because that performance
comes at an environmental cost, namely increased emissions of HC1 and S02 compared to what
lime kilns equipped with wet scrubbers would emit. There is some support for the idea that if an
ostensibly best performing pollution control device creates potentially counterproductive
environmental effects, such performance need no longer be considered "best" due to the
counterproductive effects, and could justify differentiation in the form of subcategorization (See
66 FR at 3187 [January 12, 2001]). Commenters suggest that is the case here due to increased
emissions of HC1 and S02 which would result (they maintain) if lime kilns replace wet scrubbers
with baghouses.
First, it is not clear the commenters' starting premise, that baghouses are either needed or
will be used to achieve the PM standard, is correct. We continue to believe that a properly
designed and operated wet scrubber can meet the proposed PM emission limit of 0.12 lb/tsf. The
reasons for this are specifically addressed in our following response to comments on achievable
wet scrubber performance.
The commenters provided no data to refute that application of a 0.12 lb/tsf PM emission
limit represents best control of HAP emissions if we do not create any kiln subcategories. We
note that as part of their comments, they claim that the higher temperatures of dry PM controls
result in metals vaporizing and passing through the PM control. However, the data provided in
their comment does not substantiate their claim, and analyses performed for the Hazardous
Waste Combustor NESHAP indicate that all but a few percent of the metals in question exit the
kiln as solid particulate (see pages 1-1 through 1-4 in the Draft Technical Support Document for
HWC MACT Standards, February 1996, which are in the docket for this rulemaking). The same
comments stated that the increased PM emissions control that results from baghouses could be
more than offset by increased fugitive dust emissions from dry systems. We know of no data
that supports that statement. Dry dust systems can be designed and operated to minimize
fugitive emissions. Moreover, we do know that fabric filters are commonly used when high
levels of PM control are desired in many industries. In none of these cases do we know of any
instance where less efficient wet scrubbers were applied in lieu of more efficient fabric filters
due to concerns of fugitive dust emission from the collected particulate.
We also evaluated the concerns that switching to dry controls may result in increases of
S02 and HC1 emissions. However, our analysis indicates that the extent to which S02 and HC1
emissions actually increase may have been overstated by the commenter. EPA estimates that if
all facilities currently using wet scrubbers switched to dry controls, HC1 emissions would
increase by approximately 1,310 tpy (vs 1,800 tpy estimated by the commenter) and S02
emissions would increase by about 1,830 tpy (vs 2,900 tpy estimated by the commenter). (See
the memorandum "Environmental Impacts of Decision on Best Control for Wet-Scrubber
controlled Kilns" in the docket for this final rule).
3-4
-------
We do not regard either level of increased HC1 emissions as significant. We modeled
this emissions increase as part of our determination (pursuant to section 112 (d) (4)) that
emissions of HC1 from lime kilns are below an HC1 risk threshold, with an ample margin of
safety. See FR 78054 - 78057 and the risk analysis in the docket for the final rule. Given this
determination, we cannot view these HC1 increases as being so significant as to call into question
whether the best-performing sources with respect to HAP metal reductions are in fact "best
performing".
The commenters also cite projected increases in the criteria pollutant S02 that may result
from our decision not to subcategorize by existing control device. They did not initially address
the reductions in PM emissions resulting for the decision that 0.12 lb/tsf represents best-control
for scrubber equipped kilns. We estimate that nearly 1,080 tpy of additional PM would be
removed if all existing kilns were to meet a standard of 0.12 lb/tsf, of which approximately 1.6
tpy are metal HAP. Although EPA may not promulgate standards for non-HAP under section
112 (d), Congress expected reductions in emissions of criteria pollutants such as PM to be a
benefit of the MACT program. 5 Legislative History at 8512 (Senate Report) ("When
establishing technology-based standards under this subsection, the Administrator may consider
the benefits which result from control of air pollutants that are not listed but the emissions of
which are, nevertheless, reduced by control technologies or practices necessary to meet the
prescribed limitation [going on to give VOC reductions as an example].") In comparison to
estimates of increased emissions of S02 by either the commenter or EPA, the decrease in
captured PM emissions (and attendant decrease in capture of non-mercury metal HAP) if all
existing kilns are not required to meet a standard of 0.12 lb/tsf is significant.
There is a further consideration, however. Based on the available size distribution data
from Compilation of Air Pollutant Emission Factors, AP-42, Fifth Edition, Volume I: Stationary
Point and Area Sources, 73 percent of the PM emitted directly by lime kilns is coarse PM (PM in
the size range of 2.5 to 10 micrometers). Some of the S02 emitted to the atmosphere undergoes
chemical reactions to form fine PM. [See generally the respective Criteria Documents for PM
(EPA/600/P-95/001aF-cf. 3v, 1996) and S02 (EPA/600/8-82-029aF-cF. 3v., 1982 and addenda)].
Thus, in assessing whether some potential factor might justify a decision that kilns with dry
systems are not best performing, some comparison of coarse verses fine PM emissions is needed.
If we retain a single PM emission limit of 0.12 lb/tsf for all existing kilns, total PM
emissions would be reduced (compared to separate standards for kilns with wet scrubbers and
dry controls) by an additional 1,080 tpy. Of that number, 630 tpy is fine PM and 450 is coarse
PM. The potential amount of increased S02 emissions is 1,830 try. A portion of this 1,830 tpy
of S02 will be converted in the atmosphere to produce 1,270 tpy of fine PM. Therefore, the
incremental impact of a single PM standard of 0.12 lb/tsf for both wet scrubbers and dry controls
would be an increase of 640 (1,270-630) tpy in fine PM emissions, and a decrease of 450 tpy in
coarse PM emissions. This assumes that all facilities that currently have wet scrubbers switch to
dry controls, and that 46 percent of the S02 converts to fine PM. The 46 percent conversion
estimate used by the commenter is consistent with information in the respective Criteria
Documents for PM and S02 referenced above.
3-5
-------
As recently summarized by EPA (68 FR 28339, May 23, 2003), scientific studies show
ambient PM (both fine and coarse) is associated with a series of adverse health effects. Fine PM
is associated with increases in daily mortality. Coarse PM is more strongly linked to morbidity
(e.g. hospital admissions). See generally the respective Criteria Documents for PM (EPA/600/P-
95/00 laF-cF. 3v, 1996) and S02 (EPA/600/8-82-029aF-cF. 3v., 1982 and addenda). Therefore,
it is difficult to make comparisons between the relative benefits of reducing emissions of fine
and coarse PM.
We view this situation as equivocal: it is unclear which of these types of performance is
best since on the one hand there is reduced emissions of HAP metals and coarse PM but
foregone control of S02 and sulfate (fine) PM, and, for kilns controlled with wet systems, the
converse. In this situation, and based on these facts, which, with current analytic tools seem to
us to be largely in equipoise, we are not prepared to view either wet or dry systems as best
performing and instead are promulgating a separate PM standard for each.
We emphasize that considerations of risk and relative environmental benefits are
normally irrelevant to MACT floor determinations [unless expressly authorized by statute, as in
Clean Air Act (CAA) section 112(d)(4) as applied in the final rule], since floor standards must
reflect the performance of the specified number of best performing sources. See National Lime
Ass'n v. EPA. 233 F. 3d at 640 (considerations of cost and de minimis risk cannot be considered
in making MACT floor determinations). We considered these factors in the final rule solely for
the purpose of evaluating the commenters' claim that sources using wet and dry control systems
should be evaluated separately for MACT floor purposes due to environmental benefits and
disbenefits associated with wet and dry control systems. Given the equivocal facts here, we are
not prepared to state that either the wet or dry systems are "best performing" for existing kilns,
and are instead promulgating separate PM standards (as a surrogate for non-mercury HAP
metals) for each.
Comment
Commenter IV-D-04 stated that wet scrubbers cannot meet the proposed PM emission
limit of 0.12 lb/tsf. They claimed that a wet scrubber manufacturer will only guarantee this limit
if less than 1 percent of the particles to be removed are less than 1 micrometer in diameter. The
commenter stated that EPA assumes that the average mass diameter of particles in lime kiln gas
effluent is 2 micrometers, and that this assumption is based on a single reference, and that
reference was actually fugitive lime dust, not lime kiln particulate. They further claimed that
volatilization and homogenous nucleation of potassium chloride particles in the gas stream
generates particles in the 0.1 to 0.5 micrometers size range. "As particle size decreases below 1
micrometer, inertial compaction becomes decreasingly effective. Above 0.1 micrometers,
Brownian displacement is ineffective. In the range between 0.1 and 0.5 micrometers, neither of
these two main particle capture mechanisms relied upon in wet scrubber design is very
effective." The commenter presents a recent example of a scrubber installation and an analysis
of Docket item No. II-B-126 (an EPA memo titled "Technical Feasibility of Venturi Scrubbers in
Meeting 0.12 lb/ton Outlet PM Level") to demonstrate this point.
3-6
-------
Commenter IV-D-02 claimed that a scrubber performance efficiency of 99.9 percent will
be required to meet the 0.0072 grain/dry standard cubic foot (gr/dscf) particulate concentration
which they claimed corresponds to the proposed PM emission limit of 0.12 lb/tsf. The
commenter's environmental consultant advised that it is unlikely a wet scrubber with a 35-inch
water gauge (w.g.) pressure drop could achieve this level of performance with the facility's
current inlet exhaust particulate loading.
Response
We continue to believe that a properly designed and operated venturi wet scrubber with a
35 inch w.g. pressure drop can achieve the proposed PM emission limit of 0.12 lb/tsf. The
commenter argues that EPA's basis for a 35 inch w.g. scrubber pressure drop to obtain 99.94
percent efficiency is not realistic. They base their arguments on perceived improper use of
estimation procedures, incorrect choice of typical particle size, faulty vendor estimates, and lack
of knowledge regarding measurements of an existing scrubber in Virginia.
This commenter misunderstands the manner in which EPA evaluated venturi scrubber
performance. The particle size data used in the preliminary estimate came from several sources,
not one as suggested by the commenter. Second, of the various data available, we chose the data
on fugitive lime dust PM because they had the smallest mean particle size diameter, which was 2
micrometers. This was a very conservative assumption, because, for a normal particle size
distribution, the wet scrubber pressure drop required to meet a specific removal efficiency
increases as the mean particle size decreases. All the other data presented in our analysis
indicates that the mean particle size for lime kiln particulate is much larger than the 2
micrometers we chose. For example, the particle size data reported in EPA's AP-42, Section
11.15, Lime Manufacturing, indicate that the mean particle size for uncontrolled kilns is over 15
micrometers.
We used this atypically small particle size as a median diameter to input into simplified
methods to estimate the pressure drop required to meet a scrubber PM removal efficiency of
99.94 percent (the removal efficiency required to meet a 0.12 lb/tsf PM emission limit).
The first commenter also asserted that the pressure drop estimation procedure we used is
inappropriate. The commenter stated that in determining the required wet scrubber pressure drop
in a particular application, the particulate must be assigned to a set of discrete size ranges and the
scrubber efficiency evaluated at a specific pressure drop size range by size range. These results
must then be used to calculate the overall PM removal efficiency.
We agree that, if data are available, the procedure described above is more accurate.
However, we note that the commenter did not provide any particle size data that would either
substantiate the contentions, or allow us to use the procedure claimed to be more accurate. In the
absence of the type of data described by the commenter, the conservative and simplified
3-7
-------
approach to estimating wet scrubber pressure drop used here, that is, using a conservative
estimate of median particle size, is reasonable. Therefore, based on the data available, the
procedure we have used to estimate what scrubber pressure drop, on average, will be necessary
to meet a 0.12 lb/tsf PM emission limit, is sufficiently accurate. As explained, this procedure
results in an estimate that a 35-inch w.g. pressure drop is needed.
However, as a check, we also contacted wet scrubber vendors to provide more rigorous
estimates to confirm whether the 35-inch w.g. pressure drop value was valid. One vendor, using
proprietary data, estimated a 35-inch w.g. pressure drop was sufficient provided there is
sufficient moisture in the gas stream ahead of the venturi throat to increase particle size so that
the venturi scrubber can remove the PM. We believe that the fact that two independent estimates
produced the same result verifies our estimation procedure.
If necessary, cooling water can be added prior to the venturi in order to sufficiently
increase particle size. In some cases, more water can be added through a packed bed. Note that
the cooled gas stream would have lower gas volume resulting in smaller fan, motor, and duct
requirements.2
This same commenter presented data from a wet scrubber controlled calcimatic lime kiln
indicating that potassium chloride in limestone leads to generation of a significant quantity of
particles in the 0.1 to 0.3 micrometer range, which are difficult to collect in a wet scrubber. This
wet scrubber was designed for an outlet concentration of 0.01 gr/dscf at a pressure drop of 22
inches w.g., but, after modifications, could only reach about 0.04 gr/dscf. The commenter
asserted that it is "often the case" that 10 percent of the inlet loading to a lime kiln scrubber is in
the 0.1 to 0.3 micrometer range. The commenter later provided data they claim supports this
statement. The commenter cited this example to call into question EPA's conclusion that a wet
scrubber with a pressure drop of 35 inches of water can meet a 0.12 lb/tsf PM emission limit.
First, and most important, the example cited by the commenter can in no way be used to
make generalized industry-wide conclusions about the ability of wet scrubbers to meet a 0.12
lb/tsf PM emission limit. The commenter did not include in their comments information from a
BACT analysis for the Global facility (Global Chemstone Corporation, March 15, 2002) written
by the same person who wrote the commenter's statement regarding pressure drop and particle
size. This BACT analysis states that the calcimatic kiln being controlled has different
characteristics than the more conventionally found rotary kiln. Statements about the scrubber
2 Packed beds were not specifically costed out in our model kilns, but the additional cost for the
packed bed (if ever needed) would be offset by the reduced cost of the fan, motor, and ductwork
as well as reduced annual cost for power that result from the effect of gas cooling which reduces
gas volume. However, some plants could possibly upgrade their existing scrubber for about 40
percent of the cost of a new wet scrubber. In these cases, our model plant costs overstate capital
cost by about 150 percent. On a nationwide basis, our scrubber costs probably overstate what
actual costs are likely to be.
3-8
-------
and kiln indicate that they present a unique case, not likely to be found at other lime kiln sites.
Examples are,
"... the existing total filterable particulate matter emission limit of 0.022 grains/dry
standard cubic foot is too low given the unique issues associated with operating a
calcimatic kiln with a moderate-to-high potassium content limestone feed [363 parts per
million (ppm)]." (p.l of the analysis)
"Global has a Calcimatic lime kiln, which is fundamentally different from a traditional
rotary lime kiln." (p. 2 of the analysis)
"However, there are significant differences between a rotary lime kiln and calcimatic
lime kiln." (p. 5 of the analysis)
"The dominant characteristic of rotary lime kilns is the high concentration of particulate
matter entrained in the gas stream due to the direct contact of the hot flue gas with the
tumbling feed materials within the kiln. This particulate matter provides the surface area
necessary for the condensation (heterogeneous nucleation) of vapor phase materials.
Potassium and other vapor phase materials in the kiln gas stream will concentrate
primarily in moderate-to-large particulate matter that can be controlled with high
efficiency wet scrubber systems." (p. 5 of the analysis)
"... the particulate matter loadings from a calcimatic kiln are lower than those of a rotary
kiln. Due to the low particulate matter concentrations in the calcimatic kiln effluent gas
streams, a significant fraction of the vapor phase potassium and other vapor phase
material can condense (homogeneously nucleate) to form submicrometer sized particles
after the gas stream exits the kiln. Therefore, emission figures based on rotary kiln
control equipment are not relevant to calcimatic lime kilns." (p. 6 of the analysis)
"... a reverse air fabric filter with membrane bags is technically infeasible at Global. The
problems inherently involved with submicrometer sized potassium chloride also preclude
the use of (1) reverse air fabric filters with conventional woven filter media and (2) pulse
jet fabric filters with membrane-coated felted bags." (p, 8 of the analysis, note that fabric
filters have been used to control lime kilns for decades)
"The Calcimatic kiln is more vulnerable to homogeneous nucleation than rotary lime
kilns due to the low particulate matter concentrations in the Calcimatic kiln effluent gas
stream." (p. 11 of the analysis).
Based on these comments, it is clear that the situation at the Global Stone Chemstone
facility is unique, and not applicable to the industry as a whole.
3-9
-------
However, the information presented here could be used to imply that wet scrubbers are
not as effective on calcimatic kilns as on to rotary kilns. (Note that calcimatic kilns make up only
about 6 percent of the currently operating lime kilns). However, as we have never claimed that a
wet scrubber with a 22 inch w.g. pressure drop can meet the proposed PM emission limit of 0.12
lb/tsf (as explained above, we estimated that a 35 inch w.g. pressure drop will be necessary), we
do not see how the example cited by the commenter is relevant. Also, given the fact that the
particle size data appear to have been collected after the scrubber was installed, we conclude that
that there are certain site specific factors that were not accounted for in the scrubber design.
However, this commenter goes on to claim that the wet scrubber in question could not
meet a PM emission limit of 0.12 lb/tsf even if the pressure drop were increased to 35 inches.
They claimed this is because of the presence of a large amount of particles in the 0.1 to 0.5
micrometer size range at this particular facility, and that particles in this size range are the most
difficult to collect with a wet scrubber. However, it should still be possible to meet 0.12 lb/tsf
even in this case, though additional costs for water injection or for steam injection (to provide
additional moisture in the exhaust gas) may be required. In that case, if a facility was required to
meet a 0.12 lb/tsf PM emission limit, it might opt for a fabric filter. Our model plant costing
does not purport to predict how any individual kiln might react. Finally, based on our
determination that facilities that currently have wet scrubbers meeting 0.6 lb/tsf are
environmentally equivalent to dry controls meeting 0.12 lb/tsf, the fact that this facility claimed
it cannot meet 0.12 lb/tsf is moot.
Another point made by this commenter concerned the choice of gas flow rate used for
scrubber-equipped kiln modeling. From Information Collection Request (ICR) data available to
EPA, exhaust gas flow rates for scrubber-equipped kilns were found to vary from about 100,000
actual cubic feet per minute (acfm)/ton of lime to about 310,000 acfm/ton of lime. The EPA
chose an intermediate value of 165,000 acfm/ton of lime for modeling purposes (see docket item
II-B-14). There was no intent to choose either the lowest or the highest flow rate in the database
because on average, an intermediate value would be more representative. This commenter also
stated that the value of the PM concentration in the flue gas was incorrect because we did not
account for the water vapor from evaporation in the scrubber, and used a 800°F flue gas
temperature. The concentration calculation was used to define inlet conditions to the model
scrubber so that scrubber sizing and water requirements could be estimated by a vendor. The
estimated PM concentration was not used in determining scrubber pressure drops, or in
estimating costs.
The second commenter supplied no data to support the assertion by their consultant that a
35-inch w.g. pressure drop scrubber could not achieve the required efficiency. Our responses
above also address their comment.
Comment
3-10
-------
Commenter IV-D-04 provided an analysis of wet scrubber feasibility prepared by a
consultant to supplement their summarized comment above. This analysis raised several
additional points not included in our previous comment response. These are as follows:
It took issue with the statement in Docket item No. II-B-126 "Scrubbers have been
reported to collect submicron particles as low as 0.006 grains per dry standard cubic foot
in EPA's Handbook of Control Technologies for Hazardous Air Pollutants. The
commenter pointed out that this statement concerned hydrosonic scrubbers.
It claimed the calculated PM concentration at the scrubber outlet equivalent to 0.12 lb/tsf
was incorrect.
The inlet mass loading data are limited.
The static pressure drop estimate is not applicable to bimodal distributions common to
lime kilns.
Response
The commenter correctly points out that the statement above related to hydrosonic
scrubbers. However, this statement was only intended as an introductory statement in the memo.
We based our analysis on our estimation of the performance of venturi scrubbers. Therefore, the
comment, though correct, is irrelevant.
The calculated PM concentration was not used to estimate the required scrubber
efficiency. The required scrubber efficiency was estimated based on mass emission rates.
Therefore, the concentration value has no relevance to our estimate of the scrubber pressure drop
required to meet a PM emission limit of 0.12 lb/tsf.
We agree that the inlet loading PM data are limited. However, we believe that we have
chosen a conservative value. We note that this commenter provided data elsewhere in their
comments that indicated our estimate of 160 to 175 lb/tsf may be very conservative.
The commenter did not provide any data to support the statement that the particle size
distribution from lime kilns is bimodal in the size range of 0.1 to 0.5 microns. The aerodynamic
particle sizing information from the scrubber at the Virginia site does not cover the range below
0.5 microns. Apparently the commenter is basing the statement that the particle size distribution
is bimodal on data from wet-scrubber controlled lime kilns with high sodium feed rates in the
pulp and paper industry. However, there are no data that substantiate the commenter's
statements for either calcimatic lime kilns, such as the kiln in Virginia, or for rotary lime kilns
not in the pulp and paper industry, which make up the bulk of the kiln population.
3-11
-------
Comment
Commenter IV-D-04 contends that EPA asserts incorrectly that lime plants will choose
high efficiency venturi scrubbers to replace their current wet scrubbers because high efficiency
venturi scrubbers have lower capital costs and sometimes lower annual costs than baghouses.
They further stated that, according to Docket Item No. II-F-7, five of the six model kilns the
Agency examined had much higher annualized costs for high efficiency venturi scrubbers than
for baghouses. This commenter submitted a manufacturer's cost proposal that shows a scrubber
with a 35-inch w.g. pressure drop costs substantially more than EPA estimates. They conclude
from this that lime kilns will be forced to use baghouses, with attendant increases in HC1 and
S02 emissions. Another commenter (IV-D-22) stated that the cost for the installation of a
baghouse will be greater than EPA estimated due to the location of existing equipment in the
area where the collector should be located, construction of the duct collector in a congested area
with plant operations, and accessibility to existing lime kiln dust handling systems.
Commenter IV-D-22 commented that the cost for the installation of a baghouse will be
more costly than EPA estimated due to the location of existing equipment in the area where the
collector should be located, construction of the duct collector in a congested area with plant
operations, and accessibility to existing lime kiln dust handling systems.
Response
Regarding modeled high costs for scrubbers compared to baghouses, individual models
may show this characteristic. However, the distribution of kiln sizes in the lime industry and the
allocation of model plants to those kilns show that estimated nationwide costs for installing
venturi scrubbers is about the same as for installing baghouses. Tables 1 and 2 illustrate this
condition.
In all cases, the estimated capital cost for each model scrubber is less than its
corresponding model baghouse. Five of the nine models have estimated annual scrubber costs
that exceed the annual baghouse costs by amounts ranging from 4 percent to 75 percent.
However, when model costs are converted to national costs based on kiln type and size, the
tables show a lower estimated total annual cost for upgrades with venturi scrubbers than for
upgrades with baghouses. For both types of control system, costs for any specific plant may be
more or less than the value shown by the model used to estimate nationwide cost. The plant is
expected to buy whatever system its management believes is in the best business interests of the
owners, but in the aggregate, the tables indicate that estimated annual cost for control systems is
lower if all plants replace existing equipment with venturi scrubbers rather than baghouses.
However, based on the specific model kiln costs, we still find that at least some kilns would
replace wet systems with dry if required to meet a uniform PM emissions limit of 0.12 lb/tsf.
We developed these model plants based on actual facility data from the industry
responses to questionnaires. Our model plants cover the various types of kilns currently in
3-12
-------
service, and a range of sizes of facilities. Other than the comment on gas flow rates previously
addressed, we received no comments that the model kilns are not representative of the industry.
The model plant parameters are generally based on the mid-range of the characteristics of the
segment of the existing kiln population that they are designed to represent. Because they are
based on mid-range values, comparing them to actual costs at any one specific kilns may result
in a higher or lower cost estimate. However, there is no reasonable way to model every
individual kiln or account for every site-specific factor. Using model kilns based on mid-range
values provides an accurate estimate of the nationwide costs, even if costs for some sites may
deviate from the modeled estimates.
Regarding a cost proposal substantially higher than EPA's estimate for a scrubber with
35-inch w.g. pressure drop, it is expected that some proposals will have higher costs, and some
lower, than costs estimated by EPA's methodology. The methodology is based on estimates of
basic equipment costs, and factors to calculate direct and indirect capital costs that constitute
total capital investment. Unit costs are applied to labor, utilities, waste disposal, and other
operating and maintenance costs to obtain direct annual costs. Indirect annualized costs based on
capital recovery and other service charges are also estimated and added to direct annual costs to
obtain total annual cost. The methodology is explained in the EPA Air Pollution Control Cost
Manual. For establishing projected control costs in the lime industry, the methodology was
applied to model plants having characteristics based on industry responses to questionnaires.
These model estimates account for all of the components needed to build and operate a wet
scrubber. However, the specific cost of any one component in our models, which are based on
industry averages, may be different from those for a specific plant. Overall, our model plant
approach still provides a reasonable estimate on which to base regulatory decisions.
The previous response also applies to the comment regarding installation of a fabric filter
that is projected to cost more than the model cost obtained by EPA. The cost estimation
methodology for lime plants contains factors that account for demolition and salvage of existing
equipment that will have to be replaced by the new control system. A retrofit factor is also
included to account for difficulties in replacing existing equipment with new equipment in an
existing plant. As stated in the previous paragraph, estimates made with model plants will not
address each specific case found in industry. However, nationwide costs are reasonably
estimated, and small business concerns, where they may appear, are addressed within the
regulation.
3-13
-------
Table 1. Cost to Upgrade Existing Kilns Equipped with Wet Scrubbers to Venturi Scrubbers
Model
No. of kilns w/ws*
Model Kiln Capital Cost
Model Kiln Annual Cost
Total Capital Cost
Total Annual Cost ($1000)
kiln
($1000/kiln)*
($1000/kiln)*
($1000)
A
4
252
113
1,008
452
B
5
288
131
1,440
655
C
14
520
293
7,280
4,102
D
11
663
403
7,293
4,433
G
1
301
146
301
146
H
1
499
278
499
278
1
0
682
420
0
0
L
1
351
180
351
180
M
4
177
64
708
256
Preliminary
41
18,880
10,502
Total
TOTAL (assuming 70% are major sources and 90% of scrubbers will upgrade)
11,894
6,616
*Source: Economic Impact Analysis for the Lime Manufacturing MACT Standard, ll-A-44.
Table 2. Cost to Upgrade Existing Kilns Equipped with Wet Scrubbers to Baghouses (w/ water sprays)
Model
No. of kilns w/ws*
Model Kiln Capital Cost
Model Kiln Annual Cost
Total Capital Cost
Total Annual Cost ($1000)
kiln
($1000/kiln)*
($1000/kiln)*
($1000)
A*
4
790
204
3,160
816
B
5
859
225
4,295
1,125
C
14
1488
274
20,832
3,836
D
11
2824
386
31,064
4,246
G
1
558
90
558
90
H
1
984
71
984
71
1
0
2395
131
0
0
L*
1
665
237
665
210
M*
4
569
197
2,276
788
Preliminary
41
63,834
11,182
Total
TOTAL (assuming 70% are major sources and 90% of scrubbers will upgrade)
40,215
7,045
* Model plant costs for models A, L, and M are the cost of baghouses plus cost of water injection (see ll-B-111)
3-14
-------
Comment
Several commenters (IV-D-04, IV-D-06, and IV-D-12) stated that not establishing a
subcategory for scrubber-equipped kilns will adversely affect small businesses. Commenter IV-
D-04 states the annualized cost of upgrading all scrubbers is $9.45 million, based on the
distribution of kilns and EPA's estimate of total annualized costs in Docket Item No. II-B-127.
According to the commenter, EPA predicts that upgrading these kilns will reduce HAP metals by
3.1 tpy, resulting in a cost effectiveness of $3.0 million/ton of metal HAP. The commenter
stated that EPA's assumption that 30 percent of lime plants are area sources and won't be
affected by the rule reduces the removal of metal HAP's attributed to upgrading scrubber-
equipped kilns to 2.2 tpy (although the commenter stated that EPA has provided no support for
the assumption that 30 percent of lime plants are area sources).
Commenter IV-D-12 notes that EPA's estimated annualized cost for the commenter to
install baghouses is $2,236,000, which equates to $9.3 million per ton of particulate HAP
control.
Response
The comments state that requiring kilns currently controlled by wet scrubbers to meet the
proposed 0.12 lb. tsf PM emission limit is too costly considering the environmental benefits.
However, section 112 of the Clean Air Act precludes us from considering cost when calculating
MACT floors. Therefore, none of the cost issues discussed above are sufficient to support a
separate subcategory for existing kilns with wet scrubbers. The commenters point to no process
or other difference which would prevent them from achieving the level of performance attainable
by the average of the best performing existing sources, nor suggest that those best performing
sources are unrepresentative of other kilns in terms of process, baseline emissions, or other
factors which traditionally would justify a separate subcategory.
Though costs cannot not be a consideration here, our estimate shows a cost of $7.15
million to upgrade all scrubbers, versus the $9.45 million figure provided by the commenter.
Our estimate assumes 70 percent of kilns are located at major sources and 90 percent of
scrubbers will require an upgrade.
We also note that, in addition to decreasing the emissions of HAP metals, significant PM
emission reductions will result from upgrading existing scrubbers. Upgrading the existing
scrubber equipped kilns will reduce PM emissions by approximately 1,300 tons/year for an
approximate cost effectiveness of $5,500/ton of PM. The cost per ton of reduction is reasonable
compared to other PM control regulations recently promulgated or proposed (such as the non-
road diesel engine rule). We acknowledge that, in individual cases, costs and cost effectiveness
will vary and in some cases be higher or lower depending on individual circumstances.
3-15
-------
In addition, in our cost analysis, we assumed that 90 percent of existing scrubbers would
be replaced, and included all the capital cost of replacement in the cost of the final rule. This is
probably an overly conservative way of estimating costs. In reality, it is reasonable to assume
that, on average, the existing scrubbers have only 50 percent of their useful life remaining.
Because we allocated all of the capital cost of a new scrubber to the cost of the rule, our costs are
conservative.
Comment
One commenter (IV-D-21) objects to EPA's rationale of using PM as a surrogate for
controlling toxic emissions. The commenter states that if EPA has sufficient data to indicate that
toxic emissions from lime kilns are an ambient air problem, then the regulation should focus on
reducing gaseous emissions such as HC1.
Response
By limiting emissions of PM, the rule will reduce emissions of non-volatile and semi-
volatile metal HAP, which are a subset of PM. As stated in the preamble to the proposed rule,
air pollution controls for HAP metals are the same as the PM controls used by the lime
manufacturing industry, i.e., FF, ESP, and wet scrubbers. These controls capture non-volatile
and semi-volatile metal HAP non-preferentially along with other PM, thus making PM an
acceptable indicator of these HAP metals. Quite simply, "PM control technology
indiscriminately captures HAP metals along with other particulates....". National Lime Ass'n v.
EPA. 233 F. 3d at 639. Consequently, it is an appropriate indicator when the technical basis of
the standard is performance of back-end particulate control technology. Id Another reason for
using a surrogate is the lower cost of emissions testing and monitoring for PM as compared to
the cost of emissions testing and monitoring for multiple metal HAP that would be required to
demonstrate compliance. Because PM control devices control metal HAP to roughly the same
efficiency, and because of the associated cost savings associated with emissions testing and
monitoring, the Agency has promulgated several other NESHAP where PM is a surrogate for
non-volatile and semi-volatile metal HAP.
Regarding the commenter's second point concerning regulating emissions of HC1, the
preamble to the proposed rule explained in detail the Agency's decision not to regulate HC1
emissions from lime kilns. To summarize that discussion, the EPA determined that, under the
authority of section 112(d)(4) of the CAA, no further control was necessary because HC1 is a
threshold pollutant and HC1 levels emitted from lime kilns are below the threshold value with an
ample margin of safety, after considering potential threats to both humans and the environment.
3-16
-------
Comment
One commenter (IV-D-12) requests a subcategory for their unique mine chamber exhaust
control. Exhaust gas from the kilns is vented to an underground mine chamber which is unlike a
baghouse or conventional scrubber, but functions like a scrubber due to moisture in the mine.
The typical exhaust temperatures at the exhaust stack range from 107 to 112 degrees F - well
below 400 degrees F (a work practice temperature standard that had been considered by EPA for
HC1 control, but was rejected). In addition, the condensation moisture assures saturation of the
exhaust stream and the contact time exceeds those possible under any conventional pollution
control device. The commenter also points out that the mine is not subject to mechanical failures
or operator induced malfunctions as are conventional control devices. According to the
commenter, stack testing done in 1995 demonstrated removal efficiencies of 98.9 percent for
PM, 99.1 percent for S02, 52.3 percent for NOx, and 71.9 percent for CO.
According to the commenter, the cooling of exhaust gases and subsequent condensation
of water and the presence of entrained water droplets makes continuous opacity monitoring
infeasible, which is why visible emission readings are currently performed. The visible emission
readings taken in the last several years show little variation and no exceedances of the applicable
new source performance standards (NSPS) subpart HH kiln opacity standard of 15 percent.
Response
The commenter is requesting that their use of a mine as a PM control device for their kiln
exhausts be placed in a subcategory for wet scrubbers. But, as discussed above in a response to a
previous comment, EPA has decided that it will not develop a separate subcategory for wet
scrubbers. Subcategorization is normally based on differences in manufacturing process,
emission characteristics, or technical feasibility and is not justified solely on the basis of a
different type of control device being used. Under this principal, the use of a mine as an
emission control device does not warrant a separate subcategory.
3.2 PM Emission Limit for Scrubber-Equipped Kilns
Comment
In response to EPA's request for comments on what the MACT floor limit should be if
there was a subcategory for scrubber-equipped kilns, three commenters (IV-D-04, IV-D-06, and
IV-D-16) state that the PM emission limit for scrubber-equipped kilns should be 0.6 lb/tsf. This
is based on actual operational data and permit limits for the top 12 percent of the existing kilns
that use wet scrubbers for PM emissions control. These data were compiled by the National
Lime Association (NLA) and provided to EPA.
3-17
-------
Response
As discussed above, EPA has determined that subcategorizing by existing controls is not
appropriate. However, we have also determined that from an overall environmental standpoint, a
wet scrubber meeting an emission limit of 0.60 lb/tsf could be considered best control if the
environmental disbenefits of switching to a dry control are considered. Therefore, we have set
an emission limit of 0.60 lb/tsf for kilns that currently have wet scrubbers for PM control.
3.3 PM Emission Limit for New Kilns
Comment
One commenter (IV-D-04) states that the PM emission limit for new lime kilns should be
0.12 lb/tsf, the same as the emission limit for existing kilns. The commenter notes that the
proposed limit is based on two three-hour test runs at one plant. According to the commenter,
EPA recognized in the proposal preamble that 3-hour test results are just a snapshot in time, and
should not be used as the basis for establishing an enforceable standard, and that EPA expressly
rejected such an approach when establishing the MACT floor for existing kilns.
The commenter believes that data in the docket shows that 0.10 lb/tsf is not continuously
achievable by lime kilns, and EPA should not establish a separate PM limit for new lime kilns.
Another commenter (IV-D-16) believes 0.10 lbs PM/tsf for a new kiln is too restrictive,
and EPA does not have adequate data to determine that a baghouse or scrubber equipped kiln
could achieve this low level of emissions on a sustained basis.
Response
The approach to which the commenter refers whereby EPA rejected the use of the
"average or mean" in establishing the MACT floor for existing sources, did not refer to the
average of individual test runs as implied by the comment. Rather, it refers to EPA's decision to
use the median (instead of a simple mean) of the top performing 12 percent to set the MACT
floor. Furthermore, as an indication of the achievability of the technology over the long-term,
EPA chose to rely on State-imposed permit limits (in conjunction with emissions test data
showing that those permit limits are representative of actual performance) in arriving at the
MACT floor emission limit.
In test data cited by the commenter, the three-run averages for two sets of emissions tests
for the kiln used to set the new source MACT PM limit, are below (0.079 and 0.091 lb/tsf) the
proposed PM limit of 0.1 lb/tsf for new lime kilns. The commenter noted that one of the test
3-18
-------
runs was at the proposed 0.1 lb/tsf PM limit and that the proposed 0.1 lb/tsf limit was, therefore,
inappropriate. (It is, of course, reasonable for EPA to establish a standard based on the same
methodology that will be used for complying with that standard. Cf. Chemical Waste
Management v. EPA. 976 F. 2d 2, 34 (D.C. Cir. 1992) (facially reasonable to require compliance
based on methodology used to establish the standard). EPA notes that compliance with emission
limits is normally based on a three-run average which can accommodate occasional elevated
results as long as the average is at or below the established limit. Furthermore, the emission test
results for five of the six top performing kilns were 0.0091, 0.013, 0.026, 0.027, and 0.091 lb/tsf
and are below the 0.10 lb/tsf PM emissions limit. These results adequately account for operating
variability and indicate that any new kiln using well designed and operated control devices can
meet the 0.1 lb/tsf limit. Based on this, we see no basis to conclude that a 0.10 lb/tsf PM
emission limit is not achievable or appropriate.
Comment
One commenter (IV-D-06) claimed that the proposed NESHAP will require the
replacement of their two wet scrubbers with baghouses. They claim there is no space for FF
retrofit, and that converting to baghouses will trigger prevention of significant deterioration
(PSD) non-attainment review due to increased S02 emissions.
Response
While we recognize that a facility may (or may not) have site-specific space restrictions,
we have, on average, adequately accounted for these factors by incorporating cost analysis
factors to account for retrofit and equipment demolition. We have also allowed existing
facilities 3 years to comply with the final NESHAP. This will provide sufficient time for
facilities to replace or upgrade existing equipment during scheduled outages. The averaging
provisions in the final NESHAP also provide facilities with additional flexibility concerning
replacement or upgrade of existing equipment.
Requiring an existing kiln equipped with a wet scrubber to upgrade their PM controls
will not necessarily trigger new source review (NSR). First, we have written the final rule to
allow existing kilns equipped with wet scrubbers to meet a 0.60 lb/tsf emission limit, rather then
the 0.12 lb/tsf PM emission limit. No commenters have raised any cost or technical issues to
indicate that a PM emission limit of 0.60 lb/tsf will force any facilities to choose to convert
existing kilns from wet scrubbers to dry controls. Therefore, the facility can choose to replace or
upgrade their existing scrubbers, which means there will be no S02 (or other collateral pollutant)
emissions increase to trigger NSR requirements. Second, if they choose to use a baghouse, they
may be able to avoid NSR by qualifying for a pollution control project exclusion (67 FR 80186).
3-19
-------
3.4 PM Emission Limit - General
Comment
One commenter (IV-D-14) believes the particulate matter emission limits proposed for
lime manufacturing kilns and coolers do not represent the maximum achievable control
technology and are much less stringent than the limits actually required by the Clean Air Act.
The commenter notes that the proposed rule discredits performance test data which demonstrate
that particulate emissions of less than half the proposed standard for existing plants are routinely
achieved by claiming they may not be consistently achievable, but EPA has provided no
statistics. The commenter claims that EPA has chosen instead to base the standards on permit
limits, but has selectively eliminated from consideration those permits calling for stringent
controls which are currently in place. The commenter gives the examples of Continental Lime
which is in compliance with a BACT limit for PM emissions of 0.05 lb/ton limestone, and
Western Lime which is in compliance with a permit limit for PM emissions of 0.06 lb/ton
limestone. The commenter states, "Considering just these plants, a recalculation of the existing
plant limit leads to a standard of 0.09 to 0.11 lb/ton. The new source MACT floor should be no
less stringent than 0.05 lb/ton. However, if EPA chooses to establish a MACT floor on the basis
of permit limits, it should consider all of the permit data in its possession (and available to it
through delegated State agencies). It is likely that a lower emission limit would represent the
MACT floor and that still lower beyond-the floor alternatives would result."
The commenter notes that if performance data do not represent achievable emission
limits, EPA should consider design standards based on air-to-cloth ratios. The commenter also
believes the proposed particulate emission limits for grinders, conveyors, and bins are also based
on data which overstate emissions (in nearly all cases) and do not represent MACT. The
commenter believes EPA should examine actual performance test data test or actual permit
limitations.
Response
The kilns to which the commenter refers are Graymont Western Lime (formerly
Continental Lime in Broadwater County, Montana and Western Lime in Green Bay, Wisconsin.
The permit limits cited by the commenter were apparently reported on EPA Technology Transfer
Network (TTN) website. We contacted the Montana Department of Environment and found that
the limit for the kiln in question is actually 0.5 lb/tsf and not 0.05 lb/tsf as reported on the TTN
website. Also, the complete permit for Western Lime was located on the Wisconsin Department
of Natural Resources website, which showed the permit limit for the kiln in question as being
0.12 lb/tsf rather than the 0.058 lb/tsf as reported on the TTN website. Based on the correct PM
permit limits for these two lime sources, our conclusions regarding MACT PM limits for
existing and new sources are still appropriate. (As the response to a previous comment shows,
these permit limits are also representative of actual performance.)
3-20
-------
The floor for grinders, conveyors, and bins are based on the existing new source
performance standards. We have no data to support a different floor.
3.5 Monitoring Requirements for Scrubber-Equipped Kilns
Comment
Several commenters (IV-D-04, IV-D-06, and IV-D-22) comment that the rule should
allow the use of scrubber pump amperage and other indirect measures of liquid flow rate.
Commenter IV-D-04 takes issue with EPA's claim that pressure drop and liquid flow rate are
most often monitored scrubber parameters. The commenter states that liquid flow rate is not the
most common scrubber parameter measured, in part, because the NSPS for lime manufacturing
requires the measurement of the liquid supply pressure to the control device. The commenter
also states that most permits for lime plants do not require that flow rate be measured directly,
and provided permit requirements at existing plants to demonstrate this. They note that of the 8
lime plants that do have flow-related permit limits, 5 of them are provided the option of
monitoring scrubber pump amperage instead of flow rate. The commenter also notes that of the
four example CAM plans for scrubber-equipped kilns in EPA's CAM guidance, none includes
liquid flow rate among the parameters to be monitored. The commenter recommends that the
rule allow liquid flow rate to be measured directly (e.g., with a flow meter) or indirectly (e.g., by
monitoring scrubber exhaust temperature).
One commenter (IV-D-04) believes the rule should allow the use of parameters other
than gas pressure drop if pre-approval from the permitting authority is obtained. The commenter
states that monitoring gas stream pressure drop is not appropriate for all kiln types. They use the
example of double shaft kilns, for which scrubber pressure drop is measurable only during the
fuel firing cycle, returning to zero during the reversing cycle. In addition, the commenter notes
that, "like the MACT standard for lime kilns at pulp and paper mills, the lime MACT should be
revised to allow the use of parameters other than gas pressure drop if pre-approval from the state
permitting agency is obtained."
Response
The first commenter is correct in stating that the NSPS requires that the liquid supply
pressure to the control device be monitored and not liquid flow rate. However, we note that
scrubber-equipped lime kilns in other industries have requirements to monitor liquid flow rates
to the scrubber. For example, the NESHAP for chemical recovery combustion sources at kraft,
soda, sulfite, and stand-alone semi-chemical pulp mills requires that lime kilns (used in
reprocessing black liquor at pulp and paper facilities) which are equipped with a wet scrubber
must continuously monitor the pressure drop as well as the scrubbing liquid flow rate. (See 66
FR 3196)
3-21
-------
The rule as proposed does allow the owner or operator to monitor alternative control
device operating parameters subject to approval by the Administrator. Under the proposed rule,
if you wish to request the use of an alternative monitoring method, you must submit a request
following the procedures in General Provisions section 63.8(f). In order to make this more
obvious, we explicitly cross-reference this provision in the monitoring requirements of the rule.
However, major modifications to monitoring requirements are not delegated. Therefore,
decisions on this issue must be made on a case-by-case basis by the Administrator as set out in
63.8(f).
3.6 Monitoring Requirements for Kilns With Dry Air Pollution Control Devices
Comment
Several commenters (IV-D-04, IV-D-17, and IV-D-16) request that the proposed
requirements in § 63.7113(d)(1) and (e)(1) that BLD and particulate monitoring devices (PMD)
be certified by the manufacturer to be capable of detecting PM at or below 10 mg/m3 should be
deleted because no specifications exist on which to base the certification. Commenter IV-D-04
notes that unlike COMS, there are no technical specifications, regulations, or checks and
balances to verify claims made by the BLD or PMD vendor or manufacturer.
Response
We disagree. The EPA is currently aware of instruments certified to meet levels of 0.2
mg/m3. Specifications on which to base certifications are available in EPA guidance documents.
Comment
One commenter (IV-D-04) states that the requirements in §63.7113(d)(6) and (e)(6) that
BLD and PMD be "installed, operated, adjusted, and maintained so that they follow the
manufacturer's written specifications and recommendations" should be revised to require
adherence to the manufacturer's written specifications and operating procedures (rather than
"recommendations") except where other procedures are determined to be more appropriate and
are described and justified in the plant's operations, maintenance and monitoring (OM&M) plan.
The commenter noted that there is no industry-based consensus or EPA specifications, and
"there certainly will be applications where the manufacturer's procedures or recommendations
may not be appropriate or where a better approach can be used."
3-22
-------
Response
We agree that there may be cases where other procedures are more appropriate then the
manufacturers recommendations. For this reason, we allow alternative monitoring under the
procedures in 63.8 (f). However, in general, we believe adherence to written specifications and
recommendations is more appropriate. Therefore we have retained the requirements in
§63.7113(d)(6) and (e)(6) in the final rule.
Comment
Several commenters (IV-D-04, IV-D-18, IV-D-17, IV-D-05, IV-D-16, IV-D-13, and IV-
D-20) believe COMS should be retained as a monitoring option. Commenters IV-D-04, IV-D-
17, and IV-D-16 state that COMS are sensitive enough to detect changes in fabric filter
performance at concentrations corresponding to the proposed PM limit and, unlike BLD, are
subject to specifications that enable third party evaluation of their performance. Commenter IV-
D-04 notes that COMS have been used to detect relative changes in particle mass loading for
many years at lime plants. This commenter also states that, "actual COM data acquired at lime
kilns with high efficiency filters demonstrate that COM data are appropriate and effective for
detecting gradual changes in fabric filter performance over time." According to this commenter,
EPA has suggested that BLD are preferable to COMS because they are more sensitive, but that
all lime kilns currently equipped with COMS are capable of meeting the sensitivity requirement
proposed for BLD.
Commenter IV-D-04 also noted that although the presence of bias makes any
determination of absolute opacity below 10 percent inaccurate, "it has no affect on the ability of
a COMS to detect changes in opacity at a particular facility — and hence changes in control
equipment performance over time. Detecting relative changes in mass loading is all that BLD
are capable of doing (preamble at 78062). COMS are more than adequate to detect changes in
the opacity level necessary to detect minor variations in fabric filter or ESP performance well
before catastrophic failures occur."
In addition, several commenters (IV-D-04, IV-D-18, IV-D-17, IV-D-05, and IV-D-16)
state that requiring only BLD would create unnecessary and duplicative monitoring since other
CAA rules and permits require COMS. Since the CAM rule requires that if a COM is required
under the CAA or state or local law, it must be used to satisfy the requirements of the CAM rule,
requiring BLD or PMD would result in duplicative monitoring.
Response
EPA proposed that COMS be allowed as an alternative to BLD, PM detectors, and
electrical current and voltage readings for ESP for monitoring continuous compliance with the
proposed PM emission limits for lime kilns and coolers. EPA also solicited comments on
3-23
-------
requiring the use of PM detectors and BLD in lieu of COMS. In response, several commenters
stated their support for retaining COMS as a monitoring option. The commenters argued that
COMS have a long history of use by the lime manufacturing industry and that they are effective
for detecting changes in control device performance over time. As a result, EPA sees no reason
to not allow the use of COMS.
The commenters also claimed that requiring BLD or PM CEM would be redundant
because the CAM rule requires COMS. The CAM rule only applies where a control device is
used to comply with an emission limit proposed prior to November 15, 1990. Therefore, the
CAM rule will not apply to monitoring requirements related to compliance with the Lime
Manufacturing NESHAP.
Comment
One commenter (IV-D-04) commented that EPA should clarify that only COMS
installed, relocated, or substantially refurbished after February 6, 2001 must comply with the
version of PS-1 revised on August 10, 2000. Because many COMS at lime plants were certified
to meet the PS-1 requirements in effect at the time of installation, the commenter states that it is
"impractical and cost prohibitive to remove these units from service for several months and
return them for testing to the COM manufacturer. The cost to re-certify a single monitor would
exceed the cost of a new monitor due to the test regime specified in the ASTM standard."
Response
The proposed rule states that you must install, operate, and maintain COMS as required
by 40 CFR part 63, subpart A, General Provisions and according to PS-1 of Appendix B to part
60. The current version of PS-1 itself specifies that it only applies to COMS installed, relocated,
or substantially refurbished after February 6, 2001. However, this could result in the rule being
unclear on the requirements for COMS installed on or before February 6, 2001. Therefore, we
have clarified the rule to state that COMS installed, relocated, or substantially refurbished after
February 6, 2001, must meet the requirements of PS-1 as revised on August 10, 2000. Any
COMS installed on or before February 6, 2001, should continue to meet the requirements in
effect at the time of installation unless the local regulatory agency specifically required
recertification of the COMS in question.
Comment
One commenter (IV-D-19) states that opacity does not correlate to PM mass emissions.
The commenter notes the EPA has stated on several occasions that STET systems (COMS) can
determine opacity, but COMS cannot determine PM emissions. And if particle density changes
but the particle size remains the same, opacity will not change while the mass emission rate will
change in proportion to the density change. The commenter agrees that PM is a technically
3-24
-------
sound surrogate for HAP metals, but disagrees that opacity serves as a surrogate for HAP metals
as stated in the proposal preamble.
The commenter does not believe that a COMS can be used to evaluate the continuous
compliance status of kilns, coolers, or materials processing operations (MPO) that have a mass
emission limit. The commenter is not aware of any data that shows a definitive link between
opacity and mass emissions except in very limited and controlled situations. In addition, the
commenter does not understand how a 15 percent 6-minute average opacity limit can be
correlated to a 3-hour rolling average PM emission limit of lb/ton of stone feed.
The commenter states, "the error in using a COMS to determine compliance with the PM
mass emission limit can be demonstrated from the data presented to you by the National Lime
Association. If one of the top performing lime kilns has a measured emission rate well below the
0.12 lb/ton of feed emission limit, but has a reportable opacity between 10 and 15 percent, there
can be no correlation."
The commenter believed a better alternative is to use a monitor that measures PM mass
emissions in units that are directly related to the mass emission limit. The EPA should consider
using a PM continuous emissions monitor (CEM) for continuous compliance determination of
kiln, cooler, and MPO emissions with or without a PM CEM performance specification. The
commenter notes that EPA's stated reluctance to use a PM CEM in the absence of performance
specifications is inconsistent with the remainder of the standard, since the use of bag leak
detector systems (BLDS) and a PM detector are proposed without performance specifications.
The commenter also notes that an extractive type PM CEM designed to operate in wet exhaust
streams would provide a direct indication of compliance for wet scrubbers.
Response
We agree that a COMS cannot directly measure PM emissions. However, we believe
that a properly calibrated and maintained COMS is sufficient to demonstrate long term PM
control device performance. The purpose of the monitor is to demonstrate, with reasonable
certainty, that the PM control device is operating as well as it did during the PM emission test
used to demonstrate compliance. Though we agree that a continuous monitor that directly
measures PM emissions would provide direct evidence if the PM emission limit is exceeded, we
have no data to demonstrate that this would result in additional PM emissions reductions
compared to monitoring the control device with a COMS. Therefore, requiring facilities to
remove their currently installed COMS and replace them with a PM CEM only adds additional
costs with no quantitative environmental benefit. We also note that we know of no lime kiln that
currently has an operating PM CEM.
Comment
3-25
-------
One commenter (IV-D-19) provided comments on the cost of PM CEM. The commenter
provides cost data for several monitor technologies, and notes that the list price is only part of
the cost of installing and using a PM CEM.
Response
We thank the commenter for this additional data. However, because we are not requiring
PM CEM as a monitoring option, we are assuming that facilities will opt to use a COMS due to
the lower costs. Because most existing facilities already have a COMS installed, we have not
revised our rule costs to incorporate this data.
Comment
On commenter (IV-D-19) provided comments on the relation of a PM CEM requirement
to the proposed PM emission limits. The commenter sees no reason to change the proposed level
or averaging time of the PM emission limit if a CEM is used. The commenter suggests having a
lime manufacturing plant purchase, install, and operate a PM CEM to monitor emissions over a
period of time (e.g., 6 months) to confirm the proposed limit and averaging interval.
Response
Since we believe that the monitoring options of the proposed rule are adequate to show
good operation of emissions control devices, it is not necessary to perform testing using a PM
CEM.
Comment
One commenter (IV-D-04) states that PM CEM should not be required in the lime
MACT because EPA has no data on which to establish an achievable CEM-based PM limit.
Response
We disagree. Because PM CEM directly measure the PM emission rate, the appropriate
limit would be 0.12 lb/tsf (existing sources) or 0.10 lb/tsf (new sources). However, we are not
requiring PM CEM, they are an option.
Comment
3-26
-------
Commenter IV-D-19 believes that EPA needs to include beta gauge and optical
scintillation along with light scattering technology for the PM detector.
Response
The final rule allows use of PM CEM as an option. Should PM CEM be used, they must
meet appropriate performance specifications. Any type of PM CEM, including those that rely on
beta gauges or optical scintillation, that meets the performance specifications may be used.
Comment
Several commenters (IV-D-04, IV-D-18, IV-D-06, IV-D-11, and IV-D-09) believe EPA
Method 9 should be allowed for positive pressure baghouses. According to commenter IV-D-04,
the BLD Guidance Document recognizes that requiring BLD would be very costly, and states
that the document does not apply to this type of baghouse (BLD Guidance Document, pg 2).
This commenter gives the example of a small business (Austin White Lime) that would be
required to have a BLD for each of the 8 compartments in its baghouse under the proposed rule,
and whose Title V permit allows Method 9 monitoring for the baghouse. According to
commenter IV-D-09, the associated costs of installing a separate BLD or PM CEM sensor on
each discharge or new common stack could easily exceed $1,000,000. The commenter notes
that, "baghouse pressure differential readings, together with fan amperage and daily visible
emission notations (Method 22) will provide the necessary performance assurance with ample
and timely indication of baghouse failures or malfunctions."
Response
We acknowledge that there are precedents for the use of alternatives to COMS, BLD, and
PM CEM on positive pressure baghouses that have multiple stacks. The NESHAP for portland
cement, an industry that has similarities to the lime manufacturing industry, allows the use of
method 9 opacity monitoring for kilns having control devices with multiple stacks. For existing
lime kilns controlled by positive pressure control devices having multiple stacks, the final rule
provides the option of using method 9 for daily opacity monitoring. We are also revising the
rule to allow Method 9 monitoring if the control device vents through a monovent, or if
installation of a COM in accordance with PS-1 is infeasible.
Comment
One commenter (IV-D-04) believes that COMS should be used as the monitoring option
for ESP. The commenter claims that using PM detectors on ESP is problematic, as is parametric
monitoring if the ESP is over-designed such that operating at or above the voltage levels
measured during the performance test is not needed to maintain compliance with the PM limit.
3-27
-------
One commenter (IV-D-19) comments that there are no data to support the use of ESP
operating parameters to determining compliance. The commenter notes that until the lime
industry collects this data in three-hour rolling averages for extended periods of time in a
reportable format that can be analyzed by others, no conclusions can be made regarding the
usefulness of ESP operating parameters in determining compliance or mass emissions.
Response
We agree with the commenters that COMS and PM CEM should be sued to monitor the
performance of an ESP. Therefore we have modified the final rule ro require a COMS or PM
CEM to be used to demonstrate continuous compliance. Because we are now requiring COMS
or PM CEM, monitoring ESP voltage and current is not be required.
Comment
One commenter (IV-D-09) believes that stack testing on positive pressure baghouses
should not be required. The commenter notes that the eight stacks at each of its positive pressure
baghouses are nearly inaccessible and too short to be properly tested, and "attempting a
performance test ...would be dangerous without significant additional improvements..., " and,
"the cost of all of these improvements could exceed $1,000,000." The commenter believes
positive pressure baghouses with multiple inaccessible stacks should be given the option to
perform a 3-hour Method 9 performance test when operated in conjunction with performance
monitoring.
Response
EPA regulations generally require, at a minimum, initial performance testing to
demonstrate compliance with emission limits for affected sources. Further, the General
Provisions section 63.7(d) require sources to provide adequate testing. Positive pressure
baghouses are used in many of the industries that are subject to EPA regulations and are required
to conduct performance testing in order to demonstrate compliance. EPA has developed
procedures for testing PM emissions from various configurations of positive pressure baghouses,
which are contained in Method 5D of 40 CFR part 60, appendix A. Therefore, we have decided
not to alter the compliance requirements for sources using positive pressure baghouses.
Comment
One commenter (IV-D-19) does not understand how an alarm level will be set for a PM
detector or a BLD system that will have any correlation to a PM emission limit in units of lb/ton
of stone feed. The commenter states that an extensive amount of testing over varying conditions
at or near the limit would be needed to establish the response of BLD or PM detectors as a
3-28
-------
function of actual PM emissions. The commenter also notes that EPA's bag leak detector
guidance does not provide such procedures.
Response
We have not stated that BLD limits and PM limits correlate. BLD are calibrated at the
time the PM emission test is run. The BLD then detects changes in performance verses
performance during the PM emission test.
3.7 Operating Parameters
Comment
One commenter (IV-D-04) states that a single excursion from operating parameters
recorded during a 3-hour compliance test should not constitute a violation. The commenter
believes that, "the NSPS kilns are the lime industry's top performers, and their monitoring
regime should be the benchmark against which monitoring under the MACT rule is prescribed."
Since a violation under the NSPS does not occur unless the parameter is greater than 30 percent
below the rates established during the performance test, the commenter recommends a 30
percent "buffer" between the permit limit and the 3-hour average recorded during the compliance
test. Or, "alternatively, like the Pulp and Paper MACT, the rule should specify that a violation
of the standard does not occur unless 6 or more 3-hour average parameter values are recorded
outside the established range within the 6 month reporting period."
The commenter notes that EPA's CAM guidance document states, "Use of only 3 hours
of parameter data may not be sufficient to fully characterize parameter values during normal
operation." The commenter also noted that language in the proposal preamble cautions against
developing enforceable emission standards based on 3-hour compliance tests. The commenter
also notes that none of the CAM plans for scrubbers base a permit limit on the 3-hour average
reading that occurred during a compliance test, and two of the plans allow a 15 percent buffer to
account for variability.
The commenter provided gas pressure drop readings and concurrent PM test data for
three kilns, and noted that for each of them, gas pressure drop during one or more 1-hour runs
was below the proposed 3-hour average. The commenter stated that under the proposed rules,
these readings below the 3-hour average would constitute a violation.
The commenter also believes the rule should provide an exemption from the PM
emission limit during performance testing. The commenter stated, "plant operators may need to
conduct a series of performance tests to determine the minimum pressure drop and liquid flow
3-29
-------
rate levels that will assure compliance for each set of operating conditions used for a particular
kiln. Results for these tests are not available until post-test laboratory analyses are completed."
Response
Each owner/operator is required to define the compliance parameters to be monitored in
their OM&M plan. Then, during the initial performance tests they are required to monitor and
establish the value or range of the parameters. EPA believes that the initial performance tests
will be designed to give the owner or operator flexibility in adhering to the established range for
monitored parameters. These values must be reported in the results of the test and in the
notification of compliance status to the permitting authority and must be approved by the
permitting authority. During subsequent operations, if the monitored parameters exceed the
values or fall outside the range determined during the initial performance test, it is a violation of
the operating requirements of the standard (unless it is the result of a malfunction to which the
facility responds to in accordance with the startup, shut down, and malfunction [SSM] plan). We
have has no basis for allowing deviations before considering the facility to be in violation. The
owner/operator has ample opportunity to establish a range for the operating parameters and must
thereafter operate within that range.
Most operating parameters are calculated as 3-hour averages. This is generally consistent
with performance test times. Thus, a one-hour period of insufficient gas pressure drop would
not, by itself, be considered a deviation.
Comment
One commenter (IV-D-04) stated that the accuracy criteria in Table 4 (items 8, 9, 13, and
14) for devices used to measure operating parameters should be deleted. The commenter notes
that the lime MACT is unique because it includes both accuracy and minimum tolerance criteria
for measurement devices; other recent MACT rules contain only tolerance criteria. In addition,
the commenter states that EPA has offered no explanation for the proposed accuracy criteria in
Table 4. The commenter believes it is the precision of a device (i.e., the value of a parameter
relative to what was recorded during the performance test) that is important in the context of the
lime MACT, not its accuracy.
Response
We disagree. We commonly include accuracy criteria in MACT standards for devices
used to measure operating parameters. We continue to believe that monitoring devices of known
accuracies should be used.
3.8 Opacity Limit
3-30
-------
Comment
In response to EPA's request for comments on the appropriate opacity limit (EPA is
considering an opacity of 10 to 15 percent), several commenters (IV-D-04, IV-D-18, IV-D-17,
and IV-D-16) agree that the opacity standard for lime kilns should be 15 percent as proposed.
One commenter (IV-D-04) provided additional data in the form of opacity data from four kilns.
According to this commenter, the opacity data for selected kilns (Docket Item No. I-B-125) are
not reliable for establishing an opacity standard because they are from visible emission data
collected for brief periods of time under poor viewing conditions.
Response
Based on information considered prior to proposal as well as additional information
supplied by commenters, EPA is retaining the 15 percent opacity limit for sources controlled
using FF and ESP. Information considered by EPA in proposing the opacity limit suggested that
the average opacity permit limit of the top performing lime kilns was 15 percent. Information
provided by the commenters supporting the proposed opacity limit indicated that opacity levels
may vary between 10 and 15 percent even for well operated and maintained control devices. No
information supporting a more stringent opacity limit than the one proposed was submitted by
any commenter. Therefore, EPA finds no justification for a limit more stringent than the
proposed 15 percent.
Comment
Several commenters (IV-D-04, IV-D-18, IV-D-17, IV-D-07, IV-D-05, and IV-D-13)
state that the rule should specify a time period during which opacity readings greater than 15
percent are not considered a violation. One commenter (IV-D-04) requests at a minimum that the
rule state that opacity readings greater than 15 percent for less than 1 percent of the reporting
period are not considered to be a violation.
One commenter (IV-D-07) noted that it operates two of the top six performers in the
industry, and it is impossible not to have occasional readings that would be violations if there
were no allowances for them. The commenter's State permits allow 1 percent of operating time
per quarter to exceed the opacity limit.
Commenter IV-D-05 requests that the final rule state that a particulate matter violation
has occurred if: (1) opacity readings exceed the 15 percent opacity limit for 6 percent or more of
the operating time in a quarter or, (2) opacity readings exceed the 15 percent opacity limit for 20
total hours in a quarter, whichever time frame is longer.
3-31
-------
Commenters IV-D-04 and IV-D-05 refer to the Pulp and Paper MACT as an example of
an existing rule with such an exemption.
Response
We find no justification to support allowing excursions above the 15 percent opacity
limit. We believe that well operated and maintained control devices will typically operate at
opacity levels much lower than 15 percent. Other NESHAP, including the portland cement
NESHAP, contain opacity limits from which no exceedances are allowed.
In response to the commenters' concerns about occasional excursions above the opacity
limit, there are times when opacity levels above 15 percent are not considered to be a violation of
the rule. These include periods when a control device malfunctions, or is in a period of startup
or shutdown. If opacity levels exceed 15 percent as a result of a control device startup,
shutdown, or malfunction, it would not be considered a violation of the opacity limit. The same
is true during periods when a monitoring system malfunctions or is being calibrated; provided
the startup, shutdown, and malfunction (SSM) plan is followed. Information supplied by one
commenter showed opacity readings for several kilns over several days. Nearly all of the
readings were well below the 15 percent limit with just a few exceptions for each kiln. The
commenter that supplied the opacity readings was asked to supply additional information
regarding the opacity excursions above 15 percent. In each instance, the high opacity reading
was explained by a startup, shutdown, or malfunction of the control device or by a
malfunctioning monitor or a monitoring system that was undergoing calibration, none of which
would be considered a violation of the opacity limit. As a result, we believe that well operated
and maintained control devices can meet the opacity limit, and that occasional excursions above
the limit due to control device or monitoring system malfunction will not be a violation of the
rule provided the SSM is followed.
3.9 Stone Handling Operations
Comment
Several commenters (IV-D-04, IV-D-17, and IV-D-11) request that the terms "raw
material processing" or "material processing operation (MPO)" be replaced with the term
"processed stone handling" since the only ingredient used to make lime is processed limestone.
Commenter IV-D-11 noted that the use of "raw material" could be construed to include
unprocessed stone and thus potentially include stone handling operations in the quarry.
Commenter IV-D-09 notes "the exclusion of storage piles from this proposed NESHAP
is consistent with other standards such as the Portland Cement Manufacturing MACT, Integrated
Iron and Steel Manufacturing MACT and the Non-metallic Minerals Processing Plants (Subpart
3-32
-------
000) NSPS." The commenter believes that storage piles prior to any processing operation are a
continuation of raw material handling and should be exempt from the MACT.
Commenter IV-D-11 also requests that the rule clarify that material stockpiles are not
regulated. The commenter notes that this exclusion has been granted in other MACT's such as
the Cement MACT, and the nonmetallic mineral processing NSPS.
Response
The rule has been revised to clarify that the first emission unit subject to the rule is
"processed limestone storage" immediately before the kiln, and we have changed the term
"materials processing operation" to "processed stone handling" in the final rule.
Comment
Several commenters (IV-D-04, IV-D-17, IV-D-16, and IV-D-11) commented that the
definition of "storage bin" in § 63.7143 should be revised to clarify that bins are manmade
enclosures used to store processed stone feed for the kiln. Commenter IV-D-04 noted that this is
consistent with the definition in the cement MACT, and that it clarifies that storage piles are not
subject to the rule. Commenters IV-D-04 and IV-D-16 recommend the following amended
definition: "A manmade enclosure for storage of processed limestone prior to its use as lime kiln
feed."
Response
The rule has been revised to clarify that storage bins refer to manmade bins for storage of
processed limestone.
Comment
One commenter (IV-D-04) believes the rule should not require the use of proposed
Methods 203 A, 203B, and/or 203C to measure opacity from fugitive sources. The commenter
notes that the preamble does not state why Method 9 is unsuitable, "nor does it specify how
Methods 203 A,B, or C would be used and what the corresponding opacity limit would be.
Response
We requested comment in the proposed rule on the use of proposed Methods 203 A,
203B, and/or 203C to measure opacity from fugitive sources. We never stated that Method 9
3-33
-------
was unsuitable. We were concerned that because Method 9 was developed to measure emissions
from stacks, and an alternative method might be more suitable.
We have decided not to allow the use of proposed Methods 203 A, 203B, and/or 203C in
the final rule. However, we have added additional language to the final rule to explain how to
use Method 9 to determine opacity from fugitive sources. This language was taken directly from
40 CFR 60, subpart OOO, Standards of Performance for Nonmetallic Mineral Processing Plants.
Comment
Two commenters (IV-D-04 and IV-D-17) request section 63.7120(b) be revised to clarify
that periodic (not continuous) monitoring is required for stone handling units.
Response
EPA has revised the rule to clarify that stone handling operations are generally subject to
monthly visible emissions testing and not continuous monitoring.
3.10 Test Methods
Comment
Several commenters (IV-D-04, IV-D-05, and IV-D-08) asked that EPA delete the bias
and precision criteria for HC1 test results using ASTM Method. Commenter IV-D-04 states, "the
commercial lime industry has invested substantial resources to develop a method to address
EPA's concern that Method 26 understates HC1 emissions. An ASTM committee has concluded
that Method D 6735-01 generates data that has no statistically significant bias. It is arbitrary and
unreasonable for the Agency to burden the commercial lime industry - half of which is
comprised of small businesses - with a requirement to conduct analyte spiking to prove that HC1
test results are unbiased. Likewise, requiring paired runs to comply with a precision criteria is
not warranted because (1) EPA has never raised any concerns about the precision of impinger
generated HC1 data, and (2) ASTM Committee D22 on Sampling and Testing of Atmospheres
approved the impinger-based method after evaluating the precision of the data generated with the
method. Accordingly, the bias and precision criteria for HC1 test results obtained using ASTM
Method D6735-01 should be deleted from the rule." Commenter IV-D-05 notes that these
criteria are not found in the ASTM method.
Response
3-34
-------
The ASTM validation proved that, when ASTM Method D6735-01 is conducted
properly, it is possible that data are not biased. However, because the ASTM method lacks
precision, the power of this conclusion is not very strong. The validation does not assure that
future applications of the method run by another group will also be unbiased. Sampling for HC1
in a lime plant environment is difficult. The purpose of analyte spiking is to assure that a
representative sample is transported from the stack through the probe and filter to the impingers
regardless of who is performing the test. Analyte spiking is an option in the ASTM procedure;
because of the importance of the decision made as result of this data, we chose to exercise that
quality assurance option. We agree with the commenters that analyte spiking is a complex
procedure. However, it is a requirement of the Method 320 instrumental FTIR method which we
believe is superior. There is no reason to allow less quality assurance for the inferior method.
However, we agree with the commenter that for the purpose of determining major/area status, it
is not necessary to reject a set of data if the apparent bias does not change the major/area source
determination.
With respect to the commenters' suggestion that the precision criterion be deleted, EPA
disagrees. It is an option in the ASTM procedure which EPA chose to require. Reasonable
agreement between paired samples is considered a strong indication that the method is being
consistently operated, recovered and analyzed.
One commenter (IV-D-04) cited section 6.3.1.2 (of EPA Method 301) as the EPA
precision criteria for an acceptable alternative method. They mistakenly thought the proposed
rule would impose a much more stringent precision criteria (i.e. 50 percent RSD). In fact, when
an acceptable method already exists (as in this case), Section 6.2 of EPA Method 301 applies,
which requires a candidate alternative method to be at least as precise as the method for which it
is intended to be substituted. However, because the industry is reluctant to embrace the use of
the acceptable method (EPA Method 320 or 321), we waived the requirement that the alternative
method to be at least as precise (as allowed by section 1.1.1 of EPA Method 301). Therefore, in
order to allow the use of the less precise ASTM method, we instead chose to have the testers
demonstrate their ability to get reasonable agreement at the actual facility. Note 3 in section
4.1.4 of the ASTM method reinforces this approach in that it states "The particulate matter of
mineral calcining facilities adsorbed HC1 and HF to varying degrees. The amount of adsorption
depends on process parameters and the physical/chemical properties of the dust."
Comment
Commenters IV-D-05 and IV-D-08 note that ASTM Method D6735-01 includes optional
paired train and spiking procedures. The commenter is unclear what basis EPA has used to make
these optional procedures mandatory.
Response
3-35
-------
We disagree. Exactly because it is an option to use paired trains and analyte spiking
procedures in the ASTM method, EPA clarified that for this application those options were
exercised. Section 5.4 of the ASTM method states "the test specific precision may be determined
by conducting paired runs." Paired runs aid in identifying possible suspect data and provide
backup in the event one train is invalidated. Performing paired runs depends on the test specific
data quality objectives. Section 5.3 has a similar statement regarding analyte spiking. In the
bigger picture, if the only options available to us were to accept or reject an ASTM method with
no alterations, we would have to reject many more ASTM methods for relatively easily corrected
reasons. This would severely limit our good-faith implementation of the National Technology
Transfer and Advancement Act.
Comment
Commenter IV-D-08 notes that there is no discussion in the proposal preamble
explaining where the 20 percent limit for absolute relative standard deviation (RSD) came from.
The commenter requests that the requirement be deleted, since the ASTM method does not
mention criteria for accepting or rejecting a test, and since the ASTM results are completely
consistent with EPA's own RSD criteria of 50 percent. If EPA insists on including the paired
train requirement and an associated RSD criterion, the commenter believes a ppm difference
should be considered, particularly for low concentration sources. The commenter suggests that
if the RSD exceeds 50 percent, the results continue to be valid provided the absolute difference
between the two train values is 3 ppm or less.
Commenter IV-D-05 requests that the ASTM method be incorporated by reference as it is
written.
Response
We agree with the commentator that it may not be necessary to reject data with poor
precision if it doesn't affect the major/area source decision. However, we disagree that the way
to attain this goal is to provide an absolute value that is acceptable regardless of how close to the
action level the data are.
Comment
Commenter IV-D-08 stated that the regulation would require four valid runs, which are
too many, noting that nearly all methods require three runs and not four. The commenters states
that the ASTM method defines the test to be three or more runs and that EPA should have the
same wording. Commenter IV-D-05 asked that if a spiking procedure and spike recovery
criterion are included in the final rule, that we confirm that the three test-run requirement of the
ASTM method is met with two non-spiked runs and a third, spiked run.
3-36
-------
Response
As discussed above, EPA disagrees that we are confined to exactly what the ASTM
method states. In this case the ASTM method merely sets a minimum of three. However, in this
case EPA, agrees that three or more runs may constitute a test.
Our intent is that two of the three runs be made with paired trains, and one run be made
with paired trains where one train is spiked.
Comment
One commenter (IV-D-10) believes ASTM D6735-01 to be a "primitive" or inferior
measurement technique, and EPA Method 320 (FTIR) to be a clearly superior method for HC1
measurements. The commenter notes that EPA Method 320 directly measures HC1, while
D6735-01 measures total chloride. There is no need to conduct indirect measurements, since
there is a reliable test method that directly measures HC1.
The commenter states that EPA Method 320 has a typical RSD of approximately 1 to 2
percent compared to 30 to 50 percent RSD for D6735-01. The commenter provides cost data
for EPA Method 320, and notes that the methods (EPA Method 320 and ASTM D6735-01) have
virtually identical costs. The commenter urges EPA to require the use of EPA Method 320 for
lime HC1 measurements.
Response
We agree that FTIR is the best choice of test method to measure emissions of HC1, and
note that the cost estimate provided by the commenter is close to the EPA estimate of cost to
perform the ASTM method. However, we feel that the EPA can use data from a properly
conducted ASTM D6735-01 test to determine if the source qualifies as an area source. We agree
that there is a potential to incorrectly conclude that the source is major, when in fact, it is not.
3.11 Economic Impact Analysis
Comment
One commenter (IV-D-06) claimed that the econimic impacts analysis (EIA) neglected to
include some significant costs of implementing the proposed rule, including the cost of
dismantling existing equipment, lost sales during downtime, and the cost of re-hiring personnel
after plant modifications if scrubbers must be replaced. The commenter also noted that
3-37
-------
maintenance and supervisory personnel currently do not work evening and weekend shifts, but
would likely be required in the event of failure of the recommended monitoring equipment.
Another commenter (IV-D-02) believes EPA's estimated $1.17 per ton of lime cost
estimate is low, and the cost to a typical lime producer will be significantly higher. In particular,
the commenter noted that the additional power required for high pressure drop scrubbers alone
would be approximately $1.30 per ton of produced lime. In addition, EPA's estimated
equipment costs appear to be low.
Response
The commenters believe either that implementation costs (IV-D-06) or power costs (IV-
D-02) are not properly represented. As discussed in the response to comments regarding a
separate subcategory for scrubbers, estimated implementation costs used for the EPA model
plants include costs for demolition of existing equipment and credits for salvage value. Because
plants have a three-year period in which to comply with the proposed rule, it is expected that
scheduled downtime will be used for disconnecting an existing scrubber and connecting a new
scrubber. As a general practice, building a new scrubber while the existing scrubber remains in
operation is preferable to taking the associated kiln out of service for an extended period of time
and losing production from the kiln. The plant is expected to use its labor force in the manner
normally found for planned downtime. Such labor costs (or savings) would not be attributable to
compliance with the proposed rule.
Power costs for new scrubbers are costed incrementally, i.e., costs are estimated for the
difference between 35 in. w.g. (new scrubbers) and 14 in. w.g. (existing scrubbers) as used in
Docket Item No. II-B-122. For individual model kilns, summing the power costs and dividing
by the model's production rate gives estimated incremental power costs ranging from $0.82 to
$1.47/ton of lime. On a nationwide basis, aggregating the model kiln costs apportioned among
the affected kiln population provides average costs as estimated by EPA.
Comment
Commenter IV-D-04 stated that the EIA is seriously flawed because it assumes lime
producers can pass control costs through to consumers. The commenter maintains that lime
producers cannot raise prices. The reasons include a highly competitive market due to
overcapacity, competition from unregulated sources, competitive substitutes exist for most key
markets, and significant market resistance. The commenter also claimed that recent history
proves that prices cannot be increased. Finally the commenter stated that because the price
increase assumed by EPA is erroneous, EPA's prediction that only two lime plants will close
seriously understates the impact. One other commenter (IV-D-12), also stated that they could
not increase prices.
3-38
-------
Response
The fact that many lime plants are currently operating at less than full capacity implies
that their supply curves should be relatively elastic (flat) at current production levels because
lime producers can fairly easily change output without running into capacity constraints. Based
on standard economic theory, under perfect competition the supply curve for each firm is equal
to their marginal cost curve and each producer chooses to produce the quantity where their
marginal costs are equal to the market price in order to maximize their profits. The market
supply curve is the sum of the supply curves of the individual firms and will reflect the
responsiveness of the individual firms to changes in price. Assuming that the lime industry is
very competitive (as argued by IV-D-04) and has substantial overcapacity implies that the
industry marginal cost curve (and the market supply curve) should be relatively flat at current
production levels. To the extent that the costs of the lime manufacturing MACT standard
increase the marginal costs of lime production3, having a very elastic (flat) supply curve is a
textbook case where the majority of the costs are passed on to consumers. A highly competitive
market implies, by definition, that individual producers cannot unilaterally increase their prices
without losing most, if not all, of their customers. It does not imply that the market price will not
increase in response to a general increase in the cost of lime production due to environmental
regulations.
It is certainly true that foreign lime suppliers (including suppliers located in Mexico) gain
because this regulation applies only to domestic lime producers. The EIA assumed that import
supply was quite elastic (elasticity of 7.0) and foreign suppliers are shown to increase their
output and their profits in the EIA. However, imports of lime account for an extremely tiny
share of the lime market prior to the regulation (about 1 percent nationally) and even a fairly
large percentage increase in imports shows up as a very small change in absolute terms. The fact
that imports account for such a small share of the U.S. lime market implies that transportation
costs are too high for imported lime to be competitive in the majority of the U.S. The barrier to
imports caused by these high transportation costs is expected to prevent significant replacement
of domestic lime with imported lime.
Commenter IV-D-04 states that commercial lime producers would be unable to raise
prices for lime they sell to paper plants during times when the paper plants' captive lime
facilities are off-line because the paper plants would simply manufacture and store more lime at
their own facilities at much lower cost. It is possible that this would happen to some extent,
although paper plants already have the option to produce more lime captively. The fact that they
purchase commercial lime currently suggests that they value the option of buying on the spot
market and may not substantially change their inventory practices unless there is a sufficient
increase in price. In addition, it is true that captive lime kilns that are exempt from this MACT
standard (e.g., kilns that are covered by other regulations, such as those in the beet sugar
3A substantial portion of the cost associated with this rule is for kiln retrofits with capital
equipment such that compliance capital expenditures may not vary greatly with output.
However, maintenance costs and compliance capital depreciation are expected to vary directly
with output such that the marginal costs of production are increased.
3-39
-------
industry) could possibly begin to operate during their current down times and sell lime into the
commercial market. However, the fact that they do not currently find it worthwhile to do so
again implies that the price increase would have to be large enough for them to significantly
change their current operating procedures.
To examine the historical supply responsiveness in the lime market, we estimated the
supply elasticity for lime using data from 1983-2001. These estimates capture the overall
change in the quantity of lime supplied in response to a change in the real (inflation-adjusted)
price of lime, including any entry or exit of captive suppliers from the market. Based on
estimates obtained from the econometric model, the domestic lime supply elasticity was 1.24 at
the average price and quantity for the period and 0.98 using the lime price and quantity for 1997,
the baseline year for the EIA. The value for the baseline year implies that a 1 percent increase in
price would lead lime producers to increase their lime production by 0.98 percent, other things
being equal.
For the lime price to remain constant due to entry into the commercial market by captive
suppliers, that entry would need to be sufficient that it led to the market supply curve being
perfectly elastic. There is no evidence for a perfectly elastic market supply curve due to large-
scale entry based on historical estimates of the responsiveness of lime supply to changes in real
price.
There are substitutes for lime in many of the markets in which it competes, such as
crushed limestone, caustic soda, soda ash, and other products. However, unless the alternatives
are perfect substitutes, this does not imply that the price of lime will not increase in response to
an increase in production costs. The products have different attributes and different prices and
are clearly not perfect substitutes. In addition to estimating the overall responsiveness of supply,
we also estimated the responsiveness of the quantity of lime demanded using an econometric
model and data from 1983-2001. In that model, an inflation-adjusted price for alkalies and
chlorine was used to represent the effect of the price of lime substitutes on the quantity of lime
demanded.4 The results show that the cross-price elasticity of lime with respect to the alkalies
and chlorine price is 0.30 for the baseline year. This suggests that a 1 percent increase in the
alkalies and chlorine price would result in a 0.30 percent increase in the quantity of lime
demanded, implying that these products are substitutes. However, they are not perfect
substitutes. While it is expected that there would be some substitution of alternative products for
lime in response to a higher lime price, it is highly unlikely that other products would be treated
as perfect substitutes for lime such that the market lime price literally could not increase without
all buyers substituting other products.
4Many alternative substitutes were used in alternative model specifications, but the model with
the alkalies and chlorine price was the preferred model based on model fit and consistency with
economic theory.
3-40
-------
Assuming that lime is a highly competitive market, economic theory tells us that
individual lime producers should not be able to unilaterally increase the price of their
output without losing most, if not all, their customers. If a firm raises their selling price
above the market price, their customers are expected to shift all of their business to firms
selling at the market price. However, this does not imply that the market price for lime
cannot increase in response to changes in production costs that shift the market supply
curve.
To test the responsiveness of lime demand to price, we estimated the demand elasticity
for lime using an econometric model of the industry and data from 1983-2001. The results of the
model suggest that the demand for lime is elastic, with an elasticity of -1.43 at the average price
and quantity over the 1983-2001 period. Based on the price and quantity in the baseline year,
1997, the elasticity for that year is calculated to be -1.14. The value for the baseline year means
that the quantity of lime demanded would fall by 1.14 percent in response to a 1 percent increase
in price based on historical data. A claim that the lime price literally cannot increase due to high
demand responsiveness implies that the market demand curve is perfectly flat5 (elasticity is equal
to negative infinity), which is clearly not supported by empirical evidence. The fact that some
firms are tied into long term contracts at a fixed price may delay increases in price in some
markets until the contracts expire, but prices are still expected to increase following expiration of
existing contracts.
The fact that lime prices have not increased in recent years despite plant closures and
increases in real prices in no way implies that those events do not exert upward pressure on
prices. The relevant comparison is the price with and without those events, not before and after
they occur. It is expected that prices would have been even lower if there had not been closures
and increases in input prices.
As outlined in the responses to these comments, there is no evidence to support the claim
that the assumption that lime price will increase is erroneous, and that the estimated economic
impact of the rule is understated.
Comment
Two commenters (IV-D-04, and IV-D-20) claimed that the EPA's sensitivity analysis
does not adequately respond to the small business advocacy review (SBAR) Panel's
recommendations.
Response
5Unlike the market demand curve, the demand curve facing an individual firm in a perfectly
competitive market is expected to be perfectly flat (demand elasticity of negative infinity).
3-41
-------
In response to industry comments and to verify the reasonableness of the EIA results,
market supply and demand elasticities were estimated for lime. The values from the preferred
model for 1997 are very close to the primary elasticities used in the main text of the EIA and are
well within the range of elasticities used in the sensitivity analysis in Appendix B. In addition to
the preferred model, numerous alternative models were estimated. As with any modeling
exercise, there were some differences in results across different model specifications. However,
the results were generally similar across specifications and there were no cases in which the
estimated supply or demand elasticity fell outside the ranges currently used in the Appendix B
sensitivity analysis included in the EIA. Thus, the current analysis adequately responds to
SBAR panel recommendations that a reasonable sensitivity analysis be employed and the
empirical evidence is supportive of the current scenario presented in the Section 4 of the EIA.
Comment
One commenter (IV-D-04) stated that the EPA economic model for the lime market
assumes a nationally perfectly competitive market, but lime prices are primarily dictated by large
producers who sell capacity regardless of price.
Response
This comment suggests that large lime producers have market power and therefore face a
downward sloping demand curve and have some ability to set prices. If large lime producers do
possess market power, then profit-maximizing behavior would imply that they would restrict
output below the levels expected under perfect competition in order to increase market price to
the point that their marginal revenue is equal to their marginal cost. The large producers may
have lower marginal costs such that the resulting price makes it difficult for the small producers
that take the market price as given to remain in business. However, the presence of market
power in the lime industry would tend to increase prices relative to the perfectly competitive
case, not decrease them.
Comment
One commenter ( IV-D-06) stated they sell 40 to 50 percent of their lime to the soil
stabilization market, an area that has encountered competition from liquid petroleum byproducts.
These low-cost, marginally effective products have made substantial progress entering this
market with an inexpensive, unproven product and any increase in lime costs will reduce the
competitiveness of lime in this market.
Response
3-42
-------
An increase in the price of lime relative to substitutes will tend to reduce the quantity of
lime purchased. However, to the extent that liquid petroleum byproducts are marginally
effective and unproven, they are presumably not close substitutes for lime.
Comment
One commenter (IV-D-06) claims that although EPA has indicated its rule will have
larger impacts on small businesses than large ones, the disparity is even greater than EPA
estimates. The reductions in pre-tax earnings presented in the EIA understate losses for small
firms because the costs of implementation will be higher than EPA estimates and the price of
lime will not increase. They also state that even if only 2 to 3 of the 14 small lime firms close,
that would still be closure of 14% to 21% of the small lime firms in the domestic industry. This
seems to be such a significant economic impact that it should encourage the EPA to seriously
consider additional ways to minimize the impact on small businesses.
Response
It is unclear what the basis for the first part of this comment is (it seems the same claims
they are making for small firms would also apply to large firms). As far as the second part, to
the extent that actual costs differ from EPA estimates, it is possible that the actual losses
experienced by firms would be higher or lower than presented in the EIA. However, the costs of
implementation currently used for analysis reflect EPA's best estimate of actual costs. The
assertion that lime prices cannot increase in response to an increase in production costs is not
credible (see comments above).
We also disagree that the number of small firms at risk of closure, 2 to 3, can be
considered a significant number in the context of SBAR. In any case, we have seriously
considered ways to minimize the impact on small businesses based on comments from industry
and has substantially reduced the costs of this rule relative to pervious versions. As previously
discussed, EPA, along with the Small Business Administration and the Office of Management
and Budget, convened a panel under the authority of SBAR to talk with small business
representatives on how to mitigate potential impacts to small businesses associated with the lime
manufacturing NESHAP. This panel yielded a report that included many recommendations on
how potential impacts to small businesses from the proposed rulel could be mitigated. These
recommendations are also reflected in the final rule.
Comment
One commenter (IV-D-06) claims that the EIA assumption that production costs will go
down with decreases in production is wrong. The commenter claimed that they maintains a
close watch on cost of production stated in cost per ton of lime produced. Simple economics
3-43
-------
dictates that an increase in dollar cost associated with a decrease in production results in a higher
cost of production, which is a higher cost per ton of lime.
Response
Presumably, the commenter's statement that there is an increase in dollar cost refers to
the compliance costs, but those costs are shown separately from production costs in the EIA.
The EIA does not say that the total cost per ton of lime including compliance costs will decrease,
it indicates that total production costs (not including compliance costs) will decrease in response
to a reduction in use of variable inputs as production falls.
Comment
One commenter (IV-D-06) stated they are already trying to control costs while over the
past several years our workers' compensation insurance, group health insurance, natural gas,
maintenance parts, and liability insurance costs have continued to increase by double digit
percentages, while sales prices have not increased to this extreme. Trade and economic
predictions indicate more similar increases can be expected. These factors should be important
considerations of the economic analysis.
Response
EPA gathered baseline data for the lime industry to inform the economic analysis. It is
possible that market conditions for the lime industry have worsened over time. However, the
focus of the economic analysis is the incremental impacts of the lime manufacturing MACT
standard, which may be relatively small compared with other changes that are occurring within
the lime industry.
3.12 Risk Analysis
Comment
One commenter (IV-D-01) was concerned over EPA's use of the Acute Exposure
Guideline Level (AEGL) in assessing the health risk associated with HC1. While not directly
objecting to the conclusions reached by EPA, the commenter noted that the intended use of the
AEGL, according to the National Research Council, is in conjunction with "once in a lifetime"
exposures for emergency exposures ranging from 10 minutes to 8 hours. Because the AEGL
values are intended to be used in conjunction with a single lifetime exposure, they can be higher
than short term limits recommended for populations with repeated exposures. It is not clear in
the description of the NLA analysis, if NLA in their use of AEGL was contemplating a once in a
3-44
-------
lifetime exposure or whether exposures would be occurring repeatedly. The commenter stated
that EPA should explicitly state how they believe AEGL values should be used in their risk
assessment process and what are the possible exposure levels to the public. The commenter was
also troubled by the use in the rationale of both the RfC (estimated daily exposure that over a
lifetime is not likely to result in significant non cancer effect in humans) and the AEGL (once in
a lifetime exposure).
The commenter asked that EPA clarify their position on the use of AEGL values for
environmental risk assessments, and whether its use represents a "reasonable methodology" and
"consistent with EPA methodology" as claimed in the preamble.
Response
This analysis evaluated only risks based on offsite exposure to the general population.
EPA does not assess risks associated with occupational exposures, nor do we have the authority
to address occupational issues.
In our evaluation of the NLA risk analysis, EPA reviewed the available acute dose-
response values for this compound. Among these, the CalEPA reference exposure level (REL)
and AEGL-1 values (2.1 and 2.7 mg/m3, respectively) were found to be the most health-
protective. Since these benchmarks were effectively the same, and AEGL values are products of
a Federal effort in which EPA participates, we gave priority to the AEGL. Therefore, the
AEGL-1 selected for analysis represented the most appropriate value.
Although there are relatively protective acute dose-response assessment values for
hydrochloric acid, protective assessments for other hazardous air pollutants may not exist. In
such cases we may have to rely on less-conservative analyses, perhaps based on endpoints at
mild effect levels. Nevertheless, EPA believes that the use of less-protective values (e.g.,
AEGL-2) is better than omitting acute exposure from the risk assessment.
3.13 Applicability
Comment
One commenter (IV-D-03) requested that the rule be revised to exempt lime kilns
operated in conjunction with public water treatment facilities. The commenter explained that the
kilns are used to recover the calcium carbonate sludge from their water softening process. The
recovered lime is used on-site in the water treatment process thereby reducing demand for these
resources, avoiding the air emissions that would otherwise be associated with the transport of
this material to a disposal site, and reducing the amount of solid waste generated. Under the
existing language of the rule, the lime recovery kilns located at the water treatment plant would
3-45
-------
be subject to the rule because the treatment facility operates air stripping towers which are
permitted to emit over 10 tons of a single HAP and over 25 tons per year of multiple HAP.
The commenter states that unlike lime kilns that process raw limestone materials, which
contain metal HAP, the sludge from the water softening process, the only calcareous material fed
to the kilns, does not contain metal HAP in any appreciable quantity. The elements and
compounds found in the calcareous sludge from the water softening process are representative of
those found in the treatment plants raw water source and the chemicals that are added as part of
the water softening process. As the public's water supply, neither the source water, sodium
silicate, nor the lime added prior to the kiln should have significant levels of metal HAP.
The commenter also stated that EPA did not review lime kilns at water treatment
facilities in developing the proposed rule. Nor did the NLA include these lime kilns in their
modeling of HC1 emissions. The inclusion of lime kilns operated as part of water treatment
facilities would be arbitrary and capricious and not based on any research or analysis.
The commenter suggests EPA revise 63.7081(a) to include public water treatment plants
in the list of lime producing facilities exempt from the rule. The commenter also suggests
revising 63.7143 by modifying the definition of "Stone feed" by adding the following: "Stone
feed does not include calcareous solids that are generated by and recovered in the water
softening process at water treatment plants."
Response
The lime recovery operation described by the commenter was not the intended target of
the rule. The EPA has revised the rule to exclude lime kilns that are operated as part of a
publicly owned water treatment facility to recover lime from water softening processes.
3.14 General Comments
Comment
One commenter (IV-D-16) notes that in the proposal preamble, EPA asked for comments
on a two-tiered approach for compliance monitoring. The commenter agrees that the rule should
contain a threshold limit to initiate corrective action (which if corrected within a given time
period would not constitute a violation). In many cases, entering a baghouse to perform
maintenance on a broken bag poses a significant hazard to the maintenance personnel due to
temperature, lack of oxygen, and high levels of sulfur oxides and nitrogen oxides. Thus the time
period for corrective action must consider the critical time necessary to cool down and purge a
compartment, so that the danger to maintenance personnel can be minimized.
3-46
-------
Response
For baghouses monitored using BLD systems and for ESP monitored using PM CEM,
alarm systems may be used to alert operators when preset levels have been have been exceeded
indicating that corrective actions may be necessary. Only after the alarm time has exceeded 5
percent of the total operating time in a 6-month period has a violation of an operating limit
occurred. Alarm time is counted by how long it takes it takes from the onset of a BLD or PM
CEM alarm to the time that corrective action is initiated. By not counting the time required to
correct the problem, the plant owner or operator can take the time necessary to address the cause
of the alarm in a safe and efficient manner without incurring additional alarm time. Procedures
to be followed in inspecting control devices for the cause of alarms or malfunctions should be
incorporated into a facility's OM&M plan and SSM plan.
For other parametric compliance monitoring options, the rule requires that the
owner/operator define the compliance parameters to be monitored in their OM&M plan. During
the initial performance tests, they are required to monitor and establish the value or range of the
parameters. EPA believes that the initial performance tests will be designed to give the owner or
operator flexibility in adhering to the established range for monitored parameters. These values
must be reported in the results of the test and notification of compliance status to the permitting
authority and must be approved by the permitting authority. During subsequent operations, if the
monitored parameters exceed the values or fall outside the range determined during the initial
performance test, it is a violation of the operating requirements of the standard (unless it is the
result of a malfunction to which the facility responds to in accordance with the SSM plan). The
EPA has no basis for allowing deviations before considering the facility to be in violation. The
owner/operator has ample opportunity to establish a range for the operating parameters and must
thereafter operate within that range.
Comment
Several commenters (IV-D-04, IV-D-05, and IV-D-08) believe the rule should not
require HC1 testing of all kilns. Commenter IV-D-04 believes the rule should provide state
agencies with the discretion to not require HC1 testing of all kilns. Commenters IV-D-05
Western and IV-D-08 Engineering believe the rule itself should allow flexibility in HC1 testing.
The commenters note that in recent years, many lime plants have been forced to idle or
infrequently operate kilns at operating plants due to increased fuel cost, reduced customer
demand, etc., and start up of every kiln for the purpose of conducting HC1 testing would require
significant expenditures. Commenter IV-D-04 notes that this would also result in PM and other
emissions that otherwise would not be generated. Commenter IV-D-04 uses the example of an
idled kiln of the same type as an active kiln at the plant, fired with the same fuel and stone; it
may be reasonable for a state agency to conclude that HC1 emissions testing from the active kiln
can be used to estimate emissions from the idled kiln. Furthermore, the commenters note that
testing may not be warranted on kilns that represent small fractions of the plant's total capacity.
3-47
-------
As a result, Commenter IV-D-04 requests that the rule be revised to provide state agencies with
the discretion to determine whether testing of all kilns at a lime plant is necessary in order to
demonstrate that a plant is an area source. Commenter IV-D-05 requests that the rule state that
plants which desire to claim area source status be required to test a sufficient number of kilns so
that 70 percent or more of the potential production capacity of the plant is tested for HC1. Or, as
an alternative, EPA could allow HC1 test results from one kiln to be extrapolated to estimate
emissions from comparable kilns at the same plant.
Response
A lime manufacturing plant's status as an area source or a major source is not addressed
by this rule. If there are questions regarding whether a source is an area source or a major
source, the owner or operator should refer to the General Provisions, in particular, section 63.2,
and contact the appropriate permitting authority if additional clarification is needed. We believe
this is the best approach for major source determinations because every site may have different
issues, and attempting to address these site specific issues in this rule would be inappropriate.
We believe allowing the permitting authority to evaluate these issues on a case-by-case basis is
the best approach.
Comment
Several commenters (IV-D-04, IV-D-17, IV-D-11, and IV-D-13) state that the rule
should allow lime plants to use the following emission factors to estimate HAP metal emissions:
0.15 percent of PM emissions from kiln stacks and 0.07 percent from stone handling operations.
Commenters IV-D-04 and IV-D-17 believe that the quantity of metal HAP emitted from lime
plants is sufficiently low that it is highly unlikely that such metals will be present in quantities
that would have any effect on whether a commercial lime plant exceeds the 25 ton per year
major source threshold for all HAP. According to commenter IV-D-04, Docket Item No. II-B-43
establishes the HAP metal content of lime kiln emissions to be 0.15 percent, and Docket Item
No. II-B-77 establishes the HAP metal content of fugitive stone emissions to be 0.07 percent.
Commenter IV-D-16 Refractories also commented that the rule should allow for the use of
emissions factors for HAP metals.
Commenter IV-D-04 notes that the cement MACT allows cement plants to assume that
HAP metals emissions constitute 1 percent of PM emissions for the purpose of area source
determinations. This commenter believes the quantities of metal HAP from lime plants are so
small they will not affect the area/major source determination, and thus measurement of the
metal emissions is unnecessary.
Response
3-48
-------
The rule, as proposed, contained no requirement to measure HAP metals in the emissions
from kilns or stone handling operations and we see no reason to add such provisions. The final
rule does not require that lime manufacturing plants perform emissions testing for HAP metals.
Therefore, we see no reason to establish a specific ratio in the rule. However, nothing in the rule
would preclude these ratios from being used.
Comment
Several commenters (IV-D-04, IV-D-16, IV-D-11, and IV-D-13) comment that testing
for HAP organics should not be required because it is highly unlikely that such compounds will
be present in quantities that would have any effect on whether a commercial lime plant exceeds
the 25 ton per year major source threshold for all HAP. Commenter IV-D-04 notes that lime
kilns are not expected to emit significant quantities of organic HAP because inorganic limestone
is the only feed material processed, and EPA's testing of 10 commercial lime kilns detected
virtually no organic HAP (Docket Item No. II-B-121). This commenter also notes that lime kilns
at sugar beet facilities have been exempted from the lime MACT, and no analysis of organic
emissions from lime kilns at sugar beet facilities has been conducted (Docket Item Nos. II-B-44
and 81).
Response
EPA agrees with the commenters, the rule does not require that emission testing for HAP
organics be conducted.
Comment
One commenter (IV-D-04) states the requirement in § 63.7142(b) to use FTIR, Method
18 or ASTM D6420-99 if organics testing is conducted for area source determinations should be
deleted. The commenter notes that the proposed requirement creates the inference that organics
testing should be conducted, and that testing beyond the 36 volatile organics specified in the
ASTM method is necessary, when the chance that organics will cause lime plants to be a major
source is highly remote.
Response
As stated in the previous response, the final rule does not require emission testing for
HAP organics. However, we do not believe it is appropriate to specifically state that organics
testing is not required. It is possible that a permitting authority may wish to have a facility
perform organic HAP testing due to site specific considerations. If this occurs, we believe that
EPA-approved test methods should be used.
3-49
-------
Comment
One commenter (IV-D-02) believes the final rule should include an alternative limit
based on the quantity of HAP metals released. For example, " a lime manufacturing plant with a
dry collector and a particulate release greater than 0.12 lb/tsf should not be required to meet a
MACT surrogate particulate value when its release of HAP metals is equal to or less than that of
EPA's basis for establishing the 0.12 lb PM/tsf limit...a lime plant that has a stone feed stock low
in heavy metals and/or fires with natural gas can have a lower HAP metals release than a plant
meeting the proposed MACT particulate limit. Since the intent of the proposed regulation is to
control the release of HAP metals, whether a company complies via operating practices or
particulate control equipment should be irrelevant." In the case of wet scrubbers, the commenter
recommends an alternative HAP limit of 1.5 lbs/1000 tons of stone feed (based on a particulate
release of 1.0 lb/tsf and a HAP metals concentration of 1,500 ppm in the particulate).
Response
We do not have, and the commenter did not provide, sufficient HAP metal emission
information on which to establish a metal HAP emission limit as an alternative to the PM
emission limit. We commonly use PM emissions as a surrogate for metal HAP emissions,
finding that controlling PM emissions will control the non-volatile and semi-volatile metal HAP.
At-the-stack controls used at lime manufacturing plants, i.e., baghouses, ESPs, and wet
scrubbers, capture non-volatile and semi-volatile metal HAP non-preferentially along with other
PM, making PM a permissible indicator for HAP metals. We believe using PM as a surrogate is
preferred by most sources as the cost of emissions testing and monitoring that would be required
to demonstrate compliance with the otherwise numerous standards that would apply to individual
HAP metals would be much higher than testing for and monitoring for PM emissions.
Comment
Several commenters (IV-D-04, IV-D-16, IV-D-11, and IV-D-13) state that performance
testing should be conducted under "representative" conditions rather than under the "highest
production level reasonably expected to occur." Commenter IV-D-11 notes that there are
inconsistencies between what is proposed in Table 4 of the rule and what is required under the
General Provisions at 40CFR 63.7(e)(1). EPA has recently amended the Cement MACT to fix
similar inconsistencies, and the commenter suggests the Lime MACT be similarly revised.
Response
We have changed the requirement in the final rule to require testing under representative
conditions, which is in agreement with the language in the General Provisions.
3-50
-------
Comment
One commenter (IV-D-04) notes that reports should be required only for malfunctions,
not startups and shutdowns. The commenter states that in recent proposed amendments to the
General Provisions in Subpart A of Part 63, EPA has determined that requiring reporting of
startups and shutdowns is "unnecessary and burdensome" (67 Fed. Reg. 72,881).
Response
Section 63.7121(b) requires that deviations from operating, VE, and opacity limits,
including periods of startup, shutdown, or malfunction be reported. EPA intends for this
requirement to be consistent with the General Provisions and has revised this paragraph to clarify
that reports of startups, shutdowns, and malfunctions are to be made in accordance with section
63.10(d) of the General Provisions.
Comment
Two commenters (IV-D-04 and IV-D-16) believe the final rule should provide a risk-
based exemption from the entire rule (not just from HC1 standards) for plants at which modeled
risks are below health based thresholds. Commenter IV-D-04 notes that EPA recently solicited
comment on providing risk-based exemptions in proposed MACT standards for several sources
categories. This commenter strongly supports the view that such exemptions should be provided
in MACT standards that impose substantial costs while achieving negligible reductions in risks
to public health, and believes the lime MACT fits this description.
Response
Other than the decision to not regulate emissions of HC1 from lime manufacturing, EPA
did not consider and did not request comments on providing risk-based exemptions for lime
manufacturing facilities. Although EPA is aware that risk-based exemptions were being
discussed in other proposed rules, no decisions have been made by the Agency regarding risk-
based exemptions and application to industry groups or individual plants. Due to the uncertainty
of how these exemptions would be structured, we do not believe that it would be appropriate to
include these site specific risk based exemptions in the final rule. We also believe that including
such a substantive change in the final rule without allowing the general public an opportunity to
comment would be a violation of the Administrative Procedures Act, especially in light of the
fact that their inclusion in other proposed rules have generated significant negative comment
from the public.
3-51
-------
Comment
One commenter (IV-D-04) states the Benefits Analysis is based on inaccurate
assumptions, and presents conclusions regarding reductions in metal HAP that are greatly
overstated.
The commenter submitted comments to the Small Business Advocacy Review (SBAR)
Panel explaining that the emission factor for uncontrolled existing kilns is 27 times smaller than
the emission factor used by EPA to calculate baseline emissions. The commenter notes that its
emission factor is based on actual test data from these kilns.
The commenter also notes that the emission factor for existing uncontrolled stone
handling operations is also overstated; it was derived using AP-42 emission factors with "E"
ratings. The commenter states that it also presented to the SBAR Panel a more reliable emission
factor for these units that is rated "C" and was revised in 1995.
In addition, the commenter claims that EPA overstated the amount of new capacity and
the emissions from new rotary kilns. The commenter states, "EPA should either reflect (our)
estimates in the preamble to the final rule, or provide a reasoned response to our comments that
EPA's estimates are overstated... we believe the best estimate of metal HAP reductions is 3.5 tons
(7,000 pounds) per year. Based on the 56 lime plants predicted to be subject to the MACT rule,
this translates into an annual reduction in metal HAP per lime plant of 124 pounds.
Response
EPA reviewed the new information on PM emissions presented by the commenter as well
as their calculations of baseline emissions and emission reductions resulting from the rule and
finds them to be reasonable. The preamble to the final rule reflects these revised estimates.
However, even with these revised estimates, we still note that the estimated benefits are still
significantly higher than the costs for the final rule.
Comment
One commenter (IV-D-22) states that EPA should clarify that PM from nuisance dust
collectors that service coolers are not included in the collective PM limit for kilns/coolers. The
commenter notes that nuisance dust collectors, commonly associated with coolers, (called the
cooler duct collector in some plants) collect dust generated in the transfer from the cooler to the
material handling system, and PM from the nuisance dust collector should not be included in the
kiln PM limit, but instead should be exempt as part of the lime handling system.
3-52
-------
Response
The standard only applies to air used for cooling that is exhausted from the cooler to the
kiln or that is exhausted separately to the atmosphere. Operations that process only lime
product, lime kiln dust, or fuel are excluded from coverage. The rule was revised to clarify that
dust emitted in the transfer from the cooler to the material handling system is not included in the
calculation of the PM emission rate.
Comment
One commenter (Docket No. OAR-2002-0052, item 530) states that clean air is
essential, and requests that the companies that pollute the air be made to clean it up.
Response
EPA acknowledges the comment and believes this and other standards promulgated
under the CAA are effective at reducing HAP emissions.
Comment
One commenter (IV-D-15) requests that EPA withdraw the proposed Lime
manufacturing NESHAP and publish a new draft that is substantially more environmentally
protective of PM10.
Response
The degree to which emissions of any substance can be reduced by regulation are a
reflection of the statutory basis of those regulations. Under section 112 of the CAA, EPA is
directed to establish MACT standards which are based on the emission reduction being achieved
by the top performing, currently operating facilities. The standard for lime manufacturing plants
used this approach and we believe that the result of the standard will be a substantial reduction in
total PM emissions (over 3,880 tons per year), including PM10 emissions.
3-53
------- |